{
  "best_metric": 8.593125499600319,
  "best_model_checkpoint": "kotoba_v2_enc_logs_epoch2_2/checkpoint-20500",
  "epoch": 0.08395413705427207,
  "eval_steps": 500,
  "global_step": 24500,
  "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
    },
    {
      "epoch": 0.018881114088532207,
      "grad_norm": 1.0215845108032227,
      "learning_rate": 1e-05,
      "loss": 0.2368,
      "step": 5510
    },
    {
      "epoch": 0.018915381083248237,
      "grad_norm": 1.0062447786331177,
      "learning_rate": 1e-05,
      "loss": 0.2308,
      "step": 5520
    },
    {
      "epoch": 0.018949648077964266,
      "grad_norm": 1.223649024963379,
      "learning_rate": 1e-05,
      "loss": 0.2409,
      "step": 5530
    },
    {
      "epoch": 0.018983915072680296,
      "grad_norm": 1.2076172828674316,
      "learning_rate": 1e-05,
      "loss": 0.2236,
      "step": 5540
    },
    {
      "epoch": 0.019018182067396326,
      "grad_norm": 1.154416561126709,
      "learning_rate": 1e-05,
      "loss": 0.2419,
      "step": 5550
    },
    {
      "epoch": 0.019052449062112356,
      "grad_norm": 1.284858226776123,
      "learning_rate": 1e-05,
      "loss": 0.2321,
      "step": 5560
    },
    {
      "epoch": 0.019086716056828382,
      "grad_norm": 1.0406948328018188,
      "learning_rate": 1e-05,
      "loss": 0.2485,
      "step": 5570
    },
    {
      "epoch": 0.019120983051544412,
      "grad_norm": 1.1980571746826172,
      "learning_rate": 1e-05,
      "loss": 0.2274,
      "step": 5580
    },
    {
      "epoch": 0.019155250046260442,
      "grad_norm": 1.073560357093811,
      "learning_rate": 1e-05,
      "loss": 0.2498,
      "step": 5590
    },
    {
      "epoch": 0.019189517040976472,
      "grad_norm": 1.0982617139816284,
      "learning_rate": 1e-05,
      "loss": 0.2391,
      "step": 5600
    },
    {
      "epoch": 0.019223784035692502,
      "grad_norm": 1.015085220336914,
      "learning_rate": 1e-05,
      "loss": 0.2269,
      "step": 5610
    },
    {
      "epoch": 0.01925805103040853,
      "grad_norm": 1.238585352897644,
      "learning_rate": 1e-05,
      "loss": 0.2428,
      "step": 5620
    },
    {
      "epoch": 0.01929231802512456,
      "grad_norm": 1.3326079845428467,
      "learning_rate": 1e-05,
      "loss": 0.25,
      "step": 5630
    },
    {
      "epoch": 0.01932658501984059,
      "grad_norm": 1.1263608932495117,
      "learning_rate": 1e-05,
      "loss": 0.234,
      "step": 5640
    },
    {
      "epoch": 0.019360852014556618,
      "grad_norm": 1.083595633506775,
      "learning_rate": 1e-05,
      "loss": 0.2504,
      "step": 5650
    },
    {
      "epoch": 0.019395119009272647,
      "grad_norm": 1.0787022113800049,
      "learning_rate": 1e-05,
      "loss": 0.2248,
      "step": 5660
    },
    {
      "epoch": 0.019429386003988677,
      "grad_norm": 1.312565803527832,
      "learning_rate": 1e-05,
      "loss": 0.263,
      "step": 5670
    },
    {
      "epoch": 0.019463652998704707,
      "grad_norm": 1.0305407047271729,
      "learning_rate": 1e-05,
      "loss": 0.2358,
      "step": 5680
    },
    {
      "epoch": 0.019497919993420737,
      "grad_norm": 1.0905306339263916,
      "learning_rate": 1e-05,
      "loss": 0.2358,
      "step": 5690
    },
    {
      "epoch": 0.019532186988136767,
      "grad_norm": 1.1105730533599854,
      "learning_rate": 1e-05,
      "loss": 0.2371,
      "step": 5700
    },
    {
      "epoch": 0.019566453982852797,
      "grad_norm": 1.1664555072784424,
      "learning_rate": 1e-05,
      "loss": 0.244,
      "step": 5710
    },
    {
      "epoch": 0.019600720977568827,
      "grad_norm": 1.0702719688415527,
      "learning_rate": 1e-05,
      "loss": 0.2305,
      "step": 5720
    },
    {
      "epoch": 0.019634987972284856,
      "grad_norm": 1.0736626386642456,
      "learning_rate": 1e-05,
      "loss": 0.2406,
      "step": 5730
    },
    {
      "epoch": 0.019669254967000883,
      "grad_norm": 1.0510461330413818,
      "learning_rate": 1e-05,
      "loss": 0.2335,
      "step": 5740
    },
    {
      "epoch": 0.019703521961716913,
      "grad_norm": 1.0435370206832886,
      "learning_rate": 1e-05,
      "loss": 0.2211,
      "step": 5750
    },
    {
      "epoch": 0.019737788956432942,
      "grad_norm": 1.2461049556732178,
      "learning_rate": 1e-05,
      "loss": 0.2188,
      "step": 5760
    },
    {
      "epoch": 0.019772055951148972,
      "grad_norm": 1.0351046323776245,
      "learning_rate": 1e-05,
      "loss": 0.2269,
      "step": 5770
    },
    {
      "epoch": 0.019806322945865002,
      "grad_norm": 1.124671459197998,
      "learning_rate": 1e-05,
      "loss": 0.2284,
      "step": 5780
    },
    {
      "epoch": 0.019840589940581032,
      "grad_norm": 1.145488977432251,
      "learning_rate": 1e-05,
      "loss": 0.2415,
      "step": 5790
    },
    {
      "epoch": 0.019874856935297062,
      "grad_norm": 1.1410046815872192,
      "learning_rate": 1e-05,
      "loss": 0.2296,
      "step": 5800
    },
    {
      "epoch": 0.01990912393001309,
      "grad_norm": 1.2782517671585083,
      "learning_rate": 1e-05,
      "loss": 0.2367,
      "step": 5810
    },
    {
      "epoch": 0.019943390924729118,
      "grad_norm": 1.204562783241272,
      "learning_rate": 1e-05,
      "loss": 0.2289,
      "step": 5820
    },
    {
      "epoch": 0.019977657919445148,
      "grad_norm": 1.1141811609268188,
      "learning_rate": 1e-05,
      "loss": 0.2223,
      "step": 5830
    },
    {
      "epoch": 0.020011924914161178,
      "grad_norm": 1.1790316104888916,
      "learning_rate": 1e-05,
      "loss": 0.2308,
      "step": 5840
    },
    {
      "epoch": 0.020046191908877208,
      "grad_norm": 1.0944266319274902,
      "learning_rate": 1e-05,
      "loss": 0.2366,
      "step": 5850
    },
    {
      "epoch": 0.020080458903593237,
      "grad_norm": 1.0892263650894165,
      "learning_rate": 1e-05,
      "loss": 0.2384,
      "step": 5860
    },
    {
      "epoch": 0.020114725898309267,
      "grad_norm": 1.1419873237609863,
      "learning_rate": 1e-05,
      "loss": 0.2414,
      "step": 5870
    },
    {
      "epoch": 0.020148992893025297,
      "grad_norm": 1.2230783700942993,
      "learning_rate": 1e-05,
      "loss": 0.2394,
      "step": 5880
    },
    {
      "epoch": 0.020183259887741327,
      "grad_norm": 1.1309173107147217,
      "learning_rate": 1e-05,
      "loss": 0.2561,
      "step": 5890
    },
    {
      "epoch": 0.020217526882457353,
      "grad_norm": 1.2405802011489868,
      "learning_rate": 1e-05,
      "loss": 0.259,
      "step": 5900
    },
    {
      "epoch": 0.020251793877173383,
      "grad_norm": 1.2853388786315918,
      "learning_rate": 1e-05,
      "loss": 0.2668,
      "step": 5910
    },
    {
      "epoch": 0.020286060871889413,
      "grad_norm": 1.299046277999878,
      "learning_rate": 1e-05,
      "loss": 0.251,
      "step": 5920
    },
    {
      "epoch": 0.020320327866605443,
      "grad_norm": 1.142052173614502,
      "learning_rate": 1e-05,
      "loss": 0.2655,
      "step": 5930
    },
    {
      "epoch": 0.020354594861321473,
      "grad_norm": 1.3770766258239746,
      "learning_rate": 1e-05,
      "loss": 0.2508,
      "step": 5940
    },
    {
      "epoch": 0.020388861856037502,
      "grad_norm": 1.1458237171173096,
      "learning_rate": 1e-05,
      "loss": 0.2742,
      "step": 5950
    },
    {
      "epoch": 0.020423128850753532,
      "grad_norm": 1.3130786418914795,
      "learning_rate": 1e-05,
      "loss": 0.2514,
      "step": 5960
    },
    {
      "epoch": 0.020457395845469562,
      "grad_norm": 1.2816088199615479,
      "learning_rate": 1e-05,
      "loss": 0.2593,
      "step": 5970
    },
    {
      "epoch": 0.02049166284018559,
      "grad_norm": 1.0405460596084595,
      "learning_rate": 1e-05,
      "loss": 0.2608,
      "step": 5980
    },
    {
      "epoch": 0.02052592983490162,
      "grad_norm": 1.2035329341888428,
      "learning_rate": 1e-05,
      "loss": 0.2558,
      "step": 5990
    },
    {
      "epoch": 0.020560196829617648,
      "grad_norm": 1.0495450496673584,
      "learning_rate": 1e-05,
      "loss": 0.2468,
      "step": 6000
    },
    {
      "epoch": 0.020560196829617648,
      "eval_cer": 13.079908273064033,
      "eval_loss": 0.2540421485900879,
      "eval_normalized_cer": 9.292565947242206,
      "eval_runtime": 227.4153,
      "eval_samples_per_second": 2.251,
      "eval_steps_per_second": 0.035,
      "step": 6000
    },
    {
      "epoch": 0.020594463824333678,
      "grad_norm": 1.1614056825637817,
      "learning_rate": 1e-05,
      "loss": 0.2527,
      "step": 6010
    },
    {
      "epoch": 0.020628730819049708,
      "grad_norm": 1.1835705041885376,
      "learning_rate": 1e-05,
      "loss": 0.2592,
      "step": 6020
    },
    {
      "epoch": 0.020662997813765738,
      "grad_norm": 1.1335136890411377,
      "learning_rate": 1e-05,
      "loss": 0.2727,
      "step": 6030
    },
    {
      "epoch": 0.020697264808481768,
      "grad_norm": 1.052079439163208,
      "learning_rate": 1e-05,
      "loss": 0.2514,
      "step": 6040
    },
    {
      "epoch": 0.020731531803197797,
      "grad_norm": 1.096330165863037,
      "learning_rate": 1e-05,
      "loss": 0.2684,
      "step": 6050
    },
    {
      "epoch": 0.020765798797913824,
      "grad_norm": 1.2359880208969116,
      "learning_rate": 1e-05,
      "loss": 0.2638,
      "step": 6060
    },
    {
      "epoch": 0.020800065792629854,
      "grad_norm": 1.2259430885314941,
      "learning_rate": 1e-05,
      "loss": 0.2488,
      "step": 6070
    },
    {
      "epoch": 0.020834332787345883,
      "grad_norm": 1.0531619787216187,
      "learning_rate": 1e-05,
      "loss": 0.2584,
      "step": 6080
    },
    {
      "epoch": 0.020868599782061913,
      "grad_norm": 1.1754058599472046,
      "learning_rate": 1e-05,
      "loss": 0.254,
      "step": 6090
    },
    {
      "epoch": 0.020902866776777943,
      "grad_norm": 1.0922538042068481,
      "learning_rate": 1e-05,
      "loss": 0.2522,
      "step": 6100
    },
    {
      "epoch": 0.020937133771493973,
      "grad_norm": 1.1970179080963135,
      "learning_rate": 1e-05,
      "loss": 0.267,
      "step": 6110
    },
    {
      "epoch": 0.020971400766210003,
      "grad_norm": 1.2625236511230469,
      "learning_rate": 1e-05,
      "loss": 0.2379,
      "step": 6120
    },
    {
      "epoch": 0.021005667760926033,
      "grad_norm": 1.152846336364746,
      "learning_rate": 1e-05,
      "loss": 0.2429,
      "step": 6130
    },
    {
      "epoch": 0.02103993475564206,
      "grad_norm": 1.1184160709381104,
      "learning_rate": 1e-05,
      "loss": 0.2566,
      "step": 6140
    },
    {
      "epoch": 0.02107420175035809,
      "grad_norm": 1.1153484582901,
      "learning_rate": 1e-05,
      "loss": 0.2583,
      "step": 6150
    },
    {
      "epoch": 0.02110846874507412,
      "grad_norm": 1.2822504043579102,
      "learning_rate": 1e-05,
      "loss": 0.2535,
      "step": 6160
    },
    {
      "epoch": 0.02114273573979015,
      "grad_norm": 1.1332992315292358,
      "learning_rate": 1e-05,
      "loss": 0.2799,
      "step": 6170
    },
    {
      "epoch": 0.02117700273450618,
      "grad_norm": 1.0284112691879272,
      "learning_rate": 1e-05,
      "loss": 0.2458,
      "step": 6180
    },
    {
      "epoch": 0.02121126972922221,
      "grad_norm": 1.1097975969314575,
      "learning_rate": 1e-05,
      "loss": 0.2513,
      "step": 6190
    },
    {
      "epoch": 0.021245536723938238,
      "grad_norm": 1.168990969657898,
      "learning_rate": 1e-05,
      "loss": 0.2843,
      "step": 6200
    },
    {
      "epoch": 0.021279803718654268,
      "grad_norm": 0.9956926107406616,
      "learning_rate": 1e-05,
      "loss": 0.247,
      "step": 6210
    },
    {
      "epoch": 0.021314070713370294,
      "grad_norm": 1.2191492319107056,
      "learning_rate": 1e-05,
      "loss": 0.2608,
      "step": 6220
    },
    {
      "epoch": 0.021348337708086324,
      "grad_norm": 1.0872688293457031,
      "learning_rate": 1e-05,
      "loss": 0.2463,
      "step": 6230
    },
    {
      "epoch": 0.021382604702802354,
      "grad_norm": 1.0746614933013916,
      "learning_rate": 1e-05,
      "loss": 0.244,
      "step": 6240
    },
    {
      "epoch": 0.021416871697518384,
      "grad_norm": 1.1560328006744385,
      "learning_rate": 1e-05,
      "loss": 0.2639,
      "step": 6250
    },
    {
      "epoch": 0.021451138692234414,
      "grad_norm": 1.1529641151428223,
      "learning_rate": 1e-05,
      "loss": 0.2585,
      "step": 6260
    },
    {
      "epoch": 0.021485405686950444,
      "grad_norm": 1.0708386898040771,
      "learning_rate": 1e-05,
      "loss": 0.2669,
      "step": 6270
    },
    {
      "epoch": 0.021519672681666473,
      "grad_norm": 1.208079218864441,
      "learning_rate": 1e-05,
      "loss": 0.2436,
      "step": 6280
    },
    {
      "epoch": 0.021553939676382503,
      "grad_norm": 1.1871508359909058,
      "learning_rate": 1e-05,
      "loss": 0.2655,
      "step": 6290
    },
    {
      "epoch": 0.02158820667109853,
      "grad_norm": 1.0997953414916992,
      "learning_rate": 1e-05,
      "loss": 0.2578,
      "step": 6300
    },
    {
      "epoch": 0.02162247366581456,
      "grad_norm": 1.2404417991638184,
      "learning_rate": 1e-05,
      "loss": 0.2726,
      "step": 6310
    },
    {
      "epoch": 0.02165674066053059,
      "grad_norm": 1.1724058389663696,
      "learning_rate": 1e-05,
      "loss": 0.2611,
      "step": 6320
    },
    {
      "epoch": 0.02169100765524662,
      "grad_norm": 1.124932885169983,
      "learning_rate": 1e-05,
      "loss": 0.2582,
      "step": 6330
    },
    {
      "epoch": 0.02172527464996265,
      "grad_norm": 1.129584550857544,
      "learning_rate": 1e-05,
      "loss": 0.2651,
