{
  "best_metric": 0.3695652173913043,
  "best_model_checkpoint": "swinv2-tiny-patch4-window8-256-DMAE-da3-colab/checkpoint-1485",
  "epoch": 117.33333333333333,
  "eval_steps": 500,
  "global_step": 2640,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.4444444444444444,
      "grad_norm": 3.825187921524048,
      "learning_rate": 0.0009962121212121213,
      "loss": 1.5683,
      "step": 10
    },
    {
      "epoch": 0.8888888888888888,
      "grad_norm": 5.65834903717041,
      "learning_rate": 0.0009924242424242424,
      "loss": 1.3523,
      "step": 20
    },
    {
      "epoch": 0.9777777777777777,
      "eval_accuracy": 0.32608695652173914,
      "eval_loss": 1.4024076461791992,
      "eval_runtime": 1.0691,
      "eval_samples_per_second": 43.028,
      "eval_steps_per_second": 2.806,
      "step": 22
    },
    {
      "epoch": 1.3333333333333333,
      "grad_norm": 10.152283668518066,
      "learning_rate": 0.0009886363636363636,
      "loss": 1.2778,
      "step": 30
    },
    {
      "epoch": 1.7777777777777777,
      "grad_norm": 3.099470376968384,
      "learning_rate": 0.000984848484848485,
      "loss": 1.3805,
      "step": 40
    },
    {
      "epoch": 2.0,
      "eval_accuracy": 0.2608695652173913,
      "eval_loss": 1.3775243759155273,
      "eval_runtime": 0.9313,
      "eval_samples_per_second": 49.394,
      "eval_steps_per_second": 3.221,
      "step": 45
    },
    {
      "epoch": 2.2222222222222223,
      "grad_norm": 4.7406511306762695,
      "learning_rate": 0.0009810606060606062,
      "loss": 1.3708,
      "step": 50
    },
    {
      "epoch": 2.6666666666666665,
      "grad_norm": 3.888740301132202,
      "learning_rate": 0.0009772727272727272,
      "loss": 1.3221,
      "step": 60
    },
    {
      "epoch": 2.977777777777778,
      "eval_accuracy": 0.30434782608695654,
      "eval_loss": 1.4418785572052002,
      "eval_runtime": 0.8857,
      "eval_samples_per_second": 51.938,
      "eval_steps_per_second": 3.387,
      "step": 67
    },
    {
      "epoch": 3.111111111111111,
      "grad_norm": 2.308217763900757,
      "learning_rate": 0.0009734848484848485,
      "loss": 1.2355,
      "step": 70
    },
    {
      "epoch": 3.5555555555555554,
      "grad_norm": 5.7985053062438965,
      "learning_rate": 0.0009696969696969698,
      "loss": 1.2409,
      "step": 80
    },
    {
      "epoch": 4.0,
      "grad_norm": 110.87601470947266,
      "learning_rate": 0.000965909090909091,
      "loss": 1.297,
      "step": 90
    },
    {
      "epoch": 4.0,
      "eval_accuracy": 0.32608695652173914,
      "eval_loss": 1.3581539392471313,
      "eval_runtime": 0.8843,
      "eval_samples_per_second": 52.019,
      "eval_steps_per_second": 3.393,
      "step": 90
    },
    {
      "epoch": 4.444444444444445,
      "grad_norm": 4.244718551635742,
      "learning_rate": 0.0009621212121212122,
      "loss": 1.3056,
      "step": 100
    },
    {
      "epoch": 4.888888888888889,
      "grad_norm": 4.412840843200684,
      "learning_rate": 0.0009583333333333334,
      "loss": 1.353,
      "step": 110
    },
    {
      "epoch": 4.977777777777778,
      "eval_accuracy": 0.34782608695652173,
      "eval_loss": 1.3405784368515015,
      "eval_runtime": 0.8733,
      "eval_samples_per_second": 52.672,
      "eval_steps_per_second": 3.435,
      "step": 112
    },
    {
      "epoch": 5.333333333333333,
      "grad_norm": 2.0132997035980225,
      "learning_rate": 0.0009545454545454546,
      "loss": 1.3115,
      "step": 120
    },
    {
      "epoch": 5.777777777777778,
      "grad_norm": 3.7579846382141113,
      "learning_rate": 0.0009507575757575758,
      "loss": 1.2627,
      "step": 130
    },
    {
      "epoch": 6.0,
      "eval_accuracy": 0.15217391304347827,
      "eval_loss": 1.3823662996292114,
      "eval_runtime": 0.8973,
      "eval_samples_per_second": 51.264,
      "eval_steps_per_second": 3.343,
      "step": 135
    },
    {
      "epoch": 6.222222222222222,
      "grad_norm": 3.731177806854248,
      "learning_rate": 0.000946969696969697,
      "loss": 1.2545,
      "step": 140
    },
    {
      "epoch": 6.666666666666667,
      "grad_norm": 3.8037898540496826,
      "learning_rate": 0.0009431818181818183,
      "loss": 1.3006,
      "step": 150
    },
    {
      "epoch": 6.977777777777778,
      "eval_accuracy": 0.15217391304347827,
      "eval_loss": 1.4008022546768188,
      "eval_runtime": 1.1693,
      "eval_samples_per_second": 39.34,
      "eval_steps_per_second": 2.566,
      "step": 157
    },
    {
      "epoch": 7.111111111111111,
      "grad_norm": 2.627021074295044,
      "learning_rate": 0.0009393939393939394,
      "loss": 1.2558,
      "step": 160
    },
    {
      "epoch": 7.555555555555555,
      "grad_norm": 2.487823247909546,
      "learning_rate": 0.0009356060606060606,
      "loss": 1.216,
      "step": 170
    },
    {
      "epoch": 8.0,
      "grad_norm": 2.525559902191162,
      "learning_rate": 0.0009318181818181818,
      "loss": 1.2438,
      "step": 180
    },
    {
      "epoch": 8.0,
      "eval_accuracy": 0.32608695652173914,
      "eval_loss": 1.3769112825393677,
      "eval_runtime": 0.9024,
      "eval_samples_per_second": 50.973,
      "eval_steps_per_second": 3.324,
      "step": 180
    },
    {
      "epoch": 8.444444444444445,
      "grad_norm": 9.093722343444824,
      "learning_rate": 0.000928030303030303,
      "loss": 1.2023,
      "step": 190
    },
    {
      "epoch": 8.88888888888889,
      "grad_norm": 7.364065170288086,
      "learning_rate": 0.0009242424242424242,
      "loss": 1.222,
      "step": 200
    },
    {
      "epoch": 8.977777777777778,
      "eval_accuracy": 0.30434782608695654,
      "eval_loss": 1.421162486076355,
      "eval_runtime": 0.8938,
      "eval_samples_per_second": 51.468,
      "eval_steps_per_second": 3.357,
      "step": 202
    },
    {
      "epoch": 9.333333333333334,
      "grad_norm": 3.667402744293213,
      "learning_rate": 0.0009204545454545455,
      "loss": 1.2186,
      "step": 210
    },
    {
      "epoch": 9.777777777777779,
      "grad_norm": 10.466683387756348,
      "learning_rate": 0.0009166666666666666,
      "loss": 1.2221,
      "step": 220
    },
    {
      "epoch": 10.0,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.4223273992538452,
      "eval_runtime": 0.8601,
      "eval_samples_per_second": 53.48,
      "eval_steps_per_second": 3.488,
      "step": 225
    },
    {
      "epoch": 10.222222222222221,
      "grad_norm": 13.964035987854004,
      "learning_rate": 0.0009128787878787878,
      "loss": 1.2394,
      "step": 230
    },
    {
      "epoch": 10.666666666666666,
      "grad_norm": 4.716994762420654,
      "learning_rate": 0.0009090909090909091,
      "loss": 1.2262,
      "step": 240
    },
    {
      "epoch": 10.977777777777778,
      "eval_accuracy": 0.2608695652173913,
      "eval_loss": 1.4154268503189087,
      "eval_runtime": 1.2346,
      "eval_samples_per_second": 37.258,
      "eval_steps_per_second": 2.43,
      "step": 247
    },
    {
      "epoch": 11.11111111111111,
      "grad_norm": 11.025979995727539,
      "learning_rate": 0.0009053030303030303,
      "loss": 1.2196,
      "step": 250
    },
    {
      "epoch": 11.555555555555555,
      "grad_norm": 22.928346633911133,
      "learning_rate": 0.0009015151515151515,
      "loss": 1.2131,
      "step": 260
    },
    {
      "epoch": 12.0,
      "grad_norm": 7.0689263343811035,
      "learning_rate": 0.0008977272727272727,
      "loss": 1.2381,
      "step": 270
    },
    {
      "epoch": 12.0,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.3327127695083618,
      "eval_runtime": 0.884,
      "eval_samples_per_second": 52.039,
      "eval_steps_per_second": 3.394,
      "step": 270
    },
    {
      "epoch": 12.444444444444445,
      "grad_norm": 3.511734962463379,
      "learning_rate": 0.000893939393939394,
      "loss": 1.1634,
      "step": 280
    },
    {
      "epoch": 12.88888888888889,
      "grad_norm": 33.742347717285156,
      "learning_rate": 0.0008901515151515151,
      "loss": 1.227,
      "step": 290
    },
    {
      "epoch": 12.977777777777778,
      "eval_accuracy": 0.2826086956521739,
      "eval_loss": 1.288680076599121,
      "eval_runtime": 0.8658,
      "eval_samples_per_second": 53.132,
      "eval_steps_per_second": 3.465,
      "step": 292
    },
    {
      "epoch": 13.333333333333334,
      "grad_norm": 6.175626277923584,
      "learning_rate": 0.0008863636363636364,
      "loss": 1.2082,
      "step": 300
    },
    {
      "epoch": 13.777777777777779,
      "grad_norm": 11.057406425476074,
      "learning_rate": 0.0008825757575757576,
      "loss": 1.2158,
      "step": 310
    },
    {
      "epoch": 14.0,
      "eval_accuracy": 0.2608695652173913,
      "eval_loss": 1.3465280532836914,
      "eval_runtime": 0.872,
      "eval_samples_per_second": 52.749,
      "eval_steps_per_second": 3.44,
      "step": 315
    },
    {
      "epoch": 14.222222222222221,
      "grad_norm": 37.48932647705078,
      "learning_rate": 0.0008787878787878789,
      "loss": 1.2026,
      "step": 320
    },
    {
      "epoch": 14.666666666666666,
      "grad_norm": 4.4914960861206055,
      "learning_rate": 0.000875,
      "loss": 1.2174,
      "step": 330
    },
    {
      "epoch": 14.977777777777778,
