{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 23.88344877000239,
  "global_step": 100000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.0,
      "learning_rate": 1e-05,
      "loss": 5.8028,
      "step": 10
    },
    {
      "epoch": 0.0,
      "learning_rate": 1e-05,
      "loss": 5.7562,
      "step": 20
    },
    {
      "epoch": 0.01,
      "learning_rate": 1e-05,
      "loss": 4.8255,
      "step": 30
    },
    {
      "epoch": 0.01,
      "learning_rate": 1e-05,
      "loss": 4.4343,
      "step": 40
    },
    {
      "epoch": 0.01,
      "learning_rate": 1e-05,
      "loss": 4.2783,
      "step": 50
    },
    {
      "epoch": 0.01,
      "learning_rate": 1e-05,
      "loss": 4.0117,
      "step": 60
    },
    {
      "epoch": 0.02,
      "learning_rate": 1e-05,
      "loss": 3.8073,
      "step": 70
    },
    {
      "epoch": 0.02,
      "learning_rate": 1e-05,
      "loss": 3.6936,
      "step": 80
    },
    {
      "epoch": 0.02,
      "learning_rate": 1e-05,
      "loss": 3.5248,
      "step": 90
    },
    {
      "epoch": 0.02,
      "learning_rate": 1e-05,
      "loss": 3.4096,
      "step": 100
    },
    {
      "epoch": 0.02,
      "eval_accuracy": 0.0,
      "eval_loss": 3.1817097663879395,
      "eval_runtime": 196.4966,
      "eval_samples_per_second": 3.481,
      "step": 100
    },
    {
      "epoch": 0.03,
      "learning_rate": 1e-05,
      "loss": 3.3652,
      "step": 110
    },
    {
      "epoch": 0.03,
      "learning_rate": 1e-05,
      "loss": 3.3114,
      "step": 120
    },
    {
      "epoch": 0.03,
      "learning_rate": 1e-05,
      "loss": 3.2772,
      "step": 130
    },
    {
      "epoch": 0.03,
      "learning_rate": 1e-05,
      "loss": 3.1814,
      "step": 140
    },
    {
      "epoch": 0.04,
      "learning_rate": 1e-05,
      "loss": 3.2805,
      "step": 150
    },
    {
      "epoch": 0.04,
      "learning_rate": 1e-05,
      "loss": 3.1881,
      "step": 160
    },
    {
      "epoch": 0.04,
      "learning_rate": 1e-05,
      "loss": 3.008,
      "step": 170
    },
    {
      "epoch": 0.04,
      "learning_rate": 1e-05,
      "loss": 3.138,
      "step": 180
    },
    {
      "epoch": 0.05,
      "learning_rate": 1e-05,
      "loss": 2.894,
      "step": 190
    },
    {
      "epoch": 0.05,
      "learning_rate": 1e-05,
      "loss": 2.924,
      "step": 200
    },
    {
      "epoch": 0.05,
      "eval_accuracy": 0.0043859649122807015,
      "eval_loss": 2.702718734741211,
      "eval_runtime": 172.3465,
      "eval_samples_per_second": 3.969,
      "step": 200
    },
    {
      "epoch": 0.05,
      "learning_rate": 1e-05,
      "loss": 2.8152,
      "step": 210
    },
    {
      "epoch": 0.05,
      "learning_rate": 1e-05,
      "loss": 2.8987,
      "step": 220
    },
    {
      "epoch": 0.05,
      "learning_rate": 1e-05,
      "loss": 2.734,
      "step": 230
    },
    {
      "epoch": 0.06,
      "learning_rate": 1e-05,
      "loss": 2.9393,
      "step": 240
    },
    {
      "epoch": 0.06,
      "learning_rate": 1e-05,
      "loss": 2.8134,
      "step": 250
    },
    {
      "epoch": 0.06,
      "learning_rate": 1e-05,
      "loss": 2.6527,
      "step": 260
    },
    {
      "epoch": 0.06,
      "learning_rate": 1e-05,
      "loss": 2.747,
      "step": 270
    },
    {
      "epoch": 0.07,
      "learning_rate": 1e-05,
      "loss": 3.1505,
      "step": 280
    },
    {
      "epoch": 0.07,
      "learning_rate": 1e-05,
      "loss": 2.6755,
      "step": 290
    },
    {
      "epoch": 0.07,
      "learning_rate": 1e-05,
      "loss": 2.7765,
      "step": 300
    },
    {
      "epoch": 0.07,
      "eval_accuracy": 0.02631578947368421,
      "eval_loss": 2.4588308334350586,
      "eval_runtime": 102.5903,
      "eval_samples_per_second": 6.667,
      "step": 300
    },
    {
      "epoch": 0.07,
      "learning_rate": 1e-05,
      "loss": 2.6404,
      "step": 310
    },
    {
      "epoch": 0.08,
      "learning_rate": 1e-05,
      "loss": 2.5899,
      "step": 320
    },
    {
      "epoch": 0.08,
      "learning_rate": 1e-05,
      "loss": 2.6505,
      "step": 330
    },
    {
      "epoch": 0.08,
      "learning_rate": 1e-05,
      "loss": 2.7317,
      "step": 340
    },
    {
      "epoch": 0.08,
      "learning_rate": 1e-05,
      "loss": 2.6394,
      "step": 350
    },
    {
      "epoch": 0.09,
      "learning_rate": 1e-05,
      "loss": 2.6274,
      "step": 360
    },
    {
      "epoch": 0.09,
      "learning_rate": 1e-05,
      "loss": 2.5362,
      "step": 370
    },
    {
      "epoch": 0.09,
      "learning_rate": 1e-05,
      "loss": 2.6822,
      "step": 380
    },
    {
      "epoch": 0.09,
      "learning_rate": 1e-05,
      "loss": 2.6067,
      "step": 390
    },
    {
      "epoch": 0.1,
      "learning_rate": 1e-05,
      "loss": 2.4318,
      "step": 400
    },
    {
      "epoch": 0.1,
      "eval_accuracy": 0.07602339181286549,
      "eval_loss": 2.3081016540527344,
      "eval_runtime": 165.6934,
      "eval_samples_per_second": 4.128,
      "step": 400
    },
    {
      "epoch": 0.1,
      "learning_rate": 1e-05,
      "loss": 2.4746,
      "step": 410
    },
    {
      "epoch": 0.1,
      "learning_rate": 1e-05,
      "loss": 2.4729,
      "step": 420
    },
    {
      "epoch": 0.1,
      "learning_rate": 1e-05,
      "loss": 2.6352,
      "step": 430
    },
    {
      "epoch": 0.11,
      "learning_rate": 1e-05,
      "loss": 2.5389,
      "step": 440
    },
    {
      "epoch": 0.11,
      "learning_rate": 1e-05,
      "loss": 2.4172,
      "step": 450
    },
    {
      "epoch": 0.11,
      "learning_rate": 1e-05,
      "loss": 2.4387,
      "step": 460
    },
    {
      "epoch": 0.11,
      "learning_rate": 1e-05,
      "loss": 2.3282,
      "step": 470
    },
    {
      "epoch": 0.11,
      "learning_rate": 1e-05,
      "loss": 2.5716,
      "step": 480
    },
    {
      "epoch": 0.12,
      "learning_rate": 1e-05,
      "loss": 2.5506,
      "step": 490
    },
    {
      "epoch": 0.12,
      "learning_rate": 1e-05,
      "loss": 2.5316,
      "step": 500
    },
    {
      "epoch": 0.12,
      "eval_accuracy": 0.12573099415204678,
      "eval_loss": 2.199833631515503,
      "eval_runtime": 195.2506,
      "eval_samples_per_second": 3.503,
      "step": 500
    },
    {
      "epoch": 0.12,
      "learning_rate": 1e-05,
      "loss": 2.3459,
      "step": 510
    },
    {
      "epoch": 0.12,
      "learning_rate": 1e-05,
      "loss": 2.4506,
      "step": 520
    },
    {
      "epoch": 0.13,
      "learning_rate": 1e-05,
      "loss": 2.3967,
      "step": 530
    },
    {
      "epoch": 0.13,
      "learning_rate": 1e-05,
      "loss": 2.322,
      "step": 540
    },
    {
      "epoch": 0.13,
      "learning_rate": 1e-05,
      "loss": 2.215,
      "step": 550
    },
    {
      "epoch": 0.13,
      "learning_rate": 1e-05,
      "loss": 2.4587,
      "step": 560
    },
    {
      "epoch": 0.14,
      "learning_rate": 1e-05,
      "loss": 2.3163,
      "step": 570
    },
    {
      "epoch": 0.14,
      "learning_rate": 1e-05,
      "loss": 2.577,
      "step": 580
    },
    {
      "epoch": 0.14,
      "learning_rate": 1e-05,
      "loss": 2.5107,
      "step": 590
    },
    {
      "epoch": 0.14,
      "learning_rate": 1e-05,
      "loss": 2.4594,
      "step": 600
    },
    {
      "epoch": 0.14,
      "eval_accuracy": 0.16666666666666666,
      "eval_loss": 2.1160104274749756,
      "eval_runtime": 196.3352,
      "eval_samples_per_second": 3.484,
      "step": 600
    },
    {
      "epoch": 0.15,
      "learning_rate": 1e-05,
      "loss": 2.3088,
      "step": 610
    },
    {
      "epoch": 0.15,
      "learning_rate": 1e-05,
      "loss": 2.358,
      "step": 620
    },
    {
      "epoch": 0.15,
      "learning_rate": 1e-05,
      "loss": 2.251,
      "step": 630
    },
    {
      "epoch": 0.15,
      "learning_rate": 1e-05,
      "loss": 2.2678,
      "step": 640
    },
    {
      "epoch": 0.16,
      "learning_rate": 1e-05,
      "loss": 2.3647,
      "step": 650
    },
    {
      "epoch": 0.16,
      "learning_rate": 1e-05,
      "loss": 2.268,
      "step": 660
    },
    {
      "epoch": 0.16,
      "learning_rate": 1e-05,
      "loss": 2.275,
      "step": 670
    },
    {
      "epoch": 0.16,
      "learning_rate": 1e-05,
      "loss": 2.2154,
      "step": 680
    },
    {
      "epoch": 0.16,
      "learning_rate": 1e-05,
      "loss": 2.3606,
      "step": 690
    },
    {
      "epoch": 0.17,
      "learning_rate": 1e-05,
      "loss": 2.2819,
      "step": 700
    },
    {
      "epoch": 0.17,
      "eval_accuracy": 0.195906432748538,
      "eval_loss": 2.058662176132202,
      "eval_runtime": 194.7654,
      "eval_samples_per_second": 3.512,
      "step": 700
    },
    {
      "epoch": 0.17,
      "learning_rate": 1e-05,
      "loss": 2.2181,
      "step": 710
    },
    {
      "epoch": 0.17,
      "learning_rate": 1e-05,
      "loss": 2.2065,
      "step": 720
    },
    {
      "epoch": 0.17,
      "learning_rate": 1e-05,
      "loss": 2.1479,
      "step": 730
    },
    {
      "epoch": 0.18,
      "learning_rate": 1e-05,
      "loss": 2.2029,
      "step": 740
    },
    {
      "epoch": 0.18,
      "learning_rate": 1e-05,
      "loss": 2.1592,
      "step": 750
    },
    {
      "epoch": 0.18,
      "learning_rate": 1e-05,
      "loss": 2.1916,
      "step": 760
    },
    {
      "epoch": 0.18,
      "learning_rate": 1e-05,
      "loss": 2.3466,
      "step": 770
    },
    {
      "epoch": 0.19,
      "learning_rate": 1e-05,
      "loss": 2.2752,
      "step": 780
    },
    {
      "epoch": 0.19,
      "learning_rate": 1e-05,
      "loss": 2.286,
      "step": 790
    },
    {
      "epoch": 0.19,
      "learning_rate": 1e-05,
      "loss": 2.122,
      "step": 800
    },
    {
      "epoch": 0.19,
      "eval_accuracy": 0.2412280701754386,
      "eval_loss": 2.0073294639587402,
      "eval_runtime": 196.4527,
      "eval_samples_per_second": 3.482,
      "step": 800
    },
    {
      "epoch": 0.19,
      "learning_rate": 1e-05,
      "loss": 2.2788,
      "step": 810
    },
    {
      "epoch": 0.2,
      "learning_rate": 1e-05,
      "loss": 2.2722,
      "step": 820
    },
    {
      "epoch": 0.2,
      "learning_rate": 1e-05,
      "loss": 2.2069,
      "step": 830
    },
    {
      "epoch": 0.2,
      "learning_rate": 1e-05,
      "loss": 2.2058,
      "step": 840
    },
    {
      "epoch": 0.2,
      "learning_rate": 1e-05,
      "loss": 2.2787,
      "step": 850
    },
    {
      "epoch": 0.21,
      "learning_rate": 1e-05,
      "loss": 2.0642,
      "step": 860
    },
    {
      "epoch": 0.21,
      "learning_rate": 1e-05,
      "loss": 2.0778,
      "step": 870
    },
    {
      "epoch": 0.21,
      "learning_rate": 1e-05,
      "loss": 2.3254,
      "step": 880
    },
    {
      "epoch": 0.21,
      "learning_rate": 1e-05,
      "loss": 2.1736,
      "step": 890
    },
    {
      "epoch": 0.21,
      "learning_rate": 1e-05,
      "loss": 2.0812,
      "step": 900
    },
    {
      "epoch": 0.21,
      "eval_accuracy": 0.2646198830409357,
      "eval_loss": 1.9639685153961182,
      "eval_runtime": 195.3389,
      "eval_samples_per_second": 3.502,
      "step": 900
    },
    {
      "epoch": 0.22,
      "learning_rate": 1e-05,
      "loss": 2.2958,
      "step": 910
    },
    {
      "epoch": 0.22,
      "learning_rate": 1e-05,
      "loss": 2.1834,
      "step": 920
    },
    {
      "epoch": 0.22,
      "learning_rate": 1e-05,
      "loss": 2.1188,
      "step": 930
    },
    {
      "epoch": 0.22,
      "learning_rate": 1e-05,
      "loss": 2.1533,
      "step": 940
    },
    {
      "epoch": 0.23,
      "learning_rate": 1e-05,
      "loss": 2.1699,
      "step": 950
    },
    {
      "epoch": 0.23,
      "learning_rate": 1e-05,
      "loss": 2.2046,
      "step": 960
    },
    {
      "epoch": 0.23,
      "learning_rate": 1e-05,
      "loss": 2.0988,
      "step": 970
    },
    {
      "epoch": 0.23,
      "learning_rate": 1e-05,
      "loss": 1.998,
      "step": 980
    },
    {
      "epoch": 0.24,
      "learning_rate": 1e-05,
      "loss": 2.0502,
      "step": 990
    },
    {
      "epoch": 0.24,
      "learning_rate": 1e-05,
      "loss": 2.0505,
      "step": 1000
    },
    {
      "epoch": 0.24,
      "eval_accuracy": 0.2894736842105263,
      "eval_loss": 1.9279896020889282,
      "eval_runtime": 166.1021,
      "eval_samples_per_second": 4.118,
      "step": 1000
    },
    {
      "epoch": 0.24,
      "learning_rate": 1e-05,
      "loss": 2.1506,
      "step": 1010
    },
    {
      "epoch": 0.24,
      "learning_rate": 1e-05,
      "loss": 2.0055,
      "step": 1020
    },
    {
      "epoch": 0.25,
      "learning_rate": 1e-05,
      "loss": 1.9845,
      "step": 1030
    },
    {
      "epoch": 0.25,
      "learning_rate": 1e-05,
      "loss": 1.9575,
      "step": 1040
    },
    {
      "epoch": 0.25,
      "learning_rate": 1e-05,
      "loss": 2.0788,
      "step": 1050
    },
    {
      "epoch": 0.25,
      "learning_rate": 1e-05,
      "loss": 2.1169,
      "step": 1060
    },
    {
      "epoch": 0.26,
      "learning_rate": 1e-05,
      "loss": 2.0941,
      "step": 1070
    },
    {
      "epoch": 0.26,
      "learning_rate": 1e-05,
      "loss": 2.2198,
      "step": 1080
    },
    {
      "epoch": 0.26,
      "learning_rate": 1e-05,
      "loss": 2.0716,
      "step": 1090
    },
    {
      "epoch": 0.26,
      "learning_rate": 1e-05,
      "loss": 1.9843,
      "step": 1100
    },
    {
      "epoch": 0.26,
      "eval_accuracy": 0.31140350877192985,
      "eval_loss": 1.8962386846542358,
      "eval_runtime": 165.9557,
      "eval_samples_per_second": 4.122,
      "step": 1100
    },
    {
      "epoch": 0.27,
      "learning_rate": 1e-05,
      "loss": 2.0203,
      "step": 1110
    },
    {
      "epoch": 0.27,
      "learning_rate": 1e-05,
      "loss": 2.0442,
      "step": 1120
    },
    {
      "epoch": 0.27,
      "learning_rate": 1e-05,
      "loss": 2.1095,
      "step": 1130
    },
    {
      "epoch": 0.27,
      "learning_rate": 1e-05,
      "loss": 1.8696,
      "step": 1140
    },
    {
      "epoch": 0.27,
      "learning_rate": 1e-05,
      "loss": 1.9007,
      "step": 1150
    },
    {
      "epoch": 0.28,
      "learning_rate": 1e-05,
      "loss": 2.0702,
      "step": 1160
    },
    {
      "epoch": 0.28,
      "learning_rate": 1e-05,
      "loss": 2.0819,
      "step": 1170
    },
    {
      "epoch": 0.28,
      "learning_rate": 1e-05,
      "loss": 2.0223,
      "step": 1180
    },
    {
      "epoch": 0.28,
      "learning_rate": 1e-05,
      "loss": 1.9159,
      "step": 1190
    },
    {
      "epoch": 0.29,
      "learning_rate": 1e-05,
      "loss": 1.9619,
      "step": 1200
    },
    {
      "epoch": 0.29,
      "eval_accuracy": 0.3201754385964912,
      "eval_loss": 1.8673094511032104,
      "eval_runtime": 196.7341,
      "eval_samples_per_second": 3.477,
      "step": 1200
    },
    {
      "epoch": 0.29,
      "learning_rate": 1e-05,
      "loss": 1.9303,
      "step": 1210
    },
    {
      "epoch": 0.29,
      "learning_rate": 1e-05,
      "loss": 1.9364,
      "step": 1220
    },
    {
      "epoch": 0.29,
      "learning_rate": 1e-05,
      "loss": 2.1095,
      "step": 1230
    },
    {
      "epoch": 0.3,
      "learning_rate": 1e-05,
      "loss": 2.1308,
      "step": 1240
    },
    {
      "epoch": 0.3,
      "learning_rate": 1e-05,
      "loss": 1.8673,
      "step": 1250
    },
    {
      "epoch": 0.3,
      "learning_rate": 1e-05,
      "loss": 2.0606,
      "step": 1260
    },
    {
      "epoch": 0.3,
      "learning_rate": 1e-05,
      "loss": 2.005,
      "step": 1270
    },
    {
      "epoch": 0.31,
      "learning_rate": 1e-05,
      "loss": 2.0499,
      "step": 1280
    },
    {
      "epoch": 0.31,
      "learning_rate": 1e-05,
      "loss": 1.9239,
      "step": 1290
    },
    {
      "epoch": 0.31,
      "learning_rate": 1e-05,
      "loss": 1.9639,
      "step": 1300
    },
    {
      "epoch": 0.31,
      "eval_accuracy": 0.3406432748538012,
      "eval_loss": 1.8431274890899658,
      "eval_runtime": 134.2472,
      "eval_samples_per_second": 5.095,
      "step": 1300
    },
    {
      "epoch": 0.31,
      "learning_rate": 1e-05,
      "loss": 2.0496,
      "step": 1310
    },
    {
      "epoch": 0.32,
      "learning_rate": 1e-05,
      "loss": 2.0673,
      "step": 1320
    },
    {
      "epoch": 0.32,
      "learning_rate": 1e-05,
      "loss": 2.0125,
      "step": 1330
    },
    {
      "epoch": 0.32,
      "learning_rate": 1e-05,
      "loss": 1.901,
      "step": 1340
    },
    {
      "epoch": 0.32,
      "learning_rate": 1e-05,
      "loss": 1.9696,
      "step": 1350
    },
    {
      "epoch": 0.32,
      "learning_rate": 1e-05,
      "loss": 1.8992,
      "step": 1360
    },
    {
      "epoch": 0.33,
      "learning_rate": 1e-05,
      "loss": 2.0119,
      "step": 1370
    },
    {
      "epoch": 0.33,
      "learning_rate": 1e-05,
      "loss": 1.978,
      "step": 1380
    },
    {
      "epoch": 0.33,
      "learning_rate": 1e-05,
      "loss": 2.0195,
      "step": 1390
    },
    {
      "epoch": 0.33,
      "learning_rate": 1e-05,
      "loss": 2.0739,
      "step": 1400
    },
    {
      "epoch": 0.33,
      "eval_accuracy": 0.3742690058479532,
      "eval_loss": 1.8214607238769531,
      "eval_runtime": 104.8858,
      "eval_samples_per_second": 6.521,
      "step": 1400
    },
    {
      "epoch": 0.34,
      "learning_rate": 1e-05,
      "loss": 1.8597,
      "step": 1410
    },
    {
      "epoch": 0.34,
      "learning_rate": 1e-05,
      "loss": 2.0765,
      "step": 1420
    },
    {
      "epoch": 0.34,
      "learning_rate": 1e-05,
      "loss": 1.9985,
      "step": 1430
    },
    {
      "epoch": 0.34,
      "learning_rate": 1e-05,
      "loss": 1.8301,
      "step": 1440
    },
    {
      "epoch": 0.35,
      "learning_rate": 1e-05,
      "loss": 1.871,
      "step": 1450
    },
    {
      "epoch": 0.35,
      "learning_rate": 1e-05,
      "loss": 2.0006,
      "step": 1460
    },
    {
      "epoch": 0.35,
      "learning_rate": 1e-05,
      "loss": 1.954,
      "step": 1470
    },
    {
      "epoch": 0.35,
      "learning_rate": 1e-05,
      "loss": 1.9459,
      "step": 1480
    },
    {
      "epoch": 0.36,
      "learning_rate": 1e-05,
      "loss": 1.9745,
      "step": 1490
    },
    {
      "epoch": 0.36,
      "learning_rate": 1e-05,
      "loss": 1.8441,
      "step": 1500
    },
    {
      "epoch": 0.36,
      "eval_accuracy": 0.38596491228070173,
      "eval_loss": 1.8044822216033936,
      "eval_runtime": 104.4458,
      "eval_samples_per_second": 6.549,
      "step": 1500
    },
    {
      "epoch": 0.36,
      "learning_rate": 1e-05,
      "loss": 1.8921,
      "step": 1510
    },
    {
      "epoch": 0.36,
      "learning_rate": 1e-05,
      "loss": 1.8944,
      "step": 1520
    },
    {
      "epoch": 0.37,
      "learning_rate": 1e-05,
      "loss": 1.9011,
      "step": 1530
    },
    {
      "epoch": 0.37,
      "learning_rate": 1e-05,
      "loss": 1.8769,
      "step": 1540
    },
    {
      "epoch": 0.37,
      "learning_rate": 1e-05,
      "loss": 1.9468,
      "step": 1550
    },
    {
      "epoch": 0.37,
      "learning_rate": 1e-05,
      "loss": 1.8655,
      "step": 1560
    },
    {
      "epoch": 0.37,
      "learning_rate": 1e-05,
      "loss": 1.9192,
      "step": 1570
    },
    {
      "epoch": 0.38,
      "learning_rate": 1e-05,
      "loss": 1.9333,
      "step": 1580
    },
    {
      "epoch": 0.38,
      "learning_rate": 1e-05,
      "loss": 1.8474,
      "step": 1590
    },
    {
      "epoch": 0.38,
      "learning_rate": 1e-05,
      "loss": 1.9127,
      "step": 1600
    },
    {
      "epoch": 0.38,
      "eval_accuracy": 0.3932748538011696,
      "eval_loss": 1.7868342399597168,
      "eval_runtime": 104.5599,
      "eval_samples_per_second": 6.542,
      "step": 1600
    },
    {
      "epoch": 0.38,
      "learning_rate": 1e-05,
      "loss": 1.9016,
      "step": 1610
    },
    {
      "epoch": 0.39,
      "learning_rate": 1e-05,
      "loss": 1.9871,
      "step": 1620
    },
    {
      "epoch": 0.39,
      "learning_rate": 1e-05,
      "loss": 1.9364,
      "step": 1630
    },
    {
      "epoch": 0.39,
      "learning_rate": 1e-05,
      "loss": 1.9199,
      "step": 1640
    },
    {
      "epoch": 0.39,
      "learning_rate": 1e-05,
      "loss": 1.9942,
      "step": 1650
    },
    {
      "epoch": 0.4,
      "learning_rate": 1e-05,
      "loss": 1.865,
      "step": 1660
    },
    {
      "epoch": 0.4,
      "learning_rate": 1e-05,
      "loss": 1.8613,
      "step": 1670
    },
    {
      "epoch": 0.4,
      "learning_rate": 1e-05,
      "loss": 1.897,
      "step": 1680
    },
    {
      "epoch": 0.4,
      "learning_rate": 1e-05,
      "loss": 1.8116,
      "step": 1690
    },
    {
      "epoch": 0.41,
      "learning_rate": 1e-05,
      "loss": 1.8866,
      "step": 1700
    },
    {
      "epoch": 0.41,
      "eval_accuracy": 0.391812865497076,
      "eval_loss": 1.7732278108596802,
      "eval_runtime": 96.261,
      "eval_samples_per_second": 7.106,
      "step": 1700
    },
    {
      "epoch": 0.41,
      "learning_rate": 1e-05,
      "loss": 1.8727,
      "step": 1710
    },
    {
      "epoch": 0.41,
      "learning_rate": 1e-05,
      "loss": 1.8713,
      "step": 1720
    },
    {
      "epoch": 0.41,
      "learning_rate": 1e-05,
      "loss": 1.8907,
      "step": 1730
    },
    {
      "epoch": 0.42,
      "learning_rate": 1e-05,
      "loss": 1.9466,
      "step": 1740
    },
    {
      "epoch": 0.42,
      "learning_rate": 1e-05,
      "loss": 1.8995,
      "step": 1750
    },
    {
      "epoch": 0.42,
      "learning_rate": 1e-05,
      "loss": 2.0104,
      "step": 1760
    },
    {
      "epoch": 0.42,
      "learning_rate": 1e-05,
      "loss": 1.8043,
      "step": 1770
    },
    {
      "epoch": 0.43,
      "learning_rate": 1e-05,
      "loss": 1.9651,
      "step": 1780
    },
    {
      "epoch": 0.43,
      "learning_rate": 1e-05,
      "loss": 1.8012,
      "step": 1790
    },
    {
      "epoch": 0.43,
      "learning_rate": 1e-05,
      "loss": 1.8971,
      "step": 1800
    },
    {
      "epoch": 0.43,
      "eval_accuracy": 0.4064327485380117,
      "eval_loss": 1.758914828300476,
      "eval_runtime": 104.047,
      "eval_samples_per_second": 6.574,
      "step": 1800
    },
    {
      "epoch": 0.43,
      "learning_rate": 1e-05,
      "loss": 1.9777,
      "step": 1810
    },
    {
      "epoch": 0.43,
      "learning_rate": 1e-05,
      "loss": 1.8106,
      "step": 1820
    },
    {
      "epoch": 0.44,
      "learning_rate": 1e-05,
      "loss": 1.8128,
      "step": 1830
    },
    {
      "epoch": 0.44,
      "learning_rate": 1e-05,
      "loss": 1.8316,
      "step": 1840
    },
    {
      "epoch": 0.44,
      "learning_rate": 1e-05,
      "loss": 1.9456,
      "step": 1850
    },
    {
      "epoch": 0.44,
      "learning_rate": 1e-05,
      "loss": 1.9863,
      "step": 1860
    },
    {
      "epoch": 0.45,
      "learning_rate": 1e-05,
      "loss": 1.7982,
      "step": 1870
    },
    {
      "epoch": 0.45,
      "learning_rate": 1e-05,
      "loss": 1.8757,
      "step": 1880
    },
    {
      "epoch": 0.45,
      "learning_rate": 1e-05,
      "loss": 1.8086,
      "step": 1890
    },
    {
      "epoch": 0.45,
      "learning_rate": 1e-05,
      "loss": 1.8534,
      "step": 1900
    },
    {
      "epoch": 0.45,
      "eval_accuracy": 0.41228070175438597,
      "eval_loss": 1.7410695552825928,
      "eval_runtime": 104.1098,
      "eval_samples_per_second": 6.57,
      "step": 1900
    },
    {
      "epoch": 0.46,
      "learning_rate": 1e-05,
      "loss": 1.8921,
      "step": 1910
    },
    {
      "epoch": 0.46,
      "learning_rate": 1e-05,
      "loss": 1.968,
      "step": 1920
    },
    {
      "epoch": 0.46,
      "learning_rate": 1e-05,
      "loss": 1.8086,
      "step": 1930
    },
    {
      "epoch": 0.46,
      "learning_rate": 1e-05,
      "loss": 1.7912,
      "step": 1940
    },
    {
      "epoch": 0.47,
      "learning_rate": 1e-05,
      "loss": 1.8809,
      "step": 1950
    },
    {
      "epoch": 0.47,
      "learning_rate": 1e-05,
      "loss": 1.8652,
      "step": 1960
    },
    {
      "epoch": 0.47,
      "learning_rate": 1e-05,
      "loss": 2.1881,
      "step": 1970
    },
    {
      "epoch": 0.47,
      "learning_rate": 1e-05,
      "loss": 1.8155,
      "step": 1980
    },
    {
      "epoch": 0.48,
      "learning_rate": 1e-05,
      "loss": 1.8745,
      "step": 1990
    },
    {
      "epoch": 0.48,
      "learning_rate": 1e-05,
      "loss": 1.8658,
      "step": 2000
    },
    {
      "epoch": 0.48,
      "eval_accuracy": 0.4195906432748538,
      "eval_loss": 1.7358570098876953,
      "eval_runtime": 103.6997,
      "eval_samples_per_second": 6.596,
      "step": 2000
    },
    {
      "epoch": 0.48,
      "learning_rate": 1e-05,
      "loss": 1.8854,
      "step": 2010
    },
    {
      "epoch": 0.48,
      "learning_rate": 1e-05,
      "loss": 1.7842,
      "step": 2020
    },
    {
      "epoch": 0.48,
      "learning_rate": 1e-05,
      "loss": 1.8688,
      "step": 2030
    },
    {
      "epoch": 0.49,
      "learning_rate": 1e-05,
      "loss": 1.7826,
      "step": 2040
    },
    {
      "epoch": 0.49,
      "learning_rate": 1e-05,
      "loss": 1.916,
      "step": 2050
    },
    {
      "epoch": 0.49,
      "learning_rate": 1e-05,
      "loss": 1.8807,
      "step": 2060
    },
    {
      "epoch": 0.49,
      "learning_rate": 1e-05,
      "loss": 1.8288,
      "step": 2070
    },
    {
      "epoch": 0.5,
      "learning_rate": 1e-05,
      "loss": 1.8808,
      "step": 2080
    },
    {
      "epoch": 0.5,
      "learning_rate": 1e-05,
      "loss": 1.7723,
      "step": 2090
    },
    {
      "epoch": 0.5,
      "learning_rate": 1e-05,
      "loss": 1.8761,
      "step": 2100
    },
    {
      "epoch": 0.5,
      "eval_accuracy": 0.4269005847953216,
      "eval_loss": 1.7221930027008057,
      "eval_runtime": 103.5291,
      "eval_samples_per_second": 6.607,
      "step": 2100
    },
    {
      "epoch": 0.5,
      "learning_rate": 1e-05,
      "loss": 1.7881,
      "step": 2110
    },
    {
      "epoch": 0.51,
      "learning_rate": 1e-05,
      "loss": 1.82,
      "step": 2120
    },
    {
      "epoch": 0.51,
      "learning_rate": 1e-05,
      "loss": 1.7952,
      "step": 2130
    },
    {
      "epoch": 0.51,
      "learning_rate": 1e-05,
      "loss": 1.7979,
      "step": 2140
    },
    {
      "epoch": 0.51,
      "learning_rate": 1e-05,
      "loss": 1.8364,
      "step": 2150
    },
    {
      "epoch": 0.52,
      "learning_rate": 1e-05,
      "loss": 1.8128,
      "step": 2160
    },
    {
      "epoch": 0.52,
      "learning_rate": 1e-05,
      "loss": 1.8586,
      "step": 2170
    },
    {
      "epoch": 0.52,
      "learning_rate": 1e-05,
      "loss": 1.772,
      "step": 2180
    },
    {
      "epoch": 0.52,
      "learning_rate": 1e-05,
      "loss": 1.9196,
      "step": 2190
    },
    {
      "epoch": 0.53,
      "learning_rate": 1e-05,
      "loss": 1.8123,
      "step": 2200
    },
    {
      "epoch": 0.53,
      "eval_accuracy": 0.4269005847953216,
      "eval_loss": 1.7114078998565674,
      "eval_runtime": 103.6616,
      "eval_samples_per_second": 6.598,
      "step": 2200
    },
    {
      "epoch": 0.53,
      "learning_rate": 1e-05,
      "loss": 1.8765,
      "step": 2210
    },
    {
      "epoch": 0.53,
      "learning_rate": 1e-05,
      "loss": 1.8477,
      "step": 2220
    },
    {
      "epoch": 0.53,
      "learning_rate": 1e-05,
      "loss": 1.8256,
      "step": 2230
    },
    {
      "epoch": 0.53,
      "learning_rate": 1e-05,
      "loss": 1.8625,
      "step": 2240
    },
    {
      "epoch": 0.54,
      "learning_rate": 1e-05,
      "loss": 1.7471,
      "step": 2250
    },
    {
      "epoch": 0.54,
      "learning_rate": 1e-05,
      "loss": 1.827,
      "step": 2260
    },
    {
      "epoch": 0.54,
      "learning_rate": 1e-05,
      "loss": 1.9129,
      "step": 2270
    },
    {
      "epoch": 0.54,
      "learning_rate": 1e-05,
      "loss": 1.8397,
      "step": 2280
    },
    {
      "epoch": 0.55,
      "learning_rate": 1e-05,
      "loss": 1.8185,
      "step": 2290
    },
    {
      "epoch": 0.55,
      "learning_rate": 1e-05,
      "loss": 1.9446,
      "step": 2300
    },
    {
      "epoch": 0.55,
      "eval_accuracy": 0.43567251461988304,
      "eval_loss": 1.7006075382232666,
      "eval_runtime": 102.9944,
      "eval_samples_per_second": 6.641,
      "step": 2300
    },
    {
      "epoch": 0.55,
      "learning_rate": 1e-05,
      "loss": 1.8025,
      "step": 2310
    },
    {
      "epoch": 0.55,
      "learning_rate": 1e-05,
      "loss": 1.8305,
      "step": 2320
    },
    {
      "epoch": 0.56,
      "learning_rate": 1e-05,
      "loss": 1.7644,
      "step": 2330
    },
    {
      "epoch": 0.56,
      "learning_rate": 1e-05,
      "loss": 1.8361,
      "step": 2340
    },
    {
      "epoch": 0.56,
      "learning_rate": 1e-05,
      "loss": 1.7868,
      "step": 2350
    },
    {
      "epoch": 0.56,
      "learning_rate": 1e-05,
      "loss": 1.8593,
      "step": 2360
    },
    {
      "epoch": 0.57,
      "learning_rate": 1e-05,
      "loss": 1.8839,
      "step": 2370
    },
    {
      "epoch": 0.57,
      "learning_rate": 1e-05,
      "loss": 1.7811,
      "step": 2380
    },
    {
      "epoch": 0.57,
      "learning_rate": 1e-05,
      "loss": 1.8103,
      "step": 2390
    },
    {
      "epoch": 0.57,
      "learning_rate": 1e-05,
      "loss": 1.8119,
      "step": 2400
    },
    {
      "epoch": 0.57,
      "eval_accuracy": 0.4415204678362573,
      "eval_loss": 1.6913487911224365,
      "eval_runtime": 102.5309,
      "eval_samples_per_second": 6.671,
      "step": 2400
    },
    {
      "epoch": 0.58,
      "learning_rate": 1e-05,
      "loss": 1.7505,
      "step": 2410
    },
    {
      "epoch": 0.58,
      "learning_rate": 1e-05,
      "loss": 1.7734,
      "step": 2420
    },
    {
      "epoch": 0.58,
      "learning_rate": 1e-05,
      "loss": 1.7873,
      "step": 2430
    },
    {
      "epoch": 0.58,
      "learning_rate": 1e-05,
      "loss": 1.8405,
      "step": 2440
    },
    {
      "epoch": 0.59,
      "learning_rate": 1e-05,
      "loss": 1.767,
      "step": 2450
    },
    {
      "epoch": 0.59,
      "learning_rate": 1e-05,
      "loss": 1.8,
      "step": 2460
    },
    {
      "epoch": 0.59,
      "learning_rate": 1e-05,
      "loss": 1.8127,
      "step": 2470
    },
    {
      "epoch": 0.59,
      "learning_rate": 1e-05,
      "loss": 1.7989,
      "step": 2480
    },
    {
      "epoch": 0.59,
      "learning_rate": 1e-05,
      "loss": 1.8115,
      "step": 2490
    },
    {
      "epoch": 0.6,
      "learning_rate": 1e-05,
      "loss": 1.7976,
      "step": 2500
    },
    {
      "epoch": 0.6,
      "eval_accuracy": 0.45614035087719296,
      "eval_loss": 1.685728907585144,
      "eval_runtime": 71.5626,
      "eval_samples_per_second": 9.558,
      "step": 2500
    },
    {
      "epoch": 0.6,
      "learning_rate": 1e-05,
      "loss": 1.7376,
      "step": 2510
    },
    {
      "epoch": 0.6,
      "learning_rate": 1e-05,
      "loss": 1.6886,
      "step": 2520
    },
    {
      "epoch": 0.6,
      "learning_rate": 1e-05,
      "loss": 1.806,
      "step": 2530
    },
    {
      "epoch": 0.61,
      "learning_rate": 1e-05,
      "loss": 1.7771,
      "step": 2540
    },
    {
      "epoch": 0.61,
      "learning_rate": 1e-05,
      "loss": 1.7164,
      "step": 2550
    },
    {
      "epoch": 0.61,
      "learning_rate": 1e-05,
      "loss": 1.8224,
      "step": 2560
    },
    {
      "epoch": 0.61,
      "learning_rate": 1e-05,
      "loss": 1.8404,
      "step": 2570
    },
    {
      "epoch": 0.62,
      "learning_rate": 1e-05,
      "loss": 1.7256,
      "step": 2580
    },
    {
      "epoch": 0.62,
      "learning_rate": 1e-05,
      "loss": 1.7847,
      "step": 2590
    },
    {
      "epoch": 0.62,
      "learning_rate": 1e-05,
      "loss": 1.6924,
      "step": 2600
    },
    {
      "epoch": 0.62,
      "eval_accuracy": 0.4649122807017544,
      "eval_loss": 1.6748087406158447,
      "eval_runtime": 56.7294,
      "eval_samples_per_second": 12.057,
      "step": 2600
    },
    {
      "epoch": 0.62,
      "learning_rate": 1e-05,
      "loss": 1.7252,
      "step": 2610
    },
    {
      "epoch": 0.63,
      "learning_rate": 1e-05,
      "loss": 1.7002,
      "step": 2620
    },
    {
      "epoch": 0.63,
      "learning_rate": 1e-05,
      "loss": 1.7566,
      "step": 2630
    },
    {
      "epoch": 0.63,
      "learning_rate": 1e-05,
      "loss": 1.7364,
      "step": 2640
    },
    {
      "epoch": 0.63,
      "learning_rate": 1e-05,
      "loss": 1.7749,
      "step": 2650
    },
    {
      "epoch": 0.64,
      "learning_rate": 1e-05,
      "loss": 1.7941,
      "step": 2660
    },
    {
      "epoch": 0.64,
      "learning_rate": 1e-05,
      "loss": 1.7753,
      "step": 2670
    },
    {
      "epoch": 0.64,
      "learning_rate": 1e-05,
      "loss": 1.7482,
      "step": 2680
    },
    {
      "epoch": 0.64,
      "learning_rate": 1e-05,
      "loss": 1.742,
      "step": 2690
    },
    {
      "epoch": 0.64,
      "learning_rate": 1e-05,
      "loss": 1.8214,
      "step": 2700
    },
    {
      "epoch": 0.64,
      "eval_accuracy": 0.47076023391812866,
      "eval_loss": 1.670298457145691,
      "eval_runtime": 71.7103,
      "eval_samples_per_second": 9.538,
      "step": 2700
    },
    {
      "epoch": 0.65,
      "learning_rate": 1e-05,
      "loss": 1.7248,
      "step": 2710
    },
    {
      "epoch": 0.65,
      "learning_rate": 1e-05,
      "loss": 1.7592,
      "step": 2720
    },
    {
      "epoch": 0.65,
      "learning_rate": 1e-05,
      "loss": 1.8303,
      "step": 2730
    },
    {
      "epoch": 0.65,
      "learning_rate": 1e-05,
      "loss": 1.7499,
      "step": 2740
    },
    {
      "epoch": 0.66,
      "learning_rate": 1e-05,
      "loss": 1.7656,
      "step": 2750
    },
    {
      "epoch": 0.66,
      "learning_rate": 1e-05,
      "loss": 1.7542,
      "step": 2760
    },
    {
      "epoch": 0.66,
      "learning_rate": 1e-05,
      "loss": 1.8009,
      "step": 2770
    },
    {
      "epoch": 0.66,
      "learning_rate": 1e-05,
      "loss": 1.7808,
      "step": 2780
    },
    {
      "epoch": 0.67,
      "learning_rate": 1e-05,
      "loss": 1.7339,
      "step": 2790
    },
    {
      "epoch": 0.67,
      "learning_rate": 1e-05,
      "loss": 1.7612,
      "step": 2800
    },
    {
      "epoch": 0.67,
      "eval_accuracy": 0.48391812865497075,
      "eval_loss": 1.6625455617904663,
      "eval_runtime": 80.1539,
      "eval_samples_per_second": 8.534,
      "step": 2800
    },
    {
      "epoch": 0.67,
      "learning_rate": 1e-05,
      "loss": 1.7074,
      "step": 2810
    },
    {
      "epoch": 0.67,
      "learning_rate": 1e-05,
      "loss": 1.792,
      "step": 2820
    },
    {
      "epoch": 0.68,
      "learning_rate": 1e-05,
      "loss": 1.7253,
      "step": 2830
    },
    {
      "epoch": 0.68,
      "learning_rate": 1e-05,
      "loss": 1.841,
      "step": 2840
    },
    {
      "epoch": 0.68,
      "learning_rate": 1e-05,
      "loss": 1.6939,
      "step": 2850
    },
    {
      "epoch": 0.68,
      "learning_rate": 1e-05,
      "loss": 1.7947,
      "step": 2860
    },
    {
      "epoch": 0.69,
      "learning_rate": 1e-05,
      "loss": 1.7101,
      "step": 2870
    },
    {
      "epoch": 0.69,
      "learning_rate": 1e-05,
      "loss": 1.7323,
      "step": 2880
    },
    {
      "epoch": 0.69,
      "learning_rate": 1e-05,
      "loss": 1.7972,
      "step": 2890
    },
    {
      "epoch": 0.69,
      "learning_rate": 1e-05,
      "loss": 1.7758,
      "step": 2900
    },
    {
      "epoch": 0.69,
      "eval_accuracy": 0.47514619883040937,
      "eval_loss": 1.6513162851333618,
      "eval_runtime": 80.0058,
      "eval_samples_per_second": 8.549,
      "step": 2900
    },
    {
      "epoch": 0.7,
      "learning_rate": 1e-05,
      "loss": 1.7389,
      "step": 2910
    },
    {
      "epoch": 0.7,
      "learning_rate": 1e-05,
      "loss": 1.791,
      "step": 2920
    },
    {
      "epoch": 0.7,
      "learning_rate": 1e-05,
      "loss": 1.7669,
      "step": 2930
    },
    {
      "epoch": 0.7,
      "learning_rate": 1e-05,
      "loss": 1.7677,
      "step": 2940
    },
    {
      "epoch": 0.7,
      "learning_rate": 1e-05,
      "loss": 1.7082,
      "step": 2950
    },
    {
      "epoch": 0.71,
      "learning_rate": 1e-05,
      "loss": 1.7173,
      "step": 2960
    },
    {
      "epoch": 0.71,
      "learning_rate": 1e-05,
      "loss": 1.729,
      "step": 2970
    },
    {
      "epoch": 0.71,
      "learning_rate": 1e-05,
      "loss": 1.6807,
      "step": 2980
    },
    {
      "epoch": 0.71,
      "learning_rate": 1e-05,
      "loss": 1.8942,
      "step": 2990
    },
    {
      "epoch": 0.72,
      "learning_rate": 1e-05,
      "loss": 1.7794,
      "step": 3000
    },
    {
      "epoch": 0.72,
      "eval_accuracy": 0.4853801169590643,
      "eval_loss": 1.6467478275299072,
      "eval_runtime": 81.0699,
      "eval_samples_per_second": 8.437,
      "step": 3000
    },
    {
      "epoch": 0.72,
      "learning_rate": 1e-05,
      "loss": 1.7973,
      "step": 3010
    },
    {
      "epoch": 0.72,
      "learning_rate": 1e-05,
      "loss": 1.7659,
      "step": 3020
    },
    {
      "epoch": 0.72,
      "learning_rate": 1e-05,
      "loss": 1.7583,
      "step": 3030
    },
    {
      "epoch": 0.73,
      "learning_rate": 1e-05,
      "loss": 1.7761,
      "step": 3040
    },
    {
      "epoch": 0.73,
      "learning_rate": 1e-05,
      "loss": 1.8136,
      "step": 3050
    },
    {
      "epoch": 0.73,
      "learning_rate": 1e-05,
      "loss": 1.6649,
      "step": 3060
    },
    {
      "epoch": 0.73,
      "learning_rate": 1e-05,
      "loss": 1.7221,
      "step": 3070
    },
    {
      "epoch": 0.74,
      "learning_rate": 1e-05,
      "loss": 1.6792,
      "step": 3080
    },
    {
      "epoch": 0.74,
      "learning_rate": 1e-05,
      "loss": 1.7354,
      "step": 3090
    },
    {
      "epoch": 0.74,
      "learning_rate": 1e-05,
      "loss": 1.7973,
      "step": 3100
    },
    {
      "epoch": 0.74,
      "eval_accuracy": 0.49707602339181284,
      "eval_loss": 1.6398402452468872,
      "eval_runtime": 57.9169,
      "eval_samples_per_second": 11.81,
      "step": 3100
    },
    {
      "epoch": 0.74,
      "learning_rate": 1e-05,
      "loss": 1.7468,
      "step": 3110
    },
    {
      "epoch": 0.75,
      "learning_rate": 1e-05,
      "loss": 1.6785,
      "step": 3120
    },
    {
      "epoch": 0.75,
      "learning_rate": 1e-05,
      "loss": 1.695,
      "step": 3130
    },
    {
      "epoch": 0.75,
      "learning_rate": 1e-05,
      "loss": 1.6795,
      "step": 3140
    },
    {
      "epoch": 0.75,
      "learning_rate": 1e-05,
      "loss": 1.7642,
      "step": 3150
    },
    {
      "epoch": 0.75,
      "learning_rate": 1e-05,
      "loss": 1.734,
      "step": 3160
    },
    {
      "epoch": 0.76,
      "learning_rate": 1e-05,
      "loss": 1.7136,
      "step": 3170
    },
    {
      "epoch": 0.76,
      "learning_rate": 1e-05,
      "loss": 1.7177,
      "step": 3180
    },
    {
      "epoch": 0.76,
      "learning_rate": 1e-05,
      "loss": 1.6806,
      "step": 3190
    },
    {
      "epoch": 0.76,
      "learning_rate": 1e-05,
      "loss": 1.789,
      "step": 3200
    },
    {
      "epoch": 0.76,
      "eval_accuracy": 0.4956140350877193,
      "eval_loss": 1.6345957517623901,
      "eval_runtime": 44.4888,
      "eval_samples_per_second": 15.375,
      "step": 3200
    },
    {
      "epoch": 0.77,
      "learning_rate": 1e-05,
      "loss": 1.7254,
      "step": 3210
    },
    {
      "epoch": 0.77,
      "learning_rate": 1e-05,
      "loss": 1.7461,
      "step": 3220
    },
    {
      "epoch": 0.77,
      "learning_rate": 1e-05,
      "loss": 1.7187,
      "step": 3230
    },
    {
      "epoch": 0.77,
      "learning_rate": 1e-05,
      "loss": 1.7273,
      "step": 3240
    },
    {
      "epoch": 0.78,
      "learning_rate": 1e-05,
      "loss": 1.6796,
      "step": 3250
    },
    {
      "epoch": 0.78,
      "learning_rate": 1e-05,
      "loss": 1.6988,
      "step": 3260
    },
    {
      "epoch": 0.78,
      "learning_rate": 1e-05,
      "loss": 1.711,
      "step": 3270
    },
    {
      "epoch": 0.78,
      "learning_rate": 1e-05,
      "loss": 1.7209,
      "step": 3280
    },
    {
      "epoch": 0.79,
      "learning_rate": 1e-05,
      "loss": 1.6965,
      "step": 3290
    },
    {
      "epoch": 0.79,
      "learning_rate": 1e-05,
      "loss": 1.6884,
      "step": 3300
    },
    {
      "epoch": 0.79,
      "eval_accuracy": 0.5102339181286549,
      "eval_loss": 1.6261690855026245,
      "eval_runtime": 47.78,
      "eval_samples_per_second": 14.316,
      "step": 3300
    },
    {
      "epoch": 0.79,
      "learning_rate": 1e-05,
      "loss": 1.6857,
      "step": 3310
    },
    {
      "epoch": 0.79,
      "learning_rate": 1e-05,
      "loss": 1.696,
      "step": 3320
    },
    {
      "epoch": 0.8,
      "learning_rate": 1e-05,
      "loss": 1.7406,
      "step": 3330
    },
    {
      "epoch": 0.8,
      "learning_rate": 1e-05,
      "loss": 1.7856,
      "step": 3340
    },
    {
      "epoch": 0.8,
      "learning_rate": 1e-05,
      "loss": 1.6767,
      "step": 3350
    },
    {
      "epoch": 0.8,
      "learning_rate": 1e-05,
      "loss": 1.7496,
      "step": 3360
    },
    {
      "epoch": 0.8,
      "learning_rate": 1e-05,
      "loss": 1.749,
      "step": 3370
    },
    {
      "epoch": 0.81,
      "learning_rate": 1e-05,
      "loss": 1.6974,
      "step": 3380
    },
    {
      "epoch": 0.81,
      "learning_rate": 1e-05,
      "loss": 1.7174,
      "step": 3390
    },
    {
      "epoch": 0.81,
      "learning_rate": 1e-05,
      "loss": 1.715,
      "step": 3400
    },
    {
      "epoch": 0.81,
      "eval_accuracy": 0.5248538011695907,
      "eval_loss": 1.6260160207748413,
      "eval_runtime": 40.0278,
      "eval_samples_per_second": 17.088,
      "step": 3400
    },
    {
      "epoch": 0.81,
      "learning_rate": 1e-05,
      "loss": 1.7355,
      "step": 3410
    },
    {
      "epoch": 0.82,
      "learning_rate": 1e-05,
      "loss": 1.691,
      "step": 3420
    },
    {
      "epoch": 0.82,
      "learning_rate": 1e-05,
      "loss": 1.7381,
      "step": 3430
    },
    {
      "epoch": 0.82,
      "learning_rate": 1e-05,
      "loss": 1.7021,
      "step": 3440
    },
    {
      "epoch": 0.82,
      "learning_rate": 1e-05,
      "loss": 1.7303,
      "step": 3450
    },
    {
      "epoch": 0.83,
      "learning_rate": 1e-05,
      "loss": 1.665,
      "step": 3460
    },
    {
      "epoch": 0.83,
      "learning_rate": 1e-05,
      "loss": 1.7162,
      "step": 3470
    },
    {
      "epoch": 0.83,
      "learning_rate": 1e-05,
      "loss": 1.7308,
      "step": 3480
    },
    {
      "epoch": 0.83,
      "learning_rate": 1e-05,
      "loss": 1.7716,
      "step": 3490
    },
    {
      "epoch": 0.84,
      "learning_rate": 1e-05,
      "loss": 1.7359,
      "step": 3500
    },
    {
      "epoch": 0.84,
      "eval_accuracy": 0.5219298245614035,
      "eval_loss": 1.6194047927856445,
      "eval_runtime": 39.7729,
      "eval_samples_per_second": 17.198,
      "step": 3500
    },
    {
      "epoch": 0.84,
      "learning_rate": 1e-05,
      "loss": 1.7375,
      "step": 3510
    },
    {
      "epoch": 0.84,
      "learning_rate": 1e-05,
      "loss": 1.7265,
      "step": 3520
    },
    {
      "epoch": 0.84,
      "learning_rate": 1e-05,
      "loss": 1.6752,
      "step": 3530
    },
    {
      "epoch": 0.85,
      "learning_rate": 1e-05,
      "loss": 1.7195,
      "step": 3540
    },
    {
      "epoch": 0.85,
      "learning_rate": 1e-05,
      "loss": 1.7207,
      "step": 3550
    },
    {
      "epoch": 0.85,
      "learning_rate": 1e-05,
      "loss": 1.8081,
      "step": 3560
    },
    {
      "epoch": 0.85,
      "learning_rate": 1e-05,
      "loss": 1.6767,
      "step": 3570
    },
    {
      "epoch": 0.86,
      "learning_rate": 1e-05,
      "loss": 1.673,
      "step": 3580
    },
    {
      "epoch": 0.86,
      "learning_rate": 1e-05,
      "loss": 1.7715,
      "step": 3590
    },
    {
      "epoch": 0.86,
      "learning_rate": 1e-05,
      "loss": 1.6727,
      "step": 3600
    },
    {
      "epoch": 0.86,
      "eval_accuracy": 0.5409356725146199,
      "eval_loss": 1.613726019859314,
      "eval_runtime": 40.5697,
      "eval_samples_per_second": 16.86,
      "step": 3600
    },
    {
      "epoch": 0.86,
      "learning_rate": 1e-05,
      "loss": 1.6119,
      "step": 3610
    },
    {
      "epoch": 0.86,
      "learning_rate": 1e-05,
      "loss": 1.7694,
      "step": 3620
    },
    {
      "epoch": 0.87,
      "learning_rate": 1e-05,
      "loss": 1.6488,
      "step": 3630
    },
    {
      "epoch": 0.87,
      "learning_rate": 1e-05,
      "loss": 1.7007,
      "step": 3640
    },
    {
      "epoch": 0.87,
      "learning_rate": 1e-05,
      "loss": 1.7506,
      "step": 3650
    },
    {
      "epoch": 0.87,
      "learning_rate": 1e-05,
      "loss": 1.6749,
      "step": 3660
    },
    {
      "epoch": 0.88,
      "learning_rate": 1e-05,
      "loss": 1.7541,
      "step": 3670
    },
    {
      "epoch": 0.88,
      "learning_rate": 1e-05,
      "loss": 1.7384,
      "step": 3680
    },
    {
      "epoch": 0.88,
      "learning_rate": 1e-05,
      "loss": 1.7159,
      "step": 3690
    },
    {
      "epoch": 0.88,
      "learning_rate": 1e-05,
      "loss": 1.7272,
      "step": 3700
    },
    {
      "epoch": 0.88,
      "eval_accuracy": 0.5467836257309941,
      "eval_loss": 1.6103622913360596,
      "eval_runtime": 67.6243,
      "eval_samples_per_second": 10.115,
      "step": 3700
    },
    {
      "epoch": 0.89,
      "learning_rate": 1e-05,
      "loss": 1.6904,
      "step": 3710
    },
    {
      "epoch": 0.89,
      "learning_rate": 1e-05,
      "loss": 1.6705,
      "step": 3720
    },
    {
      "epoch": 0.89,
      "learning_rate": 1e-05,
      "loss": 1.6999,
      "step": 3730
    },
    {
      "epoch": 0.89,
      "learning_rate": 1e-05,
      "loss": 1.6837,
      "step": 3740
    },
    {
      "epoch": 0.9,
      "learning_rate": 1e-05,
      "loss": 1.6929,
      "step": 3750
    },
    {
      "epoch": 0.9,
      "learning_rate": 1e-05,
      "loss": 1.7638,
      "step": 3760
    },
    {
      "epoch": 0.9,
      "learning_rate": 1e-05,
      "loss": 1.6914,
      "step": 3770
    },
    {
      "epoch": 0.9,
      "learning_rate": 1e-05,
      "loss": 1.7398,
      "step": 3780
    },
    {
      "epoch": 0.91,
      "learning_rate": 1e-05,
      "loss": 1.7062,
      "step": 3790
    },
    {
      "epoch": 0.91,
      "learning_rate": 1e-05,
      "loss": 1.6286,
      "step": 3800
    },
    {
      "epoch": 0.91,
      "eval_accuracy": 0.5526315789473685,
      "eval_loss": 1.604485034942627,
      "eval_runtime": 38.1834,
      "eval_samples_per_second": 17.914,
      "step": 3800
    },
    {
      "epoch": 0.91,
      "learning_rate": 1e-05,
      "loss": 1.7133,
      "step": 3810
    },
    {
      "epoch": 0.91,
      "learning_rate": 1e-05,
      "loss": 1.7036,
      "step": 3820
    },
    {
      "epoch": 0.91,
      "learning_rate": 1e-05,
      "loss": 1.7221,
      "step": 3830
    },
    {
      "epoch": 0.92,
      "learning_rate": 1e-05,
      "loss": 1.6587,
      "step": 3840
    },
    {
      "epoch": 0.92,
      "learning_rate": 1e-05,
      "loss": 1.6693,
      "step": 3850
    },
    {
      "epoch": 0.92,
      "learning_rate": 1e-05,
      "loss": 1.7608,
      "step": 3860
    },
    {
      "epoch": 0.92,
      "learning_rate": 1e-05,
      "loss": 1.6798,
      "step": 3870
    },
    {
      "epoch": 0.93,
      "learning_rate": 1e-05,
      "loss": 1.7112,
      "step": 3880
    },
    {
      "epoch": 0.93,
      "learning_rate": 1e-05,
      "loss": 1.6338,
      "step": 3890
    },
    {
      "epoch": 0.93,
      "learning_rate": 1e-05,
      "loss": 1.6995,
      "step": 3900
    },
    {
      "epoch": 0.93,
      "eval_accuracy": 0.5511695906432749,
      "eval_loss": 1.6006512641906738,
      "eval_runtime": 80.5206,
      "eval_samples_per_second": 8.495,
      "step": 3900
    },
    {
      "epoch": 0.93,
      "learning_rate": 1e-05,
      "loss": 1.646,
      "step": 3910
    },
    {
      "epoch": 0.94,
      "learning_rate": 1e-05,
      "loss": 1.6668,
      "step": 3920
    },
    {
      "epoch": 0.94,
      "learning_rate": 1e-05,
      "loss": 1.6662,
      "step": 3930
    },
    {
      "epoch": 0.94,
      "learning_rate": 1e-05,
      "loss": 1.6455,
      "step": 3940
    },
    {
      "epoch": 0.94,
      "learning_rate": 1e-05,
      "loss": 1.6475,
      "step": 3950
    },
    {
      "epoch": 0.95,
      "learning_rate": 1e-05,
      "loss": 1.652,
      "step": 3960
    },
    {
      "epoch": 0.95,
      "learning_rate": 1e-05,
      "loss": 1.6624,
      "step": 3970
    },
    {
      "epoch": 0.95,
      "learning_rate": 1e-05,
      "loss": 1.675,
      "step": 3980
    },
    {
      "epoch": 0.95,
      "learning_rate": 1e-05,
      "loss": 1.6769,
      "step": 3990
    },
    {
      "epoch": 0.96,
      "learning_rate": 1e-05,
      "loss": 1.6511,
      "step": 4000
    },
    {
      "epoch": 0.96,
      "eval_accuracy": 0.5584795321637427,
      "eval_loss": 1.5977518558502197,
      "eval_runtime": 80.5314,
      "eval_samples_per_second": 8.494,
      "step": 4000
    },
    {
      "epoch": 0.96,
      "learning_rate": 1e-05,
      "loss": 1.7124,
      "step": 4010
    },
    {
      "epoch": 0.96,
      "learning_rate": 1e-05,
      "loss": 1.6538,
      "step": 4020
    },
    {
      "epoch": 0.96,
      "learning_rate": 1e-05,
      "loss": 1.6516,
      "step": 4030
    },
    {
      "epoch": 0.96,
      "learning_rate": 1e-05,
      "loss": 1.7222,
      "step": 4040
    },
    {
      "epoch": 0.97,
      "learning_rate": 1e-05,
      "loss": 1.6766,
      "step": 4050
    },
    {
      "epoch": 0.97,
      "learning_rate": 1e-05,
      "loss": 1.6965,
      "step": 4060
    },
    {
      "epoch": 0.97,
      "learning_rate": 1e-05,
      "loss": 1.661,
      "step": 4070
    },
    {
      "epoch": 0.97,
      "learning_rate": 1e-05,
      "loss": 1.6252,
      "step": 4080
    },
    {
      "epoch": 0.98,
      "learning_rate": 1e-05,
      "loss": 1.7716,
      "step": 4090
    },
    {
      "epoch": 0.98,
      "learning_rate": 1e-05,
      "loss": 1.6556,
      "step": 4100
    },
    {
      "epoch": 0.98,
      "eval_accuracy": 0.5467836257309941,
      "eval_loss": 1.5935206413269043,
      "eval_runtime": 56.9293,
      "eval_samples_per_second": 12.015,
      "step": 4100
    },
    {
      "epoch": 0.98,
      "learning_rate": 1e-05,
      "loss": 1.6476,
      "step": 4110
    },
    {
      "epoch": 0.98,
      "learning_rate": 1e-05,
      "loss": 1.6643,
      "step": 4120
    },
    {
      "epoch": 0.99,
      "learning_rate": 1e-05,
      "loss": 1.694,
      "step": 4130
    },
    {
      "epoch": 0.99,
      "learning_rate": 1e-05,
      "loss": 1.5937,
      "step": 4140
    },
    {
      "epoch": 0.99,
      "learning_rate": 1e-05,
      "loss": 1.6661,
      "step": 4150
    },
    {
      "epoch": 0.99,
      "learning_rate": 1e-05,
      "loss": 1.6887,
      "step": 4160
    },
    {
      "epoch": 1.0,
      "learning_rate": 1e-05,
      "loss": 1.6639,
      "step": 4170
    },
    {
      "epoch": 1.0,
      "learning_rate": 1e-05,
      "loss": 1.7488,
      "step": 4180
    },
    {
      "epoch": 1.0,
      "learning_rate": 1e-05,
      "loss": 1.7166,
      "step": 4190
    },
    {
      "epoch": 1.0,
      "learning_rate": 1e-05,
      "loss": 1.6229,
      "step": 4200
    },
    {
      "epoch": 1.0,
      "eval_accuracy": 0.5628654970760234,
      "eval_loss": 1.5908117294311523,
      "eval_runtime": 34.4693,
      "eval_samples_per_second": 19.844,
      "step": 4200
    },
    {
      "epoch": 1.01,
      "learning_rate": 1e-05,
      "loss": 1.6645,
      "step": 4210
    },
    {
      "epoch": 1.01,
      "learning_rate": 1e-05,
      "loss": 1.6295,
      "step": 4220
    },
    {
      "epoch": 1.01,
      "learning_rate": 1e-05,
      "loss": 1.6781,
      "step": 4230
    },
    {
      "epoch": 1.01,
      "learning_rate": 1e-05,
      "loss": 1.6955,
      "step": 4240
    },
    {
      "epoch": 1.02,
      "learning_rate": 1e-05,
      "loss": 1.676,
      "step": 4250
    },
    {
      "epoch": 1.02,
      "learning_rate": 1e-05,
      "loss": 1.6307,
      "step": 4260
    },
    {
      "epoch": 1.02,
      "learning_rate": 1e-05,
      "loss": 1.6611,
      "step": 4270
    },
    {
      "epoch": 1.02,
      "learning_rate": 1e-05,
      "loss": 1.6944,
      "step": 4280
    },
    {
      "epoch": 1.02,
      "learning_rate": 1e-05,
      "loss": 1.6761,
      "step": 4290
    },
    {
      "epoch": 1.03,
      "learning_rate": 1e-05,
      "loss": 1.7008,
      "step": 4300
    },
    {
      "epoch": 1.03,
      "eval_accuracy": 0.5672514619883041,
      "eval_loss": 1.5872185230255127,
      "eval_runtime": 34.5391,
      "eval_samples_per_second": 19.804,
      "step": 4300
    },
    {
      "epoch": 1.03,
      "learning_rate": 1e-05,
      "loss": 1.6164,
      "step": 4310
    },
    {
      "epoch": 1.03,
      "learning_rate": 1e-05,
      "loss": 1.7107,
      "step": 4320
    },
    {
      "epoch": 1.03,
      "learning_rate": 1e-05,
      "loss": 1.6673,
      "step": 4330
    },
    {
      "epoch": 1.04,
      "learning_rate": 1e-05,
      "loss": 1.7692,
      "step": 4340
    },
    {
      "epoch": 1.04,
      "learning_rate": 1e-05,
      "loss": 1.6795,
      "step": 4350
    },
    {
      "epoch": 1.04,
      "learning_rate": 1e-05,
      "loss": 1.6159,
      "step": 4360
    },
    {
      "epoch": 1.04,
      "learning_rate": 1e-05,
      "loss": 1.6332,
      "step": 4370
    },
    {
      "epoch": 1.05,
      "learning_rate": 1e-05,
      "loss": 1.6142,
      "step": 4380
    },
    {
      "epoch": 1.05,
      "learning_rate": 1e-05,
      "loss": 1.6162,
      "step": 4390
    },
    {
      "epoch": 1.05,
      "learning_rate": 1e-05,
      "loss": 1.634,
      "step": 4400
    },
    {
      "epoch": 1.05,
      "eval_accuracy": 0.5701754385964912,
      "eval_loss": 1.5835323333740234,
      "eval_runtime": 35.024,
      "eval_samples_per_second": 19.529,
      "step": 4400
    },
    {
      "epoch": 1.05,
      "learning_rate": 1e-05,
      "loss": 1.7822,
      "step": 4410
    },
    {
      "epoch": 1.06,
      "learning_rate": 1e-05,
      "loss": 1.6762,
      "step": 4420
    },
    {
      "epoch": 1.06,
      "learning_rate": 1e-05,
      "loss": 1.6369,
      "step": 4430
    },
    {
      "epoch": 1.06,
      "learning_rate": 1e-05,
      "loss": 1.6421,
      "step": 4440
    },
    {
      "epoch": 1.06,
      "learning_rate": 1e-05,
      "loss": 1.6287,
      "step": 4450
    },
    {
      "epoch": 1.07,
      "learning_rate": 1e-05,
      "loss": 1.668,
      "step": 4460
    },
    {
      "epoch": 1.07,
      "learning_rate": 1e-05,
      "loss": 1.7114,
      "step": 4470
    },
    {
      "epoch": 1.07,
      "learning_rate": 1e-05,
      "loss": 1.6354,
      "step": 4480
    },
    {
      "epoch": 1.07,
      "learning_rate": 1e-05,
      "loss": 1.6433,
      "step": 4490
    },
    {
      "epoch": 1.07,
      "learning_rate": 1e-05,
      "loss": 1.6726,
      "step": 4500
    },
    {
      "epoch": 1.07,
      "eval_accuracy": 0.5818713450292398,
      "eval_loss": 1.583084225654602,
      "eval_runtime": 34.4975,
      "eval_samples_per_second": 19.828,
      "step": 4500
    },
    {
      "epoch": 1.08,
      "learning_rate": 1e-05,
      "loss": 1.6329,
      "step": 4510
    },
    {
      "epoch": 1.08,
      "learning_rate": 1e-05,
      "loss": 1.6735,
      "step": 4520
    },
    {
      "epoch": 1.08,
      "learning_rate": 1e-05,
      "loss": 1.6542,
      "step": 4530
    },
    {
      "epoch": 1.08,
      "learning_rate": 1e-05,
      "loss": 1.6851,
      "step": 4540
    },
    {
      "epoch": 1.09,
      "learning_rate": 1e-05,
      "loss": 1.6437,
      "step": 4550
    },
    {
      "epoch": 1.09,
      "learning_rate": 1e-05,
      "loss": 1.6574,
      "step": 4560
    },
    {
      "epoch": 1.09,
      "learning_rate": 1e-05,
      "loss": 1.6645,
      "step": 4570
    },
    {
      "epoch": 1.09,
      "learning_rate": 1e-05,
      "loss": 1.6545,
      "step": 4580
    },
    {
      "epoch": 1.1,
      "learning_rate": 1e-05,
      "loss": 1.5898,
      "step": 4590
    },
    {
      "epoch": 1.1,
      "learning_rate": 1e-05,
      "loss": 1.6588,
      "step": 4600
    },
    {
      "epoch": 1.1,
      "eval_accuracy": 0.5804093567251462,
      "eval_loss": 1.5767793655395508,
      "eval_runtime": 55.9702,
      "eval_samples_per_second": 12.221,
      "step": 4600
    },
    {
      "epoch": 1.1,
      "learning_rate": 1e-05,
      "loss": 1.6521,
      "step": 4610
    },
    {
      "epoch": 1.1,
      "learning_rate": 1e-05,
      "loss": 1.6468,
      "step": 4620
    },
    {
      "epoch": 1.11,
      "learning_rate": 1e-05,
      "loss": 1.5635,
      "step": 4630
    },
    {
      "epoch": 1.11,
      "learning_rate": 1e-05,
      "loss": 1.707,
      "step": 4640
    },
    {
      "epoch": 1.11,
      "learning_rate": 1e-05,
      "loss": 1.6135,
      "step": 4650
    },
    {
      "epoch": 1.11,
      "learning_rate": 1e-05,
      "loss": 1.6397,
      "step": 4660
    },
    {
      "epoch": 1.12,
      "learning_rate": 1e-05,
      "loss": 1.6524,
      "step": 4670
    },
    {
      "epoch": 1.12,
      "learning_rate": 1e-05,
      "loss": 1.6533,
      "step": 4680
    },
    {
      "epoch": 1.12,
      "learning_rate": 1e-05,
      "loss": 1.6875,
      "step": 4690
    },
    {
      "epoch": 1.12,
      "learning_rate": 1e-05,
      "loss": 1.6354,
      "step": 4700
    },
    {
      "epoch": 1.12,
      "eval_accuracy": 0.5906432748538012,
      "eval_loss": 1.5773786306381226,
      "eval_runtime": 34.8845,
      "eval_samples_per_second": 19.608,
      "step": 4700
    },
    {
      "epoch": 1.12,
      "learning_rate": 1e-05,
      "loss": 1.6097,
      "step": 4710
    },
    {
      "epoch": 1.13,
      "learning_rate": 1e-05,
      "loss": 1.6508,
      "step": 4720
    },
    {
      "epoch": 1.13,
      "learning_rate": 1e-05,
      "loss": 1.6442,
      "step": 4730
    },
    {
      "epoch": 1.13,
      "learning_rate": 1e-05,
      "loss": 1.5847,
      "step": 4740
    },
    {
      "epoch": 1.13,
      "learning_rate": 1e-05,
      "loss": 1.6333,
      "step": 4750
    },
    {
      "epoch": 1.14,
      "learning_rate": 1e-05,
      "loss": 1.6666,
      "step": 4760
    },
    {
      "epoch": 1.14,
      "learning_rate": 1e-05,
      "loss": 1.6556,
      "step": 4770
    },
    {
      "epoch": 1.14,
      "learning_rate": 1e-05,
      "loss": 1.6214,
      "step": 4780
    },
    {
      "epoch": 1.14,
      "learning_rate": 1e-05,
      "loss": 1.6546,
      "step": 4790
    },
    {
      "epoch": 1.15,
      "learning_rate": 1e-05,
      "loss": 1.6673,
      "step": 4800
    },
    {
      "epoch": 1.15,
      "eval_accuracy": 0.5891812865497076,
      "eval_loss": 1.5723403692245483,
      "eval_runtime": 57.739,
      "eval_samples_per_second": 11.846,
      "step": 4800
    },
    {
      "epoch": 1.15,
      "learning_rate": 1e-05,
      "loss": 1.6037,
      "step": 4810
    },
    {
      "epoch": 1.15,
      "learning_rate": 1e-05,
      "loss": 1.6368,
      "step": 4820
    },
    {
      "epoch": 1.15,
      "learning_rate": 1e-05,
      "loss": 1.6934,
      "step": 4830
    },
    {
      "epoch": 1.16,
      "learning_rate": 1e-05,
      "loss": 1.6427,
      "step": 4840
    },
    {
      "epoch": 1.16,
      "learning_rate": 1e-05,
      "loss": 1.6577,
      "step": 4850
    },
    {
      "epoch": 1.16,
      "learning_rate": 1e-05,
      "loss": 1.6128,
      "step": 4860
    },
    {
      "epoch": 1.16,
      "learning_rate": 1e-05,
      "loss": 1.66,
      "step": 4870
    },
    {
      "epoch": 1.17,
      "learning_rate": 1e-05,
      "loss": 1.6683,
      "step": 4880
    },
    {
      "epoch": 1.17,
      "learning_rate": 1e-05,
      "loss": 1.6239,
      "step": 4890
    },
    {
      "epoch": 1.17,
      "learning_rate": 1e-05,
      "loss": 1.6461,
      "step": 4900
    },
    {
      "epoch": 1.17,
      "eval_accuracy": 0.5935672514619883,
      "eval_loss": 1.5713655948638916,
      "eval_runtime": 58.6019,
      "eval_samples_per_second": 11.672,
      "step": 4900
    },
    {
      "epoch": 1.17,
      "learning_rate": 1e-05,
      "loss": 1.5995,
      "step": 4910
    },
    {
      "epoch": 1.18,
      "learning_rate": 1e-05,
      "loss": 1.6972,
      "step": 4920
    },
    {
      "epoch": 1.18,
      "learning_rate": 1e-05,
      "loss": 1.6515,
      "step": 4930
    },
    {
      "epoch": 1.18,
      "learning_rate": 1e-05,
      "loss": 1.6318,
      "step": 4940
    },
    {
      "epoch": 1.18,
      "learning_rate": 1e-05,
      "loss": 1.6486,
      "step": 4950
    },
    {
      "epoch": 1.18,
      "learning_rate": 1e-05,
      "loss": 1.6462,
      "step": 4960
    },
    {
      "epoch": 1.19,
      "learning_rate": 1e-05,
      "loss": 1.6663,
      "step": 4970
    },
    {
      "epoch": 1.19,
      "learning_rate": 1e-05,
      "loss": 1.6264,
      "step": 4980
    },
    {
      "epoch": 1.19,
      "learning_rate": 1e-05,
      "loss": 1.5958,
      "step": 4990
    },
    {
      "epoch": 1.19,
      "learning_rate": 1e-05,
      "loss": 1.6257,
      "step": 5000
    },
    {
      "epoch": 1.19,
      "eval_accuracy": 0.6052631578947368,
      "eval_loss": 1.5704431533813477,
      "eval_runtime": 57.5606,
      "eval_samples_per_second": 11.883,
      "step": 5000
    },
    {
      "epoch": 1.2,
      "learning_rate": 1e-05,
      "loss": 1.6078,
      "step": 5010
    },
    {
      "epoch": 1.2,
      "learning_rate": 1e-05,
      "loss": 1.6132,
      "step": 5020
    },
    {
      "epoch": 1.2,
      "learning_rate": 1e-05,
      "loss": 1.6339,
      "step": 5030
    },
    {
      "epoch": 1.2,
      "learning_rate": 1e-05,
      "loss": 1.8676,
      "step": 5040
    },
    {
      "epoch": 1.21,
      "learning_rate": 1e-05,
      "loss": 1.5769,
      "step": 5050
    },
    {
      "epoch": 1.21,
      "learning_rate": 1e-05,
      "loss": 1.6117,
      "step": 5060
    },
    {
      "epoch": 1.21,
      "learning_rate": 1e-05,
      "loss": 1.6908,
      "step": 5070
    },
    {
      "epoch": 1.21,
      "learning_rate": 1e-05,
      "loss": 1.6154,
      "step": 5080
    },
    {
      "epoch": 1.22,
      "learning_rate": 1e-05,
      "loss": 1.6461,
      "step": 5090
    },
    {
      "epoch": 1.22,
      "learning_rate": 1e-05,
      "loss": 1.5952,
      "step": 5100
    },
    {
      "epoch": 1.22,
      "eval_accuracy": 0.5994152046783626,
      "eval_loss": 1.5670478343963623,
      "eval_runtime": 35.5048,
      "eval_samples_per_second": 19.265,
      "step": 5100
    },
    {
      "epoch": 1.22,
      "learning_rate": 1e-05,
      "loss": 1.6733,
      "step": 5110
    },
    {
      "epoch": 1.22,
      "learning_rate": 1e-05,
      "loss": 1.638,
      "step": 5120
    },
    {
      "epoch": 1.23,
      "learning_rate": 1e-05,
      "loss": 1.6315,
      "step": 5130
    },
    {
      "epoch": 1.23,
      "learning_rate": 1e-05,
      "loss": 1.5857,
      "step": 5140
    },
    {
      "epoch": 1.23,
      "learning_rate": 1e-05,
      "loss": 1.602,
      "step": 5150
    },
    {
      "epoch": 1.23,
      "learning_rate": 1e-05,
      "loss": 1.6279,
      "step": 5160
    },
    {
      "epoch": 1.23,
      "learning_rate": 1e-05,
      "loss": 1.5984,
      "step": 5170
    },
    {
      "epoch": 1.24,
      "learning_rate": 1e-05,
      "loss": 1.6493,
      "step": 5180
    },
    {
      "epoch": 1.24,
      "learning_rate": 1e-05,
      "loss": 1.6578,
      "step": 5190
    },
    {
      "epoch": 1.24,
      "learning_rate": 1e-05,
      "loss": 1.6344,
      "step": 5200
    },
    {
      "epoch": 1.24,
      "eval_accuracy": 0.6038011695906432,
      "eval_loss": 1.563653588294983,
      "eval_runtime": 35.105,
      "eval_samples_per_second": 19.484,
      "step": 5200
    },
    {
      "epoch": 1.24,
      "learning_rate": 1e-05,
      "loss": 1.6391,
      "step": 5210
    },
    {
      "epoch": 1.25,
      "learning_rate": 1e-05,
      "loss": 1.6231,
      "step": 5220
    },
    {
      "epoch": 1.25,
      "learning_rate": 1e-05,
      "loss": 1.5709,
      "step": 5230
    },
    {
      "epoch": 1.25,
      "learning_rate": 1e-05,
      "loss": 1.6697,
      "step": 5240
    },
    {
      "epoch": 1.25,
      "learning_rate": 1e-05,
      "loss": 1.6112,
      "step": 5250
    },
    {
      "epoch": 1.26,
      "learning_rate": 1e-05,
      "loss": 1.6486,
      "step": 5260
    },
    {
      "epoch": 1.26,
      "learning_rate": 1e-05,
      "loss": 1.6347,
      "step": 5270
    },
    {
      "epoch": 1.26,
      "learning_rate": 1e-05,
      "loss": 1.65,
      "step": 5280
    },
    {
      "epoch": 1.26,
      "learning_rate": 1e-05,
      "loss": 1.6172,
      "step": 5290
    },
    {
      "epoch": 1.27,
      "learning_rate": 1e-05,
      "loss": 1.6056,
      "step": 5300
    },
    {
      "epoch": 1.27,
      "eval_accuracy": 0.6096491228070176,
      "eval_loss": 1.5590335130691528,
      "eval_runtime": 57.9621,
      "eval_samples_per_second": 11.801,
      "step": 5300
    },
    {
      "epoch": 1.27,
      "learning_rate": 1e-05,
      "loss": 1.6593,
      "step": 5310
    },
    {
      "epoch": 1.27,
      "learning_rate": 1e-05,
      "loss": 1.5804,
      "step": 5320
    },
    {
      "epoch": 1.27,
      "learning_rate": 1e-05,
      "loss": 1.6312,
      "step": 5330
    },
    {
      "epoch": 1.28,
      "learning_rate": 1e-05,
      "loss": 1.6532,
      "step": 5340
    },
    {
      "epoch": 1.28,
      "learning_rate": 1e-05,
      "loss": 1.6529,
      "step": 5350
    },
    {
      "epoch": 1.28,
      "learning_rate": 1e-05,
      "loss": 1.5996,
      "step": 5360
    },
    {
      "epoch": 1.28,
      "learning_rate": 1e-05,
      "loss": 1.5919,
      "step": 5370
    },
    {
      "epoch": 1.28,
      "learning_rate": 1e-05,
      "loss": 1.6223,
      "step": 5380
    },
    {
      "epoch": 1.29,
      "learning_rate": 1e-05,
      "loss": 1.6104,
      "step": 5390
    },
    {
      "epoch": 1.29,
      "learning_rate": 1e-05,
      "loss": 1.6231,
      "step": 5400
    },
    {
      "epoch": 1.29,
      "eval_accuracy": 0.6096491228070176,
      "eval_loss": 1.5562858581542969,
      "eval_runtime": 34.7228,
      "eval_samples_per_second": 19.699,
      "step": 5400
    },
    {
      "epoch": 1.29,
      "learning_rate": 1e-05,
      "loss": 1.7054,
      "step": 5410
    },
    {
      "epoch": 1.29,
      "learning_rate": 1e-05,
      "loss": 1.6112,
      "step": 5420
    },
    {
      "epoch": 1.3,
      "learning_rate": 1e-05,
      "loss": 1.6693,
      "step": 5430
    },
    {
      "epoch": 1.3,
      "learning_rate": 1e-05,
      "loss": 1.6364,
      "step": 5440
    },
    {
      "epoch": 1.3,
      "learning_rate": 1e-05,
      "loss": 1.6055,
      "step": 5450
    },
    {
      "epoch": 1.3,
      "learning_rate": 1e-05,
      "loss": 1.621,
      "step": 5460
    },
    {
      "epoch": 1.31,
      "learning_rate": 1e-05,
      "loss": 1.6822,
      "step": 5470
    },
    {
      "epoch": 1.31,
      "learning_rate": 1e-05,
      "loss": 1.6302,
      "step": 5480
    },
    {
      "epoch": 1.31,
      "learning_rate": 1e-05,
      "loss": 1.5743,
      "step": 5490
    },
    {
      "epoch": 1.31,
      "learning_rate": 1e-05,
      "loss": 1.6087,
      "step": 5500
    },
    {
      "epoch": 1.31,
      "eval_accuracy": 0.6125730994152047,
      "eval_loss": 1.555943489074707,
      "eval_runtime": 34.6877,
      "eval_samples_per_second": 19.719,
      "step": 5500
    },
    {
      "epoch": 1.32,
      "learning_rate": 1e-05,
      "loss": 1.6143,
      "step": 5510
    },
    {
      "epoch": 1.32,
      "learning_rate": 1e-05,
      "loss": 1.6162,
      "step": 5520
    },
    {
      "epoch": 1.32,
      "learning_rate": 1e-05,
      "loss": 1.6181,
      "step": 5530
    },
    {
      "epoch": 1.32,
      "learning_rate": 1e-05,
      "loss": 1.6232,
      "step": 5540
    },
    {
      "epoch": 1.33,
      "learning_rate": 1e-05,
      "loss": 1.6062,
      "step": 5550
    },
    {
      "epoch": 1.33,
      "learning_rate": 1e-05,
      "loss": 1.6494,
      "step": 5560
    },
    {
      "epoch": 1.33,
      "learning_rate": 1e-05,
      "loss": 1.5934,
      "step": 5570
    },
    {
      "epoch": 1.33,
      "learning_rate": 1e-05,
      "loss": 1.6011,
      "step": 5580
    },
    {
      "epoch": 1.34,
      "learning_rate": 1e-05,
      "loss": 1.6418,
      "step": 5590
    },
    {
      "epoch": 1.34,
      "learning_rate": 1e-05,
      "loss": 1.6727,
      "step": 5600
    },
    {
      "epoch": 1.34,
      "eval_accuracy": 0.618421052631579,
      "eval_loss": 1.555796504020691,
      "eval_runtime": 34.757,
      "eval_samples_per_second": 19.679,
      "step": 5600
    },
    {
      "epoch": 1.34,
      "learning_rate": 1e-05,
      "loss": 1.6349,
      "step": 5610
    },
    {
      "epoch": 1.34,
      "learning_rate": 1e-05,
      "loss": 1.6455,
      "step": 5620
    },
    {
      "epoch": 1.34,
      "learning_rate": 1e-05,
      "loss": 1.6595,
      "step": 5630
    },
    {
      "epoch": 1.35,
      "learning_rate": 1e-05,
      "loss": 1.6067,
      "step": 5640
    },
    {
      "epoch": 1.35,
      "learning_rate": 1e-05,
      "loss": 1.6035,
      "step": 5650
    },
    {
      "epoch": 1.35,
      "learning_rate": 1e-05,
      "loss": 1.6136,
      "step": 5660
    },
    {
      "epoch": 1.35,
      "learning_rate": 1e-05,
      "loss": 1.6171,
      "step": 5670
    },
    {
      "epoch": 1.36,
      "learning_rate": 1e-05,
      "loss": 1.5877,
      "step": 5680
    },
    {
      "epoch": 1.36,
      "learning_rate": 1e-05,
      "loss": 1.6416,
      "step": 5690
    },
    {
      "epoch": 1.36,
      "learning_rate": 1e-05,
      "loss": 1.6213,
      "step": 5700
    },
    {
      "epoch": 1.36,
      "eval_accuracy": 0.6213450292397661,
      "eval_loss": 1.5521104335784912,
      "eval_runtime": 34.6843,
      "eval_samples_per_second": 19.721,
      "step": 5700
    },
    {
      "epoch": 1.36,
      "learning_rate": 1e-05,
      "loss": 1.5656,
      "step": 5710
    },
    {
      "epoch": 1.37,
      "learning_rate": 1e-05,
      "loss": 1.6008,
      "step": 5720
    },
    {
      "epoch": 1.37,
      "learning_rate": 1e-05,
      "loss": 1.6234,
      "step": 5730
    },
    {
      "epoch": 1.37,
      "learning_rate": 1e-05,
      "loss": 1.5987,
      "step": 5740
    },
    {
      "epoch": 1.37,
      "learning_rate": 1e-05,
      "loss": 1.6028,
      "step": 5750
    },
    {
      "epoch": 1.38,
      "learning_rate": 1e-05,
      "loss": 1.6332,
      "step": 5760
    },
    {
      "epoch": 1.38,
      "learning_rate": 1e-05,
      "loss": 1.6257,
      "step": 5770
    },
    {
      "epoch": 1.38,
      "learning_rate": 1e-05,
      "loss": 1.6143,
      "step": 5780
    },
    {
      "epoch": 1.38,
      "learning_rate": 1e-05,
      "loss": 1.6087,
      "step": 5790
    },
    {
      "epoch": 1.39,
      "learning_rate": 1e-05,
      "loss": 1.6456,
      "step": 5800
    },
    {
      "epoch": 1.39,
      "eval_accuracy": 0.6242690058479532,
      "eval_loss": 1.550268530845642,
      "eval_runtime": 34.6765,
      "eval_samples_per_second": 19.725,
      "step": 5800
    },
    {
      "epoch": 1.39,
      "learning_rate": 1e-05,
      "loss": 1.5754,
      "step": 5810
    },
    {
      "epoch": 1.39,
      "learning_rate": 1e-05,
      "loss": 1.6146,
      "step": 5820
    },
    {
      "epoch": 1.39,
      "learning_rate": 1e-05,
      "loss": 1.5973,
      "step": 5830
    },
    {
      "epoch": 1.39,
      "learning_rate": 1e-05,
      "loss": 1.6048,
      "step": 5840
    },
    {
      "epoch": 1.4,
      "learning_rate": 1e-05,
      "loss": 1.6173,
      "step": 5850
    },
    {
      "epoch": 1.4,
      "learning_rate": 1e-05,
      "loss": 1.5787,
      "step": 5860
    },
    {
      "epoch": 1.4,
      "learning_rate": 1e-05,
      "loss": 1.5697,
      "step": 5870
    },
    {
      "epoch": 1.4,
      "learning_rate": 1e-05,
      "loss": 1.559,
      "step": 5880
    },
    {
      "epoch": 1.41,
      "learning_rate": 1e-05,
      "loss": 1.6387,
      "step": 5890
    },
    {
      "epoch": 1.41,
      "learning_rate": 1e-05,
      "loss": 1.5955,
      "step": 5900
    },
    {
      "epoch": 1.41,
      "eval_accuracy": 0.6286549707602339,
      "eval_loss": 1.5467467308044434,
      "eval_runtime": 34.5958,
      "eval_samples_per_second": 19.771,
      "step": 5900
    },
    {
      "epoch": 1.41,
      "learning_rate": 1e-05,
      "loss": 1.6907,
      "step": 5910
    },
    {
      "epoch": 1.41,
      "learning_rate": 1e-05,
      "loss": 1.6002,
      "step": 5920
    },
    {
      "epoch": 1.42,
      "learning_rate": 1e-05,
      "loss": 1.6729,
      "step": 5930
    },
    {
      "epoch": 1.42,
      "learning_rate": 1e-05,
      "loss": 1.641,
      "step": 5940
    },
    {
      "epoch": 1.42,
      "learning_rate": 1e-05,
      "loss": 1.5827,
      "step": 5950
    },
    {
      "epoch": 1.42,
      "learning_rate": 1e-05,
      "loss": 1.5752,
      "step": 5960
    },
    {
      "epoch": 1.43,
      "learning_rate": 1e-05,
      "loss": 1.6204,
      "step": 5970
    },
    {
      "epoch": 1.43,
      "learning_rate": 1e-05,
      "loss": 1.573,
      "step": 5980
    },
    {
      "epoch": 1.43,
      "learning_rate": 1e-05,
      "loss": 1.5926,
      "step": 5990
    },
    {
      "epoch": 1.43,
      "learning_rate": 1e-05,
      "loss": 1.6065,
      "step": 6000
    },
    {
      "epoch": 1.43,
      "eval_accuracy": 0.631578947368421,
      "eval_loss": 1.5451860427856445,
      "eval_runtime": 34.5798,
      "eval_samples_per_second": 19.78,
      "step": 6000
    },
    {
      "epoch": 1.44,
      "learning_rate": 1e-05,
      "loss": 1.6267,
      "step": 6010
    },
    {
      "epoch": 1.44,
      "learning_rate": 1e-05,
      "loss": 1.6619,
      "step": 6020
    },
    {
      "epoch": 1.44,
      "learning_rate": 1e-05,
      "loss": 1.6224,
      "step": 6030
    },
    {
      "epoch": 1.44,
      "learning_rate": 1e-05,
      "loss": 1.5886,
      "step": 6040
    },
    {
      "epoch": 1.44,
      "learning_rate": 1e-05,
      "loss": 1.6171,
      "step": 6050
    },
    {
      "epoch": 1.45,
      "learning_rate": 1e-05,
      "loss": 1.5648,
      "step": 6060
    },
    {
      "epoch": 1.45,
      "learning_rate": 1e-05,
      "loss": 1.6121,
      "step": 6070
    },
    {
      "epoch": 1.45,
      "learning_rate": 1e-05,
      "loss": 1.5892,
      "step": 6080
    },
    {
      "epoch": 1.45,
      "learning_rate": 1e-05,
      "loss": 1.5754,
      "step": 6090
    },
    {
      "epoch": 1.46,
      "learning_rate": 1e-05,
      "loss": 1.609,
      "step": 6100
    },
    {
      "epoch": 1.46,
      "eval_accuracy": 0.6388888888888888,
      "eval_loss": 1.541480302810669,
      "eval_runtime": 34.4856,
      "eval_samples_per_second": 19.834,
      "step": 6100
    },
    {
      "epoch": 1.46,
      "learning_rate": 1e-05,
      "loss": 1.6276,
      "step": 6110
    },
    {
      "epoch": 1.46,
      "learning_rate": 1e-05,
      "loss": 1.5804,
      "step": 6120
    },
    {
      "epoch": 1.46,
      "learning_rate": 1e-05,
      "loss": 1.5928,
      "step": 6130
    },
    {
      "epoch": 1.47,
      "learning_rate": 1e-05,
      "loss": 1.5657,
      "step": 6140
    },
    {
      "epoch": 1.47,
      "learning_rate": 1e-05,
      "loss": 1.6135,
      "step": 6150
    },
    {
      "epoch": 1.47,
      "learning_rate": 1e-05,
      "loss": 1.5772,
      "step": 6160
    },
    {
      "epoch": 1.47,
      "learning_rate": 1e-05,
      "loss": 1.5753,
      "step": 6170
    },
    {
      "epoch": 1.48,
      "learning_rate": 1e-05,
      "loss": 1.6105,
      "step": 6180
    },
    {
      "epoch": 1.48,
      "learning_rate": 1e-05,
      "loss": 1.6058,
      "step": 6190
    },
    {
      "epoch": 1.48,
      "learning_rate": 1e-05,
      "loss": 1.5994,
      "step": 6200
    },
    {
      "epoch": 1.48,
      "eval_accuracy": 0.6447368421052632,
      "eval_loss": 1.5405205488204956,
      "eval_runtime": 34.191,
      "eval_samples_per_second": 20.005,
      "step": 6200
    },
    {
      "epoch": 1.48,
      "learning_rate": 1e-05,
      "loss": 1.5782,
      "step": 6210
    },
    {
      "epoch": 1.49,
      "learning_rate": 1e-05,
      "loss": 1.613,
      "step": 6220
    },
    {
      "epoch": 1.49,
      "learning_rate": 1e-05,
      "loss": 1.575,
      "step": 6230
    },
    {
      "epoch": 1.49,
      "learning_rate": 1e-05,
      "loss": 1.5687,
      "step": 6240
    },
    {
      "epoch": 1.49,
      "learning_rate": 1e-05,
      "loss": 1.5386,
      "step": 6250
    },
    {
      "epoch": 1.5,
      "learning_rate": 1e-05,
      "loss": 1.6008,
      "step": 6260
    },
    {
      "epoch": 1.5,
      "learning_rate": 1e-05,
      "loss": 1.6237,
      "step": 6270
    },
    {
      "epoch": 1.5,
      "learning_rate": 1e-05,
      "loss": 1.6042,
      "step": 6280
    },
    {
      "epoch": 1.5,
      "learning_rate": 1e-05,
      "loss": 1.5737,
      "step": 6290
    },
    {
      "epoch": 1.5,
      "learning_rate": 1e-05,
      "loss": 1.5631,
      "step": 6300
    },
    {
      "epoch": 1.5,
      "eval_accuracy": 0.6447368421052632,
      "eval_loss": 1.537993311882019,
      "eval_runtime": 34.7474,
      "eval_samples_per_second": 19.685,
      "step": 6300
    },
    {
      "epoch": 1.51,
      "learning_rate": 1e-05,
      "loss": 1.5926,
      "step": 6310
    },
    {
      "epoch": 1.51,
      "learning_rate": 1e-05,
      "loss": 1.5789,
      "step": 6320
    },
    {
      "epoch": 1.51,
      "learning_rate": 1e-05,
      "loss": 1.605,
      "step": 6330
    },
    {
      "epoch": 1.51,
      "learning_rate": 1e-05,
      "loss": 1.5818,
      "step": 6340
    },
    {
      "epoch": 1.52,
      "learning_rate": 1e-05,
      "loss": 1.63,
      "step": 6350
    },
    {
      "epoch": 1.52,
      "learning_rate": 1e-05,
      "loss": 1.589,
      "step": 6360
    },
    {
      "epoch": 1.52,
      "learning_rate": 1e-05,
      "loss": 1.5864,
      "step": 6370
    },
    {
      "epoch": 1.52,
      "learning_rate": 1e-05,
      "loss": 1.6001,
      "step": 6380
    },
    {
      "epoch": 1.53,
      "learning_rate": 1e-05,
      "loss": 1.5736,
      "step": 6390
    },
    {
      "epoch": 1.53,
      "learning_rate": 1e-05,
      "loss": 1.5911,
      "step": 6400
    },
    {
      "epoch": 1.53,
      "eval_accuracy": 0.6461988304093568,
      "eval_loss": 1.5372425317764282,
      "eval_runtime": 34.6006,
      "eval_samples_per_second": 19.768,
      "step": 6400
    },
    {
      "epoch": 1.53,
      "learning_rate": 1e-05,
      "loss": 1.5966,
      "step": 6410
    },
    {
      "epoch": 1.53,
      "learning_rate": 1e-05,
      "loss": 1.6098,
      "step": 6420
    },
    {
      "epoch": 1.54,
      "learning_rate": 1e-05,
      "loss": 1.6106,
      "step": 6430
    },
    {
      "epoch": 1.54,
      "learning_rate": 1e-05,
      "loss": 1.5854,
      "step": 6440
    },
    {
      "epoch": 1.54,
      "learning_rate": 1e-05,
      "loss": 1.5437,
      "step": 6450
    },
    {
      "epoch": 1.54,
      "learning_rate": 1e-05,
      "loss": 1.5917,
      "step": 6460
    },
    {
      "epoch": 1.55,
      "learning_rate": 1e-05,
      "loss": 1.6419,
      "step": 6470
    },
    {
      "epoch": 1.55,
      "learning_rate": 1e-05,
      "loss": 1.6031,
      "step": 6480
    },
    {
      "epoch": 1.55,
      "learning_rate": 1e-05,
      "loss": 1.5589,
      "step": 6490
    },
    {
      "epoch": 1.55,
      "learning_rate": 1e-05,
      "loss": 1.6031,
      "step": 6500
    },
    {
      "epoch": 1.55,
      "eval_accuracy": 0.6461988304093568,
      "eval_loss": 1.5356801748275757,
      "eval_runtime": 34.6813,
      "eval_samples_per_second": 19.722,
      "step": 6500
    },
    {
      "epoch": 1.55,
      "learning_rate": 1e-05,
      "loss": 1.5535,
      "step": 6510
    },
    {
      "epoch": 1.56,
      "learning_rate": 1e-05,
      "loss": 1.5876,
      "step": 6520
    },
    {
      "epoch": 1.56,
      "learning_rate": 1e-05,
      "loss": 1.5522,
      "step": 6530
    },
    {
      "epoch": 1.56,
      "learning_rate": 1e-05,
      "loss": 1.5686,
      "step": 6540
    },
    {
      "epoch": 1.56,
      "learning_rate": 1e-05,
      "loss": 1.6025,
      "step": 6550
    },
    {
      "epoch": 1.57,
      "learning_rate": 1e-05,
      "loss": 1.6016,
      "step": 6560
    },
    {
      "epoch": 1.57,
      "learning_rate": 1e-05,
      "loss": 1.5972,
      "step": 6570
    },
    {
      "epoch": 1.57,
      "learning_rate": 1e-05,
      "loss": 1.5791,
      "step": 6580
    },
    {
      "epoch": 1.57,
      "learning_rate": 1e-05,
      "loss": 1.6114,
      "step": 6590
    },
    {
      "epoch": 1.58,
      "learning_rate": 1e-05,
      "loss": 1.626,
      "step": 6600
    },
    {
      "epoch": 1.58,
      "eval_accuracy": 0.6505847953216374,
      "eval_loss": 1.535131573677063,
      "eval_runtime": 34.6112,
      "eval_samples_per_second": 19.762,
      "step": 6600
    },
    {
      "epoch": 1.58,
      "learning_rate": 1e-05,
      "loss": 1.5833,
      "step": 6610
    },
    {
      "epoch": 1.58,
      "learning_rate": 1e-05,
      "loss": 1.5618,
      "step": 6620
    },
    {
      "epoch": 1.58,
      "learning_rate": 1e-05,
      "loss": 1.5826,
      "step": 6630
    },
    {
      "epoch": 1.59,
      "learning_rate": 1e-05,
      "loss": 1.559,
      "step": 6640
    },
    {
      "epoch": 1.59,
      "learning_rate": 1e-05,
      "loss": 1.556,
      "step": 6650
    },
    {
      "epoch": 1.59,
      "learning_rate": 1e-05,
      "loss": 1.5854,
      "step": 6660
    },
    {
      "epoch": 1.59,
      "learning_rate": 1e-05,
      "loss": 1.5896,
      "step": 6670
    },
    {
      "epoch": 1.6,
      "learning_rate": 1e-05,
      "loss": 1.5953,
      "step": 6680
    },
    {
      "epoch": 1.6,
      "learning_rate": 1e-05,
      "loss": 1.5987,
      "step": 6690
    },
    {
      "epoch": 1.6,
      "learning_rate": 1e-05,
      "loss": 1.5737,
      "step": 6700
    },
    {
      "epoch": 1.6,
      "eval_accuracy": 0.6549707602339181,
      "eval_loss": 1.5337629318237305,
      "eval_runtime": 35.6589,
      "eval_samples_per_second": 19.182,
      "step": 6700
    },
    {
      "epoch": 1.6,
      "learning_rate": 1e-05,
      "loss": 1.5827,
      "step": 6710
    },
    {
      "epoch": 1.6,
      "learning_rate": 1e-05,
      "loss": 1.5615,
      "step": 6720
    },
    {
      "epoch": 1.61,
      "learning_rate": 1e-05,
      "loss": 1.5761,
      "step": 6730
    },
    {
      "epoch": 1.61,
      "learning_rate": 1e-05,
      "loss": 1.5477,
      "step": 6740
    },
    {
      "epoch": 1.61,
      "learning_rate": 1e-05,
      "loss": 1.597,
      "step": 6750
    },
    {
      "epoch": 1.61,
      "learning_rate": 1e-05,
      "loss": 1.6125,
      "step": 6760
    },
    {
      "epoch": 1.62,
      "learning_rate": 1e-05,
      "loss": 1.5987,
      "step": 6770
    },
    {
      "epoch": 1.62,
      "learning_rate": 1e-05,
      "loss": 1.5885,
      "step": 6780
    },
    {
      "epoch": 1.62,
      "learning_rate": 1e-05,
      "loss": 1.5686,
      "step": 6790
    },
    {
      "epoch": 1.62,
      "learning_rate": 1e-05,
      "loss": 1.5914,
      "step": 6800
    },
    {
      "epoch": 1.62,
      "eval_accuracy": 0.6695906432748538,
      "eval_loss": 1.53092622756958,
      "eval_runtime": 34.0747,
      "eval_samples_per_second": 20.074,
      "step": 6800
    },
    {
      "epoch": 1.63,
      "learning_rate": 1e-05,
      "loss": 1.6045,
      "step": 6810
    },
    {
      "epoch": 1.63,
      "learning_rate": 1e-05,
      "loss": 1.5339,
      "step": 6820
    },
    {
      "epoch": 1.63,
      "learning_rate": 1e-05,
      "loss": 1.59,
      "step": 6830
    },
    {
      "epoch": 1.63,
      "learning_rate": 1e-05,
      "loss": 1.5689,
      "step": 6840
    },
    {
      "epoch": 1.64,
      "learning_rate": 1e-05,
      "loss": 1.6256,
      "step": 6850
    },
    {
      "epoch": 1.64,
      "learning_rate": 1e-05,
      "loss": 1.5947,
      "step": 6860
    },
    {
      "epoch": 1.64,
      "learning_rate": 1e-05,
      "loss": 1.5617,
      "step": 6870
    },
    {
      "epoch": 1.64,
      "learning_rate": 1e-05,
      "loss": 1.583,
      "step": 6880
    },
    {
      "epoch": 1.65,
      "learning_rate": 1e-05,
      "loss": 1.5754,
      "step": 6890
    },
    {
      "epoch": 1.65,
      "learning_rate": 1e-05,
      "loss": 1.5706,
      "step": 6900
    },
    {
      "epoch": 1.65,
      "eval_accuracy": 0.6666666666666666,
      "eval_loss": 1.5304936170578003,
      "eval_runtime": 33.306,
      "eval_samples_per_second": 20.537,
      "step": 6900
    },
    {
      "epoch": 1.65,
      "learning_rate": 1e-05,
      "loss": 1.5838,
      "step": 6910
    },
    {
      "epoch": 1.65,
      "learning_rate": 1e-05,
      "loss": 1.537,
      "step": 6920
    },
    {
      "epoch": 1.66,
      "learning_rate": 1e-05,
      "loss": 1.5742,
      "step": 6930
    },
    {
      "epoch": 1.66,
      "learning_rate": 1e-05,
      "loss": 1.6072,
      "step": 6940
    },
    {
      "epoch": 1.66,
      "learning_rate": 1e-05,
      "loss": 1.5707,
      "step": 6950
    },
    {
      "epoch": 1.66,
      "learning_rate": 1e-05,
      "loss": 1.5779,
      "step": 6960
    },
    {
      "epoch": 1.66,
      "learning_rate": 1e-05,
      "loss": 1.5708,
      "step": 6970
    },
    {
      "epoch": 1.67,
      "learning_rate": 1e-05,
      "loss": 1.5523,
      "step": 6980
    },
    {
      "epoch": 1.67,
      "learning_rate": 1e-05,
      "loss": 1.6126,
      "step": 6990
    },
    {
      "epoch": 1.67,
      "learning_rate": 1e-05,
      "loss": 1.6147,
      "step": 7000
    },
    {
      "epoch": 1.67,
      "eval_accuracy": 0.672514619883041,
      "eval_loss": 1.5283931493759155,
      "eval_runtime": 34.1978,
      "eval_samples_per_second": 20.001,
      "step": 7000
    },
    {
      "epoch": 1.67,
      "learning_rate": 1e-05,
      "loss": 1.5797,
      "step": 7010
    },
    {
      "epoch": 1.68,
      "learning_rate": 1e-05,
      "loss": 1.5694,
      "step": 7020
    },
    {
      "epoch": 1.68,
      "learning_rate": 1e-05,
      "loss": 1.5662,
      "step": 7030
    },
    {
      "epoch": 1.68,
      "learning_rate": 1e-05,
      "loss": 1.6063,
      "step": 7040
    },
    {
      "epoch": 1.68,
      "learning_rate": 1e-05,
      "loss": 1.5582,
      "step": 7050
    },
    {
      "epoch": 1.69,
      "learning_rate": 1e-05,
      "loss": 1.56,
      "step": 7060
    },
    {
      "epoch": 1.69,
      "learning_rate": 1e-05,
      "loss": 1.5823,
      "step": 7070
    },
    {
      "epoch": 1.69,
      "learning_rate": 1e-05,
      "loss": 1.5773,
      "step": 7080
    },
    {
      "epoch": 1.69,
      "learning_rate": 1e-05,
      "loss": 1.5467,
      "step": 7090
    },
    {
      "epoch": 1.7,
      "learning_rate": 1e-05,
      "loss": 1.5704,
      "step": 7100
    },
    {
      "epoch": 1.7,
      "eval_accuracy": 0.6827485380116959,
      "eval_loss": 1.524511694908142,
      "eval_runtime": 34.6458,
      "eval_samples_per_second": 19.743,
      "step": 7100
    },
    {
      "epoch": 1.7,
      "learning_rate": 1e-05,
      "loss": 1.5524,
      "step": 7110
    },
    {
      "epoch": 1.7,
      "learning_rate": 1e-05,
      "loss": 1.5642,
      "step": 7120
    },
    {
      "epoch": 1.7,
      "learning_rate": 1e-05,
      "loss": 1.5777,
      "step": 7130
    },
    {
      "epoch": 1.71,
      "learning_rate": 1e-05,
      "loss": 1.5738,
      "step": 7140
    },
    {
      "epoch": 1.71,
      "learning_rate": 1e-05,
      "loss": 1.569,
      "step": 7150
    },
    {
      "epoch": 1.71,
      "learning_rate": 1e-05,
      "loss": 1.5929,
      "step": 7160
    },
    {
      "epoch": 1.71,
      "learning_rate": 1e-05,
      "loss": 1.594,
      "step": 7170
    },
    {
      "epoch": 1.71,
      "learning_rate": 1e-05,
      "loss": 1.5767,
      "step": 7180
    },
    {
      "epoch": 1.72,
      "learning_rate": 1e-05,
      "loss": 1.5693,
      "step": 7190
    },
    {
      "epoch": 1.72,
      "learning_rate": 1e-05,
      "loss": 1.6038,
      "step": 7200
    },
    {
      "epoch": 1.72,
      "eval_accuracy": 0.6812865497076024,
      "eval_loss": 1.5216485261917114,
      "eval_runtime": 34.528,
      "eval_samples_per_second": 19.81,
      "step": 7200
    },
    {
      "epoch": 1.72,
      "learning_rate": 1e-05,
      "loss": 1.589,
      "step": 7210
    },
    {
      "epoch": 1.72,
      "learning_rate": 1e-05,
      "loss": 1.5774,
      "step": 7220
    },
    {
      "epoch": 1.73,
      "learning_rate": 1e-05,
      "loss": 1.55,
      "step": 7230
    },
    {
      "epoch": 1.73,
      "learning_rate": 1e-05,
      "loss": 1.5583,
      "step": 7240
    },
    {
      "epoch": 1.73,
      "learning_rate": 1e-05,
      "loss": 1.6211,
      "step": 7250
    },
    {
      "epoch": 1.73,
      "learning_rate": 1e-05,
      "loss": 1.5945,
      "step": 7260
    },
    {
      "epoch": 1.74,
      "learning_rate": 1e-05,
      "loss": 1.5829,
      "step": 7270
    },
    {
      "epoch": 1.74,
      "learning_rate": 1e-05,
      "loss": 1.6268,
      "step": 7280
    },
    {
      "epoch": 1.74,
      "learning_rate": 1e-05,
      "loss": 1.5486,
      "step": 7290
    },
    {
      "epoch": 1.74,
      "learning_rate": 1e-05,
      "loss": 1.5905,
      "step": 7300
    },
    {
      "epoch": 1.74,
      "eval_accuracy": 0.6871345029239766,
      "eval_loss": 1.5239547491073608,
      "eval_runtime": 34.2224,
      "eval_samples_per_second": 19.987,
      "step": 7300
    },
    {
      "epoch": 1.75,
      "learning_rate": 1e-05,
      "loss": 1.589,
      "step": 7310
    },
    {
      "epoch": 1.75,
      "learning_rate": 1e-05,
      "loss": 1.6047,
      "step": 7320
    },
    {
      "epoch": 1.75,
      "learning_rate": 1e-05,
      "loss": 1.6366,
      "step": 7330
    },
    {
      "epoch": 1.75,
      "learning_rate": 1e-05,
      "loss": 1.559,
      "step": 7340
    },
    {
      "epoch": 1.76,
      "learning_rate": 1e-05,
      "loss": 1.5759,
      "step": 7350
    },
    {
      "epoch": 1.76,
      "learning_rate": 1e-05,
      "loss": 1.565,
      "step": 7360
    },
    {
      "epoch": 1.76,
      "learning_rate": 1e-05,
      "loss": 1.544,
      "step": 7370
    },
    {
      "epoch": 1.76,
      "learning_rate": 1e-05,
      "loss": 1.5692,
      "step": 7380
    },
    {
      "epoch": 1.76,
      "learning_rate": 1e-05,
      "loss": 1.5461,
      "step": 7390
    },
    {
      "epoch": 1.77,
      "learning_rate": 1e-05,
      "loss": 1.5586,
      "step": 7400
    },
    {
      "epoch": 1.77,
      "eval_accuracy": 0.7046783625730995,
      "eval_loss": 1.5223685503005981,
      "eval_runtime": 34.1302,
      "eval_samples_per_second": 20.041,
      "step": 7400
    },
    {
      "epoch": 1.77,
      "learning_rate": 1e-05,
      "loss": 1.5571,
      "step": 7410
    },
    {
      "epoch": 1.77,
      "learning_rate": 1e-05,
      "loss": 1.5674,
      "step": 7420
    },
    {
      "epoch": 1.77,
      "learning_rate": 1e-05,
      "loss": 1.528,
      "step": 7430
    },
    {
      "epoch": 1.78,
      "learning_rate": 1e-05,
      "loss": 1.5337,
      "step": 7440
    },
    {
      "epoch": 1.78,
      "learning_rate": 1e-05,
      "loss": 1.5674,
      "step": 7450
    },
    {
      "epoch": 1.78,
      "learning_rate": 1e-05,
      "loss": 1.5823,
      "step": 7460
    },
    {
      "epoch": 1.78,
      "learning_rate": 1e-05,
      "loss": 1.5431,
      "step": 7470
    },
    {
      "epoch": 1.79,
      "learning_rate": 1e-05,
      "loss": 1.6037,
      "step": 7480
    },
    {
      "epoch": 1.79,
      "learning_rate": 1e-05,
      "loss": 1.5576,
      "step": 7490
    },
    {
      "epoch": 1.79,
      "learning_rate": 1e-05,
      "loss": 1.5841,
      "step": 7500
    },
    {
      "epoch": 1.79,
      "eval_accuracy": 0.7149122807017544,
      "eval_loss": 1.519676923751831,
      "eval_runtime": 34.5922,
      "eval_samples_per_second": 19.773,
      "step": 7500
    },
    {
      "epoch": 1.79,
      "learning_rate": 1e-05,
      "loss": 1.5837,
      "step": 7510
    },
    {
      "epoch": 1.8,
      "learning_rate": 1e-05,
      "loss": 1.588,
      "step": 7520
    },
    {
      "epoch": 1.8,
      "learning_rate": 1e-05,
      "loss": 1.5921,
      "step": 7530
    },
    {
      "epoch": 1.8,
      "learning_rate": 1e-05,
      "loss": 1.5599,
      "step": 7540
    },
    {
      "epoch": 1.8,
      "learning_rate": 1e-05,
      "loss": 1.5122,
      "step": 7550
    },
    {
      "epoch": 1.81,
      "learning_rate": 1e-05,
      "loss": 1.5498,
      "step": 7560
    },
    {
      "epoch": 1.81,
      "learning_rate": 1e-05,
      "loss": 1.5614,
      "step": 7570
    },
    {
      "epoch": 1.81,
      "learning_rate": 1e-05,
      "loss": 1.561,
      "step": 7580
    },
    {
      "epoch": 1.81,
      "learning_rate": 1e-05,
      "loss": 1.5762,
      "step": 7590
    },
    {
      "epoch": 1.82,
      "learning_rate": 1e-05,
      "loss": 1.5752,
      "step": 7600
    },
    {
      "epoch": 1.82,
      "eval_accuracy": 0.7090643274853801,
      "eval_loss": 1.5196900367736816,
      "eval_runtime": 34.8624,
      "eval_samples_per_second": 19.62,
      "step": 7600
    },
    {
      "epoch": 1.82,
      "learning_rate": 1e-05,
      "loss": 1.5574,
      "step": 7610
    },
    {
      "epoch": 1.82,
      "learning_rate": 1e-05,
      "loss": 1.5624,
      "step": 7620
    },
    {
      "epoch": 1.82,
      "learning_rate": 1e-05,
      "loss": 1.5876,
      "step": 7630
    },
    {
      "epoch": 1.82,
      "learning_rate": 1e-05,
      "loss": 1.5428,
      "step": 7640
    },
    {
      "epoch": 1.83,
      "learning_rate": 1e-05,
      "loss": 1.5791,
      "step": 7650
    },
    {
      "epoch": 1.83,
      "learning_rate": 1e-05,
      "loss": 1.6025,
      "step": 7660
    },
    {
      "epoch": 1.83,
      "learning_rate": 1e-05,
      "loss": 1.5659,
      "step": 7670
    },
    {
      "epoch": 1.83,
      "learning_rate": 1e-05,
      "loss": 1.5231,
      "step": 7680
    },
    {
      "epoch": 1.84,
      "learning_rate": 1e-05,
      "loss": 1.5401,
      "step": 7690
    },
    {
      "epoch": 1.84,
      "learning_rate": 1e-05,
      "loss": 1.5527,
      "step": 7700
    },
    {
      "epoch": 1.84,
      "eval_accuracy": 0.7178362573099415,
      "eval_loss": 1.51457941532135,
      "eval_runtime": 36.0842,
      "eval_samples_per_second": 18.956,
      "step": 7700
    },
    {
      "epoch": 1.84,
      "learning_rate": 1e-05,
      "loss": 1.5558,
      "step": 7710
    },
    {
      "epoch": 1.84,
      "learning_rate": 1e-05,
      "loss": 1.5305,
      "step": 7720
    },
    {
      "epoch": 1.85,
      "learning_rate": 1e-05,
      "loss": 1.5454,
      "step": 7730
    },
    {
      "epoch": 1.85,
      "learning_rate": 1e-05,
      "loss": 1.5514,
      "step": 7740
    },
    {
      "epoch": 1.85,
      "learning_rate": 1e-05,
      "loss": 1.5928,
      "step": 7750
    },
    {
      "epoch": 1.85,
      "learning_rate": 1e-05,
      "loss": 1.5553,
      "step": 7760
    },
    {
      "epoch": 1.86,
      "learning_rate": 1e-05,
      "loss": 1.5617,
      "step": 7770
    },
    {
      "epoch": 1.86,
      "learning_rate": 1e-05,
      "loss": 1.5419,
      "step": 7780
    },
    {
      "epoch": 1.86,
      "learning_rate": 1e-05,
      "loss": 1.5465,
      "step": 7790
    },
    {
      "epoch": 1.86,
      "learning_rate": 1e-05,
      "loss": 1.5477,
      "step": 7800
    },
    {
      "epoch": 1.86,
      "eval_accuracy": 0.7251461988304093,
      "eval_loss": 1.5144470930099487,
      "eval_runtime": 36.0302,
      "eval_samples_per_second": 18.984,
      "step": 7800
    },
    {
      "epoch": 1.87,
      "learning_rate": 1e-05,
      "loss": 1.5714,
      "step": 7810
    },
    {
      "epoch": 1.87,
      "learning_rate": 1e-05,
      "loss": 1.586,
      "step": 7820
    },
    {
      "epoch": 1.87,
      "learning_rate": 1e-05,
      "loss": 1.5648,
      "step": 7830
    },
    {
      "epoch": 1.87,
      "learning_rate": 1e-05,
      "loss": 1.57,
      "step": 7840
    },
    {
      "epoch": 1.87,
      "learning_rate": 1e-05,
      "loss": 1.6407,
      "step": 7850
    },
    {
      "epoch": 1.88,
      "learning_rate": 1e-05,
      "loss": 1.5492,
      "step": 7860
    },
    {
      "epoch": 1.88,
      "learning_rate": 1e-05,
      "loss": 1.5798,
      "step": 7870
    },
    {
      "epoch": 1.88,
      "learning_rate": 1e-05,
      "loss": 1.5854,
      "step": 7880
    },
    {
      "epoch": 1.88,
      "learning_rate": 1e-05,
      "loss": 1.5548,
      "step": 7890
    },
    {
      "epoch": 1.89,
      "learning_rate": 1e-05,
      "loss": 1.5643,
      "step": 7900
    },
    {
      "epoch": 1.89,
      "eval_accuracy": 0.7339181286549707,
      "eval_loss": 1.5125890970230103,
      "eval_runtime": 34.8503,
      "eval_samples_per_second": 19.627,
      "step": 7900
    },
    {
      "epoch": 1.89,
      "learning_rate": 1e-05,
      "loss": 1.568,
      "step": 7910
    },
    {
      "epoch": 1.89,
      "learning_rate": 1e-05,
      "loss": 1.5564,
      "step": 7920
    },
    {
      "epoch": 1.89,
      "learning_rate": 1e-05,
      "loss": 1.5316,
      "step": 7930
    },
    {
      "epoch": 1.9,
      "learning_rate": 1e-05,
      "loss": 1.599,
      "step": 7940
    },
    {
      "epoch": 1.9,
      "learning_rate": 1e-05,
      "loss": 1.5607,
      "step": 7950
    },
    {
      "epoch": 1.9,
      "learning_rate": 1e-05,
      "loss": 1.5495,
      "step": 7960
    },
    {
      "epoch": 1.9,
      "learning_rate": 1e-05,
      "loss": 1.5594,
      "step": 7970
    },
    {
      "epoch": 1.91,
      "learning_rate": 1e-05,
      "loss": 1.5358,
      "step": 7980
    },
    {
      "epoch": 1.91,
      "learning_rate": 1e-05,
      "loss": 1.522,
      "step": 7990
    },
    {
      "epoch": 1.91,
      "learning_rate": 1e-05,
      "loss": 1.5474,
      "step": 8000
    },
    {
      "epoch": 1.91,
      "eval_accuracy": 0.7397660818713451,
      "eval_loss": 1.5123405456542969,
      "eval_runtime": 34.7172,
      "eval_samples_per_second": 19.702,
      "step": 8000
    },
    {
      "epoch": 1.91,
      "learning_rate": 1e-05,
      "loss": 1.5772,
      "step": 8010
    },
    {
      "epoch": 1.92,
      "learning_rate": 1e-05,
      "loss": 1.5911,
      "step": 8020
    },
    {
      "epoch": 1.92,
      "learning_rate": 1e-05,
      "loss": 1.5357,
      "step": 8030
    },
    {
      "epoch": 1.92,
      "learning_rate": 1e-05,
      "loss": 1.559,
      "step": 8040
    },
    {
      "epoch": 1.92,
      "learning_rate": 1e-05,
      "loss": 1.5917,
      "step": 8050
    },
    {
      "epoch": 1.93,
      "learning_rate": 1e-05,
      "loss": 1.5703,
      "step": 8060
    },
    {
      "epoch": 1.93,
      "learning_rate": 1e-05,
      "loss": 1.5337,
      "step": 8070
    },
    {
      "epoch": 1.93,
      "learning_rate": 1e-05,
      "loss": 1.5659,
      "step": 8080
    },
    {
      "epoch": 1.93,
      "learning_rate": 1e-05,
      "loss": 1.5446,
      "step": 8090
    },
    {
      "epoch": 1.93,
      "learning_rate": 1e-05,
      "loss": 1.5747,
      "step": 8100
    },
    {
      "epoch": 1.93,
      "eval_accuracy": 0.7412280701754386,
      "eval_loss": 1.5112054347991943,
      "eval_runtime": 34.6273,
      "eval_samples_per_second": 19.753,
      "step": 8100
    },
    {
      "epoch": 1.94,
      "learning_rate": 1e-05,
      "loss": 1.543,
      "step": 8110
    },
    {
      "epoch": 1.94,
      "learning_rate": 1e-05,
      "loss": 1.5603,
      "step": 8120
    },
    {
      "epoch": 1.94,
      "learning_rate": 1e-05,
      "loss": 1.5475,
      "step": 8130
    },
    {
      "epoch": 1.94,
      "learning_rate": 1e-05,
      "loss": 1.5435,
      "step": 8140
    },
    {
      "epoch": 1.95,
      "learning_rate": 1e-05,
      "loss": 1.5448,
      "step": 8150
    },
    {
      "epoch": 1.95,
      "learning_rate": 1e-05,
      "loss": 1.5657,
      "step": 8160
    },
    {
      "epoch": 1.95,
      "learning_rate": 1e-05,
      "loss": 1.5495,
      "step": 8170
    },
    {
      "epoch": 1.95,
      "learning_rate": 1e-05,
      "loss": 1.5351,
      "step": 8180
    },
    {
      "epoch": 1.96,
      "learning_rate": 1e-05,
      "loss": 1.5682,
      "step": 8190
    },
    {
      "epoch": 1.96,
      "learning_rate": 1e-05,
      "loss": 1.5665,
      "step": 8200
    },
    {
      "epoch": 1.96,
      "eval_accuracy": 0.7543859649122807,
      "eval_loss": 1.5080853700637817,
      "eval_runtime": 34.6489,
      "eval_samples_per_second": 19.741,
      "step": 8200
    },
    {
      "epoch": 1.96,
      "learning_rate": 1e-05,
      "loss": 1.5694,
      "step": 8210
    },
    {
      "epoch": 1.96,
      "learning_rate": 1e-05,
      "loss": 1.5196,
      "step": 8220
    },
    {
      "epoch": 1.97,
      "learning_rate": 1e-05,
      "loss": 1.5841,
      "step": 8230
    },
    {
      "epoch": 1.97,
      "learning_rate": 1e-05,
      "loss": 1.5466,
      "step": 8240
    },
    {
      "epoch": 1.97,
      "learning_rate": 1e-05,
      "loss": 1.548,
      "step": 8250
    },
    {
      "epoch": 1.97,
      "learning_rate": 1e-05,
      "loss": 1.5523,
      "step": 8260
    },
    {
      "epoch": 1.98,
      "learning_rate": 1e-05,
      "loss": 1.5517,
      "step": 8270
    },
    {
      "epoch": 1.98,
      "learning_rate": 1e-05,
      "loss": 1.5999,
      "step": 8280
    },
    {
      "epoch": 1.98,
      "learning_rate": 1e-05,
      "loss": 1.5451,
      "step": 8290
    },
    {
      "epoch": 1.98,
      "learning_rate": 1e-05,
      "loss": 1.5899,
      "step": 8300
    },
    {
      "epoch": 1.98,
      "eval_accuracy": 0.7602339181286549,
      "eval_loss": 1.5067552328109741,
      "eval_runtime": 34.3477,
      "eval_samples_per_second": 19.914,
      "step": 8300
    },
    {
      "epoch": 1.98,
      "learning_rate": 1e-05,
      "loss": 1.5498,
      "step": 8310
    },
    {
      "epoch": 1.99,
      "learning_rate": 1e-05,
      "loss": 1.5688,
      "step": 8320
    },
    {
      "epoch": 1.99,
      "learning_rate": 1e-05,
      "loss": 1.557,
      "step": 8330
    },
    {
      "epoch": 1.99,
      "learning_rate": 1e-05,
      "loss": 1.5724,
      "step": 8340
    },
    {
      "epoch": 1.99,
      "learning_rate": 1e-05,
      "loss": 1.579,
      "step": 8350
    },
    {
      "epoch": 2.0,
      "learning_rate": 1e-05,
      "loss": 1.5611,
      "step": 8360
    },
    {
      "epoch": 2.0,
      "learning_rate": 1e-05,
      "loss": 1.5488,
      "step": 8370
    },
    {
      "epoch": 2.0,
      "learning_rate": 1e-05,
      "loss": 1.5548,
      "step": 8380
    },
    {
      "epoch": 2.0,
      "learning_rate": 1e-05,
      "loss": 1.5695,
      "step": 8390
    },
    {
      "epoch": 2.01,
      "learning_rate": 1e-05,
      "loss": 1.5603,
      "step": 8400
    },
    {
      "epoch": 2.01,
      "eval_accuracy": 0.7631578947368421,
      "eval_loss": 1.5068939924240112,
      "eval_runtime": 34.4662,
      "eval_samples_per_second": 19.846,
      "step": 8400
    },
    {
      "epoch": 2.01,
      "learning_rate": 1e-05,
      "loss": 1.5094,
      "step": 8410
    },
    {
      "epoch": 2.01,
      "learning_rate": 1e-05,
      "loss": 1.5109,
      "step": 8420
    },
    {
      "epoch": 2.01,
      "learning_rate": 1e-05,
      "loss": 1.5386,
      "step": 8430
    },
    {
      "epoch": 2.02,
      "learning_rate": 1e-05,
      "loss": 1.5469,
      "step": 8440
    },
    {
      "epoch": 2.02,
      "learning_rate": 1e-05,
      "loss": 1.514,
      "step": 8450
    },
    {
      "epoch": 2.02,
      "learning_rate": 1e-05,
      "loss": 1.544,
      "step": 8460
    },
    {
      "epoch": 2.02,
      "learning_rate": 1e-05,
      "loss": 1.5502,
      "step": 8470
    },
    {
      "epoch": 2.03,
      "learning_rate": 1e-05,
      "loss": 1.5255,
      "step": 8480
    },
    {
      "epoch": 2.03,
      "learning_rate": 1e-05,
      "loss": 1.5654,
      "step": 8490
    },
    {
      "epoch": 2.03,
      "learning_rate": 1e-05,
      "loss": 1.5427,
      "step": 8500
    },
    {
      "epoch": 2.03,
      "eval_accuracy": 0.7792397660818714,
      "eval_loss": 1.5058183670043945,
      "eval_runtime": 34.7336,
      "eval_samples_per_second": 19.693,
      "step": 8500
    },
    {
      "epoch": 2.03,
      "learning_rate": 1e-05,
      "loss": 1.5643,
      "step": 8510
    },
    {
      "epoch": 2.03,
      "learning_rate": 1e-05,
      "loss": 1.55,
      "step": 8520
    },
    {
      "epoch": 2.04,
      "learning_rate": 1e-05,
      "loss": 1.5278,
      "step": 8530
    },
    {
      "epoch": 2.04,
      "learning_rate": 1e-05,
      "loss": 1.5713,
      "step": 8540
    },
    {
      "epoch": 2.04,
      "learning_rate": 1e-05,
      "loss": 1.5296,
      "step": 8550
    },
    {
      "epoch": 2.04,
      "learning_rate": 1e-05,
      "loss": 1.5387,
      "step": 8560
    },
    {
      "epoch": 2.05,
      "learning_rate": 1e-05,
      "loss": 1.5572,
      "step": 8570
    },
    {
      "epoch": 2.05,
      "learning_rate": 1e-05,
      "loss": 1.5479,
      "step": 8580
    },
    {
      "epoch": 2.05,
      "learning_rate": 1e-05,
      "loss": 1.5585,
      "step": 8590
    },
    {
      "epoch": 2.05,
      "learning_rate": 1e-05,
      "loss": 1.5107,
      "step": 8600
    },
    {
      "epoch": 2.05,
      "eval_accuracy": 0.7675438596491229,
      "eval_loss": 1.5047212839126587,
      "eval_runtime": 34.6014,
      "eval_samples_per_second": 19.768,
      "step": 8600
    },
    {
      "epoch": 2.06,
      "learning_rate": 1e-05,
      "loss": 1.5428,
      "step": 8610
    },
    {
      "epoch": 2.06,
      "learning_rate": 1e-05,
      "loss": 1.5309,
      "step": 8620
    },
    {
      "epoch": 2.06,
      "learning_rate": 1e-05,
      "loss": 1.5702,
      "step": 8630
    },
    {
      "epoch": 2.06,
      "learning_rate": 1e-05,
      "loss": 1.5318,
      "step": 8640
    },
    {
      "epoch": 2.07,
      "learning_rate": 1e-05,
      "loss": 1.5824,
      "step": 8650
    },
    {
      "epoch": 2.07,
      "learning_rate": 1e-05,
      "loss": 1.5271,
      "step": 8660
    },
    {
      "epoch": 2.07,
      "learning_rate": 1e-05,
      "loss": 1.5496,
      "step": 8670
    },
    {
      "epoch": 2.07,
      "learning_rate": 1e-05,
      "loss": 1.5475,
      "step": 8680
    },
    {
      "epoch": 2.08,
      "learning_rate": 1e-05,
      "loss": 1.5647,
      "step": 8690
    },
    {
      "epoch": 2.08,
      "learning_rate": 1e-05,
      "loss": 1.5632,
      "step": 8700
    },
    {
      "epoch": 2.08,
      "eval_accuracy": 0.7865497076023392,
      "eval_loss": 1.5037449598312378,
      "eval_runtime": 34.3273,
      "eval_samples_per_second": 19.926,
      "step": 8700
    },
    {
      "epoch": 2.08,
      "learning_rate": 1e-05,
      "loss": 1.5804,
      "step": 8710
    },
    {
      "epoch": 2.08,
      "learning_rate": 1e-05,
      "loss": 1.5351,
      "step": 8720
    },
    {
      "epoch": 2.09,
      "learning_rate": 1e-05,
      "loss": 1.5146,
      "step": 8730
    },
    {
      "epoch": 2.09,
      "learning_rate": 1e-05,
      "loss": 1.5216,
      "step": 8740
    },
    {
      "epoch": 2.09,
      "learning_rate": 1e-05,
      "loss": 1.5616,
      "step": 8750
    },
    {
      "epoch": 2.09,
      "learning_rate": 1e-05,
      "loss": 1.5474,
      "step": 8760
    },
    {
      "epoch": 2.09,
      "learning_rate": 1e-05,
      "loss": 1.532,
      "step": 8770
    },
    {
      "epoch": 2.1,
      "learning_rate": 1e-05,
      "loss": 1.5545,
      "step": 8780
    },
    {
      "epoch": 2.1,
      "learning_rate": 1e-05,
      "loss": 1.5098,
      "step": 8790
    },
    {
      "epoch": 2.1,
      "learning_rate": 1e-05,
      "loss": 1.5595,
      "step": 8800
    },
    {
      "epoch": 2.1,
      "eval_accuracy": 0.793859649122807,
      "eval_loss": 1.502069115638733,
      "eval_runtime": 34.8044,
      "eval_samples_per_second": 19.653,
      "step": 8800
    },
    {
      "epoch": 2.1,
      "learning_rate": 1e-05,
      "loss": 1.5411,
      "step": 8810
    },
    {
      "epoch": 2.11,
      "learning_rate": 1e-05,
      "loss": 1.561,
      "step": 8820
    },
    {
      "epoch": 2.11,
      "learning_rate": 1e-05,
      "loss": 1.5379,
      "step": 8830
    },
    {
      "epoch": 2.11,
      "learning_rate": 1e-05,
      "loss": 1.5278,
      "step": 8840
    },
    {
      "epoch": 2.11,
      "learning_rate": 1e-05,
      "loss": 1.5542,
      "step": 8850
    },
    {
      "epoch": 2.12,
      "learning_rate": 1e-05,
      "loss": 1.5577,
      "step": 8860
    },
    {
      "epoch": 2.12,
      "learning_rate": 1e-05,
      "loss": 1.5354,
      "step": 8870
    },
    {
      "epoch": 2.12,
      "learning_rate": 1e-05,
      "loss": 1.5502,
      "step": 8880
    },
    {
      "epoch": 2.12,
      "learning_rate": 1e-05,
      "loss": 1.5187,
      "step": 8890
    },
    {
      "epoch": 2.13,
      "learning_rate": 1e-05,
      "loss": 1.529,
      "step": 8900
    },
    {
      "epoch": 2.13,
      "eval_accuracy": 0.7880116959064327,
      "eval_loss": 1.4987003803253174,
      "eval_runtime": 34.8732,
      "eval_samples_per_second": 19.614,
      "step": 8900
    },
    {
      "epoch": 2.13,
      "learning_rate": 1e-05,
      "loss": 1.5515,
      "step": 8910
    },
    {
      "epoch": 2.13,
      "learning_rate": 1e-05,
      "loss": 1.5415,
      "step": 8920
    },
    {
      "epoch": 2.13,
      "learning_rate": 1e-05,
      "loss": 1.5608,
      "step": 8930
    },
    {
      "epoch": 2.14,
      "learning_rate": 1e-05,
      "loss": 1.5678,
      "step": 8940
    },
    {
      "epoch": 2.14,
      "learning_rate": 1e-05,
      "loss": 1.5457,
      "step": 8950
    },
    {
      "epoch": 2.14,
      "learning_rate": 1e-05,
      "loss": 1.5736,
      "step": 8960
    },
    {
      "epoch": 2.14,
      "learning_rate": 1e-05,
      "loss": 1.5329,
      "step": 8970
    },
    {
      "epoch": 2.14,
      "learning_rate": 1e-05,
      "loss": 1.5195,
      "step": 8980
    },
    {
      "epoch": 2.15,
      "learning_rate": 1e-05,
      "loss": 1.5501,
      "step": 8990
    },
    {
      "epoch": 2.15,
      "learning_rate": 1e-05,
      "loss": 1.5475,
      "step": 9000
    },
    {
      "epoch": 2.15,
      "eval_accuracy": 0.8011695906432749,
      "eval_loss": 1.4988216161727905,
      "eval_runtime": 34.6258,
      "eval_samples_per_second": 19.754,
      "step": 9000
    },
    {
      "epoch": 2.15,
      "learning_rate": 1e-05,
      "loss": 1.5678,
      "step": 9010
    },
    {
      "epoch": 2.15,
      "learning_rate": 1e-05,
      "loss": 1.5716,
      "step": 9020
    },
    {
      "epoch": 2.16,
      "learning_rate": 1e-05,
      "loss": 1.5362,
      "step": 9030
    },
    {
      "epoch": 2.16,
      "learning_rate": 1e-05,
      "loss": 1.5737,
      "step": 9040
    },
    {
      "epoch": 2.16,
      "learning_rate": 1e-05,
      "loss": 1.4938,
      "step": 9050
    },
    {
      "epoch": 2.16,
      "learning_rate": 1e-05,
      "loss": 1.5182,
      "step": 9060
    },
    {
      "epoch": 2.17,
      "learning_rate": 1e-05,
      "loss": 1.5063,
      "step": 9070
    },
    {
      "epoch": 2.17,
      "learning_rate": 1e-05,
      "loss": 1.5251,
      "step": 9080
    },
    {
      "epoch": 2.17,
      "learning_rate": 1e-05,
      "loss": 1.5348,
      "step": 9090
    },
    {
      "epoch": 2.17,
      "learning_rate": 1e-05,
      "loss": 1.5205,
      "step": 9100
    },
    {
      "epoch": 2.17,
      "eval_accuracy": 0.8084795321637427,
      "eval_loss": 1.4990298748016357,
      "eval_runtime": 34.6116,
      "eval_samples_per_second": 19.762,
      "step": 9100
    },
    {
      "epoch": 2.18,
      "learning_rate": 1e-05,
      "loss": 1.5534,
      "step": 9110
    },
    {
      "epoch": 2.18,
      "learning_rate": 1e-05,
      "loss": 1.5237,
      "step": 9120
    },
    {
      "epoch": 2.18,
      "learning_rate": 1e-05,
      "loss": 1.5645,
      "step": 9130
    },
    {
      "epoch": 2.18,
      "learning_rate": 1e-05,
      "loss": 1.5213,
      "step": 9140
    },
    {
      "epoch": 2.19,
      "learning_rate": 1e-05,
      "loss": 1.5397,
      "step": 9150
    },
    {
      "epoch": 2.19,
      "learning_rate": 1e-05,
      "loss": 1.5319,
      "step": 9160
    },
    {
      "epoch": 2.19,
      "learning_rate": 1e-05,
      "loss": 1.5117,
      "step": 9170
    },
    {
      "epoch": 2.19,
      "learning_rate": 1e-05,
      "loss": 1.5614,
      "step": 9180
    },
    {
      "epoch": 2.19,
      "learning_rate": 1e-05,
      "loss": 1.5658,
      "step": 9190
    },
    {
      "epoch": 2.2,
      "learning_rate": 1e-05,
      "loss": 1.519,
      "step": 9200
    },
    {
      "epoch": 2.2,
      "eval_accuracy": 0.8201754385964912,
      "eval_loss": 1.4966936111450195,
      "eval_runtime": 34.7477,
      "eval_samples_per_second": 19.685,
      "step": 9200
    },
    {
      "epoch": 2.2,
      "learning_rate": 1e-05,
      "loss": 1.5227,
      "step": 9210
    },
    {
      "epoch": 2.2,
      "learning_rate": 1e-05,
      "loss": 1.5199,
      "step": 9220
    },
    {
      "epoch": 2.2,
      "learning_rate": 1e-05,
      "loss": 1.5655,
      "step": 9230
    },
    {
      "epoch": 2.21,
      "learning_rate": 1e-05,
      "loss": 1.5406,
      "step": 9240
    },
    {
      "epoch": 2.21,
      "learning_rate": 1e-05,
      "loss": 1.5301,
      "step": 9250
    },
    {
      "epoch": 2.21,
      "learning_rate": 1e-05,
      "loss": 1.556,
      "step": 9260
    },
    {
      "epoch": 2.21,
      "learning_rate": 1e-05,
      "loss": 1.5685,
      "step": 9270
    },
    {
      "epoch": 2.22,
      "learning_rate": 1e-05,
      "loss": 1.5711,
      "step": 9280
    },
    {
      "epoch": 2.22,
      "learning_rate": 1e-05,
      "loss": 1.5504,
      "step": 9290
    },
    {
      "epoch": 2.22,
      "learning_rate": 1e-05,
      "loss": 1.4967,
      "step": 9300
    },
    {
      "epoch": 2.22,
      "eval_accuracy": 0.8099415204678363,
      "eval_loss": 1.495988130569458,
      "eval_runtime": 34.8716,
      "eval_samples_per_second": 19.615,
      "step": 9300
    },
    {
      "epoch": 2.22,
      "learning_rate": 1e-05,
      "loss": 1.5207,
      "step": 9310
    },
    {
      "epoch": 2.23,
      "learning_rate": 1e-05,
      "loss": 1.5287,
      "step": 9320
    },
    {
      "epoch": 2.23,
      "learning_rate": 1e-05,
      "loss": 1.5546,
      "step": 9330
    },
    {
      "epoch": 2.23,
      "learning_rate": 1e-05,
      "loss": 1.5412,
      "step": 9340
    },
    {
      "epoch": 2.23,
      "learning_rate": 1e-05,
      "loss": 1.5401,
      "step": 9350
    },
    {
      "epoch": 2.24,
      "learning_rate": 1e-05,
      "loss": 1.5276,
      "step": 9360
    },
    {
      "epoch": 2.24,
      "learning_rate": 1e-05,
      "loss": 1.5296,
      "step": 9370
    },
    {
      "epoch": 2.24,
      "learning_rate": 1e-05,
      "loss": 1.5507,
      "step": 9380
    },
    {
      "epoch": 2.24,
      "learning_rate": 1e-05,
      "loss": 1.542,
      "step": 9390
    },
    {
      "epoch": 2.25,
      "learning_rate": 1e-05,
      "loss": 1.5498,
      "step": 9400
    },
    {
      "epoch": 2.25,
      "eval_accuracy": 0.827485380116959,
      "eval_loss": 1.4945071935653687,
      "eval_runtime": 34.5905,
      "eval_samples_per_second": 19.774,
      "step": 9400
    },
    {
      "epoch": 2.25,
      "learning_rate": 1e-05,
      "loss": 1.5548,
      "step": 9410
    },
    {
      "epoch": 2.25,
      "learning_rate": 1e-05,
      "loss": 1.5612,
      "step": 9420
    },
    {
      "epoch": 2.25,
      "learning_rate": 1e-05,
      "loss": 1.5301,
      "step": 9430
    },
    {
      "epoch": 2.25,
      "learning_rate": 1e-05,
      "loss": 1.5313,
      "step": 9440
    },
    {
      "epoch": 2.26,
      "learning_rate": 1e-05,
      "loss": 1.5339,
      "step": 9450
    },
    {
      "epoch": 2.26,
      "learning_rate": 1e-05,
      "loss": 1.5417,
      "step": 9460
    },
    {
      "epoch": 2.26,
      "learning_rate": 1e-05,
      "loss": 1.5151,
      "step": 9470
    },
    {
      "epoch": 2.26,
      "learning_rate": 1e-05,
      "loss": 1.517,
      "step": 9480
    },
    {
      "epoch": 2.27,
      "learning_rate": 1e-05,
      "loss": 1.4997,
      "step": 9490
    },
    {
      "epoch": 2.27,
      "learning_rate": 1e-05,
      "loss": 1.5254,
      "step": 9500
    },
    {
      "epoch": 2.27,
      "eval_accuracy": 0.8260233918128655,
      "eval_loss": 1.4944781064987183,
      "eval_runtime": 34.5208,
      "eval_samples_per_second": 19.814,
      "step": 9500
    },
    {
      "epoch": 2.27,
      "learning_rate": 1e-05,
      "loss": 1.522,
      "step": 9510
    },
    {
      "epoch": 2.27,
      "learning_rate": 1e-05,
      "loss": 1.5297,
      "step": 9520
    },
    {
      "epoch": 2.28,
      "learning_rate": 1e-05,
      "loss": 1.5578,
      "step": 9530
    },
    {
      "epoch": 2.28,
      "learning_rate": 1e-05,
      "loss": 1.5354,
      "step": 9540
    },
    {
      "epoch": 2.28,
      "learning_rate": 1e-05,
      "loss": 1.5272,
      "step": 9550
    },
    {
      "epoch": 2.28,
      "learning_rate": 1e-05,
      "loss": 1.5469,
      "step": 9560
    },
    {
      "epoch": 2.29,
      "learning_rate": 1e-05,
      "loss": 1.5535,
      "step": 9570
    },
    {
      "epoch": 2.29,
      "learning_rate": 1e-05,
      "loss": 1.5167,
      "step": 9580
    },
    {
      "epoch": 2.29,
      "learning_rate": 1e-05,
      "loss": 1.5568,
      "step": 9590
    },
    {
      "epoch": 2.29,
      "learning_rate": 1e-05,
      "loss": 1.5192,
      "step": 9600
    },
    {
      "epoch": 2.29,
      "eval_accuracy": 0.8391812865497076,
      "eval_loss": 1.4938260316848755,
      "eval_runtime": 34.782,
      "eval_samples_per_second": 19.665,
      "step": 9600
    },
    {
      "epoch": 2.3,
      "learning_rate": 1e-05,
      "loss": 1.58,
      "step": 9610
    },
    {
      "epoch": 2.3,
      "learning_rate": 1e-05,
      "loss": 1.5247,
      "step": 9620
    },
    {
      "epoch": 2.3,
      "learning_rate": 1e-05,
      "loss": 1.5309,
      "step": 9630
    },
    {
      "epoch": 2.3,
      "learning_rate": 1e-05,
      "loss": 1.5332,
      "step": 9640
    },
    {
      "epoch": 2.3,
      "learning_rate": 1e-05,
      "loss": 1.5427,
      "step": 9650
    },
    {
      "epoch": 2.31,
      "learning_rate": 1e-05,
      "loss": 1.5211,
      "step": 9660
    },
    {
      "epoch": 2.31,
      "learning_rate": 1e-05,
      "loss": 1.5599,
      "step": 9670
    },
    {
      "epoch": 2.31,
      "learning_rate": 1e-05,
      "loss": 1.5181,
      "step": 9680
    },
    {
      "epoch": 2.31,
      "learning_rate": 1e-05,
      "loss": 1.5177,
      "step": 9690
    },
    {
      "epoch": 2.32,
      "learning_rate": 1e-05,
      "loss": 1.5205,
      "step": 9700
    },
    {
      "epoch": 2.32,
      "eval_accuracy": 0.847953216374269,
      "eval_loss": 1.490982174873352,
      "eval_runtime": 34.4325,
      "eval_samples_per_second": 19.865,
      "step": 9700
    },
    {
      "epoch": 2.32,
      "learning_rate": 1e-05,
      "loss": 1.5101,
      "step": 9710
    },
    {
      "epoch": 2.32,
      "learning_rate": 1e-05,
      "loss": 1.5197,
      "step": 9720
    },
    {
      "epoch": 2.32,
      "learning_rate": 1e-05,
      "loss": 1.5401,
      "step": 9730
    },
    {
      "epoch": 2.33,
      "learning_rate": 1e-05,
      "loss": 1.542,
      "step": 9740
    },
    {
      "epoch": 2.33,
      "learning_rate": 1e-05,
      "loss": 1.504,
      "step": 9750
    },
    {
      "epoch": 2.33,
      "learning_rate": 1e-05,
      "loss": 1.5102,
      "step": 9760
    },
    {
      "epoch": 2.33,
      "learning_rate": 1e-05,
      "loss": 1.536,
      "step": 9770
    },
    {
      "epoch": 2.34,
      "learning_rate": 1e-05,
      "loss": 1.5149,
      "step": 9780
    },
    {
      "epoch": 2.34,
      "learning_rate": 1e-05,
      "loss": 1.5244,
      "step": 9790
    },
    {
      "epoch": 2.34,
      "learning_rate": 1e-05,
      "loss": 1.5334,
      "step": 9800
    },
    {
      "epoch": 2.34,
      "eval_accuracy": 0.8625730994152047,
      "eval_loss": 1.4914577007293701,
      "eval_runtime": 34.8387,
      "eval_samples_per_second": 19.633,
      "step": 9800
    },
    {
      "epoch": 2.34,
      "learning_rate": 1e-05,
      "loss": 1.5462,
      "step": 9810
    },
    {
      "epoch": 2.35,
      "learning_rate": 1e-05,
      "loss": 1.5355,
      "step": 9820
    },
    {
      "epoch": 2.35,
      "learning_rate": 1e-05,
      "loss": 1.5263,
      "step": 9830
    },
    {
      "epoch": 2.35,
      "learning_rate": 1e-05,
      "loss": 1.5235,
      "step": 9840
    },
    {
      "epoch": 2.35,
      "learning_rate": 1e-05,
      "loss": 1.5181,
      "step": 9850
    },
    {
      "epoch": 2.35,
      "learning_rate": 1e-05,
      "loss": 1.5342,
      "step": 9860
    },
    {
      "epoch": 2.36,
      "learning_rate": 1e-05,
      "loss": 1.5412,
      "step": 9870
    },
    {
      "epoch": 2.36,
      "learning_rate": 1e-05,
      "loss": 1.5289,
      "step": 9880
    },
    {
      "epoch": 2.36,
      "learning_rate": 1e-05,
      "loss": 1.5338,
      "step": 9890
    },
    {
      "epoch": 2.36,
      "learning_rate": 1e-05,
      "loss": 1.5253,
      "step": 9900
    },
    {
      "epoch": 2.36,
      "eval_accuracy": 0.8611111111111112,
      "eval_loss": 1.4913311004638672,
      "eval_runtime": 34.5428,
      "eval_samples_per_second": 19.802,
      "step": 9900
    },
    {
      "epoch": 2.37,
      "learning_rate": 1e-05,
      "loss": 1.5367,
      "step": 9910
    },
    {
      "epoch": 2.37,
      "learning_rate": 1e-05,
      "loss": 1.515,
      "step": 9920
    },
    {
      "epoch": 2.37,
      "learning_rate": 1e-05,
      "loss": 1.5659,
      "step": 9930
    },
    {
      "epoch": 2.37,
      "learning_rate": 1e-05,
      "loss": 1.5261,
      "step": 9940
    },
    {
      "epoch": 2.38,
      "learning_rate": 1e-05,
      "loss": 1.5349,
      "step": 9950
    },
    {
      "epoch": 2.38,
      "learning_rate": 1e-05,
      "loss": 1.5145,
      "step": 9960
    },
    {
      "epoch": 2.38,
      "learning_rate": 1e-05,
      "loss": 1.5189,
      "step": 9970
    },
    {
      "epoch": 2.38,
      "learning_rate": 1e-05,
      "loss": 1.5357,
      "step": 9980
    },
    {
      "epoch": 2.39,
      "learning_rate": 1e-05,
      "loss": 1.5292,
      "step": 9990
    },
    {
      "epoch": 2.39,
      "learning_rate": 1e-05,
      "loss": 1.5056,
      "step": 10000
    },
    {
      "epoch": 2.39,
      "eval_accuracy": 0.8713450292397661,
      "eval_loss": 1.487051010131836,
      "eval_runtime": 34.6498,
      "eval_samples_per_second": 19.74,
      "step": 10000
    },
    {
      "epoch": 2.39,
      "learning_rate": 1e-05,
      "loss": 1.5302,
      "step": 10010
    },
    {
      "epoch": 2.39,
      "learning_rate": 1e-05,
      "loss": 1.5548,
      "step": 10020
    },
    {
      "epoch": 2.4,
      "learning_rate": 1e-05,
      "loss": 1.4915,
      "step": 10030
    },
    {
      "epoch": 2.4,
      "learning_rate": 1e-05,
      "loss": 1.5239,
      "step": 10040
    },
    {
      "epoch": 2.4,
      "learning_rate": 1e-05,
      "loss": 1.5366,
      "step": 10050
    },
    {
      "epoch": 2.4,
      "learning_rate": 1e-05,
      "loss": 1.5136,
      "step": 10060
    },
    {
      "epoch": 2.41,
      "learning_rate": 1e-05,
      "loss": 1.5459,
      "step": 10070
    },
    {
      "epoch": 2.41,
      "learning_rate": 1e-05,
      "loss": 1.5278,
      "step": 10080
    },
    {
      "epoch": 2.41,
      "learning_rate": 1e-05,
      "loss": 1.5418,
      "step": 10090
    },
    {
      "epoch": 2.41,
      "learning_rate": 1e-05,
      "loss": 1.5457,
      "step": 10100
    },
    {
      "epoch": 2.41,
      "eval_accuracy": 0.8845029239766082,
      "eval_loss": 1.487339973449707,
      "eval_runtime": 34.721,
      "eval_samples_per_second": 19.7,
      "step": 10100
    },
    {
      "epoch": 2.41,
      "learning_rate": 1e-05,
      "loss": 1.5291,
      "step": 10110
    },
    {
      "epoch": 2.42,
      "learning_rate": 1e-05,
      "loss": 1.5353,
      "step": 10120
    },
    {
      "epoch": 2.42,
      "learning_rate": 1e-05,
      "loss": 1.5072,
      "step": 10130
    },
    {
      "epoch": 2.42,
      "learning_rate": 1e-05,
      "loss": 1.5309,
      "step": 10140
    },
    {
      "epoch": 2.42,
      "learning_rate": 1e-05,
      "loss": 1.5198,
      "step": 10150
    },
    {
      "epoch": 2.43,
      "learning_rate": 1e-05,
      "loss": 1.5198,
      "step": 10160
    },
    {
      "epoch": 2.43,
      "learning_rate": 1e-05,
      "loss": 1.4959,
      "step": 10170
    },
    {
      "epoch": 2.43,
      "learning_rate": 1e-05,
      "loss": 1.5183,
      "step": 10180
    },
    {
      "epoch": 2.43,
      "learning_rate": 1e-05,
      "loss": 1.5421,
      "step": 10190
    },
    {
      "epoch": 2.44,
      "learning_rate": 1e-05,
      "loss": 1.5362,
      "step": 10200
    },
    {
      "epoch": 2.44,
      "eval_accuracy": 0.8801169590643275,
      "eval_loss": 1.4870365858078003,
      "eval_runtime": 34.577,
      "eval_samples_per_second": 19.782,
      "step": 10200
    },
    {
      "epoch": 2.44,
      "learning_rate": 1e-05,
      "loss": 1.5426,
      "step": 10210
    },
    {
      "epoch": 2.44,
      "learning_rate": 1e-05,
      "loss": 1.4953,
      "step": 10220
    },
    {
      "epoch": 2.44,
      "learning_rate": 1e-05,
      "loss": 1.5335,
      "step": 10230
    },
    {
      "epoch": 2.45,
      "learning_rate": 1e-05,
      "loss": 1.5147,
      "step": 10240
    },
    {
      "epoch": 2.45,
      "learning_rate": 1e-05,
      "loss": 1.5194,
      "step": 10250
    },
    {
      "epoch": 2.45,
      "learning_rate": 1e-05,
      "loss": 1.5367,
      "step": 10260
    },
    {
      "epoch": 2.45,
      "learning_rate": 1e-05,
      "loss": 1.5414,
      "step": 10270
    },
    {
      "epoch": 2.46,
      "learning_rate": 1e-05,
      "loss": 1.5387,
      "step": 10280
    },
    {
      "epoch": 2.46,
      "learning_rate": 1e-05,
      "loss": 1.5279,
      "step": 10290
    },
    {
      "epoch": 2.46,
      "learning_rate": 1e-05,
      "loss": 1.4977,
      "step": 10300
    },
    {
      "epoch": 2.46,
      "eval_accuracy": 0.8786549707602339,
      "eval_loss": 1.4872841835021973,
      "eval_runtime": 34.4067,
      "eval_samples_per_second": 19.88,
      "step": 10300
    },
    {
      "epoch": 2.46,
      "learning_rate": 1e-05,
      "loss": 1.5359,
      "step": 10310
    },
    {
      "epoch": 2.46,
      "learning_rate": 1e-05,
      "loss": 1.5265,
      "step": 10320
    },
    {
      "epoch": 2.47,
      "learning_rate": 1e-05,
      "loss": 1.5063,
      "step": 10330
    },
    {
      "epoch": 2.47,
      "learning_rate": 1e-05,
      "loss": 1.5379,
      "step": 10340
    },
    {
      "epoch": 2.47,
      "learning_rate": 1e-05,
      "loss": 1.5275,
      "step": 10350
    },
    {
      "epoch": 2.47,
      "learning_rate": 1e-05,
      "loss": 1.5162,
      "step": 10360
    },
    {
      "epoch": 2.48,
      "learning_rate": 1e-05,
      "loss": 1.5118,
      "step": 10370
    },
    {
      "epoch": 2.48,
      "learning_rate": 1e-05,
      "loss": 1.5072,
      "step": 10380
    },
    {
      "epoch": 2.48,
      "learning_rate": 1e-05,
      "loss": 1.5159,
      "step": 10390
    },
    {
      "epoch": 2.48,
      "learning_rate": 1e-05,
      "loss": 1.4885,
      "step": 10400
    },
    {
      "epoch": 2.48,
      "eval_accuracy": 0.8903508771929824,
      "eval_loss": 1.4847382307052612,
      "eval_runtime": 34.5259,
      "eval_samples_per_second": 19.811,
      "step": 10400
    },
    {
      "epoch": 2.49,
      "learning_rate": 1e-05,
      "loss": 1.5418,
      "step": 10410
    },
    {
      "epoch": 2.49,
      "learning_rate": 1e-05,
      "loss": 1.5177,
      "step": 10420
    },
    {
      "epoch": 2.49,
      "learning_rate": 1e-05,
      "loss": 1.5198,
      "step": 10430
    },
    {
      "epoch": 2.49,
      "learning_rate": 1e-05,
      "loss": 1.5342,
      "step": 10440
    },
    {
      "epoch": 2.5,
      "learning_rate": 1e-05,
      "loss": 1.5359,
      "step": 10450
    },
    {
      "epoch": 2.5,
      "learning_rate": 1e-05,
      "loss": 1.5385,
      "step": 10460
    },
    {
      "epoch": 2.5,
      "learning_rate": 1e-05,
      "loss": 1.5426,
      "step": 10470
    },
    {
      "epoch": 2.5,
      "learning_rate": 1e-05,
      "loss": 1.5099,
      "step": 10480
    },
    {
      "epoch": 2.51,
      "learning_rate": 1e-05,
      "loss": 1.5215,
      "step": 10490
    },
    {
      "epoch": 2.51,
      "learning_rate": 1e-05,
      "loss": 1.5139,
      "step": 10500
    },
    {
      "epoch": 2.51,
      "eval_accuracy": 0.8991228070175439,
      "eval_loss": 1.4848551750183105,
      "eval_runtime": 34.705,
      "eval_samples_per_second": 19.709,
      "step": 10500
    },
    {
      "epoch": 2.51,
      "learning_rate": 1e-05,
      "loss": 1.5206,
      "step": 10510
    },
    {
      "epoch": 2.51,
      "learning_rate": 1e-05,
      "loss": 1.5249,
      "step": 10520
    },
    {
      "epoch": 2.51,
      "learning_rate": 1e-05,
      "loss": 1.5353,
      "step": 10530
    },
    {
      "epoch": 2.52,
      "learning_rate": 1e-05,
      "loss": 1.5287,
      "step": 10540
    },
    {
      "epoch": 2.52,
      "learning_rate": 1e-05,
      "loss": 1.5123,
      "step": 10550
    },
    {
      "epoch": 2.52,
      "learning_rate": 1e-05,
      "loss": 1.5349,
      "step": 10560
    },
    {
      "epoch": 2.52,
      "learning_rate": 1e-05,
      "loss": 1.5302,
      "step": 10570
    },
    {
      "epoch": 2.53,
      "learning_rate": 1e-05,
      "loss": 1.4932,
      "step": 10580
    },
    {
      "epoch": 2.53,
      "learning_rate": 1e-05,
      "loss": 1.5293,
      "step": 10590
    },
    {
      "epoch": 2.53,
      "learning_rate": 1e-05,
      "loss": 1.5378,
      "step": 10600
    },
    {
      "epoch": 2.53,
      "eval_accuracy": 0.8961988304093568,
      "eval_loss": 1.4837596416473389,
      "eval_runtime": 34.5593,
      "eval_samples_per_second": 19.792,
      "step": 10600
    },
    {
      "epoch": 2.53,
      "learning_rate": 1e-05,
      "loss": 1.5423,
      "step": 10610
    },
    {
      "epoch": 2.54,
      "learning_rate": 1e-05,
      "loss": 1.5088,
      "step": 10620
    },
    {
      "epoch": 2.54,
      "learning_rate": 1e-05,
      "loss": 1.5157,
      "step": 10630
    },
    {
      "epoch": 2.54,
      "learning_rate": 1e-05,
      "loss": 1.5192,
      "step": 10640
    },
    {
      "epoch": 2.54,
      "learning_rate": 1e-05,
      "loss": 1.516,
      "step": 10650
    },
    {
      "epoch": 2.55,
      "learning_rate": 1e-05,
      "loss": 1.5256,
      "step": 10660
    },
    {
      "epoch": 2.55,
      "learning_rate": 1e-05,
      "loss": 1.5497,
      "step": 10670
    },
    {
      "epoch": 2.55,
      "learning_rate": 1e-05,
      "loss": 1.5132,
      "step": 10680
    },
    {
      "epoch": 2.55,
      "learning_rate": 1e-05,
      "loss": 1.5045,
      "step": 10690
    },
    {
      "epoch": 2.56,
      "learning_rate": 1e-05,
      "loss": 1.5365,
      "step": 10700
    },
    {
      "epoch": 2.56,
      "eval_accuracy": 0.8976608187134503,
      "eval_loss": 1.4838720560073853,
      "eval_runtime": 34.5164,
      "eval_samples_per_second": 19.817,
      "step": 10700
    },
    {
      "epoch": 2.56,
      "learning_rate": 1e-05,
      "loss": 1.5267,
      "step": 10710
    },
    {
      "epoch": 2.56,
      "learning_rate": 1e-05,
      "loss": 1.532,
      "step": 10720
    },
    {
      "epoch": 2.56,
      "learning_rate": 1e-05,
      "loss": 1.5096,
      "step": 10730
    },
    {
      "epoch": 2.57,
      "learning_rate": 1e-05,
      "loss": 1.5122,
      "step": 10740
    },
    {
      "epoch": 2.57,
      "learning_rate": 1e-05,
      "loss": 1.5174,
      "step": 10750
    },
    {
      "epoch": 2.57,
      "learning_rate": 1e-05,
      "loss": 1.5326,
      "step": 10760
    },
    {
      "epoch": 2.57,
      "learning_rate": 1e-05,
      "loss": 1.5177,
      "step": 10770
    },
    {
      "epoch": 2.57,
      "learning_rate": 1e-05,
      "loss": 1.5334,
      "step": 10780
    },
    {
      "epoch": 2.58,
      "learning_rate": 1e-05,
      "loss": 1.5376,
      "step": 10790
    },
    {
      "epoch": 2.58,
      "learning_rate": 1e-05,
      "loss": 1.5218,
      "step": 10800
    },
    {
      "epoch": 2.58,
      "eval_accuracy": 0.9035087719298246,
      "eval_loss": 1.4828499555587769,
      "eval_runtime": 34.5692,
      "eval_samples_per_second": 19.786,
      "step": 10800
    },
    {
      "epoch": 2.58,
      "learning_rate": 1e-05,
      "loss": 1.5266,
      "step": 10810
    },
    {
      "epoch": 2.58,
      "learning_rate": 1e-05,
      "loss": 1.5437,
      "step": 10820
    },
    {
      "epoch": 2.59,
      "learning_rate": 1e-05,
      "loss": 1.4883,
      "step": 10830
    },
    {
      "epoch": 2.59,
      "learning_rate": 1e-05,
      "loss": 1.511,
      "step": 10840
    },
    {
      "epoch": 2.59,
      "learning_rate": 1e-05,
      "loss": 1.5131,
      "step": 10850
    },
    {
      "epoch": 2.59,
      "learning_rate": 1e-05,
      "loss": 1.5304,
      "step": 10860
    },
    {
      "epoch": 2.6,
      "learning_rate": 1e-05,
      "loss": 1.5458,
      "step": 10870
    },
    {
      "epoch": 2.6,
      "learning_rate": 1e-05,
      "loss": 1.5058,
      "step": 10880
    },
    {
      "epoch": 2.6,
      "learning_rate": 1e-05,
      "loss": 1.514,
      "step": 10890
    },
    {
      "epoch": 2.6,
      "learning_rate": 1e-05,
      "loss": 1.5279,
      "step": 10900
    },
    {
      "epoch": 2.6,
      "eval_accuracy": 0.9078947368421053,
      "eval_loss": 1.482170581817627,
      "eval_runtime": 34.2834,
      "eval_samples_per_second": 19.951,
      "step": 10900
    },
    {
      "epoch": 2.61,
      "learning_rate": 1e-05,
      "loss": 1.5246,
      "step": 10910
    },
    {
      "epoch": 2.61,
      "learning_rate": 1e-05,
      "loss": 1.5055,
      "step": 10920
    },
    {
      "epoch": 2.61,
      "learning_rate": 1e-05,
      "loss": 1.5732,
      "step": 10930
    },
    {
      "epoch": 2.61,
      "learning_rate": 1e-05,
      "loss": 1.525,
      "step": 10940
    },
    {
      "epoch": 2.62,
      "learning_rate": 1e-05,
      "loss": 1.5413,
      "step": 10950
    },
    {
      "epoch": 2.62,
      "learning_rate": 1e-05,
      "loss": 1.5177,
      "step": 10960
    },
    {
      "epoch": 2.62,
      "learning_rate": 1e-05,
      "loss": 1.5277,
      "step": 10970
    },
    {
      "epoch": 2.62,
      "learning_rate": 1e-05,
      "loss": 1.5245,
      "step": 10980
    },
    {
      "epoch": 2.62,
      "learning_rate": 1e-05,
      "loss": 1.49,
      "step": 10990
    },
    {
      "epoch": 2.63,
      "learning_rate": 1e-05,
      "loss": 1.509,
      "step": 11000
    },
    {
      "epoch": 2.63,
      "eval_accuracy": 0.9195906432748538,
      "eval_loss": 1.4814108610153198,
      "eval_runtime": 34.615,
      "eval_samples_per_second": 19.76,
      "step": 11000
    },
    {
      "epoch": 2.63,
      "learning_rate": 1e-05,
      "loss": 1.5222,
      "step": 11010
    },
    {
      "epoch": 2.63,
      "learning_rate": 1e-05,
      "loss": 1.5011,
      "step": 11020
    },
    {
      "epoch": 2.63,
      "learning_rate": 1e-05,
      "loss": 1.5009,
      "step": 11030
    },
    {
      "epoch": 2.64,
      "learning_rate": 1e-05,
      "loss": 1.5126,
      "step": 11040
    },
    {
      "epoch": 2.64,
      "learning_rate": 1e-05,
      "loss": 1.5154,
      "step": 11050
    },
    {
      "epoch": 2.64,
      "learning_rate": 1e-05,
      "loss": 1.5163,
      "step": 11060
    },
    {
      "epoch": 2.64,
      "learning_rate": 1e-05,
      "loss": 1.5052,
      "step": 11070
    },
    {
      "epoch": 2.65,
      "learning_rate": 1e-05,
      "loss": 1.5281,
      "step": 11080
    },
    {
      "epoch": 2.65,
      "learning_rate": 1e-05,
      "loss": 1.5058,
      "step": 11090
    },
    {
      "epoch": 2.65,
      "learning_rate": 1e-05,
      "loss": 1.486,
      "step": 11100
    },
    {
      "epoch": 2.65,
      "eval_accuracy": 0.9137426900584795,
      "eval_loss": 1.4804058074951172,
      "eval_runtime": 34.6679,
      "eval_samples_per_second": 19.73,
      "step": 11100
    },
    {
      "epoch": 2.65,
      "learning_rate": 1e-05,
      "loss": 1.5023,
      "step": 11110
    },
    {
      "epoch": 2.66,
      "learning_rate": 1e-05,
      "loss": 1.5163,
      "step": 11120
    },
    {
      "epoch": 2.66,
      "learning_rate": 1e-05,
      "loss": 1.495,
      "step": 11130
    },
    {
      "epoch": 2.66,
      "learning_rate": 1e-05,
      "loss": 1.5079,
      "step": 11140
    },
    {
      "epoch": 2.66,
      "learning_rate": 1e-05,
      "loss": 1.5116,
      "step": 11150
    },
    {
      "epoch": 2.67,
      "learning_rate": 1e-05,
      "loss": 1.5293,
      "step": 11160
    },
    {
      "epoch": 2.67,
      "learning_rate": 1e-05,
      "loss": 1.5018,
      "step": 11170
    },
    {
      "epoch": 2.67,
      "learning_rate": 1e-05,
      "loss": 1.5132,
      "step": 11180
    },
    {
      "epoch": 2.67,
      "learning_rate": 1e-05,
      "loss": 1.5358,
      "step": 11190
    },
    {
      "epoch": 2.67,
      "learning_rate": 1e-05,
      "loss": 1.5215,
      "step": 11200
    },
    {
      "epoch": 2.67,
      "eval_accuracy": 0.9122807017543859,
      "eval_loss": 1.4792189598083496,
      "eval_runtime": 34.3842,
      "eval_samples_per_second": 19.893,
      "step": 11200
    },
    {
      "epoch": 2.68,
      "learning_rate": 1e-05,
      "loss": 1.4745,
      "step": 11210
    },
    {
      "epoch": 2.68,
      "learning_rate": 1e-05,
      "loss": 1.4945,
      "step": 11220
    },
    {
      "epoch": 2.68,
      "learning_rate": 1e-05,
      "loss": 1.5236,
      "step": 11230
    },
    {
      "epoch": 2.68,
      "learning_rate": 1e-05,
      "loss": 1.5462,
      "step": 11240
    },
    {
      "epoch": 2.69,
      "learning_rate": 1e-05,
      "loss": 1.4989,
      "step": 11250
    },
    {
      "epoch": 2.69,
      "learning_rate": 1e-05,
      "loss": 1.5129,
      "step": 11260
    },
    {
      "epoch": 2.69,
      "learning_rate": 1e-05,
      "loss": 1.5241,
      "step": 11270
    },
    {
      "epoch": 2.69,
      "learning_rate": 1e-05,
      "loss": 1.5015,
      "step": 11280
    },
    {
      "epoch": 2.7,
      "learning_rate": 1e-05,
      "loss": 1.4968,
      "step": 11290
    },
    {
      "epoch": 2.7,
      "learning_rate": 1e-05,
      "loss": 1.5029,
      "step": 11300
    },
    {
      "epoch": 2.7,
      "eval_accuracy": 0.9137426900584795,
      "eval_loss": 1.481207013130188,
      "eval_runtime": 34.6057,
      "eval_samples_per_second": 19.766,
      "step": 11300
    },
    {
      "epoch": 2.7,
      "learning_rate": 1e-05,
      "loss": 1.5242,
      "step": 11310
    },
    {
      "epoch": 2.7,
      "learning_rate": 1e-05,
      "loss": 1.5007,
      "step": 11320
    },
    {
      "epoch": 2.71,
      "learning_rate": 1e-05,
      "loss": 1.5197,
      "step": 11330
    },
    {
      "epoch": 2.71,
      "learning_rate": 1e-05,
      "loss": 1.4982,
      "step": 11340
    },
    {
      "epoch": 2.71,
      "learning_rate": 1e-05,
      "loss": 1.5174,
      "step": 11350
    },
    {
      "epoch": 2.71,
      "learning_rate": 1e-05,
      "loss": 1.5041,
      "step": 11360
    },
    {
      "epoch": 2.72,
      "learning_rate": 1e-05,
      "loss": 1.5359,
      "step": 11370
    },
    {
      "epoch": 2.72,
      "learning_rate": 1e-05,
      "loss": 1.53,
      "step": 11380
    },
    {
      "epoch": 2.72,
      "learning_rate": 1e-05,
      "loss": 1.5177,
      "step": 11390
    },
    {
      "epoch": 2.72,
      "learning_rate": 1e-05,
      "loss": 1.5126,
      "step": 11400
    },
    {
      "epoch": 2.72,
      "eval_accuracy": 0.9181286549707602,
      "eval_loss": 1.477165937423706,
      "eval_runtime": 34.6419,
      "eval_samples_per_second": 19.745,
      "step": 11400
    },
    {
      "epoch": 2.73,
      "learning_rate": 1e-05,
      "loss": 1.5058,
      "step": 11410
    },
    {
      "epoch": 2.73,
      "learning_rate": 1e-05,
      "loss": 1.4924,
      "step": 11420
    },
    {
      "epoch": 2.73,
      "learning_rate": 1e-05,
      "loss": 1.5118,
      "step": 11430
    },
    {
      "epoch": 2.73,
      "learning_rate": 1e-05,
      "loss": 1.5232,
      "step": 11440
    },
    {
      "epoch": 2.73,
      "learning_rate": 1e-05,
      "loss": 1.494,
      "step": 11450
    },
    {
      "epoch": 2.74,
      "learning_rate": 1e-05,
      "loss": 1.4976,
      "step": 11460
    },
    {
      "epoch": 2.74,
      "learning_rate": 1e-05,
      "loss": 1.5096,
      "step": 11470
    },
    {
      "epoch": 2.74,
      "learning_rate": 1e-05,
      "loss": 1.5019,
      "step": 11480
    },
    {
      "epoch": 2.74,
      "learning_rate": 1e-05,
      "loss": 1.5096,
      "step": 11490
    },
    {
      "epoch": 2.75,
      "learning_rate": 1e-05,
      "loss": 1.5134,
      "step": 11500
    },
    {
      "epoch": 2.75,
      "eval_accuracy": 0.9210526315789473,
      "eval_loss": 1.4787001609802246,
      "eval_runtime": 34.3113,
      "eval_samples_per_second": 19.935,
      "step": 11500
    },
    {
      "epoch": 2.75,
      "learning_rate": 1e-05,
      "loss": 1.4901,
      "step": 11510
    },
    {
      "epoch": 2.75,
      "learning_rate": 1e-05,
      "loss": 1.5114,
      "step": 11520
    },
    {
      "epoch": 2.75,
      "learning_rate": 1e-05,
      "loss": 1.5221,
      "step": 11530
    },
    {
      "epoch": 2.76,
      "learning_rate": 1e-05,
      "loss": 1.523,
      "step": 11540
    },
    {
      "epoch": 2.76,
      "learning_rate": 1e-05,
      "loss": 1.5115,
      "step": 11550
    },
    {
      "epoch": 2.76,
      "learning_rate": 1e-05,
      "loss": 1.5396,
      "step": 11560
    },
    {
      "epoch": 2.76,
      "learning_rate": 1e-05,
      "loss": 1.5044,
      "step": 11570
    },
    {
      "epoch": 2.77,
      "learning_rate": 1e-05,
      "loss": 1.4972,
      "step": 11580
    },
    {
      "epoch": 2.77,
      "learning_rate": 1e-05,
      "loss": 1.5292,
      "step": 11590
    },
    {
      "epoch": 2.77,
      "learning_rate": 1e-05,
      "loss": 1.5099,
      "step": 11600
    },
    {
      "epoch": 2.77,
      "eval_accuracy": 0.9210526315789473,
      "eval_loss": 1.4754481315612793,
      "eval_runtime": 34.4683,
      "eval_samples_per_second": 19.844,
      "step": 11600
    },
    {
      "epoch": 2.77,
      "learning_rate": 1e-05,
      "loss": 1.5091,
      "step": 11610
    },
    {
      "epoch": 2.78,
      "learning_rate": 1e-05,
      "loss": 1.5279,
      "step": 11620
    },
    {
      "epoch": 2.78,
      "learning_rate": 1e-05,
      "loss": 1.5001,
      "step": 11630
    },
    {
      "epoch": 2.78,
      "learning_rate": 1e-05,
      "loss": 1.519,
      "step": 11640
    },
    {
      "epoch": 2.78,
      "learning_rate": 1e-05,
      "loss": 1.526,
      "step": 11650
    },
    {
      "epoch": 2.78,
      "learning_rate": 1e-05,
      "loss": 1.486,
      "step": 11660
    },
    {
      "epoch": 2.79,
      "learning_rate": 1e-05,
      "loss": 1.4788,
      "step": 11670
    },
    {
      "epoch": 2.79,
      "learning_rate": 1e-05,
      "loss": 1.5,
      "step": 11680
    },
    {
      "epoch": 2.79,
      "learning_rate": 1e-05,
      "loss": 1.5145,
      "step": 11690
    },
    {
      "epoch": 2.79,
      "learning_rate": 1e-05,
      "loss": 1.5236,
      "step": 11700
    },
    {
      "epoch": 2.79,
      "eval_accuracy": 0.9254385964912281,
      "eval_loss": 1.4765918254852295,
      "eval_runtime": 34.5843,
      "eval_samples_per_second": 19.778,
      "step": 11700
    },
    {
      "epoch": 2.8,
      "learning_rate": 1e-05,
      "loss": 1.4946,
      "step": 11710
    },
    {
      "epoch": 2.8,
      "learning_rate": 1e-05,
      "loss": 1.5124,
      "step": 11720
    },
    {
      "epoch": 2.8,
      "learning_rate": 1e-05,
      "loss": 1.4859,
      "step": 11730
    },
    {
      "epoch": 2.8,
      "learning_rate": 1e-05,
      "loss": 1.5072,
      "step": 11740
    },
    {
      "epoch": 2.81,
      "learning_rate": 1e-05,
      "loss": 1.5035,
      "step": 11750
    },
    {
      "epoch": 2.81,
      "learning_rate": 1e-05,
      "loss": 1.5206,
      "step": 11760
    },
    {
      "epoch": 2.81,
      "learning_rate": 1e-05,
      "loss": 1.5386,
      "step": 11770
    },
    {
      "epoch": 2.81,
      "learning_rate": 1e-05,
      "loss": 1.479,
      "step": 11780
    },
    {
      "epoch": 2.82,
      "learning_rate": 1e-05,
      "loss": 1.5092,
      "step": 11790
    },
    {
      "epoch": 2.82,
      "learning_rate": 1e-05,
      "loss": 1.5142,
      "step": 11800
    },
    {
      "epoch": 2.82,
      "eval_accuracy": 0.9283625730994152,
      "eval_loss": 1.4759526252746582,
      "eval_runtime": 34.5416,
      "eval_samples_per_second": 19.802,
      "step": 11800
    },
    {
      "epoch": 2.82,
      "learning_rate": 1e-05,
      "loss": 1.507,
      "step": 11810
    },
    {
      "epoch": 2.82,
      "learning_rate": 1e-05,
      "loss": 1.5161,
      "step": 11820
    },
    {
      "epoch": 2.83,
      "learning_rate": 1e-05,
      "loss": 1.4855,
      "step": 11830
    },
    {
      "epoch": 2.83,
      "learning_rate": 1e-05,
      "loss": 1.5089,
      "step": 11840
    },
    {
      "epoch": 2.83,
      "learning_rate": 1e-05,
      "loss": 1.5129,
      "step": 11850
    },
    {
      "epoch": 2.83,
      "learning_rate": 1e-05,
      "loss": 1.5029,
      "step": 11860
    },
    {
      "epoch": 2.83,
      "learning_rate": 1e-05,
      "loss": 1.4955,
      "step": 11870
    },
    {
      "epoch": 2.84,
      "learning_rate": 1e-05,
      "loss": 1.5149,
      "step": 11880
    },
    {
      "epoch": 2.84,
      "learning_rate": 1e-05,
      "loss": 1.5092,
      "step": 11890
    },
    {
      "epoch": 2.84,
      "learning_rate": 1e-05,
      "loss": 1.4913,
      "step": 11900
    },
    {
      "epoch": 2.84,
      "eval_accuracy": 0.9269005847953217,
      "eval_loss": 1.4763075113296509,
      "eval_runtime": 34.4595,
      "eval_samples_per_second": 19.849,
      "step": 11900
    },
    {
      "epoch": 2.84,
      "learning_rate": 1e-05,
      "loss": 1.5192,
      "step": 11910
    },
    {
      "epoch": 2.85,
      "learning_rate": 1e-05,
      "loss": 1.4957,
      "step": 11920
