{
  "best_metric": 8.912869704236611,
  "best_model_checkpoint": "kotoba_v2_enc_logs_epoch2_2/checkpoint-4000",
  "epoch": 0.018846847093816177,
  "eval_steps": 500,
  "global_step": 5500,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 3.4266994716029415e-05,
      "grad_norm": 1.0561553239822388,
      "learning_rate": 1e-05,
      "loss": 0.2361,
      "step": 10
    },
    {
      "epoch": 6.853398943205883e-05,
      "grad_norm": 1.1626238822937012,
      "learning_rate": 1e-05,
      "loss": 0.2265,
      "step": 20
    },
    {
      "epoch": 0.00010280098414808825,
      "grad_norm": 0.9845689535140991,
      "learning_rate": 1e-05,
      "loss": 0.2279,
      "step": 30
    },
    {
      "epoch": 0.00013706797886411766,
      "grad_norm": 1.142356276512146,
      "learning_rate": 1e-05,
      "loss": 0.2382,
      "step": 40
    },
    {
      "epoch": 0.00017133497358014707,
      "grad_norm": 1.0053240060806274,
      "learning_rate": 1e-05,
      "loss": 0.2473,
      "step": 50
    },
    {
      "epoch": 0.0002056019682961765,
      "grad_norm": 1.1098105907440186,
      "learning_rate": 1e-05,
      "loss": 0.2438,
      "step": 60
    },
    {
      "epoch": 0.0002398689630122059,
      "grad_norm": 1.191983699798584,
      "learning_rate": 1e-05,
      "loss": 0.2293,
      "step": 70
    },
    {
      "epoch": 0.0002741359577282353,
      "grad_norm": 1.1295104026794434,
      "learning_rate": 1e-05,
      "loss": 0.2362,
      "step": 80
    },
    {
      "epoch": 0.0003084029524442647,
      "grad_norm": 1.037972092628479,
      "learning_rate": 1e-05,
      "loss": 0.2455,
      "step": 90
    },
    {
      "epoch": 0.00034266994716029413,
      "grad_norm": 1.1975648403167725,
      "learning_rate": 1e-05,
      "loss": 0.2459,
      "step": 100
    },
    {
      "epoch": 0.00037693694187632354,
      "grad_norm": 1.0676342248916626,
      "learning_rate": 1e-05,
      "loss": 0.2271,
      "step": 110
    },
    {
      "epoch": 0.000411203936592353,
      "grad_norm": 1.0749495029449463,
      "learning_rate": 1e-05,
      "loss": 0.2417,
      "step": 120
    },
    {
      "epoch": 0.0004454709313083824,
      "grad_norm": 1.094260811805725,
      "learning_rate": 1e-05,
      "loss": 0.2354,
      "step": 130
    },
    {
      "epoch": 0.0004797379260244118,
      "grad_norm": 1.0395853519439697,
      "learning_rate": 1e-05,
      "loss": 0.2381,
      "step": 140
    },
    {
      "epoch": 0.0005140049207404412,
      "grad_norm": 1.2008885145187378,
      "learning_rate": 1e-05,
      "loss": 0.2354,
      "step": 150
    },
    {
      "epoch": 0.0005482719154564706,
      "grad_norm": 1.0647832155227661,
      "learning_rate": 1e-05,
      "loss": 0.2321,
      "step": 160
    },
    {
      "epoch": 0.0005825389101725,
      "grad_norm": 1.327071189880371,
      "learning_rate": 1e-05,
      "loss": 0.238,
      "step": 170
    },
    {
      "epoch": 0.0006168059048885295,
      "grad_norm": 1.1184055805206299,
      "learning_rate": 1e-05,
      "loss": 0.2242,
      "step": 180
    },
    {
      "epoch": 0.0006510728996045589,
      "grad_norm": 1.2512784004211426,
      "learning_rate": 1e-05,
      "loss": 0.2437,
      "step": 190
    },
    {
      "epoch": 0.0006853398943205883,
      "grad_norm": 1.0614465475082397,
      "learning_rate": 1e-05,
      "loss": 0.2382,
      "step": 200
    },
    {
      "epoch": 0.0007196068890366177,
      "grad_norm": 1.0607149600982666,
      "learning_rate": 1e-05,
      "loss": 0.2381,
      "step": 210
    },
    {
      "epoch": 0.0007538738837526471,
      "grad_norm": 1.0422028303146362,
      "learning_rate": 1e-05,
      "loss": 0.2294,
      "step": 220
    },
    {
      "epoch": 0.0007881408784686765,
      "grad_norm": 1.0162984132766724,
      "learning_rate": 1e-05,
      "loss": 0.2275,
      "step": 230
    },
    {
      "epoch": 0.000822407873184706,
      "grad_norm": 1.1085543632507324,
      "learning_rate": 1e-05,
      "loss": 0.2161,
      "step": 240
    },
    {
      "epoch": 0.0008566748679007354,
      "grad_norm": 1.1854636669158936,
      "learning_rate": 1e-05,
      "loss": 0.2382,
      "step": 250
    },
    {
      "epoch": 0.0008909418626167648,
      "grad_norm": 1.40137779712677,
      "learning_rate": 1e-05,
      "loss": 0.2579,
      "step": 260
    },
    {
      "epoch": 0.0009252088573327942,
      "grad_norm": 1.0814112424850464,
      "learning_rate": 1e-05,
      "loss": 0.2612,
      "step": 270
    },
    {
      "epoch": 0.0009594758520488236,
      "grad_norm": 1.083736538887024,
      "learning_rate": 1e-05,
      "loss": 0.2711,
      "step": 280
    },
    {
      "epoch": 0.000993742846764853,
      "grad_norm": 1.0861411094665527,
      "learning_rate": 1e-05,
      "loss": 0.2642,
      "step": 290
    },
    {
      "epoch": 0.0010280098414808825,
      "grad_norm": 1.1141265630722046,
      "learning_rate": 1e-05,
      "loss": 0.2585,
      "step": 300
    },
    {
      "epoch": 0.0010622768361969119,
      "grad_norm": 1.326241374015808,
      "learning_rate": 1e-05,
      "loss": 0.2858,
      "step": 310
    },
    {
      "epoch": 0.0010965438309129413,
      "grad_norm": 1.393750786781311,
      "learning_rate": 1e-05,
      "loss": 0.2635,
      "step": 320
    },
    {
      "epoch": 0.0011308108256289707,
      "grad_norm": 1.0851459503173828,
      "learning_rate": 1e-05,
      "loss": 0.2565,
      "step": 330
    },
    {
      "epoch": 0.001165077820345,
      "grad_norm": 1.2323757410049438,
      "learning_rate": 1e-05,
      "loss": 0.2465,
      "step": 340
    },
    {
      "epoch": 0.0011993448150610295,
      "grad_norm": 1.376953125,
      "learning_rate": 1e-05,
      "loss": 0.2671,
      "step": 350
    },
    {
      "epoch": 0.001233611809777059,
      "grad_norm": 1.084592580795288,
      "learning_rate": 1e-05,
      "loss": 0.2643,
      "step": 360
    },
    {
      "epoch": 0.0012678788044930883,
      "grad_norm": 1.2907005548477173,
      "learning_rate": 1e-05,
      "loss": 0.2584,
      "step": 370
    },
    {
      "epoch": 0.0013021457992091177,
      "grad_norm": 1.0698130130767822,
      "learning_rate": 1e-05,
      "loss": 0.2526,
      "step": 380
    },
    {
      "epoch": 0.0013364127939251471,
      "grad_norm": 1.1399807929992676,
      "learning_rate": 1e-05,
      "loss": 0.2759,
      "step": 390
    },
    {
      "epoch": 0.0013706797886411765,
      "grad_norm": 1.1480791568756104,
      "learning_rate": 1e-05,
      "loss": 0.2499,
      "step": 400
    },
    {
      "epoch": 0.001404946783357206,
      "grad_norm": 1.3095237016677856,
      "learning_rate": 1e-05,
      "loss": 0.2536,
      "step": 410
    },
    {
      "epoch": 0.0014392137780732353,
      "grad_norm": 1.068246841430664,
      "learning_rate": 1e-05,
      "loss": 0.2604,
      "step": 420
    },
    {
      "epoch": 0.0014734807727892648,
      "grad_norm": 1.2310419082641602,
      "learning_rate": 1e-05,
      "loss": 0.2632,
      "step": 430
    },
    {
      "epoch": 0.0015077477675052942,
      "grad_norm": 1.161867380142212,
      "learning_rate": 1e-05,
      "loss": 0.2584,
      "step": 440
    },
    {
      "epoch": 0.0015420147622213236,
      "grad_norm": 1.1461217403411865,
      "learning_rate": 1e-05,
      "loss": 0.2592,
      "step": 450
    },
    {
      "epoch": 0.001576281756937353,
      "grad_norm": 1.3006030321121216,
      "learning_rate": 1e-05,
      "loss": 0.2607,
      "step": 460
    },
    {
      "epoch": 0.0016105487516533824,
      "grad_norm": 1.1223125457763672,
      "learning_rate": 1e-05,
      "loss": 0.2433,
      "step": 470
    },
    {
      "epoch": 0.001644815746369412,
      "grad_norm": 1.2909380197525024,
      "learning_rate": 1e-05,
      "loss": 0.2693,
      "step": 480
    },
    {
      "epoch": 0.0016790827410854414,
      "grad_norm": 1.2270597219467163,
      "learning_rate": 1e-05,
      "loss": 0.2661,
      "step": 490
    },
    {
      "epoch": 0.0017133497358014708,
      "grad_norm": 1.1439770460128784,
      "learning_rate": 1e-05,
      "loss": 0.2517,
      "step": 500
    },
    {
      "epoch": 0.0017133497358014708,
      "eval_cer": 13.0358087846181,
      "eval_loss": 0.25224336981773376,
      "eval_normalized_cer": 9.4224620303757,
      "eval_runtime": 227.2174,
      "eval_samples_per_second": 2.253,
      "eval_steps_per_second": 0.035,
      "step": 500
    },
    {
      "epoch": 0.0017476167305175002,
      "grad_norm": 1.1377454996109009,
      "learning_rate": 1e-05,
      "loss": 0.2579,
      "step": 510
    },
    {
      "epoch": 0.0017818837252335296,
      "grad_norm": 1.2096498012542725,
      "learning_rate": 1e-05,
      "loss": 0.2727,
      "step": 520
    },
    {
      "epoch": 0.001816150719949559,
      "grad_norm": 1.187213659286499,
      "learning_rate": 1e-05,
      "loss": 0.2562,
      "step": 530
    },
    {
      "epoch": 0.0018504177146655885,
      "grad_norm": 0.969393253326416,
      "learning_rate": 1e-05,
      "loss": 0.2378,
      "step": 540
    },
    {
      "epoch": 0.0018846847093816179,
      "grad_norm": 0.9745528697967529,
      "learning_rate": 1e-05,
      "loss": 0.2774,
      "step": 550
    },
    {
      "epoch": 0.0019189517040976473,
      "grad_norm": 1.0725352764129639,
      "learning_rate": 1e-05,
      "loss": 0.2541,
      "step": 560
    },
    {
      "epoch": 0.0019532186988136767,
      "grad_norm": 1.217871904373169,
      "learning_rate": 1e-05,
      "loss": 0.2395,
      "step": 570
    },
    {
      "epoch": 0.001987485693529706,
      "grad_norm": 1.3582627773284912,
      "learning_rate": 1e-05,
      "loss": 0.2594,
      "step": 580
    },
    {
      "epoch": 0.0020217526882457355,
      "grad_norm": 1.2415379285812378,
      "learning_rate": 1e-05,
      "loss": 0.2582,
      "step": 590
    },
    {
      "epoch": 0.002056019682961765,
      "grad_norm": 0.9810131192207336,
      "learning_rate": 1e-05,
      "loss": 0.2284,
      "step": 600
    },
    {
      "epoch": 0.0020902866776777943,
      "grad_norm": 0.9806564450263977,
      "learning_rate": 1e-05,
      "loss": 0.2688,
      "step": 610
    },
    {
      "epoch": 0.0021245536723938237,
      "grad_norm": 1.2755467891693115,
      "learning_rate": 1e-05,
      "loss": 0.2591,
      "step": 620
    },
    {
      "epoch": 0.002158820667109853,
      "grad_norm": 0.9300326704978943,
      "learning_rate": 1e-05,
      "loss": 0.2444,
      "step": 630
    },
    {
      "epoch": 0.0021930876618258825,
      "grad_norm": 1.1276524066925049,
      "learning_rate": 1e-05,
      "loss": 0.236,
      "step": 640
    },
    {
      "epoch": 0.002227354656541912,
      "grad_norm": 1.1786876916885376,
      "learning_rate": 1e-05,
      "loss": 0.2443,
      "step": 650
    },
    {
      "epoch": 0.0022616216512579414,
      "grad_norm": 1.1702712774276733,
      "learning_rate": 1e-05,
      "loss": 0.2627,
      "step": 660
    },
    {
      "epoch": 0.0022958886459739708,
      "grad_norm": 1.2837899923324585,
      "learning_rate": 1e-05,
      "loss": 0.2378,
      "step": 670
    },
    {
      "epoch": 0.00233015564069,
      "grad_norm": 1.0623608827590942,
      "learning_rate": 1e-05,
      "loss": 0.2491,
      "step": 680
    },
    {
      "epoch": 0.0023644226354060296,
      "grad_norm": 1.1288243532180786,