      "eval_accuracy": 0.30434782608695654,
      "eval_loss": 1.34762704372406,
      "eval_runtime": 1.1749,
      "eval_samples_per_second": 39.152,
      "eval_steps_per_second": 2.553,
      "step": 337
    },
    {
      "epoch": 15.11111111111111,
      "grad_norm": 3.3884503841400146,
      "learning_rate": 0.0008712121212121212,
      "loss": 1.1947,
      "step": 340
    },
    {
      "epoch": 15.555555555555555,
      "grad_norm": 3.722648859024048,
      "learning_rate": 0.0008674242424242425,
      "loss": 1.2054,
      "step": 350
    },
    {
      "epoch": 16.0,
      "grad_norm": 15.185981750488281,
      "learning_rate": 0.0008636363636363636,
      "loss": 1.1767,
      "step": 360
    },
    {
      "epoch": 16.0,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.4023534059524536,
      "eval_runtime": 0.8914,
      "eval_samples_per_second": 51.601,
      "eval_steps_per_second": 3.365,
      "step": 360
    },
    {
      "epoch": 16.444444444444443,
      "grad_norm": 4.847630500793457,
      "learning_rate": 0.0008598484848484849,
      "loss": 1.1721,
      "step": 370
    },
    {
      "epoch": 16.88888888888889,
      "grad_norm": 3.731781482696533,
      "learning_rate": 0.0008560606060606061,
      "loss": 1.2067,
      "step": 380
    },
    {
      "epoch": 16.977777777777778,
      "eval_accuracy": 0.17391304347826086,
      "eval_loss": 1.3664109706878662,
      "eval_runtime": 0.8715,
      "eval_samples_per_second": 52.783,
      "eval_steps_per_second": 3.442,
      "step": 382
    },
    {
      "epoch": 17.333333333333332,
      "grad_norm": 3.2619216442108154,
      "learning_rate": 0.0008522727272727273,
      "loss": 1.1866,
      "step": 390
    },
    {
      "epoch": 17.77777777777778,
      "grad_norm": 11.475706100463867,
      "learning_rate": 0.0008484848484848485,
      "loss": 1.2303,
      "step": 400
    },
    {
      "epoch": 18.0,
      "eval_accuracy": 0.2826086956521739,
      "eval_loss": 1.4259557723999023,
      "eval_runtime": 0.8714,
      "eval_samples_per_second": 52.786,
      "eval_steps_per_second": 3.443,
      "step": 405
    },
    {
      "epoch": 18.22222222222222,
      "grad_norm": 18.98031997680664,
      "learning_rate": 0.0008446969696969698,
      "loss": 1.1939,
      "step": 410
    },
    {
      "epoch": 18.666666666666668,
      "grad_norm": 2.759476900100708,
      "learning_rate": 0.000840909090909091,
      "loss": 1.222,
      "step": 420
    },
    {
      "epoch": 18.977777777777778,
      "eval_accuracy": 0.17391304347826086,
      "eval_loss": 1.480705976486206,
      "eval_runtime": 1.1937,
      "eval_samples_per_second": 38.535,
      "eval_steps_per_second": 2.513,
      "step": 427
    },
    {
      "epoch": 19.11111111111111,
      "grad_norm": 4.408325672149658,
      "learning_rate": 0.0008371212121212122,
      "loss": 1.1725,
      "step": 430
    },
    {
      "epoch": 19.555555555555557,
      "grad_norm": 2.795828104019165,
      "learning_rate": 0.0008333333333333334,
      "loss": 1.1577,
      "step": 440
    },
    {
      "epoch": 20.0,
      "grad_norm": 8.652217864990234,
      "learning_rate": 0.0008295454545454546,
      "loss": 1.2026,
      "step": 450
    },
    {
      "epoch": 20.0,
      "eval_accuracy": 0.17391304347826086,
      "eval_loss": 1.3851475715637207,
      "eval_runtime": 0.8768,
      "eval_samples_per_second": 52.465,
      "eval_steps_per_second": 3.422,
      "step": 450
    },
    {
      "epoch": 20.444444444444443,
      "grad_norm": 4.27825927734375,
      "learning_rate": 0.0008257575757575758,
      "loss": 1.1934,
      "step": 460
    },
    {
      "epoch": 20.88888888888889,
      "grad_norm": 3.146332263946533,
      "learning_rate": 0.000821969696969697,
      "loss": 1.2185,
      "step": 470
    },
    {
      "epoch": 20.977777777777778,
      "eval_accuracy": 0.2608695652173913,
      "eval_loss": 1.3213552236557007,
      "eval_runtime": 0.889,
      "eval_samples_per_second": 51.741,
      "eval_steps_per_second": 3.374,
      "step": 472
    },
    {
      "epoch": 21.333333333333332,
      "grad_norm": 6.067198753356934,
      "learning_rate": 0.0008181818181818183,
      "loss": 1.2416,
      "step": 480
    },
    {
      "epoch": 21.77777777777778,
      "grad_norm": 1.9041274785995483,
      "learning_rate": 0.0008143939393939394,
      "loss": 1.2773,
      "step": 490
    },
    {
      "epoch": 22.0,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.4403716325759888,
      "eval_runtime": 0.9305,
      "eval_samples_per_second": 49.436,
      "eval_steps_per_second": 3.224,
      "step": 495
    },
    {
      "epoch": 22.22222222222222,
      "grad_norm": 3.5751218795776367,
      "learning_rate": 0.0008106060606060606,
      "loss": 1.2544,
      "step": 500
    },
    {
      "epoch": 22.666666666666668,
      "grad_norm": 6.662022590637207,
      "learning_rate": 0.0008068181818181818,
      "loss": 1.227,
      "step": 510
    },
    {
      "epoch": 22.977777777777778,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.453503966331482,
      "eval_runtime": 1.2139,
      "eval_samples_per_second": 37.894,
      "eval_steps_per_second": 2.471,
      "step": 517
    },
    {
      "epoch": 23.11111111111111,
      "grad_norm": 2.344249725341797,
      "learning_rate": 0.000803030303030303,
      "loss": 1.1912,
      "step": 520
    },
    {
      "epoch": 23.555555555555557,
      "grad_norm": 5.093134880065918,
      "learning_rate": 0.0007992424242424242,
      "loss": 1.211,
      "step": 530
    },
    {
      "epoch": 24.0,
      "grad_norm": 3.9864282608032227,
      "learning_rate": 0.0007954545454545455,
      "loss": 1.2032,
      "step": 540
    },
    {
      "epoch": 24.0,
      "eval_accuracy": 0.30434782608695654,
      "eval_loss": 1.3966683149337769,
      "eval_runtime": 0.8732,
      "eval_samples_per_second": 52.682,
      "eval_steps_per_second": 3.436,
      "step": 540
    },
    {
      "epoch": 24.444444444444443,
      "grad_norm": 5.3367438316345215,
      "learning_rate": 0.0007916666666666666,
      "loss": 1.2051,
      "step": 550
    },
    {
      "epoch": 24.88888888888889,
      "grad_norm": 3.129544973373413,
      "learning_rate": 0.0007878787878787878,
      "loss": 1.2223,
      "step": 560
    },
    {
      "epoch": 24.977777777777778,
      "eval_accuracy": 0.32608695652173914,
      "eval_loss": 1.408994436264038,
      "eval_runtime": 0.8542,
      "eval_samples_per_second": 53.849,
      "eval_steps_per_second": 3.512,
      "step": 562
    },
    {
      "epoch": 25.333333333333332,
      "grad_norm": 2.691143751144409,
      "learning_rate": 0.0007840909090909091,
      "loss": 1.2041,
      "step": 570
    },
    {
      "epoch": 25.77777777777778,
      "grad_norm": 9.64910888671875,
      "learning_rate": 0.0007803030303030303,
      "loss": 1.2527,
      "step": 580
    },
    {
      "epoch": 26.0,
      "eval_accuracy": 0.2608695652173913,
      "eval_loss": 1.4858431816101074,
      "eval_runtime": 0.9084,
      "eval_samples_per_second": 50.639,
      "eval_steps_per_second": 3.303,
      "step": 585
    },
    {
      "epoch": 26.22222222222222,
      "grad_norm": 2.8862998485565186,
      "learning_rate": 0.0007765151515151515,
      "loss": 1.1968,
      "step": 590
    },
    {
      "epoch": 26.666666666666668,
      "grad_norm": 2.620485544204712,
      "learning_rate": 0.0007727272727272727,
      "loss": 1.2203,
      "step": 600
    },
    {
      "epoch": 26.977777777777778,
      "eval_accuracy": 0.17391304347826086,
      "eval_loss": 1.4366178512573242,
      "eval_runtime": 0.888,
      "eval_samples_per_second": 51.8,
      "eval_steps_per_second": 3.378,
      "step": 607
    },
    {
      "epoch": 27.11111111111111,
      "grad_norm": 1.672736406326294,
      "learning_rate": 0.000768939393939394,
      "loss": 1.182,
      "step": 610
    },
    {
      "epoch": 27.555555555555557,
      "grad_norm": 4.047964096069336,
      "learning_rate": 0.0007651515151515151,
      "loss": 1.1824,
      "step": 620
    },
    {
      "epoch": 28.0,
      "grad_norm": 7.370471477508545,
      "learning_rate": 0.0007613636363636364,
      "loss": 1.1993,
      "step": 630
    },
    {
      "epoch": 28.0,
      "eval_accuracy": 0.2608695652173913,
      "eval_loss": 1.4055887460708618,
      "eval_runtime": 0.8703,
      "eval_samples_per_second": 52.856,
      "eval_steps_per_second": 3.447,
      "step": 630
    },
    {
      "epoch": 28.444444444444443,
      "grad_norm": 2.3139352798461914,
      "learning_rate": 0.0007575757575757576,
      "loss": 1.1908,
      "step": 640
    },
    {
      "epoch": 28.88888888888889,
      "grad_norm": 2.4274730682373047,
      "learning_rate": 0.0007537878787878788,
      "loss": 1.2014,
      "step": 650
    },
    {
      "epoch": 28.977777777777778,
      "eval_accuracy": 0.30434782608695654,
      "eval_loss": 1.3755403757095337,
      "eval_runtime": 0.8706,
      "eval_samples_per_second": 52.837,
      "eval_steps_per_second": 3.446,
      "step": 652
    },
    {
      "epoch": 29.333333333333332,
      "grad_norm": 2.012920618057251,
      "learning_rate": 0.00075,
      "loss": 1.2134,
      "step": 660
    },
    {
      "epoch": 29.77777777777778,