    },
    {
      "epoch": 2.85,
      "learning_rate": 1e-05,
      "loss": 1.4995,
      "step": 11930
    },
    {
      "epoch": 2.85,
      "learning_rate": 1e-05,
      "loss": 1.4929,
      "step": 11940
    },
    {
      "epoch": 2.85,
      "learning_rate": 1e-05,
      "loss": 1.4828,
      "step": 11950
    },
    {
      "epoch": 2.86,
      "learning_rate": 1e-05,
      "loss": 1.5106,
      "step": 11960
    },
    {
      "epoch": 2.86,
      "learning_rate": 1e-05,
      "loss": 1.4979,
      "step": 11970
    },
    {
      "epoch": 2.86,
      "learning_rate": 1e-05,
      "loss": 1.5293,
      "step": 11980
    },
    {
      "epoch": 2.86,
      "learning_rate": 1e-05,
      "loss": 1.5051,
      "step": 11990
    },
    {
      "epoch": 2.87,
      "learning_rate": 1e-05,
      "loss": 1.504,
      "step": 12000
    },
    {
      "epoch": 2.87,
      "eval_accuracy": 0.9327485380116959,
      "eval_loss": 1.4737495183944702,
      "eval_runtime": 34.5194,
      "eval_samples_per_second": 19.815,
      "step": 12000
    },
    {
      "epoch": 2.87,
      "learning_rate": 1e-05,
      "loss": 1.5172,
      "step": 12010
    },
    {
      "epoch": 2.87,
      "learning_rate": 1e-05,
      "loss": 1.4881,
      "step": 12020
    },
    {
      "epoch": 2.87,
      "learning_rate": 1e-05,
      "loss": 1.5011,
      "step": 12030
    },
    {
      "epoch": 2.88,
      "learning_rate": 1e-05,
      "loss": 1.5084,
      "step": 12040
    },
    {
      "epoch": 2.88,
      "learning_rate": 1e-05,
      "loss": 1.504,
      "step": 12050
    },
    {
      "epoch": 2.88,
      "learning_rate": 1e-05,
      "loss": 1.4743,
      "step": 12060
    },
    {
      "epoch": 2.88,
      "learning_rate": 1e-05,
      "loss": 1.4883,
      "step": 12070
    },
    {
      "epoch": 2.89,
      "learning_rate": 1e-05,
      "loss": 1.5069,
      "step": 12080
    },
    {
      "epoch": 2.89,
      "learning_rate": 1e-05,
      "loss": 1.5069,
      "step": 12090
    },
    {
      "epoch": 2.89,
      "learning_rate": 1e-05,
      "loss": 1.493,
      "step": 12100
    },
    {
      "epoch": 2.89,
      "eval_accuracy": 0.9371345029239766,
      "eval_loss": 1.4727365970611572,
      "eval_runtime": 34.5692,
      "eval_samples_per_second": 19.786,
      "step": 12100
    },
    {
      "epoch": 2.89,
      "learning_rate": 1e-05,
      "loss": 1.5191,
      "step": 12110
    },
    {
      "epoch": 2.89,
      "learning_rate": 1e-05,
      "loss": 1.5129,
      "step": 12120
    },
    {
      "epoch": 2.9,
      "learning_rate": 1e-05,
      "loss": 1.5002,
      "step": 12130
    },
    {
      "epoch": 2.9,
      "learning_rate": 1e-05,
      "loss": 1.494,
      "step": 12140
    },
    {
      "epoch": 2.9,
      "learning_rate": 1e-05,
      "loss": 1.5189,
      "step": 12150
    },
    {
      "epoch": 2.9,
      "learning_rate": 1e-05,
      "loss": 1.5022,
      "step": 12160
    },
    {
      "epoch": 2.91,
      "learning_rate": 1e-05,
      "loss": 1.5066,
      "step": 12170
    },
    {
      "epoch": 2.91,
      "learning_rate": 1e-05,
      "loss": 1.5186,
      "step": 12180
    },
    {
      "epoch": 2.91,
      "learning_rate": 1e-05,
      "loss": 1.4923,
      "step": 12190
    },
    {
      "epoch": 2.91,
      "learning_rate": 1e-05,
      "loss": 1.5294,
      "step": 12200
    },
    {
      "epoch": 2.91,
      "eval_accuracy": 0.9342105263157895,
      "eval_loss": 1.4733854532241821,
      "eval_runtime": 34.467,
      "eval_samples_per_second": 19.845,
      "step": 12200
    },
    {
      "epoch": 2.92,
      "learning_rate": 1e-05,
      "loss": 1.5128,
      "step": 12210
    },
    {
      "epoch": 2.92,
      "learning_rate": 1e-05,
      "loss": 1.5094,
      "step": 12220
    },
    {
      "epoch": 2.92,
      "learning_rate": 1e-05,
      "loss": 1.5101,
      "step": 12230
    },
    {
      "epoch": 2.92,
      "learning_rate": 1e-05,
      "loss": 1.486,
      "step": 12240
    },
    {
      "epoch": 2.93,
      "learning_rate": 1e-05,
      "loss": 1.5191,
      "step": 12250
    },
    {
      "epoch": 2.93,
      "learning_rate": 1e-05,
      "loss": 1.4961,
      "step": 12260
    },
    {
      "epoch": 2.93,
      "learning_rate": 1e-05,
      "loss": 1.5318,
      "step": 12270
    },
    {
      "epoch": 2.93,
      "learning_rate": 1e-05,
      "loss": 1.5049,
      "step": 12280
    },
    {
      "epoch": 2.94,
      "learning_rate": 1e-05,
      "loss": 1.5043,
      "step": 12290
    },
    {
      "epoch": 2.94,
      "learning_rate": 1e-05,
      "loss": 1.5031,
      "step": 12300
    },
    {
      "epoch": 2.94,
      "eval_accuracy": 0.9400584795321637,
      "eval_loss": 1.473747968673706,
      "eval_runtime": 34.5925,
      "eval_samples_per_second": 19.773,
      "step": 12300
    },
    {
      "epoch": 2.94,
      "learning_rate": 1e-05,
      "loss": 1.5079,
      "step": 12310
    },
    {
      "epoch": 2.94,
      "learning_rate": 1e-05,
      "loss": 1.5018,
      "step": 12320
    },
    {
      "epoch": 2.94,
      "learning_rate": 1e-05,
      "loss": 1.4927,
      "step": 12330
    },
    {
      "epoch": 2.95,
      "learning_rate": 1e-05,
      "loss": 1.5313,
      "step": 12340
    },
    {
      "epoch": 2.95,
      "learning_rate": 1e-05,
      "loss": 1.4929,
      "step": 12350
    },
    {
      "epoch": 2.95,
      "learning_rate": 1e-05,
      "loss": 1.4884,
      "step": 12360
    },
    {
      "epoch": 2.95,
      "learning_rate": 1e-05,
      "loss": 1.4888,
      "step": 12370
    },
    {
      "epoch": 2.96,
      "learning_rate": 1e-05,
      "loss": 1.5102,
      "step": 12380
    },
    {
      "epoch": 2.96,
      "learning_rate": 1e-05,
      "loss": 1.4947,
      "step": 12390
    },
    {
      "epoch": 2.96,
      "learning_rate": 1e-05,
      "loss": 1.4907,
      "step": 12400
    },
    {
      "epoch": 2.96,
      "eval_accuracy": 0.945906432748538,
      "eval_loss": 1.473435640335083,
      "eval_runtime": 34.7294,
      "eval_samples_per_second": 19.695,
      "step": 12400
    },
    {
      "epoch": 2.96,
      "learning_rate": 1e-05,
      "loss": 1.4989,
      "step": 12410
    },
    {
      "epoch": 2.97,
      "learning_rate": 1e-05,
      "loss": 1.4948,
      "step": 12420
    },
    {
      "epoch": 2.97,
      "learning_rate": 1e-05,
      "loss": 1.4876,
      "step": 12430
    },
    {
      "epoch": 2.97,
      "learning_rate": 1e-05,
      "loss": 1.4944,
      "step": 12440
    },
    {
      "epoch": 2.97,
      "learning_rate": 1e-05,
      "loss": 1.5044,
      "step": 12450
    },
    {
      "epoch": 2.98,
      "learning_rate": 1e-05,
      "loss": 1.5046,
      "step": 12460
    },
    {
      "epoch": 2.98,
      "learning_rate": 1e-05,
      "loss": 1.4792,
      "step": 12470
    },
    {
      "epoch": 2.98,
      "learning_rate": 1e-05,
      "loss": 1.4981,
      "step": 12480
    },
    {
      "epoch": 2.98,
      "learning_rate": 1e-05,
      "loss": 1.5085,
      "step": 12490
    },
    {
      "epoch": 2.99,
      "learning_rate": 1e-05,
      "loss": 1.4982,
      "step": 12500
    },
    {
      "epoch": 2.99,
      "eval_accuracy": 0.945906432748538,
      "eval_loss": 1.4736099243164062,
      "eval_runtime": 34.4881,
      "eval_samples_per_second": 19.833,
      "step": 12500
    },
    {
      "epoch": 2.99,
      "learning_rate": 1e-05,
      "loss": 1.4917,
      "step": 12510
    },
    {
      "epoch": 2.99,
      "learning_rate": 1e-05,
      "loss": 1.5102,
      "step": 12520
    },
    {
      "epoch": 2.99,
      "learning_rate": 1e-05,
      "loss": 1.4763,
      "step": 12530
    },
    {
      "epoch": 2.99,
      "learning_rate": 1e-05,
      "loss": 1.4867,
      "step": 12540
    },
    {
      "epoch": 3.0,
      "learning_rate": 1e-05,
      "loss": 1.5083,
      "step": 12550
    },
    {
      "epoch": 3.0,
      "learning_rate": 1e-05,
      "loss": 1.5044,
      "step": 12560
    },
    {
      "epoch": 3.0,
      "learning_rate": 1e-05,
      "loss": 1.502,
      "step": 12570
    },
    {
      "epoch": 3.0,
      "learning_rate": 1e-05,
      "loss": 1.4932,
      "step": 12580
    },
    {
      "epoch": 3.01,
      "learning_rate": 1e-05,
      "loss": 1.4862,
      "step": 12590
    },
    {
      "epoch": 3.01,
      "learning_rate": 1e-05,
      "loss": 1.4949,
      "step": 12600
    },
    {
      "epoch": 3.01,
      "eval_accuracy": 0.9400584795321637,
      "eval_loss": 1.470678687095642,
      "eval_runtime": 34.3012,
      "eval_samples_per_second": 19.941,
      "step": 12600
    },
    {
      "epoch": 3.01,
      "learning_rate": 1e-05,
      "loss": 1.5157,
      "step": 12610
    },
    {
      "epoch": 3.01,
      "learning_rate": 1e-05,
      "loss": 1.5062,
      "step": 12620
    },
    {
      "epoch": 3.02,
      "learning_rate": 1e-05,
      "loss": 1.4882,
      "step": 12630
    },
    {
      "epoch": 3.02,
      "learning_rate": 1e-05,
      "loss": 1.5041,
      "step": 12640
    },
    {
      "epoch": 3.02,
      "learning_rate": 1e-05,
      "loss": 1.5003,
      "step": 12650
    },
    {
      "epoch": 3.02,
      "learning_rate": 1e-05,
      "loss": 1.5033,
      "step": 12660
    },
    {
      "epoch": 3.03,
      "learning_rate": 1e-05,
      "loss": 1.4919,
      "step": 12670
    },
    {
      "epoch": 3.03,
      "learning_rate": 1e-05,
      "loss": 1.5056,
      "step": 12680
    },
    {
      "epoch": 3.03,
      "learning_rate": 1e-05,
      "loss": 1.5044,
      "step": 12690
    },
    {
      "epoch": 3.03,
      "learning_rate": 1e-05,
      "loss": 1.4918,
      "step": 12700
    },
    {
      "epoch": 3.03,
      "eval_accuracy": 0.9371345029239766,
      "eval_loss": 1.469843864440918,
      "eval_runtime": 34.6138,
      "eval_samples_per_second": 19.761,
      "step": 12700
    },
    {
      "epoch": 3.04,
      "learning_rate": 1e-05,
      "loss": 1.4897,
      "step": 12710
    },
    {
      "epoch": 3.04,
      "learning_rate": 1e-05,
      "loss": 1.4849,
      "step": 12720
    },
    {
      "epoch": 3.04,
      "learning_rate": 1e-05,
      "loss": 1.5061,
      "step": 12730
    },
    {
      "epoch": 3.04,
      "learning_rate": 1e-05,
      "loss": 1.5001,
      "step": 12740
    },
    {
      "epoch": 3.05,
      "learning_rate": 1e-05,
      "loss": 1.4997,
      "step": 12750
    },
    {
      "epoch": 3.05,
      "learning_rate": 1e-05,
      "loss": 1.4954,
      "step": 12760
    },
    {
      "epoch": 3.05,
      "learning_rate": 1e-05,
      "loss": 1.5565,
      "step": 12770
    },
    {
      "epoch": 3.05,
      "learning_rate": 1e-05,
      "loss": 1.5144,
      "step": 12780
    },
    {
      "epoch": 3.05,
      "learning_rate": 1e-05,
      "loss": 1.4737,
      "step": 12790
    },
    {
      "epoch": 3.06,
      "learning_rate": 1e-05,
      "loss": 1.5049,
      "step": 12800
    },
    {
      "epoch": 3.06,
      "eval_accuracy": 0.9415204678362573,
      "eval_loss": 1.4709651470184326,
      "eval_runtime": 34.3962,
      "eval_samples_per_second": 19.886,
      "step": 12800
    },
    {
      "epoch": 3.06,
      "learning_rate": 1e-05,
      "loss": 1.4999,
      "step": 12810
    },
    {
      "epoch": 3.06,
      "learning_rate": 1e-05,
      "loss": 1.504,
      "step": 12820
    },
    {
      "epoch": 3.06,
      "learning_rate": 1e-05,
      "loss": 1.4871,
      "step": 12830
    },
    {
      "epoch": 3.07,
      "learning_rate": 1e-05,
      "loss": 1.4814,
      "step": 12840
    },
    {
      "epoch": 3.07,
      "learning_rate": 1e-05,
      "loss": 1.5025,
      "step": 12850
    },
    {
      "epoch": 3.07,
      "learning_rate": 1e-05,
      "loss": 1.5198,
      "step": 12860
    },
    {
      "epoch": 3.07,
      "learning_rate": 1e-05,
      "loss": 1.4886,
      "step": 12870
    },
    {
      "epoch": 3.08,
      "learning_rate": 1e-05,
      "loss": 1.4894,
      "step": 12880
    },
    {
      "epoch": 3.08,
      "learning_rate": 1e-05,
      "loss": 1.5,
      "step": 12890
    },
    {
      "epoch": 3.08,
      "learning_rate": 1e-05,
      "loss": 1.4921,
      "step": 12900
    },
    {
      "epoch": 3.08,
      "eval_accuracy": 0.9444444444444444,
      "eval_loss": 1.4683705568313599,
      "eval_runtime": 34.4352,
      "eval_samples_per_second": 19.863,
      "step": 12900
    },
    {
      "epoch": 3.08,
      "learning_rate": 1e-05,
      "loss": 1.5121,
      "step": 12910
    },
    {
      "epoch": 3.09,
      "learning_rate": 1e-05,
      "loss": 1.4906,
      "step": 12920
    },
    {
      "epoch": 3.09,
      "learning_rate": 1e-05,
      "loss": 1.5069,
      "step": 12930
    },
    {
      "epoch": 3.09,
      "learning_rate": 1e-05,
      "loss": 1.4996,
      "step": 12940
    },
    {
      "epoch": 3.09,
      "learning_rate": 1e-05,
      "loss": 1.4931,
      "step": 12950
    },
    {
      "epoch": 3.1,
      "learning_rate": 1e-05,
      "loss": 1.4888,
      "step": 12960
    },
    {
      "epoch": 3.1,
      "learning_rate": 1e-05,
      "loss": 1.4862,
      "step": 12970
    },
    {
      "epoch": 3.1,
      "learning_rate": 1e-05,
      "loss": 1.5011,
      "step": 12980
    },
    {
      "epoch": 3.1,
      "learning_rate": 1e-05,
      "loss": 1.5037,
      "step": 12990
    },
    {
      "epoch": 3.1,
      "learning_rate": 1e-05,
      "loss": 1.4844,
      "step": 13000
    },
    {
      "epoch": 3.1,
      "eval_accuracy": 0.9429824561403509,
      "eval_loss": 1.4690443277359009,
      "eval_runtime": 34.4562,
      "eval_samples_per_second": 19.851,
      "step": 13000
    },
    {
      "epoch": 3.11,
      "learning_rate": 1e-05,
      "loss": 1.4838,
      "step": 13010
    },
    {
      "epoch": 3.11,
      "learning_rate": 1e-05,
      "loss": 1.4863,
      "step": 13020
    },
    {
      "epoch": 3.11,
      "learning_rate": 1e-05,
      "loss": 1.4775,
      "step": 13030
    },
    {
      "epoch": 3.11,
      "learning_rate": 1e-05,
      "loss": 1.5221,
      "step": 13040
    },
    {
      "epoch": 3.12,
      "learning_rate": 1e-05,
      "loss": 1.4914,
      "step": 13050
    },
    {
      "epoch": 3.12,
      "learning_rate": 1e-05,
      "loss": 1.5119,
      "step": 13060
    },
    {
      "epoch": 3.12,
      "learning_rate": 1e-05,
      "loss": 1.5084,
      "step": 13070
    },
    {
      "epoch": 3.12,
      "learning_rate": 1e-05,
      "loss": 1.4959,
      "step": 13080
    },
    {
      "epoch": 3.13,
      "learning_rate": 1e-05,
      "loss": 1.505,
      "step": 13090
    },
    {
      "epoch": 3.13,
      "learning_rate": 1e-05,
      "loss": 1.4945,
      "step": 13100
    },
    {
      "epoch": 3.13,
      "eval_accuracy": 0.945906432748538,
      "eval_loss": 1.4675788879394531,
      "eval_runtime": 34.6843,
      "eval_samples_per_second": 19.721,
      "step": 13100
    },
    {
      "epoch": 3.13,
      "learning_rate": 1e-05,
      "loss": 1.4849,
      "step": 13110
    },
    {
      "epoch": 3.13,
      "learning_rate": 1e-05,
      "loss": 1.4842,
      "step": 13120
    },
    {
      "epoch": 3.14,
      "learning_rate": 1e-05,
      "loss": 1.4964,
      "step": 13130
    },
    {
      "epoch": 3.14,
      "learning_rate": 1e-05,
      "loss": 1.4944,
      "step": 13140
    },
    {
      "epoch": 3.14,
      "learning_rate": 1e-05,
      "loss": 1.5508,
      "step": 13150
    },
    {
      "epoch": 3.14,
      "learning_rate": 1e-05,
      "loss": 1.4887,
      "step": 13160
    },
    {
      "epoch": 3.15,
      "learning_rate": 1e-05,
      "loss": 1.4934,
      "step": 13170
    },
    {
      "epoch": 3.15,
      "learning_rate": 1e-05,
      "loss": 1.4822,
      "step": 13180
    },
    {
      "epoch": 3.15,
      "learning_rate": 1e-05,
      "loss": 1.4798,
      "step": 13190
    },
    {
      "epoch": 3.15,
      "learning_rate": 1e-05,
      "loss": 1.4788,
      "step": 13200
    },
    {
      "epoch": 3.15,
      "eval_accuracy": 0.9473684210526315,
      "eval_loss": 1.4664546251296997,
      "eval_runtime": 34.4452,
      "eval_samples_per_second": 19.858,
      "step": 13200
    },
    {
      "epoch": 3.16,
      "learning_rate": 1e-05,
      "loss": 1.4788,
      "step": 13210
    },
    {
      "epoch": 3.16,
      "learning_rate": 1e-05,
      "loss": 1.4818,
      "step": 13220
    },
    {
      "epoch": 3.16,
      "learning_rate": 1e-05,
      "loss": 1.5156,
      "step": 13230
    },
    {
      "epoch": 3.16,
      "learning_rate": 1e-05,
      "loss": 1.5028,
      "step": 13240
    },
    {
      "epoch": 3.16,
      "learning_rate": 1e-05,
      "loss": 1.4972,
      "step": 13250
    },
    {
      "epoch": 3.17,
      "learning_rate": 1e-05,
      "loss": 1.4885,
      "step": 13260
    },
    {
      "epoch": 3.17,
      "learning_rate": 1e-05,
      "loss": 1.4767,
      "step": 13270
    },
    {
      "epoch": 3.17,
      "learning_rate": 1e-05,
      "loss": 1.5007,
      "step": 13280
    },
    {
      "epoch": 3.17,
      "learning_rate": 1e-05,
      "loss": 1.5051,
      "step": 13290
    },
    {
      "epoch": 3.18,
      "learning_rate": 1e-05,
      "loss": 1.4873,
      "step": 13300
    },
    {
      "epoch": 3.18,
      "eval_accuracy": 0.9400584795321637,
      "eval_loss": 1.4688345193862915,
      "eval_runtime": 34.449,
      "eval_samples_per_second": 19.855,
      "step": 13300
    },
    {
      "epoch": 3.18,
      "learning_rate": 1e-05,
      "loss": 1.4943,
      "step": 13310
    },
    {
      "epoch": 3.18,
      "learning_rate": 1e-05,
      "loss": 1.5059,
      "step": 13320
    },
    {
      "epoch": 3.18,
      "learning_rate": 1e-05,
      "loss": 1.4691,
      "step": 13330
    },
    {
      "epoch": 3.19,
      "learning_rate": 1e-05,
      "loss": 1.4927,
      "step": 13340
    },
    {
      "epoch": 3.19,
      "learning_rate": 1e-05,
      "loss": 1.5154,
      "step": 13350
    },
    {
      "epoch": 3.19,
      "learning_rate": 1e-05,
      "loss": 1.4971,
      "step": 13360
    },
    {
      "epoch": 3.19,
      "learning_rate": 1e-05,
      "loss": 1.4743,
      "step": 13370
    },
    {
      "epoch": 3.2,
      "learning_rate": 1e-05,
      "loss": 1.4609,
      "step": 13380
    },
    {
      "epoch": 3.2,
      "learning_rate": 1e-05,
      "loss": 1.4728,
      "step": 13390
    },
    {
      "epoch": 3.2,
      "learning_rate": 1e-05,
      "loss": 1.4823,
      "step": 13400
    },
    {
      "epoch": 3.2,
      "eval_accuracy": 0.9473684210526315,
      "eval_loss": 1.467768669128418,
      "eval_runtime": 34.2852,
      "eval_samples_per_second": 19.95,
      "step": 13400
    },
    {
      "epoch": 3.2,
      "learning_rate": 1e-05,
      "loss": 1.4997,
      "step": 13410
    },
    {
      "epoch": 3.21,
      "learning_rate": 1e-05,
      "loss": 1.4704,
      "step": 13420
    },
    {
      "epoch": 3.21,
      "learning_rate": 1e-05,
      "loss": 1.4818,
      "step": 13430
    },
    {
      "epoch": 3.21,
      "learning_rate": 1e-05,
      "loss": 1.517,
      "step": 13440
    },
    {
      "epoch": 3.21,
      "learning_rate": 1e-05,
      "loss": 1.5066,
      "step": 13450
    },
    {
      "epoch": 3.21,
      "learning_rate": 1e-05,
      "loss": 1.491,
      "step": 13460
    },
    {
      "epoch": 3.22,
      "learning_rate": 1e-05,
      "loss": 1.5016,
      "step": 13470
    },
    {
      "epoch": 3.22,
      "learning_rate": 1e-05,
      "loss": 1.4954,
      "step": 13480
    },
    {
      "epoch": 3.22,
      "learning_rate": 1e-05,
      "loss": 1.4871,
      "step": 13490
    },
    {
      "epoch": 3.22,
      "learning_rate": 1e-05,
      "loss": 1.4882,
      "step": 13500
    },
    {
      "epoch": 3.22,
      "eval_accuracy": 0.9488304093567251,
      "eval_loss": 1.4654582738876343,
      "eval_runtime": 34.2554,
      "eval_samples_per_second": 19.968,
      "step": 13500
    },
    {
      "epoch": 3.23,
      "learning_rate": 1e-05,
      "loss": 1.5075,
      "step": 13510
    },
    {
      "epoch": 3.23,
      "learning_rate": 1e-05,
      "loss": 1.489,
      "step": 13520
    },
    {
      "epoch": 3.23,
      "learning_rate": 1e-05,
      "loss": 1.5079,
      "step": 13530
    },
    {
      "epoch": 3.23,
      "learning_rate": 1e-05,
      "loss": 1.5024,
      "step": 13540
    },
    {
      "epoch": 3.24,
      "learning_rate": 1e-05,
      "loss": 1.5116,
      "step": 13550
    },
    {
      "epoch": 3.24,
      "learning_rate": 1e-05,
      "loss": 1.4953,
      "step": 13560
    },
    {
      "epoch": 3.24,
      "learning_rate": 1e-05,
      "loss": 1.481,
      "step": 13570
    },
    {
      "epoch": 3.24,
      "learning_rate": 1e-05,
      "loss": 1.4875,
      "step": 13580
    },
    {
      "epoch": 3.25,
      "learning_rate": 1e-05,
      "loss": 1.5115,
      "step": 13590
    },
    {
      "epoch": 3.25,
      "learning_rate": 1e-05,
      "loss": 1.4779,
      "step": 13600
    },
    {
      "epoch": 3.25,
      "eval_accuracy": 0.9576023391812866,
      "eval_loss": 1.466334342956543,
      "eval_runtime": 34.5789,
      "eval_samples_per_second": 19.781,
      "step": 13600
    },
    {
      "epoch": 3.25,
      "learning_rate": 1e-05,
      "loss": 1.5016,
      "step": 13610
    },
    {
      "epoch": 3.25,
      "learning_rate": 1e-05,
      "loss": 1.49,
      "step": 13620
    },
    {
      "epoch": 3.26,
      "learning_rate": 1e-05,
      "loss": 1.4841,
      "step": 13630
    },
    {
      "epoch": 3.26,
      "learning_rate": 1e-05,
      "loss": 1.5025,
      "step": 13640
    },
    {
      "epoch": 3.26,
      "learning_rate": 1e-05,
      "loss": 1.4881,
      "step": 13650
    },
    {
      "epoch": 3.26,
      "learning_rate": 1e-05,
      "loss": 1.4769,
      "step": 13660
    },
    {
      "epoch": 3.26,
      "learning_rate": 1e-05,
      "loss": 1.4789,
      "step": 13670
    },
    {
      "epoch": 3.27,
      "learning_rate": 1e-05,
      "loss": 1.4925,
      "step": 13680
    },
    {
      "epoch": 3.27,
      "learning_rate": 1e-05,
      "loss": 1.4823,
      "step": 13690
    },
    {
      "epoch": 3.27,
      "learning_rate": 1e-05,
      "loss": 1.4727,
      "step": 13700
    },
    {
      "epoch": 3.27,
      "eval_accuracy": 0.9576023391812866,
      "eval_loss": 1.4664592742919922,
      "eval_runtime": 34.3375,
      "eval_samples_per_second": 19.92,
      "step": 13700
    },
    {
      "epoch": 3.27,
      "learning_rate": 1e-05,
      "loss": 1.4998,
      "step": 13710
    },
    {
      "epoch": 3.28,
      "learning_rate": 1e-05,
      "loss": 1.4939,
      "step": 13720
    },
    {
      "epoch": 3.28,
      "learning_rate": 1e-05,
      "loss": 1.5096,
      "step": 13730
    },
    {
      "epoch": 3.28,
      "learning_rate": 1e-05,
      "loss": 1.5038,
      "step": 13740
    },
    {
      "epoch": 3.28,
      "learning_rate": 1e-05,
      "loss": 1.5007,
      "step": 13750
    },
    {
      "epoch": 3.29,
      "learning_rate": 1e-05,
      "loss": 1.4995,
      "step": 13760
    },
    {
      "epoch": 3.29,
      "learning_rate": 1e-05,
      "loss": 1.4779,
      "step": 13770
    },
    {
      "epoch": 3.29,
      "learning_rate": 1e-05,
      "loss": 1.4656,
      "step": 13780
    },
    {
      "epoch": 3.29,
      "learning_rate": 1e-05,
      "loss": 1.4839,
      "step": 13790
    },
    {
      "epoch": 3.3,
      "learning_rate": 1e-05,
      "loss": 1.4876,
      "step": 13800
    },
    {
      "epoch": 3.3,
      "eval_accuracy": 0.9619883040935673,
      "eval_loss": 1.4666510820388794,
      "eval_runtime": 34.5356,
      "eval_samples_per_second": 19.806,
      "step": 13800
    },
    {
      "epoch": 3.3,
      "learning_rate": 1e-05,
      "loss": 1.4985,
      "step": 13810
    },
    {
      "epoch": 3.3,
      "learning_rate": 1e-05,
      "loss": 1.4916,
      "step": 13820
    },
    {
      "epoch": 3.3,
      "learning_rate": 1e-05,
      "loss": 1.4813,
      "step": 13830
    },
    {
      "epoch": 3.31,
      "learning_rate": 1e-05,
      "loss": 1.4888,
      "step": 13840
    },
    {
      "epoch": 3.31,
      "learning_rate": 1e-05,
      "loss": 1.4801,
      "step": 13850
    },
    {
      "epoch": 3.31,
      "learning_rate": 1e-05,
      "loss": 1.4889,
      "step": 13860
    },
    {
      "epoch": 3.31,
      "learning_rate": 1e-05,
      "loss": 1.4786,
      "step": 13870
    },
    {
      "epoch": 3.32,
      "learning_rate": 1e-05,
      "loss": 1.5284,
      "step": 13880
    },
    {
      "epoch": 3.32,
      "learning_rate": 1e-05,
      "loss": 1.4968,
      "step": 13890
    },
    {
      "epoch": 3.32,
      "learning_rate": 1e-05,
      "loss": 1.4881,
      "step": 13900
    },
    {
      "epoch": 3.32,
      "eval_accuracy": 0.9590643274853801,
      "eval_loss": 1.4671493768692017,
      "eval_runtime": 34.6506,
      "eval_samples_per_second": 19.74,
      "step": 13900
    },
    {
      "epoch": 3.32,
      "learning_rate": 1e-05,
      "loss": 1.4848,
      "step": 13910
    },
    {
      "epoch": 3.32,
      "learning_rate": 1e-05,
      "loss": 1.482,
      "step": 13920
    },
    {
      "epoch": 3.33,
      "learning_rate": 1e-05,
      "loss": 1.4884,
      "step": 13930
    },
    {
      "epoch": 3.33,
      "learning_rate": 1e-05,
      "loss": 1.471,
      "step": 13940
    },
    {
      "epoch": 3.33,
      "learning_rate": 1e-05,
      "loss": 1.4881,
      "step": 13950
    },
    {
      "epoch": 3.33,
      "learning_rate": 1e-05,
      "loss": 1.4915,
      "step": 13960
    },
    {
      "epoch": 3.34,
      "learning_rate": 1e-05,
      "loss": 1.4696,
      "step": 13970
    },
    {
      "epoch": 3.34,
      "learning_rate": 1e-05,
      "loss": 1.52,
      "step": 13980
    },
    {
      "epoch": 3.34,
      "learning_rate": 1e-05,
      "loss": 1.4777,
      "step": 13990
    },
    {
      "epoch": 3.34,
      "learning_rate": 1e-05,
      "loss": 1.4843,
      "step": 14000
    },
    {
      "epoch": 3.34,
      "eval_accuracy": 0.9634502923976608,
      "eval_loss": 1.4656871557235718,
      "eval_runtime": 34.6072,
      "eval_samples_per_second": 19.765,
      "step": 14000
    },
    {
      "epoch": 3.35,
      "learning_rate": 1e-05,
      "loss": 1.4852,
      "step": 14010
    },
    {
      "epoch": 3.35,
      "learning_rate": 1e-05,
      "loss": 1.497,
      "step": 14020
    },
    {
      "epoch": 3.35,
      "learning_rate": 1e-05,
      "loss": 1.5015,
      "step": 14030
    },
    {
      "epoch": 3.35,
      "learning_rate": 1e-05,
      "loss": 1.485,
      "step": 14040
    },
    {
      "epoch": 3.36,
      "learning_rate": 1e-05,
      "loss": 1.4789,
      "step": 14050
    },
    {
      "epoch": 3.36,
      "learning_rate": 1e-05,
      "loss": 1.482,
      "step": 14060
    },
    {
      "epoch": 3.36,
      "learning_rate": 1e-05,
      "loss": 1.4838,
      "step": 14070
    },
    {
      "epoch": 3.36,
      "learning_rate": 1e-05,
      "loss": 1.4833,
      "step": 14080
    },
    {
      "epoch": 3.37,
      "learning_rate": 1e-05,
      "loss": 1.4807,
      "step": 14090
    },
    {
      "epoch": 3.37,
      "learning_rate": 1e-05,
      "loss": 1.4994,
      "step": 14100
    },
    {
      "epoch": 3.37,
      "eval_accuracy": 0.9605263157894737,
      "eval_loss": 1.4645426273345947,
      "eval_runtime": 34.5892,
      "eval_samples_per_second": 19.775,
      "step": 14100
    },
    {
      "epoch": 3.37,
      "learning_rate": 1e-05,
      "loss": 1.4687,
      "step": 14110
    },
    {
      "epoch": 3.37,
      "learning_rate": 1e-05,
      "loss": 1.4985,
      "step": 14120
    },
    {
      "epoch": 3.37,
      "learning_rate": 1e-05,
      "loss": 1.4885,
      "step": 14130
    },
    {
      "epoch": 3.38,
      "learning_rate": 1e-05,
      "loss": 1.5022,
      "step": 14140
    },
    {
      "epoch": 3.38,
      "learning_rate": 1e-05,
      "loss": 1.4965,
      "step": 14150
    },
    {
      "epoch": 3.38,
      "learning_rate": 1e-05,
      "loss": 1.4941,
      "step": 14160
    },
    {
      "epoch": 3.38,
      "learning_rate": 1e-05,
      "loss": 1.4719,
      "step": 14170
    },
    {
      "epoch": 3.39,
      "learning_rate": 1e-05,
      "loss": 1.4898,
      "step": 14180
    },
    {
      "epoch": 3.39,
      "learning_rate": 1e-05,
      "loss": 1.5021,
      "step": 14190
    },
    {
      "epoch": 3.39,
      "learning_rate": 1e-05,
      "loss": 1.4964,
      "step": 14200
    },
    {
      "epoch": 3.39,
      "eval_accuracy": 0.9619883040935673,
      "eval_loss": 1.4642037153244019,
      "eval_runtime": 34.4892,
      "eval_samples_per_second": 19.832,
      "step": 14200
    },
    {
      "epoch": 3.39,
      "learning_rate": 1e-05,
      "loss": 1.4888,
      "step": 14210
    },
    {
      "epoch": 3.4,
      "learning_rate": 1e-05,
      "loss": 1.4739,
      "step": 14220
    },
    {
      "epoch": 3.4,
      "learning_rate": 1e-05,
      "loss": 1.5043,
      "step": 14230
    },
    {
      "epoch": 3.4,
      "learning_rate": 1e-05,
      "loss": 1.4767,
      "step": 14240
    },
    {
      "epoch": 3.4,
      "learning_rate": 1e-05,
      "loss": 1.4904,
      "step": 14250
    },
    {
      "epoch": 3.41,
      "learning_rate": 1e-05,
      "loss": 1.4858,
      "step": 14260
    },
    {
      "epoch": 3.41,
      "learning_rate": 1e-05,
      "loss": 1.4695,
      "step": 14270
    },
    {
      "epoch": 3.41,
      "learning_rate": 1e-05,
      "loss": 1.4821,
      "step": 14280
    },
    {
      "epoch": 3.41,
      "learning_rate": 1e-05,
      "loss": 1.4777,
      "step": 14290
    },
    {
      "epoch": 3.42,
      "learning_rate": 1e-05,
      "loss": 1.4793,
      "step": 14300
    },
    {
      "epoch": 3.42,
      "eval_accuracy": 0.966374269005848,
      "eval_loss": 1.4635285139083862,
      "eval_runtime": 34.2848,
      "eval_samples_per_second": 19.951,
      "step": 14300
    },
    {
      "epoch": 3.42,
      "learning_rate": 1e-05,
      "loss": 1.4803,
      "step": 14310
    },
    {
      "epoch": 3.42,
      "learning_rate": 1e-05,
      "loss": 1.4828,
      "step": 14320
    },
    {
      "epoch": 3.42,
      "learning_rate": 1e-05,
      "loss": 1.503,
      "step": 14330
    },
    {
      "epoch": 3.42,
      "learning_rate": 1e-05,
      "loss": 1.4881,
      "step": 14340
    },
    {
      "epoch": 3.43,
      "learning_rate": 1e-05,
      "loss": 1.4953,
      "step": 14350
    },
    {
      "epoch": 3.43,
      "learning_rate": 1e-05,
      "loss": 1.4979,
      "step": 14360
    },
    {
      "epoch": 3.43,
      "learning_rate": 1e-05,
      "loss": 1.5232,
      "step": 14370
    },
    {
      "epoch": 3.43,
      "learning_rate": 1e-05,
      "loss": 1.4855,
      "step": 14380
    },
    {
      "epoch": 3.44,
      "learning_rate": 1e-05,
      "loss": 1.488,
      "step": 14390
    },
    {
      "epoch": 3.44,
      "learning_rate": 1e-05,
      "loss": 1.4828,
      "step": 14400
    },
    {
      "epoch": 3.44,
      "eval_accuracy": 0.9605263157894737,
      "eval_loss": 1.4646635055541992,
      "eval_runtime": 34.4497,
      "eval_samples_per_second": 19.855,
      "step": 14400
    },
    {
      "epoch": 3.44,
      "learning_rate": 1e-05,
      "loss": 1.4831,
      "step": 14410
    },
    {
      "epoch": 3.44,
      "learning_rate": 1e-05,
      "loss": 1.5016,
      "step": 14420
    },
    {
      "epoch": 3.45,
      "learning_rate": 1e-05,
      "loss": 1.4922,
      "step": 14430
    },
    {
      "epoch": 3.45,
      "learning_rate": 1e-05,
      "loss": 1.4961,
      "step": 14440
    },
    {
      "epoch": 3.45,
      "learning_rate": 1e-05,
      "loss": 1.4864,
      "step": 14450
    },
    {
      "epoch": 3.45,
      "learning_rate": 1e-05,
      "loss": 1.5,
      "step": 14460
    },
    {
      "epoch": 3.46,
      "learning_rate": 1e-05,
      "loss": 1.4962,
      "step": 14470
    },
    {
      "epoch": 3.46,
      "learning_rate": 1e-05,
      "loss": 1.5027,
      "step": 14480
    },
    {
      "epoch": 3.46,
      "learning_rate": 1e-05,
      "loss": 1.4815,
      "step": 14490
    },
    {
      "epoch": 3.46,
      "learning_rate": 1e-05,
      "loss": 1.4895,
      "step": 14500
    },
    {
      "epoch": 3.46,
      "eval_accuracy": 0.9619883040935673,
      "eval_loss": 1.4644132852554321,
      "eval_runtime": 34.5619,
      "eval_samples_per_second": 19.791,
      "step": 14500
    },
    {
      "epoch": 3.47,
      "learning_rate": 1e-05,
      "loss": 1.4939,
      "step": 14510
    },
    {
      "epoch": 3.47,
      "learning_rate": 1e-05,
      "loss": 1.4743,
      "step": 14520
    },
    {
      "epoch": 3.47,
      "learning_rate": 1e-05,
      "loss": 1.4916,
      "step": 14530
    },
    {
      "epoch": 3.47,
      "learning_rate": 1e-05,
      "loss": 1.4733,
      "step": 14540
    },
    {
      "epoch": 3.48,
      "learning_rate": 1e-05,
      "loss": 1.4771,
      "step": 14550
    },
    {
      "epoch": 3.48,
      "learning_rate": 1e-05,
      "loss": 1.4936,
      "step": 14560
    },
    {
      "epoch": 3.48,
      "learning_rate": 1e-05,
      "loss": 1.4982,
      "step": 14570
    },
    {
      "epoch": 3.48,
      "learning_rate": 1e-05,
      "loss": 1.479,
      "step": 14580
    },
    {
      "epoch": 3.48,
      "learning_rate": 1e-05,
      "loss": 1.4686,
      "step": 14590
    },
    {
      "epoch": 3.49,
      "learning_rate": 1e-05,
      "loss": 1.4983,
      "step": 14600
    },
    {
      "epoch": 3.49,
      "eval_accuracy": 0.9634502923976608,
      "eval_loss": 1.4656448364257812,
      "eval_runtime": 34.6544,
      "eval_samples_per_second": 19.738,
      "step": 14600
    },
    {
      "epoch": 3.49,
      "learning_rate": 1e-05,
      "loss": 1.469,
      "step": 14610
    },
    {
      "epoch": 3.49,
      "learning_rate": 1e-05,
      "loss": 1.4769,
      "step": 14620
    },
    {
      "epoch": 3.49,
      "learning_rate": 1e-05,
      "loss": 1.4839,
      "step": 14630
    },
    {
      "epoch": 3.5,
      "learning_rate": 1e-05,
      "loss": 1.4835,
      "step": 14640
    },
    {
      "epoch": 3.5,
      "learning_rate": 1e-05,
      "loss": 1.4957,
      "step": 14650
    },
    {
      "epoch": 3.5,
      "learning_rate": 1e-05,
      "loss": 1.4758,
      "step": 14660
    },
    {
      "epoch": 3.5,
      "learning_rate": 1e-05,
      "loss": 1.5071,
      "step": 14670
    },
    {
      "epoch": 3.51,
      "learning_rate": 1e-05,
      "loss": 1.4765,
      "step": 14680
    },
    {
      "epoch": 3.51,
      "learning_rate": 1e-05,
      "loss": 1.4766,
      "step": 14690
    },
    {
      "epoch": 3.51,
      "learning_rate": 1e-05,
      "loss": 1.4998,
      "step": 14700
    },
    {
      "epoch": 3.51,
      "eval_accuracy": 0.966374269005848,
      "eval_loss": 1.4635447263717651,
      "eval_runtime": 34.7171,
      "eval_samples_per_second": 19.702,
      "step": 14700
    },
    {
      "epoch": 3.51,
      "learning_rate": 1e-05,
      "loss": 1.4777,
      "step": 14710
    },
    {
      "epoch": 3.52,
      "learning_rate": 1e-05,
      "loss": 1.4935,
      "step": 14720
    },
    {
      "epoch": 3.52,
      "learning_rate": 1e-05,
      "loss": 1.4688,
      "step": 14730
    },
    {
      "epoch": 3.52,
      "learning_rate": 1e-05,
      "loss": 1.4833,
      "step": 14740
    },
    {
      "epoch": 3.52,
      "learning_rate": 1e-05,
      "loss": 1.5011,
      "step": 14750
    },
    {
      "epoch": 3.53,
      "learning_rate": 1e-05,
      "loss": 1.5626,
      "step": 14760
    },
    {
      "epoch": 3.53,
      "learning_rate": 1e-05,
      "loss": 1.4705,
      "step": 14770
    },
    {
      "epoch": 3.53,
      "learning_rate": 1e-05,
      "loss": 1.469,
      "step": 14780
    },
    {
      "epoch": 3.53,
      "learning_rate": 1e-05,
      "loss": 1.4767,
      "step": 14790
    },
    {
      "epoch": 3.53,
      "learning_rate": 1e-05,
      "loss": 1.4888,
      "step": 14800
    },
    {
      "epoch": 3.53,
      "eval_accuracy": 0.9634502923976608,
      "eval_loss": 1.4628511667251587,
      "eval_runtime": 34.4366,
      "eval_samples_per_second": 19.863,
      "step": 14800
    },
    {
      "epoch": 3.54,
      "learning_rate": 1e-05,
      "loss": 1.5053,
      "step": 14810
    },
    {
      "epoch": 3.54,
      "learning_rate": 1e-05,
      "loss": 1.4797,
      "step": 14820
    },
    {
      "epoch": 3.54,
      "learning_rate": 1e-05,
      "loss": 1.4788,
      "step": 14830
    },
    {
      "epoch": 3.54,
      "learning_rate": 1e-05,
      "loss": 1.4667,
      "step": 14840
    },
    {
      "epoch": 3.55,
      "learning_rate": 1e-05,
      "loss": 1.4795,
      "step": 14850
    },
    {
      "epoch": 3.55,
      "learning_rate": 1e-05,
      "loss": 1.4803,
      "step": 14860
    },
    {
      "epoch": 3.55,
      "learning_rate": 1e-05,
      "loss": 1.4885,
      "step": 14870
    },
    {
      "epoch": 3.55,
      "learning_rate": 1e-05,
      "loss": 1.4948,
      "step": 14880
    },
    {
      "epoch": 3.56,
      "learning_rate": 1e-05,
      "loss": 1.4883,
      "step": 14890
    },
    {
      "epoch": 3.56,
      "learning_rate": 1e-05,
      "loss": 1.4733,
      "step": 14900
    },
    {
      "epoch": 3.56,
      "eval_accuracy": 0.9619883040935673,
      "eval_loss": 1.4612964391708374,
      "eval_runtime": 34.4562,
      "eval_samples_per_second": 19.851,
      "step": 14900
    },
    {
      "epoch": 3.56,
      "learning_rate": 1e-05,
      "loss": 1.4891,
      "step": 14910
    },
    {
      "epoch": 3.56,
      "learning_rate": 1e-05,
      "loss": 1.4852,
      "step": 14920
    },
    {
      "epoch": 3.57,
      "learning_rate": 1e-05,
      "loss": 1.4934,
      "step": 14930
    },
    {
      "epoch": 3.57,
      "learning_rate": 1e-05,
      "loss": 1.4916,
      "step": 14940
    },
    {
      "epoch": 3.57,
      "learning_rate": 1e-05,
      "loss": 1.4734,
      "step": 14950
    },
    {
      "epoch": 3.57,
      "learning_rate": 1e-05,
      "loss": 1.4701,
      "step": 14960
    },
    {
      "epoch": 3.58,
      "learning_rate": 1e-05,
      "loss": 1.4816,
      "step": 14970
    },
    {
      "epoch": 3.58,
      "learning_rate": 1e-05,
      "loss": 1.4755,
      "step": 14980
    },
    {
      "epoch": 3.58,
      "learning_rate": 1e-05,
      "loss": 1.4766,
      "step": 14990
    },
    {
      "epoch": 3.58,
      "learning_rate": 1e-05,
      "loss": 1.4841,
      "step": 15000
    },
    {
      "epoch": 3.58,
      "eval_accuracy": 0.9619883040935673,
      "eval_loss": 1.4618583917617798,
      "eval_runtime": 34.5441,
      "eval_samples_per_second": 19.801,
      "step": 15000
    },
    {
      "epoch": 3.58,
      "learning_rate": 1e-05,
      "loss": 1.4921,
      "step": 15010
    },
    {
      "epoch": 3.59,
      "learning_rate": 1e-05,
      "loss": 1.4759,
      "step": 15020
    },
    {
      "epoch": 3.59,
      "learning_rate": 1e-05,
      "loss": 1.496,
      "step": 15030
    },
    {
      "epoch": 3.59,
      "learning_rate": 1e-05,
      "loss": 1.4858,
      "step": 15040
    },
    {
      "epoch": 3.59,
      "learning_rate": 1e-05,
      "loss": 1.4848,
      "step": 15050
    },
    {
      "epoch": 3.6,
      "learning_rate": 1e-05,
      "loss": 1.535,
      "step": 15060
    },
    {
      "epoch": 3.6,
      "learning_rate": 1e-05,
      "loss": 1.4901,
      "step": 15070
    },
    {
      "epoch": 3.6,
      "learning_rate": 1e-05,
      "loss": 1.4633,
      "step": 15080
    },
    {
      "epoch": 3.6,
      "learning_rate": 1e-05,
      "loss": 1.5394,
      "step": 15090
    },
    {
      "epoch": 3.61,
      "learning_rate": 1e-05,
      "loss": 1.4572,
      "step": 15100
    },
    {
      "epoch": 3.61,
      "eval_accuracy": 0.9634502923976608,
      "eval_loss": 1.4611775875091553,
      "eval_runtime": 34.4509,
      "eval_samples_per_second": 19.854,
      "step": 15100
    },
    {
      "epoch": 3.61,
      "learning_rate": 1e-05,
      "loss": 1.4982,
      "step": 15110
    },
    {
      "epoch": 3.61,
      "learning_rate": 1e-05,
      "loss": 1.4831,
      "step": 15120
    },
    {
      "epoch": 3.61,
      "learning_rate": 1e-05,
      "loss": 1.4749,
      "step": 15130
    },
    {
      "epoch": 3.62,
      "learning_rate": 1e-05,
      "loss": 1.4742,
      "step": 15140
    },
    {
      "epoch": 3.62,
      "learning_rate": 1e-05,
      "loss": 1.4889,
      "step": 15150
    },
    {
      "epoch": 3.62,
      "learning_rate": 1e-05,
      "loss": 1.48,
      "step": 15160
    },
    {
      "epoch": 3.62,
      "learning_rate": 1e-05,
      "loss": 1.4824,
      "step": 15170
    },
    {
      "epoch": 3.63,
      "learning_rate": 1e-05,
      "loss": 1.4735,
      "step": 15180
    },
    {
      "epoch": 3.63,
      "learning_rate": 1e-05,
      "loss": 1.4758,
      "step": 15190
    },
    {
      "epoch": 3.63,
      "learning_rate": 1e-05,
      "loss": 1.4887,
      "step": 15200
    },
    {
      "epoch": 3.63,
      "eval_accuracy": 0.9634502923976608,
      "eval_loss": 1.4607446193695068,
      "eval_runtime": 34.554,
      "eval_samples_per_second": 19.795,
      "step": 15200
    },
    {
      "epoch": 3.63,
      "learning_rate": 1e-05,
      "loss": 1.4801,
      "step": 15210
    },
    {
      "epoch": 3.64,
      "learning_rate": 1e-05,
      "loss": 1.4823,
      "step": 15220
    },
    {
      "epoch": 3.64,
      "learning_rate": 1e-05,
      "loss": 1.4736,
      "step": 15230
    },
    {
      "epoch": 3.64,
      "learning_rate": 1e-05,
      "loss": 1.4888,
      "step": 15240
    },
    {
      "epoch": 3.64,
      "learning_rate": 1e-05,
      "loss": 1.476,
      "step": 15250
    },
    {
      "epoch": 3.64,
      "learning_rate": 1e-05,
      "loss": 1.4917,
      "step": 15260
    },
    {
      "epoch": 3.65,
      "learning_rate": 1e-05,
      "loss": 1.4697,
      "step": 15270
    },
    {
      "epoch": 3.65,
      "learning_rate": 1e-05,
      "loss": 1.4932,
      "step": 15280
    },
    {
      "epoch": 3.65,
      "learning_rate": 1e-05,
      "loss": 1.4926,
      "step": 15290
    },
    {
      "epoch": 3.65,
      "learning_rate": 1e-05,
      "loss": 1.4851,
      "step": 15300
    },
    {
      "epoch": 3.65,
      "eval_accuracy": 0.9634502923976608,
      "eval_loss": 1.4600787162780762,
      "eval_runtime": 34.3378,
      "eval_samples_per_second": 19.92,
      "step": 15300
    },
    {
      "epoch": 3.66,
      "learning_rate": 1e-05,
      "loss": 1.4841,
      "step": 15310
    },
    {
      "epoch": 3.66,
      "learning_rate": 1e-05,
      "loss": 1.4667,
      "step": 15320
    },
    {
      "epoch": 3.66,
      "learning_rate": 1e-05,
      "loss": 1.4878,
      "step": 15330
    },
    {
      "epoch": 3.66,
      "learning_rate": 1e-05,
      "loss": 1.4817,
      "step": 15340
    },
    {
      "epoch": 3.67,
      "learning_rate": 1e-05,
      "loss": 1.4755,
      "step": 15350
    },
    {
      "epoch": 3.67,
      "learning_rate": 1e-05,
      "loss": 1.4765,
      "step": 15360
    },
    {
      "epoch": 3.67,
      "learning_rate": 1e-05,
      "loss": 1.4715,
      "step": 15370
    },
    {
      "epoch": 3.67,
      "learning_rate": 1e-05,
      "loss": 1.4746,
      "step": 15380
    },
    {
      "epoch": 3.68,
      "learning_rate": 1e-05,
      "loss": 1.5101,
      "step": 15390
    },
    {
      "epoch": 3.68,
      "learning_rate": 1e-05,
      "loss": 1.4777,
      "step": 15400
    },
    {
      "epoch": 3.68,
      "eval_accuracy": 0.9576023391812866,
      "eval_loss": 1.4621694087982178,
      "eval_runtime": 34.5018,
      "eval_samples_per_second": 19.825,
      "step": 15400
    },
    {
      "epoch": 3.68,
      "learning_rate": 1e-05,
      "loss": 1.4732,
      "step": 15410
    },
    {
      "epoch": 3.68,
      "learning_rate": 1e-05,
      "loss": 1.4951,
      "step": 15420
    },
    {
      "epoch": 3.69,
      "learning_rate": 1e-05,
      "loss": 1.4767,
      "step": 15430
    },
    {
      "epoch": 3.69,
      "learning_rate": 1e-05,
      "loss": 1.4694,
      "step": 15440
    },
    {
      "epoch": 3.69,
      "learning_rate": 1e-05,
      "loss": 1.4763,
      "step": 15450
    },
    {
      "epoch": 3.69,
      "learning_rate": 1e-05,
      "loss": 1.5,
      "step": 15460
    },
    {
      "epoch": 3.69,
      "learning_rate": 1e-05,
      "loss": 1.484,
      "step": 15470
    },
    {
      "epoch": 3.7,
      "learning_rate": 1e-05,
      "loss": 1.4835,
      "step": 15480
    },
    {
      "epoch": 3.7,
      "learning_rate": 1e-05,
      "loss": 1.5006,
      "step": 15490
    },
    {
      "epoch": 3.7,
      "learning_rate": 1e-05,
      "loss": 1.4918,
      "step": 15500
    },
    {
      "epoch": 3.7,
      "eval_accuracy": 0.9649122807017544,
      "eval_loss": 1.460999608039856,
      "eval_runtime": 34.4342,
      "eval_samples_per_second": 19.864,
      "step": 15500
    },
    {
      "epoch": 3.7,
      "learning_rate": 1e-05,
      "loss": 1.4644,
      "step": 15510
    },
    {
      "epoch": 3.71,
      "learning_rate": 1e-05,
      "loss": 1.4797,
      "step": 15520
    },
    {
      "epoch": 3.71,
      "learning_rate": 1e-05,
      "loss": 1.4725,
      "step": 15530
    },
    {
      "epoch": 3.71,
      "learning_rate": 1e-05,
      "loss": 1.4763,
      "step": 15540
    },
    {
      "epoch": 3.71,
      "learning_rate": 1e-05,
      "loss": 1.4933,
      "step": 15550
    },
    {
      "epoch": 3.72,
      "learning_rate": 1e-05,
      "loss": 1.4849,
      "step": 15560
    },
    {
      "epoch": 3.72,
      "learning_rate": 1e-05,
      "loss": 1.4848,
      "step": 15570
    },
    {
      "epoch": 3.72,
      "learning_rate": 1e-05,
      "loss": 1.4635,
      "step": 15580
    },
    {
      "epoch": 3.72,
      "learning_rate": 1e-05,
      "loss": 1.4659,
      "step": 15590
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 1.4613,
      "step": 15600
    },
    {
      "epoch": 3.73,
      "eval_accuracy": 0.966374269005848,
      "eval_loss": 1.4581016302108765,
      "eval_runtime": 34.4544,
      "eval_samples_per_second": 19.852,
      "step": 15600
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 1.4892,
      "step": 15610
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 1.4753,
      "step": 15620
    },
    {
      "epoch": 3.73,
      "learning_rate": 1e-05,
      "loss": 1.4784,
      "step": 15630
    },
    {
      "epoch": 3.74,
      "learning_rate": 1e-05,
      "loss": 1.4758,
      "step": 15640
    },
    {
      "epoch": 3.74,
      "learning_rate": 1e-05,
      "loss": 1.4768,
      "step": 15650
    },
    {
      "epoch": 3.74,
      "learning_rate": 1e-05,
      "loss": 1.4827,
      "step": 15660
    },
    {
      "epoch": 3.74,
      "learning_rate": 1e-05,
      "loss": 1.4669,
      "step": 15670
    },
    {
      "epoch": 3.74,
      "learning_rate": 1e-05,
      "loss": 1.4683,
      "step": 15680
    },
    {
      "epoch": 3.75,
      "learning_rate": 1e-05,
      "loss": 1.4809,
      "step": 15690
    },
    {
      "epoch": 3.75,
      "learning_rate": 1e-05,
      "loss": 1.4801,
      "step": 15700
    },
    {
      "epoch": 3.75,
      "eval_accuracy": 0.966374269005848,
      "eval_loss": 1.459914207458496,
      "eval_runtime": 34.0456,
      "eval_samples_per_second": 20.091,
      "step": 15700
    },
    {
      "epoch": 3.75,
      "learning_rate": 1e-05,
      "loss": 1.4961,
      "step": 15710
    },
    {
      "epoch": 3.75,
      "learning_rate": 1e-05,
      "loss": 1.4895,
      "step": 15720
    },
    {
      "epoch": 3.76,
      "learning_rate": 1e-05,
      "loss": 1.4655,
      "step": 15730
    },
    {
      "epoch": 3.76,
      "learning_rate": 1e-05,
      "loss": 1.4674,
      "step": 15740
    },
    {
      "epoch": 3.76,
      "learning_rate": 1e-05,
      "loss": 1.4598,
      "step": 15750
    },
    {
      "epoch": 3.76,
      "learning_rate": 1e-05,
      "loss": 1.4807,
      "step": 15760
    },
    {
      "epoch": 3.77,
      "learning_rate": 1e-05,
      "loss": 1.4757,
      "step": 15770
    },
    {
      "epoch": 3.77,
      "learning_rate": 1e-05,
      "loss": 1.4823,
      "step": 15780
    },
    {
      "epoch": 3.77,
      "learning_rate": 1e-05,
      "loss": 1.4606,
      "step": 15790
    },
    {
      "epoch": 3.77,
      "learning_rate": 1e-05,
      "loss": 1.4769,
      "step": 15800
    },
    {
      "epoch": 3.77,
      "eval_accuracy": 0.9619883040935673,
      "eval_loss": 1.4586738348007202,
      "eval_runtime": 34.4962,
      "eval_samples_per_second": 19.828,
      "step": 15800
    },
    {
      "epoch": 3.78,
      "learning_rate": 1e-05,
      "loss": 1.479,
      "step": 15810
    },
    {
      "epoch": 3.78,
      "learning_rate": 1e-05,
      "loss": 1.4638,
      "step": 15820
    },
    {
      "epoch": 3.78,
      "learning_rate": 1e-05,
      "loss": 1.4744,
      "step": 15830
    },
    {
      "epoch": 3.78,
      "learning_rate": 1e-05,
      "loss": 1.475,
      "step": 15840
    },
    {
      "epoch": 3.79,
      "learning_rate": 1e-05,
      "loss": 1.4809,
      "step": 15850
    },
    {
      "epoch": 3.79,
      "learning_rate": 1e-05,
      "loss": 1.4879,
      "step": 15860
    },
    {
      "epoch": 3.79,
      "learning_rate": 1e-05,
      "loss": 1.4638,
      "step": 15870
    },
    {
      "epoch": 3.79,
      "learning_rate": 1e-05,
      "loss": 1.4943,
      "step": 15880
    },
    {
      "epoch": 3.8,
      "learning_rate": 1e-05,
      "loss": 1.4876,
      "step": 15890
    },
    {
      "epoch": 3.8,
      "learning_rate": 1e-05,
      "loss": 1.4849,
      "step": 15900
    },
    {
      "epoch": 3.8,
      "eval_accuracy": 0.966374269005848,
      "eval_loss": 1.4557596445083618,
      "eval_runtime": 34.419,
      "eval_samples_per_second": 19.873,
      "step": 15900
    },
    {
      "epoch": 3.8,
      "learning_rate": 1e-05,
      "loss": 1.4694,
      "step": 15910
    },
    {
      "epoch": 3.8,
      "learning_rate": 1e-05,
      "loss": 1.4717,
      "step": 15920
    },
    {
      "epoch": 3.8,
      "learning_rate": 1e-05,
      "loss": 1.4719,
      "step": 15930
    },
    {
      "epoch": 3.81,
      "learning_rate": 1e-05,
      "loss": 1.4632,
      "step": 15940
    },
    {
      "epoch": 3.81,
      "learning_rate": 1e-05,
      "loss": 1.465,
      "step": 15950
    },
    {
      "epoch": 3.81,
      "learning_rate": 1e-05,
      "loss": 1.4864,
      "step": 15960
    },
    {
      "epoch": 3.81,
      "learning_rate": 1e-05,
      "loss": 1.4818,
      "step": 15970
    },
    {
      "epoch": 3.82,
      "learning_rate": 1e-05,
      "loss": 1.4791,
      "step": 15980
    },
    {
      "epoch": 3.82,
      "learning_rate": 1e-05,
      "loss": 1.4692,
      "step": 15990
    },
    {
      "epoch": 3.82,
      "learning_rate": 1e-05,
      "loss": 1.4932,
      "step": 16000
    },
    {
      "epoch": 3.82,
      "eval_accuracy": 0.9634502923976608,
      "eval_loss": 1.458377718925476,
      "eval_runtime": 34.5997,
      "eval_samples_per_second": 19.769,
      "step": 16000
    },
    {
      "epoch": 3.82,
      "learning_rate": 1e-05,
      "loss": 1.4764,
      "step": 16010
    },
    {
      "epoch": 3.83,
      "learning_rate": 1e-05,
      "loss": 1.4758,
      "step": 16020
    },
    {
      "epoch": 3.83,
      "learning_rate": 1e-05,
      "loss": 1.4726,
      "step": 16030
    },
    {
      "epoch": 3.83,
      "learning_rate": 1e-05,
      "loss": 1.4742,
      "step": 16040
    },
    {
      "epoch": 3.83,
      "learning_rate": 1e-05,
      "loss": 1.4616,
      "step": 16050
    },
    {
      "epoch": 3.84,
      "learning_rate": 1e-05,
      "loss": 1.4694,
      "step": 16060
    },
    {
      "epoch": 3.84,
      "learning_rate": 1e-05,
      "loss": 1.4656,
      "step": 16070
    },
    {
      "epoch": 3.84,
      "learning_rate": 1e-05,
      "loss": 1.4694,
      "step": 16080
    },
    {
      "epoch": 3.84,
      "learning_rate": 1e-05,
      "loss": 1.4797,
      "step": 16090
    },
    {
      "epoch": 3.85,
      "learning_rate": 1e-05,
      "loss": 1.485,
      "step": 16100
    },
    {
      "epoch": 3.85,
      "eval_accuracy": 0.9678362573099415,
      "eval_loss": 1.4578373432159424,
      "eval_runtime": 34.2852,
      "eval_samples_per_second": 19.95,
      "step": 16100
    },
    {
      "epoch": 3.85,
      "learning_rate": 1e-05,
      "loss": 1.5491,
      "step": 16110
    },
    {
      "epoch": 3.85,
      "learning_rate": 1e-05,
      "loss": 1.4781,
      "step": 16120
    },
    {
      "epoch": 3.85,
      "learning_rate": 1e-05,
      "loss": 1.4795,
      "step": 16130
    },
    {
      "epoch": 3.85,
      "learning_rate": 1e-05,
      "loss": 1.4771,
      "step": 16140
    },
    {
      "epoch": 3.86,
      "learning_rate": 1e-05,
      "loss": 1.4961,
      "step": 16150
    },
    {
      "epoch": 3.86,
      "learning_rate": 1e-05,
      "loss": 1.4649,
      "step": 16160
    },
    {
      "epoch": 3.86,
      "learning_rate": 1e-05,
      "loss": 1.4725,
      "step": 16170
    },
    {
      "epoch": 3.86,
      "learning_rate": 1e-05,
      "loss": 1.4812,
      "step": 16180
    },
    {
      "epoch": 3.87,
      "learning_rate": 1e-05,
      "loss": 1.4807,
      "step": 16190
    },
    {
      "epoch": 3.87,
      "learning_rate": 1e-05,
      "loss": 1.4651,
      "step": 16200
    },
    {
      "epoch": 3.87,
      "eval_accuracy": 0.9678362573099415,
      "eval_loss": 1.4592318534851074,
      "eval_runtime": 34.2412,
      "eval_samples_per_second": 19.976,
      "step": 16200
    },
    {
      "epoch": 3.87,
      "learning_rate": 1e-05,
      "loss": 1.4719,
      "step": 16210
    },
    {
      "epoch": 3.87,
      "learning_rate": 1e-05,
      "loss": 1.4904,
      "step": 16220
    },
    {
      "epoch": 3.88,
      "learning_rate": 1e-05,
      "loss": 1.4695,
      "step": 16230
    },
    {
      "epoch": 3.88,
      "learning_rate": 1e-05,
      "loss": 1.4792,
      "step": 16240
    },
    {
      "epoch": 3.88,
      "learning_rate": 1e-05,
      "loss": 1.4854,
      "step": 16250
    },
    {
      "epoch": 3.88,
      "learning_rate": 1e-05,
      "loss": 1.4818,
      "step": 16260
    },
    {
      "epoch": 3.89,
      "learning_rate": 1e-05,
      "loss": 1.4725,
      "step": 16270
    },
    {
      "epoch": 3.89,
      "learning_rate": 1e-05,
      "loss": 1.4749,
      "step": 16280
    },
    {
      "epoch": 3.89,
      "learning_rate": 1e-05,
      "loss": 1.473,
      "step": 16290
    },
    {
      "epoch": 3.89,
      "learning_rate": 1e-05,
      "loss": 1.4786,
      "step": 16300
    },
    {
      "epoch": 3.89,
      "eval_accuracy": 0.9649122807017544,
      "eval_loss": 1.4578384160995483,
      "eval_runtime": 34.5779,
      "eval_samples_per_second": 19.781,
      "step": 16300
    },
    {
      "epoch": 3.9,
      "learning_rate": 1e-05,
      "loss": 1.4739,
      "step": 16310
    },
    {
      "epoch": 3.9,
      "learning_rate": 1e-05,
      "loss": 1.4659,
      "step": 16320
    },
    {
      "epoch": 3.9,
      "learning_rate": 1e-05,
      "loss": 1.4854,
      "step": 16330
    },
    {
      "epoch": 3.9,
      "learning_rate": 1e-05,
      "loss": 1.4842,
      "step": 16340
    },
    {
      "epoch": 3.9,
      "learning_rate": 1e-05,
      "loss": 1.4621,
      "step": 16350
    },
    {
      "epoch": 3.91,
      "learning_rate": 1e-05,
      "loss": 1.4549,
      "step": 16360
    },
    {
      "epoch": 3.91,
      "learning_rate": 1e-05,
      "loss": 1.4693,
      "step": 16370
    },
    {
      "epoch": 3.91,
      "learning_rate": 1e-05,
      "loss": 1.4736,
      "step": 16380
    },
    {
      "epoch": 3.91,
      "learning_rate": 1e-05,
      "loss": 1.4614,
      "step": 16390
    },
    {
      "epoch": 3.92,
      "learning_rate": 1e-05,
      "loss": 1.4644,
      "step": 16400
    },
    {
      "epoch": 3.92,
      "eval_accuracy": 0.9678362573099415,
      "eval_loss": 1.456149935722351,
      "eval_runtime": 34.5112,
      "eval_samples_per_second": 19.82,
      "step": 16400
    },
    {
      "epoch": 3.92,
      "learning_rate": 1e-05,
      "loss": 1.4651,
      "step": 16410
    },
    {
      "epoch": 3.92,
      "learning_rate": 1e-05,
      "loss": 1.4704,
      "step": 16420
    },
    {
      "epoch": 3.92,
      "learning_rate": 1e-05,
      "loss": 1.4736,
      "step": 16430
    },
    {
      "epoch": 3.93,
      "learning_rate": 1e-05,
      "loss": 1.4694,
      "step": 16440
    },
    {
      "epoch": 3.93,
      "learning_rate": 1e-05,
      "loss": 1.494,
      "step": 16450
    },
    {
      "epoch": 3.93,
      "learning_rate": 1e-05,
      "loss": 1.4678,
      "step": 16460
    },
    {
      "epoch": 3.93,
      "learning_rate": 1e-05,
      "loss": 1.4655,
      "step": 16470
    },
    {
      "epoch": 3.94,
      "learning_rate": 1e-05,
      "loss": 1.4833,
      "step": 16480
    },
    {
      "epoch": 3.94,
      "learning_rate": 1e-05,
      "loss": 1.4679,
      "step": 16490
    },
    {
      "epoch": 3.94,
      "learning_rate": 1e-05,
      "loss": 1.4713,
      "step": 16500
    },
    {
      "epoch": 3.94,
      "eval_accuracy": 0.9619883040935673,
      "eval_loss": 1.456910252571106,
      "eval_runtime": 34.3649,
      "eval_samples_per_second": 19.904,
      "step": 16500
    },
    {
      "epoch": 3.94,
      "learning_rate": 1e-05,
      "loss": 1.4785,
      "step": 16510
    },
    {
      "epoch": 3.95,
      "learning_rate": 1e-05,
      "loss": 1.4689,
      "step": 16520
    },
    {
      "epoch": 3.95,
      "learning_rate": 1e-05,
      "loss": 1.4665,
      "step": 16530
    },
    {
      "epoch": 3.95,
      "learning_rate": 1e-05,
      "loss": 1.4767,
      "step": 16540
    },
    {
      "epoch": 3.95,
      "learning_rate": 1e-05,
      "loss": 1.4866,
      "step": 16550
    },
    {
      "epoch": 3.96,
      "learning_rate": 1e-05,
      "loss": 1.4779,
      "step": 16560
    },
    {
      "epoch": 3.96,
      "learning_rate": 1e-05,
      "loss": 1.469,
      "step": 16570
    },
    {
      "epoch": 3.96,
      "learning_rate": 1e-05,
      "loss": 1.4686,
      "step": 16580
    },
    {
      "epoch": 3.96,
      "learning_rate": 1e-05,
      "loss": 1.479,
      "step": 16590
    },
    {
      "epoch": 3.96,
      "learning_rate": 1e-05,
      "loss": 1.4687,
      "step": 16600
    },
    {
      "epoch": 3.96,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.4559948444366455,
      "eval_runtime": 34.5888,
      "eval_samples_per_second": 19.775,
      "step": 16600
    },
    {
      "epoch": 3.97,
      "learning_rate": 1e-05,
      "loss": 1.4738,
      "step": 16610
    },
    {
      "epoch": 3.97,
      "learning_rate": 1e-05,
      "loss": 1.4654,
      "step": 16620
    },
    {
      "epoch": 3.97,
      "learning_rate": 1e-05,
      "loss": 1.4686,
      "step": 16630
    },
    {
      "epoch": 3.97,
      "learning_rate": 1e-05,
      "loss": 1.4644,
      "step": 16640
    },
    {
      "epoch": 3.98,
      "learning_rate": 1e-05,
      "loss": 1.4832,
      "step": 16650
    },
    {
      "epoch": 3.98,
      "learning_rate": 1e-05,
      "loss": 1.486,
      "step": 16660
    },
    {
      "epoch": 3.98,
      "learning_rate": 1e-05,
      "loss": 1.4666,
      "step": 16670
    },
    {
      "epoch": 3.98,
      "learning_rate": 1e-05,
      "loss": 1.4595,
      "step": 16680
    },
    {
      "epoch": 3.99,
      "learning_rate": 1e-05,
      "loss": 1.4672,
      "step": 16690
    },
    {
      "epoch": 3.99,
      "learning_rate": 1e-05,
      "loss": 1.5376,
      "step": 16700
    },
    {
      "epoch": 3.99,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.4550962448120117,
      "eval_runtime": 34.5348,
      "eval_samples_per_second": 19.806,
      "step": 16700
    },
    {
      "epoch": 3.99,
      "learning_rate": 1e-05,
      "loss": 1.4622,
      "step": 16710
    },
    {
      "epoch": 3.99,
      "learning_rate": 1e-05,
      "loss": 1.477,
      "step": 16720
    },
    {
      "epoch": 4.0,
      "learning_rate": 1e-05,
      "loss": 1.4998,
      "step": 16730
    },
    {
      "epoch": 4.0,
      "learning_rate": 1e-05,
      "loss": 1.4689,
      "step": 16740
    },
    {
      "epoch": 4.0,
      "learning_rate": 1e-05,
      "loss": 1.4602,
      "step": 16750
    },
    {
      "epoch": 4.0,
      "learning_rate": 1e-05,
      "loss": 1.4736,
      "step": 16760
    },
    {
      "epoch": 4.01,
      "learning_rate": 1e-05,
      "loss": 1.465,
      "step": 16770
    },
    {
      "epoch": 4.01,
      "learning_rate": 1e-05,
      "loss": 1.4656,
      "step": 16780
    },
    {
      "epoch": 4.01,
      "learning_rate": 1e-05,
      "loss": 1.4782,
      "step": 16790
    },
    {
      "epoch": 4.01,
      "learning_rate": 1e-05,
      "loss": 1.4835,
      "step": 16800
    },
    {
      "epoch": 4.01,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.4553635120391846,
      "eval_runtime": 34.5551,
      "eval_samples_per_second": 19.794,
      "step": 16800
    },
    {
      "epoch": 4.01,
      "learning_rate": 1e-05,
      "loss": 1.5024,
      "step": 16810
    },
    {
      "epoch": 4.02,
      "learning_rate": 1e-05,
      "loss": 1.4719,
      "step": 16820
    },
    {
      "epoch": 4.02,
      "learning_rate": 1e-05,
      "loss": 1.471,
      "step": 16830
    },
    {
      "epoch": 4.02,
      "learning_rate": 1e-05,
      "loss": 1.5067,
      "step": 16840
    },
    {
      "epoch": 4.02,
      "learning_rate": 1e-05,
      "loss": 1.4721,
      "step": 16850
    },
    {
      "epoch": 4.03,
      "learning_rate": 1e-05,
      "loss": 1.4555,
      "step": 16860
    },
    {
      "epoch": 4.03,
      "learning_rate": 1e-05,
      "loss": 1.5034,
      "step": 16870
    },
    {
      "epoch": 4.03,
      "learning_rate": 1e-05,
      "loss": 1.4678,
      "step": 16880
    },
    {
      "epoch": 4.03,
      "learning_rate": 1e-05,
      "loss": 1.4824,
      "step": 16890
    },
    {
      "epoch": 4.04,
      "learning_rate": 1e-05,
      "loss": 1.4677,
      "step": 16900
    },
    {
      "epoch": 4.04,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.455155611038208,
      "eval_runtime": 34.387,
      "eval_samples_per_second": 19.891,
      "step": 16900
    },
    {
      "epoch": 4.04,
      "learning_rate": 1e-05,
      "loss": 1.4655,
      "step": 16910
    },
    {
      "epoch": 4.04,
      "learning_rate": 1e-05,
      "loss": 1.4771,
      "step": 16920
    },
    {
      "epoch": 4.04,
      "learning_rate": 1e-05,
      "loss": 1.4757,
      "step": 16930
    },
    {
      "epoch": 4.05,
      "learning_rate": 1e-05,
      "loss": 1.4788,
      "step": 16940
    },
    {
      "epoch": 4.05,
      "learning_rate": 1e-05,
      "loss": 1.4774,
      "step": 16950
    },
    {
      "epoch": 4.05,
      "learning_rate": 1e-05,
      "loss": 1.4761,
      "step": 16960
    },
    {
      "epoch": 4.05,
      "learning_rate": 1e-05,
      "loss": 1.4768,
      "step": 16970
    },
    {
      "epoch": 4.06,
      "learning_rate": 1e-05,
      "loss": 1.4659,
      "step": 16980
    },
    {
      "epoch": 4.06,
      "learning_rate": 1e-05,
      "loss": 1.4636,
      "step": 16990
    },
    {
      "epoch": 4.06,
      "learning_rate": 1e-05,
      "loss": 1.472,
      "step": 17000
    },
    {
      "epoch": 4.06,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.4546358585357666,
      "eval_runtime": 34.4688,
      "eval_samples_per_second": 19.844,
      "step": 17000
    },
    {
      "epoch": 4.06,
      "learning_rate": 1e-05,
      "loss": 1.47,
      "step": 17010
    },
    {
      "epoch": 4.06,
      "learning_rate": 1e-05,
      "loss": 1.4829,
      "step": 17020
    },
    {
      "epoch": 4.07,
      "learning_rate": 1e-05,
      "loss": 1.4868,
      "step": 17030
    },
    {
      "epoch": 4.07,
      "learning_rate": 1e-05,
      "loss": 1.472,
      "step": 17040
    },
    {
      "epoch": 4.07,
      "learning_rate": 1e-05,
      "loss": 1.468,
      "step": 17050
    },
    {
      "epoch": 4.07,
      "learning_rate": 1e-05,
      "loss": 1.4742,
      "step": 17060
    },
    {
      "epoch": 4.08,
      "learning_rate": 1e-05,
      "loss": 1.4642,
      "step": 17070
    },
    {
      "epoch": 4.08,
      "learning_rate": 1e-05,
      "loss": 1.4723,
      "step": 17080
    },
    {
      "epoch": 4.08,
      "learning_rate": 1e-05,
      "loss": 1.4593,
      "step": 17090
    },
    {
      "epoch": 4.08,
      "learning_rate": 1e-05,
      "loss": 1.471,
      "step": 17100
    },
    {
      "epoch": 4.08,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.4545882940292358,
      "eval_runtime": 34.5111,
      "eval_samples_per_second": 19.82,
      "step": 17100
    },
    {
      "epoch": 4.09,
      "learning_rate": 1e-05,
      "loss": 1.4571,
      "step": 17110
    },
    {
      "epoch": 4.09,
      "learning_rate": 1e-05,
      "loss": 1.4653,
      "step": 17120
    },
    {
      "epoch": 4.09,
      "learning_rate": 1e-05,
      "loss": 1.4599,
      "step": 17130
    },
    {
      "epoch": 4.09,
      "learning_rate": 1e-05,
      "loss": 1.4625,
      "step": 17140
    },
    {
      "epoch": 4.1,
      "learning_rate": 1e-05,
      "loss": 1.4575,
      "step": 17150
    },
    {
      "epoch": 4.1,
      "learning_rate": 1e-05,
      "loss": 1.4662,
      "step": 17160
    },
    {
      "epoch": 4.1,
      "learning_rate": 1e-05,
      "loss": 1.4717,
      "step": 17170
    },
    {
      "epoch": 4.1,
      "learning_rate": 1e-05,
      "loss": 1.4443,
      "step": 17180
    },
    {
      "epoch": 4.11,
      "learning_rate": 1e-05,
      "loss": 1.5192,
      "step": 17190
    },
    {
      "epoch": 4.11,
      "learning_rate": 1e-05,
      "loss": 1.4574,
      "step": 17200
    },
    {
      "epoch": 4.11,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.4538660049438477,
      "eval_runtime": 34.3367,
      "eval_samples_per_second": 19.92,
      "step": 17200
    },
    {
      "epoch": 4.11,
      "learning_rate": 1e-05,
      "loss": 1.4724,
      "step": 17210
    },
    {
      "epoch": 4.11,
      "learning_rate": 1e-05,
      "loss": 1.4693,
      "step": 17220
    },
    {
      "epoch": 4.12,
      "learning_rate": 1e-05,
      "loss": 1.4694,
      "step": 17230
    },
    {
      "epoch": 4.12,
      "learning_rate": 1e-05,
      "loss": 1.4646,
      "step": 17240
    },
    {
      "epoch": 4.12,
      "learning_rate": 1e-05,
      "loss": 1.4768,
      "step": 17250
    },
    {
      "epoch": 4.12,
      "learning_rate": 1e-05,
      "loss": 1.4669,
      "step": 17260
    },
    {
      "epoch": 4.12,
      "learning_rate": 1e-05,
      "loss": 1.4603,
      "step": 17270
    },
    {
      "epoch": 4.13,
      "learning_rate": 1e-05,
      "loss": 1.4694,
      "step": 17280
    },
    {
      "epoch": 4.13,
      "learning_rate": 1e-05,
      "loss": 1.4683,
      "step": 17290
    },
    {
      "epoch": 4.13,
      "learning_rate": 1e-05,
      "loss": 1.4614,
      "step": 17300
    },
    {
      "epoch": 4.13,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.4537320137023926,
      "eval_runtime": 34.219,
      "eval_samples_per_second": 19.989,
      "step": 17300
    },
    {
      "epoch": 4.13,
      "learning_rate": 1e-05,
      "loss": 1.4768,
      "step": 17310
    },
    {
      "epoch": 4.14,
      "learning_rate": 1e-05,
      "loss": 1.4699,
      "step": 17320
    },
    {
      "epoch": 4.14,
      "learning_rate": 1e-05,
      "loss": 1.4714,
      "step": 17330
    },
    {
      "epoch": 4.14,
      "learning_rate": 1e-05,
      "loss": 1.4719,
      "step": 17340
    },
    {
      "epoch": 4.14,
      "learning_rate": 1e-05,
      "loss": 1.4666,
      "step": 17350
    },
    {
      "epoch": 4.15,
      "learning_rate": 1e-05,
      "loss": 1.4644,
      "step": 17360
    },
    {
      "epoch": 4.15,
      "learning_rate": 1e-05,
      "loss": 1.4695,
      "step": 17370
    },
    {
      "epoch": 4.15,
      "learning_rate": 1e-05,
      "loss": 1.4642,
      "step": 17380
    },
    {
      "epoch": 4.15,
      "learning_rate": 1e-05,
      "loss": 1.4672,
      "step": 17390
    },
    {
      "epoch": 4.16,
      "learning_rate": 1e-05,
      "loss": 1.4715,
      "step": 17400
    },
    {
      "epoch": 4.16,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.4529050588607788,
      "eval_runtime": 34.4369,
      "eval_samples_per_second": 19.862,
      "step": 17400
    },
    {
      "epoch": 4.16,
      "learning_rate": 1e-05,
      "loss": 1.4688,
      "step": 17410
    },
    {
      "epoch": 4.16,
      "learning_rate": 1e-05,
      "loss": 1.4569,
      "step": 17420
    },
    {
      "epoch": 4.16,
      "learning_rate": 1e-05,
      "loss": 1.4669,
      "step": 17430
    },
    {
      "epoch": 4.17,
      "learning_rate": 1e-05,
      "loss": 1.4935,
      "step": 17440
    },
    {
      "epoch": 4.17,
      "learning_rate": 1e-05,
      "loss": 1.4589,
      "step": 17450
    },
    {
      "epoch": 4.17,
      "learning_rate": 1e-05,
      "loss": 1.474,
      "step": 17460
    },
    {
      "epoch": 4.17,
      "learning_rate": 1e-05,
      "loss": 1.4606,
      "step": 17470
    },
    {
      "epoch": 4.17,
      "learning_rate": 1e-05,
      "loss": 1.4745,
      "step": 17480
    },
    {
      "epoch": 4.18,
      "learning_rate": 1e-05,
      "loss": 1.47,
      "step": 17490
    },
    {
      "epoch": 4.18,
      "learning_rate": 1e-05,
      "loss": 1.4751,
      "step": 17500
    },
    {
      "epoch": 4.18,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.4525117874145508,
      "eval_runtime": 34.6747,
      "eval_samples_per_second": 19.726,
      "step": 17500
    },
    {
      "epoch": 4.18,
      "learning_rate": 1e-05,
      "loss": 1.4784,
      "step": 17510
    },
    {
      "epoch": 4.18,
      "learning_rate": 1e-05,
      "loss": 1.5045,
      "step": 17520
    },
    {
      "epoch": 4.19,
      "learning_rate": 1e-05,
      "loss": 1.4807,
      "step": 17530
    },
    {
      "epoch": 4.19,
      "learning_rate": 1e-05,
      "loss": 1.4769,
      "step": 17540
    },
    {
      "epoch": 4.19,
      "learning_rate": 1e-05,
      "loss": 1.4674,
      "step": 17550
    },
    {
      "epoch": 4.19,
      "learning_rate": 1e-05,
      "loss": 1.4711,
      "step": 17560
    },
    {
      "epoch": 4.2,
      "learning_rate": 1e-05,
      "loss": 1.4739,
      "step": 17570
    },
    {
      "epoch": 4.2,
      "learning_rate": 1e-05,
      "loss": 1.4585,
      "step": 17580
    },
    {
      "epoch": 4.2,
      "learning_rate": 1e-05,
      "loss": 1.4704,
      "step": 17590
    },
    {
      "epoch": 4.2,
      "learning_rate": 1e-05,
      "loss": 1.4748,
      "step": 17600
    },
    {
      "epoch": 4.2,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.4517972469329834,
      "eval_runtime": 34.6089,
      "eval_samples_per_second": 19.764,
      "step": 17600
    },
    {
      "epoch": 4.21,
      "learning_rate": 1e-05,
      "loss": 1.4699,
      "step": 17610
    },
    {
      "epoch": 4.21,
      "learning_rate": 1e-05,
      "loss": 1.4719,
      "step": 17620
    },
    {
      "epoch": 4.21,
      "learning_rate": 1e-05,
      "loss": 1.455,
      "step": 17630
    },
    {
      "epoch": 4.21,
      "learning_rate": 1e-05,
      "loss": 1.4588,
      "step": 17640
    },
    {
      "epoch": 4.22,
      "learning_rate": 1e-05,
      "loss": 1.4539,
      "step": 17650
    },
    {
      "epoch": 4.22,
      "learning_rate": 1e-05,
      "loss": 1.4715,
      "step": 17660
    },
    {
      "epoch": 4.22,
      "learning_rate": 1e-05,
      "loss": 1.4574,
      "step": 17670
    },
    {
      "epoch": 4.22,
      "learning_rate": 1e-05,
      "loss": 1.4634,
      "step": 17680
    },
    {
      "epoch": 4.22,
      "learning_rate": 1e-05,
      "loss": 1.4771,
      "step": 17690
    },
    {
      "epoch": 4.23,
      "learning_rate": 1e-05,
      "loss": 1.47,
      "step": 17700
    },
    {
      "epoch": 4.23,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.4515776634216309,
      "eval_runtime": 34.4259,
      "eval_samples_per_second": 19.869,
      "step": 17700
    },
    {
      "epoch": 4.23,
      "learning_rate": 1e-05,
      "loss": 1.458,
      "step": 17710
    },
    {
      "epoch": 4.23,
      "learning_rate": 1e-05,
      "loss": 1.4645,
      "step": 17720
    },
    {
      "epoch": 4.23,
      "learning_rate": 1e-05,
      "loss": 1.5146,
      "step": 17730
    },
    {
      "epoch": 4.24,
      "learning_rate": 1e-05,
      "loss": 1.4635,
      "step": 17740
    },
    {
      "epoch": 4.24,
      "learning_rate": 1e-05,
      "loss": 1.4641,
      "step": 17750
    },
    {
      "epoch": 4.24,
      "learning_rate": 1e-05,
      "loss": 1.4696,
      "step": 17760
    },
    {
      "epoch": 4.24,
      "learning_rate": 1e-05,
      "loss": 1.4667,
      "step": 17770
    },
    {
      "epoch": 4.25,
      "learning_rate": 1e-05,
      "loss": 1.4645,
      "step": 17780
    },
    {
      "epoch": 4.25,
      "learning_rate": 1e-05,
      "loss": 1.4853,
      "step": 17790
    },
    {
      "epoch": 4.25,
      "learning_rate": 1e-05,
      "loss": 1.4856,
      "step": 17800
    },
    {
      "epoch": 4.25,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.4530291557312012,
      "eval_runtime": 34.6302,
      "eval_samples_per_second": 19.752,
      "step": 17800
    },
    {
      "epoch": 4.25,
      "learning_rate": 1e-05,
      "loss": 1.4702,
      "step": 17810
    },
    {
      "epoch": 4.26,
      "learning_rate": 1e-05,
      "loss": 1.47,
      "step": 17820
    },
    {
      "epoch": 4.26,
      "learning_rate": 1e-05,
      "loss": 1.4667,
      "step": 17830
    },
    {
      "epoch": 4.26,
      "learning_rate": 1e-05,
      "loss": 1.4537,
      "step": 17840
    },
    {
      "epoch": 4.26,
      "learning_rate": 1e-05,
      "loss": 1.4692,
      "step": 17850
    },
    {
      "epoch": 4.27,
      "learning_rate": 1e-05,
      "loss": 1.4651,
      "step": 17860
    },
    {
      "epoch": 4.27,
      "learning_rate": 1e-05,
      "loss": 1.4647,
      "step": 17870
    },
    {
      "epoch": 4.27,
      "learning_rate": 1e-05,
      "loss": 1.4688,
      "step": 17880
    },
    {
      "epoch": 4.27,
      "learning_rate": 1e-05,
      "loss": 1.4855,
      "step": 17890
    },
    {
      "epoch": 4.28,
      "learning_rate": 1e-05,
      "loss": 1.5259,
      "step": 17900
    },
    {
      "epoch": 4.28,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.4523541927337646,
      "eval_runtime": 34.5624,
      "eval_samples_per_second": 19.79,
      "step": 17900
    },
    {
      "epoch": 4.28,
      "learning_rate": 1e-05,
      "loss": 1.4533,
      "step": 17910
    },
    {
      "epoch": 4.28,
      "learning_rate": 1e-05,
      "loss": 1.4672,
      "step": 17920
    },
    {
      "epoch": 4.28,
      "learning_rate": 1e-05,
      "loss": 1.4736,
      "step": 17930
    },
    {
      "epoch": 4.28,
      "learning_rate": 1e-05,
      "loss": 1.4637,
      "step": 17940
    },
    {
      "epoch": 4.29,
      "learning_rate": 1e-05,
      "loss": 1.4703,
      "step": 17950
    },
    {
      "epoch": 4.29,
      "learning_rate": 1e-05,
      "loss": 1.4628,
      "step": 17960
    },
    {
      "epoch": 4.29,
      "learning_rate": 1e-05,
      "loss": 1.4629,
      "step": 17970
    },
    {
      "epoch": 4.29,
      "learning_rate": 1e-05,
      "loss": 1.4692,
      "step": 17980
    },
    {
      "epoch": 4.3,
      "learning_rate": 1e-05,
      "loss": 1.4713,
      "step": 17990
    },
    {
      "epoch": 4.3,
      "learning_rate": 1e-05,
      "loss": 1.4747,
      "step": 18000
    },
    {
      "epoch": 4.3,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.4521273374557495,
      "eval_runtime": 34.2955,
      "eval_samples_per_second": 19.944,
      "step": 18000
    },
    {
      "epoch": 4.3,
      "learning_rate": 1e-05,
      "loss": 1.4637,
      "step": 18010
    },
    {
      "epoch": 4.3,
      "learning_rate": 1e-05,
      "loss": 1.469,
      "step": 18020
    },
    {
      "epoch": 4.31,
      "learning_rate": 1e-05,
      "loss": 1.4803,
      "step": 18030
    },
    {
      "epoch": 4.31,
      "learning_rate": 1e-05,
      "loss": 1.4557,
      "step": 18040
    },
    {
      "epoch": 4.31,
      "learning_rate": 1e-05,
      "loss": 1.4709,
      "step": 18050
    },
    {
      "epoch": 4.31,
      "learning_rate": 1e-05,
      "loss": 1.4639,
      "step": 18060
    },
    {
      "epoch": 4.32,
      "learning_rate": 1e-05,
      "loss": 1.4543,
      "step": 18070
    },
    {
      "epoch": 4.32,
      "learning_rate": 1e-05,
      "loss": 1.4606,
      "step": 18080
    },
    {
      "epoch": 4.32,
      "learning_rate": 1e-05,
      "loss": 1.4687,
      "step": 18090
    },
    {
      "epoch": 4.32,
      "learning_rate": 1e-05,
      "loss": 1.4672,
      "step": 18100
    },
    {
      "epoch": 4.32,
      "eval_accuracy": 0.9678362573099415,
      "eval_loss": 1.4522207975387573,
      "eval_runtime": 34.3997,
      "eval_samples_per_second": 19.884,
      "step": 18100
    },
    {
      "epoch": 4.33,
      "learning_rate": 1e-05,
      "loss": 1.4714,
      "step": 18110
    },
    {
      "epoch": 4.33,
      "learning_rate": 1e-05,
      "loss": 1.4807,
      "step": 18120
    },
    {
      "epoch": 4.33,
      "learning_rate": 1e-05,
      "loss": 1.4643,
      "step": 18130
    },
    {
      "epoch": 4.33,
      "learning_rate": 1e-05,
      "loss": 1.4545,
      "step": 18140
    },
    {
      "epoch": 4.33,
      "learning_rate": 1e-05,
      "loss": 1.4624,
      "step": 18150
    },
    {
      "epoch": 4.34,
      "learning_rate": 1e-05,
      "loss": 1.4713,
      "step": 18160
    },
    {
      "epoch": 4.34,
      "learning_rate": 1e-05,
      "loss": 1.4601,
      "step": 18170
    },
    {
      "epoch": 4.34,
      "learning_rate": 1e-05,
      "loss": 1.4603,
      "step": 18180
    },
    {
      "epoch": 4.34,
      "learning_rate": 1e-05,
      "loss": 1.4734,
      "step": 18190
    },
    {
      "epoch": 4.35,
      "learning_rate": 1e-05,
      "loss": 1.4585,
      "step": 18200
    },
    {
      "epoch": 4.35,
      "eval_accuracy": 0.9634502923976608,
      "eval_loss": 1.4528049230575562,
      "eval_runtime": 34.5242,
      "eval_samples_per_second": 19.812,
      "step": 18200
    },
    {
      "epoch": 4.35,
      "learning_rate": 1e-05,
      "loss": 1.4792,
      "step": 18210
    },
    {
      "epoch": 4.35,
      "learning_rate": 1e-05,
      "loss": 1.4818,
      "step": 18220
    },
    {
      "epoch": 4.35,
      "learning_rate": 1e-05,
      "loss": 1.4613,
      "step": 18230
    },
    {
      "epoch": 4.36,
      "learning_rate": 1e-05,
      "loss": 1.4627,
      "step": 18240
    },
    {
      "epoch": 4.36,
      "learning_rate": 1e-05,
      "loss": 1.4656,
      "step": 18250
    },
    {
      "epoch": 4.36,
      "learning_rate": 1e-05,
      "loss": 1.4704,
      "step": 18260
    },
    {
      "epoch": 4.36,
      "learning_rate": 1e-05,
      "loss": 1.4658,
      "step": 18270
    },
    {
      "epoch": 4.37,
      "learning_rate": 1e-05,
      "loss": 1.4765,
      "step": 18280
    },
    {
      "epoch": 4.37,
      "learning_rate": 1e-05,
      "loss": 1.4664,
      "step": 18290
    },
    {
      "epoch": 4.37,
      "learning_rate": 1e-05,
      "loss": 1.4675,
      "step": 18300
    },
    {
      "epoch": 4.37,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.4518284797668457,
      "eval_runtime": 34.3958,
      "eval_samples_per_second": 19.886,
      "step": 18300
    },
    {
      "epoch": 4.37,
      "learning_rate": 1e-05,
      "loss": 1.4738,
      "step": 18310
    },
    {
      "epoch": 4.38,
      "learning_rate": 1e-05,
      "loss": 1.4722,
      "step": 18320
    },
    {
      "epoch": 4.38,
      "learning_rate": 1e-05,
      "loss": 1.4736,
      "step": 18330
    },
    {
      "epoch": 4.38,
      "learning_rate": 1e-05,
      "loss": 1.474,
      "step": 18340
    },
    {
      "epoch": 4.38,
      "learning_rate": 1e-05,
      "loss": 1.4601,
      "step": 18350
    },
    {
      "epoch": 4.39,
      "learning_rate": 1e-05,
      "loss": 1.461,
      "step": 18360
    },
    {
      "epoch": 4.39,
      "learning_rate": 1e-05,
      "loss": 1.4549,
      "step": 18370
    },
    {
      "epoch": 4.39,
      "learning_rate": 1e-05,
      "loss": 1.4616,
      "step": 18380
    },
    {
      "epoch": 4.39,
      "learning_rate": 1e-05,
      "loss": 1.4686,
      "step": 18390
    },
    {
      "epoch": 4.39,
      "learning_rate": 1e-05,
      "loss": 1.4629,
      "step": 18400
    },
    {
      "epoch": 4.39,
      "eval_accuracy": 0.966374269005848,
      "eval_loss": 1.4509459733963013,
      "eval_runtime": 34.3099,
      "eval_samples_per_second": 19.936,
      "step": 18400
    },
    {
      "epoch": 4.4,
      "learning_rate": 1e-05,
      "loss": 1.4562,
      "step": 18410
    },
    {
      "epoch": 4.4,
      "learning_rate": 1e-05,
      "loss": 1.4597,
      "step": 18420
    },
    {
      "epoch": 4.4,
      "learning_rate": 1e-05,
      "loss": 1.4531,
      "step": 18430
    },
    {
      "epoch": 4.4,
      "learning_rate": 1e-05,
      "loss": 1.4641,
      "step": 18440
    },
    {
      "epoch": 4.41,
      "learning_rate": 1e-05,
      "loss": 1.4682,
      "step": 18450
    },
    {
      "epoch": 4.41,
      "learning_rate": 1e-05,
      "loss": 1.4613,
      "step": 18460
    },
    {
      "epoch": 4.41,
      "learning_rate": 1e-05,
      "loss": 1.4721,
      "step": 18470
    },
    {
      "epoch": 4.41,
      "learning_rate": 1e-05,
      "loss": 1.5142,
      "step": 18480
    },
    {
      "epoch": 4.42,
      "learning_rate": 1e-05,
      "loss": 1.4681,
      "step": 18490
    },
    {
      "epoch": 4.42,
      "learning_rate": 1e-05,
      "loss": 1.4568,
      "step": 18500
    },
    {
      "epoch": 4.42,
      "eval_accuracy": 0.9678362573099415,
      "eval_loss": 1.4500266313552856,
      "eval_runtime": 34.5392,
      "eval_samples_per_second": 19.804,
      "step": 18500
    },
    {
      "epoch": 4.42,
      "learning_rate": 1e-05,
      "loss": 1.4735,
      "step": 18510
    },
    {
      "epoch": 4.42,
      "learning_rate": 1e-05,
      "loss": 1.4651,
      "step": 18520
    },
    {
      "epoch": 4.43,
      "learning_rate": 1e-05,
      "loss": 1.4646,
      "step": 18530
    },
    {
      "epoch": 4.43,
      "learning_rate": 1e-05,
      "loss": 1.4589,
      "step": 18540
    },
    {
      "epoch": 4.43,
      "learning_rate": 1e-05,
      "loss": 1.4783,
      "step": 18550
    },
    {
      "epoch": 4.43,
      "learning_rate": 1e-05,
      "loss": 1.4698,
      "step": 18560
    },
    {
      "epoch": 4.44,
      "learning_rate": 1e-05,
      "loss": 1.4511,
      "step": 18570
    },
    {
      "epoch": 4.44,
      "learning_rate": 1e-05,
      "loss": 1.4525,
      "step": 18580
    },
    {
      "epoch": 4.44,
      "learning_rate": 1e-05,
      "loss": 1.4652,
      "step": 18590
    },
    {
      "epoch": 4.44,
      "learning_rate": 1e-05,
      "loss": 1.4743,
      "step": 18600
    },
    {
      "epoch": 4.44,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.4490563869476318,
      "eval_runtime": 34.3288,
      "eval_samples_per_second": 19.925,
      "step": 18600
    },
    {
      "epoch": 4.44,
      "learning_rate": 1e-05,
      "loss": 1.4519,
      "step": 18610
    },
    {
      "epoch": 4.45,
      "learning_rate": 1e-05,
      "loss": 1.4695,
      "step": 18620
    },
    {
      "epoch": 4.45,
      "learning_rate": 1e-05,
      "loss": 1.4512,
      "step": 18630
    },
    {
      "epoch": 4.45,
      "learning_rate": 1e-05,
      "loss": 1.4631,
      "step": 18640
    },
    {
      "epoch": 4.45,
      "learning_rate": 1e-05,
      "loss": 1.4711,
      "step": 18650
    },
    {
      "epoch": 4.46,
      "learning_rate": 1e-05,
      "loss": 1.4663,
      "step": 18660
    },
    {
      "epoch": 4.46,
      "learning_rate": 1e-05,
      "loss": 1.4638,
      "step": 18670
    },
    {
      "epoch": 4.46,
      "learning_rate": 1e-05,
      "loss": 1.464,
      "step": 18680
    },
    {
      "epoch": 4.46,
      "learning_rate": 1e-05,
      "loss": 1.4618,
      "step": 18690
    },
    {
      "epoch": 4.47,
      "learning_rate": 1e-05,
      "loss": 1.4727,
      "step": 18700
    },
    {
      "epoch": 4.47,
      "eval_accuracy": 0.966374269005848,
      "eval_loss": 1.4488704204559326,
      "eval_runtime": 34.4895,
      "eval_samples_per_second": 19.832,
      "step": 18700
    },
    {
      "epoch": 4.47,
      "learning_rate": 1e-05,
      "loss": 1.4792,
      "step": 18710
    },
    {
      "epoch": 4.47,
      "learning_rate": 1e-05,
      "loss": 1.4505,
      "step": 18720
    },
    {
      "epoch": 4.47,
      "learning_rate": 1e-05,
      "loss": 1.477,
      "step": 18730
    },
    {
      "epoch": 4.48,
      "learning_rate": 1e-05,
      "loss": 1.4563,
      "step": 18740
    },
    {
      "epoch": 4.48,
      "learning_rate": 1e-05,
      "loss": 1.4601,
      "step": 18750
    },
    {
      "epoch": 4.48,
      "learning_rate": 1e-05,
      "loss": 1.4647,
      "step": 18760
    },
    {
      "epoch": 4.48,
      "learning_rate": 1e-05,
      "loss": 1.4609,
      "step": 18770
    },
    {
      "epoch": 4.49,
      "learning_rate": 1e-05,
      "loss": 1.4732,
      "step": 18780
    },
    {
      "epoch": 4.49,
      "learning_rate": 1e-05,
      "loss": 1.4657,
      "step": 18790
    },
    {
      "epoch": 4.49,
      "learning_rate": 1e-05,
      "loss": 1.4768,
      "step": 18800
    },
    {
      "epoch": 4.49,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.4500631093978882,
      "eval_runtime": 34.6735,
      "eval_samples_per_second": 19.727,
      "step": 18800
    },
    {
      "epoch": 4.49,
      "learning_rate": 1e-05,
      "loss": 1.474,
      "step": 18810
    },
    {
      "epoch": 4.49,
      "learning_rate": 1e-05,
      "loss": 1.4498,
      "step": 18820
    },
    {
      "epoch": 4.5,
      "learning_rate": 1e-05,
      "loss": 1.4717,
      "step": 18830
    },
    {
      "epoch": 4.5,
      "learning_rate": 1e-05,
      "loss": 1.4607,
      "step": 18840
    },
    {
      "epoch": 4.5,
      "learning_rate": 1e-05,
      "loss": 1.4716,
      "step": 18850
    },
    {
      "epoch": 4.5,
      "learning_rate": 1e-05,
      "loss": 1.4706,
      "step": 18860
    },
    {
      "epoch": 4.51,
      "learning_rate": 1e-05,
      "loss": 1.4576,
      "step": 18870
    },
    {
      "epoch": 4.51,
      "learning_rate": 1e-05,
      "loss": 1.4669,
      "step": 18880
    },
    {
      "epoch": 4.51,
      "learning_rate": 1e-05,
      "loss": 1.4623,
      "step": 18890
    },
    {
      "epoch": 4.51,
      "learning_rate": 1e-05,
      "loss": 1.4655,
      "step": 18900
    },
    {
      "epoch": 4.51,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.450200080871582,
      "eval_runtime": 34.405,
      "eval_samples_per_second": 19.881,
      "step": 18900
    },
    {
      "epoch": 4.52,
      "learning_rate": 1e-05,
      "loss": 1.4519,
      "step": 18910
    },
    {
      "epoch": 4.52,
      "learning_rate": 1e-05,
      "loss": 1.4578,
      "step": 18920
    },
    {
      "epoch": 4.52,
      "learning_rate": 1e-05,
      "loss": 1.4676,
      "step": 18930
    },
    {
      "epoch": 4.52,
      "learning_rate": 1e-05,
      "loss": 1.4574,
      "step": 18940
    },
    {
      "epoch": 4.53,
      "learning_rate": 1e-05,
      "loss": 1.4627,
      "step": 18950
    },
    {
      "epoch": 4.53,
      "learning_rate": 1e-05,
      "loss": 1.4694,
      "step": 18960
    },
    {
      "epoch": 4.53,
      "learning_rate": 1e-05,
      "loss": 1.4823,
      "step": 18970
    },
    {
      "epoch": 4.53,
      "learning_rate": 1e-05,
      "loss": 1.4538,
      "step": 18980
    },
    {
      "epoch": 4.54,
      "learning_rate": 1e-05,
      "loss": 1.4626,
      "step": 18990
    },
    {
      "epoch": 4.54,
      "learning_rate": 1e-05,
      "loss": 1.4647,
      "step": 19000
    },
    {
      "epoch": 4.54,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.449528694152832,
      "eval_runtime": 34.5612,
      "eval_samples_per_second": 19.791,
      "step": 19000
    },
    {
      "epoch": 4.54,
      "learning_rate": 1e-05,
      "loss": 1.4664,
      "step": 19010
    },
    {
      "epoch": 4.54,
      "learning_rate": 1e-05,
      "loss": 1.5007,
      "step": 19020
    },
    {
      "epoch": 4.55,
      "learning_rate": 1e-05,
      "loss": 1.4548,
      "step": 19030
    },
    {
      "epoch": 4.55,
      "learning_rate": 1e-05,
      "loss": 1.4633,
      "step": 19040
    },
    {
      "epoch": 4.55,
      "learning_rate": 1e-05,
      "loss": 1.4587,
      "step": 19050
    },
    {
      "epoch": 4.55,
      "learning_rate": 1e-05,
      "loss": 1.4703,
      "step": 19060
    },
    {
      "epoch": 4.55,
      "learning_rate": 1e-05,
      "loss": 1.4685,
      "step": 19070
    },
    {
      "epoch": 4.56,
      "learning_rate": 1e-05,
      "loss": 1.4597,
      "step": 19080
    },
    {
      "epoch": 4.56,
      "learning_rate": 1e-05,
      "loss": 1.4526,
      "step": 19090
    },
    {
      "epoch": 4.56,
      "learning_rate": 1e-05,
      "loss": 1.4544,
      "step": 19100
    },
    {
      "epoch": 4.56,
      "eval_accuracy": 0.9722222222222222,
      "eval_loss": 1.4472013711929321,
      "eval_runtime": 34.0584,
      "eval_samples_per_second": 20.083,
      "step": 19100
    },
    {
      "epoch": 4.56,
      "learning_rate": 1e-05,
      "loss": 1.4773,
      "step": 19110
    },
    {
      "epoch": 4.57,
      "learning_rate": 1e-05,
      "loss": 1.4617,
      "step": 19120
    },
    {
      "epoch": 4.57,
      "learning_rate": 1e-05,
      "loss": 1.4616,
      "step": 19130
    },
    {
      "epoch": 4.57,
      "learning_rate": 1e-05,
      "loss": 1.4556,
      "step": 19140
    },
    {
      "epoch": 4.57,
      "learning_rate": 1e-05,
      "loss": 1.463,
      "step": 19150
    },
    {
      "epoch": 4.58,
      "learning_rate": 1e-05,
      "loss": 1.4611,
      "step": 19160
    },
    {
      "epoch": 4.58,
      "learning_rate": 1e-05,
      "loss": 1.4722,
      "step": 19170
    },
    {
      "epoch": 4.58,
      "learning_rate": 1e-05,
      "loss": 1.4893,
      "step": 19180
    },
    {
      "epoch": 4.58,
      "learning_rate": 1e-05,
      "loss": 1.4544,
      "step": 19190
    },
    {
      "epoch": 4.59,
      "learning_rate": 1e-05,
      "loss": 1.4595,
      "step": 19200
    },
    {
      "epoch": 4.59,
      "eval_accuracy": 0.9678362573099415,
      "eval_loss": 1.4480645656585693,
      "eval_runtime": 34.672,
      "eval_samples_per_second": 19.728,
      "step": 19200
    },
    {
      "epoch": 4.59,
      "learning_rate": 1e-05,
      "loss": 1.4746,
      "step": 19210
    },
    {
      "epoch": 4.59,
      "learning_rate": 1e-05,
      "loss": 1.4459,
      "step": 19220
    },
    {
      "epoch": 4.59,
      "learning_rate": 1e-05,
      "loss": 1.4737,
      "step": 19230
    },
    {
      "epoch": 4.6,
      "learning_rate": 1e-05,
      "loss": 1.4601,
      "step": 19240
    },
    {
      "epoch": 4.6,
      "learning_rate": 1e-05,
      "loss": 1.4588,
      "step": 19250
    },
    {
      "epoch": 4.6,
      "learning_rate": 1e-05,
      "loss": 1.4604,
      "step": 19260
    },
    {
      "epoch": 4.6,
      "learning_rate": 1e-05,
      "loss": 1.4683,
      "step": 19270
    },
    {
      "epoch": 4.6,
      "learning_rate": 1e-05,
      "loss": 1.4683,
      "step": 19280
    },
    {
      "epoch": 4.61,
      "learning_rate": 1e-05,
      "loss": 1.4653,
      "step": 19290
    },
    {
      "epoch": 4.61,
      "learning_rate": 1e-05,
      "loss": 1.4532,
      "step": 19300
    },
    {
      "epoch": 4.61,
      "eval_accuracy": 0.9722222222222222,
      "eval_loss": 1.4475253820419312,
      "eval_runtime": 34.5731,
      "eval_samples_per_second": 19.784,
      "step": 19300
    },
    {
      "epoch": 4.61,
      "learning_rate": 1e-05,
      "loss": 1.4602,
      "step": 19310
    },
    {
      "epoch": 4.61,
      "learning_rate": 1e-05,
      "loss": 1.4574,
      "step": 19320
    },
    {
      "epoch": 4.62,
      "learning_rate": 1e-05,
      "loss": 1.453,
      "step": 19330
    },
    {
      "epoch": 4.62,
      "learning_rate": 1e-05,
      "loss": 1.4635,
      "step": 19340
    },
    {
      "epoch": 4.62,
      "learning_rate": 1e-05,
      "loss": 1.474,
      "step": 19350
    },
    {
      "epoch": 4.62,
      "learning_rate": 1e-05,
      "loss": 1.4745,
      "step": 19360
    },
    {
      "epoch": 4.63,
      "learning_rate": 1e-05,
      "loss": 1.4703,
      "step": 19370
    },
    {
      "epoch": 4.63,
      "learning_rate": 1e-05,
      "loss": 1.449,
      "step": 19380
    },
    {
      "epoch": 4.63,
      "learning_rate": 1e-05,
      "loss": 1.4578,
      "step": 19390
    },
    {
      "epoch": 4.63,
      "learning_rate": 1e-05,
      "loss": 1.4678,
      "step": 19400
    },
    {
      "epoch": 4.63,
      "eval_accuracy": 0.9736842105263158,
      "eval_loss": 1.4481322765350342,
      "eval_runtime": 34.5509,
      "eval_samples_per_second": 19.797,
      "step": 19400
    },
    {
      "epoch": 4.64,
      "learning_rate": 1e-05,
      "loss": 1.4755,
      "step": 19410
    },
    {
      "epoch": 4.64,
      "learning_rate": 1e-05,
      "loss": 1.4612,
      "step": 19420
    },
    {
      "epoch": 4.64,
      "learning_rate": 1e-05,
      "loss": 1.466,
      "step": 19430
    },
    {
      "epoch": 4.64,
      "learning_rate": 1e-05,
      "loss": 1.4696,
      "step": 19440
    },
    {
      "epoch": 4.65,
      "learning_rate": 1e-05,
      "loss": 1.4528,
      "step": 19450
    },
    {
      "epoch": 4.65,
      "learning_rate": 1e-05,
      "loss": 1.4608,
      "step": 19460
    },
    {
      "epoch": 4.65,
      "learning_rate": 1e-05,
      "loss": 1.4596,
      "step": 19470
    },
    {
      "epoch": 4.65,
      "learning_rate": 1e-05,
      "loss": 1.4558,
      "step": 19480
    },
    {
      "epoch": 4.65,
      "learning_rate": 1e-05,
      "loss": 1.4566,
      "step": 19490
    },
    {
      "epoch": 4.66,
      "learning_rate": 1e-05,
      "loss": 1.461,
      "step": 19500
    },
    {
      "epoch": 4.66,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.4471712112426758,
      "eval_runtime": 34.5202,
      "eval_samples_per_second": 19.814,
      "step": 19500
    },
    {
      "epoch": 4.66,
      "learning_rate": 1e-05,
      "loss": 1.4597,
      "step": 19510
    },
    {
      "epoch": 4.66,
      "learning_rate": 1e-05,
      "loss": 1.4684,
      "step": 19520
    },
    {
      "epoch": 4.66,
      "learning_rate": 1e-05,
      "loss": 1.4748,
      "step": 19530
    },
    {
      "epoch": 4.67,
      "learning_rate": 1e-05,
      "loss": 1.4629,
      "step": 19540
    },
    {
      "epoch": 4.67,
      "learning_rate": 1e-05,
      "loss": 1.4978,
      "step": 19550
    },
    {
      "epoch": 4.67,
      "learning_rate": 1e-05,
      "loss": 1.4544,
      "step": 19560
    },
    {
      "epoch": 4.67,
      "learning_rate": 1e-05,
      "loss": 1.4617,
      "step": 19570
    },
    {
      "epoch": 4.68,
      "learning_rate": 1e-05,
      "loss": 1.4689,
      "step": 19580
    },
    {
      "epoch": 4.68,
      "learning_rate": 1e-05,
      "loss": 1.4791,
      "step": 19590
    },
    {
      "epoch": 4.68,
      "learning_rate": 1e-05,
      "loss": 1.4583,
      "step": 19600
    },
    {
      "epoch": 4.68,
      "eval_accuracy": 0.9736842105263158,
      "eval_loss": 1.4475111961364746,
      "eval_runtime": 34.8916,
      "eval_samples_per_second": 19.604,
      "step": 19600
    },
    {
      "epoch": 4.68,
      "learning_rate": 1e-05,
      "loss": 1.459,
      "step": 19610
    },
    {
      "epoch": 4.69,
      "learning_rate": 1e-05,
      "loss": 1.4543,
      "step": 19620
    },
    {
      "epoch": 4.69,
      "learning_rate": 1e-05,
      "loss": 1.4491,
      "step": 19630
    },
    {
      "epoch": 4.69,
      "learning_rate": 1e-05,
      "loss": 1.4677,
      "step": 19640
    },
    {
      "epoch": 4.69,
      "learning_rate": 1e-05,
      "loss": 1.4436,
      "step": 19650
    },
    {
      "epoch": 4.7,
      "learning_rate": 1e-05,
      "loss": 1.4596,
      "step": 19660
    },
    {
      "epoch": 4.7,
      "learning_rate": 1e-05,
      "loss": 1.458,
      "step": 19670
    },
    {
      "epoch": 4.7,
      "learning_rate": 1e-05,
      "loss": 1.4729,
      "step": 19680
    },
    {
      "epoch": 4.7,
      "learning_rate": 1e-05,
      "loss": 1.4575,
      "step": 19690
    },
    {
      "epoch": 4.71,
      "learning_rate": 1e-05,
      "loss": 1.4655,
      "step": 19700
    },
    {
      "epoch": 4.71,
      "eval_accuracy": 0.9722222222222222,
      "eval_loss": 1.4485477209091187,
      "eval_runtime": 35.9019,
      "eval_samples_per_second": 19.052,
      "step": 19700
    },
    {
      "epoch": 4.71,
      "learning_rate": 1e-05,
      "loss": 1.4523,
      "step": 19710
    },
    {
      "epoch": 4.71,
      "learning_rate": 1e-05,
      "loss": 1.4546,
      "step": 19720
    },
    {
      "epoch": 4.71,
      "learning_rate": 1e-05,
      "loss": 1.4665,
      "step": 19730
    },
    {
      "epoch": 4.71,
      "learning_rate": 1e-05,
      "loss": 1.4525,
      "step": 19740
    },
    {
      "epoch": 4.72,
      "learning_rate": 1e-05,
      "loss": 1.4639,
      "step": 19750
    },
    {
      "epoch": 4.72,
      "learning_rate": 1e-05,
      "loss": 1.4637,
      "step": 19760
    },
    {
      "epoch": 4.72,
      "learning_rate": 1e-05,
      "loss": 1.4473,
      "step": 19770
    },
    {
      "epoch": 4.72,
      "learning_rate": 1e-05,
      "loss": 1.4634,
      "step": 19780
    },
    {
      "epoch": 4.73,
      "learning_rate": 1e-05,
      "loss": 1.4548,
      "step": 19790
    },
    {
      "epoch": 4.73,
      "learning_rate": 1e-05,
      "loss": 1.463,
      "step": 19800
    },
    {
      "epoch": 4.73,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.447371244430542,
      "eval_runtime": 34.5737,
      "eval_samples_per_second": 19.784,
      "step": 19800
    },
    {
      "epoch": 4.73,
      "learning_rate": 1e-05,
      "loss": 1.4586,
      "step": 19810
    },
    {
      "epoch": 4.73,
      "learning_rate": 1e-05,
      "loss": 1.4566,
      "step": 19820
    },
    {
      "epoch": 4.74,
      "learning_rate": 1e-05,
      "loss": 1.4675,
      "step": 19830
    },
    {
      "epoch": 4.74,
      "learning_rate": 1e-05,
      "loss": 1.4632,
      "step": 19840
    },
    {
      "epoch": 4.74,
      "learning_rate": 1e-05,
      "loss": 1.4763,
      "step": 19850
    },
    {
      "epoch": 4.74,
      "learning_rate": 1e-05,
      "loss": 1.467,
      "step": 19860
    },
    {
      "epoch": 4.75,
      "learning_rate": 1e-05,
      "loss": 1.4623,
      "step": 19870
    },
    {
      "epoch": 4.75,
      "learning_rate": 1e-05,
      "loss": 1.4564,
      "step": 19880
    },
    {
      "epoch": 4.75,
      "learning_rate": 1e-05,
      "loss": 1.4621,
      "step": 19890
    },
    {
      "epoch": 4.75,
      "learning_rate": 1e-05,
      "loss": 1.4535,
      "step": 19900
    },
    {
      "epoch": 4.75,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.446195363998413,
      "eval_runtime": 34.1379,
      "eval_samples_per_second": 20.036,
      "step": 19900
    },
    {
      "epoch": 4.76,
      "learning_rate": 1e-05,
      "loss": 1.4522,
      "step": 19910
    },
    {
      "epoch": 4.76,
      "learning_rate": 1e-05,
      "loss": 1.4568,
      "step": 19920
    },
    {
      "epoch": 4.76,
      "learning_rate": 1e-05,
      "loss": 1.4595,
      "step": 19930
    },
    {
      "epoch": 4.76,
      "learning_rate": 1e-05,
      "loss": 1.4611,
      "step": 19940
    },
    {
      "epoch": 4.76,
      "learning_rate": 1e-05,
      "loss": 1.4694,
      "step": 19950
    },
    {
      "epoch": 4.77,
      "learning_rate": 1e-05,
      "loss": 1.4557,
      "step": 19960
    },
    {
      "epoch": 4.77,
      "learning_rate": 1e-05,
      "loss": 1.4586,
      "step": 19970
    },
    {
      "epoch": 4.77,
      "learning_rate": 1e-05,
      "loss": 1.4601,
      "step": 19980
    },
    {
      "epoch": 4.77,
      "learning_rate": 1e-05,
      "loss": 1.4508,
      "step": 19990
    },
    {
      "epoch": 4.78,
      "learning_rate": 1e-05,
      "loss": 1.4474,
      "step": 20000
    },
    {
      "epoch": 4.78,
      "eval_accuracy": 0.9722222222222222,
      "eval_loss": 1.4467116594314575,
      "eval_runtime": 34.6522,
      "eval_samples_per_second": 19.739,
      "step": 20000
    },
    {
      "epoch": 4.78,
      "learning_rate": 1e-05,
      "loss": 1.4826,
      "step": 20010
    },
    {
      "epoch": 4.78,
      "learning_rate": 1e-05,
      "loss": 1.46,
      "step": 20020
    },
    {
      "epoch": 4.78,
      "learning_rate": 1e-05,
      "loss": 1.4599,
      "step": 20030
    },
    {
      "epoch": 4.79,
      "learning_rate": 1e-05,
      "loss": 1.4671,
      "step": 20040
    },
    {
      "epoch": 4.79,
      "learning_rate": 1e-05,
      "loss": 1.4792,
      "step": 20050
    },
    {
      "epoch": 4.79,
      "learning_rate": 1e-05,
      "loss": 1.4629,
      "step": 20060
    },
    {
      "epoch": 4.79,
      "learning_rate": 1e-05,
      "loss": 1.4787,
      "step": 20070
    },
    {
      "epoch": 4.8,
      "learning_rate": 1e-05,
      "loss": 1.4654,
      "step": 20080
    },
    {
      "epoch": 4.8,
      "learning_rate": 1e-05,
      "loss": 1.4632,
      "step": 20090
    },
    {
      "epoch": 4.8,
      "learning_rate": 1e-05,
      "loss": 1.4528,
      "step": 20100
    },
    {
      "epoch": 4.8,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.4465885162353516,
      "eval_runtime": 34.2255,
      "eval_samples_per_second": 19.985,
      "step": 20100
    },
    {
      "epoch": 4.8,
      "learning_rate": 1e-05,
      "loss": 1.4578,
      "step": 20110
    },
    {
      "epoch": 4.81,
      "learning_rate": 1e-05,
      "loss": 1.4632,
      "step": 20120
    },
    {
      "epoch": 4.81,
      "learning_rate": 1e-05,
      "loss": 1.4639,
      "step": 20130
    },
    {
      "epoch": 4.81,
      "learning_rate": 1e-05,
      "loss": 1.4629,
      "step": 20140
    },
    {
      "epoch": 4.81,
      "learning_rate": 1e-05,
      "loss": 1.4551,
      "step": 20150
    },
    {
      "epoch": 4.81,
      "learning_rate": 1e-05,
      "loss": 1.4705,
      "step": 20160
    },
    {
      "epoch": 4.82,
      "learning_rate": 1e-05,
      "loss": 1.4491,
      "step": 20170
    },
    {
      "epoch": 4.82,
      "learning_rate": 1e-05,
      "loss": 1.4546,
      "step": 20180
    },
    {
      "epoch": 4.82,
      "learning_rate": 1e-05,
      "loss": 1.4585,
      "step": 20190
    },
    {
      "epoch": 4.82,
      "learning_rate": 1e-05,
      "loss": 1.4651,
      "step": 20200
    },
    {
      "epoch": 4.82,
      "eval_accuracy": 0.9751461988304093,
      "eval_loss": 1.4467846155166626,
      "eval_runtime": 34.4003,
      "eval_samples_per_second": 19.884,
      "step": 20200
    },
    {
      "epoch": 4.83,
      "learning_rate": 1e-05,
      "loss": 1.4553,
      "step": 20210
    },
    {
      "epoch": 4.83,
      "learning_rate": 1e-05,
      "loss": 1.457,
      "step": 20220
    },
    {
      "epoch": 4.83,
      "learning_rate": 1e-05,
      "loss": 1.4633,
      "step": 20230
    },
    {
      "epoch": 4.83,
      "learning_rate": 1e-05,
      "loss": 1.4566,
      "step": 20240
    },
    {
      "epoch": 4.84,
      "learning_rate": 1e-05,
      "loss": 1.4604,
      "step": 20250
    },
    {
      "epoch": 4.84,
      "learning_rate": 1e-05,
      "loss": 1.4749,
      "step": 20260
    },
    {
      "epoch": 4.84,
      "learning_rate": 1e-05,
      "loss": 1.464,
      "step": 20270
    },
    {
      "epoch": 4.84,
      "learning_rate": 1e-05,
      "loss": 1.4604,
      "step": 20280
    },
    {
      "epoch": 4.85,
      "learning_rate": 1e-05,
      "loss": 1.4631,
      "step": 20290
    },
    {
      "epoch": 4.85,
      "learning_rate": 1e-05,
      "loss": 1.4596,
      "step": 20300
    },
    {
      "epoch": 4.85,
      "eval_accuracy": 0.9678362573099415,
      "eval_loss": 1.4461092948913574,
      "eval_runtime": 34.5405,
      "eval_samples_per_second": 19.803,
      "step": 20300
    },
    {
      "epoch": 4.85,
      "learning_rate": 1e-05,
      "loss": 1.4681,
      "step": 20310
    },
    {
      "epoch": 4.85,
      "learning_rate": 1e-05,
      "loss": 1.4748,
      "step": 20320
    },
    {
      "epoch": 4.86,
      "learning_rate": 1e-05,
      "loss": 1.471,
      "step": 20330
    },
    {
      "epoch": 4.86,
      "learning_rate": 1e-05,
      "loss": 1.4605,
      "step": 20340
    },
    {
      "epoch": 4.86,
      "learning_rate": 1e-05,
      "loss": 1.4654,
      "step": 20350
    },
    {
      "epoch": 4.86,
      "learning_rate": 1e-05,
      "loss": 1.4652,
      "step": 20360
    },
    {
      "epoch": 4.87,
      "learning_rate": 1e-05,
      "loss": 1.4708,
      "step": 20370
    },
    {
      "epoch": 4.87,
      "learning_rate": 1e-05,
      "loss": 1.467,
      "step": 20380
    },
    {
      "epoch": 4.87,
      "learning_rate": 1e-05,
      "loss": 1.4574,
      "step": 20390
    },
    {
      "epoch": 4.87,
      "learning_rate": 1e-05,
      "loss": 1.4644,
      "step": 20400
    },
    {
      "epoch": 4.87,
      "eval_accuracy": 0.9736842105263158,
      "eval_loss": 1.4442626237869263,
      "eval_runtime": 34.5692,
      "eval_samples_per_second": 19.786,
      "step": 20400
    },
    {
      "epoch": 4.87,
      "learning_rate": 1e-05,
      "loss": 1.4613,
      "step": 20410
    },
    {
      "epoch": 4.88,
      "learning_rate": 1e-05,
      "loss": 1.4664,
      "step": 20420
    },
    {
      "epoch": 4.88,
      "learning_rate": 1e-05,
      "loss": 1.4648,
      "step": 20430
    },
    {
      "epoch": 4.88,
      "learning_rate": 1e-05,
      "loss": 1.4554,
      "step": 20440
    },
    {
      "epoch": 4.88,
      "learning_rate": 1e-05,
      "loss": 1.4568,
      "step": 20450
    },
    {
      "epoch": 4.89,
      "learning_rate": 1e-05,
      "loss": 1.4496,
      "step": 20460
    },
    {
      "epoch": 4.89,
      "learning_rate": 1e-05,
      "loss": 1.4579,
      "step": 20470
    },
    {
      "epoch": 4.89,
      "learning_rate": 1e-05,
      "loss": 1.4544,
      "step": 20480
    },
    {
      "epoch": 4.89,
      "learning_rate": 1e-05,
      "loss": 1.4653,
      "step": 20490
    },
    {
      "epoch": 4.9,
      "learning_rate": 1e-05,
      "loss": 1.4531,
      "step": 20500
    },
    {
      "epoch": 4.9,
      "eval_accuracy": 0.9722222222222222,
      "eval_loss": 1.4470142126083374,
      "eval_runtime": 34.5079,
      "eval_samples_per_second": 19.822,
      "step": 20500
    },
    {
      "epoch": 4.9,
      "learning_rate": 1e-05,
      "loss": 1.4596,
      "step": 20510
    },
    {
      "epoch": 4.9,
      "learning_rate": 1e-05,
      "loss": 1.4459,
      "step": 20520
    },
    {
      "epoch": 4.9,
      "learning_rate": 1e-05,
      "loss": 1.4646,
      "step": 20530
    },
    {
      "epoch": 4.91,
      "learning_rate": 1e-05,
      "loss": 1.4498,
      "step": 20540
    },
    {
      "epoch": 4.91,
      "learning_rate": 1e-05,
      "loss": 1.4631,
      "step": 20550
    },
    {
      "epoch": 4.91,
      "learning_rate": 1e-05,
      "loss": 1.4782,
      "step": 20560
    },
    {
      "epoch": 4.91,
      "learning_rate": 1e-05,
      "loss": 1.4637,
      "step": 20570
    },
    {
      "epoch": 4.92,
      "learning_rate": 1e-05,
      "loss": 1.4666,
      "step": 20580
    },
    {
      "epoch": 4.92,
      "learning_rate": 1e-05,
      "loss": 1.4546,
      "step": 20590
    },
    {
      "epoch": 4.92,
      "learning_rate": 1e-05,
      "loss": 1.4572,
      "step": 20600
    },
    {
      "epoch": 4.92,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.445745587348938,
      "eval_runtime": 34.4427,
      "eval_samples_per_second": 19.859,
      "step": 20600
    },
    {
      "epoch": 4.92,
      "learning_rate": 1e-05,
      "loss": 1.4497,
      "step": 20610
    },
    {
      "epoch": 4.92,
      "learning_rate": 1e-05,
      "loss": 1.4679,
      "step": 20620
    },
    {
      "epoch": 4.93,
      "learning_rate": 1e-05,
      "loss": 1.4547,
      "step": 20630
    },
    {
      "epoch": 4.93,
      "learning_rate": 1e-05,
      "loss": 1.4715,
      "step": 20640
    },
    {
      "epoch": 4.93,
      "learning_rate": 1e-05,
      "loss": 1.4836,
      "step": 20650
    },
    {
      "epoch": 4.93,
      "learning_rate": 1e-05,
      "loss": 1.4526,
      "step": 20660
    },
    {
      "epoch": 4.94,
      "learning_rate": 1e-05,
      "loss": 1.4544,
      "step": 20670
    },
    {
      "epoch": 4.94,
      "learning_rate": 1e-05,
      "loss": 1.446,
      "step": 20680
    },
    {
      "epoch": 4.94,
      "learning_rate": 1e-05,
      "loss": 1.4575,
      "step": 20690
    },
    {
      "epoch": 4.94,
      "learning_rate": 1e-05,
      "loss": 1.4685,
      "step": 20700
    },
    {
      "epoch": 4.94,
      "eval_accuracy": 0.9722222222222222,
      "eval_loss": 1.4458080530166626,
      "eval_runtime": 34.5918,
      "eval_samples_per_second": 19.773,
      "step": 20700
    },
    {
      "epoch": 4.95,
      "learning_rate": 1e-05,
      "loss": 1.4458,
      "step": 20710
    },
    {
      "epoch": 4.95,
      "learning_rate": 1e-05,
      "loss": 1.4477,
      "step": 20720
    },
    {
      "epoch": 4.95,
      "learning_rate": 1e-05,
      "loss": 1.4902,
      "step": 20730
    },
    {
      "epoch": 4.95,
      "learning_rate": 1e-05,
      "loss": 1.4511,
      "step": 20740
    },
    {
      "epoch": 4.96,
      "learning_rate": 1e-05,
      "loss": 1.4608,
      "step": 20750
    },
    {
      "epoch": 4.96,
      "learning_rate": 1e-05,
      "loss": 1.4567,
      "step": 20760
    },
    {
      "epoch": 4.96,
      "learning_rate": 1e-05,
      "loss": 1.4545,
      "step": 20770
    },
    {
      "epoch": 4.96,
      "learning_rate": 1e-05,
      "loss": 1.4538,
      "step": 20780
    },
    {
      "epoch": 4.97,
      "learning_rate": 1e-05,
      "loss": 1.4547,
      "step": 20790
    },
    {
      "epoch": 4.97,
      "learning_rate": 1e-05,
      "loss": 1.4594,
      "step": 20800
    },
    {
      "epoch": 4.97,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.4443880319595337,
      "eval_runtime": 34.4442,
      "eval_samples_per_second": 19.858,
      "step": 20800
    },
    {
      "epoch": 4.97,
      "learning_rate": 1e-05,
      "loss": 1.4635,
      "step": 20810
    },
    {
      "epoch": 4.97,
      "learning_rate": 1e-05,
      "loss": 1.4575,
      "step": 20820
    },
    {
      "epoch": 4.97,
      "learning_rate": 1e-05,
      "loss": 1.4583,
      "step": 20830
    },
    {
      "epoch": 4.98,
      "learning_rate": 1e-05,
      "loss": 1.4492,
      "step": 20840
    },
    {
      "epoch": 4.98,
      "learning_rate": 1e-05,
      "loss": 1.4568,
      "step": 20850
    },
    {
      "epoch": 4.98,
      "learning_rate": 1e-05,
      "loss": 1.4596,
      "step": 20860
    },
    {
      "epoch": 4.98,
      "learning_rate": 1e-05,
      "loss": 1.4596,
      "step": 20870
    },
    {
      "epoch": 4.99,
      "learning_rate": 1e-05,
      "loss": 1.4618,
      "step": 20880
    },
    {
      "epoch": 4.99,
      "learning_rate": 1e-05,
      "loss": 1.455,
      "step": 20890
    },
    {
      "epoch": 4.99,
      "learning_rate": 1e-05,
      "loss": 1.465,
      "step": 20900
    },
    {
      "epoch": 4.99,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.446098804473877,
      "eval_runtime": 34.3969,
      "eval_samples_per_second": 19.886,
      "step": 20900
    },
    {
      "epoch": 4.99,
      "learning_rate": 1e-05,
      "loss": 1.4487,
      "step": 20910
    },
    {
      "epoch": 5.0,
      "learning_rate": 1e-05,
      "loss": 1.4549,
      "step": 20920
    },
    {
      "epoch": 5.0,
      "learning_rate": 1e-05,
      "loss": 1.4523,
      "step": 20930
    },
    {
      "epoch": 5.0,
      "learning_rate": 1e-05,
      "loss": 1.4458,
      "step": 20940
    },
    {
      "epoch": 5.0,
      "learning_rate": 1e-05,
      "loss": 1.4507,
      "step": 20950
    },
    {
      "epoch": 5.01,
      "learning_rate": 1e-05,
      "loss": 1.4609,
      "step": 20960
    },
    {
      "epoch": 5.01,
      "learning_rate": 1e-05,
      "loss": 1.4482,
      "step": 20970
    },
    {
      "epoch": 5.01,
      "learning_rate": 1e-05,
      "loss": 1.4569,
      "step": 20980
    },
    {
      "epoch": 5.01,
      "learning_rate": 1e-05,
      "loss": 1.4417,
      "step": 20990
    },
    {
      "epoch": 5.02,
      "learning_rate": 1e-05,
      "loss": 1.458,
      "step": 21000
    },
    {
      "epoch": 5.02,
      "eval_accuracy": 0.9736842105263158,
      "eval_loss": 1.4458869695663452,
      "eval_runtime": 34.5972,
      "eval_samples_per_second": 19.77,
      "step": 21000
    },
    {
      "epoch": 5.02,
      "learning_rate": 1e-05,
      "loss": 1.4588,
      "step": 21010
    },
    {
      "epoch": 5.02,
      "learning_rate": 1e-05,
      "loss": 1.4554,
      "step": 21020
    },
    {
      "epoch": 5.02,
      "learning_rate": 1e-05,
      "loss": 1.4634,
      "step": 21030
    },
    {
      "epoch": 5.03,
      "learning_rate": 1e-05,
      "loss": 1.4516,
      "step": 21040
    },
    {
      "epoch": 5.03,
      "learning_rate": 1e-05,
      "loss": 1.4544,
      "step": 21050
    },
    {
      "epoch": 5.03,
      "learning_rate": 1e-05,
      "loss": 1.469,
      "step": 21060
    },
    {
      "epoch": 5.03,
      "learning_rate": 1e-05,
      "loss": 1.4575,
      "step": 21070
    },
    {
      "epoch": 5.03,
      "learning_rate": 1e-05,
      "loss": 1.4562,
      "step": 21080
    },
    {
      "epoch": 5.04,
      "learning_rate": 1e-05,
      "loss": 1.4507,
      "step": 21090
    },
    {
      "epoch": 5.04,
      "learning_rate": 1e-05,
      "loss": 1.4608,
      "step": 21100
    },
    {
      "epoch": 5.04,
      "eval_accuracy": 0.9736842105263158,
      "eval_loss": 1.4431785345077515,
      "eval_runtime": 34.5696,
      "eval_samples_per_second": 19.786,
      "step": 21100
    },
    {
      "epoch": 5.04,
      "learning_rate": 1e-05,
      "loss": 1.4562,
      "step": 21110
    },
    {
      "epoch": 5.04,
      "learning_rate": 1e-05,
      "loss": 1.4591,
      "step": 21120
    },
    {
      "epoch": 5.05,
      "learning_rate": 1e-05,
      "loss": 1.4408,
      "step": 21130
    },
    {
      "epoch": 5.05,
      "learning_rate": 1e-05,
      "loss": 1.4455,
      "step": 21140
    },
    {
      "epoch": 5.05,
      "learning_rate": 1e-05,
      "loss": 1.4427,
      "step": 21150
    },
    {
      "epoch": 5.05,
      "learning_rate": 1e-05,
      "loss": 1.4456,
      "step": 21160
    },
    {
      "epoch": 5.06,
      "learning_rate": 1e-05,
      "loss": 1.4483,
      "step": 21170
    },
    {
      "epoch": 5.06,
      "learning_rate": 1e-05,
      "loss": 1.4483,
      "step": 21180
    },
    {
      "epoch": 5.06,
      "learning_rate": 1e-05,
      "loss": 1.4728,
      "step": 21190
    },
    {
      "epoch": 5.06,
      "learning_rate": 1e-05,
      "loss": 1.4484,
      "step": 21200
    },
    {
      "epoch": 5.06,
      "eval_accuracy": 0.9751461988304093,
      "eval_loss": 1.4455170631408691,
      "eval_runtime": 34.3062,
      "eval_samples_per_second": 19.938,
      "step": 21200
    },
    {
      "epoch": 5.07,
      "learning_rate": 1e-05,
      "loss": 1.4464,
      "step": 21210
    },
    {
      "epoch": 5.07,
      "learning_rate": 1e-05,
      "loss": 1.4545,
      "step": 21220
    },
    {
      "epoch": 5.07,
      "learning_rate": 1e-05,
      "loss": 1.4719,
      "step": 21230
    },
    {
      "epoch": 5.07,
      "learning_rate": 1e-05,
      "loss": 1.4453,
      "step": 21240
    },
    {
      "epoch": 5.08,
      "learning_rate": 1e-05,
      "loss": 1.456,
      "step": 21250
    },
    {
      "epoch": 5.08,
      "learning_rate": 1e-05,
      "loss": 1.4577,
      "step": 21260
    },
    {
      "epoch": 5.08,
      "learning_rate": 1e-05,
      "loss": 1.4492,
      "step": 21270
    },
    {
      "epoch": 5.08,
      "learning_rate": 1e-05,
      "loss": 1.4677,
      "step": 21280
    },
    {
      "epoch": 5.08,
      "learning_rate": 1e-05,
      "loss": 1.4545,
      "step": 21290
    },
    {
      "epoch": 5.09,
      "learning_rate": 1e-05,
      "loss": 1.4624,
      "step": 21300
    },
    {
      "epoch": 5.09,
      "eval_accuracy": 0.9722222222222222,
      "eval_loss": 1.4430495500564575,
      "eval_runtime": 34.526,
      "eval_samples_per_second": 19.811,
      "step": 21300
    },
    {
      "epoch": 5.09,
      "learning_rate": 1e-05,
      "loss": 1.4591,
      "step": 21310
    },
    {
      "epoch": 5.09,
      "learning_rate": 1e-05,
      "loss": 1.4538,
      "step": 21320
    },
    {
      "epoch": 5.09,
      "learning_rate": 1e-05,
      "loss": 1.4539,
      "step": 21330
    },
    {
      "epoch": 5.1,
      "learning_rate": 1e-05,
      "loss": 1.4454,
      "step": 21340
    },
    {
      "epoch": 5.1,
      "learning_rate": 1e-05,
      "loss": 1.4574,
      "step": 21350
    },
    {
      "epoch": 5.1,
      "learning_rate": 1e-05,
      "loss": 1.4629,
      "step": 21360
    },
    {
      "epoch": 5.1,
      "learning_rate": 1e-05,
      "loss": 1.4544,
      "step": 21370
    },
    {
      "epoch": 5.11,
      "learning_rate": 1e-05,
      "loss": 1.4573,
      "step": 21380
    },
    {
      "epoch": 5.11,
      "learning_rate": 1e-05,
      "loss": 1.4613,
      "step": 21390
    },
    {
      "epoch": 5.11,
      "learning_rate": 1e-05,
      "loss": 1.4518,
      "step": 21400
    },
    {
      "epoch": 5.11,
      "eval_accuracy": 0.9692982456140351,
      "eval_loss": 1.4429206848144531,
      "eval_runtime": 34.27,
      "eval_samples_per_second": 19.959,
      "step": 21400
    },
    {
      "epoch": 5.11,
      "learning_rate": 1e-05,
      "loss": 1.4585,
      "step": 21410
    },
    {
      "epoch": 5.12,
      "learning_rate": 1e-05,
      "loss": 1.4545,
      "step": 21420
    },
    {
      "epoch": 5.12,
      "learning_rate": 1e-05,
      "loss": 1.4534,
      "step": 21430
    },
    {
      "epoch": 5.12,
      "learning_rate": 1e-05,
      "loss": 1.4521,
      "step": 21440
    },
    {
      "epoch": 5.12,
      "learning_rate": 1e-05,
      "loss": 1.4511,
      "step": 21450
    },
    {
      "epoch": 5.13,
      "learning_rate": 1e-05,
      "loss": 1.4443,
      "step": 21460
    },
    {
      "epoch": 5.13,
      "learning_rate": 1e-05,
      "loss": 1.4598,
      "step": 21470
    },
    {
      "epoch": 5.13,
      "learning_rate": 1e-05,
      "loss": 1.4502,
      "step": 21480
    },
    {
      "epoch": 5.13,
      "learning_rate": 1e-05,
      "loss": 1.4764,
      "step": 21490
    },
    {
      "epoch": 5.13,
      "learning_rate": 1e-05,
      "loss": 1.4514,
      "step": 21500
    },
    {
      "epoch": 5.13,
      "eval_accuracy": 0.9722222222222222,
      "eval_loss": 1.443873643875122,
      "eval_runtime": 34.4985,
      "eval_samples_per_second": 19.827,
      "step": 21500
    },
    {
      "epoch": 5.14,
      "learning_rate": 1e-05,
      "loss": 1.47,
      "step": 21510
    },
    {
      "epoch": 5.14,
      "learning_rate": 1e-05,
      "loss": 1.4537,
      "step": 21520
    },
    {
      "epoch": 5.14,
      "learning_rate": 1e-05,
      "loss": 1.4638,
      "step": 21530
    },
    {
      "epoch": 5.14,
      "learning_rate": 1e-05,
      "loss": 1.4419,
      "step": 21540
    },
    {
      "epoch": 5.15,
      "learning_rate": 1e-05,
      "loss": 1.4548,
      "step": 21550
    },
    {
      "epoch": 5.15,
      "learning_rate": 1e-05,
      "loss": 1.4508,
      "step": 21560
    },
    {
      "epoch": 5.15,
      "learning_rate": 1e-05,
      "loss": 1.4581,
      "step": 21570
    },
    {
      "epoch": 5.15,
      "learning_rate": 1e-05,
      "loss": 1.4698,
      "step": 21580
    },
    {
      "epoch": 5.16,
      "learning_rate": 1e-05,
      "loss": 1.4516,
      "step": 21590
    },
    {
      "epoch": 5.16,
      "learning_rate": 1e-05,
      "loss": 1.4542,
      "step": 21600
    },
    {
      "epoch": 5.16,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.4424933195114136,
      "eval_runtime": 34.6782,
      "eval_samples_per_second": 19.724,
      "step": 21600
    },
    {
      "epoch": 5.16,
      "learning_rate": 1e-05,
      "loss": 1.4524,
      "step": 21610
    },
    {
      "epoch": 5.16,
      "learning_rate": 1e-05,
      "loss": 1.4494,
      "step": 21620
    },
    {
      "epoch": 5.17,
      "learning_rate": 1e-05,
      "loss": 1.4476,
      "step": 21630
    },
    {
      "epoch": 5.17,
      "learning_rate": 1e-05,
      "loss": 1.4439,
      "step": 21640
    },
    {
      "epoch": 5.17,
      "learning_rate": 1e-05,
      "loss": 1.4603,
      "step": 21650
    },
    {
      "epoch": 5.17,
      "learning_rate": 1e-05,
      "loss": 1.4529,
      "step": 21660
    },
    {
      "epoch": 5.18,
      "learning_rate": 1e-05,
      "loss": 1.4555,
      "step": 21670
    },
    {
      "epoch": 5.18,
      "learning_rate": 1e-05,
      "loss": 1.4616,
      "step": 21680
    },
    {
      "epoch": 5.18,
      "learning_rate": 1e-05,
      "loss": 1.452,
      "step": 21690
    },
    {
      "epoch": 5.18,
      "learning_rate": 1e-05,
      "loss": 1.4616,
      "step": 21700
    },
    {
      "epoch": 5.18,
      "eval_accuracy": 0.9722222222222222,
      "eval_loss": 1.4428153038024902,
      "eval_runtime": 34.5439,
      "eval_samples_per_second": 19.801,
      "step": 21700
    },
    {
      "epoch": 5.19,
      "learning_rate": 1e-05,
      "loss": 1.4621,
      "step": 21710
    },
    {
      "epoch": 5.19,
      "learning_rate": 1e-05,
      "loss": 1.4586,
      "step": 21720
    },
    {
      "epoch": 5.19,
      "learning_rate": 1e-05,
      "loss": 1.4495,
      "step": 21730
    },
    {
      "epoch": 5.19,
      "learning_rate": 1e-05,
      "loss": 1.459,
      "step": 21740
    },
    {
      "epoch": 5.19,
      "learning_rate": 1e-05,
      "loss": 1.4523,
      "step": 21750
    },
    {
      "epoch": 5.2,
      "learning_rate": 1e-05,
      "loss": 1.4648,
      "step": 21760
    },
    {
      "epoch": 5.2,
      "learning_rate": 1e-05,
      "loss": 1.5084,
      "step": 21770
    },
    {
      "epoch": 5.2,
      "learning_rate": 1e-05,
      "loss": 1.4434,
      "step": 21780
    },
    {
      "epoch": 5.2,
      "learning_rate": 1e-05,
      "loss": 1.4498,
      "step": 21790
    },
    {
      "epoch": 5.21,
      "learning_rate": 1e-05,
      "loss": 1.4661,
      "step": 21800
    },
    {
      "epoch": 5.21,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4446927309036255,
      "eval_runtime": 34.4667,
      "eval_samples_per_second": 19.845,
      "step": 21800
    },
    {
      "epoch": 5.21,
      "learning_rate": 1e-05,
      "loss": 1.4541,
      "step": 21810
    },
    {
      "epoch": 5.21,
      "learning_rate": 1e-05,
      "loss": 1.4493,
      "step": 21820
    },
    {
      "epoch": 5.21,
      "learning_rate": 1e-05,
      "loss": 1.4498,
      "step": 21830
    },
    {
      "epoch": 5.22,
      "learning_rate": 1e-05,
      "loss": 1.4537,
      "step": 21840
    },
    {
      "epoch": 5.22,
      "learning_rate": 1e-05,
      "loss": 1.4524,
      "step": 21850
    },
    {
      "epoch": 5.22,
      "learning_rate": 1e-05,
      "loss": 1.4577,
      "step": 21860
    },
    {
      "epoch": 5.22,
      "learning_rate": 1e-05,
      "loss": 1.4539,
      "step": 21870
    },
    {
      "epoch": 5.23,
      "learning_rate": 1e-05,
      "loss": 1.4449,
      "step": 21880
    },
    {
      "epoch": 5.23,
      "learning_rate": 1e-05,
      "loss": 1.4585,
      "step": 21890
    },
    {
      "epoch": 5.23,
      "learning_rate": 1e-05,
      "loss": 1.4482,
      "step": 21900
    },
    {
      "epoch": 5.23,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4424163103103638,
      "eval_runtime": 34.282,
      "eval_samples_per_second": 19.952,
      "step": 21900
    },
    {
      "epoch": 5.23,
      "learning_rate": 1e-05,
      "loss": 1.4603,
      "step": 21910
    },
    {
      "epoch": 5.24,
      "learning_rate": 1e-05,
      "loss": 1.4567,
      "step": 21920
    },
    {
      "epoch": 5.24,
      "learning_rate": 1e-05,
      "loss": 1.4523,
      "step": 21930
    },
    {
      "epoch": 5.24,
      "learning_rate": 1e-05,
      "loss": 1.449,
      "step": 21940
    },
    {
      "epoch": 5.24,
      "learning_rate": 1e-05,
      "loss": 1.4464,
      "step": 21950
    },
    {
      "epoch": 5.24,
      "learning_rate": 1e-05,
      "loss": 1.4582,
      "step": 21960
    },
    {
      "epoch": 5.25,
      "learning_rate": 1e-05,
      "loss": 1.4727,
      "step": 21970
    },
    {
      "epoch": 5.25,
      "learning_rate": 1e-05,
      "loss": 1.4495,
      "step": 21980
    },
    {
      "epoch": 5.25,
      "learning_rate": 1e-05,
      "loss": 1.4491,
      "step": 21990
    },
    {
      "epoch": 5.25,
      "learning_rate": 1e-05,
      "loss": 1.4496,
      "step": 22000
    },
    {
      "epoch": 5.25,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4427037239074707,
      "eval_runtime": 34.6227,
      "eval_samples_per_second": 19.756,
      "step": 22000
    },
    {
      "epoch": 5.26,
      "learning_rate": 1e-05,
      "loss": 1.4542,
      "step": 22010
    },
    {
      "epoch": 5.26,
      "learning_rate": 1e-05,
      "loss": 1.4461,
      "step": 22020
    },
    {
      "epoch": 5.26,
      "learning_rate": 1e-05,
      "loss": 1.4511,
      "step": 22030
    },
    {
      "epoch": 5.26,
      "learning_rate": 1e-05,
      "loss": 1.4517,
      "step": 22040
    },
    {
      "epoch": 5.27,
      "learning_rate": 1e-05,
      "loss": 1.4585,
      "step": 22050
    },
    {
      "epoch": 5.27,
      "learning_rate": 1e-05,
      "loss": 1.4495,
      "step": 22060
    },
    {
      "epoch": 5.27,
      "learning_rate": 1e-05,
      "loss": 1.4597,
      "step": 22070
    },
    {
      "epoch": 5.27,
      "learning_rate": 1e-05,
      "loss": 1.4531,
      "step": 22080
    },
    {
      "epoch": 5.28,
      "learning_rate": 1e-05,
      "loss": 1.4482,
      "step": 22090
    },
    {
      "epoch": 5.28,
      "learning_rate": 1e-05,
      "loss": 1.445,
      "step": 22100
    },
    {
      "epoch": 5.28,
      "eval_accuracy": 0.9751461988304093,
      "eval_loss": 1.4398616552352905,
      "eval_runtime": 34.4788,
      "eval_samples_per_second": 19.838,
      "step": 22100
    },
    {
      "epoch": 5.28,
      "learning_rate": 1e-05,
      "loss": 1.4559,
      "step": 22110
    },
    {
      "epoch": 5.28,
      "learning_rate": 1e-05,
      "loss": 1.4498,
      "step": 22120
    },
    {
      "epoch": 5.29,
      "learning_rate": 1e-05,
      "loss": 1.4574,
      "step": 22130
    },
    {
      "epoch": 5.29,
      "learning_rate": 1e-05,
      "loss": 1.4482,
      "step": 22140
    },
    {
      "epoch": 5.29,
      "learning_rate": 1e-05,
      "loss": 1.453,
      "step": 22150
    },
    {
      "epoch": 5.29,
      "learning_rate": 1e-05,
      "loss": 1.4563,
      "step": 22160
    },
    {
      "epoch": 5.29,
      "learning_rate": 1e-05,
      "loss": 1.4488,
      "step": 22170
    },
    {
      "epoch": 5.3,
      "learning_rate": 1e-05,
      "loss": 1.4514,
      "step": 22180
    },
    {
      "epoch": 5.3,
      "learning_rate": 1e-05,
      "loss": 1.4558,
      "step": 22190
    },
    {
      "epoch": 5.3,
      "learning_rate": 1e-05,
      "loss": 1.4672,
      "step": 22200
    },
    {
      "epoch": 5.3,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4410799741744995,
      "eval_runtime": 34.4635,
      "eval_samples_per_second": 19.847,
      "step": 22200
    },
    {
      "epoch": 5.3,
      "learning_rate": 1e-05,
      "loss": 1.4512,
      "step": 22210
    },
    {
      "epoch": 5.31,
      "learning_rate": 1e-05,
      "loss": 1.445,
      "step": 22220
    },
    {
      "epoch": 5.31,
      "learning_rate": 1e-05,
      "loss": 1.4577,
      "step": 22230
    },
    {
      "epoch": 5.31,
      "learning_rate": 1e-05,
      "loss": 1.4571,
      "step": 22240
    },
    {
      "epoch": 5.31,
      "learning_rate": 1e-05,
      "loss": 1.4472,
      "step": 22250
    },
    {
      "epoch": 5.32,
      "learning_rate": 1e-05,
      "loss": 1.4645,
      "step": 22260
    },
    {
      "epoch": 5.32,
      "learning_rate": 1e-05,
      "loss": 1.451,
      "step": 22270
    },
    {
      "epoch": 5.32,
      "learning_rate": 1e-05,
      "loss": 1.4534,
      "step": 22280
    },
    {
      "epoch": 5.32,
      "learning_rate": 1e-05,
      "loss": 1.4418,
      "step": 22290
    },
    {
      "epoch": 5.33,
      "learning_rate": 1e-05,
      "loss": 1.4498,
      "step": 22300
    },
    {
      "epoch": 5.33,
      "eval_accuracy": 0.9751461988304093,
      "eval_loss": 1.4414217472076416,
      "eval_runtime": 35.4507,
      "eval_samples_per_second": 19.294,
      "step": 22300
    },
    {
      "epoch": 5.33,
      "learning_rate": 1e-05,
      "loss": 1.4506,
      "step": 22310
    },
    {
      "epoch": 5.33,
      "learning_rate": 1e-05,
      "loss": 1.4411,
      "step": 22320
    },
    {
      "epoch": 5.33,
      "learning_rate": 1e-05,
      "loss": 1.4713,
      "step": 22330
    },
    {
      "epoch": 5.34,
      "learning_rate": 1e-05,
      "loss": 1.4583,
      "step": 22340
    },
    {
      "epoch": 5.34,
      "learning_rate": 1e-05,
      "loss": 1.4594,
      "step": 22350
    },
    {
      "epoch": 5.34,
      "learning_rate": 1e-05,
      "loss": 1.4549,
      "step": 22360
    },
    {
      "epoch": 5.34,
      "learning_rate": 1e-05,
      "loss": 1.4589,
      "step": 22370
    },
    {
      "epoch": 5.35,
      "learning_rate": 1e-05,
      "loss": 1.4501,
      "step": 22380
    },
    {
      "epoch": 5.35,
      "learning_rate": 1e-05,
      "loss": 1.449,
      "step": 22390
    },
    {
      "epoch": 5.35,
      "learning_rate": 1e-05,
      "loss": 1.4581,
      "step": 22400
    },
    {
      "epoch": 5.35,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4441415071487427,
      "eval_runtime": 34.4587,
      "eval_samples_per_second": 19.85,
      "step": 22400
    },
    {
      "epoch": 5.35,
      "learning_rate": 1e-05,
      "loss": 1.4447,
      "step": 22410
    },
    {
      "epoch": 5.35,
      "learning_rate": 1e-05,
      "loss": 1.4462,
      "step": 22420
    },
    {
      "epoch": 5.36,
      "learning_rate": 1e-05,
      "loss": 1.4489,
      "step": 22430
    },
    {
      "epoch": 5.36,
      "learning_rate": 1e-05,
      "loss": 1.4545,
      "step": 22440
    },
    {
      "epoch": 5.36,
      "learning_rate": 1e-05,
      "loss": 1.447,
      "step": 22450
    },
    {
      "epoch": 5.36,
      "learning_rate": 1e-05,
      "loss": 1.4545,
      "step": 22460
    },
    {
      "epoch": 5.37,
      "learning_rate": 1e-05,
      "loss": 1.4484,
      "step": 22470
    },
    {
      "epoch": 5.37,
      "learning_rate": 1e-05,
      "loss": 1.4524,
      "step": 22480
    },
    {
      "epoch": 5.37,
      "learning_rate": 1e-05,
      "loss": 1.45,
      "step": 22490
    },
    {
      "epoch": 5.37,
      "learning_rate": 1e-05,
      "loss": 1.4614,
      "step": 22500
    },
    {
      "epoch": 5.37,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4406715631484985,
      "eval_runtime": 34.5629,
      "eval_samples_per_second": 19.79,
      "step": 22500
    },
    {
      "epoch": 5.38,
      "learning_rate": 1e-05,
      "loss": 1.443,
      "step": 22510
    },
    {
      "epoch": 5.38,
      "learning_rate": 1e-05,
      "loss": 1.4489,
      "step": 22520
    },
    {
      "epoch": 5.38,
      "learning_rate": 1e-05,
      "loss": 1.4497,
      "step": 22530
    },
    {
      "epoch": 5.38,
      "learning_rate": 1e-05,
      "loss": 1.4561,
      "step": 22540
    },
    {
      "epoch": 5.39,
      "learning_rate": 1e-05,
      "loss": 1.4492,
      "step": 22550
    },
    {
      "epoch": 5.39,
      "learning_rate": 1e-05,
      "loss": 1.4546,
      "step": 22560
    },
    {
      "epoch": 5.39,
      "learning_rate": 1e-05,
      "loss": 1.4526,
      "step": 22570
    },
    {
      "epoch": 5.39,
      "learning_rate": 1e-05,
      "loss": 1.4566,
      "step": 22580
    },
    {
      "epoch": 5.4,
      "learning_rate": 1e-05,
      "loss": 1.4492,
      "step": 22590
    },
    {
      "epoch": 5.4,
      "learning_rate": 1e-05,
      "loss": 1.4494,
      "step": 22600
    },
    {
      "epoch": 5.4,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4418083429336548,
      "eval_runtime": 34.61,
      "eval_samples_per_second": 19.763,
      "step": 22600
    },
    {
      "epoch": 5.4,
      "learning_rate": 1e-05,
      "loss": 1.462,
      "step": 22610
    },
    {
      "epoch": 5.4,
      "learning_rate": 1e-05,
      "loss": 1.4433,
      "step": 22620
    },
    {
      "epoch": 5.4,
      "learning_rate": 1e-05,
      "loss": 1.452,
      "step": 22630
    },
    {
      "epoch": 5.41,
      "learning_rate": 1e-05,
      "loss": 1.4463,
      "step": 22640
    },
    {
      "epoch": 5.41,
      "learning_rate": 1e-05,
      "loss": 1.454,
      "step": 22650
    },
    {
      "epoch": 5.41,
      "learning_rate": 1e-05,
      "loss": 1.4493,
      "step": 22660
    },
    {
      "epoch": 5.41,
      "learning_rate": 1e-05,
      "loss": 1.4501,
      "step": 22670
    },
    {
      "epoch": 5.42,
      "learning_rate": 1e-05,
      "loss": 1.4488,
      "step": 22680
    },
    {
      "epoch": 5.42,
      "learning_rate": 1e-05,
      "loss": 1.4455,
      "step": 22690
    },
    {
      "epoch": 5.42,
      "learning_rate": 1e-05,
      "loss": 1.443,
      "step": 22700
    },
    {
      "epoch": 5.42,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4407645463943481,
      "eval_runtime": 34.4767,
      "eval_samples_per_second": 19.84,
      "step": 22700
    },
    {
      "epoch": 5.42,
      "learning_rate": 1e-05,
      "loss": 1.4414,
      "step": 22710
    },
    {
      "epoch": 5.43,
      "learning_rate": 1e-05,
      "loss": 1.4431,
      "step": 22720
    },
    {
      "epoch": 5.43,
      "learning_rate": 1e-05,
      "loss": 1.4441,
      "step": 22730
    },
    {
      "epoch": 5.43,
      "learning_rate": 1e-05,
      "loss": 1.4528,
      "step": 22740
    },
    {
      "epoch": 5.43,
      "learning_rate": 1e-05,
      "loss": 1.4541,
      "step": 22750
    },
    {
      "epoch": 5.44,
      "learning_rate": 1e-05,
      "loss": 1.4563,
      "step": 22760
    },
    {
      "epoch": 5.44,
      "learning_rate": 1e-05,
      "loss": 1.4563,
      "step": 22770
    },
    {
      "epoch": 5.44,
      "learning_rate": 1e-05,
      "loss": 1.4718,
      "step": 22780
    },
    {
      "epoch": 5.44,
      "learning_rate": 1e-05,
      "loss": 1.4521,
      "step": 22790
    },
    {
      "epoch": 5.45,
      "learning_rate": 1e-05,
      "loss": 1.4513,
      "step": 22800
    },
    {
      "epoch": 5.45,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.440168857574463,
      "eval_runtime": 34.3143,
      "eval_samples_per_second": 19.933,
      "step": 22800
    },
    {
      "epoch": 5.45,
      "learning_rate": 1e-05,
      "loss": 1.4617,
      "step": 22810
    },
    {
      "epoch": 5.45,
      "learning_rate": 1e-05,
      "loss": 1.4451,
      "step": 22820
    },
    {
      "epoch": 5.45,
      "learning_rate": 1e-05,
      "loss": 1.4529,
      "step": 22830
    },
    {
      "epoch": 5.45,
      "learning_rate": 1e-05,
      "loss": 1.4531,
      "step": 22840
    },
    {
      "epoch": 5.46,
      "learning_rate": 1e-05,
      "loss": 1.4506,
      "step": 22850
    },
    {
      "epoch": 5.46,
      "learning_rate": 1e-05,
      "loss": 1.4475,
      "step": 22860
    },
    {
      "epoch": 5.46,
      "learning_rate": 1e-05,
      "loss": 1.4509,
      "step": 22870
    },
    {
      "epoch": 5.46,
      "learning_rate": 1e-05,
      "loss": 1.4459,
      "step": 22880
    },
    {
      "epoch": 5.47,
      "learning_rate": 1e-05,
      "loss": 1.4911,
      "step": 22890
    },
    {
      "epoch": 5.47,
      "learning_rate": 1e-05,
      "loss": 1.4543,
      "step": 22900
    },
    {
      "epoch": 5.47,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4413350820541382,
      "eval_runtime": 34.5993,
      "eval_samples_per_second": 19.769,
      "step": 22900
    },
    {
      "epoch": 5.47,
      "learning_rate": 1e-05,
      "loss": 1.4484,
      "step": 22910
    },
    {
      "epoch": 5.47,
      "learning_rate": 1e-05,
      "loss": 1.4434,
      "step": 22920
    },
    {
      "epoch": 5.48,
      "learning_rate": 1e-05,
      "loss": 1.4449,
      "step": 22930
    },
    {
      "epoch": 5.48,
      "learning_rate": 1e-05,
      "loss": 1.4452,
      "step": 22940
    },
    {
      "epoch": 5.48,
      "learning_rate": 1e-05,
      "loss": 1.4454,
      "step": 22950
    },
    {
      "epoch": 5.48,
      "learning_rate": 1e-05,
      "loss": 1.4633,
      "step": 22960
    },
    {
      "epoch": 5.49,
      "learning_rate": 1e-05,
      "loss": 1.4522,
      "step": 22970
    },
    {
      "epoch": 5.49,
      "learning_rate": 1e-05,
      "loss": 1.4614,
      "step": 22980
    },
    {
      "epoch": 5.49,
      "learning_rate": 1e-05,
      "loss": 1.4516,
      "step": 22990
    },
    {
      "epoch": 5.49,
      "learning_rate": 1e-05,
      "loss": 1.4439,
      "step": 23000
    },
    {
      "epoch": 5.49,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4398645162582397,
      "eval_runtime": 34.2918,
      "eval_samples_per_second": 19.946,
      "step": 23000
    },
    {
      "epoch": 5.5,
      "learning_rate": 1e-05,
      "loss": 1.4542,
      "step": 23010
    },
    {
      "epoch": 5.5,
      "learning_rate": 1e-05,
      "loss": 1.4494,
      "step": 23020
    },
    {
      "epoch": 5.5,
      "learning_rate": 1e-05,
      "loss": 1.442,
      "step": 23030
    },
    {
      "epoch": 5.5,
      "learning_rate": 1e-05,
      "loss": 1.4542,
      "step": 23040
    },
    {
      "epoch": 5.51,
      "learning_rate": 1e-05,
      "loss": 1.4549,
      "step": 23050
    },
    {
      "epoch": 5.51,
      "learning_rate": 1e-05,
      "loss": 1.456,
      "step": 23060
    },
    {
      "epoch": 5.51,
      "learning_rate": 1e-05,
      "loss": 1.4376,
      "step": 23070
    },
    {
      "epoch": 5.51,
      "learning_rate": 1e-05,
      "loss": 1.4486,
      "step": 23080
    },
    {
      "epoch": 5.51,
      "learning_rate": 1e-05,
      "loss": 1.4507,
      "step": 23090
    },
    {
      "epoch": 5.52,
      "learning_rate": 1e-05,
      "loss": 1.4517,
      "step": 23100
    },
    {
      "epoch": 5.52,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.440263032913208,
      "eval_runtime": 34.5264,
      "eval_samples_per_second": 19.811,
      "step": 23100
    },
    {
      "epoch": 5.52,
      "learning_rate": 1e-05,
      "loss": 1.4505,
      "step": 23110
    },
    {
      "epoch": 5.52,
      "learning_rate": 1e-05,
      "loss": 1.4487,
      "step": 23120
    },
    {
      "epoch": 5.52,
      "learning_rate": 1e-05,
      "loss": 1.4708,
      "step": 23130
    },
    {
      "epoch": 5.53,
      "learning_rate": 1e-05,
      "loss": 1.4517,
      "step": 23140
    },
    {
      "epoch": 5.53,
      "learning_rate": 1e-05,
      "loss": 1.4368,
      "step": 23150
    },
    {
      "epoch": 5.53,
      "learning_rate": 1e-05,
      "loss": 1.4425,
      "step": 23160
    },
    {
      "epoch": 5.53,
      "learning_rate": 1e-05,
      "loss": 1.451,
      "step": 23170
    },
    {
      "epoch": 5.54,
      "learning_rate": 1e-05,
      "loss": 1.4541,
      "step": 23180
    },
    {
      "epoch": 5.54,
      "learning_rate": 1e-05,
      "loss": 1.4481,
      "step": 23190
    },
    {
      "epoch": 5.54,
      "learning_rate": 1e-05,
      "loss": 1.4508,
      "step": 23200
    },
    {
      "epoch": 5.54,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4389944076538086,
      "eval_runtime": 34.3792,
      "eval_samples_per_second": 19.896,
      "step": 23200
    },
    {
      "epoch": 5.54,
      "learning_rate": 1e-05,
      "loss": 1.4536,
      "step": 23210
    },
    {
      "epoch": 5.55,
      "learning_rate": 1e-05,
      "loss": 1.4484,
      "step": 23220
    },
    {
      "epoch": 5.55,
      "learning_rate": 1e-05,
      "loss": 1.4609,
      "step": 23230
    },
    {
      "epoch": 5.55,
      "learning_rate": 1e-05,
      "loss": 1.473,
      "step": 23240
    },
    {
      "epoch": 5.55,
      "learning_rate": 1e-05,
      "loss": 1.4487,
      "step": 23250
    },
    {
      "epoch": 5.56,
      "learning_rate": 1e-05,
      "loss": 1.4487,
      "step": 23260
    },
    {
      "epoch": 5.56,
      "learning_rate": 1e-05,
      "loss": 1.4494,
      "step": 23270
    },
    {
      "epoch": 5.56,
      "learning_rate": 1e-05,
      "loss": 1.4514,
      "step": 23280
    },
    {
      "epoch": 5.56,
      "learning_rate": 1e-05,
      "loss": 1.4459,
      "step": 23290
    },
    {
      "epoch": 5.56,
      "learning_rate": 1e-05,
      "loss": 1.467,
      "step": 23300
    },
    {
      "epoch": 5.56,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.439513087272644,
      "eval_runtime": 34.498,
      "eval_samples_per_second": 19.827,
      "step": 23300
    },
    {
      "epoch": 5.57,
      "learning_rate": 1e-05,
      "loss": 1.4451,
      "step": 23310
    },
    {
      "epoch": 5.57,
      "learning_rate": 1e-05,
      "loss": 1.4436,
      "step": 23320
    },
    {
      "epoch": 5.57,
      "learning_rate": 1e-05,
      "loss": 1.4561,
      "step": 23330
    },
    {
      "epoch": 5.57,
      "learning_rate": 1e-05,
      "loss": 1.4455,
      "step": 23340
    },
    {
      "epoch": 5.58,
      "learning_rate": 1e-05,
      "loss": 1.445,
      "step": 23350
    },
    {
      "epoch": 5.58,
      "learning_rate": 1e-05,
      "loss": 1.4408,
      "step": 23360
    },
    {
      "epoch": 5.58,
      "learning_rate": 1e-05,
      "loss": 1.4405,
      "step": 23370
    },
    {
      "epoch": 5.58,
      "learning_rate": 1e-05,
      "loss": 1.4484,
      "step": 23380
    },
    {
      "epoch": 5.59,
      "learning_rate": 1e-05,
      "loss": 1.4475,
      "step": 23390
    },
    {
      "epoch": 5.59,
      "learning_rate": 1e-05,
      "loss": 1.4497,
      "step": 23400
    },
    {
      "epoch": 5.59,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.439876914024353,
      "eval_runtime": 34.4177,
      "eval_samples_per_second": 19.873,
      "step": 23400
    },
    {
      "epoch": 5.59,
      "learning_rate": 1e-05,
      "loss": 1.4654,
      "step": 23410
    },
    {
      "epoch": 5.59,
      "learning_rate": 1e-05,
      "loss": 1.4404,
      "step": 23420
    },
    {
      "epoch": 5.6,
      "learning_rate": 1e-05,
      "loss": 1.4538,
      "step": 23430
    },
    {
      "epoch": 5.6,
      "learning_rate": 1e-05,
      "loss": 1.4466,
      "step": 23440
    },
    {
      "epoch": 5.6,
      "learning_rate": 1e-05,
      "loss": 1.4505,
      "step": 23450
    },
    {
      "epoch": 5.6,
      "learning_rate": 1e-05,
      "loss": 1.4513,
      "step": 23460
    },
    {
      "epoch": 5.61,
      "learning_rate": 1e-05,
      "loss": 1.4543,
      "step": 23470
    },
    {
      "epoch": 5.61,
      "learning_rate": 1e-05,
      "loss": 1.5763,
      "step": 23480
    },
    {
      "epoch": 5.61,
      "learning_rate": 1e-05,
      "loss": 1.4578,
      "step": 23490
    },
    {
      "epoch": 5.61,
      "learning_rate": 1e-05,
      "loss": 1.4454,
      "step": 23500
    },
    {
      "epoch": 5.61,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4397681951522827,
      "eval_runtime": 34.4966,
      "eval_samples_per_second": 19.828,
      "step": 23500
    },
    {
      "epoch": 5.61,
      "learning_rate": 1e-05,
      "loss": 1.4548,
      "step": 23510
    },
    {
      "epoch": 5.62,
      "learning_rate": 1e-05,
      "loss": 1.4413,
      "step": 23520
    },
    {
      "epoch": 5.62,
      "learning_rate": 1e-05,
      "loss": 1.4472,
      "step": 23530
    },
    {
      "epoch": 5.62,
      "learning_rate": 1e-05,
      "loss": 1.4439,
      "step": 23540
    },
    {
      "epoch": 5.62,
      "learning_rate": 1e-05,
      "loss": 1.4414,
      "step": 23550
    },
    {
      "epoch": 5.63,
      "learning_rate": 1e-05,
      "loss": 1.4522,
      "step": 23560
    },
    {
      "epoch": 5.63,
      "learning_rate": 1e-05,
      "loss": 1.4603,
      "step": 23570
    },
    {
      "epoch": 5.63,
      "learning_rate": 1e-05,
      "loss": 1.4474,
      "step": 23580
    },
    {
      "epoch": 5.63,
      "learning_rate": 1e-05,
      "loss": 1.4454,
      "step": 23590
    },
    {
      "epoch": 5.64,
      "learning_rate": 1e-05,
      "loss": 1.4502,
      "step": 23600
    },
    {
      "epoch": 5.64,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4391789436340332,
      "eval_runtime": 34.4416,
      "eval_samples_per_second": 19.86,
      "step": 23600
    },
    {
      "epoch": 5.64,
      "learning_rate": 1e-05,
      "loss": 1.4591,
      "step": 23610
    },
    {
      "epoch": 5.64,
      "learning_rate": 1e-05,
      "loss": 1.4742,
      "step": 23620
    },
    {
      "epoch": 5.64,
      "learning_rate": 1e-05,
      "loss": 1.4515,
      "step": 23630
    },
    {
      "epoch": 5.65,
      "learning_rate": 1e-05,
      "loss": 1.4453,
      "step": 23640
    },
    {
      "epoch": 5.65,
      "learning_rate": 1e-05,
      "loss": 1.4909,
      "step": 23650
    },
    {
      "epoch": 5.65,
      "learning_rate": 1e-05,
      "loss": 1.4468,
      "step": 23660
    },
    {
      "epoch": 5.65,
      "learning_rate": 1e-05,
      "loss": 1.4553,
      "step": 23670
    },
    {
      "epoch": 5.66,
      "learning_rate": 1e-05,
      "loss": 1.4537,
      "step": 23680
    },
    {
      "epoch": 5.66,
      "learning_rate": 1e-05,
      "loss": 1.4522,
      "step": 23690
    },
    {
      "epoch": 5.66,
      "learning_rate": 1e-05,
      "loss": 1.4468,
      "step": 23700
    },
    {
      "epoch": 5.66,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4397222995758057,
      "eval_runtime": 34.3782,
      "eval_samples_per_second": 19.896,
      "step": 23700
    },
    {
      "epoch": 5.66,
      "learning_rate": 1e-05,
      "loss": 1.459,
      "step": 23710
    },
    {
      "epoch": 5.67,
      "learning_rate": 1e-05,
      "loss": 1.4653,
      "step": 23720
    },
    {
      "epoch": 5.67,
      "learning_rate": 1e-05,
      "loss": 1.4382,
      "step": 23730
    },
    {
      "epoch": 5.67,
      "learning_rate": 1e-05,
      "loss": 1.4437,
      "step": 23740
    },
    {
      "epoch": 5.67,
      "learning_rate": 1e-05,
      "loss": 1.4455,
      "step": 23750
    },
    {
      "epoch": 5.67,
      "learning_rate": 1e-05,
      "loss": 1.4474,
      "step": 23760
    },
    {
      "epoch": 5.68,
      "learning_rate": 1e-05,
      "loss": 1.4537,
      "step": 23770
    },
    {
      "epoch": 5.68,
      "learning_rate": 1e-05,
      "loss": 1.45,
      "step": 23780
    },
    {
      "epoch": 5.68,
      "learning_rate": 1e-05,
      "loss": 1.4547,
      "step": 23790
    },
    {
      "epoch": 5.68,
      "learning_rate": 1e-05,
      "loss": 1.4446,
      "step": 23800
    },
    {
      "epoch": 5.68,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4379268884658813,