      "learning_rate": 1e-05,
      "loss": 0.2773,
      "step": 690
    },
    {
      "epoch": 0.002398689630122059,
      "grad_norm": 1.0192692279815674,
      "learning_rate": 1e-05,
      "loss": 0.2492,
      "step": 700
    },
    {
      "epoch": 0.0024329566248380884,
      "grad_norm": 1.2274680137634277,
      "learning_rate": 1e-05,
      "loss": 0.2345,
      "step": 710
    },
    {
      "epoch": 0.002467223619554118,
      "grad_norm": 1.240645170211792,
      "learning_rate": 1e-05,
      "loss": 0.2624,
      "step": 720
    },
    {
      "epoch": 0.002501490614270147,
      "grad_norm": 1.0681366920471191,
      "learning_rate": 1e-05,
      "loss": 0.2553,
      "step": 730
    },
    {
      "epoch": 0.0025357576089861766,
      "grad_norm": 1.0161867141723633,
      "learning_rate": 1e-05,
      "loss": 0.2547,
      "step": 740
    },
    {
      "epoch": 0.002570024603702206,
      "grad_norm": 1.2384017705917358,
      "learning_rate": 1e-05,
      "loss": 0.2449,
      "step": 750
    },
    {
      "epoch": 0.0026042915984182354,
      "grad_norm": 1.1739261150360107,
      "learning_rate": 1e-05,
      "loss": 0.2523,
      "step": 760
    },
    {
      "epoch": 0.002638558593134265,
      "grad_norm": 1.0396535396575928,
      "learning_rate": 1e-05,
      "loss": 0.2535,
      "step": 770
    },
    {
      "epoch": 0.0026728255878502943,
      "grad_norm": 1.14767324924469,
      "learning_rate": 1e-05,
      "loss": 0.2594,
      "step": 780
    },
    {
      "epoch": 0.0027070925825663237,
      "grad_norm": 1.1783303022384644,
      "learning_rate": 1e-05,
      "loss": 0.2546,
      "step": 790
    },
    {
      "epoch": 0.002741359577282353,
      "grad_norm": 1.1065645217895508,
      "learning_rate": 1e-05,
      "loss": 0.2547,
      "step": 800
    },
    {
      "epoch": 0.0027756265719983825,
      "grad_norm": 1.256645917892456,
      "learning_rate": 1e-05,
      "loss": 0.2548,
      "step": 810
    },
    {
      "epoch": 0.002809893566714412,
      "grad_norm": 1.058158278465271,
      "learning_rate": 1e-05,
      "loss": 0.257,
      "step": 820
    },
    {
      "epoch": 0.0028441605614304413,
      "grad_norm": 1.0647656917572021,
      "learning_rate": 1e-05,
      "loss": 0.2479,
      "step": 830
    },
    {
      "epoch": 0.0028784275561464707,
      "grad_norm": 1.1984691619873047,
      "learning_rate": 1e-05,
      "loss": 0.2503,
      "step": 840
    },
    {
      "epoch": 0.0029126945508625,
      "grad_norm": 1.1380070447921753,
      "learning_rate": 1e-05,
      "loss": 0.245,
      "step": 850
    },
    {
      "epoch": 0.0029469615455785295,
      "grad_norm": 1.2131065130233765,
      "learning_rate": 1e-05,
      "loss": 0.242,
      "step": 860
    },
    {
      "epoch": 0.002981228540294559,
      "grad_norm": 1.1822234392166138,
      "learning_rate": 1e-05,
      "loss": 0.2613,
      "step": 870
    },
    {
      "epoch": 0.0030154955350105883,
      "grad_norm": 1.0591018199920654,
      "learning_rate": 1e-05,
      "loss": 0.2654,
      "step": 880
    },
    {
      "epoch": 0.0030497625297266177,
      "grad_norm": 1.2318428754806519,
      "learning_rate": 1e-05,
      "loss": 0.2525,
      "step": 890
    },
    {
      "epoch": 0.003084029524442647,
      "grad_norm": 1.0146839618682861,
      "learning_rate": 1e-05,
      "loss": 0.2609,
      "step": 900
    },
    {
      "epoch": 0.0031182965191586766,
      "grad_norm": 1.1508561372756958,
      "learning_rate": 1e-05,
      "loss": 0.2541,
      "step": 910
    },
    {
      "epoch": 0.003152563513874706,
      "grad_norm": 1.1494849920272827,
      "learning_rate": 1e-05,
      "loss": 0.2461,
      "step": 920
    },
    {
      "epoch": 0.0031868305085907354,
      "grad_norm": 1.2423807382583618,
      "learning_rate": 1e-05,
      "loss": 0.2573,
      "step": 930
    },
    {
      "epoch": 0.0032210975033067648,
      "grad_norm": 1.2714438438415527,
      "learning_rate": 1e-05,
      "loss": 0.2545,
      "step": 940
    },
    {
      "epoch": 0.0032553644980227946,
      "grad_norm": 1.2088007926940918,
      "learning_rate": 1e-05,
      "loss": 0.2773,
      "step": 950
    },
    {
      "epoch": 0.003289631492738824,
      "grad_norm": 1.0737963914871216,
      "learning_rate": 1e-05,
      "loss": 0.2495,
      "step": 960
    },
    {
      "epoch": 0.0033238984874548534,
      "grad_norm": 1.0942472219467163,
      "learning_rate": 1e-05,
      "loss": 0.2401,
      "step": 970
    },
    {
      "epoch": 0.003358165482170883,
      "grad_norm": 1.1282986402511597,
      "learning_rate": 1e-05,
      "loss": 0.2638,
      "step": 980
    },
    {
      "epoch": 0.0033924324768869123,
      "grad_norm": 1.0762425661087036,
      "learning_rate": 1e-05,
      "loss": 0.2619,
      "step": 990
    },
    {
      "epoch": 0.0034266994716029417,
      "grad_norm": 1.09200119972229,
      "learning_rate": 1e-05,
      "loss": 0.2464,
      "step": 1000
    },
    {
      "epoch": 0.0034266994716029417,
      "eval_cer": 13.80313988357735,
      "eval_loss": 0.25397512316703796,
      "eval_normalized_cer": 9.952038369304557,
      "eval_runtime": 227.5088,
      "eval_samples_per_second": 2.25,
      "eval_steps_per_second": 0.035,
      "step": 1000
    },
    {
      "epoch": 0.003460966466318971,
      "grad_norm": 0.9681844711303711,
      "learning_rate": 1e-05,
      "loss": 0.2567,
      "step": 1010
    },
    {
      "epoch": 0.0034952334610350005,
      "grad_norm": 1.0064711570739746,
      "learning_rate": 1e-05,
      "loss": 0.2514,
      "step": 1020
    },
    {
      "epoch": 0.00352950045575103,
      "grad_norm": 1.190294623374939,
      "learning_rate": 1e-05,
      "loss": 0.2654,
      "step": 1030
    },
    {
      "epoch": 0.0035637674504670593,
      "grad_norm": 1.332492709159851,
      "learning_rate": 1e-05,
      "loss": 0.2725,
      "step": 1040
    },
    {
      "epoch": 0.0035980344451830887,
      "grad_norm": 1.1110397577285767,
      "learning_rate": 1e-05,
      "loss": 0.2504,
      "step": 1050
    },
    {
      "epoch": 0.003632301439899118,
      "grad_norm": 1.2327215671539307,
      "learning_rate": 1e-05,
      "loss": 0.2733,
      "step": 1060
    },
    {
      "epoch": 0.0036665684346151475,
      "grad_norm": 1.1694815158843994,
      "learning_rate": 1e-05,
      "loss": 0.2611,
      "step": 1070
    },
    {
      "epoch": 0.003700835429331177,
      "grad_norm": 1.212570309638977,
      "learning_rate": 1e-05,
      "loss": 0.2556,
      "step": 1080
    },
    {
      "epoch": 0.0037351024240472063,
      "grad_norm": 1.1467297077178955,
      "learning_rate": 1e-05,
      "loss": 0.2485,
      "step": 1090
    },
    {
      "epoch": 0.0037693694187632357,
      "grad_norm": 0.9628469347953796,
      "learning_rate": 1e-05,
      "loss": 0.2523,
      "step": 1100
    },
    {
      "epoch": 0.003803636413479265,
      "grad_norm": 1.1593494415283203,
      "learning_rate": 1e-05,
      "loss": 0.2635,
      "step": 1110
    },
    {
      "epoch": 0.0038379034081952946,
      "grad_norm": 1.1376386880874634,
      "learning_rate": 1e-05,
      "loss": 0.2504,
      "step": 1120
    },
    {
      "epoch": 0.003872170402911324,
      "grad_norm": 1.129338026046753,
      "learning_rate": 1e-05,
      "loss": 0.2601,
      "step": 1130
    },
    {
      "epoch": 0.003906437397627353,
      "grad_norm": 1.0889575481414795,
      "learning_rate": 1e-05,
      "loss": 0.2455,
      "step": 1140
    },
    {
      "epoch": 0.003940704392343382,
      "grad_norm": 1.1437270641326904,
      "learning_rate": 1e-05,
      "loss": 0.253,
      "step": 1150
    },
    {
      "epoch": 0.003974971387059412,
      "grad_norm": 1.0283392667770386,
      "learning_rate": 1e-05,
      "loss": 0.2507,
      "step": 1160
    },
    {
      "epoch": 0.004009238381775441,
      "grad_norm": 1.130747675895691,
      "learning_rate": 1e-05,
      "loss": 0.2715,
      "step": 1170
    },
    {
      "epoch": 0.004043505376491471,
      "grad_norm": 1.3483778238296509,
      "learning_rate": 1e-05,
      "loss": 0.2742,
      "step": 1180
    },
    {
      "epoch": 0.0040777723712075,
      "grad_norm": 1.0879924297332764,
      "learning_rate": 1e-05,
      "loss": 0.2641,
      "step": 1190
    },
    {
      "epoch": 0.00411203936592353,
      "grad_norm": 1.1242927312850952,
      "learning_rate": 1e-05,
      "loss": 0.2586,
      "step": 1200
    },
    {
      "epoch": 0.004146306360639559,
      "grad_norm": 1.0185858011245728,
      "learning_rate": 1e-05,
      "loss": 0.2465,
      "step": 1210
    },
    {
      "epoch": 0.004180573355355589,
      "grad_norm": 0.9555259943008423,
      "learning_rate": 1e-05,
      "loss": 0.2528,
      "step": 1220
    },
    {
      "epoch": 0.004214840350071618,
      "grad_norm": 1.210371971130371,
      "learning_rate": 1e-05,
      "loss": 0.2613,
      "step": 1230
    },
    {
      "epoch": 0.0042491073447876474,
      "grad_norm": 1.1261368989944458,
      "learning_rate": 1e-05,
      "loss": 0.2551,
      "step": 1240
    },
    {
      "epoch": 0.004283374339503676,
      "grad_norm": 1.2142603397369385,
      "learning_rate": 1e-05,
      "loss": 0.264,
      "step": 1250
    },
    {
      "epoch": 0.004317641334219706,
      "grad_norm": 1.057758092880249,
      "learning_rate": 1e-05,
      "loss": 0.2587,
      "step": 1260
    },
    {
      "epoch": 0.004351908328935736,
      "grad_norm": 1.0871245861053467,
      "learning_rate": 1e-05,
      "loss": 0.2549,
      "step": 1270
    },
    {
      "epoch": 0.004386175323651765,
      "grad_norm": 1.1214648485183716,
      "learning_rate": 1e-05,
      "loss": 0.2582,
      "step": 1280
    },
    {
      "epoch": 0.004420442318367795,
      "grad_norm": 1.0265707969665527,
      "learning_rate": 1e-05,
      "loss": 0.2123,
      "step": 1290
    },
    {
      "epoch": 0.004454709313083824,
      "grad_norm": 1.1180216073989868,
      "learning_rate": 1e-05,
      "loss": 0.2245,
      "step": 1300
    },
    {
      "epoch": 0.004488976307799854,
      "grad_norm": 1.028238296508789,
      "learning_rate": 1e-05,
      "loss": 0.2118,
      "step": 1310
    },
    {
      "epoch": 0.004523243302515883,
      "grad_norm": 1.0321682691574097,
      "learning_rate": 1e-05,
      "loss": 0.2196,
      "step": 1320
    },
    {
      "epoch": 0.0045575102972319126,
      "grad_norm": 1.1180269718170166,
      "learning_rate": 1e-05,
      "loss": 0.2403,
      "step": 1330
    },
    {
      "epoch": 0.0045917772919479415,
      "grad_norm": 1.079560399055481,
      "learning_rate": 1e-05,
      "loss": 0.2309,
      "step": 1340
    },
    {
      "epoch": 0.004626044286663971,
      "grad_norm": 1.0062284469604492,
      "learning_rate": 1e-05,
      "loss": 0.228,
      "step": 1350
    },
    {
      "epoch": 0.00466031128138,
      "grad_norm": 1.1098395586013794,
      "learning_rate": 1e-05,
      "loss": 0.2435,
      "step": 1360
    },
    {
      "epoch": 0.00469457827609603,
      "grad_norm": 1.0619688034057617,
      "learning_rate": 1e-05,
      "loss": 0.2342,
      "step": 1370
    },
    {
      "epoch": 0.004728845270812059,
      "grad_norm": 1.1943925619125366,
      "learning_rate": 1e-05,
      "loss": 0.2315,
      "step": 1380
    },
    {
      "epoch": 0.004763112265528089,
      "grad_norm": 1.0958552360534668,