      "grad_norm": 3.179375171661377,
      "learning_rate": 0.0007462121212121212,
      "loss": 1.2027,
      "step": 670
    },
    {
      "epoch": 30.0,
      "eval_accuracy": 0.2608695652173913,
      "eval_loss": 1.457945466041565,
      "eval_runtime": 1.1362,
      "eval_samples_per_second": 40.485,
      "eval_steps_per_second": 2.64,
      "step": 675
    },
    {
      "epoch": 30.22222222222222,
      "grad_norm": 3.6860008239746094,
      "learning_rate": 0.0007424242424242425,
      "loss": 1.2086,
      "step": 680
    },
    {
      "epoch": 30.666666666666668,
      "grad_norm": 1.5310014486312866,
      "learning_rate": 0.0007386363636363636,
      "loss": 1.1961,
      "step": 690
    },
    {
      "epoch": 30.977777777777778,
      "eval_accuracy": 0.2608695652173913,
      "eval_loss": 1.4524133205413818,
      "eval_runtime": 0.9718,
      "eval_samples_per_second": 47.335,
      "eval_steps_per_second": 3.087,
      "step": 697
    },
    {
      "epoch": 31.11111111111111,
      "grad_norm": 24.489850997924805,
      "learning_rate": 0.0007348484848484849,
      "loss": 1.2059,
      "step": 700
    },
    {
      "epoch": 31.555555555555557,
      "grad_norm": 2.539658308029175,
      "learning_rate": 0.0007310606060606061,
      "loss": 1.1874,
      "step": 710
    },
    {
      "epoch": 32.0,
      "grad_norm": 2.7552974224090576,
      "learning_rate": 0.0007272727272727273,
      "loss": 1.1939,
      "step": 720
    },
    {
      "epoch": 32.0,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.448819875717163,
      "eval_runtime": 0.8862,
      "eval_samples_per_second": 51.905,
      "eval_steps_per_second": 3.385,
      "step": 720
    },
    {
      "epoch": 32.44444444444444,
      "grad_norm": 3.2986652851104736,
      "learning_rate": 0.0007234848484848485,
      "loss": 1.2196,
      "step": 730
    },
    {
      "epoch": 32.888888888888886,
      "grad_norm": 2.5736711025238037,
      "learning_rate": 0.0007196969696969698,
      "loss": 1.1889,
      "step": 740
    },
    {
      "epoch": 32.977777777777774,
      "eval_accuracy": 0.15217391304347827,
      "eval_loss": 1.456831693649292,
      "eval_runtime": 0.8853,
      "eval_samples_per_second": 51.958,
      "eval_steps_per_second": 3.389,
      "step": 742
    },
    {
      "epoch": 33.333333333333336,
      "grad_norm": 5.722994327545166,
      "learning_rate": 0.0007159090909090909,
      "loss": 1.178,
      "step": 750
    },
    {
      "epoch": 33.77777777777778,
      "grad_norm": 1.9777653217315674,
      "learning_rate": 0.0007121212121212122,
      "loss": 1.1871,
      "step": 760
    },
    {
      "epoch": 34.0,
      "eval_accuracy": 0.32608695652173914,
      "eval_loss": 1.3814184665679932,
      "eval_runtime": 0.8797,
      "eval_samples_per_second": 52.292,
      "eval_steps_per_second": 3.41,
      "step": 765
    },
    {
      "epoch": 34.22222222222222,
      "grad_norm": 2.5516679286956787,
      "learning_rate": 0.0007083333333333334,
      "loss": 1.2329,
      "step": 770
    },
    {
      "epoch": 34.666666666666664,
      "grad_norm": 3.3046047687530518,
      "learning_rate": 0.0007045454545454546,
      "loss": 1.1778,
      "step": 780
    },
    {
      "epoch": 34.977777777777774,
      "eval_accuracy": 0.13043478260869565,
      "eval_loss": 1.44027578830719,
      "eval_runtime": 1.196,
      "eval_samples_per_second": 38.461,
      "eval_steps_per_second": 2.508,
      "step": 787
    },
    {
      "epoch": 35.111111111111114,
      "grad_norm": 1.8858623504638672,
      "learning_rate": 0.0007007575757575758,
      "loss": 1.2006,
      "step": 790
    },
    {
      "epoch": 35.55555555555556,
      "grad_norm": 5.07556676864624,
      "learning_rate": 0.000696969696969697,
      "loss": 1.2925,
      "step": 800
    },
    {
      "epoch": 36.0,
      "grad_norm": 5.501034259796143,
      "learning_rate": 0.0006931818181818183,
      "loss": 1.2404,
      "step": 810
    },
    {
      "epoch": 36.0,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.4436728954315186,
      "eval_runtime": 0.9177,
      "eval_samples_per_second": 50.125,
      "eval_steps_per_second": 3.269,
      "step": 810
    },
    {
      "epoch": 36.44444444444444,
      "grad_norm": 2.563979387283325,
      "learning_rate": 0.0006893939393939394,
      "loss": 1.222,
      "step": 820
    },
    {
      "epoch": 36.888888888888886,
      "grad_norm": 2.562971591949463,
      "learning_rate": 0.0006856060606060606,
      "loss": 1.197,
      "step": 830
    },
    {
      "epoch": 36.977777777777774,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.476518988609314,
      "eval_runtime": 0.8797,
      "eval_samples_per_second": 52.292,
      "eval_steps_per_second": 3.41,
      "step": 832
    },
    {
      "epoch": 37.333333333333336,
      "grad_norm": 4.266129016876221,
      "learning_rate": 0.0006818181818181818,
      "loss": 1.1733,
      "step": 840
    },
    {
      "epoch": 37.77777777777778,
      "grad_norm": 2.0012102127075195,
      "learning_rate": 0.000678030303030303,
      "loss": 1.2161,
      "step": 850
    },
    {
      "epoch": 38.0,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.3720359802246094,
      "eval_runtime": 0.8896,
      "eval_samples_per_second": 51.708,
      "eval_steps_per_second": 3.372,
      "step": 855
    },
    {
      "epoch": 38.22222222222222,
      "grad_norm": 5.82732629776001,
      "learning_rate": 0.0006742424242424242,
      "loss": 1.194,
      "step": 860
    },
    {
      "epoch": 38.666666666666664,
      "grad_norm": 3.146446466445923,
      "learning_rate": 0.0006704545454545455,
      "loss": 1.221,
      "step": 870
    },
    {
      "epoch": 38.977777777777774,
      "eval_accuracy": 0.34782608695652173,
      "eval_loss": 1.3749516010284424,
      "eval_runtime": 1.1844,
      "eval_samples_per_second": 38.839,
      "eval_steps_per_second": 2.533,
      "step": 877
    },
    {
      "epoch": 39.111111111111114,
      "grad_norm": 2.4704267978668213,
      "learning_rate": 0.0006666666666666666,
      "loss": 1.2082,
      "step": 880
    },
    {
      "epoch": 39.55555555555556,
      "grad_norm": 3.377894878387451,
      "learning_rate": 0.0006628787878787878,
      "loss": 1.2203,
      "step": 890
    },
    {
      "epoch": 40.0,
      "grad_norm": 3.2576770782470703,
      "learning_rate": 0.0006590909090909091,
      "loss": 1.229,
      "step": 900
    },
    {
      "epoch": 40.0,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.3404773473739624,
      "eval_runtime": 0.876,
      "eval_samples_per_second": 52.511,
      "eval_steps_per_second": 3.425,
      "step": 900
    },
    {
      "epoch": 40.44444444444444,
      "grad_norm": 2.0540192127227783,
      "learning_rate": 0.0006553030303030303,
      "loss": 1.2222,
      "step": 910
    },
    {
      "epoch": 40.888888888888886,
      "grad_norm": 2.336094617843628,
      "learning_rate": 0.0006515151515151515,
      "loss": 1.2046,
      "step": 920
    },
    {
      "epoch": 40.977777777777774,
      "eval_accuracy": 0.2608695652173913,
      "eval_loss": 1.4231040477752686,
      "eval_runtime": 0.899,
      "eval_samples_per_second": 51.168,
      "eval_steps_per_second": 3.337,
      "step": 922
    },
    {
      "epoch": 41.333333333333336,
      "grad_norm": 3.401913642883301,
      "learning_rate": 0.0006477272727272727,
      "loss": 1.2028,
      "step": 930
    },
    {
      "epoch": 41.77777777777778,
      "grad_norm": 1.3715537786483765,
      "learning_rate": 0.000643939393939394,
      "loss": 1.2077,
      "step": 940
    },
    {
      "epoch": 42.0,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.4383732080459595,
      "eval_runtime": 1.0424,
      "eval_samples_per_second": 44.128,
      "eval_steps_per_second": 2.878,
      "step": 945
    },
    {
      "epoch": 42.22222222222222,
      "grad_norm": 3.548417806625366,
      "learning_rate": 0.0006401515151515151,
      "loss": 1.2126,
      "step": 950
    },
    {
      "epoch": 42.666666666666664,
      "grad_norm": 15.91876220703125,
      "learning_rate": 0.0006363636363636364,
      "loss": 1.1865,
      "step": 960
    },
    {
      "epoch": 42.977777777777774,
      "eval_accuracy": 0.2608695652173913,
      "eval_loss": 1.4346174001693726,
      "eval_runtime": 0.9529,
      "eval_samples_per_second": 48.275,
      "eval_steps_per_second": 3.148,
      "step": 967
    },
    {
      "epoch": 43.111111111111114,
      "grad_norm": 3.9566547870635986,
      "learning_rate": 0.0006325757575757576,
      "loss": 1.175,
      "step": 970
    },
    {
      "epoch": 43.55555555555556,
      "grad_norm": 2.891098976135254,
      "learning_rate": 0.0006287878787878788,
      "loss": 1.2044,
      "step": 980
    },
    {
      "epoch": 44.0,
      "grad_norm": 1.5360569953918457,
      "learning_rate": 0.000625,
      "loss": 1.1882,
      "step": 990
    },
    {
      "epoch": 44.0,
      "eval_accuracy": 0.2826086956521739,
      "eval_loss": 1.367881417274475,
      "eval_runtime": 0.8863,
      "eval_samples_per_second": 51.903,
      "eval_steps_per_second": 3.385,
      "step": 990
    },
    {
      "epoch": 44.44444444444444,
      "grad_norm": 2.5285489559173584,
      "learning_rate": 0.0006212121212121212,