      "eval_runtime": 34.4244,
      "eval_samples_per_second": 19.87,
      "step": 23800
    },
    {
      "epoch": 5.69,
      "learning_rate": 1e-05,
      "loss": 1.4482,
      "step": 23810
    },
    {
      "epoch": 5.69,
      "learning_rate": 1e-05,
      "loss": 1.4521,
      "step": 23820
    },
    {
      "epoch": 5.69,
      "learning_rate": 1e-05,
      "loss": 1.4436,
      "step": 23830
    },
    {
      "epoch": 5.69,
      "learning_rate": 1e-05,
      "loss": 1.4449,
      "step": 23840
    },
    {
      "epoch": 5.7,
      "learning_rate": 1e-05,
      "loss": 1.4499,
      "step": 23850
    },
    {
      "epoch": 5.7,
      "learning_rate": 1e-05,
      "loss": 1.4381,
      "step": 23860
    },
    {
      "epoch": 5.7,
      "learning_rate": 1e-05,
      "loss": 1.5044,
      "step": 23870
    },
    {
      "epoch": 5.7,
      "learning_rate": 1e-05,
      "loss": 1.4402,
      "step": 23880
    },
    {
      "epoch": 5.71,
      "learning_rate": 1e-05,
      "loss": 1.4588,
      "step": 23890
    },
    {
      "epoch": 5.71,
      "learning_rate": 1e-05,
      "loss": 1.4428,
      "step": 23900
    },
    {
      "epoch": 5.71,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4379096031188965,
      "eval_runtime": 34.5111,
      "eval_samples_per_second": 19.82,
      "step": 23900
    },
    {
      "epoch": 5.71,
      "learning_rate": 1e-05,
      "loss": 1.4373,
      "step": 23910
    },
    {
      "epoch": 5.71,
      "learning_rate": 1e-05,
      "loss": 1.4562,
      "step": 23920
    },
    {
      "epoch": 5.72,
      "learning_rate": 1e-05,
      "loss": 1.4472,
      "step": 23930
    },
    {
      "epoch": 5.72,
      "learning_rate": 1e-05,
      "loss": 1.4518,
      "step": 23940
    },
    {
      "epoch": 5.72,
      "learning_rate": 1e-05,
      "loss": 1.445,
      "step": 23950
    },
    {
      "epoch": 5.72,
      "learning_rate": 1e-05,
      "loss": 1.4468,
      "step": 23960
    },
    {
      "epoch": 5.72,
      "learning_rate": 1e-05,
      "loss": 1.4445,
      "step": 23970
    },
    {
      "epoch": 5.73,
      "learning_rate": 1e-05,
      "loss": 1.4493,
      "step": 23980
    },
    {
      "epoch": 5.73,
      "learning_rate": 1e-05,
      "loss": 1.4453,
      "step": 23990
    },
    {
      "epoch": 5.73,
      "learning_rate": 1e-05,
      "loss": 1.44,
      "step": 24000
    },
    {
      "epoch": 5.73,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4376565217971802,
      "eval_runtime": 34.4217,
      "eval_samples_per_second": 19.871,
      "step": 24000
    },
    {
      "epoch": 5.73,
      "learning_rate": 1e-05,
      "loss": 1.4443,
      "step": 24010
    },
    {
      "epoch": 5.74,
      "learning_rate": 1e-05,
      "loss": 1.4467,
      "step": 24020
    },
    {
      "epoch": 5.74,
      "learning_rate": 1e-05,
      "loss": 1.4626,
      "step": 24030
    },
    {
      "epoch": 5.74,
      "learning_rate": 1e-05,
      "loss": 1.453,
      "step": 24040
    },
    {
      "epoch": 5.74,
      "learning_rate": 1e-05,
      "loss": 1.4402,
      "step": 24050
    },
    {
      "epoch": 5.75,
      "learning_rate": 1e-05,
      "loss": 1.4373,
      "step": 24060
    },
    {
      "epoch": 5.75,
      "learning_rate": 1e-05,
      "loss": 1.4492,
      "step": 24070
    },
    {
      "epoch": 5.75,
      "learning_rate": 1e-05,
      "loss": 1.4494,
      "step": 24080
    },
    {
      "epoch": 5.75,
      "learning_rate": 1e-05,
      "loss": 1.4425,
      "step": 24090
    },
    {
      "epoch": 5.76,
      "learning_rate": 1e-05,
      "loss": 1.4536,
      "step": 24100
    },
    {
      "epoch": 5.76,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4388165473937988,
      "eval_runtime": 34.4312,
      "eval_samples_per_second": 19.866,
      "step": 24100
    },
    {
      "epoch": 5.76,
      "learning_rate": 1e-05,
      "loss": 1.4508,
      "step": 24110
    },
    {
      "epoch": 5.76,
      "learning_rate": 1e-05,
      "loss": 1.4495,
      "step": 24120
    },
    {
      "epoch": 5.76,
      "learning_rate": 1e-05,
      "loss": 1.4417,
      "step": 24130
    },
    {
      "epoch": 5.77,
      "learning_rate": 1e-05,
      "loss": 1.4457,
      "step": 24140
    },
    {
      "epoch": 5.77,
      "learning_rate": 1e-05,
      "loss": 1.4498,
      "step": 24150
    },
    {
      "epoch": 5.77,
      "learning_rate": 1e-05,
      "loss": 1.4585,
      "step": 24160
    },
    {
      "epoch": 5.77,
      "learning_rate": 1e-05,
      "loss": 1.4409,
      "step": 24170
    },
    {
      "epoch": 5.78,
      "learning_rate": 1e-05,
      "loss": 1.4532,
      "step": 24180
    },
    {
      "epoch": 5.78,
      "learning_rate": 1e-05,
      "loss": 1.4485,
      "step": 24190
    },
    {
      "epoch": 5.78,
      "learning_rate": 1e-05,
      "loss": 1.4418,
      "step": 24200
    },
    {
      "epoch": 5.78,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.438604712486267,
      "eval_runtime": 34.7512,
      "eval_samples_per_second": 19.683,
      "step": 24200
    },
    {
      "epoch": 5.78,
      "learning_rate": 1e-05,
      "loss": 1.4419,
      "step": 24210
    },
    {
      "epoch": 5.78,
      "learning_rate": 1e-05,
      "loss": 1.4385,
      "step": 24220
    },
    {
      "epoch": 5.79,
      "learning_rate": 1e-05,
      "loss": 1.4386,
      "step": 24230
    },
    {
      "epoch": 5.79,
      "learning_rate": 1e-05,
      "loss": 1.4385,
      "step": 24240
    },
    {
      "epoch": 5.79,
      "learning_rate": 1e-05,
      "loss": 1.4438,
      "step": 24250
    },
    {
      "epoch": 5.79,
      "learning_rate": 1e-05,
      "loss": 1.4452,
      "step": 24260
    },
    {
      "epoch": 5.8,
      "learning_rate": 1e-05,
      "loss": 1.4575,
      "step": 24270
    },
    {
      "epoch": 5.8,
      "learning_rate": 1e-05,
      "loss": 1.4435,
      "step": 24280
    },
    {
      "epoch": 5.8,
      "learning_rate": 1e-05,
      "loss": 1.4458,
      "step": 24290
    },
    {
      "epoch": 5.8,
      "learning_rate": 1e-05,
      "loss": 1.4463,
      "step": 24300
    },
    {
      "epoch": 5.8,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4374828338623047,
      "eval_runtime": 34.4395,
      "eval_samples_per_second": 19.861,
      "step": 24300
    },
    {
      "epoch": 5.81,
      "learning_rate": 1e-05,
      "loss": 1.431,
      "step": 24310
    },
    {
      "epoch": 5.81,
      "learning_rate": 1e-05,
      "loss": 1.457,
      "step": 24320
    },
    {
      "epoch": 5.81,
      "learning_rate": 1e-05,
      "loss": 1.4492,
      "step": 24330
    },
    {
      "epoch": 5.81,
      "learning_rate": 1e-05,
      "loss": 1.4512,
      "step": 24340
    },
    {
      "epoch": 5.82,
      "learning_rate": 1e-05,
      "loss": 1.4516,
      "step": 24350
    },
    {
      "epoch": 5.82,
      "learning_rate": 1e-05,
      "loss": 1.4547,
      "step": 24360
    },
    {
      "epoch": 5.82,
      "learning_rate": 1e-05,
      "loss": 1.4461,
      "step": 24370
    },
    {
      "epoch": 5.82,
      "learning_rate": 1e-05,
      "loss": 1.4475,
      "step": 24380
    },
    {
      "epoch": 5.83,
      "learning_rate": 1e-05,
      "loss": 1.4484,
      "step": 24390
    },
    {
      "epoch": 5.83,
      "learning_rate": 1e-05,
      "loss": 1.4425,
      "step": 24400
    },
    {
      "epoch": 5.83,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4364237785339355,
      "eval_runtime": 34.5908,
      "eval_samples_per_second": 19.774,
      "step": 24400
    },
    {
      "epoch": 5.83,
      "learning_rate": 1e-05,
      "loss": 1.4416,
      "step": 24410
    },
    {
      "epoch": 5.83,
      "learning_rate": 1e-05,
      "loss": 1.4524,
      "step": 24420
    },
    {
      "epoch": 5.83,
      "learning_rate": 1e-05,
      "loss": 1.4515,
      "step": 24430
    },
    {
      "epoch": 5.84,
      "learning_rate": 1e-05,
      "loss": 1.451,
      "step": 24440
    },
    {
      "epoch": 5.84,
      "learning_rate": 1e-05,
      "loss": 1.4461,
      "step": 24450
    },
    {
      "epoch": 5.84,
      "learning_rate": 1e-05,
      "loss": 1.4588,
      "step": 24460
    },
    {
      "epoch": 5.84,
      "learning_rate": 1e-05,
      "loss": 1.44,
      "step": 24470
    },
    {
      "epoch": 5.85,
      "learning_rate": 1e-05,
      "loss": 1.4452,
      "step": 24480
    },
    {
      "epoch": 5.85,
      "learning_rate": 1e-05,
      "loss": 1.4465,
      "step": 24490
    },
    {
      "epoch": 5.85,
      "learning_rate": 1e-05,
      "loss": 1.462,
      "step": 24500
    },
    {
      "epoch": 5.85,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4378098249435425,
      "eval_runtime": 34.7023,
      "eval_samples_per_second": 19.71,
      "step": 24500
    },
    {
      "epoch": 5.85,
      "learning_rate": 1e-05,
      "loss": 1.4485,
      "step": 24510
    },
    {
      "epoch": 5.86,
      "learning_rate": 1e-05,
      "loss": 1.4466,
      "step": 24520
    },
    {
      "epoch": 5.86,
      "learning_rate": 1e-05,
      "loss": 1.4526,
      "step": 24530
    },
    {
      "epoch": 5.86,
      "learning_rate": 1e-05,
      "loss": 1.4468,
      "step": 24540
    },
    {
      "epoch": 5.86,
      "learning_rate": 1e-05,
      "loss": 1.4421,
      "step": 24550
    },
    {
      "epoch": 5.87,
      "learning_rate": 1e-05,
      "loss": 1.4439,
      "step": 24560
    },
    {
      "epoch": 5.87,
      "learning_rate": 1e-05,
      "loss": 1.4393,
      "step": 24570
    },
    {
      "epoch": 5.87,
      "learning_rate": 1e-05,
      "loss": 1.4438,
      "step": 24580
    },
    {
      "epoch": 5.87,
      "learning_rate": 1e-05,
      "loss": 1.4474,
      "step": 24590
    },
    {
      "epoch": 5.88,
      "learning_rate": 1e-05,
      "loss": 1.4463,
      "step": 24600
    },
    {
      "epoch": 5.88,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.436104416847229,
      "eval_runtime": 34.3759,
      "eval_samples_per_second": 19.898,
      "step": 24600
    },
    {
      "epoch": 5.88,
      "learning_rate": 1e-05,
      "loss": 1.4505,
      "step": 24610
    },
    {
      "epoch": 5.88,
      "learning_rate": 1e-05,
      "loss": 1.4478,
      "step": 24620
    },
    {
      "epoch": 5.88,
      "learning_rate": 1e-05,
      "loss": 1.4505,
      "step": 24630
    },
    {
      "epoch": 5.88,
      "learning_rate": 1e-05,
      "loss": 1.4518,
      "step": 24640
    },
    {
      "epoch": 5.89,
      "learning_rate": 1e-05,
      "loss": 1.4479,
      "step": 24650
    },
    {
      "epoch": 5.89,
      "learning_rate": 1e-05,
      "loss": 1.4541,
      "step": 24660
    },
    {
      "epoch": 5.89,
      "learning_rate": 1e-05,
      "loss": 1.4442,
      "step": 24670
    },
    {
      "epoch": 5.89,
      "learning_rate": 1e-05,
      "loss": 1.4537,
      "step": 24680
    },
    {
      "epoch": 5.9,
      "learning_rate": 1e-05,
      "loss": 1.4496,
      "step": 24690
    },
    {
      "epoch": 5.9,
      "learning_rate": 1e-05,
      "loss": 1.4435,
      "step": 24700
    },
    {
      "epoch": 5.9,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4375194311141968,
      "eval_runtime": 34.4139,
      "eval_samples_per_second": 19.876,
      "step": 24700
    },
    {
      "epoch": 5.9,
      "learning_rate": 1e-05,
      "loss": 1.4445,
      "step": 24710
    },
    {
      "epoch": 5.9,
      "learning_rate": 1e-05,
      "loss": 1.4552,
      "step": 24720
    },
    {
      "epoch": 5.91,
      "learning_rate": 1e-05,
      "loss": 1.4419,
      "step": 24730
    },
    {
      "epoch": 5.91,
      "learning_rate": 1e-05,
      "loss": 1.4462,
      "step": 24740
    },
    {
      "epoch": 5.91,
      "learning_rate": 1e-05,
      "loss": 1.4429,
      "step": 24750
    },
    {
      "epoch": 5.91,
      "learning_rate": 1e-05,
      "loss": 1.4484,
      "step": 24760
    },
    {
      "epoch": 5.92,
      "learning_rate": 1e-05,
      "loss": 1.4454,
      "step": 24770
    },
    {
      "epoch": 5.92,
      "learning_rate": 1e-05,
      "loss": 1.4435,
      "step": 24780
    },
    {
      "epoch": 5.92,
      "learning_rate": 1e-05,
      "loss": 1.4435,
      "step": 24790
    },
    {
      "epoch": 5.92,
      "learning_rate": 1e-05,
      "loss": 1.4528,
      "step": 24800
    },
    {
      "epoch": 5.92,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4366196393966675,
      "eval_runtime": 34.5426,
      "eval_samples_per_second": 19.802,
      "step": 24800
    },
    {
      "epoch": 5.93,
      "learning_rate": 1e-05,
      "loss": 1.4455,
      "step": 24810
    },
    {
      "epoch": 5.93,
      "learning_rate": 1e-05,
      "loss": 1.4442,
      "step": 24820
    },
    {
      "epoch": 5.93,
      "learning_rate": 1e-05,
      "loss": 1.4434,
      "step": 24830
    },
    {
      "epoch": 5.93,
      "learning_rate": 1e-05,
      "loss": 1.4429,
      "step": 24840
    },
    {
      "epoch": 5.94,
      "learning_rate": 1e-05,
      "loss": 1.4386,
      "step": 24850
    },
    {
      "epoch": 5.94,
      "learning_rate": 1e-05,
      "loss": 1.4471,
      "step": 24860
    },
    {
      "epoch": 5.94,
      "learning_rate": 1e-05,
      "loss": 1.4453,
      "step": 24870
    },
    {
      "epoch": 5.94,
      "learning_rate": 1e-05,
      "loss": 1.4516,
      "step": 24880
    },
    {
      "epoch": 5.94,
      "learning_rate": 1e-05,
      "loss": 1.4533,
      "step": 24890
    },
    {
      "epoch": 5.95,
      "learning_rate": 1e-05,
      "loss": 1.4488,
      "step": 24900
    },
    {
      "epoch": 5.95,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4374855756759644,
      "eval_runtime": 34.4791,
      "eval_samples_per_second": 19.838,
      "step": 24900
    },
    {
      "epoch": 5.95,
      "learning_rate": 1e-05,
      "loss": 1.4428,
      "step": 24910
    },
    {
      "epoch": 5.95,
      "learning_rate": 1e-05,
      "loss": 1.4485,
      "step": 24920
    },
    {
      "epoch": 5.95,
      "learning_rate": 1e-05,
      "loss": 1.4495,
      "step": 24930
    },
    {
      "epoch": 5.96,
      "learning_rate": 1e-05,
      "loss": 1.4577,
      "step": 24940
    },
    {
      "epoch": 5.96,
      "learning_rate": 1e-05,
      "loss": 1.4487,
      "step": 24950
    },
    {
      "epoch": 5.96,
      "learning_rate": 1e-05,
      "loss": 1.451,
      "step": 24960
    },
    {
      "epoch": 5.96,
      "learning_rate": 1e-05,
      "loss": 1.4434,
      "step": 24970
    },
    {
      "epoch": 5.97,
      "learning_rate": 1e-05,
      "loss": 1.4583,
      "step": 24980
    },
    {
      "epoch": 5.97,
      "learning_rate": 1e-05,
      "loss": 1.4349,
      "step": 24990
    },
    {
      "epoch": 5.97,
      "learning_rate": 1e-05,
      "loss": 1.4465,
      "step": 25000
    },
    {
      "epoch": 5.97,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4376389980316162,
      "eval_runtime": 34.3854,
      "eval_samples_per_second": 19.892,
      "step": 25000
    },
    {
      "epoch": 5.97,
      "learning_rate": 1e-05,
      "loss": 1.4377,
      "step": 25010
    },
    {
      "epoch": 5.98,
      "learning_rate": 1e-05,
      "loss": 1.4502,
      "step": 25020
    },
    {
      "epoch": 5.98,
      "learning_rate": 1e-05,
      "loss": 1.4489,
      "step": 25030
    },
    {
      "epoch": 5.98,
      "learning_rate": 1e-05,
      "loss": 1.4439,
      "step": 25040
    },
    {
      "epoch": 5.98,
      "learning_rate": 1e-05,
      "loss": 1.4511,
      "step": 25050
    },
    {
      "epoch": 5.99,
      "learning_rate": 1e-05,
      "loss": 1.435,
      "step": 25060
    },
    {
      "epoch": 5.99,
      "learning_rate": 1e-05,
      "loss": 1.4536,
      "step": 25070
    },
    {
      "epoch": 5.99,
      "learning_rate": 1e-05,
      "loss": 1.4446,
      "step": 25080
    },
    {
      "epoch": 5.99,
      "learning_rate": 1e-05,
      "loss": 1.4381,
      "step": 25090
    },
    {
      "epoch": 5.99,
      "learning_rate": 1e-05,
      "loss": 1.4511,
      "step": 25100
    },
    {
      "epoch": 5.99,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4370065927505493,
      "eval_runtime": 34.6341,
      "eval_samples_per_second": 19.749,
      "step": 25100
    },
    {
      "epoch": 6.0,
      "learning_rate": 1e-05,
      "loss": 1.4485,
      "step": 25110
    },
    {
      "epoch": 6.0,
      "learning_rate": 1e-05,
      "loss": 1.4874,
      "step": 25120
    },
    {
      "epoch": 6.0,
      "learning_rate": 1e-05,
      "loss": 1.4495,
      "step": 25130
    },
    {
      "epoch": 6.0,
      "learning_rate": 1e-05,
      "loss": 1.4458,
      "step": 25140
    },
    {
      "epoch": 6.01,
      "learning_rate": 1e-05,
      "loss": 1.4436,
      "step": 25150
    },
    {
      "epoch": 6.01,
      "learning_rate": 1e-05,
      "loss": 1.4401,
      "step": 25160
    },
    {
      "epoch": 6.01,
      "learning_rate": 1e-05,
      "loss": 1.4524,
      "step": 25170
    },
    {
      "epoch": 6.01,
      "learning_rate": 1e-05,
      "loss": 1.4429,
      "step": 25180
    },
    {
      "epoch": 6.02,
      "learning_rate": 1e-05,
      "loss": 1.4413,
      "step": 25190
    },
    {
      "epoch": 6.02,
      "learning_rate": 1e-05,
      "loss": 1.4447,
      "step": 25200
    },
    {
      "epoch": 6.02,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4363834857940674,
      "eval_runtime": 34.4134,
      "eval_samples_per_second": 19.876,
      "step": 25200
    },
    {
      "epoch": 6.02,
      "learning_rate": 1e-05,
      "loss": 1.4529,
      "step": 25210
    },
    {
      "epoch": 6.02,
      "learning_rate": 1e-05,
      "loss": 1.4371,
      "step": 25220
    },
    {
      "epoch": 6.03,
      "learning_rate": 1e-05,
      "loss": 1.4513,
      "step": 25230
    },
    {
      "epoch": 6.03,
      "learning_rate": 1e-05,
      "loss": 1.439,
      "step": 25240
    },
    {
      "epoch": 6.03,
      "learning_rate": 1e-05,
      "loss": 1.4574,
      "step": 25250
    },
    {
      "epoch": 6.03,
      "learning_rate": 1e-05,
      "loss": 1.4441,
      "step": 25260
    },
    {
      "epoch": 6.04,
      "learning_rate": 1e-05,
      "loss": 1.4563,
      "step": 25270
    },
    {
      "epoch": 6.04,
      "learning_rate": 1e-05,
      "loss": 1.4444,
      "step": 25280
    },
    {
      "epoch": 6.04,
      "learning_rate": 1e-05,
      "loss": 1.4597,
      "step": 25290
    },
    {
      "epoch": 6.04,
      "learning_rate": 1e-05,
      "loss": 1.4403,
      "step": 25300
    },
    {
      "epoch": 6.04,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4364502429962158,
      "eval_runtime": 34.7261,
      "eval_samples_per_second": 19.697,
      "step": 25300
    },
    {
      "epoch": 6.04,
      "learning_rate": 1e-05,
      "loss": 1.4484,
      "step": 25310
    },
    {
      "epoch": 6.05,
      "learning_rate": 1e-05,
      "loss": 1.447,
      "step": 25320
    },
    {
      "epoch": 6.05,
      "learning_rate": 1e-05,
      "loss": 1.458,
      "step": 25330
    },
    {
      "epoch": 6.05,
      "learning_rate": 1e-05,
      "loss": 1.4529,
      "step": 25340
    },
    {
      "epoch": 6.05,
      "learning_rate": 1e-05,
      "loss": 1.4488,
      "step": 25350
    },
    {
      "epoch": 6.06,
      "learning_rate": 1e-05,
      "loss": 1.4419,
      "step": 25360
    },
    {
      "epoch": 6.06,
      "learning_rate": 1e-05,
      "loss": 1.4442,
      "step": 25370
    },
    {
      "epoch": 6.06,
      "learning_rate": 1e-05,
      "loss": 1.4494,
      "step": 25380
    },
    {
      "epoch": 6.06,
      "learning_rate": 1e-05,
      "loss": 1.4415,
      "step": 25390
    },
    {
      "epoch": 6.07,
      "learning_rate": 1e-05,
      "loss": 1.4369,
      "step": 25400
    },
    {
      "epoch": 6.07,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4353899955749512,
      "eval_runtime": 34.5329,
      "eval_samples_per_second": 19.807,
      "step": 25400
    },
    {
      "epoch": 6.07,
      "learning_rate": 1e-05,
      "loss": 1.4501,
      "step": 25410
    },
    {
      "epoch": 6.07,
      "learning_rate": 1e-05,
      "loss": 1.4437,
      "step": 25420
    },
    {
      "epoch": 6.07,
      "learning_rate": 1e-05,
      "loss": 1.4554,
      "step": 25430
    },
    {
      "epoch": 6.08,
      "learning_rate": 1e-05,
      "loss": 1.4399,
      "step": 25440
    },
    {
      "epoch": 6.08,
      "learning_rate": 1e-05,
      "loss": 1.4391,
      "step": 25450
    },
    {
      "epoch": 6.08,
      "learning_rate": 1e-05,
      "loss": 1.4416,
      "step": 25460
    },
    {
      "epoch": 6.08,
      "learning_rate": 1e-05,
      "loss": 1.4485,
      "step": 25470
    },
    {
      "epoch": 6.09,
      "learning_rate": 1e-05,
      "loss": 1.4359,
      "step": 25480
    },
    {
      "epoch": 6.09,
      "learning_rate": 1e-05,
      "loss": 1.4522,
      "step": 25490
    },
    {
      "epoch": 6.09,
      "learning_rate": 1e-05,
      "loss": 1.4571,
      "step": 25500
    },
    {
      "epoch": 6.09,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4365099668502808,
      "eval_runtime": 34.7054,
      "eval_samples_per_second": 19.709,
      "step": 25500
    },
    {
      "epoch": 6.09,
      "learning_rate": 1e-05,
      "loss": 1.4403,
      "step": 25510
    },
    {
      "epoch": 6.1,
      "learning_rate": 1e-05,
      "loss": 1.4425,
      "step": 25520
    },
    {
      "epoch": 6.1,
      "learning_rate": 1e-05,
      "loss": 1.441,
      "step": 25530
    },
    {
      "epoch": 6.1,
      "learning_rate": 1e-05,
      "loss": 1.4444,
      "step": 25540
    },
    {
      "epoch": 6.1,
      "learning_rate": 1e-05,
      "loss": 1.4504,
      "step": 25550
    },
    {
      "epoch": 6.1,
      "learning_rate": 1e-05,
      "loss": 1.441,
      "step": 25560
    },
    {
      "epoch": 6.11,
      "learning_rate": 1e-05,
      "loss": 1.4385,
      "step": 25570
    },
    {
      "epoch": 6.11,
      "learning_rate": 1e-05,
      "loss": 1.4566,
      "step": 25580
    },
    {
      "epoch": 6.11,
      "learning_rate": 1e-05,
      "loss": 1.4545,
      "step": 25590
    },
    {
      "epoch": 6.11,
      "learning_rate": 1e-05,
      "loss": 1.4394,
      "step": 25600
    },
    {
      "epoch": 6.11,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4362614154815674,
      "eval_runtime": 34.8241,
      "eval_samples_per_second": 19.642,
      "step": 25600
    },
    {
      "epoch": 6.12,
      "learning_rate": 1e-05,
      "loss": 1.4434,
      "step": 25610
    },
    {
      "epoch": 6.12,
      "learning_rate": 1e-05,
      "loss": 1.4456,
      "step": 25620
    },
    {
      "epoch": 6.12,
      "learning_rate": 1e-05,
      "loss": 1.4578,
      "step": 25630
    },
    {
      "epoch": 6.12,
      "learning_rate": 1e-05,
      "loss": 1.4446,
      "step": 25640
    },
    {
      "epoch": 6.13,
      "learning_rate": 1e-05,
      "loss": 1.4397,
      "step": 25650
    },
    {
      "epoch": 6.13,
      "learning_rate": 1e-05,
      "loss": 1.4405,
      "step": 25660
    },
    {
      "epoch": 6.13,
      "learning_rate": 1e-05,
      "loss": 1.4468,
      "step": 25670
    },
    {
      "epoch": 6.13,
      "learning_rate": 1e-05,
      "loss": 1.4913,
      "step": 25680
    },
    {
      "epoch": 6.14,
      "learning_rate": 1e-05,
      "loss": 1.4379,
      "step": 25690
    },
    {
      "epoch": 6.14,
      "learning_rate": 1e-05,
      "loss": 1.4357,
      "step": 25700
    },
    {
      "epoch": 6.14,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4361763000488281,
      "eval_runtime": 34.5195,
      "eval_samples_per_second": 19.815,
      "step": 25700
    },
    {
      "epoch": 6.14,
      "learning_rate": 1e-05,
      "loss": 1.4372,
      "step": 25710
    },
    {
      "epoch": 6.14,
      "learning_rate": 1e-05,
      "loss": 1.4486,
      "step": 25720
    },
    {
      "epoch": 6.15,
      "learning_rate": 1e-05,
      "loss": 1.4441,
      "step": 25730
    },
    {
      "epoch": 6.15,
      "learning_rate": 1e-05,
      "loss": 1.4421,
      "step": 25740
    },
    {
      "epoch": 6.15,
      "learning_rate": 1e-05,
      "loss": 1.4421,
      "step": 25750
    },
    {
      "epoch": 6.15,
      "learning_rate": 1e-05,
      "loss": 1.4459,
      "step": 25760
    },
    {
      "epoch": 6.15,
      "learning_rate": 1e-05,
      "loss": 1.4463,
      "step": 25770
    },
    {
      "epoch": 6.16,
      "learning_rate": 1e-05,
      "loss": 1.4392,
      "step": 25780
    },
    {
      "epoch": 6.16,
      "learning_rate": 1e-05,
      "loss": 1.4394,
      "step": 25790
    },
    {
      "epoch": 6.16,
      "learning_rate": 1e-05,
      "loss": 1.4461,
      "step": 25800
    },
    {
      "epoch": 6.16,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.436369776725769,
      "eval_runtime": 35.0164,
      "eval_samples_per_second": 19.534,
      "step": 25800
    },
    {
      "epoch": 6.16,
      "learning_rate": 1e-05,
      "loss": 1.4414,
      "step": 25810
    },
    {
      "epoch": 6.17,
      "learning_rate": 1e-05,
      "loss": 1.4393,
      "step": 25820
    },
    {
      "epoch": 6.17,
      "learning_rate": 1e-05,
      "loss": 1.441,
      "step": 25830
    },
    {
      "epoch": 6.17,
      "learning_rate": 1e-05,
      "loss": 1.447,
      "step": 25840
    },
    {
      "epoch": 6.17,
      "learning_rate": 1e-05,
      "loss": 1.4505,
      "step": 25850
    },
    {
      "epoch": 6.18,
      "learning_rate": 1e-05,
      "loss": 1.445,
      "step": 25860
    },
    {
      "epoch": 6.18,
      "learning_rate": 1e-05,
      "loss": 1.4396,
      "step": 25870
    },
    {
      "epoch": 6.18,
      "learning_rate": 1e-05,
      "loss": 1.439,
      "step": 25880
    },
    {
      "epoch": 6.18,
      "learning_rate": 1e-05,
      "loss": 1.4424,
      "step": 25890
    },
    {
      "epoch": 6.19,
      "learning_rate": 1e-05,
      "loss": 1.4492,
      "step": 25900
    },
    {
      "epoch": 6.19,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4358834028244019,
      "eval_runtime": 34.3364,
      "eval_samples_per_second": 19.921,
      "step": 25900
    },
    {
      "epoch": 6.19,
      "learning_rate": 1e-05,
      "loss": 1.4436,
      "step": 25910
    },
    {
      "epoch": 6.19,
      "learning_rate": 1e-05,
      "loss": 1.4337,
      "step": 25920
    },
    {
      "epoch": 6.19,
      "learning_rate": 1e-05,
      "loss": 1.4497,
      "step": 25930
    },
    {
      "epoch": 6.2,
      "learning_rate": 1e-05,
      "loss": 1.4531,
      "step": 25940
    },
    {
      "epoch": 6.2,
      "learning_rate": 1e-05,
      "loss": 1.4413,
      "step": 25950
    },
    {
      "epoch": 6.2,
      "learning_rate": 1e-05,
      "loss": 1.45,
      "step": 25960
    },
    {
      "epoch": 6.2,
      "learning_rate": 1e-05,
      "loss": 1.4514,
      "step": 25970
    },
    {
      "epoch": 6.2,
      "learning_rate": 1e-05,
      "loss": 1.4469,
      "step": 25980
    },
    {
      "epoch": 6.21,
      "learning_rate": 1e-05,
      "loss": 1.4398,
      "step": 25990
    },
    {
      "epoch": 6.21,
      "learning_rate": 1e-05,
      "loss": 1.4433,
      "step": 26000
    },
    {
      "epoch": 6.21,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4368946552276611,
      "eval_runtime": 34.4394,
      "eval_samples_per_second": 19.861,
      "step": 26000
    },
    {
      "epoch": 6.21,
      "learning_rate": 1e-05,
      "loss": 1.4488,
      "step": 26010
    },
    {
      "epoch": 6.21,
      "learning_rate": 1e-05,
      "loss": 1.4488,
      "step": 26020
    },
    {
      "epoch": 6.22,
      "learning_rate": 1e-05,
      "loss": 1.4503,
      "step": 26030
    },
    {
      "epoch": 6.22,
      "learning_rate": 1e-05,
      "loss": 1.4402,
      "step": 26040
    },
    {
      "epoch": 6.22,
      "learning_rate": 1e-05,
      "loss": 1.443,
      "step": 26050
    },
    {
      "epoch": 6.22,
      "learning_rate": 1e-05,
      "loss": 1.4414,
      "step": 26060
    },
    {
      "epoch": 6.23,
      "learning_rate": 1e-05,
      "loss": 1.4396,
      "step": 26070
    },
    {
      "epoch": 6.23,
      "learning_rate": 1e-05,
      "loss": 1.4423,
      "step": 26080
    },
    {
      "epoch": 6.23,
      "learning_rate": 1e-05,
      "loss": 1.4535,
      "step": 26090
    },
    {
      "epoch": 6.23,
      "learning_rate": 1e-05,
      "loss": 1.4433,
      "step": 26100
    },
    {
      "epoch": 6.23,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4354615211486816,
      "eval_runtime": 34.7271,
      "eval_samples_per_second": 19.696,
      "step": 26100
    },
    {
      "epoch": 6.24,
      "learning_rate": 1e-05,
      "loss": 1.439,
      "step": 26110
    },
    {
      "epoch": 6.24,
      "learning_rate": 1e-05,
      "loss": 1.4472,
      "step": 26120
    },
    {
      "epoch": 6.24,
      "learning_rate": 1e-05,
      "loss": 1.4475,
      "step": 26130
    },
    {
      "epoch": 6.24,
      "learning_rate": 1e-05,
      "loss": 1.4435,
      "step": 26140
    },
    {
      "epoch": 6.25,
      "learning_rate": 1e-05,
      "loss": 1.4492,
      "step": 26150
    },
    {
      "epoch": 6.25,
      "learning_rate": 1e-05,
      "loss": 1.453,
      "step": 26160
    },
    {
      "epoch": 6.25,
      "learning_rate": 1e-05,
      "loss": 1.4541,
      "step": 26170
    },
    {
      "epoch": 6.25,
      "learning_rate": 1e-05,
      "loss": 1.4484,
      "step": 26180
    },
    {
      "epoch": 6.26,
      "learning_rate": 1e-05,
      "loss": 1.4393,
      "step": 26190
    },
    {
      "epoch": 6.26,
      "learning_rate": 1e-05,
      "loss": 1.4415,
      "step": 26200
    },
    {
      "epoch": 6.26,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.434644341468811,
      "eval_runtime": 34.0632,
      "eval_samples_per_second": 20.08,
      "step": 26200
    },
    {
      "epoch": 6.26,
      "learning_rate": 1e-05,
      "loss": 1.4312,
      "step": 26210
    },
    {
      "epoch": 6.26,
      "learning_rate": 1e-05,
      "loss": 1.4549,
      "step": 26220
    },
    {
      "epoch": 6.26,
      "learning_rate": 1e-05,
      "loss": 1.4444,
      "step": 26230
    },
    {
      "epoch": 6.27,
      "learning_rate": 1e-05,
      "loss": 1.437,
      "step": 26240
    },
    {
      "epoch": 6.27,
      "learning_rate": 1e-05,
      "loss": 1.4434,
      "step": 26250
    },
    {
      "epoch": 6.27,
      "learning_rate": 1e-05,
      "loss": 1.4489,
      "step": 26260
    },
    {
      "epoch": 6.27,
      "learning_rate": 1e-05,
      "loss": 1.4375,
      "step": 26270
    },
    {
      "epoch": 6.28,
      "learning_rate": 1e-05,
      "loss": 1.4373,
      "step": 26280
    },
    {
      "epoch": 6.28,
      "learning_rate": 1e-05,
      "loss": 1.44,
      "step": 26290
    },
    {
      "epoch": 6.28,
      "learning_rate": 1e-05,
      "loss": 1.445,
      "step": 26300
    },
    {
      "epoch": 6.28,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4351495504379272,
      "eval_runtime": 34.4253,
      "eval_samples_per_second": 19.869,
      "step": 26300
    },
    {
      "epoch": 6.28,
      "learning_rate": 1e-05,
      "loss": 1.4472,
      "step": 26310
    },
    {
      "epoch": 6.29,
      "learning_rate": 1e-05,
      "loss": 1.4478,
      "step": 26320
    },
    {
      "epoch": 6.29,
      "learning_rate": 1e-05,
      "loss": 1.4468,
      "step": 26330
    },
    {
      "epoch": 6.29,
      "learning_rate": 1e-05,
      "loss": 1.4343,
      "step": 26340
    },
    {
      "epoch": 6.29,
      "learning_rate": 1e-05,
      "loss": 1.435,
      "step": 26350
    },
    {
      "epoch": 6.3,
      "learning_rate": 1e-05,
      "loss": 1.4411,
      "step": 26360
    },
    {
      "epoch": 6.3,
      "learning_rate": 1e-05,
      "loss": 1.4545,
      "step": 26370
    },
    {
      "epoch": 6.3,
      "learning_rate": 1e-05,
      "loss": 1.4415,
      "step": 26380
    },
    {
      "epoch": 6.3,
      "learning_rate": 1e-05,
      "loss": 1.4436,
      "step": 26390
    },
    {
      "epoch": 6.31,
      "learning_rate": 1e-05,
      "loss": 1.4492,
      "step": 26400
    },
    {
      "epoch": 6.31,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4349793195724487,
      "eval_runtime": 34.3393,
      "eval_samples_per_second": 19.919,
      "step": 26400
    },
    {
      "epoch": 6.31,
      "learning_rate": 1e-05,
      "loss": 1.4502,
      "step": 26410
    },
    {
      "epoch": 6.31,
      "learning_rate": 1e-05,
      "loss": 1.4402,
      "step": 26420
    },
    {
      "epoch": 6.31,
      "learning_rate": 1e-05,
      "loss": 1.4434,
      "step": 26430
    },
    {
      "epoch": 6.31,
      "learning_rate": 1e-05,
      "loss": 1.4434,
      "step": 26440
    },
    {
      "epoch": 6.32,
      "learning_rate": 1e-05,
      "loss": 1.4386,
      "step": 26450
    },
    {
      "epoch": 6.32,
      "learning_rate": 1e-05,
      "loss": 1.4404,
      "step": 26460
    },
    {
      "epoch": 6.32,
      "learning_rate": 1e-05,
      "loss": 1.4396,
      "step": 26470
    },
    {
      "epoch": 6.32,
      "learning_rate": 1e-05,
      "loss": 1.4416,
      "step": 26480
    },
    {
      "epoch": 6.33,
      "learning_rate": 1e-05,
      "loss": 1.4466,
      "step": 26490
    },
    {
      "epoch": 6.33,
      "learning_rate": 1e-05,
      "loss": 1.4422,
      "step": 26500
    },
    {
      "epoch": 6.33,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4353151321411133,
      "eval_runtime": 34.2911,
      "eval_samples_per_second": 19.947,
      "step": 26500
    },
    {
      "epoch": 6.33,
      "learning_rate": 1e-05,
      "loss": 1.4459,
      "step": 26510
    },
    {
      "epoch": 6.33,
      "learning_rate": 1e-05,
      "loss": 1.4372,
      "step": 26520
    },
    {
      "epoch": 6.34,
      "learning_rate": 1e-05,
      "loss": 1.4379,
      "step": 26530
    },
    {
      "epoch": 6.34,
      "learning_rate": 1e-05,
      "loss": 1.4465,
      "step": 26540
    },
    {
      "epoch": 6.34,
      "learning_rate": 1e-05,
      "loss": 1.4405,
      "step": 26550
    },
    {
      "epoch": 6.34,
      "learning_rate": 1e-05,
      "loss": 1.4377,
      "step": 26560
    },
    {
      "epoch": 6.35,
      "learning_rate": 1e-05,
      "loss": 1.4462,
      "step": 26570
    },
    {
      "epoch": 6.35,
      "learning_rate": 1e-05,
      "loss": 1.4369,
      "step": 26580
    },
    {
      "epoch": 6.35,
      "learning_rate": 1e-05,
      "loss": 1.4367,
      "step": 26590
    },
    {
      "epoch": 6.35,
      "learning_rate": 1e-05,
      "loss": 1.4422,
      "step": 26600
    },
    {
      "epoch": 6.35,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4341922998428345,
      "eval_runtime": 34.5217,
      "eval_samples_per_second": 19.814,
      "step": 26600
    },
    {
      "epoch": 6.36,
      "learning_rate": 1e-05,
      "loss": 1.4401,
      "step": 26610
    },
    {
      "epoch": 6.36,
      "learning_rate": 1e-05,
      "loss": 1.4349,
      "step": 26620
    },
    {
      "epoch": 6.36,
      "learning_rate": 1e-05,
      "loss": 1.4411,
      "step": 26630
    },
    {
      "epoch": 6.36,
      "learning_rate": 1e-05,
      "loss": 1.4369,
      "step": 26640
    },
    {
      "epoch": 6.36,
      "learning_rate": 1e-05,
      "loss": 1.4402,
      "step": 26650
    },
    {
      "epoch": 6.37,
      "learning_rate": 1e-05,
      "loss": 1.441,
      "step": 26660
    },
    {
      "epoch": 6.37,
      "learning_rate": 1e-05,
      "loss": 1.444,
      "step": 26670
    },
    {
      "epoch": 6.37,
      "learning_rate": 1e-05,
      "loss": 1.4413,
      "step": 26680
    },
    {
      "epoch": 6.37,
      "learning_rate": 1e-05,
      "loss": 1.4417,
      "step": 26690
    },
    {
      "epoch": 6.38,
      "learning_rate": 1e-05,
      "loss": 1.4412,
      "step": 26700
    },
    {
      "epoch": 6.38,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4348976612091064,
      "eval_runtime": 34.5379,
      "eval_samples_per_second": 19.804,
      "step": 26700
    },
    {
      "epoch": 6.38,
      "learning_rate": 1e-05,
      "loss": 1.438,
      "step": 26710
    },
    {
      "epoch": 6.38,
      "learning_rate": 1e-05,
      "loss": 1.4484,
      "step": 26720
    },
    {
      "epoch": 6.38,
      "learning_rate": 1e-05,
      "loss": 1.4376,
      "step": 26730
    },
    {
      "epoch": 6.39,
      "learning_rate": 1e-05,
      "loss": 1.4421,
      "step": 26740
    },
    {
      "epoch": 6.39,
      "learning_rate": 1e-05,
      "loss": 1.4434,
      "step": 26750
    },
    {
      "epoch": 6.39,
      "learning_rate": 1e-05,
      "loss": 1.447,
      "step": 26760
    },
    {
      "epoch": 6.39,
      "learning_rate": 1e-05,
      "loss": 1.4351,
      "step": 26770
    },
    {
      "epoch": 6.4,
      "learning_rate": 1e-05,
      "loss": 1.4388,
      "step": 26780
    },
    {
      "epoch": 6.4,
      "learning_rate": 1e-05,
      "loss": 1.448,
      "step": 26790
    },
    {
      "epoch": 6.4,
      "learning_rate": 1e-05,
      "loss": 1.4439,
      "step": 26800
    },
    {
      "epoch": 6.4,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4351774454116821,
      "eval_runtime": 34.3521,
      "eval_samples_per_second": 19.911,
      "step": 26800
    },
    {
      "epoch": 6.4,
      "learning_rate": 1e-05,
      "loss": 1.4403,
      "step": 26810
    },
    {
      "epoch": 6.41,
      "learning_rate": 1e-05,
      "loss": 1.4449,
      "step": 26820
    },
    {
      "epoch": 6.41,
      "learning_rate": 1e-05,
      "loss": 1.4418,
      "step": 26830
    },
    {
      "epoch": 6.41,
      "learning_rate": 1e-05,
      "loss": 1.4471,
      "step": 26840
    },
    {
      "epoch": 6.41,
      "learning_rate": 1e-05,
      "loss": 1.4395,
      "step": 26850
    },
    {
      "epoch": 6.42,
      "learning_rate": 1e-05,
      "loss": 1.45,
      "step": 26860
    },
    {
      "epoch": 6.42,
      "learning_rate": 1e-05,
      "loss": 1.4451,
      "step": 26870
    },
    {
      "epoch": 6.42,
      "learning_rate": 1e-05,
      "loss": 1.4428,
      "step": 26880
    },
    {
      "epoch": 6.42,
      "learning_rate": 1e-05,
      "loss": 1.4443,
      "step": 26890
    },
    {
      "epoch": 6.42,
      "learning_rate": 1e-05,
      "loss": 1.4647,
      "step": 26900
    },
    {
      "epoch": 6.42,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4335548877716064,
      "eval_runtime": 34.5875,
      "eval_samples_per_second": 19.776,
      "step": 26900
    },
    {
      "epoch": 6.43,
      "learning_rate": 1e-05,
      "loss": 1.4387,
      "step": 26910
    },
    {
      "epoch": 6.43,
      "learning_rate": 1e-05,
      "loss": 1.4372,
      "step": 26920
    },
    {
      "epoch": 6.43,
      "learning_rate": 1e-05,
      "loss": 1.4381,
      "step": 26930
    },
    {
      "epoch": 6.43,
      "learning_rate": 1e-05,
      "loss": 1.4411,
      "step": 26940
    },
    {
      "epoch": 6.44,
      "learning_rate": 1e-05,
      "loss": 1.4388,
      "step": 26950
    },
    {
      "epoch": 6.44,
      "learning_rate": 1e-05,
      "loss": 1.4463,
      "step": 26960
    },
    {
      "epoch": 6.44,
      "learning_rate": 1e-05,
      "loss": 1.4395,
      "step": 26970
    },
    {
      "epoch": 6.44,
      "learning_rate": 1e-05,
      "loss": 1.4352,
      "step": 26980
    },
    {
      "epoch": 6.45,
      "learning_rate": 1e-05,
      "loss": 1.4694,
      "step": 26990
    },
    {
      "epoch": 6.45,
      "learning_rate": 1e-05,
      "loss": 1.4499,
      "step": 27000
    },
    {
      "epoch": 6.45,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4350945949554443,
      "eval_runtime": 34.4376,
      "eval_samples_per_second": 19.862,
      "step": 27000
    },
    {
      "epoch": 6.45,
      "learning_rate": 1e-05,
      "loss": 1.4441,
      "step": 27010
    },
    {
      "epoch": 6.45,
      "learning_rate": 1e-05,
      "loss": 1.4466,
      "step": 27020
    },
    {
      "epoch": 6.46,
      "learning_rate": 1e-05,
      "loss": 1.4402,
      "step": 27030
    },
    {
      "epoch": 6.46,
      "learning_rate": 1e-05,
      "loss": 1.4444,
      "step": 27040
    },
    {
      "epoch": 6.46,
      "learning_rate": 1e-05,
      "loss": 1.4429,
      "step": 27050
    },
    {
      "epoch": 6.46,
      "learning_rate": 1e-05,
      "loss": 1.4455,
      "step": 27060
    },
    {
      "epoch": 6.47,
      "learning_rate": 1e-05,
      "loss": 1.4474,
      "step": 27070
    },
    {
      "epoch": 6.47,
      "learning_rate": 1e-05,
      "loss": 1.436,
      "step": 27080
    },
    {
      "epoch": 6.47,
      "learning_rate": 1e-05,
      "loss": 1.4354,
      "step": 27090
    },
    {
      "epoch": 6.47,
      "learning_rate": 1e-05,
      "loss": 1.4463,
      "step": 27100
    },
    {
      "epoch": 6.47,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4351810216903687,
      "eval_runtime": 34.5111,
      "eval_samples_per_second": 19.82,
      "step": 27100
    },
    {
      "epoch": 6.47,
      "learning_rate": 1e-05,
      "loss": 1.4369,
      "step": 27110
    },
    {
      "epoch": 6.48,
      "learning_rate": 1e-05,
      "loss": 1.4341,
      "step": 27120
    },
    {
      "epoch": 6.48,
      "learning_rate": 1e-05,
      "loss": 1.4415,
      "step": 27130
    },
    {
      "epoch": 6.48,
      "learning_rate": 1e-05,
      "loss": 1.4457,
      "step": 27140
    },
    {
      "epoch": 6.48,
      "learning_rate": 1e-05,
      "loss": 1.4469,
      "step": 27150
    },
    {
      "epoch": 6.49,
      "learning_rate": 1e-05,
      "loss": 1.4394,
      "step": 27160
    },
    {
      "epoch": 6.49,
      "learning_rate": 1e-05,
      "loss": 1.4342,
      "step": 27170
    },
    {
      "epoch": 6.49,
      "learning_rate": 1e-05,
      "loss": 1.4461,
      "step": 27180
    },
    {
      "epoch": 6.49,
      "learning_rate": 1e-05,
      "loss": 1.4471,
      "step": 27190
    },
    {
      "epoch": 6.5,
      "learning_rate": 1e-05,
      "loss": 1.4387,
      "step": 27200
    },
    {
      "epoch": 6.5,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4337577819824219,
      "eval_runtime": 34.5376,
      "eval_samples_per_second": 19.804,
      "step": 27200
    },
    {
      "epoch": 6.5,
      "learning_rate": 1e-05,
      "loss": 1.446,
      "step": 27210
    },
    {
      "epoch": 6.5,
      "learning_rate": 1e-05,
      "loss": 1.438,
      "step": 27220
    },
    {
      "epoch": 6.5,
      "learning_rate": 1e-05,
      "loss": 1.4363,
      "step": 27230
    },
    {
      "epoch": 6.51,
      "learning_rate": 1e-05,
      "loss": 1.4468,
      "step": 27240
    },
    {
      "epoch": 6.51,
      "learning_rate": 1e-05,
      "loss": 1.4398,
      "step": 27250
    },
    {
      "epoch": 6.51,
      "learning_rate": 1e-05,
      "loss": 1.4408,
      "step": 27260
    },
    {
      "epoch": 6.51,
      "learning_rate": 1e-05,
      "loss": 1.4419,
      "step": 27270
    },
    {
      "epoch": 6.52,
      "learning_rate": 1e-05,
      "loss": 1.4424,
      "step": 27280
    },
    {
      "epoch": 6.52,
      "learning_rate": 1e-05,
      "loss": 1.4404,
      "step": 27290
    },
    {
      "epoch": 6.52,
      "learning_rate": 1e-05,
      "loss": 1.4422,
      "step": 27300
    },
    {
      "epoch": 6.52,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4341362714767456,
      "eval_runtime": 34.6906,
      "eval_samples_per_second": 19.717,
      "step": 27300
    },
    {
      "epoch": 6.52,
      "learning_rate": 1e-05,
      "loss": 1.4452,
      "step": 27310
    },
    {
      "epoch": 6.52,
      "learning_rate": 1e-05,
      "loss": 1.4455,
      "step": 27320
    },
    {
      "epoch": 6.53,
      "learning_rate": 1e-05,
      "loss": 1.4361,
      "step": 27330
    },
    {
      "epoch": 6.53,
      "learning_rate": 1e-05,
      "loss": 1.4399,
      "step": 27340
    },
    {
      "epoch": 6.53,
      "learning_rate": 1e-05,
      "loss": 1.4335,
      "step": 27350
    },
    {
      "epoch": 6.53,
      "learning_rate": 1e-05,
      "loss": 1.4504,
      "step": 27360
    },
    {
      "epoch": 6.54,
      "learning_rate": 1e-05,
      "loss": 1.436,
      "step": 27370
    },
    {
      "epoch": 6.54,
      "learning_rate": 1e-05,
      "loss": 1.4374,
      "step": 27380
    },
    {
      "epoch": 6.54,
      "learning_rate": 1e-05,
      "loss": 1.4612,
      "step": 27390
    },
    {
      "epoch": 6.54,
      "learning_rate": 1e-05,
      "loss": 1.4348,
      "step": 27400
    },
    {
      "epoch": 6.54,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4325109720230103,
      "eval_runtime": 34.8022,
      "eval_samples_per_second": 19.654,
      "step": 27400
    },
    {
      "epoch": 6.55,
      "learning_rate": 1e-05,
      "loss": 1.4361,
      "step": 27410
    },
    {
      "epoch": 6.55,
      "learning_rate": 1e-05,
      "loss": 1.4395,
      "step": 27420
    },
    {
      "epoch": 6.55,
      "learning_rate": 1e-05,
      "loss": 1.4356,
      "step": 27430
    },
    {
      "epoch": 6.55,
      "learning_rate": 1e-05,
      "loss": 1.4473,
      "step": 27440
    },
    {
      "epoch": 6.56,
      "learning_rate": 1e-05,
      "loss": 1.4377,
      "step": 27450
    },
    {
      "epoch": 6.56,
      "learning_rate": 1e-05,
      "loss": 1.4303,
      "step": 27460
    },
    {
      "epoch": 6.56,
      "learning_rate": 1e-05,
      "loss": 1.4468,
      "step": 27470
    },
    {
      "epoch": 6.56,
      "learning_rate": 1e-05,
      "loss": 1.4373,
      "step": 27480
    },
    {
      "epoch": 6.57,
      "learning_rate": 1e-05,
      "loss": 1.4817,
      "step": 27490
    },
    {
      "epoch": 6.57,
      "learning_rate": 1e-05,
      "loss": 1.4393,
      "step": 27500
    },
    {
      "epoch": 6.57,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4341787099838257,
      "eval_runtime": 34.2806,
      "eval_samples_per_second": 19.953,
      "step": 27500
    },
    {
      "epoch": 6.57,
      "learning_rate": 1e-05,
      "loss": 1.4417,
      "step": 27510
    },
    {
      "epoch": 6.57,
      "learning_rate": 1e-05,
      "loss": 1.4382,
      "step": 27520
    },
    {
      "epoch": 6.58,
      "learning_rate": 1e-05,
      "loss": 1.4695,
      "step": 27530
    },
    {
      "epoch": 6.58,
      "learning_rate": 1e-05,
      "loss": 1.4369,
      "step": 27540
    },
    {
      "epoch": 6.58,
      "learning_rate": 1e-05,
      "loss": 1.4431,
      "step": 27550
    },
    {
      "epoch": 6.58,
      "learning_rate": 1e-05,
      "loss": 1.44,
      "step": 27560
    },
    {
      "epoch": 6.58,
      "learning_rate": 1e-05,
      "loss": 1.4423,
      "step": 27570
    },
    {
      "epoch": 6.59,
      "learning_rate": 1e-05,
      "loss": 1.4428,
      "step": 27580
    },
    {
      "epoch": 6.59,
      "learning_rate": 1e-05,
      "loss": 1.4392,
      "step": 27590
    },
    {
      "epoch": 6.59,
      "learning_rate": 1e-05,
      "loss": 1.4397,
      "step": 27600
    },
    {
      "epoch": 6.59,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4349558353424072,
      "eval_runtime": 34.1246,
      "eval_samples_per_second": 20.044,
      "step": 27600
    },
    {
      "epoch": 6.59,
      "learning_rate": 1e-05,
      "loss": 1.4412,
      "step": 27610
    },
    {
      "epoch": 6.6,
      "learning_rate": 1e-05,
      "loss": 1.4394,
      "step": 27620
    },
    {
      "epoch": 6.6,
      "learning_rate": 1e-05,
      "loss": 1.4359,
      "step": 27630
    },
    {
      "epoch": 6.6,
      "learning_rate": 1e-05,
      "loss": 1.4373,
      "step": 27640
    },
    {
      "epoch": 6.6,
      "learning_rate": 1e-05,
      "loss": 1.4444,
      "step": 27650
    },
    {
      "epoch": 6.61,
      "learning_rate": 1e-05,
      "loss": 1.4451,
      "step": 27660
    },
    {
      "epoch": 6.61,
      "learning_rate": 1e-05,
      "loss": 1.4335,
      "step": 27670
    },
    {
      "epoch": 6.61,
      "learning_rate": 1e-05,
      "loss": 1.4428,
      "step": 27680
    },
    {
      "epoch": 6.61,
      "learning_rate": 1e-05,
      "loss": 1.4397,
      "step": 27690
    },
    {
      "epoch": 6.62,
      "learning_rate": 1e-05,
      "loss": 1.4395,
      "step": 27700
    },
    {
      "epoch": 6.62,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4324934482574463,
      "eval_runtime": 34.2324,
      "eval_samples_per_second": 19.981,
      "step": 27700
    },
    {
      "epoch": 6.62,
      "learning_rate": 1e-05,
      "loss": 1.4506,
      "step": 27710
    },
    {
      "epoch": 6.62,
      "learning_rate": 1e-05,
      "loss": 1.444,
      "step": 27720
    },
    {
      "epoch": 6.62,
      "learning_rate": 1e-05,
      "loss": 1.4418,
      "step": 27730
    },
    {
      "epoch": 6.63,
      "learning_rate": 1e-05,
      "loss": 1.4485,
      "step": 27740
    },
    {
      "epoch": 6.63,
      "learning_rate": 1e-05,
      "loss": 1.4366,
      "step": 27750
    },
    {
      "epoch": 6.63,
      "learning_rate": 1e-05,
      "loss": 1.44,
      "step": 27760
    },
    {
      "epoch": 6.63,
      "learning_rate": 1e-05,
      "loss": 1.4364,
      "step": 27770
    },
    {
      "epoch": 6.63,
      "learning_rate": 1e-05,
      "loss": 1.452,
      "step": 27780
    },
    {
      "epoch": 6.64,
      "learning_rate": 1e-05,
      "loss": 1.4436,
      "step": 27790
    },
    {
      "epoch": 6.64,
      "learning_rate": 1e-05,
      "loss": 1.4365,
      "step": 27800
    },
    {
      "epoch": 6.64,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4337815046310425,
      "eval_runtime": 34.4109,
      "eval_samples_per_second": 19.877,
      "step": 27800
    },
    {
      "epoch": 6.64,
      "learning_rate": 1e-05,
      "loss": 1.441,
      "step": 27810
    },
    {
      "epoch": 6.64,
      "learning_rate": 1e-05,
      "loss": 1.4392,
      "step": 27820
    },
    {
      "epoch": 6.65,
      "learning_rate": 1e-05,
      "loss": 1.4359,
      "step": 27830
    },
    {
      "epoch": 6.65,
      "learning_rate": 1e-05,
      "loss": 1.4388,
      "step": 27840
    },
    {
      "epoch": 6.65,
      "learning_rate": 1e-05,
      "loss": 1.4416,
      "step": 27850
    },
    {
      "epoch": 6.65,
      "learning_rate": 1e-05,
      "loss": 1.4468,
      "step": 27860
    },
    {
      "epoch": 6.66,
      "learning_rate": 1e-05,
      "loss": 1.4375,
      "step": 27870
    },
    {
      "epoch": 6.66,
      "learning_rate": 1e-05,
      "loss": 1.4326,
      "step": 27880
    },
    {
      "epoch": 6.66,
      "learning_rate": 1e-05,
      "loss": 1.4342,
      "step": 27890
    },
    {
      "epoch": 6.66,
      "learning_rate": 1e-05,
      "loss": 1.4381,
      "step": 27900
    },
    {
      "epoch": 6.66,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.433609127998352,
      "eval_runtime": 34.568,
      "eval_samples_per_second": 19.787,
      "step": 27900
    },
    {
      "epoch": 6.67,
      "learning_rate": 1e-05,
      "loss": 1.4494,
      "step": 27910
    },
    {
      "epoch": 6.67,
      "learning_rate": 1e-05,
      "loss": 1.4423,
      "step": 27920
    },
    {
      "epoch": 6.67,
      "learning_rate": 1e-05,
      "loss": 1.4397,
      "step": 27930
    },
    {
      "epoch": 6.67,
      "learning_rate": 1e-05,
      "loss": 1.4389,
      "step": 27940
    },
    {
      "epoch": 6.68,
      "learning_rate": 1e-05,
      "loss": 1.4451,
      "step": 27950
    },
    {
      "epoch": 6.68,
      "learning_rate": 1e-05,
      "loss": 1.4423,
      "step": 27960
    },
    {
      "epoch": 6.68,
      "learning_rate": 1e-05,
      "loss": 1.4325,
      "step": 27970
    },
    {
      "epoch": 6.68,
      "learning_rate": 1e-05,
      "loss": 1.4411,
      "step": 27980
    },
    {
      "epoch": 6.68,
      "learning_rate": 1e-05,
      "loss": 1.449,
      "step": 27990
    },
    {
      "epoch": 6.69,
      "learning_rate": 1e-05,
      "loss": 1.4374,
      "step": 28000
    },
    {
      "epoch": 6.69,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4343637228012085,
      "eval_runtime": 34.4701,
      "eval_samples_per_second": 19.843,
      "step": 28000
    },
    {
      "epoch": 6.69,
      "learning_rate": 1e-05,
      "loss": 1.4377,
      "step": 28010
    },
    {
      "epoch": 6.69,
      "learning_rate": 1e-05,
      "loss": 1.4438,
      "step": 28020
    },
    {
      "epoch": 6.69,
      "learning_rate": 1e-05,
      "loss": 1.4555,
      "step": 28030
    },
    {
      "epoch": 6.7,
      "learning_rate": 1e-05,
      "loss": 1.4645,
      "step": 28040
    },
    {
      "epoch": 6.7,
      "learning_rate": 1e-05,
      "loss": 1.4319,
      "step": 28050
    },
    {
      "epoch": 6.7,
      "learning_rate": 1e-05,
      "loss": 1.4336,
      "step": 28060
    },
    {
      "epoch": 6.7,
      "learning_rate": 1e-05,
      "loss": 1.4422,
      "step": 28070
    },
    {
      "epoch": 6.71,
      "learning_rate": 1e-05,
      "loss": 1.43,
      "step": 28080
    },
    {
      "epoch": 6.71,
      "learning_rate": 1e-05,
      "loss": 1.4396,
      "step": 28090
    },
    {
      "epoch": 6.71,
      "learning_rate": 1e-05,
      "loss": 1.4361,
      "step": 28100
    },
    {
      "epoch": 6.71,
      "eval_accuracy": 0.9751461988304093,
      "eval_loss": 1.4335529804229736,
      "eval_runtime": 34.121,
      "eval_samples_per_second": 20.046,
      "step": 28100
    },
    {
      "epoch": 6.71,
      "learning_rate": 1e-05,
      "loss": 1.441,
      "step": 28110
    },
    {
      "epoch": 6.72,
      "learning_rate": 1e-05,
      "loss": 1.4246,
      "step": 28120
    },
    {
      "epoch": 6.72,
      "learning_rate": 1e-05,
      "loss": 1.441,
      "step": 28130
    },
    {
      "epoch": 6.72,
      "learning_rate": 1e-05,
      "loss": 1.4408,
      "step": 28140
    },
    {
      "epoch": 6.72,
      "learning_rate": 1e-05,
      "loss": 1.4381,
      "step": 28150
    },
    {
      "epoch": 6.73,
      "learning_rate": 1e-05,
      "loss": 1.4448,
      "step": 28160
    },
    {
      "epoch": 6.73,
      "learning_rate": 1e-05,
      "loss": 1.4472,
      "step": 28170
    },
    {
      "epoch": 6.73,
      "learning_rate": 1e-05,
      "loss": 1.4339,
      "step": 28180
    },
    {
      "epoch": 6.73,
      "learning_rate": 1e-05,
      "loss": 1.4466,
      "step": 28190
    },
    {
      "epoch": 6.74,
      "learning_rate": 1e-05,
      "loss": 1.4404,
      "step": 28200
    },
    {
      "epoch": 6.74,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4336799383163452,
      "eval_runtime": 34.5663,
      "eval_samples_per_second": 19.788,
      "step": 28200
    },
    {
      "epoch": 6.74,
      "learning_rate": 1e-05,
      "loss": 1.4326,
      "step": 28210
    },
    {
      "epoch": 6.74,
      "learning_rate": 1e-05,
      "loss": 1.4414,
      "step": 28220
    },
    {
      "epoch": 6.74,
      "learning_rate": 1e-05,
      "loss": 1.433,
      "step": 28230
    },
    {
      "epoch": 6.74,
      "learning_rate": 1e-05,
      "loss": 1.4343,
      "step": 28240
    },
    {
      "epoch": 6.75,
      "learning_rate": 1e-05,
      "loss": 1.448,
      "step": 28250
    },
    {
      "epoch": 6.75,
      "learning_rate": 1e-05,
      "loss": 1.4387,
      "step": 28260
    },
    {
      "epoch": 6.75,
      "learning_rate": 1e-05,
      "loss": 1.4378,
      "step": 28270
    },
    {
      "epoch": 6.75,
      "learning_rate": 1e-05,
      "loss": 1.4354,
      "step": 28280
    },
    {
      "epoch": 6.76,
      "learning_rate": 1e-05,
      "loss": 1.4425,
      "step": 28290
    },
    {
      "epoch": 6.76,
      "learning_rate": 1e-05,
      "loss": 1.4521,
      "step": 28300
    },
    {
      "epoch": 6.76,
      "eval_accuracy": 0.9751461988304093,
      "eval_loss": 1.4323303699493408,
      "eval_runtime": 34.3455,
      "eval_samples_per_second": 19.915,
      "step": 28300
    },
    {
      "epoch": 6.76,
      "learning_rate": 1e-05,
      "loss": 1.4462,
      "step": 28310
    },
    {
      "epoch": 6.76,
      "learning_rate": 1e-05,
      "loss": 1.4499,
      "step": 28320
    },
    {
      "epoch": 6.77,
      "learning_rate": 1e-05,
      "loss": 1.4446,
      "step": 28330
    },
    {
      "epoch": 6.77,
      "learning_rate": 1e-05,
      "loss": 1.445,
      "step": 28340
    },
    {
      "epoch": 6.77,
      "learning_rate": 1e-05,
      "loss": 1.4382,
      "step": 28350
    },
    {
      "epoch": 6.77,
      "learning_rate": 1e-05,
      "loss": 1.4341,
      "step": 28360
    },
    {
      "epoch": 6.78,
      "learning_rate": 1e-05,
      "loss": 1.4396,
      "step": 28370
    },
    {
      "epoch": 6.78,
      "learning_rate": 1e-05,
      "loss": 1.4493,
      "step": 28380
    },
    {
      "epoch": 6.78,
      "learning_rate": 1e-05,
      "loss": 1.4432,
      "step": 28390
    },
    {
      "epoch": 6.78,
      "learning_rate": 1e-05,
      "loss": 1.4374,
      "step": 28400
    },
    {
      "epoch": 6.78,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4327962398529053,
      "eval_runtime": 34.4902,
      "eval_samples_per_second": 19.832,
      "step": 28400
    },
    {
      "epoch": 6.79,
      "learning_rate": 1e-05,
      "loss": 1.4614,
      "step": 28410
    },
    {
      "epoch": 6.79,
      "learning_rate": 1e-05,
      "loss": 1.4381,
      "step": 28420
    },
    {
      "epoch": 6.79,
      "learning_rate": 1e-05,
      "loss": 1.4353,
      "step": 28430
    },
    {
      "epoch": 6.79,
      "learning_rate": 1e-05,
      "loss": 1.4372,
      "step": 28440
    },
    {
      "epoch": 6.79,
      "learning_rate": 1e-05,
      "loss": 1.4301,
      "step": 28450
    },
    {
      "epoch": 6.8,
      "learning_rate": 1e-05,
      "loss": 1.4414,
      "step": 28460
    },
    {
      "epoch": 6.8,
      "learning_rate": 1e-05,
      "loss": 1.437,
      "step": 28470
    },
    {
      "epoch": 6.8,
      "learning_rate": 1e-05,
      "loss": 1.4463,
      "step": 28480
    },
    {
      "epoch": 6.8,
      "learning_rate": 1e-05,
      "loss": 1.4615,
      "step": 28490
    },
    {
      "epoch": 6.81,
      "learning_rate": 1e-05,
      "loss": 1.4312,
      "step": 28500
    },
    {
      "epoch": 6.81,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.43134343624115,
      "eval_runtime": 34.5664,
      "eval_samples_per_second": 19.788,
      "step": 28500
    },
    {
      "epoch": 6.81,
      "learning_rate": 1e-05,
      "loss": 1.4333,
      "step": 28510
    },
    {
      "epoch": 6.81,
      "learning_rate": 1e-05,
      "loss": 1.4319,
      "step": 28520
    },
    {
      "epoch": 6.81,
      "learning_rate": 1e-05,
      "loss": 1.4371,
      "step": 28530
    },
    {
      "epoch": 6.82,
      "learning_rate": 1e-05,
      "loss": 1.4436,
      "step": 28540
    },
    {
      "epoch": 6.82,
      "learning_rate": 1e-05,
      "loss": 1.4373,
      "step": 28550
    },
    {
      "epoch": 6.82,
      "learning_rate": 1e-05,
      "loss": 1.4455,
      "step": 28560
    },
    {
      "epoch": 6.82,
      "learning_rate": 1e-05,
      "loss": 1.4428,
      "step": 28570
    },
    {
      "epoch": 6.83,
      "learning_rate": 1e-05,
      "loss": 1.4366,
      "step": 28580
    },
    {
      "epoch": 6.83,
      "learning_rate": 1e-05,
      "loss": 1.4528,
      "step": 28590
    },
    {
      "epoch": 6.83,
      "learning_rate": 1e-05,
      "loss": 1.4411,
      "step": 28600
    },
    {
      "epoch": 6.83,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4321269989013672,
      "eval_runtime": 34.5341,
      "eval_samples_per_second": 19.807,
      "step": 28600
    },
    {
      "epoch": 6.83,
      "learning_rate": 1e-05,
      "loss": 1.4441,
      "step": 28610
    },
    {
      "epoch": 6.84,
      "learning_rate": 1e-05,
      "loss": 1.4473,
      "step": 28620
    },
    {
      "epoch": 6.84,
      "learning_rate": 1e-05,
      "loss": 1.4389,
      "step": 28630
    },
    {
      "epoch": 6.84,
      "learning_rate": 1e-05,
      "loss": 1.4371,
      "step": 28640
    },
    {
      "epoch": 6.84,
      "learning_rate": 1e-05,
      "loss": 1.4453,
      "step": 28650
    },
    {
      "epoch": 6.84,
      "learning_rate": 1e-05,
      "loss": 1.4278,
      "step": 28660
    },
    {
      "epoch": 6.85,
      "learning_rate": 1e-05,
      "loss": 1.4476,
      "step": 28670
    },
    {
      "epoch": 6.85,
      "learning_rate": 1e-05,
      "loss": 1.4342,
      "step": 28680
    },
    {
      "epoch": 6.85,
      "learning_rate": 1e-05,
      "loss": 1.4461,
      "step": 28690
    },
    {
      "epoch": 6.85,
      "learning_rate": 1e-05,
      "loss": 1.4393,
      "step": 28700
    },
    {
      "epoch": 6.85,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4327913522720337,
      "eval_runtime": 34.5483,
      "eval_samples_per_second": 19.798,
      "step": 28700
    },
    {
      "epoch": 6.86,
      "learning_rate": 1e-05,
      "loss": 1.4474,
      "step": 28710
    },
    {
      "epoch": 6.86,
      "learning_rate": 1e-05,
      "loss": 1.4444,
      "step": 28720
    },
    {
      "epoch": 6.86,
      "learning_rate": 1e-05,
      "loss": 1.4291,
      "step": 28730
    },
    {
      "epoch": 6.86,
      "learning_rate": 1e-05,
      "loss": 1.437,
      "step": 28740
    },
    {
      "epoch": 6.87,
      "learning_rate": 1e-05,
      "loss": 1.4409,
      "step": 28750
    },
    {
      "epoch": 6.87,
      "learning_rate": 1e-05,
      "loss": 1.4492,
      "step": 28760
    },
    {
      "epoch": 6.87,
      "learning_rate": 1e-05,
      "loss": 1.4341,
      "step": 28770
    },
    {
      "epoch": 6.87,
      "learning_rate": 1e-05,
      "loss": 1.4494,
      "step": 28780
    },
    {
      "epoch": 6.88,
      "learning_rate": 1e-05,
      "loss": 1.4386,
      "step": 28790
    },
    {
      "epoch": 6.88,
      "learning_rate": 1e-05,
      "loss": 1.4331,
      "step": 28800
    },
    {
      "epoch": 6.88,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4317677021026611,
      "eval_runtime": 34.4662,
      "eval_samples_per_second": 19.846,
      "step": 28800
    },
    {
      "epoch": 6.88,
      "learning_rate": 1e-05,
      "loss": 1.443,
      "step": 28810
    },
    {
      "epoch": 6.88,
      "learning_rate": 1e-05,
      "loss": 1.4445,
      "step": 28820
    },
    {
      "epoch": 6.89,
      "learning_rate": 1e-05,
      "loss": 1.4431,
      "step": 28830
    },
    {
      "epoch": 6.89,
      "learning_rate": 1e-05,
      "loss": 1.4539,
      "step": 28840
    },
    {
      "epoch": 6.89,
      "learning_rate": 1e-05,
      "loss": 1.4339,
      "step": 28850
    },
    {
      "epoch": 6.89,
      "learning_rate": 1e-05,
      "loss": 1.4421,
      "step": 28860
    },
    {
      "epoch": 6.9,
      "learning_rate": 1e-05,
      "loss": 1.4377,
      "step": 28870
    },
    {
      "epoch": 6.9,
      "learning_rate": 1e-05,
      "loss": 1.4368,
      "step": 28880
    },
    {
      "epoch": 6.9,
      "learning_rate": 1e-05,
      "loss": 1.4411,
      "step": 28890
    },
    {
      "epoch": 6.9,
      "learning_rate": 1e-05,
      "loss": 1.4379,
      "step": 28900
    },
    {
      "epoch": 6.9,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.431147575378418,
      "eval_runtime": 34.5011,
      "eval_samples_per_second": 19.825,
      "step": 28900
    },
    {
      "epoch": 6.9,
      "learning_rate": 1e-05,
      "loss": 1.4445,
      "step": 28910
    },
    {
      "epoch": 6.91,
      "learning_rate": 1e-05,
      "loss": 1.4438,
      "step": 28920
    },
    {
      "epoch": 6.91,
      "learning_rate": 1e-05,
      "loss": 1.4373,
      "step": 28930
    },
    {
      "epoch": 6.91,
      "learning_rate": 1e-05,
      "loss": 1.4406,
      "step": 28940
    },
    {
      "epoch": 6.91,
      "learning_rate": 1e-05,
      "loss": 1.4544,
      "step": 28950
    },
    {
      "epoch": 6.92,
      "learning_rate": 1e-05,
      "loss": 1.4348,
      "step": 28960
    },
    {
      "epoch": 6.92,
      "learning_rate": 1e-05,
      "loss": 1.4346,
      "step": 28970
    },
    {
      "epoch": 6.92,
      "learning_rate": 1e-05,
      "loss": 1.4515,
      "step": 28980
    },
    {
      "epoch": 6.92,
      "learning_rate": 1e-05,
      "loss": 1.4352,
      "step": 28990
    },
    {
      "epoch": 6.93,
      "learning_rate": 1e-05,
      "loss": 1.4297,
      "step": 29000
    },
    {
      "epoch": 6.93,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4317106008529663,
      "eval_runtime": 34.5263,
      "eval_samples_per_second": 19.811,
      "step": 29000
    },
    {
      "epoch": 6.93,
      "learning_rate": 1e-05,
      "loss": 1.4375,
      "step": 29010
    },
    {
      "epoch": 6.93,
      "learning_rate": 1e-05,
      "loss": 1.434,
      "step": 29020
    },
    {
      "epoch": 6.93,
      "learning_rate": 1e-05,
      "loss": 1.439,
      "step": 29030
    },
    {
      "epoch": 6.94,
      "learning_rate": 1e-05,
      "loss": 1.4353,
      "step": 29040
    },
    {
      "epoch": 6.94,
      "learning_rate": 1e-05,
      "loss": 1.4459,
      "step": 29050
    },
    {
      "epoch": 6.94,
      "learning_rate": 1e-05,
      "loss": 1.4297,
      "step": 29060
    },
    {
      "epoch": 6.94,
      "learning_rate": 1e-05,
      "loss": 1.4388,
      "step": 29070
    },
    {
      "epoch": 6.95,
      "learning_rate": 1e-05,
      "loss": 1.4379,
      "step": 29080
    },
    {
      "epoch": 6.95,
      "learning_rate": 1e-05,
      "loss": 1.4351,
      "step": 29090
    },
    {
      "epoch": 6.95,
      "learning_rate": 1e-05,
      "loss": 1.4325,
      "step": 29100
    },
    {
      "epoch": 6.95,
      "eval_accuracy": 0.9751461988304093,
      "eval_loss": 1.431464433670044,
      "eval_runtime": 34.5939,
      "eval_samples_per_second": 19.772,
      "step": 29100
    },
    {
      "epoch": 6.95,
      "learning_rate": 1e-05,
      "loss": 1.4415,
      "step": 29110
    },
    {
      "epoch": 6.95,
      "learning_rate": 1e-05,
      "loss": 1.4329,
      "step": 29120
    },
    {
      "epoch": 6.96,
      "learning_rate": 1e-05,
      "loss": 1.4404,
      "step": 29130
    },
    {
      "epoch": 6.96,
      "learning_rate": 1e-05,
      "loss": 1.4344,
      "step": 29140
    },
    {
      "epoch": 6.96,
      "learning_rate": 1e-05,
      "loss": 1.436,
      "step": 29150
    },
    {
      "epoch": 6.96,
      "learning_rate": 1e-05,
      "loss": 1.4316,
      "step": 29160
    },
    {
      "epoch": 6.97,
      "learning_rate": 1e-05,
      "loss": 1.4635,
      "step": 29170
    },
    {
      "epoch": 6.97,
      "learning_rate": 1e-05,
      "loss": 1.448,
      "step": 29180
    },
    {
      "epoch": 6.97,
      "learning_rate": 1e-05,
      "loss": 1.4329,
      "step": 29190
    },
    {
      "epoch": 6.97,
      "learning_rate": 1e-05,
      "loss": 1.4342,
      "step": 29200
    },
    {
      "epoch": 6.97,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4320753812789917,
      "eval_runtime": 34.5955,
      "eval_samples_per_second": 19.771,
      "step": 29200
    },
    {
      "epoch": 6.98,
      "learning_rate": 1e-05,
      "loss": 1.4403,
      "step": 29210
    },
    {
      "epoch": 6.98,
      "learning_rate": 1e-05,
      "loss": 1.4346,
      "step": 29220
    },
    {
      "epoch": 6.98,
      "learning_rate": 1e-05,
      "loss": 1.4325,
      "step": 29230
    },
    {
      "epoch": 6.98,
      "learning_rate": 1e-05,
      "loss": 1.4377,
      "step": 29240
    },
    {
      "epoch": 6.99,
      "learning_rate": 1e-05,
      "loss": 1.4349,
      "step": 29250
    },
    {
      "epoch": 6.99,
      "learning_rate": 1e-05,
      "loss": 1.4366,
      "step": 29260
    },
    {
      "epoch": 6.99,
      "learning_rate": 1e-05,
      "loss": 1.4484,
      "step": 29270
    },
    {
      "epoch": 6.99,
      "learning_rate": 1e-05,
      "loss": 1.4278,
      "step": 29280
    },
    {
      "epoch": 7.0,
      "learning_rate": 1e-05,
      "loss": 1.4396,
      "step": 29290
    },
    {
      "epoch": 7.0,
      "learning_rate": 1e-05,
      "loss": 1.431,
      "step": 29300
    },
    {
      "epoch": 7.0,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4315484762191772,
      "eval_runtime": 34.4448,
      "eval_samples_per_second": 19.858,
      "step": 29300
    },
    {
      "epoch": 7.0,
      "learning_rate": 1e-05,
      "loss": 1.4677,
      "step": 29310
    },
    {
      "epoch": 7.0,
      "learning_rate": 1e-05,
      "loss": 1.4436,
      "step": 29320
    },
    {
      "epoch": 7.01,
      "learning_rate": 1e-05,
      "loss": 1.4359,
      "step": 29330
    },
    {
      "epoch": 7.01,
      "learning_rate": 1e-05,
      "loss": 1.4453,
      "step": 29340
    },
    {
      "epoch": 7.01,
      "learning_rate": 1e-05,
      "loss": 1.4331,
      "step": 29350
    },
    {
      "epoch": 7.01,
      "learning_rate": 1e-05,
      "loss": 1.4344,
      "step": 29360
    },
    {
      "epoch": 7.01,
      "learning_rate": 1e-05,
      "loss": 1.4405,
      "step": 29370
    },
    {
      "epoch": 7.02,
      "learning_rate": 1e-05,
      "loss": 1.4351,
      "step": 29380
    },
    {
      "epoch": 7.02,
      "learning_rate": 1e-05,
      "loss": 1.4248,
      "step": 29390
    },
    {
      "epoch": 7.02,
      "learning_rate": 1e-05,
      "loss": 1.4384,
      "step": 29400
    },
    {
      "epoch": 7.02,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4316810369491577,
      "eval_runtime": 34.5773,
      "eval_samples_per_second": 19.782,
      "step": 29400
    },
    {
      "epoch": 7.02,
      "learning_rate": 1e-05,
      "loss": 1.446,
      "step": 29410
    },
    {
      "epoch": 7.03,
      "learning_rate": 1e-05,
      "loss": 1.4392,
      "step": 29420
    },
    {
      "epoch": 7.03,
      "learning_rate": 1e-05,
      "loss": 1.4379,
      "step": 29430
    },
    {
      "epoch": 7.03,
      "learning_rate": 1e-05,
      "loss": 1.4393,
      "step": 29440
    },
    {
      "epoch": 7.03,
      "learning_rate": 1e-05,
      "loss": 1.4385,
      "step": 29450
    },
    {
      "epoch": 7.04,
      "learning_rate": 1e-05,
      "loss": 1.4472,
      "step": 29460
    },
    {
      "epoch": 7.04,
      "learning_rate": 1e-05,
      "loss": 1.4457,
      "step": 29470
    },
    {
      "epoch": 7.04,
      "learning_rate": 1e-05,
      "loss": 1.5308,
      "step": 29480
    },
    {
      "epoch": 7.04,
      "learning_rate": 1e-05,
      "loss": 1.4371,
      "step": 29490
    },
    {
      "epoch": 7.05,
      "learning_rate": 1e-05,
      "loss": 1.435,
      "step": 29500
    },
    {
      "epoch": 7.05,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4307992458343506,
      "eval_runtime": 34.2423,
      "eval_samples_per_second": 19.975,
      "step": 29500
    },
    {
      "epoch": 7.05,
      "learning_rate": 1e-05,
      "loss": 1.4407,
      "step": 29510
    },
    {
      "epoch": 7.05,
      "learning_rate": 1e-05,
      "loss": 1.4345,
      "step": 29520
    },
    {
      "epoch": 7.05,
      "learning_rate": 1e-05,
      "loss": 1.4449,
      "step": 29530
    },
    {
      "epoch": 7.06,
      "learning_rate": 1e-05,
      "loss": 1.4359,
      "step": 29540
    },
    {
      "epoch": 7.06,
      "learning_rate": 1e-05,
      "loss": 1.4363,
      "step": 29550
    },
    {
      "epoch": 7.06,
      "learning_rate": 1e-05,
      "loss": 1.4338,
      "step": 29560
    },
    {
      "epoch": 7.06,
      "learning_rate": 1e-05,
      "loss": 1.4414,
      "step": 29570
    },
    {
      "epoch": 7.06,
      "learning_rate": 1e-05,
      "loss": 1.4251,
      "step": 29580
    },
    {
      "epoch": 7.07,
      "learning_rate": 1e-05,
      "loss": 1.4352,
      "step": 29590
    },
    {
      "epoch": 7.07,
      "learning_rate": 1e-05,
      "loss": 1.4371,
      "step": 29600
    },
    {
      "epoch": 7.07,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4310097694396973,
      "eval_runtime": 34.7179,
      "eval_samples_per_second": 19.702,
      "step": 29600
    },
    {
      "epoch": 7.07,
      "learning_rate": 1e-05,
      "loss": 1.4533,
      "step": 29610
    },
    {
      "epoch": 7.07,
      "learning_rate": 1e-05,
      "loss": 1.4382,
      "step": 29620
    },
    {
      "epoch": 7.08,
      "learning_rate": 1e-05,
      "loss": 1.4328,
      "step": 29630
    },
    {
      "epoch": 7.08,
      "learning_rate": 1e-05,
      "loss": 1.4345,
      "step": 29640
    },
    {
      "epoch": 7.08,
      "learning_rate": 1e-05,
      "loss": 1.4259,
      "step": 29650
    },
    {
      "epoch": 7.08,
      "learning_rate": 1e-05,
      "loss": 1.4693,
      "step": 29660
    },
    {
      "epoch": 7.09,
      "learning_rate": 1e-05,
      "loss": 1.4369,
      "step": 29670
    },
    {
      "epoch": 7.09,
      "learning_rate": 1e-05,
      "loss": 1.4298,
      "step": 29680
    },
    {
      "epoch": 7.09,
      "learning_rate": 1e-05,
      "loss": 1.4418,
      "step": 29690
    },
    {
      "epoch": 7.09,
      "learning_rate": 1e-05,
      "loss": 1.4391,
      "step": 29700
    },
    {
      "epoch": 7.09,
      "eval_accuracy": 0.9736842105263158,
      "eval_loss": 1.431829810142517,
      "eval_runtime": 34.366,
      "eval_samples_per_second": 19.903,
      "step": 29700
    },
    {
      "epoch": 7.1,
      "learning_rate": 1e-05,
      "loss": 1.4424,
      "step": 29710
    },
    {
      "epoch": 7.1,
      "learning_rate": 1e-05,
      "loss": 1.4325,
      "step": 29720
    },
    {
      "epoch": 7.1,
      "learning_rate": 1e-05,
      "loss": 1.4388,
      "step": 29730
    },
    {
      "epoch": 7.1,
      "learning_rate": 1e-05,
      "loss": 1.4369,
      "step": 29740
    },
    {
      "epoch": 7.11,
      "learning_rate": 1e-05,
      "loss": 1.4361,
      "step": 29750
    },
    {
      "epoch": 7.11,
      "learning_rate": 1e-05,
      "loss": 1.4414,
      "step": 29760
    },
    {
      "epoch": 7.11,
      "learning_rate": 1e-05,
      "loss": 1.4374,
      "step": 29770
    },
    {
      "epoch": 7.11,
      "learning_rate": 1e-05,
      "loss": 1.4405,
      "step": 29780
    },
    {
      "epoch": 7.11,
      "learning_rate": 1e-05,
      "loss": 1.4448,
      "step": 29790
    },
    {
      "epoch": 7.12,
      "learning_rate": 1e-05,
      "loss": 1.4282,
      "step": 29800
    },
    {
      "epoch": 7.12,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4295769929885864,
      "eval_runtime": 34.4119,
      "eval_samples_per_second": 19.877,
      "step": 29800
    },
    {
      "epoch": 7.12,
      "learning_rate": 1e-05,
      "loss": 1.4345,
      "step": 29810
    },
    {
      "epoch": 7.12,
      "learning_rate": 1e-05,
      "loss": 1.4428,
      "step": 29820
    },
    {
      "epoch": 7.12,
      "learning_rate": 1e-05,
      "loss": 1.4374,
      "step": 29830
    },
    {
      "epoch": 7.13,
      "learning_rate": 1e-05,
      "loss": 1.4344,
      "step": 29840
    },
    {
      "epoch": 7.13,
      "learning_rate": 1e-05,
      "loss": 1.4323,
      "step": 29850
    },
    {
      "epoch": 7.13,
      "learning_rate": 1e-05,
      "loss": 1.4327,
      "step": 29860
    },
    {
      "epoch": 7.13,
      "learning_rate": 1e-05,
      "loss": 1.4252,
      "step": 29870
    },
    {
      "epoch": 7.14,
      "learning_rate": 1e-05,
      "loss": 1.4366,
      "step": 29880
    },
    {
      "epoch": 7.14,
      "learning_rate": 1e-05,
      "loss": 1.4355,
      "step": 29890
    },
    {
      "epoch": 7.14,
      "learning_rate": 1e-05,
      "loss": 1.4384,
      "step": 29900
    },
    {
      "epoch": 7.14,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4301882982254028,
      "eval_runtime": 34.5187,
      "eval_samples_per_second": 19.815,
      "step": 29900
    },
    {
      "epoch": 7.14,
      "learning_rate": 1e-05,
      "loss": 1.4353,
      "step": 29910
    },
    {
      "epoch": 7.15,
      "learning_rate": 1e-05,
      "loss": 1.446,
      "step": 29920
    },
    {
      "epoch": 7.15,
      "learning_rate": 1e-05,
      "loss": 1.4433,
      "step": 29930
    },
    {
      "epoch": 7.15,
      "learning_rate": 1e-05,
      "loss": 1.4342,
      "step": 29940
    },
    {
      "epoch": 7.15,
      "learning_rate": 1e-05,
      "loss": 1.4382,
      "step": 29950
    },
    {
      "epoch": 7.16,
      "learning_rate": 1e-05,
      "loss": 1.4448,
      "step": 29960
    },
    {
      "epoch": 7.16,
      "learning_rate": 1e-05,
      "loss": 1.4356,
      "step": 29970
    },
    {
      "epoch": 7.16,
      "learning_rate": 1e-05,
      "loss": 1.4297,
      "step": 29980
    },
    {
      "epoch": 7.16,
      "learning_rate": 1e-05,
      "loss": 1.431,
      "step": 29990
    },
    {
      "epoch": 7.17,
      "learning_rate": 1e-05,
      "loss": 1.4385,
      "step": 30000
    },
    {
      "epoch": 7.17,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4310133457183838,
      "eval_runtime": 34.4257,
      "eval_samples_per_second": 19.869,
      "step": 30000
    },
    {
      "epoch": 7.17,
      "learning_rate": 1e-05,
      "loss": 1.4409,
      "step": 30010
    },
    {
      "epoch": 7.17,
      "learning_rate": 1e-05,
      "loss": 1.4446,
      "step": 30020
    },
    {
      "epoch": 7.17,
      "learning_rate": 1e-05,
      "loss": 1.4374,
      "step": 30030
    },
    {
      "epoch": 7.17,
      "learning_rate": 1e-05,
      "loss": 1.4392,
      "step": 30040
    },
    {
      "epoch": 7.18,
      "learning_rate": 1e-05,
      "loss": 1.4333,
      "step": 30050
    },
    {
      "epoch": 7.18,
      "learning_rate": 1e-05,
      "loss": 1.4395,
      "step": 30060
    },
    {
      "epoch": 7.18,
      "learning_rate": 1e-05,
      "loss": 1.4433,
      "step": 30070
    },
    {
      "epoch": 7.18,
      "learning_rate": 1e-05,
      "loss": 1.4379,
      "step": 30080
    },
    {
      "epoch": 7.19,
      "learning_rate": 1e-05,
      "loss": 1.4478,
      "step": 30090
    },
    {
      "epoch": 7.19,
      "learning_rate": 1e-05,
      "loss": 1.4365,
      "step": 30100
    },
    {
      "epoch": 7.19,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4301698207855225,
      "eval_runtime": 34.4039,
      "eval_samples_per_second": 19.881,
      "step": 30100
    },
    {
      "epoch": 7.19,
      "learning_rate": 1e-05,
      "loss": 1.438,
      "step": 30110
    },
    {
      "epoch": 7.19,
      "learning_rate": 1e-05,
      "loss": 1.4328,
      "step": 30120
    },
    {
      "epoch": 7.2,
      "learning_rate": 1e-05,
      "loss": 1.4321,
      "step": 30130
    },
    {
      "epoch": 7.2,
      "learning_rate": 1e-05,
      "loss": 1.4362,
      "step": 30140
    },
    {
      "epoch": 7.2,
      "learning_rate": 1e-05,
      "loss": 1.4467,
      "step": 30150
    },
    {
      "epoch": 7.2,
      "learning_rate": 1e-05,
      "loss": 1.4411,
      "step": 30160
    },
    {
      "epoch": 7.21,
      "learning_rate": 1e-05,
      "loss": 1.4317,
      "step": 30170
    },
    {
      "epoch": 7.21,
      "learning_rate": 1e-05,
      "loss": 1.4337,
      "step": 30180
    },
    {
      "epoch": 7.21,
      "learning_rate": 1e-05,
      "loss": 1.4348,
      "step": 30190
    },
    {
      "epoch": 7.21,
      "learning_rate": 1e-05,
      "loss": 1.4351,
      "step": 30200
    },
    {
      "epoch": 7.21,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4308090209960938,
      "eval_runtime": 34.7197,
      "eval_samples_per_second": 19.701,
      "step": 30200
    },
    {
      "epoch": 7.22,
      "learning_rate": 1e-05,
      "loss": 1.4306,
      "step": 30210
    },
    {
      "epoch": 7.22,
      "learning_rate": 1e-05,
      "loss": 1.436,
      "step": 30220
    },
    {
      "epoch": 7.22,
      "learning_rate": 1e-05,
      "loss": 1.4443,
      "step": 30230
    },
    {
      "epoch": 7.22,
      "learning_rate": 1e-05,
      "loss": 1.4411,
      "step": 30240
    },
    {
      "epoch": 7.22,
      "learning_rate": 1e-05,
      "loss": 1.4452,
      "step": 30250
    },
    {
      "epoch": 7.23,
      "learning_rate": 1e-05,
      "loss": 1.4392,
      "step": 30260
    },
    {
      "epoch": 7.23,
      "learning_rate": 1e-05,
      "loss": 1.4355,
      "step": 30270
    },
    {
      "epoch": 7.23,
      "learning_rate": 1e-05,
      "loss": 1.4324,
      "step": 30280
    },
    {
      "epoch": 7.23,
      "learning_rate": 1e-05,
      "loss": 1.4321,
      "step": 30290
    },
    {
      "epoch": 7.24,
      "learning_rate": 1e-05,
      "loss": 1.4376,
      "step": 30300
    },
    {
      "epoch": 7.24,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4305766820907593,
      "eval_runtime": 34.6811,
      "eval_samples_per_second": 19.723,
      "step": 30300
    },
    {
      "epoch": 7.24,
      "learning_rate": 1e-05,
      "loss": 1.4316,
      "step": 30310
    },
    {
      "epoch": 7.24,
      "learning_rate": 1e-05,
      "loss": 1.438,
      "step": 30320
    },
    {
      "epoch": 7.24,
      "learning_rate": 1e-05,
      "loss": 1.4335,
      "step": 30330
    },
    {
      "epoch": 7.25,
      "learning_rate": 1e-05,
      "loss": 1.4432,
      "step": 30340
    },
    {
      "epoch": 7.25,
      "learning_rate": 1e-05,
      "loss": 1.4342,
      "step": 30350
    },
    {
      "epoch": 7.25,
      "learning_rate": 1e-05,
      "loss": 1.4488,
      "step": 30360
    },
    {
      "epoch": 7.25,
      "learning_rate": 1e-05,
      "loss": 1.4326,
      "step": 30370
    },
    {
      "epoch": 7.26,
      "learning_rate": 1e-05,
      "loss": 1.4361,
      "step": 30380
    },
    {
      "epoch": 7.26,
      "learning_rate": 1e-05,
      "loss": 1.4385,
      "step": 30390
    },
    {
      "epoch": 7.26,
      "learning_rate": 1e-05,
      "loss": 1.444,
      "step": 30400
    },
    {
      "epoch": 7.26,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4304471015930176,
      "eval_runtime": 34.4026,
      "eval_samples_per_second": 19.882,
      "step": 30400
    },
    {
      "epoch": 7.26,
      "learning_rate": 1e-05,
      "loss": 1.4525,
      "step": 30410
    },
    {
      "epoch": 7.27,
      "learning_rate": 1e-05,
      "loss": 1.4335,
      "step": 30420
    },
    {
      "epoch": 7.27,
      "learning_rate": 1e-05,
      "loss": 1.4372,
      "step": 30430
    },
    {
      "epoch": 7.27,
      "learning_rate": 1e-05,
      "loss": 1.4371,
      "step": 30440
    },
    {
      "epoch": 7.27,
      "learning_rate": 1e-05,
      "loss": 1.4358,
      "step": 30450
    },
    {
      "epoch": 7.27,
      "learning_rate": 1e-05,
      "loss": 1.4325,
      "step": 30460
    },
    {
      "epoch": 7.28,
      "learning_rate": 1e-05,
      "loss": 1.4227,
      "step": 30470
    },
    {
      "epoch": 7.28,
      "learning_rate": 1e-05,
      "loss": 1.4382,
      "step": 30480
    },
    {
      "epoch": 7.28,
      "learning_rate": 1e-05,
      "loss": 1.4513,
      "step": 30490
    },
    {
      "epoch": 7.28,
      "learning_rate": 1e-05,
      "loss": 1.4394,
      "step": 30500
    },
    {
      "epoch": 7.28,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4309790134429932,
      "eval_runtime": 34.3432,
      "eval_samples_per_second": 19.917,
      "step": 30500
    },
    {
      "epoch": 7.29,
      "learning_rate": 1e-05,
      "loss": 1.4416,
      "step": 30510
    },
    {
      "epoch": 7.29,
      "learning_rate": 1e-05,
      "loss": 1.4407,
      "step": 30520
    },
    {
      "epoch": 7.29,
      "learning_rate": 1e-05,
      "loss": 1.434,
      "step": 30530
    },
    {
      "epoch": 7.29,
      "learning_rate": 1e-05,
      "loss": 1.4375,
      "step": 30540
    },
    {
      "epoch": 7.3,
      "learning_rate": 1e-05,
      "loss": 1.4376,
      "step": 30550
    },
    {
      "epoch": 7.3,
      "learning_rate": 1e-05,
      "loss": 1.4393,
      "step": 30560
    },
    {
      "epoch": 7.3,
      "learning_rate": 1e-05,
      "loss": 1.4414,
      "step": 30570
    },
    {
      "epoch": 7.3,
      "learning_rate": 1e-05,
      "loss": 1.4376,
      "step": 30580
    },
    {
      "epoch": 7.31,
      "learning_rate": 1e-05,
      "loss": 1.4284,
      "step": 30590
    },
    {
      "epoch": 7.31,
      "learning_rate": 1e-05,
      "loss": 1.4501,
      "step": 30600
    },
    {
      "epoch": 7.31,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4284719228744507,
      "eval_runtime": 34.5392,
      "eval_samples_per_second": 19.804,
      "step": 30600
    },
    {
      "epoch": 7.31,
      "learning_rate": 1e-05,
      "loss": 1.4318,
      "step": 30610
    },
    {
      "epoch": 7.31,
      "learning_rate": 1e-05,
      "loss": 1.4831,
      "step": 30620
    },
    {
      "epoch": 7.32,
      "learning_rate": 1e-05,
      "loss": 1.4307,
      "step": 30630
    },
    {
      "epoch": 7.32,
      "learning_rate": 1e-05,
      "loss": 1.4278,
      "step": 30640
    },
    {
      "epoch": 7.32,
      "learning_rate": 1e-05,
      "loss": 1.4392,
      "step": 30650
    },
    {
      "epoch": 7.32,
      "learning_rate": 1e-05,
      "loss": 1.4411,
      "step": 30660
    },
    {
      "epoch": 7.33,
      "learning_rate": 1e-05,
      "loss": 1.4324,
      "step": 30670
    },
    {
      "epoch": 7.33,
      "learning_rate": 1e-05,
      "loss": 1.4376,
      "step": 30680
    },
    {
      "epoch": 7.33,
      "learning_rate": 1e-05,
      "loss": 1.439,
      "step": 30690
    },
    {
      "epoch": 7.33,
      "learning_rate": 1e-05,
      "loss": 1.4405,
      "step": 30700
    },
    {
      "epoch": 7.33,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4305627346038818,
      "eval_runtime": 34.6161,
      "eval_samples_per_second": 19.76,
      "step": 30700
    },
    {
      "epoch": 7.33,
      "learning_rate": 1e-05,
      "loss": 1.441,
      "step": 30710
    },
    {
      "epoch": 7.34,
      "learning_rate": 1e-05,
      "loss": 1.4286,
      "step": 30720
    },
    {
      "epoch": 7.34,
      "learning_rate": 1e-05,
      "loss": 1.4325,
      "step": 30730
    },
    {
      "epoch": 7.34,
      "learning_rate": 1e-05,
      "loss": 1.4452,
      "step": 30740
    },
    {
      "epoch": 7.34,
      "learning_rate": 1e-05,
      "loss": 1.4419,
      "step": 30750
    },
    {
      "epoch": 7.35,
      "learning_rate": 1e-05,
      "loss": 1.4309,
      "step": 30760
    },
    {
      "epoch": 7.35,
      "learning_rate": 1e-05,
      "loss": 1.4345,
      "step": 30770
    },
    {
      "epoch": 7.35,
      "learning_rate": 1e-05,
      "loss": 1.4413,
      "step": 30780
    },
    {
      "epoch": 7.35,
      "learning_rate": 1e-05,
      "loss": 1.432,
      "step": 30790
    },
    {
      "epoch": 7.36,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 30800
    },
    {
      "epoch": 7.36,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4291794300079346,
      "eval_runtime": 34.6422,
      "eval_samples_per_second": 19.745,
      "step": 30800
    },
    {
      "epoch": 7.36,
      "learning_rate": 1e-05,
      "loss": 1.4311,
      "step": 30810
    },
    {
      "epoch": 7.36,
      "learning_rate": 1e-05,
      "loss": 1.4419,
      "step": 30820
    },
    {
      "epoch": 7.36,
      "learning_rate": 1e-05,
      "loss": 1.4409,
      "step": 30830
    },
    {
      "epoch": 7.37,
      "learning_rate": 1e-05,
      "loss": 1.434,
      "step": 30840
    },
    {
      "epoch": 7.37,
      "learning_rate": 1e-05,
      "loss": 1.4356,
      "step": 30850
    },
    {
      "epoch": 7.37,
      "learning_rate": 1e-05,
      "loss": 1.4331,
      "step": 30860
    },
    {
      "epoch": 7.37,
      "learning_rate": 1e-05,
      "loss": 1.4269,
      "step": 30870
    },
    {
      "epoch": 7.38,
      "learning_rate": 1e-05,
      "loss": 1.4362,
      "step": 30880
    },
    {
      "epoch": 7.38,
      "learning_rate": 1e-05,
      "loss": 1.4311,
      "step": 30890
    },
    {
      "epoch": 7.38,
      "learning_rate": 1e-05,
      "loss": 1.4455,
      "step": 30900
    },
    {
      "epoch": 7.38,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4288409948349,
      "eval_runtime": 34.4926,
      "eval_samples_per_second": 19.83,
      "step": 30900
    },
    {
      "epoch": 7.38,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 30910
    },
    {
      "epoch": 7.38,
      "learning_rate": 1e-05,
      "loss": 1.4321,
      "step": 30920
    },
    {
      "epoch": 7.39,
      "learning_rate": 1e-05,
      "loss": 1.4359,
      "step": 30930
    },
    {
      "epoch": 7.39,
      "learning_rate": 1e-05,
      "loss": 1.4313,
      "step": 30940
    },
    {
      "epoch": 7.39,
      "learning_rate": 1e-05,
      "loss": 1.4329,
      "step": 30950
    },
    {
      "epoch": 7.39,
      "learning_rate": 1e-05,
      "loss": 1.4387,
      "step": 30960
    },
    {
      "epoch": 7.4,
      "learning_rate": 1e-05,
      "loss": 1.431,
      "step": 30970
    },
    {
      "epoch": 7.4,
      "learning_rate": 1e-05,
      "loss": 1.4364,
      "step": 30980
    },
    {
      "epoch": 7.4,
      "learning_rate": 1e-05,
      "loss": 1.426,
      "step": 30990
    },
    {
      "epoch": 7.4,
      "learning_rate": 1e-05,
      "loss": 1.4375,
      "step": 31000
    },
    {
      "epoch": 7.4,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4311327934265137,
      "eval_runtime": 34.4418,
      "eval_samples_per_second": 19.86,
      "step": 31000
    },
    {
      "epoch": 7.41,
      "learning_rate": 1e-05,
      "loss": 1.4442,
      "step": 31010
    },
    {
      "epoch": 7.41,
      "learning_rate": 1e-05,
      "loss": 1.4437,
      "step": 31020
    },
    {
      "epoch": 7.41,
      "learning_rate": 1e-05,
      "loss": 1.428,
      "step": 31030
    },
    {
      "epoch": 7.41,
      "learning_rate": 1e-05,
      "loss": 1.4314,
      "step": 31040
    },
    {
      "epoch": 7.42,
      "learning_rate": 1e-05,
      "loss": 1.438,
      "step": 31050
    },
    {
      "epoch": 7.42,
      "learning_rate": 1e-05,
      "loss": 1.4336,
      "step": 31060
    },
    {
      "epoch": 7.42,
      "learning_rate": 1e-05,
      "loss": 1.4282,
      "step": 31070
    },
    {
      "epoch": 7.42,
      "learning_rate": 1e-05,
      "loss": 1.4326,
      "step": 31080
    },
    {
      "epoch": 7.43,
      "learning_rate": 1e-05,
      "loss": 1.4323,
      "step": 31090
    },
    {
      "epoch": 7.43,
      "learning_rate": 1e-05,
      "loss": 1.4331,
      "step": 31100
    },
    {
      "epoch": 7.43,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.428904414176941,
      "eval_runtime": 34.6498,
      "eval_samples_per_second": 19.74,
      "step": 31100
    },
    {
      "epoch": 7.43,
      "learning_rate": 1e-05,
      "loss": 1.4317,
      "step": 31110
    },
    {
      "epoch": 7.43,
      "learning_rate": 1e-05,
      "loss": 1.425,
      "step": 31120
    },
    {
      "epoch": 7.43,
      "learning_rate": 1e-05,
      "loss": 1.4371,
      "step": 31130
    },
    {
      "epoch": 7.44,
      "learning_rate": 1e-05,
      "loss": 1.4326,
      "step": 31140
    },
    {
      "epoch": 7.44,
      "learning_rate": 1e-05,
      "loss": 1.4311,
      "step": 31150
    },
    {
      "epoch": 7.44,
      "learning_rate": 1e-05,
      "loss": 1.4326,
      "step": 31160
    },
    {
      "epoch": 7.44,
      "learning_rate": 1e-05,
      "loss": 1.4339,
      "step": 31170
    },
    {
      "epoch": 7.45,
      "learning_rate": 1e-05,
      "loss": 1.4492,
      "step": 31180
    },
    {
      "epoch": 7.45,
      "learning_rate": 1e-05,
      "loss": 1.4316,
      "step": 31190
    },
    {
      "epoch": 7.45,
      "learning_rate": 1e-05,
      "loss": 1.4381,
      "step": 31200
    },
    {
      "epoch": 7.45,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4297783374786377,
      "eval_runtime": 34.5251,
      "eval_samples_per_second": 19.812,
      "step": 31200
    },
    {
      "epoch": 7.45,
      "learning_rate": 1e-05,
      "loss": 1.4352,
      "step": 31210
    },
    {
      "epoch": 7.46,
      "learning_rate": 1e-05,
      "loss": 1.427,
      "step": 31220
    },
    {
      "epoch": 7.46,
      "learning_rate": 1e-05,
      "loss": 1.4338,
      "step": 31230
    },
    {
      "epoch": 7.46,
      "learning_rate": 1e-05,
      "loss": 1.4417,
      "step": 31240
    },
    {
      "epoch": 7.46,
      "learning_rate": 1e-05,
      "loss": 1.4347,
      "step": 31250
    },
    {
      "epoch": 7.47,
      "learning_rate": 1e-05,
      "loss": 1.4331,
      "step": 31260
    },
    {
      "epoch": 7.47,
      "learning_rate": 1e-05,
      "loss": 1.4305,
      "step": 31270
    },
    {
      "epoch": 7.47,
      "learning_rate": 1e-05,
      "loss": 1.4342,
      "step": 31280
    },
    {
      "epoch": 7.47,
      "learning_rate": 1e-05,
      "loss": 1.4309,
      "step": 31290
    },
    {
      "epoch": 7.48,
      "learning_rate": 1e-05,
      "loss": 1.4374,
      "step": 31300
    },
    {
      "epoch": 7.48,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4299055337905884,
      "eval_runtime": 34.8701,
      "eval_samples_per_second": 19.616,
      "step": 31300
    },
    {
      "epoch": 7.48,
      "learning_rate": 1e-05,
      "loss": 1.4413,
      "step": 31310
    },
    {
      "epoch": 7.48,
      "learning_rate": 1e-05,
      "loss": 1.4863,
      "step": 31320
    },
    {
      "epoch": 7.48,
      "learning_rate": 1e-05,
      "loss": 1.4423,
      "step": 31330
    },
    {
      "epoch": 7.49,
      "learning_rate": 1e-05,
      "loss": 1.4362,
      "step": 31340
    },
    {
      "epoch": 7.49,
      "learning_rate": 1e-05,
      "loss": 1.4381,
      "step": 31350
    },
    {
      "epoch": 7.49,
      "learning_rate": 1e-05,
      "loss": 1.4385,
      "step": 31360
    },
    {
      "epoch": 7.49,
      "learning_rate": 1e-05,
      "loss": 1.4337,
      "step": 31370
    },
    {
      "epoch": 7.49,
      "learning_rate": 1e-05,
      "loss": 1.4352,
      "step": 31380
    },
    {
      "epoch": 7.5,
      "learning_rate": 1e-05,
      "loss": 1.4335,
      "step": 31390
    },
    {
      "epoch": 7.5,
      "learning_rate": 1e-05,
      "loss": 1.432,
      "step": 31400
    },
    {
      "epoch": 7.5,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4282335042953491,
      "eval_runtime": 34.5702,
      "eval_samples_per_second": 19.786,
      "step": 31400
    },
    {
      "epoch": 7.5,
      "learning_rate": 1e-05,
      "loss": 1.4366,
      "step": 31410
    },
    {
      "epoch": 7.5,
      "learning_rate": 1e-05,
      "loss": 1.443,
      "step": 31420
    },
    {
      "epoch": 7.51,
      "learning_rate": 1e-05,
      "loss": 1.4504,
      "step": 31430
    },
    {
      "epoch": 7.51,
      "learning_rate": 1e-05,
      "loss": 1.4388,
      "step": 31440
    },
    {
      "epoch": 7.51,
      "learning_rate": 1e-05,
      "loss": 1.4303,
      "step": 31450
    },
    {
      "epoch": 7.51,
      "learning_rate": 1e-05,
      "loss": 1.433,
      "step": 31460
    },
    {
      "epoch": 7.52,
      "learning_rate": 1e-05,
      "loss": 1.4261,
      "step": 31470
    },
    {
      "epoch": 7.52,
      "learning_rate": 1e-05,
      "loss": 1.4297,
      "step": 31480
    },
    {
      "epoch": 7.52,
      "learning_rate": 1e-05,
      "loss": 1.4354,
      "step": 31490
    },
    {
      "epoch": 7.52,
      "learning_rate": 1e-05,
      "loss": 1.4326,
      "step": 31500
    },
    {
      "epoch": 7.52,
      "eval_accuracy": 0.9751461988304093,
      "eval_loss": 1.4290436506271362,
      "eval_runtime": 34.4473,
      "eval_samples_per_second": 19.856,
      "step": 31500
    },
    {
      "epoch": 7.53,
      "learning_rate": 1e-05,
      "loss": 1.4316,
      "step": 31510
    },
    {
      "epoch": 7.53,
      "learning_rate": 1e-05,
      "loss": 1.4321,
      "step": 31520
    },
    {
      "epoch": 7.53,
      "learning_rate": 1e-05,
      "loss": 1.4338,
      "step": 31530
    },
    {
      "epoch": 7.53,
      "learning_rate": 1e-05,
      "loss": 1.4426,
      "step": 31540
    },
    {
      "epoch": 7.54,
      "learning_rate": 1e-05,
      "loss": 1.4363,
      "step": 31550
    },
    {
      "epoch": 7.54,
      "learning_rate": 1e-05,
      "loss": 1.436,
      "step": 31560
    },
    {
      "epoch": 7.54,
      "learning_rate": 1e-05,
      "loss": 1.4358,
      "step": 31570
    },
    {
      "epoch": 7.54,
      "learning_rate": 1e-05,
      "loss": 1.432,
      "step": 31580
    },
    {
      "epoch": 7.54,
      "learning_rate": 1e-05,
      "loss": 1.435,
      "step": 31590
    },
    {
      "epoch": 7.55,
      "learning_rate": 1e-05,
      "loss": 1.4321,
      "step": 31600
    },
    {
      "epoch": 7.55,
      "eval_accuracy": 0.9722222222222222,
      "eval_loss": 1.4280563592910767,
      "eval_runtime": 34.4152,
      "eval_samples_per_second": 19.875,
      "step": 31600
    },
    {
      "epoch": 7.55,
      "learning_rate": 1e-05,
      "loss": 1.4662,
      "step": 31610
    },
    {
      "epoch": 7.55,
      "learning_rate": 1e-05,
      "loss": 1.4292,
      "step": 31620
    },
    {
      "epoch": 7.55,
      "learning_rate": 1e-05,
      "loss": 1.4431,
      "step": 31630
    },
    {
      "epoch": 7.56,
      "learning_rate": 1e-05,
      "loss": 1.4274,
      "step": 31640
    },
    {
      "epoch": 7.56,
      "learning_rate": 1e-05,
      "loss": 1.4397,
      "step": 31650
    },
    {
      "epoch": 7.56,
      "learning_rate": 1e-05,
      "loss": 1.4315,
      "step": 31660
    },
    {
      "epoch": 7.56,
      "learning_rate": 1e-05,
      "loss": 1.4347,
      "step": 31670
    },
    {
      "epoch": 7.57,
      "learning_rate": 1e-05,
      "loss": 1.4363,
      "step": 31680
    },
    {
      "epoch": 7.57,
      "learning_rate": 1e-05,
      "loss": 1.4332,
      "step": 31690
    },
    {
      "epoch": 7.57,
      "learning_rate": 1e-05,
      "loss": 1.431,
      "step": 31700
    },
    {
      "epoch": 7.57,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4286717176437378,
      "eval_runtime": 34.6546,
      "eval_samples_per_second": 19.738,
      "step": 31700
    },
    {
      "epoch": 7.57,
      "learning_rate": 1e-05,
      "loss": 1.4317,
      "step": 31710
    },
    {
      "epoch": 7.58,
      "learning_rate": 1e-05,
      "loss": 1.4361,
      "step": 31720
    },
    {
      "epoch": 7.58,
      "learning_rate": 1e-05,
      "loss": 1.4371,
      "step": 31730
    },
    {
      "epoch": 7.58,
      "learning_rate": 1e-05,
      "loss": 1.4434,
      "step": 31740
    },
    {
      "epoch": 7.58,
      "learning_rate": 1e-05,
      "loss": 1.4357,
      "step": 31750
    },
    {
      "epoch": 7.59,
      "learning_rate": 1e-05,
      "loss": 1.4295,
      "step": 31760
    },
    {
      "epoch": 7.59,
      "learning_rate": 1e-05,
      "loss": 1.4288,
      "step": 31770
    },
    {
      "epoch": 7.59,
      "learning_rate": 1e-05,
      "loss": 1.4464,
      "step": 31780
    },
    {
      "epoch": 7.59,
      "learning_rate": 1e-05,
      "loss": 1.4327,
      "step": 31790
    },
    {
      "epoch": 7.59,
      "learning_rate": 1e-05,
      "loss": 1.4361,
      "step": 31800
    },
    {
      "epoch": 7.59,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.427823781967163,
      "eval_runtime": 34.3964,
      "eval_samples_per_second": 19.886,
      "step": 31800
    },
    {
      "epoch": 7.6,
      "learning_rate": 1e-05,
      "loss": 1.4353,
      "step": 31810
    },
    {
      "epoch": 7.6,
      "learning_rate": 1e-05,
      "loss": 1.429,
      "step": 31820
    },
    {
      "epoch": 7.6,
      "learning_rate": 1e-05,
      "loss": 1.4328,
      "step": 31830
    },
    {
      "epoch": 7.6,
      "learning_rate": 1e-05,
      "loss": 1.4373,
      "step": 31840
    },
    {
      "epoch": 7.61,
      "learning_rate": 1e-05,
      "loss": 1.4263,
      "step": 31850
    },
    {
      "epoch": 7.61,
      "learning_rate": 1e-05,
      "loss": 1.4352,
      "step": 31860
    },
    {
      "epoch": 7.61,
      "learning_rate": 1e-05,
      "loss": 1.441,
      "step": 31870
    },
    {
      "epoch": 7.61,
      "learning_rate": 1e-05,
      "loss": 1.4343,
      "step": 31880
    },
    {
      "epoch": 7.62,
      "learning_rate": 1e-05,
      "loss": 1.4307,
      "step": 31890
    },
    {
      "epoch": 7.62,
      "learning_rate": 1e-05,
      "loss": 1.4274,
      "step": 31900
    },
    {
      "epoch": 7.62,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4292782545089722,
      "eval_runtime": 34.9165,
      "eval_samples_per_second": 19.59,
      "step": 31900
    },
    {
      "epoch": 7.62,
      "learning_rate": 1e-05,
      "loss": 1.4399,
      "step": 31910
    },
    {
      "epoch": 7.62,
      "learning_rate": 1e-05,
      "loss": 1.4293,
      "step": 31920
    },
    {
      "epoch": 7.63,
      "learning_rate": 1e-05,
      "loss": 1.4352,
      "step": 31930
    },
    {
      "epoch": 7.63,
      "learning_rate": 1e-05,
      "loss": 1.4304,
      "step": 31940
    },
    {
      "epoch": 7.63,
      "learning_rate": 1e-05,
      "loss": 1.4357,
      "step": 31950
    },
    {
      "epoch": 7.63,
      "learning_rate": 1e-05,
      "loss": 1.4402,
      "step": 31960
    },
    {
      "epoch": 7.64,
      "learning_rate": 1e-05,
      "loss": 1.4278,
      "step": 31970
    },
    {
      "epoch": 7.64,
      "learning_rate": 1e-05,
      "loss": 1.4338,
      "step": 31980
    },
    {
      "epoch": 7.64,
      "learning_rate": 1e-05,
      "loss": 1.4377,
      "step": 31990
    },
    {
      "epoch": 7.64,
      "learning_rate": 1e-05,
      "loss": 1.4314,
      "step": 32000
    },
    {
      "epoch": 7.64,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4284343719482422,
      "eval_runtime": 34.4579,
      "eval_samples_per_second": 19.85,
      "step": 32000
    },
    {
      "epoch": 7.65,
      "learning_rate": 1e-05,
      "loss": 1.4397,
      "step": 32010
    },
    {
      "epoch": 7.65,
      "learning_rate": 1e-05,
      "loss": 1.4353,
      "step": 32020
    },
    {
      "epoch": 7.65,
      "learning_rate": 1e-05,
      "loss": 1.4305,
      "step": 32030
    },
    {
      "epoch": 7.65,
      "learning_rate": 1e-05,
      "loss": 1.4313,
      "step": 32040
    },
    {
      "epoch": 7.65,
      "learning_rate": 1e-05,
      "loss": 1.4357,
      "step": 32050
    },
    {
      "epoch": 7.66,
      "learning_rate": 1e-05,
      "loss": 1.4314,
      "step": 32060
    },
    {
      "epoch": 7.66,
      "learning_rate": 1e-05,
      "loss": 1.4307,
      "step": 32070
    },
    {
      "epoch": 7.66,
      "learning_rate": 1e-05,
      "loss": 1.4403,
      "step": 32080
    },
    {
      "epoch": 7.66,
      "learning_rate": 1e-05,
      "loss": 1.4356,
      "step": 32090
    },
    {
      "epoch": 7.67,
      "learning_rate": 1e-05,
      "loss": 1.4375,
      "step": 32100
    },
    {
      "epoch": 7.67,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4285306930541992,
      "eval_runtime": 34.5787,
      "eval_samples_per_second": 19.781,
      "step": 32100
    },
    {
      "epoch": 7.67,
      "learning_rate": 1e-05,
      "loss": 1.4341,
      "step": 32110
    },
    {
      "epoch": 7.67,
      "learning_rate": 1e-05,
      "loss": 1.4308,
      "step": 32120
    },
    {
      "epoch": 7.67,
      "learning_rate": 1e-05,
      "loss": 1.4376,
      "step": 32130
    },
    {
      "epoch": 7.68,
      "learning_rate": 1e-05,
      "loss": 1.4422,
      "step": 32140
    },
    {
      "epoch": 7.68,
      "learning_rate": 1e-05,
      "loss": 1.4277,
      "step": 32150
    },
    {
      "epoch": 7.68,
      "learning_rate": 1e-05,
      "loss": 1.4348,
      "step": 32160
    },
    {
      "epoch": 7.68,
      "learning_rate": 1e-05,
      "loss": 1.4355,
      "step": 32170
    },
    {
      "epoch": 7.69,
      "learning_rate": 1e-05,
      "loss": 1.4309,
      "step": 32180
    },
    {
      "epoch": 7.69,
      "learning_rate": 1e-05,
      "loss": 1.432,
      "step": 32190
    },
    {
      "epoch": 7.69,
      "learning_rate": 1e-05,
      "loss": 1.4253,
      "step": 32200
    },
    {
      "epoch": 7.69,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.429384469985962,
      "eval_runtime": 34.8207,
      "eval_samples_per_second": 19.644,
      "step": 32200
    },
    {
      "epoch": 7.69,
      "learning_rate": 1e-05,
      "loss": 1.4326,
      "step": 32210
    },
    {
      "epoch": 7.7,
      "learning_rate": 1e-05,
      "loss": 1.4404,
      "step": 32220
    },
    {
      "epoch": 7.7,
      "learning_rate": 1e-05,
      "loss": 1.4279,
      "step": 32230
    },
    {
      "epoch": 7.7,
      "learning_rate": 1e-05,
      "loss": 1.4396,
      "step": 32240
    },
    {
      "epoch": 7.7,
      "learning_rate": 1e-05,
      "loss": 1.4307,
      "step": 32250
    },
    {
      "epoch": 7.7,
      "learning_rate": 1e-05,
      "loss": 1.4287,
      "step": 32260
    },
    {
      "epoch": 7.71,
      "learning_rate": 1e-05,
      "loss": 1.4333,
      "step": 32270
    },
    {
      "epoch": 7.71,
      "learning_rate": 1e-05,
      "loss": 1.4386,
      "step": 32280
    },
    {
      "epoch": 7.71,
      "learning_rate": 1e-05,
      "loss": 1.4348,
      "step": 32290
    },
    {
      "epoch": 7.71,
      "learning_rate": 1e-05,
      "loss": 1.4442,
      "step": 32300
    },
    {
      "epoch": 7.71,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4290417432785034,
      "eval_runtime": 34.4998,
      "eval_samples_per_second": 19.826,
      "step": 32300
    },
    {
      "epoch": 7.72,
      "learning_rate": 1e-05,
      "loss": 1.4244,
      "step": 32310
    },
    {
      "epoch": 7.72,
      "learning_rate": 1e-05,
      "loss": 1.4289,
      "step": 32320
    },
    {
      "epoch": 7.72,
      "learning_rate": 1e-05,
      "loss": 1.4354,
      "step": 32330
    },
    {
      "epoch": 7.72,
      "learning_rate": 1e-05,
      "loss": 1.4405,
      "step": 32340
    },
    {
      "epoch": 7.73,
      "learning_rate": 1e-05,
      "loss": 1.4396,
      "step": 32350
    },
    {
      "epoch": 7.73,
      "learning_rate": 1e-05,
      "loss": 1.4351,
      "step": 32360
    },
    {
      "epoch": 7.73,
      "learning_rate": 1e-05,
      "loss": 1.4297,
      "step": 32370
    },
    {
      "epoch": 7.73,
      "learning_rate": 1e-05,
      "loss": 1.4302,
      "step": 32380
    },
    {
      "epoch": 7.74,
      "learning_rate": 1e-05,
      "loss": 1.4254,
      "step": 32390
    },
    {
      "epoch": 7.74,
      "learning_rate": 1e-05,
      "loss": 1.4265,
      "step": 32400
    },
    {
      "epoch": 7.74,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.4298630952835083,
      "eval_runtime": 35.1173,
      "eval_samples_per_second": 19.478,
      "step": 32400
    },
    {
      "epoch": 7.74,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 32410
    },
    {
      "epoch": 7.74,
      "learning_rate": 1e-05,
      "loss": 1.4284,
      "step": 32420
    },
    {
      "epoch": 7.75,
      "learning_rate": 1e-05,
      "loss": 1.429,
      "step": 32430
    },
    {
      "epoch": 7.75,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 32440
    },
    {
      "epoch": 7.75,
      "learning_rate": 1e-05,
      "loss": 1.4318,
      "step": 32450
    },
    {
      "epoch": 7.75,
      "learning_rate": 1e-05,
      "loss": 1.4376,
      "step": 32460
    },
    {
      "epoch": 7.75,
      "learning_rate": 1e-05,
      "loss": 1.4347,
      "step": 32470
    },
    {
      "epoch": 7.76,
      "learning_rate": 1e-05,
      "loss": 1.4518,
      "step": 32480
    },
    {
      "epoch": 7.76,
      "learning_rate": 1e-05,
      "loss": 1.4361,
      "step": 32490
    },
    {
      "epoch": 7.76,
      "learning_rate": 1e-05,
      "loss": 1.4383,
      "step": 32500
    },
    {
      "epoch": 7.76,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4283889532089233,
      "eval_runtime": 34.9206,
      "eval_samples_per_second": 19.587,
      "step": 32500
    },
    {
      "epoch": 7.76,
      "learning_rate": 1e-05,
      "loss": 1.431,
      "step": 32510
    },
    {
      "epoch": 7.77,
      "learning_rate": 1e-05,
      "loss": 1.4291,
      "step": 32520
    },
    {
      "epoch": 7.77,
      "learning_rate": 1e-05,
      "loss": 1.4225,
      "step": 32530
    },
    {
      "epoch": 7.77,
      "learning_rate": 1e-05,
      "loss": 1.4337,
      "step": 32540
    },
    {
      "epoch": 7.77,
      "learning_rate": 1e-05,
      "loss": 1.4331,
      "step": 32550
    },
    {
      "epoch": 7.78,
      "learning_rate": 1e-05,
      "loss": 1.4489,
      "step": 32560
    },
    {
      "epoch": 7.78,
      "learning_rate": 1e-05,
      "loss": 1.4309,
      "step": 32570
    },
    {
      "epoch": 7.78,
      "learning_rate": 1e-05,
      "loss": 1.4458,
      "step": 32580
    },
    {
      "epoch": 7.78,
      "learning_rate": 1e-05,
      "loss": 1.4337,
      "step": 32590
    },
    {
      "epoch": 7.79,
      "learning_rate": 1e-05,
      "loss": 1.4376,
      "step": 32600
    },
    {
      "epoch": 7.79,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4282349348068237,
      "eval_runtime": 35.0215,
      "eval_samples_per_second": 19.531,
      "step": 32600
    },
    {
      "epoch": 7.79,
      "learning_rate": 1e-05,
      "loss": 1.4421,
      "step": 32610
    },
    {
      "epoch": 7.79,
      "learning_rate": 1e-05,
      "loss": 1.4364,
      "step": 32620
    },
    {
      "epoch": 7.79,
      "learning_rate": 1e-05,
      "loss": 1.4315,
      "step": 32630
    },
    {
      "epoch": 7.8,
      "learning_rate": 1e-05,
      "loss": 1.429,
      "step": 32640
    },
    {
      "epoch": 7.8,
      "learning_rate": 1e-05,
      "loss": 1.4298,
      "step": 32650
    },
    {
      "epoch": 7.8,
      "learning_rate": 1e-05,
      "loss": 1.4533,
      "step": 32660
    },
    {
      "epoch": 7.8,
      "learning_rate": 1e-05,
      "loss": 1.4398,
      "step": 32670
    },
    {
      "epoch": 7.81,
      "learning_rate": 1e-05,
      "loss": 1.4369,
      "step": 32680
    },
    {
      "epoch": 7.81,
      "learning_rate": 1e-05,
      "loss": 1.4297,
      "step": 32690
    },
    {
      "epoch": 7.81,
      "learning_rate": 1e-05,
      "loss": 1.4303,
      "step": 32700
    },
    {
      "epoch": 7.81,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4291337728500366,
      "eval_runtime": 34.6675,
      "eval_samples_per_second": 19.73,
      "step": 32700
    },
    {
      "epoch": 7.81,
      "learning_rate": 1e-05,
      "loss": 1.4332,
      "step": 32710
    },
    {
      "epoch": 7.81,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 32720
    },
    {
      "epoch": 7.82,
      "learning_rate": 1e-05,
      "loss": 1.432,
      "step": 32730
    },
    {
      "epoch": 7.82,
      "learning_rate": 1e-05,
      "loss": 1.4229,
      "step": 32740
    },
    {
      "epoch": 7.82,
      "learning_rate": 1e-05,
      "loss": 1.438,
      "step": 32750
    },
    {
      "epoch": 7.82,
      "learning_rate": 1e-05,
      "loss": 1.4308,
      "step": 32760
    },
    {
      "epoch": 7.83,
      "learning_rate": 1e-05,
      "loss": 1.4282,
      "step": 32770
    },
    {
      "epoch": 7.83,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 32780
    },
    {
      "epoch": 7.83,
      "learning_rate": 1e-05,
      "loss": 1.4332,
      "step": 32790
    },
    {
      "epoch": 7.83,
      "learning_rate": 1e-05,
      "loss": 1.4313,
      "step": 32800
    },
    {
      "epoch": 7.83,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4274216890335083,
      "eval_runtime": 34.3672,
      "eval_samples_per_second": 19.903,
      "step": 32800
    },
    {
      "epoch": 7.84,
      "learning_rate": 1e-05,
      "loss": 1.434,
      "step": 32810
    },
    {
      "epoch": 7.84,
      "learning_rate": 1e-05,
      "loss": 1.4391,
      "step": 32820
    },
    {
      "epoch": 7.84,
      "learning_rate": 1e-05,
      "loss": 1.432,
      "step": 32830
    },
    {
      "epoch": 7.84,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 32840
    },
    {
      "epoch": 7.85,
      "learning_rate": 1e-05,
      "loss": 1.4322,
      "step": 32850
    },
    {
      "epoch": 7.85,
      "learning_rate": 1e-05,
      "loss": 1.4251,
      "step": 32860
    },
    {
      "epoch": 7.85,
      "learning_rate": 1e-05,
      "loss": 1.4313,
      "step": 32870
    },
    {
      "epoch": 7.85,
      "learning_rate": 1e-05,
      "loss": 1.4314,
      "step": 32880
    },
    {
      "epoch": 7.86,
      "learning_rate": 1e-05,
      "loss": 1.4297,
      "step": 32890
    },
    {
      "epoch": 7.86,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 32900
    },
    {
      "epoch": 7.86,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4282222986221313,
      "eval_runtime": 34.0504,
      "eval_samples_per_second": 20.088,
      "step": 32900
    },
    {
      "epoch": 7.86,
      "learning_rate": 1e-05,
      "loss": 1.4371,
      "step": 32910
    },
    {
      "epoch": 7.86,
      "learning_rate": 1e-05,
      "loss": 1.4309,
      "step": 32920
    },
    {
      "epoch": 7.86,
      "learning_rate": 1e-05,
      "loss": 1.4273,
      "step": 32930
    },
    {
      "epoch": 7.87,
      "learning_rate": 1e-05,
      "loss": 1.422,
      "step": 32940
    },
    {
      "epoch": 7.87,
      "learning_rate": 1e-05,
      "loss": 1.4399,
      "step": 32950
    },
    {
      "epoch": 7.87,
      "learning_rate": 1e-05,
      "loss": 1.4281,
      "step": 32960
    },
    {
      "epoch": 7.87,
      "learning_rate": 1e-05,
      "loss": 1.4283,
      "step": 32970
    },
    {
      "epoch": 7.88,
      "learning_rate": 1e-05,
      "loss": 1.4292,
      "step": 32980
    },
    {
      "epoch": 7.88,
      "learning_rate": 1e-05,
      "loss": 1.4366,
      "step": 32990
    },
    {
      "epoch": 7.88,
      "learning_rate": 1e-05,
      "loss": 1.433,
      "step": 33000
    },
    {
      "epoch": 7.88,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4280920028686523,
      "eval_runtime": 34.3936,
      "eval_samples_per_second": 19.887,
      "step": 33000
    },
    {
      "epoch": 7.88,
      "learning_rate": 1e-05,
      "loss": 1.4391,
      "step": 33010
    },
    {
      "epoch": 7.89,
      "learning_rate": 1e-05,
      "loss": 1.4305,
      "step": 33020
    },
    {
      "epoch": 7.89,
      "learning_rate": 1e-05,
      "loss": 1.433,
      "step": 33030
    },
    {
      "epoch": 7.89,
      "learning_rate": 1e-05,
      "loss": 1.4356,
      "step": 33040
    },
    {
      "epoch": 7.89,
      "learning_rate": 1e-05,
      "loss": 1.4397,
      "step": 33050
    },
    {
      "epoch": 7.9,
      "learning_rate": 1e-05,
      "loss": 1.4361,
      "step": 33060
    },
    {
      "epoch": 7.9,
      "learning_rate": 1e-05,
      "loss": 1.4319,
      "step": 33070
    },
    {
      "epoch": 7.9,
      "learning_rate": 1e-05,
      "loss": 1.4318,
      "step": 33080
    },
    {
      "epoch": 7.9,
      "learning_rate": 1e-05,
      "loss": 1.4245,
      "step": 33090
    },
    {
      "epoch": 7.91,
      "learning_rate": 1e-05,
      "loss": 1.43,
      "step": 33100
    },
    {
      "epoch": 7.91,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4285892248153687,
      "eval_runtime": 34.1905,
      "eval_samples_per_second": 20.006,
      "step": 33100
    },
    {
      "epoch": 7.91,
      "learning_rate": 1e-05,
      "loss": 1.4281,
      "step": 33110
    },
    {
      "epoch": 7.91,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 33120
    },
    {
      "epoch": 7.91,
      "learning_rate": 1e-05,
      "loss": 1.4314,
      "step": 33130
    },
    {
      "epoch": 7.91,
      "learning_rate": 1e-05,
      "loss": 1.4324,
      "step": 33140
    },
    {
      "epoch": 7.92,
      "learning_rate": 1e-05,
      "loss": 1.438,
      "step": 33150
    },
    {
      "epoch": 7.92,
      "learning_rate": 1e-05,
      "loss": 1.4306,
      "step": 33160
    },
    {
      "epoch": 7.92,
      "learning_rate": 1e-05,
      "loss": 1.4345,
      "step": 33170
    },
    {
      "epoch": 7.92,
      "learning_rate": 1e-05,
      "loss": 1.4337,
      "step": 33180
    },
    {
      "epoch": 7.93,
      "learning_rate": 1e-05,
      "loss": 1.4317,
      "step": 33190
    },
    {
      "epoch": 7.93,
      "learning_rate": 1e-05,
      "loss": 1.4327,
      "step": 33200
    },
    {
      "epoch": 7.93,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.428202509880066,
      "eval_runtime": 34.3943,
      "eval_samples_per_second": 19.887,
      "step": 33200
    },
    {
      "epoch": 7.93,
      "learning_rate": 1e-05,
      "loss": 1.4292,
      "step": 33210
    },
    {
      "epoch": 7.93,
      "learning_rate": 1e-05,
      "loss": 1.4323,
      "step": 33220
    },
    {
      "epoch": 7.94,
      "learning_rate": 1e-05,
      "loss": 1.4349,
      "step": 33230
    },
    {
      "epoch": 7.94,
      "learning_rate": 1e-05,
      "loss": 1.4264,
      "step": 33240
    },
    {
      "epoch": 7.94,
      "learning_rate": 1e-05,
      "loss": 1.4582,
      "step": 33250
    },
    {
      "epoch": 7.94,
      "learning_rate": 1e-05,
      "loss": 1.4282,
      "step": 33260
    },
    {
      "epoch": 7.95,
      "learning_rate": 1e-05,
      "loss": 1.435,
      "step": 33270
    },
    {
      "epoch": 7.95,
      "learning_rate": 1e-05,
      "loss": 1.4331,
      "step": 33280
    },
    {
      "epoch": 7.95,
      "learning_rate": 1e-05,
      "loss": 1.4335,
      "step": 33290
    },
    {
      "epoch": 7.95,
      "learning_rate": 1e-05,
      "loss": 1.4288,
      "step": 33300
    },
    {
      "epoch": 7.95,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4268418550491333,
      "eval_runtime": 34.3316,
      "eval_samples_per_second": 19.923,
      "step": 33300
    },
    {
      "epoch": 7.96,
      "learning_rate": 1e-05,
      "loss": 1.4723,
      "step": 33310
    },
    {
      "epoch": 7.96,
      "learning_rate": 1e-05,
      "loss": 1.4304,
      "step": 33320
    },
    {
      "epoch": 7.96,
      "learning_rate": 1e-05,
      "loss": 1.4271,
      "step": 33330
    },
    {
      "epoch": 7.96,
      "learning_rate": 1e-05,
      "loss": 1.4271,
      "step": 33340
    },
    {
      "epoch": 7.97,
      "learning_rate": 1e-05,
      "loss": 1.4366,
      "step": 33350
    },
    {
      "epoch": 7.97,
      "learning_rate": 1e-05,
      "loss": 1.438,
      "step": 33360
    },
    {
      "epoch": 7.97,
      "learning_rate": 1e-05,
      "loss": 1.4369,
      "step": 33370
    },
    {
      "epoch": 7.97,
      "learning_rate": 1e-05,
      "loss": 1.43,
      "step": 33380
    },
    {
      "epoch": 7.97,
      "learning_rate": 1e-05,
      "loss": 1.4459,
      "step": 33390
    },
    {
      "epoch": 7.98,
      "learning_rate": 1e-05,
      "loss": 1.4366,
      "step": 33400
    },
    {
      "epoch": 7.98,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4265297651290894,
      "eval_runtime": 34.4891,
      "eval_samples_per_second": 19.832,
      "step": 33400
    },
    {
      "epoch": 7.98,
      "learning_rate": 1e-05,
      "loss": 1.4338,
      "step": 33410
    },
    {
      "epoch": 7.98,
      "learning_rate": 1e-05,
      "loss": 1.4295,
      "step": 33420
    },
    {
      "epoch": 7.98,
      "learning_rate": 1e-05,
      "loss": 1.4355,
      "step": 33430
    },
    {
      "epoch": 7.99,
      "learning_rate": 1e-05,
      "loss": 1.432,
      "step": 33440
    },
    {
      "epoch": 7.99,
      "learning_rate": 1e-05,
      "loss": 1.4354,
      "step": 33450
    },
    {
      "epoch": 7.99,
      "learning_rate": 1e-05,
      "loss": 1.4297,
      "step": 33460
    },
    {
      "epoch": 7.99,
      "learning_rate": 1e-05,
      "loss": 1.4327,
      "step": 33470
    },
    {
      "epoch": 8.0,
      "learning_rate": 1e-05,
      "loss": 1.4292,
      "step": 33480
    },
    {
      "epoch": 8.0,
      "learning_rate": 1e-05,
      "loss": 1.4295,
      "step": 33490
    },
    {
      "epoch": 8.0,
      "learning_rate": 1e-05,
      "loss": 1.4277,
      "step": 33500
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4278669357299805,
      "eval_runtime": 34.2938,
      "eval_samples_per_second": 19.945,
      "step": 33500
    },
    {
      "epoch": 8.0,
      "learning_rate": 1e-05,
      "loss": 1.4301,
      "step": 33510
    },
    {
      "epoch": 8.01,
      "learning_rate": 1e-05,
      "loss": 1.4284,
      "step": 33520
    },
    {
      "epoch": 8.01,
      "learning_rate": 1e-05,
      "loss": 1.453,
      "step": 33530
    },
    {
      "epoch": 8.01,
      "learning_rate": 1e-05,
      "loss": 1.4341,
      "step": 33540
    },
    {
      "epoch": 8.01,
      "learning_rate": 1e-05,
      "loss": 1.4317,
      "step": 33550
    },
    {
      "epoch": 8.02,
      "learning_rate": 1e-05,
      "loss": 1.4307,
      "step": 33560
    },
    {
      "epoch": 8.02,
      "learning_rate": 1e-05,
      "loss": 1.4246,
      "step": 33570
    },
    {
      "epoch": 8.02,
      "learning_rate": 1e-05,
      "loss": 1.4341,
      "step": 33580
    },
    {
      "epoch": 8.02,
      "learning_rate": 1e-05,
      "loss": 1.4264,
      "step": 33590
    },
    {
      "epoch": 8.02,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 33600
    },
    {
      "epoch": 8.02,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4285528659820557,
      "eval_runtime": 34.4982,
      "eval_samples_per_second": 19.827,
      "step": 33600
    },
    {
      "epoch": 8.03,
      "learning_rate": 1e-05,
      "loss": 1.4329,
      "step": 33610
    },
    {
      "epoch": 8.03,
      "learning_rate": 1e-05,
      "loss": 1.4328,
      "step": 33620
    },
    {
      "epoch": 8.03,
      "learning_rate": 1e-05,
      "loss": 1.4335,
      "step": 33630
    },
    {
      "epoch": 8.03,
      "learning_rate": 1e-05,
      "loss": 1.4338,
      "step": 33640
    },
    {
      "epoch": 8.04,
      "learning_rate": 1e-05,
      "loss": 1.4282,
      "step": 33650
    },
    {
      "epoch": 8.04,
      "learning_rate": 1e-05,
      "loss": 1.4337,
      "step": 33660
    },
    {
      "epoch": 8.04,
      "learning_rate": 1e-05,
      "loss": 1.4366,
      "step": 33670
    },
    {
      "epoch": 8.04,
      "learning_rate": 1e-05,
      "loss": 1.4366,
      "step": 33680
    },
    {
      "epoch": 8.05,
      "learning_rate": 1e-05,
      "loss": 1.4265,
      "step": 33690
    },
    {
      "epoch": 8.05,
      "learning_rate": 1e-05,
      "loss": 1.4373,
      "step": 33700
    },
    {
      "epoch": 8.05,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4264594316482544,
      "eval_runtime": 34.6646,
      "eval_samples_per_second": 19.732,
      "step": 33700
    },
    {
      "epoch": 8.05,
      "learning_rate": 1e-05,
      "loss": 1.4344,
      "step": 33710
    },
    {
      "epoch": 8.05,
      "learning_rate": 1e-05,
      "loss": 1.4343,
      "step": 33720
    },
    {
      "epoch": 8.06,
      "learning_rate": 1e-05,
      "loss": 1.4261,
      "step": 33730
    },
    {
      "epoch": 8.06,
      "learning_rate": 1e-05,
      "loss": 1.4253,
      "step": 33740
    },
    {
      "epoch": 8.06,
      "learning_rate": 1e-05,
      "loss": 1.4344,
      "step": 33750
    },
    {
      "epoch": 8.06,
      "learning_rate": 1e-05,
      "loss": 1.4273,
      "step": 33760
    },
    {
      "epoch": 8.07,
      "learning_rate": 1e-05,
      "loss": 1.4304,
      "step": 33770
    },
    {
      "epoch": 8.07,
      "learning_rate": 1e-05,
      "loss": 1.4323,
      "step": 33780
    },
    {
      "epoch": 8.07,
      "learning_rate": 1e-05,
      "loss": 1.4319,
      "step": 33790
    },
    {
      "epoch": 8.07,
      "learning_rate": 1e-05,
      "loss": 1.4392,
      "step": 33800
    },
    {
      "epoch": 8.07,
      "eval_accuracy": 0.9751461988304093,
      "eval_loss": 1.427140235900879,
      "eval_runtime": 34.5583,
      "eval_samples_per_second": 19.793,
      "step": 33800
    },
    {
      "epoch": 8.07,
      "learning_rate": 1e-05,
      "loss": 1.432,
      "step": 33810
    },
    {
      "epoch": 8.08,
      "learning_rate": 1e-05,
      "loss": 1.4317,
      "step": 33820
    },
    {
      "epoch": 8.08,
      "learning_rate": 1e-05,
      "loss": 1.4209,
      "step": 33830
    },
    {
      "epoch": 8.08,
      "learning_rate": 1e-05,
      "loss": 1.4322,
      "step": 33840
    },
    {
      "epoch": 8.08,
      "learning_rate": 1e-05,
      "loss": 1.426,
      "step": 33850
    },
    {
      "epoch": 8.09,
      "learning_rate": 1e-05,
      "loss": 1.4321,
      "step": 33860
    },
    {
      "epoch": 8.09,
      "learning_rate": 1e-05,
      "loss": 1.4363,
      "step": 33870
    },
    {
      "epoch": 8.09,
      "learning_rate": 1e-05,
      "loss": 1.4426,
      "step": 33880
    },
    {
      "epoch": 8.09,
      "learning_rate": 1e-05,
      "loss": 1.4304,
      "step": 33890
    },
    {
      "epoch": 8.1,
      "learning_rate": 1e-05,
      "loss": 1.4271,
      "step": 33900
    },
    {
      "epoch": 8.1,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4266947507858276,
      "eval_runtime": 34.1923,
      "eval_samples_per_second": 20.004,
      "step": 33900
    },
    {
      "epoch": 8.1,
      "learning_rate": 1e-05,
      "loss": 1.4298,
      "step": 33910
    },
    {
      "epoch": 8.1,
      "learning_rate": 1e-05,
      "loss": 1.4304,
      "step": 33920
    },
    {
      "epoch": 8.1,
      "learning_rate": 1e-05,
      "loss": 1.4379,
      "step": 33930
    },
    {
      "epoch": 8.11,
      "learning_rate": 1e-05,
      "loss": 1.4337,
      "step": 33940
    },
    {
      "epoch": 8.11,
      "learning_rate": 1e-05,
      "loss": 1.4357,
      "step": 33950
    },
    {
      "epoch": 8.11,
      "learning_rate": 1e-05,
      "loss": 1.4246,
      "step": 33960
    },
    {
      "epoch": 8.11,
      "learning_rate": 1e-05,
      "loss": 1.4348,
      "step": 33970
    },
    {
      "epoch": 8.12,
      "learning_rate": 1e-05,
      "loss": 1.441,
      "step": 33980
    },
    {
      "epoch": 8.12,
      "learning_rate": 1e-05,
      "loss": 1.428,
      "step": 33990
    },
    {
      "epoch": 8.12,
      "learning_rate": 1e-05,
      "loss": 1.437,
      "step": 34000
    },
    {
      "epoch": 8.12,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4275610446929932,
      "eval_runtime": 34.2722,
      "eval_samples_per_second": 19.958,
      "step": 34000
    },
    {
      "epoch": 8.12,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 34010
    },
    {
      "epoch": 8.13,
      "learning_rate": 1e-05,
      "loss": 1.4266,
      "step": 34020
    },
    {
      "epoch": 8.13,
      "learning_rate": 1e-05,
      "loss": 1.4339,
      "step": 34030
    },
    {
      "epoch": 8.13,
      "learning_rate": 1e-05,
      "loss": 1.4257,
      "step": 34040
    },
    {
      "epoch": 8.13,
      "learning_rate": 1e-05,
      "loss": 1.4259,
      "step": 34050
    },
    {
      "epoch": 8.13,
      "learning_rate": 1e-05,
      "loss": 1.4278,
      "step": 34060
    },
    {
      "epoch": 8.14,
      "learning_rate": 1e-05,
      "loss": 1.4442,
      "step": 34070
    },
    {
      "epoch": 8.14,
      "learning_rate": 1e-05,
      "loss": 1.432,
      "step": 34080
    },
    {
      "epoch": 8.14,
      "learning_rate": 1e-05,
      "loss": 1.4347,
      "step": 34090
    },
    {
      "epoch": 8.14,
      "learning_rate": 1e-05,
      "loss": 1.4289,
      "step": 34100
    },
    {
      "epoch": 8.14,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4270378351211548,
      "eval_runtime": 33.7768,
      "eval_samples_per_second": 20.251,
      "step": 34100
    },
    {
      "epoch": 8.15,
      "learning_rate": 1e-05,
      "loss": 1.4243,
      "step": 34110
    },
    {
      "epoch": 8.15,
      "learning_rate": 1e-05,
      "loss": 1.4308,
      "step": 34120
    },
    {
      "epoch": 8.15,
      "learning_rate": 1e-05,
      "loss": 1.4388,
      "step": 34130
    },
    {
      "epoch": 8.15,
      "learning_rate": 1e-05,
      "loss": 1.4347,
      "step": 34140
    },
    {
      "epoch": 8.16,
      "learning_rate": 1e-05,
      "loss": 1.427,
      "step": 34150
    },
    {
      "epoch": 8.16,
      "learning_rate": 1e-05,
      "loss": 1.4262,
      "step": 34160
    },
    {
      "epoch": 8.16,
      "learning_rate": 1e-05,
      "loss": 1.4315,
      "step": 34170
    },
    {
      "epoch": 8.16,
      "learning_rate": 1e-05,
      "loss": 1.4319,
      "step": 34180
    },
    {
      "epoch": 8.17,
      "learning_rate": 1e-05,
      "loss": 1.4253,
      "step": 34190
    },
    {
      "epoch": 8.17,
      "learning_rate": 1e-05,
      "loss": 1.4308,
      "step": 34200
    },
    {
      "epoch": 8.17,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4259949922561646,
      "eval_runtime": 34.082,
      "eval_samples_per_second": 20.069,
      "step": 34200
    },
    {
      "epoch": 8.17,
      "learning_rate": 1e-05,
      "loss": 1.4241,
      "step": 34210
    },
    {
      "epoch": 8.17,
      "learning_rate": 1e-05,
      "loss": 1.4285,
      "step": 34220
    },
    {
      "epoch": 8.18,
      "learning_rate": 1e-05,
      "loss": 1.4364,
      "step": 34230
    },
    {
      "epoch": 8.18,
      "learning_rate": 1e-05,
      "loss": 1.4263,
      "step": 34240
    },
    {
      "epoch": 8.18,
      "learning_rate": 1e-05,
      "loss": 1.432,
      "step": 34250
    },
    {
      "epoch": 8.18,
      "learning_rate": 1e-05,
      "loss": 1.4383,
      "step": 34260
    },
    {
      "epoch": 8.18,
      "learning_rate": 1e-05,
      "loss": 1.4308,
      "step": 34270
    },
    {
      "epoch": 8.19,
      "learning_rate": 1e-05,
      "loss": 1.4388,
      "step": 34280
    },
    {
      "epoch": 8.19,
      "learning_rate": 1e-05,
      "loss": 1.4293,
      "step": 34290
    },
    {
      "epoch": 8.19,
      "learning_rate": 1e-05,
      "loss": 1.4319,
      "step": 34300
    },
    {
      "epoch": 8.19,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4260886907577515,
      "eval_runtime": 33.8487,
      "eval_samples_per_second": 20.208,
      "step": 34300
    },
    {
      "epoch": 8.19,
      "learning_rate": 1e-05,
      "loss": 1.429,
      "step": 34310
    },
    {
      "epoch": 8.2,
      "learning_rate": 1e-05,
      "loss": 1.4281,
      "step": 34320
    },
    {
      "epoch": 8.2,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 34330
    },
    {
      "epoch": 8.2,
      "learning_rate": 1e-05,
      "loss": 1.4234,
      "step": 34340
    },
    {
      "epoch": 8.2,
      "learning_rate": 1e-05,
      "loss": 1.4275,
      "step": 34350
    },
    {
      "epoch": 8.21,
      "learning_rate": 1e-05,
      "loss": 1.4289,
      "step": 34360
    },
    {
      "epoch": 8.21,
      "learning_rate": 1e-05,
      "loss": 1.426,
      "step": 34370
    },
    {
      "epoch": 8.21,
      "learning_rate": 1e-05,
      "loss": 1.4249,
      "step": 34380
    },
    {
      "epoch": 8.21,
      "learning_rate": 1e-05,
      "loss": 1.4303,
      "step": 34390
    },
    {
      "epoch": 8.22,
      "learning_rate": 1e-05,
      "loss": 1.4308,
      "step": 34400
    },
    {
      "epoch": 8.22,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4252005815505981,
      "eval_runtime": 34.0665,
      "eval_samples_per_second": 20.078,
      "step": 34400
    },
    {
      "epoch": 8.22,
      "learning_rate": 1e-05,
      "loss": 1.4266,
      "step": 34410
    },
    {
      "epoch": 8.22,
      "learning_rate": 1e-05,
      "loss": 1.4292,
      "step": 34420
    },
    {
      "epoch": 8.22,
      "learning_rate": 1e-05,
      "loss": 1.4336,
      "step": 34430
    },
    {
      "epoch": 8.23,
      "learning_rate": 1e-05,
      "loss": 1.4252,
      "step": 34440
    },
    {
      "epoch": 8.23,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 34450
    },
    {
      "epoch": 8.23,
      "learning_rate": 1e-05,
      "loss": 1.438,
      "step": 34460
    },
    {
      "epoch": 8.23,
      "learning_rate": 1e-05,
      "loss": 1.4309,
      "step": 34470
    },
    {
      "epoch": 8.24,
      "learning_rate": 1e-05,
      "loss": 1.4364,
      "step": 34480
    },
    {
      "epoch": 8.24,
      "learning_rate": 1e-05,
      "loss": 1.426,
      "step": 34490
    },
    {
      "epoch": 8.24,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 34500
    },
    {
      "epoch": 8.24,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4260541200637817,
      "eval_runtime": 34.0152,
      "eval_samples_per_second": 20.109,
      "step": 34500
    },
    {
      "epoch": 8.24,
      "learning_rate": 1e-05,
      "loss": 1.4241,
      "step": 34510
    },
    {
      "epoch": 8.24,
      "learning_rate": 1e-05,
      "loss": 1.4383,
      "step": 34520
    },
    {
      "epoch": 8.25,
      "learning_rate": 1e-05,
      "loss": 1.4257,
      "step": 34530
    },
    {
      "epoch": 8.25,
      "learning_rate": 1e-05,
      "loss": 1.4263,
      "step": 34540
    },
    {
      "epoch": 8.25,
      "learning_rate": 1e-05,
      "loss": 1.4287,
      "step": 34550
    },
    {
      "epoch": 8.25,
      "learning_rate": 1e-05,
      "loss": 1.4338,
      "step": 34560
    },
    {
      "epoch": 8.26,
      "learning_rate": 1e-05,
      "loss": 1.4311,
      "step": 34570
    },
    {
      "epoch": 8.26,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 34580
    },
    {
      "epoch": 8.26,
      "learning_rate": 1e-05,
      "loss": 1.4248,
      "step": 34590
    },
    {
      "epoch": 8.26,
      "learning_rate": 1e-05,
      "loss": 1.4316,
      "step": 34600
    },
    {
      "epoch": 8.26,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4275717735290527,
      "eval_runtime": 34.0541,
      "eval_samples_per_second": 20.086,
      "step": 34600
    },
    {
      "epoch": 8.27,
      "learning_rate": 1e-05,
      "loss": 1.4259,
      "step": 34610
    },
    {
      "epoch": 8.27,
      "learning_rate": 1e-05,
      "loss": 1.4292,
      "step": 34620
    },
    {
      "epoch": 8.27,
      "learning_rate": 1e-05,
      "loss": 1.4267,
      "step": 34630
    },
    {
      "epoch": 8.27,
      "learning_rate": 1e-05,
      "loss": 1.4349,
      "step": 34640
    },
    {
      "epoch": 8.28,
      "learning_rate": 1e-05,
      "loss": 1.4311,
      "step": 34650
    },
    {
      "epoch": 8.28,
      "learning_rate": 1e-05,
      "loss": 1.4316,
      "step": 34660
    },
    {
      "epoch": 8.28,
      "learning_rate": 1e-05,
      "loss": 1.4316,
      "step": 34670
    },
    {
      "epoch": 8.28,
      "learning_rate": 1e-05,
      "loss": 1.4346,
      "step": 34680
    },
    {
      "epoch": 8.29,
      "learning_rate": 1e-05,
      "loss": 1.4436,
      "step": 34690
    },
    {
      "epoch": 8.29,
      "learning_rate": 1e-05,
      "loss": 1.4286,
      "step": 34700
    },
    {
      "epoch": 8.29,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.426818609237671,
      "eval_runtime": 33.9559,
      "eval_samples_per_second": 20.144,
      "step": 34700
    },
    {
      "epoch": 8.29,
      "learning_rate": 1e-05,
      "loss": 1.434,
      "step": 34710
    },
    {
      "epoch": 8.29,
      "learning_rate": 1e-05,
      "loss": 1.4244,
      "step": 34720
    },
    {
      "epoch": 8.29,
      "learning_rate": 1e-05,
      "loss": 1.425,
      "step": 34730
    },
    {
      "epoch": 8.3,
      "learning_rate": 1e-05,
      "loss": 1.4254,
      "step": 34740
    },
    {
      "epoch": 8.3,
      "learning_rate": 1e-05,
      "loss": 1.4319,
      "step": 34750
    },
    {
      "epoch": 8.3,
      "learning_rate": 1e-05,
      "loss": 1.426,
      "step": 34760
    },
    {
      "epoch": 8.3,
      "learning_rate": 1e-05,
      "loss": 1.4347,
      "step": 34770
    },
    {
      "epoch": 8.31,
      "learning_rate": 1e-05,
      "loss": 1.4352,
      "step": 34780
    },
    {
      "epoch": 8.31,
      "learning_rate": 1e-05,
      "loss": 1.4359,
      "step": 34790
    },
    {
      "epoch": 8.31,
      "learning_rate": 1e-05,
      "loss": 1.4386,
      "step": 34800
    },
    {
      "epoch": 8.31,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4272431135177612,
      "eval_runtime": 34.0721,
      "eval_samples_per_second": 20.075,
      "step": 34800
    },
    {
      "epoch": 8.31,
      "learning_rate": 1e-05,
      "loss": 1.4277,
      "step": 34810
    },
    {
      "epoch": 8.32,
      "learning_rate": 1e-05,
      "loss": 1.4315,
      "step": 34820
    },
    {
      "epoch": 8.32,
      "learning_rate": 1e-05,
      "loss": 1.4233,
      "step": 34830
    },
    {
      "epoch": 8.32,
      "learning_rate": 1e-05,
      "loss": 1.4314,
      "step": 34840
    },
    {
      "epoch": 8.32,
      "learning_rate": 1e-05,
      "loss": 1.4281,
      "step": 34850
    },
    {
      "epoch": 8.33,
      "learning_rate": 1e-05,
      "loss": 1.4247,
      "step": 34860
    },
    {
      "epoch": 8.33,
      "learning_rate": 1e-05,
      "loss": 1.435,
      "step": 34870
    },
    {
      "epoch": 8.33,
      "learning_rate": 1e-05,
      "loss": 1.4232,
      "step": 34880
    },
    {
      "epoch": 8.33,
      "learning_rate": 1e-05,
      "loss": 1.4376,
      "step": 34890
    },
    {
      "epoch": 8.34,
      "learning_rate": 1e-05,
      "loss": 1.4274,
      "step": 34900
    },
    {
      "epoch": 8.34,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.426880121231079,
      "eval_runtime": 33.9087,
      "eval_samples_per_second": 20.172,
      "step": 34900
    },
    {
      "epoch": 8.34,
      "learning_rate": 1e-05,
      "loss": 1.4333,
      "step": 34910
    },
    {
      "epoch": 8.34,
      "learning_rate": 1e-05,
      "loss": 1.4282,
      "step": 34920
    },
    {
      "epoch": 8.34,
      "learning_rate": 1e-05,
      "loss": 1.4313,
      "step": 34930
    },
    {
      "epoch": 8.34,
      "learning_rate": 1e-05,
      "loss": 1.4292,
      "step": 34940
    },
    {
      "epoch": 8.35,
      "learning_rate": 1e-05,
      "loss": 1.4256,
      "step": 34950
    },
    {
      "epoch": 8.35,
      "learning_rate": 1e-05,
      "loss": 1.4237,
      "step": 34960
    },
    {
      "epoch": 8.35,
      "learning_rate": 1e-05,
      "loss": 1.4281,
      "step": 34970
    },
    {
      "epoch": 8.35,
      "learning_rate": 1e-05,
      "loss": 1.4336,
      "step": 34980
    },
    {
      "epoch": 8.36,
      "learning_rate": 1e-05,
      "loss": 1.4318,
      "step": 34990
    },
    {
      "epoch": 8.36,
      "learning_rate": 1e-05,
      "loss": 1.4262,
      "step": 35000
    },
    {
      "epoch": 8.36,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.42515230178833,
      "eval_runtime": 34.1483,
      "eval_samples_per_second": 20.03,
      "step": 35000
    },
    {
      "epoch": 8.36,
      "learning_rate": 1e-05,
      "loss": 1.428,
      "step": 35010
    },
    {
      "epoch": 8.36,
      "learning_rate": 1e-05,
      "loss": 1.4316,
      "step": 35020
    },
    {
      "epoch": 8.37,
      "learning_rate": 1e-05,
      "loss": 1.4356,
      "step": 35030
    },
    {
      "epoch": 8.37,
      "learning_rate": 1e-05,
      "loss": 1.4295,
      "step": 35040
    },
    {
      "epoch": 8.37,
      "learning_rate": 1e-05,
      "loss": 1.4407,
      "step": 35050
    },
    {
      "epoch": 8.37,
      "learning_rate": 1e-05,
      "loss": 1.4358,
      "step": 35060
    },
    {
      "epoch": 8.38,
      "learning_rate": 1e-05,
      "loss": 1.4352,
      "step": 35070
    },
    {
      "epoch": 8.38,
      "learning_rate": 1e-05,
      "loss": 1.4275,
      "step": 35080
    },
    {
      "epoch": 8.38,
      "learning_rate": 1e-05,
      "loss": 1.4291,
      "step": 35090
    },
    {
      "epoch": 8.38,
      "learning_rate": 1e-05,
      "loss": 1.4278,
      "step": 35100
    },
    {
      "epoch": 8.38,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4260330200195312,
      "eval_runtime": 34.1119,
      "eval_samples_per_second": 20.052,
      "step": 35100
    },
    {
      "epoch": 8.39,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 35110
    },
    {
      "epoch": 8.39,
      "learning_rate": 1e-05,
      "loss": 1.4325,
      "step": 35120
    },
    {
      "epoch": 8.39,
      "learning_rate": 1e-05,
      "loss": 1.4296,
      "step": 35130
    },
    {
      "epoch": 8.39,
      "learning_rate": 1e-05,
      "loss": 1.4265,
      "step": 35140
    },
    {
      "epoch": 8.4,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 35150
    },
    {
      "epoch": 8.4,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 35160
    },
    {
      "epoch": 8.4,
      "learning_rate": 1e-05,
      "loss": 1.4299,
      "step": 35170
    },
    {
      "epoch": 8.4,
      "learning_rate": 1e-05,
      "loss": 1.4282,
      "step": 35180
    },
    {
      "epoch": 8.4,
      "learning_rate": 1e-05,
      "loss": 1.426,
      "step": 35190
    },
    {
      "epoch": 8.41,
      "learning_rate": 1e-05,
      "loss": 1.4293,
      "step": 35200
    },
    {
      "epoch": 8.41,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4258756637573242,
      "eval_runtime": 34.3539,
      "eval_samples_per_second": 19.91,
      "step": 35200
    },
    {
      "epoch": 8.41,
      "learning_rate": 1e-05,
      "loss": 1.4245,
      "step": 35210
    },
    {
      "epoch": 8.41,
      "learning_rate": 1e-05,
      "loss": 1.4368,
      "step": 35220
    },
    {
      "epoch": 8.41,
      "learning_rate": 1e-05,
      "loss": 1.4346,
      "step": 35230
    },
    {
      "epoch": 8.42,
      "learning_rate": 1e-05,
      "loss": 1.4264,
      "step": 35240
    },
    {
      "epoch": 8.42,
      "learning_rate": 1e-05,
      "loss": 1.4213,
      "step": 35250
    },
    {
      "epoch": 8.42,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 35260
    },
    {
      "epoch": 8.42,
      "learning_rate": 1e-05,
      "loss": 1.4327,
      "step": 35270
    },
    {
      "epoch": 8.43,
      "learning_rate": 1e-05,
      "loss": 1.4265,
      "step": 35280
    },
    {
      "epoch": 8.43,
      "learning_rate": 1e-05,
      "loss": 1.4307,
      "step": 35290
    },
    {
      "epoch": 8.43,
      "learning_rate": 1e-05,
      "loss": 1.4348,
      "step": 35300
    },
    {
      "epoch": 8.43,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4267126321792603,
      "eval_runtime": 33.8965,
      "eval_samples_per_second": 20.179,
      "step": 35300
    },
    {
      "epoch": 8.43,
      "learning_rate": 1e-05,
      "loss": 1.4348,
      "step": 35310
    },
    {
      "epoch": 8.44,
      "learning_rate": 1e-05,
      "loss": 1.4364,
      "step": 35320
    },
    {
      "epoch": 8.44,
      "learning_rate": 1e-05,
      "loss": 1.427,
      "step": 35330
    },
    {
      "epoch": 8.44,
      "learning_rate": 1e-05,
      "loss": 1.4303,
      "step": 35340
    },
    {
      "epoch": 8.44,
      "learning_rate": 1e-05,
      "loss": 1.4266,
      "step": 35350
    },
    {
      "epoch": 8.45,
      "learning_rate": 1e-05,
      "loss": 1.4291,
      "step": 35360
    },
    {
      "epoch": 8.45,
      "learning_rate": 1e-05,
      "loss": 1.431,
      "step": 35370
    },
    {
      "epoch": 8.45,
      "learning_rate": 1e-05,
      "loss": 1.4299,
      "step": 35380
    },
    {
      "epoch": 8.45,
      "learning_rate": 1e-05,
      "loss": 1.4242,
      "step": 35390
    },
    {
      "epoch": 8.45,
      "learning_rate": 1e-05,
      "loss": 1.4341,
      "step": 35400
    },
    {
      "epoch": 8.45,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4247909784317017,
      "eval_runtime": 34.1286,
      "eval_samples_per_second": 20.042,
      "step": 35400
    },
    {
      "epoch": 8.46,
      "learning_rate": 1e-05,
      "loss": 1.4302,
      "step": 35410
    },
    {
      "epoch": 8.46,
      "learning_rate": 1e-05,
      "loss": 1.425,
      "step": 35420
    },
    {
      "epoch": 8.46,
      "learning_rate": 1e-05,
      "loss": 1.4371,
      "step": 35430
    },
    {
      "epoch": 8.46,
      "learning_rate": 1e-05,
      "loss": 1.4314,
      "step": 35440
    },
    {
      "epoch": 8.47,
      "learning_rate": 1e-05,
      "loss": 1.4257,
      "step": 35450
    },
    {
      "epoch": 8.47,
      "learning_rate": 1e-05,
      "loss": 1.4298,
      "step": 35460
    },
    {
      "epoch": 8.47,
      "learning_rate": 1e-05,
      "loss": 1.4344,
      "step": 35470
    },
    {
      "epoch": 8.47,
      "learning_rate": 1e-05,
      "loss": 1.4271,
      "step": 35480
    },
    {
      "epoch": 8.48,
      "learning_rate": 1e-05,
      "loss": 1.4268,
      "step": 35490
    },
    {
      "epoch": 8.48,
      "learning_rate": 1e-05,
      "loss": 1.4317,
      "step": 35500
    },
    {
      "epoch": 8.48,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4251840114593506,
      "eval_runtime": 33.9451,
      "eval_samples_per_second": 20.15,
      "step": 35500
    },
    {
      "epoch": 8.48,
      "learning_rate": 1e-05,
      "loss": 1.4303,
      "step": 35510
    },
    {
      "epoch": 8.48,
      "learning_rate": 1e-05,
      "loss": 1.4317,
      "step": 35520
    },
    {
      "epoch": 8.49,
      "learning_rate": 1e-05,
      "loss": 1.4322,
      "step": 35530
    },
    {
      "epoch": 8.49,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 35540
    },
    {
      "epoch": 8.49,
      "learning_rate": 1e-05,
      "loss": 1.4299,
      "step": 35550
    },
    {
      "epoch": 8.49,
      "learning_rate": 1e-05,
      "loss": 1.4292,
      "step": 35560
    },
    {
      "epoch": 8.5,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 35570
    },
    {
      "epoch": 8.5,
      "learning_rate": 1e-05,
      "loss": 1.4263,
      "step": 35580
    },
    {
      "epoch": 8.5,
      "learning_rate": 1e-05,
      "loss": 1.4301,
      "step": 35590
    },
    {
      "epoch": 8.5,
      "learning_rate": 1e-05,
      "loss": 1.4323,
      "step": 35600
    },
    {
      "epoch": 8.5,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4257274866104126,
      "eval_runtime": 33.5067,
      "eval_samples_per_second": 20.414,
      "step": 35600
    },
    {
      "epoch": 8.5,
      "learning_rate": 1e-05,
      "loss": 1.4359,
      "step": 35610
    },
    {
      "epoch": 8.51,
      "learning_rate": 1e-05,
      "loss": 1.4333,
      "step": 35620
    },
    {
      "epoch": 8.51,
      "learning_rate": 1e-05,
      "loss": 1.4241,
      "step": 35630
    },
    {
      "epoch": 8.51,
      "learning_rate": 1e-05,
      "loss": 1.4303,
      "step": 35640
    },
    {
      "epoch": 8.51,
      "learning_rate": 1e-05,
      "loss": 1.4278,
      "step": 35650
    },
    {
      "epoch": 8.52,
      "learning_rate": 1e-05,
      "loss": 1.4301,
      "step": 35660
    },
    {
      "epoch": 8.52,
      "learning_rate": 1e-05,
      "loss": 1.4283,
      "step": 35670
    },
    {
      "epoch": 8.52,
      "learning_rate": 1e-05,
      "loss": 1.4314,
      "step": 35680
    },
    {
      "epoch": 8.52,
      "learning_rate": 1e-05,
      "loss": 1.4336,
      "step": 35690
    },
    {
      "epoch": 8.53,
      "learning_rate": 1e-05,
      "loss": 1.4259,
      "step": 35700
    },
    {
      "epoch": 8.53,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4255626201629639,
      "eval_runtime": 33.7542,
      "eval_samples_per_second": 20.264,
      "step": 35700
    },
    {
      "epoch": 8.53,
      "learning_rate": 1e-05,
      "loss": 1.4287,
      "step": 35710
    },
    {
      "epoch": 8.53,
      "learning_rate": 1e-05,
      "loss": 1.4273,
      "step": 35720
    },
    {
      "epoch": 8.53,
      "learning_rate": 1e-05,
      "loss": 1.4236,
      "step": 35730
    },
    {
      "epoch": 8.54,
      "learning_rate": 1e-05,
      "loss": 1.4252,
      "step": 35740
    },
    {
      "epoch": 8.54,
      "learning_rate": 1e-05,
      "loss": 1.4268,
      "step": 35750
    },
    {
      "epoch": 8.54,
      "learning_rate": 1e-05,