      "learning_rate": 1e-05,
      "loss": 0.2379,
      "step": 1390
    },
    {
      "epoch": 0.004797379260244118,
      "grad_norm": 1.0984197854995728,
      "learning_rate": 1e-05,
      "loss": 0.2208,
      "step": 1400
    },
    {
      "epoch": 0.004831646254960148,
      "grad_norm": 1.0741859674453735,
      "learning_rate": 1e-05,
      "loss": 0.2378,
      "step": 1410
    },
    {
      "epoch": 0.004865913249676177,
      "grad_norm": 1.1457058191299438,
      "learning_rate": 1e-05,
      "loss": 0.2516,
      "step": 1420
    },
    {
      "epoch": 0.004900180244392207,
      "grad_norm": 0.9849014282226562,
      "learning_rate": 1e-05,
      "loss": 0.2406,
      "step": 1430
    },
    {
      "epoch": 0.004934447239108236,
      "grad_norm": 1.1174912452697754,
      "learning_rate": 1e-05,
      "loss": 0.2122,
      "step": 1440
    },
    {
      "epoch": 0.0049687142338242654,
      "grad_norm": 1.0292854309082031,
      "learning_rate": 1e-05,
      "loss": 0.2349,
      "step": 1450
    },
    {
      "epoch": 0.005002981228540294,
      "grad_norm": 1.0343785285949707,
      "learning_rate": 1e-05,
      "loss": 0.2158,
      "step": 1460
    },
    {
      "epoch": 0.005037248223256324,
      "grad_norm": 1.1178008317947388,
      "learning_rate": 1e-05,
      "loss": 0.2264,
      "step": 1470
    },
    {
      "epoch": 0.005071515217972353,
      "grad_norm": 1.0238450765609741,
      "learning_rate": 1e-05,
      "loss": 0.2287,
      "step": 1480
    },
    {
      "epoch": 0.005105782212688383,
      "grad_norm": 1.1728886365890503,
      "learning_rate": 1e-05,
      "loss": 0.2373,
      "step": 1490
    },
    {
      "epoch": 0.005140049207404412,
      "grad_norm": 1.227034091949463,
      "learning_rate": 1e-05,
      "loss": 0.222,
      "step": 1500
    },
    {
      "epoch": 0.005140049207404412,
      "eval_cer": 13.150467454577527,
      "eval_loss": 0.25801682472229004,
      "eval_normalized_cer": 9.452438049560353,
      "eval_runtime": 227.9378,
      "eval_samples_per_second": 2.246,
      "eval_steps_per_second": 0.035,
      "step": 1500
    },
    {
      "epoch": 0.005174316202120442,
      "grad_norm": 1.0703920125961304,
      "learning_rate": 1e-05,
      "loss": 0.2156,
      "step": 1510
    },
    {
      "epoch": 0.005208583196836471,
      "grad_norm": 1.1343841552734375,
      "learning_rate": 1e-05,
      "loss": 0.2126,
      "step": 1520
    },
    {
      "epoch": 0.005242850191552501,
      "grad_norm": 1.1743741035461426,
      "learning_rate": 1e-05,
      "loss": 0.2491,
      "step": 1530
    },
    {
      "epoch": 0.00527711718626853,
      "grad_norm": 1.1476744413375854,
      "learning_rate": 1e-05,
      "loss": 0.236,
      "step": 1540
    },
    {
      "epoch": 0.0053113841809845595,
      "grad_norm": 1.0899590253829956,
      "learning_rate": 1e-05,
      "loss": 0.2361,
      "step": 1550
    },
    {
      "epoch": 0.0053456511757005885,
      "grad_norm": 1.0281250476837158,
      "learning_rate": 1e-05,
      "loss": 0.2226,
      "step": 1560
    },
    {
      "epoch": 0.005379918170416618,
      "grad_norm": 0.9932867884635925,
      "learning_rate": 1e-05,
      "loss": 0.2301,
      "step": 1570
    },
    {
      "epoch": 0.005414185165132647,
      "grad_norm": 1.1992309093475342,
      "learning_rate": 1e-05,
      "loss": 0.2179,
      "step": 1580
    },
    {
      "epoch": 0.005448452159848677,
      "grad_norm": 1.0017774105072021,
      "learning_rate": 1e-05,
      "loss": 0.2244,
      "step": 1590
    },
    {
      "epoch": 0.005482719154564706,
      "grad_norm": 1.0827686786651611,
      "learning_rate": 1e-05,
      "loss": 0.2313,
      "step": 1600
    },
    {
      "epoch": 0.005516986149280736,
      "grad_norm": 1.2260409593582153,
      "learning_rate": 1e-05,
      "loss": 0.229,
      "step": 1610
    },
    {
      "epoch": 0.005551253143996765,
      "grad_norm": 1.2530804872512817,
      "learning_rate": 1e-05,
      "loss": 0.2437,
      "step": 1620
    },
    {
      "epoch": 0.005585520138712795,
      "grad_norm": 1.068452000617981,
      "learning_rate": 1e-05,
      "loss": 0.2138,
      "step": 1630
    },
    {
      "epoch": 0.005619787133428824,
      "grad_norm": 1.3108712434768677,
      "learning_rate": 1e-05,
      "loss": 0.2284,
      "step": 1640
    },
    {
      "epoch": 0.005654054128144854,
      "grad_norm": 1.0919209718704224,
      "learning_rate": 1e-05,
      "loss": 0.213,
      "step": 1650
    },
    {
      "epoch": 0.005688321122860883,
      "grad_norm": 1.1530914306640625,
      "learning_rate": 1e-05,
      "loss": 0.2292,
      "step": 1660
    },
    {
      "epoch": 0.005722588117576912,
      "grad_norm": 1.084028959274292,
      "learning_rate": 1e-05,
      "loss": 0.2393,
      "step": 1670
    },
    {
      "epoch": 0.005756855112292941,
      "grad_norm": 1.247847557067871,
      "learning_rate": 1e-05,
      "loss": 0.2452,
      "step": 1680
    },
    {
      "epoch": 0.005791122107008971,
      "grad_norm": 1.03806734085083,
      "learning_rate": 1e-05,
      "loss": 0.2317,
      "step": 1690
    },
    {
      "epoch": 0.005825389101725,
      "grad_norm": 1.1643092632293701,
      "learning_rate": 1e-05,
      "loss": 0.2348,
      "step": 1700
    },
    {
      "epoch": 0.00585965609644103,
      "grad_norm": 1.1066207885742188,
      "learning_rate": 1e-05,
      "loss": 0.2348,
      "step": 1710
    },
    {
      "epoch": 0.005893923091157059,
      "grad_norm": 1.1813760995864868,
      "learning_rate": 1e-05,
      "loss": 0.2295,
      "step": 1720
    },
    {
      "epoch": 0.005928190085873089,
      "grad_norm": 1.1444518566131592,
      "learning_rate": 1e-05,
      "loss": 0.2101,
      "step": 1730
    },
    {
      "epoch": 0.005962457080589118,
      "grad_norm": 1.1485129594802856,
      "learning_rate": 1e-05,
      "loss": 0.2397,
      "step": 1740
    },
    {
      "epoch": 0.005996724075305148,
      "grad_norm": 1.1813607215881348,
      "learning_rate": 1e-05,
      "loss": 0.231,
      "step": 1750
    },
    {
      "epoch": 0.006030991070021177,
      "grad_norm": 1.4075005054473877,
      "learning_rate": 1e-05,
      "loss": 0.2306,
      "step": 1760
    },
    {
      "epoch": 0.0060652580647372065,
      "grad_norm": 1.2183804512023926,
      "learning_rate": 1e-05,
      "loss": 0.2227,
      "step": 1770
    },
    {
      "epoch": 0.0060995250594532355,
      "grad_norm": 1.3654927015304565,
      "learning_rate": 1e-05,
      "loss": 0.2341,
      "step": 1780
    },
    {
      "epoch": 0.006133792054169265,
      "grad_norm": 1.2806668281555176,
      "learning_rate": 1e-05,
      "loss": 0.2226,
      "step": 1790
    },
    {
      "epoch": 0.006168059048885294,
      "grad_norm": 1.2949618101119995,
      "learning_rate": 1e-05,
      "loss": 0.2698,
      "step": 1800
    },
    {
      "epoch": 0.006202326043601324,
      "grad_norm": 1.3080159425735474,
      "learning_rate": 1e-05,
      "loss": 0.2691,
      "step": 1810
    },
    {
      "epoch": 0.006236593038317353,
      "grad_norm": 1.1831908226013184,
      "learning_rate": 1e-05,
      "loss": 0.2644,
      "step": 1820
    },
    {
      "epoch": 0.006270860033033383,
      "grad_norm": 1.1216965913772583,
      "learning_rate": 1e-05,
      "loss": 0.2582,
      "step": 1830
    },
    {
      "epoch": 0.006305127027749412,
      "grad_norm": 1.1943161487579346,
      "learning_rate": 1e-05,
      "loss": 0.2769,
      "step": 1840
    },
    {
      "epoch": 0.006339394022465442,
      "grad_norm": 1.0856040716171265,
      "learning_rate": 1e-05,
      "loss": 0.2526,
      "step": 1850
    },
    {
      "epoch": 0.006373661017181471,
      "grad_norm": 1.1100040674209595,
      "learning_rate": 1e-05,
      "loss": 0.2576,
      "step": 1860
    },
    {
      "epoch": 0.006407928011897501,
      "grad_norm": 1.3369051218032837,
      "learning_rate": 1e-05,
      "loss": 0.2684,
      "step": 1870
    },
    {
      "epoch": 0.0064421950066135296,
      "grad_norm": 1.158797264099121,
      "learning_rate": 1e-05,
      "loss": 0.2474,
      "step": 1880
    },
    {
      "epoch": 0.006476462001329559,
      "grad_norm": 1.1821873188018799,
      "learning_rate": 1e-05,
      "loss": 0.272,
      "step": 1890
    },
    {
      "epoch": 0.006510728996045589,
      "grad_norm": 1.0739686489105225,
      "learning_rate": 1e-05,
      "loss": 0.2798,
      "step": 1900
    },
    {
      "epoch": 0.006544995990761618,
      "grad_norm": 1.0639653205871582,
      "learning_rate": 1e-05,
      "loss": 0.2682,
      "step": 1910
    },
    {
      "epoch": 0.006579262985477648,
      "grad_norm": 1.2149512767791748,
      "learning_rate": 1e-05,
      "loss": 0.2586,
      "step": 1920
    },
    {
      "epoch": 0.006613529980193677,
      "grad_norm": 1.1057014465332031,
      "learning_rate": 1e-05,
      "loss": 0.2719,
      "step": 1930
    },
    {
      "epoch": 0.006647796974909707,
      "grad_norm": 1.0929185152053833,
      "learning_rate": 1e-05,
      "loss": 0.2703,
      "step": 1940
    },
    {
      "epoch": 0.006682063969625736,
      "grad_norm": 1.0322917699813843,
      "learning_rate": 1e-05,
      "loss": 0.2477,
      "step": 1950
    },
    {
      "epoch": 0.006716330964341766,
      "grad_norm": 1.2460272312164307,
      "learning_rate": 1e-05,
      "loss": 0.2816,
      "step": 1960
    },
    {
      "epoch": 0.006750597959057795,
      "grad_norm": 1.2049859762191772,
      "learning_rate": 1e-05,
      "loss": 0.2648,
      "step": 1970
    },
    {
      "epoch": 0.0067848649537738245,
      "grad_norm": 1.1182633638381958,
      "learning_rate": 1e-05,
      "loss": 0.2549,
      "step": 1980
    },
    {
      "epoch": 0.0068191319484898535,
      "grad_norm": 1.1514990329742432,
      "learning_rate": 1e-05,
      "loss": 0.2695,
      "step": 1990
    },
    {
      "epoch": 0.006853398943205883,
      "grad_norm": 1.0150858163833618,
      "learning_rate": 1e-05,
      "loss": 0.2532,
      "step": 2000
    },
    {
      "epoch": 0.006853398943205883,
      "eval_cer": 13.565002645969306,
      "eval_loss": 0.2523655593395233,
      "eval_normalized_cer": 9.942046362909672,
      "eval_runtime": 226.5571,
      "eval_samples_per_second": 2.26,
      "eval_steps_per_second": 0.035,
      "step": 2000
    },
    {
      "epoch": 0.006887665937921912,
      "grad_norm": 1.0476700067520142,
      "learning_rate": 1e-05,
      "loss": 0.2555,
      "step": 2010
    },
    {
      "epoch": 0.006921932932637942,
      "grad_norm": 1.1178691387176514,
      "learning_rate": 1e-05,
      "loss": 0.2489,
      "step": 2020
    },
    {
      "epoch": 0.006956199927353971,
      "grad_norm": 1.2596313953399658,
      "learning_rate": 1e-05,
      "loss": 0.2884,
      "step": 2030
    },
    {
      "epoch": 0.006990466922070001,
      "grad_norm": 1.1929702758789062,
      "learning_rate": 1e-05,
      "loss": 0.262,
      "step": 2040
    },
    {
      "epoch": 0.00702473391678603,
      "grad_norm": 1.1269497871398926,
      "learning_rate": 1e-05,
      "loss": 0.2758,
      "step": 2050
    },
    {
      "epoch": 0.00705900091150206,
      "grad_norm": 1.1495511531829834,
      "learning_rate": 1e-05,
      "loss": 0.2668,
      "step": 2060
    },
    {
      "epoch": 0.007093267906218089,