      "loss": 1.1957,
      "step": 1000
    },
    {
      "epoch": 44.888888888888886,
      "grad_norm": 4.244635105133057,
      "learning_rate": 0.0006174242424242425,
      "loss": 1.2528,
      "step": 1010
    },
    {
      "epoch": 44.977777777777774,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.3451467752456665,
      "eval_runtime": 0.8705,
      "eval_samples_per_second": 52.845,
      "eval_steps_per_second": 3.446,
      "step": 1012
    },
    {
      "epoch": 45.333333333333336,
      "grad_norm": 1.46609365940094,
      "learning_rate": 0.0006136363636363636,
      "loss": 1.2534,
      "step": 1020
    },
    {
      "epoch": 45.77777777777778,
      "grad_norm": 9.328937530517578,
      "learning_rate": 0.0006098484848484849,
      "loss": 1.1836,
      "step": 1030
    },
    {
      "epoch": 46.0,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.4912604093551636,
      "eval_runtime": 0.8907,
      "eval_samples_per_second": 51.643,
      "eval_steps_per_second": 3.368,
      "step": 1035
    },
    {
      "epoch": 46.22222222222222,
      "grad_norm": 2.901005744934082,
      "learning_rate": 0.0006060606060606061,
      "loss": 1.2226,
      "step": 1040
    },
    {
      "epoch": 46.666666666666664,
      "grad_norm": 4.646663188934326,
      "learning_rate": 0.0006022727272727273,
      "loss": 1.2009,
      "step": 1050
    },
    {
      "epoch": 46.977777777777774,
      "eval_accuracy": 0.32608695652173914,
      "eval_loss": 1.4841315746307373,
      "eval_runtime": 1.1901,
      "eval_samples_per_second": 38.652,
      "eval_steps_per_second": 2.521,
      "step": 1057
    },
    {
      "epoch": 47.111111111111114,
      "grad_norm": 2.45768141746521,
      "learning_rate": 0.0005984848484848485,
      "loss": 1.219,
      "step": 1060
    },
    {
      "epoch": 47.55555555555556,
      "grad_norm": 5.484715938568115,
      "learning_rate": 0.0005946969696969698,
      "loss": 1.2043,
      "step": 1070
    },
    {
      "epoch": 48.0,
      "grad_norm": 62.28512954711914,
      "learning_rate": 0.0005909090909090909,
      "loss": 1.203,
      "step": 1080
    },
    {
      "epoch": 48.0,
      "eval_accuracy": 0.30434782608695654,
      "eval_loss": 1.4326200485229492,
      "eval_runtime": 0.8718,
      "eval_samples_per_second": 52.767,
      "eval_steps_per_second": 3.441,
      "step": 1080
    },
    {
      "epoch": 48.44444444444444,
      "grad_norm": 2.838622808456421,
      "learning_rate": 0.0005871212121212122,
      "loss": 1.2071,
      "step": 1090
    },
    {
      "epoch": 48.888888888888886,
      "grad_norm": 4.207474708557129,
      "learning_rate": 0.0005833333333333334,
      "loss": 1.1679,
      "step": 1100
    },
    {
      "epoch": 48.977777777777774,
      "eval_accuracy": 0.30434782608695654,
      "eval_loss": 1.3934518098831177,
      "eval_runtime": 0.857,
      "eval_samples_per_second": 53.673,
      "eval_steps_per_second": 3.5,
      "step": 1102
    },
    {
      "epoch": 49.333333333333336,
      "grad_norm": 2.942615270614624,
      "learning_rate": 0.0005795454545454545,
      "loss": 1.1719,
      "step": 1110
    },
    {
      "epoch": 49.77777777777778,
      "grad_norm": 54.94511032104492,
      "learning_rate": 0.0005757575757575758,
      "loss": 1.179,
      "step": 1120
    },
    {
      "epoch": 50.0,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.4185277223587036,
      "eval_runtime": 0.8854,
      "eval_samples_per_second": 51.953,
      "eval_steps_per_second": 3.388,
      "step": 1125
    },
    {
      "epoch": 50.22222222222222,
      "grad_norm": 11.695096969604492,
      "learning_rate": 0.000571969696969697,
      "loss": 1.1624,
      "step": 1130
    },
    {
      "epoch": 50.666666666666664,
      "grad_norm": 1.823878288269043,
      "learning_rate": 0.0005681818181818183,
      "loss": 1.1687,
      "step": 1140
    },
    {
      "epoch": 50.977777777777774,
      "eval_accuracy": 0.2826086956521739,
      "eval_loss": 1.3686347007751465,
      "eval_runtime": 1.2456,
      "eval_samples_per_second": 36.929,
      "eval_steps_per_second": 2.408,
      "step": 1147
    },
    {
      "epoch": 51.111111111111114,
      "grad_norm": 17.766942977905273,
      "learning_rate": 0.0005643939393939394,
      "loss": 1.17,
      "step": 1150
    },
    {
      "epoch": 51.55555555555556,
      "grad_norm": 4.572807788848877,
      "learning_rate": 0.0005606060606060606,
      "loss": 1.1363,
      "step": 1160
    },
    {
      "epoch": 52.0,
      "grad_norm": 7.582568168640137,
      "learning_rate": 0.0005568181818181818,
      "loss": 1.1779,
      "step": 1170
    },
    {
      "epoch": 52.0,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.4319127798080444,
      "eval_runtime": 0.895,
      "eval_samples_per_second": 51.396,
      "eval_steps_per_second": 3.352,
      "step": 1170
    },
    {
      "epoch": 52.44444444444444,
      "grad_norm": 2.6896812915802,
      "learning_rate": 0.000553030303030303,
      "loss": 1.1698,
      "step": 1180
    },
    {
      "epoch": 52.888888888888886,
      "grad_norm": 3.666240692138672,
      "learning_rate": 0.0005492424242424242,
      "loss": 1.1566,
      "step": 1190
    },
    {
      "epoch": 52.977777777777774,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.3800519704818726,
      "eval_runtime": 0.9305,
      "eval_samples_per_second": 49.434,
      "eval_steps_per_second": 3.224,
      "step": 1192
    },
    {
      "epoch": 53.333333333333336,
      "grad_norm": 4.159182071685791,
      "learning_rate": 0.0005454545454545455,
      "loss": 1.1785,
      "step": 1200
    },
    {
      "epoch": 53.77777777777778,
      "grad_norm": 5.972134590148926,
      "learning_rate": 0.0005416666666666666,
      "loss": 1.192,
      "step": 1210
    },
    {
      "epoch": 54.0,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.3745651245117188,
      "eval_runtime": 1.1774,
      "eval_samples_per_second": 39.07,
      "eval_steps_per_second": 2.548,
      "step": 1215
    },
    {
      "epoch": 54.22222222222222,
      "grad_norm": 5.2233428955078125,
      "learning_rate": 0.0005378787878787878,
      "loss": 1.1768,
      "step": 1220
    },
    {
      "epoch": 54.666666666666664,
      "grad_norm": 5.244997501373291,
      "learning_rate": 0.0005340909090909091,
      "loss": 1.1803,
      "step": 1230
    },
    {
      "epoch": 54.977777777777774,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.4016964435577393,
      "eval_runtime": 0.8875,
      "eval_samples_per_second": 51.83,
      "eval_steps_per_second": 3.38,
      "step": 1237
    },
    {
      "epoch": 55.111111111111114,
      "grad_norm": 4.16229772567749,
      "learning_rate": 0.0005303030303030302,
      "loss": 1.1548,
      "step": 1240
    },
    {
      "epoch": 55.55555555555556,
      "grad_norm": 3.8485047817230225,
      "learning_rate": 0.0005265151515151515,
      "loss": 1.1629,
      "step": 1250
    },
    {
      "epoch": 56.0,
      "grad_norm": 3.398857593536377,
      "learning_rate": 0.0005227272727272727,
      "loss": 1.194,
      "step": 1260
    },
    {
      "epoch": 56.0,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.4288326501846313,
      "eval_runtime": 0.8835,
      "eval_samples_per_second": 52.063,
      "eval_steps_per_second": 3.395,
      "step": 1260
    },
    {
      "epoch": 56.44444444444444,
      "grad_norm": 2.9012982845306396,
      "learning_rate": 0.000518939393939394,
      "loss": 1.1283,
      "step": 1270
    },
    {
      "epoch": 56.888888888888886,
      "grad_norm": 2.652462959289551,
      "learning_rate": 0.0005151515151515151,
      "loss": 1.1486,
      "step": 1280
    },
    {
      "epoch": 56.977777777777774,
      "eval_accuracy": 0.30434782608695654,
      "eval_loss": 1.392043113708496,
      "eval_runtime": 1.1872,
      "eval_samples_per_second": 38.747,
      "eval_steps_per_second": 2.527,
      "step": 1282
    },
    {
      "epoch": 57.333333333333336,
      "grad_norm": 16.2806453704834,
      "learning_rate": 0.0005113636363636364,
      "loss": 1.154,
      "step": 1290
    },
    {
      "epoch": 57.77777777777778,
      "grad_norm": 2.9445173740386963,
      "learning_rate": 0.0005075757575757576,
      "loss": 1.1429,
      "step": 1300
    },
    {
      "epoch": 58.0,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.461561918258667,
      "eval_runtime": 0.8733,
      "eval_samples_per_second": 52.675,
      "eval_steps_per_second": 3.435,
      "step": 1305
    },
    {
      "epoch": 58.22222222222222,
      "grad_norm": 4.916953086853027,
      "learning_rate": 0.0005037878787878788,
      "loss": 1.1694,
      "step": 1310
    },
    {
      "epoch": 58.666666666666664,
      "grad_norm": 5.632236957550049,
      "learning_rate": 0.0005,
      "loss": 1.1655,
      "step": 1320
    },
    {
      "epoch": 58.977777777777774,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.4119428396224976,
      "eval_runtime": 0.8592,
      "eval_samples_per_second": 53.535,
      "eval_steps_per_second": 3.491,
      "step": 1327
    },
    {
      "epoch": 59.111111111111114,
      "grad_norm": 10.658681869506836,
      "learning_rate": 0.0004962121212121212,
      "loss": 1.148,