      "loss": 1.4322,
      "step": 35760
    },
    {
      "epoch": 8.54,
      "learning_rate": 1e-05,
      "loss": 1.4225,
      "step": 35770
    },
    {
      "epoch": 8.55,
      "learning_rate": 1e-05,
      "loss": 1.427,
      "step": 35780
    },
    {
      "epoch": 8.55,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 35790
    },
    {
      "epoch": 8.55,
      "learning_rate": 1e-05,
      "loss": 1.4289,
      "step": 35800
    },
    {
      "epoch": 8.55,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4244177341461182,
      "eval_runtime": 33.9139,
      "eval_samples_per_second": 20.169,
      "step": 35800
    },
    {
      "epoch": 8.55,
      "learning_rate": 1e-05,
      "loss": 1.4331,
      "step": 35810
    },
    {
      "epoch": 8.56,
      "learning_rate": 1e-05,
      "loss": 1.4353,
      "step": 35820
    },
    {
      "epoch": 8.56,
      "learning_rate": 1e-05,
      "loss": 1.4336,
      "step": 35830
    },
    {
      "epoch": 8.56,
      "learning_rate": 1e-05,
      "loss": 1.4349,
      "step": 35840
    },
    {
      "epoch": 8.56,
      "learning_rate": 1e-05,
      "loss": 1.4296,
      "step": 35850
    },
    {
      "epoch": 8.56,
      "learning_rate": 1e-05,
      "loss": 1.4234,
      "step": 35860
    },
    {
      "epoch": 8.57,
      "learning_rate": 1e-05,
      "loss": 1.4234,
      "step": 35870
    },
    {
      "epoch": 8.57,
      "learning_rate": 1e-05,
      "loss": 1.4341,
      "step": 35880
    },
    {
      "epoch": 8.57,
      "learning_rate": 1e-05,
      "loss": 1.4316,
      "step": 35890
    },
    {
      "epoch": 8.57,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 35900
    },
    {
      "epoch": 8.57,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4253084659576416,
      "eval_runtime": 34.0182,
      "eval_samples_per_second": 20.107,
      "step": 35900
    },
    {
      "epoch": 8.58,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 35910
    },
    {
      "epoch": 8.58,
      "learning_rate": 1e-05,
      "loss": 1.4348,
      "step": 35920
    },
    {
      "epoch": 8.58,
      "learning_rate": 1e-05,
      "loss": 1.425,
      "step": 35930
    },
    {
      "epoch": 8.58,
      "learning_rate": 1e-05,
      "loss": 1.4307,
      "step": 35940
    },
    {
      "epoch": 8.59,
      "learning_rate": 1e-05,
      "loss": 1.426,
      "step": 35950
    },
    {
      "epoch": 8.59,
      "learning_rate": 1e-05,
      "loss": 1.4335,
      "step": 35960
    },
    {
      "epoch": 8.59,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 35970
    },
    {
      "epoch": 8.59,
      "learning_rate": 1e-05,
      "loss": 1.4313,
      "step": 35980
    },
    {
      "epoch": 8.6,
      "learning_rate": 1e-05,
      "loss": 1.4311,
      "step": 35990
    },
    {
      "epoch": 8.6,
      "learning_rate": 1e-05,
      "loss": 1.4297,
      "step": 36000
    },
    {
      "epoch": 8.6,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.425299882888794,
      "eval_runtime": 34.0085,
      "eval_samples_per_second": 20.113,
      "step": 36000
    },
    {
      "epoch": 8.6,
      "learning_rate": 1e-05,
      "loss": 1.4271,
      "step": 36010
    },
    {
      "epoch": 8.6,
      "learning_rate": 1e-05,
      "loss": 1.4328,
      "step": 36020
    },
    {
      "epoch": 8.61,
      "learning_rate": 1e-05,
      "loss": 1.4268,
      "step": 36030
    },
    {
      "epoch": 8.61,
      "learning_rate": 1e-05,
      "loss": 1.4298,
      "step": 36040
    },
    {
      "epoch": 8.61,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 36050
    },
    {
      "epoch": 8.61,
      "learning_rate": 1e-05,
      "loss": 1.4333,
      "step": 36060
    },
    {
      "epoch": 8.61,
      "learning_rate": 1e-05,
      "loss": 1.4369,
      "step": 36070
    },
    {
      "epoch": 8.62,
      "learning_rate": 1e-05,
      "loss": 1.4208,
      "step": 36080
    },
    {
      "epoch": 8.62,
      "learning_rate": 1e-05,
      "loss": 1.4322,
      "step": 36090
    },
    {
      "epoch": 8.62,
      "learning_rate": 1e-05,
      "loss": 1.4308,
      "step": 36100
    },
    {
      "epoch": 8.62,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4243108034133911,
      "eval_runtime": 33.7877,
      "eval_samples_per_second": 20.244,
      "step": 36100
    },
    {
      "epoch": 8.62,
      "learning_rate": 1e-05,
      "loss": 1.429,
      "step": 36110
    },
    {
      "epoch": 8.63,
      "learning_rate": 1e-05,
      "loss": 1.4384,
      "step": 36120
    },
    {
      "epoch": 8.63,
      "learning_rate": 1e-05,
      "loss": 1.4267,
      "step": 36130
    },
    {
      "epoch": 8.63,
      "learning_rate": 1e-05,
      "loss": 1.4295,
      "step": 36140
    },
    {
      "epoch": 8.63,
      "learning_rate": 1e-05,
      "loss": 1.425,
      "step": 36150
    },
    {
      "epoch": 8.64,
      "learning_rate": 1e-05,
      "loss": 1.4328,
      "step": 36160
    },
    {
      "epoch": 8.64,
      "learning_rate": 1e-05,
      "loss": 1.4293,
      "step": 36170
    },
    {
      "epoch": 8.64,
      "learning_rate": 1e-05,
      "loss": 1.4292,
      "step": 36180
    },
    {
      "epoch": 8.64,
      "learning_rate": 1e-05,
      "loss": 1.4363,
      "step": 36190
    },
    {
      "epoch": 8.65,
      "learning_rate": 1e-05,
      "loss": 1.4277,
      "step": 36200
    },
    {
      "epoch": 8.65,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.425471305847168,
      "eval_runtime": 33.6983,
      "eval_samples_per_second": 20.298,
      "step": 36200
    },
    {
      "epoch": 8.65,
      "learning_rate": 1e-05,
      "loss": 1.4357,
      "step": 36210
    },
    {
      "epoch": 8.65,
      "learning_rate": 1e-05,
      "loss": 1.423,
      "step": 36220
    },
    {
      "epoch": 8.65,
      "learning_rate": 1e-05,
      "loss": 1.4302,
      "step": 36230
    },
    {
      "epoch": 8.66,
      "learning_rate": 1e-05,
      "loss": 1.4289,
      "step": 36240
    },
    {
      "epoch": 8.66,
      "learning_rate": 1e-05,
      "loss": 1.4295,
      "step": 36250
    },
    {
      "epoch": 8.66,
      "learning_rate": 1e-05,
      "loss": 1.4341,
      "step": 36260
    },
    {
      "epoch": 8.66,
      "learning_rate": 1e-05,
      "loss": 1.4253,
      "step": 36270
    },
    {
      "epoch": 8.66,
      "learning_rate": 1e-05,
      "loss": 1.4332,
      "step": 36280
    },
    {
      "epoch": 8.67,
      "learning_rate": 1e-05,
      "loss": 1.4295,
      "step": 36290
    },
    {
      "epoch": 8.67,
      "learning_rate": 1e-05,
      "loss": 1.4329,
      "step": 36300
    },
    {
      "epoch": 8.67,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4241238832473755,
      "eval_runtime": 34.0611,
      "eval_samples_per_second": 20.082,
      "step": 36300
    },
    {
      "epoch": 8.67,
      "learning_rate": 1e-05,
      "loss": 1.4285,
      "step": 36310
    },
    {
      "epoch": 8.67,
      "learning_rate": 1e-05,
      "loss": 1.4259,
      "step": 36320
    },
    {
      "epoch": 8.68,
      "learning_rate": 1e-05,
      "loss": 1.4423,
      "step": 36330
    },
    {
      "epoch": 8.68,
      "learning_rate": 1e-05,
      "loss": 1.4214,
      "step": 36340
    },
    {
      "epoch": 8.68,
      "learning_rate": 1e-05,
      "loss": 1.4243,
      "step": 36350
    },
    {
      "epoch": 8.68,
      "learning_rate": 1e-05,
      "loss": 1.4277,
      "step": 36360
    },
    {
      "epoch": 8.69,
      "learning_rate": 1e-05,
      "loss": 1.4315,
      "step": 36370
    },
    {
      "epoch": 8.69,
      "learning_rate": 1e-05,
      "loss": 1.4315,
      "step": 36380
    },
    {
      "epoch": 8.69,
      "learning_rate": 1e-05,
      "loss": 1.4284,
      "step": 36390
    },
    {
      "epoch": 8.69,
      "learning_rate": 1e-05,
      "loss": 1.4235,
      "step": 36400
    },
    {
      "epoch": 8.69,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4252840280532837,
      "eval_runtime": 34.0021,
      "eval_samples_per_second": 20.116,
      "step": 36400
    },
    {
      "epoch": 8.7,
      "learning_rate": 1e-05,
      "loss": 1.4247,
      "step": 36410
    },
    {
      "epoch": 8.7,
      "learning_rate": 1e-05,
      "loss": 1.4267,
      "step": 36420
    },
    {
      "epoch": 8.7,
      "learning_rate": 1e-05,
      "loss": 1.4336,
      "step": 36430
    },
    {
      "epoch": 8.7,
      "learning_rate": 1e-05,
      "loss": 1.4283,
      "step": 36440
    },
    {
      "epoch": 8.71,
      "learning_rate": 1e-05,
      "loss": 1.4283,
      "step": 36450
    },
    {
      "epoch": 8.71,
      "learning_rate": 1e-05,
      "loss": 1.4233,
      "step": 36460
    },
    {
      "epoch": 8.71,
      "learning_rate": 1e-05,
      "loss": 1.4362,
      "step": 36470
    },
    {
      "epoch": 8.71,
      "learning_rate": 1e-05,
      "loss": 1.4277,
      "step": 36480
    },
    {
      "epoch": 8.72,
      "learning_rate": 1e-05,
      "loss": 1.4422,
      "step": 36490
    },
    {
      "epoch": 8.72,
      "learning_rate": 1e-05,
      "loss": 1.431,
      "step": 36500
    },
    {
      "epoch": 8.72,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4243203401565552,
      "eval_runtime": 33.9553,
      "eval_samples_per_second": 20.144,
      "step": 36500
    },
    {
      "epoch": 8.72,
      "learning_rate": 1e-05,
      "loss": 1.4229,
      "step": 36510
    },
    {
      "epoch": 8.72,
      "learning_rate": 1e-05,
      "loss": 1.4313,
      "step": 36520
    },
    {
      "epoch": 8.72,
      "learning_rate": 1e-05,
      "loss": 1.4254,
      "step": 36530
    },
    {
      "epoch": 8.73,
      "learning_rate": 1e-05,
      "loss": 1.4307,
      "step": 36540
    },
    {
      "epoch": 8.73,
      "learning_rate": 1e-05,
      "loss": 1.4296,
      "step": 36550
    },
    {
      "epoch": 8.73,
      "learning_rate": 1e-05,
      "loss": 1.4292,
      "step": 36560
    },
    {
      "epoch": 8.73,
      "learning_rate": 1e-05,
      "loss": 1.4356,
      "step": 36570
    },
    {
      "epoch": 8.74,
      "learning_rate": 1e-05,
      "loss": 1.4267,
      "step": 36580
    },
    {
      "epoch": 8.74,
      "learning_rate": 1e-05,
      "loss": 1.4226,
      "step": 36590
    },
    {
      "epoch": 8.74,
      "learning_rate": 1e-05,
      "loss": 1.4273,
      "step": 36600
    },
    {
      "epoch": 8.74,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4243717193603516,
      "eval_runtime": 33.9138,
      "eval_samples_per_second": 20.169,
      "step": 36600
    },
    {
      "epoch": 8.74,
      "learning_rate": 1e-05,
      "loss": 1.426,
      "step": 36610
    },
    {
      "epoch": 8.75,
      "learning_rate": 1e-05,
      "loss": 1.4251,
      "step": 36620
    },
    {
      "epoch": 8.75,
      "learning_rate": 1e-05,
      "loss": 1.4247,
      "step": 36630
    },
    {
      "epoch": 8.75,
      "learning_rate": 1e-05,
      "loss": 1.4253,
      "step": 36640
    },
    {
      "epoch": 8.75,
      "learning_rate": 1e-05,
      "loss": 1.4263,
      "step": 36650
    },
    {
      "epoch": 8.76,
      "learning_rate": 1e-05,
      "loss": 1.4262,
      "step": 36660
    },
    {
      "epoch": 8.76,
      "learning_rate": 1e-05,
      "loss": 1.4278,
      "step": 36670
    },
    {
      "epoch": 8.76,
      "learning_rate": 1e-05,
      "loss": 1.427,
      "step": 36680
    },
    {
      "epoch": 8.76,
      "learning_rate": 1e-05,
      "loss": 1.4293,
      "step": 36690
    },
    {
      "epoch": 8.77,
      "learning_rate": 1e-05,
      "loss": 1.4277,
      "step": 36700
    },
    {
      "epoch": 8.77,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4239521026611328,
      "eval_runtime": 33.9448,
      "eval_samples_per_second": 20.15,
      "step": 36700
    },
    {
      "epoch": 8.77,
      "learning_rate": 1e-05,
      "loss": 1.4241,
      "step": 36710
    },
    {
      "epoch": 8.77,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 36720
    },
    {
      "epoch": 8.77,
      "learning_rate": 1e-05,
      "loss": 1.4283,
      "step": 36730
    },
    {
      "epoch": 8.77,
      "learning_rate": 1e-05,
      "loss": 1.4289,
      "step": 36740
    },
    {
      "epoch": 8.78,
      "learning_rate": 1e-05,
      "loss": 1.4282,
      "step": 36750
    },
    {
      "epoch": 8.78,
      "learning_rate": 1e-05,
      "loss": 1.4311,
      "step": 36760
    },
    {
      "epoch": 8.78,
      "learning_rate": 1e-05,
      "loss": 1.4232,
      "step": 36770
    },
    {
      "epoch": 8.78,
      "learning_rate": 1e-05,
      "loss": 1.4244,
      "step": 36780
    },
    {
      "epoch": 8.79,
      "learning_rate": 1e-05,
      "loss": 1.4235,
      "step": 36790
    },
    {
      "epoch": 8.79,
      "learning_rate": 1e-05,
      "loss": 1.4276,
      "step": 36800
    },
    {
      "epoch": 8.79,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4242215156555176,
      "eval_runtime": 33.9764,
      "eval_samples_per_second": 20.132,
      "step": 36800
    },
    {
      "epoch": 8.79,
      "learning_rate": 1e-05,
      "loss": 1.4328,
      "step": 36810
    },
    {
      "epoch": 8.79,
      "learning_rate": 1e-05,
      "loss": 1.4548,
      "step": 36820
    },
    {
      "epoch": 8.8,
      "learning_rate": 1e-05,
      "loss": 1.4253,
      "step": 36830
    },
    {
      "epoch": 8.8,
      "learning_rate": 1e-05,
      "loss": 1.4254,
      "step": 36840
    },
    {
      "epoch": 8.8,
      "learning_rate": 1e-05,
      "loss": 1.4274,
      "step": 36850
    },
    {
      "epoch": 8.8,
      "learning_rate": 1e-05,
      "loss": 1.4332,
      "step": 36860
    },
    {
      "epoch": 8.81,
      "learning_rate": 1e-05,
      "loss": 1.425,
      "step": 36870
    },
    {
      "epoch": 8.81,
      "learning_rate": 1e-05,
      "loss": 1.4356,
      "step": 36880
    },
    {
      "epoch": 8.81,
      "learning_rate": 1e-05,
      "loss": 1.4301,
      "step": 36890
    },
    {
      "epoch": 8.81,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 36900
    },
    {
      "epoch": 8.81,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4230380058288574,
      "eval_runtime": 34.1798,
      "eval_samples_per_second": 20.012,
      "step": 36900
    },
    {
      "epoch": 8.82,
      "learning_rate": 1e-05,
      "loss": 1.4242,
      "step": 36910
    },
    {
      "epoch": 8.82,
      "learning_rate": 1e-05,
      "loss": 1.4285,
      "step": 36920
    },
    {
      "epoch": 8.82,
      "learning_rate": 1e-05,
      "loss": 1.428,
      "step": 36930
    },
    {
      "epoch": 8.82,
      "learning_rate": 1e-05,
      "loss": 1.4242,
      "step": 36940
    },
    {
      "epoch": 8.82,
      "learning_rate": 1e-05,
      "loss": 1.4263,
      "step": 36950
    },
    {
      "epoch": 8.83,
      "learning_rate": 1e-05,
      "loss": 1.4299,
      "step": 36960
    },
    {
      "epoch": 8.83,
      "learning_rate": 1e-05,
      "loss": 1.4328,
      "step": 36970
    },
    {
      "epoch": 8.83,
      "learning_rate": 1e-05,
      "loss": 1.4318,
      "step": 36980
    },
    {
      "epoch": 8.83,
      "learning_rate": 1e-05,
      "loss": 1.4256,
      "step": 36990
    },
    {
      "epoch": 8.84,
      "learning_rate": 1e-05,
      "loss": 1.4245,
      "step": 37000
    },
    {
      "epoch": 8.84,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4252007007598877,
      "eval_runtime": 33.8016,
      "eval_samples_per_second": 20.236,
      "step": 37000
    },
    {
      "epoch": 8.84,
      "learning_rate": 1e-05,
      "loss": 1.4265,
      "step": 37010
    },
    {
      "epoch": 8.84,
      "learning_rate": 1e-05,
      "loss": 1.4279,
      "step": 37020
    },
    {
      "epoch": 8.84,
      "learning_rate": 1e-05,
      "loss": 1.4251,
      "step": 37030
    },
    {
      "epoch": 8.85,
      "learning_rate": 1e-05,
      "loss": 1.4246,
      "step": 37040
    },
    {
      "epoch": 8.85,
      "learning_rate": 1e-05,
      "loss": 1.4407,
      "step": 37050
    },
    {
      "epoch": 8.85,
      "learning_rate": 1e-05,
      "loss": 1.4258,
      "step": 37060
    },
    {
      "epoch": 8.85,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 37070
    },
    {
      "epoch": 8.86,
      "learning_rate": 1e-05,
      "loss": 1.4254,
      "step": 37080
    },
    {
      "epoch": 8.86,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 37090
    },
    {
      "epoch": 8.86,
      "learning_rate": 1e-05,
      "loss": 1.4242,
      "step": 37100
    },
    {
      "epoch": 8.86,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4235261678695679,
      "eval_runtime": 33.8472,
      "eval_samples_per_second": 20.208,
      "step": 37100
    },
    {
      "epoch": 8.86,
      "learning_rate": 1e-05,
      "loss": 1.4277,
      "step": 37110
    },
    {
      "epoch": 8.87,
      "learning_rate": 1e-05,
      "loss": 1.4267,
      "step": 37120
    },
    {
      "epoch": 8.87,
      "learning_rate": 1e-05,
      "loss": 1.4302,
      "step": 37130
    },
    {
      "epoch": 8.87,
      "learning_rate": 1e-05,
      "loss": 1.4273,
      "step": 37140
    },
    {
      "epoch": 8.87,
      "learning_rate": 1e-05,
      "loss": 1.4319,
      "step": 37150
    },
    {
      "epoch": 8.88,
      "learning_rate": 1e-05,
      "loss": 1.4324,
      "step": 37160
    },
    {
      "epoch": 8.88,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 37170
    },
    {
      "epoch": 8.88,
      "learning_rate": 1e-05,
      "loss": 1.4287,
      "step": 37180
    },
    {
      "epoch": 8.88,
      "learning_rate": 1e-05,
      "loss": 1.4231,
      "step": 37190
    },
    {
      "epoch": 8.88,
      "learning_rate": 1e-05,
      "loss": 1.4293,
      "step": 37200
    },
    {
      "epoch": 8.88,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.423323631286621,
      "eval_runtime": 33.9341,
      "eval_samples_per_second": 20.157,
      "step": 37200
    },
    {
      "epoch": 8.89,
      "learning_rate": 1e-05,
      "loss": 1.4305,
      "step": 37210
    },
    {
      "epoch": 8.89,
      "learning_rate": 1e-05,
      "loss": 1.4288,
      "step": 37220
    },
    {
      "epoch": 8.89,
      "learning_rate": 1e-05,
      "loss": 1.4279,
      "step": 37230
    },
    {
      "epoch": 8.89,
      "learning_rate": 1e-05,
      "loss": 1.4213,
      "step": 37240
    },
    {
      "epoch": 8.9,
      "learning_rate": 1e-05,
      "loss": 1.4251,
      "step": 37250
    },
    {
      "epoch": 8.9,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 37260
    },
    {
      "epoch": 8.9,
      "learning_rate": 1e-05,
      "loss": 1.4232,
      "step": 37270
    },
    {
      "epoch": 8.9,
      "learning_rate": 1e-05,
      "loss": 1.4255,
      "step": 37280
    },
    {
      "epoch": 8.91,
      "learning_rate": 1e-05,
      "loss": 1.4603,
      "step": 37290
    },
    {
      "epoch": 8.91,
      "learning_rate": 1e-05,
      "loss": 1.4282,
      "step": 37300
    },
    {
      "epoch": 8.91,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4238145351409912,
      "eval_runtime": 34.201,
      "eval_samples_per_second": 19.999,
      "step": 37300
    },
    {
      "epoch": 8.91,
      "learning_rate": 1e-05,
      "loss": 1.429,
      "step": 37310
    },
    {
      "epoch": 8.91,
      "learning_rate": 1e-05,
      "loss": 1.4321,
      "step": 37320
    },
    {
      "epoch": 8.92,
      "learning_rate": 1e-05,
      "loss": 1.4283,
      "step": 37330
    },
    {
      "epoch": 8.92,
      "learning_rate": 1e-05,
      "loss": 1.4287,
      "step": 37340
    },
    {
      "epoch": 8.92,
      "learning_rate": 1e-05,
      "loss": 1.4177,
      "step": 37350
    },
    {
      "epoch": 8.92,
      "learning_rate": 1e-05,
      "loss": 1.4261,
      "step": 37360
    },
    {
      "epoch": 8.93,
      "learning_rate": 1e-05,
      "loss": 1.4258,
      "step": 37370
    },
    {
      "epoch": 8.93,
      "learning_rate": 1e-05,
      "loss": 1.4257,
      "step": 37380
    },
    {
      "epoch": 8.93,
      "learning_rate": 1e-05,
      "loss": 1.4234,
      "step": 37390
    },
    {
      "epoch": 8.93,
      "learning_rate": 1e-05,
      "loss": 1.4252,
      "step": 37400
    },
    {
      "epoch": 8.93,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4231505393981934,
      "eval_runtime": 33.9604,
      "eval_samples_per_second": 20.141,
      "step": 37400
    },
    {
      "epoch": 8.93,
      "learning_rate": 1e-05,
      "loss": 1.428,
      "step": 37410
    },
    {
      "epoch": 8.94,
      "learning_rate": 1e-05,
      "loss": 1.4218,
      "step": 37420
    },
    {
      "epoch": 8.94,
      "learning_rate": 1e-05,
      "loss": 1.4233,
      "step": 37430
    },
    {
      "epoch": 8.94,
      "learning_rate": 1e-05,
      "loss": 1.429,
      "step": 37440
    },
    {
      "epoch": 8.94,
      "learning_rate": 1e-05,
      "loss": 1.4329,
      "step": 37450
    },
    {
      "epoch": 8.95,
      "learning_rate": 1e-05,
      "loss": 1.4317,
      "step": 37460
    },
    {
      "epoch": 8.95,
      "learning_rate": 1e-05,
      "loss": 1.4212,
      "step": 37470
    },
    {
      "epoch": 8.95,
      "learning_rate": 1e-05,
      "loss": 1.4187,
      "step": 37480
    },
    {
      "epoch": 8.95,
      "learning_rate": 1e-05,
      "loss": 1.4243,
      "step": 37490
    },
    {
      "epoch": 8.96,
      "learning_rate": 1e-05,
      "loss": 1.4248,
      "step": 37500
    },
    {
      "epoch": 8.96,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.424094796180725,
      "eval_runtime": 33.8441,
      "eval_samples_per_second": 20.21,
      "step": 37500
    },
    {
      "epoch": 8.96,
      "learning_rate": 1e-05,
      "loss": 1.4333,
      "step": 37510
    },
    {
      "epoch": 8.96,
      "learning_rate": 1e-05,
      "loss": 1.4297,
      "step": 37520
    },
    {
      "epoch": 8.96,
      "learning_rate": 1e-05,
      "loss": 1.4334,
      "step": 37530
    },
    {
      "epoch": 8.97,
      "learning_rate": 1e-05,
      "loss": 1.4257,
      "step": 37540
    },
    {
      "epoch": 8.97,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 37550
    },
    {
      "epoch": 8.97,
      "learning_rate": 1e-05,
      "loss": 1.426,
      "step": 37560
    },
    {
      "epoch": 8.97,
      "learning_rate": 1e-05,
      "loss": 1.4235,
      "step": 37570
    },
    {
      "epoch": 8.98,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 37580
    },
    {
      "epoch": 8.98,
      "learning_rate": 1e-05,
      "loss": 1.4261,
      "step": 37590
    },
    {
      "epoch": 8.98,
      "learning_rate": 1e-05,
      "loss": 1.4321,
      "step": 37600
    },
    {
      "epoch": 8.98,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4246643781661987,
      "eval_runtime": 34.0486,
      "eval_samples_per_second": 20.089,
      "step": 37600
    },
    {
      "epoch": 8.98,
      "learning_rate": 1e-05,
      "loss": 1.4266,
      "step": 37610
    },
    {
      "epoch": 8.98,
      "learning_rate": 1e-05,
      "loss": 1.4366,
      "step": 37620
    },
    {
      "epoch": 8.99,
      "learning_rate": 1e-05,
      "loss": 1.4224,
      "step": 37630
    },
    {
      "epoch": 8.99,
      "learning_rate": 1e-05,
      "loss": 1.431,
      "step": 37640
    },
    {
      "epoch": 8.99,
      "learning_rate": 1e-05,
      "loss": 1.441,
      "step": 37650
    },
    {
      "epoch": 8.99,
      "learning_rate": 1e-05,
      "loss": 1.4293,
      "step": 37660
    },
    {
      "epoch": 9.0,
      "learning_rate": 1e-05,
      "loss": 1.4237,
      "step": 37670
    },
    {
      "epoch": 9.0,
      "learning_rate": 1e-05,
      "loss": 1.4328,
      "step": 37680
    },
    {
      "epoch": 9.0,
      "learning_rate": 1e-05,
      "loss": 1.4198,
      "step": 37690
    },
    {
      "epoch": 9.0,
      "learning_rate": 1e-05,
      "loss": 1.4296,
      "step": 37700
    },
    {
      "epoch": 9.0,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4234720468521118,
      "eval_runtime": 33.9503,
      "eval_samples_per_second": 20.147,
      "step": 37700
    },
    {
      "epoch": 9.01,
      "learning_rate": 1e-05,
      "loss": 1.4399,
      "step": 37710
    },
    {
      "epoch": 9.01,
      "learning_rate": 1e-05,
      "loss": 1.4304,
      "step": 37720
    },
    {
      "epoch": 9.01,
      "learning_rate": 1e-05,
      "loss": 1.4234,
      "step": 37730
    },
    {
      "epoch": 9.01,
      "learning_rate": 1e-05,
      "loss": 1.4298,
      "step": 37740
    },
    {
      "epoch": 9.02,
      "learning_rate": 1e-05,
      "loss": 1.4504,
      "step": 37750
    },
    {
      "epoch": 9.02,
      "learning_rate": 1e-05,
      "loss": 1.4237,
      "step": 37760
    },
    {
      "epoch": 9.02,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 37770
    },
    {
      "epoch": 9.02,
      "learning_rate": 1e-05,
      "loss": 1.4257,
      "step": 37780
    },
    {
      "epoch": 9.03,
      "learning_rate": 1e-05,
      "loss": 1.4198,
      "step": 37790
    },
    {
      "epoch": 9.03,
      "learning_rate": 1e-05,
      "loss": 1.432,
      "step": 37800
    },
    {
      "epoch": 9.03,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4243077039718628,
      "eval_runtime": 33.7384,
      "eval_samples_per_second": 20.274,
      "step": 37800
    },
    {
      "epoch": 9.03,
      "learning_rate": 1e-05,
      "loss": 1.4357,
      "step": 37810
    },
    {
      "epoch": 9.03,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 37820
    },
    {
      "epoch": 9.04,
      "learning_rate": 1e-05,
      "loss": 1.427,
      "step": 37830
    },
    {
      "epoch": 9.04,
      "learning_rate": 1e-05,
      "loss": 1.4258,
      "step": 37840
    },
    {
      "epoch": 9.04,
      "learning_rate": 1e-05,
      "loss": 1.4326,
      "step": 37850
    },
    {
      "epoch": 9.04,
      "learning_rate": 1e-05,
      "loss": 1.4287,
      "step": 37860
    },
    {
      "epoch": 9.04,
      "learning_rate": 1e-05,
      "loss": 1.4234,
      "step": 37870
    },
    {
      "epoch": 9.05,
      "learning_rate": 1e-05,
      "loss": 1.4232,
      "step": 37880
    },
    {
      "epoch": 9.05,
      "learning_rate": 1e-05,
      "loss": 1.4339,
      "step": 37890
    },
    {
      "epoch": 9.05,
      "learning_rate": 1e-05,
      "loss": 1.422,
      "step": 37900
    },
    {
      "epoch": 9.05,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4230388402938843,
      "eval_runtime": 33.4141,
      "eval_samples_per_second": 20.47,
      "step": 37900
    },
    {
      "epoch": 9.05,
      "learning_rate": 1e-05,
      "loss": 1.4234,
      "step": 37910
    },
    {
      "epoch": 9.06,
      "learning_rate": 1e-05,
      "loss": 1.4224,
      "step": 37920
    },
    {
      "epoch": 9.06,
      "learning_rate": 1e-05,
      "loss": 1.4239,
      "step": 37930
    },
    {
      "epoch": 9.06,
      "learning_rate": 1e-05,
      "loss": 1.4312,
      "step": 37940
    },
    {
      "epoch": 9.06,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 37950
    },
    {
      "epoch": 9.07,
      "learning_rate": 1e-05,
      "loss": 1.4331,
      "step": 37960
    },
    {
      "epoch": 9.07,
      "learning_rate": 1e-05,
      "loss": 1.4325,
      "step": 37970
    },
    {
      "epoch": 9.07,
      "learning_rate": 1e-05,
      "loss": 1.4299,
      "step": 37980
    },
    {
      "epoch": 9.07,
      "learning_rate": 1e-05,
      "loss": 1.4324,
      "step": 37990
    },
    {
      "epoch": 9.08,
      "learning_rate": 1e-05,
      "loss": 1.4297,
      "step": 38000
    },
    {
      "epoch": 9.08,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4231606721878052,
      "eval_runtime": 33.539,
      "eval_samples_per_second": 20.394,
      "step": 38000
    },
    {
      "epoch": 9.08,
      "learning_rate": 1e-05,
      "loss": 1.4221,
      "step": 38010
    },
    {
      "epoch": 9.08,
      "learning_rate": 1e-05,
      "loss": 1.4253,
      "step": 38020
    },
    {
      "epoch": 9.08,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 38030
    },
    {
      "epoch": 9.09,
      "learning_rate": 1e-05,
      "loss": 1.4338,
      "step": 38040
    },
    {
      "epoch": 9.09,
      "learning_rate": 1e-05,
      "loss": 1.4423,
      "step": 38050
    },
    {
      "epoch": 9.09,
      "learning_rate": 1e-05,
      "loss": 1.4253,
      "step": 38060
    },
    {
      "epoch": 9.09,
      "learning_rate": 1e-05,
      "loss": 1.4288,
      "step": 38070
    },
    {
      "epoch": 9.09,
      "learning_rate": 1e-05,
      "loss": 1.4276,
      "step": 38080
    },
    {
      "epoch": 9.1,
      "learning_rate": 1e-05,
      "loss": 1.4252,
      "step": 38090
    },
    {
      "epoch": 9.1,
      "learning_rate": 1e-05,
      "loss": 1.4272,
      "step": 38100
    },
    {
      "epoch": 9.1,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4239295721054077,
      "eval_runtime": 34.6154,
      "eval_samples_per_second": 19.76,
      "step": 38100
    },
    {
      "epoch": 9.1,
      "learning_rate": 1e-05,
      "loss": 1.4224,
      "step": 38110
    },
    {
      "epoch": 9.1,
      "learning_rate": 1e-05,
      "loss": 1.4247,
      "step": 38120
    },
    {
      "epoch": 9.11,
      "learning_rate": 1e-05,
      "loss": 1.4203,
      "step": 38130
    },
    {
      "epoch": 9.11,
      "learning_rate": 1e-05,
      "loss": 1.4302,
      "step": 38140
    },
    {
      "epoch": 9.11,
      "learning_rate": 1e-05,
      "loss": 1.4239,
      "step": 38150
    },
    {
      "epoch": 9.11,
      "learning_rate": 1e-05,
      "loss": 1.4254,
      "step": 38160
    },
    {
      "epoch": 9.12,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 38170
    },
    {
      "epoch": 9.12,
      "learning_rate": 1e-05,
      "loss": 1.4361,
      "step": 38180
    },
    {
      "epoch": 9.12,
      "learning_rate": 1e-05,
      "loss": 1.4273,
      "step": 38190
    },
    {
      "epoch": 9.12,
      "learning_rate": 1e-05,
      "loss": 1.4259,
      "step": 38200
    },
    {
      "epoch": 9.12,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4236606359481812,
      "eval_runtime": 33.6273,
      "eval_samples_per_second": 20.341,
      "step": 38200
    },
    {
      "epoch": 9.13,
      "learning_rate": 1e-05,
      "loss": 1.4307,
      "step": 38210
    },
    {
      "epoch": 9.13,
      "learning_rate": 1e-05,
      "loss": 1.4263,
      "step": 38220
    },
    {
      "epoch": 9.13,
      "learning_rate": 1e-05,
      "loss": 1.4366,
      "step": 38230
    },
    {
      "epoch": 9.13,
      "learning_rate": 1e-05,
      "loss": 1.4253,
      "step": 38240
    },
    {
      "epoch": 9.14,
      "learning_rate": 1e-05,
      "loss": 1.4276,
      "step": 38250
    },
    {
      "epoch": 9.14,
      "learning_rate": 1e-05,
      "loss": 1.4223,
      "step": 38260
    },
    {
      "epoch": 9.14,
      "learning_rate": 1e-05,
      "loss": 1.4245,
      "step": 38270
    },
    {
      "epoch": 9.14,
      "learning_rate": 1e-05,
      "loss": 1.4277,
      "step": 38280
    },
    {
      "epoch": 9.14,
      "learning_rate": 1e-05,
      "loss": 1.428,
      "step": 38290
    },
    {
      "epoch": 9.15,
      "learning_rate": 1e-05,
      "loss": 1.4207,
      "step": 38300
    },
    {
      "epoch": 9.15,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4224814176559448,
      "eval_runtime": 33.4723,
      "eval_samples_per_second": 20.435,
      "step": 38300
    },
    {
      "epoch": 9.15,
      "learning_rate": 1e-05,
      "loss": 1.4295,
      "step": 38310
    },
    {
      "epoch": 9.15,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 38320
    },
    {
      "epoch": 9.15,
      "learning_rate": 1e-05,
      "loss": 1.4287,
      "step": 38330
    },
    {
      "epoch": 9.16,
      "learning_rate": 1e-05,
      "loss": 1.4291,
      "step": 38340
    },
    {
      "epoch": 9.16,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 38350
    },
    {
      "epoch": 9.16,
      "learning_rate": 1e-05,
      "loss": 1.4241,
      "step": 38360
    },
    {
      "epoch": 9.16,
      "learning_rate": 1e-05,
      "loss": 1.4263,
      "step": 38370
    },
    {
      "epoch": 9.17,
      "learning_rate": 1e-05,
      "loss": 1.425,
      "step": 38380
    },
    {
      "epoch": 9.17,
      "learning_rate": 1e-05,
      "loss": 1.4263,
      "step": 38390
    },
    {
      "epoch": 9.17,
      "learning_rate": 1e-05,
      "loss": 1.4254,
      "step": 38400
    },
    {
      "epoch": 9.17,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4228925704956055,
      "eval_runtime": 33.408,
      "eval_samples_per_second": 20.474,
      "step": 38400
    },
    {
      "epoch": 9.17,
      "learning_rate": 1e-05,
      "loss": 1.4216,
      "step": 38410
    },
    {
      "epoch": 9.18,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 38420
    },
    {
      "epoch": 9.18,
      "learning_rate": 1e-05,
      "loss": 1.429,
      "step": 38430
    },
    {
      "epoch": 9.18,
      "learning_rate": 1e-05,
      "loss": 1.4281,
      "step": 38440
    },
    {
      "epoch": 9.18,
      "learning_rate": 1e-05,
      "loss": 1.4239,
      "step": 38450
    },
    {
      "epoch": 9.19,
      "learning_rate": 1e-05,
      "loss": 1.4226,
      "step": 38460
    },
    {
      "epoch": 9.19,
      "learning_rate": 1e-05,
      "loss": 1.4268,
      "step": 38470
    },
    {
      "epoch": 9.19,
      "learning_rate": 1e-05,
      "loss": 1.4339,
      "step": 38480
    },
    {
      "epoch": 9.19,
      "learning_rate": 1e-05,
      "loss": 1.4223,
      "step": 38490
    },
    {
      "epoch": 9.2,
      "learning_rate": 1e-05,
      "loss": 1.4224,
      "step": 38500
    },
    {
      "epoch": 9.2,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4230594635009766,
      "eval_runtime": 33.3622,
      "eval_samples_per_second": 20.502,
      "step": 38500
    },
    {
      "epoch": 9.2,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 38510
    },
    {
      "epoch": 9.2,
      "learning_rate": 1e-05,
      "loss": 1.4254,
      "step": 38520
    },
    {
      "epoch": 9.2,
      "learning_rate": 1e-05,
      "loss": 1.4218,
      "step": 38530
    },
    {
      "epoch": 9.2,
      "learning_rate": 1e-05,
      "loss": 1.4563,
      "step": 38540
    },
    {
      "epoch": 9.21,
      "learning_rate": 1e-05,
      "loss": 1.4243,
      "step": 38550
    },
    {
      "epoch": 9.21,
      "learning_rate": 1e-05,
      "loss": 1.4241,
      "step": 38560
    },
    {
      "epoch": 9.21,
      "learning_rate": 1e-05,
      "loss": 1.4298,
      "step": 38570
    },
    {
      "epoch": 9.21,
      "learning_rate": 1e-05,
      "loss": 1.4251,
      "step": 38580
    },
    {
      "epoch": 9.22,
      "learning_rate": 1e-05,
      "loss": 1.4332,
      "step": 38590
    },
    {
      "epoch": 9.22,
      "learning_rate": 1e-05,
      "loss": 1.4239,
      "step": 38600
    },
    {
      "epoch": 9.22,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4222257137298584,
      "eval_runtime": 33.3636,
      "eval_samples_per_second": 20.501,
      "step": 38600
    },
    {
      "epoch": 9.22,
      "learning_rate": 1e-05,
      "loss": 1.4265,
      "step": 38610
    },
    {
      "epoch": 9.22,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 38620
    },
    {
      "epoch": 9.23,
      "learning_rate": 1e-05,
      "loss": 1.4252,
      "step": 38630
    },
    {
      "epoch": 9.23,
      "learning_rate": 1e-05,
      "loss": 1.4212,
      "step": 38640
    },
    {
      "epoch": 9.23,
      "learning_rate": 1e-05,
      "loss": 1.426,
      "step": 38650
    },
    {
      "epoch": 9.23,
      "learning_rate": 1e-05,
      "loss": 1.4342,
      "step": 38660
    },
    {
      "epoch": 9.24,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 38670
    },
    {
      "epoch": 9.24,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 38680
    },
    {
      "epoch": 9.24,
      "learning_rate": 1e-05,
      "loss": 1.43,
      "step": 38690
    },
    {
      "epoch": 9.24,
      "learning_rate": 1e-05,
      "loss": 1.4269,
      "step": 38700
    },
    {
      "epoch": 9.24,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4226405620574951,
      "eval_runtime": 33.6236,
      "eval_samples_per_second": 20.343,
      "step": 38700
    },
    {
      "epoch": 9.25,
      "learning_rate": 1e-05,
      "loss": 1.4263,
      "step": 38710
    },
    {
      "epoch": 9.25,
      "learning_rate": 1e-05,
      "loss": 1.4193,
      "step": 38720
    },
    {
      "epoch": 9.25,
      "learning_rate": 1e-05,
      "loss": 1.4257,
      "step": 38730
    },
    {
      "epoch": 9.25,
      "learning_rate": 1e-05,
      "loss": 1.4289,
      "step": 38740
    },
    {
      "epoch": 9.25,
      "learning_rate": 1e-05,
      "loss": 1.4242,
      "step": 38750
    },
    {
      "epoch": 9.26,
      "learning_rate": 1e-05,
      "loss": 1.4253,
      "step": 38760
    },
    {
      "epoch": 9.26,
      "learning_rate": 1e-05,
      "loss": 1.4274,
      "step": 38770
    },
    {
      "epoch": 9.26,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 38780
    },
    {
      "epoch": 9.26,
      "learning_rate": 1e-05,
      "loss": 1.4227,
      "step": 38790
    },
    {
      "epoch": 9.27,
      "learning_rate": 1e-05,
      "loss": 1.4281,
      "step": 38800
    },
    {
      "epoch": 9.27,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4225388765335083,
      "eval_runtime": 33.4064,
      "eval_samples_per_second": 20.475,
      "step": 38800
    },
    {
      "epoch": 9.27,
      "learning_rate": 1e-05,
      "loss": 1.4254,
      "step": 38810
    },
    {
      "epoch": 9.27,
      "learning_rate": 1e-05,
      "loss": 1.4298,
      "step": 38820
    },
    {
      "epoch": 9.27,
      "learning_rate": 1e-05,
      "loss": 1.4267,
      "step": 38830
    },
    {
      "epoch": 9.28,
      "learning_rate": 1e-05,
      "loss": 1.4288,
      "step": 38840
    },
    {
      "epoch": 9.28,
      "learning_rate": 1e-05,
      "loss": 1.4231,
      "step": 38850
    },
    {
      "epoch": 9.28,
      "learning_rate": 1e-05,
      "loss": 1.4265,
      "step": 38860
    },
    {
      "epoch": 9.28,
      "learning_rate": 1e-05,
      "loss": 1.4267,
      "step": 38870
    },
    {
      "epoch": 9.29,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 38880
    },
    {
      "epoch": 9.29,
      "learning_rate": 1e-05,
      "loss": 1.4308,
      "step": 38890
    },
    {
      "epoch": 9.29,
      "learning_rate": 1e-05,
      "loss": 1.422,
      "step": 38900
    },
    {
      "epoch": 9.29,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4223017692565918,
      "eval_runtime": 33.4433,
      "eval_samples_per_second": 20.453,
      "step": 38900
    },
    {
      "epoch": 9.29,
      "learning_rate": 1e-05,
      "loss": 1.4276,
      "step": 38910
    },
    {
      "epoch": 9.3,
      "learning_rate": 1e-05,
      "loss": 1.4315,
      "step": 38920
    },
    {
      "epoch": 9.3,
      "learning_rate": 1e-05,
      "loss": 1.4278,
      "step": 38930
    },
    {
      "epoch": 9.3,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 38940
    },
    {
      "epoch": 9.3,
      "learning_rate": 1e-05,
      "loss": 1.4195,
      "step": 38950
    },
    {
      "epoch": 9.3,
      "learning_rate": 1e-05,
      "loss": 1.4245,
      "step": 38960
    },
    {
      "epoch": 9.31,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 38970
    },
    {
      "epoch": 9.31,
      "learning_rate": 1e-05,
      "loss": 1.4186,
      "step": 38980
    },
    {
      "epoch": 9.31,
      "learning_rate": 1e-05,
      "loss": 1.425,
      "step": 38990
    },
    {
      "epoch": 9.31,
      "learning_rate": 1e-05,
      "loss": 1.4209,
      "step": 39000
    },
    {
      "epoch": 9.31,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.421040654182434,
      "eval_runtime": 33.7211,
      "eval_samples_per_second": 20.284,
      "step": 39000
    },
    {
      "epoch": 9.32,
      "learning_rate": 1e-05,
      "loss": 1.4293,
      "step": 39010
    },
    {
      "epoch": 9.32,
      "learning_rate": 1e-05,
      "loss": 1.4286,
      "step": 39020
    },
    {
      "epoch": 9.32,
      "learning_rate": 1e-05,
      "loss": 1.4295,
      "step": 39030
    },
    {
      "epoch": 9.32,
      "learning_rate": 1e-05,
      "loss": 1.423,
      "step": 39040
    },
    {
      "epoch": 9.33,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 39050
    },
    {
      "epoch": 9.33,
      "learning_rate": 1e-05,
      "loss": 1.4319,
      "step": 39060
    },
    {
      "epoch": 9.33,
      "learning_rate": 1e-05,
      "loss": 1.4277,
      "step": 39070
    },
    {
      "epoch": 9.33,
      "learning_rate": 1e-05,
      "loss": 1.4237,
      "step": 39080
    },
    {
      "epoch": 9.34,
      "learning_rate": 1e-05,
      "loss": 1.4226,
      "step": 39090
    },
    {
      "epoch": 9.34,
      "learning_rate": 1e-05,
      "loss": 1.4267,
      "step": 39100
    },
    {
      "epoch": 9.34,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4223829507827759,
      "eval_runtime": 33.635,
      "eval_samples_per_second": 20.336,
      "step": 39100
    },
    {
      "epoch": 9.34,
      "learning_rate": 1e-05,
      "loss": 1.4307,
      "step": 39110
    },
    {
      "epoch": 9.34,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 39120
    },
    {
      "epoch": 9.35,
      "learning_rate": 1e-05,
      "loss": 1.4241,
      "step": 39130
    },
    {
      "epoch": 9.35,
      "learning_rate": 1e-05,
      "loss": 1.4212,
      "step": 39140
    },
    {
      "epoch": 9.35,
      "learning_rate": 1e-05,
      "loss": 1.4302,
      "step": 39150
    },
    {
      "epoch": 9.35,
      "learning_rate": 1e-05,
      "loss": 1.4325,
      "step": 39160
    },
    {
      "epoch": 9.36,
      "learning_rate": 1e-05,
      "loss": 1.4241,
      "step": 39170
    },
    {
      "epoch": 9.36,
      "learning_rate": 1e-05,
      "loss": 1.4261,
      "step": 39180
    },
    {
      "epoch": 9.36,
      "learning_rate": 1e-05,
      "loss": 1.4256,
      "step": 39190
    },
    {
      "epoch": 9.36,
      "learning_rate": 1e-05,
      "loss": 1.4216,
      "step": 39200
    },
    {
      "epoch": 9.36,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4229578971862793,
      "eval_runtime": 33.5175,
      "eval_samples_per_second": 20.407,
      "step": 39200
    },
    {
      "epoch": 9.36,
      "learning_rate": 1e-05,
      "loss": 1.4225,
      "step": 39210
    },
    {
      "epoch": 9.37,
      "learning_rate": 1e-05,
      "loss": 1.4251,
      "step": 39220
    },
    {
      "epoch": 9.37,
      "learning_rate": 1e-05,
      "loss": 1.4255,
      "step": 39230
    },
    {
      "epoch": 9.37,
      "learning_rate": 1e-05,
      "loss": 1.4371,
      "step": 39240
    },
    {
      "epoch": 9.37,
      "learning_rate": 1e-05,
      "loss": 1.4284,
      "step": 39250
    },
    {
      "epoch": 9.38,
      "learning_rate": 1e-05,
      "loss": 1.4231,
      "step": 39260
    },
    {
      "epoch": 9.38,
      "learning_rate": 1e-05,
      "loss": 1.4296,
      "step": 39270
    },
    {
      "epoch": 9.38,
      "learning_rate": 1e-05,
      "loss": 1.422,
      "step": 39280
    },
    {
      "epoch": 9.38,
      "learning_rate": 1e-05,
      "loss": 1.4245,
      "step": 39290
    },
    {
      "epoch": 9.39,
      "learning_rate": 1e-05,
      "loss": 1.422,
      "step": 39300
    },
    {
      "epoch": 9.39,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4222767353057861,
      "eval_runtime": 33.7468,
      "eval_samples_per_second": 20.269,
      "step": 39300
    },
    {
      "epoch": 9.39,
      "learning_rate": 1e-05,
      "loss": 1.4245,
      "step": 39310
    },
    {
      "epoch": 9.39,
      "learning_rate": 1e-05,
      "loss": 1.4283,
      "step": 39320
    },
    {
      "epoch": 9.39,
      "learning_rate": 1e-05,
      "loss": 1.4261,
      "step": 39330
    },
    {
      "epoch": 9.4,
      "learning_rate": 1e-05,
      "loss": 1.4265,
      "step": 39340
    },
    {
      "epoch": 9.4,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 39350
    },
    {
      "epoch": 9.4,
      "learning_rate": 1e-05,
      "loss": 1.4236,
      "step": 39360
    },
    {
      "epoch": 9.4,
      "learning_rate": 1e-05,
      "loss": 1.4352,
      "step": 39370
    },
    {
      "epoch": 9.41,
      "learning_rate": 1e-05,
      "loss": 1.4233,
      "step": 39380
    },
    {
      "epoch": 9.41,
      "learning_rate": 1e-05,
      "loss": 1.4303,
      "step": 39390
    },
    {
      "epoch": 9.41,
      "learning_rate": 1e-05,
      "loss": 1.4285,
      "step": 39400
    },
    {
      "epoch": 9.41,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4223226308822632,
      "eval_runtime": 33.7819,
      "eval_samples_per_second": 20.248,
      "step": 39400
    },
    {
      "epoch": 9.41,
      "learning_rate": 1e-05,
      "loss": 1.4256,
      "step": 39410
    },
    {
      "epoch": 9.41,
      "learning_rate": 1e-05,
      "loss": 1.4279,
      "step": 39420
    },
    {
      "epoch": 9.42,
      "learning_rate": 1e-05,
      "loss": 1.4313,
      "step": 39430
    },
    {
      "epoch": 9.42,
      "learning_rate": 1e-05,
      "loss": 1.4213,
      "step": 39440
    },
    {
      "epoch": 9.42,
      "learning_rate": 1e-05,
      "loss": 1.428,
      "step": 39450
    },
    {
      "epoch": 9.42,
      "learning_rate": 1e-05,
      "loss": 1.4258,
      "step": 39460
    },
    {
      "epoch": 9.43,
      "learning_rate": 1e-05,
      "loss": 1.4226,
      "step": 39470
    },
    {
      "epoch": 9.43,
      "learning_rate": 1e-05,
      "loss": 1.4426,
      "step": 39480
    },
    {
      "epoch": 9.43,
      "learning_rate": 1e-05,
      "loss": 1.4231,
      "step": 39490
    },
    {
      "epoch": 9.43,
      "learning_rate": 1e-05,
      "loss": 1.4286,
      "step": 39500
    },
    {
      "epoch": 9.43,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4221357107162476,
      "eval_runtime": 33.5725,
      "eval_samples_per_second": 20.374,
      "step": 39500
    },
    {
      "epoch": 9.44,
      "learning_rate": 1e-05,
      "loss": 1.4229,
      "step": 39510
    },
    {
      "epoch": 9.44,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 39520
    },
    {
      "epoch": 9.44,
      "learning_rate": 1e-05,
      "loss": 1.4232,
      "step": 39530
    },
    {
      "epoch": 9.44,
      "learning_rate": 1e-05,
      "loss": 1.4231,
      "step": 39540
    },
    {
      "epoch": 9.45,
      "learning_rate": 1e-05,
      "loss": 1.4266,
      "step": 39550
    },
    {
      "epoch": 9.45,
      "learning_rate": 1e-05,
      "loss": 1.4306,
      "step": 39560
    },
    {
      "epoch": 9.45,
      "learning_rate": 1e-05,
      "loss": 1.4299,
      "step": 39570
    },
    {
      "epoch": 9.45,
      "learning_rate": 1e-05,
      "loss": 1.4267,
      "step": 39580
    },
    {
      "epoch": 9.46,
      "learning_rate": 1e-05,
      "loss": 1.4262,
      "step": 39590
    },
    {
      "epoch": 9.46,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 39600
    },
    {
      "epoch": 9.46,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4216047525405884,
      "eval_runtime": 34.2545,
      "eval_samples_per_second": 19.968,
      "step": 39600
    },
    {
      "epoch": 9.46,
      "learning_rate": 1e-05,
      "loss": 1.4269,
      "step": 39610
    },
    {
      "epoch": 9.46,
      "learning_rate": 1e-05,
      "loss": 1.428,
      "step": 39620
    },
    {
      "epoch": 9.47,
      "learning_rate": 1e-05,
      "loss": 1.4262,
      "step": 39630
    },
    {
      "epoch": 9.47,
      "learning_rate": 1e-05,
      "loss": 1.4289,
      "step": 39640
    },
    {
      "epoch": 9.47,
      "learning_rate": 1e-05,
      "loss": 1.4237,
      "step": 39650
    },
    {
      "epoch": 9.47,
      "learning_rate": 1e-05,
      "loss": 1.4237,
      "step": 39660
    },
    {
      "epoch": 9.47,
      "learning_rate": 1e-05,
      "loss": 1.4227,
      "step": 39670
    },
    {
      "epoch": 9.48,
      "learning_rate": 1e-05,
      "loss": 1.4276,
      "step": 39680
    },
    {
      "epoch": 9.48,
      "learning_rate": 1e-05,
      "loss": 1.4256,
      "step": 39690
    },
    {
      "epoch": 9.48,
      "learning_rate": 1e-05,
      "loss": 1.4296,
      "step": 39700
    },
    {
      "epoch": 9.48,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4226685762405396,
      "eval_runtime": 33.2001,
      "eval_samples_per_second": 20.602,
      "step": 39700
    },
    {
      "epoch": 9.48,
      "learning_rate": 1e-05,
      "loss": 1.4232,
      "step": 39710
    },
    {
      "epoch": 9.49,
      "learning_rate": 1e-05,
      "loss": 1.4256,
      "step": 39720
    },
    {
      "epoch": 9.49,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 39730
    },
    {
      "epoch": 9.49,
      "learning_rate": 1e-05,
      "loss": 1.4266,
      "step": 39740
    },
    {
      "epoch": 9.49,
      "learning_rate": 1e-05,
      "loss": 1.423,
      "step": 39750
    },
    {
      "epoch": 9.5,
      "learning_rate": 1e-05,
      "loss": 1.4232,
      "step": 39760
    },
    {
      "epoch": 9.5,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 39770
    },
    {
      "epoch": 9.5,
      "learning_rate": 1e-05,
      "loss": 1.4239,
      "step": 39780
    },
    {
      "epoch": 9.5,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 39790
    },
    {
      "epoch": 9.51,
      "learning_rate": 1e-05,
      "loss": 1.4313,
      "step": 39800
    },
    {
      "epoch": 9.51,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4216502904891968,
      "eval_runtime": 33.3737,
      "eval_samples_per_second": 20.495,
      "step": 39800
    },
    {
      "epoch": 9.51,
      "learning_rate": 1e-05,
      "loss": 1.4287,
      "step": 39810
    },
    {
      "epoch": 9.51,
      "learning_rate": 1e-05,
      "loss": 1.4267,
      "step": 39820
    },
    {
      "epoch": 9.51,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 39830
    },
    {
      "epoch": 9.52,
      "learning_rate": 1e-05,
      "loss": 1.4227,
      "step": 39840
    },
    {
      "epoch": 9.52,
      "learning_rate": 1e-05,
      "loss": 1.4233,
      "step": 39850
    },
    {
      "epoch": 9.52,
      "learning_rate": 1e-05,
      "loss": 1.4235,
      "step": 39860
    },
    {
      "epoch": 9.52,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 39870
    },
    {
      "epoch": 9.52,
      "learning_rate": 1e-05,
      "loss": 1.4198,
      "step": 39880
    },
    {
      "epoch": 9.53,
      "learning_rate": 1e-05,
      "loss": 1.4246,
      "step": 39890
    },
    {
      "epoch": 9.53,
      "learning_rate": 1e-05,
      "loss": 1.4232,
      "step": 39900
    },
    {
      "epoch": 9.53,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4226136207580566,
      "eval_runtime": 34.0899,
      "eval_samples_per_second": 20.065,
      "step": 39900
    },
    {
      "epoch": 9.53,
      "learning_rate": 1e-05,
      "loss": 1.4292,
      "step": 39910
    },
    {
      "epoch": 9.53,
      "learning_rate": 1e-05,
      "loss": 1.422,
      "step": 39920
    },
    {
      "epoch": 9.54,
      "learning_rate": 1e-05,
      "loss": 1.4242,
      "step": 39930
    },
    {
      "epoch": 9.54,
      "learning_rate": 1e-05,
      "loss": 1.4214,
      "step": 39940
    },
    {
      "epoch": 9.54,
      "learning_rate": 1e-05,
      "loss": 1.4262,
      "step": 39950
    },
    {
      "epoch": 9.54,
      "learning_rate": 1e-05,
      "loss": 1.4243,
      "step": 39960
    },
    {
      "epoch": 9.55,
      "learning_rate": 1e-05,
      "loss": 1.4258,
      "step": 39970
    },
    {
      "epoch": 9.55,
      "learning_rate": 1e-05,
      "loss": 1.4257,
      "step": 39980
    },
    {
      "epoch": 9.55,
      "learning_rate": 1e-05,
      "loss": 1.4166,
      "step": 39990
    },
    {
      "epoch": 9.55,
      "learning_rate": 1e-05,
      "loss": 1.4284,
      "step": 40000
    },
    {
      "epoch": 9.55,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.422987937927246,
      "eval_runtime": 33.5402,
      "eval_samples_per_second": 20.393,
      "step": 40000
    },
    {
      "epoch": 9.56,
      "learning_rate": 1e-05,
      "loss": 1.4242,
      "step": 40010
    },
    {
      "epoch": 9.56,
      "learning_rate": 1e-05,
      "loss": 1.427,
      "step": 40020
    },
    {
      "epoch": 9.56,
      "learning_rate": 1e-05,
      "loss": 1.4195,
      "step": 40030
    },
    {
      "epoch": 9.56,
      "learning_rate": 1e-05,
      "loss": 1.422,
      "step": 40040
    },
    {
      "epoch": 9.57,
      "learning_rate": 1e-05,
      "loss": 1.4223,
      "step": 40050
    },
    {
      "epoch": 9.57,
      "learning_rate": 1e-05,
      "loss": 1.4252,
      "step": 40060
    },
    {
      "epoch": 9.57,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 40070
    },
    {
      "epoch": 9.57,
      "learning_rate": 1e-05,
      "loss": 1.425,
      "step": 40080
    },
    {
      "epoch": 9.57,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 40090
    },
    {
      "epoch": 9.58,
      "learning_rate": 1e-05,
      "loss": 1.4265,
      "step": 40100
    },
    {
      "epoch": 9.58,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4224494695663452,
      "eval_runtime": 33.3125,
      "eval_samples_per_second": 20.533,
      "step": 40100
    },
    {
      "epoch": 9.58,
      "learning_rate": 1e-05,
      "loss": 1.4253,
      "step": 40110
    },
    {
      "epoch": 9.58,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 40120
    },
    {
      "epoch": 9.58,
      "learning_rate": 1e-05,
      "loss": 1.4217,
      "step": 40130
    },
    {
      "epoch": 9.59,
      "learning_rate": 1e-05,
      "loss": 1.4251,
      "step": 40140
    },
    {
      "epoch": 9.59,
      "learning_rate": 1e-05,
      "loss": 1.4217,
      "step": 40150
    },
    {
      "epoch": 9.59,
      "learning_rate": 1e-05,
      "loss": 1.4245,
      "step": 40160
    },
    {
      "epoch": 9.59,
      "learning_rate": 1e-05,
      "loss": 1.4328,
      "step": 40170
    },
    {
      "epoch": 9.6,
      "learning_rate": 1e-05,
      "loss": 1.429,
      "step": 40180
    },
    {
      "epoch": 9.6,
      "learning_rate": 1e-05,
      "loss": 1.4272,
      "step": 40190
    },
    {
      "epoch": 9.6,
      "learning_rate": 1e-05,
      "loss": 1.43,
      "step": 40200
    },
    {
      "epoch": 9.6,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4212323427200317,
      "eval_runtime": 33.3884,
      "eval_samples_per_second": 20.486,
      "step": 40200
    },
    {
      "epoch": 9.6,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 40210
    },
    {
      "epoch": 9.61,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 40220
    },
    {
      "epoch": 9.61,
      "learning_rate": 1e-05,
      "loss": 1.4251,
      "step": 40230
    },
    {
      "epoch": 9.61,
      "learning_rate": 1e-05,
      "loss": 1.4225,
      "step": 40240
    },
    {
      "epoch": 9.61,
      "learning_rate": 1e-05,
      "loss": 1.4229,
      "step": 40250
    },
    {
      "epoch": 9.62,
      "learning_rate": 1e-05,
      "loss": 1.4329,
      "step": 40260
    },
    {
      "epoch": 9.62,
      "learning_rate": 1e-05,
      "loss": 1.4274,
      "step": 40270
    },
    {
      "epoch": 9.62,
      "learning_rate": 1e-05,
      "loss": 1.4212,
      "step": 40280
    },
    {
      "epoch": 9.62,
      "learning_rate": 1e-05,
      "loss": 1.4227,
      "step": 40290
    },
    {
      "epoch": 9.63,
      "learning_rate": 1e-05,
      "loss": 1.4331,
      "step": 40300
    },
    {
      "epoch": 9.63,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4211924076080322,
      "eval_runtime": 33.0788,
      "eval_samples_per_second": 20.678,
      "step": 40300
    },
    {
      "epoch": 9.63,
      "learning_rate": 1e-05,
      "loss": 1.4269,
      "step": 40310
    },
    {
      "epoch": 9.63,
      "learning_rate": 1e-05,
      "loss": 1.4264,
      "step": 40320
    },
    {
      "epoch": 9.63,
      "learning_rate": 1e-05,
      "loss": 1.4255,
      "step": 40330
    },
    {
      "epoch": 9.63,
      "learning_rate": 1e-05,
      "loss": 1.425,
      "step": 40340
    },
    {
      "epoch": 9.64,
      "learning_rate": 1e-05,
      "loss": 1.4258,
      "step": 40350
    },
    {
      "epoch": 9.64,
      "learning_rate": 1e-05,
      "loss": 1.4177,
      "step": 40360
    },
    {
      "epoch": 9.64,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 40370
    },
    {
      "epoch": 9.64,
      "learning_rate": 1e-05,
      "loss": 1.4238,
      "step": 40380
    },
    {
      "epoch": 9.65,
      "learning_rate": 1e-05,
      "loss": 1.4319,
      "step": 40390
    },
    {
      "epoch": 9.65,
      "learning_rate": 1e-05,
      "loss": 1.4253,
      "step": 40400
    },
    {
      "epoch": 9.65,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.421921968460083,
      "eval_runtime": 33.1988,
      "eval_samples_per_second": 20.603,
      "step": 40400
    },
    {
      "epoch": 9.65,
      "learning_rate": 1e-05,
      "loss": 1.4235,
      "step": 40410
    },
    {
      "epoch": 9.65,
      "learning_rate": 1e-05,
      "loss": 1.4245,
      "step": 40420
    },
    {
      "epoch": 9.66,
      "learning_rate": 1e-05,
      "loss": 1.4656,
      "step": 40430
    },
    {
      "epoch": 9.66,
      "learning_rate": 1e-05,
      "loss": 1.433,
      "step": 40440
    },
    {
      "epoch": 9.66,
      "learning_rate": 1e-05,
      "loss": 1.4322,
      "step": 40450
    },
    {
      "epoch": 9.66,
      "learning_rate": 1e-05,
      "loss": 1.4212,
      "step": 40460
    },
    {
      "epoch": 9.67,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 40470
    },
    {
      "epoch": 9.67,
      "learning_rate": 1e-05,
      "loss": 1.4229,
      "step": 40480
    },
    {
      "epoch": 9.67,
      "learning_rate": 1e-05,
      "loss": 1.4207,
      "step": 40490
    },
    {
      "epoch": 9.67,
      "learning_rate": 1e-05,
      "loss": 1.4276,
      "step": 40500
    },
    {
      "epoch": 9.67,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.422469139099121,
      "eval_runtime": 33.2691,
      "eval_samples_per_second": 20.56,
      "step": 40500
    },
    {
      "epoch": 9.68,
      "learning_rate": 1e-05,
      "loss": 1.4303,
      "step": 40510
    },
    {
      "epoch": 9.68,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 40520
    },
    {
      "epoch": 9.68,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 40530
    },
    {
      "epoch": 9.68,
      "learning_rate": 1e-05,
      "loss": 1.4275,
      "step": 40540
    },
    {
      "epoch": 9.68,
      "learning_rate": 1e-05,
      "loss": 1.4304,
      "step": 40550
    },
    {
      "epoch": 9.69,
      "learning_rate": 1e-05,
      "loss": 1.4244,
      "step": 40560
    },
    {
      "epoch": 9.69,
      "learning_rate": 1e-05,
      "loss": 1.4221,
      "step": 40570
    },
    {
      "epoch": 9.69,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 40580
    },
    {
      "epoch": 9.69,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 40590
    },
    {
      "epoch": 9.7,
      "learning_rate": 1e-05,
      "loss": 1.4243,
      "step": 40600
    },
    {
      "epoch": 9.7,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4217225313186646,
      "eval_runtime": 33.8383,
      "eval_samples_per_second": 20.214,
      "step": 40600
    },
    {
      "epoch": 9.7,
      "learning_rate": 1e-05,
      "loss": 1.4351,
      "step": 40610
    },
    {
      "epoch": 9.7,
      "learning_rate": 1e-05,
      "loss": 1.4616,
      "step": 40620
    },
    {
      "epoch": 9.7,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 40630
    },
    {
      "epoch": 9.71,
      "learning_rate": 1e-05,
      "loss": 1.4186,
      "step": 40640
    },
    {
      "epoch": 9.71,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 40650
    },
    {
      "epoch": 9.71,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 40660
    },
    {
      "epoch": 9.71,
      "learning_rate": 1e-05,
      "loss": 1.4243,
      "step": 40670
    },
    {
      "epoch": 9.72,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 40680
    },
    {
      "epoch": 9.72,
      "learning_rate": 1e-05,
      "loss": 1.4309,
      "step": 40690
    },
    {
      "epoch": 9.72,
      "learning_rate": 1e-05,
      "loss": 1.4239,
      "step": 40700
    },
    {
      "epoch": 9.72,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4228404760360718,
      "eval_runtime": 33.5617,
      "eval_samples_per_second": 20.38,
      "step": 40700
    },
    {
      "epoch": 9.72,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 40710
    },
    {
      "epoch": 9.73,
      "learning_rate": 1e-05,
      "loss": 1.4254,
      "step": 40720
    },
    {
      "epoch": 9.73,
      "learning_rate": 1e-05,
      "loss": 1.4229,
      "step": 40730
    },
    {
      "epoch": 9.73,
      "learning_rate": 1e-05,
      "loss": 1.4282,
      "step": 40740
    },
    {
      "epoch": 9.73,
      "learning_rate": 1e-05,
      "loss": 1.4261,
      "step": 40750
    },
    {
      "epoch": 9.73,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 40760
    },
    {
      "epoch": 9.74,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 40770
    },
    {
      "epoch": 9.74,
      "learning_rate": 1e-05,
      "loss": 1.4229,
      "step": 40780
    },
    {
      "epoch": 9.74,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 40790
    },
    {
      "epoch": 9.74,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 40800
    },
    {
      "epoch": 9.74,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.42117440700531,
      "eval_runtime": 33.4467,
      "eval_samples_per_second": 20.45,
      "step": 40800
    },
    {
      "epoch": 9.75,
      "learning_rate": 1e-05,
      "loss": 1.4241,
      "step": 40810
    },
    {
      "epoch": 9.75,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 40820
    },
    {
      "epoch": 9.75,
      "learning_rate": 1e-05,
      "loss": 1.4199,
      "step": 40830
    },
    {
      "epoch": 9.75,
      "learning_rate": 1e-05,
      "loss": 1.4234,
      "step": 40840
    },
    {
      "epoch": 9.76,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 40850
    },
    {
      "epoch": 9.76,
      "learning_rate": 1e-05,
      "loss": 1.4225,
      "step": 40860
    },
    {
      "epoch": 9.76,
      "learning_rate": 1e-05,
      "loss": 1.4284,
      "step": 40870
    },
    {
      "epoch": 9.76,
      "learning_rate": 1e-05,
      "loss": 1.4233,
      "step": 40880
    },
    {
      "epoch": 9.77,
      "learning_rate": 1e-05,
      "loss": 1.4207,
      "step": 40890
    },
    {
      "epoch": 9.77,
      "learning_rate": 1e-05,
      "loss": 1.4201,
      "step": 40900
    },
    {
      "epoch": 9.77,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4211825132369995,
      "eval_runtime": 33.316,
      "eval_samples_per_second": 20.531,
      "step": 40900
    },
    {
      "epoch": 9.77,
      "learning_rate": 1e-05,
      "loss": 1.423,
      "step": 40910
    },
    {
      "epoch": 9.77,
      "learning_rate": 1e-05,
      "loss": 1.4236,
      "step": 40920
    },
    {
      "epoch": 9.78,
      "learning_rate": 1e-05,
      "loss": 1.4225,
      "step": 40930
    },
    {
      "epoch": 9.78,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 40940
    },
    {
      "epoch": 9.78,
      "learning_rate": 1e-05,
      "loss": 1.4172,
      "step": 40950
    },
    {
      "epoch": 9.78,
      "learning_rate": 1e-05,
      "loss": 1.4284,
      "step": 40960
    },
    {
      "epoch": 9.79,
      "learning_rate": 1e-05,
      "loss": 1.4299,
      "step": 40970
    },
    {
      "epoch": 9.79,
      "learning_rate": 1e-05,
      "loss": 1.4217,
      "step": 40980
    },
    {
      "epoch": 9.79,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 40990
    },
    {
      "epoch": 9.79,
      "learning_rate": 1e-05,
      "loss": 1.4255,
      "step": 41000
    },
    {
      "epoch": 9.79,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.422580599784851,
      "eval_runtime": 33.5714,
      "eval_samples_per_second": 20.374,
      "step": 41000
    },
    {
      "epoch": 9.79,
      "learning_rate": 1e-05,
      "loss": 1.4232,
      "step": 41010
    },
    {
      "epoch": 9.8,
      "learning_rate": 1e-05,
      "loss": 1.4178,
      "step": 41020
    },
    {
      "epoch": 9.8,
      "learning_rate": 1e-05,
      "loss": 1.4279,
      "step": 41030
    },
    {
      "epoch": 9.8,
      "learning_rate": 1e-05,
      "loss": 1.4198,
      "step": 41040
    },
    {
      "epoch": 9.8,
      "learning_rate": 1e-05,
      "loss": 1.4241,
      "step": 41050
    },
    {
      "epoch": 9.81,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 41060
    },
    {
      "epoch": 9.81,
      "learning_rate": 1e-05,
      "loss": 1.4286,
      "step": 41070
    },
    {
      "epoch": 9.81,
      "learning_rate": 1e-05,
      "loss": 1.4247,
      "step": 41080
    },
    {
      "epoch": 9.81,
      "learning_rate": 1e-05,
      "loss": 1.4341,
      "step": 41090
    },
    {
      "epoch": 9.82,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 41100
    },
    {
      "epoch": 9.82,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4219818115234375,
      "eval_runtime": 33.5544,
      "eval_samples_per_second": 20.385,
      "step": 41100
    },
    {
      "epoch": 9.82,
      "learning_rate": 1e-05,
      "loss": 1.4248,
      "step": 41110
    },
    {
      "epoch": 9.82,
      "learning_rate": 1e-05,
      "loss": 1.4209,
      "step": 41120
    },
    {
      "epoch": 9.82,
      "learning_rate": 1e-05,
      "loss": 1.4231,
      "step": 41130
    },
    {
      "epoch": 9.83,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 41140
    },
    {
      "epoch": 9.83,
      "learning_rate": 1e-05,
      "loss": 1.4237,
      "step": 41150
    },
    {
      "epoch": 9.83,
      "learning_rate": 1e-05,
      "loss": 1.4233,
      "step": 41160
    },
    {
      "epoch": 9.83,
      "learning_rate": 1e-05,
      "loss": 1.4287,
      "step": 41170
    },
    {
      "epoch": 9.84,
      "learning_rate": 1e-05,
      "loss": 1.4208,
      "step": 41180
    },
    {
      "epoch": 9.84,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 41190
    },
    {
      "epoch": 9.84,
      "learning_rate": 1e-05,
      "loss": 1.4243,
      "step": 41200
    },
    {
      "epoch": 9.84,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4212608337402344,
      "eval_runtime": 33.2666,
      "eval_samples_per_second": 20.561,
      "step": 41200
    },
    {
      "epoch": 9.84,
      "learning_rate": 1e-05,
      "loss": 1.4201,
      "step": 41210
    },
    {
      "epoch": 9.84,
      "learning_rate": 1e-05,
      "loss": 1.4209,
      "step": 41220
    },
    {
      "epoch": 9.85,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 41230
    },
    {
      "epoch": 9.85,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 41240
    },
    {
      "epoch": 9.85,
      "learning_rate": 1e-05,
      "loss": 1.4288,
      "step": 41250
    },
    {
      "epoch": 9.85,
      "learning_rate": 1e-05,
      "loss": 1.4233,
      "step": 41260
    },
    {
      "epoch": 9.86,
      "learning_rate": 1e-05,
      "loss": 1.4157,
      "step": 41270
    },
    {
      "epoch": 9.86,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 41280
    },
    {
      "epoch": 9.86,
      "learning_rate": 1e-05,
      "loss": 1.4225,
      "step": 41290
    },
    {
      "epoch": 9.86,
      "learning_rate": 1e-05,
      "loss": 1.4253,
      "step": 41300
    },
    {
      "epoch": 9.86,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.422195315361023,
      "eval_runtime": 33.6187,
      "eval_samples_per_second": 20.346,
      "step": 41300
    },
    {
      "epoch": 9.87,
      "learning_rate": 1e-05,
      "loss": 1.4217,
      "step": 41310
    },
    {
      "epoch": 9.87,
      "learning_rate": 1e-05,
      "loss": 1.423,
      "step": 41320
    },
    {
      "epoch": 9.87,
      "learning_rate": 1e-05,
      "loss": 1.4233,
      "step": 41330
    },
    {
      "epoch": 9.87,
      "learning_rate": 1e-05,
      "loss": 1.4241,
      "step": 41340
    },
    {
      "epoch": 9.88,
      "learning_rate": 1e-05,
      "loss": 1.428,
      "step": 41350
    },
    {
      "epoch": 9.88,
      "learning_rate": 1e-05,
      "loss": 1.4233,
      "step": 41360
    },
    {
      "epoch": 9.88,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 41370
    },
    {
      "epoch": 9.88,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 41380
    },
    {
      "epoch": 9.89,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 41390
    },
    {
      "epoch": 9.89,
      "learning_rate": 1e-05,
      "loss": 1.4192,
      "step": 41400
    },
    {
      "epoch": 9.89,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4221385717391968,
      "eval_runtime": 33.2803,
      "eval_samples_per_second": 20.553,
      "step": 41400
    },
    {
      "epoch": 9.89,
      "learning_rate": 1e-05,
      "loss": 1.4239,
      "step": 41410
    },
    {
      "epoch": 9.89,
      "learning_rate": 1e-05,
      "loss": 1.4258,
      "step": 41420
    },
    {
      "epoch": 9.89,
      "learning_rate": 1e-05,
      "loss": 1.4226,
      "step": 41430
    },
    {
      "epoch": 9.9,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 41440
    },
    {
      "epoch": 9.9,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 41450
    },
    {
      "epoch": 9.9,
      "learning_rate": 1e-05,
      "loss": 1.4295,
      "step": 41460
    },
    {
      "epoch": 9.9,
      "learning_rate": 1e-05,
      "loss": 1.4243,
      "step": 41470
    },
    {
      "epoch": 9.91,
      "learning_rate": 1e-05,
      "loss": 1.4274,
      "step": 41480
    },
    {
      "epoch": 9.91,
      "learning_rate": 1e-05,
      "loss": 1.4193,
      "step": 41490
    },
    {
      "epoch": 9.91,
      "learning_rate": 1e-05,
      "loss": 1.4257,
      "step": 41500
    },
    {
      "epoch": 9.91,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4201841354370117,
      "eval_runtime": 33.4407,
      "eval_samples_per_second": 20.454,
      "step": 41500
    },
    {
      "epoch": 9.91,
      "learning_rate": 1e-05,
      "loss": 1.4234,
      "step": 41510
    },
    {
      "epoch": 9.92,
      "learning_rate": 1e-05,
      "loss": 1.4231,
      "step": 41520
    },
    {
      "epoch": 9.92,
      "learning_rate": 1e-05,
      "loss": 1.4292,
      "step": 41530
    },
    {
      "epoch": 9.92,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 41540
    },
    {
      "epoch": 9.92,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 41550
    },
    {
      "epoch": 9.93,
      "learning_rate": 1e-05,
      "loss": 1.4304,
      "step": 41560
    },
    {
      "epoch": 9.93,
      "learning_rate": 1e-05,
      "loss": 1.4195,
      "step": 41570
    },
    {
      "epoch": 9.93,
      "learning_rate": 1e-05,
      "loss": 1.4174,
      "step": 41580
    },
    {
      "epoch": 9.93,
      "learning_rate": 1e-05,
      "loss": 1.4208,
      "step": 41590
    },
    {
      "epoch": 9.94,
      "learning_rate": 1e-05,
      "loss": 1.422,
      "step": 41600
    },
    {
      "epoch": 9.94,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4219263792037964,
      "eval_runtime": 33.4542,
      "eval_samples_per_second": 20.446,
      "step": 41600
    },
    {
      "epoch": 9.94,
      "learning_rate": 1e-05,
      "loss": 1.4291,
      "step": 41610
    },
    {
      "epoch": 9.94,
      "learning_rate": 1e-05,
      "loss": 1.4198,
      "step": 41620
    },
    {
      "epoch": 9.94,
      "learning_rate": 1e-05,
      "loss": 1.4302,
      "step": 41630
    },
    {
      "epoch": 9.95,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 41640
    },
    {
      "epoch": 9.95,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 41650
    },
    {
      "epoch": 9.95,
      "learning_rate": 1e-05,
      "loss": 1.4242,
      "step": 41660
    },
    {
      "epoch": 9.95,
      "learning_rate": 1e-05,
      "loss": 1.4192,
      "step": 41670
    },
    {
      "epoch": 9.95,
      "learning_rate": 1e-05,
      "loss": 1.4262,
      "step": 41680
    },
    {
      "epoch": 9.96,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 41690
    },
    {
      "epoch": 9.96,
      "learning_rate": 1e-05,
      "loss": 1.4246,
      "step": 41700
    },
    {
      "epoch": 9.96,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4206942319869995,
      "eval_runtime": 33.0206,
      "eval_samples_per_second": 20.714,
      "step": 41700
    },
    {
      "epoch": 9.96,
      "learning_rate": 1e-05,
      "loss": 1.4208,
      "step": 41710
    },
    {
      "epoch": 9.96,
      "learning_rate": 1e-05,
      "loss": 1.4243,
      "step": 41720
    },
    {
      "epoch": 9.97,
      "learning_rate": 1e-05,
      "loss": 1.4236,
      "step": 41730
    },
    {
      "epoch": 9.97,
      "learning_rate": 1e-05,
      "loss": 1.4338,
      "step": 41740
    },
    {
      "epoch": 9.97,
      "learning_rate": 1e-05,
      "loss": 1.4252,
      "step": 41750
    },
    {
      "epoch": 9.97,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 41760
    },
    {
      "epoch": 9.98,
      "learning_rate": 1e-05,
      "loss": 1.4248,
      "step": 41770
    },
    {
      "epoch": 9.98,
      "learning_rate": 1e-05,
      "loss": 1.4236,
      "step": 41780
    },
    {
      "epoch": 9.98,
      "learning_rate": 1e-05,
      "loss": 1.4259,
      "step": 41790
    },
    {
      "epoch": 9.98,
      "learning_rate": 1e-05,
      "loss": 1.4231,
      "step": 41800
    },
    {
      "epoch": 9.98,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4214091300964355,
      "eval_runtime": 33.2513,
      "eval_samples_per_second": 20.571,
      "step": 41800
    },
    {
      "epoch": 9.99,
      "learning_rate": 1e-05,
      "loss": 1.4223,
      "step": 41810
    },
    {
      "epoch": 9.99,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 41820
    },
    {
      "epoch": 9.99,
      "learning_rate": 1e-05,
      "loss": 1.4302,
      "step": 41830
    },
    {
      "epoch": 9.99,
      "learning_rate": 1e-05,
      "loss": 1.4239,
      "step": 41840
    },
    {
      "epoch": 10.0,
      "learning_rate": 1e-05,
      "loss": 1.4278,
      "step": 41850
    },
    {
      "epoch": 10.0,
      "learning_rate": 1e-05,
      "loss": 1.42,
      "step": 41860
    },
    {
      "epoch": 10.0,
      "learning_rate": 1e-05,
      "loss": 1.4235,
      "step": 41870
    },
    {
      "epoch": 10.0,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 41880
    },
    {
      "epoch": 10.0,
      "learning_rate": 1e-05,
      "loss": 1.4331,
      "step": 41890
    },
    {
      "epoch": 10.01,
      "learning_rate": 1e-05,
      "loss": 1.4225,
      "step": 41900
    },
    {
      "epoch": 10.01,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.420977234840393,
      "eval_runtime": 33.6017,
      "eval_samples_per_second": 20.356,
      "step": 41900
    },
    {
      "epoch": 10.01,
      "learning_rate": 1e-05,
      "loss": 1.4229,
      "step": 41910
    },
    {
      "epoch": 10.01,
      "learning_rate": 1e-05,
      "loss": 1.4285,
      "step": 41920
    },
    {
      "epoch": 10.01,
      "learning_rate": 1e-05,
      "loss": 1.4236,
      "step": 41930
    },
    {
      "epoch": 10.02,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 41940
    },
    {
      "epoch": 10.02,
      "learning_rate": 1e-05,
      "loss": 1.437,
      "step": 41950
    },
    {
      "epoch": 10.02,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 41960
    },
    {
      "epoch": 10.02,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 41970
    },
    {
      "epoch": 10.03,
      "learning_rate": 1e-05,
      "loss": 1.4287,
      "step": 41980
    },
    {
      "epoch": 10.03,
      "learning_rate": 1e-05,
      "loss": 1.4224,
      "step": 41990
    },
    {
      "epoch": 10.03,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 42000
    },
    {
      "epoch": 10.03,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4210131168365479,
      "eval_runtime": 33.2293,
      "eval_samples_per_second": 20.584,
      "step": 42000
    },
    {
      "epoch": 10.03,
      "learning_rate": 1e-05,
      "loss": 1.4328,
      "step": 42010
    },
    {
      "epoch": 10.04,
      "learning_rate": 1e-05,
      "loss": 1.42,
      "step": 42020
    },
    {
      "epoch": 10.04,
      "learning_rate": 1e-05,
      "loss": 1.4255,
      "step": 42030
    },
    {
      "epoch": 10.04,
      "learning_rate": 1e-05,
      "loss": 1.4208,
      "step": 42040
    },
    {
      "epoch": 10.04,
      "learning_rate": 1e-05,
      "loss": 1.4229,
      "step": 42050
    },
    {
      "epoch": 10.05,
      "learning_rate": 1e-05,
      "loss": 1.4244,
      "step": 42060
    },
    {
      "epoch": 10.05,
      "learning_rate": 1e-05,
      "loss": 1.4165,
      "step": 42070
    },
    {
      "epoch": 10.05,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 42080
    },
    {
      "epoch": 10.05,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 42090
    },
    {
      "epoch": 10.05,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 42100
    },
    {
      "epoch": 10.05,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4214699268341064,
      "eval_runtime": 33.3269,
      "eval_samples_per_second": 20.524,
      "step": 42100
    },
    {
      "epoch": 10.06,
      "learning_rate": 1e-05,
      "loss": 1.4283,
      "step": 42110
    },
    {
      "epoch": 10.06,
      "learning_rate": 1e-05,
      "loss": 1.4226,
      "step": 42120
    },
    {
      "epoch": 10.06,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 42130
    },
    {
      "epoch": 10.06,
      "learning_rate": 1e-05,
      "loss": 1.4178,
      "step": 42140
    },
    {
      "epoch": 10.07,
      "learning_rate": 1e-05,
      "loss": 1.4174,
      "step": 42150
    },
    {
      "epoch": 10.07,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 42160
    },
    {
      "epoch": 10.07,
      "learning_rate": 1e-05,
      "loss": 1.4238,
      "step": 42170
    },
    {
      "epoch": 10.07,
      "learning_rate": 1e-05,
      "loss": 1.4271,
      "step": 42180
    },
    {
      "epoch": 10.08,
      "learning_rate": 1e-05,
      "loss": 1.4264,
      "step": 42190
    },
    {
      "epoch": 10.08,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 42200
    },
    {
      "epoch": 10.08,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.419441819190979,
      "eval_runtime": 33.1508,
      "eval_samples_per_second": 20.633,
      "step": 42200
    },
    {
      "epoch": 10.08,
      "learning_rate": 1e-05,
      "loss": 1.4246,
      "step": 42210
    },
    {
      "epoch": 10.08,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 42220
    },
    {
      "epoch": 10.09,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 42230
    },
    {
      "epoch": 10.09,
      "learning_rate": 1e-05,
      "loss": 1.4234,
      "step": 42240
    },
    {
      "epoch": 10.09,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 42250
    },
    {
      "epoch": 10.09,
      "learning_rate": 1e-05,
      "loss": 1.4275,
      "step": 42260
    },
    {
      "epoch": 10.1,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 42270
    },
    {
      "epoch": 10.1,
      "learning_rate": 1e-05,
      "loss": 1.4187,
      "step": 42280
    },
    {
      "epoch": 10.1,
      "learning_rate": 1e-05,
      "loss": 1.4226,
      "step": 42290
    },
    {
      "epoch": 10.1,
      "learning_rate": 1e-05,
      "loss": 1.4281,
      "step": 42300
    },
    {
      "epoch": 10.1,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4207208156585693,
      "eval_runtime": 32.8888,
      "eval_samples_per_second": 20.797,
      "step": 42300
    },
    {
      "epoch": 10.11,
      "learning_rate": 1e-05,
      "loss": 1.4221,
      "step": 42310
    },
    {
      "epoch": 10.11,
      "learning_rate": 1e-05,
      "loss": 1.4254,
      "step": 42320
    },
    {
      "epoch": 10.11,
      "learning_rate": 1e-05,
      "loss": 1.4274,
      "step": 42330
    },
    {
      "epoch": 10.11,
      "learning_rate": 1e-05,
      "loss": 1.4232,
      "step": 42340
    },
    {
      "epoch": 10.11,
      "learning_rate": 1e-05,
      "loss": 1.4218,
      "step": 42350
    },
    {
      "epoch": 10.12,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 42360
    },
    {
      "epoch": 10.12,
      "learning_rate": 1e-05,
      "loss": 1.4238,
      "step": 42370
    },
    {
      "epoch": 10.12,
      "learning_rate": 1e-05,
      "loss": 1.4187,
      "step": 42380
    },
    {
      "epoch": 10.12,
      "learning_rate": 1e-05,
      "loss": 1.4235,
      "step": 42390
    },
    {
      "epoch": 10.13,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 42400
    },
    {
      "epoch": 10.13,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4203609228134155,
      "eval_runtime": 32.9066,
      "eval_samples_per_second": 20.786,
      "step": 42400
    },
    {
      "epoch": 10.13,
      "learning_rate": 1e-05,
      "loss": 1.4246,
      "step": 42410
    },
    {
      "epoch": 10.13,
      "learning_rate": 1e-05,
      "loss": 1.4193,
      "step": 42420
    },
    {
      "epoch": 10.13,
      "learning_rate": 1e-05,
      "loss": 1.4287,
      "step": 42430
    },
    {
      "epoch": 10.14,
      "learning_rate": 1e-05,
      "loss": 1.4268,
      "step": 42440
    },
    {
      "epoch": 10.14,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 42450
    },
    {
      "epoch": 10.14,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 42460
    },
    {
      "epoch": 10.14,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 42470
    },
    {
      "epoch": 10.15,
      "learning_rate": 1e-05,
      "loss": 1.4208,
      "step": 42480
    },
    {
      "epoch": 10.15,
      "learning_rate": 1e-05,
      "loss": 1.4284,
      "step": 42490
    },
    {
      "epoch": 10.15,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 42500
    },
    {
      "epoch": 10.15,
      "eval_accuracy": 0.9868421052631579,
      "eval_loss": 1.4205641746520996,
      "eval_runtime": 33.3764,
      "eval_samples_per_second": 20.494,
      "step": 42500
    },
    {
      "epoch": 10.15,
      "learning_rate": 1e-05,
      "loss": 1.4195,
      "step": 42510
    },
    {
      "epoch": 10.16,
      "learning_rate": 1e-05,
      "loss": 1.42,
      "step": 42520
    },
    {
      "epoch": 10.16,
      "learning_rate": 1e-05,
      "loss": 1.4217,
      "step": 42530
    },
    {
      "epoch": 10.16,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 42540
    },
    {
      "epoch": 10.16,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 42550
    },
    {
      "epoch": 10.16,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 42560
    },
    {
      "epoch": 10.17,
      "learning_rate": 1e-05,
      "loss": 1.4242,
      "step": 42570
    },
    {
      "epoch": 10.17,
      "learning_rate": 1e-05,
      "loss": 1.4213,
      "step": 42580
    },
    {
      "epoch": 10.17,
      "learning_rate": 1e-05,
      "loss": 1.4177,
      "step": 42590
    },
    {
      "epoch": 10.17,
      "learning_rate": 1e-05,
      "loss": 1.4217,
      "step": 42600
    },
    {
      "epoch": 10.17,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4201807975769043,
      "eval_runtime": 33.1994,
      "eval_samples_per_second": 20.603,
      "step": 42600
    },
    {
      "epoch": 10.18,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 42610
    },
    {
      "epoch": 10.18,
      "learning_rate": 1e-05,
      "loss": 1.4277,
      "step": 42620
    },
    {
      "epoch": 10.18,
      "learning_rate": 1e-05,
      "loss": 1.4203,
      "step": 42630
    },
    {
      "epoch": 10.18,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 42640
    },
    {
      "epoch": 10.19,
      "learning_rate": 1e-05,
      "loss": 1.422,
      "step": 42650
    },
    {
      "epoch": 10.19,
      "learning_rate": 1e-05,
      "loss": 1.4207,
      "step": 42660
    },
    {
      "epoch": 10.19,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 42670
    },
    {
      "epoch": 10.19,
      "learning_rate": 1e-05,
      "loss": 1.4218,
      "step": 42680
    },
    {
      "epoch": 10.2,
      "learning_rate": 1e-05,
      "loss": 1.4224,
      "step": 42690
    },
    {
      "epoch": 10.2,
      "learning_rate": 1e-05,
      "loss": 1.423,
      "step": 42700
    },
    {
      "epoch": 10.2,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4194331169128418,
      "eval_runtime": 33.5572,
      "eval_samples_per_second": 20.383,
      "step": 42700
    },
    {
      "epoch": 10.2,
      "learning_rate": 1e-05,
      "loss": 1.4192,
      "step": 42710
    },
    {
      "epoch": 10.2,
      "learning_rate": 1e-05,
      "loss": 1.4209,
      "step": 42720
    },
    {
      "epoch": 10.21,
      "learning_rate": 1e-05,
      "loss": 1.4207,
      "step": 42730
    },
    {
      "epoch": 10.21,
      "learning_rate": 1e-05,
      "loss": 1.4283,
      "step": 42740
    },
    {
      "epoch": 10.21,
      "learning_rate": 1e-05,
      "loss": 1.4213,
      "step": 42750
    },
    {
      "epoch": 10.21,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 42760
    },
    {
      "epoch": 10.21,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 42770
    },
    {
      "epoch": 10.22,
      "learning_rate": 1e-05,
      "loss": 1.4275,
      "step": 42780
    },
    {
      "epoch": 10.22,
      "learning_rate": 1e-05,
      "loss": 1.4249,
      "step": 42790
    },
    {
      "epoch": 10.22,
      "learning_rate": 1e-05,
      "loss": 1.4249,
      "step": 42800
    },
    {
      "epoch": 10.22,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.419966459274292,
      "eval_runtime": 33.2523,
      "eval_samples_per_second": 20.57,
      "step": 42800
    },
    {
      "epoch": 10.22,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 42810
    },
    {
      "epoch": 10.23,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 42820
    },
    {
      "epoch": 10.23,
      "learning_rate": 1e-05,
      "loss": 1.4352,
      "step": 42830
    },
    {
      "epoch": 10.23,
      "learning_rate": 1e-05,
      "loss": 1.4147,
      "step": 42840
    },
    {
      "epoch": 10.23,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 42850
    },
    {
      "epoch": 10.24,
      "learning_rate": 1e-05,
      "loss": 1.426,
      "step": 42860
    },
    {
      "epoch": 10.24,
      "learning_rate": 1e-05,
      "loss": 1.4218,
      "step": 42870
    },
    {
      "epoch": 10.24,
      "learning_rate": 1e-05,
      "loss": 1.4209,
      "step": 42880
    },
    {
      "epoch": 10.24,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 42890
    },
    {
      "epoch": 10.25,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 42900
    },
    {
      "epoch": 10.25,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4201442003250122,
      "eval_runtime": 32.7213,
      "eval_samples_per_second": 20.904,
      "step": 42900
    },
    {
      "epoch": 10.25,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 42910
    },
    {
      "epoch": 10.25,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 42920
    },
    {
      "epoch": 10.25,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 42930
    },
    {
      "epoch": 10.26,
      "learning_rate": 1e-05,
      "loss": 1.4186,
      "step": 42940
    },
    {
      "epoch": 10.26,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 42950
    },
    {
      "epoch": 10.26,
      "learning_rate": 1e-05,
      "loss": 1.4176,
      "step": 42960
    },
    {
      "epoch": 10.26,
      "learning_rate": 1e-05,
      "loss": 1.425,
      "step": 42970
    },
    {
      "epoch": 10.27,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 42980
    },
    {
      "epoch": 10.27,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 42990
    },
    {
      "epoch": 10.27,
      "learning_rate": 1e-05,
      "loss": 1.4262,
      "step": 43000
    },
    {
      "epoch": 10.27,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4207741022109985,
      "eval_runtime": 32.9724,
      "eval_samples_per_second": 20.745,
      "step": 43000
    },
    {
      "epoch": 10.27,
      "learning_rate": 1e-05,
      "loss": 1.4244,
      "step": 43010
    },
    {
      "epoch": 10.27,
      "learning_rate": 1e-05,
      "loss": 1.4181,
      "step": 43020
    },
    {
      "epoch": 10.28,
      "learning_rate": 1e-05,
      "loss": 1.431,
      "step": 43030
    },
    {
      "epoch": 10.28,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 43040
    },
    {
      "epoch": 10.28,
      "learning_rate": 1e-05,
      "loss": 1.4248,
      "step": 43050
    },
    {
      "epoch": 10.28,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 43060
    },
    {
      "epoch": 10.29,
      "learning_rate": 1e-05,
      "loss": 1.4271,
      "step": 43070
    },
    {
      "epoch": 10.29,
      "learning_rate": 1e-05,
      "loss": 1.4198,
      "step": 43080
    },
    {
      "epoch": 10.29,
      "learning_rate": 1e-05,
      "loss": 1.4285,
      "step": 43090
    },
    {
      "epoch": 10.29,
      "learning_rate": 1e-05,
      "loss": 1.4246,
      "step": 43100
    },
    {
      "epoch": 10.29,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4192308187484741,
      "eval_runtime": 32.9005,
      "eval_samples_per_second": 20.79,
      "step": 43100
    },
    {
      "epoch": 10.3,
      "learning_rate": 1e-05,
      "loss": 1.4234,
      "step": 43110
    },
    {
      "epoch": 10.3,
      "learning_rate": 1e-05,
      "loss": 1.4276,
      "step": 43120
    },
    {
      "epoch": 10.3,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 43130
    },
    {
      "epoch": 10.3,
      "learning_rate": 1e-05,
      "loss": 1.4207,
      "step": 43140
    },
    {
      "epoch": 10.31,
      "learning_rate": 1e-05,
      "loss": 1.4255,
      "step": 43150
    },
    {
      "epoch": 10.31,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 43160
    },
    {
      "epoch": 10.31,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 43170
    },
    {
      "epoch": 10.31,
      "learning_rate": 1e-05,
      "loss": 1.4218,
      "step": 43180
    },
    {
      "epoch": 10.32,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 43190
    },
    {
      "epoch": 10.32,
      "learning_rate": 1e-05,
      "loss": 1.4221,
      "step": 43200
    },
    {
      "epoch": 10.32,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4196358919143677,
      "eval_runtime": 32.7831,
      "eval_samples_per_second": 20.864,
      "step": 43200
    },
    {
      "epoch": 10.32,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 43210
    },
    {
      "epoch": 10.32,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 43220
    },
    {
      "epoch": 10.32,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 43230
    },
    {
      "epoch": 10.33,
      "learning_rate": 1e-05,
      "loss": 1.4262,
      "step": 43240
    },
    {
      "epoch": 10.33,
      "learning_rate": 1e-05,
      "loss": 1.4155,
      "step": 43250
    },
    {
      "epoch": 10.33,
      "learning_rate": 1e-05,
      "loss": 1.4208,
      "step": 43260
    },
    {
      "epoch": 10.33,
      "learning_rate": 1e-05,
      "loss": 1.4273,
      "step": 43270
    },
    {
      "epoch": 10.34,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 43280
    },
    {
      "epoch": 10.34,
      "learning_rate": 1e-05,
      "loss": 1.4265,
      "step": 43290
    },
    {
      "epoch": 10.34,
      "learning_rate": 1e-05,
      "loss": 1.4257,
      "step": 43300
    },
    {
      "epoch": 10.34,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4193412065505981,
      "eval_runtime": 32.9212,
      "eval_samples_per_second": 20.777,
      "step": 43300
    },
    {
      "epoch": 10.34,
      "learning_rate": 1e-05,
      "loss": 1.4305,
      "step": 43310
    },
    {
      "epoch": 10.35,
      "learning_rate": 1e-05,
      "loss": 1.4249,
      "step": 43320
    },
    {
      "epoch": 10.35,
      "learning_rate": 1e-05,
      "loss": 1.429,
      "step": 43330
    },
    {
      "epoch": 10.35,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 43340
    },
    {
      "epoch": 10.35,
      "learning_rate": 1e-05,
      "loss": 1.4177,
      "step": 43350
    },
    {
      "epoch": 10.36,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 43360
    },
    {
      "epoch": 10.36,
      "learning_rate": 1e-05,
      "loss": 1.4216,
      "step": 43370
    },
    {
      "epoch": 10.36,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 43380
    },
    {
      "epoch": 10.36,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 43390
    },
    {
      "epoch": 10.37,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 43400
    },
    {
      "epoch": 10.37,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4191969633102417,
      "eval_runtime": 32.6248,
      "eval_samples_per_second": 20.966,
      "step": 43400
    },
    {
      "epoch": 10.37,
      "learning_rate": 1e-05,
      "loss": 1.4227,
      "step": 43410
    },
    {
      "epoch": 10.37,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 43420
    },
    {
      "epoch": 10.37,
      "learning_rate": 1e-05,
      "loss": 1.4178,
      "step": 43430
    },
    {
      "epoch": 10.37,
      "learning_rate": 1e-05,
      "loss": 1.4172,
      "step": 43440
    },
    {
      "epoch": 10.38,
      "learning_rate": 1e-05,
      "loss": 1.4192,
      "step": 43450
    },
    {
      "epoch": 10.38,
      "learning_rate": 1e-05,
      "loss": 1.429,
      "step": 43460
    },
    {
      "epoch": 10.38,
      "learning_rate": 1e-05,
      "loss": 1.4214,
      "step": 43470
    },
    {
      "epoch": 10.38,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 43480
    },
    {
      "epoch": 10.39,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 43490
    },
    {
      "epoch": 10.39,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 43500
    },
    {
      "epoch": 10.39,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4213738441467285,
      "eval_runtime": 32.5511,
      "eval_samples_per_second": 21.013,
      "step": 43500
    },
    {
      "epoch": 10.39,
      "learning_rate": 1e-05,
      "loss": 1.4245,
      "step": 43510
    },
    {
      "epoch": 10.39,
      "learning_rate": 1e-05,
      "loss": 1.4272,
      "step": 43520
    },
    {
      "epoch": 10.4,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 43530
    },
    {
      "epoch": 10.4,
      "learning_rate": 1e-05,
      "loss": 1.4282,
      "step": 43540
    },
    {
      "epoch": 10.4,
      "learning_rate": 1e-05,
      "loss": 1.4186,
      "step": 43550
    },
    {
      "epoch": 10.4,
      "learning_rate": 1e-05,
      "loss": 1.4239,
      "step": 43560
    },
    {
      "epoch": 10.41,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 43570
    },
    {
      "epoch": 10.41,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 43580
    },
    {
      "epoch": 10.41,
      "learning_rate": 1e-05,
      "loss": 1.4212,
      "step": 43590
    },
    {
      "epoch": 10.41,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 43600
    },
    {
      "epoch": 10.41,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4194648265838623,
      "eval_runtime": 32.7533,
      "eval_samples_per_second": 20.883,
      "step": 43600
    },
    {
      "epoch": 10.42,
      "learning_rate": 1e-05,
      "loss": 1.4265,
      "step": 43610
    },
    {
      "epoch": 10.42,
      "learning_rate": 1e-05,
      "loss": 1.4201,
      "step": 43620
    },
    {
      "epoch": 10.42,
      "learning_rate": 1e-05,
      "loss": 1.4233,
      "step": 43630
    },
    {
      "epoch": 10.42,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 43640
    },
    {
      "epoch": 10.43,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 43650
    },
    {
      "epoch": 10.43,
      "learning_rate": 1e-05,
      "loss": 1.4217,
      "step": 43660
    },
    {
      "epoch": 10.43,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 43670
    },
    {
      "epoch": 10.43,
      "learning_rate": 1e-05,
      "loss": 1.4244,
      "step": 43680
    },
    {
      "epoch": 10.43,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 43690
    },
    {
      "epoch": 10.44,
      "learning_rate": 1e-05,
      "loss": 1.4231,
      "step": 43700
    },
    {
      "epoch": 10.44,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4195255041122437,
      "eval_runtime": 32.6178,
      "eval_samples_per_second": 20.97,
      "step": 43700
    },
    {
      "epoch": 10.44,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 43710
    },
    {
      "epoch": 10.44,
      "learning_rate": 1e-05,
      "loss": 1.4247,
      "step": 43720
    },
    {
      "epoch": 10.44,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 43730
    },
    {
      "epoch": 10.45,
      "learning_rate": 1e-05,
      "loss": 1.4323,
      "step": 43740
    },
    {
      "epoch": 10.45,
      "learning_rate": 1e-05,
      "loss": 1.4193,
      "step": 43750
    },
    {
      "epoch": 10.45,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 43760
    },
    {
      "epoch": 10.45,
      "learning_rate": 1e-05,
      "loss": 1.4155,
      "step": 43770
    },
    {
      "epoch": 10.46,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 43780
    },
    {
      "epoch": 10.46,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 43790
    },
    {
      "epoch": 10.46,
      "learning_rate": 1e-05,
      "loss": 1.4238,
      "step": 43800
    },
    {
      "epoch": 10.46,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4194318056106567,
      "eval_runtime": 32.6973,
      "eval_samples_per_second": 20.919,
      "step": 43800
    },
    {
      "epoch": 10.46,
      "learning_rate": 1e-05,
      "loss": 1.4231,
      "step": 43810
    },
    {
      "epoch": 10.47,
      "learning_rate": 1e-05,
      "loss": 1.4166,
      "step": 43820
    },
    {
      "epoch": 10.47,
      "learning_rate": 1e-05,
      "loss": 1.4236,
      "step": 43830
    },
    {
      "epoch": 10.47,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 43840
    },
    {
      "epoch": 10.47,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 43850
    },
    {
      "epoch": 10.48,
      "learning_rate": 1e-05,
      "loss": 1.4268,
      "step": 43860
    },
    {
      "epoch": 10.48,
      "learning_rate": 1e-05,
      "loss": 1.4301,
      "step": 43870
    },
    {
      "epoch": 10.48,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 43880
    },
    {
      "epoch": 10.48,
      "learning_rate": 1e-05,
      "loss": 1.4255,
      "step": 43890
    },
    {
      "epoch": 10.48,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 43900
    },
    {
      "epoch": 10.48,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4182957410812378,
      "eval_runtime": 32.7545,
      "eval_samples_per_second": 20.883,
      "step": 43900
    },
    {
      "epoch": 10.49,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 43910
    },
    {
      "epoch": 10.49,
      "learning_rate": 1e-05,
      "loss": 1.4172,
      "step": 43920
    },
    {
      "epoch": 10.49,
      "learning_rate": 1e-05,
      "loss": 1.42,
      "step": 43930
    },
    {
      "epoch": 10.49,
      "learning_rate": 1e-05,
      "loss": 1.4297,
      "step": 43940
    },
    {
      "epoch": 10.5,
      "learning_rate": 1e-05,
      "loss": 1.4193,
      "step": 43950
    },
    {
      "epoch": 10.5,
      "learning_rate": 1e-05,
      "loss": 1.4218,
      "step": 43960
    },
    {
      "epoch": 10.5,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 43970
    },
    {
      "epoch": 10.5,
      "learning_rate": 1e-05,
      "loss": 1.4265,
      "step": 43980
    },
    {
      "epoch": 10.51,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 43990
    },
    {
      "epoch": 10.51,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 44000
    },
    {
      "epoch": 10.51,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4190998077392578,
      "eval_runtime": 32.8068,
      "eval_samples_per_second": 20.849,
      "step": 44000
    },
    {
      "epoch": 10.51,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 44010
    },
    {
      "epoch": 10.51,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 44020
    },
    {
      "epoch": 10.52,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 44030
    },
    {
      "epoch": 10.52,
      "learning_rate": 1e-05,
      "loss": 1.4213,
      "step": 44040
    },
    {
      "epoch": 10.52,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 44050
    },
    {
      "epoch": 10.52,
      "learning_rate": 1e-05,
      "loss": 1.4267,
      "step": 44060
    },
    {
      "epoch": 10.53,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 44070
    },
    {
      "epoch": 10.53,
      "learning_rate": 1e-05,
      "loss": 1.4165,
      "step": 44080
    },
    {
      "epoch": 10.53,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 44090
    },
    {
      "epoch": 10.53,
      "learning_rate": 1e-05,
      "loss": 1.4233,
      "step": 44100
    },
    {
      "epoch": 10.53,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.419988751411438,
      "eval_runtime": 32.9844,
      "eval_samples_per_second": 20.737,
      "step": 44100
    },
    {
      "epoch": 10.53,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 44110
    },
    {
      "epoch": 10.54,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 44120
    },
    {
      "epoch": 10.54,
      "learning_rate": 1e-05,
      "loss": 1.4229,
      "step": 44130
    },
    {
      "epoch": 10.54,
      "learning_rate": 1e-05,
      "loss": 1.4249,
      "step": 44140
    },
    {
      "epoch": 10.54,
      "learning_rate": 1e-05,
      "loss": 1.4239,
      "step": 44150
    },
    {
      "epoch": 10.55,
      "learning_rate": 1e-05,
      "loss": 1.4165,
      "step": 44160
    },
    {
      "epoch": 10.55,
      "learning_rate": 1e-05,
      "loss": 1.4349,
      "step": 44170
    },
    {
      "epoch": 10.55,
      "learning_rate": 1e-05,
      "loss": 1.4249,
      "step": 44180
    },
    {
      "epoch": 10.55,
      "learning_rate": 1e-05,
      "loss": 1.4209,
      "step": 44190
    },
    {
      "epoch": 10.56,
      "learning_rate": 1e-05,
      "loss": 1.4191,
      "step": 44200
    },
    {
      "epoch": 10.56,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4185675382614136,
      "eval_runtime": 32.6931,
      "eval_samples_per_second": 20.922,
      "step": 44200
    },
    {
      "epoch": 10.56,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 44210
    },
    {
      "epoch": 10.56,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 44220
    },
    {
      "epoch": 10.56,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 44230
    },
    {
      "epoch": 10.57,
      "learning_rate": 1e-05,
      "loss": 1.4256,
      "step": 44240
    },
    {
      "epoch": 10.57,
      "learning_rate": 1e-05,
      "loss": 1.4176,
      "step": 44250
    },
    {
      "epoch": 10.57,
      "learning_rate": 1e-05,
      "loss": 1.4237,
      "step": 44260
    },
    {
      "epoch": 10.57,
      "learning_rate": 1e-05,
      "loss": 1.4193,
      "step": 44270
    },
    {
      "epoch": 10.58,
      "learning_rate": 1e-05,
      "loss": 1.4176,
      "step": 44280
    },
    {
      "epoch": 10.58,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 44290
    },
    {
      "epoch": 10.58,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 44300
    },
    {
      "epoch": 10.58,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.418918490409851,
      "eval_runtime": 32.6829,
      "eval_samples_per_second": 20.928,
      "step": 44300
    },
    {
      "epoch": 10.58,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 44310
    },
    {
      "epoch": 10.59,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 44320
    },
    {
      "epoch": 10.59,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 44330
    },
    {
      "epoch": 10.59,
      "learning_rate": 1e-05,
      "loss": 1.4178,
      "step": 44340
    },
    {
      "epoch": 10.59,
      "learning_rate": 1e-05,
      "loss": 1.4248,
      "step": 44350
    },
    {
      "epoch": 10.59,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 44360
    },
    {
      "epoch": 10.6,
      "learning_rate": 1e-05,
      "loss": 1.4172,
      "step": 44370
    },
    {
      "epoch": 10.6,
      "learning_rate": 1e-05,
      "loss": 1.4187,
      "step": 44380
    },
    {
      "epoch": 10.6,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 44390
    },
    {
      "epoch": 10.6,
      "learning_rate": 1e-05,
      "loss": 1.4238,
      "step": 44400
    },
    {
      "epoch": 10.6,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4192520380020142,
      "eval_runtime": 32.8917,
      "eval_samples_per_second": 20.796,
      "step": 44400
    },
    {
      "epoch": 10.61,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 44410
    },
    {
      "epoch": 10.61,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 44420
    },
    {
      "epoch": 10.61,
      "learning_rate": 1e-05,
      "loss": 1.4217,
      "step": 44430
    },
    {
      "epoch": 10.61,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 44440
    },
    {
      "epoch": 10.62,
      "learning_rate": 1e-05,
      "loss": 1.4297,
      "step": 44450
    },
    {
      "epoch": 10.62,
      "learning_rate": 1e-05,
      "loss": 1.4177,
      "step": 44460
    },
    {
      "epoch": 10.62,
      "learning_rate": 1e-05,
      "loss": 1.4172,
      "step": 44470
    },
    {
      "epoch": 10.62,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 44480
    },
    {
      "epoch": 10.63,
      "learning_rate": 1e-05,
      "loss": 1.4187,
      "step": 44490
    },
    {
      "epoch": 10.63,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 44500
    },
    {
      "epoch": 10.63,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4202367067337036,
      "eval_runtime": 32.8436,
      "eval_samples_per_second": 20.826,
      "step": 44500
    },
    {
      "epoch": 10.63,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 44510
    },
    {
      "epoch": 10.63,
      "learning_rate": 1e-05,
      "loss": 1.4241,
      "step": 44520
    },
    {
      "epoch": 10.64,
      "learning_rate": 1e-05,
      "loss": 1.4218,
      "step": 44530
    },
    {
      "epoch": 10.64,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 44540
    },
    {
      "epoch": 10.64,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 44550
    },
    {
      "epoch": 10.64,
      "learning_rate": 1e-05,
      "loss": 1.4302,
      "step": 44560
    },
    {
      "epoch": 10.64,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 44570
    },
    {
      "epoch": 10.65,
      "learning_rate": 1e-05,
      "loss": 1.4368,
      "step": 44580
    },
    {
      "epoch": 10.65,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 44590
    },
    {
      "epoch": 10.65,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 44600
    },
    {
      "epoch": 10.65,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4197137355804443,
      "eval_runtime": 32.921,
      "eval_samples_per_second": 20.777,
      "step": 44600
    },
    {
      "epoch": 10.65,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 44610
    },
    {
      "epoch": 10.66,
      "learning_rate": 1e-05,
      "loss": 1.4187,
      "step": 44620
    },
    {
      "epoch": 10.66,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 44630
    },
    {
      "epoch": 10.66,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 44640
    },
    {
      "epoch": 10.66,
      "learning_rate": 1e-05,
      "loss": 1.4216,
      "step": 44650
    },
    {
      "epoch": 10.67,
      "learning_rate": 1e-05,
      "loss": 1.4243,
      "step": 44660
    },
    {
      "epoch": 10.67,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 44670
    },
    {
      "epoch": 10.67,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 44680
    },
    {
      "epoch": 10.67,
      "learning_rate": 1e-05,
      "loss": 1.4192,
      "step": 44690
    },
    {
      "epoch": 10.68,
      "learning_rate": 1e-05,
      "loss": 1.4218,
      "step": 44700
    },
    {
      "epoch": 10.68,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.419343113899231,
      "eval_runtime": 32.8284,
      "eval_samples_per_second": 20.836,
      "step": 44700
    },
    {
      "epoch": 10.68,
      "learning_rate": 1e-05,
      "loss": 1.4195,
      "step": 44710
    },
    {
      "epoch": 10.68,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 44720
    },
    {
      "epoch": 10.68,
      "learning_rate": 1e-05,
      "loss": 1.4214,
      "step": 44730
    },
    {
      "epoch": 10.69,
      "learning_rate": 1e-05,
      "loss": 1.429,
      "step": 44740
    },
    {
      "epoch": 10.69,
      "learning_rate": 1e-05,
      "loss": 1.4198,
      "step": 44750
    },
    {
      "epoch": 10.69,
      "learning_rate": 1e-05,
      "loss": 1.4201,
      "step": 44760
    },
    {
      "epoch": 10.69,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 44770
    },
    {
      "epoch": 10.7,
      "learning_rate": 1e-05,
      "loss": 1.4191,
      "step": 44780
    },
    {
      "epoch": 10.7,
      "learning_rate": 1e-05,
      "loss": 1.4243,
      "step": 44790
    },
    {
      "epoch": 10.7,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 44800
    },
    {
      "epoch": 10.7,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.419527292251587,
      "eval_runtime": 32.692,
      "eval_samples_per_second": 20.923,
      "step": 44800
    },
    {
      "epoch": 10.7,
      "learning_rate": 1e-05,
      "loss": 1.4198,
      "step": 44810
    },
    {
      "epoch": 10.7,
      "learning_rate": 1e-05,
      "loss": 1.4223,
      "step": 44820
    },
    {
      "epoch": 10.71,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 44830
    },
    {
      "epoch": 10.71,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 44840
    },
    {
      "epoch": 10.71,
      "learning_rate": 1e-05,
      "loss": 1.4254,
      "step": 44850
    },
    {
      "epoch": 10.71,
      "learning_rate": 1e-05,
      "loss": 1.4172,
      "step": 44860
    },
    {
      "epoch": 10.72,
      "learning_rate": 1e-05,
      "loss": 1.4216,
      "step": 44870
    },
    {
      "epoch": 10.72,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 44880
    },
    {
      "epoch": 10.72,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 44890
    },
    {
      "epoch": 10.72,
      "learning_rate": 1e-05,
      "loss": 1.4192,
      "step": 44900
    },
    {
      "epoch": 10.72,
      "eval_accuracy": 0.9868421052631579,
      "eval_loss": 1.418502688407898,
      "eval_runtime": 32.6784,
      "eval_samples_per_second": 20.931,
      "step": 44900
    },
    {
      "epoch": 10.73,
      "learning_rate": 1e-05,
      "loss": 1.4235,
      "step": 44910
    },
    {
      "epoch": 10.73,
      "learning_rate": 1e-05,
      "loss": 1.4207,
      "step": 44920
    },
    {
      "epoch": 10.73,
      "learning_rate": 1e-05,
      "loss": 1.4271,
      "step": 44930
    },
    {
      "epoch": 10.73,
      "learning_rate": 1e-05,
      "loss": 1.423,
      "step": 44940
    },
    {
      "epoch": 10.74,
      "learning_rate": 1e-05,
      "loss": 1.4261,
      "step": 44950
    },
    {
      "epoch": 10.74,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 44960
    },
    {
      "epoch": 10.74,
      "learning_rate": 1e-05,
      "loss": 1.4251,
      "step": 44970
    },
    {
      "epoch": 10.74,
      "learning_rate": 1e-05,
      "loss": 1.4315,
      "step": 44980
    },
    {
      "epoch": 10.75,
      "learning_rate": 1e-05,
      "loss": 1.42,
      "step": 44990
    },
    {
      "epoch": 10.75,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 45000
    },
    {
      "epoch": 10.75,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.419242262840271,
      "eval_runtime": 32.8224,
      "eval_samples_per_second": 20.839,
      "step": 45000
    },
    {
      "epoch": 10.75,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 45010
    },
    {
      "epoch": 10.75,
      "learning_rate": 1e-05,
      "loss": 1.4199,
      "step": 45020
    },
    {
      "epoch": 10.75,
      "learning_rate": 1e-05,
      "loss": 1.4178,
      "step": 45030
    },
    {
      "epoch": 10.76,
      "learning_rate": 1e-05,
      "loss": 1.4199,
      "step": 45040
    },
    {
      "epoch": 10.76,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 45050
    },
    {
      "epoch": 10.76,
      "learning_rate": 1e-05,
      "loss": 1.4259,
      "step": 45060
    },
    {
      "epoch": 10.76,
      "learning_rate": 1e-05,
      "loss": 1.4237,
      "step": 45070
    },
    {
      "epoch": 10.77,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 45080
    },
    {
      "epoch": 10.77,
      "learning_rate": 1e-05,
      "loss": 1.4294,
      "step": 45090
    },
    {
      "epoch": 10.77,
      "learning_rate": 1e-05,
      "loss": 1.4256,
      "step": 45100
    },
    {
      "epoch": 10.77,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4186104536056519,
      "eval_runtime": 32.9549,
      "eval_samples_per_second": 20.756,
      "step": 45100
    },
    {
      "epoch": 10.77,
      "learning_rate": 1e-05,
      "loss": 1.4229,
      "step": 45110
    },
    {
      "epoch": 10.78,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 45120
    },
    {
      "epoch": 10.78,
      "learning_rate": 1e-05,
      "loss": 1.4207,
      "step": 45130
    },
    {
      "epoch": 10.78,
      "learning_rate": 1e-05,
      "loss": 1.4237,
      "step": 45140
    },
    {
      "epoch": 10.78,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 45150
    },
    {
      "epoch": 10.79,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 45160
    },
    {
      "epoch": 10.79,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 45170
    },
    {
      "epoch": 10.79,
      "learning_rate": 1e-05,
      "loss": 1.4186,
      "step": 45180
    },
    {
      "epoch": 10.79,
      "learning_rate": 1e-05,
      "loss": 1.4186,
      "step": 45190
    },
    {
      "epoch": 10.8,
      "learning_rate": 1e-05,
      "loss": 1.4174,
      "step": 45200
    },
    {
      "epoch": 10.8,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4196614027023315,
      "eval_runtime": 32.942,
      "eval_samples_per_second": 20.764,
      "step": 45200
    },
    {
      "epoch": 10.8,
      "learning_rate": 1e-05,
      "loss": 1.4241,
      "step": 45210
    },
    {
      "epoch": 10.8,
      "learning_rate": 1e-05,
      "loss": 1.4165,
      "step": 45220
    },
    {
      "epoch": 10.8,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 45230
    },
    {
      "epoch": 10.8,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 45240
    },
    {
      "epoch": 10.81,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 45250
    },
    {
      "epoch": 10.81,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 45260
    },
    {
      "epoch": 10.81,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 45270
    },
    {
      "epoch": 10.81,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 45280
    },
    {
      "epoch": 10.82,
      "learning_rate": 1e-05,
      "loss": 1.4214,
      "step": 45290
    },
    {
      "epoch": 10.82,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 45300
    },
    {
      "epoch": 10.82,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4195103645324707,
      "eval_runtime": 32.947,
      "eval_samples_per_second": 20.761,
      "step": 45300
    },
    {
      "epoch": 10.82,
      "learning_rate": 1e-05,
      "loss": 1.4229,
      "step": 45310
    },
    {
      "epoch": 10.82,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 45320
    },
    {
      "epoch": 10.83,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 45330
    },
    {
      "epoch": 10.83,
      "learning_rate": 1e-05,
      "loss": 1.4187,
      "step": 45340
    },
    {
      "epoch": 10.83,
      "learning_rate": 1e-05,
      "loss": 1.4155,
      "step": 45350
    },
    {
      "epoch": 10.83,
      "learning_rate": 1e-05,
      "loss": 1.4217,
      "step": 45360
    },
    {
      "epoch": 10.84,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 45370
    },
    {
      "epoch": 10.84,
      "learning_rate": 1e-05,
      "loss": 1.4165,
      "step": 45380
    },
    {
      "epoch": 10.84,
      "learning_rate": 1e-05,
      "loss": 1.4214,
      "step": 45390
    },
    {
      "epoch": 10.84,
      "learning_rate": 1e-05,
      "loss": 1.4174,
      "step": 45400
    },
    {
      "epoch": 10.84,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4201809167861938,
      "eval_runtime": 33.1764,
      "eval_samples_per_second": 20.617,
      "step": 45400
    },
    {
      "epoch": 10.85,
      "learning_rate": 1e-05,
      "loss": 1.4254,
      "step": 45410
    },
    {
      "epoch": 10.85,
      "learning_rate": 1e-05,
      "loss": 1.4208,
      "step": 45420
    },
    {
      "epoch": 10.85,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 45430
    },
    {
      "epoch": 10.85,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 45440
    },
    {
      "epoch": 10.86,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 45450
    },
    {
      "epoch": 10.86,
      "learning_rate": 1e-05,
      "loss": 1.4209,
      "step": 45460
    },
    {
      "epoch": 10.86,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 45470
    },
    {
      "epoch": 10.86,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 45480
    },
    {
      "epoch": 10.86,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 45490
    },
    {
      "epoch": 10.87,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 45500
    },
    {
      "epoch": 10.87,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4177262783050537,
      "eval_runtime": 32.4393,
      "eval_samples_per_second": 21.086,
      "step": 45500
    },
    {
      "epoch": 10.87,
      "learning_rate": 1e-05,
      "loss": 1.4272,
      "step": 45510
    },
    {
      "epoch": 10.87,
      "learning_rate": 1e-05,
      "loss": 1.423,
      "step": 45520
    },
    {
      "epoch": 10.87,
      "learning_rate": 1e-05,
      "loss": 1.4225,
      "step": 45530
    },
    {
      "epoch": 10.88,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 45540
    },
    {
      "epoch": 10.88,
      "learning_rate": 1e-05,
      "loss": 1.4208,
      "step": 45550
    },
    {
      "epoch": 10.88,
      "learning_rate": 1e-05,
      "loss": 1.42,
      "step": 45560
    },
    {
      "epoch": 10.88,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 45570
    },
    {
      "epoch": 10.89,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 45580
    },
    {
      "epoch": 10.89,
      "learning_rate": 1e-05,
      "loss": 1.4212,
      "step": 45590
    },
    {
      "epoch": 10.89,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 45600
    },
    {
      "epoch": 10.89,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4183067083358765,
      "eval_runtime": 32.7152,
      "eval_samples_per_second": 20.908,
      "step": 45600
    },
    {
      "epoch": 10.89,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 45610
    },
    {
      "epoch": 10.9,
      "learning_rate": 1e-05,
      "loss": 1.4195,
      "step": 45620
    },
    {
      "epoch": 10.9,
      "learning_rate": 1e-05,
      "loss": 1.4172,
      "step": 45630
    },
    {
      "epoch": 10.9,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 45640
    },
    {
      "epoch": 10.9,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 45650
    },
    {
      "epoch": 10.91,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 45660
    },
    {
      "epoch": 10.91,
      "learning_rate": 1e-05,
      "loss": 1.4244,
      "step": 45670
    },
    {
      "epoch": 10.91,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 45680
    },
    {
      "epoch": 10.91,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 45690
    },
    {
      "epoch": 10.91,
      "learning_rate": 1e-05,
      "loss": 1.4244,
      "step": 45700
    },
    {
      "epoch": 10.91,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4196683168411255,
      "eval_runtime": 32.6946,
      "eval_samples_per_second": 20.921,
      "step": 45700
    },
    {
      "epoch": 10.92,
      "learning_rate": 1e-05,
      "loss": 1.4166,
      "step": 45710
    },
    {
      "epoch": 10.92,
      "learning_rate": 1e-05,
      "loss": 1.4195,
      "step": 45720
    },
    {
      "epoch": 10.92,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 45730
    },
    {
      "epoch": 10.92,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 45740
    },
    {
      "epoch": 10.93,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 45750
    },
    {
      "epoch": 10.93,
      "learning_rate": 1e-05,
      "loss": 1.4221,
      "step": 45760
    },
    {
      "epoch": 10.93,
      "learning_rate": 1e-05,
      "loss": 1.4676,
      "step": 45770
    },
    {
      "epoch": 10.93,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 45780
    },
    {
      "epoch": 10.94,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 45790
    },
    {
      "epoch": 10.94,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 45800
    },
    {
      "epoch": 10.94,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4188369512557983,
      "eval_runtime": 32.7582,
      "eval_samples_per_second": 20.88,
      "step": 45800
    },
    {
      "epoch": 10.94,
      "learning_rate": 1e-05,
      "loss": 1.4181,
      "step": 45810
    },
    {
      "epoch": 10.94,
      "learning_rate": 1e-05,
      "loss": 1.4186,
      "step": 45820
    },
    {
      "epoch": 10.95,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 45830
    },
    {
      "epoch": 10.95,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 45840
    },
    {
      "epoch": 10.95,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 45850
    },
    {
      "epoch": 10.95,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 45860
    },
    {
      "epoch": 10.96,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 45870
    },
    {
      "epoch": 10.96,
      "learning_rate": 1e-05,
      "loss": 1.4244,
      "step": 45880
    },
    {
      "epoch": 10.96,
      "learning_rate": 1e-05,
      "loss": 1.4224,
      "step": 45890
    },
    {
      "epoch": 10.96,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 45900
    },
    {
      "epoch": 10.96,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4190343618392944,
      "eval_runtime": 32.7016,
      "eval_samples_per_second": 20.916,
      "step": 45900
    },
    {
      "epoch": 10.96,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 45910
    },
    {
      "epoch": 10.97,
      "learning_rate": 1e-05,
      "loss": 1.4199,
      "step": 45920
    },
    {
      "epoch": 10.97,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 45930
    },
    {
      "epoch": 10.97,
      "learning_rate": 1e-05,
      "loss": 1.4256,
      "step": 45940
    },
    {
      "epoch": 10.97,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 45950
    },
    {
      "epoch": 10.98,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 45960
    },
    {
      "epoch": 10.98,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 45970
    },
    {
      "epoch": 10.98,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 45980
    },
    {
      "epoch": 10.98,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 45990
    },
    {
      "epoch": 10.99,
      "learning_rate": 1e-05,
      "loss": 1.4246,
      "step": 46000
    },
    {
      "epoch": 10.99,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4180179834365845,
      "eval_runtime": 32.6446,
      "eval_samples_per_second": 20.953,
      "step": 46000
    },
    {
      "epoch": 10.99,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 46010
    },
    {
      "epoch": 10.99,
      "learning_rate": 1e-05,
      "loss": 1.4218,
      "step": 46020
    },
    {
      "epoch": 10.99,
      "learning_rate": 1e-05,
      "loss": 1.4199,
      "step": 46030
    },
    {
      "epoch": 11.0,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 46040
    },
    {
      "epoch": 11.0,
      "learning_rate": 1e-05,
      "loss": 1.4309,
      "step": 46050
    },
    {
      "epoch": 11.0,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 46060
    },
    {
      "epoch": 11.0,
      "learning_rate": 1e-05,
      "loss": 1.42,
      "step": 46070
    },
    {
      "epoch": 11.01,
      "learning_rate": 1e-05,
      "loss": 1.4209,
      "step": 46080
    },
    {
      "epoch": 11.01,
      "learning_rate": 1e-05,
      "loss": 1.4191,
      "step": 46090
    },
    {
      "epoch": 11.01,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 46100
    },
    {
      "epoch": 11.01,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4179508686065674,
      "eval_runtime": 32.8316,
      "eval_samples_per_second": 20.834,
      "step": 46100
    },
    {
      "epoch": 11.01,
      "learning_rate": 1e-05,
      "loss": 1.4157,
      "step": 46110
    },
    {
      "epoch": 11.02,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 46120
    },
    {
      "epoch": 11.02,
      "learning_rate": 1e-05,
      "loss": 1.4191,
      "step": 46130
    },
    {
      "epoch": 11.02,
      "learning_rate": 1e-05,
      "loss": 1.4212,
      "step": 46140
    },
    {
      "epoch": 11.02,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 46150
    },
    {
      "epoch": 11.02,
      "learning_rate": 1e-05,
      "loss": 1.4201,
      "step": 46160
    },
    {
      "epoch": 11.03,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 46170
    },
    {
      "epoch": 11.03,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 46180
    },
    {
      "epoch": 11.03,
      "learning_rate": 1e-05,
      "loss": 1.4191,
      "step": 46190
    },
    {
      "epoch": 11.03,
      "learning_rate": 1e-05,
      "loss": 1.4234,
      "step": 46200
    },
    {
      "epoch": 11.03,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4179903268814087,
      "eval_runtime": 32.8504,
      "eval_samples_per_second": 20.822,
      "step": 46200
    },
    {
      "epoch": 11.04,
      "learning_rate": 1e-05,
      "loss": 1.42,
      "step": 46210
    },
    {
      "epoch": 11.04,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 46220
    },
    {
      "epoch": 11.04,
      "learning_rate": 1e-05,
      "loss": 1.4181,
      "step": 46230
    },
    {
      "epoch": 11.04,
      "learning_rate": 1e-05,
      "loss": 1.4172,
      "step": 46240
    },
    {
      "epoch": 11.05,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 46250
    },
    {
      "epoch": 11.05,
      "learning_rate": 1e-05,
      "loss": 1.4199,
      "step": 46260
    },
    {
      "epoch": 11.05,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 46270
    },
    {
      "epoch": 11.05,
      "learning_rate": 1e-05,
      "loss": 1.4249,
      "step": 46280
    },
    {
      "epoch": 11.06,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 46290
    },
    {
      "epoch": 11.06,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 46300
    },
    {
      "epoch": 11.06,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.417241096496582,
      "eval_runtime": 32.6972,
      "eval_samples_per_second": 20.919,
      "step": 46300
    },
    {
      "epoch": 11.06,
      "learning_rate": 1e-05,
      "loss": 1.4235,
      "step": 46310
    },
    {
      "epoch": 11.06,
      "learning_rate": 1e-05,
      "loss": 1.4257,
      "step": 46320
    },
    {
      "epoch": 11.07,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 46330
    },
    {
      "epoch": 11.07,
      "learning_rate": 1e-05,
      "loss": 1.4207,
      "step": 46340
    },
    {
      "epoch": 11.07,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 46350
    },
    {
      "epoch": 11.07,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 46360
    },
    {
      "epoch": 11.07,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 46370
    },
    {
      "epoch": 11.08,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 46380
    },
    {
      "epoch": 11.08,
      "learning_rate": 1e-05,
      "loss": 1.4245,
      "step": 46390
    },
    {
      "epoch": 11.08,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 46400
    },
    {
      "epoch": 11.08,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4187724590301514,
      "eval_runtime": 32.7349,
      "eval_samples_per_second": 20.895,
      "step": 46400
    },
    {
      "epoch": 11.08,
      "learning_rate": 1e-05,
      "loss": 1.4231,
      "step": 46410
    },
    {
      "epoch": 11.09,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 46420
    },
    {
      "epoch": 11.09,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 46430
    },
    {
      "epoch": 11.09,
      "learning_rate": 1e-05,
      "loss": 1.4186,
      "step": 46440
    },
    {
      "epoch": 11.09,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 46450
    },
    {
      "epoch": 11.1,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 46460
    },
    {
      "epoch": 11.1,
      "learning_rate": 1e-05,
      "loss": 1.4233,
      "step": 46470
    },
    {
      "epoch": 11.1,
      "learning_rate": 1e-05,
      "loss": 1.4236,
      "step": 46480
    },
    {
      "epoch": 11.1,
      "learning_rate": 1e-05,
      "loss": 1.425,
      "step": 46490
    },
    {
      "epoch": 11.11,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 46500
    },
    {
      "epoch": 11.11,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4178593158721924,
      "eval_runtime": 32.2563,
      "eval_samples_per_second": 21.205,
      "step": 46500
    },
    {
      "epoch": 11.11,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 46510
    },
    {
      "epoch": 11.11,
      "learning_rate": 1e-05,
      "loss": 1.4224,
      "step": 46520
    },
    {
      "epoch": 11.11,
      "learning_rate": 1e-05,
      "loss": 1.4225,
      "step": 46530
    },
    {
      "epoch": 11.12,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 46540
    },
    {
      "epoch": 11.12,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 46550
    },
    {
      "epoch": 11.12,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 46560
    },
    {
      "epoch": 11.12,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 46570
    },
    {
      "epoch": 11.12,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 46580
    },
    {
      "epoch": 11.13,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 46590
    },
    {
      "epoch": 11.13,
      "learning_rate": 1e-05,
      "loss": 1.42,
      "step": 46600
    },
    {
      "epoch": 11.13,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4178450107574463,
      "eval_runtime": 32.2948,
      "eval_samples_per_second": 21.18,
      "step": 46600
    },
    {
      "epoch": 11.13,
      "learning_rate": 1e-05,
      "loss": 1.4255,
      "step": 46610
    },
    {
      "epoch": 11.13,
      "learning_rate": 1e-05,
      "loss": 1.4298,
      "step": 46620
    },
    {
      "epoch": 11.14,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 46630
    },
    {
      "epoch": 11.14,
      "learning_rate": 1e-05,
      "loss": 1.4142,
      "step": 46640
    },
    {
      "epoch": 11.14,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 46650
    },
    {
      "epoch": 11.14,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 46660
    },
    {
      "epoch": 11.15,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 46670
    },
    {
      "epoch": 11.15,
      "learning_rate": 1e-05,
      "loss": 1.4223,
      "step": 46680
    },
    {
      "epoch": 11.15,
      "learning_rate": 1e-05,
      "loss": 1.4165,
      "step": 46690
    },
    {
      "epoch": 11.15,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 46700
    },
    {
      "epoch": 11.15,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4179433584213257,
      "eval_runtime": 32.6224,
      "eval_samples_per_second": 20.967,
      "step": 46700
    },
    {
      "epoch": 11.16,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 46710
    },
    {
      "epoch": 11.16,
      "learning_rate": 1e-05,
      "loss": 1.4227,
      "step": 46720
    },
    {
      "epoch": 11.16,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 46730
    },
    {
      "epoch": 11.16,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 46740
    },
    {
      "epoch": 11.17,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 46750
    },
    {
      "epoch": 11.17,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 46760
    },
    {
      "epoch": 11.17,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 46770
    },
    {
      "epoch": 11.17,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 46780
    },
    {
      "epoch": 11.18,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 46790
    },
    {
      "epoch": 11.18,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 46800
    },
    {
      "epoch": 11.18,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.417405128479004,
      "eval_runtime": 32.689,
      "eval_samples_per_second": 20.924,
      "step": 46800
    },
    {
      "epoch": 11.18,
      "learning_rate": 1e-05,
      "loss": 1.4217,
      "step": 46810
    },
    {
      "epoch": 11.18,
      "learning_rate": 1e-05,
      "loss": 1.4275,
      "step": 46820
    },
    {
      "epoch": 11.18,
      "learning_rate": 1e-05,
      "loss": 1.4187,
      "step": 46830
    },
    {
      "epoch": 11.19,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 46840
    },
    {
      "epoch": 11.19,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 46850
    },
    {
      "epoch": 11.19,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 46860
    },
    {
      "epoch": 11.19,
      "learning_rate": 1e-05,
      "loss": 1.4255,
      "step": 46870
    },
    {
      "epoch": 11.2,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 46880
    },
    {
      "epoch": 11.2,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 46890
    },
    {
      "epoch": 11.2,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 46900
    },
    {
      "epoch": 11.2,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4173353910446167,
      "eval_runtime": 32.6356,
      "eval_samples_per_second": 20.959,
      "step": 46900
    },
    {
      "epoch": 11.2,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 46910
    },
    {
      "epoch": 11.21,
      "learning_rate": 1e-05,
      "loss": 1.4213,
      "step": 46920
    },
    {
      "epoch": 11.21,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 46930
    },
    {
      "epoch": 11.21,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 46940
    },
    {
      "epoch": 11.21,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 46950
    },
    {
      "epoch": 11.22,
      "learning_rate": 1e-05,
      "loss": 1.4176,
      "step": 46960
    },
    {
      "epoch": 11.22,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 46970
    },
    {
      "epoch": 11.22,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 46980
    },
    {
      "epoch": 11.22,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 46990
    },
    {
      "epoch": 11.23,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 47000
    },
    {
      "epoch": 11.23,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4175370931625366,
      "eval_runtime": 33.0299,
      "eval_samples_per_second": 20.709,
      "step": 47000
    },
    {
      "epoch": 11.23,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 47010
    },
    {
      "epoch": 11.23,
      "learning_rate": 1e-05,
      "loss": 1.4267,
      "step": 47020
    },
    {
      "epoch": 11.23,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 47030
    },
    {
      "epoch": 11.23,
      "learning_rate": 1e-05,
      "loss": 1.4181,
      "step": 47040
    },
    {
      "epoch": 11.24,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 47050
    },
    {
      "epoch": 11.24,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 47060
    },
    {
      "epoch": 11.24,
      "learning_rate": 1e-05,
      "loss": 1.4209,
      "step": 47070
    },
    {
      "epoch": 11.24,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 47080
    },
    {
      "epoch": 11.25,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 47090
    },
    {
      "epoch": 11.25,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 47100
    },
    {
      "epoch": 11.25,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4175645112991333,
      "eval_runtime": 32.4769,
      "eval_samples_per_second": 21.061,
      "step": 47100
    },
    {
      "epoch": 11.25,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 47110
    },
    {
      "epoch": 11.25,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 47120
    },
    {
      "epoch": 11.26,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 47130
    },
    {
      "epoch": 11.26,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 47140
    },
    {
      "epoch": 11.26,
      "learning_rate": 1e-05,
      "loss": 1.4227,
      "step": 47150
    },
    {
      "epoch": 11.26,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 47160
    },
    {
      "epoch": 11.27,
      "learning_rate": 1e-05,
      "loss": 1.4193,
      "step": 47170
    },
    {
      "epoch": 11.27,
      "learning_rate": 1e-05,
      "loss": 1.4209,
      "step": 47180
    },
    {
      "epoch": 11.27,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 47190
    },
    {
      "epoch": 11.27,
      "learning_rate": 1e-05,
      "loss": 1.4157,
      "step": 47200
    },
    {
      "epoch": 11.27,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4174624681472778,
      "eval_runtime": 32.6239,
      "eval_samples_per_second": 20.966,
      "step": 47200
    },
    {
      "epoch": 11.28,
      "learning_rate": 1e-05,
      "loss": 1.4214,
      "step": 47210
    },
    {
      "epoch": 11.28,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 47220
    },
    {
      "epoch": 11.28,
      "learning_rate": 1e-05,
      "loss": 1.4174,
      "step": 47230
    },
    {
      "epoch": 11.28,
      "learning_rate": 1e-05,
      "loss": 1.4234,
      "step": 47240
    },
    {
      "epoch": 11.28,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 47250
    },
    {
      "epoch": 11.29,
      "learning_rate": 1e-05,
      "loss": 1.4177,
      "step": 47260
    },
    {
      "epoch": 11.29,
      "learning_rate": 1e-05,
      "loss": 1.4191,
      "step": 47270
    },
    {
      "epoch": 11.29,
      "learning_rate": 1e-05,
      "loss": 1.4318,
      "step": 47280
    },
    {
      "epoch": 11.29,
      "learning_rate": 1e-05,
      "loss": 1.4199,
      "step": 47290
    },
    {
      "epoch": 11.3,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 47300
    },
    {
      "epoch": 11.3,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4175928831100464,
      "eval_runtime": 32.6319,
      "eval_samples_per_second": 20.961,
      "step": 47300
    },
    {
      "epoch": 11.3,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 47310
    },
    {
      "epoch": 11.3,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 47320
    },
    {
      "epoch": 11.3,
      "learning_rate": 1e-05,
      "loss": 1.422,
      "step": 47330
    },
    {
      "epoch": 11.31,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 47340
    },
    {
      "epoch": 11.31,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 47350
    },
    {
      "epoch": 11.31,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 47360
    },
    {
      "epoch": 11.31,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 47370
    },
    {
      "epoch": 11.32,
      "learning_rate": 1e-05,
      "loss": 1.4222,
      "step": 47380
    },
    {
      "epoch": 11.32,
      "learning_rate": 1e-05,
      "loss": 1.424,
      "step": 47390
    },
    {
      "epoch": 11.32,
      "learning_rate": 1e-05,
      "loss": 1.4177,
      "step": 47400
    },
    {
      "epoch": 11.32,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4180570840835571,
      "eval_runtime": 32.6732,
      "eval_samples_per_second": 20.935,
      "step": 47400
    },
    {
      "epoch": 11.32,
      "learning_rate": 1e-05,
      "loss": 1.4157,
      "step": 47410
    },
    {
      "epoch": 11.33,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 47420
    },
    {
      "epoch": 11.33,
      "learning_rate": 1e-05,
      "loss": 1.4213,
      "step": 47430
    },
    {
      "epoch": 11.33,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 47440
    },
    {
      "epoch": 11.33,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 47450
    },
    {
      "epoch": 11.34,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 47460
    },
    {
      "epoch": 11.34,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 47470
    },
    {
      "epoch": 11.34,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 47480
    },
    {
      "epoch": 11.34,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 47490
    },
    {
      "epoch": 11.34,
      "learning_rate": 1e-05,
      "loss": 1.4236,
      "step": 47500
    },
    {
      "epoch": 11.34,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.417212724685669,
      "eval_runtime": 32.5131,
      "eval_samples_per_second": 21.038,
      "step": 47500
    },
    {
      "epoch": 11.35,
      "learning_rate": 1e-05,
      "loss": 1.4203,
      "step": 47510
    },
    {
      "epoch": 11.35,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 47520
    },
    {
      "epoch": 11.35,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 47530
    },
    {
      "epoch": 11.35,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 47540
    },
    {
      "epoch": 11.36,
      "learning_rate": 1e-05,
      "loss": 1.4177,
      "step": 47550
    },
    {
      "epoch": 11.36,
      "learning_rate": 1e-05,
      "loss": 1.4217,
      "step": 47560
    },
    {
      "epoch": 11.36,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 47570
    },
    {
      "epoch": 11.36,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 47580
    },
    {
      "epoch": 11.37,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 47590
    },
    {
      "epoch": 11.37,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 47600
    },
    {
      "epoch": 11.37,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.416948676109314,
      "eval_runtime": 32.4877,
      "eval_samples_per_second": 21.054,
      "step": 47600
    },
    {
      "epoch": 11.37,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 47610
    },
    {
      "epoch": 11.37,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 47620
    },
    {
      "epoch": 11.38,
      "learning_rate": 1e-05,
      "loss": 1.4216,
      "step": 47630
    },
    {
      "epoch": 11.38,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 47640
    },
    {
      "epoch": 11.38,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 47650
    },
    {
      "epoch": 11.38,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 47660
    },
    {
      "epoch": 11.39,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 47670
    },
    {
      "epoch": 11.39,
      "learning_rate": 1e-05,
      "loss": 1.4232,
      "step": 47680
    },
    {
      "epoch": 11.39,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 47690
    },
    {
      "epoch": 11.39,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 47700
    },
    {
      "epoch": 11.39,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4173504114151,
      "eval_runtime": 32.5985,
      "eval_samples_per_second": 20.983,
      "step": 47700
    },
    {
      "epoch": 11.39,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 47710
    },
    {
      "epoch": 11.4,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 47720
    },
    {
      "epoch": 11.4,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 47730
    },
    {
      "epoch": 11.4,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 47740
    },
    {
      "epoch": 11.4,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 47750
    },
    {
      "epoch": 11.41,
      "learning_rate": 1e-05,
      "loss": 1.4213,
      "step": 47760
    },
    {
      "epoch": 11.41,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 47770
    },
    {
      "epoch": 11.41,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 47780
    },
    {
      "epoch": 11.41,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 47790
    },
    {
      "epoch": 11.42,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 47800
    },
    {
      "epoch": 11.42,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4174085855484009,
      "eval_runtime": 32.71,
      "eval_samples_per_second": 20.911,
      "step": 47800
    },
    {
      "epoch": 11.42,
      "learning_rate": 1e-05,
      "loss": 1.4217,
      "step": 47810
    },
    {
      "epoch": 11.42,
      "learning_rate": 1e-05,
      "loss": 1.4223,
      "step": 47820
    },
    {
      "epoch": 11.42,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 47830
    },
    {
      "epoch": 11.43,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 47840
    },
    {
      "epoch": 11.43,
      "learning_rate": 1e-05,
      "loss": 1.4177,
      "step": 47850
    },
    {
      "epoch": 11.43,
      "learning_rate": 1e-05,
      "loss": 1.4209,
      "step": 47860
    },
    {
      "epoch": 11.43,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 47870
    },
    {
      "epoch": 11.44,
      "learning_rate": 1e-05,
      "loss": 1.4231,
      "step": 47880
    },
    {
      "epoch": 11.44,
      "learning_rate": 1e-05,
      "loss": 1.4193,
      "step": 47890
    },
    {
      "epoch": 11.44,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 47900
    },
    {
      "epoch": 11.44,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4169846773147583,
      "eval_runtime": 32.7118,
      "eval_samples_per_second": 20.91,
      "step": 47900
    },
    {
      "epoch": 11.44,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 47910
    },
    {
      "epoch": 11.44,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 47920
    },
    {
      "epoch": 11.45,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 47930
    },
    {
      "epoch": 11.45,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 47940
    },
    {
      "epoch": 11.45,
      "learning_rate": 1e-05,
      "loss": 1.4192,
      "step": 47950
    },
    {
      "epoch": 11.45,
      "learning_rate": 1e-05,
      "loss": 1.4212,
      "step": 47960
    },
    {
      "epoch": 11.46,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 47970
    },
    {
      "epoch": 11.46,
      "learning_rate": 1e-05,
      "loss": 1.4193,
      "step": 47980
    },
    {
      "epoch": 11.46,
      "learning_rate": 1e-05,
      "loss": 1.4223,
      "step": 47990
    },
    {
      "epoch": 11.46,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 48000
    },
    {
      "epoch": 11.46,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4165723323822021,
      "eval_runtime": 32.4412,
      "eval_samples_per_second": 21.084,
      "step": 48000
    },
    {
      "epoch": 11.47,
      "learning_rate": 1e-05,
      "loss": 1.4187,
      "step": 48010
    },
    {
      "epoch": 11.47,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 48020
    },
    {