      "grad_norm": 1.0648061037063599,
      "learning_rate": 1e-05,
      "loss": 0.2548,
      "step": 2070
    },
    {
      "epoch": 0.007127534900934119,
      "grad_norm": 1.3193435668945312,
      "learning_rate": 1e-05,
      "loss": 0.2743,
      "step": 2080
    },
    {
      "epoch": 0.0071618018956501476,
      "grad_norm": 1.2877907752990723,
      "learning_rate": 1e-05,
      "loss": 0.248,
      "step": 2090
    },
    {
      "epoch": 0.007196068890366177,
      "grad_norm": 1.2012474536895752,
      "learning_rate": 1e-05,
      "loss": 0.2662,
      "step": 2100
    },
    {
      "epoch": 0.007230335885082206,
      "grad_norm": 1.1491566896438599,
      "learning_rate": 1e-05,
      "loss": 0.2666,
      "step": 2110
    },
    {
      "epoch": 0.007264602879798236,
      "grad_norm": 1.1861019134521484,
      "learning_rate": 1e-05,
      "loss": 0.2618,
      "step": 2120
    },
    {
      "epoch": 0.007298869874514265,
      "grad_norm": 1.123963713645935,
      "learning_rate": 1e-05,
      "loss": 0.2646,
      "step": 2130
    },
    {
      "epoch": 0.007333136869230295,
      "grad_norm": 1.2697441577911377,
      "learning_rate": 1e-05,
      "loss": 0.2713,
      "step": 2140
    },
    {
      "epoch": 0.007367403863946324,
      "grad_norm": 0.9741083383560181,
      "learning_rate": 1e-05,
      "loss": 0.2463,
      "step": 2150
    },
    {
      "epoch": 0.007401670858662354,
      "grad_norm": 1.0292670726776123,
      "learning_rate": 1e-05,
      "loss": 0.2542,
      "step": 2160
    },
    {
      "epoch": 0.007435937853378383,
      "grad_norm": 1.0958001613616943,
      "learning_rate": 1e-05,
      "loss": 0.2463,
      "step": 2170
    },
    {
      "epoch": 0.007470204848094413,
      "grad_norm": 1.166869044303894,
      "learning_rate": 1e-05,
      "loss": 0.2454,
      "step": 2180
    },
    {
      "epoch": 0.007504471842810442,
      "grad_norm": 1.2552424669265747,
      "learning_rate": 1e-05,
      "loss": 0.2498,
      "step": 2190
    },
    {
      "epoch": 0.0075387388375264715,
      "grad_norm": 1.1589868068695068,
      "learning_rate": 1e-05,
      "loss": 0.2659,
      "step": 2200
    },
    {
      "epoch": 0.0075730058322425004,
      "grad_norm": 1.1640287637710571,
      "learning_rate": 1e-05,
      "loss": 0.257,
      "step": 2210
    },
    {
      "epoch": 0.00760727282695853,
      "grad_norm": 1.0953587293624878,
      "learning_rate": 1e-05,
      "loss": 0.2444,
      "step": 2220
    },
    {
      "epoch": 0.007641539821674559,
      "grad_norm": 1.2174441814422607,
      "learning_rate": 1e-05,
      "loss": 0.2626,
      "step": 2230
    },
    {
      "epoch": 0.007675806816390589,
      "grad_norm": 1.1194220781326294,
      "learning_rate": 1e-05,
      "loss": 0.241,
      "step": 2240
    },
    {
      "epoch": 0.007710073811106618,
      "grad_norm": 1.0677419900894165,
      "learning_rate": 1e-05,
      "loss": 0.2718,
      "step": 2250
    },
    {
      "epoch": 0.007744340805822648,
      "grad_norm": 1.0956069231033325,
      "learning_rate": 1e-05,
      "loss": 0.2493,
      "step": 2260
    },
    {
      "epoch": 0.007778607800538677,
      "grad_norm": 1.1772819757461548,
      "learning_rate": 1e-05,
      "loss": 0.2614,
      "step": 2270
    },
    {
      "epoch": 0.007812874795254707,
      "grad_norm": 1.0341110229492188,
      "learning_rate": 1e-05,
      "loss": 0.2488,
      "step": 2280
    },
    {
      "epoch": 0.007847141789970737,
      "grad_norm": 1.174186110496521,
      "learning_rate": 1e-05,
      "loss": 0.2542,
      "step": 2290
    },
    {
      "epoch": 0.007881408784686765,
      "grad_norm": 0.9867792725563049,
      "learning_rate": 1e-05,
      "loss": 0.2582,
      "step": 2300
    },
    {
      "epoch": 0.007915675779402795,
      "grad_norm": 1.1443661451339722,
      "learning_rate": 1e-05,
      "loss": 0.2331,
      "step": 2310
    },
    {
      "epoch": 0.007949942774118824,
      "grad_norm": 1.117896318435669,
      "learning_rate": 1e-05,
      "loss": 0.2277,
      "step": 2320
    },
    {
      "epoch": 0.007984209768834854,
      "grad_norm": 1.13510000705719,
      "learning_rate": 1e-05,
      "loss": 0.2137,
      "step": 2330
    },
    {
      "epoch": 0.008018476763550882,
      "grad_norm": 0.9749162793159485,
      "learning_rate": 1e-05,
      "loss": 0.2161,
      "step": 2340
    },
    {
      "epoch": 0.008052743758266912,
      "grad_norm": 1.1519534587860107,
      "learning_rate": 1e-05,
      "loss": 0.2254,
      "step": 2350
    },
    {
      "epoch": 0.008087010752982942,
      "grad_norm": 1.0861778259277344,
      "learning_rate": 1e-05,
      "loss": 0.2153,
      "step": 2360
    },
    {
      "epoch": 0.008121277747698972,
      "grad_norm": 1.0184444189071655,
      "learning_rate": 1e-05,
      "loss": 0.2066,
      "step": 2370
    },
    {
      "epoch": 0.008155544742415,
      "grad_norm": 1.0581239461898804,
      "learning_rate": 1e-05,
      "loss": 0.2243,
      "step": 2380
    },
    {
      "epoch": 0.00818981173713103,
      "grad_norm": 0.9954540729522705,
      "learning_rate": 1e-05,
      "loss": 0.2171,
      "step": 2390
    },
    {
      "epoch": 0.00822407873184706,
      "grad_norm": 1.121960163116455,
      "learning_rate": 1e-05,
      "loss": 0.2216,
      "step": 2400
    },
    {
      "epoch": 0.00825834572656309,
      "grad_norm": 1.097725510597229,
      "learning_rate": 1e-05,
      "loss": 0.2142,
      "step": 2410
    },
    {
      "epoch": 0.008292612721279118,
      "grad_norm": 1.0566459894180298,
      "learning_rate": 1e-05,
      "loss": 0.2272,
      "step": 2420
    },
    {
      "epoch": 0.008326879715995147,
      "grad_norm": 1.0077927112579346,
      "learning_rate": 1e-05,
      "loss": 0.211,
      "step": 2430
    },
    {
      "epoch": 0.008361146710711177,
      "grad_norm": 1.176035761833191,
      "learning_rate": 1e-05,
      "loss": 0.2125,
      "step": 2440
    },
    {
      "epoch": 0.008395413705427207,
      "grad_norm": 1.0064568519592285,
      "learning_rate": 1e-05,
      "loss": 0.2066,
      "step": 2450
    },
    {
      "epoch": 0.008429680700143235,
      "grad_norm": 1.1852171421051025,
      "learning_rate": 1e-05,
      "loss": 0.2087,
      "step": 2460
    },
    {
      "epoch": 0.008463947694859265,
      "grad_norm": 0.9580971002578735,
      "learning_rate": 1e-05,
      "loss": 0.2172,
      "step": 2470
    },
    {
      "epoch": 0.008498214689575295,
      "grad_norm": 1.1230813264846802,
      "learning_rate": 1e-05,
      "loss": 0.2104,
      "step": 2480
    },
    {
      "epoch": 0.008532481684291325,
      "grad_norm": 1.1891340017318726,
      "learning_rate": 1e-05,
      "loss": 0.229,
      "step": 2490
    },
    {
      "epoch": 0.008566748679007353,
      "grad_norm": 1.2579045295715332,
      "learning_rate": 1e-05,
      "loss": 0.2109,
      "step": 2500
    },
    {
      "epoch": 0.008566748679007353,
      "eval_cer": 13.300405715293703,
      "eval_loss": 0.26059621572494507,
      "eval_normalized_cer": 9.502398081534773,
      "eval_runtime": 226.5522,
      "eval_samples_per_second": 2.26,
      "eval_steps_per_second": 0.035,
      "step": 2500
    },
    {
      "epoch": 0.008601015673723383,
      "grad_norm": 1.0522507429122925,
      "learning_rate": 1e-05,
      "loss": 0.2154,
      "step": 2510
    },
    {
      "epoch": 0.008635282668439413,
      "grad_norm": 1.0875492095947266,
      "learning_rate": 1e-05,
      "loss": 0.2251,
      "step": 2520
    },
    {
      "epoch": 0.008669549663155442,
      "grad_norm": 1.0868346691131592,
      "learning_rate": 1e-05,
      "loss": 0.2086,
      "step": 2530
    },
    {
      "epoch": 0.008703816657871472,
      "grad_norm": 1.0993175506591797,
      "learning_rate": 1e-05,
      "loss": 0.205,
      "step": 2540
    },
    {
      "epoch": 0.0087380836525875,
      "grad_norm": 1.0495941638946533,
      "learning_rate": 1e-05,
      "loss": 0.2135,
      "step": 2550
    },
    {
      "epoch": 0.00877235064730353,
      "grad_norm": 1.0326807498931885,
      "learning_rate": 1e-05,
      "loss": 0.2105,
      "step": 2560
    },
    {
      "epoch": 0.00880661764201956,
      "grad_norm": 1.0804367065429688,
      "learning_rate": 1e-05,
      "loss": 0.2438,
      "step": 2570
    },
    {
      "epoch": 0.00884088463673559,
      "grad_norm": 1.0738023519515991,
      "learning_rate": 1e-05,
      "loss": 0.2537,
      "step": 2580
    },
    {
      "epoch": 0.008875151631451618,
      "grad_norm": 1.1695871353149414,
      "learning_rate": 1e-05,
      "loss": 0.2518,
      "step": 2590
    },
    {
      "epoch": 0.008909418626167648,
      "grad_norm": 1.155653476715088,
      "learning_rate": 1e-05,
      "loss": 0.2592,
      "step": 2600
    },
    {
      "epoch": 0.008943685620883678,
      "grad_norm": 1.1516027450561523,
      "learning_rate": 1e-05,
      "loss": 0.2387,
      "step": 2610
    },
    {
      "epoch": 0.008977952615599707,
      "grad_norm": 1.2618260383605957,
      "learning_rate": 1e-05,
      "loss": 0.2638,
      "step": 2620
    },
    {
      "epoch": 0.009012219610315736,
      "grad_norm": 1.2422987222671509,
      "learning_rate": 1e-05,
      "loss": 0.2459,
      "step": 2630
    },
    {
      "epoch": 0.009046486605031765,
      "grad_norm": 1.1460082530975342,
      "learning_rate": 1e-05,
      "loss": 0.2509,
      "step": 2640
    },
    {
      "epoch": 0.009080753599747795,
      "grad_norm": 1.2502261400222778,
      "learning_rate": 1e-05,
      "loss": 0.2595,
      "step": 2650
    },
    {
      "epoch": 0.009115020594463825,
      "grad_norm": 1.139840006828308,
      "learning_rate": 1e-05,
      "loss": 0.255,
      "step": 2660
    },
    {
      "epoch": 0.009149287589179853,
      "grad_norm": 1.3247896432876587,
      "learning_rate": 1e-05,
      "loss": 0.2721,
      "step": 2670
    },
    {
      "epoch": 0.009183554583895883,
      "grad_norm": 1.1355103254318237,
      "learning_rate": 1e-05,
      "loss": 0.2604,
      "step": 2680
    },
    {
      "epoch": 0.009217821578611913,
      "grad_norm": 1.106541633605957,
      "learning_rate": 1e-05,
      "loss": 0.2374,
      "step": 2690
    },
    {
      "epoch": 0.009252088573327943,
      "grad_norm": 1.2375975847244263,
      "learning_rate": 1e-05,
      "loss": 0.2719,
      "step": 2700
    },
    {
      "epoch": 0.00928635556804397,
      "grad_norm": 1.1048275232315063,
      "learning_rate": 1e-05,
      "loss": 0.2791,
      "step": 2710
    },
    {
      "epoch": 0.00932062256276,
      "grad_norm": 0.9889766573905945,
      "learning_rate": 1e-05,
      "loss": 0.2457,
      "step": 2720
    },
    {
      "epoch": 0.00935488955747603,
      "grad_norm": 1.1566202640533447,
      "learning_rate": 1e-05,
      "loss": 0.252,
      "step": 2730
    },
    {
      "epoch": 0.00938915655219206,
      "grad_norm": 1.1586074829101562,
      "learning_rate": 1e-05,
      "loss": 0.2517,
      "step": 2740
    },
    {
      "epoch": 0.009423423546908088,
      "grad_norm": 0.990419328212738,
      "learning_rate": 1e-05,
      "loss": 0.2572,
      "step": 2750
    },
    {
      "epoch": 0.009457690541624118,
      "grad_norm": 1.1101089715957642,
      "learning_rate": 1e-05,
      "loss": 0.2525,
      "step": 2760
    },
    {