      "step": 1330
    },
    {
      "epoch": 59.55555555555556,
      "grad_norm": 4.212299823760986,
      "learning_rate": 0.0004924242424242425,
      "loss": 1.1508,
      "step": 1340
    },
    {
      "epoch": 60.0,
      "grad_norm": 4.59652853012085,
      "learning_rate": 0.0004886363636363636,
      "loss": 1.1697,
      "step": 1350
    },
    {
      "epoch": 60.0,
      "eval_accuracy": 0.2608695652173913,
      "eval_loss": 1.3811644315719604,
      "eval_runtime": 0.9127,
      "eval_samples_per_second": 50.4,
      "eval_steps_per_second": 3.287,
      "step": 1350
    },
    {
      "epoch": 60.44444444444444,
      "grad_norm": 3.656782865524292,
      "learning_rate": 0.0004848484848484849,
      "loss": 1.1312,
      "step": 1360
    },
    {
      "epoch": 60.888888888888886,
      "grad_norm": 9.660019874572754,
      "learning_rate": 0.0004810606060606061,
      "loss": 1.1898,
      "step": 1370
    },
    {
      "epoch": 60.977777777777774,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.4008588790893555,
      "eval_runtime": 1.1831,
      "eval_samples_per_second": 38.883,
      "eval_steps_per_second": 2.536,
      "step": 1372
    },
    {
      "epoch": 61.333333333333336,
      "grad_norm": 5.424683570861816,
      "learning_rate": 0.0004772727272727273,
      "loss": 1.2188,
      "step": 1380
    },
    {
      "epoch": 61.77777777777778,
      "grad_norm": 3.420642375946045,
      "learning_rate": 0.0004734848484848485,
      "loss": 1.1882,
      "step": 1390
    },
    {
      "epoch": 62.0,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.422127604484558,
      "eval_runtime": 1.0461,
      "eval_samples_per_second": 43.972,
      "eval_steps_per_second": 2.868,
      "step": 1395
    },
    {
      "epoch": 62.22222222222222,
      "grad_norm": 37.69420623779297,
      "learning_rate": 0.0004696969696969697,
      "loss": 1.1428,
      "step": 1400
    },
    {
      "epoch": 62.666666666666664,
      "grad_norm": 6.073638439178467,
      "learning_rate": 0.0004659090909090909,
      "loss": 1.134,
      "step": 1410
    },
    {
      "epoch": 62.977777777777774,
      "eval_accuracy": 0.2608695652173913,
      "eval_loss": 1.618972897529602,
      "eval_runtime": 0.888,
      "eval_samples_per_second": 51.801,
      "eval_steps_per_second": 3.378,
      "step": 1417
    },
    {
      "epoch": 63.111111111111114,
      "grad_norm": 14.145001411437988,
      "learning_rate": 0.0004621212121212121,
      "loss": 1.177,
      "step": 1420
    },
    {
      "epoch": 63.55555555555556,
      "grad_norm": 3.415473222732544,
      "learning_rate": 0.0004583333333333333,
      "loss": 1.1739,
      "step": 1430
    },
    {
      "epoch": 64.0,
      "grad_norm": 9.372400283813477,
      "learning_rate": 0.00045454545454545455,
      "loss": 1.1748,
      "step": 1440
    },
    {
      "epoch": 64.0,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.4336298704147339,
      "eval_runtime": 0.9015,
      "eval_samples_per_second": 51.025,
      "eval_steps_per_second": 3.328,
      "step": 1440
    },
    {
      "epoch": 64.44444444444444,
      "grad_norm": 3.4770920276641846,
      "learning_rate": 0.00045075757575757577,
      "loss": 1.1419,
      "step": 1450
    },
    {
      "epoch": 64.88888888888889,
      "grad_norm": 17.83055877685547,
      "learning_rate": 0.000446969696969697,
      "loss": 1.1439,
      "step": 1460
    },
    {
      "epoch": 64.97777777777777,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.3744150400161743,
      "eval_runtime": 1.1623,
      "eval_samples_per_second": 39.576,
      "eval_steps_per_second": 2.581,
      "step": 1462
    },
    {
      "epoch": 65.33333333333333,
      "grad_norm": 5.167716026306152,
      "learning_rate": 0.0004431818181818182,
      "loss": 1.1155,
      "step": 1470
    },
    {
      "epoch": 65.77777777777777,
      "grad_norm": 2.334927558898926,
      "learning_rate": 0.0004393939393939394,
      "loss": 1.1585,
      "step": 1480
    },
    {
      "epoch": 66.0,
      "eval_accuracy": 0.3695652173913043,
      "eval_loss": 1.3992067575454712,
      "eval_runtime": 0.8747,
      "eval_samples_per_second": 52.591,
      "eval_steps_per_second": 3.43,
      "step": 1485
    },
    {
      "epoch": 66.22222222222223,
      "grad_norm": 4.668221473693848,
      "learning_rate": 0.0004356060606060606,
      "loss": 1.136,
      "step": 1490
    },
    {
      "epoch": 66.66666666666667,
      "grad_norm": 2.9979135990142822,
      "learning_rate": 0.0004318181818181818,
      "loss": 1.1344,
      "step": 1500
    },
    {
      "epoch": 66.97777777777777,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.3951774835586548,
      "eval_runtime": 0.8935,
      "eval_samples_per_second": 51.481,
      "eval_steps_per_second": 3.357,
      "step": 1507
    },
    {
      "epoch": 67.11111111111111,
      "grad_norm": 2.9879891872406006,
      "learning_rate": 0.00042803030303030303,
      "loss": 1.1615,
      "step": 1510
    },
    {
      "epoch": 67.55555555555556,
      "grad_norm": 3.322258710861206,
      "learning_rate": 0.00042424242424242425,
      "loss": 1.1635,
      "step": 1520
    },
    {
      "epoch": 68.0,
      "grad_norm": 4.408764362335205,
      "learning_rate": 0.0004204545454545455,
      "loss": 1.1374,
      "step": 1530
    },
    {
      "epoch": 68.0,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.3666102886199951,
      "eval_runtime": 0.8532,
      "eval_samples_per_second": 53.917,
      "eval_steps_per_second": 3.516,
      "step": 1530
    },
    {
      "epoch": 68.44444444444444,
      "grad_norm": 18.494497299194336,
      "learning_rate": 0.0004166666666666667,
      "loss": 1.126,
      "step": 1540
    },
    {
      "epoch": 68.88888888888889,
      "grad_norm": 6.762816905975342,
      "learning_rate": 0.0004128787878787879,
      "loss": 1.1252,
      "step": 1550
    },
    {
      "epoch": 68.97777777777777,
      "eval_accuracy": 0.2826086956521739,
      "eval_loss": 1.3704602718353271,
      "eval_runtime": 1.2029,
      "eval_samples_per_second": 38.24,
      "eval_steps_per_second": 2.494,
      "step": 1552
    },
    {
      "epoch": 69.33333333333333,
      "grad_norm": 4.585610389709473,
      "learning_rate": 0.00040909090909090913,
      "loss": 1.1272,
      "step": 1560
    },
    {
      "epoch": 69.77777777777777,
      "grad_norm": 10.724448204040527,
      "learning_rate": 0.0004053030303030303,
      "loss": 1.1339,
      "step": 1570
    },
    {
      "epoch": 70.0,
      "eval_accuracy": 0.2826086956521739,
      "eval_loss": 1.3982820510864258,
      "eval_runtime": 1.0724,
      "eval_samples_per_second": 42.893,
      "eval_steps_per_second": 2.797,
      "step": 1575
    },
    {
      "epoch": 70.22222222222223,
      "grad_norm": 5.506129741668701,
      "learning_rate": 0.0004015151515151515,
      "loss": 1.1491,
      "step": 1580
    },
    {
      "epoch": 70.66666666666667,
      "grad_norm": 17.69223976135254,
      "learning_rate": 0.00039772727272727274,
      "loss": 1.1344,
      "step": 1590
    },
    {
      "epoch": 70.97777777777777,
      "eval_accuracy": 0.30434782608695654,
      "eval_loss": 1.3792437314987183,
      "eval_runtime": 0.8913,
      "eval_samples_per_second": 51.609,
      "eval_steps_per_second": 3.366,
      "step": 1597
    },
    {
      "epoch": 71.11111111111111,
      "grad_norm": 10.686148643493652,
      "learning_rate": 0.0003939393939393939,
      "loss": 1.1495,
      "step": 1600
    },
    {
      "epoch": 71.55555555555556,
      "grad_norm": 2.353846549987793,
      "learning_rate": 0.0003901515151515151,
      "loss": 1.1566,
      "step": 1610
    },
    {
      "epoch": 72.0,
      "grad_norm": 3.250394821166992,
      "learning_rate": 0.00038636363636363635,
      "loss": 1.1343,
      "step": 1620
    },
    {
      "epoch": 72.0,
      "eval_accuracy": 0.2826086956521739,
      "eval_loss": 1.4466689825057983,
      "eval_runtime": 0.891,
      "eval_samples_per_second": 51.629,
      "eval_steps_per_second": 3.367,
      "step": 1620
    },
    {
      "epoch": 72.44444444444444,
      "grad_norm": 6.152634143829346,
      "learning_rate": 0.00038257575757575757,
      "loss": 1.1417,
      "step": 1630
    },
    {
      "epoch": 72.88888888888889,
      "grad_norm": 93.5090560913086,
      "learning_rate": 0.0003787878787878788,
      "loss": 1.1555,
      "step": 1640
    },
    {
      "epoch": 72.97777777777777,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.4822701215744019,
      "eval_runtime": 1.1509,
      "eval_samples_per_second": 39.968,
      "eval_steps_per_second": 2.607,
      "step": 1642
    },
    {
      "epoch": 73.33333333333333,
      "grad_norm": 90.8385238647461,
      "learning_rate": 0.000375,
      "loss": 1.1227,
      "step": 1650
    },
    {
      "epoch": 73.77777777777777,
      "grad_norm": 3.00873064994812,
      "learning_rate": 0.00037121212121212123,
      "loss": 1.1329,
      "step": 1660
    },
    {
      "epoch": 74.0,
      "eval_accuracy": 0.15217391304347827,
      "eval_loss": 1.5136324167251587,
      "eval_runtime": 1.0615,
      "eval_samples_per_second": 43.334,
      "eval_steps_per_second": 2.826,
      "step": 1665
    },
    {