      "epoch": 11.47,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 48030
    },
    {
      "epoch": 11.47,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 48040
    },
    {
      "epoch": 11.48,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 48050
    },
    {
      "epoch": 11.48,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 48060
    },
    {
      "epoch": 11.48,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 48070
    },
    {
      "epoch": 11.48,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 48080
    },
    {
      "epoch": 11.49,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 48090
    },
    {
      "epoch": 11.49,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 48100
    },
    {
      "epoch": 11.49,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4163296222686768,
      "eval_runtime": 32.7143,
      "eval_samples_per_second": 20.908,
      "step": 48100
    },
    {
      "epoch": 11.49,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 48110
    },
    {
      "epoch": 11.49,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 48120
    },
    {
      "epoch": 11.5,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 48130
    },
    {
      "epoch": 11.5,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 48140
    },
    {
      "epoch": 11.5,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 48150
    },
    {
      "epoch": 11.5,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 48160
    },
    {
      "epoch": 11.5,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 48170
    },
    {
      "epoch": 11.51,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 48180
    },
    {
      "epoch": 11.51,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 48190
    },
    {
      "epoch": 11.51,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 48200
    },
    {
      "epoch": 11.51,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4170728921890259,
      "eval_runtime": 32.9054,
      "eval_samples_per_second": 20.787,
      "step": 48200
    },
    {
      "epoch": 11.51,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 48210
    },
    {
      "epoch": 11.52,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 48220
    },
    {
      "epoch": 11.52,
      "learning_rate": 1e-05,
      "loss": 1.4209,
      "step": 48230
    },
    {
      "epoch": 11.52,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 48240
    },
    {
      "epoch": 11.52,
      "learning_rate": 1e-05,
      "loss": 1.4157,
      "step": 48250
    },
    {
      "epoch": 11.53,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 48260
    },
    {
      "epoch": 11.53,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 48270
    },
    {
      "epoch": 11.53,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 48280
    },
    {
      "epoch": 11.53,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 48290
    },
    {
      "epoch": 11.54,
      "learning_rate": 1e-05,
      "loss": 1.4178,
      "step": 48300
    },
    {
      "epoch": 11.54,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4161524772644043,
      "eval_runtime": 33.5785,
      "eval_samples_per_second": 20.37,
      "step": 48300
    },
    {
      "epoch": 11.54,
      "learning_rate": 1e-05,
      "loss": 1.4218,
      "step": 48310
    },
    {
      "epoch": 11.54,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 48320
    },
    {
      "epoch": 11.54,
      "learning_rate": 1e-05,
      "loss": 1.4247,
      "step": 48330
    },
    {
      "epoch": 11.55,
      "learning_rate": 1e-05,
      "loss": 1.4191,
      "step": 48340
    },
    {
      "epoch": 11.55,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 48350
    },
    {
      "epoch": 11.55,
      "learning_rate": 1e-05,
      "loss": 1.4223,
      "step": 48360
    },
    {
      "epoch": 11.55,
      "learning_rate": 1e-05,
      "loss": 1.4177,
      "step": 48370
    },
    {
      "epoch": 11.55,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 48380
    },
    {
      "epoch": 11.56,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 48390
    },
    {
      "epoch": 11.56,
      "learning_rate": 1e-05,
      "loss": 1.4243,
      "step": 48400
    },
    {
      "epoch": 11.56,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4168245792388916,
      "eval_runtime": 32.4386,
      "eval_samples_per_second": 21.086,
      "step": 48400
    },
    {
      "epoch": 11.56,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 48410
    },
    {
      "epoch": 11.56,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 48420
    },
    {
      "epoch": 11.57,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 48430
    },
    {
      "epoch": 11.57,
      "learning_rate": 1e-05,
      "loss": 1.4181,
      "step": 48440
    },
    {
      "epoch": 11.57,
      "learning_rate": 1e-05,
      "loss": 1.4165,
      "step": 48450
    },
    {
      "epoch": 11.57,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 48460
    },
    {
      "epoch": 11.58,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 48470
    },
    {
      "epoch": 11.58,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 48480
    },
    {
      "epoch": 11.58,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 48490
    },
    {
      "epoch": 11.58,
      "learning_rate": 1e-05,
      "loss": 1.4214,
      "step": 48500
    },
    {
      "epoch": 11.58,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4175502061843872,
      "eval_runtime": 32.1582,
      "eval_samples_per_second": 21.27,
      "step": 48500
    },
    {
      "epoch": 11.59,
      "learning_rate": 1e-05,
      "loss": 1.4192,
      "step": 48510
    },
    {
      "epoch": 11.59,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 48520
    },
    {
      "epoch": 11.59,
      "learning_rate": 1e-05,
      "loss": 1.4246,
      "step": 48530
    },
    {
      "epoch": 11.59,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 48540
    },
    {
      "epoch": 11.6,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 48550
    },
    {
      "epoch": 11.6,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 48560
    },
    {
      "epoch": 11.6,
      "learning_rate": 1e-05,
      "loss": 1.4187,
      "step": 48570
    },
    {
      "epoch": 11.6,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 48580
    },
    {
      "epoch": 11.6,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 48590
    },
    {
      "epoch": 11.61,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 48600
    },
    {
      "epoch": 11.61,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4177041053771973,
      "eval_runtime": 32.6706,
      "eval_samples_per_second": 20.936,
      "step": 48600
    },
    {
      "epoch": 11.61,
      "learning_rate": 1e-05,
      "loss": 1.4239,
      "step": 48610
    },
    {
      "epoch": 11.61,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 48620
    },
    {
      "epoch": 11.61,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 48630
    },
    {
      "epoch": 11.62,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 48640
    },
    {
      "epoch": 11.62,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 48650
    },
    {
      "epoch": 11.62,
      "learning_rate": 1e-05,
      "loss": 1.4198,
      "step": 48660
    },
    {
      "epoch": 11.62,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 48670
    },
    {
      "epoch": 11.63,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 48680
    },
    {
      "epoch": 11.63,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 48690
    },
    {
      "epoch": 11.63,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 48700
    },
    {
      "epoch": 11.63,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.41704261302948,
      "eval_runtime": 33.1785,
      "eval_samples_per_second": 20.616,
      "step": 48700
    },
    {
      "epoch": 11.63,
      "learning_rate": 1e-05,
      "loss": 1.4203,
      "step": 48710
    },
    {
      "epoch": 11.64,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 48720
    },
    {
      "epoch": 11.64,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 48730
    },
    {
      "epoch": 11.64,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 48740
    },
    {
      "epoch": 11.64,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 48750
    },
    {
      "epoch": 11.65,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 48760
    },
    {
      "epoch": 11.65,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 48770
    },
    {
      "epoch": 11.65,
      "learning_rate": 1e-05,
      "loss": 1.4245,
      "step": 48780
    },
    {
      "epoch": 11.65,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 48790
    },
    {
      "epoch": 11.66,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 48800
    },
    {
      "epoch": 11.66,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.417678713798523,
      "eval_runtime": 32.6225,
      "eval_samples_per_second": 20.967,
      "step": 48800
    },
    {
      "epoch": 11.66,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 48810
    },
    {
      "epoch": 11.66,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 48820
    },
    {
      "epoch": 11.66,
      "learning_rate": 1e-05,
      "loss": 1.4178,
      "step": 48830
    },
    {
      "epoch": 11.66,
      "learning_rate": 1e-05,
      "loss": 1.4208,
      "step": 48840
    },
    {
      "epoch": 11.67,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 48850
    },
    {
      "epoch": 11.67,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 48860
    },
    {
      "epoch": 11.67,
      "learning_rate": 1e-05,
      "loss": 1.4181,
      "step": 48870
    },
    {
      "epoch": 11.67,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 48880
    },
    {
      "epoch": 11.68,
      "learning_rate": 1e-05,
      "loss": 1.4244,
      "step": 48890
    },
    {
      "epoch": 11.68,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 48900
    },
    {
      "epoch": 11.68,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.416495680809021,
      "eval_runtime": 32.7023,
      "eval_samples_per_second": 20.916,
      "step": 48900
    },
    {
      "epoch": 11.68,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 48910
    },
    {
      "epoch": 11.68,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 48920
    },
    {
      "epoch": 11.69,
      "learning_rate": 1e-05,
      "loss": 1.4255,
      "step": 48930
    },
    {
      "epoch": 11.69,
      "learning_rate": 1e-05,
      "loss": 1.4177,
      "step": 48940
    },
    {
      "epoch": 11.69,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 48950
    },
    {
      "epoch": 11.69,
      "learning_rate": 1e-05,
      "loss": 1.4203,
      "step": 48960
    },
    {
      "epoch": 11.7,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 48970
    },
    {
      "epoch": 11.7,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 48980
    },
    {
      "epoch": 11.7,
      "learning_rate": 1e-05,
      "loss": 1.4203,
      "step": 48990
    },
    {
      "epoch": 11.7,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 49000
    },
    {
      "epoch": 11.7,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4169459342956543,
      "eval_runtime": 32.5546,
      "eval_samples_per_second": 21.011,
      "step": 49000
    },
    {
      "epoch": 11.71,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 49010
    },
    {
      "epoch": 11.71,
      "learning_rate": 1e-05,
      "loss": 1.4336,
      "step": 49020
    },
    {
      "epoch": 11.71,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 49030
    },
    {
      "epoch": 11.71,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 49040
    },
    {
      "epoch": 11.71,
      "learning_rate": 1e-05,
      "loss": 1.423,
      "step": 49050
    },
    {
      "epoch": 11.72,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 49060
    },
    {
      "epoch": 11.72,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 49070
    },
    {
      "epoch": 11.72,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 49080
    },
    {
      "epoch": 11.72,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 49090
    },
    {
      "epoch": 11.73,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 49100
    },
    {
      "epoch": 11.73,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4164036512374878,
      "eval_runtime": 32.4479,
      "eval_samples_per_second": 21.08,
      "step": 49100
    },
    {
      "epoch": 11.73,
      "learning_rate": 1e-05,
      "loss": 1.4228,
      "step": 49110
    },
    {
      "epoch": 11.73,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 49120
    },
    {
      "epoch": 11.73,
      "learning_rate": 1e-05,
      "loss": 1.4313,
      "step": 49130
    },
    {
      "epoch": 11.74,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 49140
    },
    {
      "epoch": 11.74,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 49150
    },
    {
      "epoch": 11.74,
      "learning_rate": 1e-05,
      "loss": 1.4207,
      "step": 49160
    },
    {
      "epoch": 11.74,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 49170
    },
    {
      "epoch": 11.75,
      "learning_rate": 1e-05,
      "loss": 1.4157,
      "step": 49180
    },
    {
      "epoch": 11.75,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 49190
    },
    {
      "epoch": 11.75,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 49200
    },
    {
      "epoch": 11.75,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4158340692520142,
      "eval_runtime": 33.323,
      "eval_samples_per_second": 20.526,
      "step": 49200
    },
    {
      "epoch": 11.75,
      "learning_rate": 1e-05,
      "loss": 1.4201,
      "step": 49210
    },
    {
      "epoch": 11.76,
      "learning_rate": 1e-05,
      "loss": 1.4247,
      "step": 49220
    },
    {
      "epoch": 11.76,
      "learning_rate": 1e-05,
      "loss": 1.4195,
      "step": 49230
    },
    {
      "epoch": 11.76,
      "learning_rate": 1e-05,
      "loss": 1.4136,
      "step": 49240
    },
    {
      "epoch": 11.76,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 49250
    },
    {
      "epoch": 11.76,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 49260
    },
    {
      "epoch": 11.77,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 49270
    },
    {
      "epoch": 11.77,
      "learning_rate": 1e-05,
      "loss": 1.445,
      "step": 49280
    },
    {
      "epoch": 11.77,
      "learning_rate": 1e-05,
      "loss": 1.4186,
      "step": 49290
    },
    {
      "epoch": 11.77,
      "learning_rate": 1e-05,
      "loss": 1.4301,
      "step": 49300
    },
    {
      "epoch": 11.77,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4164360761642456,
      "eval_runtime": 32.7613,
      "eval_samples_per_second": 20.878,
      "step": 49300
    },
    {
      "epoch": 11.78,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 49310
    },
    {
      "epoch": 11.78,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 49320
    },
    {
      "epoch": 11.78,
      "learning_rate": 1e-05,
      "loss": 1.4193,
      "step": 49330
    },
    {
      "epoch": 11.78,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 49340
    },
    {
      "epoch": 11.79,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 49350
    },
    {
      "epoch": 11.79,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 49360
    },
    {
      "epoch": 11.79,
      "learning_rate": 1e-05,
      "loss": 1.4166,
      "step": 49370
    },
    {
      "epoch": 11.79,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 49380
    },
    {
      "epoch": 11.8,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 49390
    },
    {
      "epoch": 11.8,
      "learning_rate": 1e-05,
      "loss": 1.4178,
      "step": 49400
    },
    {
      "epoch": 11.8,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4154199361801147,
      "eval_runtime": 32.4075,
      "eval_samples_per_second": 21.106,
      "step": 49400
    },
    {
      "epoch": 11.8,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 49410
    },
    {
      "epoch": 11.8,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 49420
    },
    {
      "epoch": 11.81,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 49430
    },
    {
      "epoch": 11.81,
      "learning_rate": 1e-05,
      "loss": 1.4237,
      "step": 49440
    },
    {
      "epoch": 11.81,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 49450
    },
    {
      "epoch": 11.81,
      "learning_rate": 1e-05,
      "loss": 1.4192,
      "step": 49460
    },
    {
      "epoch": 11.82,
      "learning_rate": 1e-05,
      "loss": 1.4203,
      "step": 49470
    },
    {
      "epoch": 11.82,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 49480
    },
    {
      "epoch": 11.82,
      "learning_rate": 1e-05,
      "loss": 1.42,
      "step": 49490
    },
    {
      "epoch": 11.82,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 49500
    },
    {
      "epoch": 11.82,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4154736995697021,
      "eval_runtime": 33.0086,
      "eval_samples_per_second": 20.722,
      "step": 49500
    },
    {
      "epoch": 11.82,
      "learning_rate": 1e-05,
      "loss": 1.4147,
      "step": 49510
    },
    {
      "epoch": 11.83,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 49520
    },
    {
      "epoch": 11.83,
      "learning_rate": 1e-05,
      "loss": 1.423,
      "step": 49530
    },
    {
      "epoch": 11.83,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 49540
    },
    {
      "epoch": 11.83,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 49550
    },
    {
      "epoch": 11.84,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 49560
    },
    {
      "epoch": 11.84,
      "learning_rate": 1e-05,
      "loss": 1.4172,
      "step": 49570
    },
    {
      "epoch": 11.84,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 49580
    },
    {
      "epoch": 11.84,
      "learning_rate": 1e-05,
      "loss": 1.4255,
      "step": 49590
    },
    {
      "epoch": 11.85,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 49600
    },
    {
      "epoch": 11.85,
      "eval_accuracy": 0.9868421052631579,
      "eval_loss": 1.4154924154281616,
      "eval_runtime": 32.4219,
      "eval_samples_per_second": 21.097,
      "step": 49600
    },
    {
      "epoch": 11.85,
      "learning_rate": 1e-05,
      "loss": 1.4113,
      "step": 49610
    },
    {
      "epoch": 11.85,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 49620
    },
    {
      "epoch": 11.85,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 49630
    },
    {
      "epoch": 11.86,
      "learning_rate": 1e-05,
      "loss": 1.422,
      "step": 49640
    },
    {
      "epoch": 11.86,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 49650
    },
    {
      "epoch": 11.86,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 49660
    },
    {
      "epoch": 11.86,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 49670
    },
    {
      "epoch": 11.87,
      "learning_rate": 1e-05,
      "loss": 1.4235,
      "step": 49680
    },
    {
      "epoch": 11.87,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 49690
    },
    {
      "epoch": 11.87,
      "learning_rate": 1e-05,
      "loss": 1.4176,
      "step": 49700
    },
    {
      "epoch": 11.87,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4161971807479858,
      "eval_runtime": 32.3217,
      "eval_samples_per_second": 21.162,
      "step": 49700
    },
    {
      "epoch": 11.87,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 49710
    },
    {
      "epoch": 11.87,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 49720
    },
    {
      "epoch": 11.88,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 49730
    },
    {
      "epoch": 11.88,
      "learning_rate": 1e-05,
      "loss": 1.4235,
      "step": 49740
    },
    {
      "epoch": 11.88,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 49750
    },
    {
      "epoch": 11.88,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 49760
    },
    {
      "epoch": 11.89,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 49770
    },
    {
      "epoch": 11.89,
      "learning_rate": 1e-05,
      "loss": 1.4187,
      "step": 49780
    },
    {
      "epoch": 11.89,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 49790
    },
    {
      "epoch": 11.89,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 49800
    },
    {
      "epoch": 11.89,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.416424036026001,
      "eval_runtime": 32.4828,
      "eval_samples_per_second": 21.057,
      "step": 49800
    },
    {
      "epoch": 11.9,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 49810
    },
    {
      "epoch": 11.9,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 49820
    },
    {
      "epoch": 11.9,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 49830
    },
    {
      "epoch": 11.9,
      "learning_rate": 1e-05,
      "loss": 1.4332,
      "step": 49840
    },
    {
      "epoch": 11.91,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 49850
    },
    {
      "epoch": 11.91,
      "learning_rate": 1e-05,
      "loss": 1.4187,
      "step": 49860
    },
    {
      "epoch": 11.91,
      "learning_rate": 1e-05,
      "loss": 1.4142,
      "step": 49870
    },
    {
      "epoch": 11.91,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 49880
    },
    {
      "epoch": 11.92,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 49890
    },
    {
      "epoch": 11.92,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 49900
    },
    {
      "epoch": 11.92,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4171574115753174,
      "eval_runtime": 32.7731,
      "eval_samples_per_second": 20.871,
      "step": 49900
    },
    {
      "epoch": 11.92,
      "learning_rate": 1e-05,
      "loss": 1.4201,
      "step": 49910
    },
    {
      "epoch": 11.92,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 49920
    },
    {
      "epoch": 11.93,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 49930
    },
    {
      "epoch": 11.93,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 49940
    },
    {
      "epoch": 11.93,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 49950
    },
    {
      "epoch": 11.93,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 49960
    },
    {
      "epoch": 11.93,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 49970
    },
    {
      "epoch": 11.94,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 49980
    },
    {
      "epoch": 11.94,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 49990
    },
    {
      "epoch": 11.94,
      "learning_rate": 1e-05,
      "loss": 1.4142,
      "step": 50000
    },
    {
      "epoch": 11.94,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.415612816810608,
      "eval_runtime": 34.5911,
      "eval_samples_per_second": 19.774,
      "step": 50000
    },
    {
      "epoch": 11.94,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 50010
    },
    {
      "epoch": 11.95,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 50020
    },
    {
      "epoch": 11.95,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 50030
    },
    {
      "epoch": 11.95,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 50040
    },
    {
      "epoch": 11.95,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 50050
    },
    {
      "epoch": 11.96,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 50060
    },
    {
      "epoch": 11.96,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 50070
    },
    {
      "epoch": 11.96,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 50080
    },
    {
      "epoch": 11.96,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 50090
    },
    {
      "epoch": 11.97,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 50100
    },
    {
      "epoch": 11.97,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4166507720947266,
      "eval_runtime": 32.803,
      "eval_samples_per_second": 20.852,
      "step": 50100
    },
    {
      "epoch": 11.97,
      "learning_rate": 1e-05,
      "loss": 1.4201,
      "step": 50110
    },
    {
      "epoch": 11.97,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 50120
    },
    {
      "epoch": 11.97,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 50130
    },
    {
      "epoch": 11.98,
      "learning_rate": 1e-05,
      "loss": 1.4136,
      "step": 50140
    },
    {
      "epoch": 11.98,
      "learning_rate": 1e-05,
      "loss": 1.4187,
      "step": 50150
    },
    {
      "epoch": 11.98,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 50160
    },
    {
      "epoch": 11.98,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 50170
    },
    {
      "epoch": 11.98,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 50180
    },
    {
      "epoch": 11.99,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 50190
    },
    {
      "epoch": 11.99,
      "learning_rate": 1e-05,
      "loss": 1.4174,
      "step": 50200
    },
    {
      "epoch": 11.99,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4160645008087158,
      "eval_runtime": 32.8484,
      "eval_samples_per_second": 20.823,
      "step": 50200
    },
    {
      "epoch": 11.99,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 50210
    },
    {
      "epoch": 11.99,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 50220
    },
    {
      "epoch": 12.0,
      "learning_rate": 1e-05,
      "loss": 1.4186,
      "step": 50230
    },
    {
      "epoch": 12.0,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 50240
    },
    {
      "epoch": 12.0,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 50250
    },
    {
      "epoch": 12.0,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 50260
    },
    {
      "epoch": 12.01,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 50270
    },
    {
      "epoch": 12.01,
      "learning_rate": 1e-05,
      "loss": 1.4207,
      "step": 50280
    },
    {
      "epoch": 12.01,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 50290
    },
    {
      "epoch": 12.01,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 50300
    },
    {
      "epoch": 12.01,
      "eval_accuracy": 0.9868421052631579,
      "eval_loss": 1.4168813228607178,
      "eval_runtime": 32.5128,
      "eval_samples_per_second": 21.038,
      "step": 50300
    },
    {
      "epoch": 12.02,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 50310
    },
    {
      "epoch": 12.02,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 50320
    },
    {
      "epoch": 12.02,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 50330
    },
    {
      "epoch": 12.02,
      "learning_rate": 1e-05,
      "loss": 1.4178,
      "step": 50340
    },
    {
      "epoch": 12.03,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 50350
    },
    {
      "epoch": 12.03,
      "learning_rate": 1e-05,
      "loss": 1.4157,
      "step": 50360
    },
    {
      "epoch": 12.03,
      "learning_rate": 1e-05,
      "loss": 1.4213,
      "step": 50370
    },
    {
      "epoch": 12.03,
      "learning_rate": 1e-05,
      "loss": 1.4211,
      "step": 50380
    },
    {
      "epoch": 12.03,
      "learning_rate": 1e-05,
      "loss": 1.4203,
      "step": 50390
    },
    {
      "epoch": 12.04,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 50400
    },
    {
      "epoch": 12.04,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4161556959152222,
      "eval_runtime": 32.9014,
      "eval_samples_per_second": 20.789,
      "step": 50400
    },
    {
      "epoch": 12.04,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 50410
    },
    {
      "epoch": 12.04,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 50420
    },
    {
      "epoch": 12.04,
      "learning_rate": 1e-05,
      "loss": 1.4192,
      "step": 50430
    },
    {
      "epoch": 12.05,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 50440
    },
    {
      "epoch": 12.05,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 50450
    },
    {
      "epoch": 12.05,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 50460
    },
    {
      "epoch": 12.05,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 50470
    },
    {
      "epoch": 12.06,
      "learning_rate": 1e-05,
      "loss": 1.4136,
      "step": 50480
    },
    {
      "epoch": 12.06,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 50490
    },
    {
      "epoch": 12.06,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 50500
    },
    {
      "epoch": 12.06,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4148224592208862,
      "eval_runtime": 32.5946,
      "eval_samples_per_second": 20.985,
      "step": 50500
    },
    {
      "epoch": 12.06,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 50510
    },
    {
      "epoch": 12.07,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 50520
    },
    {
      "epoch": 12.07,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 50530
    },
    {
      "epoch": 12.07,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 50540
    },
    {
      "epoch": 12.07,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 50550
    },
    {
      "epoch": 12.08,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 50560
    },
    {
      "epoch": 12.08,
      "learning_rate": 1e-05,
      "loss": 1.4374,
      "step": 50570
    },
    {
      "epoch": 12.08,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 50580
    },
    {
      "epoch": 12.08,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 50590
    },
    {
      "epoch": 12.09,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 50600
    },
    {
      "epoch": 12.09,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4149800539016724,
      "eval_runtime": 32.417,
      "eval_samples_per_second": 21.1,
      "step": 50600
    },
    {
      "epoch": 12.09,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 50610
    },
    {
      "epoch": 12.09,
      "learning_rate": 1e-05,
      "loss": 1.4178,
      "step": 50620
    },
    {
      "epoch": 12.09,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 50630
    },
    {
      "epoch": 12.09,
      "learning_rate": 1e-05,
      "loss": 1.4252,
      "step": 50640
    },
    {
      "epoch": 12.1,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 50650
    },
    {
      "epoch": 12.1,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 50660
    },
    {
      "epoch": 12.1,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 50670
    },
    {
      "epoch": 12.1,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 50680
    },
    {
      "epoch": 12.11,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 50690
    },
    {
      "epoch": 12.11,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 50700
    },
    {
      "epoch": 12.11,
      "eval_accuracy": 0.9868421052631579,
      "eval_loss": 1.4151371717453003,
      "eval_runtime": 32.6541,
      "eval_samples_per_second": 20.947,
      "step": 50700
    },
    {
      "epoch": 12.11,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 50710
    },
    {
      "epoch": 12.11,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 50720
    },
    {
      "epoch": 12.12,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 50730
    },
    {
      "epoch": 12.12,
      "learning_rate": 1e-05,
      "loss": 1.4181,
      "step": 50740
    },
    {
      "epoch": 12.12,
      "learning_rate": 1e-05,
      "loss": 1.4192,
      "step": 50750
    },
    {
      "epoch": 12.12,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 50760
    },
    {
      "epoch": 12.13,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 50770
    },
    {
      "epoch": 12.13,
      "learning_rate": 1e-05,
      "loss": 1.4236,
      "step": 50780
    },
    {
      "epoch": 12.13,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 50790
    },
    {
      "epoch": 12.13,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 50800
    },
    {
      "epoch": 12.13,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.415820598602295,
      "eval_runtime": 32.5451,
      "eval_samples_per_second": 21.017,
      "step": 50800
    },
    {
      "epoch": 12.14,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 50810
    },
    {
      "epoch": 12.14,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 50820
    },
    {
      "epoch": 12.14,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 50830
    },
    {
      "epoch": 12.14,
      "learning_rate": 1e-05,
      "loss": 1.4204,
      "step": 50840
    },
    {
      "epoch": 12.14,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 50850
    },
    {
      "epoch": 12.15,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 50860
    },
    {
      "epoch": 12.15,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 50870
    },
    {
      "epoch": 12.15,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 50880
    },
    {
      "epoch": 12.15,
      "learning_rate": 1e-05,
      "loss": 1.4214,
      "step": 50890
    },
    {
      "epoch": 12.16,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 50900
    },
    {
      "epoch": 12.16,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4155985116958618,
      "eval_runtime": 32.577,
      "eval_samples_per_second": 20.996,
      "step": 50900
    },
    {
      "epoch": 12.16,
      "learning_rate": 1e-05,
      "loss": 1.4291,
      "step": 50910
    },
    {
      "epoch": 12.16,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 50920
    },
    {
      "epoch": 12.16,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 50930
    },
    {
      "epoch": 12.17,
      "learning_rate": 1e-05,
      "loss": 1.4215,
      "step": 50940
    },
    {
      "epoch": 12.17,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 50950
    },
    {
      "epoch": 12.17,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 50960
    },
    {
      "epoch": 12.17,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 50970
    },
    {
      "epoch": 12.18,
      "learning_rate": 1e-05,
      "loss": 1.4147,
      "step": 50980
    },
    {
      "epoch": 12.18,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 50990
    },
    {
      "epoch": 12.18,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 51000
    },
    {
      "epoch": 12.18,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4157471656799316,
      "eval_runtime": 32.497,
      "eval_samples_per_second": 21.048,
      "step": 51000
    },
    {
      "epoch": 12.18,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 51010
    },
    {
      "epoch": 12.19,
      "learning_rate": 1e-05,
      "loss": 1.4198,
      "step": 51020
    },
    {
      "epoch": 12.19,
      "learning_rate": 1e-05,
      "loss": 1.4202,
      "step": 51030
    },
    {
      "epoch": 12.19,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 51040
    },
    {
      "epoch": 12.19,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 51050
    },
    {
      "epoch": 12.19,
      "learning_rate": 1e-05,
      "loss": 1.4172,
      "step": 51060
    },
    {
      "epoch": 12.2,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 51070
    },
    {
      "epoch": 12.2,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 51080
    },
    {
      "epoch": 12.2,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 51090
    },
    {
      "epoch": 12.2,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 51100
    },
    {
      "epoch": 12.2,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4140287637710571,
      "eval_runtime": 32.7327,
      "eval_samples_per_second": 20.897,
      "step": 51100
    },
    {
      "epoch": 12.21,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 51110
    },
    {
      "epoch": 12.21,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 51120
    },
    {
      "epoch": 12.21,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 51130
    },
    {
      "epoch": 12.21,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 51140
    },
    {
      "epoch": 12.22,
      "learning_rate": 1e-05,
      "loss": 1.4176,
      "step": 51150
    },
    {
      "epoch": 12.22,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 51160
    },
    {
      "epoch": 12.22,
      "learning_rate": 1e-05,
      "loss": 1.4166,
      "step": 51170
    },
    {
      "epoch": 12.22,
      "learning_rate": 1e-05,
      "loss": 1.4142,
      "step": 51180
    },
    {
      "epoch": 12.23,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 51190
    },
    {
      "epoch": 12.23,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 51200
    },
    {
      "epoch": 12.23,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4159153699874878,
      "eval_runtime": 32.6732,
      "eval_samples_per_second": 20.935,
      "step": 51200
    },
    {
      "epoch": 12.23,
      "learning_rate": 1e-05,
      "loss": 1.4136,
      "step": 51210
    },
    {
      "epoch": 12.23,
      "learning_rate": 1e-05,
      "loss": 1.4176,
      "step": 51220
    },
    {
      "epoch": 12.24,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 51230
    },
    {
      "epoch": 12.24,
      "learning_rate": 1e-05,
      "loss": 1.4181,
      "step": 51240
    },
    {
      "epoch": 12.24,
      "learning_rate": 1e-05,
      "loss": 1.4186,
      "step": 51250
    },
    {
      "epoch": 12.24,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 51260
    },
    {
      "epoch": 12.25,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 51270
    },
    {
      "epoch": 12.25,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 51280
    },
    {
      "epoch": 12.25,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 51290
    },
    {
      "epoch": 12.25,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 51300
    },
    {
      "epoch": 12.25,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4153108596801758,
      "eval_runtime": 32.1681,
      "eval_samples_per_second": 21.263,
      "step": 51300
    },
    {
      "epoch": 12.25,
      "learning_rate": 1e-05,
      "loss": 1.4191,
      "step": 51310
    },
    {
      "epoch": 12.26,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 51320
    },
    {
      "epoch": 12.26,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 51330
    },
    {
      "epoch": 12.26,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 51340
    },
    {
      "epoch": 12.26,
      "learning_rate": 1e-05,
      "loss": 1.4193,
      "step": 51350
    },
    {
      "epoch": 12.27,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 51360
    },
    {
      "epoch": 12.27,
      "learning_rate": 1e-05,
      "loss": 1.4113,
      "step": 51370
    },
    {
      "epoch": 12.27,
      "learning_rate": 1e-05,
      "loss": 1.4166,
      "step": 51380
    },
    {
      "epoch": 12.27,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 51390
    },
    {
      "epoch": 12.28,
      "learning_rate": 1e-05,
      "loss": 1.4213,
      "step": 51400
    },
    {
      "epoch": 12.28,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4152796268463135,
      "eval_runtime": 32.4019,
      "eval_samples_per_second": 21.11,
      "step": 51400
    },
    {
      "epoch": 12.28,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 51410
    },
    {
      "epoch": 12.28,
      "learning_rate": 1e-05,
      "loss": 1.4142,
      "step": 51420
    },
    {
      "epoch": 12.28,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 51430
    },
    {
      "epoch": 12.29,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 51440
    },
    {
      "epoch": 12.29,
      "learning_rate": 1e-05,
      "loss": 1.4181,
      "step": 51450
    },
    {
      "epoch": 12.29,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 51460
    },
    {
      "epoch": 12.29,
      "learning_rate": 1e-05,
      "loss": 1.4178,
      "step": 51470
    },
    {
      "epoch": 12.3,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 51480
    },
    {
      "epoch": 12.3,
      "learning_rate": 1e-05,
      "loss": 1.4227,
      "step": 51490
    },
    {
      "epoch": 12.3,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 51500
    },
    {
      "epoch": 12.3,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4151064157485962,
      "eval_runtime": 32.7857,
      "eval_samples_per_second": 20.863,
      "step": 51500
    },
    {
      "epoch": 12.3,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 51510
    },
    {
      "epoch": 12.3,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 51520
    },
    {
      "epoch": 12.31,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 51530
    },
    {
      "epoch": 12.31,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 51540
    },
    {
      "epoch": 12.31,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 51550
    },
    {
      "epoch": 12.31,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 51560
    },
    {
      "epoch": 12.32,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 51570
    },
    {
      "epoch": 12.32,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 51580
    },
    {
      "epoch": 12.32,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 51590
    },
    {
      "epoch": 12.32,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 51600
    },
    {
      "epoch": 12.32,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4156683683395386,
      "eval_runtime": 32.9694,
      "eval_samples_per_second": 20.746,
      "step": 51600
    },
    {
      "epoch": 12.33,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 51610
    },
    {
      "epoch": 12.33,
      "learning_rate": 1e-05,
      "loss": 1.4166,
      "step": 51620
    },
    {
      "epoch": 12.33,
      "learning_rate": 1e-05,
      "loss": 1.4157,
      "step": 51630
    },
    {
      "epoch": 12.33,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 51640
    },
    {
      "epoch": 12.34,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 51650
    },
    {
      "epoch": 12.34,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 51660
    },
    {
      "epoch": 12.34,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 51670
    },
    {
      "epoch": 12.34,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 51680
    },
    {
      "epoch": 12.35,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 51690
    },
    {
      "epoch": 12.35,
      "learning_rate": 1e-05,
      "loss": 1.4147,
      "step": 51700
    },
    {
      "epoch": 12.35,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.415432095527649,
      "eval_runtime": 32.6796,
      "eval_samples_per_second": 20.931,
      "step": 51700
    },
    {
      "epoch": 12.35,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 51710
    },
    {
      "epoch": 12.35,
      "learning_rate": 1e-05,
      "loss": 1.4137,
      "step": 51720
    },
    {
      "epoch": 12.35,
      "learning_rate": 1e-05,
      "loss": 1.42,
      "step": 51730
    },
    {
      "epoch": 12.36,
      "learning_rate": 1e-05,
      "loss": 1.4136,
      "step": 51740
    },
    {
      "epoch": 12.36,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 51750
    },
    {
      "epoch": 12.36,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 51760
    },
    {
      "epoch": 12.36,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 51770
    },
    {
      "epoch": 12.37,
      "learning_rate": 1e-05,
      "loss": 1.4177,
      "step": 51780
    },
    {
      "epoch": 12.37,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 51790
    },
    {
      "epoch": 12.37,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 51800
    },
    {
      "epoch": 12.37,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4165705442428589,
      "eval_runtime": 32.7278,
      "eval_samples_per_second": 20.9,
      "step": 51800
    },
    {
      "epoch": 12.37,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 51810
    },
    {
      "epoch": 12.38,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 51820
    },
    {
      "epoch": 12.38,
      "learning_rate": 1e-05,
      "loss": 1.4241,
      "step": 51830
    },
    {
      "epoch": 12.38,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 51840
    },
    {
      "epoch": 12.38,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 51850
    },
    {
      "epoch": 12.39,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 51860
    },
    {
      "epoch": 12.39,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 51870
    },
    {
      "epoch": 12.39,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 51880
    },
    {
      "epoch": 12.39,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 51890
    },
    {
      "epoch": 12.4,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 51900
    },
    {
      "epoch": 12.4,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.415600299835205,
      "eval_runtime": 32.5562,
      "eval_samples_per_second": 21.01,
      "step": 51900
    },
    {
      "epoch": 12.4,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 51910
    },
    {
      "epoch": 12.4,
      "learning_rate": 1e-05,
      "loss": 1.4157,
      "step": 51920
    },
    {
      "epoch": 12.4,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 51930
    },
    {
      "epoch": 12.41,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 51940
    },
    {
      "epoch": 12.41,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 51950
    },
    {
      "epoch": 12.41,
      "learning_rate": 1e-05,
      "loss": 1.4176,
      "step": 51960
    },
    {
      "epoch": 12.41,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 51970
    },
    {
      "epoch": 12.41,
      "learning_rate": 1e-05,
      "loss": 1.423,
      "step": 51980
    },
    {
      "epoch": 12.42,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 51990
    },
    {
      "epoch": 12.42,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 52000
    },
    {
      "epoch": 12.42,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4152588844299316,
      "eval_runtime": 32.5657,
      "eval_samples_per_second": 21.004,
      "step": 52000
    },
    {
      "epoch": 12.42,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 52010
    },
    {
      "epoch": 12.42,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 52020
    },
    {
      "epoch": 12.43,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 52030
    },
    {
      "epoch": 12.43,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 52040
    },
    {
      "epoch": 12.43,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 52050
    },
    {
      "epoch": 12.43,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 52060
    },
    {
      "epoch": 12.44,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 52070
    },
    {
      "epoch": 12.44,
      "learning_rate": 1e-05,
      "loss": 1.4147,
      "step": 52080
    },
    {
      "epoch": 12.44,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 52090
    },
    {
      "epoch": 12.44,
      "learning_rate": 1e-05,
      "loss": 1.4219,
      "step": 52100
    },
    {
      "epoch": 12.44,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4159847497940063,
      "eval_runtime": 32.1937,
      "eval_samples_per_second": 21.246,
      "step": 52100
    },
    {
      "epoch": 12.45,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 52110
    },
    {
      "epoch": 12.45,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 52120
    },
    {
      "epoch": 12.45,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 52130
    },
    {
      "epoch": 12.45,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 52140
    },
    {
      "epoch": 12.46,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 52150
    },
    {
      "epoch": 12.46,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 52160
    },
    {
      "epoch": 12.46,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 52170
    },
    {
      "epoch": 12.46,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 52180
    },
    {
      "epoch": 12.46,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 52190
    },
    {
      "epoch": 12.47,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 52200
    },
    {
      "epoch": 12.47,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4150850772857666,
      "eval_runtime": 32.6325,
      "eval_samples_per_second": 20.961,
      "step": 52200
    },
    {
      "epoch": 12.47,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 52210
    },
    {
      "epoch": 12.47,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 52220
    },
    {
      "epoch": 12.47,
      "learning_rate": 1e-05,
      "loss": 1.4155,
      "step": 52230
    },
    {
      "epoch": 12.48,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 52240
    },
    {
      "epoch": 12.48,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 52250
    },
    {
      "epoch": 12.48,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 52260
    },
    {
      "epoch": 12.48,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 52270
    },
    {
      "epoch": 12.49,
      "learning_rate": 1e-05,
      "loss": 1.4137,
      "step": 52280
    },
    {
      "epoch": 12.49,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 52290
    },
    {
      "epoch": 12.49,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 52300
    },
    {
      "epoch": 12.49,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4158320426940918,
      "eval_runtime": 32.2908,
      "eval_samples_per_second": 21.183,
      "step": 52300
    },
    {
      "epoch": 12.49,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 52310
    },
    {
      "epoch": 12.5,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 52320
    },
    {
      "epoch": 12.5,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 52330
    },
    {
      "epoch": 12.5,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 52340
    },
    {
      "epoch": 12.5,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 52350
    },
    {
      "epoch": 12.51,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 52360
    },
    {
      "epoch": 12.51,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 52370
    },
    {
      "epoch": 12.51,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 52380
    },
    {
      "epoch": 12.51,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 52390
    },
    {
      "epoch": 12.51,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 52400
    },
    {
      "epoch": 12.51,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.415065884590149,
      "eval_runtime": 32.32,
      "eval_samples_per_second": 21.163,
      "step": 52400
    },
    {
      "epoch": 12.52,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 52410
    },
    {
      "epoch": 12.52,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 52420
    },
    {
      "epoch": 12.52,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 52430
    },
    {
      "epoch": 12.52,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 52440
    },
    {
      "epoch": 12.53,
      "learning_rate": 1e-05,
      "loss": 1.4142,
      "step": 52450
    },
    {
      "epoch": 12.53,
      "learning_rate": 1e-05,
      "loss": 1.4223,
      "step": 52460
    },
    {
      "epoch": 12.53,
      "learning_rate": 1e-05,
      "loss": 1.4209,
      "step": 52470
    },
    {
      "epoch": 12.53,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 52480
    },
    {
      "epoch": 12.54,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 52490
    },
    {
      "epoch": 12.54,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 52500
    },
    {
      "epoch": 12.54,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4152445793151855,
      "eval_runtime": 33.0112,
      "eval_samples_per_second": 20.72,
      "step": 52500
    },
    {
      "epoch": 12.54,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 52510
    },
    {
      "epoch": 12.54,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 52520
    },
    {
      "epoch": 12.55,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 52530
    },
    {
      "epoch": 12.55,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 52540
    },
    {
      "epoch": 12.55,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 52550
    },
    {
      "epoch": 12.55,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 52560
    },
    {
      "epoch": 12.56,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 52570
    },
    {
      "epoch": 12.56,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 52580
    },
    {
      "epoch": 12.56,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 52590
    },
    {
      "epoch": 12.56,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 52600
    },
    {
      "epoch": 12.56,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.414404273033142,
      "eval_runtime": 32.6284,
      "eval_samples_per_second": 20.963,
      "step": 52600
    },
    {
      "epoch": 12.57,
      "learning_rate": 1e-05,
      "loss": 1.4113,
      "step": 52610
    },
    {
      "epoch": 12.57,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 52620
    },
    {
      "epoch": 12.57,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 52630
    },
    {
      "epoch": 12.57,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 52640
    },
    {
      "epoch": 12.57,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 52650
    },
    {
      "epoch": 12.58,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 52660
    },
    {
      "epoch": 12.58,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 52670
    },
    {
      "epoch": 12.58,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 52680
    },
    {
      "epoch": 12.58,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 52690
    },
    {
      "epoch": 12.59,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 52700
    },
    {
      "epoch": 12.59,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4150224924087524,
      "eval_runtime": 32.5782,
      "eval_samples_per_second": 20.996,
      "step": 52700
    },
    {
      "epoch": 12.59,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 52710
    },
    {
      "epoch": 12.59,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 52720
    },
    {
      "epoch": 12.59,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 52730
    },
    {
      "epoch": 12.6,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 52740
    },
    {
      "epoch": 12.6,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 52750
    },
    {
      "epoch": 12.6,
      "learning_rate": 1e-05,
      "loss": 1.4199,
      "step": 52760
    },
    {
      "epoch": 12.6,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 52770
    },
    {
      "epoch": 12.61,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 52780
    },
    {
      "epoch": 12.61,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 52790
    },
    {
      "epoch": 12.61,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 52800
    },
    {
      "epoch": 12.61,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4155086278915405,
      "eval_runtime": 32.5724,
      "eval_samples_per_second": 20.999,
      "step": 52800
    },
    {
      "epoch": 12.61,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 52810
    },
    {
      "epoch": 12.62,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 52820
    },
    {
      "epoch": 12.62,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 52830
    },
    {
      "epoch": 12.62,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 52840
    },
    {
      "epoch": 12.62,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 52850
    },
    {
      "epoch": 12.62,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 52860
    },
    {
      "epoch": 12.63,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 52870
    },
    {
      "epoch": 12.63,
      "learning_rate": 1e-05,
      "loss": 1.4201,
      "step": 52880
    },
    {
      "epoch": 12.63,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 52890
    },
    {
      "epoch": 12.63,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 52900
    },
    {
      "epoch": 12.63,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4155402183532715,
      "eval_runtime": 32.6272,
      "eval_samples_per_second": 20.964,
      "step": 52900
    },
    {
      "epoch": 12.64,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 52910
    },
    {
      "epoch": 12.64,
      "learning_rate": 1e-05,
      "loss": 1.4196,
      "step": 52920
    },
    {
      "epoch": 12.64,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 52930
    },
    {
      "epoch": 12.64,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 52940
    },
    {
      "epoch": 12.65,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 52950
    },
    {
      "epoch": 12.65,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 52960
    },
    {
      "epoch": 12.65,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 52970
    },
    {
      "epoch": 12.65,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 52980
    },
    {
      "epoch": 12.66,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 52990
    },
    {
      "epoch": 12.66,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 53000
    },
    {
      "epoch": 12.66,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4146751165390015,
      "eval_runtime": 33.9629,
      "eval_samples_per_second": 20.14,
      "step": 53000
    },
    {
      "epoch": 12.66,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 53010
    },
    {
      "epoch": 12.66,
      "learning_rate": 1e-05,
      "loss": 1.4155,
      "step": 53020
    },
    {
      "epoch": 12.67,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 53030
    },
    {
      "epoch": 12.67,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 53040
    },
    {
      "epoch": 12.67,
      "learning_rate": 1e-05,
      "loss": 1.4193,
      "step": 53050
    },
    {
      "epoch": 12.67,
      "learning_rate": 1e-05,
      "loss": 1.4191,
      "step": 53060
    },
    {
      "epoch": 12.67,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 53070
    },
    {
      "epoch": 12.68,
      "learning_rate": 1e-05,
      "loss": 1.4177,
      "step": 53080
    },
    {
      "epoch": 12.68,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 53090
    },
    {
      "epoch": 12.68,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 53100
    },
    {
      "epoch": 12.68,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4149746894836426,
      "eval_runtime": 32.6661,
      "eval_samples_per_second": 20.939,
      "step": 53100
    },
    {
      "epoch": 12.68,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 53110
    },
    {
      "epoch": 12.69,
      "learning_rate": 1e-05,
      "loss": 1.4176,
      "step": 53120
    },
    {
      "epoch": 12.69,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 53130
    },
    {
      "epoch": 12.69,
      "learning_rate": 1e-05,
      "loss": 1.4252,
      "step": 53140
    },
    {
      "epoch": 12.69,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 53150
    },
    {
      "epoch": 12.7,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 53160
    },
    {
      "epoch": 12.7,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 53170
    },
    {
      "epoch": 12.7,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 53180
    },
    {
      "epoch": 12.7,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 53190
    },
    {
      "epoch": 12.71,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 53200
    },
    {
      "epoch": 12.71,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4147149324417114,
      "eval_runtime": 32.4809,
      "eval_samples_per_second": 21.059,
      "step": 53200
    },
    {
      "epoch": 12.71,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 53210
    },
    {
      "epoch": 12.71,
      "learning_rate": 1e-05,
      "loss": 1.4195,
      "step": 53220
    },
    {
      "epoch": 12.71,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 53230
    },
    {
      "epoch": 12.72,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 53240
    },
    {
      "epoch": 12.72,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 53250
    },
    {
      "epoch": 12.72,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 53260
    },
    {
      "epoch": 12.72,
      "learning_rate": 1e-05,
      "loss": 1.413,
      "step": 53270
    },
    {
      "epoch": 12.73,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 53280
    },
    {
      "epoch": 12.73,
      "learning_rate": 1e-05,
      "loss": 1.4256,
      "step": 53290
    },
    {
      "epoch": 12.73,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 53300
    },
    {
      "epoch": 12.73,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4151628017425537,
      "eval_runtime": 32.5977,
      "eval_samples_per_second": 20.983,
      "step": 53300
    },
    {
      "epoch": 12.73,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 53310
    },
    {
      "epoch": 12.73,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 53320
    },
    {
      "epoch": 12.74,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 53330
    },
    {
      "epoch": 12.74,
      "learning_rate": 1e-05,
      "loss": 1.4147,
      "step": 53340
    },
    {
      "epoch": 12.74,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 53350
    },
    {
      "epoch": 12.74,
      "learning_rate": 1e-05,
      "loss": 1.4195,
      "step": 53360
    },
    {
      "epoch": 12.75,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 53370
    },
    {
      "epoch": 12.75,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 53380
    },
    {
      "epoch": 12.75,
      "learning_rate": 1e-05,
      "loss": 1.4174,
      "step": 53390
    },
    {
      "epoch": 12.75,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 53400
    },
    {
      "epoch": 12.75,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4145185947418213,
      "eval_runtime": 32.5769,
      "eval_samples_per_second": 20.996,
      "step": 53400
    },
    {
      "epoch": 12.76,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 53410
    },
    {
      "epoch": 12.76,
      "learning_rate": 1e-05,
      "loss": 1.4214,
      "step": 53420
    },
    {
      "epoch": 12.76,
      "learning_rate": 1e-05,
      "loss": 1.4157,
      "step": 53430
    },
    {
      "epoch": 12.76,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 53440
    },
    {
      "epoch": 12.77,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 53450
    },
    {
      "epoch": 12.77,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 53460
    },
    {
      "epoch": 12.77,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 53470
    },
    {
      "epoch": 12.77,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 53480
    },
    {
      "epoch": 12.78,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 53490
    },
    {
      "epoch": 12.78,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 53500
    },
    {
      "epoch": 12.78,
      "eval_accuracy": 0.9868421052631579,
      "eval_loss": 1.4147355556488037,
      "eval_runtime": 32.6641,
      "eval_samples_per_second": 20.94,
      "step": 53500
    },
    {
      "epoch": 12.78,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 53510
    },
    {
      "epoch": 12.78,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 53520
    },
    {
      "epoch": 12.78,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 53530
    },
    {
      "epoch": 12.79,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 53540
    },
    {
      "epoch": 12.79,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 53550
    },
    {
      "epoch": 12.79,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 53560
    },
    {
      "epoch": 12.79,
      "learning_rate": 1e-05,
      "loss": 1.4166,
      "step": 53570
    },
    {
      "epoch": 12.8,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 53580
    },
    {
      "epoch": 12.8,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 53590
    },
    {
      "epoch": 12.8,
      "learning_rate": 1e-05,
      "loss": 1.413,
      "step": 53600
    },
    {
      "epoch": 12.8,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4152791500091553,
      "eval_runtime": 32.6421,
      "eval_samples_per_second": 20.955,
      "step": 53600
    },
    {
      "epoch": 12.8,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 53610
    },
    {
      "epoch": 12.81,
      "learning_rate": 1e-05,
      "loss": 1.4178,
      "step": 53620
    },
    {
      "epoch": 12.81,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 53630
    },
    {
      "epoch": 12.81,
      "learning_rate": 1e-05,
      "loss": 1.4246,
      "step": 53640
    },
    {
      "epoch": 12.81,
      "learning_rate": 1e-05,
      "loss": 1.4155,
      "step": 53650
    },
    {
      "epoch": 12.82,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 53660
    },
    {
      "epoch": 12.82,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 53670
    },
    {
      "epoch": 12.82,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 53680
    },
    {
      "epoch": 12.82,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 53690
    },
    {
      "epoch": 12.83,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 53700
    },
    {
      "epoch": 12.83,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4143058061599731,
      "eval_runtime": 32.642,
      "eval_samples_per_second": 20.955,
      "step": 53700
    },
    {
      "epoch": 12.83,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 53710
    },
    {
      "epoch": 12.83,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 53720
    },
    {
      "epoch": 12.83,
      "learning_rate": 1e-05,
      "loss": 1.4192,
      "step": 53730
    },
    {
      "epoch": 12.83,
      "learning_rate": 1e-05,
      "loss": 1.4176,
      "step": 53740
    },
    {
      "epoch": 12.84,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 53750
    },
    {
      "epoch": 12.84,
      "learning_rate": 1e-05,
      "loss": 1.4113,
      "step": 53760
    },
    {
      "epoch": 12.84,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 53770
    },
    {
      "epoch": 12.84,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 53780
    },
    {
      "epoch": 12.85,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 53790
    },
    {
      "epoch": 12.85,
      "learning_rate": 1e-05,
      "loss": 1.4136,
      "step": 53800
    },
    {
      "epoch": 12.85,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4144463539123535,
      "eval_runtime": 32.7163,
      "eval_samples_per_second": 20.907,
      "step": 53800
    },
    {
      "epoch": 12.85,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 53810
    },
    {
      "epoch": 12.85,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 53820
    },
    {
      "epoch": 12.86,
      "learning_rate": 1e-05,
      "loss": 1.4441,
      "step": 53830
    },
    {
      "epoch": 12.86,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 53840
    },
    {
      "epoch": 12.86,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 53850
    },
    {
      "epoch": 12.86,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 53860
    },
    {
      "epoch": 12.87,
      "learning_rate": 1e-05,
      "loss": 1.4191,
      "step": 53870
    },
    {
      "epoch": 12.87,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 53880
    },
    {
      "epoch": 12.87,
      "learning_rate": 1e-05,
      "loss": 1.4346,
      "step": 53890
    },
    {
      "epoch": 12.87,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 53900
    },
    {
      "epoch": 12.87,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4151360988616943,
      "eval_runtime": 32.5483,
      "eval_samples_per_second": 21.015,
      "step": 53900
    },
    {
      "epoch": 12.88,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 53910
    },
    {
      "epoch": 12.88,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 53920
    },
    {
      "epoch": 12.88,
      "learning_rate": 1e-05,
      "loss": 1.4328,
      "step": 53930
    },
    {
      "epoch": 12.88,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 53940
    },
    {
      "epoch": 12.89,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 53950
    },
    {
      "epoch": 12.89,
      "learning_rate": 1e-05,
      "loss": 1.4208,
      "step": 53960
    },
    {
      "epoch": 12.89,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 53970
    },
    {
      "epoch": 12.89,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 53980
    },
    {
      "epoch": 12.89,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 53990
    },
    {
      "epoch": 12.9,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 54000
    },
    {
      "epoch": 12.9,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4138646125793457,
      "eval_runtime": 32.1513,
      "eval_samples_per_second": 21.274,
      "step": 54000
    },
    {
      "epoch": 12.9,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 54010
    },
    {
      "epoch": 12.9,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 54020
    },
    {
      "epoch": 12.9,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 54030
    },
    {
      "epoch": 12.91,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 54040
    },
    {
      "epoch": 12.91,
      "learning_rate": 1e-05,
      "loss": 1.4137,
      "step": 54050
    },
    {
      "epoch": 12.91,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 54060
    },
    {
      "epoch": 12.91,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 54070
    },
    {
      "epoch": 12.92,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 54080
    },
    {
      "epoch": 12.92,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 54090
    },
    {
      "epoch": 12.92,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 54100
    },
    {
      "epoch": 12.92,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4148656129837036,
      "eval_runtime": 32.2083,
      "eval_samples_per_second": 21.237,
      "step": 54100
    },
    {
      "epoch": 12.92,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 54110
    },
    {
      "epoch": 12.93,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 54120
    },
    {
      "epoch": 12.93,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 54130
    },
    {
      "epoch": 12.93,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 54140
    },
    {
      "epoch": 12.93,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 54150
    },
    {
      "epoch": 12.94,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 54160
    },
    {
      "epoch": 12.94,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 54170
    },
    {
      "epoch": 12.94,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 54180
    },
    {
      "epoch": 12.94,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 54190
    },
    {
      "epoch": 12.94,
      "learning_rate": 1e-05,
      "loss": 1.4157,
      "step": 54200
    },
    {
      "epoch": 12.94,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4133814573287964,
      "eval_runtime": 32.4093,
      "eval_samples_per_second": 21.105,
      "step": 54200
    },
    {
      "epoch": 12.95,
      "learning_rate": 1e-05,
      "loss": 1.4191,
      "step": 54210
    },
    {
      "epoch": 12.95,
      "learning_rate": 1e-05,
      "loss": 1.4234,
      "step": 54220
    },
    {
      "epoch": 12.95,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 54230
    },
    {
      "epoch": 12.95,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 54240
    },
    {
      "epoch": 12.96,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 54250
    },
    {
      "epoch": 12.96,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 54260
    },
    {
      "epoch": 12.96,
      "learning_rate": 1e-05,
      "loss": 1.4165,
      "step": 54270
    },
    {
      "epoch": 12.96,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 54280
    },
    {
      "epoch": 12.97,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 54290
    },
    {
      "epoch": 12.97,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 54300
    },
    {
      "epoch": 12.97,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.413649082183838,
      "eval_runtime": 32.8506,
      "eval_samples_per_second": 20.822,
      "step": 54300
    },
    {
      "epoch": 12.97,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 54310
    },
    {
      "epoch": 12.97,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 54320
    },
    {
      "epoch": 12.98,
      "learning_rate": 1e-05,
      "loss": 1.4172,
      "step": 54330
    },
    {
      "epoch": 12.98,
      "learning_rate": 1e-05,
      "loss": 1.4224,
      "step": 54340
    },
    {
      "epoch": 12.98,
      "learning_rate": 1e-05,
      "loss": 1.4176,
      "step": 54350
    },
    {
      "epoch": 12.98,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 54360
    },
    {
      "epoch": 12.99,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 54370
    },
    {
      "epoch": 12.99,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 54380
    },
    {
      "epoch": 12.99,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 54390
    },
    {
      "epoch": 12.99,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 54400
    },
    {
      "epoch": 12.99,
      "eval_accuracy": 0.9868421052631579,
      "eval_loss": 1.4142950773239136,
      "eval_runtime": 32.3649,
      "eval_samples_per_second": 21.134,
      "step": 54400
    },
    {
      "epoch": 12.99,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 54410
    },
    {
      "epoch": 13.0,
      "learning_rate": 1e-05,
      "loss": 1.4198,
      "step": 54420
    },
    {
      "epoch": 13.0,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 54430
    },
    {
      "epoch": 13.0,
      "learning_rate": 1e-05,
      "loss": 1.4181,
      "step": 54440
    },
    {
      "epoch": 13.0,
      "learning_rate": 1e-05,
      "loss": 1.4237,
      "step": 54450
    },
    {
      "epoch": 13.01,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 54460
    },
    {
      "epoch": 13.01,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 54470
    },
    {
      "epoch": 13.01,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 54480
    },
    {
      "epoch": 13.01,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 54490
    },
    {
      "epoch": 13.02,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 54500
    },
    {
      "epoch": 13.02,
      "eval_accuracy": 0.9868421052631579,
      "eval_loss": 1.4145641326904297,
      "eval_runtime": 32.6656,
      "eval_samples_per_second": 20.939,
      "step": 54500
    },
    {
      "epoch": 13.02,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 54510
    },
    {
      "epoch": 13.02,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 54520
    },
    {
      "epoch": 13.02,
      "learning_rate": 1e-05,
      "loss": 1.4166,
      "step": 54530
    },
    {
      "epoch": 13.03,
      "learning_rate": 1e-05,
      "loss": 1.4113,
      "step": 54540
    },
    {
      "epoch": 13.03,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 54550
    },
    {
      "epoch": 13.03,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 54560
    },
    {
      "epoch": 13.03,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 54570
    },
    {
      "epoch": 13.04,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 54580
    },
    {
      "epoch": 13.04,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 54590
    },
    {
      "epoch": 13.04,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 54600
    },
    {
      "epoch": 13.04,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4146991968154907,
      "eval_runtime": 32.4742,
      "eval_samples_per_second": 21.063,
      "step": 54600
    },
    {
      "epoch": 13.04,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 54610
    },
    {
      "epoch": 13.05,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 54620
    },
    {
      "epoch": 13.05,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 54630
    },
    {
      "epoch": 13.05,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 54640
    },
    {
      "epoch": 13.05,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 54650
    },
    {
      "epoch": 13.05,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 54660
    },
    {
      "epoch": 13.06,
      "learning_rate": 1e-05,
      "loss": 1.4174,
      "step": 54670
    },
    {
      "epoch": 13.06,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 54680
    },
    {
      "epoch": 13.06,
      "learning_rate": 1e-05,
      "loss": 1.4193,
      "step": 54690
    },
    {
      "epoch": 13.06,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 54700
    },
    {
      "epoch": 13.06,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4145737886428833,
      "eval_runtime": 32.4606,
      "eval_samples_per_second": 21.072,
      "step": 54700
    },
    {
      "epoch": 13.07,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 54710
    },
    {
      "epoch": 13.07,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 54720
    },
    {
      "epoch": 13.07,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 54730
    },
    {
      "epoch": 13.07,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 54740
    },
    {
      "epoch": 13.08,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 54750
    },
    {
      "epoch": 13.08,
      "learning_rate": 1e-05,
      "loss": 1.4136,
      "step": 54760
    },
    {
      "epoch": 13.08,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 54770
    },
    {
      "epoch": 13.08,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 54780
    },
    {
      "epoch": 13.09,
      "learning_rate": 1e-05,
      "loss": 1.4172,
      "step": 54790
    },
    {
      "epoch": 13.09,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 54800
    },
    {
      "epoch": 13.09,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4147101640701294,
      "eval_runtime": 32.6277,
      "eval_samples_per_second": 20.964,
      "step": 54800
    },
    {
      "epoch": 13.09,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 54810
    },
    {
      "epoch": 13.09,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 54820
    },
    {
      "epoch": 13.1,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 54830
    },
    {
      "epoch": 13.1,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 54840
    },
    {
      "epoch": 13.1,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 54850
    },
    {
      "epoch": 13.1,
      "learning_rate": 1e-05,
      "loss": 1.4242,
      "step": 54860
    },
    {
      "epoch": 13.1,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 54870
    },
    {
      "epoch": 13.11,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 54880
    },
    {
      "epoch": 13.11,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 54890
    },
    {
      "epoch": 13.11,
      "learning_rate": 1e-05,
      "loss": 1.4142,
      "step": 54900
    },
    {
      "epoch": 13.11,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4139537811279297,
      "eval_runtime": 32.4232,
      "eval_samples_per_second": 21.096,
      "step": 54900
    },
    {
      "epoch": 13.11,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 54910
    },
    {
      "epoch": 13.12,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 54920
    },
    {
      "epoch": 13.12,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 54930
    },
    {
      "epoch": 13.12,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 54940
    },
    {
      "epoch": 13.12,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 54950
    },
    {
      "epoch": 13.13,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 54960
    },
    {
      "epoch": 13.13,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 54970
    },
    {
      "epoch": 13.13,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 54980
    },
    {
      "epoch": 13.13,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 54990
    },
    {
      "epoch": 13.14,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 55000
    },
    {
      "epoch": 13.14,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.414209246635437,
      "eval_runtime": 32.3916,
      "eval_samples_per_second": 21.117,
      "step": 55000
    },
    {
      "epoch": 13.14,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 55010
    },
    {
      "epoch": 13.14,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 55020
    },
    {
      "epoch": 13.14,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 55030
    },
    {
      "epoch": 13.15,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 55040
    },
    {
      "epoch": 13.15,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 55050
    },
    {
      "epoch": 13.15,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 55060
    },
    {
      "epoch": 13.15,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 55070
    },
    {
      "epoch": 13.16,
      "learning_rate": 1e-05,
      "loss": 1.4136,
      "step": 55080
    },
    {
      "epoch": 13.16,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 55090
    },
    {
      "epoch": 13.16,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 55100
    },
    {
      "epoch": 13.16,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4134875535964966,
      "eval_runtime": 32.2618,
      "eval_samples_per_second": 21.202,
      "step": 55100
    },
    {
      "epoch": 13.16,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 55110
    },
    {
      "epoch": 13.16,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 55120
    },
    {
      "epoch": 13.17,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 55130
    },
    {
      "epoch": 13.17,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 55140
    },
    {
      "epoch": 13.17,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 55150
    },
    {
      "epoch": 13.17,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 55160
    },
    {
      "epoch": 13.18,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 55170
    },
    {
      "epoch": 13.18,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 55180
    },
    {
      "epoch": 13.18,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 55190
    },
    {
      "epoch": 13.18,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 55200
    },
    {
      "epoch": 13.18,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.412827491760254,
      "eval_runtime": 32.1614,
      "eval_samples_per_second": 21.268,
      "step": 55200
    },
    {
      "epoch": 13.19,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 55210
    },
    {
      "epoch": 13.19,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 55220
    },
    {
      "epoch": 13.19,
      "learning_rate": 1e-05,
      "loss": 1.4191,
      "step": 55230
    },
    {
      "epoch": 13.19,
      "learning_rate": 1e-05,
      "loss": 1.4201,
      "step": 55240
    },
    {
      "epoch": 13.2,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 55250
    },
    {
      "epoch": 13.2,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 55260
    },
    {
      "epoch": 13.2,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 55270
    },
    {
      "epoch": 13.2,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 55280
    },
    {
      "epoch": 13.21,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 55290
    },
    {
      "epoch": 13.21,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 55300
    },
    {
      "epoch": 13.21,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4142930507659912,
      "eval_runtime": 32.1504,
      "eval_samples_per_second": 21.275,
      "step": 55300
    },
    {
      "epoch": 13.21,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 55310
    },
    {
      "epoch": 13.21,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 55320
    },
    {
      "epoch": 13.21,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 55330
    },
    {
      "epoch": 13.22,
      "learning_rate": 1e-05,
      "loss": 1.4238,
      "step": 55340
    },
    {
      "epoch": 13.22,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 55350
    },
    {
      "epoch": 13.22,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 55360
    },
    {
      "epoch": 13.22,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 55370
    },
    {
      "epoch": 13.23,
      "learning_rate": 1e-05,
      "loss": 1.4225,
      "step": 55380
    },
    {
      "epoch": 13.23,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 55390
    },
    {
      "epoch": 13.23,
      "learning_rate": 1e-05,
      "loss": 1.4184,
      "step": 55400
    },
    {
      "epoch": 13.23,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4135112762451172,
      "eval_runtime": 32.5174,
      "eval_samples_per_second": 21.035,
      "step": 55400
    },
    {
      "epoch": 13.23,
      "learning_rate": 1e-05,
      "loss": 1.4197,
      "step": 55410
    },
    {
      "epoch": 13.24,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 55420
    },
    {
      "epoch": 13.24,
      "learning_rate": 1e-05,
      "loss": 1.4155,
      "step": 55430
    },
    {
      "epoch": 13.24,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 55440
    },
    {
      "epoch": 13.24,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 55450
    },
    {
      "epoch": 13.25,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 55460
    },
    {
      "epoch": 13.25,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 55470
    },
    {
      "epoch": 13.25,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 55480
    },
    {
      "epoch": 13.25,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 55490
    },
    {
      "epoch": 13.26,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 55500
    },
    {
      "epoch": 13.26,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4141385555267334,
      "eval_runtime": 32.5553,
      "eval_samples_per_second": 21.01,
      "step": 55500
    },
    {
      "epoch": 13.26,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 55510
    },
    {
      "epoch": 13.26,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 55520
    },
    {
      "epoch": 13.26,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 55530
    },
    {
      "epoch": 13.26,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 55540
    },
    {
      "epoch": 13.27,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 55550
    },
    {
      "epoch": 13.27,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 55560
    },
    {
      "epoch": 13.27,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 55570
    },
    {
      "epoch": 13.27,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 55580
    },
    {
      "epoch": 13.28,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 55590
    },
    {
      "epoch": 13.28,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 55600
    },
    {
      "epoch": 13.28,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4135775566101074,
      "eval_runtime": 32.5315,
      "eval_samples_per_second": 21.026,
      "step": 55600
    },
    {
      "epoch": 13.28,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 55610
    },
    {
      "epoch": 13.28,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 55620
    },
    {
      "epoch": 13.29,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 55630
    },
    {
      "epoch": 13.29,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 55640
    },
    {
      "epoch": 13.29,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 55650
    },
    {
      "epoch": 13.29,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 55660
    },
    {
      "epoch": 13.3,
      "learning_rate": 1e-05,
      "loss": 1.419,
      "step": 55670
    },
    {
      "epoch": 13.3,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 55680
    },
    {
      "epoch": 13.3,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 55690
    },
    {
      "epoch": 13.3,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 55700
    },
    {
      "epoch": 13.3,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4132215976715088,
      "eval_runtime": 32.5584,
      "eval_samples_per_second": 21.008,
      "step": 55700
    },
    {
      "epoch": 13.31,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 55710
    },
    {
      "epoch": 13.31,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 55720
    },
    {
      "epoch": 13.31,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 55730
    },
    {
      "epoch": 13.31,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 55740
    },
    {
      "epoch": 13.32,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 55750
    },
    {
      "epoch": 13.32,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 55760
    },
    {
      "epoch": 13.32,
      "learning_rate": 1e-05,
      "loss": 1.4136,
      "step": 55770
    },
    {
      "epoch": 13.32,
      "learning_rate": 1e-05,
      "loss": 1.4136,
      "step": 55780
    },
    {
      "epoch": 13.32,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 55790
    },
    {
      "epoch": 13.33,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 55800
    },
    {
      "epoch": 13.33,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4136059284210205,
      "eval_runtime": 32.5936,
      "eval_samples_per_second": 20.986,
      "step": 55800
    },
    {
      "epoch": 13.33,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 55810
    },
    {
      "epoch": 13.33,
      "learning_rate": 1e-05,
      "loss": 1.4165,
      "step": 55820
    },
    {
      "epoch": 13.33,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 55830
    },
    {
      "epoch": 13.34,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 55840
    },
    {
      "epoch": 13.34,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 55850
    },
    {
      "epoch": 13.34,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 55860
    },
    {
      "epoch": 13.34,
      "learning_rate": 1e-05,
      "loss": 1.4193,
      "step": 55870
    },
    {
      "epoch": 13.35,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 55880
    },
    {
      "epoch": 13.35,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 55890
    },
    {
      "epoch": 13.35,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 55900
    },
    {
      "epoch": 13.35,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4131513833999634,
      "eval_runtime": 32.5507,
      "eval_samples_per_second": 21.013,
      "step": 55900
    },
    {
      "epoch": 13.35,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 55910
    },
    {
      "epoch": 13.36,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 55920
    },
    {
      "epoch": 13.36,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 55930
    },
    {
      "epoch": 13.36,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 55940
    },
    {
      "epoch": 13.36,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 55950
    },
    {
      "epoch": 13.37,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 55960
    },
    {
      "epoch": 13.37,
      "learning_rate": 1e-05,
      "loss": 1.4349,
      "step": 55970
    },
    {
      "epoch": 13.37,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 55980
    },
    {
      "epoch": 13.37,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 55990
    },
    {
      "epoch": 13.37,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 56000
    },
    {
      "epoch": 13.37,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4128382205963135,
      "eval_runtime": 32.6403,
      "eval_samples_per_second": 20.956,
      "step": 56000
    },
    {
      "epoch": 13.38,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 56010
    },
    {
      "epoch": 13.38,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 56020
    },
    {
      "epoch": 13.38,
      "learning_rate": 1e-05,
      "loss": 1.4286,
      "step": 56030
    },
    {
      "epoch": 13.38,
      "learning_rate": 1e-05,
      "loss": 1.4195,
      "step": 56040
    },
    {
      "epoch": 13.39,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 56050
    },
    {
      "epoch": 13.39,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 56060
    },
    {
      "epoch": 13.39,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 56070
    },
    {
      "epoch": 13.39,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 56080
    },
    {
      "epoch": 13.4,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 56090
    },
    {
      "epoch": 13.4,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 56100
    },
    {
      "epoch": 13.4,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4146016836166382,
      "eval_runtime": 32.8853,
      "eval_samples_per_second": 20.8,
      "step": 56100
    },
    {
      "epoch": 13.4,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 56110
    },
    {
      "epoch": 13.4,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 56120
    },
    {
      "epoch": 13.41,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 56130
    },
    {
      "epoch": 13.41,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 56140
    },
    {
      "epoch": 13.41,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 56150
    },
    {
      "epoch": 13.41,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 56160
    },
    {
      "epoch": 13.42,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 56170
    },
    {
      "epoch": 13.42,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 56180
    },
    {
      "epoch": 13.42,
      "learning_rate": 1e-05,
      "loss": 1.4157,
      "step": 56190
    },
    {
      "epoch": 13.42,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 56200
    },
    {
      "epoch": 13.42,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4132403135299683,
      "eval_runtime": 32.6905,
      "eval_samples_per_second": 20.923,
      "step": 56200
    },
    {
      "epoch": 13.42,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 56210
    },
    {
      "epoch": 13.43,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 56220
    },
    {
      "epoch": 13.43,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 56230
    },
    {
      "epoch": 13.43,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 56240
    },
    {
      "epoch": 13.43,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 56250
    },
    {
      "epoch": 13.44,
      "learning_rate": 1e-05,
      "loss": 1.4272,
      "step": 56260
    },
    {
      "epoch": 13.44,
      "learning_rate": 1e-05,
      "loss": 1.4172,
      "step": 56270
    },
    {
      "epoch": 13.44,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 56280
    },
    {
      "epoch": 13.44,
      "learning_rate": 1e-05,
      "loss": 1.4191,
      "step": 56290
    },
    {
      "epoch": 13.45,
      "learning_rate": 1e-05,
      "loss": 1.4166,
      "step": 56300
    },
    {
      "epoch": 13.45,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4142793416976929,
      "eval_runtime": 32.4977,
      "eval_samples_per_second": 21.048,
      "step": 56300
    },
    {
      "epoch": 13.45,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 56310
    },
    {
      "epoch": 13.45,
      "learning_rate": 1e-05,
      "loss": 1.4113,
      "step": 56320
    },
    {
      "epoch": 13.45,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 56330
    },
    {
      "epoch": 13.46,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 56340
    },
    {
      "epoch": 13.46,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 56350
    },
    {
      "epoch": 13.46,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 56360
    },
    {
      "epoch": 13.46,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 56370
    },
    {
      "epoch": 13.47,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 56380
    },
    {
      "epoch": 13.47,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 56390
    },
    {
      "epoch": 13.47,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 56400
    },
    {
      "epoch": 13.47,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4134175777435303,
      "eval_runtime": 32.3349,
      "eval_samples_per_second": 21.154,
      "step": 56400
    },
    {
      "epoch": 13.47,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 56410
    },
    {
      "epoch": 13.48,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 56420
    },
    {
      "epoch": 13.48,
      "learning_rate": 1e-05,
      "loss": 1.4142,
      "step": 56430
    },
    {
      "epoch": 13.48,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 56440
    },
    {
      "epoch": 13.48,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 56450
    },
    {
      "epoch": 13.48,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 56460
    },
    {
      "epoch": 13.49,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 56470
    },
    {
      "epoch": 13.49,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 56480
    },
    {
      "epoch": 13.49,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 56490
    },
    {
      "epoch": 13.49,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 56500
    },
    {
      "epoch": 13.49,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4133881330490112,
      "eval_runtime": 32.5941,
      "eval_samples_per_second": 20.985,
      "step": 56500
    },
    {
      "epoch": 13.5,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 56510
    },
    {
      "epoch": 13.5,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 56520
    },
    {
      "epoch": 13.5,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 56530
    },
    {
      "epoch": 13.5,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 56540
    },
    {
      "epoch": 13.51,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 56550
    },
    {
      "epoch": 13.51,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 56560
    },
    {
      "epoch": 13.51,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 56570
    },
    {
      "epoch": 13.51,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 56580
    },
    {
      "epoch": 13.52,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 56590
    },
    {
      "epoch": 13.52,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 56600
    },
    {
      "epoch": 13.52,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4143002033233643,
      "eval_runtime": 32.4014,
      "eval_samples_per_second": 21.11,
      "step": 56600
    },
    {
      "epoch": 13.52,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 56610
    },
    {
      "epoch": 13.52,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 56620
    },
    {
      "epoch": 13.53,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 56630
    },
    {
      "epoch": 13.53,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 56640
    },
    {
      "epoch": 13.53,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 56650
    },
    {
      "epoch": 13.53,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 56660
    },
    {
      "epoch": 13.53,
      "learning_rate": 1e-05,
      "loss": 1.4113,
      "step": 56670
    },
    {
      "epoch": 13.54,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 56680
    },
    {
      "epoch": 13.54,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 56690
    },
    {
      "epoch": 13.54,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 56700
    },
    {
      "epoch": 13.54,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4129122495651245,
      "eval_runtime": 32.7168,
      "eval_samples_per_second": 20.907,
      "step": 56700
    },
    {
      "epoch": 13.54,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 56710
    },
    {
      "epoch": 13.55,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 56720
    },
    {
      "epoch": 13.55,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 56730
    },
    {
      "epoch": 13.55,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 56740
    },
    {
      "epoch": 13.55,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 56750
    },
    {
      "epoch": 13.56,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 56760
    },
    {
      "epoch": 13.56,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 56770
    },
    {
      "epoch": 13.56,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 56780
    },
    {
      "epoch": 13.56,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 56790
    },
    {
      "epoch": 13.57,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 56800
    },
    {
      "epoch": 13.57,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4128296375274658,
      "eval_runtime": 32.7162,
      "eval_samples_per_second": 20.907,
      "step": 56800
    },
    {
      "epoch": 13.57,
      "learning_rate": 1e-05,
      "loss": 1.4182,
      "step": 56810
    },
    {
      "epoch": 13.57,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 56820
    },
    {
      "epoch": 13.57,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 56830
    },
    {
      "epoch": 13.58,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 56840
    },
    {
      "epoch": 13.58,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 56850
    },
    {
      "epoch": 13.58,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 56860
    },
    {
      "epoch": 13.58,
      "learning_rate": 1e-05,
      "loss": 1.4142,
      "step": 56870
    },
    {
      "epoch": 13.58,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 56880
    },
    {
      "epoch": 13.59,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 56890
    },
    {
      "epoch": 13.59,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 56900
    },
    {
      "epoch": 13.59,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4138309955596924,
      "eval_runtime": 32.3432,
      "eval_samples_per_second": 21.148,
      "step": 56900
    },
    {
      "epoch": 13.59,
      "learning_rate": 1e-05,
      "loss": 1.4113,
      "step": 56910
    },
    {
      "epoch": 13.59,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 56920
    },
    {
      "epoch": 13.6,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 56930
    },
    {
      "epoch": 13.6,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 56940
    },
    {
      "epoch": 13.6,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 56950
    },
    {
      "epoch": 13.6,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 56960
    },
    {
      "epoch": 13.61,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 56970
    },
    {
      "epoch": 13.61,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 56980
    },
    {
      "epoch": 13.61,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 56990
    },
    {
      "epoch": 13.61,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 57000
    },
    {
      "epoch": 13.61,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4133875370025635,
      "eval_runtime": 32.6093,
      "eval_samples_per_second": 20.976,
      "step": 57000
    },
    {
      "epoch": 13.62,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 57010
    },
    {
      "epoch": 13.62,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 57020
    },
    {
      "epoch": 13.62,
      "learning_rate": 1e-05,
      "loss": 1.4203,
      "step": 57030
    },
    {
      "epoch": 13.62,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 57040
    },
    {
      "epoch": 13.63,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 57050
    },
    {
      "epoch": 13.63,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 57060
    },
    {
      "epoch": 13.63,
      "learning_rate": 1e-05,
      "loss": 1.4459,
      "step": 57070
    },
    {
      "epoch": 13.63,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 57080
    },
    {
      "epoch": 13.64,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 57090
    },
    {
      "epoch": 13.64,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 57100
    },
    {
      "epoch": 13.64,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4135257005691528,
      "eval_runtime": 32.438,
      "eval_samples_per_second": 21.086,
      "step": 57100
    },
    {
      "epoch": 13.64,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 57110
    },
    {
      "epoch": 13.64,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 57120
    },
    {
      "epoch": 13.64,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 57130
    },
    {
      "epoch": 13.65,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 57140
    },
    {
      "epoch": 13.65,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 57150
    },
    {
      "epoch": 13.65,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 57160
    },
    {
      "epoch": 13.65,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 57170
    },
    {
      "epoch": 13.66,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 57180
    },
    {
      "epoch": 13.66,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 57190
    },
    {
      "epoch": 13.66,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 57200
    },
    {
      "epoch": 13.66,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4130704402923584,
      "eval_runtime": 32.4609,
      "eval_samples_per_second": 21.072,
      "step": 57200
    },
    {
      "epoch": 13.66,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 57210
    },
    {
      "epoch": 13.67,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 57220
    },
    {
      "epoch": 13.67,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 57230
    },
    {
      "epoch": 13.67,
      "learning_rate": 1e-05,
      "loss": 1.4278,
      "step": 57240
    },
    {
      "epoch": 13.67,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 57250
    },
    {
      "epoch": 13.68,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 57260
    },
    {
      "epoch": 13.68,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 57270
    },
    {
      "epoch": 13.68,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 57280
    },
    {
      "epoch": 13.68,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 57290
    },
    {
      "epoch": 13.69,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 57300
    },
    {
      "epoch": 13.69,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4127709865570068,
      "eval_runtime": 32.7644,
      "eval_samples_per_second": 20.876,
      "step": 57300
    },
    {
      "epoch": 13.69,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 57310
    },
    {
      "epoch": 13.69,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 57320
    },
    {
      "epoch": 13.69,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 57330
    },
    {
      "epoch": 13.69,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 57340
    },
    {
      "epoch": 13.7,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 57350
    },
    {
      "epoch": 13.7,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 57360
    },
    {
      "epoch": 13.7,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 57370
    },
    {
      "epoch": 13.7,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 57380
    },
    {
      "epoch": 13.71,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 57390
    },
    {
      "epoch": 13.71,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 57400
    },
    {
      "epoch": 13.71,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4132970571517944,
      "eval_runtime": 32.5251,
      "eval_samples_per_second": 21.03,
      "step": 57400
    },
    {
      "epoch": 13.71,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 57410
    },
    {
      "epoch": 13.71,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 57420
    },
    {
      "epoch": 13.72,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 57430
    },
    {
      "epoch": 13.72,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 57440
    },
    {
      "epoch": 13.72,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 57450
    },
    {
      "epoch": 13.72,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 57460
    },
    {
      "epoch": 13.73,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 57470
    },
    {
      "epoch": 13.73,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 57480
    },
    {
      "epoch": 13.73,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 57490
    },
    {
      "epoch": 13.73,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 57500
    },
    {
      "epoch": 13.73,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4125943183898926,
      "eval_runtime": 32.9833,
      "eval_samples_per_second": 20.738,
      "step": 57500
    },
    {
      "epoch": 13.74,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 57510
    },
    {
      "epoch": 13.74,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 57520
    },
    {
      "epoch": 13.74,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 57530
    },
    {
      "epoch": 13.74,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 57540
    },
    {
      "epoch": 13.74,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 57550
    },
    {
      "epoch": 13.75,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 57560
    },
    {
      "epoch": 13.75,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 57570
    },
    {
      "epoch": 13.75,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 57580
    },
    {
      "epoch": 13.75,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 57590
    },
    {
      "epoch": 13.76,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 57600
    },
    {
      "epoch": 13.76,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4122618436813354,
      "eval_runtime": 32.2704,
      "eval_samples_per_second": 21.196,
      "step": 57600
    },
    {
      "epoch": 13.76,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 57610
    },
    {
      "epoch": 13.76,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 57620
    },
    {
      "epoch": 13.76,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 57630
    },
    {
      "epoch": 13.77,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 57640
    },
    {
      "epoch": 13.77,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 57650
    },
    {
      "epoch": 13.77,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 57660
    },
    {
      "epoch": 13.77,
      "learning_rate": 1e-05,
      "loss": 1.4181,
      "step": 57670
    },
    {
      "epoch": 13.78,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 57680
    },
    {
      "epoch": 13.78,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 57690
    },
    {
      "epoch": 13.78,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 57700
    },
    {
      "epoch": 13.78,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4131001234054565,
      "eval_runtime": 32.4385,
      "eval_samples_per_second": 21.086,
      "step": 57700
    },
    {
      "epoch": 13.78,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 57710
    },
    {
      "epoch": 13.79,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 57720
    },
    {
      "epoch": 13.79,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 57730
    },
    {
      "epoch": 13.79,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 57740
    },
    {
      "epoch": 13.79,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 57750
    },
    {
      "epoch": 13.8,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 57760
    },
    {
      "epoch": 13.8,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 57770
    },
    {
      "epoch": 13.8,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 57780
    },
    {
      "epoch": 13.8,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 57790
    },
    {
      "epoch": 13.8,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 57800
    },
    {
      "epoch": 13.8,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.412663459777832,
      "eval_runtime": 32.6951,
      "eval_samples_per_second": 20.921,
      "step": 57800
    },
    {
      "epoch": 13.81,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 57810
    },
    {
      "epoch": 13.81,
      "learning_rate": 1e-05,
      "loss": 1.4328,
      "step": 57820
    },
    {
      "epoch": 13.81,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 57830
    },
    {
      "epoch": 13.81,
      "learning_rate": 1e-05,
      "loss": 1.4194,
      "step": 57840
    },
    {
      "epoch": 13.82,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 57850
    },
    {
      "epoch": 13.82,
      "learning_rate": 1e-05,
      "loss": 1.4174,
      "step": 57860
    },
    {
      "epoch": 13.82,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 57870
    },
    {
      "epoch": 13.82,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 57880
    },
    {
      "epoch": 13.83,
      "learning_rate": 1e-05,
      "loss": 1.4181,
      "step": 57890
    },
    {
      "epoch": 13.83,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 57900
    },
    {
      "epoch": 13.83,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4125932455062866,
      "eval_runtime": 32.1863,
      "eval_samples_per_second": 21.251,
      "step": 57900
    },
    {
      "epoch": 13.83,
      "learning_rate": 1e-05,
      "loss": 1.4176,
      "step": 57910
    },
    {
      "epoch": 13.83,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 57920
    },
    {
      "epoch": 13.84,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 57930
    },
    {
      "epoch": 13.84,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 57940
    },
    {
      "epoch": 13.84,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 57950
    },
    {
      "epoch": 13.84,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 57960
    },
    {
      "epoch": 13.85,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 57970
    },
    {
      "epoch": 13.85,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 57980
    },
    {
      "epoch": 13.85,
      "learning_rate": 1e-05,
      "loss": 1.4164,
      "step": 57990
    },
    {
      "epoch": 13.85,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 58000
    },
    {
      "epoch": 13.85,
      "eval_accuracy": 0.9868421052631579,
      "eval_loss": 1.4125031232833862,
      "eval_runtime": 32.6699,
      "eval_samples_per_second": 20.937,
      "step": 58000
    },
    {
      "epoch": 13.85,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 58010
    },
    {
      "epoch": 13.86,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 58020
    },
    {
      "epoch": 13.86,
      "learning_rate": 1e-05,
      "loss": 1.4198,
      "step": 58030
    },
    {
      "epoch": 13.86,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 58040
    },
    {
      "epoch": 13.86,
      "learning_rate": 1e-05,
      "loss": 1.4189,
      "step": 58050
    },
    {
      "epoch": 13.87,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 58060
    },
    {
      "epoch": 13.87,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 58070
    },
    {
      "epoch": 13.87,
      "learning_rate": 1e-05,
      "loss": 1.4142,
      "step": 58080
    },
    {
      "epoch": 13.87,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 58090
    },
    {
      "epoch": 13.88,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 58100
    },
    {
      "epoch": 13.88,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.413333773612976,
      "eval_runtime": 32.5543,
      "eval_samples_per_second": 21.011,
      "step": 58100
    },
    {
      "epoch": 13.88,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 58110
    },
    {
      "epoch": 13.88,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 58120
    },
    {
      "epoch": 13.88,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 58130
    },
    {
      "epoch": 13.89,
      "learning_rate": 1e-05,
      "loss": 1.426,
      "step": 58140
    },
    {
      "epoch": 13.89,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 58150
    },
    {
      "epoch": 13.89,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 58160
    },
    {
      "epoch": 13.89,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 58170
    },
    {
      "epoch": 13.9,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 58180
    },
    {
      "epoch": 13.9,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 58190
    },
    {
      "epoch": 13.9,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 58200
    },
    {
      "epoch": 13.9,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4121471643447876,
      "eval_runtime": 32.6443,
      "eval_samples_per_second": 20.953,
      "step": 58200
    },
    {
      "epoch": 13.9,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 58210
    },
    {
      "epoch": 13.9,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 58220
    },
    {
      "epoch": 13.91,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 58230
    },
    {
      "epoch": 13.91,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 58240
    },
    {
      "epoch": 13.91,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 58250
    },
    {
      "epoch": 13.91,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 58260
    },
    {
      "epoch": 13.92,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 58270
    },
    {
      "epoch": 13.92,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 58280
    },
    {
      "epoch": 13.92,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 58290
    },
    {
      "epoch": 13.92,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 58300
    },
    {
      "epoch": 13.92,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.412467360496521,
      "eval_runtime": 32.6276,
      "eval_samples_per_second": 20.964,
      "step": 58300
    },
    {
      "epoch": 13.93,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 58310
    },
    {
      "epoch": 13.93,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 58320
    },
    {
      "epoch": 13.93,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 58330
    },
    {
      "epoch": 13.93,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 58340
    },
    {
      "epoch": 13.94,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 58350
    },
    {
      "epoch": 13.94,
      "learning_rate": 1e-05,
      "loss": 1.4174,
      "step": 58360
    },
    {
      "epoch": 13.94,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 58370
    },
    {
      "epoch": 13.94,
      "learning_rate": 1e-05,
      "loss": 1.4147,
      "step": 58380
    },
    {
      "epoch": 13.95,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 58390
    },
    {
      "epoch": 13.95,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 58400
    },
    {
      "epoch": 13.95,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4126015901565552,
      "eval_runtime": 32.4755,
      "eval_samples_per_second": 21.062,
      "step": 58400
    },
    {
      "epoch": 13.95,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 58410
    },
    {
      "epoch": 13.95,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 58420
    },
    {
      "epoch": 13.96,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 58430
    },
    {
      "epoch": 13.96,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 58440
    },
    {
      "epoch": 13.96,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 58450
    },
    {
      "epoch": 13.96,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 58460
    },
    {
      "epoch": 13.96,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 58470
    },
    {
      "epoch": 13.97,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 58480
    },
    {
      "epoch": 13.97,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 58490
    },
    {
      "epoch": 13.97,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 58500
    },
    {
      "epoch": 13.97,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4121787548065186,
      "eval_runtime": 32.5482,
      "eval_samples_per_second": 21.015,
      "step": 58500
    },
    {
      "epoch": 13.97,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 58510
    },
    {
      "epoch": 13.98,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 58520
    },
    {
      "epoch": 13.98,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 58530
    },
    {
      "epoch": 13.98,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 58540
    },
    {
      "epoch": 13.98,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 58550
    },
    {
      "epoch": 13.99,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 58560
    },
    {
      "epoch": 13.99,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 58570
    },
    {
      "epoch": 13.99,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 58580
    },
    {
      "epoch": 13.99,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 58590
    },
    {
      "epoch": 14.0,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 58600
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4139076471328735,
      "eval_runtime": 32.2433,
      "eval_samples_per_second": 21.214,
      "step": 58600
    },
    {
      "epoch": 14.0,
      "learning_rate": 1e-05,
      "loss": 1.4325,
      "step": 58610
    },
    {
      "epoch": 14.0,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 58620
    },
    {
      "epoch": 14.0,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 58630
    },
    {
      "epoch": 14.01,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 58640
    },
    {
      "epoch": 14.01,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 58650
    },
    {
      "epoch": 14.01,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 58660
    },
    {
      "epoch": 14.01,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 58670
    },
    {
      "epoch": 14.01,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 58680
    },
    {
      "epoch": 14.02,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 58690
    },
    {
      "epoch": 14.02,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 58700
    },
    {
      "epoch": 14.02,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4116936922073364,
      "eval_runtime": 32.6637,
      "eval_samples_per_second": 20.941,
      "step": 58700
    },
    {
      "epoch": 14.02,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 58710
    },
    {
      "epoch": 14.02,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 58720
    },
    {
      "epoch": 14.03,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 58730
    },
    {
      "epoch": 14.03,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 58740
    },
    {
      "epoch": 14.03,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 58750
    },
    {
      "epoch": 14.03,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 58760
    },
    {
      "epoch": 14.04,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 58770
    },
    {
      "epoch": 14.04,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 58780
    },
    {
      "epoch": 14.04,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 58790
    },
    {
      "epoch": 14.04,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 58800
    },
    {
      "epoch": 14.04,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4127209186553955,
      "eval_runtime": 32.251,
      "eval_samples_per_second": 21.209,
      "step": 58800
    },
    {
      "epoch": 14.05,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 58810
    },
    {
      "epoch": 14.05,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 58820
    },
    {
      "epoch": 14.05,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 58830
    },
    {
      "epoch": 14.05,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 58840
    },
    {
      "epoch": 14.06,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 58850
    },
    {
      "epoch": 14.06,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 58860
    },
    {
      "epoch": 14.06,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 58870
    },
    {
      "epoch": 14.06,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 58880
    },
    {
      "epoch": 14.06,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 58890
    },
    {
      "epoch": 14.07,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 58900
    },
    {
      "epoch": 14.07,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4110255241394043,
      "eval_runtime": 32.4092,
      "eval_samples_per_second": 21.105,
      "step": 58900
    },
    {
      "epoch": 14.07,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 58910
    },
    {
      "epoch": 14.07,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 58920
    },
    {
      "epoch": 14.07,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 58930
    },
    {
      "epoch": 14.08,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 58940
    },
    {
      "epoch": 14.08,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 58950
    },
    {
      "epoch": 14.08,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 58960
    },
    {
      "epoch": 14.08,
      "learning_rate": 1e-05,
      "loss": 1.4147,
      "step": 58970
    },
    {
      "epoch": 14.09,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 58980
    },
    {
      "epoch": 14.09,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 58990
    },
    {
      "epoch": 14.09,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 59000
    },
    {
      "epoch": 14.09,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4123581647872925,
      "eval_runtime": 32.669,
      "eval_samples_per_second": 20.937,
      "step": 59000
    },
    {
      "epoch": 14.09,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 59010
    },
    {
      "epoch": 14.1,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 59020
    },
    {
      "epoch": 14.1,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 59030
    },
    {
      "epoch": 14.1,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 59040
    },
    {
      "epoch": 14.1,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 59050
    },
    {
      "epoch": 14.11,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 59060
    },
    {
      "epoch": 14.11,
      "learning_rate": 1e-05,
      "loss": 1.4137,
      "step": 59070
    },
    {
      "epoch": 14.11,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 59080
    },
    {
      "epoch": 14.11,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 59090
    },
    {
      "epoch": 14.12,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 59100
    },
    {
      "epoch": 14.12,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4114679098129272,
      "eval_runtime": 32.5063,
      "eval_samples_per_second": 21.042,
      "step": 59100
    },
    {
      "epoch": 14.12,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 59110
    },
    {
      "epoch": 14.12,
      "learning_rate": 1e-05,
      "loss": 1.4162,
      "step": 59120
    },
    {
      "epoch": 14.12,
      "learning_rate": 1e-05,
      "loss": 1.4155,
      "step": 59130
    },
    {
      "epoch": 14.12,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 59140
    },
    {
      "epoch": 14.13,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 59150
    },
    {
      "epoch": 14.13,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 59160
    },
    {
      "epoch": 14.13,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 59170
    },
    {
      "epoch": 14.13,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 59180
    },
    {
      "epoch": 14.14,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 59190
    },
    {
      "epoch": 14.14,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 59200
    },
    {
      "epoch": 14.14,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4120044708251953,
      "eval_runtime": 32.6403,
      "eval_samples_per_second": 20.956,
      "step": 59200
    },
    {
      "epoch": 14.14,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 59210
    },
    {
      "epoch": 14.14,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 59220
    },
    {
      "epoch": 14.15,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 59230
    },
    {
      "epoch": 14.15,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 59240
    },
    {
      "epoch": 14.15,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 59250
    },
    {
      "epoch": 14.15,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 59260
    },
    {
      "epoch": 14.16,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 59270
    },
    {
      "epoch": 14.16,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 59280
    },
    {
      "epoch": 14.16,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 59290
    },
    {
      "epoch": 14.16,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 59300
    },
    {
      "epoch": 14.16,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4111289978027344,
      "eval_runtime": 32.643,
      "eval_samples_per_second": 20.954,
      "step": 59300
    },
    {
      "epoch": 14.17,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 59310
    },
    {
      "epoch": 14.17,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 59320
    },
    {
      "epoch": 14.17,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 59330
    },
    {
      "epoch": 14.17,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 59340
    },
    {
      "epoch": 14.17,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 59350
    },
    {
      "epoch": 14.18,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 59360
    },
    {
      "epoch": 14.18,
      "learning_rate": 1e-05,
      "loss": 1.4113,
      "step": 59370
    },
    {
      "epoch": 14.18,
      "learning_rate": 1e-05,
      "loss": 1.4113,
      "step": 59380
    },
    {
      "epoch": 14.18,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 59390
    },
    {
      "epoch": 14.19,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 59400
    },
    {
      "epoch": 14.19,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4128690958023071,
      "eval_runtime": 32.6155,
      "eval_samples_per_second": 20.972,
      "step": 59400
    },
    {
      "epoch": 14.19,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 59410
    },
    {
      "epoch": 14.19,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 59420
    },
    {
      "epoch": 14.19,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 59430
    },
    {
      "epoch": 14.2,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 59440
    },
    {
      "epoch": 14.2,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 59450
    },
    {
      "epoch": 14.2,
      "learning_rate": 1e-05,
      "loss": 1.4157,
      "step": 59460
    },
    {
      "epoch": 14.2,
      "learning_rate": 1e-05,
      "loss": 1.4147,
      "step": 59470
    },
    {
      "epoch": 14.21,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 59480
    },
    {
      "epoch": 14.21,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 59490
    },
    {
      "epoch": 14.21,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 59500
    },
    {
      "epoch": 14.21,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4115397930145264,
      "eval_runtime": 32.5629,
      "eval_samples_per_second": 21.005,
      "step": 59500
    },
    {
      "epoch": 14.21,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 59510
    },
    {
      "epoch": 14.22,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 59520
    },
    {
      "epoch": 14.22,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 59530
    },
    {
      "epoch": 14.22,
      "learning_rate": 1e-05,
      "loss": 1.4136,
      "step": 59540
    },
    {
      "epoch": 14.22,
      "learning_rate": 1e-05,
      "loss": 1.4176,
      "step": 59550
    },
    {
      "epoch": 14.22,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 59560
    },
    {
      "epoch": 14.23,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 59570
    },
    {
      "epoch": 14.23,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 59580
    },
    {
      "epoch": 14.23,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 59590
    },
    {
      "epoch": 14.23,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 59600
    },
    {
      "epoch": 14.23,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.411746859550476,
      "eval_runtime": 32.5741,
      "eval_samples_per_second": 20.998,
      "step": 59600
    },
    {
      "epoch": 14.24,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 59610
    },
    {
      "epoch": 14.24,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 59620
    },
    {
      "epoch": 14.24,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 59630
    },
    {
      "epoch": 14.24,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 59640
    },
    {
      "epoch": 14.25,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 59650
    },
    {
      "epoch": 14.25,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 59660
    },
    {
      "epoch": 14.25,
      "learning_rate": 1e-05,
      "loss": 1.4178,
      "step": 59670
    },
    {
      "epoch": 14.25,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 59680
    },
    {
      "epoch": 14.26,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 59690
    },
    {
      "epoch": 14.26,
      "learning_rate": 1e-05,
      "loss": 1.4185,
      "step": 59700
    },
    {
      "epoch": 14.26,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4122834205627441,
      "eval_runtime": 32.6992,
      "eval_samples_per_second": 20.918,
      "step": 59700
    },
    {
      "epoch": 14.26,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 59710
    },
    {
      "epoch": 14.26,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 59720
    },
    {
      "epoch": 14.27,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 59730
    },
    {
      "epoch": 14.27,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 59740
    },
    {
      "epoch": 14.27,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 59750
    },
    {
      "epoch": 14.27,
      "learning_rate": 1e-05,
      "loss": 1.4206,
      "step": 59760
    },
    {
      "epoch": 14.28,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 59770
    },
    {
      "epoch": 14.28,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 59780
    },
    {
      "epoch": 14.28,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 59790
    },
    {
      "epoch": 14.28,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 59800
    },
    {
      "epoch": 14.28,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4117363691329956,
      "eval_runtime": 32.6342,
      "eval_samples_per_second": 20.96,
      "step": 59800
    },
    {
      "epoch": 14.28,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 59810
    },
    {
      "epoch": 14.29,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 59820
    },
    {
      "epoch": 14.29,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 59830
    },
    {
      "epoch": 14.29,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 59840
    },
    {
      "epoch": 14.29,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 59850
    },
    {
      "epoch": 14.3,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 59860
    },
    {
      "epoch": 14.3,
      "learning_rate": 1e-05,
      "loss": 1.413,
      "step": 59870
    },
    {
      "epoch": 14.3,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 59880
    },
    {
      "epoch": 14.3,
      "learning_rate": 1e-05,
      "loss": 1.413,
      "step": 59890
    },
    {
      "epoch": 14.31,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 59900
    },
    {
      "epoch": 14.31,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.412007212638855,
      "eval_runtime": 32.5853,
      "eval_samples_per_second": 20.991,
      "step": 59900
    },
    {
      "epoch": 14.31,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 59910
    },
    {
      "epoch": 14.31,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 59920
    },
    {
      "epoch": 14.31,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 59930
    },
    {
      "epoch": 14.32,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 59940
    },
    {
      "epoch": 14.32,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 59950
    },
    {
      "epoch": 14.32,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 59960
    },
    {
      "epoch": 14.32,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 59970
    },
    {
      "epoch": 14.33,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 59980
    },
    {
      "epoch": 14.33,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 59990
    },
    {
      "epoch": 14.33,
      "learning_rate": 1e-05,
      "loss": 1.4147,
      "step": 60000
    },
    {
      "epoch": 14.33,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.412237524986267,
      "eval_runtime": 32.7566,
      "eval_samples_per_second": 20.881,
      "step": 60000
    },
    {
      "epoch": 14.33,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 60010
    },
    {
      "epoch": 14.33,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 60020
    },
    {
      "epoch": 14.34,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 60030
    },
    {
      "epoch": 14.34,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 60040
    },
    {
      "epoch": 14.34,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 60050
    },
    {
      "epoch": 14.34,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 60060
    },
    {
      "epoch": 14.35,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 60070
    },
    {
      "epoch": 14.35,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 60080
    },
    {
      "epoch": 14.35,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 60090
    },
    {
      "epoch": 14.35,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 60100
    },
    {
      "epoch": 14.35,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4120662212371826,
      "eval_runtime": 32.7169,
      "eval_samples_per_second": 20.907,
      "step": 60100
    },
    {
      "epoch": 14.36,
      "learning_rate": 1e-05,
      "loss": 1.418,
      "step": 60110
    },
    {
      "epoch": 14.36,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 60120
    },
    {
      "epoch": 14.36,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 60130
    },
    {
      "epoch": 14.36,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 60140
    },
    {
      "epoch": 14.37,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 60150
    },
    {
      "epoch": 14.37,
      "learning_rate": 1e-05,
      "loss": 1.4113,
      "step": 60160
    },
    {
      "epoch": 14.37,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 60170
    },
    {
      "epoch": 14.37,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 60180
    },
    {
      "epoch": 14.38,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 60190
    },
    {
      "epoch": 14.38,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 60200
    },
    {
      "epoch": 14.38,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4119693040847778,
      "eval_runtime": 32.4659,
      "eval_samples_per_second": 21.068,
      "step": 60200
    },
    {
      "epoch": 14.38,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 60210
    },
    {
      "epoch": 14.38,
      "learning_rate": 1e-05,
      "loss": 1.4179,
      "step": 60220
    },
    {
      "epoch": 14.39,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 60230
    },
    {
      "epoch": 14.39,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 60240
    },
    {
      "epoch": 14.39,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 60250
    },
    {
      "epoch": 14.39,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 60260
    },
    {
      "epoch": 14.39,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 60270
    },
    {
      "epoch": 14.4,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 60280
    },
    {
      "epoch": 14.4,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 60290
    },
    {
      "epoch": 14.4,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 60300
    },
    {
      "epoch": 14.4,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4117532968521118,
      "eval_runtime": 32.7989,
      "eval_samples_per_second": 20.854,
      "step": 60300
    },
    {
      "epoch": 14.4,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 60310
    },
    {
      "epoch": 14.41,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 60320
    },
    {
      "epoch": 14.41,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 60330
    },
    {
      "epoch": 14.41,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 60340
    },
    {
      "epoch": 14.41,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 60350
    },
    {
      "epoch": 14.42,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 60360
    },
    {
      "epoch": 14.42,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 60370
    },
    {
      "epoch": 14.42,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 60380
    },
    {
      "epoch": 14.42,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 60390
    },
    {
      "epoch": 14.43,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 60400
    },
    {
      "epoch": 14.43,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4119771718978882,
      "eval_runtime": 33.1736,
      "eval_samples_per_second": 20.619,
      "step": 60400
    },
    {
      "epoch": 14.43,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 60410
    },
    {
      "epoch": 14.43,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 60420
    },
    {
      "epoch": 14.43,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 60430
    },
    {
      "epoch": 14.44,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 60440
    },
    {
      "epoch": 14.44,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 60450
    },
    {
      "epoch": 14.44,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 60460
    },
    {
      "epoch": 14.44,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 60470
    },
    {
      "epoch": 14.44,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 60480
    },
    {
      "epoch": 14.45,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 60490
    },
    {
      "epoch": 14.45,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 60500
    },
    {
      "epoch": 14.45,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4123705625534058,
      "eval_runtime": 32.5237,
      "eval_samples_per_second": 21.031,
      "step": 60500
    },
    {
      "epoch": 14.45,
      "learning_rate": 1e-05,
      "loss": 1.413,
      "step": 60510
    },
    {
      "epoch": 14.45,
      "learning_rate": 1e-05,
      "loss": 1.4153,
      "step": 60520
    },
    {
      "epoch": 14.46,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 60530
    },
    {
      "epoch": 14.46,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 60540
    },
    {
      "epoch": 14.46,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 60550
    },
    {
      "epoch": 14.46,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 60560
    },
    {
      "epoch": 14.47,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 60570
    },
    {
      "epoch": 14.47,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 60580
    },
    {
      "epoch": 14.47,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 60590
    },
    {
      "epoch": 14.47,
      "learning_rate": 1e-05,
      "loss": 1.413,
      "step": 60600
    },
    {
      "epoch": 14.47,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4119683504104614,
      "eval_runtime": 32.6437,
      "eval_samples_per_second": 20.954,
      "step": 60600
    },
    {
      "epoch": 14.48,
      "learning_rate": 1e-05,
      "loss": 1.4166,
      "step": 60610
    },
    {
      "epoch": 14.48,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 60620
    },
    {
      "epoch": 14.48,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 60630
    },
    {
      "epoch": 14.48,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 60640
    },
    {
      "epoch": 14.49,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 60650
    },
    {
      "epoch": 14.49,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 60660
    },
    {
      "epoch": 14.49,
      "learning_rate": 1e-05,
      "loss": 1.4156,
      "step": 60670
    },
    {
      "epoch": 14.49,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 60680
    },
    {
      "epoch": 14.49,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 60690
    },
    {
      "epoch": 14.5,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 60700
    },
    {
      "epoch": 14.5,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.411042332649231,
      "eval_runtime": 32.4916,
      "eval_samples_per_second": 21.052,
      "step": 60700
    },
    {
      "epoch": 14.5,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 60710
    },
    {
      "epoch": 14.5,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 60720
    },
    {
      "epoch": 14.5,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 60730
    },
    {
      "epoch": 14.51,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 60740
    },
    {
      "epoch": 14.51,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 60750
    },
    {
      "epoch": 14.51,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 60760
    },
    {
      "epoch": 14.51,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 60770
    },
    {
      "epoch": 14.52,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 60780
    },
    {
      "epoch": 14.52,
      "learning_rate": 1e-05,
      "loss": 1.4136,
      "step": 60790
    },
    {
      "epoch": 14.52,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 60800
    },
    {
      "epoch": 14.52,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.411872386932373,
      "eval_runtime": 32.2034,
      "eval_samples_per_second": 21.24,
      "step": 60800
    },
    {
      "epoch": 14.52,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 60810
    },
    {
      "epoch": 14.53,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 60820
    },
    {
      "epoch": 14.53,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 60830
    },
    {
      "epoch": 14.53,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 60840
    },
    {
      "epoch": 14.53,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 60850
    },
    {
      "epoch": 14.54,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 60860
    },
    {
      "epoch": 14.54,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 60870
    },
    {
      "epoch": 14.54,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 60880
    },
    {
      "epoch": 14.54,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 60890
    },
    {
      "epoch": 14.55,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 60900
    },
    {
      "epoch": 14.55,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4111363887786865,
      "eval_runtime": 32.3386,
      "eval_samples_per_second": 21.151,
      "step": 60900
    },
    {
      "epoch": 14.55,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 60910
    },
    {
      "epoch": 14.55,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 60920
    },
    {