      "epoch": 0.009491957536340148,
      "grad_norm": 1.0488269329071045,
      "learning_rate": 1e-05,
      "loss": 0.2452,
      "step": 2770
    },
    {
      "epoch": 0.009526224531056178,
      "grad_norm": 1.1127737760543823,
      "learning_rate": 1e-05,
      "loss": 0.2578,
      "step": 2780
    },
    {
      "epoch": 0.009560491525772206,
      "grad_norm": 1.2353262901306152,
      "learning_rate": 1e-05,
      "loss": 0.2412,
      "step": 2790
    },
    {
      "epoch": 0.009594758520488236,
      "grad_norm": 1.1262571811676025,
      "learning_rate": 1e-05,
      "loss": 0.2438,
      "step": 2800
    },
    {
      "epoch": 0.009629025515204266,
      "grad_norm": 1.294323205947876,
      "learning_rate": 1e-05,
      "loss": 0.2512,
      "step": 2810
    },
    {
      "epoch": 0.009663292509920296,
      "grad_norm": 1.0706703662872314,
      "learning_rate": 1e-05,
      "loss": 0.2595,
      "step": 2820
    },
    {
      "epoch": 0.009697559504636324,
      "grad_norm": 1.0089077949523926,
      "learning_rate": 1e-05,
      "loss": 0.2522,
      "step": 2830
    },
    {
      "epoch": 0.009731826499352354,
      "grad_norm": 0.9697763323783875,
      "learning_rate": 1e-05,
      "loss": 0.2684,
      "step": 2840
    },
    {
      "epoch": 0.009766093494068383,
      "grad_norm": 1.1122509241104126,
      "learning_rate": 1e-05,
      "loss": 0.2629,
      "step": 2850
    },
    {
      "epoch": 0.009800360488784413,
      "grad_norm": 1.0381057262420654,
      "learning_rate": 1e-05,
      "loss": 0.2482,
      "step": 2860
    },
    {
      "epoch": 0.009834627483500441,
      "grad_norm": 1.126947045326233,
      "learning_rate": 1e-05,
      "loss": 0.2674,
      "step": 2870
    },
    {
      "epoch": 0.009868894478216471,
      "grad_norm": 1.0714973211288452,
      "learning_rate": 1e-05,
      "loss": 0.2634,
      "step": 2880
    },
    {
      "epoch": 0.009903161472932501,
      "grad_norm": 1.0942039489746094,
      "learning_rate": 1e-05,
      "loss": 0.2751,
      "step": 2890
    },
    {
      "epoch": 0.009937428467648531,
      "grad_norm": 1.1503955125808716,
      "learning_rate": 1e-05,
      "loss": 0.272,
      "step": 2900
    },
    {
      "epoch": 0.009971695462364559,
      "grad_norm": 1.1912988424301147,
      "learning_rate": 1e-05,
      "loss": 0.2645,
      "step": 2910
    },
    {
      "epoch": 0.010005962457080589,
      "grad_norm": 1.0941249132156372,
      "learning_rate": 1e-05,
      "loss": 0.2531,
      "step": 2920
    },
    {
      "epoch": 0.010040229451796619,
      "grad_norm": 1.2545968294143677,
      "learning_rate": 1e-05,
      "loss": 0.2562,
      "step": 2930
    },
    {
      "epoch": 0.010074496446512649,
      "grad_norm": 1.3605022430419922,
      "learning_rate": 1e-05,
      "loss": 0.2601,
      "step": 2940
    },
    {
      "epoch": 0.010108763441228677,
      "grad_norm": 1.0911775827407837,
      "learning_rate": 1e-05,
      "loss": 0.2605,
      "step": 2950
    },
    {
      "epoch": 0.010143030435944706,
      "grad_norm": 1.133867859840393,
      "learning_rate": 1e-05,
      "loss": 0.2554,
      "step": 2960
    },
    {
      "epoch": 0.010177297430660736,
      "grad_norm": 1.2511764764785767,
      "learning_rate": 1e-05,
      "loss": 0.2658,
      "step": 2970
    },
    {
      "epoch": 0.010211564425376766,
      "grad_norm": 1.1705303192138672,
      "learning_rate": 1e-05,
      "loss": 0.2737,
      "step": 2980
    },
    {
      "epoch": 0.010245831420092794,
      "grad_norm": 1.132071614265442,
      "learning_rate": 1e-05,
      "loss": 0.2665,
      "step": 2990
    },
    {
      "epoch": 0.010280098414808824,
      "grad_norm": 1.2301791906356812,
      "learning_rate": 1e-05,
      "loss": 0.2645,
      "step": 3000
    },
    {
      "epoch": 0.010280098414808824,
      "eval_cer": 12.938789910037043,
      "eval_loss": 0.2511608302593231,
      "eval_normalized_cer": 9.152677857713828,
      "eval_runtime": 227.4553,
      "eval_samples_per_second": 2.251,
      "eval_steps_per_second": 0.035,
      "step": 3000
    },
    {
      "epoch": 0.010314365409524854,
      "grad_norm": 1.1527032852172852,
      "learning_rate": 1e-05,
      "loss": 0.2508,
      "step": 3010
    },
    {
      "epoch": 0.010348632404240884,
      "grad_norm": 1.1162952184677124,
      "learning_rate": 1e-05,
      "loss": 0.2728,
      "step": 3020
    },
    {
      "epoch": 0.010382899398956912,
      "grad_norm": 1.062084436416626,
      "learning_rate": 1e-05,
      "loss": 0.2496,
      "step": 3030
    },
    {
      "epoch": 0.010417166393672942,
      "grad_norm": 1.1536457538604736,
      "learning_rate": 1e-05,
      "loss": 0.2633,
      "step": 3040
    },
    {
      "epoch": 0.010451433388388972,
      "grad_norm": 1.2096189260482788,
      "learning_rate": 1e-05,
      "loss": 0.2498,
      "step": 3050
    },
    {
      "epoch": 0.010485700383105001,
      "grad_norm": 0.9950299263000488,
      "learning_rate": 1e-05,
      "loss": 0.246,
      "step": 3060
    },
    {
      "epoch": 0.01051996737782103,
      "grad_norm": 1.0628243684768677,
      "learning_rate": 1e-05,
      "loss": 0.2544,
      "step": 3070
    },
    {
      "epoch": 0.01055423437253706,
      "grad_norm": 1.042555570602417,
      "learning_rate": 1e-05,
      "loss": 0.2401,
      "step": 3080
    },
    {
      "epoch": 0.01058850136725309,
      "grad_norm": 1.22646164894104,
      "learning_rate": 1e-05,
      "loss": 0.2503,
      "step": 3090
    },
    {
      "epoch": 0.010622768361969119,
      "grad_norm": 1.0862691402435303,
      "learning_rate": 1e-05,
      "loss": 0.2508,
      "step": 3100
    },
    {
      "epoch": 0.010657035356685147,
      "grad_norm": 1.148868203163147,
      "learning_rate": 1e-05,
      "loss": 0.2526,
      "step": 3110
    },
    {
      "epoch": 0.010691302351401177,
      "grad_norm": 1.1677169799804688,
      "learning_rate": 1e-05,
      "loss": 0.2481,
      "step": 3120
    },
    {
      "epoch": 0.010725569346117207,
      "grad_norm": 0.990696132183075,
      "learning_rate": 1e-05,
      "loss": 0.2421,
      "step": 3130
    },
    {
      "epoch": 0.010759836340833237,
      "grad_norm": 1.2869263887405396,
      "learning_rate": 1e-05,
      "loss": 0.2463,
      "step": 3140
    },
    {
      "epoch": 0.010794103335549265,
      "grad_norm": 1.0741721391677856,
      "learning_rate": 1e-05,
      "loss": 0.2617,
      "step": 3150
    },
    {
      "epoch": 0.010828370330265295,
      "grad_norm": 1.103102445602417,
      "learning_rate": 1e-05,
      "loss": 0.2442,
      "step": 3160
    },
    {
      "epoch": 0.010862637324981324,
      "grad_norm": 1.2562378644943237,
      "learning_rate": 1e-05,
      "loss": 0.2589,
      "step": 3170
    },
    {
      "epoch": 0.010896904319697354,
      "grad_norm": 1.2153191566467285,
      "learning_rate": 1e-05,
      "loss": 0.2417,
      "step": 3180
    },
    {
      "epoch": 0.010931171314413384,
      "grad_norm": 1.0507330894470215,
      "learning_rate": 1e-05,
      "loss": 0.2607,
      "step": 3190
    },
    {
      "epoch": 0.010965438309129412,
      "grad_norm": 1.1882787942886353,
      "learning_rate": 1e-05,
      "loss": 0.2469,
      "step": 3200
    },
    {
      "epoch": 0.010999705303845442,
      "grad_norm": 1.1394702196121216,
      "learning_rate": 1e-05,
      "loss": 0.2574,
      "step": 3210
    },
    {
      "epoch": 0.011033972298561472,
      "grad_norm": 1.2482614517211914,
      "learning_rate": 1e-05,
      "loss": 0.2456,
      "step": 3220
    },
    {
      "epoch": 0.011068239293277502,
      "grad_norm": 1.0362995862960815,
      "learning_rate": 1e-05,
      "loss": 0.2589,
      "step": 3230
    },
    {
      "epoch": 0.01110250628799353,
      "grad_norm": 1.1730456352233887,
      "learning_rate": 1e-05,
      "loss": 0.2497,
      "step": 3240
    },
    {
      "epoch": 0.01113677328270956,
      "grad_norm": 1.1563142538070679,
      "learning_rate": 1e-05,
      "loss": 0.2439,
      "step": 3250
    },
    {
      "epoch": 0.01117104027742559,
      "grad_norm": 1.1030769348144531,
      "learning_rate": 1e-05,
      "loss": 0.2671,
      "step": 3260
    },
    {
      "epoch": 0.01120530727214162,
      "grad_norm": 1.1719223260879517,
      "learning_rate": 1e-05,
      "loss": 0.2501,
      "step": 3270
    },
    {
      "epoch": 0.011239574266857648,
      "grad_norm": 1.1840440034866333,
      "learning_rate": 1e-05,
      "loss": 0.2643,
      "step": 3280
    },
    {
      "epoch": 0.011273841261573677,
      "grad_norm": 1.1928170919418335,
      "learning_rate": 1e-05,
      "loss": 0.2629,
      "step": 3290
    },
    {
      "epoch": 0.011308108256289707,
      "grad_norm": 1.0311812162399292,
      "learning_rate": 1e-05,
      "loss": 0.2552,
      "step": 3300
    },
    {
      "epoch": 0.011342375251005737,
      "grad_norm": 1.1625889539718628,
      "learning_rate": 1e-05,
      "loss": 0.2561,
      "step": 3310
    },
    {
      "epoch": 0.011376642245721765,
      "grad_norm": 1.0287625789642334,
      "learning_rate": 1e-05,
      "loss": 0.2341,
      "step": 3320
    },
    {
      "epoch": 0.011410909240437795,
      "grad_norm": 1.1310815811157227,
      "learning_rate": 1e-05,
      "loss": 0.2554,
      "step": 3330
    },
    {
      "epoch": 0.011445176235153825,
      "grad_norm": 1.1266168355941772,
      "learning_rate": 1e-05,
      "loss": 0.234,
      "step": 3340
    },
    {
      "epoch": 0.011479443229869855,
      "grad_norm": 1.1979014873504639,
      "learning_rate": 1e-05,
      "loss": 0.2559,
      "step": 3350
    },
    {
      "epoch": 0.011513710224585883,
      "grad_norm": 1.0378515720367432,
      "learning_rate": 1e-05,
      "loss": 0.2502,
      "step": 3360
    },
    {
      "epoch": 0.011547977219301913,
      "grad_norm": 1.1832512617111206,
      "learning_rate": 1e-05,
      "loss": 0.236,
      "step": 3370
    },
    {
      "epoch": 0.011582244214017942,
      "grad_norm": 0.9605569839477539,
      "learning_rate": 1e-05,
      "loss": 0.2349,
      "step": 3380
    },
    {
      "epoch": 0.011616511208733972,
      "grad_norm": 1.0463056564331055,
      "learning_rate": 1e-05,
      "loss": 0.2328,
      "step": 3390
    },
    {
      "epoch": 0.01165077820345,
      "grad_norm": 1.1021932363510132,
      "learning_rate": 1e-05,
      "loss": 0.2383,
      "step": 3400
    },
    {
      "epoch": 0.01168504519816603,
      "grad_norm": 1.040493130683899,
      "learning_rate": 1e-05,
      "loss": 0.2374,
      "step": 3410
    },
    {
      "epoch": 0.01171931219288206,
      "grad_norm": 1.1483063697814941,
      "learning_rate": 1e-05,
      "loss": 0.2398,
      "step": 3420
    },
    {
      "epoch": 0.01175357918759809,
      "grad_norm": 1.0316531658172607,
      "learning_rate": 1e-05,
      "loss": 0.2329,
      "step": 3430
    },
    {
      "epoch": 0.011787846182314118,
      "grad_norm": 1.1677886247634888,
      "learning_rate": 1e-05,
      "loss": 0.2493,
      "step": 3440
    },
    {
      "epoch": 0.011822113177030148,
      "grad_norm": 1.2078930139541626,
      "learning_rate": 1e-05,
      "loss": 0.2337,
      "step": 3450
    },
    {
      "epoch": 0.011856380171746178,
      "grad_norm": 1.178202509880066,
      "learning_rate": 1e-05,
      "loss": 0.239,