      "step": 6340
    },
    {
      "epoch": 0.02175954164467868,
      "grad_norm": 1.1869479417800903,
      "learning_rate": 1e-05,
      "loss": 0.2451,
      "step": 6350
    },
    {
      "epoch": 0.02179380863939471,
      "grad_norm": 1.1753504276275635,
      "learning_rate": 1e-05,
      "loss": 0.2509,
      "step": 6360
    },
    {
      "epoch": 0.02182807563411074,
      "grad_norm": 1.1704761981964111,
      "learning_rate": 1e-05,
      "loss": 0.2614,
      "step": 6370
    },
    {
      "epoch": 0.02186234262882677,
      "grad_norm": 1.347970724105835,
      "learning_rate": 1e-05,
      "loss": 0.253,
      "step": 6380
    },
    {
      "epoch": 0.021896609623542795,
      "grad_norm": 1.0677597522735596,
      "learning_rate": 1e-05,
      "loss": 0.2539,
      "step": 6390
    },
    {
      "epoch": 0.021930876618258825,
      "grad_norm": 1.1567541360855103,
      "learning_rate": 1e-05,
      "loss": 0.2621,
      "step": 6400
    },
    {
      "epoch": 0.021965143612974854,
      "grad_norm": 1.1231553554534912,
      "learning_rate": 1e-05,
      "loss": 0.2453,
      "step": 6410
    },
    {
      "epoch": 0.021999410607690884,
      "grad_norm": 1.0485198497772217,
      "learning_rate": 1e-05,
      "loss": 0.2503,
      "step": 6420
    },
    {
      "epoch": 0.022033677602406914,
      "grad_norm": 1.12228262424469,
      "learning_rate": 1e-05,
      "loss": 0.2488,
      "step": 6430
    },
    {
      "epoch": 0.022067944597122944,
      "grad_norm": 1.2610136270523071,
      "learning_rate": 1e-05,
      "loss": 0.2445,
      "step": 6440
    },
    {
      "epoch": 0.022102211591838974,
      "grad_norm": 0.9546436071395874,
      "learning_rate": 1e-05,
      "loss": 0.226,
      "step": 6450
    },
    {
      "epoch": 0.022136478586555004,
      "grad_norm": 1.3363466262817383,
      "learning_rate": 1e-05,
      "loss": 0.2489,
      "step": 6460
    },
    {
      "epoch": 0.02217074558127103,
      "grad_norm": 1.1454704999923706,
      "learning_rate": 1e-05,
      "loss": 0.2434,
      "step": 6470
    },
    {
      "epoch": 0.02220501257598706,
      "grad_norm": 1.1578549146652222,
      "learning_rate": 1e-05,
      "loss": 0.2549,
      "step": 6480
    },
    {
      "epoch": 0.02223927957070309,
      "grad_norm": 1.096081018447876,
      "learning_rate": 1e-05,
      "loss": 0.2472,
      "step": 6490
    },
    {
      "epoch": 0.02227354656541912,
      "grad_norm": 1.2388731241226196,
      "learning_rate": 1e-05,
      "loss": 0.2457,
      "step": 6500
    },
    {
      "epoch": 0.02227354656541912,
      "eval_cer": 12.929970012347859,
      "eval_loss": 0.24839338660240173,
      "eval_normalized_cer": 9.242605915267786,
      "eval_runtime": 227.6401,
      "eval_samples_per_second": 2.249,
      "eval_steps_per_second": 0.035,
      "step": 6500
    },
    {
      "epoch": 0.02230781356013515,
      "grad_norm": 1.0306715965270996,
      "learning_rate": 1e-05,
      "loss": 0.2393,
      "step": 6510
    },
    {
      "epoch": 0.02234208055485118,
      "grad_norm": 1.1339504718780518,
      "learning_rate": 1e-05,
      "loss": 0.2563,
      "step": 6520
    },
    {
      "epoch": 0.02237634754956721,
      "grad_norm": 0.912266731262207,
      "learning_rate": 1e-05,
      "loss": 0.2465,
      "step": 6530
    },
    {
      "epoch": 0.02241061454428324,
      "grad_norm": 1.1917020082473755,
      "learning_rate": 1e-05,
      "loss": 0.2395,
      "step": 6540
    },
    {
      "epoch": 0.022444881538999265,
      "grad_norm": 1.248515248298645,
      "learning_rate": 1e-05,
      "loss": 0.2479,
      "step": 6550
    },
    {
      "epoch": 0.022479148533715295,
      "grad_norm": 1.180799961090088,
      "learning_rate": 1e-05,
      "loss": 0.2616,
      "step": 6560
    },
    {
      "epoch": 0.022513415528431325,
      "grad_norm": 1.0700205564498901,
      "learning_rate": 1e-05,
      "loss": 0.2401,
      "step": 6570
    },
    {
      "epoch": 0.022547682523147355,
      "grad_norm": 1.1814614534378052,
      "learning_rate": 1e-05,
      "loss": 0.2471,
      "step": 6580
    },
    {
      "epoch": 0.022581949517863385,
      "grad_norm": 1.3973134756088257,
      "learning_rate": 1e-05,
      "loss": 0.2383,
      "step": 6590
    },
    {
      "epoch": 0.022616216512579414,
      "grad_norm": 1.244265079498291,
      "learning_rate": 1e-05,
      "loss": 0.2548,
      "step": 6600
    },
    {
      "epoch": 0.022650483507295444,
      "grad_norm": 1.1685833930969238,
      "learning_rate": 1e-05,
      "loss": 0.2499,
      "step": 6610
    },
    {
      "epoch": 0.022684750502011474,
      "grad_norm": 1.1566667556762695,
      "learning_rate": 1e-05,
      "loss": 0.2443,
      "step": 6620
    },
    {
      "epoch": 0.0227190174967275,
      "grad_norm": 1.0241929292678833,
      "learning_rate": 1e-05,
      "loss": 0.2412,
      "step": 6630
    },
    {
      "epoch": 0.02275328449144353,
      "grad_norm": 1.0359474420547485,
      "learning_rate": 1e-05,
      "loss": 0.2374,
      "step": 6640
    },
    {
      "epoch": 0.02278755148615956,
      "grad_norm": 1.040810227394104,
      "learning_rate": 1e-05,
      "loss": 0.2254,
      "step": 6650
    },
    {
      "epoch": 0.02282181848087559,
      "grad_norm": 1.0343252420425415,
      "learning_rate": 1e-05,
      "loss": 0.2366,
      "step": 6660
    },
    {
      "epoch": 0.02285608547559162,
      "grad_norm": 1.052739143371582,
      "learning_rate": 1e-05,
      "loss": 0.2273,
      "step": 6670
    },
    {
      "epoch": 0.02289035247030765,
      "grad_norm": 1.0414966344833374,
      "learning_rate": 1e-05,
      "loss": 0.2082,
      "step": 6680
    },
    {
      "epoch": 0.02292461946502368,
      "grad_norm": 1.2340532541275024,
      "learning_rate": 1e-05,
      "loss": 0.2241,
      "step": 6690
    },
    {
      "epoch": 0.02295888645973971,
      "grad_norm": 0.9693310260772705,
      "learning_rate": 1e-05,
      "loss": 0.2322,
      "step": 6700
    },
    {
      "epoch": 0.022993153454455736,
      "grad_norm": 1.103025197982788,
      "learning_rate": 1e-05,
      "loss": 0.236,
      "step": 6710
    },
    {
      "epoch": 0.023027420449171766,
      "grad_norm": 1.119689702987671,
      "learning_rate": 1e-05,
      "loss": 0.214,
      "step": 6720
    },
    {
      "epoch": 0.023061687443887795,
      "grad_norm": 0.93172287940979,
      "learning_rate": 1e-05,
      "loss": 0.2094,
      "step": 6730
    },
    {
      "epoch": 0.023095954438603825,
      "grad_norm": 1.0207446813583374,
      "learning_rate": 1e-05,
      "loss": 0.2238,
      "step": 6740
    },
    {
      "epoch": 0.023130221433319855,
      "grad_norm": 1.200201392173767,
      "learning_rate": 1e-05,
      "loss": 0.218,
      "step": 6750
    },
    {
      "epoch": 0.023164488428035885,
      "grad_norm": 1.1485291719436646,
      "learning_rate": 1e-05,
      "loss": 0.2314,
      "step": 6760
    },
    {
      "epoch": 0.023198755422751915,
      "grad_norm": 1.2236285209655762,
      "learning_rate": 1e-05,
      "loss": 0.2326,
      "step": 6770
    },
    {
      "epoch": 0.023233022417467945,
      "grad_norm": 1.1756523847579956,
      "learning_rate": 1e-05,
      "loss": 0.2122,
      "step": 6780
    },
    {
      "epoch": 0.02326728941218397,
      "grad_norm": 1.0356839895248413,
      "learning_rate": 1e-05,
      "loss": 0.2078,
      "step": 6790
    },
    {
      "epoch": 0.0233015564069,
      "grad_norm": 1.1896883249282837,
      "learning_rate": 1e-05,
      "loss": 0.2072,
      "step": 6800
    },
    {
      "epoch": 0.02333582340161603,
      "grad_norm": 1.1080976724624634,
      "learning_rate": 1e-05,
      "loss": 0.2127,
      "step": 6810
    },
    {
      "epoch": 0.02337009039633206,
      "grad_norm": 1.128263235092163,
      "learning_rate": 1e-05,
      "loss": 0.2282,
      "step": 6820
    },
    {
      "epoch": 0.02340435739104809,
      "grad_norm": 1.0398188829421997,
      "learning_rate": 1e-05,
      "loss": 0.2095,
      "step": 6830
    },
    {
      "epoch": 0.02343862438576412,
      "grad_norm": 1.1791975498199463,
      "learning_rate": 1e-05,
      "loss": 0.2216,
      "step": 6840
    },
    {
      "epoch": 0.02347289138048015,
      "grad_norm": 1.1444710493087769,
      "learning_rate": 1e-05,
      "loss": 0.2447,
      "step": 6850
    },
    {
      "epoch": 0.02350715837519618,
      "grad_norm": 1.136607050895691,
      "learning_rate": 1e-05,
      "loss": 0.2093,
      "step": 6860
    },
    {
      "epoch": 0.023541425369912206,
      "grad_norm": 1.0915231704711914,
      "learning_rate": 1e-05,
      "loss": 0.2128,
      "step": 6870
    },
    {
      "epoch": 0.023575692364628236,
      "grad_norm": 1.0416276454925537,
      "learning_rate": 1e-05,
      "loss": 0.2092,
      "step": 6880
    },
    {
      "epoch": 0.023609959359344266,
      "grad_norm": 1.3693732023239136,
      "learning_rate": 1e-05,
      "loss": 0.2137,
      "step": 6890
    },
    {
      "epoch": 0.023644226354060296,
      "grad_norm": 1.1747677326202393,
      "learning_rate": 1e-05,
      "loss": 0.2215,
      "step": 6900
    },
    {
      "epoch": 0.023678493348776326,
      "grad_norm": 1.1593588590621948,
      "learning_rate": 1e-05,
      "loss": 0.2234,
      "step": 6910
    },
    {
      "epoch": 0.023712760343492355,
      "grad_norm": 1.2322016954421997,
      "learning_rate": 1e-05,
      "loss": 0.2437,
      "step": 6920
    },
    {
      "epoch": 0.023747027338208385,
      "grad_norm": 1.167648196220398,
      "learning_rate": 1e-05,
      "loss": 0.2461,
      "step": 6930
    },
    {
      "epoch": 0.023781294332924415,
      "grad_norm": 1.0984666347503662,
      "learning_rate": 1e-05,
      "loss": 0.2584,
      "step": 6940
    },
    {
      "epoch": 0.023815561327640445,
      "grad_norm": 1.1234291791915894,
      "learning_rate": 1e-05,
      "loss": 0.2532,
      "step": 6950
    },
    {
      "epoch": 0.02384982832235647,
      "grad_norm": 1.2158063650131226,
      "learning_rate": 1e-05,
      "loss": 0.2567,
      "step": 6960
    },
    {
      "epoch": 0.0238840953170725,
      "grad_norm": 1.0958101749420166,
      "learning_rate": 1e-05,
      "loss": 0.2387,
      "step": 6970
    },
    {
      "epoch": 0.02391836231178853,
      "grad_norm": 1.1536844968795776,
      "learning_rate": 1e-05,
      "loss": 0.2712,
      "step": 6980
    },
    {
      "epoch": 0.02395262930650456,
      "grad_norm": 1.2437007427215576,
      "learning_rate": 1e-05,
      "loss": 0.2563,
      "step": 6990
    },
    {
      "epoch": 0.02398689630122059,
      "grad_norm": 1.0884592533111572,
      "learning_rate": 1e-05,
      "loss": 0.2379,
      "step": 7000
    },
    {
      "epoch": 0.02398689630122059,
      "eval_cer": 13.079908273064033,
      "eval_loss": 0.2514401376247406,
      "eval_normalized_cer": 9.622302158273381,
      "eval_runtime": 227.6705,
      "eval_samples_per_second": 2.249,
      "eval_steps_per_second": 0.035,
      "step": 7000
    },
    {
      "epoch": 0.02402116329593662,
      "grad_norm": 1.2332980632781982,
      "learning_rate": 1e-05,
      "loss": 0.2543,
      "step": 7010
    },
    {
      "epoch": 0.02405543029065265,
      "grad_norm": 1.1041260957717896,
      "learning_rate": 1e-05,
      "loss": 0.2663,
      "step": 7020
    },
    {
      "epoch": 0.02408969728536868,
      "grad_norm": 1.1479183435440063,
      "learning_rate": 1e-05,
      "loss": 0.2528,
      "step": 7030
    },
    {
      "epoch": 0.024123964280084707,
      "grad_norm": 1.103766918182373,
      "learning_rate": 1e-05,
      "loss": 0.2336,
      "step": 7040
    },
    {
      "epoch": 0.024158231274800736,
      "grad_norm": 1.238996148109436,
      "learning_rate": 1e-05,
      "loss": 0.2436,
      "step": 7050
    },
    {
      "epoch": 0.024192498269516766,
      "grad_norm": 1.2652095556259155,
      "learning_rate": 1e-05,
      "loss": 0.2464,
      "step": 7060
    },
    {
      "epoch": 0.024226765264232796,
      "grad_norm": 1.180665373802185,
      "learning_rate": 1e-05,
      "loss": 0.2541,
      "step": 7070
    },
    {
      "epoch": 0.024261032258948826,
      "grad_norm": 1.1601506471633911,
      "learning_rate": 1e-05,
      "loss": 0.2508,
      "step": 7080
    },
    {
      "epoch": 0.024295299253664856,
      "grad_norm": 1.257034420967102,
      "learning_rate": 1e-05,
      "loss": 0.2446,
      "step": 7090
    },
    {
      "epoch": 0.024329566248380886,
      "grad_norm": 1.0813285112380981,
      "learning_rate": 1e-05,
      "loss": 0.2546,
      "step": 7100
    },
    {
      "epoch": 0.024363833243096916,
      "grad_norm": 1.1124157905578613,
      "learning_rate": 1e-05,
      "loss": 0.2379,
      "step": 7110
    },
    {
      "epoch": 0.024398100237812942,
      "grad_norm": 1.0615211725234985,
      "learning_rate": 1e-05,
      "loss": 0.253,
      "step": 7120
    },
    {
      "epoch": 0.024432367232528972,
      "grad_norm": 1.185677409172058,
      "learning_rate": 1e-05,
      "loss": 0.2383,
      "step": 7130
    },
    {
      "epoch": 0.024466634227245,
      "grad_norm": 1.1810061931610107,
      "learning_rate": 1e-05,
      "loss": 0.2603,
      "step": 7140
    },
    {
      "epoch": 0.02450090122196103,
      "grad_norm": 1.155860424041748,
      "learning_rate": 1e-05,
      "loss": 0.2434,
      "step": 7150
    },
    {
      "epoch": 0.02453516821667706,
      "grad_norm": 1.113008737564087,
      "learning_rate": 1e-05,
      "loss": 0.2529,
      "step": 7160
    },
    {
      "epoch": 0.02456943521139309,
      "grad_norm": 1.1276872158050537,
      "learning_rate": 1e-05,
      "loss": 0.2265,
      "step": 7170
    },
    {
      "epoch": 0.02460370220610912,
      "grad_norm": 1.149792194366455,