      "epoch": 74.22222222222223,
      "grad_norm": 9.816000938415527,
      "learning_rate": 0.00036742424242424245,
      "loss": 1.1719,
      "step": 1670
    },
    {
      "epoch": 74.66666666666667,
      "grad_norm": 2.1428558826446533,
      "learning_rate": 0.00036363636363636367,
      "loss": 1.1513,
      "step": 1680
    },
    {
      "epoch": 74.97777777777777,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.479099988937378,
      "eval_runtime": 0.8852,
      "eval_samples_per_second": 51.967,
      "eval_steps_per_second": 3.389,
      "step": 1687
    },
    {
      "epoch": 75.11111111111111,
      "grad_norm": 4.258810997009277,
      "learning_rate": 0.0003598484848484849,
      "loss": 1.1449,
      "step": 1690
    },
    {
      "epoch": 75.55555555555556,
      "grad_norm": 3.7126126289367676,
      "learning_rate": 0.0003560606060606061,
      "loss": 1.1289,
      "step": 1700
    },
    {
      "epoch": 76.0,
      "grad_norm": 4.913658142089844,
      "learning_rate": 0.0003522727272727273,
      "loss": 1.1278,
      "step": 1710
    },
    {
      "epoch": 76.0,
      "eval_accuracy": 0.2608695652173913,
      "eval_loss": 1.4527482986450195,
      "eval_runtime": 0.8695,
      "eval_samples_per_second": 52.906,
      "eval_steps_per_second": 3.45,
      "step": 1710
    },
    {
      "epoch": 76.44444444444444,
      "grad_norm": 5.169142723083496,
      "learning_rate": 0.0003484848484848485,
      "loss": 1.1212,
      "step": 1720
    },
    {
      "epoch": 76.88888888888889,
      "grad_norm": 6.800394058227539,
      "learning_rate": 0.0003446969696969697,
      "loss": 1.0956,
      "step": 1730
    },
    {
      "epoch": 76.97777777777777,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.4839533567428589,
      "eval_runtime": 0.9059,
      "eval_samples_per_second": 50.779,
      "eval_steps_per_second": 3.312,
      "step": 1732
    },
    {
      "epoch": 77.33333333333333,
      "grad_norm": 4.102919578552246,
      "learning_rate": 0.0003409090909090909,
      "loss": 1.1466,
      "step": 1740
    },
    {
      "epoch": 77.77777777777777,
      "grad_norm": 4.560825824737549,
      "learning_rate": 0.0003371212121212121,
      "loss": 1.1131,
      "step": 1750
    },
    {
      "epoch": 78.0,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.4900346994400024,
      "eval_runtime": 1.1932,
      "eval_samples_per_second": 38.552,
      "eval_steps_per_second": 2.514,
      "step": 1755
    },
    {
      "epoch": 78.22222222222223,
      "grad_norm": 1.923434853553772,
      "learning_rate": 0.0003333333333333333,
      "loss": 1.1285,
      "step": 1760
    },
    {
      "epoch": 78.66666666666667,
      "grad_norm": 4.646895885467529,
      "learning_rate": 0.00032954545454545454,
      "loss": 1.1376,
      "step": 1770
    },
    {
      "epoch": 78.97777777777777,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.5395020246505737,
      "eval_runtime": 0.8752,
      "eval_samples_per_second": 52.557,
      "eval_steps_per_second": 3.428,
      "step": 1777
    },
    {
      "epoch": 79.11111111111111,
      "grad_norm": 8.433492660522461,
      "learning_rate": 0.00032575757575757576,
      "loss": 1.1072,
      "step": 1780
    },
    {
      "epoch": 79.55555555555556,
      "grad_norm": 5.669383525848389,
      "learning_rate": 0.000321969696969697,
      "loss": 1.1135,
      "step": 1790
    },
    {
      "epoch": 80.0,
      "grad_norm": 18.24361801147461,
      "learning_rate": 0.0003181818181818182,
      "loss": 1.0883,
      "step": 1800
    },
    {
      "epoch": 80.0,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.5037870407104492,
      "eval_runtime": 0.8647,
      "eval_samples_per_second": 53.198,
      "eval_steps_per_second": 3.469,
      "step": 1800
    },
    {
      "epoch": 80.44444444444444,
      "grad_norm": 8.68807315826416,
      "learning_rate": 0.0003143939393939394,
      "loss": 1.0899,
      "step": 1810
    },
    {
      "epoch": 80.88888888888889,
      "grad_norm": 39.473899841308594,
      "learning_rate": 0.0003106060606060606,
      "loss": 1.1017,
      "step": 1820
    },
    {
      "epoch": 80.97777777777777,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.5392367839813232,
      "eval_runtime": 0.8702,
      "eval_samples_per_second": 52.864,
      "eval_steps_per_second": 3.448,
      "step": 1822
    },
    {
      "epoch": 81.33333333333333,
      "grad_norm": 3.8075191974639893,
      "learning_rate": 0.0003068181818181818,
      "loss": 1.0607,
      "step": 1830
    },
    {
      "epoch": 81.77777777777777,
      "grad_norm": 4.356723308563232,
      "learning_rate": 0.00030303030303030303,
      "loss": 1.1608,
      "step": 1840
    },
    {
      "epoch": 82.0,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.4875361919403076,
      "eval_runtime": 1.2071,
      "eval_samples_per_second": 38.106,
      "eval_steps_per_second": 2.485,
      "step": 1845
    },
    {
      "epoch": 82.22222222222223,
      "grad_norm": 4.517791748046875,
      "learning_rate": 0.00029924242424242425,
      "loss": 1.1144,
      "step": 1850
    },
    {
      "epoch": 82.66666666666667,
      "grad_norm": 4.995427131652832,
      "learning_rate": 0.00029545454545454547,
      "loss": 1.1308,
      "step": 1860
    },
    {
      "epoch": 82.97777777777777,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.5079646110534668,
      "eval_runtime": 0.8614,
      "eval_samples_per_second": 53.403,
      "eval_steps_per_second": 3.483,
      "step": 1867
    },
    {
      "epoch": 83.11111111111111,
      "grad_norm": 4.479354381561279,
      "learning_rate": 0.0002916666666666667,
      "loss": 1.0821,
      "step": 1870
    },
    {
      "epoch": 83.55555555555556,
      "grad_norm": 4.680623531341553,
      "learning_rate": 0.0002878787878787879,
      "loss": 1.0904,
      "step": 1880
    },
    {
      "epoch": 84.0,
      "grad_norm": 11.278671264648438,
      "learning_rate": 0.00028409090909090913,
      "loss": 1.1382,
      "step": 1890
    },
    {
      "epoch": 84.0,
      "eval_accuracy": 0.17391304347826086,
      "eval_loss": 1.4835433959960938,
      "eval_runtime": 0.8759,
      "eval_samples_per_second": 52.52,
      "eval_steps_per_second": 3.425,
      "step": 1890
    },
    {
      "epoch": 84.44444444444444,
      "grad_norm": 4.505593776702881,
      "learning_rate": 0.0002803030303030303,
      "loss": 1.0869,
      "step": 1900
    },
    {
      "epoch": 84.88888888888889,
      "grad_norm": 3.064387083053589,
      "learning_rate": 0.0002765151515151515,
      "loss": 1.1195,
      "step": 1910
    },
    {
      "epoch": 84.97777777777777,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.4076049327850342,
      "eval_runtime": 0.865,
      "eval_samples_per_second": 53.179,
      "eval_steps_per_second": 3.468,
      "step": 1912
    },
    {
      "epoch": 85.33333333333333,
      "grad_norm": 4.745396137237549,
      "learning_rate": 0.00027272727272727274,
      "loss": 1.1153,
      "step": 1920
    },
    {
      "epoch": 85.77777777777777,
      "grad_norm": 14.576072692871094,
      "learning_rate": 0.0002689393939393939,
      "loss": 1.1149,
      "step": 1930
    },
    {
      "epoch": 86.0,
      "eval_accuracy": 0.17391304347826086,
      "eval_loss": 1.4840431213378906,
      "eval_runtime": 1.1314,
      "eval_samples_per_second": 40.656,
      "eval_steps_per_second": 2.651,
      "step": 1935
    },
    {
      "epoch": 86.22222222222223,
      "grad_norm": 4.840237617492676,
      "learning_rate": 0.0002651515151515151,
      "loss": 1.1314,
      "step": 1940
    },
    {
      "epoch": 86.66666666666667,
      "grad_norm": 7.4581756591796875,
      "learning_rate": 0.00026136363636363634,
      "loss": 1.1344,
      "step": 1950
    },
    {
      "epoch": 86.97777777777777,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.473250150680542,
      "eval_runtime": 1.086,
      "eval_samples_per_second": 42.357,
      "eval_steps_per_second": 2.762,
      "step": 1957
    },
    {
      "epoch": 87.11111111111111,
      "grad_norm": 4.788184642791748,
      "learning_rate": 0.00025757575757575756,
      "loss": 1.0985,
      "step": 1960
    },
    {
      "epoch": 87.55555555555556,
      "grad_norm": 7.187996864318848,
      "learning_rate": 0.0002537878787878788,
      "loss": 1.1018,
      "step": 1970
    },
    {
      "epoch": 88.0,
      "grad_norm": 6.393486976623535,
      "learning_rate": 0.00025,
      "loss": 1.1268,
      "step": 1980
    },
    {
      "epoch": 88.0,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.4446380138397217,
      "eval_runtime": 0.877,
      "eval_samples_per_second": 52.451,
      "eval_steps_per_second": 3.421,
      "step": 1980
    },
    {
      "epoch": 88.44444444444444,
      "grad_norm": 19.996049880981445,
      "learning_rate": 0.0002462121212121212,
      "loss": 1.115,
      "step": 1990
    },
    {
      "epoch": 88.88888888888889,
      "grad_norm": 11.436882019042969,
      "learning_rate": 0.00024242424242424245,
      "loss": 1.1267,
      "step": 2000
    },
    {
      "epoch": 88.97777777777777,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.4359674453735352,
      "eval_runtime": 0.8812,