      "step": 3460
    },
    {
      "epoch": 0.011890647166462208,
      "grad_norm": 1.0453248023986816,
      "learning_rate": 1e-05,
      "loss": 0.2233,
      "step": 3470
    },
    {
      "epoch": 0.011924914161178236,
      "grad_norm": 1.0171067714691162,
      "learning_rate": 1e-05,
      "loss": 0.2338,
      "step": 3480
    },
    {
      "epoch": 0.011959181155894266,
      "grad_norm": 1.051792860031128,
      "learning_rate": 1e-05,
      "loss": 0.2394,
      "step": 3490
    },
    {
      "epoch": 0.011993448150610295,
      "grad_norm": 1.1237847805023193,
      "learning_rate": 1e-05,
      "loss": 0.2428,
      "step": 3500
    },
    {
      "epoch": 0.011993448150610295,
      "eval_cer": 13.071088375374845,
      "eval_loss": 0.25454944372177124,
      "eval_normalized_cer": 9.542366107114308,
      "eval_runtime": 228.9468,
      "eval_samples_per_second": 2.236,
      "eval_steps_per_second": 0.035,
      "step": 3500
    },
    {
      "epoch": 0.012027715145326325,
      "grad_norm": 1.1366350650787354,
      "learning_rate": 1e-05,
      "loss": 0.2353,
      "step": 3510
    },
    {
      "epoch": 0.012061982140042353,
      "grad_norm": 1.136927604675293,
      "learning_rate": 1e-05,
      "loss": 0.2358,
      "step": 3520
    },
    {
      "epoch": 0.012096249134758383,
      "grad_norm": 1.1875656843185425,
      "learning_rate": 1e-05,
      "loss": 0.2305,
      "step": 3530
    },
    {
      "epoch": 0.012130516129474413,
      "grad_norm": 1.2016057968139648,
      "learning_rate": 1e-05,
      "loss": 0.2435,
      "step": 3540
    },
    {
      "epoch": 0.012164783124190443,
      "grad_norm": 1.209622859954834,
      "learning_rate": 1e-05,
      "loss": 0.2361,
      "step": 3550
    },
    {
      "epoch": 0.012199050118906471,
      "grad_norm": 1.0696970224380493,
      "learning_rate": 1e-05,
      "loss": 0.2385,
      "step": 3560
    },
    {
      "epoch": 0.0122333171136225,
      "grad_norm": 1.2674167156219482,
      "learning_rate": 1e-05,
      "loss": 0.243,
      "step": 3570
    },
    {
      "epoch": 0.01226758410833853,
      "grad_norm": 1.2928141355514526,
      "learning_rate": 1e-05,
      "loss": 0.2491,
      "step": 3580
    },
    {
      "epoch": 0.01230185110305456,
      "grad_norm": 1.0642272233963013,
      "learning_rate": 1e-05,
      "loss": 0.2356,
      "step": 3590
    },
    {
      "epoch": 0.012336118097770589,
      "grad_norm": 1.0935972929000854,
      "learning_rate": 1e-05,
      "loss": 0.2389,
      "step": 3600
    },
    {
      "epoch": 0.012370385092486618,
      "grad_norm": 1.180668830871582,
      "learning_rate": 1e-05,
      "loss": 0.2409,
      "step": 3610
    },
    {
      "epoch": 0.012404652087202648,
      "grad_norm": 1.2312487363815308,
      "learning_rate": 1e-05,
      "loss": 0.2478,
      "step": 3620
    },
    {
      "epoch": 0.012438919081918678,
      "grad_norm": 0.947522759437561,
      "learning_rate": 1e-05,
      "loss": 0.2281,
      "step": 3630
    },
    {
      "epoch": 0.012473186076634706,
      "grad_norm": 1.0618727207183838,
      "learning_rate": 1e-05,
      "loss": 0.2423,
      "step": 3640
    },
    {
      "epoch": 0.012507453071350736,
      "grad_norm": 1.0766098499298096,
      "learning_rate": 1e-05,
      "loss": 0.2364,
      "step": 3650
    },
    {
      "epoch": 0.012541720066066766,
      "grad_norm": 1.1174747943878174,
      "learning_rate": 1e-05,
      "loss": 0.238,
      "step": 3660
    },
    {
      "epoch": 0.012575987060782796,
      "grad_norm": 1.1940118074417114,
      "learning_rate": 1e-05,
      "loss": 0.2212,
      "step": 3670
    },
    {
      "epoch": 0.012610254055498824,
      "grad_norm": 1.1407246589660645,
      "learning_rate": 1e-05,
      "loss": 0.2423,
      "step": 3680
    },
    {
      "epoch": 0.012644521050214854,
      "grad_norm": 1.2646050453186035,
      "learning_rate": 1e-05,
      "loss": 0.2252,
      "step": 3690
    },
    {
      "epoch": 0.012678788044930884,
      "grad_norm": 1.130337119102478,
      "learning_rate": 1e-05,
      "loss": 0.2131,
      "step": 3700
    },
    {
      "epoch": 0.012713055039646913,
      "grad_norm": 1.1432557106018066,
      "learning_rate": 1e-05,
      "loss": 0.2386,
      "step": 3710
    },
    {
      "epoch": 0.012747322034362941,
      "grad_norm": 1.1370545625686646,
      "learning_rate": 1e-05,
      "loss": 0.2347,
      "step": 3720
    },
    {
      "epoch": 0.012781589029078971,
      "grad_norm": 1.3126403093338013,
      "learning_rate": 1e-05,
      "loss": 0.2159,
      "step": 3730
    },
    {
      "epoch": 0.012815856023795001,
      "grad_norm": 1.2375295162200928,
      "learning_rate": 1e-05,
      "loss": 0.2275,
      "step": 3740
    },
    {
      "epoch": 0.012850123018511031,
      "grad_norm": 1.0877372026443481,
      "learning_rate": 1e-05,
      "loss": 0.2201,
      "step": 3750
    },
    {
      "epoch": 0.012884390013227059,
      "grad_norm": 1.1122978925704956,
      "learning_rate": 1e-05,
      "loss": 0.229,
      "step": 3760
    },
    {
      "epoch": 0.012918657007943089,
      "grad_norm": 1.0270159244537354,
      "learning_rate": 1e-05,
      "loss": 0.2313,
      "step": 3770
    },
    {
      "epoch": 0.012952924002659119,
      "grad_norm": 1.1370947360992432,
      "learning_rate": 1e-05,
      "loss": 0.229,
      "step": 3780
    },
    {
      "epoch": 0.012987190997375149,
      "grad_norm": 1.2888813018798828,
      "learning_rate": 1e-05,
      "loss": 0.2384,
      "step": 3790
    },
    {
      "epoch": 0.013021457992091178,
      "grad_norm": 1.2443634271621704,
      "learning_rate": 1e-05,
      "loss": 0.2218,
      "step": 3800
    },
    {
      "epoch": 0.013055724986807207,
      "grad_norm": 1.1919447183609009,
      "learning_rate": 1e-05,
      "loss": 0.2277,
      "step": 3810
    },
    {
      "epoch": 0.013089991981523236,
      "grad_norm": 1.140600562095642,
      "learning_rate": 1e-05,
      "loss": 0.2317,
      "step": 3820
    },
    {
      "epoch": 0.013124258976239266,
      "grad_norm": 1.074697494506836,
      "learning_rate": 1e-05,
      "loss": 0.2273,
      "step": 3830
    },
    {
      "epoch": 0.013158525970955296,
      "grad_norm": 1.1003391742706299,
      "learning_rate": 1e-05,
      "loss": 0.2217,
      "step": 3840
    },
    {
      "epoch": 0.013192792965671324,
      "grad_norm": 1.1427338123321533,
      "learning_rate": 1e-05,
      "loss": 0.2377,
      "step": 3850
    },
    {
      "epoch": 0.013227059960387354,
      "grad_norm": 1.0806514024734497,
      "learning_rate": 1e-05,
      "loss": 0.2332,
      "step": 3860
    },
    {
      "epoch": 0.013261326955103384,
      "grad_norm": 1.1547067165374756,
      "learning_rate": 1e-05,
      "loss": 0.2306,
      "step": 3870
    },
    {
      "epoch": 0.013295593949819414,
      "grad_norm": 1.2483099699020386,
      "learning_rate": 1e-05,
      "loss": 0.2166,
      "step": 3880
    },
    {
      "epoch": 0.013329860944535442,
      "grad_norm": 1.096939206123352,
      "learning_rate": 1e-05,
      "loss": 0.2253,
      "step": 3890
    },
    {
      "epoch": 0.013364127939251472,
      "grad_norm": 1.1876115798950195,
      "learning_rate": 1e-05,
      "loss": 0.2377,
      "step": 3900
    },
    {
      "epoch": 0.013398394933967502,
      "grad_norm": 1.1380902528762817,
      "learning_rate": 1e-05,
      "loss": 0.2256,
      "step": 3910
    },
    {
      "epoch": 0.013432661928683531,
      "grad_norm": 1.0738089084625244,
      "learning_rate": 1e-05,
      "loss": 0.2307,
      "step": 3920
    },
    {
      "epoch": 0.01346692892339956,
      "grad_norm": 1.0351170301437378,
      "learning_rate": 1e-05,
      "loss": 0.2296,
      "step": 3930
    },
    {
      "epoch": 0.01350119591811559,
      "grad_norm": 1.2752678394317627,
      "learning_rate": 1e-05,
      "loss": 0.2462,
      "step": 3940
    },
    {
      "epoch": 0.01353546291283162,
      "grad_norm": 1.2618532180786133,
      "learning_rate": 1e-05,
      "loss": 0.2364,
      "step": 3950
    },
    {
      "epoch": 0.013569729907547649,
      "grad_norm": 1.1907076835632324,
      "learning_rate": 1e-05,
      "loss": 0.2397,
      "step": 3960
    },
    {
      "epoch": 0.013603996902263677,
      "grad_norm": 0.9435076117515564,
      "learning_rate": 1e-05,
      "loss": 0.2391,
      "step": 3970
    },
    {
      "epoch": 0.013638263896979707,
      "grad_norm": 1.0608407258987427,
      "learning_rate": 1e-05,
      "loss": 0.2241,
      "step": 3980
    },
    {
      "epoch": 0.013672530891695737,
      "grad_norm": 1.0729584693908691,
      "learning_rate": 1e-05,
      "loss": 0.2237,
      "step": 3990
    },
    {
      "epoch": 0.013706797886411767,
      "grad_norm": 1.2006182670593262,
      "learning_rate": 1e-05,
      "loss": 0.2386,
      "step": 4000
    },
    {
      "epoch": 0.013706797886411767,
      "eval_cer": 12.594813900158758,
      "eval_loss": 0.25156331062316895,
      "eval_normalized_cer": 8.912869704236611,
      "eval_runtime": 228.7977,
      "eval_samples_per_second": 2.238,
      "eval_steps_per_second": 0.035,
      "step": 4000
    },
    {
      "epoch": 0.013741064881127795,
      "grad_norm": 1.2020457983016968,
      "learning_rate": 1e-05,
      "loss": 0.2318,
      "step": 4010
    },
    {
      "epoch": 0.013775331875843825,
      "grad_norm": 1.0251790285110474,
      "learning_rate": 1e-05,
      "loss": 0.248,
      "step": 4020
    },
    {
      "epoch": 0.013809598870559854,
      "grad_norm": 1.160437822341919,
      "learning_rate": 1e-05,
      "loss": 0.2385,
      "step": 4030
    },
    {
      "epoch": 0.013843865865275884,
      "grad_norm": 1.025770664215088,
      "learning_rate": 1e-05,
      "loss": 0.2293,
      "step": 4040
    },
    {
      "epoch": 0.013878132859991912,
      "grad_norm": 1.111954689025879,
      "learning_rate": 1e-05,
      "loss": 0.2377,
      "step": 4050
    },
    {
      "epoch": 0.013912399854707942,
      "grad_norm": 1.0644809007644653,
      "learning_rate": 1e-05,
      "loss": 0.2195,
      "step": 4060
    },
    {
      "epoch": 0.013946666849423972,
      "grad_norm": 1.2926712036132812,
      "learning_rate": 1e-05,
      "loss": 0.2508,
      "step": 4070
    },
    {
      "epoch": 0.013980933844140002,
      "grad_norm": 1.2169601917266846,
      "learning_rate": 1e-05,
      "loss": 0.2401,
      "step": 4080
    },
    {
      "epoch": 0.01401520083885603,
      "grad_norm": 1.1396681070327759,
      "learning_rate": 1e-05,
      "loss": 0.2305,
      "step": 4090
    },
    {
      "epoch": 0.01404946783357206,
      "grad_norm": 1.2242721319198608,
      "learning_rate": 1e-05,
      "loss": 0.2301,
      "step": 4100
    },
    {
      "epoch": 0.01408373482828809,
      "grad_norm": 1.195324420928955,
      "learning_rate": 1e-05,
      "loss": 0.2368,
      "step": 4110
    },
    {
      "epoch": 0.01411800182300412,
      "grad_norm": 1.2345412969589233,
      "learning_rate": 1e-05,
      "loss": 0.2301,
      "step": 4120
    },
    {
      "epoch": 0.014152268817720148,
      "grad_norm": 1.1502156257629395,
      "learning_rate": 1e-05,
      "loss": 0.2327,
      "step": 4130
    },
    {
      "epoch": 0.014186535812436177,
      "grad_norm": 1.2128121852874756,
      "learning_rate": 1e-05,