      "learning_rate": 1e-05,
      "loss": 0.2349,
      "step": 7180
    },
    {
      "epoch": 0.02463796920082515,
      "grad_norm": 1.1619532108306885,
      "learning_rate": 1e-05,
      "loss": 0.2336,
      "step": 7190
    },
    {
      "epoch": 0.024672236195541177,
      "grad_norm": 1.0760303735733032,
      "learning_rate": 1e-05,
      "loss": 0.2315,
      "step": 7200
    },
    {
      "epoch": 0.024706503190257207,
      "grad_norm": 1.2807782888412476,
      "learning_rate": 1e-05,
      "loss": 0.2382,
      "step": 7210
    },
    {
      "epoch": 0.024740770184973237,
      "grad_norm": 1.0910037755966187,
      "learning_rate": 1e-05,
      "loss": 0.2333,
      "step": 7220
    },
    {
      "epoch": 0.024775037179689267,
      "grad_norm": 1.2938390970230103,
      "learning_rate": 1e-05,
      "loss": 0.2147,
      "step": 7230
    },
    {
      "epoch": 0.024809304174405297,
      "grad_norm": 1.185542106628418,
      "learning_rate": 1e-05,
      "loss": 0.2232,
      "step": 7240
    },
    {
      "epoch": 0.024843571169121326,
      "grad_norm": 1.0598995685577393,
      "learning_rate": 1e-05,
      "loss": 0.2278,
      "step": 7250
    },
    {
      "epoch": 0.024877838163837356,
      "grad_norm": 1.1860477924346924,
      "learning_rate": 1e-05,
      "loss": 0.2179,
      "step": 7260
    },
    {
      "epoch": 0.024912105158553386,
      "grad_norm": 1.1935844421386719,
      "learning_rate": 1e-05,
      "loss": 0.238,
      "step": 7270
    },
    {
      "epoch": 0.024946372153269412,
      "grad_norm": 1.0449039936065674,
      "learning_rate": 1e-05,
      "loss": 0.2307,
      "step": 7280
    },
    {
      "epoch": 0.024980639147985442,
      "grad_norm": 1.0651369094848633,
      "learning_rate": 1e-05,
      "loss": 0.2379,
      "step": 7290
    },
    {
      "epoch": 0.025014906142701472,
      "grad_norm": 1.0416852235794067,
      "learning_rate": 1e-05,
      "loss": 0.2208,
      "step": 7300
    },
    {
      "epoch": 0.025049173137417502,
      "grad_norm": 1.0064860582351685,
      "learning_rate": 1e-05,
      "loss": 0.2227,
      "step": 7310
    },
    {
      "epoch": 0.025083440132133532,
      "grad_norm": 1.0357342958450317,
      "learning_rate": 1e-05,
      "loss": 0.22,
      "step": 7320
    },
    {
      "epoch": 0.02511770712684956,
      "grad_norm": 1.019918441772461,
      "learning_rate": 1e-05,
      "loss": 0.2396,
      "step": 7330
    },
    {
      "epoch": 0.02515197412156559,
      "grad_norm": 1.0327798128128052,
      "learning_rate": 1e-05,
      "loss": 0.2118,
      "step": 7340
    },
    {
      "epoch": 0.02518624111628162,
      "grad_norm": 0.9973874092102051,
      "learning_rate": 1e-05,
      "loss": 0.2275,
      "step": 7350
    },
    {
      "epoch": 0.025220508110997648,
      "grad_norm": 1.093544840812683,
      "learning_rate": 1e-05,
      "loss": 0.2214,
      "step": 7360
    },
    {
      "epoch": 0.025254775105713678,
      "grad_norm": 1.118829369544983,
      "learning_rate": 1e-05,
      "loss": 0.237,
      "step": 7370
    },
    {
      "epoch": 0.025289042100429707,
      "grad_norm": 1.2009224891662598,
      "learning_rate": 1e-05,
      "loss": 0.2447,
      "step": 7380
    },
    {
      "epoch": 0.025323309095145737,
      "grad_norm": 1.1427584886550903,
      "learning_rate": 1e-05,
      "loss": 0.234,
      "step": 7390
    },
    {
      "epoch": 0.025357576089861767,
      "grad_norm": 0.9685842394828796,
      "learning_rate": 1e-05,
      "loss": 0.2231,
      "step": 7400
    },
    {
      "epoch": 0.025391843084577797,
      "grad_norm": 1.165501356124878,
      "learning_rate": 1e-05,
      "loss": 0.2139,
      "step": 7410
    },
    {
      "epoch": 0.025426110079293827,
      "grad_norm": 1.4023411273956299,
      "learning_rate": 1e-05,
      "loss": 0.236,
      "step": 7420
    },
    {
      "epoch": 0.025460377074009857,
      "grad_norm": 1.218546748161316,
      "learning_rate": 1e-05,
      "loss": 0.2433,
      "step": 7430
    },
    {
      "epoch": 0.025494644068725883,
      "grad_norm": 1.4930671453475952,
      "learning_rate": 1e-05,
      "loss": 0.2466,
      "step": 7440
    },
    {
      "epoch": 0.025528911063441913,
      "grad_norm": 1.145317554473877,
      "learning_rate": 1e-05,
      "loss": 0.2535,
      "step": 7450
    },
    {
      "epoch": 0.025563178058157943,
      "grad_norm": 1.2366299629211426,
      "learning_rate": 1e-05,
      "loss": 0.2606,
      "step": 7460
    },
    {
      "epoch": 0.025597445052873972,
      "grad_norm": 1.0542744398117065,
      "learning_rate": 1e-05,
      "loss": 0.2493,
      "step": 7470
    },
    {
      "epoch": 0.025631712047590002,
      "grad_norm": 1.2272337675094604,
      "learning_rate": 1e-05,
      "loss": 0.2537,
      "step": 7480
    },
    {
      "epoch": 0.025665979042306032,
      "grad_norm": 1.169912576675415,
      "learning_rate": 1e-05,
      "loss": 0.2581,
      "step": 7490
    },
    {
      "epoch": 0.025700246037022062,
      "grad_norm": 1.1997913122177124,
      "learning_rate": 1e-05,
      "loss": 0.2547,
      "step": 7500
    },
    {
      "epoch": 0.025700246037022062,
      "eval_cer": 12.859410830834362,
      "eval_loss": 0.2470153123140335,
      "eval_normalized_cer": 9.162669864108713,
      "eval_runtime": 227.7782,
      "eval_samples_per_second": 2.248,
      "eval_steps_per_second": 0.035,
      "step": 7500
    },
    {
      "epoch": 0.025734513031738092,
      "grad_norm": 1.0920944213867188,
      "learning_rate": 1e-05,
      "loss": 0.2498,
      "step": 7510
    },
    {
      "epoch": 0.025768780026454118,
      "grad_norm": 1.349660038948059,
      "learning_rate": 1e-05,
      "loss": 0.2591,
      "step": 7520
    },
    {
      "epoch": 0.025803047021170148,
      "grad_norm": 1.0097490549087524,
      "learning_rate": 1e-05,
      "loss": 0.2514,
      "step": 7530
    },
    {
      "epoch": 0.025837314015886178,
      "grad_norm": 1.118241548538208,
      "learning_rate": 1e-05,
      "loss": 0.2603,
      "step": 7540
    },
    {
      "epoch": 0.025871581010602208,
      "grad_norm": 1.078802466392517,
      "learning_rate": 1e-05,
      "loss": 0.2532,
      "step": 7550
    },
    {
      "epoch": 0.025905848005318238,
      "grad_norm": 1.0794482231140137,
      "learning_rate": 1e-05,
      "loss": 0.2521,
      "step": 7560
    },
    {
      "epoch": 0.025940115000034267,
      "grad_norm": 1.130106806755066,
      "learning_rate": 1e-05,
      "loss": 0.2574,
      "step": 7570
    },
    {
      "epoch": 0.025974381994750297,
      "grad_norm": 1.112724781036377,
      "learning_rate": 1e-05,
      "loss": 0.253,
      "step": 7580
    },
    {
      "epoch": 0.026008648989466327,
      "grad_norm": 1.2646088600158691,
      "learning_rate": 1e-05,
      "loss": 0.2548,
      "step": 7590
    },
    {
      "epoch": 0.026042915984182357,
      "grad_norm": 1.1961979866027832,
      "learning_rate": 1e-05,
      "loss": 0.2548,
      "step": 7600
    },
    {
      "epoch": 0.026077182978898383,
      "grad_norm": 1.2568695545196533,
      "learning_rate": 1e-05,
      "loss": 0.245,
      "step": 7610
    },
    {
      "epoch": 0.026111449973614413,
      "grad_norm": 1.0233054161071777,
      "learning_rate": 1e-05,
      "loss": 0.2429,
      "step": 7620
    },
    {
      "epoch": 0.026145716968330443,
      "grad_norm": 1.4355731010437012,
      "learning_rate": 1e-05,
      "loss": 0.2623,
      "step": 7630
    },
    {
      "epoch": 0.026179983963046473,
      "grad_norm": 0.9781149625778198,
      "learning_rate": 1e-05,
      "loss": 0.2436,
      "step": 7640
    },
    {
      "epoch": 0.026214250957762503,
      "grad_norm": 1.085255742073059,
      "learning_rate": 1e-05,
      "loss": 0.2475,
      "step": 7650
    },
    {
      "epoch": 0.026248517952478533,
      "grad_norm": 1.0647081136703491,
      "learning_rate": 1e-05,
      "loss": 0.2596,
      "step": 7660
    },
    {
      "epoch": 0.026282784947194562,
      "grad_norm": 1.3411939144134521,
      "learning_rate": 1e-05,
      "loss": 0.2444,
      "step": 7670
    },
    {
      "epoch": 0.026317051941910592,
      "grad_norm": 1.0778676271438599,
      "learning_rate": 1e-05,
      "loss": 0.2499,
      "step": 7680
    },
    {
      "epoch": 0.02635131893662662,
      "grad_norm": 1.1606541872024536,
      "learning_rate": 1e-05,
      "loss": 0.2537,
      "step": 7690
    },
    {
      "epoch": 0.02638558593134265,
      "grad_norm": 1.0706511735916138,
      "learning_rate": 1e-05,
      "loss": 0.2324,
      "step": 7700
    },
    {
      "epoch": 0.02641985292605868,
      "grad_norm": 1.2074836492538452,
      "learning_rate": 1e-05,
      "loss": 0.2487,
      "step": 7710
    },
    {
      "epoch": 0.026454119920774708,
      "grad_norm": 1.0147804021835327,
      "learning_rate": 1e-05,
      "loss": 0.2202,
      "step": 7720
    },
    {
      "epoch": 0.026488386915490738,
      "grad_norm": 1.1806961297988892,
      "learning_rate": 1e-05,
      "loss": 0.2464,
      "step": 7730
    },
    {
      "epoch": 0.026522653910206768,
      "grad_norm": 1.1552751064300537,
      "learning_rate": 1e-05,
      "loss": 0.2244,
      "step": 7740
    },
    {
      "epoch": 0.026556920904922798,
      "grad_norm": 1.115871548652649,
      "learning_rate": 1e-05,
      "loss": 0.2389,
      "step": 7750
    },
    {
      "epoch": 0.026591187899638827,
      "grad_norm": 1.0924640893936157,
      "learning_rate": 1e-05,
      "loss": 0.2237,
      "step": 7760
    },
    {
      "epoch": 0.026625454894354854,
      "grad_norm": 1.021644115447998,
      "learning_rate": 1e-05,
      "loss": 0.2257,
      "step": 7770
    },
    {
      "epoch": 0.026659721889070884,
      "grad_norm": 1.1757131814956665,
      "learning_rate": 1e-05,
      "loss": 0.2278,
      "step": 7780
    },
    {
      "epoch": 0.026693988883786914,
      "grad_norm": 1.1914074420928955,
      "learning_rate": 1e-05,
      "loss": 0.2266,
      "step": 7790
    },
    {
      "epoch": 0.026728255878502943,
      "grad_norm": 1.0416505336761475,
      "learning_rate": 1e-05,
      "loss": 0.2273,
      "step": 7800
    },
    {
      "epoch": 0.026762522873218973,
      "grad_norm": 1.0241059064865112,
      "learning_rate": 1e-05,
      "loss": 0.2342,
      "step": 7810
    },
    {
      "epoch": 0.026796789867935003,
      "grad_norm": 1.133334994316101,
      "learning_rate": 1e-05,
      "loss": 0.2303,
      "step": 7820
    },
    {
      "epoch": 0.026831056862651033,
      "grad_norm": 1.1711792945861816,
      "learning_rate": 1e-05,
      "loss": 0.2333,
      "step": 7830
    },
    {
      "epoch": 0.026865323857367063,
      "grad_norm": 1.1120338439941406,
      "learning_rate": 1e-05,
      "loss": 0.2474,
      "step": 7840
    },
    {
      "epoch": 0.02689959085208309,
      "grad_norm": 1.1995311975479126,
      "learning_rate": 1e-05,
      "loss": 0.2472,
      "step": 7850
    },
    {
      "epoch": 0.02693385784679912,
      "grad_norm": 1.1725718975067139,
      "learning_rate": 1e-05,
      "loss": 0.2361,
      "step": 7860
    },
    {
      "epoch": 0.02696812484151515,
      "grad_norm": 0.9564438462257385,
      "learning_rate": 1e-05,
      "loss": 0.2266,
      "step": 7870
    },
    {
      "epoch": 0.02700239183623118,
      "grad_norm": 1.140692114830017,
      "learning_rate": 1e-05,
      "loss": 0.2319,
      "step": 7880
    },
    {
      "epoch": 0.02703665883094721,
      "grad_norm": 1.0812654495239258,
      "learning_rate": 1e-05,
      "loss": 0.2434,
      "step": 7890
    },
    {
      "epoch": 0.02707092582566324,
      "grad_norm": 1.179500937461853,
      "learning_rate": 1e-05,
      "loss": 0.2191,
      "step": 7900
    },
    {
      "epoch": 0.027105192820379268,
      "grad_norm": 1.1073647737503052,
      "learning_rate": 1e-05,
      "loss": 0.2315,
      "step": 7910
    },
    {
      "epoch": 0.027139459815095298,
      "grad_norm": 1.093070387840271,
      "learning_rate": 1e-05,
      "loss": 0.2256,
      "step": 7920
    },
    {
      "epoch": 0.027173726809811324,
      "grad_norm": 1.2253212928771973,
      "learning_rate": 1e-05,
      "loss": 0.2413,
      "step": 7930
    },
    {
      "epoch": 0.027207993804527354,
      "grad_norm": 1.1531736850738525,
      "learning_rate": 1e-05,
      "loss": 0.2514,
      "step": 7940
    },
    {
      "epoch": 0.027242260799243384,
      "grad_norm": 1.0366076231002808,
      "learning_rate": 1e-05,
      "loss": 0.2475,
      "step": 7950
    },
    {
      "epoch": 0.027276527793959414,
      "grad_norm": 1.1657369136810303,
      "learning_rate": 1e-05,
      "loss": 0.2475,
      "step": 7960
    },
    {
      "epoch": 0.027310794788675444,
      "grad_norm": 1.3050105571746826,
      "learning_rate": 1e-05,
      "loss": 0.2704,
      "step": 7970
    },
    {
      "epoch": 0.027345061783391474,
      "grad_norm": 1.1378298997879028,
      "learning_rate": 1e-05,
      "loss": 0.2481,
      "step": 7980
    },
    {
      "epoch": 0.027379328778107503,
      "grad_norm": 1.1434043645858765,
      "learning_rate": 1e-05,
      "loss": 0.2671,
      "step": 7990
    },
    {
      "epoch": 0.027413595772823533,
      "grad_norm": 1.0899518728256226,
      "learning_rate": 1e-05,
      "loss": 0.2573,
      "step": 8000
    },
    {
      "epoch": 0.027413595772823533,
      "eval_cer": 12.903510319280295,
      "eval_loss": 0.2475583553314209,
      "eval_normalized_cer": 9.362509992006395,
      "eval_runtime": 228.4278,
      "eval_samples_per_second": 2.241,
      "eval_steps_per_second": 0.035,
      "step": 8000
    },
    {
      "epoch": 0.02744786276753956,
      "grad_norm": 0.970212459564209,
      "learning_rate": 1e-05,
      "loss": 0.238,
      "step": 8010