      "eval_samples_per_second": 52.199,
      "eval_steps_per_second": 3.404,
      "step": 2002
    },
    {
      "epoch": 89.33333333333333,
      "grad_norm": 7.150528430938721,
      "learning_rate": 0.00023863636363636364,
      "loss": 1.1335,
      "step": 2010
    },
    {
      "epoch": 89.77777777777777,
      "grad_norm": 142.4888153076172,
      "learning_rate": 0.00023484848484848486,
      "loss": 1.1034,
      "step": 2020
    },
    {
      "epoch": 90.0,
      "eval_accuracy": 0.15217391304347827,
      "eval_loss": 1.4328769445419312,
      "eval_runtime": 0.8938,
      "eval_samples_per_second": 51.465,
      "eval_steps_per_second": 3.356,
      "step": 2025
    },
    {
      "epoch": 90.22222222222223,
      "grad_norm": 13.855992317199707,
      "learning_rate": 0.00023106060606060605,
      "loss": 1.0987,
      "step": 2030
    },
    {
      "epoch": 90.66666666666667,
      "grad_norm": 4.523609638214111,
      "learning_rate": 0.00022727272727272727,
      "loss": 1.1113,
      "step": 2040
    },
    {
      "epoch": 90.97777777777777,
      "eval_accuracy": 0.17391304347826086,
      "eval_loss": 1.4670028686523438,
      "eval_runtime": 1.1907,
      "eval_samples_per_second": 38.631,
      "eval_steps_per_second": 2.519,
      "step": 2047
    },
    {
      "epoch": 91.11111111111111,
      "grad_norm": 8.413890838623047,
      "learning_rate": 0.0002234848484848485,
      "loss": 1.0848,
      "step": 2050
    },
    {
      "epoch": 91.55555555555556,
      "grad_norm": 2.8552653789520264,
      "learning_rate": 0.0002196969696969697,
      "loss": 1.0788,
      "step": 2060
    },
    {
      "epoch": 92.0,
      "grad_norm": 22.520193099975586,
      "learning_rate": 0.0002159090909090909,
      "loss": 1.0957,
      "step": 2070
    },
    {
      "epoch": 92.0,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.4802157878875732,
      "eval_runtime": 0.8728,
      "eval_samples_per_second": 52.706,
      "eval_steps_per_second": 3.437,
      "step": 2070
    },
    {
      "epoch": 92.44444444444444,
      "grad_norm": 17.830493927001953,
      "learning_rate": 0.00021212121212121213,
      "loss": 1.122,
      "step": 2080
    },
    {
      "epoch": 92.88888888888889,
      "grad_norm": 4.072429656982422,
      "learning_rate": 0.00020833333333333335,
      "loss": 1.1227,
      "step": 2090
    },
    {
      "epoch": 92.97777777777777,
      "eval_accuracy": 0.17391304347826086,
      "eval_loss": 1.4715131521224976,
      "eval_runtime": 0.8786,
      "eval_samples_per_second": 52.357,
      "eval_steps_per_second": 3.415,
      "step": 2092
    },
    {
      "epoch": 93.33333333333333,
      "grad_norm": 4.416510581970215,
      "learning_rate": 0.00020454545454545457,
      "loss": 1.0755,
      "step": 2100
    },
    {
      "epoch": 93.77777777777777,
      "grad_norm": 19.229928970336914,
      "learning_rate": 0.00020075757575757576,
      "loss": 1.1083,
      "step": 2110
    },
    {
      "epoch": 94.0,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.4812626838684082,
      "eval_runtime": 0.8656,
      "eval_samples_per_second": 53.14,
      "eval_steps_per_second": 3.466,
      "step": 2115
    },
    {
      "epoch": 94.22222222222223,
      "grad_norm": 4.718553066253662,
      "learning_rate": 0.00019696969696969695,
      "loss": 1.0757,
      "step": 2120
    },
    {
      "epoch": 94.66666666666667,
      "grad_norm": 5.6447601318359375,
      "learning_rate": 0.00019318181818181817,
      "loss": 1.0583,
      "step": 2130
    },
    {
      "epoch": 94.97777777777777,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.520257830619812,
      "eval_runtime": 1.1597,
      "eval_samples_per_second": 39.665,
      "eval_steps_per_second": 2.587,
      "step": 2137
    },
    {
      "epoch": 95.11111111111111,
      "grad_norm": 5.209888935089111,
      "learning_rate": 0.0001893939393939394,
      "loss": 1.098,
      "step": 2140
    },
    {
      "epoch": 95.55555555555556,
      "grad_norm": 7.493114471435547,
      "learning_rate": 0.00018560606060606061,
      "loss": 1.0796,
      "step": 2150
    },
    {
      "epoch": 96.0,
      "grad_norm": 11.232746124267578,
      "learning_rate": 0.00018181818181818183,
      "loss": 1.093,
      "step": 2160
    },
    {
      "epoch": 96.0,
      "eval_accuracy": 0.17391304347826086,
      "eval_loss": 1.5394465923309326,
      "eval_runtime": 0.8817,
      "eval_samples_per_second": 52.171,
      "eval_steps_per_second": 3.402,
      "step": 2160
    },
    {
      "epoch": 96.44444444444444,
      "grad_norm": 9.968954086303711,
      "learning_rate": 0.00017803030303030305,
      "loss": 1.0958,
      "step": 2170
    },
    {
      "epoch": 96.88888888888889,
      "grad_norm": 15.155268669128418,
      "learning_rate": 0.00017424242424242425,
      "loss": 1.0809,
      "step": 2180
    },
    {
      "epoch": 96.97777777777777,
      "eval_accuracy": 0.17391304347826086,
      "eval_loss": 1.4620193243026733,
      "eval_runtime": 0.8656,
      "eval_samples_per_second": 53.142,
      "eval_steps_per_second": 3.466,
      "step": 2182
    },
    {
      "epoch": 97.33333333333333,
      "grad_norm": 10.245019912719727,
      "learning_rate": 0.00017045454545454544,
      "loss": 1.0822,
      "step": 2190
    },
    {
      "epoch": 97.77777777777777,
      "grad_norm": 56.624778747558594,
      "learning_rate": 0.00016666666666666666,
      "loss": 1.0888,
      "step": 2200
    },
    {
      "epoch": 98.0,
      "eval_accuracy": 0.17391304347826086,
      "eval_loss": 1.4407347440719604,
      "eval_runtime": 0.8737,
      "eval_samples_per_second": 52.651,
      "eval_steps_per_second": 3.434,
      "step": 2205
    },
    {
      "epoch": 98.22222222222223,
      "grad_norm": 16.125377655029297,
      "learning_rate": 0.00016287878787878788,
      "loss": 1.0803,
      "step": 2210
    },
    {
      "epoch": 98.66666666666667,
      "grad_norm": 4.7502546310424805,
      "learning_rate": 0.0001590909090909091,
      "loss": 1.1292,
      "step": 2220
    },
    {
      "epoch": 98.97777777777777,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.4577943086624146,
      "eval_runtime": 1.2602,
      "eval_samples_per_second": 36.503,
      "eval_steps_per_second": 2.381,
      "step": 2227
    },
    {
      "epoch": 99.11111111111111,
      "grad_norm": 12.011200904846191,
      "learning_rate": 0.0001553030303030303,
      "loss": 1.076,
      "step": 2230
    },
    {
      "epoch": 99.55555555555556,
      "grad_norm": 10.061297416687012,
      "learning_rate": 0.00015151515151515152,
      "loss": 1.0789,
      "step": 2240
    },
    {
      "epoch": 100.0,
      "grad_norm": 7.683816432952881,
      "learning_rate": 0.00014772727272727274,
      "loss": 1.0754,
      "step": 2250
    },
    {
      "epoch": 100.0,
      "eval_accuracy": 0.17391304347826086,
      "eval_loss": 1.5031030178070068,
      "eval_runtime": 0.8791,
      "eval_samples_per_second": 52.328,
      "eval_steps_per_second": 3.413,
      "step": 2250
    },
    {
      "epoch": 100.44444444444444,
      "grad_norm": 10.789087295532227,
      "learning_rate": 0.00014393939393939396,
      "loss": 1.0662,
      "step": 2260
    },
    {
      "epoch": 100.88888888888889,
      "grad_norm": 14.93548583984375,
      "learning_rate": 0.00014015151515151515,
      "loss": 1.0817,
      "step": 2270
    },
    {
      "epoch": 100.97777777777777,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.4460813999176025,
      "eval_runtime": 0.8764,
      "eval_samples_per_second": 52.489,
      "eval_steps_per_second": 3.423,
      "step": 2272
    },
    {
      "epoch": 101.33333333333333,
      "grad_norm": 7.292104244232178,
      "learning_rate": 0.00013636363636363637,
      "loss": 1.1087,
      "step": 2280
    },
    {
      "epoch": 101.77777777777777,
      "grad_norm": 21.396413803100586,
      "learning_rate": 0.00013257575757575756,
      "loss": 1.0671,
      "step": 2290
    },
    {
      "epoch": 102.0,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.4722661972045898,
      "eval_runtime": 0.8883,
      "eval_samples_per_second": 51.786,
      "eval_steps_per_second": 3.377,
      "step": 2295
    },
    {
      "epoch": 102.22222222222223,
      "grad_norm": 7.56445837020874,
      "learning_rate": 0.00012878787878787878,
      "loss": 1.0837,
      "step": 2300
    },
    {
      "epoch": 102.66666666666667,
      "grad_norm": 12.246611595153809,
      "learning_rate": 0.000125,
      "loss": 1.0815,
      "step": 2310
    },
    {
      "epoch": 102.97777777777777,
      "eval_accuracy": 0.1956521739130435,
      "eval_loss": 1.4988662004470825,
      "eval_runtime": 1.1771,
      "eval_samples_per_second": 39.079,
      "eval_steps_per_second": 2.549,
      "step": 2317
    },
    {
      "epoch": 103.11111111111111,
      "grad_norm": 10.69598388671875,
      "learning_rate": 0.00012121212121212122,
      "loss": 1.0852,
      "step": 2320
    },
    {
      "epoch": 103.55555555555556,
      "grad_norm": 9.647980690002441,
      "learning_rate": 0.00011742424242424243,
      "loss": 1.076,
      "step": 2330
    },
    {
      "epoch": 104.0,
      "grad_norm": 38.92795944213867,
      "learning_rate": 0.00011363636363636364,