      "loss": 0.2458,
      "step": 4140
    },
    {
      "epoch": 0.014220802807152207,
      "grad_norm": 1.2618858814239502,
      "learning_rate": 1e-05,
      "loss": 0.231,
      "step": 4150
    },
    {
      "epoch": 0.014255069801868237,
      "grad_norm": 1.0879299640655518,
      "learning_rate": 1e-05,
      "loss": 0.2302,
      "step": 4160
    },
    {
      "epoch": 0.014289336796584265,
      "grad_norm": 0.9794358015060425,
      "learning_rate": 1e-05,
      "loss": 0.239,
      "step": 4170
    },
    {
      "epoch": 0.014323603791300295,
      "grad_norm": 1.1454006433486938,
      "learning_rate": 1e-05,
      "loss": 0.2328,
      "step": 4180
    },
    {
      "epoch": 0.014357870786016325,
      "grad_norm": 1.223686933517456,
      "learning_rate": 1e-05,
      "loss": 0.2211,
      "step": 4190
    },
    {
      "epoch": 0.014392137780732355,
      "grad_norm": 1.1423155069351196,
      "learning_rate": 1e-05,
      "loss": 0.2391,
      "step": 4200
    },
    {
      "epoch": 0.014426404775448383,
      "grad_norm": 1.1027394533157349,
      "learning_rate": 1e-05,
      "loss": 0.2279,
      "step": 4210
    },
    {
      "epoch": 0.014460671770164413,
      "grad_norm": 1.1777397394180298,
      "learning_rate": 1e-05,
      "loss": 0.2293,
      "step": 4220
    },
    {
      "epoch": 0.014494938764880443,
      "grad_norm": 1.01688551902771,
      "learning_rate": 1e-05,
      "loss": 0.2275,
      "step": 4230
    },
    {
      "epoch": 0.014529205759596472,
      "grad_norm": 1.1520488262176514,
      "learning_rate": 1e-05,
      "loss": 0.2301,
      "step": 4240
    },
    {
      "epoch": 0.0145634727543125,
      "grad_norm": 1.2820484638214111,
      "learning_rate": 1e-05,
      "loss": 0.2205,
      "step": 4250
    },
    {
      "epoch": 0.01459773974902853,
      "grad_norm": 1.169291377067566,
      "learning_rate": 1e-05,
      "loss": 0.2389,
      "step": 4260
    },
    {
      "epoch": 0.01463200674374456,
      "grad_norm": 1.1135886907577515,
      "learning_rate": 1e-05,
      "loss": 0.2384,
      "step": 4270
    },
    {
      "epoch": 0.01466627373846059,
      "grad_norm": 1.0846205949783325,
      "learning_rate": 1e-05,
      "loss": 0.223,
      "step": 4280
    },
    {
      "epoch": 0.014700540733176618,
      "grad_norm": 0.981488049030304,
      "learning_rate": 1e-05,
      "loss": 0.2092,
      "step": 4290
    },
    {
      "epoch": 0.014734807727892648,
      "grad_norm": 1.0437407493591309,
      "learning_rate": 1e-05,
      "loss": 0.2293,
      "step": 4300
    },
    {
      "epoch": 0.014769074722608678,
      "grad_norm": 1.005792260169983,
      "learning_rate": 1e-05,
      "loss": 0.2286,
      "step": 4310
    },
    {
      "epoch": 0.014803341717324708,
      "grad_norm": 1.1903142929077148,
      "learning_rate": 1e-05,
      "loss": 0.231,
      "step": 4320
    },
    {
      "epoch": 0.014837608712040736,
      "grad_norm": 1.1308993101119995,
      "learning_rate": 1e-05,
      "loss": 0.2458,
      "step": 4330
    },
    {
      "epoch": 0.014871875706756766,
      "grad_norm": 1.0948210954666138,
      "learning_rate": 1e-05,
      "loss": 0.213,
      "step": 4340
    },
    {
      "epoch": 0.014906142701472795,
      "grad_norm": 1.2674663066864014,
      "learning_rate": 1e-05,
      "loss": 0.2432,
      "step": 4350
    },
    {
      "epoch": 0.014940409696188825,
      "grad_norm": 1.4228485822677612,
      "learning_rate": 1e-05,
      "loss": 0.2491,
      "step": 4360
    },
    {
      "epoch": 0.014974676690904853,
      "grad_norm": 1.1533160209655762,
      "learning_rate": 1e-05,
      "loss": 0.2485,
      "step": 4370
    },
    {
      "epoch": 0.015008943685620883,
      "grad_norm": 1.1454424858093262,
      "learning_rate": 1e-05,
      "loss": 0.2635,
      "step": 4380
    },
    {
      "epoch": 0.015043210680336913,
      "grad_norm": 1.2944281101226807,
      "learning_rate": 1e-05,
      "loss": 0.2651,
      "step": 4390
    },
    {
      "epoch": 0.015077477675052943,
      "grad_norm": 1.2148584127426147,
      "learning_rate": 1e-05,
      "loss": 0.2694,
      "step": 4400
    },
    {
      "epoch": 0.015111744669768971,
      "grad_norm": 1.091282844543457,
      "learning_rate": 1e-05,
      "loss": 0.2672,
      "step": 4410
    },
    {
      "epoch": 0.015146011664485001,
      "grad_norm": 1.2254445552825928,
      "learning_rate": 1e-05,
      "loss": 0.2583,
      "step": 4420
    },
    {
      "epoch": 0.01518027865920103,
      "grad_norm": 1.367516279220581,
      "learning_rate": 1e-05,
      "loss": 0.2586,
      "step": 4430
    },
    {
      "epoch": 0.01521454565391706,
      "grad_norm": 1.1858383417129517,
      "learning_rate": 1e-05,
      "loss": 0.2764,
      "step": 4440
    },
    {
      "epoch": 0.01524881264863309,
      "grad_norm": 1.1331857442855835,
      "learning_rate": 1e-05,
      "loss": 0.2577,
      "step": 4450
    },
    {
      "epoch": 0.015283079643349119,
      "grad_norm": 1.2343239784240723,
      "learning_rate": 1e-05,
      "loss": 0.2661,
      "step": 4460
    },
    {
      "epoch": 0.015317346638065148,
      "grad_norm": 1.0893656015396118,
      "learning_rate": 1e-05,
      "loss": 0.2538,
      "step": 4470
    },
    {
      "epoch": 0.015351613632781178,
      "grad_norm": 1.1467857360839844,
      "learning_rate": 1e-05,
      "loss": 0.2496,
      "step": 4480
    },
    {
      "epoch": 0.015385880627497208,
      "grad_norm": 1.2753335237503052,
      "learning_rate": 1e-05,
      "loss": 0.2797,
      "step": 4490
    },
    {
      "epoch": 0.015420147622213236,
      "grad_norm": 1.1355762481689453,
      "learning_rate": 1e-05,
      "loss": 0.2672,
      "step": 4500
    },
    {
      "epoch": 0.015420147622213236,
      "eval_cer": 13.159287352266713,
      "eval_loss": 0.24996142089366913,
      "eval_normalized_cer": 9.59232613908873,
      "eval_runtime": 228.0477,
      "eval_samples_per_second": 2.245,
      "eval_steps_per_second": 0.035,
      "step": 4500
    },
    {
      "epoch": 0.015454414616929266,
      "grad_norm": 1.2256762981414795,
      "learning_rate": 1e-05,
      "loss": 0.2662,
      "step": 4510
    },
    {
      "epoch": 0.015488681611645296,
      "grad_norm": 1.0631389617919922,
      "learning_rate": 1e-05,
      "loss": 0.2596,
      "step": 4520
    },
    {
      "epoch": 0.015522948606361326,
      "grad_norm": 1.0759390592575073,
      "learning_rate": 1e-05,
      "loss": 0.2553,
      "step": 4530
    },
    {
      "epoch": 0.015557215601077354,
      "grad_norm": 1.1867231130599976,
      "learning_rate": 1e-05,
      "loss": 0.2498,
      "step": 4540
    },
    {
      "epoch": 0.015591482595793384,
      "grad_norm": 1.1203633546829224,
      "learning_rate": 1e-05,
      "loss": 0.2732,
      "step": 4550
    },
    {
      "epoch": 0.015625749590509413,
      "grad_norm": 1.1223920583724976,
      "learning_rate": 1e-05,
      "loss": 0.2535,
      "step": 4560
    },
    {
      "epoch": 0.015660016585225443,
      "grad_norm": 1.066497564315796,
      "learning_rate": 1e-05,
      "loss": 0.2456,
      "step": 4570
    },
    {
      "epoch": 0.015694283579941473,
      "grad_norm": 1.2520133256912231,
      "learning_rate": 1e-05,
      "loss": 0.2558,
      "step": 4580
    },
    {
      "epoch": 0.015728550574657503,
      "grad_norm": 1.3602423667907715,
      "learning_rate": 1e-05,
      "loss": 0.2698,
      "step": 4590
    },
    {
      "epoch": 0.01576281756937353,
      "grad_norm": 1.1748729944229126,
      "learning_rate": 1e-05,
      "loss": 0.2621,
      "step": 4600
    },
    {
      "epoch": 0.01579708456408956,
      "grad_norm": 0.9431802034378052,
      "learning_rate": 1e-05,
      "loss": 0.2433,
      "step": 4610
    },
    {
      "epoch": 0.01583135155880559,
      "grad_norm": 1.0146753787994385,
      "learning_rate": 1e-05,
      "loss": 0.239,
      "step": 4620
    },
    {
      "epoch": 0.01586561855352162,
      "grad_norm": 1.1340891122817993,
      "learning_rate": 1e-05,
      "loss": 0.2437,
      "step": 4630
    },
    {
      "epoch": 0.01589988554823765,
      "grad_norm": 1.1456454992294312,
      "learning_rate": 1e-05,
      "loss": 0.2307,
      "step": 4640
    },
    {
      "epoch": 0.01593415254295368,
      "grad_norm": 1.1026827096939087,
      "learning_rate": 1e-05,
      "loss": 0.2295,
      "step": 4650
    },
    {
      "epoch": 0.01596841953766971,
      "grad_norm": 1.2215088605880737,
      "learning_rate": 1e-05,
      "loss": 0.245,
      "step": 4660
    },
    {
      "epoch": 0.01600268653238574,
      "grad_norm": 1.1760615110397339,
      "learning_rate": 1e-05,
      "loss": 0.2461,
      "step": 4670
    },
    {
      "epoch": 0.016036953527101765,
      "grad_norm": 1.1690876483917236,
      "learning_rate": 1e-05,
      "loss": 0.2282,
      "step": 4680
    },
    {
      "epoch": 0.016071220521817794,
      "grad_norm": 1.182026743888855,
      "learning_rate": 1e-05,
      "loss": 0.2351,
      "step": 4690
    },
    {
      "epoch": 0.016105487516533824,
      "grad_norm": 1.0182474851608276,
      "learning_rate": 1e-05,
      "loss": 0.2284,
      "step": 4700
    },
    {
      "epoch": 0.016139754511249854,
      "grad_norm": 1.2531431913375854,
      "learning_rate": 1e-05,
      "loss": 0.244,
      "step": 4710
    },
    {
      "epoch": 0.016174021505965884,
      "grad_norm": 0.9633692502975464,
      "learning_rate": 1e-05,
      "loss": 0.2297,
      "step": 4720
    },
    {
      "epoch": 0.016208288500681914,
      "grad_norm": 1.1144667863845825,
      "learning_rate": 1e-05,
      "loss": 0.2475,
      "step": 4730
    },
    {
      "epoch": 0.016242555495397944,
      "grad_norm": 1.0768555402755737,
      "learning_rate": 1e-05,
      "loss": 0.2216,
      "step": 4740
    },
    {
      "epoch": 0.016276822490113974,
      "grad_norm": 1.2052035331726074,
      "learning_rate": 1e-05,
      "loss": 0.2278,
      "step": 4750
    },
    {
      "epoch": 0.01631108948483,
      "grad_norm": 1.0291496515274048,
      "learning_rate": 1e-05,
      "loss": 0.2226,
      "step": 4760
    },
    {
      "epoch": 0.01634535647954603,
      "grad_norm": 1.2100346088409424,
      "learning_rate": 1e-05,
      "loss": 0.2278,
      "step": 4770
    },
    {
      "epoch": 0.01637962347426206,
      "grad_norm": 1.214861273765564,
      "learning_rate": 1e-05,
      "loss": 0.2313,
      "step": 4780
    },
    {
      "epoch": 0.01641389046897809,
      "grad_norm": 1.137210726737976,
      "learning_rate": 1e-05,
      "loss": 0.2235,
      "step": 4790
    },
    {
      "epoch": 0.01644815746369412,
      "grad_norm": 1.046673059463501,
      "learning_rate": 1e-05,
      "loss": 0.2231,
      "step": 4800
    },
    {
      "epoch": 0.01648242445841015,
      "grad_norm": 1.08164644241333,
      "learning_rate": 1e-05,
      "loss": 0.2235,
      "step": 4810
    },
    {
      "epoch": 0.01651669145312618,
      "grad_norm": 1.1432491540908813,
      "learning_rate": 1e-05,
      "loss": 0.246,
      "step": 4820
    },
    {
      "epoch": 0.01655095844784221,
      "grad_norm": 1.1684173345565796,
      "learning_rate": 1e-05,
      "loss": 0.218,
      "step": 4830
    },
    {
      "epoch": 0.016585225442558235,
      "grad_norm": 1.0895615816116333,
      "learning_rate": 1e-05,