    },
    {
      "epoch": 0.02748212976225559,
      "grad_norm": 1.0460257530212402,
      "learning_rate": 1e-05,
      "loss": 0.2606,
      "step": 8020
    },
    {
      "epoch": 0.02751639675697162,
      "grad_norm": 1.116742491722107,
      "learning_rate": 1e-05,
      "loss": 0.2571,
      "step": 8030
    },
    {
      "epoch": 0.02755066375168765,
      "grad_norm": 1.2562140226364136,
      "learning_rate": 1e-05,
      "loss": 0.2561,
      "step": 8040
    },
    {
      "epoch": 0.02758493074640368,
      "grad_norm": 1.2499713897705078,
      "learning_rate": 1e-05,
      "loss": 0.2683,
      "step": 8050
    },
    {
      "epoch": 0.02761919774111971,
      "grad_norm": 1.151715874671936,
      "learning_rate": 1e-05,
      "loss": 0.2463,
      "step": 8060
    },
    {
      "epoch": 0.02765346473583574,
      "grad_norm": 1.2527892589569092,
      "learning_rate": 1e-05,
      "loss": 0.261,
      "step": 8070
    },
    {
      "epoch": 0.02768773173055177,
      "grad_norm": 1.1776025295257568,
      "learning_rate": 1e-05,
      "loss": 0.2616,
      "step": 8080
    },
    {
      "epoch": 0.027721998725267795,
      "grad_norm": 1.1632285118103027,
      "learning_rate": 1e-05,
      "loss": 0.2508,
      "step": 8090
    },
    {
      "epoch": 0.027756265719983825,
      "grad_norm": 1.3266422748565674,
      "learning_rate": 1e-05,
      "loss": 0.2667,
      "step": 8100
    },
    {
      "epoch": 0.027790532714699855,
      "grad_norm": 1.240424633026123,
      "learning_rate": 1e-05,
      "loss": 0.2582,
      "step": 8110
    },
    {
      "epoch": 0.027824799709415884,
      "grad_norm": 1.1874525547027588,
      "learning_rate": 1e-05,
      "loss": 0.2505,
      "step": 8120
    },
    {
      "epoch": 0.027859066704131914,
      "grad_norm": 1.0850279331207275,
      "learning_rate": 1e-05,
      "loss": 0.2556,
      "step": 8130
    },
    {
      "epoch": 0.027893333698847944,
      "grad_norm": 1.203342318534851,
      "learning_rate": 1e-05,
      "loss": 0.2526,
      "step": 8140
    },
    {
      "epoch": 0.027927600693563974,
      "grad_norm": 0.9685319066047668,
      "learning_rate": 1e-05,
      "loss": 0.2614,
      "step": 8150
    },
    {
      "epoch": 0.027961867688280004,
      "grad_norm": 1.020749807357788,
      "learning_rate": 1e-05,
      "loss": 0.2763,
      "step": 8160
    },
    {
      "epoch": 0.02799613468299603,
      "grad_norm": 1.1530399322509766,
      "learning_rate": 1e-05,
      "loss": 0.2544,
      "step": 8170
    },
    {
      "epoch": 0.02803040167771206,
      "grad_norm": 1.0800687074661255,
      "learning_rate": 1e-05,
      "loss": 0.2628,
      "step": 8180
    },
    {
      "epoch": 0.02806466867242809,
      "grad_norm": 1.1825618743896484,
      "learning_rate": 1e-05,
      "loss": 0.2524,
      "step": 8190
    },
    {
      "epoch": 0.02809893566714412,
      "grad_norm": 1.176870346069336,
      "learning_rate": 1e-05,
      "loss": 0.2401,
      "step": 8200
    },
    {
      "epoch": 0.02813320266186015,
      "grad_norm": 1.19720458984375,
      "learning_rate": 1e-05,
      "loss": 0.2801,
      "step": 8210
    },
    {
      "epoch": 0.02816746965657618,
      "grad_norm": 1.0634618997573853,
      "learning_rate": 1e-05,
      "loss": 0.2607,
      "step": 8220
    },
    {
      "epoch": 0.02820173665129221,
      "grad_norm": 1.1780894994735718,
      "learning_rate": 1e-05,
      "loss": 0.2558,
      "step": 8230
    },
    {
      "epoch": 0.02823600364600824,
      "grad_norm": 1.18949294090271,
      "learning_rate": 1e-05,
      "loss": 0.2432,
      "step": 8240
    },
    {
      "epoch": 0.02827027064072427,
      "grad_norm": 1.3350197076797485,
      "learning_rate": 1e-05,
      "loss": 0.2644,
      "step": 8250
    },
    {
      "epoch": 0.028304537635440295,
      "grad_norm": 1.1507694721221924,
      "learning_rate": 1e-05,
      "loss": 0.254,
      "step": 8260
    },
    {
      "epoch": 0.028338804630156325,
      "grad_norm": 1.0806615352630615,
      "learning_rate": 1e-05,
      "loss": 0.2479,
      "step": 8270
    },
    {
      "epoch": 0.028373071624872355,
      "grad_norm": 1.1201471090316772,
      "learning_rate": 1e-05,
      "loss": 0.2553,
      "step": 8280
    },
    {
      "epoch": 0.028407338619588385,
      "grad_norm": 1.0681666135787964,
      "learning_rate": 1e-05,
      "loss": 0.258,
      "step": 8290
    },
    {
      "epoch": 0.028441605614304415,
      "grad_norm": 1.0958445072174072,
      "learning_rate": 1e-05,
      "loss": 0.2502,
      "step": 8300
    },
    {
      "epoch": 0.028475872609020444,
      "grad_norm": 1.165635585784912,
      "learning_rate": 1e-05,
      "loss": 0.2642,
      "step": 8310
    },
    {
      "epoch": 0.028510139603736474,
      "grad_norm": 0.9674690961837769,
      "learning_rate": 1e-05,
      "loss": 0.2385,
      "step": 8320
    },
    {
      "epoch": 0.028544406598452504,
      "grad_norm": 1.239996314048767,
      "learning_rate": 1e-05,
      "loss": 0.2706,
      "step": 8330
    },
    {
      "epoch": 0.02857867359316853,
      "grad_norm": 1.0063962936401367,
      "learning_rate": 1e-05,
      "loss": 0.2448,
      "step": 8340
    },
    {
      "epoch": 0.02861294058788456,
      "grad_norm": 1.0466179847717285,
      "learning_rate": 1e-05,
      "loss": 0.2452,
      "step": 8350
    },
    {
      "epoch": 0.02864720758260059,
      "grad_norm": Infinity,
      "learning_rate": 1e-05,
      "loss": 0.2595,
      "step": 8360
    },
    {
      "epoch": 0.02868147457731662,
      "grad_norm": 1.1461595296859741,
      "learning_rate": 1e-05,
      "loss": 0.2515,
      "step": 8370
    },
    {
      "epoch": 0.02871574157203265,
      "grad_norm": 1.2697845697402954,
      "learning_rate": 1e-05,
      "loss": 0.2641,
      "step": 8380
    },
    {
      "epoch": 0.02875000856674868,
      "grad_norm": 1.2665945291519165,
      "learning_rate": 1e-05,
      "loss": 0.2613,
      "step": 8390
    },
    {
      "epoch": 0.02878427556146471,
      "grad_norm": 1.1350281238555908,
      "learning_rate": 1e-05,
      "loss": 0.2524,
      "step": 8400
    },
    {
      "epoch": 0.02881854255618074,
      "grad_norm": 1.0341808795928955,
      "learning_rate": 1e-05,
      "loss": 0.2466,
      "step": 8410
    },
    {
      "epoch": 0.028852809550896766,
      "grad_norm": 1.1108484268188477,
      "learning_rate": 1e-05,
      "loss": 0.2471,
      "step": 8420
    },
    {
      "epoch": 0.028887076545612796,
      "grad_norm": 1.059414029121399,
      "learning_rate": 1e-05,
      "loss": 0.2695,
      "step": 8430
    },
    {
      "epoch": 0.028921343540328825,
      "grad_norm": 1.0888679027557373,
      "learning_rate": 1e-05,
      "loss": 0.2683,
      "step": 8440
    },
    {
      "epoch": 0.028955610535044855,
      "grad_norm": 1.1649068593978882,
      "learning_rate": 1e-05,
      "loss": 0.2485,
      "step": 8450
    },
    {
      "epoch": 0.028989877529760885,
      "grad_norm": 1.218563199043274,
      "learning_rate": 1e-05,
      "loss": 0.2456,
      "step": 8460
    },
    {
      "epoch": 0.029024144524476915,
      "grad_norm": 1.3558833599090576,
      "learning_rate": 1e-05,
      "loss": 0.2517,
      "step": 8470
    },
    {
      "epoch": 0.029058411519192945,
      "grad_norm": 1.2579597234725952,
      "learning_rate": 1e-05,
      "loss": 0.2516,
      "step": 8480
    },
    {
      "epoch": 0.029092678513908975,
      "grad_norm": 1.185253381729126,
      "learning_rate": 1e-05,
      "loss": 0.2475,
      "step": 8490
    },
    {
      "epoch": 0.029126945508625,
      "grad_norm": 1.1937752962112427,
      "learning_rate": 1e-05,
      "loss": 0.2654,
      "step": 8500
    },
    {
      "epoch": 0.029126945508625,
      "eval_cer": 12.89469042159111,
      "eval_loss": 0.2503049969673157,
      "eval_normalized_cer": 8.952837729816148,
      "eval_runtime": 229.0216,
      "eval_samples_per_second": 2.236,
      "eval_steps_per_second": 0.035,
      "step": 8500
    },
    {
      "epoch": 0.02916121250334103,
      "grad_norm": 1.1005933284759521,
      "learning_rate": 1e-05,
      "loss": 0.247,
      "step": 8510
    },
    {
      "epoch": 0.02919547949805706,
      "grad_norm": 1.0437865257263184,
      "learning_rate": 1e-05,
      "loss": 0.2492,
      "step": 8520
    },
    {
      "epoch": 0.02922974649277309,
      "grad_norm": 1.0478579998016357,
      "learning_rate": 1e-05,
      "loss": 0.2508,
      "step": 8530
    },
    {
      "epoch": 0.02926401348748912,
      "grad_norm": 1.0615030527114868,
      "learning_rate": 1e-05,
      "loss": 0.2305,
      "step": 8540
    },
    {
      "epoch": 0.02929828048220515,
      "grad_norm": 1.105209469795227,
      "learning_rate": 1e-05,
      "loss": 0.2572,
      "step": 8550
    },
    {
      "epoch": 0.02933254747692118,
      "grad_norm": 1.139857530593872,
      "learning_rate": 1e-05,
      "loss": 0.237,
      "step": 8560
    },
    {
      "epoch": 0.02936681447163721,
      "grad_norm": 1.0326822996139526,
      "learning_rate": 1e-05,
      "loss": 0.2587,
      "step": 8570
    },
    {
      "epoch": 0.029401081466353236,
      "grad_norm": 1.4446253776550293,
      "learning_rate": 1e-05,
      "loss": 0.2488,
      "step": 8580
    },
    {
      "epoch": 0.029435348461069266,
      "grad_norm": 1.070698857307434,
      "learning_rate": 1e-05,
      "loss": 0.236,
      "step": 8590
    },
    {
      "epoch": 0.029469615455785296,
      "grad_norm": 1.119545817375183,
      "learning_rate": 1e-05,
      "loss": 0.2406,
      "step": 8600
    },
    {
      "epoch": 0.029503882450501326,
      "grad_norm": 1.1146609783172607,
      "learning_rate": 1e-05,
      "loss": 0.2429,
      "step": 8610
    },
    {
      "epoch": 0.029538149445217356,
      "grad_norm": 1.107639193534851,
      "learning_rate": 1e-05,
      "loss": 0.2412,
      "step": 8620
    },
    {
      "epoch": 0.029572416439933386,
      "grad_norm": 1.0722100734710693,
      "learning_rate": 1e-05,
      "loss": 0.2444,
      "step": 8630
    },
    {
      "epoch": 0.029606683434649415,
      "grad_norm": 1.3313097953796387,
      "learning_rate": 1e-05,
      "loss": 0.2482,
      "step": 8640
    },
    {
      "epoch": 0.029640950429365445,
      "grad_norm": 1.13486647605896,
      "learning_rate": 1e-05,
      "loss": 0.2397,
      "step": 8650
    },
    {
      "epoch": 0.02967521742408147,
      "grad_norm": 1.1610273122787476,
      "learning_rate": 1e-05,
      "loss": 0.2594,
      "step": 8660
    },
    {
      "epoch": 0.0297094844187975,
      "grad_norm": 1.3555855751037598,
      "learning_rate": 1e-05,
      "loss": 0.2521,
      "step": 8670
    },
    {
      "epoch": 0.02974375141351353,
      "grad_norm": 1.2834869623184204,
      "learning_rate": 1e-05,
      "loss": 0.2638,
      "step": 8680
    },
    {
      "epoch": 0.02977801840822956,
      "grad_norm": 1.1000789403915405,
      "learning_rate": 1e-05,
      "loss": 0.2371,
      "step": 8690
    },
    {
      "epoch": 0.02981228540294559,
      "grad_norm": 0.9901896119117737,
      "learning_rate": 1e-05,
      "loss": 0.2535,
      "step": 8700
    },
    {
      "epoch": 0.02984655239766162,
      "grad_norm": 1.0514518022537231,
      "learning_rate": 1e-05,
      "loss": 0.2443,
      "step": 8710
    },
    {
      "epoch": 0.02988081939237765,
      "grad_norm": 1.1127166748046875,
      "learning_rate": 1e-05,
      "loss": 0.2259,
      "step": 8720
    },
    {
      "epoch": 0.02991508638709368,
      "grad_norm": 1.0674943923950195,
      "learning_rate": 1e-05,
      "loss": 0.2309,
      "step": 8730
    },
    {
      "epoch": 0.029949353381809707,
      "grad_norm": 1.0397884845733643,
      "learning_rate": 1e-05,
      "loss": 0.2411,
      "step": 8740
    },
    {
      "epoch": 0.029983620376525737,
      "grad_norm": 1.2052630186080933,
      "learning_rate": 1e-05,
      "loss": 0.2294,
      "step": 8750
    },
    {
      "epoch": 0.030017887371241767,
      "grad_norm": 1.1350561380386353,
      "learning_rate": 1e-05,
      "loss": 0.2264,
      "step": 8760
    },
    {
      "epoch": 0.030052154365957796,
      "grad_norm": 1.1187571287155151,
      "learning_rate": 1e-05,
      "loss": 0.2325,
      "step": 8770
    },
    {
      "epoch": 0.030086421360673826,
      "grad_norm": 1.0860145092010498,
      "learning_rate": 1e-05,
      "loss": 0.2328,
      "step": 8780
    },
    {
      "epoch": 0.030120688355389856,
      "grad_norm": 1.1102906465530396,
      "learning_rate": 1e-05,
      "loss": 0.2271,
      "step": 8790
    },
    {
      "epoch": 0.030154955350105886,
      "grad_norm": 1.0239520072937012,
      "learning_rate": 1e-05,
      "loss": 0.2489,
      "step": 8800
    },
    {
      "epoch": 0.030189222344821916,
      "grad_norm": 1.0980205535888672,
      "learning_rate": 1e-05,
      "loss": 0.2314,
      "step": 8810
    },
    {
      "epoch": 0.030223489339537942,
      "grad_norm": 1.1887843608856201,
      "learning_rate": 1e-05,
      "loss": 0.24,
      "step": 8820
    },
    {
      "epoch": 0.030257756334253972,
      "grad_norm": 1.2101106643676758,
      "learning_rate": 1e-05,
      "loss": 0.2353,
      "step": 8830
    },
    {
      "epoch": 0.030292023328970002,
      "grad_norm": 1.1793345212936401,
      "learning_rate": 1e-05,
      "loss": 0.2482,
      "step": 8840
    },
    {
      "epoch": 0.03032629032368603,
      "grad_norm": 1.1983866691589355,
      "learning_rate": 1e-05,
      "loss": 0.2441,
      "step": 8850
    },
    {
      "epoch": 0.03036055731840206,
      "grad_norm": 0.9888906478881836,
      "learning_rate": 1e-05,
      "loss": 0.2271,
      "step": 8860
    },
    {
      "epoch": 0.03039482431311809,
      "grad_norm": 1.217824101448059,