      "epoch": 14.55,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 60930
    },
    {
      "epoch": 14.55,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 60940
    },
    {
      "epoch": 14.56,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 60950
    },
    {
      "epoch": 14.56,
      "learning_rate": 1e-05,
      "loss": 1.413,
      "step": 60960
    },
    {
      "epoch": 14.56,
      "learning_rate": 1e-05,
      "loss": 1.4155,
      "step": 60970
    },
    {
      "epoch": 14.56,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 60980
    },
    {
      "epoch": 14.57,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 60990
    },
    {
      "epoch": 14.57,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 61000
    },
    {
      "epoch": 14.57,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4118964672088623,
      "eval_runtime": 32.6978,
      "eval_samples_per_second": 20.919,
      "step": 61000
    },
    {
      "epoch": 14.57,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 61010
    },
    {
      "epoch": 14.57,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 61020
    },
    {
      "epoch": 14.58,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 61030
    },
    {
      "epoch": 14.58,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 61040
    },
    {
      "epoch": 14.58,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 61050
    },
    {
      "epoch": 14.58,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 61060
    },
    {
      "epoch": 14.59,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 61070
    },
    {
      "epoch": 14.59,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 61080
    },
    {
      "epoch": 14.59,
      "learning_rate": 1e-05,
      "loss": 1.4157,
      "step": 61090
    },
    {
      "epoch": 14.59,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 61100
    },
    {
      "epoch": 14.59,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4113800525665283,
      "eval_runtime": 32.5542,
      "eval_samples_per_second": 21.011,
      "step": 61100
    },
    {
      "epoch": 14.6,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 61110
    },
    {
      "epoch": 14.6,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 61120
    },
    {
      "epoch": 14.6,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 61130
    },
    {
      "epoch": 14.6,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 61140
    },
    {
      "epoch": 14.6,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 61150
    },
    {
      "epoch": 14.61,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 61160
    },
    {
      "epoch": 14.61,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 61170
    },
    {
      "epoch": 14.61,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 61180
    },
    {
      "epoch": 14.61,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 61190
    },
    {
      "epoch": 14.62,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 61200
    },
    {
      "epoch": 14.62,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4116570949554443,
      "eval_runtime": 32.572,
      "eval_samples_per_second": 21.0,
      "step": 61200
    },
    {
      "epoch": 14.62,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 61210
    },
    {
      "epoch": 14.62,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 61220
    },
    {
      "epoch": 14.62,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 61230
    },
    {
      "epoch": 14.63,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 61240
    },
    {
      "epoch": 14.63,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 61250
    },
    {
      "epoch": 14.63,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 61260
    },
    {
      "epoch": 14.63,
      "learning_rate": 1e-05,
      "loss": 1.4142,
      "step": 61270
    },
    {
      "epoch": 14.64,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 61280
    },
    {
      "epoch": 14.64,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 61290
    },
    {
      "epoch": 14.64,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 61300
    },
    {
      "epoch": 14.64,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4101588726043701,
      "eval_runtime": 32.5926,
      "eval_samples_per_second": 20.986,
      "step": 61300
    },
    {
      "epoch": 14.64,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 61310
    },
    {
      "epoch": 14.65,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 61320
    },
    {
      "epoch": 14.65,
      "learning_rate": 1e-05,
      "loss": 1.4142,
      "step": 61330
    },
    {
      "epoch": 14.65,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 61340
    },
    {
      "epoch": 14.65,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 61350
    },
    {
      "epoch": 14.65,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 61360
    },
    {
      "epoch": 14.66,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 61370
    },
    {
      "epoch": 14.66,
      "learning_rate": 1e-05,
      "loss": 1.4113,
      "step": 61380
    },
    {
      "epoch": 14.66,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 61390
    },
    {
      "epoch": 14.66,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 61400
    },
    {
      "epoch": 14.66,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4116535186767578,
      "eval_runtime": 32.6266,
      "eval_samples_per_second": 20.965,
      "step": 61400
    },
    {
      "epoch": 14.67,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 61410
    },
    {
      "epoch": 14.67,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 61420
    },
    {
      "epoch": 14.67,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 61430
    },
    {
      "epoch": 14.67,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 61440
    },
    {
      "epoch": 14.68,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 61450
    },
    {
      "epoch": 14.68,
      "learning_rate": 1e-05,
      "loss": 1.4188,
      "step": 61460
    },
    {
      "epoch": 14.68,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 61470
    },
    {
      "epoch": 14.68,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 61480
    },
    {
      "epoch": 14.69,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 61490
    },
    {
      "epoch": 14.69,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 61500
    },
    {
      "epoch": 14.69,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.411902666091919,
      "eval_runtime": 32.3614,
      "eval_samples_per_second": 21.136,
      "step": 61500
    },
    {
      "epoch": 14.69,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 61510
    },
    {
      "epoch": 14.69,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 61520
    },
    {
      "epoch": 14.7,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 61530
    },
    {
      "epoch": 14.7,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 61540
    },
    {
      "epoch": 14.7,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 61550
    },
    {
      "epoch": 14.7,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 61560
    },
    {
      "epoch": 14.71,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 61570
    },
    {
      "epoch": 14.71,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 61580
    },
    {
      "epoch": 14.71,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 61590
    },
    {
      "epoch": 14.71,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 61600
    },
    {
      "epoch": 14.71,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4106053113937378,
      "eval_runtime": 32.65,
      "eval_samples_per_second": 20.949,
      "step": 61600
    },
    {
      "epoch": 14.71,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 61610
    },
    {
      "epoch": 14.72,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 61620
    },
    {
      "epoch": 14.72,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 61630
    },
    {
      "epoch": 14.72,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 61640
    },
    {
      "epoch": 14.72,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 61650
    },
    {
      "epoch": 14.73,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 61660
    },
    {
      "epoch": 14.73,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 61670
    },
    {
      "epoch": 14.73,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 61680
    },
    {
      "epoch": 14.73,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 61690
    },
    {
      "epoch": 14.74,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 61700
    },
    {
      "epoch": 14.74,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4115220308303833,
      "eval_runtime": 32.586,
      "eval_samples_per_second": 20.991,
      "step": 61700
    },
    {
      "epoch": 14.74,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 61710
    },
    {
      "epoch": 14.74,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 61720
    },
    {
      "epoch": 14.74,
      "learning_rate": 1e-05,
      "loss": 1.4183,
      "step": 61730
    },
    {
      "epoch": 14.75,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 61740
    },
    {
      "epoch": 14.75,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 61750
    },
    {
      "epoch": 14.75,
      "learning_rate": 1e-05,
      "loss": 1.5106,
      "step": 61760
    },
    {
      "epoch": 14.75,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 61770
    },
    {
      "epoch": 14.76,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 61780
    },
    {
      "epoch": 14.76,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 61790
    },
    {
      "epoch": 14.76,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 61800
    },
    {
      "epoch": 14.76,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.410914659500122,
      "eval_runtime": 32.4165,
      "eval_samples_per_second": 21.1,
      "step": 61800
    },
    {
      "epoch": 14.76,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 61810
    },
    {
      "epoch": 14.76,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 61820
    },
    {
      "epoch": 14.77,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 61830
    },
    {
      "epoch": 14.77,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 61840
    },
    {
      "epoch": 14.77,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 61850
    },
    {
      "epoch": 14.77,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 61860
    },
    {
      "epoch": 14.78,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 61870
    },
    {
      "epoch": 14.78,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 61880
    },
    {
      "epoch": 14.78,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 61890
    },
    {
      "epoch": 14.78,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 61900
    },
    {
      "epoch": 14.78,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4118598699569702,
      "eval_runtime": 32.8984,
      "eval_samples_per_second": 20.791,
      "step": 61900
    },
    {
      "epoch": 14.79,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 61910
    },
    {
      "epoch": 14.79,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 61920
    },
    {
      "epoch": 14.79,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 61930
    },
    {
      "epoch": 14.79,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 61940
    },
    {
      "epoch": 14.8,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 61950
    },
    {
      "epoch": 14.8,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 61960
    },
    {
      "epoch": 14.8,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 61970
    },
    {
      "epoch": 14.8,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 61980
    },
    {
      "epoch": 14.81,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 61990
    },
    {
      "epoch": 14.81,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 62000
    },
    {
      "epoch": 14.81,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.411170482635498,
      "eval_runtime": 32.5406,
      "eval_samples_per_second": 21.02,
      "step": 62000
    },
    {
      "epoch": 14.81,
      "learning_rate": 1e-05,
      "loss": 1.4159,
      "step": 62010
    },
    {
      "epoch": 14.81,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 62020
    },
    {
      "epoch": 14.81,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 62030
    },
    {
      "epoch": 14.82,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 62040
    },
    {
      "epoch": 14.82,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 62050
    },
    {
      "epoch": 14.82,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 62060
    },
    {
      "epoch": 14.82,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 62070
    },
    {
      "epoch": 14.83,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 62080
    },
    {
      "epoch": 14.83,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 62090
    },
    {
      "epoch": 14.83,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 62100
    },
    {
      "epoch": 14.83,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4114642143249512,
      "eval_runtime": 32.2225,
      "eval_samples_per_second": 21.227,
      "step": 62100
    },
    {
      "epoch": 14.83,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 62110
    },
    {
      "epoch": 14.84,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 62120
    },
    {
      "epoch": 14.84,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 62130
    },
    {
      "epoch": 14.84,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 62140
    },
    {
      "epoch": 14.84,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 62150
    },
    {
      "epoch": 14.85,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 62160
    },
    {
      "epoch": 14.85,
      "learning_rate": 1e-05,
      "loss": 1.4205,
      "step": 62170
    },
    {
      "epoch": 14.85,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 62180
    },
    {
      "epoch": 14.85,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 62190
    },
    {
      "epoch": 14.86,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 62200
    },
    {
      "epoch": 14.86,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4111398458480835,
      "eval_runtime": 32.8465,
      "eval_samples_per_second": 20.824,
      "step": 62200
    },
    {
      "epoch": 14.86,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 62210
    },
    {
      "epoch": 14.86,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 62220
    },
    {
      "epoch": 14.86,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 62230
    },
    {
      "epoch": 14.87,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 62240
    },
    {
      "epoch": 14.87,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 62250
    },
    {
      "epoch": 14.87,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 62260
    },
    {
      "epoch": 14.87,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 62270
    },
    {
      "epoch": 14.87,
      "learning_rate": 1e-05,
      "loss": 1.4142,
      "step": 62280
    },
    {
      "epoch": 14.88,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 62290
    },
    {
      "epoch": 14.88,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 62300
    },
    {
      "epoch": 14.88,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4125407934188843,
      "eval_runtime": 32.8315,
      "eval_samples_per_second": 20.834,
      "step": 62300
    },
    {
      "epoch": 14.88,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 62310
    },
    {
      "epoch": 14.88,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 62320
    },
    {
      "epoch": 14.89,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 62330
    },
    {
      "epoch": 14.89,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 62340
    },
    {
      "epoch": 14.89,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 62350
    },
    {
      "epoch": 14.89,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 62360
    },
    {
      "epoch": 14.9,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 62370
    },
    {
      "epoch": 14.9,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 62380
    },
    {
      "epoch": 14.9,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 62390
    },
    {
      "epoch": 14.9,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 62400
    },
    {
      "epoch": 14.9,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4110536575317383,
      "eval_runtime": 32.7332,
      "eval_samples_per_second": 20.896,
      "step": 62400
    },
    {
      "epoch": 14.91,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 62410
    },
    {
      "epoch": 14.91,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 62420
    },
    {
      "epoch": 14.91,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 62430
    },
    {
      "epoch": 14.91,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 62440
    },
    {
      "epoch": 14.92,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 62450
    },
    {
      "epoch": 14.92,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 62460
    },
    {
      "epoch": 14.92,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 62470
    },
    {
      "epoch": 14.92,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 62480
    },
    {
      "epoch": 14.92,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 62490
    },
    {
      "epoch": 14.93,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 62500
    },
    {
      "epoch": 14.93,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.409896731376648,
      "eval_runtime": 32.8423,
      "eval_samples_per_second": 20.827,
      "step": 62500
    },
    {
      "epoch": 14.93,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 62510
    },
    {
      "epoch": 14.93,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 62520
    },
    {
      "epoch": 14.93,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 62530
    },
    {
      "epoch": 14.94,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 62540
    },
    {
      "epoch": 14.94,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 62550
    },
    {
      "epoch": 14.94,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 62560
    },
    {
      "epoch": 14.94,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 62570
    },
    {
      "epoch": 14.95,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 62580
    },
    {
      "epoch": 14.95,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 62590
    },
    {
      "epoch": 14.95,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 62600
    },
    {
      "epoch": 14.95,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4107626676559448,
      "eval_runtime": 32.3542,
      "eval_samples_per_second": 21.141,
      "step": 62600
    },
    {
      "epoch": 14.95,
      "learning_rate": 1e-05,
      "loss": 1.4163,
      "step": 62610
    },
    {
      "epoch": 14.96,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 62620
    },
    {
      "epoch": 14.96,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 62630
    },
    {
      "epoch": 14.96,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 62640
    },
    {
      "epoch": 14.96,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 62650
    },
    {
      "epoch": 14.97,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 62660
    },
    {
      "epoch": 14.97,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 62670
    },
    {
      "epoch": 14.97,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 62680
    },
    {
      "epoch": 14.97,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 62690
    },
    {
      "epoch": 14.97,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 62700
    },
    {
      "epoch": 14.97,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4110610485076904,
      "eval_runtime": 32.6096,
      "eval_samples_per_second": 20.975,
      "step": 62700
    },
    {
      "epoch": 14.98,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 62710
    },
    {
      "epoch": 14.98,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 62720
    },
    {
      "epoch": 14.98,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 62730
    },
    {
      "epoch": 14.98,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 62740
    },
    {
      "epoch": 14.99,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 62750
    },
    {
      "epoch": 14.99,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 62760
    },
    {
      "epoch": 14.99,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 62770
    },
    {
      "epoch": 14.99,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 62780
    },
    {
      "epoch": 15.0,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 62790
    },
    {
      "epoch": 15.0,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 62800
    },
    {
      "epoch": 15.0,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4105955362319946,
      "eval_runtime": 32.6268,
      "eval_samples_per_second": 20.964,
      "step": 62800
    },
    {
      "epoch": 15.0,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 62810
    },
    {
      "epoch": 15.0,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 62820
    },
    {
      "epoch": 15.01,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 62830
    },
    {
      "epoch": 15.01,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 62840
    },
    {
      "epoch": 15.01,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 62850
    },
    {
      "epoch": 15.01,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 62860
    },
    {
      "epoch": 15.02,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 62870
    },
    {
      "epoch": 15.02,
      "learning_rate": 1e-05,
      "loss": 1.413,
      "step": 62880
    },
    {
      "epoch": 15.02,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 62890
    },
    {
      "epoch": 15.02,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 62900
    },
    {
      "epoch": 15.02,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4107818603515625,
      "eval_runtime": 32.2059,
      "eval_samples_per_second": 21.238,
      "step": 62900
    },
    {
      "epoch": 15.03,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 62910
    },
    {
      "epoch": 15.03,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 62920
    },
    {
      "epoch": 15.03,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 62930
    },
    {
      "epoch": 15.03,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 62940
    },
    {
      "epoch": 15.03,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 62950
    },
    {
      "epoch": 15.04,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 62960
    },
    {
      "epoch": 15.04,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 62970
    },
    {
      "epoch": 15.04,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 62980
    },
    {
      "epoch": 15.04,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 62990
    },
    {
      "epoch": 15.05,
      "learning_rate": 1e-05,
      "loss": 1.4136,
      "step": 63000
    },
    {
      "epoch": 15.05,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4107779264450073,
      "eval_runtime": 32.6441,
      "eval_samples_per_second": 20.953,
      "step": 63000
    },
    {
      "epoch": 15.05,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 63010
    },
    {
      "epoch": 15.05,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 63020
    },
    {
      "epoch": 15.05,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 63030
    },
    {
      "epoch": 15.06,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 63040
    },
    {
      "epoch": 15.06,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 63050
    },
    {
      "epoch": 15.06,
      "learning_rate": 1e-05,
      "loss": 1.413,
      "step": 63060
    },
    {
      "epoch": 15.06,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 63070
    },
    {
      "epoch": 15.07,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 63080
    },
    {
      "epoch": 15.07,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 63090
    },
    {
      "epoch": 15.07,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 63100
    },
    {
      "epoch": 15.07,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4107062816619873,
      "eval_runtime": 32.2665,
      "eval_samples_per_second": 21.198,
      "step": 63100
    },
    {
      "epoch": 15.07,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 63110
    },
    {
      "epoch": 15.08,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 63120
    },
    {
      "epoch": 15.08,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 63130
    },
    {
      "epoch": 15.08,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 63140
    },
    {
      "epoch": 15.08,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 63150
    },
    {
      "epoch": 15.08,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 63160
    },
    {
      "epoch": 15.09,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 63170
    },
    {
      "epoch": 15.09,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 63180
    },
    {
      "epoch": 15.09,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 63190
    },
    {
      "epoch": 15.09,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 63200
    },
    {
      "epoch": 15.09,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4098058938980103,
      "eval_runtime": 32.662,
      "eval_samples_per_second": 20.942,
      "step": 63200
    },
    {
      "epoch": 15.1,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 63210
    },
    {
      "epoch": 15.1,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 63220
    },
    {
      "epoch": 15.1,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 63230
    },
    {
      "epoch": 15.1,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 63240
    },
    {
      "epoch": 15.11,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 63250
    },
    {
      "epoch": 15.11,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 63260
    },
    {
      "epoch": 15.11,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 63270
    },
    {
      "epoch": 15.11,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 63280
    },
    {
      "epoch": 15.12,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 63290
    },
    {
      "epoch": 15.12,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 63300
    },
    {
      "epoch": 15.12,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4108083248138428,
      "eval_runtime": 32.5158,
      "eval_samples_per_second": 21.036,
      "step": 63300
    },
    {
      "epoch": 15.12,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 63310
    },
    {
      "epoch": 15.12,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 63320
    },
    {
      "epoch": 15.13,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 63330
    },
    {
      "epoch": 15.13,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 63340
    },
    {
      "epoch": 15.13,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 63350
    },
    {
      "epoch": 15.13,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 63360
    },
    {
      "epoch": 15.13,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 63370
    },
    {
      "epoch": 15.14,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 63380
    },
    {
      "epoch": 15.14,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 63390
    },
    {
      "epoch": 15.14,
      "learning_rate": 1e-05,
      "loss": 1.413,
      "step": 63400
    },
    {
      "epoch": 15.14,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4107009172439575,
      "eval_runtime": 32.5803,
      "eval_samples_per_second": 20.994,
      "step": 63400
    },
    {
      "epoch": 15.14,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 63410
    },
    {
      "epoch": 15.15,
      "learning_rate": 1e-05,
      "loss": 1.421,
      "step": 63420
    },
    {
      "epoch": 15.15,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 63430
    },
    {
      "epoch": 15.15,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 63440
    },
    {
      "epoch": 15.15,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 63450
    },
    {
      "epoch": 15.16,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 63460
    },
    {
      "epoch": 15.16,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 63470
    },
    {
      "epoch": 15.16,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 63480
    },
    {
      "epoch": 15.16,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 63490
    },
    {
      "epoch": 15.17,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 63500
    },
    {
      "epoch": 15.17,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4104158878326416,
      "eval_runtime": 32.5026,
      "eval_samples_per_second": 21.044,
      "step": 63500
    },
    {
      "epoch": 15.17,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 63510
    },
    {
      "epoch": 15.17,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 63520
    },
    {
      "epoch": 15.17,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 63530
    },
    {
      "epoch": 15.18,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 63540
    },
    {
      "epoch": 15.18,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 63550
    },
    {
      "epoch": 15.18,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 63560
    },
    {
      "epoch": 15.18,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 63570
    },
    {
      "epoch": 15.19,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 63580
    },
    {
      "epoch": 15.19,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 63590
    },
    {
      "epoch": 15.19,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 63600
    },
    {
      "epoch": 15.19,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4109668731689453,
      "eval_runtime": 32.313,
      "eval_samples_per_second": 21.168,
      "step": 63600
    },
    {
      "epoch": 15.19,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 63610
    },
    {
      "epoch": 15.19,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 63620
    },
    {
      "epoch": 15.2,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 63630
    },
    {
      "epoch": 15.2,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 63640
    },
    {
      "epoch": 15.2,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 63650
    },
    {
      "epoch": 15.2,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 63660
    },
    {
      "epoch": 15.21,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 63670
    },
    {
      "epoch": 15.21,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 63680
    },
    {
      "epoch": 15.21,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 63690
    },
    {
      "epoch": 15.21,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 63700
    },
    {
      "epoch": 15.21,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.409905195236206,
      "eval_runtime": 32.4644,
      "eval_samples_per_second": 21.069,
      "step": 63700
    },
    {
      "epoch": 15.22,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 63710
    },
    {
      "epoch": 15.22,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 63720
    },
    {
      "epoch": 15.22,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 63730
    },
    {
      "epoch": 15.22,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 63740
    },
    {
      "epoch": 15.23,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 63750
    },
    {
      "epoch": 15.23,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 63760
    },
    {
      "epoch": 15.23,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 63770
    },
    {
      "epoch": 15.23,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 63780
    },
    {
      "epoch": 15.24,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 63790
    },
    {
      "epoch": 15.24,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 63800
    },
    {
      "epoch": 15.24,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4093185663223267,
      "eval_runtime": 32.7352,
      "eval_samples_per_second": 20.895,
      "step": 63800
    },
    {
      "epoch": 15.24,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 63810
    },
    {
      "epoch": 15.24,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 63820
    },
    {
      "epoch": 15.24,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 63830
    },
    {
      "epoch": 15.25,
      "learning_rate": 1e-05,
      "loss": 1.4235,
      "step": 63840
    },
    {
      "epoch": 15.25,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 63850
    },
    {
      "epoch": 15.25,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 63860
    },
    {
      "epoch": 15.25,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 63870
    },
    {
      "epoch": 15.26,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 63880
    },
    {
      "epoch": 15.26,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 63890
    },
    {
      "epoch": 15.26,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 63900
    },
    {
      "epoch": 15.26,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4092108011245728,
      "eval_runtime": 32.5855,
      "eval_samples_per_second": 20.991,
      "step": 63900
    },
    {
      "epoch": 15.26,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 63910
    },
    {
      "epoch": 15.27,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 63920
    },
    {
      "epoch": 15.27,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 63930
    },
    {
      "epoch": 15.27,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 63940
    },
    {
      "epoch": 15.27,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 63950
    },
    {
      "epoch": 15.28,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 63960
    },
    {
      "epoch": 15.28,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 63970
    },
    {
      "epoch": 15.28,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 63980
    },
    {
      "epoch": 15.28,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 63990
    },
    {
      "epoch": 15.29,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 64000
    },
    {
      "epoch": 15.29,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.409868597984314,
      "eval_runtime": 32.7529,
      "eval_samples_per_second": 20.884,
      "step": 64000
    },
    {
      "epoch": 15.29,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 64010
    },
    {
      "epoch": 15.29,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 64020
    },
    {
      "epoch": 15.29,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 64030
    },
    {
      "epoch": 15.29,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 64040
    },
    {
      "epoch": 15.3,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 64050
    },
    {
      "epoch": 15.3,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 64060
    },
    {
      "epoch": 15.3,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 64070
    },
    {
      "epoch": 15.3,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 64080
    },
    {
      "epoch": 15.31,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 64090
    },
    {
      "epoch": 15.31,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 64100
    },
    {
      "epoch": 15.31,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4107277393341064,
      "eval_runtime": 32.7407,
      "eval_samples_per_second": 20.891,
      "step": 64100
    },
    {
      "epoch": 15.31,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 64110
    },
    {
      "epoch": 15.31,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 64120
    },
    {
      "epoch": 15.32,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 64130
    },
    {
      "epoch": 15.32,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 64140
    },
    {
      "epoch": 15.32,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 64150
    },
    {
      "epoch": 15.32,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 64160
    },
    {
      "epoch": 15.33,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 64170
    },
    {
      "epoch": 15.33,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 64180
    },
    {
      "epoch": 15.33,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 64190
    },
    {
      "epoch": 15.33,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 64200
    },
    {
      "epoch": 15.33,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4094059467315674,
      "eval_runtime": 32.7534,
      "eval_samples_per_second": 20.883,
      "step": 64200
    },
    {
      "epoch": 15.34,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 64210
    },
    {
      "epoch": 15.34,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 64220
    },
    {
      "epoch": 15.34,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 64230
    },
    {
      "epoch": 15.34,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 64240
    },
    {
      "epoch": 15.35,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 64250
    },
    {
      "epoch": 15.35,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 64260
    },
    {
      "epoch": 15.35,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 64270
    },
    {
      "epoch": 15.35,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 64280
    },
    {
      "epoch": 15.35,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 64290
    },
    {
      "epoch": 15.36,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 64300
    },
    {
      "epoch": 15.36,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4105119705200195,
      "eval_runtime": 32.5701,
      "eval_samples_per_second": 21.001,
      "step": 64300
    },
    {
      "epoch": 15.36,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 64310
    },
    {
      "epoch": 15.36,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 64320
    },
    {
      "epoch": 15.36,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 64330
    },
    {
      "epoch": 15.37,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 64340
    },
    {
      "epoch": 15.37,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 64350
    },
    {
      "epoch": 15.37,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 64360
    },
    {
      "epoch": 15.37,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 64370
    },
    {
      "epoch": 15.38,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 64380
    },
    {
      "epoch": 15.38,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 64390
    },
    {
      "epoch": 15.38,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 64400
    },
    {
      "epoch": 15.38,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4108392000198364,
      "eval_runtime": 32.5391,
      "eval_samples_per_second": 21.021,
      "step": 64400
    },
    {
      "epoch": 15.38,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 64410
    },
    {
      "epoch": 15.39,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 64420
    },
    {
      "epoch": 15.39,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 64430
    },
    {
      "epoch": 15.39,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 64440
    },
    {
      "epoch": 15.39,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 64450
    },
    {
      "epoch": 15.4,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 64460
    },
    {
      "epoch": 15.4,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 64470
    },
    {
      "epoch": 15.4,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 64480
    },
    {
      "epoch": 15.4,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 64490
    },
    {
      "epoch": 15.4,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 64500
    },
    {
      "epoch": 15.4,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.409493088722229,
      "eval_runtime": 32.3375,
      "eval_samples_per_second": 21.152,
      "step": 64500
    },
    {
      "epoch": 15.41,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 64510
    },
    {
      "epoch": 15.41,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 64520
    },
    {
      "epoch": 15.41,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 64530
    },
    {
      "epoch": 15.41,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 64540
    },
    {
      "epoch": 15.42,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 64550
    },
    {
      "epoch": 15.42,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 64560
    },
    {
      "epoch": 15.42,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 64570
    },
    {
      "epoch": 15.42,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 64580
    },
    {
      "epoch": 15.43,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 64590
    },
    {
      "epoch": 15.43,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 64600
    },
    {
      "epoch": 15.43,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4100251197814941,
      "eval_runtime": 32.7762,
      "eval_samples_per_second": 20.869,
      "step": 64600
    },
    {
      "epoch": 15.43,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 64610
    },
    {
      "epoch": 15.43,
      "learning_rate": 1e-05,
      "loss": 1.4166,
      "step": 64620
    },
    {
      "epoch": 15.44,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 64630
    },
    {
      "epoch": 15.44,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 64640
    },
    {
      "epoch": 15.44,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 64650
    },
    {
      "epoch": 15.44,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 64660
    },
    {
      "epoch": 15.45,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 64670
    },
    {
      "epoch": 15.45,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 64680
    },
    {
      "epoch": 15.45,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 64690
    },
    {
      "epoch": 15.45,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 64700
    },
    {
      "epoch": 15.45,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4104334115982056,
      "eval_runtime": 32.6266,
      "eval_samples_per_second": 20.965,
      "step": 64700
    },
    {
      "epoch": 15.45,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 64710
    },
    {
      "epoch": 15.46,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 64720
    },
    {
      "epoch": 15.46,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 64730
    },
    {
      "epoch": 15.46,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 64740
    },
    {
      "epoch": 15.46,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 64750
    },
    {
      "epoch": 15.47,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 64760
    },
    {
      "epoch": 15.47,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 64770
    },
    {
      "epoch": 15.47,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 64780
    },
    {
      "epoch": 15.47,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 64790
    },
    {
      "epoch": 15.48,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 64800
    },
    {
      "epoch": 15.48,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.410092830657959,
      "eval_runtime": 32.5417,
      "eval_samples_per_second": 21.019,
      "step": 64800
    },
    {
      "epoch": 15.48,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 64810
    },
    {
      "epoch": 15.48,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 64820
    },
    {
      "epoch": 15.48,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 64830
    },
    {
      "epoch": 15.49,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 64840
    },
    {
      "epoch": 15.49,
      "learning_rate": 1e-05,
      "loss": 1.4155,
      "step": 64850
    },
    {
      "epoch": 15.49,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 64860
    },
    {
      "epoch": 15.49,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 64870
    },
    {
      "epoch": 15.5,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 64880
    },
    {
      "epoch": 15.5,
      "learning_rate": 1e-05,
      "loss": 1.4141,
      "step": 64890
    },
    {
      "epoch": 15.5,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 64900
    },
    {
      "epoch": 15.5,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4104114770889282,
      "eval_runtime": 32.8884,
      "eval_samples_per_second": 20.798,
      "step": 64900
    },
    {
      "epoch": 15.5,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 64910
    },
    {
      "epoch": 15.51,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 64920
    },
    {
      "epoch": 15.51,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 64930
    },
    {
      "epoch": 15.51,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 64940
    },
    {
      "epoch": 15.51,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 64950
    },
    {
      "epoch": 15.51,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 64960
    },
    {
      "epoch": 15.52,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 64970
    },
    {
      "epoch": 15.52,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 64980
    },
    {
      "epoch": 15.52,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 64990
    },
    {
      "epoch": 15.52,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 65000
    },
    {
      "epoch": 15.52,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.40932035446167,
      "eval_runtime": 32.7637,
      "eval_samples_per_second": 20.877,
      "step": 65000
    },
    {
      "epoch": 15.53,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 65010
    },
    {
      "epoch": 15.53,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 65020
    },
    {
      "epoch": 15.53,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 65030
    },
    {
      "epoch": 15.53,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 65040
    },
    {
      "epoch": 15.54,
      "learning_rate": 1e-05,
      "loss": 1.4143,
      "step": 65050
    },
    {
      "epoch": 15.54,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 65060
    },
    {
      "epoch": 15.54,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 65070
    },
    {
      "epoch": 15.54,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 65080
    },
    {
      "epoch": 15.55,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 65090
    },
    {
      "epoch": 15.55,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 65100
    },
    {
      "epoch": 15.55,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4099156856536865,
      "eval_runtime": 32.5179,
      "eval_samples_per_second": 21.035,
      "step": 65100
    },
    {
      "epoch": 15.55,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 65110
    },
    {
      "epoch": 15.55,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 65120
    },
    {
      "epoch": 15.56,
      "learning_rate": 1e-05,
      "loss": 1.4137,
      "step": 65130
    },
    {
      "epoch": 15.56,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 65140
    },
    {
      "epoch": 15.56,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 65150
    },
    {
      "epoch": 15.56,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 65160
    },
    {
      "epoch": 15.56,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 65170
    },
    {
      "epoch": 15.57,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 65180
    },
    {
      "epoch": 15.57,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 65190
    },
    {
      "epoch": 15.57,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 65200
    },
    {
      "epoch": 15.57,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4102753400802612,
      "eval_runtime": 32.5437,
      "eval_samples_per_second": 21.018,
      "step": 65200
    },
    {
      "epoch": 15.57,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 65210
    },
    {
      "epoch": 15.58,
      "learning_rate": 1e-05,
      "loss": 1.4147,
      "step": 65220
    },
    {
      "epoch": 15.58,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 65230
    },
    {
      "epoch": 15.58,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 65240
    },
    {
      "epoch": 15.58,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 65250
    },
    {
      "epoch": 15.59,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 65260
    },
    {
      "epoch": 15.59,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 65270
    },
    {
      "epoch": 15.59,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 65280
    },
    {
      "epoch": 15.59,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 65290
    },
    {
      "epoch": 15.6,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 65300
    },
    {
      "epoch": 15.6,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.410377860069275,
      "eval_runtime": 32.4986,
      "eval_samples_per_second": 21.047,
      "step": 65300
    },
    {
      "epoch": 15.6,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 65310
    },
    {
      "epoch": 15.6,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 65320
    },
    {
      "epoch": 15.6,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 65330
    },
    {
      "epoch": 15.61,
      "learning_rate": 1e-05,
      "loss": 1.4169,
      "step": 65340
    },
    {
      "epoch": 15.61,
      "learning_rate": 1e-05,
      "loss": 1.4144,
      "step": 65350
    },
    {
      "epoch": 15.61,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 65360
    },
    {
      "epoch": 15.61,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 65370
    },
    {
      "epoch": 15.61,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 65380
    },
    {
      "epoch": 15.62,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 65390
    },
    {
      "epoch": 15.62,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 65400
    },
    {
      "epoch": 15.62,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4114139080047607,
      "eval_runtime": 32.7289,
      "eval_samples_per_second": 20.899,
      "step": 65400
    },
    {
      "epoch": 15.62,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 65410
    },
    {
      "epoch": 15.62,
      "learning_rate": 1e-05,
      "loss": 1.4137,
      "step": 65420
    },
    {
      "epoch": 15.63,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 65430
    },
    {
      "epoch": 15.63,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 65440
    },
    {
      "epoch": 15.63,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 65450
    },
    {
      "epoch": 15.63,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 65460
    },
    {
      "epoch": 15.64,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 65470
    },
    {
      "epoch": 15.64,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 65480
    },
    {
      "epoch": 15.64,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 65490
    },
    {
      "epoch": 15.64,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 65500
    },
    {
      "epoch": 15.64,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4096927642822266,
      "eval_runtime": 32.403,
      "eval_samples_per_second": 21.109,
      "step": 65500
    },
    {
      "epoch": 15.65,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 65510
    },
    {
      "epoch": 15.65,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 65520
    },
    {
      "epoch": 15.65,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 65530
    },
    {
      "epoch": 15.65,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 65540
    },
    {
      "epoch": 15.66,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 65550
    },
    {
      "epoch": 15.66,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 65560
    },
    {
      "epoch": 15.66,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 65570
    },
    {
      "epoch": 15.66,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 65580
    },
    {
      "epoch": 15.67,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 65590
    },
    {
      "epoch": 15.67,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 65600
    },
    {
      "epoch": 15.67,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4101439714431763,
      "eval_runtime": 32.4446,
      "eval_samples_per_second": 21.082,
      "step": 65600
    },
    {
      "epoch": 15.67,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 65610
    },
    {
      "epoch": 15.67,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 65620
    },
    {
      "epoch": 15.67,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 65630
    },
    {
      "epoch": 15.68,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 65640
    },
    {
      "epoch": 15.68,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 65650
    },
    {
      "epoch": 15.68,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 65660
    },
    {
      "epoch": 15.68,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 65670
    },
    {
      "epoch": 15.69,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 65680
    },
    {
      "epoch": 15.69,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 65690
    },
    {
      "epoch": 15.69,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 65700
    },
    {
      "epoch": 15.69,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.410506248474121,
      "eval_runtime": 32.4385,
      "eval_samples_per_second": 21.086,
      "step": 65700
    },
    {
      "epoch": 15.69,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 65710
    },
    {
      "epoch": 15.7,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 65720
    },
    {
      "epoch": 15.7,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 65730
    },
    {
      "epoch": 15.7,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 65740
    },
    {
      "epoch": 15.7,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 65750
    },
    {
      "epoch": 15.71,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 65760
    },
    {
      "epoch": 15.71,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 65770
    },
    {
      "epoch": 15.71,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 65780
    },
    {
      "epoch": 15.71,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 65790
    },
    {
      "epoch": 15.72,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 65800
    },
    {
      "epoch": 15.72,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4113539457321167,
      "eval_runtime": 32.5756,
      "eval_samples_per_second": 20.997,
      "step": 65800
    },
    {
      "epoch": 15.72,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 65810
    },
    {
      "epoch": 15.72,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 65820
    },
    {
      "epoch": 15.72,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 65830
    },
    {
      "epoch": 15.72,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 65840
    },
    {
      "epoch": 15.73,
      "learning_rate": 1e-05,
      "loss": 1.416,
      "step": 65850
    },
    {
      "epoch": 15.73,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 65860
    },
    {
      "epoch": 15.73,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 65870
    },
    {
      "epoch": 15.73,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 65880
    },
    {
      "epoch": 15.74,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 65890
    },
    {
      "epoch": 15.74,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 65900
    },
    {
      "epoch": 15.74,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4110034704208374,
      "eval_runtime": 32.3667,
      "eval_samples_per_second": 21.133,
      "step": 65900
    },
    {
      "epoch": 15.74,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 65910
    },
    {
      "epoch": 15.74,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 65920
    },
    {
      "epoch": 15.75,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 65930
    },
    {
      "epoch": 15.75,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 65940
    },
    {
      "epoch": 15.75,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 65950
    },
    {
      "epoch": 15.75,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 65960
    },
    {
      "epoch": 15.76,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 65970
    },
    {
      "epoch": 15.76,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 65980
    },
    {
      "epoch": 15.76,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 65990
    },
    {
      "epoch": 15.76,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 66000
    },
    {
      "epoch": 15.76,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4111065864562988,
      "eval_runtime": 32.4605,
      "eval_samples_per_second": 21.072,
      "step": 66000
    },
    {
      "epoch": 15.77,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 66010
    },
    {
      "epoch": 15.77,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 66020
    },
    {
      "epoch": 15.77,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 66030
    },
    {
      "epoch": 15.77,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 66040
    },
    {
      "epoch": 15.78,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 66050
    },
    {
      "epoch": 15.78,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 66060
    },
    {
      "epoch": 15.78,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 66070
    },
    {
      "epoch": 15.78,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 66080
    },
    {
      "epoch": 15.78,
      "learning_rate": 1e-05,
      "loss": 1.4158,
      "step": 66090
    },
    {
      "epoch": 15.79,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 66100
    },
    {
      "epoch": 15.79,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.410598874092102,
      "eval_runtime": 32.2162,
      "eval_samples_per_second": 21.232,
      "step": 66100
    },
    {
      "epoch": 15.79,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 66110
    },
    {
      "epoch": 15.79,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 66120
    },
    {
      "epoch": 15.79,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 66130
    },
    {
      "epoch": 15.8,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 66140
    },
    {
      "epoch": 15.8,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 66150
    },
    {
      "epoch": 15.8,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 66160
    },
    {
      "epoch": 15.8,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 66170
    },
    {
      "epoch": 15.81,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 66180
    },
    {
      "epoch": 15.81,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 66190
    },
    {
      "epoch": 15.81,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 66200
    },
    {
      "epoch": 15.81,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4091147184371948,
      "eval_runtime": 32.683,
      "eval_samples_per_second": 20.928,
      "step": 66200
    },
    {
      "epoch": 15.81,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 66210
    },
    {
      "epoch": 15.82,
      "learning_rate": 1e-05,
      "loss": 1.4171,
      "step": 66220
    },
    {
      "epoch": 15.82,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 66230
    },
    {
      "epoch": 15.82,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 66240
    },
    {
      "epoch": 15.82,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 66250
    },
    {
      "epoch": 15.83,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 66260
    },
    {
      "epoch": 15.83,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 66270
    },
    {
      "epoch": 15.83,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 66280
    },
    {
      "epoch": 15.83,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 66290
    },
    {
      "epoch": 15.83,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 66300
    },
    {
      "epoch": 15.83,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.41048002243042,
      "eval_runtime": 32.6707,
      "eval_samples_per_second": 20.936,
      "step": 66300
    },
    {
      "epoch": 15.84,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 66310
    },
    {
      "epoch": 15.84,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 66320
    },
    {
      "epoch": 15.84,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 66330
    },
    {
      "epoch": 15.84,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 66340
    },
    {
      "epoch": 15.85,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 66350
    },
    {
      "epoch": 15.85,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 66360
    },
    {
      "epoch": 15.85,
      "learning_rate": 1e-05,
      "loss": 1.4147,
      "step": 66370
    },
    {
      "epoch": 15.85,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 66380
    },
    {
      "epoch": 15.86,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 66390
    },
    {
      "epoch": 15.86,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 66400
    },
    {
      "epoch": 15.86,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4098267555236816,
      "eval_runtime": 32.4074,
      "eval_samples_per_second": 21.106,
      "step": 66400
    },
    {
      "epoch": 15.86,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 66410
    },
    {
      "epoch": 15.86,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 66420
    },
    {
      "epoch": 15.87,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 66430
    },
    {
      "epoch": 15.87,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 66440
    },
    {
      "epoch": 15.87,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 66450
    },
    {
      "epoch": 15.87,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 66460
    },
    {
      "epoch": 15.88,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 66470
    },
    {
      "epoch": 15.88,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 66480
    },
    {
      "epoch": 15.88,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 66490
    },
    {
      "epoch": 15.88,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 66500
    },
    {
      "epoch": 15.88,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.409828782081604,
      "eval_runtime": 32.4765,
      "eval_samples_per_second": 21.061,
      "step": 66500
    },
    {
      "epoch": 15.88,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 66510
    },
    {
      "epoch": 15.89,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 66520
    },
    {
      "epoch": 15.89,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 66530
    },
    {
      "epoch": 15.89,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 66540
    },
    {
      "epoch": 15.89,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 66550
    },
    {
      "epoch": 15.9,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 66560
    },
    {
      "epoch": 15.9,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 66570
    },
    {
      "epoch": 15.9,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 66580
    },
    {
      "epoch": 15.9,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 66590
    },
    {
      "epoch": 15.91,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 66600
    },
    {
      "epoch": 15.91,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4099314212799072,
      "eval_runtime": 32.4955,
      "eval_samples_per_second": 21.049,
      "step": 66600
    },
    {
      "epoch": 15.91,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 66610
    },
    {
      "epoch": 15.91,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 66620
    },
    {
      "epoch": 15.91,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 66630
    },
    {
      "epoch": 15.92,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 66640
    },
    {
      "epoch": 15.92,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 66650
    },
    {
      "epoch": 15.92,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 66660
    },
    {
      "epoch": 15.92,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 66670
    },
    {
      "epoch": 15.93,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 66680
    },
    {
      "epoch": 15.93,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 66690
    },
    {
      "epoch": 15.93,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 66700
    },
    {
      "epoch": 15.93,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4088943004608154,
      "eval_runtime": 32.1511,
      "eval_samples_per_second": 21.275,
      "step": 66700
    },
    {
      "epoch": 15.93,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 66710
    },
    {
      "epoch": 15.94,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 66720
    },
    {
      "epoch": 15.94,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 66730
    },
    {
      "epoch": 15.94,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 66740
    },
    {
      "epoch": 15.94,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 66750
    },
    {
      "epoch": 15.94,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 66760
    },
    {
      "epoch": 15.95,
      "learning_rate": 1e-05,
      "loss": 1.4155,
      "step": 66770
    },
    {
      "epoch": 15.95,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 66780
    },
    {
      "epoch": 15.95,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 66790
    },
    {
      "epoch": 15.95,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 66800
    },
    {
      "epoch": 15.95,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.41061270236969,
      "eval_runtime": 32.4696,
      "eval_samples_per_second": 21.066,
      "step": 66800
    },
    {
      "epoch": 15.96,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 66810
    },
    {
      "epoch": 15.96,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 66820
    },
    {
      "epoch": 15.96,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 66830
    },
    {
      "epoch": 15.96,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 66840
    },
    {
      "epoch": 15.97,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 66850
    },
    {
      "epoch": 15.97,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 66860
    },
    {
      "epoch": 15.97,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 66870
    },
    {
      "epoch": 15.97,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 66880
    },
    {
      "epoch": 15.98,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 66890
    },
    {
      "epoch": 15.98,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 66900
    },
    {
      "epoch": 15.98,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.40971040725708,
      "eval_runtime": 32.7186,
      "eval_samples_per_second": 20.906,
      "step": 66900
    },
    {
      "epoch": 15.98,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 66910
    },
    {
      "epoch": 15.98,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 66920
    },
    {
      "epoch": 15.99,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 66930
    },
    {
      "epoch": 15.99,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 66940
    },
    {
      "epoch": 15.99,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 66950
    },
    {
      "epoch": 15.99,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 66960
    },
    {
      "epoch": 15.99,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 66970
    },
    {
      "epoch": 16.0,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 66980
    },
    {
      "epoch": 16.0,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 66990
    },
    {
      "epoch": 16.0,
      "learning_rate": 1e-05,
      "loss": 1.4173,
      "step": 67000
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4096649885177612,
      "eval_runtime": 33.6136,
      "eval_samples_per_second": 20.349,
      "step": 67000
    },
    {
      "epoch": 16.0,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 67010
    },
    {
      "epoch": 16.01,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 67020
    },
    {
      "epoch": 16.01,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 67030
    },
    {
      "epoch": 16.01,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 67040
    },
    {
      "epoch": 16.01,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 67050
    },
    {
      "epoch": 16.02,
      "learning_rate": 1e-05,
      "loss": 1.4149,
      "step": 67060
    },
    {
      "epoch": 16.02,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 67070
    },
    {
      "epoch": 16.02,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 67080
    },
    {
      "epoch": 16.02,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 67090
    },
    {
      "epoch": 16.03,
      "learning_rate": 1e-05,
      "loss": 1.4118,
      "step": 67100
    },
    {
      "epoch": 16.03,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4094399213790894,
      "eval_runtime": 32.3483,
      "eval_samples_per_second": 21.145,
      "step": 67100
    },
    {
      "epoch": 16.03,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 67110
    },
    {
      "epoch": 16.03,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 67120
    },
    {
      "epoch": 16.03,
      "learning_rate": 1e-05,
      "loss": 1.4155,
      "step": 67130
    },
    {
      "epoch": 16.04,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 67140
    },
    {
      "epoch": 16.04,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 67150
    },
    {
      "epoch": 16.04,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 67160
    },
    {
      "epoch": 16.04,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 67170
    },
    {
      "epoch": 16.04,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 67180
    },
    {
      "epoch": 16.05,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 67190
    },
    {
      "epoch": 16.05,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 67200
    },
    {
      "epoch": 16.05,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4092910289764404,
      "eval_runtime": 32.7048,
      "eval_samples_per_second": 20.914,
      "step": 67200
    },
    {
      "epoch": 16.05,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 67210
    },
    {
      "epoch": 16.05,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 67220
    },
    {
      "epoch": 16.06,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 67230
    },
    {
      "epoch": 16.06,
      "learning_rate": 1e-05,
      "loss": 1.4152,
      "step": 67240
    },
    {
      "epoch": 16.06,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 67250
    },
    {
      "epoch": 16.06,
      "learning_rate": 1e-05,
      "loss": 1.4172,
      "step": 67260
    },
    {
      "epoch": 16.07,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 67270
    },
    {
      "epoch": 16.07,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 67280
    },
    {
      "epoch": 16.07,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 67290
    },
    {
      "epoch": 16.07,
      "learning_rate": 1e-05,
      "loss": 1.4137,
      "step": 67300
    },
    {
      "epoch": 16.07,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.409538984298706,
      "eval_runtime": 32.1369,
      "eval_samples_per_second": 21.284,
      "step": 67300
    },
    {
      "epoch": 16.08,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 67310
    },
    {
      "epoch": 16.08,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 67320
    },
    {
      "epoch": 16.08,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 67330
    },
    {
      "epoch": 16.08,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 67340
    },
    {
      "epoch": 16.09,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 67350
    },
    {
      "epoch": 16.09,
      "learning_rate": 1e-05,
      "loss": 1.4136,
      "step": 67360
    },
    {
      "epoch": 16.09,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 67370
    },
    {
      "epoch": 16.09,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 67380
    },
    {
      "epoch": 16.1,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 67390
    },
    {
      "epoch": 16.1,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 67400
    },
    {
      "epoch": 16.1,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4092941284179688,
      "eval_runtime": 32.5338,
      "eval_samples_per_second": 21.024,
      "step": 67400
    },
    {
      "epoch": 16.1,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 67410
    },
    {
      "epoch": 16.1,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 67420
    },
    {
      "epoch": 16.1,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 67430
    },
    {
      "epoch": 16.11,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 67440
    },
    {
      "epoch": 16.11,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 67450
    },
    {
      "epoch": 16.11,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 67460
    },
    {
      "epoch": 16.11,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 67470
    },
    {
      "epoch": 16.12,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 67480
    },
    {
      "epoch": 16.12,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 67490
    },
    {
      "epoch": 16.12,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 67500
    },
    {
      "epoch": 16.12,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4090691804885864,
      "eval_runtime": 32.8805,
      "eval_samples_per_second": 20.803,
      "step": 67500
    },
    {
      "epoch": 16.12,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 67510
    },
    {
      "epoch": 16.13,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 67520
    },
    {
      "epoch": 16.13,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 67530
    },
    {
      "epoch": 16.13,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 67540
    },
    {
      "epoch": 16.13,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 67550
    },
    {
      "epoch": 16.14,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 67560
    },
    {
      "epoch": 16.14,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 67570
    },
    {
      "epoch": 16.14,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 67580
    },
    {
      "epoch": 16.14,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 67590
    },
    {
      "epoch": 16.15,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 67600
    },
    {
      "epoch": 16.15,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4090977907180786,
      "eval_runtime": 32.3075,
      "eval_samples_per_second": 21.172,
      "step": 67600
    },
    {
      "epoch": 16.15,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 67610
    },
    {
      "epoch": 16.15,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 67620
    },
    {
      "epoch": 16.15,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 67630
    },
    {
      "epoch": 16.15,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 67640
    },
    {
      "epoch": 16.16,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 67650
    },
    {
      "epoch": 16.16,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 67660
    },
    {
      "epoch": 16.16,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 67670
    },
    {
      "epoch": 16.16,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 67680
    },
    {
      "epoch": 16.17,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 67690
    },
    {
      "epoch": 16.17,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 67700
    },
    {
      "epoch": 16.17,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4089992046356201,
      "eval_runtime": 32.4233,
      "eval_samples_per_second": 21.096,
      "step": 67700
    },
    {
      "epoch": 16.17,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 67710
    },
    {
      "epoch": 16.17,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 67720
    },
    {
      "epoch": 16.18,
      "learning_rate": 1e-05,
      "loss": 1.4168,
      "step": 67730
    },
    {
      "epoch": 16.18,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 67740
    },
    {
      "epoch": 16.18,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 67750
    },
    {
      "epoch": 16.18,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 67760
    },
    {
      "epoch": 16.19,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 67770
    },
    {
      "epoch": 16.19,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 67780
    },
    {
      "epoch": 16.19,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 67790
    },
    {
      "epoch": 16.19,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 67800
    },
    {
      "epoch": 16.19,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4093211889266968,
      "eval_runtime": 32.3353,
      "eval_samples_per_second": 21.153,
      "step": 67800
    },
    {
      "epoch": 16.2,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 67810
    },
    {
      "epoch": 16.2,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 67820
    },
    {
      "epoch": 16.2,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 67830
    },
    {
      "epoch": 16.2,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 67840
    },
    {
      "epoch": 16.2,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 67850
    },
    {
      "epoch": 16.21,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 67860
    },
    {
      "epoch": 16.21,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 67870
    },
    {
      "epoch": 16.21,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 67880
    },
    {
      "epoch": 16.21,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 67890
    },
    {
      "epoch": 16.22,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 67900
    },
    {
      "epoch": 16.22,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4086116552352905,
      "eval_runtime": 32.5715,
      "eval_samples_per_second": 21.0,
      "step": 67900
    },
    {
      "epoch": 16.22,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 67910
    },
    {
      "epoch": 16.22,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 67920
    },
    {
      "epoch": 16.22,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 67930
    },
    {
      "epoch": 16.23,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 67940
    },
    {
      "epoch": 16.23,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 67950
    },
    {
      "epoch": 16.23,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 67960
    },
    {
      "epoch": 16.23,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 67970
    },
    {
      "epoch": 16.24,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 67980
    },
    {
      "epoch": 16.24,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 67990
    },
    {
      "epoch": 16.24,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 68000
    },
    {
      "epoch": 16.24,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.409299373626709,
      "eval_runtime": 32.2629,
      "eval_samples_per_second": 21.201,
      "step": 68000
    },
    {
      "epoch": 16.24,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 68010
    },
    {
      "epoch": 16.25,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 68020
    },
    {
      "epoch": 16.25,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 68030
    },
    {
      "epoch": 16.25,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 68040
    },
    {
      "epoch": 16.25,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 68050
    },
    {
      "epoch": 16.26,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 68060
    },
    {
      "epoch": 16.26,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 68070
    },
    {
      "epoch": 16.26,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 68080
    },
    {
      "epoch": 16.26,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 68090
    },
    {
      "epoch": 16.26,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 68100
    },
    {
      "epoch": 16.26,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.409034013748169,
      "eval_runtime": 32.1107,
      "eval_samples_per_second": 21.301,
      "step": 68100
    },
    {
      "epoch": 16.27,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 68110
    },
    {
      "epoch": 16.27,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 68120
    },
    {
      "epoch": 16.27,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 68130
    },
    {
      "epoch": 16.27,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 68140
    },
    {
      "epoch": 16.28,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 68150
    },
    {
      "epoch": 16.28,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 68160
    },
    {
      "epoch": 16.28,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 68170
    },
    {
      "epoch": 16.28,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 68180
    },
    {
      "epoch": 16.29,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 68190
    },
    {
      "epoch": 16.29,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 68200
    },
    {
      "epoch": 16.29,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4093016386032104,
      "eval_runtime": 32.5904,
      "eval_samples_per_second": 20.988,
      "step": 68200
    },
    {
      "epoch": 16.29,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 68210
    },
    {
      "epoch": 16.29,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 68220
    },
    {
      "epoch": 16.3,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 68230
    },
    {
      "epoch": 16.3,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 68240
    },
    {
      "epoch": 16.3,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 68250
    },
    {
      "epoch": 16.3,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 68260
    },
    {
      "epoch": 16.31,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 68270
    },
    {
      "epoch": 16.31,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 68280
    },
    {
      "epoch": 16.31,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 68290
    },
    {
      "epoch": 16.31,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 68300
    },
    {
      "epoch": 16.31,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4097189903259277,
      "eval_runtime": 32.3704,
      "eval_samples_per_second": 21.13,
      "step": 68300
    },
    {
      "epoch": 16.31,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 68310
    },
    {
      "epoch": 16.32,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 68320
    },
    {
      "epoch": 16.32,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 68330
    },
    {
      "epoch": 16.32,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 68340
    },
    {
      "epoch": 16.32,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 68350
    },
    {
      "epoch": 16.33,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 68360
    },
    {
      "epoch": 16.33,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 68370
    },
    {
      "epoch": 16.33,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 68380
    },
    {
      "epoch": 16.33,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 68390
    },
    {
      "epoch": 16.34,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 68400
    },
    {
      "epoch": 16.34,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4088335037231445,
      "eval_runtime": 32.6356,
      "eval_samples_per_second": 20.959,
      "step": 68400
    },
    {
      "epoch": 16.34,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 68410
    },
    {
      "epoch": 16.34,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 68420
    },
    {
      "epoch": 16.34,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 68430
    },
    {
      "epoch": 16.35,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 68440
    },
    {
      "epoch": 16.35,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 68450
    },
    {
      "epoch": 16.35,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 68460
    },
    {
      "epoch": 16.35,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 68470
    },
    {
      "epoch": 16.36,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 68480
    },
    {
      "epoch": 16.36,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 68490
    },
    {
      "epoch": 16.36,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 68500
    },
    {
      "epoch": 16.36,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4086474180221558,
      "eval_runtime": 32.6484,
      "eval_samples_per_second": 20.95,
      "step": 68500
    },
    {
      "epoch": 16.36,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 68510
    },
    {
      "epoch": 16.36,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 68520
    },
    {
      "epoch": 16.37,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 68530
    },
    {
      "epoch": 16.37,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 68540
    },
    {
      "epoch": 16.37,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 68550
    },
    {
      "epoch": 16.37,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 68560
    },
    {
      "epoch": 16.38,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 68570
    },
    {
      "epoch": 16.38,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 68580
    },
    {
      "epoch": 16.38,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 68590
    },
    {
      "epoch": 16.38,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 68600
    },
    {
      "epoch": 16.38,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4088088274002075,
      "eval_runtime": 32.6548,
      "eval_samples_per_second": 20.946,
      "step": 68600
    },
    {
      "epoch": 16.39,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 68610
    },
    {
      "epoch": 16.39,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 68620
    },
    {
      "epoch": 16.39,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 68630
    },
    {
      "epoch": 16.39,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 68640
    },
    {
      "epoch": 16.4,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 68650
    },
    {
      "epoch": 16.4,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 68660
    },
    {
      "epoch": 16.4,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 68670
    },
    {
      "epoch": 16.4,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 68680
    },
    {
      "epoch": 16.41,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 68690
    },
    {
      "epoch": 16.41,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 68700
    },
    {
      "epoch": 16.41,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4095522165298462,
      "eval_runtime": 32.9958,
      "eval_samples_per_second": 20.73,
      "step": 68700
    },
    {
      "epoch": 16.41,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 68710
    },
    {
      "epoch": 16.41,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 68720
    },
    {
      "epoch": 16.42,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 68730
    },
    {
      "epoch": 16.42,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 68740
    },
    {
      "epoch": 16.42,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 68750
    },
    {
      "epoch": 16.42,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 68760
    },
    {
      "epoch": 16.42,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 68770
    },
    {
      "epoch": 16.43,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 68780
    },
    {
      "epoch": 16.43,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 68790
    },
    {
      "epoch": 16.43,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 68800
    },
    {
      "epoch": 16.43,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4076615571975708,
      "eval_runtime": 32.611,
      "eval_samples_per_second": 20.975,
      "step": 68800
    },
    {
      "epoch": 16.43,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 68810
    },
    {
      "epoch": 16.44,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 68820
    },
    {
      "epoch": 16.44,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 68830
    },
    {
      "epoch": 16.44,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 68840
    },
    {
      "epoch": 16.44,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 68850
    },
    {
      "epoch": 16.45,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 68860
    },
    {
      "epoch": 16.45,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 68870
    },
    {
      "epoch": 16.45,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 68880
    },
    {
      "epoch": 16.45,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 68890
    },
    {
      "epoch": 16.46,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 68900
    },
    {
      "epoch": 16.46,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4082424640655518,
      "eval_runtime": 32.8721,
      "eval_samples_per_second": 20.808,
      "step": 68900
    },
    {
      "epoch": 16.46,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 68910
    },
    {
      "epoch": 16.46,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 68920
    },
    {
      "epoch": 16.46,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 68930
    },
    {
      "epoch": 16.47,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 68940
    },
    {
      "epoch": 16.47,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 68950
    },
    {
      "epoch": 16.47,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 68960
    },
    {
      "epoch": 16.47,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 68970
    },
    {
      "epoch": 16.47,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 68980
    },
    {
      "epoch": 16.48,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 68990
    },
    {
      "epoch": 16.48,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 69000
    },
    {
      "epoch": 16.48,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4094878435134888,
      "eval_runtime": 32.5925,
      "eval_samples_per_second": 20.986,
      "step": 69000
    },
    {
      "epoch": 16.48,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 69010
    },
    {
      "epoch": 16.48,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 69020
    },
    {
      "epoch": 16.49,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 69030
    },
    {
      "epoch": 16.49,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 69040
    },
    {
      "epoch": 16.49,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 69050
    },
    {
      "epoch": 16.49,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 69060
    },
    {
      "epoch": 16.5,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 69070
    },
    {
      "epoch": 16.5,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 69080
    },
    {
      "epoch": 16.5,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 69090
    },
    {
      "epoch": 16.5,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 69100
    },
    {
      "epoch": 16.5,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4093513488769531,
      "eval_runtime": 32.4987,
      "eval_samples_per_second": 21.047,
      "step": 69100
    },
    {
      "epoch": 16.51,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 69110
    },
    {
      "epoch": 16.51,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 69120
    },
    {
      "epoch": 16.51,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 69130
    },
    {
      "epoch": 16.51,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 69140
    },
    {
      "epoch": 16.52,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 69150
    },
    {
      "epoch": 16.52,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 69160
    },
    {
      "epoch": 16.52,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 69170
    },
    {
      "epoch": 16.52,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 69180
    },
    {
      "epoch": 16.52,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 69190
    },
    {
      "epoch": 16.53,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 69200
    },
    {
      "epoch": 16.53,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.408308982849121,
      "eval_runtime": 32.404,
      "eval_samples_per_second": 21.108,
      "step": 69200
    },
    {
      "epoch": 16.53,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 69210
    },
    {
      "epoch": 16.53,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 69220
    },
    {
      "epoch": 16.53,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 69230
    },
    {
      "epoch": 16.54,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 69240
    },
    {
      "epoch": 16.54,
      "learning_rate": 1e-05,
      "loss": 1.4113,
      "step": 69250
    },
    {
      "epoch": 16.54,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 69260
    },
    {
      "epoch": 16.54,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 69270
    },
    {
      "epoch": 16.55,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 69280
    },
    {
      "epoch": 16.55,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 69290
    },
    {
      "epoch": 16.55,
      "learning_rate": 1e-05,
      "loss": 1.4155,
      "step": 69300
    },
    {
      "epoch": 16.55,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.408079981803894,
      "eval_runtime": 32.6511,
      "eval_samples_per_second": 20.949,
      "step": 69300
    },
    {
      "epoch": 16.55,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 69310
    },
    {
      "epoch": 16.56,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 69320
    },
    {
      "epoch": 16.56,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 69330
    },
    {
      "epoch": 16.56,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 69340
    },
    {
      "epoch": 16.56,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 69350
    },
    {
      "epoch": 16.57,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 69360
    },
    {
      "epoch": 16.57,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 69370
    },
    {
      "epoch": 16.57,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 69380
    },
    {
      "epoch": 16.57,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 69390
    },
    {
      "epoch": 16.58,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 69400
    },
    {
      "epoch": 16.58,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.408424735069275,
      "eval_runtime": 32.3036,
      "eval_samples_per_second": 21.174,
      "step": 69400
    },
    {
      "epoch": 16.58,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 69410
    },
    {
      "epoch": 16.58,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 69420
    },
    {
      "epoch": 16.58,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 69430
    },
    {
      "epoch": 16.58,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 69440
    },
    {
      "epoch": 16.59,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 69450
    },
    {
      "epoch": 16.59,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 69460
    },
    {
      "epoch": 16.59,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 69470
    },
    {
      "epoch": 16.59,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 69480
    },
    {
      "epoch": 16.6,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 69490
    },
    {
      "epoch": 16.6,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 69500
    },
    {
      "epoch": 16.6,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4088363647460938,
      "eval_runtime": 32.5531,
      "eval_samples_per_second": 21.012,
      "step": 69500
    },
    {
      "epoch": 16.6,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 69510
    },
    {
      "epoch": 16.6,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 69520
    },
    {
      "epoch": 16.61,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 69530
    },
    {
      "epoch": 16.61,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 69540
    },
    {
      "epoch": 16.61,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 69550
    },
    {
      "epoch": 16.61,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 69560
    },
    {
      "epoch": 16.62,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 69570
    },
    {
      "epoch": 16.62,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 69580
    },
    {
      "epoch": 16.62,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 69590
    },
    {
      "epoch": 16.62,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 69600
    },
    {
      "epoch": 16.62,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4083976745605469,
      "eval_runtime": 33.454,
      "eval_samples_per_second": 20.446,
      "step": 69600
    },
    {
      "epoch": 16.63,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 69610
    },
    {
      "epoch": 16.63,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 69620
    },
    {
      "epoch": 16.63,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 69630
    },
    {
      "epoch": 16.63,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 69640
    },
    {
      "epoch": 16.63,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 69650
    },
    {
      "epoch": 16.64,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 69660
    },
    {
      "epoch": 16.64,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 69670
    },
    {
      "epoch": 16.64,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 69680
    },
    {
      "epoch": 16.64,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 69690
    },
    {
      "epoch": 16.65,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 69700
    },
    {
      "epoch": 16.65,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.408567190170288,
      "eval_runtime": 32.4957,
      "eval_samples_per_second": 21.049,
      "step": 69700
    },
    {
      "epoch": 16.65,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 69710
    },
    {
      "epoch": 16.65,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 69720
    },
    {
      "epoch": 16.65,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 69730
    },
    {
      "epoch": 16.66,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 69740
    },
    {
      "epoch": 16.66,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 69750
    },
    {
      "epoch": 16.66,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 69760
    },
    {
      "epoch": 16.66,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 69770
    },
    {
      "epoch": 16.67,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 69780
    },
    {
      "epoch": 16.67,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 69790
    },
    {
      "epoch": 16.67,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 69800
    },
    {
      "epoch": 16.67,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4089163541793823,
      "eval_runtime": 32.5779,
      "eval_samples_per_second": 20.996,
      "step": 69800
    },
    {
      "epoch": 16.67,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 69810
    },
    {
      "epoch": 16.68,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 69820
    },
    {
      "epoch": 16.68,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 69830
    },
    {
      "epoch": 16.68,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 69840
    },
    {
      "epoch": 16.68,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 69850
    },
    {
      "epoch": 16.68,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 69860
    },
    {
      "epoch": 16.69,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 69870
    },
    {
      "epoch": 16.69,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 69880
    },
    {
      "epoch": 16.69,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 69890
    },
    {
      "epoch": 16.69,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 69900
    },
    {
      "epoch": 16.69,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4085941314697266,
      "eval_runtime": 32.5077,
      "eval_samples_per_second": 21.041,
      "step": 69900
    },
    {
      "epoch": 16.7,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 69910
    },
    {
      "epoch": 16.7,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 69920
    },
    {
      "epoch": 16.7,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 69930
    },
    {
      "epoch": 16.7,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 69940
    },
    {
      "epoch": 16.71,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 69950
    },
    {
      "epoch": 16.71,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 69960
    },
    {
      "epoch": 16.71,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 69970
    },
    {
      "epoch": 16.71,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 69980
    },
    {
      "epoch": 16.72,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 69990
    },
    {
      "epoch": 16.72,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 70000
    },
    {
      "epoch": 16.72,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.408667802810669,
      "eval_runtime": 32.2111,
      "eval_samples_per_second": 21.235,
      "step": 70000
    },
    {
      "epoch": 16.72,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 70010
    },
    {
      "epoch": 16.72,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 70020
    },
    {
      "epoch": 16.73,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 70030
    },
    {
      "epoch": 16.73,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 70040
    },
    {
      "epoch": 16.73,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 70050
    },
    {
      "epoch": 16.73,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 70060
    },
    {
      "epoch": 16.74,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 70070
    },
    {
      "epoch": 16.74,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 70080
    },
    {
      "epoch": 16.74,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 70090
    },
    {
      "epoch": 16.74,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 70100
    },
    {
      "epoch": 16.74,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4088702201843262,
      "eval_runtime": 32.6158,
      "eval_samples_per_second": 20.971,
      "step": 70100
    },
    {
      "epoch": 16.74,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 70110
    },
    {
      "epoch": 16.75,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 70120
    },
    {
      "epoch": 16.75,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 70130
    },
    {
      "epoch": 16.75,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 70140
    },
    {
      "epoch": 16.75,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 70150
    },
    {
      "epoch": 16.76,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 70160
    },
    {
      "epoch": 16.76,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 70170
    },
    {
      "epoch": 16.76,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 70180
    },
    {
      "epoch": 16.76,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 70190
    },
    {
      "epoch": 16.77,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 70200
    },
    {
      "epoch": 16.77,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4087518453598022,
      "eval_runtime": 32.4824,
      "eval_samples_per_second": 21.058,
      "step": 70200
    },
    {
      "epoch": 16.77,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 70210
    },
    {
      "epoch": 16.77,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 70220
    },
    {
      "epoch": 16.77,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 70230
    },
    {
      "epoch": 16.78,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 70240
    },
    {
      "epoch": 16.78,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 70250
    },
    {
      "epoch": 16.78,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 70260
    },
    {
      "epoch": 16.78,
      "learning_rate": 1e-05,
      "loss": 1.4116,
      "step": 70270
    },
    {
      "epoch": 16.79,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 70280
    },
    {
      "epoch": 16.79,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 70290
    },
    {
      "epoch": 16.79,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 70300
    },
    {
      "epoch": 16.79,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4079718589782715,
      "eval_runtime": 32.5058,
      "eval_samples_per_second": 21.042,
      "step": 70300
    },
    {
      "epoch": 16.79,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 70310
    },
    {
      "epoch": 16.79,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 70320
    },
    {
      "epoch": 16.8,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 70330
    },
    {
      "epoch": 16.8,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 70340
    },
    {
      "epoch": 16.8,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 70350
    },
    {
      "epoch": 16.8,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 70360
    },
    {
      "epoch": 16.81,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 70370
    },
    {
      "epoch": 16.81,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 70380
    },
    {
      "epoch": 16.81,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 70390
    },
    {
      "epoch": 16.81,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 70400
    },
    {
      "epoch": 16.81,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4090598821640015,
      "eval_runtime": 32.6403,
      "eval_samples_per_second": 20.956,
      "step": 70400
    },
    {
      "epoch": 16.82,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 70410
    },
    {
      "epoch": 16.82,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 70420
    },
    {
      "epoch": 16.82,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 70430
    },
    {
      "epoch": 16.82,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 70440
    },
    {
      "epoch": 16.83,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 70450
    },
    {
      "epoch": 16.83,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 70460
    },
    {
      "epoch": 16.83,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 70470
    },
    {
      "epoch": 16.83,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 70480
    },
    {
      "epoch": 16.84,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 70490
    },
    {
      "epoch": 16.84,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 70500
    },
    {
      "epoch": 16.84,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4087483882904053,
      "eval_runtime": 32.4511,
      "eval_samples_per_second": 21.078,
      "step": 70500
    },
    {
      "epoch": 16.84,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 70510
    },
    {
      "epoch": 16.84,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 70520
    },
    {
      "epoch": 16.84,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 70530
    },
    {
      "epoch": 16.85,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 70540
    },
    {
      "epoch": 16.85,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 70550
    },
    {
      "epoch": 16.85,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 70560
    },
    {
      "epoch": 16.85,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 70570
    },
    {
      "epoch": 16.86,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 70580
    },
    {
      "epoch": 16.86,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 70590
    },
    {
      "epoch": 16.86,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 70600
    },
    {
      "epoch": 16.86,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.40784752368927,
      "eval_runtime": 32.1687,
      "eval_samples_per_second": 21.263,
      "step": 70600
    },
    {
      "epoch": 16.86,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 70610
    },
    {
      "epoch": 16.87,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 70620
    },
    {
      "epoch": 16.87,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 70630
    },
    {
      "epoch": 16.87,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 70640
    },
    {
      "epoch": 16.87,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 70650
    },
    {
      "epoch": 16.88,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 70660
    },
    {
      "epoch": 16.88,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 70670
    },
    {
      "epoch": 16.88,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 70680
    },
    {
      "epoch": 16.88,
      "learning_rate": 1e-05,
      "loss": 1.414,
      "step": 70690
    },
    {
      "epoch": 16.89,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 70700
    },
    {
      "epoch": 16.89,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4082626104354858,
      "eval_runtime": 32.5675,
      "eval_samples_per_second": 21.003,
      "step": 70700
    },
    {
      "epoch": 16.89,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 70710
    },
    {
      "epoch": 16.89,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 70720
    },
    {
      "epoch": 16.89,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 70730
    },
    {
      "epoch": 16.9,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 70740
    },
    {
      "epoch": 16.9,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 70750
    },
    {
      "epoch": 16.9,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 70760
    },
    {
      "epoch": 16.9,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 70770
    },
    {
      "epoch": 16.9,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 70780
    },
    {
      "epoch": 16.91,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 70790
    },
    {
      "epoch": 16.91,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 70800
    },
    {
      "epoch": 16.91,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4078891277313232,
      "eval_runtime": 32.5647,
      "eval_samples_per_second": 21.004,
      "step": 70800
    },
    {
      "epoch": 16.91,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 70810
    },
    {
      "epoch": 16.91,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 70820
    },
    {
      "epoch": 16.92,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 70830
    },
    {
      "epoch": 16.92,
      "learning_rate": 1e-05,
      "loss": 1.4127,
      "step": 70840
    },
    {
      "epoch": 16.92,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 70850
    },
    {
      "epoch": 16.92,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 70860
    },
    {
      "epoch": 16.93,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 70870
    },
    {
      "epoch": 16.93,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 70880
    },
    {
      "epoch": 16.93,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 70890
    },
    {
      "epoch": 16.93,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 70900
    },
    {
      "epoch": 16.93,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4078854322433472,
      "eval_runtime": 32.5171,
      "eval_samples_per_second": 21.035,
      "step": 70900
    },
    {
      "epoch": 16.94,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 70910
    },
    {
      "epoch": 16.94,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 70920
    },
    {
      "epoch": 16.94,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 70930
    },
    {
      "epoch": 16.94,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 70940
    },
    {
      "epoch": 16.95,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 70950
    },
    {
      "epoch": 16.95,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 70960
    },
    {
      "epoch": 16.95,
      "learning_rate": 1e-05,
      "loss": 1.4125,
      "step": 70970
    },
    {
      "epoch": 16.95,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 70980
    },
    {
      "epoch": 16.95,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 70990
    },
    {
      "epoch": 16.96,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 71000
    },
    {
      "epoch": 16.96,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4079331159591675,
      "eval_runtime": 32.643,
      "eval_samples_per_second": 20.954,
      "step": 71000
    },
    {
      "epoch": 16.96,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 71010
    },
    {
      "epoch": 16.96,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 71020
    },
    {
      "epoch": 16.96,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 71030
    },
    {
      "epoch": 16.97,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 71040
    },
    {
      "epoch": 16.97,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 71050
    },
    {
      "epoch": 16.97,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 71060
    },
    {
      "epoch": 16.97,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 71070
    },
    {
      "epoch": 16.98,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 71080
    },
    {
      "epoch": 16.98,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 71090
    },
    {
      "epoch": 16.98,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 71100
    },
    {
      "epoch": 16.98,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4081017971038818,
      "eval_runtime": 32.5066,
      "eval_samples_per_second": 21.042,
      "step": 71100
    },
    {
      "epoch": 16.98,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 71110
    },
    {
      "epoch": 16.99,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 71120
    },
    {
      "epoch": 16.99,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 71130
    },
    {
      "epoch": 16.99,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 71140
    },
    {
      "epoch": 16.99,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 71150
    },
    {
      "epoch": 17.0,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 71160
    },
    {
      "epoch": 17.0,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 71170
    },
    {
      "epoch": 17.0,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 71180
    },
    {
      "epoch": 17.0,
      "learning_rate": 1e-05,
      "loss": 1.4146,
      "step": 71190
    },
    {
      "epoch": 17.01,
      "learning_rate": 1e-05,
      "loss": 1.412,
      "step": 71200
    },
    {
      "epoch": 17.01,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4079827070236206,
      "eval_runtime": 32.8537,
      "eval_samples_per_second": 20.82,
      "step": 71200
    },
    {
      "epoch": 17.01,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 71210
    },
    {
      "epoch": 17.01,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 71220
    },
    {
      "epoch": 17.01,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 71230
    },
    {
      "epoch": 17.01,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 71240
    },
    {
      "epoch": 17.02,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 71250
    },
    {
      "epoch": 17.02,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 71260
    },
    {
      "epoch": 17.02,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 71270
    },
    {
      "epoch": 17.02,
      "learning_rate": 1e-05,
      "loss": 1.4121,
      "step": 71280
    },
    {
      "epoch": 17.03,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 71290
    },
    {
      "epoch": 17.03,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 71300
    },
    {
      "epoch": 17.03,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4080770015716553,
      "eval_runtime": 32.6746,
      "eval_samples_per_second": 20.934,
      "step": 71300
    },
    {
      "epoch": 17.03,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 71310
    },
    {
      "epoch": 17.03,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 71320
    },
    {
      "epoch": 17.04,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 71330
    },
    {
      "epoch": 17.04,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 71340
    },
    {
      "epoch": 17.04,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 71350
    },
    {
      "epoch": 17.04,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 71360
    },
    {
      "epoch": 17.05,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 71370
    },
    {
      "epoch": 17.05,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 71380
    },
    {
      "epoch": 17.05,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 71390
    },
    {
      "epoch": 17.05,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 71400
    },
    {
      "epoch": 17.05,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4078388214111328,
      "eval_runtime": 32.5804,
      "eval_samples_per_second": 20.994,
      "step": 71400
    },
    {
      "epoch": 17.06,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 71410
    },
    {
      "epoch": 17.06,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 71420
    },
    {
      "epoch": 17.06,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 71430
    },
    {
      "epoch": 17.06,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 71440
    },
    {
      "epoch": 17.06,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 71450
    },
    {
      "epoch": 17.07,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 71460
    },
    {
      "epoch": 17.07,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 71470
    },
    {
      "epoch": 17.07,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 71480
    },
    {
      "epoch": 17.07,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 71490
    },
    {
      "epoch": 17.08,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 71500
    },
    {
      "epoch": 17.08,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4077728986740112,
      "eval_runtime": 32.5989,
      "eval_samples_per_second": 20.982,
      "step": 71500
    },
    {
      "epoch": 17.08,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 71510
    },
    {
      "epoch": 17.08,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 71520
    },
    {
      "epoch": 17.08,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 71530
    },
    {
      "epoch": 17.09,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 71540
    },
    {
      "epoch": 17.09,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 71550
    },
    {
      "epoch": 17.09,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 71560
    },
    {
      "epoch": 17.09,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 71570
    },
    {
      "epoch": 17.1,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 71580
    },
    {
      "epoch": 17.1,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 71590
    },
    {
      "epoch": 17.1,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 71600
    },
    {
      "epoch": 17.1,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4086986780166626,
      "eval_runtime": 32.7946,
      "eval_samples_per_second": 20.857,
      "step": 71600
    },
    {
      "epoch": 17.1,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 71610
    },
    {
      "epoch": 17.11,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 71620
    },
    {
      "epoch": 17.11,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 71630
    },
    {
      "epoch": 17.11,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 71640
    },
    {
      "epoch": 17.11,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 71650
    },
    {
      "epoch": 17.11,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 71660
    },
    {
      "epoch": 17.12,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 71670
    },
    {
      "epoch": 17.12,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 71680
    },
    {
      "epoch": 17.12,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 71690
    },
    {
      "epoch": 17.12,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 71700
    },
    {
      "epoch": 17.12,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4081013202667236,
      "eval_runtime": 32.684,
      "eval_samples_per_second": 20.928,
      "step": 71700
    },
    {
      "epoch": 17.13,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 71710
    },
    {
      "epoch": 17.13,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 71720
    },
    {
      "epoch": 17.13,
      "learning_rate": 1e-05,
      "loss": 1.413,
      "step": 71730
    },
    {
      "epoch": 17.13,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 71740
    },
    {
      "epoch": 17.14,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 71750
    },
    {
      "epoch": 17.14,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 71760
    },
    {
      "epoch": 17.14,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 71770
    },
    {
      "epoch": 17.14,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 71780
    },
    {
      "epoch": 17.15,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 71790
    },
    {
      "epoch": 17.15,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 71800
    },
    {
      "epoch": 17.15,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.40831458568573,
      "eval_runtime": 32.6113,
      "eval_samples_per_second": 20.974,
      "step": 71800
    },
    {
      "epoch": 17.15,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 71810
    },
    {
      "epoch": 17.15,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 71820
    },
    {
      "epoch": 17.16,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 71830
    },
    {
      "epoch": 17.16,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 71840
    },
    {
      "epoch": 17.16,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 71850
    },
    {
      "epoch": 17.16,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 71860
    },
    {
      "epoch": 17.17,
      "learning_rate": 1e-05,
      "loss": 1.4126,
      "step": 71870
    },
    {
      "epoch": 17.17,
      "learning_rate": 1e-05,
      "loss": 1.4124,
      "step": 71880
    },
    {
      "epoch": 17.17,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 71890
    },
    {
      "epoch": 17.17,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 71900
    },
    {
      "epoch": 17.17,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4083073139190674,
      "eval_runtime": 32.2207,
      "eval_samples_per_second": 21.229,
      "step": 71900
    },
    {
      "epoch": 17.17,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 71910
    },
    {
      "epoch": 17.18,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 71920
    },
    {
      "epoch": 17.18,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 71930
    },
    {
      "epoch": 17.18,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 71940
    },
    {
      "epoch": 17.18,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 71950
    },
    {
      "epoch": 17.19,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 71960
    },
    {
      "epoch": 17.19,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 71970
    },
    {
      "epoch": 17.19,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 71980
    },
    {
      "epoch": 17.19,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 71990
    },
    {
      "epoch": 17.2,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 72000
    },
    {
      "epoch": 17.2,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4079726934432983,
      "eval_runtime": 32.6162,
      "eval_samples_per_second": 20.971,
      "step": 72000
    },
    {
      "epoch": 17.2,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 72010
    },
    {
      "epoch": 17.2,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 72020
    },
    {
      "epoch": 17.2,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 72030
    },
    {
      "epoch": 17.21,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 72040
    },
    {
      "epoch": 17.21,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 72050
    },
    {
      "epoch": 17.21,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 72060
    },
    {
      "epoch": 17.21,
      "learning_rate": 1e-05,
      "loss": 1.4147,
      "step": 72070
    },
    {
      "epoch": 17.22,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 72080
    },
    {
      "epoch": 17.22,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 72090
    },
    {
      "epoch": 17.22,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 72100
    },
    {
      "epoch": 17.22,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4078065156936646,
      "eval_runtime": 32.2352,
      "eval_samples_per_second": 21.219,
      "step": 72100
    },
    {
      "epoch": 17.22,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 72110
    },
    {
      "epoch": 17.22,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 72120
    },
    {
      "epoch": 17.23,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 72130
    },
    {
      "epoch": 17.23,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 72140
    },
    {
      "epoch": 17.23,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 72150
    },
    {
      "epoch": 17.23,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 72160
    },
    {
      "epoch": 17.24,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 72170
    },
    {
      "epoch": 17.24,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 72180
    },
    {
      "epoch": 17.24,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 72190
    },
    {
      "epoch": 17.24,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 72200
    },
    {
      "epoch": 17.24,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.407731533050537,
      "eval_runtime": 32.245,
      "eval_samples_per_second": 21.213,
      "step": 72200
    },
    {
      "epoch": 17.25,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 72210
    },
    {
      "epoch": 17.25,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 72220
    },
    {
      "epoch": 17.25,
      "learning_rate": 1e-05,
      "loss": 1.4148,
      "step": 72230
    },
    {
      "epoch": 17.25,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 72240
    },
    {
      "epoch": 17.26,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 72250
    },
    {
      "epoch": 17.26,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 72260
    },
    {
      "epoch": 17.26,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 72270
    },
    {
      "epoch": 17.26,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 72280
    },
    {
      "epoch": 17.27,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 72290
    },
    {
      "epoch": 17.27,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 72300
    },
    {
      "epoch": 17.27,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4076803922653198,
      "eval_runtime": 32.8016,
      "eval_samples_per_second": 20.853,
      "step": 72300
    },
    {
      "epoch": 17.27,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 72310
    },
    {
      "epoch": 17.27,
      "learning_rate": 1e-05,
      "loss": 1.4142,
      "step": 72320
    },
    {
      "epoch": 17.27,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 72330
    },
    {
      "epoch": 17.28,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 72340
    },
    {
      "epoch": 17.28,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 72350
    },
    {
      "epoch": 17.28,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 72360
    },
    {
      "epoch": 17.28,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 72370
    },
    {
      "epoch": 17.29,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 72380
    },
    {
      "epoch": 17.29,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 72390
    },
    {
      "epoch": 17.29,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 72400
    },
    {
      "epoch": 17.29,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4082051515579224,
      "eval_runtime": 32.5932,
      "eval_samples_per_second": 20.986,
      "step": 72400
    },
    {
      "epoch": 17.29,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 72410
    },
    {
      "epoch": 17.3,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 72420
    },
    {
      "epoch": 17.3,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 72430
    },
    {
      "epoch": 17.3,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 72440
    },
    {
      "epoch": 17.3,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 72450
    },
    {
      "epoch": 17.31,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 72460
    },
    {
      "epoch": 17.31,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 72470
    },
    {
      "epoch": 17.31,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 72480
    },
    {
      "epoch": 17.31,
      "learning_rate": 1e-05,
      "loss": 1.4108,
      "step": 72490
    },
    {
      "epoch": 17.32,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 72500
    },
    {
      "epoch": 17.32,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4075567722320557,
      "eval_runtime": 32.6113,
      "eval_samples_per_second": 20.974,
      "step": 72500
    },
    {
      "epoch": 17.32,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 72510
    },
    {
      "epoch": 17.32,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 72520
    },
    {
      "epoch": 17.32,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 72530
    },
    {
      "epoch": 17.33,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 72540
    },
    {
      "epoch": 17.33,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 72550
    },
    {
      "epoch": 17.33,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 72560
    },
    {
      "epoch": 17.33,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 72570
    },
    {
      "epoch": 17.33,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 72580
    },
    {
      "epoch": 17.34,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 72590
    },
    {
      "epoch": 17.34,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 72600
    },
    {
      "epoch": 17.34,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4082262516021729,
      "eval_runtime": 32.8133,
      "eval_samples_per_second": 20.845,
      "step": 72600
    },
    {
      "epoch": 17.34,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 72610
    },
    {
      "epoch": 17.34,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 72620
    },
    {
      "epoch": 17.35,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 72630
    },
    {
      "epoch": 17.35,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 72640
    },
    {
      "epoch": 17.35,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 72650
    },
    {
      "epoch": 17.35,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 72660
    },
    {
      "epoch": 17.36,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 72670
    },
    {
      "epoch": 17.36,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 72680
    },
    {
      "epoch": 17.36,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 72690
    },
    {
      "epoch": 17.36,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 72700
    },
    {
      "epoch": 17.36,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4084804058074951,
      "eval_runtime": 32.577,
      "eval_samples_per_second": 20.996,
      "step": 72700
    },
    {
      "epoch": 17.37,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 72710
    },
    {
      "epoch": 17.37,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 72720
    },
    {
      "epoch": 17.37,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 72730
    },
    {
      "epoch": 17.37,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 72740
    },
    {
      "epoch": 17.38,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 72750
    },
    {
      "epoch": 17.38,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 72760
    },
    {
      "epoch": 17.38,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 72770
    },
    {
      "epoch": 17.38,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 72780
    },
    {
      "epoch": 17.38,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 72790
    },
    {
      "epoch": 17.39,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 72800
    },
    {
      "epoch": 17.39,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4078093767166138,
      "eval_runtime": 32.6319,
      "eval_samples_per_second": 20.961,
      "step": 72800
    },
    {
      "epoch": 17.39,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 72810
    },
    {
      "epoch": 17.39,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 72820
    },
    {
      "epoch": 17.39,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 72830
    },
    {
      "epoch": 17.4,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 72840
    },
    {
      "epoch": 17.4,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 72850
    },
    {
      "epoch": 17.4,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 72860
    },
    {
      "epoch": 17.4,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 72870
    },
    {
      "epoch": 17.41,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 72880
    },
    {
      "epoch": 17.41,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 72890
    },
    {
      "epoch": 17.41,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 72900
    },
    {
      "epoch": 17.41,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.40753972530365,
      "eval_runtime": 32.2385,
      "eval_samples_per_second": 21.217,
      "step": 72900
    },
    {
      "epoch": 17.41,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 72910
    },
    {
      "epoch": 17.42,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 72920
    },
    {
      "epoch": 17.42,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 72930
    },
    {
      "epoch": 17.42,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 72940
    },
    {
      "epoch": 17.42,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 72950
    },
    {
      "epoch": 17.43,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 72960
    },
    {
      "epoch": 17.43,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 72970
    },
    {
      "epoch": 17.43,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 72980
    },
    {
      "epoch": 17.43,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 72990
    },
    {
      "epoch": 17.43,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 73000
    },
    {
      "epoch": 17.43,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4077459573745728,
      "eval_runtime": 32.5947,
      "eval_samples_per_second": 20.985,
      "step": 73000
    },
    {
      "epoch": 17.44,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 73010
    },
    {
      "epoch": 17.44,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 73020
    },
    {
      "epoch": 17.44,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 73030
    },
    {
      "epoch": 17.44,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 73040
    },
    {
      "epoch": 17.45,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 73050
    },
    {
      "epoch": 17.45,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 73060
    },
    {
      "epoch": 17.45,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 73070
    },
    {
      "epoch": 17.45,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 73080
    },
    {
      "epoch": 17.46,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 73090
    },
    {
      "epoch": 17.46,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 73100
    },
    {
      "epoch": 17.46,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4073456525802612,
      "eval_runtime": 32.8126,
      "eval_samples_per_second": 20.846,
      "step": 73100
    },
    {
      "epoch": 17.46,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 73110
    },
    {
      "epoch": 17.46,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 73120
    },
    {
      "epoch": 17.47,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 73130
    },
    {
      "epoch": 17.47,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 73140
    },
    {
      "epoch": 17.47,
      "learning_rate": 1e-05,
      "loss": 1.4114,
      "step": 73150
    },
    {
      "epoch": 17.47,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 73160
    },
    {
      "epoch": 17.48,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 73170
    },
    {
      "epoch": 17.48,
      "learning_rate": 1e-05,
      "loss": 1.4403,
      "step": 73180
    },
    {
      "epoch": 17.48,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 73190
    },
    {
      "epoch": 17.48,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 73200
    },
    {
      "epoch": 17.48,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.407393455505371,
      "eval_runtime": 32.5051,
      "eval_samples_per_second": 21.043,
      "step": 73200
    },
    {
      "epoch": 17.49,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 73210
    },
    {
      "epoch": 17.49,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 73220
    },
    {
      "epoch": 17.49,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 73230
    },
    {
      "epoch": 17.49,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 73240
    },
    {
      "epoch": 17.49,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 73250
    },
    {
      "epoch": 17.5,
      "learning_rate": 1e-05,
      "loss": 1.4135,
      "step": 73260
    },
    {
      "epoch": 17.5,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 73270
    },
    {
      "epoch": 17.5,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 73280
    },
    {
      "epoch": 17.5,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 73290
    },
    {
      "epoch": 17.51,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 73300
    },
    {
      "epoch": 17.51,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4076951742172241,
      "eval_runtime": 32.5848,
      "eval_samples_per_second": 20.991,
      "step": 73300
    },
    {
      "epoch": 17.51,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 73310
    },
    {
      "epoch": 17.51,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 73320
    },
    {
      "epoch": 17.51,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 73330
    },
    {
      "epoch": 17.52,
      "learning_rate": 1e-05,
      "loss": 1.4115,
      "step": 73340
    },
    {
      "epoch": 17.52,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 73350
    },
    {
      "epoch": 17.52,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 73360
    },
    {
      "epoch": 17.52,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 73370
    },
    {
      "epoch": 17.53,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 73380
    },
    {
      "epoch": 17.53,
      "learning_rate": 1e-05,
      "loss": 1.4099,
      "step": 73390
    },
    {
      "epoch": 17.53,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 73400
    },
    {
      "epoch": 17.53,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4078466892242432,
      "eval_runtime": 32.4564,
      "eval_samples_per_second": 21.074,
      "step": 73400
    },
    {
      "epoch": 17.53,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 73410
    },
    {
      "epoch": 17.54,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 73420
    },
    {
      "epoch": 17.54,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 73430
    },
    {
      "epoch": 17.54,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 73440
    },
    {
      "epoch": 17.54,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 73450
    },
    {
      "epoch": 17.54,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 73460
    },
    {
      "epoch": 17.55,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 73470
    },
    {
      "epoch": 17.55,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 73480
    },
    {
      "epoch": 17.55,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 73490
    },
    {
      "epoch": 17.55,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 73500
    },
    {
      "epoch": 17.55,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4086544513702393,
      "eval_runtime": 32.4521,
      "eval_samples_per_second": 21.077,
      "step": 73500
    },
    {
      "epoch": 17.56,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 73510
    },
    {
      "epoch": 17.56,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 73520
    },
    {
      "epoch": 17.56,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 73530
    },
    {
      "epoch": 17.56,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 73540
    },
    {
      "epoch": 17.57,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 73550
    },
    {
      "epoch": 17.57,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 73560
    },
    {
      "epoch": 17.57,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 73570
    },
    {
      "epoch": 17.57,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 73580
    },
    {
      "epoch": 17.58,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 73590
    },
    {
      "epoch": 17.58,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 73600
    },
    {
      "epoch": 17.58,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4075180292129517,
      "eval_runtime": 32.5666,
      "eval_samples_per_second": 21.003,
      "step": 73600
    },
    {
      "epoch": 17.58,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 73610
    },
    {
      "epoch": 17.58,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 73620
    },
    {
      "epoch": 17.59,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 73630
    },
    {
      "epoch": 17.59,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 73640
    },
    {
      "epoch": 17.59,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 73650
    },
    {
      "epoch": 17.59,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 73660
    },
    {
      "epoch": 17.59,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 73670
    },
    {
      "epoch": 17.6,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 73680
    },
    {
      "epoch": 17.6,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 73690
    },
    {
      "epoch": 17.6,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 73700
    },
    {
      "epoch": 17.6,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4073693752288818,
      "eval_runtime": 32.6307,
      "eval_samples_per_second": 20.962,
      "step": 73700
    },
    {
      "epoch": 17.6,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 73710
    },
    {
      "epoch": 17.61,
      "learning_rate": 1e-05,
      "loss": 1.4145,
      "step": 73720
    },
    {
      "epoch": 17.61,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 73730
    },
    {
      "epoch": 17.61,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 73740
    },
    {
      "epoch": 17.61,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 73750
    },
    {
      "epoch": 17.62,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 73760
    },
    {
      "epoch": 17.62,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 73770
    },
    {
      "epoch": 17.62,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 73780
    },
    {
      "epoch": 17.62,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 73790
    },
    {
      "epoch": 17.63,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 73800
    },
    {
      "epoch": 17.63,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.407542109489441,
      "eval_runtime": 32.6546,
      "eval_samples_per_second": 20.947,
      "step": 73800
    },
    {
      "epoch": 17.63,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 73810
    },
    {
      "epoch": 17.63,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 73820
    },
    {
      "epoch": 17.63,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 73830
    },
    {
      "epoch": 17.64,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 73840
    },
    {
      "epoch": 17.64,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 73850
    },
    {
      "epoch": 17.64,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 73860
    },
    {
      "epoch": 17.64,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 73870
    },
    {
      "epoch": 17.65,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 73880
    },
    {
      "epoch": 17.65,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 73890
    },
    {
      "epoch": 17.65,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 73900
    },
    {
      "epoch": 17.65,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.407906174659729,
      "eval_runtime": 32.6676,
      "eval_samples_per_second": 20.938,
      "step": 73900
    },
    {
      "epoch": 17.65,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 73910
    },
    {
      "epoch": 17.65,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 73920
    },
    {
      "epoch": 17.66,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 73930
    },
    {
      "epoch": 17.66,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 73940
    },
    {
      "epoch": 17.66,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 73950
    },
    {
      "epoch": 17.66,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 73960
    },
    {
      "epoch": 17.67,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 73970
    },
    {
      "epoch": 17.67,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 73980
    },
    {
      "epoch": 17.67,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 73990
    },
    {
      "epoch": 17.67,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 74000
    },
    {
      "epoch": 17.67,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.407265305519104,
      "eval_runtime": 32.6713,
      "eval_samples_per_second": 20.936,
      "step": 74000
    },
    {
      "epoch": 17.68,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 74010
    },
    {
      "epoch": 17.68,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 74020
    },
    {
      "epoch": 17.68,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 74030
    },
    {