      "loss": 1.0967,
      "step": 2340
    },
    {
      "epoch": 104.0,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.465432047843933,
      "eval_runtime": 0.8692,
      "eval_samples_per_second": 52.922,
      "eval_steps_per_second": 3.451,
      "step": 2340
    },
    {
      "epoch": 104.44444444444444,
      "grad_norm": 9.53869915008545,
      "learning_rate": 0.00010984848484848486,
      "loss": 1.0838,
      "step": 2350
    },
    {
      "epoch": 104.88888888888889,
      "grad_norm": 5.269750118255615,
      "learning_rate": 0.00010606060606060606,
      "loss": 1.091,
      "step": 2360
    },
    {
      "epoch": 104.97777777777777,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.4559190273284912,
      "eval_runtime": 0.8756,
      "eval_samples_per_second": 52.535,
      "eval_steps_per_second": 3.426,
      "step": 2362
    },
    {
      "epoch": 105.33333333333333,
      "grad_norm": 12.788183212280273,
      "learning_rate": 0.00010227272727272728,
      "loss": 1.1085,
      "step": 2370
    },
    {
      "epoch": 105.77777777777777,
      "grad_norm": 7.0092058181762695,
      "learning_rate": 9.848484848484848e-05,
      "loss": 1.0895,
      "step": 2380
    },
    {
      "epoch": 106.0,
      "eval_accuracy": 0.2826086956521739,
      "eval_loss": 1.4221450090408325,
      "eval_runtime": 1.1631,
      "eval_samples_per_second": 39.549,
      "eval_steps_per_second": 2.579,
      "step": 2385
    },
    {
      "epoch": 106.22222222222223,
      "grad_norm": 12.933279037475586,
      "learning_rate": 9.46969696969697e-05,
      "loss": 1.1548,
      "step": 2390
    },
    {
      "epoch": 106.66666666666667,
      "grad_norm": 4.65310001373291,
      "learning_rate": 9.090909090909092e-05,
      "loss": 1.0847,
      "step": 2400
    },
    {
      "epoch": 106.97777777777777,
      "eval_accuracy": 0.2826086956521739,
      "eval_loss": 1.4292521476745605,
      "eval_runtime": 0.8992,
      "eval_samples_per_second": 51.156,
      "eval_steps_per_second": 3.336,
      "step": 2407
    },
    {
      "epoch": 107.11111111111111,
      "grad_norm": 7.354497909545898,
      "learning_rate": 8.712121212121212e-05,
      "loss": 1.0996,
      "step": 2410
    },
    {
      "epoch": 107.55555555555556,
      "grad_norm": 5.702237129211426,
      "learning_rate": 8.333333333333333e-05,
      "loss": 1.0883,
      "step": 2420
    },
    {
      "epoch": 108.0,
      "grad_norm": 4.873330116271973,
      "learning_rate": 7.954545454545455e-05,
      "loss": 1.102,
      "step": 2430
    },
    {
      "epoch": 108.0,
      "eval_accuracy": 0.2391304347826087,
      "eval_loss": 1.4582384824752808,
      "eval_runtime": 0.9053,
      "eval_samples_per_second": 50.81,
      "eval_steps_per_second": 3.314,
      "step": 2430
    },
    {
      "epoch": 108.44444444444444,
      "grad_norm": 78.408447265625,
      "learning_rate": 7.575757575757576e-05,
      "loss": 1.1048,
      "step": 2440
    },
    {
      "epoch": 108.88888888888889,
      "grad_norm": 6.651626110076904,
      "learning_rate": 7.196969696969698e-05,
      "loss": 1.0404,
      "step": 2450
    },
    {
      "epoch": 108.97777777777777,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.4655812978744507,
      "eval_runtime": 0.8722,
      "eval_samples_per_second": 52.741,
      "eval_steps_per_second": 3.44,
      "step": 2452
    },
    {
      "epoch": 109.33333333333333,
      "grad_norm": 10.666740417480469,
      "learning_rate": 6.818181818181818e-05,
      "loss": 1.0799,
      "step": 2460
    },
    {
      "epoch": 109.77777777777777,
      "grad_norm": 6.093382358551025,
      "learning_rate": 6.439393939393939e-05,
      "loss": 1.0488,
      "step": 2470
    },
    {
      "epoch": 110.0,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.489004373550415,
      "eval_runtime": 1.0926,
      "eval_samples_per_second": 42.101,
      "eval_steps_per_second": 2.746,
      "step": 2475
    },
    {
      "epoch": 110.22222222222223,
      "grad_norm": 5.012928009033203,
      "learning_rate": 6.060606060606061e-05,
      "loss": 1.1006,
      "step": 2480
    },
    {
      "epoch": 110.66666666666667,
      "grad_norm": 8.174079895019531,
      "learning_rate": 5.681818181818182e-05,
      "loss": 1.0966,
      "step": 2490
    },
    {
      "epoch": 110.97777777777777,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.4631787538528442,
      "eval_runtime": 0.8917,
      "eval_samples_per_second": 51.586,
      "eval_steps_per_second": 3.364,
      "step": 2497
    },
    {
      "epoch": 111.11111111111111,
      "grad_norm": 4.650646686553955,
      "learning_rate": 5.303030303030303e-05,
      "loss": 1.0629,
      "step": 2500
    },
    {
      "epoch": 111.55555555555556,
      "grad_norm": 5.132219314575195,
      "learning_rate": 4.924242424242424e-05,
      "loss": 1.0835,
      "step": 2510
    },
    {
      "epoch": 112.0,
      "grad_norm": 8.792770385742188,
      "learning_rate": 4.545454545454546e-05,
      "loss": 1.0901,
      "step": 2520
    },
    {
      "epoch": 112.0,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.4494850635528564,
      "eval_runtime": 0.872,
      "eval_samples_per_second": 52.75,
      "eval_steps_per_second": 3.44,
      "step": 2520
    },
    {
      "epoch": 112.44444444444444,
      "grad_norm": 6.029526710510254,
      "learning_rate": 4.1666666666666665e-05,
      "loss": 1.0764,
      "step": 2530
    },
    {
      "epoch": 112.88888888888889,
      "grad_norm": 5.1245927810668945,
      "learning_rate": 3.787878787878788e-05,
      "loss": 1.1008,
      "step": 2540
    },
    {
      "epoch": 112.97777777777777,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.4332908391952515,
      "eval_runtime": 0.9004,
      "eval_samples_per_second": 51.09,
      "eval_steps_per_second": 3.332,
      "step": 2542
    },
    {
      "epoch": 113.33333333333333,
      "grad_norm": 11.329251289367676,
      "learning_rate": 3.409090909090909e-05,
      "loss": 1.0763,
      "step": 2550
    },
    {
      "epoch": 113.77777777777777,
      "grad_norm": 3.865112066268921,
      "learning_rate": 3.0303030303030306e-05,
      "loss": 1.0884,
      "step": 2560
    },
    {
      "epoch": 114.0,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.4406064748764038,
      "eval_runtime": 0.9189,
      "eval_samples_per_second": 50.057,
      "eval_steps_per_second": 3.265,
      "step": 2565
    },
    {
      "epoch": 114.22222222222223,
      "grad_norm": 7.575568675994873,
      "learning_rate": 2.6515151515151516e-05,
      "loss": 1.088,
      "step": 2570
    },
    {
      "epoch": 114.66666666666667,
      "grad_norm": 7.7806620597839355,
      "learning_rate": 2.272727272727273e-05,
      "loss": 1.0889,
      "step": 2580
    },
    {
      "epoch": 114.97777777777777,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.447421908378601,
      "eval_runtime": 0.9635,
      "eval_samples_per_second": 47.742,
      "eval_steps_per_second": 3.114,
      "step": 2587
    },
    {
      "epoch": 115.11111111111111,
      "grad_norm": 20.770198822021484,
      "learning_rate": 1.893939393939394e-05,
      "loss": 1.0754,
      "step": 2590
    },
    {
      "epoch": 115.55555555555556,
      "grad_norm": 5.939628601074219,
      "learning_rate": 1.5151515151515153e-05,
      "loss": 1.07,
      "step": 2600
    },
    {
      "epoch": 116.0,
      "grad_norm": 35.29814147949219,
      "learning_rate": 1.1363636363636365e-05,
      "loss": 1.0729,
      "step": 2610
    },
    {
      "epoch": 116.0,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.4561296701431274,
      "eval_runtime": 0.899,
      "eval_samples_per_second": 51.169,
      "eval_steps_per_second": 3.337,
      "step": 2610
    },
    {
      "epoch": 116.44444444444444,
      "grad_norm": 24.017641067504883,
      "learning_rate": 7.5757575757575764e-06,
      "loss": 1.0589,
      "step": 2620
    },
    {
      "epoch": 116.88888888888889,
      "grad_norm": 14.974956512451172,
      "learning_rate": 3.7878787878787882e-06,
      "loss": 1.0671,
      "step": 2630
    },
    {
      "epoch": 116.97777777777777,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.4538123607635498,
      "eval_runtime": 0.8903,
      "eval_samples_per_second": 51.667,
      "eval_steps_per_second": 3.37,
      "step": 2632
    },
    {
      "epoch": 117.33333333333333,
      "grad_norm": 8.030067443847656,
      "learning_rate": 0.0,
      "loss": 1.0937,
      "step": 2640
    },
    {
      "epoch": 117.33333333333333,
      "eval_accuracy": 0.21739130434782608,
      "eval_loss": 1.453188180923462,
      "eval_runtime": 1.73,
      "eval_samples_per_second": 26.589,
      "eval_steps_per_second": 1.734,
      "step": 2640
    },
    {
      "epoch": 117.33333333333333,
      "step": 2640,
      "total_flos": 5.466852859010089e+18,
      "train_loss": 1.1631801536588957,
      "train_runtime": 5282.7516,
      "train_samples_per_second": 32.529,
      "train_steps_per_second": 0.5
    }
  ],
  "logging_steps": 10,
  "max_steps": 2640,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 120,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 5.466852859010089e+18,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}