      "learning_rate": 1e-05,
      "loss": 0.2298,
      "step": 8870
    },
    {
      "epoch": 0.03042909130783412,
      "grad_norm": 0.9851268529891968,
      "learning_rate": 1e-05,
      "loss": 0.2318,
      "step": 8880
    },
    {
      "epoch": 0.03046335830255015,
      "grad_norm": 1.0329748392105103,
      "learning_rate": 1e-05,
      "loss": 0.2654,
      "step": 8890
    },
    {
      "epoch": 0.03049762529726618,
      "grad_norm": 1.067325234413147,
      "learning_rate": 1e-05,
      "loss": 0.23,
      "step": 8900
    },
    {
      "epoch": 0.030531892291982207,
      "grad_norm": 1.2636964321136475,
      "learning_rate": 1e-05,
      "loss": 0.2303,
      "step": 8910
    },
    {
      "epoch": 0.030566159286698237,
      "grad_norm": 1.1565788984298706,
      "learning_rate": 1e-05,
      "loss": 0.2494,
      "step": 8920
    },
    {
      "epoch": 0.030600426281414267,
      "grad_norm": 1.2197197675704956,
      "learning_rate": 1e-05,
      "loss": 0.2338,
      "step": 8930
    },
    {
      "epoch": 0.030634693276130297,
      "grad_norm": 1.1062088012695312,
      "learning_rate": 1e-05,
      "loss": 0.2157,
      "step": 8940
    },
    {
      "epoch": 0.030668960270846327,
      "grad_norm": 1.107677936553955,
      "learning_rate": 1e-05,
      "loss": 0.2387,
      "step": 8950
    },
    {
      "epoch": 0.030703227265562356,
      "grad_norm": 1.1791733503341675,
      "learning_rate": 1e-05,
      "loss": 0.2337,
      "step": 8960
    },
    {
      "epoch": 0.030737494260278386,
      "grad_norm": 1.1337239742279053,
      "learning_rate": 1e-05,
      "loss": 0.2127,
      "step": 8970
    },
    {
      "epoch": 0.030771761254994416,
      "grad_norm": 1.0418322086334229,
      "learning_rate": 1e-05,
      "loss": 0.222,
      "step": 8980
    },
    {
      "epoch": 0.030806028249710442,
      "grad_norm": 1.1591708660125732,
      "learning_rate": 1e-05,
      "loss": 0.2388,
      "step": 8990
    },
    {
      "epoch": 0.030840295244426472,
      "grad_norm": 1.0103886127471924,
      "learning_rate": 1e-05,
      "loss": 0.2264,
      "step": 9000
    },
    {
      "epoch": 0.030840295244426472,
      "eval_cer": 13.273946022226143,
      "eval_loss": 0.2514854073524475,
      "eval_normalized_cer": 9.492406075139888,
      "eval_runtime": 227.5807,
      "eval_samples_per_second": 2.25,
      "eval_steps_per_second": 0.035,
      "step": 9000
    },
    {
      "epoch": 0.030874562239142502,
      "grad_norm": 1.15034019947052,
      "learning_rate": 1e-05,
      "loss": 0.2215,
      "step": 9010
    },
    {
      "epoch": 0.030908829233858532,
      "grad_norm": 1.183698296546936,
      "learning_rate": 1e-05,
      "loss": 0.2235,
      "step": 9020
    },
    {
      "epoch": 0.030943096228574562,
      "grad_norm": 1.1930736303329468,
      "learning_rate": 1e-05,
      "loss": 0.25,
      "step": 9030
    },
    {
      "epoch": 0.03097736322329059,
      "grad_norm": 1.0650999546051025,
      "learning_rate": 1e-05,
      "loss": 0.224,
      "step": 9040
    },
    {
      "epoch": 0.03101163021800662,
      "grad_norm": 1.0613574981689453,
      "learning_rate": 1e-05,
      "loss": 0.2379,
      "step": 9050
    },
    {
      "epoch": 0.03104589721272265,
      "grad_norm": 1.3004292249679565,
      "learning_rate": 1e-05,
      "loss": 0.2286,
      "step": 9060
    },
    {
      "epoch": 0.031080164207438678,
      "grad_norm": 1.144573450088501,
      "learning_rate": 1e-05,
      "loss": 0.2375,
      "step": 9070
    },
    {
      "epoch": 0.031114431202154708,
      "grad_norm": 1.3552353382110596,
      "learning_rate": 1e-05,
      "loss": 0.216,
      "step": 9080
    },
    {
      "epoch": 0.031148698196870737,
      "grad_norm": 1.139901041984558,
      "learning_rate": 1e-05,
      "loss": 0.2377,
      "step": 9090
    },
    {
      "epoch": 0.031182965191586767,
      "grad_norm": 1.179685354232788,
      "learning_rate": 1e-05,
      "loss": 0.2389,
      "step": 9100
    },
    {
      "epoch": 0.031217232186302797,
      "grad_norm": 1.1494848728179932,
      "learning_rate": 1e-05,
      "loss": 0.2275,
      "step": 9110
    },
    {
      "epoch": 0.03125149918101883,
      "grad_norm": 1.0726871490478516,
      "learning_rate": 1e-05,
      "loss": 0.211,
      "step": 9120
    },
    {
      "epoch": 0.03128576617573486,
      "grad_norm": 1.254655361175537,
      "learning_rate": 1e-05,
      "loss": 0.2332,
      "step": 9130
    },
    {
      "epoch": 0.03132003317045089,
      "grad_norm": 1.1774569749832153,
      "learning_rate": 1e-05,
      "loss": 0.2441,
      "step": 9140
    },
    {
      "epoch": 0.031354300165166916,
      "grad_norm": 1.0810356140136719,
      "learning_rate": 1e-05,
      "loss": 0.2354,
      "step": 9150
    },
    {
      "epoch": 0.031388567159882946,
      "grad_norm": 1.0976495742797852,
      "learning_rate": 1e-05,
      "loss": 0.2347,
      "step": 9160
    },
    {
      "epoch": 0.031422834154598976,
      "grad_norm": 1.2417839765548706,
      "learning_rate": 1e-05,
      "loss": 0.2261,
      "step": 9170
    },
    {
      "epoch": 0.031457101149315006,
      "grad_norm": 1.0022953748703003,
      "learning_rate": 1e-05,
      "loss": 0.2277,
      "step": 9180
    },
    {
      "epoch": 0.03149136814403103,
      "grad_norm": 1.1461567878723145,
      "learning_rate": 1e-05,
      "loss": 0.218,
      "step": 9190
    },
    {
      "epoch": 0.03152563513874706,
      "grad_norm": 1.1877334117889404,
      "learning_rate": 1e-05,
      "loss": 0.2195,
      "step": 9200
    },
    {
      "epoch": 0.03155990213346309,
      "grad_norm": 1.1513786315917969,
      "learning_rate": 1e-05,
      "loss": 0.2354,
      "step": 9210
    },
    {
      "epoch": 0.03159416912817912,
      "grad_norm": 1.057938814163208,
      "learning_rate": 1e-05,
      "loss": 0.2428,
      "step": 9220
    },
    {
      "epoch": 0.03162843612289515,
      "grad_norm": 1.0752719640731812,
      "learning_rate": 1e-05,
      "loss": 0.2499,
      "step": 9230
    },
    {
      "epoch": 0.03166270311761118,
      "grad_norm": 1.1784312725067139,
      "learning_rate": 1e-05,
      "loss": 0.2556,
      "step": 9240
    },
    {
      "epoch": 0.03169697011232721,
      "grad_norm": 1.0176231861114502,
      "learning_rate": 1e-05,
      "loss": 0.2552,
      "step": 9250
    },
    {
      "epoch": 0.03173123710704324,
      "grad_norm": 1.0849392414093018,
      "learning_rate": 1e-05,
      "loss": 0.2483,
      "step": 9260
    },
    {
      "epoch": 0.03176550410175927,
      "grad_norm": 1.0042351484298706,
      "learning_rate": 1e-05,
      "loss": 0.2472,
      "step": 9270
    },
    {
      "epoch": 0.0317997710964753,
      "grad_norm": 1.0480408668518066,
      "learning_rate": 1e-05,
      "loss": 0.238,
      "step": 9280
    },
    {
      "epoch": 0.03183403809119133,
      "grad_norm": 1.0602933168411255,
      "learning_rate": 1e-05,
      "loss": 0.2657,
      "step": 9290
    },
    {
      "epoch": 0.03186830508590736,
      "grad_norm": 1.1704037189483643,
      "learning_rate": 1e-05,
      "loss": 0.2634,
      "step": 9300
    },
    {
      "epoch": 0.03190257208062339,
      "grad_norm": 1.2454304695129395,
      "learning_rate": 1e-05,
      "loss": 0.2648,
      "step": 9310
    },
    {
      "epoch": 0.03193683907533942,
      "grad_norm": 1.0540211200714111,
      "learning_rate": 1e-05,
      "loss": 0.2462,
      "step": 9320
    },
    {
      "epoch": 0.03197110607005545,
      "grad_norm": 1.1440715789794922,
      "learning_rate": 1e-05,
      "loss": 0.2609,
      "step": 9330
    },
    {
      "epoch": 0.03200537306477148,
      "grad_norm": 1.0083932876586914,
      "learning_rate": 1e-05,
      "loss": 0.2422,
      "step": 9340
    },
    {
      "epoch": 0.0320396400594875,
      "grad_norm": 1.0180490016937256,
      "learning_rate": 1e-05,
      "loss": 0.2402,
      "step": 9350
    },
    {
      "epoch": 0.03207390705420353,
      "grad_norm": 1.1158274412155151,
      "learning_rate": 1e-05,
      "loss": 0.2378,
      "step": 9360
    },
    {
      "epoch": 0.03210817404891956,
      "grad_norm": 1.2014826536178589,
      "learning_rate": 1e-05,
      "loss": 0.2433,
      "step": 9370
    },
    {
      "epoch": 0.03214244104363559,
      "grad_norm": 1.1604617834091187,
      "learning_rate": 1e-05,
      "loss": 0.2396,
      "step": 9380
    },
    {
      "epoch": 0.03217670803835162,
      "grad_norm": 1.1088517904281616,
      "learning_rate": 1e-05,
      "loss": 0.2364,
      "step": 9390
    },
    {
      "epoch": 0.03221097503306765,
      "grad_norm": 1.1905596256256104,
      "learning_rate": 1e-05,
      "loss": 0.2335,
      "step": 9400
    },
    {
      "epoch": 0.03224524202778368,
      "grad_norm": 1.2237385511398315,
      "learning_rate": 1e-05,
      "loss": 0.2522,
      "step": 9410
    },
    {
      "epoch": 0.03227950902249971,
      "grad_norm": 1.0582191944122314,
      "learning_rate": 1e-05,
      "loss": 0.2467,
      "step": 9420
    },
    {
      "epoch": 0.03231377601721574,
      "grad_norm": 1.2065699100494385,
      "learning_rate": 1e-05,
      "loss": 0.2359,
      "step": 9430
    },
    {
      "epoch": 0.03234804301193177,
      "grad_norm": 1.1399251222610474,
      "learning_rate": 1e-05,
      "loss": 0.2507,
      "step": 9440
    },
    {
      "epoch": 0.0323823100066478,
      "grad_norm": 1.070038914680481,
      "learning_rate": 1e-05,
      "loss": 0.2732,
      "step": 9450
    },
    {
      "epoch": 0.03241657700136383,
      "grad_norm": 1.0940920114517212,
      "learning_rate": 1e-05,
      "loss": 0.2609,
      "step": 9460
    },
    {
      "epoch": 0.03245084399607986,
      "grad_norm": 1.2042659521102905,
      "learning_rate": 1e-05,
      "loss": 0.2511,
      "step": 9470
    },
    {
      "epoch": 0.03248511099079589,
      "grad_norm": 1.0865731239318848,
      "learning_rate": 1e-05,
      "loss": 0.2693,
      "step": 9480
    },
    {
      "epoch": 0.03251937798551192,
      "grad_norm": 1.0884779691696167,
      "learning_rate": 1e-05,
      "loss": 0.2303,
      "step": 9490
    },
    {
      "epoch": 0.03255364498022795,
      "grad_norm": 1.0557243824005127,
      "learning_rate": 1e-05,
      "loss": 0.2611,
      "step": 9500
    },
    {
      "epoch": 0.03255364498022795,
      "eval_cer": 13.300405715293703,
      "eval_loss": 0.2491595298051834,
      "eval_normalized_cer": 9.722222222222223,
      "eval_runtime": 227.5398,
      "eval_samples_per_second": 2.25,
      "eval_steps_per_second": 0.035,
      "step": 9500
    },
    {
      "epoch": 0.03258791197494398,
      "grad_norm": 1.1241854429244995,
      "learning_rate": 1e-05,
      "loss": 0.2531,
      "step": 9510
    },
    {
      "epoch": 0.03262217896966,
      "grad_norm": 1.091977834701538,
      "learning_rate": 1e-05,
      "loss": 0.2437,
      "step": 9520
    },
    {
      "epoch": 0.03265644596437603,
      "grad_norm": 1.0969996452331543,
      "learning_rate": 1e-05,
      "loss": 0.2595,
      "step": 9530
    },
    {
      "epoch": 0.03269071295909206,
      "grad_norm": 1.1597386598587036,
      "learning_rate": 1e-05,
      "loss": 0.246,
      "step": 9540
    },
    {
      "epoch": 0.03272497995380809,
      "grad_norm": 1.0741667747497559,
      "learning_rate": 1e-05,
      "loss": 0.2435,
      "step": 9550
    },
    {
      "epoch": 0.03275924694852412,
      "grad_norm": 1.1600459814071655,
      "learning_rate": 1e-05,
      "loss": 0.2476,
      "step": 9560
    },
    {
      "epoch": 0.03279351394324015,
      "grad_norm": 1.0636577606201172,
      "learning_rate": 1e-05,
      "loss": 0.2323,
      "step": 9570
    },
    {
      "epoch": 0.03282778093795618,
      "grad_norm": 1.0010998249053955,
      "learning_rate": 1e-05,
      "loss": 0.2436,
      "step": 9580
    },
    {
      "epoch": 0.03286204793267221,
      "grad_norm": 1.1286782026290894,
      "learning_rate": 1e-05,
      "loss": 0.2585,
      "step": 9590
    },
    {
      "epoch": 0.03289631492738824,
      "grad_norm": 1.1575970649719238,
      "learning_rate": 1e-05,
      "loss": 0.2529,
      "step": 9600
    },
    {
      "epoch": 0.03293058192210427,
      "grad_norm": 1.2235829830169678,
      "learning_rate": 1e-05,
      "loss": 0.2617,
      "step": 9610
    },
    {
      "epoch": 0.0329648489168203,
      "grad_norm": 1.203371286392212,
      "learning_rate": 1e-05,
      "loss": 0.2299,
      "step": 9620
    },
    {
      "epoch": 0.03299911591153633,
      "grad_norm": 1.2438814640045166,
      "learning_rate": 1e-05,
      "loss": 0.2544,
      "step": 9630
    },
    {
      "epoch": 0.03303338290625236,
      "grad_norm": 1.2102582454681396,
      "learning_rate": 1e-05,
      "loss": 0.2415,
      "step": 9640
    },
    {
      "epoch": 0.03306764990096839,
      "grad_norm": 0.9984006285667419,
      "learning_rate": 1e-05,
      "loss": 0.2633,
      "step": 9650
    },
    {
      "epoch": 0.03310191689568442,
      "grad_norm": 1.0693376064300537,
      "learning_rate": 1e-05,
      "loss": 0.2394,
      "step": 9660
    },
    {
      "epoch": 0.03313618389040045,
      "grad_norm": 1.2770649194717407,
      "learning_rate": 1e-05,
      "loss": 0.2577,
      "step": 9670
    },
    {
      "epoch": 0.03317045088511647,
      "grad_norm": 1.3751499652862549,
      "learning_rate": 1e-05,
      "loss": 0.2595,
      "step": 9680
    },
    {
      "epoch": 0.0332047178798325,
      "grad_norm": 1.247056484222412,
      "learning_rate": 1e-05,
      "loss": 0.2526,
      "step": 9690
    },
    {
      "epoch": 0.03323898487454853,
      "grad_norm": 1.2248870134353638,
      "learning_rate": 1e-05,
      "loss": 0.2609,
      "step": 9700
    },
    {
      "epoch": 0.03327325186926456,