      "epoch": 17.68,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 74040
    },
    {
      "epoch": 17.69,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 74050
    },
    {
      "epoch": 17.69,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 74060
    },
    {
      "epoch": 17.69,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 74070
    },
    {
      "epoch": 17.69,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 74080
    },
    {
      "epoch": 17.7,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 74090
    },
    {
      "epoch": 17.7,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 74100
    },
    {
      "epoch": 17.7,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4077496528625488,
      "eval_runtime": 32.4961,
      "eval_samples_per_second": 21.049,
      "step": 74100
    },
    {
      "epoch": 17.7,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 74110
    },
    {
      "epoch": 17.7,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 74120
    },
    {
      "epoch": 17.7,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 74130
    },
    {
      "epoch": 17.71,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 74140
    },
    {
      "epoch": 17.71,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 74150
    },
    {
      "epoch": 17.71,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 74160
    },
    {
      "epoch": 17.71,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 74170
    },
    {
      "epoch": 17.72,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 74180
    },
    {
      "epoch": 17.72,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 74190
    },
    {
      "epoch": 17.72,
      "learning_rate": 1e-05,
      "loss": 1.4167,
      "step": 74200
    },
    {
      "epoch": 17.72,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4074476957321167,
      "eval_runtime": 32.2518,
      "eval_samples_per_second": 21.208,
      "step": 74200
    },
    {
      "epoch": 17.72,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 74210
    },
    {
      "epoch": 17.73,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 74220
    },
    {
      "epoch": 17.73,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 74230
    },
    {
      "epoch": 17.73,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 74240
    },
    {
      "epoch": 17.73,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 74250
    },
    {
      "epoch": 17.74,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 74260
    },
    {
      "epoch": 17.74,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 74270
    },
    {
      "epoch": 17.74,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 74280
    },
    {
      "epoch": 17.74,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 74290
    },
    {
      "epoch": 17.75,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 74300
    },
    {
      "epoch": 17.75,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4071044921875,
      "eval_runtime": 32.6936,
      "eval_samples_per_second": 20.922,
      "step": 74300
    },
    {
      "epoch": 17.75,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 74310
    },
    {
      "epoch": 17.75,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 74320
    },
    {
      "epoch": 17.75,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 74330
    },
    {
      "epoch": 17.75,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 74340
    },
    {
      "epoch": 17.76,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 74350
    },
    {
      "epoch": 17.76,
      "learning_rate": 1e-05,
      "loss": 1.4098,
      "step": 74360
    },
    {
      "epoch": 17.76,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 74370
    },
    {
      "epoch": 17.76,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 74380
    },
    {
      "epoch": 17.77,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 74390
    },
    {
      "epoch": 17.77,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 74400
    },
    {
      "epoch": 17.77,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4074187278747559,
      "eval_runtime": 32.3922,
      "eval_samples_per_second": 21.116,
      "step": 74400
    },
    {
      "epoch": 17.77,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 74410
    },
    {
      "epoch": 17.77,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 74420
    },
    {
      "epoch": 17.78,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 74430
    },
    {
      "epoch": 17.78,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 74440
    },
    {
      "epoch": 17.78,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 74450
    },
    {
      "epoch": 17.78,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 74460
    },
    {
      "epoch": 17.79,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 74470
    },
    {
      "epoch": 17.79,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 74480
    },
    {
      "epoch": 17.79,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 74490
    },
    {
      "epoch": 17.79,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 74500
    },
    {
      "epoch": 17.79,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4072380065917969,
      "eval_runtime": 32.512,
      "eval_samples_per_second": 21.038,
      "step": 74500
    },
    {
      "epoch": 17.8,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 74510
    },
    {
      "epoch": 17.8,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 74520
    },
    {
      "epoch": 17.8,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 74530
    },
    {
      "epoch": 17.8,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 74540
    },
    {
      "epoch": 17.81,
      "learning_rate": 1e-05,
      "loss": 1.4177,
      "step": 74550
    },
    {
      "epoch": 17.81,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 74560
    },
    {
      "epoch": 17.81,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 74570
    },
    {
      "epoch": 17.81,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 74580
    },
    {
      "epoch": 17.81,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 74590
    },
    {
      "epoch": 17.82,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 74600
    },
    {
      "epoch": 17.82,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.407659888267517,
      "eval_runtime": 32.7261,
      "eval_samples_per_second": 20.901,
      "step": 74600
    },
    {
      "epoch": 17.82,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 74610
    },
    {
      "epoch": 17.82,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 74620
    },
    {
      "epoch": 17.82,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 74630
    },
    {
      "epoch": 17.83,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 74640
    },
    {
      "epoch": 17.83,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 74650
    },
    {
      "epoch": 17.83,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 74660
    },
    {
      "epoch": 17.83,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 74670
    },
    {
      "epoch": 17.84,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 74680
    },
    {
      "epoch": 17.84,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 74690
    },
    {
      "epoch": 17.84,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 74700
    },
    {
      "epoch": 17.84,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4066169261932373,
      "eval_runtime": 32.5935,
      "eval_samples_per_second": 20.986,
      "step": 74700
    },
    {
      "epoch": 17.84,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 74710
    },
    {
      "epoch": 17.85,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 74720
    },
    {
      "epoch": 17.85,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 74730
    },
    {
      "epoch": 17.85,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 74740
    },
    {
      "epoch": 17.85,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 74750
    },
    {
      "epoch": 17.86,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 74760
    },
    {
      "epoch": 17.86,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 74770
    },
    {
      "epoch": 17.86,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 74780
    },
    {
      "epoch": 17.86,
      "learning_rate": 1e-05,
      "loss": 1.4154,
      "step": 74790
    },
    {
      "epoch": 17.86,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 74800
    },
    {
      "epoch": 17.86,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.407680630683899,
      "eval_runtime": 32.3525,
      "eval_samples_per_second": 21.142,
      "step": 74800
    },
    {
      "epoch": 17.87,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 74810
    },
    {
      "epoch": 17.87,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 74820
    },
    {
      "epoch": 17.87,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 74830
    },
    {
      "epoch": 17.87,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 74840
    },
    {
      "epoch": 17.88,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 74850
    },
    {
      "epoch": 17.88,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 74860
    },
    {
      "epoch": 17.88,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 74870
    },
    {
      "epoch": 17.88,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 74880
    },
    {
      "epoch": 17.89,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 74890
    },
    {
      "epoch": 17.89,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 74900
    },
    {
      "epoch": 17.89,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4068771600723267,
      "eval_runtime": 32.7272,
      "eval_samples_per_second": 20.9,
      "step": 74900
    },
    {
      "epoch": 17.89,
      "learning_rate": 1e-05,
      "loss": 1.4129,
      "step": 74910
    },
    {
      "epoch": 17.89,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 74920
    },
    {
      "epoch": 17.9,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 74930
    },
    {
      "epoch": 17.9,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 74940
    },
    {
      "epoch": 17.9,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 74950
    },
    {
      "epoch": 17.9,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 74960
    },
    {
      "epoch": 17.91,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 74970
    },
    {
      "epoch": 17.91,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 74980
    },
    {
      "epoch": 17.91,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 74990
    },
    {
      "epoch": 17.91,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 75000
    },
    {
      "epoch": 17.91,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4065440893173218,
      "eval_runtime": 32.6816,
      "eval_samples_per_second": 20.929,
      "step": 75000
    },
    {
      "epoch": 17.91,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 75010
    },
    {
      "epoch": 17.92,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 75020
    },
    {
      "epoch": 17.92,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 75030
    },
    {
      "epoch": 17.92,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 75040
    },
    {
      "epoch": 17.92,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 75050
    },
    {
      "epoch": 17.93,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 75060
    },
    {
      "epoch": 17.93,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 75070
    },
    {
      "epoch": 17.93,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 75080
    },
    {
      "epoch": 17.93,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 75090
    },
    {
      "epoch": 17.94,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 75100
    },
    {
      "epoch": 17.94,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4071919918060303,
      "eval_runtime": 32.5575,
      "eval_samples_per_second": 21.009,
      "step": 75100
    },
    {
      "epoch": 17.94,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 75110
    },
    {
      "epoch": 17.94,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 75120
    },
    {
      "epoch": 17.94,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 75130
    },
    {
      "epoch": 17.95,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 75140
    },
    {
      "epoch": 17.95,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 75150
    },
    {
      "epoch": 17.95,
      "learning_rate": 1e-05,
      "loss": 1.4122,
      "step": 75160
    },
    {
      "epoch": 17.95,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 75170
    },
    {
      "epoch": 17.96,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 75180
    },
    {
      "epoch": 17.96,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 75190
    },
    {
      "epoch": 17.96,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 75200
    },
    {
      "epoch": 17.96,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.407349705696106,
      "eval_runtime": 32.519,
      "eval_samples_per_second": 21.034,
      "step": 75200
    },
    {
      "epoch": 17.96,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 75210
    },
    {
      "epoch": 17.97,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 75220
    },
    {
      "epoch": 17.97,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 75230
    },
    {
      "epoch": 17.97,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 75240
    },
    {
      "epoch": 17.97,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 75250
    },
    {
      "epoch": 17.97,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 75260
    },
    {
      "epoch": 17.98,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 75270
    },
    {
      "epoch": 17.98,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 75280
    },
    {
      "epoch": 17.98,
      "learning_rate": 1e-05,
      "loss": 1.4132,
      "step": 75290
    },
    {
      "epoch": 17.98,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 75300
    },
    {
      "epoch": 17.98,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.407129168510437,
      "eval_runtime": 32.4643,
      "eval_samples_per_second": 21.069,
      "step": 75300
    },
    {
      "epoch": 17.99,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 75310
    },
    {
      "epoch": 17.99,
      "learning_rate": 1e-05,
      "loss": 1.4117,
      "step": 75320
    },
    {
      "epoch": 17.99,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 75330
    },
    {
      "epoch": 17.99,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 75340
    },
    {
      "epoch": 18.0,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 75350
    },
    {
      "epoch": 18.0,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 75360
    },
    {
      "epoch": 18.0,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 75370
    },
    {
      "epoch": 18.0,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 75380
    },
    {
      "epoch": 18.01,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 75390
    },
    {
      "epoch": 18.01,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 75400
    },
    {
      "epoch": 18.01,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4068926572799683,
      "eval_runtime": 32.5058,
      "eval_samples_per_second": 21.042,
      "step": 75400
    },
    {
      "epoch": 18.01,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 75410
    },
    {
      "epoch": 18.01,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 75420
    },
    {
      "epoch": 18.02,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 75430
    },
    {
      "epoch": 18.02,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 75440
    },
    {
      "epoch": 18.02,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 75450
    },
    {
      "epoch": 18.02,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 75460
    },
    {
      "epoch": 18.02,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 75470
    },
    {
      "epoch": 18.03,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 75480
    },
    {
      "epoch": 18.03,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 75490
    },
    {
      "epoch": 18.03,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 75500
    },
    {
      "epoch": 18.03,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4067552089691162,
      "eval_runtime": 32.3146,
      "eval_samples_per_second": 21.167,
      "step": 75500
    },
    {
      "epoch": 18.03,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 75510
    },
    {
      "epoch": 18.04,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 75520
    },
    {
      "epoch": 18.04,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 75530
    },
    {
      "epoch": 18.04,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 75540
    },
    {
      "epoch": 18.04,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 75550
    },
    {
      "epoch": 18.05,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 75560
    },
    {
      "epoch": 18.05,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 75570
    },
    {
      "epoch": 18.05,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 75580
    },
    {
      "epoch": 18.05,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 75590
    },
    {
      "epoch": 18.06,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 75600
    },
    {
      "epoch": 18.06,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4072239398956299,
      "eval_runtime": 32.6039,
      "eval_samples_per_second": 20.979,
      "step": 75600
    },
    {
      "epoch": 18.06,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 75610
    },
    {
      "epoch": 18.06,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 75620
    },
    {
      "epoch": 18.06,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 75630
    },
    {
      "epoch": 18.07,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 75640
    },
    {
      "epoch": 18.07,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 75650
    },
    {
      "epoch": 18.07,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 75660
    },
    {
      "epoch": 18.07,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 75670
    },
    {
      "epoch": 18.07,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 75680
    },
    {
      "epoch": 18.08,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 75690
    },
    {
      "epoch": 18.08,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 75700
    },
    {
      "epoch": 18.08,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4066331386566162,
      "eval_runtime": 32.2992,
      "eval_samples_per_second": 21.177,
      "step": 75700
    },
    {
      "epoch": 18.08,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 75710
    },
    {
      "epoch": 18.08,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 75720
    },
    {
      "epoch": 18.09,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 75730
    },
    {
      "epoch": 18.09,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 75740
    },
    {
      "epoch": 18.09,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 75750
    },
    {
      "epoch": 18.09,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 75760
    },
    {
      "epoch": 18.1,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 75770
    },
    {
      "epoch": 18.1,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 75780
    },
    {
      "epoch": 18.1,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 75790
    },
    {
      "epoch": 18.1,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 75800
    },
    {
      "epoch": 18.1,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4064147472381592,
      "eval_runtime": 32.6354,
      "eval_samples_per_second": 20.959,
      "step": 75800
    },
    {
      "epoch": 18.11,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 75810
    },
    {
      "epoch": 18.11,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 75820
    },
    {
      "epoch": 18.11,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 75830
    },
    {
      "epoch": 18.11,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 75840
    },
    {
      "epoch": 18.12,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 75850
    },
    {
      "epoch": 18.12,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 75860
    },
    {
      "epoch": 18.12,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 75870
    },
    {
      "epoch": 18.12,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 75880
    },
    {
      "epoch": 18.13,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 75890
    },
    {
      "epoch": 18.13,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 75900
    },
    {
      "epoch": 18.13,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4064769744873047,
      "eval_runtime": 32.286,
      "eval_samples_per_second": 21.186,
      "step": 75900
    },
    {
      "epoch": 18.13,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 75910
    },
    {
      "epoch": 18.13,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 75920
    },
    {
      "epoch": 18.13,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 75930
    },
    {
      "epoch": 18.14,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 75940
    },
    {
      "epoch": 18.14,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 75950
    },
    {
      "epoch": 18.14,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 75960
    },
    {
      "epoch": 18.14,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 75970
    },
    {
      "epoch": 18.15,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 75980
    },
    {
      "epoch": 18.15,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 75990
    },
    {
      "epoch": 18.15,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 76000
    },
    {
      "epoch": 18.15,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.406875491142273,
      "eval_runtime": 32.374,
      "eval_samples_per_second": 21.128,
      "step": 76000
    },
    {
      "epoch": 18.15,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 76010
    },
    {
      "epoch": 18.16,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 76020
    },
    {
      "epoch": 18.16,
      "learning_rate": 1e-05,
      "loss": 1.4077,
      "step": 76030
    },
    {
      "epoch": 18.16,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 76040
    },
    {
      "epoch": 18.16,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 76050
    },
    {
      "epoch": 18.17,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 76060
    },
    {
      "epoch": 18.17,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 76070
    },
    {
      "epoch": 18.17,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 76080
    },
    {
      "epoch": 18.17,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 76090
    },
    {
      "epoch": 18.18,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 76100
    },
    {
      "epoch": 18.18,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4069095849990845,
      "eval_runtime": 32.5025,
      "eval_samples_per_second": 21.045,
      "step": 76100
    },
    {
      "epoch": 18.18,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 76110
    },
    {
      "epoch": 18.18,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 76120
    },
    {
      "epoch": 18.18,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 76130
    },
    {
      "epoch": 18.18,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 76140
    },
    {
      "epoch": 18.19,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 76150
    },
    {
      "epoch": 18.19,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 76160
    },
    {
      "epoch": 18.19,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 76170
    },
    {
      "epoch": 18.19,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 76180
    },
    {
      "epoch": 18.2,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 76190
    },
    {
      "epoch": 18.2,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 76200
    },
    {
      "epoch": 18.2,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4067914485931396,
      "eval_runtime": 32.5579,
      "eval_samples_per_second": 21.009,
      "step": 76200
    },
    {
      "epoch": 18.2,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 76210
    },
    {
      "epoch": 18.2,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 76220
    },
    {
      "epoch": 18.21,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 76230
    },
    {
      "epoch": 18.21,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 76240
    },
    {
      "epoch": 18.21,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 76250
    },
    {
      "epoch": 18.21,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 76260
    },
    {
      "epoch": 18.22,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 76270
    },
    {
      "epoch": 18.22,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 76280
    },
    {
      "epoch": 18.22,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 76290
    },
    {
      "epoch": 18.22,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 76300
    },
    {
      "epoch": 18.22,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.406306266784668,
      "eval_runtime": 32.3585,
      "eval_samples_per_second": 21.138,
      "step": 76300
    },
    {
      "epoch": 18.23,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 76310
    },
    {
      "epoch": 18.23,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 76320
    },
    {
      "epoch": 18.23,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 76330
    },
    {
      "epoch": 18.23,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 76340
    },
    {
      "epoch": 18.24,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 76350
    },
    {
      "epoch": 18.24,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 76360
    },
    {
      "epoch": 18.24,
      "learning_rate": 1e-05,
      "loss": 1.4291,
      "step": 76370
    },
    {
      "epoch": 18.24,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 76380
    },
    {
      "epoch": 18.24,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 76390
    },
    {
      "epoch": 18.25,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 76400
    },
    {
      "epoch": 18.25,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.40743088722229,
      "eval_runtime": 32.4808,
      "eval_samples_per_second": 21.059,
      "step": 76400
    },
    {
      "epoch": 18.25,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 76410
    },
    {
      "epoch": 18.25,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 76420
    },
    {
      "epoch": 18.25,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 76430
    },
    {
      "epoch": 18.26,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 76440
    },
    {
      "epoch": 18.26,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 76450
    },
    {
      "epoch": 18.26,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 76460
    },
    {
      "epoch": 18.26,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 76470
    },
    {
      "epoch": 18.27,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 76480
    },
    {
      "epoch": 18.27,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 76490
    },
    {
      "epoch": 18.27,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 76500
    },
    {
      "epoch": 18.27,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4061472415924072,
      "eval_runtime": 32.3282,
      "eval_samples_per_second": 21.158,
      "step": 76500
    },
    {
      "epoch": 18.27,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 76510
    },
    {
      "epoch": 18.28,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 76520
    },
    {
      "epoch": 18.28,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 76530
    },
    {
      "epoch": 18.28,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 76540
    },
    {
      "epoch": 18.28,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 76550
    },
    {
      "epoch": 18.29,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 76560
    },
    {
      "epoch": 18.29,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 76570
    },
    {
      "epoch": 18.29,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 76580
    },
    {
      "epoch": 18.29,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 76590
    },
    {
      "epoch": 18.29,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 76600
    },
    {
      "epoch": 18.29,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.406335473060608,
      "eval_runtime": 32.3824,
      "eval_samples_per_second": 21.123,
      "step": 76600
    },
    {
      "epoch": 18.3,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 76610
    },
    {
      "epoch": 18.3,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 76620
    },
    {
      "epoch": 18.3,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 76630
    },
    {
      "epoch": 18.3,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 76640
    },
    {
      "epoch": 18.31,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 76650
    },
    {
      "epoch": 18.31,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 76660
    },
    {
      "epoch": 18.31,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 76670
    },
    {
      "epoch": 18.31,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 76680
    },
    {
      "epoch": 18.32,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 76690
    },
    {
      "epoch": 18.32,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 76700
    },
    {
      "epoch": 18.32,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4065312147140503,
      "eval_runtime": 31.9985,
      "eval_samples_per_second": 21.376,
      "step": 76700
    },
    {
      "epoch": 18.32,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 76710
    },
    {
      "epoch": 18.32,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 76720
    },
    {
      "epoch": 18.33,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 76730
    },
    {
      "epoch": 18.33,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 76740
    },
    {
      "epoch": 18.33,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 76750
    },
    {
      "epoch": 18.33,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 76760
    },
    {
      "epoch": 18.34,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 76770
    },
    {
      "epoch": 18.34,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 76780
    },
    {
      "epoch": 18.34,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 76790
    },
    {
      "epoch": 18.34,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 76800
    },
    {
      "epoch": 18.34,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4057836532592773,
      "eval_runtime": 31.9088,
      "eval_samples_per_second": 21.436,
      "step": 76800
    },
    {
      "epoch": 18.34,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 76810
    },
    {
      "epoch": 18.35,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 76820
    },
    {
      "epoch": 18.35,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 76830
    },
    {
      "epoch": 18.35,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 76840
    },
    {
      "epoch": 18.35,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 76850
    },
    {
      "epoch": 18.36,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 76860
    },
    {
      "epoch": 18.36,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 76870
    },
    {
      "epoch": 18.36,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 76880
    },
    {
      "epoch": 18.36,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 76890
    },
    {
      "epoch": 18.37,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 76900
    },
    {
      "epoch": 18.37,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4067527055740356,
      "eval_runtime": 32.2296,
      "eval_samples_per_second": 21.223,
      "step": 76900
    },
    {
      "epoch": 18.37,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 76910
    },
    {
      "epoch": 18.37,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 76920
    },
    {
      "epoch": 18.37,
      "learning_rate": 1e-05,
      "loss": 1.4119,
      "step": 76930
    },
    {
      "epoch": 18.38,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 76940
    },
    {
      "epoch": 18.38,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 76950
    },
    {
      "epoch": 18.38,
      "learning_rate": 1e-05,
      "loss": 1.4106,
      "step": 76960
    },
    {
      "epoch": 18.38,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 76970
    },
    {
      "epoch": 18.39,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 76980
    },
    {
      "epoch": 18.39,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 76990
    },
    {
      "epoch": 18.39,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 77000
    },
    {
      "epoch": 18.39,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.406126856803894,
      "eval_runtime": 32.3872,
      "eval_samples_per_second": 21.119,
      "step": 77000
    },
    {
      "epoch": 18.39,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 77010
    },
    {
      "epoch": 18.4,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 77020
    },
    {
      "epoch": 18.4,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 77030
    },
    {
      "epoch": 18.4,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 77040
    },
    {
      "epoch": 18.4,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 77050
    },
    {
      "epoch": 18.4,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 77060
    },
    {
      "epoch": 18.41,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 77070
    },
    {
      "epoch": 18.41,
      "learning_rate": 1e-05,
      "loss": 1.4091,
      "step": 77080
    },
    {
      "epoch": 18.41,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 77090
    },
    {
      "epoch": 18.41,
      "learning_rate": 1e-05,
      "loss": 1.4093,
      "step": 77100
    },
    {
      "epoch": 18.41,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4061468839645386,
      "eval_runtime": 32.2803,
      "eval_samples_per_second": 21.189,
      "step": 77100
    },
    {
      "epoch": 18.42,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 77110
    },
    {
      "epoch": 18.42,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 77120
    },
    {
      "epoch": 18.42,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 77130
    },
    {
      "epoch": 18.42,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 77140
    },
    {
      "epoch": 18.43,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 77150
    },
    {
      "epoch": 18.43,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 77160
    },
    {
      "epoch": 18.43,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 77170
    },
    {
      "epoch": 18.43,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 77180
    },
    {
      "epoch": 18.44,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 77190
    },
    {
      "epoch": 18.44,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 77200
    },
    {
      "epoch": 18.44,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.407877802848816,
      "eval_runtime": 32.2454,
      "eval_samples_per_second": 21.212,
      "step": 77200
    },
    {
      "epoch": 18.44,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 77210
    },
    {
      "epoch": 18.44,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 77220
    },
    {
      "epoch": 18.45,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 77230
    },
    {
      "epoch": 18.45,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 77240
    },
    {
      "epoch": 18.45,
      "learning_rate": 1e-05,
      "loss": 1.4128,
      "step": 77250
    },
    {
      "epoch": 18.45,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 77260
    },
    {
      "epoch": 18.45,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 77270
    },
    {
      "epoch": 18.46,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 77280
    },
    {
      "epoch": 18.46,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 77290
    },
    {
      "epoch": 18.46,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 77300
    },
    {
      "epoch": 18.46,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4061646461486816,
      "eval_runtime": 32.259,
      "eval_samples_per_second": 21.203,
      "step": 77300
    },
    {
      "epoch": 18.46,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 77310
    },
    {
      "epoch": 18.47,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 77320
    },
    {
      "epoch": 18.47,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 77330
    },
    {
      "epoch": 18.47,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 77340
    },
    {
      "epoch": 18.47,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 77350
    },
    {
      "epoch": 18.48,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 77360
    },
    {
      "epoch": 18.48,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 77370
    },
    {
      "epoch": 18.48,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 77380
    },
    {
      "epoch": 18.48,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 77390
    },
    {
      "epoch": 18.49,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 77400
    },
    {
      "epoch": 18.49,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4062161445617676,
      "eval_runtime": 32.824,
      "eval_samples_per_second": 20.838,
      "step": 77400
    },
    {
      "epoch": 18.49,
      "learning_rate": 1e-05,
      "loss": 1.4225,
      "step": 77410
    },
    {
      "epoch": 18.49,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 77420
    },
    {
      "epoch": 18.49,
      "learning_rate": 1e-05,
      "loss": 1.4095,
      "step": 77430
    },
    {
      "epoch": 18.5,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 77440
    },
    {
      "epoch": 18.5,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 77450
    },
    {
      "epoch": 18.5,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 77460
    },
    {
      "epoch": 18.5,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 77470
    },
    {
      "epoch": 18.5,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 77480
    },
    {
      "epoch": 18.51,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 77490
    },
    {
      "epoch": 18.51,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 77500
    },
    {
      "epoch": 18.51,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.406540870666504,
      "eval_runtime": 32.2551,
      "eval_samples_per_second": 21.206,
      "step": 77500
    },
    {
      "epoch": 18.51,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 77510
    },
    {
      "epoch": 18.51,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 77520
    },
    {
      "epoch": 18.52,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 77530
    },
    {
      "epoch": 18.52,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 77540
    },
    {
      "epoch": 18.52,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 77550
    },
    {
      "epoch": 18.52,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 77560
    },
    {
      "epoch": 18.53,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 77570
    },
    {
      "epoch": 18.53,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 77580
    },
    {
      "epoch": 18.53,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 77590
    },
    {
      "epoch": 18.53,
      "learning_rate": 1e-05,
      "loss": 1.4094,
      "step": 77600
    },
    {
      "epoch": 18.53,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4065123796463013,
      "eval_runtime": 32.3637,
      "eval_samples_per_second": 21.135,
      "step": 77600
    },
    {
      "epoch": 18.54,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 77610
    },
    {
      "epoch": 18.54,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 77620
    },
    {
      "epoch": 18.54,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 77630
    },
    {
      "epoch": 18.54,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 77640
    },
    {
      "epoch": 18.55,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 77650
    },
    {
      "epoch": 18.55,
      "learning_rate": 1e-05,
      "loss": 1.417,
      "step": 77660
    },
    {
      "epoch": 18.55,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 77670
    },
    {
      "epoch": 18.55,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 77680
    },
    {
      "epoch": 18.56,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 77690
    },
    {
      "epoch": 18.56,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 77700
    },
    {
      "epoch": 18.56,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.406859040260315,
      "eval_runtime": 32.265,
      "eval_samples_per_second": 21.199,
      "step": 77700
    },
    {
      "epoch": 18.56,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 77710
    },
    {
      "epoch": 18.56,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 77720
    },
    {
      "epoch": 18.56,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 77730
    },
    {
      "epoch": 18.57,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 77740
    },
    {
      "epoch": 18.57,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 77750
    },
    {
      "epoch": 18.57,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 77760
    },
    {
      "epoch": 18.57,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 77770
    },
    {
      "epoch": 18.58,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 77780
    },
    {
      "epoch": 18.58,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 77790
    },
    {
      "epoch": 18.58,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 77800
    },
    {
      "epoch": 18.58,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4067898988723755,
      "eval_runtime": 31.9175,
      "eval_samples_per_second": 21.43,
      "step": 77800
    },
    {
      "epoch": 18.58,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 77810
    },
    {
      "epoch": 18.59,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 77820
    },
    {
      "epoch": 18.59,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 77830
    },
    {
      "epoch": 18.59,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 77840
    },
    {
      "epoch": 18.59,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 77850
    },
    {
      "epoch": 18.6,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 77860
    },
    {
      "epoch": 18.6,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 77870
    },
    {
      "epoch": 18.6,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 77880
    },
    {
      "epoch": 18.6,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 77890
    },
    {
      "epoch": 18.61,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 77900
    },
    {
      "epoch": 18.61,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4059085845947266,
      "eval_runtime": 32.0119,
      "eval_samples_per_second": 21.367,
      "step": 77900
    },
    {
      "epoch": 18.61,
      "learning_rate": 1e-05,
      "loss": 1.4097,
      "step": 77910
    },
    {
      "epoch": 18.61,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 77920
    },
    {
      "epoch": 18.61,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 77930
    },
    {
      "epoch": 18.61,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 77940
    },
    {
      "epoch": 18.62,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 77950
    },
    {
      "epoch": 18.62,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 77960
    },
    {
      "epoch": 18.62,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 77970
    },
    {
      "epoch": 18.62,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 77980
    },
    {
      "epoch": 18.63,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 77990
    },
    {
      "epoch": 18.63,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 78000
    },
    {
      "epoch": 18.63,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4058723449707031,
      "eval_runtime": 31.7699,
      "eval_samples_per_second": 21.53,
      "step": 78000
    },
    {
      "epoch": 18.63,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 78010
    },
    {
      "epoch": 18.63,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 78020
    },
    {
      "epoch": 18.64,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 78030
    },
    {
      "epoch": 18.64,
      "learning_rate": 1e-05,
      "loss": 1.4074,
      "step": 78040
    },
    {
      "epoch": 18.64,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 78050
    },
    {
      "epoch": 18.64,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 78060
    },
    {
      "epoch": 18.65,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 78070
    },
    {
      "epoch": 18.65,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 78080
    },
    {
      "epoch": 18.65,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 78090
    },
    {
      "epoch": 18.65,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 78100
    },
    {
      "epoch": 18.65,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4063739776611328,
      "eval_runtime": 32.5285,
      "eval_samples_per_second": 21.028,
      "step": 78100
    },
    {
      "epoch": 18.66,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 78110
    },
    {
      "epoch": 18.66,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 78120
    },
    {
      "epoch": 18.66,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 78130
    },
    {
      "epoch": 18.66,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 78140
    },
    {
      "epoch": 18.66,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 78150
    },
    {
      "epoch": 18.67,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 78160
    },
    {
      "epoch": 18.67,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 78170
    },
    {
      "epoch": 18.67,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 78180
    },
    {
      "epoch": 18.67,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 78190
    },
    {
      "epoch": 18.68,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 78200
    },
    {
      "epoch": 18.68,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4060211181640625,
      "eval_runtime": 32.0043,
      "eval_samples_per_second": 21.372,
      "step": 78200
    },
    {
      "epoch": 18.68,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 78210
    },
    {
      "epoch": 18.68,
      "learning_rate": 1e-05,
      "loss": 1.4088,
      "step": 78220
    },
    {
      "epoch": 18.68,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 78230
    },
    {
      "epoch": 18.69,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 78240
    },
    {
      "epoch": 18.69,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 78250
    },
    {
      "epoch": 18.69,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 78260
    },
    {
      "epoch": 18.69,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 78270
    },
    {
      "epoch": 18.7,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 78280
    },
    {
      "epoch": 18.7,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 78290
    },
    {
      "epoch": 18.7,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 78300
    },
    {
      "epoch": 18.7,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4056260585784912,
      "eval_runtime": 32.3571,
      "eval_samples_per_second": 21.139,
      "step": 78300
    },
    {
      "epoch": 18.7,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 78310
    },
    {
      "epoch": 18.71,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 78320
    },
    {
      "epoch": 18.71,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 78330
    },
    {
      "epoch": 18.71,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 78340
    },
    {
      "epoch": 18.71,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 78350
    },
    {
      "epoch": 18.72,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 78360
    },
    {
      "epoch": 18.72,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 78370
    },
    {
      "epoch": 18.72,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 78380
    },
    {
      "epoch": 18.72,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 78390
    },
    {
      "epoch": 18.72,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 78400
    },
    {
      "epoch": 18.72,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4065412282943726,
      "eval_runtime": 32.1385,
      "eval_samples_per_second": 21.283,
      "step": 78400
    },
    {
      "epoch": 18.73,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 78410
    },
    {
      "epoch": 18.73,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 78420
    },
    {
      "epoch": 18.73,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 78430
    },
    {
      "epoch": 18.73,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 78440
    },
    {
      "epoch": 18.74,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 78450
    },
    {
      "epoch": 18.74,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 78460
    },
    {
      "epoch": 18.74,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 78470
    },
    {
      "epoch": 18.74,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 78480
    },
    {
      "epoch": 18.75,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 78490
    },
    {
      "epoch": 18.75,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 78500
    },
    {
      "epoch": 18.75,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.406038522720337,
      "eval_runtime": 32.4956,
      "eval_samples_per_second": 21.049,
      "step": 78500
    },
    {
      "epoch": 18.75,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 78510
    },
    {
      "epoch": 18.75,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 78520
    },
    {
      "epoch": 18.76,
      "learning_rate": 1e-05,
      "loss": 1.4313,
      "step": 78530
    },
    {
      "epoch": 18.76,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 78540
    },
    {
      "epoch": 18.76,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 78550
    },
    {
      "epoch": 18.76,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 78560
    },
    {
      "epoch": 18.77,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 78570
    },
    {
      "epoch": 18.77,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 78580
    },
    {
      "epoch": 18.77,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 78590
    },
    {
      "epoch": 18.77,
      "learning_rate": 1e-05,
      "loss": 1.4161,
      "step": 78600
    },
    {
      "epoch": 18.77,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4065577983856201,
      "eval_runtime": 32.0164,
      "eval_samples_per_second": 21.364,
      "step": 78600
    },
    {
      "epoch": 18.77,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 78610
    },
    {
      "epoch": 18.78,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 78620
    },
    {
      "epoch": 18.78,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 78630
    },
    {
      "epoch": 18.78,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 78640
    },
    {
      "epoch": 18.78,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 78650
    },
    {
      "epoch": 18.79,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 78660
    },
    {
      "epoch": 18.79,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 78670
    },
    {
      "epoch": 18.79,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 78680
    },
    {
      "epoch": 18.79,
      "learning_rate": 1e-05,
      "loss": 1.4111,
      "step": 78690
    },
    {
      "epoch": 18.8,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 78700
    },
    {
      "epoch": 18.8,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4059374332427979,
      "eval_runtime": 32.0256,
      "eval_samples_per_second": 21.358,
      "step": 78700
    },
    {
      "epoch": 18.8,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 78710
    },
    {
      "epoch": 18.8,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 78720
    },
    {
      "epoch": 18.8,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 78730
    },
    {
      "epoch": 18.81,
      "learning_rate": 1e-05,
      "loss": 1.398,
      "step": 78740
    },
    {
      "epoch": 18.81,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 78750
    },
    {
      "epoch": 18.81,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 78760
    },
    {
      "epoch": 18.81,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 78770
    },
    {
      "epoch": 18.82,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 78780
    },
    {
      "epoch": 18.82,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 78790
    },
    {
      "epoch": 18.82,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 78800
    },
    {
      "epoch": 18.82,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4051377773284912,
      "eval_runtime": 32.2339,
      "eval_samples_per_second": 21.22,
      "step": 78800
    },
    {
      "epoch": 18.82,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 78810
    },
    {
      "epoch": 18.82,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 78820
    },
    {
      "epoch": 18.83,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 78830
    },
    {
      "epoch": 18.83,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 78840
    },
    {
      "epoch": 18.83,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 78850
    },
    {
      "epoch": 18.83,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 78860
    },
    {
      "epoch": 18.84,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 78870
    },
    {
      "epoch": 18.84,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 78880
    },
    {
      "epoch": 18.84,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 78890
    },
    {
      "epoch": 18.84,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 78900
    },
    {
      "epoch": 18.84,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.405678391456604,
      "eval_runtime": 31.8962,
      "eval_samples_per_second": 21.445,
      "step": 78900
    },
    {
      "epoch": 18.85,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 78910
    },
    {
      "epoch": 18.85,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 78920
    },
    {
      "epoch": 18.85,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 78930
    },
    {
      "epoch": 18.85,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 78940
    },
    {
      "epoch": 18.86,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 78950
    },
    {
      "epoch": 18.86,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 78960
    },
    {
      "epoch": 18.86,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 78970
    },
    {
      "epoch": 18.86,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 78980
    },
    {
      "epoch": 18.87,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 78990
    },
    {
      "epoch": 18.87,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 79000
    },
    {
      "epoch": 18.87,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.406604290008545,
      "eval_runtime": 32.0812,
      "eval_samples_per_second": 21.321,
      "step": 79000
    },
    {
      "epoch": 18.87,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 79010
    },
    {
      "epoch": 18.87,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 79020
    },
    {
      "epoch": 18.88,
      "learning_rate": 1e-05,
      "loss": 1.4105,
      "step": 79030
    },
    {
      "epoch": 18.88,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 79040
    },
    {
      "epoch": 18.88,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 79050
    },
    {
      "epoch": 18.88,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 79060
    },
    {
      "epoch": 18.88,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 79070
    },
    {
      "epoch": 18.89,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 79080
    },
    {
      "epoch": 18.89,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 79090
    },
    {
      "epoch": 18.89,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 79100
    },
    {
      "epoch": 18.89,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4057018756866455,
      "eval_runtime": 31.6457,
      "eval_samples_per_second": 21.614,
      "step": 79100
    },
    {
      "epoch": 18.89,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 79110
    },
    {
      "epoch": 18.9,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 79120
    },
    {
      "epoch": 18.9,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 79130
    },
    {
      "epoch": 18.9,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 79140
    },
    {
      "epoch": 18.9,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 79150
    },
    {
      "epoch": 18.91,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 79160
    },
    {
      "epoch": 18.91,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 79170
    },
    {
      "epoch": 18.91,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 79180
    },
    {
      "epoch": 18.91,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 79190
    },
    {
      "epoch": 18.92,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 79200
    },
    {
      "epoch": 18.92,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4057966470718384,
      "eval_runtime": 31.9341,
      "eval_samples_per_second": 21.419,
      "step": 79200
    },
    {
      "epoch": 18.92,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 79210
    },
    {
      "epoch": 18.92,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 79220
    },
    {
      "epoch": 18.92,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 79230
    },
    {
      "epoch": 18.93,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 79240
    },
    {
      "epoch": 18.93,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 79250
    },
    {
      "epoch": 18.93,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 79260
    },
    {
      "epoch": 18.93,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 79270
    },
    {
      "epoch": 18.93,
      "learning_rate": 1e-05,
      "loss": 1.4079,
      "step": 79280
    },
    {
      "epoch": 18.94,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 79290
    },
    {
      "epoch": 18.94,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 79300
    },
    {
      "epoch": 18.94,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4063727855682373,
      "eval_runtime": 31.8344,
      "eval_samples_per_second": 21.486,
      "step": 79300
    },
    {
      "epoch": 18.94,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 79310
    },
    {
      "epoch": 18.94,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 79320
    },
    {
      "epoch": 18.95,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 79330
    },
    {
      "epoch": 18.95,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 79340
    },
    {
      "epoch": 18.95,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 79350
    },
    {
      "epoch": 18.95,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 79360
    },
    {
      "epoch": 18.96,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 79370
    },
    {
      "epoch": 18.96,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 79380
    },
    {
      "epoch": 18.96,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 79390
    },
    {
      "epoch": 18.96,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 79400
    },
    {
      "epoch": 18.96,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4066298007965088,
      "eval_runtime": 32.0374,
      "eval_samples_per_second": 21.35,
      "step": 79400
    },
    {
      "epoch": 18.97,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 79410
    },
    {
      "epoch": 18.97,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 79420
    },
    {
      "epoch": 18.97,
      "learning_rate": 1e-05,
      "loss": 1.4082,
      "step": 79430
    },
    {
      "epoch": 18.97,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 79440
    },
    {
      "epoch": 18.98,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 79450
    },
    {
      "epoch": 18.98,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 79460
    },
    {
      "epoch": 18.98,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 79470
    },
    {
      "epoch": 18.98,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 79480
    },
    {
      "epoch": 18.98,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 79490
    },
    {
      "epoch": 18.99,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 79500
    },
    {
      "epoch": 18.99,
      "eval_accuracy": 0.9868421052631579,
      "eval_loss": 1.406396746635437,
      "eval_runtime": 31.9752,
      "eval_samples_per_second": 21.392,
      "step": 79500
    },
    {
      "epoch": 18.99,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 79510
    },
    {
      "epoch": 18.99,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 79520
    },
    {
      "epoch": 18.99,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 79530
    },
    {
      "epoch": 19.0,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 79540
    },
    {
      "epoch": 19.0,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 79550
    },
    {
      "epoch": 19.0,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 79560
    },
    {
      "epoch": 19.0,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 79570
    },
    {
      "epoch": 19.01,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 79580
    },
    {
      "epoch": 19.01,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 79590
    },
    {
      "epoch": 19.01,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 79600
    },
    {
      "epoch": 19.01,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4056220054626465,
      "eval_runtime": 32.2127,
      "eval_samples_per_second": 21.234,
      "step": 79600
    },
    {
      "epoch": 19.01,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 79610
    },
    {
      "epoch": 19.02,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 79620
    },
    {
      "epoch": 19.02,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 79630
    },
    {
      "epoch": 19.02,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 79640
    },
    {
      "epoch": 19.02,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 79650
    },
    {
      "epoch": 19.03,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 79660
    },
    {
      "epoch": 19.03,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 79670
    },
    {
      "epoch": 19.03,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 79680
    },
    {
      "epoch": 19.03,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 79690
    },
    {
      "epoch": 19.04,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 79700
    },
    {
      "epoch": 19.04,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4058862924575806,
      "eval_runtime": 31.9534,
      "eval_samples_per_second": 21.406,
      "step": 79700
    },
    {
      "epoch": 19.04,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 79710
    },
    {
      "epoch": 19.04,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 79720
    },
    {
      "epoch": 19.04,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 79730
    },
    {
      "epoch": 19.04,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 79740
    },
    {
      "epoch": 19.05,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 79750
    },
    {
      "epoch": 19.05,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 79760
    },
    {
      "epoch": 19.05,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 79770
    },
    {
      "epoch": 19.05,
      "learning_rate": 1e-05,
      "loss": 1.4326,
      "step": 79780
    },
    {
      "epoch": 19.06,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 79790
    },
    {
      "epoch": 19.06,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 79800
    },
    {
      "epoch": 19.06,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4062235355377197,
      "eval_runtime": 31.9163,
      "eval_samples_per_second": 21.431,
      "step": 79800
    },
    {
      "epoch": 19.06,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 79810
    },
    {
      "epoch": 19.06,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 79820
    },
    {
      "epoch": 19.07,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 79830
    },
    {
      "epoch": 19.07,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 79840
    },
    {
      "epoch": 19.07,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 79850
    },
    {
      "epoch": 19.07,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 79860
    },
    {
      "epoch": 19.08,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 79870
    },
    {
      "epoch": 19.08,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 79880
    },
    {
      "epoch": 19.08,
      "learning_rate": 1e-05,
      "loss": 1.413,
      "step": 79890
    },
    {
      "epoch": 19.08,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 79900
    },
    {
      "epoch": 19.08,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4053875207901,
      "eval_runtime": 32.1168,
      "eval_samples_per_second": 21.297,
      "step": 79900
    },
    {
      "epoch": 19.09,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 79910
    },
    {
      "epoch": 19.09,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 79920
    },
    {
      "epoch": 19.09,
      "learning_rate": 1e-05,
      "loss": 1.415,
      "step": 79930
    },
    {
      "epoch": 19.09,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 79940
    },
    {
      "epoch": 19.09,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 79950
    },
    {
      "epoch": 19.1,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 79960
    },
    {
      "epoch": 19.1,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 79970
    },
    {
      "epoch": 19.1,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 79980
    },
    {
      "epoch": 19.1,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 79990
    },
    {
      "epoch": 19.11,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 80000
    },
    {
      "epoch": 19.11,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.405471682548523,
      "eval_runtime": 32.1151,
      "eval_samples_per_second": 21.298,
      "step": 80000
    },
    {
      "epoch": 19.11,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 80010
    },
    {
      "epoch": 19.11,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 80020
    },
    {
      "epoch": 19.11,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 80030
    },
    {
      "epoch": 19.12,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 80040
    },
    {
      "epoch": 19.12,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 80050
    },
    {
      "epoch": 19.12,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 80060
    },
    {
      "epoch": 19.12,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 80070
    },
    {
      "epoch": 19.13,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 80080
    },
    {
      "epoch": 19.13,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 80090
    },
    {
      "epoch": 19.13,
      "learning_rate": 1e-05,
      "loss": 1.4109,
      "step": 80100
    },
    {
      "epoch": 19.13,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.406091332435608,
      "eval_runtime": 32.0201,
      "eval_samples_per_second": 21.362,
      "step": 80100
    },
    {
      "epoch": 19.13,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 80110
    },
    {
      "epoch": 19.14,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 80120
    },
    {
      "epoch": 19.14,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 80130
    },
    {
      "epoch": 19.14,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 80140
    },
    {
      "epoch": 19.14,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 80150
    },
    {
      "epoch": 19.14,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 80160
    },
    {
      "epoch": 19.15,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 80170
    },
    {
      "epoch": 19.15,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 80180
    },
    {
      "epoch": 19.15,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 80190
    },
    {
      "epoch": 19.15,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 80200
    },
    {
      "epoch": 19.15,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4058481454849243,
      "eval_runtime": 31.9031,
      "eval_samples_per_second": 21.44,
      "step": 80200
    },
    {
      "epoch": 19.16,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 80210
    },
    {
      "epoch": 19.16,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 80220
    },
    {
      "epoch": 19.16,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 80230
    },
    {
      "epoch": 19.16,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 80240
    },
    {
      "epoch": 19.17,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 80250
    },
    {
      "epoch": 19.17,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 80260
    },
    {
      "epoch": 19.17,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 80270
    },
    {
      "epoch": 19.17,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 80280
    },
    {
      "epoch": 19.18,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 80290
    },
    {
      "epoch": 19.18,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 80300
    },
    {
      "epoch": 19.18,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4060370922088623,
      "eval_runtime": 31.8468,
      "eval_samples_per_second": 21.478,
      "step": 80300
    },
    {
      "epoch": 19.18,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 80310
    },
    {
      "epoch": 19.18,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 80320
    },
    {
      "epoch": 19.19,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 80330
    },
    {
      "epoch": 19.19,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 80340
    },
    {
      "epoch": 19.19,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 80350
    },
    {
      "epoch": 19.19,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 80360
    },
    {
      "epoch": 19.2,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 80370
    },
    {
      "epoch": 19.2,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 80380
    },
    {
      "epoch": 19.2,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 80390
    },
    {
      "epoch": 19.2,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 80400
    },
    {
      "epoch": 19.2,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4053332805633545,
      "eval_runtime": 32.0556,
      "eval_samples_per_second": 21.338,
      "step": 80400
    },
    {
      "epoch": 19.2,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 80410
    },
    {
      "epoch": 19.21,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 80420
    },
    {
      "epoch": 19.21,
      "learning_rate": 1e-05,
      "loss": 1.4151,
      "step": 80430
    },
    {
      "epoch": 19.21,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 80440
    },
    {
      "epoch": 19.21,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 80450
    },
    {
      "epoch": 19.22,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 80460
    },
    {
      "epoch": 19.22,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 80470
    },
    {
      "epoch": 19.22,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 80480
    },
    {
      "epoch": 19.22,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 80490
    },
    {
      "epoch": 19.23,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 80500
    },
    {
      "epoch": 19.23,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4056593179702759,
      "eval_runtime": 32.1628,
      "eval_samples_per_second": 21.267,
      "step": 80500
    },
    {
      "epoch": 19.23,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 80510
    },
    {
      "epoch": 19.23,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 80520
    },
    {
      "epoch": 19.23,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 80530
    },
    {
      "epoch": 19.24,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 80540
    },
    {
      "epoch": 19.24,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 80550
    },
    {
      "epoch": 19.24,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 80560
    },
    {
      "epoch": 19.24,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 80570
    },
    {
      "epoch": 19.25,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 80580
    },
    {
      "epoch": 19.25,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 80590
    },
    {
      "epoch": 19.25,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 80600
    },
    {
      "epoch": 19.25,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.406029224395752,
      "eval_runtime": 31.9422,
      "eval_samples_per_second": 21.414,
      "step": 80600
    },
    {
      "epoch": 19.25,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 80610
    },
    {
      "epoch": 19.25,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 80620
    },
    {
      "epoch": 19.26,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 80630
    },
    {
      "epoch": 19.26,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 80640
    },
    {
      "epoch": 19.26,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 80650
    },
    {
      "epoch": 19.26,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 80660
    },
    {
      "epoch": 19.27,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 80670
    },
    {
      "epoch": 19.27,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 80680
    },
    {
      "epoch": 19.27,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 80690
    },
    {
      "epoch": 19.27,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 80700
    },
    {
      "epoch": 19.27,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.405454158782959,
      "eval_runtime": 31.7227,
      "eval_samples_per_second": 21.562,
      "step": 80700
    },
    {
      "epoch": 19.28,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 80710
    },
    {
      "epoch": 19.28,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 80720
    },
    {
      "epoch": 19.28,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 80730
    },
    {
      "epoch": 19.28,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 80740
    },
    {
      "epoch": 19.29,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 80750
    },
    {
      "epoch": 19.29,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 80760
    },
    {
      "epoch": 19.29,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 80770
    },
    {
      "epoch": 19.29,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 80780
    },
    {
      "epoch": 19.3,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 80790
    },
    {
      "epoch": 19.3,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 80800
    },
    {
      "epoch": 19.3,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.40590500831604,
      "eval_runtime": 32.0023,
      "eval_samples_per_second": 21.373,
      "step": 80800
    },
    {
      "epoch": 19.3,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 80810
    },
    {
      "epoch": 19.3,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 80820
    },
    {
      "epoch": 19.3,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 80830
    },
    {
      "epoch": 19.31,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 80840
    },
    {
      "epoch": 19.31,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 80850
    },
    {
      "epoch": 19.31,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 80860
    },
    {
      "epoch": 19.31,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 80870
    },
    {
      "epoch": 19.32,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 80880
    },
    {
      "epoch": 19.32,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 80890
    },
    {
      "epoch": 19.32,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 80900
    },
    {
      "epoch": 19.32,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4053395986557007,
      "eval_runtime": 32.0565,
      "eval_samples_per_second": 21.337,
      "step": 80900
    },
    {
      "epoch": 19.32,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 80910
    },
    {
      "epoch": 19.33,
      "learning_rate": 1e-05,
      "loss": 1.4092,
      "step": 80920
    },
    {
      "epoch": 19.33,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 80930
    },
    {
      "epoch": 19.33,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 80940
    },
    {
      "epoch": 19.33,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 80950
    },
    {
      "epoch": 19.34,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 80960
    },
    {
      "epoch": 19.34,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 80970
    },
    {
      "epoch": 19.34,
      "learning_rate": 1e-05,
      "loss": 1.411,
      "step": 80980
    },
    {
      "epoch": 19.34,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 80990
    },
    {
      "epoch": 19.35,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 81000
    },
    {
      "epoch": 19.35,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4058483839035034,
      "eval_runtime": 32.2015,
      "eval_samples_per_second": 21.241,
      "step": 81000
    },
    {
      "epoch": 19.35,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 81010
    },
    {
      "epoch": 19.35,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 81020
    },
    {
      "epoch": 19.35,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 81030
    },
    {
      "epoch": 19.36,
      "learning_rate": 1e-05,
      "loss": 1.4133,
      "step": 81040
    },
    {
      "epoch": 19.36,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 81050
    },
    {
      "epoch": 19.36,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 81060
    },
    {
      "epoch": 19.36,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 81070
    },
    {
      "epoch": 19.36,
      "learning_rate": 1e-05,
      "loss": 1.4086,
      "step": 81080
    },
    {
      "epoch": 19.37,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 81090
    },
    {
      "epoch": 19.37,
      "learning_rate": 1e-05,
      "loss": 1.4107,
      "step": 81100
    },
    {
      "epoch": 19.37,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4054343700408936,
      "eval_runtime": 32.4165,
      "eval_samples_per_second": 21.1,
      "step": 81100
    },
    {
      "epoch": 19.37,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 81110
    },
    {
      "epoch": 19.37,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 81120
    },
    {
      "epoch": 19.38,
      "learning_rate": 1e-05,
      "loss": 1.4123,
      "step": 81130
    },
    {
      "epoch": 19.38,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 81140
    },
    {
      "epoch": 19.38,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 81150
    },
    {
      "epoch": 19.38,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 81160
    },
    {
      "epoch": 19.39,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 81170
    },
    {
      "epoch": 19.39,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 81180
    },
    {
      "epoch": 19.39,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 81190
    },
    {
      "epoch": 19.39,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 81200
    },
    {
      "epoch": 19.39,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4052964448928833,
      "eval_runtime": 32.2918,
      "eval_samples_per_second": 21.182,
      "step": 81200
    },
    {
      "epoch": 19.4,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 81210
    },
    {
      "epoch": 19.4,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 81220
    },
    {
      "epoch": 19.4,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 81230
    },
    {
      "epoch": 19.4,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 81240
    },
    {
      "epoch": 19.41,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 81250
    },
    {
      "epoch": 19.41,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 81260
    },
    {
      "epoch": 19.41,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 81270
    },
    {
      "epoch": 19.41,
      "learning_rate": 1e-05,
      "loss": 1.4175,
      "step": 81280
    },
    {
      "epoch": 19.41,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 81290
    },
    {
      "epoch": 19.42,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 81300
    },
    {
      "epoch": 19.42,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.406449556350708,
      "eval_runtime": 32.0317,
      "eval_samples_per_second": 21.354,
      "step": 81300
    },
    {
      "epoch": 19.42,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 81310
    },
    {
      "epoch": 19.42,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 81320
    },
    {
      "epoch": 19.42,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 81330
    },
    {
      "epoch": 19.43,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 81340
    },
    {
      "epoch": 19.43,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 81350
    },
    {
      "epoch": 19.43,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 81360
    },
    {
      "epoch": 19.43,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 81370
    },
    {
      "epoch": 19.44,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 81380
    },
    {
      "epoch": 19.44,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 81390
    },
    {
      "epoch": 19.44,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 81400
    },
    {
      "epoch": 19.44,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4048855304718018,
      "eval_runtime": 32.1836,
      "eval_samples_per_second": 21.253,
      "step": 81400
    },
    {
      "epoch": 19.44,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 81410
    },
    {
      "epoch": 19.45,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 81420
    },
    {
      "epoch": 19.45,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 81430
    },
    {
      "epoch": 19.45,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 81440
    },
    {
      "epoch": 19.45,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 81450
    },
    {
      "epoch": 19.46,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 81460
    },
    {
      "epoch": 19.46,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 81470
    },
    {
      "epoch": 19.46,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 81480
    },
    {
      "epoch": 19.46,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 81490
    },
    {
      "epoch": 19.47,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 81500
    },
    {
      "epoch": 19.47,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4052932262420654,
      "eval_runtime": 32.0406,
      "eval_samples_per_second": 21.348,
      "step": 81500
    },
    {
      "epoch": 19.47,
      "learning_rate": 1e-05,
      "loss": 1.4096,
      "step": 81510
    },
    {
      "epoch": 19.47,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 81520
    },
    {
      "epoch": 19.47,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 81530
    },
    {
      "epoch": 19.47,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 81540
    },
    {
      "epoch": 19.48,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 81550
    },
    {
      "epoch": 19.48,
      "learning_rate": 1e-05,
      "loss": 1.3986,
      "step": 81560
    },
    {
      "epoch": 19.48,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 81570
    },
    {
      "epoch": 19.48,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 81580
    },
    {
      "epoch": 19.49,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 81590
    },
    {
      "epoch": 19.49,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 81600
    },
    {
      "epoch": 19.49,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.404403805732727,
      "eval_runtime": 32.2322,
      "eval_samples_per_second": 21.221,
      "step": 81600
    },
    {
      "epoch": 19.49,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 81610
    },
    {
      "epoch": 19.49,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 81620
    },
    {
      "epoch": 19.5,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 81630
    },
    {
      "epoch": 19.5,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 81640
    },
    {
      "epoch": 19.5,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 81650
    },
    {
      "epoch": 19.5,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 81660
    },
    {
      "epoch": 19.51,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 81670
    },
    {
      "epoch": 19.51,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 81680
    },
    {
      "epoch": 19.51,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 81690
    },
    {
      "epoch": 19.51,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 81700
    },
    {
      "epoch": 19.51,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4057925939559937,
      "eval_runtime": 32.1917,
      "eval_samples_per_second": 21.248,
      "step": 81700
    },
    {
      "epoch": 19.52,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 81710
    },
    {
      "epoch": 19.52,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 81720
    },
    {
      "epoch": 19.52,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 81730
    },
    {
      "epoch": 19.52,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 81740
    },
    {
      "epoch": 19.52,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 81750
    },
    {
      "epoch": 19.53,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 81760
    },
    {
      "epoch": 19.53,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 81770
    },
    {
      "epoch": 19.53,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 81780
    },
    {
      "epoch": 19.53,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 81790
    },
    {
      "epoch": 19.54,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 81800
    },
    {
      "epoch": 19.54,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4056130647659302,
      "eval_runtime": 31.9491,
      "eval_samples_per_second": 21.409,
      "step": 81800
    },
    {
      "epoch": 19.54,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 81810
    },
    {
      "epoch": 19.54,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 81820
    },
    {
      "epoch": 19.54,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 81830
    },
    {
      "epoch": 19.55,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 81840
    },
    {
      "epoch": 19.55,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 81850
    },
    {
      "epoch": 19.55,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 81860
    },
    {
      "epoch": 19.55,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 81870
    },
    {
      "epoch": 19.56,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 81880
    },
    {
      "epoch": 19.56,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 81890
    },
    {
      "epoch": 19.56,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 81900
    },
    {
      "epoch": 19.56,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4051417112350464,
      "eval_runtime": 32.7612,
      "eval_samples_per_second": 20.878,
      "step": 81900
    },
    {
      "epoch": 19.56,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 81910
    },
    {
      "epoch": 19.57,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 81920
    },
    {
      "epoch": 19.57,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 81930
    },
    {
      "epoch": 19.57,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 81940
    },
    {
      "epoch": 19.57,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 81950
    },
    {
      "epoch": 19.57,
      "learning_rate": 1e-05,
      "loss": 1.41,
      "step": 81960
    },
    {
      "epoch": 19.58,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 81970
    },
    {
      "epoch": 19.58,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 81980
    },
    {
      "epoch": 19.58,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 81990
    },
    {
      "epoch": 19.58,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 82000
    },
    {
      "epoch": 19.58,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.405766487121582,
      "eval_runtime": 32.3122,
      "eval_samples_per_second": 21.168,
      "step": 82000
    },
    {
      "epoch": 19.59,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 82010
    },
    {
      "epoch": 19.59,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 82020
    },
    {
      "epoch": 19.59,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 82030
    },
    {
      "epoch": 19.59,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 82040
    },
    {
      "epoch": 19.6,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 82050
    },
    {
      "epoch": 19.6,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 82060
    },
    {
      "epoch": 19.6,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 82070
    },
    {
      "epoch": 19.6,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 82080
    },
    {
      "epoch": 19.61,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 82090
    },
    {
      "epoch": 19.61,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 82100
    },
    {
      "epoch": 19.61,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.405072808265686,
      "eval_runtime": 32.7859,
      "eval_samples_per_second": 20.863,
      "step": 82100
    },
    {
      "epoch": 19.61,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 82110
    },
    {
      "epoch": 19.61,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 82120
    },
    {
      "epoch": 19.62,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 82130
    },
    {
      "epoch": 19.62,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 82140
    },
    {
      "epoch": 19.62,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 82150
    },
    {
      "epoch": 19.62,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 82160
    },
    {
      "epoch": 19.63,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 82170
    },
    {
      "epoch": 19.63,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 82180
    },
    {
      "epoch": 19.63,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 82190
    },
    {
      "epoch": 19.63,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 82200
    },
    {
      "epoch": 19.63,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4050078392028809,
      "eval_runtime": 32.0993,
      "eval_samples_per_second": 21.309,
      "step": 82200
    },
    {
      "epoch": 19.63,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 82210
    },
    {
      "epoch": 19.64,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 82220
    },
    {
      "epoch": 19.64,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 82230
    },
    {
      "epoch": 19.64,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 82240
    },
    {
      "epoch": 19.64,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 82250
    },
    {
      "epoch": 19.65,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 82260
    },
    {
      "epoch": 19.65,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 82270
    },
    {
      "epoch": 19.65,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 82280
    },
    {
      "epoch": 19.65,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 82290
    },
    {
      "epoch": 19.66,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 82300
    },
    {
      "epoch": 19.66,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4045101404190063,
      "eval_runtime": 31.9884,
      "eval_samples_per_second": 21.383,
      "step": 82300
    },
    {
      "epoch": 19.66,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 82310
    },
    {
      "epoch": 19.66,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 82320
    },
    {
      "epoch": 19.66,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 82330
    },
    {
      "epoch": 19.67,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 82340
    },
    {
      "epoch": 19.67,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 82350
    },
    {
      "epoch": 19.67,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 82360
    },
    {
      "epoch": 19.67,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 82370
    },
    {
      "epoch": 19.68,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 82380
    },
    {
      "epoch": 19.68,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 82390
    },
    {
      "epoch": 19.68,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 82400
    },
    {
      "epoch": 19.68,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4049320220947266,
      "eval_runtime": 32.0369,
      "eval_samples_per_second": 21.35,
      "step": 82400
    },
    {
      "epoch": 19.68,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 82410
    },
    {
      "epoch": 19.68,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 82420
    },
    {
      "epoch": 19.69,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 82430
    },
    {
      "epoch": 19.69,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 82440
    },
    {
      "epoch": 19.69,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 82450
    },
    {
      "epoch": 19.69,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 82460
    },
    {
      "epoch": 19.7,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 82470
    },
    {
      "epoch": 19.7,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 82480
    },
    {
      "epoch": 19.7,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 82490
    },
    {
      "epoch": 19.7,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 82500
    },
    {
      "epoch": 19.7,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4043877124786377,
      "eval_runtime": 31.9485,
      "eval_samples_per_second": 21.409,
      "step": 82500
    },
    {
      "epoch": 19.71,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 82510
    },
    {
      "epoch": 19.71,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 82520
    },
    {
      "epoch": 19.71,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 82530
    },
    {
      "epoch": 19.71,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 82540
    },
    {
      "epoch": 19.72,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 82550
    },
    {
      "epoch": 19.72,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 82560
    },
    {
      "epoch": 19.72,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 82570
    },
    {
      "epoch": 19.72,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 82580
    },
    {
      "epoch": 19.73,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 82590
    },
    {
      "epoch": 19.73,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 82600
    },
    {
      "epoch": 19.73,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4045636653900146,
      "eval_runtime": 32.5363,
      "eval_samples_per_second": 21.023,
      "step": 82600
    },
    {
      "epoch": 19.73,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 82610
    },
    {
      "epoch": 19.73,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 82620
    },
    {
      "epoch": 19.73,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 82630
    },
    {
      "epoch": 19.74,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 82640
    },
    {
      "epoch": 19.74,
      "learning_rate": 1e-05,
      "loss": 1.4085,
      "step": 82650
    },
    {
      "epoch": 19.74,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 82660
    },
    {
      "epoch": 19.74,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 82670
    },
    {
      "epoch": 19.75,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 82680
    },
    {
      "epoch": 19.75,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 82690
    },
    {
      "epoch": 19.75,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 82700
    },
    {
      "epoch": 19.75,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.404991626739502,
      "eval_runtime": 31.7347,
      "eval_samples_per_second": 21.554,
      "step": 82700
    },
    {
      "epoch": 19.75,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 82710
    },
    {
      "epoch": 19.76,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 82720
    },
    {
      "epoch": 19.76,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 82730
    },
    {
      "epoch": 19.76,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 82740
    },
    {
      "epoch": 19.76,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 82750
    },
    {
      "epoch": 19.77,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 82760
    },
    {
      "epoch": 19.77,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 82770
    },
    {
      "epoch": 19.77,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 82780
    },
    {
      "epoch": 19.77,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 82790
    },
    {
      "epoch": 19.78,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 82800
    },
    {
      "epoch": 19.78,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.405390739440918,
      "eval_runtime": 32.1999,
      "eval_samples_per_second": 21.242,
      "step": 82800
    },
    {
      "epoch": 19.78,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 82810
    },
    {
      "epoch": 19.78,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 82820
    },
    {
      "epoch": 19.78,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 82830
    },
    {
      "epoch": 19.79,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 82840
    },
    {
      "epoch": 19.79,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 82850
    },
    {
      "epoch": 19.79,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 82860
    },
    {
      "epoch": 19.79,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 82870
    },
    {
      "epoch": 19.79,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 82880
    },
    {
      "epoch": 19.8,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 82890
    },
    {
      "epoch": 19.8,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 82900
    },
    {
      "epoch": 19.8,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4051998853683472,
      "eval_runtime": 32.1568,
      "eval_samples_per_second": 21.271,
      "step": 82900
    },
    {
      "epoch": 19.8,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 82910
    },
    {
      "epoch": 19.8,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 82920
    },
    {
      "epoch": 19.81,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 82930
    },
    {
      "epoch": 19.81,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 82940
    },
    {
      "epoch": 19.81,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 82950
    },
    {
      "epoch": 19.81,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 82960
    },
    {
      "epoch": 19.82,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 82970
    },
    {
      "epoch": 19.82,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 82980
    },
    {
      "epoch": 19.82,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 82990
    },
    {
      "epoch": 19.82,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 83000
    },
    {
      "epoch": 19.82,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4053828716278076,
      "eval_runtime": 31.9023,
      "eval_samples_per_second": 21.44,
      "step": 83000
    },
    {
      "epoch": 19.83,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 83010
    },
    {
      "epoch": 19.83,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 83020
    },
    {
      "epoch": 19.83,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 83030
    },
    {
      "epoch": 19.83,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 83040
    },
    {
      "epoch": 19.84,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 83050
    },
    {
      "epoch": 19.84,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 83060
    },
    {
      "epoch": 19.84,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 83070
    },
    {
      "epoch": 19.84,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 83080
    },
    {
      "epoch": 19.84,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 83090
    },
    {
      "epoch": 19.85,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 83100
    },
    {
      "epoch": 19.85,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4048792123794556,
      "eval_runtime": 32.0288,
      "eval_samples_per_second": 21.356,
      "step": 83100
    },
    {
      "epoch": 19.85,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 83110
    },
    {
      "epoch": 19.85,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 83120
    },
    {
      "epoch": 19.85,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 83130
    },
    {
      "epoch": 19.86,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 83140
    },
    {
      "epoch": 19.86,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 83150
    },
    {
      "epoch": 19.86,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 83160
    },
    {
      "epoch": 19.86,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 83170
    },
    {
      "epoch": 19.87,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 83180
    },
    {
      "epoch": 19.87,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 83190
    },
    {
      "epoch": 19.87,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 83200
    },
    {
      "epoch": 19.87,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4047346115112305,
      "eval_runtime": 32.1337,
      "eval_samples_per_second": 21.286,
      "step": 83200
    },
    {
      "epoch": 19.87,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 83210
    },
    {
      "epoch": 19.88,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 83220
    },
    {
      "epoch": 19.88,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 83230
    },
    {
      "epoch": 19.88,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 83240
    },
    {
      "epoch": 19.88,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 83250
    },
    {
      "epoch": 19.89,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 83260
    },
    {
      "epoch": 19.89,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 83270
    },
    {
      "epoch": 19.89,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 83280
    },
    {
      "epoch": 19.89,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 83290
    },
    {
      "epoch": 19.89,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 83300
    },
    {
      "epoch": 19.89,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.405104637145996,
      "eval_runtime": 32.0062,
      "eval_samples_per_second": 21.371,
      "step": 83300
    },
    {
      "epoch": 19.9,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 83310
    },
    {
      "epoch": 19.9,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 83320
    },
    {
      "epoch": 19.9,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 83330
    },
    {
      "epoch": 19.9,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 83340
    },
    {
      "epoch": 19.91,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 83350
    },
    {
      "epoch": 19.91,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 83360
    },
    {
      "epoch": 19.91,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 83370
    },
    {
      "epoch": 19.91,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 83380
    },
    {
      "epoch": 19.92,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 83390
    },
    {
      "epoch": 19.92,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 83400
    },
    {
      "epoch": 19.92,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4050507545471191,
      "eval_runtime": 31.9535,
      "eval_samples_per_second": 21.406,
      "step": 83400
    },
    {
      "epoch": 19.92,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 83410
    },
    {
      "epoch": 19.92,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 83420
    },
    {
      "epoch": 19.93,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 83430
    },
    {
      "epoch": 19.93,
      "learning_rate": 1e-05,
      "loss": 1.3983,
      "step": 83440
    },
    {
      "epoch": 19.93,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 83450
    },
    {
      "epoch": 19.93,
      "learning_rate": 1e-05,
      "loss": 1.406,
      "step": 83460
    },
    {
      "epoch": 19.94,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 83470
    },
    {
      "epoch": 19.94,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 83480
    },
    {
      "epoch": 19.94,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 83490
    },
    {
      "epoch": 19.94,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 83500
    },
    {
      "epoch": 19.94,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4053587913513184,
      "eval_runtime": 32.4156,
      "eval_samples_per_second": 21.101,
      "step": 83500
    },
    {
      "epoch": 19.95,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 83510
    },
    {
      "epoch": 19.95,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 83520
    },
    {
      "epoch": 19.95,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 83530
    },
    {
      "epoch": 19.95,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 83540
    },
    {
      "epoch": 19.95,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 83550
    },
    {
      "epoch": 19.96,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 83560
    },
    {
      "epoch": 19.96,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 83570
    },
    {
      "epoch": 19.96,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 83580
    },
    {
      "epoch": 19.96,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 83590
    },
    {
      "epoch": 19.97,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 83600
    },
    {
      "epoch": 19.97,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4045401811599731,
      "eval_runtime": 32.1086,
      "eval_samples_per_second": 21.303,
      "step": 83600
    },
    {
      "epoch": 19.97,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 83610
    },
    {
      "epoch": 19.97,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 83620
    },
    {
      "epoch": 19.97,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 83630
    },
    {
      "epoch": 19.98,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 83640
    },
    {
      "epoch": 19.98,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 83650
    },
    {
      "epoch": 19.98,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 83660
    },
    {
      "epoch": 19.98,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 83670
    },
    {
      "epoch": 19.99,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 83680
    },
    {
      "epoch": 19.99,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 83690
    },
    {
      "epoch": 19.99,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 83700
    },
    {
      "epoch": 19.99,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4042599201202393,
      "eval_runtime": 32.0983,
      "eval_samples_per_second": 21.31,
      "step": 83700
    },
    {
      "epoch": 19.99,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 83710
    },
    {
      "epoch": 20.0,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 83720
    },
    {
      "epoch": 20.0,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 83730
    },
    {
      "epoch": 20.0,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 83740
    },
    {
      "epoch": 20.0,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 83750
    },
    {
      "epoch": 20.0,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 83760
    },
    {
      "epoch": 20.01,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 83770
    },
    {
      "epoch": 20.01,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 83780
    },
    {
      "epoch": 20.01,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 83790
    },
    {
      "epoch": 20.01,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 83800
    },
    {
      "epoch": 20.01,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4048717021942139,
      "eval_runtime": 31.6524,
      "eval_samples_per_second": 21.61,
      "step": 83800
    },
    {
      "epoch": 20.02,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 83810
    },
    {
      "epoch": 20.02,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 83820
    },
    {
      "epoch": 20.02,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 83830
    },
    {
      "epoch": 20.02,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 83840
    },
    {
      "epoch": 20.03,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 83850
    },
    {
      "epoch": 20.03,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 83860
    },
    {
      "epoch": 20.03,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 83870
    },
    {
      "epoch": 20.03,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 83880
    },
    {
      "epoch": 20.04,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 83890
    },
    {
      "epoch": 20.04,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 83900
    },
    {
      "epoch": 20.04,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.404159426689148,
      "eval_runtime": 31.9947,
      "eval_samples_per_second": 21.379,
      "step": 83900
    },
    {
      "epoch": 20.04,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 83910
    },
    {
      "epoch": 20.04,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 83920
    },
    {
      "epoch": 20.05,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 83930
    },
    {
      "epoch": 20.05,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 83940
    },
    {
      "epoch": 20.05,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 83950
    },
    {
      "epoch": 20.05,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 83960
    },
    {
      "epoch": 20.05,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 83970
    },
    {
      "epoch": 20.06,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 83980
    },
    {
      "epoch": 20.06,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 83990
    },
    {
      "epoch": 20.06,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 84000
    },
    {
      "epoch": 20.06,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4052512645721436,
      "eval_runtime": 31.8059,
      "eval_samples_per_second": 21.505,
      "step": 84000
    },
    {
      "epoch": 20.06,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 84010
    },
    {
      "epoch": 20.07,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 84020
    },
    {
      "epoch": 20.07,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 84030
    },
    {
      "epoch": 20.07,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 84040
    },
    {
      "epoch": 20.07,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 84050
    },
    {
      "epoch": 20.08,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 84060
    },
    {
      "epoch": 20.08,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 84070
    },
    {
      "epoch": 20.08,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 84080
    },
    {
      "epoch": 20.08,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 84090
    },
    {
      "epoch": 20.09,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 84100
    },
    {
      "epoch": 20.09,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4047669172286987,
      "eval_runtime": 32.4149,
      "eval_samples_per_second": 21.101,
      "step": 84100
    },
    {
      "epoch": 20.09,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 84110
    },
    {
      "epoch": 20.09,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 84120
    },
    {
      "epoch": 20.09,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 84130
    },
    {
      "epoch": 20.1,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 84140
    },
    {
      "epoch": 20.1,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 84150
    },
    {
      "epoch": 20.1,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 84160
    },
    {
      "epoch": 20.1,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 84170
    },
    {
      "epoch": 20.11,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 84180
    },
    {
      "epoch": 20.11,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 84190
    },
    {
      "epoch": 20.11,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 84200
    },
    {
      "epoch": 20.11,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4045883417129517,
      "eval_runtime": 31.8968,
      "eval_samples_per_second": 21.444,
      "step": 84200
    },
    {
      "epoch": 20.11,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 84210
    },
    {
      "epoch": 20.11,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 84220
    },
    {
      "epoch": 20.12,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 84230
    },
    {
      "epoch": 20.12,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 84240
    },
    {
      "epoch": 20.12,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 84250
    },
    {
      "epoch": 20.12,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 84260
    },
    {
      "epoch": 20.13,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 84270
    },
    {
      "epoch": 20.13,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 84280
    },
    {
      "epoch": 20.13,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 84290
    },
    {
      "epoch": 20.13,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 84300
    },
    {
      "epoch": 20.13,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4052505493164062,
      "eval_runtime": 32.1206,
      "eval_samples_per_second": 21.295,
      "step": 84300
    },
    {
      "epoch": 20.14,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 84310
    },
    {
      "epoch": 20.14,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 84320
    },
    {
      "epoch": 20.14,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 84330
    },
    {
      "epoch": 20.14,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 84340
    },
    {
      "epoch": 20.15,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 84350
    },
    {
      "epoch": 20.15,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 84360
    },
    {
      "epoch": 20.15,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 84370
    },
    {
      "epoch": 20.15,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 84380
    },
    {
      "epoch": 20.16,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 84390
    },
    {
      "epoch": 20.16,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 84400
    },
    {
      "epoch": 20.16,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.404160737991333,
      "eval_runtime": 31.9371,
      "eval_samples_per_second": 21.417,
      "step": 84400
    },
    {
      "epoch": 20.16,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 84410
    },
    {
      "epoch": 20.16,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 84420
    },
    {
      "epoch": 20.16,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 84430
    },
    {
      "epoch": 20.17,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 84440
    },
    {
      "epoch": 20.17,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 84450
    },
    {
      "epoch": 20.17,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 84460
    },
    {
      "epoch": 20.17,
      "learning_rate": 1e-05,
      "loss": 1.4101,
      "step": 84470
    },
    {
      "epoch": 20.18,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 84480
    },
    {
      "epoch": 20.18,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 84490
    },
    {
      "epoch": 20.18,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 84500
    },
    {
      "epoch": 20.18,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4046443700790405,
      "eval_runtime": 31.9491,
      "eval_samples_per_second": 21.409,
      "step": 84500
    },
    {
      "epoch": 20.18,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 84510
    },
    {
      "epoch": 20.19,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 84520
    },
    {
      "epoch": 20.19,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 84530
    },
    {
      "epoch": 20.19,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 84540
    },
    {
      "epoch": 20.19,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 84550
    },
    {
      "epoch": 20.2,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 84560
    },
    {
      "epoch": 20.2,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 84570
    },
    {
      "epoch": 20.2,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 84580
    },
    {
      "epoch": 20.2,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 84590
    },
    {
      "epoch": 20.21,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 84600
    },
    {
      "epoch": 20.21,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4044551849365234,
      "eval_runtime": 31.9532,
      "eval_samples_per_second": 21.406,
      "step": 84600
    },
    {
      "epoch": 20.21,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 84610
    },
    {
      "epoch": 20.21,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 84620
    },
    {
      "epoch": 20.21,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 84630
    },
    {
      "epoch": 20.21,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 84640
    },
    {
      "epoch": 20.22,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 84650
    },
    {
      "epoch": 20.22,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 84660
    },
    {
      "epoch": 20.22,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 84670
    },
    {
      "epoch": 20.22,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 84680
    },
    {
      "epoch": 20.23,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 84690
    },
    {
      "epoch": 20.23,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 84700
    },
    {
      "epoch": 20.23,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.405288577079773,
      "eval_runtime": 31.9253,
      "eval_samples_per_second": 21.425,
      "step": 84700
    },
    {
      "epoch": 20.23,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 84710
    },
    {
      "epoch": 20.23,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 84720
    },
    {
      "epoch": 20.24,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 84730
    },
    {
      "epoch": 20.24,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 84740
    },
    {
      "epoch": 20.24,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 84750
    },
    {
      "epoch": 20.24,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 84760
    },
    {
      "epoch": 20.25,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 84770
    },
    {
      "epoch": 20.25,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 84780
    },
    {
      "epoch": 20.25,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 84790
    },
    {
      "epoch": 20.25,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 84800
    },
    {
      "epoch": 20.25,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4046915769577026,
      "eval_runtime": 32.0669,
      "eval_samples_per_second": 21.33,
      "step": 84800
    },
    {
      "epoch": 20.26,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 84810
    },
    {
      "epoch": 20.26,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 84820
    },
    {
      "epoch": 20.26,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 84830
    },
    {
      "epoch": 20.26,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 84840
    },
    {
      "epoch": 20.27,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 84850
    },
    {
      "epoch": 20.27,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 84860
    },
    {
      "epoch": 20.27,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 84870
    },
    {
      "epoch": 20.27,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 84880
    },
    {
      "epoch": 20.27,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 84890
    },
    {
      "epoch": 20.28,
      "learning_rate": 1e-05,
      "loss": 1.408,
      "step": 84900
    },
    {
      "epoch": 20.28,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4052220582962036,
      "eval_runtime": 31.8513,
      "eval_samples_per_second": 21.475,
      "step": 84900
    },
    {
      "epoch": 20.28,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 84910
    },
    {
      "epoch": 20.28,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 84920
    },
    {
      "epoch": 20.28,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 84930
    },
    {
      "epoch": 20.29,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 84940
    },
    {
      "epoch": 20.29,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 84950
    },
    {
      "epoch": 20.29,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 84960
    },
    {
      "epoch": 20.29,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 84970
    },
    {
      "epoch": 20.3,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 84980
    },
    {
      "epoch": 20.3,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 84990
    },
    {
      "epoch": 20.3,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 85000
    },
    {
      "epoch": 20.3,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4047048091888428,
      "eval_runtime": 31.6344,
      "eval_samples_per_second": 21.622,
      "step": 85000
    },
    {
      "epoch": 20.3,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 85010
    },
    {
      "epoch": 20.31,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 85020
    },
    {
      "epoch": 20.31,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 85030
    },
    {
      "epoch": 20.31,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 85040
    },
    {
      "epoch": 20.31,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 85050
    },
    {
      "epoch": 20.32,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 85060
    },
    {
      "epoch": 20.32,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 85070
    },
    {
      "epoch": 20.32,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 85080
    },
    {
      "epoch": 20.32,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 85090
    },
    {
      "epoch": 20.32,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 85100
    },
    {
      "epoch": 20.32,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4048186540603638,
      "eval_runtime": 32.7018,
      "eval_samples_per_second": 20.916,
      "step": 85100
    },
    {
      "epoch": 20.33,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 85110
    },
    {
      "epoch": 20.33,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 85120
    },
    {
      "epoch": 20.33,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 85130
    },
    {
      "epoch": 20.33,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 85140
    },
    {
      "epoch": 20.34,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 85150
    },
    {
      "epoch": 20.34,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 85160
    },
    {
      "epoch": 20.34,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 85170
    },
    {
      "epoch": 20.34,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 85180
    },
    {
      "epoch": 20.35,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 85190
    },
    {
      "epoch": 20.35,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 85200
    },
    {
      "epoch": 20.35,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4050184488296509,
      "eval_runtime": 32.2437,
      "eval_samples_per_second": 21.213,
      "step": 85200
    },
    {
      "epoch": 20.35,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 85210
    },
    {
      "epoch": 20.35,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 85220
    },
    {
      "epoch": 20.36,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 85230
    },
    {
      "epoch": 20.36,
      "learning_rate": 1e-05,
      "loss": 1.4075,
      "step": 85240
    },
    {
      "epoch": 20.36,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 85250
    },
    {
      "epoch": 20.36,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 85260
    },
    {
      "epoch": 20.37,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 85270
    },
    {
      "epoch": 20.37,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 85280
    },
    {
      "epoch": 20.37,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 85290
    },
    {
      "epoch": 20.37,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 85300
    },
    {
      "epoch": 20.37,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4049818515777588,
      "eval_runtime": 32.4213,
      "eval_samples_per_second": 21.097,
      "step": 85300
    },
    {
      "epoch": 20.37,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 85310
    },
    {
      "epoch": 20.38,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 85320
    },
    {
      "epoch": 20.38,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 85330
    },
    {
      "epoch": 20.38,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 85340
    },
    {
      "epoch": 20.38,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 85350
    },
    {
      "epoch": 20.39,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 85360
    },
    {
      "epoch": 20.39,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 85370
    },
    {
      "epoch": 20.39,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 85380
    },
    {
      "epoch": 20.39,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 85390
    },
    {
      "epoch": 20.4,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 85400
    },
    {
      "epoch": 20.4,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4039840698242188,
      "eval_runtime": 31.7216,
      "eval_samples_per_second": 21.563,
      "step": 85400
    },
    {
      "epoch": 20.4,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 85410
    },
    {
      "epoch": 20.4,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 85420
    },
    {
      "epoch": 20.4,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 85430
    },
    {
      "epoch": 20.41,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 85440
    },
    {
      "epoch": 20.41,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 85450
    },
    {
      "epoch": 20.41,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 85460
    },
    {
      "epoch": 20.41,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 85470
    },
    {
      "epoch": 20.42,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 85480
    },
    {
      "epoch": 20.42,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 85490
    },
    {
      "epoch": 20.42,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 85500
    },
    {
      "epoch": 20.42,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4038488864898682,
      "eval_runtime": 32.1914,
      "eval_samples_per_second": 21.248,
      "step": 85500
    },
    {
      "epoch": 20.42,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 85510
    },
    {
      "epoch": 20.43,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 85520
    },
    {
      "epoch": 20.43,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 85530
    },
    {
      "epoch": 20.43,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 85540
    },
    {
      "epoch": 20.43,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 85550
    },
    {
      "epoch": 20.43,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 85560
    },
    {
      "epoch": 20.44,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 85570
    },
    {
      "epoch": 20.44,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 85580
    },
    {
      "epoch": 20.44,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 85590
    },
    {
      "epoch": 20.44,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 85600
    },
    {
      "epoch": 20.44,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4044910669326782,
      "eval_runtime": 32.0816,
      "eval_samples_per_second": 21.321,
      "step": 85600
    },
    {
      "epoch": 20.45,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 85610
    },
    {
      "epoch": 20.45,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 85620
    },
    {
      "epoch": 20.45,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 85630
    },
    {
      "epoch": 20.45,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 85640
    },
    {
      "epoch": 20.46,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 85650
    },
    {
      "epoch": 20.46,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 85660
    },
    {
      "epoch": 20.46,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 85670
    },
    {
      "epoch": 20.46,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 85680
    },
    {
      "epoch": 20.47,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 85690
    },
    {
      "epoch": 20.47,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 85700
    },
    {
      "epoch": 20.47,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.403740644454956,
      "eval_runtime": 32.0343,
      "eval_samples_per_second": 21.352,
      "step": 85700
    },
    {
      "epoch": 20.47,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 85710
    },
    {
      "epoch": 20.47,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 85720
    },
    {
      "epoch": 20.48,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 85730
    },
    {
      "epoch": 20.48,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 85740
    },
    {
      "epoch": 20.48,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 85750
    },
    {
      "epoch": 20.48,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 85760
    },
    {
      "epoch": 20.48,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 85770
    },
    {
      "epoch": 20.49,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 85780
    },
    {
      "epoch": 20.49,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 85790
    },
    {
      "epoch": 20.49,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 85800
    },
    {
      "epoch": 20.49,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4044698476791382,
      "eval_runtime": 31.9529,
      "eval_samples_per_second": 21.406,
      "step": 85800
    },
    {
      "epoch": 20.49,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 85810
    },
    {
      "epoch": 20.5,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 85820
    },
    {
      "epoch": 20.5,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 85830
    },
    {
      "epoch": 20.5,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 85840
    },
    {
      "epoch": 20.5,
      "learning_rate": 1e-05,
      "loss": 1.4102,
      "step": 85850
    },
    {
      "epoch": 20.51,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 85860
    },
    {
      "epoch": 20.51,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 85870
    },
    {
      "epoch": 20.51,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 85880
    },
    {
      "epoch": 20.51,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 85890
    },
    {
      "epoch": 20.52,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 85900
    },
    {
      "epoch": 20.52,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.404272198677063,
      "eval_runtime": 32.2128,
      "eval_samples_per_second": 21.234,
      "step": 85900
    },
    {
      "epoch": 20.52,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 85910
    },
    {
      "epoch": 20.52,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 85920
    },
    {
      "epoch": 20.52,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 85930
    },
    {
      "epoch": 20.53,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 85940
    },
    {
      "epoch": 20.53,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 85950
    },
    {
      "epoch": 20.53,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 85960
    },
    {
      "epoch": 20.53,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 85970
    },
    {
      "epoch": 20.53,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 85980
    },
    {
      "epoch": 20.54,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 85990
    },
    {
      "epoch": 20.54,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 86000
    },
    {
      "epoch": 20.54,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4048305749893188,
      "eval_runtime": 32.0305,
      "eval_samples_per_second": 21.355,
      "step": 86000
    },
    {
      "epoch": 20.54,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 86010
    },
    {
      "epoch": 20.54,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 86020
    },
    {
      "epoch": 20.55,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 86030
    },
    {
      "epoch": 20.55,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 86040
    },
    {
      "epoch": 20.55,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 86050
    },
    {
      "epoch": 20.55,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 86060
    },
    {
      "epoch": 20.56,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 86070
    },
    {
      "epoch": 20.56,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 86080
    },
    {
      "epoch": 20.56,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 86090
    },
    {
      "epoch": 20.56,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 86100
    },
    {
      "epoch": 20.56,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4046504497528076,
      "eval_runtime": 32.3185,
      "eval_samples_per_second": 21.164,
      "step": 86100
    },
    {
      "epoch": 20.57,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 86110
    },
    {
      "epoch": 20.57,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 86120
    },
    {
      "epoch": 20.57,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 86130
    },
    {
      "epoch": 20.57,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 86140
    },
    {
      "epoch": 20.58,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 86150
    },
    {
      "epoch": 20.58,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 86160
    },
    {
      "epoch": 20.58,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 86170
    },
    {
      "epoch": 20.58,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 86180
    },
    {
      "epoch": 20.59,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 86190
    },
    {
      "epoch": 20.59,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 86200
    },
    {
      "epoch": 20.59,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4032198190689087,
      "eval_runtime": 32.107,
      "eval_samples_per_second": 21.304,
      "step": 86200
    },
    {
      "epoch": 20.59,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 86210
    },
    {
      "epoch": 20.59,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 86220
    },
    {
      "epoch": 20.59,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 86230
    },
    {
      "epoch": 20.6,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 86240
    },
    {
      "epoch": 20.6,
      "learning_rate": 1e-05,
      "loss": 1.4131,
      "step": 86250
    },
    {
      "epoch": 20.6,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 86260
    },
    {
      "epoch": 20.6,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 86270
    },
    {
      "epoch": 20.61,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 86280
    },
    {
      "epoch": 20.61,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 86290
    },
    {
      "epoch": 20.61,
      "learning_rate": 1e-05,
      "loss": 1.4134,
      "step": 86300
    },
    {
      "epoch": 20.61,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4045143127441406,
      "eval_runtime": 32.082,
      "eval_samples_per_second": 21.32,
      "step": 86300
    },
    {
      "epoch": 20.61,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 86310
    },
    {
      "epoch": 20.62,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 86320
    },
    {
      "epoch": 20.62,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 86330
    },
    {
      "epoch": 20.62,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 86340
    },
    {
      "epoch": 20.62,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 86350
    },
    {
      "epoch": 20.63,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 86360
    },
    {
      "epoch": 20.63,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 86370
    },
    {
      "epoch": 20.63,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 86380
    },
    {
      "epoch": 20.63,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 86390
    },
    {
      "epoch": 20.64,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 86400
    },
    {
      "epoch": 20.64,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4050848484039307,
      "eval_runtime": 32.0816,
      "eval_samples_per_second": 21.321,
      "step": 86400
    },
    {
      "epoch": 20.64,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 86410
    },
    {
      "epoch": 20.64,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 86420
    },
    {
      "epoch": 20.64,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 86430
    },
    {
      "epoch": 20.64,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 86440
    },
    {
      "epoch": 20.65,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 86450
    },
    {
      "epoch": 20.65,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 86460
    },
    {
      "epoch": 20.65,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 86470
    },
    {
      "epoch": 20.65,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 86480
    },
    {
      "epoch": 20.66,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 86490
    },
    {
      "epoch": 20.66,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 86500
    },
    {
      "epoch": 20.66,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4041001796722412,
      "eval_runtime": 32.0879,
      "eval_samples_per_second": 21.316,
      "step": 86500
    },
    {
      "epoch": 20.66,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 86510
    },
    {
      "epoch": 20.66,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 86520
    },
    {
      "epoch": 20.67,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 86530
    },
    {
      "epoch": 20.67,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 86540
    },
    {
      "epoch": 20.67,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 86550
    },
    {
      "epoch": 20.67,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 86560
    },
    {
      "epoch": 20.68,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 86570
    },
    {
      "epoch": 20.68,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 86580
    },
    {
      "epoch": 20.68,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 86590
    },
    {
      "epoch": 20.68,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 86600
    },
    {
      "epoch": 20.68,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4050424098968506,
      "eval_runtime": 32.1029,
      "eval_samples_per_second": 21.306,
      "step": 86600
    },
    {
      "epoch": 20.69,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 86610
    },
    {
      "epoch": 20.69,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 86620
    },
    {
      "epoch": 20.69,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 86630
    },
    {
      "epoch": 20.69,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 86640
    },
    {
      "epoch": 20.7,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 86650
    },
    {
      "epoch": 20.7,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 86660
    },
    {
      "epoch": 20.7,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 86670
    },
    {
      "epoch": 20.7,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 86680
    },
    {
      "epoch": 20.7,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 86690
    },
    {
      "epoch": 20.71,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 86700
    },
    {
      "epoch": 20.71,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4040746688842773,
      "eval_runtime": 32.1454,
      "eval_samples_per_second": 21.278,
      "step": 86700
    },
    {
      "epoch": 20.71,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 86710
    },
    {
      "epoch": 20.71,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 86720
    },
    {
      "epoch": 20.71,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 86730
    },
    {
      "epoch": 20.72,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 86740
    },
    {
      "epoch": 20.72,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 86750
    },
    {
      "epoch": 20.72,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 86760
    },
    {
      "epoch": 20.72,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 86770
    },
    {
      "epoch": 20.73,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 86780
    },
    {
      "epoch": 20.73,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 86790
    },
    {
      "epoch": 20.73,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 86800
    },
    {
      "epoch": 20.73,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4038856029510498,
      "eval_runtime": 32.2579,
      "eval_samples_per_second": 21.204,
      "step": 86800
    },
    {
      "epoch": 20.73,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 86810
    },
    {
      "epoch": 20.74,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 86820
    },
    {
      "epoch": 20.74,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 86830
    },
    {
      "epoch": 20.74,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 86840
    },
    {
      "epoch": 20.74,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 86850
    },
    {
      "epoch": 20.75,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 86860
    },
    {
      "epoch": 20.75,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 86870
    },
    {
      "epoch": 20.75,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 86880
    },
    {
      "epoch": 20.75,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 86890
    },
    {
      "epoch": 20.75,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 86900
    },
    {
      "epoch": 20.75,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.404587984085083,
      "eval_runtime": 31.9877,
      "eval_samples_per_second": 21.383,
      "step": 86900
    },
    {
      "epoch": 20.76,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 86910
    },
    {
      "epoch": 20.76,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 86920
    },
    {
      "epoch": 20.76,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 86930
    },
    {
      "epoch": 20.76,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 86940
    },
    {
      "epoch": 20.77,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 86950
    },
    {
      "epoch": 20.77,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 86960
    },
    {
      "epoch": 20.77,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 86970
    },
    {
      "epoch": 20.77,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 86980
    },
    {
      "epoch": 20.78,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 86990
    },
    {
      "epoch": 20.78,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 87000
    },
    {
      "epoch": 20.78,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4037946462631226,
      "eval_runtime": 31.9379,
      "eval_samples_per_second": 21.417,
      "step": 87000
    },
    {
      "epoch": 20.78,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 87010
    },
    {
      "epoch": 20.78,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 87020
    },
    {
      "epoch": 20.79,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 87030
    },
    {
      "epoch": 20.79,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 87040
    },
    {
      "epoch": 20.79,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 87050
    },
    {
      "epoch": 20.79,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 87060
    },
    {
      "epoch": 20.8,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 87070
    },
    {
      "epoch": 20.8,
      "learning_rate": 1e-05,
      "loss": 1.3962,
      "step": 87080
    },
    {
      "epoch": 20.8,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 87090
    },
    {
      "epoch": 20.8,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 87100
    },
    {
      "epoch": 20.8,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4045172929763794,
      "eval_runtime": 31.9541,
      "eval_samples_per_second": 21.406,
      "step": 87100
    },
    {
      "epoch": 20.8,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 87110
    },
    {
      "epoch": 20.81,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 87120
    },
    {
      "epoch": 20.81,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 87130
    },
    {
      "epoch": 20.81,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 87140
    },
    {
      "epoch": 20.81,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 87150
    },
    {