      "loss": 0.2109,
      "step": 4840
    },
    {
      "epoch": 0.016619492437274265,
      "grad_norm": 1.1505770683288574,
      "learning_rate": 1e-05,
      "loss": 0.2283,
      "step": 4850
    },
    {
      "epoch": 0.016653759431990295,
      "grad_norm": 1.3385730981826782,
      "learning_rate": 1e-05,
      "loss": 0.2344,
      "step": 4860
    },
    {
      "epoch": 0.016688026426706325,
      "grad_norm": 1.109035611152649,
      "learning_rate": 1e-05,
      "loss": 0.2558,
      "step": 4870
    },
    {
      "epoch": 0.016722293421422355,
      "grad_norm": 1.1834880113601685,
      "learning_rate": 1e-05,
      "loss": 0.2247,
      "step": 4880
    },
    {
      "epoch": 0.016756560416138384,
      "grad_norm": 1.2369152307510376,
      "learning_rate": 1e-05,
      "loss": 0.2449,
      "step": 4890
    },
    {
      "epoch": 0.016790827410854414,
      "grad_norm": 1.131173014640808,
      "learning_rate": 1e-05,
      "loss": 0.2458,
      "step": 4900
    },
    {
      "epoch": 0.016825094405570444,
      "grad_norm": 1.1100351810455322,
      "learning_rate": 1e-05,
      "loss": 0.2523,
      "step": 4910
    },
    {
      "epoch": 0.01685936140028647,
      "grad_norm": 1.1857340335845947,
      "learning_rate": 1e-05,
      "loss": 0.2523,
      "step": 4920
    },
    {
      "epoch": 0.0168936283950025,
      "grad_norm": 1.1568819284439087,
      "learning_rate": 1e-05,
      "loss": 0.2549,
      "step": 4930
    },
    {
      "epoch": 0.01692789538971853,
      "grad_norm": 1.104872465133667,
      "learning_rate": 1e-05,
      "loss": 0.2449,
      "step": 4940
    },
    {
      "epoch": 0.01696216238443456,
      "grad_norm": 1.0907660722732544,
      "learning_rate": 1e-05,
      "loss": 0.2496,
      "step": 4950
    },
    {
      "epoch": 0.01699642937915059,
      "grad_norm": 1.1100903749465942,
      "learning_rate": 1e-05,
      "loss": 0.239,
      "step": 4960
    },
    {
      "epoch": 0.01703069637386662,
      "grad_norm": 1.141200065612793,
      "learning_rate": 1e-05,
      "loss": 0.2459,
      "step": 4970
    },
    {
      "epoch": 0.01706496336858265,
      "grad_norm": 1.2853361368179321,
      "learning_rate": 1e-05,
      "loss": 0.2452,
      "step": 4980
    },
    {
      "epoch": 0.01709923036329868,
      "grad_norm": 1.1542645692825317,
      "learning_rate": 1e-05,
      "loss": 0.2635,
      "step": 4990
    },
    {
      "epoch": 0.017133497358014706,
      "grad_norm": 1.2022640705108643,
      "learning_rate": 1e-05,
      "loss": 0.2371,
      "step": 5000
    },
    {
      "epoch": 0.017133497358014706,
      "eval_cer": 12.92115011465867,
      "eval_loss": 0.2521001100540161,
      "eval_normalized_cer": 9.30255795363709,
      "eval_runtime": 227.4868,
      "eval_samples_per_second": 2.251,
      "eval_steps_per_second": 0.035,
      "step": 5000
    },
    {
      "epoch": 0.017167764352730736,
      "grad_norm": 1.0765001773834229,
      "learning_rate": 1e-05,
      "loss": 0.2455,
      "step": 5010
    },
    {
      "epoch": 0.017202031347446765,
      "grad_norm": 1.0711493492126465,
      "learning_rate": 1e-05,
      "loss": 0.2422,
      "step": 5020
    },
    {
      "epoch": 0.017236298342162795,
      "grad_norm": 1.0719484090805054,
      "learning_rate": 1e-05,
      "loss": 0.2531,
      "step": 5030
    },
    {
      "epoch": 0.017270565336878825,
      "grad_norm": 1.1884721517562866,
      "learning_rate": 1e-05,
      "loss": 0.2508,
      "step": 5040
    },
    {
      "epoch": 0.017304832331594855,
      "grad_norm": 1.068827509880066,
      "learning_rate": 1e-05,
      "loss": 0.2474,
      "step": 5050
    },
    {
      "epoch": 0.017339099326310885,
      "grad_norm": 1.1308655738830566,
      "learning_rate": 1e-05,
      "loss": 0.2627,
      "step": 5060
    },
    {
      "epoch": 0.017373366321026915,
      "grad_norm": 1.1527314186096191,
      "learning_rate": 1e-05,
      "loss": 0.2535,
      "step": 5070
    },
    {
      "epoch": 0.017407633315742944,
      "grad_norm": 1.1800657510757446,
      "learning_rate": 1e-05,
      "loss": 0.2587,
      "step": 5080
    },
    {
      "epoch": 0.01744190031045897,
      "grad_norm": 1.095189094543457,
      "learning_rate": 1e-05,
      "loss": 0.2424,
      "step": 5090
    },
    {
      "epoch": 0.017476167305175,
      "grad_norm": 1.109617829322815,
      "learning_rate": 1e-05,
      "loss": 0.2543,
      "step": 5100
    },
    {
      "epoch": 0.01751043429989103,
      "grad_norm": 1.2110544443130493,
      "learning_rate": 1e-05,
      "loss": 0.2687,
      "step": 5110
    },
    {
      "epoch": 0.01754470129460706,
      "grad_norm": 1.0466723442077637,
      "learning_rate": 1e-05,
      "loss": 0.2424,
      "step": 5120
    },
    {
      "epoch": 0.01757896828932309,
      "grad_norm": 1.2060648202896118,
      "learning_rate": 1e-05,
      "loss": 0.2337,
      "step": 5130
    },
    {
      "epoch": 0.01761323528403912,
      "grad_norm": 1.203142762184143,
      "learning_rate": 1e-05,
      "loss": 0.2556,
      "step": 5140
    },
    {
      "epoch": 0.01764750227875515,
      "grad_norm": 1.0751283168792725,
      "learning_rate": 1e-05,
      "loss": 0.2235,
      "step": 5150
    },
    {
      "epoch": 0.01768176927347118,
      "grad_norm": 1.1377781629562378,
      "learning_rate": 1e-05,
      "loss": 0.2448,
      "step": 5160
    },
    {
      "epoch": 0.017716036268187206,
      "grad_norm": 1.147454023361206,
      "learning_rate": 1e-05,
      "loss": 0.2172,
      "step": 5170
    },
    {
      "epoch": 0.017750303262903236,
      "grad_norm": 1.129897952079773,
      "learning_rate": 1e-05,
      "loss": 0.2418,
      "step": 5180
    },
    {
      "epoch": 0.017784570257619266,
      "grad_norm": 1.1261131763458252,
      "learning_rate": 1e-05,
      "loss": 0.2328,
      "step": 5190
    },
    {
      "epoch": 0.017818837252335296,
      "grad_norm": 1.0794824361801147,
      "learning_rate": 1e-05,
      "loss": 0.2546,
      "step": 5200
    },
    {
      "epoch": 0.017853104247051325,
      "grad_norm": 1.1870142221450806,
      "learning_rate": 1e-05,
      "loss": 0.249,
      "step": 5210
    },
    {
      "epoch": 0.017887371241767355,
      "grad_norm": 1.0414400100708008,
      "learning_rate": 1e-05,
      "loss": 0.2285,
      "step": 5220
    },
    {
      "epoch": 0.017921638236483385,
      "grad_norm": 1.173405647277832,
      "learning_rate": 1e-05,
      "loss": 0.2529,
      "step": 5230
    },
    {
      "epoch": 0.017955905231199415,
      "grad_norm": 1.039650797843933,
      "learning_rate": 1e-05,
      "loss": 0.2321,
      "step": 5240
    },
    {
      "epoch": 0.01799017222591544,
      "grad_norm": 1.0359266996383667,
      "learning_rate": 1e-05,
      "loss": 0.2433,
      "step": 5250
    },
    {
      "epoch": 0.01802443922063147,
      "grad_norm": 1.0630840063095093,
      "learning_rate": 1e-05,
      "loss": 0.2117,
      "step": 5260
    },
    {
      "epoch": 0.0180587062153475,
      "grad_norm": 1.0937180519104004,
      "learning_rate": 1e-05,
      "loss": 0.2454,
      "step": 5270
    },
    {
      "epoch": 0.01809297321006353,
      "grad_norm": 1.1015993356704712,
      "learning_rate": 1e-05,
      "loss": 0.238,
      "step": 5280
    },
    {
      "epoch": 0.01812724020477956,
      "grad_norm": 1.060584545135498,
      "learning_rate": 1e-05,
      "loss": 0.2475,
      "step": 5290
    },
    {
      "epoch": 0.01816150719949559,
      "grad_norm": 1.1389795541763306,
      "learning_rate": 1e-05,
      "loss": 0.233,
      "step": 5300
    },
    {
      "epoch": 0.01819577419421162,
      "grad_norm": 1.0018917322158813,
      "learning_rate": 1e-05,
      "loss": 0.2453,
      "step": 5310
    },
    {
      "epoch": 0.01823004118892765,
      "grad_norm": 1.0546092987060547,
      "learning_rate": 1e-05,
      "loss": 0.2333,
      "step": 5320
    },
    {
      "epoch": 0.018264308183643677,
      "grad_norm": 1.1121848821640015,
      "learning_rate": 1e-05,
      "loss": 0.2317,
      "step": 5330
    },
    {
      "epoch": 0.018298575178359706,
      "grad_norm": 1.1613191366195679,
      "learning_rate": 1e-05,
      "loss": 0.2549,
      "step": 5340
    },
    {
      "epoch": 0.018332842173075736,
      "grad_norm": 1.1250524520874023,
      "learning_rate": 1e-05,
      "loss": 0.2471,
      "step": 5350
    },
    {
      "epoch": 0.018367109167791766,
      "grad_norm": 1.0905226469039917,
      "learning_rate": 1e-05,
      "loss": 0.229,
      "step": 5360
    },
    {
      "epoch": 0.018401376162507796,
      "grad_norm": 0.9885173439979553,
      "learning_rate": 1e-05,
      "loss": 0.2542,
      "step": 5370
    },
    {
      "epoch": 0.018435643157223826,
      "grad_norm": 1.288758635520935,
      "learning_rate": 1e-05,
      "loss": 0.2472,
      "step": 5380
    },
    {
      "epoch": 0.018469910151939856,
      "grad_norm": 1.2433462142944336,
      "learning_rate": 1e-05,
      "loss": 0.2427,
      "step": 5390
    },
    {
      "epoch": 0.018504177146655885,
      "grad_norm": 1.2367336750030518,
      "learning_rate": 1e-05,
      "loss": 0.2511,
      "step": 5400
    },
    {
      "epoch": 0.018538444141371912,
      "grad_norm": 1.1871395111083984,
      "learning_rate": 1e-05,
      "loss": 0.2276,
      "step": 5410
    },
    {
      "epoch": 0.01857271113608794,
      "grad_norm": 0.9569379091262817,
      "learning_rate": 1e-05,
      "loss": 0.2475,
      "step": 5420
    },
    {
      "epoch": 0.01860697813080397,
      "grad_norm": 1.1487014293670654,
      "learning_rate": 1e-05,
      "loss": 0.2295,
      "step": 5430
    },
    {
      "epoch": 0.01864124512552,
      "grad_norm": 1.0800844430923462,
      "learning_rate": 1e-05,
      "loss": 0.2247,
      "step": 5440
    },
    {
      "epoch": 0.01867551212023603,
      "grad_norm": 1.1834380626678467,
      "learning_rate": 1e-05,
      "loss": 0.226,
      "step": 5450
    },
    {
      "epoch": 0.01870977911495206,
      "grad_norm": 1.0035191774368286,
      "learning_rate": 1e-05,
      "loss": 0.2414,
      "step": 5460
    },
    {
      "epoch": 0.01874404610966809,
      "grad_norm": 1.0685466527938843,
      "learning_rate": 1e-05,
      "loss": 0.2449,
      "step": 5470
    },
    {
      "epoch": 0.01877831310438412,
      "grad_norm": 1.1921565532684326,
      "learning_rate": 1e-05,
      "loss": 0.2419,
      "step": 5480
    },
    {
      "epoch": 0.018812580099100147,
      "grad_norm": 1.1201281547546387,
      "learning_rate": 1e-05,
      "loss": 0.255,
      "step": 5490
    },
    {
      "epoch": 0.018846847093816177,
      "grad_norm": 1.1162866353988647,
      "learning_rate": 1e-05,
      "loss": 0.2426,
      "step": 5500
    },
    {
      "epoch": 0.018846847093816177,
      "eval_cer": 13.238666431469396,
      "eval_loss": 0.25262224674224854,
      "eval_normalized_cer": 9.562350119904076,
      "eval_runtime": 229.0802,
      "eval_samples_per_second": 2.235,
      "eval_steps_per_second": 0.035,
      "step": 5500
    }
  ],
  "logging_steps": 10,
  "max_steps": 291826,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 9223372036854775807,
  "save_steps": 500,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.11517466886144e+21,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": null
}