      "grad_norm": 1.0467997789382935,
      "learning_rate": 1e-05,
      "loss": 0.2353,
      "step": 9710
    },
    {
      "epoch": 0.03330751886398059,
      "grad_norm": 1.2705328464508057,
      "learning_rate": 1e-05,
      "loss": 0.2483,
      "step": 9720
    },
    {
      "epoch": 0.03334178585869662,
      "grad_norm": 1.0360983610153198,
      "learning_rate": 1e-05,
      "loss": 0.2497,
      "step": 9730
    },
    {
      "epoch": 0.03337605285341265,
      "grad_norm": 1.1409697532653809,
      "learning_rate": 1e-05,
      "loss": 0.2506,
      "step": 9740
    },
    {
      "epoch": 0.03341031984812868,
      "grad_norm": 1.0656344890594482,
      "learning_rate": 1e-05,
      "loss": 0.252,
      "step": 9750
    },
    {
      "epoch": 0.03344458684284471,
      "grad_norm": 1.1420108079910278,
      "learning_rate": 1e-05,
      "loss": 0.2383,
      "step": 9760
    },
    {
      "epoch": 0.03347885383756074,
      "grad_norm": 1.0792863368988037,
      "learning_rate": 1e-05,
      "loss": 0.2461,
      "step": 9770
    },
    {
      "epoch": 0.03351312083227677,
      "grad_norm": 1.1297261714935303,
      "learning_rate": 1e-05,
      "loss": 0.2501,
      "step": 9780
    },
    {
      "epoch": 0.0335473878269928,
      "grad_norm": 1.2627495527267456,
      "learning_rate": 1e-05,
      "loss": 0.2452,
      "step": 9790
    },
    {
      "epoch": 0.03358165482170883,
      "grad_norm": 0.982812762260437,
      "learning_rate": 1e-05,
      "loss": 0.2443,
      "step": 9800
    },
    {
      "epoch": 0.03361592181642486,
      "grad_norm": 1.048766016960144,
      "learning_rate": 1e-05,
      "loss": 0.2531,
      "step": 9810
    },
    {
      "epoch": 0.03365018881114089,
      "grad_norm": 1.141780972480774,
      "learning_rate": 1e-05,
      "loss": 0.2271,
      "step": 9820
    },
    {
      "epoch": 0.03368445580585692,
      "grad_norm": 1.101762056350708,
      "learning_rate": 1e-05,
      "loss": 0.2338,
      "step": 9830
    },
    {
      "epoch": 0.03371872280057294,
      "grad_norm": 1.0150196552276611,
      "learning_rate": 1e-05,
      "loss": 0.2404,
      "step": 9840
    },
    {
      "epoch": 0.03375298979528897,
      "grad_norm": 1.1550086736679077,
      "learning_rate": 1e-05,
      "loss": 0.2479,
      "step": 9850
    },
    {
      "epoch": 0.033787256790005,
      "grad_norm": 1.1246519088745117,
      "learning_rate": 1e-05,
      "loss": 0.2525,
      "step": 9860
    },
    {
      "epoch": 0.03382152378472103,
      "grad_norm": 0.9673643708229065,
      "learning_rate": 1e-05,
      "loss": 0.2453,
      "step": 9870
    },
    {
      "epoch": 0.03385579077943706,
      "grad_norm": 1.019649863243103,
      "learning_rate": 1e-05,
      "loss": 0.2407,
      "step": 9880
    },
    {
      "epoch": 0.03389005777415309,
      "grad_norm": 1.1477577686309814,
      "learning_rate": 1e-05,
      "loss": 0.2374,
      "step": 9890
    },
    {
      "epoch": 0.03392432476886912,
      "grad_norm": 0.9760174751281738,
      "learning_rate": 1e-05,
      "loss": 0.248,
      "step": 9900
    },
    {
      "epoch": 0.03395859176358515,
      "grad_norm": 1.1654585599899292,
      "learning_rate": 1e-05,
      "loss": 0.253,
      "step": 9910
    },
    {
      "epoch": 0.03399285875830118,
      "grad_norm": 1.0818895101547241,
      "learning_rate": 1e-05,
      "loss": 0.2397,
      "step": 9920
    },
    {
      "epoch": 0.03402712575301721,
      "grad_norm": 1.1635690927505493,
      "learning_rate": 1e-05,
      "loss": 0.2539,
      "step": 9930
    },
    {
      "epoch": 0.03406139274773324,
      "grad_norm": 1.0819408893585205,
      "learning_rate": 1e-05,
      "loss": 0.2352,
      "step": 9940
    },
    {
      "epoch": 0.03409565974244927,
      "grad_norm": 1.0151749849319458,
      "learning_rate": 1e-05,
      "loss": 0.2378,
      "step": 9950
    },
    {
      "epoch": 0.0341299267371653,
      "grad_norm": 1.092203140258789,
      "learning_rate": 1e-05,
      "loss": 0.2406,
      "step": 9960
    },
    {
      "epoch": 0.03416419373188133,
      "grad_norm": 1.1004047393798828,
      "learning_rate": 1e-05,
      "loss": 0.2375,
      "step": 9970
    },
    {
      "epoch": 0.03419846072659736,
      "grad_norm": 1.2596560716629028,
      "learning_rate": 1e-05,
      "loss": 0.2601,
      "step": 9980
    },
    {
      "epoch": 0.03423272772131339,
      "grad_norm": 1.0936402082443237,
      "learning_rate": 1e-05,
      "loss": 0.2451,
      "step": 9990
    },
    {
      "epoch": 0.03426699471602941,
      "grad_norm": 1.1660488843917847,
      "learning_rate": 1e-05,
      "loss": 0.2523,
      "step": 10000
    },
    {
      "epoch": 0.03426699471602941,
      "eval_cer": 12.656553183983066,
      "eval_loss": 0.24734708666801453,
      "eval_normalized_cer": 8.912869704236611,
      "eval_runtime": 226.9924,
      "eval_samples_per_second": 2.256,
      "eval_steps_per_second": 0.035,
      "step": 10000
    },
    {
      "epoch": 0.03430126171074544,
      "grad_norm": 1.0807442665100098,
      "learning_rate": 1e-05,
      "loss": 0.2583,
      "step": 10010
    },
    {
      "epoch": 0.03433552870546147,
      "grad_norm": 1.17780339717865,
      "learning_rate": 1e-05,
      "loss": 0.2685,
      "step": 10020
    },
    {
      "epoch": 0.0343697957001775,
      "grad_norm": 0.98011314868927,
      "learning_rate": 1e-05,
      "loss": 0.2445,
      "step": 10030
    },
    {
      "epoch": 0.03440406269489353,
      "grad_norm": 1.3025845289230347,
      "learning_rate": 1e-05,
      "loss": 0.2526,
      "step": 10040
    },
    {
      "epoch": 0.03443832968960956,
      "grad_norm": 1.4656189680099487,
      "learning_rate": 1e-05,
      "loss": 0.2618,
      "step": 10050
    },
    {
      "epoch": 0.03447259668432559,
      "grad_norm": 1.2372117042541504,
      "learning_rate": 1e-05,
      "loss": 0.2545,
      "step": 10060
    },
    {
      "epoch": 0.03450686367904162,
      "grad_norm": 1.1028844118118286,
      "learning_rate": 1e-05,
      "loss": 0.2511,
      "step": 10070
    },
    {
      "epoch": 0.03454113067375765,
      "grad_norm": 1.2402809858322144,
      "learning_rate": 1e-05,
      "loss": 0.2507,
      "step": 10080
    },
    {
      "epoch": 0.03457539766847368,
      "grad_norm": 1.1127125024795532,
      "learning_rate": 1e-05,
      "loss": 0.2634,
      "step": 10090
    },
    {
      "epoch": 0.03460966466318971,
      "grad_norm": 1.1486737728118896,
      "learning_rate": 1e-05,
      "loss": 0.2495,
      "step": 10100
    },
    {
      "epoch": 0.03464393165790574,
      "grad_norm": 1.0792872905731201,
      "learning_rate": 1e-05,
      "loss": 0.2347,
      "step": 10110
    },
    {
      "epoch": 0.03467819865262177,
      "grad_norm": 1.0346540212631226,
      "learning_rate": 1e-05,
      "loss": 0.2468,
      "step": 10120
    },
    {
      "epoch": 0.0347124656473378,
      "grad_norm": 1.165614128112793,
      "learning_rate": 1e-05,
      "loss": 0.2419,
      "step": 10130
    },
    {
      "epoch": 0.03474673264205383,
      "grad_norm": 1.1619865894317627,
      "learning_rate": 1e-05,
      "loss": 0.2428,
      "step": 10140
    },
    {
      "epoch": 0.03478099963676986,
      "grad_norm": 1.2419089078903198,
      "learning_rate": 1e-05,
      "loss": 0.2652,
      "step": 10150
    },
    {
      "epoch": 0.03481526663148589,
      "grad_norm": 1.0978246927261353,
      "learning_rate": 1e-05,
      "loss": 0.2396,
      "step": 10160
    },
    {
      "epoch": 0.03484953362620191,
      "grad_norm": 0.9916633367538452,
      "learning_rate": 1e-05,
      "loss": 0.2603,
      "step": 10170
    },
    {
      "epoch": 0.03488380062091794,
      "grad_norm": 1.026318907737732,
      "learning_rate": 1e-05,
      "loss": 0.2564,
      "step": 10180
    },
    {
      "epoch": 0.03491806761563397,
      "grad_norm": 1.1151725053787231,
      "learning_rate": 1e-05,
      "loss": 0.2587,
      "step": 10190
    },
    {
      "epoch": 0.03495233461035,
      "grad_norm": 1.1420174837112427,
      "learning_rate": 1e-05,
      "loss": 0.2695,
      "step": 10200
    },
    {
      "epoch": 0.03498660160506603,
      "grad_norm": 1.0311987400054932,
      "learning_rate": 1e-05,
      "loss": 0.2574,
      "step": 10210
    },
    {
      "epoch": 0.03502086859978206,
      "grad_norm": 0.9759404063224792,
      "learning_rate": 1e-05,
      "loss": 0.2398,
      "step": 10220
    },
    {
      "epoch": 0.03505513559449809,
      "grad_norm": 1.5086033344268799,
      "learning_rate": 1e-05,
      "loss": 0.2589,
      "step": 10230
    },
    {
      "epoch": 0.03508940258921412,
      "grad_norm": 1.167893409729004,
      "learning_rate": 1e-05,
      "loss": 0.2536,
      "step": 10240
    },
    {
      "epoch": 0.03512366958393015,
      "grad_norm": 1.0426411628723145,
      "learning_rate": 1e-05,
      "loss": 0.237,
      "step": 10250
    },
    {
      "epoch": 0.03515793657864618,
      "grad_norm": 1.4766713380813599,
      "learning_rate": 1e-05,
      "loss": 0.2363,
      "step": 10260
    },
    {
      "epoch": 0.03519220357336221,
      "grad_norm": 1.019641399383545,
      "learning_rate": 1e-05,
      "loss": 0.2281,
      "step": 10270
    },
    {
      "epoch": 0.03522647056807824,
      "grad_norm": 1.1627217531204224,
      "learning_rate": 1e-05,
      "loss": 0.2524,
      "step": 10280
    },
    {
      "epoch": 0.03526073756279427,
      "grad_norm": 1.165414571762085,
      "learning_rate": 1e-05,
      "loss": 0.2437,
      "step": 10290
    },
    {
      "epoch": 0.0352950045575103,
      "grad_norm": 1.1816645860671997,
      "learning_rate": 1e-05,
      "loss": 0.2471,
      "step": 10300
    },
    {
      "epoch": 0.03532927155222633,
      "grad_norm": 1.0116764307022095,
      "learning_rate": 1e-05,
      "loss": 0.226,
      "step": 10310
    },
    {
      "epoch": 0.03536353854694236,
      "grad_norm": 1.2257869243621826,
      "learning_rate": 1e-05,
      "loss": 0.2441,
      "step": 10320
    },
    {
      "epoch": 0.03539780554165838,
      "grad_norm": 1.1971989870071411,
      "learning_rate": 1e-05,
      "loss": 0.2268,
      "step": 10330
    },
    {
      "epoch": 0.03543207253637441,
      "grad_norm": 1.057354211807251,
      "learning_rate": 1e-05,
      "loss": 0.2142,
      "step": 10340
    },
    {
      "epoch": 0.03546633953109044,
      "grad_norm": 1.2593644857406616,
      "learning_rate": 1e-05,
      "loss": 0.2149,
      "step": 10350
    },
    {
      "epoch": 0.03550060652580647,
      "grad_norm": 1.0903648138046265,
      "learning_rate": 1e-05,
      "loss": 0.2494,
      "step": 10360
    },
    {
      "epoch": 0.0355348735205225,
      "grad_norm": 1.2079240083694458,
      "learning_rate": 1e-05,
      "loss": 0.2273,
      "step": 10370
    },
    {
      "epoch": 0.03556914051523853,
      "grad_norm": 1.206696629524231,
      "learning_rate": 1e-05,
      "loss": 0.2142,
      "step": 10380
    },
    {
      "epoch": 0.03560340750995456,
      "grad_norm": 1.2176302671432495,
      "learning_rate": 1e-05,
      "loss": 0.2277,
      "step": 10390
    },
    {
      "epoch": 0.03563767450467059,
      "grad_norm": 0.9780252575874329,
      "learning_rate": 1e-05,
      "loss": 0.2089,
      "step": 10400
    },
    {
      "epoch": 0.03567194149938662,
      "grad_norm": 1.1160544157028198,
      "learning_rate": 1e-05,
      "loss": 0.2268,
      "step": 10410
    },
    {
      "epoch": 0.03570620849410265,
      "grad_norm": 1.1231842041015625,
      "learning_rate": 1e-05,
      "loss": 0.2385,
      "step": 10420
    },
    {
      "epoch": 0.03574047548881868,
      "grad_norm": 1.142675518989563,
      "learning_rate": 1e-05,
      "loss": 0.2386,
      "step": 10430
    },
    {
      "epoch": 0.03577474248353471,
      "grad_norm": 1.1563806533813477,
      "learning_rate": 1e-05,
      "loss": 0.2217,
      "step": 10440
    },
    {
      "epoch": 0.03580900947825074,
      "grad_norm": 1.0655251741409302,
      "learning_rate": 1e-05,
      "loss": 0.2207,
      "step": 10450
    },
    {
      "epoch": 0.03584327647296677,
      "grad_norm": 1.1469309329986572,
      "learning_rate": 1e-05,
      "loss": 0.2211,
      "step": 10460
    },
    {
      "epoch": 0.0358775434676828,
      "grad_norm": 1.182896375656128,
      "learning_rate": 1e-05,
      "loss": 0.2316,
      "step": 10470
    },
    {
      "epoch": 0.03591181046239883,
      "grad_norm": 1.018953800201416,
      "learning_rate": 1e-05,
      "loss": 0.2308,
      "step": 10480
    },
    {
      "epoch": 0.03594607745711485,
      "grad_norm": 1.1578072309494019,
      "learning_rate": 1e-05,
      "loss": 0.2272,
      "step": 10490
    },
    {
      "epoch": 0.03598034445183088,
      "grad_norm": 1.1483505964279175,
      "learning_rate": 1e-05,
      "loss": 0.2469,
      "step": 10500
    },
    {
      "epoch": 0.03598034445183088,
      "eval_cer": 13.026988886928912,
      "eval_loss": 0.24956555664539337,
      "eval_normalized_cer": 9.442446043165468,
      "eval_runtime": 227.6026,
      "eval_samples_per_second": 2.25,
      "eval_steps_per_second": 0.035,
      "step": 10500
    },
    {
      "epoch": 0.03601461144654691,
      "grad_norm": 1.1269420385360718,
      "learning_rate": 1e-05,
      "loss": 0.2482,
      "step": 10510
    },
    {
      "epoch": 0.03604887844126294,
      "grad_norm": 1.0399614572525024,
      "learning_rate": 1e-05,
      "loss": 0.2473,
      "step": 10520
    },
    {
      "epoch": 0.03608314543597897,
      "grad_norm": 1.11722731590271,
      "learning_rate": 1e-05,
      "loss": 0.2566,
      "step": 10530
    },
    {
      "epoch": 0.036117412430695,
      "grad_norm": 1.0773251056671143,
      "learning_rate": 1e-05,
      "loss": 0.2479,
      "step": 10540