      "epoch": 20.82,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 87160
    },
    {
      "epoch": 20.82,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 87170
    },
    {
      "epoch": 20.82,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 87180
    },
    {
      "epoch": 20.82,
      "learning_rate": 1e-05,
      "loss": 1.398,
      "step": 87190
    },
    {
      "epoch": 20.83,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 87200
    },
    {
      "epoch": 20.83,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4044653177261353,
      "eval_runtime": 31.9933,
      "eval_samples_per_second": 21.379,
      "step": 87200
    },
    {
      "epoch": 20.83,
      "learning_rate": 1e-05,
      "loss": 1.4139,
      "step": 87210
    },
    {
      "epoch": 20.83,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 87220
    },
    {
      "epoch": 20.83,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 87230
    },
    {
      "epoch": 20.84,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 87240
    },
    {
      "epoch": 20.84,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 87250
    },
    {
      "epoch": 20.84,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 87260
    },
    {
      "epoch": 20.84,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 87270
    },
    {
      "epoch": 20.85,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 87280
    },
    {
      "epoch": 20.85,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 87290
    },
    {
      "epoch": 20.85,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 87300
    },
    {
      "epoch": 20.85,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4047579765319824,
      "eval_runtime": 32.0619,
      "eval_samples_per_second": 21.334,
      "step": 87300
    },
    {
      "epoch": 20.85,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 87310
    },
    {
      "epoch": 20.86,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 87320
    },
    {
      "epoch": 20.86,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 87330
    },
    {
      "epoch": 20.86,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 87340
    },
    {
      "epoch": 20.86,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 87350
    },
    {
      "epoch": 20.86,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 87360
    },
    {
      "epoch": 20.87,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 87370
    },
    {
      "epoch": 20.87,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 87380
    },
    {
      "epoch": 20.87,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 87390
    },
    {
      "epoch": 20.87,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 87400
    },
    {
      "epoch": 20.87,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4036284685134888,
      "eval_runtime": 31.6014,
      "eval_samples_per_second": 21.645,
      "step": 87400
    },
    {
      "epoch": 20.88,
      "learning_rate": 1e-05,
      "loss": 1.3986,
      "step": 87410
    },
    {
      "epoch": 20.88,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 87420
    },
    {
      "epoch": 20.88,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 87430
    },
    {
      "epoch": 20.88,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 87440
    },
    {
      "epoch": 20.89,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 87450
    },
    {
      "epoch": 20.89,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 87460
    },
    {
      "epoch": 20.89,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 87470
    },
    {
      "epoch": 20.89,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 87480
    },
    {
      "epoch": 20.9,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 87490
    },
    {
      "epoch": 20.9,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 87500
    },
    {
      "epoch": 20.9,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.404073715209961,
      "eval_runtime": 32.0489,
      "eval_samples_per_second": 21.342,
      "step": 87500
    },
    {
      "epoch": 20.9,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 87510
    },
    {
      "epoch": 20.9,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 87520
    },
    {
      "epoch": 20.91,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 87530
    },
    {
      "epoch": 20.91,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 87540
    },
    {
      "epoch": 20.91,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 87550
    },
    {
      "epoch": 20.91,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 87560
    },
    {
      "epoch": 20.91,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 87570
    },
    {
      "epoch": 20.92,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 87580
    },
    {
      "epoch": 20.92,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 87590
    },
    {
      "epoch": 20.92,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 87600
    },
    {
      "epoch": 20.92,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4039759635925293,
      "eval_runtime": 32.1313,
      "eval_samples_per_second": 21.288,
      "step": 87600
    },
    {
      "epoch": 20.92,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 87610
    },
    {
      "epoch": 20.93,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 87620
    },
    {
      "epoch": 20.93,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 87630
    },
    {
      "epoch": 20.93,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 87640
    },
    {
      "epoch": 20.93,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 87650
    },
    {
      "epoch": 20.94,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 87660
    },
    {
      "epoch": 20.94,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 87670
    },
    {
      "epoch": 20.94,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 87680
    },
    {
      "epoch": 20.94,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 87690
    },
    {
      "epoch": 20.95,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 87700
    },
    {
      "epoch": 20.95,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4035841226577759,
      "eval_runtime": 32.0797,
      "eval_samples_per_second": 21.322,
      "step": 87700
    },
    {
      "epoch": 20.95,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 87710
    },
    {
      "epoch": 20.95,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 87720
    },
    {
      "epoch": 20.95,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 87730
    },
    {
      "epoch": 20.96,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 87740
    },
    {
      "epoch": 20.96,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 87750
    },
    {
      "epoch": 20.96,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 87760
    },
    {
      "epoch": 20.96,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 87770
    },
    {
      "epoch": 20.96,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 87780
    },
    {
      "epoch": 20.97,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 87790
    },
    {
      "epoch": 20.97,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 87800
    },
    {
      "epoch": 20.97,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4036061763763428,
      "eval_runtime": 31.7524,
      "eval_samples_per_second": 21.542,
      "step": 87800
    },
    {
      "epoch": 20.97,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 87810
    },
    {
      "epoch": 20.97,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 87820
    },
    {
      "epoch": 20.98,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 87830
    },
    {
      "epoch": 20.98,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 87840
    },
    {
      "epoch": 20.98,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 87850
    },
    {
      "epoch": 20.98,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 87860
    },
    {
      "epoch": 20.99,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 87870
    },
    {
      "epoch": 20.99,
      "learning_rate": 1e-05,
      "loss": 1.4087,
      "step": 87880
    },
    {
      "epoch": 20.99,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 87890
    },
    {
      "epoch": 20.99,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 87900
    },
    {
      "epoch": 20.99,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4044063091278076,
      "eval_runtime": 31.7445,
      "eval_samples_per_second": 21.547,
      "step": 87900
    },
    {
      "epoch": 21.0,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 87910
    },
    {
      "epoch": 21.0,
      "learning_rate": 1e-05,
      "loss": 1.407,
      "step": 87920
    },
    {
      "epoch": 21.0,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 87930
    },
    {
      "epoch": 21.0,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 87940
    },
    {
      "epoch": 21.01,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 87950
    },
    {
      "epoch": 21.01,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 87960
    },
    {
      "epoch": 21.01,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 87970
    },
    {
      "epoch": 21.01,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 87980
    },
    {
      "epoch": 21.02,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 87990
    },
    {
      "epoch": 21.02,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 88000
    },
    {
      "epoch": 21.02,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4033160209655762,
      "eval_runtime": 32.0014,
      "eval_samples_per_second": 21.374,
      "step": 88000
    },
    {
      "epoch": 21.02,
      "learning_rate": 1e-05,
      "loss": 1.4059,
      "step": 88010
    },
    {
      "epoch": 21.02,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 88020
    },
    {
      "epoch": 21.02,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 88030
    },
    {
      "epoch": 21.03,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 88040
    },
    {
      "epoch": 21.03,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 88050
    },
    {
      "epoch": 21.03,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 88060
    },
    {
      "epoch": 21.03,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 88070
    },
    {
      "epoch": 21.04,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 88080
    },
    {
      "epoch": 21.04,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 88090
    },
    {
      "epoch": 21.04,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 88100
    },
    {
      "epoch": 21.04,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4038594961166382,
      "eval_runtime": 32.1816,
      "eval_samples_per_second": 21.254,
      "step": 88100
    },
    {
      "epoch": 21.04,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 88110
    },
    {
      "epoch": 21.05,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 88120
    },
    {
      "epoch": 21.05,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 88130
    },
    {
      "epoch": 21.05,
      "learning_rate": 1e-05,
      "loss": 1.3983,
      "step": 88140
    },
    {
      "epoch": 21.05,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 88150
    },
    {
      "epoch": 21.06,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 88160
    },
    {
      "epoch": 21.06,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 88170
    },
    {
      "epoch": 21.06,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 88180
    },
    {
      "epoch": 21.06,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 88190
    },
    {
      "epoch": 21.07,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 88200
    },
    {
      "epoch": 21.07,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.40395987033844,
      "eval_runtime": 32.0183,
      "eval_samples_per_second": 21.363,
      "step": 88200
    },
    {
      "epoch": 21.07,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 88210
    },
    {
      "epoch": 21.07,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 88220
    },
    {
      "epoch": 21.07,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 88230
    },
    {
      "epoch": 21.07,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 88240
    },
    {
      "epoch": 21.08,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 88250
    },
    {
      "epoch": 21.08,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 88260
    },
    {
      "epoch": 21.08,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 88270
    },
    {
      "epoch": 21.08,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 88280
    },
    {
      "epoch": 21.09,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 88290
    },
    {
      "epoch": 21.09,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 88300
    },
    {
      "epoch": 21.09,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4045109748840332,
      "eval_runtime": 31.8821,
      "eval_samples_per_second": 21.454,
      "step": 88300
    },
    {
      "epoch": 21.09,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 88310
    },
    {
      "epoch": 21.09,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 88320
    },
    {
      "epoch": 21.1,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 88330
    },
    {
      "epoch": 21.1,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 88340
    },
    {
      "epoch": 21.1,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 88350
    },
    {
      "epoch": 21.1,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 88360
    },
    {
      "epoch": 21.11,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 88370
    },
    {
      "epoch": 21.11,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 88380
    },
    {
      "epoch": 21.11,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 88390
    },
    {
      "epoch": 21.11,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 88400
    },
    {
      "epoch": 21.11,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4033379554748535,
      "eval_runtime": 32.0181,
      "eval_samples_per_second": 21.363,
      "step": 88400
    },
    {
      "epoch": 21.12,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 88410
    },
    {
      "epoch": 21.12,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 88420
    },
    {
      "epoch": 21.12,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 88430
    },
    {
      "epoch": 21.12,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 88440
    },
    {
      "epoch": 21.12,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 88450
    },
    {
      "epoch": 21.13,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 88460
    },
    {
      "epoch": 21.13,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 88470
    },
    {
      "epoch": 21.13,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 88480
    },
    {
      "epoch": 21.13,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 88490
    },
    {
      "epoch": 21.14,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 88500
    },
    {
      "epoch": 21.14,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.403783917427063,
      "eval_runtime": 31.933,
      "eval_samples_per_second": 21.42,
      "step": 88500
    },
    {
      "epoch": 21.14,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 88510
    },
    {
      "epoch": 21.14,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 88520
    },
    {
      "epoch": 21.14,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 88530
    },
    {
      "epoch": 21.15,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 88540
    },
    {
      "epoch": 21.15,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 88550
    },
    {
      "epoch": 21.15,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 88560
    },
    {
      "epoch": 21.15,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 88570
    },
    {
      "epoch": 21.16,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 88580
    },
    {
      "epoch": 21.16,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 88590
    },
    {
      "epoch": 21.16,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 88600
    },
    {
      "epoch": 21.16,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4043481349945068,
      "eval_runtime": 32.0351,
      "eval_samples_per_second": 21.352,
      "step": 88600
    },
    {
      "epoch": 21.16,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 88610
    },
    {
      "epoch": 21.17,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 88620
    },
    {
      "epoch": 21.17,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 88630
    },
    {
      "epoch": 21.17,
      "learning_rate": 1e-05,
      "loss": 1.4112,
      "step": 88640
    },
    {
      "epoch": 21.17,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 88650
    },
    {
      "epoch": 21.18,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 88660
    },
    {
      "epoch": 21.18,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 88670
    },
    {
      "epoch": 21.18,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 88680
    },
    {
      "epoch": 21.18,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 88690
    },
    {
      "epoch": 21.18,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 88700
    },
    {
      "epoch": 21.18,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4034467935562134,
      "eval_runtime": 31.7522,
      "eval_samples_per_second": 21.542,
      "step": 88700
    },
    {
      "epoch": 21.19,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 88710
    },
    {
      "epoch": 21.19,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 88720
    },
    {
      "epoch": 21.19,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 88730
    },
    {
      "epoch": 21.19,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 88740
    },
    {
      "epoch": 21.2,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 88750
    },
    {
      "epoch": 21.2,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 88760
    },
    {
      "epoch": 21.2,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 88770
    },
    {
      "epoch": 21.2,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 88780
    },
    {
      "epoch": 21.21,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 88790
    },
    {
      "epoch": 21.21,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 88800
    },
    {
      "epoch": 21.21,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4027494192123413,
      "eval_runtime": 31.9383,
      "eval_samples_per_second": 21.416,
      "step": 88800
    },
    {
      "epoch": 21.21,
      "learning_rate": 1e-05,
      "loss": 1.4076,
      "step": 88810
    },
    {
      "epoch": 21.21,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 88820
    },
    {
      "epoch": 21.22,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 88830
    },
    {
      "epoch": 21.22,
      "learning_rate": 1e-05,
      "loss": 1.4061,
      "step": 88840
    },
    {
      "epoch": 21.22,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 88850
    },
    {
      "epoch": 21.22,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 88860
    },
    {
      "epoch": 21.23,
      "learning_rate": 1e-05,
      "loss": 1.3981,
      "step": 88870
    },
    {
      "epoch": 21.23,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 88880
    },
    {
      "epoch": 21.23,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 88890
    },
    {
      "epoch": 21.23,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 88900
    },
    {
      "epoch": 21.23,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4031667709350586,
      "eval_runtime": 32.0587,
      "eval_samples_per_second": 21.336,
      "step": 88900
    },
    {
      "epoch": 21.23,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 88910
    },
    {
      "epoch": 21.24,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 88920
    },
    {
      "epoch": 21.24,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 88930
    },
    {
      "epoch": 21.24,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 88940
    },
    {
      "epoch": 21.24,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 88950
    },
    {
      "epoch": 21.25,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 88960
    },
    {
      "epoch": 21.25,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 88970
    },
    {
      "epoch": 21.25,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 88980
    },
    {
      "epoch": 21.25,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 88990
    },
    {
      "epoch": 21.26,
      "learning_rate": 1e-05,
      "loss": 1.3985,
      "step": 89000
    },
    {
      "epoch": 21.26,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4031285047531128,
      "eval_runtime": 32.0481,
      "eval_samples_per_second": 21.343,
      "step": 89000
    },
    {
      "epoch": 21.26,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 89010
    },
    {
      "epoch": 21.26,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 89020
    },
    {
      "epoch": 21.26,
      "learning_rate": 1e-05,
      "loss": 1.3985,
      "step": 89030
    },
    {
      "epoch": 21.27,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 89040
    },
    {
      "epoch": 21.27,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 89050
    },
    {
      "epoch": 21.27,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 89060
    },
    {
      "epoch": 21.27,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 89070
    },
    {
      "epoch": 21.28,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 89080
    },
    {
      "epoch": 21.28,
      "learning_rate": 1e-05,
      "loss": 1.3983,
      "step": 89090
    },
    {
      "epoch": 21.28,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 89100
    },
    {
      "epoch": 21.28,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4028685092926025,
      "eval_runtime": 31.8403,
      "eval_samples_per_second": 21.482,
      "step": 89100
    },
    {
      "epoch": 21.28,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 89110
    },
    {
      "epoch": 21.28,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 89120
    },
    {
      "epoch": 21.29,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 89130
    },
    {
      "epoch": 21.29,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 89140
    },
    {
      "epoch": 21.29,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 89150
    },
    {
      "epoch": 21.29,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 89160
    },
    {
      "epoch": 21.3,
      "learning_rate": 1e-05,
      "loss": 1.4068,
      "step": 89170
    },
    {
      "epoch": 21.3,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 89180
    },
    {
      "epoch": 21.3,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 89190
    },
    {
      "epoch": 21.3,
      "learning_rate": 1e-05,
      "loss": 1.3981,
      "step": 89200
    },
    {
      "epoch": 21.3,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4037227630615234,
      "eval_runtime": 31.9665,
      "eval_samples_per_second": 21.397,
      "step": 89200
    },
    {
      "epoch": 21.31,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 89210
    },
    {
      "epoch": 21.31,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 89220
    },
    {
      "epoch": 21.31,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 89230
    },
    {
      "epoch": 21.31,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 89240
    },
    {
      "epoch": 21.32,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 89250
    },
    {
      "epoch": 21.32,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 89260
    },
    {
      "epoch": 21.32,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 89270
    },
    {
      "epoch": 21.32,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 89280
    },
    {
      "epoch": 21.33,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 89290
    },
    {
      "epoch": 21.33,
      "learning_rate": 1e-05,
      "loss": 1.3976,
      "step": 89300
    },
    {
      "epoch": 21.33,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4039889574050903,
      "eval_runtime": 32.1738,
      "eval_samples_per_second": 21.26,
      "step": 89300
    },
    {
      "epoch": 21.33,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 89310
    },
    {
      "epoch": 21.33,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 89320
    },
    {
      "epoch": 21.34,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 89330
    },
    {
      "epoch": 21.34,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 89340
    },
    {
      "epoch": 21.34,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 89350
    },
    {
      "epoch": 21.34,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 89360
    },
    {
      "epoch": 21.34,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 89370
    },
    {
      "epoch": 21.35,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 89380
    },
    {
      "epoch": 21.35,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 89390
    },
    {
      "epoch": 21.35,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 89400
    },
    {
      "epoch": 21.35,
      "eval_accuracy": 0.9751461988304093,
      "eval_loss": 1.4036519527435303,
      "eval_runtime": 32.618,
      "eval_samples_per_second": 20.97,
      "step": 89400
    },
    {
      "epoch": 21.35,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 89410
    },
    {
      "epoch": 21.36,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 89420
    },
    {
      "epoch": 21.36,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 89430
    },
    {
      "epoch": 21.36,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 89440
    },
    {
      "epoch": 21.36,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 89450
    },
    {
      "epoch": 21.37,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 89460
    },
    {
      "epoch": 21.37,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 89470
    },
    {
      "epoch": 21.37,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 89480
    },
    {
      "epoch": 21.37,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 89490
    },
    {
      "epoch": 21.38,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 89500
    },
    {
      "epoch": 21.38,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4048292636871338,
      "eval_runtime": 32.2316,
      "eval_samples_per_second": 21.221,
      "step": 89500
    },
    {
      "epoch": 21.38,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 89510
    },
    {
      "epoch": 21.38,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 89520
    },
    {
      "epoch": 21.38,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 89530
    },
    {
      "epoch": 21.39,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 89540
    },
    {
      "epoch": 21.39,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 89550
    },
    {
      "epoch": 21.39,
      "learning_rate": 1e-05,
      "loss": 1.3982,
      "step": 89560
    },
    {
      "epoch": 21.39,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 89570
    },
    {
      "epoch": 21.39,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 89580
    },
    {
      "epoch": 21.4,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 89590
    },
    {
      "epoch": 21.4,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 89600
    },
    {
      "epoch": 21.4,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.403419852256775,
      "eval_runtime": 32.6888,
      "eval_samples_per_second": 20.925,
      "step": 89600
    },
    {
      "epoch": 21.4,
      "learning_rate": 1e-05,
      "loss": 1.3985,
      "step": 89610
    },
    {
      "epoch": 21.4,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 89620
    },
    {
      "epoch": 21.41,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 89630
    },
    {
      "epoch": 21.41,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 89640
    },
    {
      "epoch": 21.41,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 89650
    },
    {
      "epoch": 21.41,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 89660
    },
    {
      "epoch": 21.42,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 89670
    },
    {
      "epoch": 21.42,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 89680
    },
    {
      "epoch": 21.42,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 89690
    },
    {
      "epoch": 21.42,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 89700
    },
    {
      "epoch": 21.42,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4031355381011963,
      "eval_runtime": 31.9953,
      "eval_samples_per_second": 21.378,
      "step": 89700
    },
    {
      "epoch": 21.43,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 89710
    },
    {
      "epoch": 21.43,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 89720
    },
    {
      "epoch": 21.43,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 89730
    },
    {
      "epoch": 21.43,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 89740
    },
    {
      "epoch": 21.44,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 89750
    },
    {
      "epoch": 21.44,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 89760
    },
    {
      "epoch": 21.44,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 89770
    },
    {
      "epoch": 21.44,
      "learning_rate": 1e-05,
      "loss": 1.4071,
      "step": 89780
    },
    {
      "epoch": 21.44,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 89790
    },
    {
      "epoch": 21.45,
      "learning_rate": 1e-05,
      "loss": 1.3972,
      "step": 89800
    },
    {
      "epoch": 21.45,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4034134149551392,
      "eval_runtime": 32.2768,
      "eval_samples_per_second": 21.192,
      "step": 89800
    },
    {
      "epoch": 21.45,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 89810
    },
    {
      "epoch": 21.45,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 89820
    },
    {
      "epoch": 21.45,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 89830
    },
    {
      "epoch": 21.46,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 89840
    },
    {
      "epoch": 21.46,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 89850
    },
    {
      "epoch": 21.46,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 89860
    },
    {
      "epoch": 21.46,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 89870
    },
    {
      "epoch": 21.47,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 89880
    },
    {
      "epoch": 21.47,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 89890
    },
    {
      "epoch": 21.47,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 89900
    },
    {
      "epoch": 21.47,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4042439460754395,
      "eval_runtime": 32.042,
      "eval_samples_per_second": 21.347,
      "step": 89900
    },
    {
      "epoch": 21.47,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 89910
    },
    {
      "epoch": 21.48,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 89920
    },
    {
      "epoch": 21.48,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 89930
    },
    {
      "epoch": 21.48,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 89940
    },
    {
      "epoch": 21.48,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 89950
    },
    {
      "epoch": 21.49,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 89960
    },
    {
      "epoch": 21.49,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 89970
    },
    {
      "epoch": 21.49,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 89980
    },
    {
      "epoch": 21.49,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 89990
    },
    {
      "epoch": 21.5,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 90000
    },
    {
      "epoch": 21.5,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4031835794448853,
      "eval_runtime": 31.7647,
      "eval_samples_per_second": 21.533,
      "step": 90000
    },
    {
      "epoch": 21.5,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 90010
    },
    {
      "epoch": 21.5,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 90020
    },
    {
      "epoch": 21.5,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 90030
    },
    {
      "epoch": 21.5,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 90040
    },
    {
      "epoch": 21.51,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 90050
    },
    {
      "epoch": 21.51,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 90060
    },
    {
      "epoch": 21.51,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 90070
    },
    {
      "epoch": 21.51,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 90080
    },
    {
      "epoch": 21.52,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 90090
    },
    {
      "epoch": 21.52,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 90100
    },
    {
      "epoch": 21.52,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4033814668655396,
      "eval_runtime": 32.1846,
      "eval_samples_per_second": 21.252,
      "step": 90100
    },
    {
      "epoch": 21.52,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 90110
    },
    {
      "epoch": 21.52,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 90120
    },
    {
      "epoch": 21.53,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 90130
    },
    {
      "epoch": 21.53,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 90140
    },
    {
      "epoch": 21.53,
      "learning_rate": 1e-05,
      "loss": 1.3977,
      "step": 90150
    },
    {
      "epoch": 21.53,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 90160
    },
    {
      "epoch": 21.54,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 90170
    },
    {
      "epoch": 21.54,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 90180
    },
    {
      "epoch": 21.54,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 90190
    },
    {
      "epoch": 21.54,
      "learning_rate": 1e-05,
      "loss": 1.3987,
      "step": 90200
    },
    {
      "epoch": 21.54,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4035471677780151,
      "eval_runtime": 31.9679,
      "eval_samples_per_second": 21.396,
      "step": 90200
    },
    {
      "epoch": 21.55,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 90210
    },
    {
      "epoch": 21.55,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 90220
    },
    {
      "epoch": 21.55,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 90230
    },
    {
      "epoch": 21.55,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 90240
    },
    {
      "epoch": 21.55,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 90250
    },
    {
      "epoch": 21.56,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 90260
    },
    {
      "epoch": 21.56,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 90270
    },
    {
      "epoch": 21.56,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 90280
    },
    {
      "epoch": 21.56,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 90290
    },
    {
      "epoch": 21.57,
      "learning_rate": 1e-05,
      "loss": 1.4138,
      "step": 90300
    },
    {
      "epoch": 21.57,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.403402328491211,
      "eval_runtime": 32.0538,
      "eval_samples_per_second": 21.339,
      "step": 90300
    },
    {
      "epoch": 21.57,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 90310
    },
    {
      "epoch": 21.57,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 90320
    },
    {
      "epoch": 21.57,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 90330
    },
    {
      "epoch": 21.58,
      "learning_rate": 1e-05,
      "loss": 1.4063,
      "step": 90340
    },
    {
      "epoch": 21.58,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 90350
    },
    {
      "epoch": 21.58,
      "learning_rate": 1e-05,
      "loss": 1.4084,
      "step": 90360
    },
    {
      "epoch": 21.58,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 90370
    },
    {
      "epoch": 21.59,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 90380
    },
    {
      "epoch": 21.59,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 90390
    },
    {
      "epoch": 21.59,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 90400
    },
    {
      "epoch": 21.59,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.403896689414978,
      "eval_runtime": 32.1022,
      "eval_samples_per_second": 21.307,
      "step": 90400
    },
    {
      "epoch": 21.59,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 90410
    },
    {
      "epoch": 21.6,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 90420
    },
    {
      "epoch": 21.6,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 90430
    },
    {
      "epoch": 21.6,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 90440
    },
    {
      "epoch": 21.6,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 90450
    },
    {
      "epoch": 21.6,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 90460
    },
    {
      "epoch": 21.61,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 90470
    },
    {
      "epoch": 21.61,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 90480
    },
    {
      "epoch": 21.61,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 90490
    },
    {
      "epoch": 21.61,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 90500
    },
    {
      "epoch": 21.61,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4031612873077393,
      "eval_runtime": 31.6812,
      "eval_samples_per_second": 21.59,
      "step": 90500
    },
    {
      "epoch": 21.62,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 90510
    },
    {
      "epoch": 21.62,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 90520
    },
    {
      "epoch": 21.62,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 90530
    },
    {
      "epoch": 21.62,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 90540
    },
    {
      "epoch": 21.63,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 90550
    },
    {
      "epoch": 21.63,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 90560
    },
    {
      "epoch": 21.63,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 90570
    },
    {
      "epoch": 21.63,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 90580
    },
    {
      "epoch": 21.64,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 90590
    },
    {
      "epoch": 21.64,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 90600
    },
    {
      "epoch": 21.64,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.402550458908081,
      "eval_runtime": 31.8225,
      "eval_samples_per_second": 21.494,
      "step": 90600
    },
    {
      "epoch": 21.64,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 90610
    },
    {
      "epoch": 21.64,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 90620
    },
    {
      "epoch": 21.65,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 90630
    },
    {
      "epoch": 21.65,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 90640
    },
    {
      "epoch": 21.65,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 90650
    },
    {
      "epoch": 21.65,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 90660
    },
    {
      "epoch": 21.66,
      "learning_rate": 1e-05,
      "loss": 1.3984,
      "step": 90670
    },
    {
      "epoch": 21.66,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 90680
    },
    {
      "epoch": 21.66,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 90690
    },
    {
      "epoch": 21.66,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 90700
    },
    {
      "epoch": 21.66,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4038612842559814,
      "eval_runtime": 32.1533,
      "eval_samples_per_second": 21.273,
      "step": 90700
    },
    {
      "epoch": 21.66,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 90710
    },
    {
      "epoch": 21.67,
      "learning_rate": 1e-05,
      "loss": 1.4051,
      "step": 90720
    },
    {
      "epoch": 21.67,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 90730
    },
    {
      "epoch": 21.67,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 90740
    },
    {
      "epoch": 21.67,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 90750
    },
    {
      "epoch": 21.68,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 90760
    },
    {
      "epoch": 21.68,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 90770
    },
    {
      "epoch": 21.68,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 90780
    },
    {
      "epoch": 21.68,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 90790
    },
    {
      "epoch": 21.69,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 90800
    },
    {
      "epoch": 21.69,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4030394554138184,
      "eval_runtime": 32.0011,
      "eval_samples_per_second": 21.374,
      "step": 90800
    },
    {
      "epoch": 21.69,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 90810
    },
    {
      "epoch": 21.69,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 90820
    },
    {
      "epoch": 21.69,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 90830
    },
    {
      "epoch": 21.7,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 90840
    },
    {
      "epoch": 21.7,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 90850
    },
    {
      "epoch": 21.7,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 90860
    },
    {
      "epoch": 21.7,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 90870
    },
    {
      "epoch": 21.71,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 90880
    },
    {
      "epoch": 21.71,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 90890
    },
    {
      "epoch": 21.71,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 90900
    },
    {
      "epoch": 21.71,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4034457206726074,
      "eval_runtime": 32.1173,
      "eval_samples_per_second": 21.297,
      "step": 90900
    },
    {
      "epoch": 21.71,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 90910
    },
    {
      "epoch": 21.71,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 90920
    },
    {
      "epoch": 21.72,
      "learning_rate": 1e-05,
      "loss": 1.4065,
      "step": 90930
    },
    {
      "epoch": 21.72,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 90940
    },
    {
      "epoch": 21.72,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 90950
    },
    {
      "epoch": 21.72,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 90960
    },
    {
      "epoch": 21.73,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 90970
    },
    {
      "epoch": 21.73,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 90980
    },
    {
      "epoch": 21.73,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 90990
    },
    {
      "epoch": 21.73,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 91000
    },
    {
      "epoch": 21.73,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4038020372390747,
      "eval_runtime": 32.1435,
      "eval_samples_per_second": 21.28,
      "step": 91000
    },
    {
      "epoch": 21.74,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 91010
    },
    {
      "epoch": 21.74,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 91020
    },
    {
      "epoch": 21.74,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 91030
    },
    {
      "epoch": 21.74,
      "learning_rate": 1e-05,
      "loss": 1.3972,
      "step": 91040
    },
    {
      "epoch": 21.75,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 91050
    },
    {
      "epoch": 21.75,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 91060
    },
    {
      "epoch": 21.75,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 91070
    },
    {
      "epoch": 21.75,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 91080
    },
    {
      "epoch": 21.76,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 91090
    },
    {
      "epoch": 21.76,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 91100
    },
    {
      "epoch": 21.76,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.403339147567749,
      "eval_runtime": 31.7216,
      "eval_samples_per_second": 21.563,
      "step": 91100
    },
    {
      "epoch": 21.76,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 91110
    },
    {
      "epoch": 21.76,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 91120
    },
    {
      "epoch": 21.76,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 91130
    },
    {
      "epoch": 21.77,
      "learning_rate": 1e-05,
      "loss": 1.3986,
      "step": 91140
    },
    {
      "epoch": 21.77,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 91150
    },
    {
      "epoch": 21.77,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 91160
    },
    {
      "epoch": 21.77,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 91170
    },
    {
      "epoch": 21.78,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 91180
    },
    {
      "epoch": 21.78,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 91190
    },
    {
      "epoch": 21.78,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 91200
    },
    {
      "epoch": 21.78,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4031703472137451,
      "eval_runtime": 31.7914,
      "eval_samples_per_second": 21.515,
      "step": 91200
    },
    {
      "epoch": 21.78,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 91210
    },
    {
      "epoch": 21.79,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 91220
    },
    {
      "epoch": 21.79,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 91230
    },
    {
      "epoch": 21.79,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 91240
    },
    {
      "epoch": 21.79,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 91250
    },
    {
      "epoch": 21.8,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 91260
    },
    {
      "epoch": 21.8,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 91270
    },
    {
      "epoch": 21.8,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 91280
    },
    {
      "epoch": 21.8,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 91290
    },
    {
      "epoch": 21.81,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 91300
    },
    {
      "epoch": 21.81,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4038935899734497,
      "eval_runtime": 31.9,
      "eval_samples_per_second": 21.442,
      "step": 91300
    },
    {
      "epoch": 21.81,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 91310
    },
    {
      "epoch": 21.81,
      "learning_rate": 1e-05,
      "loss": 1.4043,
      "step": 91320
    },
    {
      "epoch": 21.81,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 91330
    },
    {
      "epoch": 21.82,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 91340
    },
    {
      "epoch": 21.82,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 91350
    },
    {
      "epoch": 21.82,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 91360
    },
    {
      "epoch": 21.82,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 91370
    },
    {
      "epoch": 21.82,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 91380
    },
    {
      "epoch": 21.83,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 91390
    },
    {
      "epoch": 21.83,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 91400
    },
    {
      "epoch": 21.83,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4032721519470215,
      "eval_runtime": 32.0967,
      "eval_samples_per_second": 21.311,
      "step": 91400
    },
    {
      "epoch": 21.83,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 91410
    },
    {
      "epoch": 21.83,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 91420
    },
    {
      "epoch": 21.84,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 91430
    },
    {
      "epoch": 21.84,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 91440
    },
    {
      "epoch": 21.84,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 91450
    },
    {
      "epoch": 21.84,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 91460
    },
    {
      "epoch": 21.85,
      "learning_rate": 1e-05,
      "loss": 1.3972,
      "step": 91470
    },
    {
      "epoch": 21.85,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 91480
    },
    {
      "epoch": 21.85,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 91490
    },
    {
      "epoch": 21.85,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 91500
    },
    {
      "epoch": 21.85,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4028382301330566,
      "eval_runtime": 32.049,
      "eval_samples_per_second": 21.342,
      "step": 91500
    },
    {
      "epoch": 21.86,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 91510
    },
    {
      "epoch": 21.86,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 91520
    },
    {
      "epoch": 21.86,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 91530
    },
    {
      "epoch": 21.86,
      "learning_rate": 1e-05,
      "loss": 1.4078,
      "step": 91540
    },
    {
      "epoch": 21.87,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 91550
    },
    {
      "epoch": 21.87,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 91560
    },
    {
      "epoch": 21.87,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 91570
    },
    {
      "epoch": 21.87,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 91580
    },
    {
      "epoch": 21.87,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 91590
    },
    {
      "epoch": 21.88,
      "learning_rate": 1e-05,
      "loss": 1.3987,
      "step": 91600
    },
    {
      "epoch": 21.88,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.403244137763977,
      "eval_runtime": 32.2569,
      "eval_samples_per_second": 21.205,
      "step": 91600
    },
    {
      "epoch": 21.88,
      "learning_rate": 1e-05,
      "loss": 1.4056,
      "step": 91610
    },
    {
      "epoch": 21.88,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 91620
    },
    {
      "epoch": 21.88,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 91630
    },
    {
      "epoch": 21.89,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 91640
    },
    {
      "epoch": 21.89,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 91650
    },
    {
      "epoch": 21.89,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 91660
    },
    {
      "epoch": 21.89,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 91670
    },
    {
      "epoch": 21.9,
      "learning_rate": 1e-05,
      "loss": 1.4058,
      "step": 91680
    },
    {
      "epoch": 21.9,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 91690
    },
    {
      "epoch": 21.9,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 91700
    },
    {
      "epoch": 21.9,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4032223224639893,
      "eval_runtime": 32.1142,
      "eval_samples_per_second": 21.299,
      "step": 91700
    },
    {
      "epoch": 21.9,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 91710
    },
    {
      "epoch": 21.91,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 91720
    },
    {
      "epoch": 21.91,
      "learning_rate": 1e-05,
      "loss": 1.3985,
      "step": 91730
    },
    {
      "epoch": 21.91,
      "learning_rate": 1e-05,
      "loss": 1.4064,
      "step": 91740
    },
    {
      "epoch": 21.91,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 91750
    },
    {
      "epoch": 21.92,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 91760
    },
    {
      "epoch": 21.92,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 91770
    },
    {
      "epoch": 21.92,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 91780
    },
    {
      "epoch": 21.92,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 91790
    },
    {
      "epoch": 21.93,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 91800
    },
    {
      "epoch": 21.93,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4027495384216309,
      "eval_runtime": 32.4254,
      "eval_samples_per_second": 21.095,
      "step": 91800
    },
    {
      "epoch": 21.93,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 91810
    },
    {
      "epoch": 21.93,
      "learning_rate": 1e-05,
      "loss": 1.3987,
      "step": 91820
    },
    {
      "epoch": 21.93,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 91830
    },
    {
      "epoch": 21.93,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 91840
    },
    {
      "epoch": 21.94,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 91850
    },
    {
      "epoch": 21.94,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 91860
    },
    {
      "epoch": 21.94,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 91870
    },
    {
      "epoch": 21.94,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 91880
    },
    {
      "epoch": 21.95,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 91890
    },
    {
      "epoch": 21.95,
      "learning_rate": 1e-05,
      "loss": 1.3984,
      "step": 91900
    },
    {
      "epoch": 21.95,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4033875465393066,
      "eval_runtime": 31.9246,
      "eval_samples_per_second": 21.425,
      "step": 91900
    },
    {
      "epoch": 21.95,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 91910
    },
    {
      "epoch": 21.95,
      "learning_rate": 1e-05,
      "loss": 1.4057,
      "step": 91920
    },
    {
      "epoch": 21.96,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 91930
    },
    {
      "epoch": 21.96,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 91940
    },
    {
      "epoch": 21.96,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 91950
    },
    {
      "epoch": 21.96,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 91960
    },
    {
      "epoch": 21.97,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 91970
    },
    {
      "epoch": 21.97,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 91980
    },
    {
      "epoch": 21.97,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 91990
    },
    {
      "epoch": 21.97,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 92000
    },
    {
      "epoch": 21.97,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4039982557296753,
      "eval_runtime": 31.8015,
      "eval_samples_per_second": 21.508,
      "step": 92000
    },
    {
      "epoch": 21.98,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 92010
    },
    {
      "epoch": 21.98,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 92020
    },
    {
      "epoch": 21.98,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 92030
    },
    {
      "epoch": 21.98,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 92040
    },
    {
      "epoch": 21.98,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 92050
    },
    {
      "epoch": 21.99,
      "learning_rate": 1e-05,
      "loss": 1.3983,
      "step": 92060
    },
    {
      "epoch": 21.99,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 92070
    },
    {
      "epoch": 21.99,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 92080
    },
    {
      "epoch": 21.99,
      "learning_rate": 1e-05,
      "loss": 1.3985,
      "step": 92090
    },
    {
      "epoch": 22.0,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 92100
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4027361869812012,
      "eval_runtime": 32.1725,
      "eval_samples_per_second": 21.26,
      "step": 92100
    },
    {
      "epoch": 22.0,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 92110
    },
    {
      "epoch": 22.0,
      "learning_rate": 1e-05,
      "loss": 1.3986,
      "step": 92120
    },
    {
      "epoch": 22.0,
      "learning_rate": 1e-05,
      "loss": 1.3981,
      "step": 92130
    },
    {
      "epoch": 22.01,
      "learning_rate": 1e-05,
      "loss": 1.405,
      "step": 92140
    },
    {
      "epoch": 22.01,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 92150
    },
    {
      "epoch": 22.01,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 92160
    },
    {
      "epoch": 22.01,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 92170
    },
    {
      "epoch": 22.02,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 92180
    },
    {
      "epoch": 22.02,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 92190
    },
    {
      "epoch": 22.02,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 92200
    },
    {
      "epoch": 22.02,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4028061628341675,
      "eval_runtime": 31.8643,
      "eval_samples_per_second": 21.466,
      "step": 92200
    },
    {
      "epoch": 22.02,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 92210
    },
    {
      "epoch": 22.03,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 92220
    },
    {
      "epoch": 22.03,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 92230
    },
    {
      "epoch": 22.03,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 92240
    },
    {
      "epoch": 22.03,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 92250
    },
    {
      "epoch": 22.03,
      "learning_rate": 1e-05,
      "loss": 1.3982,
      "step": 92260
    },
    {
      "epoch": 22.04,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 92270
    },
    {
      "epoch": 22.04,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 92280
    },
    {
      "epoch": 22.04,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 92290
    },
    {
      "epoch": 22.04,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 92300
    },
    {
      "epoch": 22.04,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4033621549606323,
      "eval_runtime": 32.0946,
      "eval_samples_per_second": 21.312,
      "step": 92300
    },
    {
      "epoch": 22.05,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 92310
    },
    {
      "epoch": 22.05,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 92320
    },
    {
      "epoch": 22.05,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 92330
    },
    {
      "epoch": 22.05,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 92340
    },
    {
      "epoch": 22.06,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 92350
    },
    {
      "epoch": 22.06,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 92360
    },
    {
      "epoch": 22.06,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 92370
    },
    {
      "epoch": 22.06,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 92380
    },
    {
      "epoch": 22.07,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 92390
    },
    {
      "epoch": 22.07,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 92400
    },
    {
      "epoch": 22.07,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4031598567962646,
      "eval_runtime": 31.9859,
      "eval_samples_per_second": 21.384,
      "step": 92400
    },
    {
      "epoch": 22.07,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 92410
    },
    {
      "epoch": 22.07,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 92420
    },
    {
      "epoch": 22.08,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 92430
    },
    {
      "epoch": 22.08,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 92440
    },
    {
      "epoch": 22.08,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 92450
    },
    {
      "epoch": 22.08,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 92460
    },
    {
      "epoch": 22.09,
      "learning_rate": 1e-05,
      "loss": 1.4089,
      "step": 92470
    },
    {
      "epoch": 22.09,
      "learning_rate": 1e-05,
      "loss": 1.3983,
      "step": 92480
    },
    {
      "epoch": 22.09,
      "learning_rate": 1e-05,
      "loss": 1.3964,
      "step": 92490
    },
    {
      "epoch": 22.09,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 92500
    },
    {
      "epoch": 22.09,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4017704725265503,
      "eval_runtime": 32.1556,
      "eval_samples_per_second": 21.272,
      "step": 92500
    },
    {
      "epoch": 22.09,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 92510
    },
    {
      "epoch": 22.1,
      "learning_rate": 1e-05,
      "loss": 1.3959,
      "step": 92520
    },
    {
      "epoch": 22.1,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 92530
    },
    {
      "epoch": 22.1,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 92540
    },
    {
      "epoch": 22.1,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 92550
    },
    {
      "epoch": 22.11,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 92560
    },
    {
      "epoch": 22.11,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 92570
    },
    {
      "epoch": 22.11,
      "learning_rate": 1e-05,
      "loss": 1.3963,
      "step": 92580
    },
    {
      "epoch": 22.11,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 92590
    },
    {
      "epoch": 22.12,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 92600
    },
    {
      "epoch": 22.12,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4032747745513916,
      "eval_runtime": 32.1581,
      "eval_samples_per_second": 21.27,
      "step": 92600
    },
    {
      "epoch": 22.12,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 92610
    },
    {
      "epoch": 22.12,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 92620
    },
    {
      "epoch": 22.12,
      "learning_rate": 1e-05,
      "loss": 1.3981,
      "step": 92630
    },
    {
      "epoch": 22.13,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 92640
    },
    {
      "epoch": 22.13,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 92650
    },
    {
      "epoch": 22.13,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 92660
    },
    {
      "epoch": 22.13,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 92670
    },
    {
      "epoch": 22.14,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 92680
    },
    {
      "epoch": 22.14,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 92690
    },
    {
      "epoch": 22.14,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 92700
    },
    {
      "epoch": 22.14,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4032331705093384,
      "eval_runtime": 32.0429,
      "eval_samples_per_second": 21.346,
      "step": 92700
    },
    {
      "epoch": 22.14,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 92710
    },
    {
      "epoch": 22.14,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 92720
    },
    {
      "epoch": 22.15,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 92730
    },
    {
      "epoch": 22.15,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 92740
    },
    {
      "epoch": 22.15,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 92750
    },
    {
      "epoch": 22.15,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 92760
    },
    {
      "epoch": 22.16,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 92770
    },
    {
      "epoch": 22.16,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 92780
    },
    {
      "epoch": 22.16,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 92790
    },
    {
      "epoch": 22.16,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 92800
    },
    {
      "epoch": 22.16,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4034435749053955,
      "eval_runtime": 31.7313,
      "eval_samples_per_second": 21.556,
      "step": 92800
    },
    {
      "epoch": 22.17,
      "learning_rate": 1e-05,
      "loss": 1.4052,
      "step": 92810
    },
    {
      "epoch": 22.17,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 92820
    },
    {
      "epoch": 22.17,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 92830
    },
    {
      "epoch": 22.17,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 92840
    },
    {
      "epoch": 22.18,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 92850
    },
    {
      "epoch": 22.18,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 92860
    },
    {
      "epoch": 22.18,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 92870
    },
    {
      "epoch": 22.18,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 92880
    },
    {
      "epoch": 22.19,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 92890
    },
    {
      "epoch": 22.19,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 92900
    },
    {
      "epoch": 22.19,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4030529260635376,
      "eval_runtime": 32.1173,
      "eval_samples_per_second": 21.297,
      "step": 92900
    },
    {
      "epoch": 22.19,
      "learning_rate": 1e-05,
      "loss": 1.3963,
      "step": 92910
    },
    {
      "epoch": 22.19,
      "learning_rate": 1e-05,
      "loss": 1.3978,
      "step": 92920
    },
    {
      "epoch": 22.19,
      "learning_rate": 1e-05,
      "loss": 1.4081,
      "step": 92930
    },
    {
      "epoch": 22.2,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 92940
    },
    {
      "epoch": 22.2,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 92950
    },
    {
      "epoch": 22.2,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 92960
    },
    {
      "epoch": 22.2,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 92970
    },
    {
      "epoch": 22.21,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 92980
    },
    {
      "epoch": 22.21,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 92990
    },
    {
      "epoch": 22.21,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 93000
    },
    {
      "epoch": 22.21,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4034454822540283,
      "eval_runtime": 31.8364,
      "eval_samples_per_second": 21.485,
      "step": 93000
    },
    {
      "epoch": 22.21,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 93010
    },
    {
      "epoch": 22.22,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 93020
    },
    {
      "epoch": 22.22,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 93030
    },
    {
      "epoch": 22.22,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 93040
    },
    {
      "epoch": 22.22,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 93050
    },
    {
      "epoch": 22.23,
      "learning_rate": 1e-05,
      "loss": 1.3973,
      "step": 93060
    },
    {
      "epoch": 22.23,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 93070
    },
    {
      "epoch": 22.23,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 93080
    },
    {
      "epoch": 22.23,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 93090
    },
    {
      "epoch": 22.24,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 93100
    },
    {
      "epoch": 22.24,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4029674530029297,
      "eval_runtime": 31.8721,
      "eval_samples_per_second": 21.461,
      "step": 93100
    },
    {
      "epoch": 22.24,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 93110
    },
    {
      "epoch": 22.24,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 93120
    },
    {
      "epoch": 22.24,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 93130
    },
    {
      "epoch": 22.25,
      "learning_rate": 1e-05,
      "loss": 1.3982,
      "step": 93140
    },
    {
      "epoch": 22.25,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 93150
    },
    {
      "epoch": 22.25,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 93160
    },
    {
      "epoch": 22.25,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 93170
    },
    {
      "epoch": 22.25,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 93180
    },
    {
      "epoch": 22.26,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 93190
    },
    {
      "epoch": 22.26,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 93200
    },
    {
      "epoch": 22.26,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4029181003570557,
      "eval_runtime": 32.071,
      "eval_samples_per_second": 21.328,
      "step": 93200
    },
    {
      "epoch": 22.26,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 93210
    },
    {
      "epoch": 22.26,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 93220
    },
    {
      "epoch": 22.27,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 93230
    },
    {
      "epoch": 22.27,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 93240
    },
    {
      "epoch": 22.27,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 93250
    },
    {
      "epoch": 22.27,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 93260
    },
    {
      "epoch": 22.28,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 93270
    },
    {
      "epoch": 22.28,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 93280
    },
    {
      "epoch": 22.28,
      "learning_rate": 1e-05,
      "loss": 1.3984,
      "step": 93290
    },
    {
      "epoch": 22.28,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 93300
    },
    {
      "epoch": 22.28,
      "eval_accuracy": 0.97953216374269,
      "eval_loss": 1.4027594327926636,
      "eval_runtime": 32.4164,
      "eval_samples_per_second": 21.1,
      "step": 93300
    },
    {
      "epoch": 22.29,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 93310
    },
    {
      "epoch": 22.29,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 93320
    },
    {
      "epoch": 22.29,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 93330
    },
    {
      "epoch": 22.29,
      "learning_rate": 1e-05,
      "loss": 1.3974,
      "step": 93340
    },
    {
      "epoch": 22.3,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 93350
    },
    {
      "epoch": 22.3,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 93360
    },
    {
      "epoch": 22.3,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 93370
    },
    {
      "epoch": 22.3,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 93380
    },
    {
      "epoch": 22.3,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 93390
    },
    {
      "epoch": 22.31,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 93400
    },
    {
      "epoch": 22.31,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4037140607833862,
      "eval_runtime": 32.3224,
      "eval_samples_per_second": 21.162,
      "step": 93400
    },
    {
      "epoch": 22.31,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 93410
    },
    {
      "epoch": 22.31,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 93420
    },
    {
      "epoch": 22.31,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 93430
    },
    {
      "epoch": 22.32,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 93440
    },
    {
      "epoch": 22.32,
      "learning_rate": 1e-05,
      "loss": 1.4103,
      "step": 93450
    },
    {
      "epoch": 22.32,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 93460
    },
    {
      "epoch": 22.32,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 93470
    },
    {
      "epoch": 22.33,
      "learning_rate": 1e-05,
      "loss": 1.3983,
      "step": 93480
    },
    {
      "epoch": 22.33,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 93490
    },
    {
      "epoch": 22.33,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 93500
    },
    {
      "epoch": 22.33,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4026715755462646,
      "eval_runtime": 31.9862,
      "eval_samples_per_second": 21.384,
      "step": 93500
    },
    {
      "epoch": 22.33,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 93510
    },
    {
      "epoch": 22.34,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 93520
    },
    {
      "epoch": 22.34,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 93530
    },
    {
      "epoch": 22.34,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 93540
    },
    {
      "epoch": 22.34,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 93550
    },
    {
      "epoch": 22.35,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 93560
    },
    {
      "epoch": 22.35,
      "learning_rate": 1e-05,
      "loss": 1.3983,
      "step": 93570
    },
    {
      "epoch": 22.35,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 93580
    },
    {
      "epoch": 22.35,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 93590
    },
    {
      "epoch": 22.35,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 93600
    },
    {
      "epoch": 22.35,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4025092124938965,
      "eval_runtime": 32.2067,
      "eval_samples_per_second": 21.238,
      "step": 93600
    },
    {
      "epoch": 22.36,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 93610
    },
    {
      "epoch": 22.36,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 93620
    },
    {
      "epoch": 22.36,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 93630
    },
    {
      "epoch": 22.36,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 93640
    },
    {
      "epoch": 22.37,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 93650
    },
    {
      "epoch": 22.37,
      "learning_rate": 1e-05,
      "loss": 1.4073,
      "step": 93660
    },
    {
      "epoch": 22.37,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 93670
    },
    {
      "epoch": 22.37,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 93680
    },
    {
      "epoch": 22.38,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 93690
    },
    {
      "epoch": 22.38,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 93700
    },
    {
      "epoch": 22.38,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4024063348770142,
      "eval_runtime": 31.9711,
      "eval_samples_per_second": 21.394,
      "step": 93700
    },
    {
      "epoch": 22.38,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 93710
    },
    {
      "epoch": 22.38,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 93720
    },
    {
      "epoch": 22.39,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 93730
    },
    {
      "epoch": 22.39,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 93740
    },
    {
      "epoch": 22.39,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 93750
    },
    {
      "epoch": 22.39,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 93760
    },
    {
      "epoch": 22.4,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 93770
    },
    {
      "epoch": 22.4,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 93780
    },
    {
      "epoch": 22.4,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 93790
    },
    {
      "epoch": 22.4,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 93800
    },
    {
      "epoch": 22.4,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.402715802192688,
      "eval_runtime": 32.2502,
      "eval_samples_per_second": 21.209,
      "step": 93800
    },
    {
      "epoch": 22.41,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 93810
    },
    {
      "epoch": 22.41,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 93820
    },
    {
      "epoch": 22.41,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 93830
    },
    {
      "epoch": 22.41,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 93840
    },
    {
      "epoch": 22.41,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 93850
    },
    {
      "epoch": 22.42,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 93860
    },
    {
      "epoch": 22.42,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 93870
    },
    {
      "epoch": 22.42,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 93880
    },
    {
      "epoch": 22.42,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 93890
    },
    {
      "epoch": 22.43,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 93900
    },
    {
      "epoch": 22.43,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.403295874595642,
      "eval_runtime": 32.0402,
      "eval_samples_per_second": 21.348,
      "step": 93900
    },
    {
      "epoch": 22.43,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 93910
    },
    {
      "epoch": 22.43,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 93920
    },
    {
      "epoch": 22.43,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 93930
    },
    {
      "epoch": 22.44,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 93940
    },
    {
      "epoch": 22.44,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 93950
    },
    {
      "epoch": 22.44,
      "learning_rate": 1e-05,
      "loss": 1.4069,
      "step": 93960
    },
    {
      "epoch": 22.44,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 93970
    },
    {
      "epoch": 22.45,
      "learning_rate": 1e-05,
      "loss": 1.4083,
      "step": 93980
    },
    {
      "epoch": 22.45,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 93990
    },
    {
      "epoch": 22.45,
      "learning_rate": 1e-05,
      "loss": 1.3975,
      "step": 94000
    },
    {
      "epoch": 22.45,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4029632806777954,
      "eval_runtime": 32.08,
      "eval_samples_per_second": 21.322,
      "step": 94000
    },
    {
      "epoch": 22.45,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 94010
    },
    {
      "epoch": 22.46,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 94020
    },
    {
      "epoch": 22.46,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 94030
    },
    {
      "epoch": 22.46,
      "learning_rate": 1e-05,
      "loss": 1.3986,
      "step": 94040
    },
    {
      "epoch": 22.46,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 94050
    },
    {
      "epoch": 22.46,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 94060
    },
    {
      "epoch": 22.47,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 94070
    },
    {
      "epoch": 22.47,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 94080
    },
    {
      "epoch": 22.47,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 94090
    },
    {
      "epoch": 22.47,
      "learning_rate": 1e-05,
      "loss": 1.3969,
      "step": 94100
    },
    {
      "epoch": 22.47,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4030697345733643,
      "eval_runtime": 32.2905,
      "eval_samples_per_second": 21.183,
      "step": 94100
    },
    {
      "epoch": 22.48,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 94110
    },
    {
      "epoch": 22.48,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 94120
    },
    {
      "epoch": 22.48,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 94130
    },
    {
      "epoch": 22.48,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 94140
    },
    {
      "epoch": 22.49,
      "learning_rate": 1e-05,
      "loss": 1.3985,
      "step": 94150
    },
    {
      "epoch": 22.49,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 94160
    },
    {
      "epoch": 22.49,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 94170
    },
    {
      "epoch": 22.49,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 94180
    },
    {
      "epoch": 22.5,
      "learning_rate": 1e-05,
      "loss": 1.4054,
      "step": 94190
    },
    {
      "epoch": 22.5,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 94200
    },
    {
      "epoch": 22.5,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4023584127426147,
      "eval_runtime": 31.7717,
      "eval_samples_per_second": 21.529,
      "step": 94200
    },
    {
      "epoch": 22.5,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 94210
    },
    {
      "epoch": 22.5,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 94220
    },
    {
      "epoch": 22.51,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 94230
    },
    {
      "epoch": 22.51,
      "learning_rate": 1e-05,
      "loss": 1.4062,
      "step": 94240
    },
    {
      "epoch": 22.51,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 94250
    },
    {
      "epoch": 22.51,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 94260
    },
    {
      "epoch": 22.51,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 94270
    },
    {
      "epoch": 22.52,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 94280
    },
    {
      "epoch": 22.52,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 94290
    },
    {
      "epoch": 22.52,
      "learning_rate": 1e-05,
      "loss": 1.3987,
      "step": 94300
    },
    {
      "epoch": 22.52,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.403527855873108,
      "eval_runtime": 33.5418,
      "eval_samples_per_second": 20.392,
      "step": 94300
    },
    {
      "epoch": 22.52,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 94310
    },
    {
      "epoch": 22.53,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 94320
    },
    {
      "epoch": 22.53,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 94330
    },
    {
      "epoch": 22.53,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 94340
    },
    {
      "epoch": 22.53,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 94350
    },
    {
      "epoch": 22.54,
      "learning_rate": 1e-05,
      "loss": 1.398,
      "step": 94360
    },
    {
      "epoch": 22.54,
      "learning_rate": 1e-05,
      "loss": 1.398,
      "step": 94370
    },
    {
      "epoch": 22.54,
      "learning_rate": 1e-05,
      "loss": 1.3972,
      "step": 94380
    },
    {
      "epoch": 22.54,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 94390
    },
    {
      "epoch": 22.55,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 94400
    },
    {
      "epoch": 22.55,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.402438998222351,
      "eval_runtime": 32.1625,
      "eval_samples_per_second": 21.267,
      "step": 94400
    },
    {
      "epoch": 22.55,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 94410
    },
    {
      "epoch": 22.55,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 94420
    },
    {
      "epoch": 22.55,
      "learning_rate": 1e-05,
      "loss": 1.3984,
      "step": 94430
    },
    {
      "epoch": 22.56,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 94440
    },
    {
      "epoch": 22.56,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 94450
    },
    {
      "epoch": 22.56,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 94460
    },
    {
      "epoch": 22.56,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 94470
    },
    {
      "epoch": 22.57,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 94480
    },
    {
      "epoch": 22.57,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 94490
    },
    {
      "epoch": 22.57,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 94500
    },
    {
      "epoch": 22.57,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.402205228805542,
      "eval_runtime": 31.955,
      "eval_samples_per_second": 21.405,
      "step": 94500
    },
    {
      "epoch": 22.57,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 94510
    },
    {
      "epoch": 22.57,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 94520
    },
    {
      "epoch": 22.58,
      "learning_rate": 1e-05,
      "loss": 1.4067,
      "step": 94530
    },
    {
      "epoch": 22.58,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 94540
    },
    {
      "epoch": 22.58,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 94550
    },
    {
      "epoch": 22.58,
      "learning_rate": 1e-05,
      "loss": 1.3987,
      "step": 94560
    },
    {
      "epoch": 22.59,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 94570
    },
    {
      "epoch": 22.59,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 94580
    },
    {
      "epoch": 22.59,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 94590
    },
    {
      "epoch": 22.59,
      "learning_rate": 1e-05,
      "loss": 1.398,
      "step": 94600
    },
    {
      "epoch": 22.59,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4023343324661255,
      "eval_runtime": 32.031,
      "eval_samples_per_second": 21.354,
      "step": 94600
    },
    {
      "epoch": 22.6,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 94610
    },
    {
      "epoch": 22.6,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 94620
    },
    {
      "epoch": 22.6,
      "learning_rate": 1e-05,
      "loss": 1.398,
      "step": 94630
    },
    {
      "epoch": 22.6,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 94640
    },
    {
      "epoch": 22.61,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 94650
    },
    {
      "epoch": 22.61,
      "learning_rate": 1e-05,
      "loss": 1.4039,
      "step": 94660
    },
    {
      "epoch": 22.61,
      "learning_rate": 1e-05,
      "loss": 1.3977,
      "step": 94670
    },
    {
      "epoch": 22.61,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 94680
    },
    {
      "epoch": 22.62,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 94690
    },
    {
      "epoch": 22.62,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 94700
    },
    {
      "epoch": 22.62,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4025273323059082,
      "eval_runtime": 31.8332,
      "eval_samples_per_second": 21.487,
      "step": 94700
    },
    {
      "epoch": 22.62,
      "learning_rate": 1e-05,
      "loss": 1.4035,
      "step": 94710
    },
    {
      "epoch": 22.62,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 94720
    },
    {
      "epoch": 22.62,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 94730
    },
    {
      "epoch": 22.63,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 94740
    },
    {
      "epoch": 22.63,
      "learning_rate": 1e-05,
      "loss": 1.4036,
      "step": 94750
    },
    {
      "epoch": 22.63,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 94760
    },
    {
      "epoch": 22.63,
      "learning_rate": 1e-05,
      "loss": 1.3972,
      "step": 94770
    },
    {
      "epoch": 22.64,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 94780
    },
    {
      "epoch": 22.64,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 94790
    },
    {
      "epoch": 22.64,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 94800
    },
    {
      "epoch": 22.64,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.4016388654708862,
      "eval_runtime": 31.6677,
      "eval_samples_per_second": 21.599,
      "step": 94800
    },
    {
      "epoch": 22.64,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 94810
    },
    {
      "epoch": 22.65,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 94820
    },
    {
      "epoch": 22.65,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 94830
    },
    {
      "epoch": 22.65,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 94840
    },
    {
      "epoch": 22.65,
      "learning_rate": 1e-05,
      "loss": 1.3971,
      "step": 94850
    },
    {
      "epoch": 22.66,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 94860
    },
    {
      "epoch": 22.66,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 94870
    },
    {
      "epoch": 22.66,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 94880
    },
    {
      "epoch": 22.66,
      "learning_rate": 1e-05,
      "loss": 1.398,
      "step": 94890
    },
    {
      "epoch": 22.67,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 94900
    },
    {
      "epoch": 22.67,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4016250371932983,
      "eval_runtime": 32.078,
      "eval_samples_per_second": 21.323,
      "step": 94900
    },
    {
      "epoch": 22.67,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 94910
    },
    {
      "epoch": 22.67,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 94920
    },
    {
      "epoch": 22.67,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 94930
    },
    {
      "epoch": 22.67,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 94940
    },
    {
      "epoch": 22.68,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 94950
    },
    {
      "epoch": 22.68,
      "learning_rate": 1e-05,
      "loss": 1.3982,
      "step": 94960
    },
    {
      "epoch": 22.68,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 94970
    },
    {
      "epoch": 22.68,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 94980
    },
    {
      "epoch": 22.69,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 94990
    },
    {
      "epoch": 22.69,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 95000
    },
    {
      "epoch": 22.69,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4026910066604614,
      "eval_runtime": 32.0349,
      "eval_samples_per_second": 21.352,
      "step": 95000
    },
    {
      "epoch": 22.69,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 95010
    },
    {
      "epoch": 22.69,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 95020
    },
    {
      "epoch": 22.7,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 95030
    },
    {
      "epoch": 22.7,
      "learning_rate": 1e-05,
      "loss": 1.3987,
      "step": 95040
    },
    {
      "epoch": 22.7,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 95050
    },
    {
      "epoch": 22.7,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 95060
    },
    {
      "epoch": 22.71,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 95070
    },
    {
      "epoch": 22.71,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 95080
    },
    {
      "epoch": 22.71,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 95090
    },
    {
      "epoch": 22.71,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 95100
    },
    {
      "epoch": 22.71,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4020196199417114,
      "eval_runtime": 32.2755,
      "eval_samples_per_second": 21.193,
      "step": 95100
    },
    {
      "epoch": 22.72,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 95110
    },
    {
      "epoch": 22.72,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 95120
    },
    {
      "epoch": 22.72,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 95130
    },
    {
      "epoch": 22.72,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 95140
    },
    {
      "epoch": 22.73,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 95150
    },
    {
      "epoch": 22.73,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 95160
    },
    {
      "epoch": 22.73,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 95170
    },
    {
      "epoch": 22.73,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 95180
    },
    {
      "epoch": 22.73,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 95190
    },
    {
      "epoch": 22.74,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 95200
    },
    {
      "epoch": 22.74,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.403035283088684,
      "eval_runtime": 32.1979,
      "eval_samples_per_second": 21.244,
      "step": 95200
    },
    {
      "epoch": 22.74,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 95210
    },
    {
      "epoch": 22.74,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 95220
    },
    {
      "epoch": 22.74,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 95230
    },
    {
      "epoch": 22.75,
      "learning_rate": 1e-05,
      "loss": 1.3968,
      "step": 95240
    },
    {
      "epoch": 22.75,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 95250
    },
    {
      "epoch": 22.75,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 95260
    },
    {
      "epoch": 22.75,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 95270
    },
    {
      "epoch": 22.76,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 95280
    },
    {
      "epoch": 22.76,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 95290
    },
    {
      "epoch": 22.76,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 95300
    },
    {
      "epoch": 22.76,
      "eval_accuracy": 0.9766081871345029,
      "eval_loss": 1.402931571006775,
      "eval_runtime": 31.9628,
      "eval_samples_per_second": 21.4,
      "step": 95300
    },
    {
      "epoch": 22.76,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 95310
    },
    {
      "epoch": 22.77,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 95320
    },
    {
      "epoch": 22.77,
      "learning_rate": 1e-05,
      "loss": 1.3981,
      "step": 95330
    },
    {
      "epoch": 22.77,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 95340
    },
    {
      "epoch": 22.77,
      "learning_rate": 1e-05,
      "loss": 1.3987,
      "step": 95350
    },
    {
      "epoch": 22.78,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 95360
    },
    {
      "epoch": 22.78,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 95370
    },
    {
      "epoch": 22.78,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 95380
    },
    {
      "epoch": 22.78,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 95390
    },
    {
      "epoch": 22.78,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 95400
    },
    {
      "epoch": 22.78,
      "eval_accuracy": 0.9707602339181286,
      "eval_loss": 1.4023799896240234,
      "eval_runtime": 32.5105,
      "eval_samples_per_second": 21.039,
      "step": 95400
    },
    {
      "epoch": 22.79,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 95410
    },
    {
      "epoch": 22.79,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 95420
    },
    {
      "epoch": 22.79,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 95430
    },
    {
      "epoch": 22.79,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 95440
    },
    {
      "epoch": 22.8,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 95450
    },
    {
      "epoch": 22.8,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 95460
    },
    {
      "epoch": 22.8,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 95470
    },
    {
      "epoch": 22.8,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 95480
    },
    {
      "epoch": 22.81,
      "learning_rate": 1e-05,
      "loss": 1.3986,
      "step": 95490
    },
    {
      "epoch": 22.81,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 95500
    },
    {
      "epoch": 22.81,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4021605253219604,
      "eval_runtime": 32.1797,
      "eval_samples_per_second": 21.256,
      "step": 95500
    },
    {
      "epoch": 22.81,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 95510
    },
    {
      "epoch": 22.81,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 95520
    },
    {
      "epoch": 22.82,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 95530
    },
    {
      "epoch": 22.82,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 95540
    },
    {
      "epoch": 22.82,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 95550
    },
    {
      "epoch": 22.82,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 95560
    },
    {
      "epoch": 22.83,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 95570
    },
    {
      "epoch": 22.83,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 95580
    },
    {
      "epoch": 22.83,
      "learning_rate": 1e-05,
      "loss": 1.3987,
      "step": 95590
    },
    {
      "epoch": 22.83,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 95600
    },
    {
      "epoch": 22.83,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4017155170440674,
      "eval_runtime": 31.9828,
      "eval_samples_per_second": 21.387,
      "step": 95600
    },
    {
      "epoch": 22.83,
      "learning_rate": 1e-05,
      "loss": 1.3968,
      "step": 95610
    },
    {
      "epoch": 22.84,
      "learning_rate": 1e-05,
      "loss": 1.4072,
      "step": 95620
    },
    {
      "epoch": 22.84,
      "learning_rate": 1e-05,
      "loss": 1.397,
      "step": 95630
    },
    {
      "epoch": 22.84,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 95640
    },
    {
      "epoch": 22.84,
      "learning_rate": 1e-05,
      "loss": 1.398,
      "step": 95650
    },
    {
      "epoch": 22.85,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 95660
    },
    {
      "epoch": 22.85,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 95670
    },
    {
      "epoch": 22.85,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 95680
    },
    {
      "epoch": 22.85,
      "learning_rate": 1e-05,
      "loss": 1.4034,
      "step": 95690
    },
    {
      "epoch": 22.86,
      "learning_rate": 1e-05,
      "loss": 1.4045,
      "step": 95700
    },
    {
      "epoch": 22.86,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4021021127700806,
      "eval_runtime": 32.2911,
      "eval_samples_per_second": 21.182,
      "step": 95700
    },
    {
      "epoch": 22.86,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 95710
    },
    {
      "epoch": 22.86,
      "learning_rate": 1e-05,
      "loss": 1.3985,
      "step": 95720
    },
    {
      "epoch": 22.86,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 95730
    },
    {
      "epoch": 22.87,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 95740
    },
    {
      "epoch": 22.87,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 95750
    },
    {
      "epoch": 22.87,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 95760
    },
    {
      "epoch": 22.87,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 95770
    },
    {
      "epoch": 22.88,
      "learning_rate": 1e-05,
      "loss": 1.4037,
      "step": 95780
    },
    {
      "epoch": 22.88,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 95790
    },
    {
      "epoch": 22.88,
      "learning_rate": 1e-05,
      "loss": 1.3969,
      "step": 95800
    },
    {
      "epoch": 22.88,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4023940563201904,
      "eval_runtime": 32.0449,
      "eval_samples_per_second": 21.345,
      "step": 95800
    },
    {
      "epoch": 22.88,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 95810
    },
    {
      "epoch": 22.89,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 95820
    },
    {
      "epoch": 22.89,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 95830
    },
    {
      "epoch": 22.89,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 95840
    },
    {
      "epoch": 22.89,
      "learning_rate": 1e-05,
      "loss": 1.3961,
      "step": 95850
    },
    {
      "epoch": 22.89,
      "learning_rate": 1e-05,
      "loss": 1.3978,
      "step": 95860
    },
    {
      "epoch": 22.9,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 95870
    },
    {
      "epoch": 22.9,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 95880
    },
    {
      "epoch": 22.9,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 95890
    },
    {
      "epoch": 22.9,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 95900
    },
    {
      "epoch": 22.9,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4016518592834473,
      "eval_runtime": 31.9399,
      "eval_samples_per_second": 21.415,
      "step": 95900
    },
    {
      "epoch": 22.91,
      "learning_rate": 1e-05,
      "loss": 1.4041,
      "step": 95910
    },
    {
      "epoch": 22.91,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 95920
    },
    {
      "epoch": 22.91,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 95930
    },
    {
      "epoch": 22.91,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 95940
    },
    {
      "epoch": 22.92,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 95950
    },
    {
      "epoch": 22.92,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 95960
    },
    {
      "epoch": 22.92,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 95970
    },
    {
      "epoch": 22.92,
      "learning_rate": 1e-05,
      "loss": 1.4055,
      "step": 95980
    },
    {
      "epoch": 22.93,
      "learning_rate": 1e-05,
      "loss": 1.3972,
      "step": 95990
    },
    {
      "epoch": 22.93,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 96000
    },
    {
      "epoch": 22.93,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4024536609649658,
      "eval_runtime": 32.0906,
      "eval_samples_per_second": 21.315,
      "step": 96000
    },
    {
      "epoch": 22.93,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 96010
    },
    {
      "epoch": 22.93,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 96020
    },
    {
      "epoch": 22.94,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 96030
    },
    {
      "epoch": 22.94,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 96040
    },
    {
      "epoch": 22.94,
      "learning_rate": 1e-05,
      "loss": 1.3982,
      "step": 96050
    },
    {
      "epoch": 22.94,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 96060
    },
    {
      "epoch": 22.94,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 96070
    },
    {
      "epoch": 22.95,
      "learning_rate": 1e-05,
      "loss": 1.3973,
      "step": 96080
    },
    {
      "epoch": 22.95,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 96090
    },
    {
      "epoch": 22.95,
      "learning_rate": 1e-05,
      "loss": 1.3978,
      "step": 96100
    },
    {
      "epoch": 22.95,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4014848470687866,
      "eval_runtime": 32.5159,
      "eval_samples_per_second": 21.036,
      "step": 96100
    },
    {
      "epoch": 22.95,
      "learning_rate": 1e-05,
      "loss": 1.4026,
      "step": 96110
    },
    {
      "epoch": 22.96,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 96120
    },
    {
      "epoch": 22.96,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 96130
    },
    {
      "epoch": 22.96,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 96140
    },
    {
      "epoch": 22.96,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 96150
    },
    {
      "epoch": 22.97,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 96160
    },
    {
      "epoch": 22.97,
      "learning_rate": 1e-05,
      "loss": 1.4044,
      "step": 96170
    },
    {
      "epoch": 22.97,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 96180
    },
    {
      "epoch": 22.97,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 96190
    },
    {
      "epoch": 22.98,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 96200
    },
    {
      "epoch": 22.98,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4017298221588135,
      "eval_runtime": 32.0839,
      "eval_samples_per_second": 21.319,
      "step": 96200
    },
    {
      "epoch": 22.98,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 96210
    },
    {
      "epoch": 22.98,
      "learning_rate": 1e-05,
      "loss": 1.3975,
      "step": 96220
    },
    {
      "epoch": 22.98,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 96230
    },
    {
      "epoch": 22.99,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 96240
    },
    {
      "epoch": 22.99,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 96250
    },
    {
      "epoch": 22.99,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 96260
    },
    {
      "epoch": 22.99,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 96270
    },
    {
      "epoch": 22.99,
      "learning_rate": 1e-05,
      "loss": 1.3982,
      "step": 96280
    },
    {
      "epoch": 23.0,
      "learning_rate": 1e-05,
      "loss": 1.3982,
      "step": 96290
    },
    {
      "epoch": 23.0,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 96300
    },
    {
      "epoch": 23.0,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.402190089225769,
      "eval_runtime": 32.2038,
      "eval_samples_per_second": 21.24,
      "step": 96300
    },
    {
      "epoch": 23.0,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 96310
    },
    {
      "epoch": 23.0,
      "learning_rate": 1e-05,
      "loss": 1.3981,
      "step": 96320
    },
    {
      "epoch": 23.01,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 96330
    },
    {
      "epoch": 23.01,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 96340
    },
    {
      "epoch": 23.01,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 96350
    },
    {
      "epoch": 23.01,
      "learning_rate": 1e-05,
      "loss": 1.3982,
      "step": 96360
    },
    {
      "epoch": 23.02,
      "learning_rate": 1e-05,
      "loss": 1.3975,
      "step": 96370
    },
    {
      "epoch": 23.02,
      "learning_rate": 1e-05,
      "loss": 1.398,
      "step": 96380
    },
    {
      "epoch": 23.02,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 96390
    },
    {
      "epoch": 23.02,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 96400
    },
    {
      "epoch": 23.02,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.402673363685608,
      "eval_runtime": 31.844,
      "eval_samples_per_second": 21.48,
      "step": 96400
    },
    {
      "epoch": 23.03,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 96410
    },
    {
      "epoch": 23.03,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 96420
    },
    {
      "epoch": 23.03,
      "learning_rate": 1e-05,
      "loss": 1.4046,
      "step": 96430
    },
    {
      "epoch": 23.03,
      "learning_rate": 1e-05,
      "loss": 1.3967,
      "step": 96440
    },
    {
      "epoch": 23.04,
      "learning_rate": 1e-05,
      "loss": 1.3987,
      "step": 96450
    },
    {
      "epoch": 23.04,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 96460
    },
    {
      "epoch": 23.04,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 96470
    },
    {
      "epoch": 23.04,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 96480
    },
    {
      "epoch": 23.05,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 96490
    },
    {
      "epoch": 23.05,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 96500
    },
    {
      "epoch": 23.05,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4025592803955078,
      "eval_runtime": 31.7936,
      "eval_samples_per_second": 21.514,
      "step": 96500
    },
    {
      "epoch": 23.05,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 96510
    },
    {
      "epoch": 23.05,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 96520
    },
    {
      "epoch": 23.05,
      "learning_rate": 1e-05,
      "loss": 1.3987,
      "step": 96530
    },
    {
      "epoch": 23.06,
      "learning_rate": 1e-05,
      "loss": 1.4042,
      "step": 96540
    },
    {
      "epoch": 23.06,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 96550
    },
    {
      "epoch": 23.06,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 96560
    },
    {
      "epoch": 23.06,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 96570
    },
    {
      "epoch": 23.07,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 96580
    },
    {
      "epoch": 23.07,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 96590
    },
    {
      "epoch": 23.07,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 96600
    },
    {
      "epoch": 23.07,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.401560664176941,
      "eval_runtime": 31.9678,
      "eval_samples_per_second": 21.397,
      "step": 96600
    },
    {
      "epoch": 23.07,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 96610
    },
    {
      "epoch": 23.08,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 96620
    },
    {
      "epoch": 23.08,
      "learning_rate": 1e-05,
      "loss": 1.3977,
      "step": 96630
    },
    {
      "epoch": 23.08,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 96640
    },
    {
      "epoch": 23.08,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 96650
    },
    {
      "epoch": 23.09,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 96660
    },
    {
      "epoch": 23.09,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 96670
    },
    {
      "epoch": 23.09,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 96680
    },
    {
      "epoch": 23.09,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 96690
    },
    {
      "epoch": 23.1,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 96700
    },
    {
      "epoch": 23.1,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.402352213859558,
      "eval_runtime": 31.8852,
      "eval_samples_per_second": 21.452,
      "step": 96700
    },
    {
      "epoch": 23.1,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 96710
    },
    {
      "epoch": 23.1,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 96720
    },
    {
      "epoch": 23.1,
      "learning_rate": 1e-05,
      "loss": 1.4047,
      "step": 96730
    },
    {
      "epoch": 23.1,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 96740
    },
    {
      "epoch": 23.11,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 96750
    },
    {
      "epoch": 23.11,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 96760
    },
    {
      "epoch": 23.11,
      "learning_rate": 1e-05,
      "loss": 1.3985,
      "step": 96770
    },
    {
      "epoch": 23.11,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 96780
    },
    {
      "epoch": 23.12,
      "learning_rate": 1e-05,
      "loss": 1.3986,
      "step": 96790
    },
    {
      "epoch": 23.12,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 96800
    },
    {
      "epoch": 23.12,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4018499851226807,
      "eval_runtime": 31.9173,
      "eval_samples_per_second": 21.43,
      "step": 96800
    },
    {
      "epoch": 23.12,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 96810
    },
    {
      "epoch": 23.12,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 96820
    },
    {
      "epoch": 23.13,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 96830
    },
    {
      "epoch": 23.13,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 96840
    },
    {
      "epoch": 23.13,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 96850
    },
    {
      "epoch": 23.13,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 96860
    },
    {
      "epoch": 23.14,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 96870
    },
    {
      "epoch": 23.14,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 96880
    },
    {
      "epoch": 23.14,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 96890
    },
    {
      "epoch": 23.14,
      "learning_rate": 1e-05,
      "loss": 1.3967,
      "step": 96900
    },
    {
      "epoch": 23.14,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4011881351470947,
      "eval_runtime": 31.9787,
      "eval_samples_per_second": 21.389,
      "step": 96900
    },
    {
      "epoch": 23.15,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 96910
    },
    {
      "epoch": 23.15,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 96920
    },
    {
      "epoch": 23.15,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 96930
    },
    {
      "epoch": 23.15,
      "learning_rate": 1e-05,
      "loss": 1.3987,
      "step": 96940
    },
    {
      "epoch": 23.16,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 96950
    },
    {
      "epoch": 23.16,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 96960
    },
    {
      "epoch": 23.16,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 96970
    },
    {
      "epoch": 23.16,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 96980
    },
    {
      "epoch": 23.16,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 96990
    },
    {
      "epoch": 23.17,
      "learning_rate": 1e-05,
      "loss": 1.396,
      "step": 97000
    },
    {
      "epoch": 23.17,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4024661779403687,
      "eval_runtime": 32.24,
      "eval_samples_per_second": 21.216,
      "step": 97000
    },
    {
      "epoch": 23.17,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 97010
    },
    {
      "epoch": 23.17,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 97020
    },
    {
      "epoch": 23.17,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 97030
    },
    {
      "epoch": 23.18,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 97040
    },
    {
      "epoch": 23.18,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 97050
    },
    {
      "epoch": 23.18,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 97060
    },
    {
      "epoch": 23.18,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 97070
    },
    {
      "epoch": 23.19,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 97080
    },
    {
      "epoch": 23.19,
      "learning_rate": 1e-05,
      "loss": 1.398,
      "step": 97090
    },
    {
      "epoch": 23.19,
      "learning_rate": 1e-05,
      "loss": 1.3957,
      "step": 97100
    },
    {
      "epoch": 23.19,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4020050764083862,
      "eval_runtime": 32.3318,
      "eval_samples_per_second": 21.156,
      "step": 97100
    },
    {
      "epoch": 23.19,
      "learning_rate": 1e-05,
      "loss": 1.3987,
      "step": 97110
    },
    {
      "epoch": 23.2,
      "learning_rate": 1e-05,
      "loss": 1.3978,
      "step": 97120
    },
    {
      "epoch": 23.2,
      "learning_rate": 1e-05,
      "loss": 1.3981,
      "step": 97130
    },
    {
      "epoch": 23.2,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 97140
    },
    {
      "epoch": 23.2,
      "learning_rate": 1e-05,
      "loss": 1.4029,
      "step": 97150
    },
    {
      "epoch": 23.21,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 97160
    },
    {
      "epoch": 23.21,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 97170
    },
    {
      "epoch": 23.21,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 97180
    },
    {
      "epoch": 23.21,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 97190
    },
    {
      "epoch": 23.21,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 97200
    },
    {
      "epoch": 23.21,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.402198314666748,
      "eval_runtime": 32.0678,
      "eval_samples_per_second": 21.33,
      "step": 97200
    },
    {
      "epoch": 23.22,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 97210
    },
    {
      "epoch": 23.22,
      "learning_rate": 1e-05,
      "loss": 1.3981,
      "step": 97220
    },
    {
      "epoch": 23.22,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 97230
    },
    {
      "epoch": 23.22,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 97240
    },
    {
      "epoch": 23.23,
      "learning_rate": 1e-05,
      "loss": 1.3981,
      "step": 97250
    },
    {
      "epoch": 23.23,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 97260
    },
    {
      "epoch": 23.23,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 97270
    },
    {
      "epoch": 23.23,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 97280
    },
    {
      "epoch": 23.24,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 97290
    },
    {
      "epoch": 23.24,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 97300
    },
    {
      "epoch": 23.24,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.40275239944458,
      "eval_runtime": 32.0308,
      "eval_samples_per_second": 21.354,
      "step": 97300
    },
    {
      "epoch": 23.24,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 97310
    },
    {
      "epoch": 23.24,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 97320
    },
    {
      "epoch": 23.25,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 97330
    },
    {
      "epoch": 23.25,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 97340
    },
    {
      "epoch": 23.25,
      "learning_rate": 1e-05,
      "loss": 1.4038,
      "step": 97350
    },
    {
      "epoch": 23.25,
      "learning_rate": 1e-05,
      "loss": 1.3978,
      "step": 97360
    },
    {
      "epoch": 23.26,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 97370
    },
    {
      "epoch": 23.26,
      "learning_rate": 1e-05,
      "loss": 1.3969,
      "step": 97380
    },
    {
      "epoch": 23.26,
      "learning_rate": 1e-05,
      "loss": 1.3983,
      "step": 97390
    },
    {
      "epoch": 23.26,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 97400
    },
    {
      "epoch": 23.26,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4023730754852295,
      "eval_runtime": 31.9377,
      "eval_samples_per_second": 21.417,
      "step": 97400
    },
    {
      "epoch": 23.26,
      "learning_rate": 1e-05,
      "loss": 1.4048,
      "step": 97410
    },
    {
      "epoch": 23.27,
      "learning_rate": 1e-05,
      "loss": 1.4013,
      "step": 97420
    },
    {
      "epoch": 23.27,
      "learning_rate": 1e-05,
      "loss": 1.409,
      "step": 97430
    },
    {
      "epoch": 23.27,
      "learning_rate": 1e-05,
      "loss": 1.3975,
      "step": 97440
    },
    {
      "epoch": 23.27,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 97450
    },
    {
      "epoch": 23.28,
      "learning_rate": 1e-05,
      "loss": 1.3984,
      "step": 97460
    },
    {
      "epoch": 23.28,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 97470
    },
    {
      "epoch": 23.28,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 97480
    },
    {
      "epoch": 23.28,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 97490
    },
    {
      "epoch": 23.29,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 97500
    },
    {
      "epoch": 23.29,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4014173746109009,
      "eval_runtime": 31.8212,
      "eval_samples_per_second": 21.495,
      "step": 97500
    },
    {
      "epoch": 23.29,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 97510
    },
    {
      "epoch": 23.29,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 97520
    },
    {
      "epoch": 23.29,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 97530
    },
    {
      "epoch": 23.3,
      "learning_rate": 1e-05,
      "loss": 1.4049,
      "step": 97540
    },
    {
      "epoch": 23.3,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 97550
    },
    {
      "epoch": 23.3,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 97560
    },
    {
      "epoch": 23.3,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 97570
    },
    {
      "epoch": 23.31,
      "learning_rate": 1e-05,
      "loss": 1.3985,
      "step": 97580
    },
    {
      "epoch": 23.31,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 97590
    },
    {
      "epoch": 23.31,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 97600
    },
    {
      "epoch": 23.31,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4022215604782104,
      "eval_runtime": 32.1579,
      "eval_samples_per_second": 21.27,
      "step": 97600
    },
    {
      "epoch": 23.31,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 97610
    },
    {
      "epoch": 23.32,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 97620
    },
    {
      "epoch": 23.32,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 97630
    },
    {
      "epoch": 23.32,
      "learning_rate": 1e-05,
      "loss": 1.4028,
      "step": 97640
    },
    {
      "epoch": 23.32,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 97650
    },
    {
      "epoch": 23.32,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 97660
    },
    {
      "epoch": 23.33,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 97670
    },
    {
      "epoch": 23.33,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 97680
    },
    {
      "epoch": 23.33,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 97690
    },
    {
      "epoch": 23.33,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 97700
    },
    {
      "epoch": 23.33,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4012483358383179,
      "eval_runtime": 32.9624,
      "eval_samples_per_second": 20.751,
      "step": 97700
    },
    {
      "epoch": 23.34,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 97710
    },
    {
      "epoch": 23.34,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 97720
    },
    {
      "epoch": 23.34,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 97730
    },
    {
      "epoch": 23.34,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 97740
    },
    {
      "epoch": 23.35,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 97750
    },
    {
      "epoch": 23.35,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 97760
    },
    {
      "epoch": 23.35,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 97770
    },
    {
      "epoch": 23.35,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 97780
    },
    {
      "epoch": 23.36,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 97790
    },
    {
      "epoch": 23.36,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 97800
    },
    {
      "epoch": 23.36,
      "eval_accuracy": 0.9853801169590644,
      "eval_loss": 1.4020562171936035,
      "eval_runtime": 32.0577,
      "eval_samples_per_second": 21.337,
      "step": 97800
    },
    {
      "epoch": 23.36,
      "learning_rate": 1e-05,
      "loss": 1.3977,
      "step": 97810
    },
    {
      "epoch": 23.36,
      "learning_rate": 1e-05,
      "loss": 1.398,
      "step": 97820
    },
    {
      "epoch": 23.37,
      "learning_rate": 1e-05,
      "loss": 1.396,
      "step": 97830
    },
    {
      "epoch": 23.37,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 97840
    },
    {
      "epoch": 23.37,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 97850
    },
    {
      "epoch": 23.37,
      "learning_rate": 1e-05,
      "loss": 1.3982,
      "step": 97860
    },
    {
      "epoch": 23.37,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 97870
    },
    {
      "epoch": 23.38,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 97880
    },
    {
      "epoch": 23.38,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 97890
    },
    {
      "epoch": 23.38,
      "learning_rate": 1e-05,
      "loss": 1.3973,
      "step": 97900
    },
    {
      "epoch": 23.38,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.4015063047409058,
      "eval_runtime": 31.9873,
      "eval_samples_per_second": 21.383,
      "step": 97900
    },
    {
      "epoch": 23.38,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 97910
    },
    {
      "epoch": 23.39,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 97920
    },
    {
      "epoch": 23.39,
      "learning_rate": 1e-05,
      "loss": 1.398,
      "step": 97930
    },
    {
      "epoch": 23.39,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 97940
    },
    {
      "epoch": 23.39,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 97950
    },
    {
      "epoch": 23.4,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 97960
    },
    {
      "epoch": 23.4,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 97970
    },
    {
      "epoch": 23.4,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 97980
    },
    {
      "epoch": 23.4,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 97990
    },
    {
      "epoch": 23.41,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 98000
    },
    {
      "epoch": 23.41,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.401679515838623,
      "eval_runtime": 32.0363,
      "eval_samples_per_second": 21.351,
      "step": 98000
    },
    {
      "epoch": 23.41,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 98010
    },
    {
      "epoch": 23.41,
      "learning_rate": 1e-05,
      "loss": 1.3983,
      "step": 98020
    },
    {
      "epoch": 23.41,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 98030
    },
    {
      "epoch": 23.42,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 98040
    },
    {
      "epoch": 23.42,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 98050
    },
    {
      "epoch": 23.42,
      "learning_rate": 1e-05,
      "loss": 1.3982,
      "step": 98060
    },
    {
      "epoch": 23.42,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 98070
    },
    {
      "epoch": 23.42,
      "learning_rate": 1e-05,
      "loss": 1.3973,
      "step": 98080
    },
    {
      "epoch": 23.43,
      "learning_rate": 1e-05,
      "loss": 1.397,
      "step": 98090
    },
    {
      "epoch": 23.43,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 98100
    },
    {
      "epoch": 23.43,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.40183687210083,
      "eval_runtime": 32.1584,
      "eval_samples_per_second": 21.27,
      "step": 98100
    },
    {
      "epoch": 23.43,
      "learning_rate": 1e-05,
      "loss": 1.3983,
      "step": 98110
    },
    {
      "epoch": 23.43,
      "learning_rate": 1e-05,
      "loss": 1.4033,
      "step": 98120
    },
    {
      "epoch": 23.44,
      "learning_rate": 1e-05,
      "loss": 1.3976,
      "step": 98130
    },
    {
      "epoch": 23.44,
      "learning_rate": 1e-05,
      "loss": 1.4027,
      "step": 98140
    },
    {
      "epoch": 23.44,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 98150
    },
    {
      "epoch": 23.44,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 98160
    },
    {
      "epoch": 23.45,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 98170
    },
    {
      "epoch": 23.45,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 98180
    },
    {
      "epoch": 23.45,
      "learning_rate": 1e-05,
      "loss": 1.3984,
      "step": 98190
    },
    {
      "epoch": 23.45,
      "learning_rate": 1e-05,
      "loss": 1.3976,
      "step": 98200
    },
    {
      "epoch": 23.45,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4020135402679443,
      "eval_runtime": 31.9708,
      "eval_samples_per_second": 21.395,
      "step": 98200
    },
    {
      "epoch": 23.46,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 98210
    },
    {
      "epoch": 23.46,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 98220
    },
    {
      "epoch": 23.46,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 98230
    },
    {
      "epoch": 23.46,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 98240
    },
    {
      "epoch": 23.47,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 98250
    },
    {
      "epoch": 23.47,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 98260
    },
    {
      "epoch": 23.47,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 98270
    },
    {
      "epoch": 23.47,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 98280
    },
    {
      "epoch": 23.48,
      "learning_rate": 1e-05,
      "loss": 1.3971,
      "step": 98290
    },
    {
      "epoch": 23.48,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 98300
    },
    {
      "epoch": 23.48,
      "eval_accuracy": 0.9839181286549707,
      "eval_loss": 1.400738000869751,
      "eval_runtime": 32.0836,
      "eval_samples_per_second": 21.319,
      "step": 98300
    },
    {
      "epoch": 23.48,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 98310
    },
    {
      "epoch": 23.48,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 98320
    },
    {
      "epoch": 23.48,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 98330
    },
    {
      "epoch": 23.49,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 98340
    },
    {
      "epoch": 23.49,
      "learning_rate": 1e-05,
      "loss": 1.3982,
      "step": 98350
    },
    {
      "epoch": 23.49,
      "learning_rate": 1e-05,
      "loss": 1.3953,
      "step": 98360
    },
    {
      "epoch": 23.49,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 98370
    },
    {
      "epoch": 23.5,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 98380
    },
    {
      "epoch": 23.5,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 98390
    },
    {
      "epoch": 23.5,
      "learning_rate": 1e-05,
      "loss": 1.3984,
      "step": 98400
    },
    {
      "epoch": 23.5,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4017508029937744,
      "eval_runtime": 31.6114,
      "eval_samples_per_second": 21.638,
      "step": 98400
    },
    {
      "epoch": 23.5,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 98410
    },
    {
      "epoch": 23.51,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 98420
    },
    {
      "epoch": 23.51,
      "learning_rate": 1e-05,
      "loss": 1.3972,
      "step": 98430
    },
    {
      "epoch": 23.51,
      "learning_rate": 1e-05,
      "loss": 1.3983,
      "step": 98440
    },
    {
      "epoch": 23.51,
      "learning_rate": 1e-05,
      "loss": 1.3986,
      "step": 98450
    },
    {
      "epoch": 23.52,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 98460
    },
    {
      "epoch": 23.52,
      "learning_rate": 1e-05,
      "loss": 1.4053,
      "step": 98470
    },
    {
      "epoch": 23.52,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 98480
    },
    {
      "epoch": 23.52,
      "learning_rate": 1e-05,
      "loss": 1.3972,
      "step": 98490
    },
    {
      "epoch": 23.53,
      "learning_rate": 1e-05,
      "loss": 1.3963,
      "step": 98500
    },
    {
      "epoch": 23.53,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4011480808258057,
      "eval_runtime": 32.1645,
      "eval_samples_per_second": 21.266,
      "step": 98500
    },
    {
      "epoch": 23.53,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 98510
    },
    {
      "epoch": 23.53,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 98520
    },
    {
      "epoch": 23.53,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 98530
    },
    {
      "epoch": 23.53,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 98540
    },
    {
      "epoch": 23.54,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 98550
    },
    {
      "epoch": 23.54,
      "learning_rate": 1e-05,
      "loss": 1.3984,
      "step": 98560
    },
    {
      "epoch": 23.54,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 98570
    },
    {
      "epoch": 23.54,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 98580
    },
    {
      "epoch": 23.55,
      "learning_rate": 1e-05,
      "loss": 1.3986,
      "step": 98590
    },
    {
      "epoch": 23.55,
      "learning_rate": 1e-05,
      "loss": 1.3987,
      "step": 98600
    },
    {
      "epoch": 23.55,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4012123346328735,
      "eval_runtime": 32.0556,
      "eval_samples_per_second": 21.338,
      "step": 98600
    },
    {
      "epoch": 23.55,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 98610
    },
    {
      "epoch": 23.55,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 98620
    },
    {
      "epoch": 23.56,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 98630
    },
    {
      "epoch": 23.56,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 98640
    },
    {
      "epoch": 23.56,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 98650
    },
    {
      "epoch": 23.56,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 98660
    },
    {
      "epoch": 23.57,
      "learning_rate": 1e-05,
      "loss": 1.4104,
      "step": 98670
    },
    {
      "epoch": 23.57,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 98680
    },
    {
      "epoch": 23.57,
      "learning_rate": 1e-05,
      "loss": 1.3989,
      "step": 98690
    },
    {
      "epoch": 23.57,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 98700
    },
    {
      "epoch": 23.57,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4015426635742188,
      "eval_runtime": 32.0558,
      "eval_samples_per_second": 21.338,
      "step": 98700
    },
    {
      "epoch": 23.58,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 98710
    },
    {
      "epoch": 23.58,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 98720
    },
    {
      "epoch": 23.58,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 98730
    },
    {
      "epoch": 23.58,
      "learning_rate": 1e-05,
      "loss": 1.3978,
      "step": 98740
    },
    {
      "epoch": 23.58,
      "learning_rate": 1e-05,
      "loss": 1.3985,
      "step": 98750
    },
    {
      "epoch": 23.59,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 98760
    },
    {
      "epoch": 23.59,
      "learning_rate": 1e-05,
      "loss": 1.3981,
      "step": 98770
    },
    {
      "epoch": 23.59,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 98780
    },
    {
      "epoch": 23.59,
      "learning_rate": 1e-05,
      "loss": 1.3984,
      "step": 98790
    },
    {
      "epoch": 23.6,
      "learning_rate": 1e-05,
      "loss": 1.4017,
      "step": 98800
    },
    {
      "epoch": 23.6,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4011799097061157,
      "eval_runtime": 32.3863,
      "eval_samples_per_second": 21.12,
      "step": 98800
    },
    {
      "epoch": 23.6,
      "learning_rate": 1e-05,
      "loss": 1.4022,
      "step": 98810
    },
    {
      "epoch": 23.6,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 98820
    },
    {
      "epoch": 23.6,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 98830
    },
    {
      "epoch": 23.61,
      "learning_rate": 1e-05,
      "loss": 1.3966,
      "step": 98840
    },
    {
      "epoch": 23.61,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 98850
    },
    {
      "epoch": 23.61,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 98860
    },
    {
      "epoch": 23.61,
      "learning_rate": 1e-05,
      "loss": 1.403,
      "step": 98870
    },
    {
      "epoch": 23.62,
      "learning_rate": 1e-05,
      "loss": 1.3976,
      "step": 98880
    },
    {
      "epoch": 23.62,
      "learning_rate": 1e-05,
      "loss": 1.4025,
      "step": 98890
    },
    {
      "epoch": 23.62,
      "learning_rate": 1e-05,
      "loss": 1.3977,
      "step": 98900
    },
    {
      "epoch": 23.62,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4011541604995728,
      "eval_runtime": 32.1076,
      "eval_samples_per_second": 21.303,
      "step": 98900
    },
    {
      "epoch": 23.62,
      "learning_rate": 1e-05,
      "loss": 1.3984,
      "step": 98910
    },
    {
      "epoch": 23.63,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 98920
    },
    {
      "epoch": 23.63,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 98930
    },
    {
      "epoch": 23.63,
      "learning_rate": 1e-05,
      "loss": 1.3975,
      "step": 98940
    },
    {
      "epoch": 23.63,
      "learning_rate": 1e-05,
      "loss": 1.3973,
      "step": 98950
    },
    {
      "epoch": 23.64,
      "learning_rate": 1e-05,
      "loss": 1.3984,
      "step": 98960
    },
    {
      "epoch": 23.64,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 98970
    },
    {
      "epoch": 23.64,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 98980
    },
    {
      "epoch": 23.64,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 98990
    },
    {
      "epoch": 23.64,
      "learning_rate": 1e-05,
      "loss": 1.4032,
      "step": 99000
    },
    {
      "epoch": 23.64,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.40212082862854,
      "eval_runtime": 32.3619,
      "eval_samples_per_second": 21.136,
      "step": 99000
    },
    {
      "epoch": 23.65,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 99010
    },
    {
      "epoch": 23.65,
      "learning_rate": 1e-05,
      "loss": 1.4066,
      "step": 99020
    },
    {
      "epoch": 23.65,
      "learning_rate": 1e-05,
      "loss": 1.4009,
      "step": 99030
    },
    {
      "epoch": 23.65,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 99040
    },
    {
      "epoch": 23.66,
      "learning_rate": 1e-05,
      "loss": 1.3965,
      "step": 99050
    },
    {
      "epoch": 23.66,
      "learning_rate": 1e-05,
      "loss": 1.398,
      "step": 99060
    },
    {
      "epoch": 23.66,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 99070
    },
    {
      "epoch": 23.66,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 99080
    },
    {
      "epoch": 23.67,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 99090
    },
    {
      "epoch": 23.67,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 99100
    },
    {
      "epoch": 23.67,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4017229080200195,
      "eval_runtime": 31.9802,
      "eval_samples_per_second": 21.388,
      "step": 99100
    },
    {
      "epoch": 23.67,
      "learning_rate": 1e-05,
      "loss": 1.4007,
      "step": 99110
    },
    {
      "epoch": 23.67,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 99120
    },
    {
      "epoch": 23.68,
      "learning_rate": 1e-05,
      "loss": 1.3975,
      "step": 99130
    },
    {
      "epoch": 23.68,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 99140
    },
    {
      "epoch": 23.68,
      "learning_rate": 1e-05,
      "loss": 1.3973,
      "step": 99150
    },
    {
      "epoch": 23.68,
      "learning_rate": 1e-05,
      "loss": 1.4012,
      "step": 99160
    },
    {
      "epoch": 23.69,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 99170
    },
    {
      "epoch": 23.69,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 99180
    },
    {
      "epoch": 23.69,
      "learning_rate": 1e-05,
      "loss": 1.3958,
      "step": 99190
    },
    {
      "epoch": 23.69,
      "learning_rate": 1e-05,
      "loss": 1.3953,
      "step": 99200
    },
    {
      "epoch": 23.69,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4024063348770142,
      "eval_runtime": 31.9432,
      "eval_samples_per_second": 21.413,
      "step": 99200
    },
    {
      "epoch": 23.69,
      "learning_rate": 1e-05,
      "loss": 1.3997,
      "step": 99210
    },
    {
      "epoch": 23.7,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 99220
    },
    {
      "epoch": 23.7,
      "learning_rate": 1e-05,
      "loss": 1.3974,
      "step": 99230
    },
    {
      "epoch": 23.7,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 99240
    },
    {
      "epoch": 23.7,
      "learning_rate": 1e-05,
      "loss": 1.404,
      "step": 99250
    },
    {
      "epoch": 23.71,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 99260
    },
    {
      "epoch": 23.71,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 99270
    },
    {
      "epoch": 23.71,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 99280
    },
    {
      "epoch": 23.71,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 99290
    },
    {
      "epoch": 23.72,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 99300
    },
    {
      "epoch": 23.72,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4012871980667114,
      "eval_runtime": 32.0534,
      "eval_samples_per_second": 21.339,
      "step": 99300
    },
    {
      "epoch": 23.72,
      "learning_rate": 1e-05,
      "loss": 1.4024,
      "step": 99310
    },
    {
      "epoch": 23.72,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 99320
    },
    {
      "epoch": 23.72,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 99330
    },
    {
      "epoch": 23.73,
      "learning_rate": 1e-05,
      "loss": 1.3975,
      "step": 99340
    },
    {
      "epoch": 23.73,
      "learning_rate": 1e-05,
      "loss": 1.3974,
      "step": 99350
    },
    {
      "epoch": 23.73,
      "learning_rate": 1e-05,
      "loss": 1.3996,
      "step": 99360
    },
    {
      "epoch": 23.73,
      "learning_rate": 1e-05,
      "loss": 1.3977,
      "step": 99370
    },
    {
      "epoch": 23.74,
      "learning_rate": 1e-05,
      "loss": 1.399,
      "step": 99380
    },
    {
      "epoch": 23.74,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 99390
    },
    {
      "epoch": 23.74,
      "learning_rate": 1e-05,
      "loss": 1.3964,
      "step": 99400
    },
    {
      "epoch": 23.74,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.401084303855896,
      "eval_runtime": 32.1245,
      "eval_samples_per_second": 21.292,
      "step": 99400
    },
    {
      "epoch": 23.74,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 99410
    },
    {
      "epoch": 23.74,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 99420
    },
    {
      "epoch": 23.75,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 99430
    },
    {
      "epoch": 23.75,
      "learning_rate": 1e-05,
      "loss": 1.4006,
      "step": 99440
    },
    {
      "epoch": 23.75,
      "learning_rate": 1e-05,
      "loss": 1.3987,
      "step": 99450
    },
    {
      "epoch": 23.75,
      "learning_rate": 1e-05,
      "loss": 1.3973,
      "step": 99460
    },
    {
      "epoch": 23.76,
      "learning_rate": 1e-05,
      "loss": 1.3999,
      "step": 99470
    },
    {
      "epoch": 23.76,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 99480
    },
    {
      "epoch": 23.76,
      "learning_rate": 1e-05,
      "loss": 1.4,
      "step": 99490
    },
    {
      "epoch": 23.76,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 99500
    },
    {
      "epoch": 23.76,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.401384711265564,
      "eval_runtime": 31.7885,
      "eval_samples_per_second": 21.517,
      "step": 99500
    },
    {
      "epoch": 23.77,
      "learning_rate": 1e-05,
      "loss": 1.402,
      "step": 99510
    },
    {
      "epoch": 23.77,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 99520
    },
    {
      "epoch": 23.77,
      "learning_rate": 1e-05,
      "loss": 1.4003,
      "step": 99530
    },
    {
      "epoch": 23.77,
      "learning_rate": 1e-05,
      "loss": 1.3969,
      "step": 99540
    },
    {
      "epoch": 23.78,
      "learning_rate": 1e-05,
      "loss": 1.3998,
      "step": 99550
    },
    {
      "epoch": 23.78,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 99560
    },
    {
      "epoch": 23.78,
      "learning_rate": 1e-05,
      "loss": 1.4015,
      "step": 99570
    },
    {
      "epoch": 23.78,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 99580
    },
    {
      "epoch": 23.79,
      "learning_rate": 1e-05,
      "loss": 1.3981,
      "step": 99590
    },
    {
      "epoch": 23.79,
      "learning_rate": 1e-05,
      "loss": 1.4002,
      "step": 99600
    },
    {
      "epoch": 23.79,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4022239446640015,
      "eval_runtime": 32.1168,
      "eval_samples_per_second": 21.297,
      "step": 99600
    },
    {
      "epoch": 23.79,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 99610
    },
    {
      "epoch": 23.79,
      "learning_rate": 1e-05,
      "loss": 1.4031,
      "step": 99620
    },
    {
      "epoch": 23.8,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 99630
    },
    {
      "epoch": 23.8,
      "learning_rate": 1e-05,
      "loss": 1.4019,
      "step": 99640
    },
    {
      "epoch": 23.8,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 99650
    },
    {
      "epoch": 23.8,
      "learning_rate": 1e-05,
      "loss": 1.3991,
      "step": 99660
    },
    {
      "epoch": 23.8,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 99670
    },
    {
      "epoch": 23.81,
      "learning_rate": 1e-05,
      "loss": 1.3983,
      "step": 99680
    },
    {
      "epoch": 23.81,
      "learning_rate": 1e-05,
      "loss": 1.3979,
      "step": 99690
    },
    {
      "epoch": 23.81,
      "learning_rate": 1e-05,
      "loss": 1.3973,
      "step": 99700
    },
    {
      "epoch": 23.81,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4020286798477173,
      "eval_runtime": 31.8695,
      "eval_samples_per_second": 21.463,
      "step": 99700
    },
    {
      "epoch": 23.81,
      "learning_rate": 1e-05,
      "loss": 1.4023,
      "step": 99710
    },
    {
      "epoch": 23.82,
      "learning_rate": 1e-05,
      "loss": 1.3992,
      "step": 99720
    },
    {
      "epoch": 23.82,
      "learning_rate": 1e-05,
      "loss": 1.4005,
      "step": 99730
    },
    {
      "epoch": 23.82,
      "learning_rate": 1e-05,
      "loss": 1.401,
      "step": 99740
    },
    {
      "epoch": 23.82,
      "learning_rate": 1e-05,
      "loss": 1.3957,
      "step": 99750
    },
    {
      "epoch": 23.83,
      "learning_rate": 1e-05,
      "loss": 1.4008,
      "step": 99760
    },
    {
      "epoch": 23.83,
      "learning_rate": 1e-05,
      "loss": 1.3971,
      "step": 99770
    },
    {
      "epoch": 23.83,
      "learning_rate": 1e-05,
      "loss": 1.3995,
      "step": 99780
    },
    {
      "epoch": 23.83,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 99790
    },
    {
      "epoch": 23.84,
      "learning_rate": 1e-05,
      "loss": 1.3976,
      "step": 99800
    },
    {
      "epoch": 23.84,
      "eval_accuracy": 0.9824561403508771,
      "eval_loss": 1.4012902975082397,
      "eval_runtime": 31.6664,
      "eval_samples_per_second": 21.6,
      "step": 99800
    },
    {
      "epoch": 23.84,
      "learning_rate": 1e-05,
      "loss": 1.3975,
      "step": 99810
    },
    {
      "epoch": 23.84,
      "learning_rate": 1e-05,
      "loss": 1.4018,
      "step": 99820
    },
    {
      "epoch": 23.84,
      "learning_rate": 1e-05,
      "loss": 1.3984,
      "step": 99830
    },
    {
      "epoch": 23.85,
      "learning_rate": 1e-05,
      "loss": 1.4021,
      "step": 99840
    },
    {
      "epoch": 23.85,
      "learning_rate": 1e-05,
      "loss": 1.3961,
      "step": 99850
    },
    {
      "epoch": 23.85,
      "learning_rate": 1e-05,
      "loss": 1.3978,
      "step": 99860
    },
    {
      "epoch": 23.85,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 99870
    },
    {
      "epoch": 23.85,
      "learning_rate": 1e-05,
      "loss": 1.4016,
      "step": 99880
    },
    {
      "epoch": 23.86,
      "learning_rate": 1e-05,
      "loss": 1.3994,
      "step": 99890
    },
    {
      "epoch": 23.86,
      "learning_rate": 1e-05,
      "loss": 1.4001,
      "step": 99900
    },
    {
      "epoch": 23.86,
      "eval_accuracy": 0.9780701754385965,
      "eval_loss": 1.4017143249511719,
      "eval_runtime": 32.2215,
      "eval_samples_per_second": 21.228,
      "step": 99900
    },
    {
      "epoch": 23.86,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 99910
    },
    {
      "epoch": 23.86,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 99920
    },
    {
      "epoch": 23.87,
      "learning_rate": 1e-05,
      "loss": 1.3993,
      "step": 99930
    },
    {
      "epoch": 23.87,
      "learning_rate": 1e-05,
      "loss": 1.4004,
      "step": 99940
    },
    {
      "epoch": 23.87,
      "learning_rate": 1e-05,
      "loss": 1.3986,
      "step": 99950
    },
    {
      "epoch": 23.87,
      "learning_rate": 1e-05,
      "loss": 1.3988,
      "step": 99960
    },
    {
      "epoch": 23.88,
      "learning_rate": 1e-05,
      "loss": 1.4014,
      "step": 99970
    },
    {
      "epoch": 23.88,
      "learning_rate": 1e-05,
      "loss": 1.3981,
      "step": 99980
    },
    {
      "epoch": 23.88,
      "learning_rate": 1e-05,
      "loss": 1.3956,
      "step": 99990
    },
    {
      "epoch": 23.88,
      "learning_rate": 1e-05,
      "loss": 1.4011,
      "step": 100000
    },
    {
      "epoch": 23.88,
      "eval_accuracy": 0.9809941520467836,
      "eval_loss": 1.4024540185928345,
      "eval_runtime": 32.4674,
      "eval_samples_per_second": 21.067,
      "step": 100000
    }
  ],
  "max_steps": 100000,
  "num_train_epochs": 24,
  "total_flos": 1.1674782886841549e+17,
  "trial_name": null,
  "trial_params": null
}