    },
    {
      "epoch": 0.03615167942541103,
      "grad_norm": 1.0123059749603271,
      "learning_rate": 1e-05,
      "loss": 0.2527,
      "step": 10550
    },
    {
      "epoch": 0.03618594642012706,
      "grad_norm": 1.14670991897583,
      "learning_rate": 1e-05,
      "loss": 0.2474,
      "step": 10560
    },
    {
      "epoch": 0.03622021341484309,
      "grad_norm": 1.021543025970459,
      "learning_rate": 1e-05,
      "loss": 0.2314,
      "step": 10570
    },
    {
      "epoch": 0.03625448040955912,
      "grad_norm": 1.1485329866409302,
      "learning_rate": 1e-05,
      "loss": 0.2603,
      "step": 10580
    },
    {
      "epoch": 0.03628874740427515,
      "grad_norm": 1.239241600036621,
      "learning_rate": 1e-05,
      "loss": 0.2479,
      "step": 10590
    },
    {
      "epoch": 0.03632301439899118,
      "grad_norm": 1.2692322731018066,
      "learning_rate": 1e-05,
      "loss": 0.2411,
      "step": 10600
    },
    {
      "epoch": 0.03635728139370721,
      "grad_norm": 1.1440175771713257,
      "learning_rate": 1e-05,
      "loss": 0.256,
      "step": 10610
    },
    {
      "epoch": 0.03639154838842324,
      "grad_norm": 1.0153664350509644,
      "learning_rate": 1e-05,
      "loss": 0.2468,
      "step": 10620
    },
    {
      "epoch": 0.03642581538313927,
      "grad_norm": 1.0046017169952393,
      "learning_rate": 1e-05,
      "loss": 0.2463,
      "step": 10630
    },
    {
      "epoch": 0.0364600823778553,
      "grad_norm": 1.023366928100586,
      "learning_rate": 1e-05,
      "loss": 0.2499,
      "step": 10640
    },
    {
      "epoch": 0.03649434937257132,
      "grad_norm": 1.1663336753845215,
      "learning_rate": 1e-05,
      "loss": 0.2476,
      "step": 10650
    },
    {
      "epoch": 0.03652861636728735,
      "grad_norm": 1.2302120923995972,
      "learning_rate": 1e-05,
      "loss": 0.2469,
      "step": 10660
    },
    {
      "epoch": 0.03656288336200338,
      "grad_norm": 1.0369274616241455,
      "learning_rate": 1e-05,
      "loss": 0.2468,
      "step": 10670
    },
    {
      "epoch": 0.03659715035671941,
      "grad_norm": 1.3913209438323975,
      "learning_rate": 1e-05,
      "loss": 0.2499,
      "step": 10680
    },
    {
      "epoch": 0.03663141735143544,
      "grad_norm": 1.0193636417388916,
      "learning_rate": 1e-05,
      "loss": 0.25,
      "step": 10690
    },
    {
      "epoch": 0.03666568434615147,
      "grad_norm": 1.2572286128997803,
      "learning_rate": 1e-05,
      "loss": 0.2441,
      "step": 10700
    },
    {
      "epoch": 0.0366999513408675,
      "grad_norm": 1.1679438352584839,
      "learning_rate": 1e-05,
      "loss": 0.2462,
      "step": 10710
    },
    {
      "epoch": 0.03673421833558353,
      "grad_norm": 1.1959030628204346,
      "learning_rate": 1e-05,
      "loss": 0.2641,
      "step": 10720
    },
    {
      "epoch": 0.03676848533029956,
      "grad_norm": 1.1327241659164429,
      "learning_rate": 1e-05,
      "loss": 0.2538,
      "step": 10730
    },
    {
      "epoch": 0.03680275232501559,
      "grad_norm": 1.0999104976654053,
      "learning_rate": 1e-05,
      "loss": 0.2492,
      "step": 10740
    },
    {
      "epoch": 0.03683701931973162,
      "grad_norm": 1.1578527688980103,
      "learning_rate": 1e-05,
      "loss": 0.2506,
      "step": 10750
    },
    {
      "epoch": 0.03687128631444765,
      "grad_norm": 1.243034839630127,
      "learning_rate": 1e-05,
      "loss": 0.2404,
      "step": 10760
    },
    {
      "epoch": 0.03690555330916368,
      "grad_norm": 1.1157968044281006,
      "learning_rate": 1e-05,
      "loss": 0.2597,
      "step": 10770
    },
    {
      "epoch": 0.03693982030387971,
      "grad_norm": 1.0121145248413086,
      "learning_rate": 1e-05,
      "loss": 0.2457,
      "step": 10780
    },
    {
      "epoch": 0.03697408729859574,
      "grad_norm": 1.1467009782791138,
      "learning_rate": 1e-05,
      "loss": 0.2714,
      "step": 10790
    },
    {
      "epoch": 0.03700835429331177,
      "grad_norm": 1.1445188522338867,
      "learning_rate": 1e-05,
      "loss": 0.2652,
      "step": 10800
    },
    {
      "epoch": 0.0370426212880278,
      "grad_norm": 0.9861304759979248,
      "learning_rate": 1e-05,
      "loss": 0.2408,
      "step": 10810
    },
    {
      "epoch": 0.037076888282743824,
      "grad_norm": 1.1524399518966675,
      "learning_rate": 1e-05,
      "loss": 0.2645,
      "step": 10820
    },
    {
      "epoch": 0.037111155277459854,
      "grad_norm": 1.4263722896575928,
      "learning_rate": 1e-05,
      "loss": 0.2364,
      "step": 10830
    },
    {
      "epoch": 0.03714542227217588,
      "grad_norm": 1.057739019393921,
      "learning_rate": 1e-05,
      "loss": 0.2526,
      "step": 10840
    },
    {
      "epoch": 0.03717968926689191,
      "grad_norm": 1.232234239578247,
      "learning_rate": 1e-05,
      "loss": 0.2605,
      "step": 10850
    },
    {
      "epoch": 0.03721395626160794,
      "grad_norm": 1.2277271747589111,
      "learning_rate": 1e-05,
      "loss": 0.2646,
      "step": 10860
    },
    {
      "epoch": 0.03724822325632397,
      "grad_norm": 1.1447609663009644,
      "learning_rate": 1e-05,
      "loss": 0.2508,
      "step": 10870
    },
    {
      "epoch": 0.03728249025104,
      "grad_norm": 1.5821764469146729,
      "learning_rate": 1e-05,
      "loss": 0.259,
      "step": 10880
    },
    {
      "epoch": 0.03731675724575603,
      "grad_norm": 1.0948309898376465,
      "learning_rate": 1e-05,
      "loss": 0.2502,
      "step": 10890
    },
    {
      "epoch": 0.03735102424047206,
      "grad_norm": 1.0639638900756836,
      "learning_rate": 1e-05,
      "loss": 0.2796,
      "step": 10900
    },
    {
      "epoch": 0.03738529123518809,
      "grad_norm": 1.0613322257995605,
      "learning_rate": 1e-05,
      "loss": 0.2606,
      "step": 10910
    },
    {
      "epoch": 0.03741955822990412,
      "grad_norm": 1.1825395822525024,
      "learning_rate": 1e-05,
      "loss": 0.2382,
      "step": 10920
    },
    {
      "epoch": 0.03745382522462015,
      "grad_norm": 1.1198760271072388,
      "learning_rate": 1e-05,
      "loss": 0.2601,
      "step": 10930
    },
    {
      "epoch": 0.03748809221933618,
      "grad_norm": 0.9922842383384705,
      "learning_rate": 1e-05,
      "loss": 0.2567,
      "step": 10940
    },
    {
      "epoch": 0.03752235921405221,
      "grad_norm": 1.0075231790542603,
      "learning_rate": 1e-05,
      "loss": 0.2373,
      "step": 10950
    },
    {
      "epoch": 0.03755662620876824,
      "grad_norm": 1.2739824056625366,
      "learning_rate": 1e-05,
      "loss": 0.2571,
      "step": 10960
    },
    {
      "epoch": 0.03759089320348427,
      "grad_norm": 1.049249529838562,
      "learning_rate": 1e-05,
      "loss": 0.2505,
      "step": 10970
    },
    {
      "epoch": 0.037625160198200294,
      "grad_norm": 1.0130621194839478,
      "learning_rate": 1e-05,
      "loss": 0.246,
      "step": 10980
    },
    {
      "epoch": 0.037659427192916324,
      "grad_norm": 1.143740177154541,
      "learning_rate": 1e-05,
      "loss": 0.2584,
      "step": 10990
    },
    {
      "epoch": 0.037693694187632354,
      "grad_norm": 1.1482734680175781,
      "learning_rate": 1e-05,
      "loss": 0.245,
      "step": 11000
    },
    {
      "epoch": 0.037693694187632354,
      "eval_cer": 13.150467454577527,
      "eval_loss": 0.24854739010334015,
      "eval_normalized_cer": 9.492406075139888,
      "eval_runtime": 227.9,
      "eval_samples_per_second": 2.247,
      "eval_steps_per_second": 0.035,
      "step": 11000
    },
    {
      "epoch": 0.037727961182348384,
      "grad_norm": 1.175059199333191,
      "learning_rate": 1e-05,
      "loss": 0.235,
      "step": 11010
    },
    {
      "epoch": 0.037762228177064414,
      "grad_norm": 1.06391441822052,
      "learning_rate": 1e-05,
      "loss": 0.2548,
      "step": 11020
    },
    {
      "epoch": 0.03779649517178044,
      "grad_norm": 1.1072652339935303,
      "learning_rate": 1e-05,
      "loss": 0.2563,
      "step": 11030
    },
    {
      "epoch": 0.03783076216649647,
      "grad_norm": 1.0364381074905396,
      "learning_rate": 1e-05,
      "loss": 0.2353,
      "step": 11040
    },
    {
      "epoch": 0.0378650291612125,
      "grad_norm": 1.120888113975525,
      "learning_rate": 1e-05,
      "loss": 0.2564,
      "step": 11050
    },
    {
      "epoch": 0.03789929615592853,
      "grad_norm": 1.1769522428512573,
      "learning_rate": 1e-05,
      "loss": 0.2485,
      "step": 11060
    },
    {
      "epoch": 0.03793356315064456,
      "grad_norm": 1.103209137916565,
      "learning_rate": 1e-05,
      "loss": 0.2579,
      "step": 11070
    },
    {
      "epoch": 0.03796783014536059,
      "grad_norm": 1.191256046295166,
      "learning_rate": 1e-05,
      "loss": 0.258,
      "step": 11080
    },
    {
      "epoch": 0.03800209714007662,
      "grad_norm": 1.103756070137024,
      "learning_rate": 1e-05,
      "loss": 0.2542,
      "step": 11090
    },
    {
      "epoch": 0.03803636413479265,
      "grad_norm": 1.2097468376159668,
      "learning_rate": 1e-05,
      "loss": 0.2581,
      "step": 11100
    },
    {
      "epoch": 0.03807063112950868,
      "grad_norm": 1.0787367820739746,
      "learning_rate": 1e-05,
      "loss": 0.2428,
      "step": 11110
    },
    {
      "epoch": 0.03810489812422471,
      "grad_norm": 1.0501831769943237,
      "learning_rate": 1e-05,
      "loss": 0.2585,
      "step": 11120
    },
    {
      "epoch": 0.03813916511894074,
      "grad_norm": 1.1210603713989258,
      "learning_rate": 1e-05,
      "loss": 0.2435,
      "step": 11130
    },
    {
      "epoch": 0.038173432113656765,
      "grad_norm": 1.0846295356750488,
      "learning_rate": 1e-05,
      "loss": 0.2525,
      "step": 11140
    },
    {
      "epoch": 0.038207699108372795,
      "grad_norm": 1.176488995552063,
      "learning_rate": 1e-05,
      "loss": 0.2499,
      "step": 11150
    },
    {
      "epoch": 0.038241966103088824,
      "grad_norm": 1.034157395362854,
      "learning_rate": 1e-05,
      "loss": 0.2701,
      "step": 11160
    },
    {
      "epoch": 0.038276233097804854,
      "grad_norm": 1.0763425827026367,
      "learning_rate": 1e-05,
      "loss": 0.2327,
      "step": 11170
    },
    {
      "epoch": 0.038310500092520884,
      "grad_norm": 1.010324478149414,
      "learning_rate": 1e-05,
      "loss": 0.2506,
      "step": 11180
    },
    {
      "epoch": 0.038344767087236914,
      "grad_norm": 1.0999796390533447,
      "learning_rate": 1e-05,
      "loss": 0.2482,
      "step": 11190
    },
    {
      "epoch": 0.038379034081952944,
      "grad_norm": 1.2314294576644897,
      "learning_rate": 1e-05,
      "loss": 0.2413,
      "step": 11200
    },
    {
      "epoch": 0.038413301076668974,
      "grad_norm": 1.109063982963562,
      "learning_rate": 1e-05,
      "loss": 0.2433,
      "step": 11210
    },
    {
      "epoch": 0.038447568071385003,
      "grad_norm": 1.0760470628738403,
      "learning_rate": 1e-05,
      "loss": 0.2572,
      "step": 11220
    },
    {
      "epoch": 0.03848183506610103,
      "grad_norm": 1.2139952182769775,
      "learning_rate": 1e-05,
      "loss": 0.2542,
      "step": 11230
    },
    {
      "epoch": 0.03851610206081706,
      "grad_norm": 1.0825960636138916,
      "learning_rate": 1e-05,
      "loss": 0.2446,
      "step": 11240
    },
    {
      "epoch": 0.03855036905553309,
      "grad_norm": 1.1650110483169556,
      "learning_rate": 1e-05,
      "loss": 0.2486,
      "step": 11250
    },
    {
      "epoch": 0.03858463605024912,
      "grad_norm": 1.074236512184143,
      "learning_rate": 1e-05,
      "loss": 0.2623,
      "step": 11260
    },
    {
      "epoch": 0.03861890304496515,
      "grad_norm": 1.0651731491088867,
      "learning_rate": 1e-05,
      "loss": 0.2409,
      "step": 11270
    },
    {
      "epoch": 0.03865317003968118,
      "grad_norm": 1.0689282417297363,
      "learning_rate": 1e-05,
      "loss": 0.2329,
      "step": 11280
    },
    {
      "epoch": 0.03868743703439721,
      "grad_norm": 1.1548572778701782,
      "learning_rate": 1e-05,
      "loss": 0.2572,
      "step": 11290
    },
    {
      "epoch": 0.038721704029113235,
      "grad_norm": 1.1205992698669434,
      "learning_rate": 1e-05,
      "loss": 0.2424,
      "step": 11300
    },
    {
      "epoch": 0.038755971023829265,
      "grad_norm": 1.0806999206542969,
      "learning_rate": 1e-05,
      "loss": 0.2336,
      "step": 11310
    },
    {
      "epoch": 0.038790238018545295,
      "grad_norm": 1.0330880880355835,
      "learning_rate": 1e-05,
      "loss": 0.2371,
      "step": 11320
    },
    {
      "epoch": 0.038824505013261325,
      "grad_norm": 1.2580816745758057,
      "learning_rate": 1e-05,
      "loss": 0.2297,
      "step": 11330
    },
    {
      "epoch": 0.038858772007977355,
      "grad_norm": 1.3312656879425049,
      "learning_rate": 1e-05,
      "loss": 0.2478,
      "step": 11340
    },
    {
      "epoch": 0.038893039002693384,
      "grad_norm": 1.0043836832046509,
      "learning_rate": 1e-05,
      "loss": 0.2388,
      "step": 11350
    },
    {
      "epoch": 0.038927305997409414,
      "grad_norm": 1.0665231943130493,
      "learning_rate": 1e-05,
      "loss": 0.234,
      "step": 11360
    },
    {
      "epoch": 0.038961572992125444,
      "grad_norm": 1.1114041805267334,
      "learning_rate": 1e-05,
      "loss": 0.2407,
      "step": 11370
    },
    {
      "epoch": 0.038995839986841474,
      "grad_norm": 1.043134093284607,
      "learning_rate": 1e-05,
      "loss": 0.2352,
      "step": 11380
    },
    {
      "epoch": 0.039030106981557504,
      "grad_norm": 1.1435351371765137,
      "learning_rate": 1e-05,
      "loss": 0.2347,
      "step": 11390
    },
    {
      "epoch": 0.039064373976273534,
      "grad_norm": 1.2625036239624023,
      "learning_rate": 1e-05,
      "loss": 0.2515,
      "step": 11400
    },
    {
      "epoch": 0.039098640970989564,
      "grad_norm": 1.101953148841858,
      "learning_rate": 1e-05,
      "loss": 0.2354,
      "step": 11410
    },
    {
      "epoch": 0.03913290796570559,
      "grad_norm": 1.0906771421432495,
      "learning_rate": 1e-05,
      "loss": 0.2145,
      "step": 11420
    },
    {
      "epoch": 0.03916717496042162,
      "grad_norm": 1.2161511182785034,
      "learning_rate": 1e-05,
      "loss": 0.2366,
      "step": 11430
    },
    {
      "epoch": 0.03920144195513765,
      "grad_norm": 1.0494539737701416,
      "learning_rate": 1e-05,
      "loss": 0.2445,
      "step": 11440
    },
    {
      "epoch": 0.03923570894985368,
      "grad_norm": 1.098476529121399,
      "learning_rate": 1e-05,
      "loss": 0.2392,
      "step": 11450
    },
    {
      "epoch": 0.03926997594456971,
      "grad_norm": 1.1904308795928955,
      "learning_rate": 1e-05,
      "loss": 0.2452,
      "step": 11460
    },
    {
      "epoch": 0.039304242939285736,
      "grad_norm": 1.1967114210128784,
      "learning_rate": 1e-05,
      "loss": 0.2489,
      "step": 11470
    },
    {
      "epoch": 0.039338509934001765,
      "grad_norm": 1.2335082292556763,
      "learning_rate": 1e-05,
      "loss": 0.2428,
      "step": 11480
    },
    {
      "epoch": 0.039372776928717795,
      "grad_norm": 1.0949335098266602,
      "learning_rate": 1e-05,
      "loss": 0.2433,
      "step": 11490
    },
    {
      "epoch": 0.039407043923433825,
      "grad_norm": 0.9655303955078125,
      "learning_rate": 1e-05,
      "loss": 0.2298,
      "step": 11500
    },
    {
      "epoch": 0.039407043923433825,
      "eval_cer": 13.238666431469396,
      "eval_loss": 0.24733339250087738,
      "eval_normalized_cer": 9.532374100719425,
      "eval_runtime": 228.8524,
      "eval_samples_per_second": 2.237,
      "eval_steps_per_second": 0.035,
      "step": 11500
    },
    {
      "epoch": 0.039441310918149855,
      "grad_norm": 1.303244709968567,
      "learning_rate": 1e-05,
      "loss": 0.2312,
      "step": 11510
    },
    {
      "epoch": 0.039475577912865885,
      "grad_norm": 1.2272698879241943,
      "learning_rate": 1e-05,
      "loss": 0.2371,
      "step": 11520
    },
    {
      "epoch": 0.039509844907581915,
      "grad_norm": 1.2555683851242065,
      "learning_rate": 1e-05,
      "loss": 0.248,
      "step": 11530
    },
    {
      "epoch": 0.039544111902297945,
      "grad_norm": 1.2972025871276855,
      "learning_rate": 1e-05,
      "loss": 0.2637,
      "step": 11540
    },
    {
      "epoch": 0.039578378897013974,
      "grad_norm": 0.9733885526657104,
      "learning_rate": 1e-05,
      "loss": 0.2397,
      "step": 11550
    },
    {
      "epoch": 0.039612645891730004,
      "grad_norm": 1.1637110710144043,
      "learning_rate": 1e-05,
      "loss": 0.2468,
      "step": 11560
    },
    {
      "epoch": 0.039646912886446034,
      "grad_norm": 1.2355756759643555,
      "learning_rate": 1e-05,
      "loss": 0.2407,
      "step": 11570
    },
    {
      "epoch": 0.039681179881162064,
      "grad_norm": 1.1079312562942505,
      "learning_rate": 1e-05,
      "loss": 0.2511,
      "step": 11580
    },
    {
      "epoch": 0.039715446875878094,
      "grad_norm": 1.13614821434021,
      "learning_rate": 1e-05,
      "loss": 0.2442,
      "step": 11590
    },
    {
      "epoch": 0.039749713870594124,
      "grad_norm": 1.2050237655639648,
      "learning_rate": 1e-05,
      "loss": 0.2505,
      "step": 11600
    },
    {
      "epoch": 0.03978398086531015,
      "grad_norm": 1.020393967628479,
      "learning_rate": 1e-05,
      "loss": 0.2305,
      "step": 11610
    },
    {
      "epoch": 0.03981824786002618,
      "grad_norm": 1.088463544845581,
      "learning_rate": 1e-05,
      "loss": 0.2542,
      "step": 11620
    },
    {
      "epoch": 0.039852514854742206,
      "grad_norm": 1.121472716331482,
      "learning_rate": 1e-05,
      "loss": 0.2489,
      "step": 11630
    },
    {
      "epoch": 0.039886781849458236,
      "grad_norm": 1.109485149383545,
      "learning_rate": 1e-05,
      "loss": 0.2488,
      "step": 11640
    },
    {
      "epoch": 0.039921048844174266,
      "grad_norm": 1.0824321508407593,
      "learning_rate": 1e-05,
      "loss": 0.2395,
      "step": 11650
    },
    {
      "epoch": 0.039955315838890296,
      "grad_norm": 1.1438390016555786,
      "learning_rate": 1e-05,
      "loss": 0.2318,
      "step": 11660
    },
    {
      "epoch": 0.039989582833606326,
      "grad_norm": 1.005821943283081,
      "learning_rate": 1e-05,
      "loss": 0.2482,
      "step": 11670
    },
    {
      "epoch": 0.040023849828322355,
      "grad_norm": 1.187921166419983,
      "learning_rate": 1e-05,
      "loss": 0.2479,
      "step": 11680
    },
    {
      "epoch": 0.040058116823038385,
      "grad_norm": 1.0112143754959106,
      "learning_rate": 1e-05,
      "loss": 0.2555,
      "step": 11690
    },
    {
      "epoch": 0.040092383817754415,
      "grad_norm": 1.1568365097045898,
      "learning_rate": 1e-05,
      "loss": 0.2483,
      "step": 11700
    },
    {
      "epoch": 0.040126650812470445,
      "grad_norm": 1.1201119422912598,
      "learning_rate": 1e-05,
      "loss": 0.256,
      "step": 11710
    },
    {
      "epoch": 0.040160917807186475,
      "grad_norm": 1.1281081438064575,
      "learning_rate": 1e-05,
      "loss": 0.2462,
      "step": 11720
    },
    {
      "epoch": 0.040195184801902505,
      "grad_norm": 0.9730721116065979,
      "learning_rate": 1e-05,
      "loss": 0.2534,
      "step": 11730
    },
    {
      "epoch": 0.040229451796618534,
      "grad_norm": 1.1069108247756958,
      "learning_rate": 1e-05,
      "loss": 0.2497,
      "step": 11740
    },
    {
      "epoch": 0.040263718791334564,
      "grad_norm": 1.1095935106277466,
      "learning_rate": 1e-05,
      "loss": 0.2448,
      "step": 11750
    },
    {
      "epoch": 0.040297985786050594,
      "grad_norm": 0.9690611958503723,
      "learning_rate": 1e-05,
      "loss": 0.2448,
      "step": 11760
    },
    {
      "epoch": 0.040332252780766624,
      "grad_norm": 1.0263612270355225,
      "learning_rate": 1e-05,
      "loss": 0.2555,
      "step": 11770
    },
    {
      "epoch": 0.040366519775482654,
      "grad_norm": 1.0731168985366821,
      "learning_rate": 1e-05,
      "loss": 0.2321,
      "step": 11780
    },
    {
      "epoch": 0.04040078677019868,
      "grad_norm": 1.1446433067321777,
      "learning_rate": 1e-05,
      "loss": 0.239,
      "step": 11790
    },
    {
      "epoch": 0.04043505376491471,
      "grad_norm": 1.0776352882385254,
      "learning_rate": 1e-05,
      "loss": 0.2255,
      "step": 11800
    },
    {
      "epoch": 0.040469320759630736,
      "grad_norm": 0.9721156358718872,
      "learning_rate": 1e-05,
      "loss": 0.2234,
      "step": 11810
    },
    {
      "epoch": 0.040503587754346766,
      "grad_norm": 0.9534703493118286,
      "learning_rate": 1e-05,
      "loss": 0.2163,
      "step": 11820
    },
    {
      "epoch": 0.040537854749062796,
      "grad_norm": 1.0248794555664062,
      "learning_rate": 1e-05,
      "loss": 0.2406,
      "step": 11830
    },
    {
      "epoch": 0.040572121743778826,
      "grad_norm": 1.1740145683288574,
      "learning_rate": 1e-05,
      "loss": 0.2394,
      "step": 11840
    },
    {
      "epoch": 0.040606388738494856,
      "grad_norm": 1.1622172594070435,
      "learning_rate": 1e-05,
      "loss": 0.2387,
      "step": 11850
    },
    {
      "epoch": 0.040640655733210886,
      "grad_norm": 1.0684759616851807,
      "learning_rate": 1e-05,
      "loss": 0.2196,
      "step": 11860
    },
    {
      "epoch": 0.040674922727926915,
      "grad_norm": 1.024851679801941,
      "learning_rate": 1e-05,
      "loss": 0.2178,
      "step": 11870
    },
    {
      "epoch": 0.040709189722642945,
      "grad_norm": 1.2293421030044556,
      "learning_rate": 1e-05,
      "loss": 0.2372,
      "step": 11880
    },
    {
      "epoch": 0.040743456717358975,
      "grad_norm": 1.2226061820983887,
      "learning_rate": 1e-05,
      "loss": 0.2466,
      "step": 11890
    },
    {
      "epoch": 0.040777723712075005,
      "grad_norm": 1.0775419473648071,
      "learning_rate": 1e-05,
      "loss": 0.221,
      "step": 11900
    },
    {
      "epoch": 0.040811990706791035,
      "grad_norm": 1.0354384183883667,
      "learning_rate": 1e-05,
      "loss": 0.2338,
      "step": 11910
    },
    {
      "epoch": 0.040846257701507065,
      "grad_norm": 0.9725399613380432,
      "learning_rate": 1e-05,
      "loss": 0.2312,
      "step": 11920
    },
    {
      "epoch": 0.040880524696223094,
      "grad_norm": 0.9638645648956299,
      "learning_rate": 1e-05,
      "loss": 0.2238,
      "step": 11930
    },
    {
      "epoch": 0.040914791690939124,
      "grad_norm": 1.1646082401275635,
      "learning_rate": 1e-05,
      "loss": 0.2339,
      "step": 11940
    },
    {
      "epoch": 0.04094905868565515,
      "grad_norm": 1.049614429473877,
      "learning_rate": 1e-05,
      "loss": 0.229,
      "step": 11950
    },
    {
      "epoch": 0.04098332568037118,
      "grad_norm": 1.1187442541122437,
      "learning_rate": 1e-05,
      "loss": 0.2437,
      "step": 11960
    },
    {
      "epoch": 0.04101759267508721,
      "grad_norm": 1.1472731828689575,
      "learning_rate": 1e-05,
      "loss": 0.2416,
      "step": 11970
    },
    {
      "epoch": 0.04105185966980324,
      "grad_norm": 1.0694329738616943,
      "learning_rate": 1e-05,
      "loss": 0.2308,
      "step": 11980
    },
    {
      "epoch": 0.04108612666451927,
      "grad_norm": 0.9863060116767883,
      "learning_rate": 1e-05,
      "loss": 0.2258,
      "step": 11990
    },
    {
      "epoch": 0.041120393659235296,
      "grad_norm": 1.1150392293930054,
      "learning_rate": 1e-05,
      "loss": 0.2342,
      "step": 12000
    },
    {
      "epoch": 0.041120393659235296,
      "eval_cer": 12.797671547010056,
      "eval_loss": 0.2504700720310211,
      "eval_normalized_cer": 9.152677857713828,
      "eval_runtime": 227.2028,
      "eval_samples_per_second": 2.253,
      "eval_steps_per_second": 0.035,
      "step": 12000
    },
    {
      "epoch": 0.041154660653951326,
      "grad_norm": 1.0455725193023682,
      "learning_rate": 1e-05,
      "loss": 0.2263,
      "step": 12010
    },
    {
      "epoch": 0.041188927648667356,
      "grad_norm": 1.2993946075439453,
      "learning_rate": 1e-05,
      "loss": 0.2273,
      "step": 12020
    },
    {
      "epoch": 0.041223194643383386,
      "grad_norm": 1.159058690071106,
      "learning_rate": 1e-05,
      "loss": 0.2259,
      "step": 12030
    },
    {
      "epoch": 0.041257461638099416,
      "grad_norm": 1.1908732652664185,
      "learning_rate": 1e-05,
      "loss": 0.2489,
      "step": 12040
    },
    {
      "epoch": 0.041291728632815446,
      "grad_norm": 1.3122719526290894,
      "learning_rate": 1e-05,
      "loss": 0.2544,
      "step": 12050
    },
    {
      "epoch": 0.041325995627531475,
      "grad_norm": 1.1521992683410645,
      "learning_rate": 1e-05,
      "loss": 0.2496,
      "step": 12060
    },
    {
      "epoch": 0.041360262622247505,
      "grad_norm": 1.1007260084152222,
      "learning_rate": 1e-05,
      "loss": 0.2453,
      "step": 12070
    },
    {
      "epoch": 0.041394529616963535,
      "grad_norm": 1.1719632148742676,
      "learning_rate": 1e-05,
      "loss": 0.2419,
      "step": 12080
    },
    {
      "epoch": 0.041428796611679565,
      "grad_norm": 1.088536024093628,
      "learning_rate": 1e-05,
      "loss": 0.2408,
      "step": 12090
    },
    {
      "epoch": 0.041463063606395595,
      "grad_norm": 1.1596314907073975,
      "learning_rate": 1e-05,
      "loss": 0.2463,
      "step": 12100
    },
    {
      "epoch": 0.041497330601111625,
      "grad_norm": 1.1144077777862549,
      "learning_rate": 1e-05,
      "loss": 0.2411,
      "step": 12110
    },
    {
      "epoch": 0.04153159759582765,
      "grad_norm": 1.2945681810379028,
      "learning_rate": 1e-05,
      "loss": 0.2381,
      "step": 12120
    },
    {
      "epoch": 0.04156586459054368,
      "grad_norm": 1.0997275114059448,
      "learning_rate": 1e-05,
      "loss": 0.2704,
      "step": 12130
    },
    {
      "epoch": 0.04160013158525971,
      "grad_norm": 1.0945874452590942,
      "learning_rate": 1e-05,
      "loss": 0.2249,
      "step": 12140
    },
    {
      "epoch": 0.04163439857997574,
      "grad_norm": 1.2051665782928467,
      "learning_rate": 1e-05,
      "loss": 0.2442,
      "step": 12150
    },
    {
      "epoch": 0.04166866557469177,
      "grad_norm": 0.9178060293197632,
      "learning_rate": 1e-05,
      "loss": 0.2519,
      "step": 12160
    },
    {
      "epoch": 0.0417029325694078,
      "grad_norm": 1.1225532293319702,
      "learning_rate": 1e-05,
      "loss": 0.2425,
      "step": 12170
    },
    {
      "epoch": 0.04173719956412383,
      "grad_norm": 1.0947092771530151,
      "learning_rate": 1e-05,
      "loss": 0.2305,
      "step": 12180
    },
    {
      "epoch": 0.041771466558839856,
      "grad_norm": 1.0374338626861572,
      "learning_rate": 1e-05,
      "loss": 0.2445,
      "step": 12190
    },
    {
      "epoch": 0.041805733553555886,
      "grad_norm": 1.1471805572509766,
      "learning_rate": 1e-05,
      "loss": 0.2447,
      "step": 12200
    },
    {
      "epoch": 0.041840000548271916,
      "grad_norm": 1.1241774559020996,
      "learning_rate": 1e-05,
      "loss": 0.2235,
      "step": 12210
    },
    {
      "epoch": 0.041874267542987946,
      "grad_norm": 1.243691086769104,
      "learning_rate": 1e-05,
      "loss": 0.2409,
      "step": 12220
    },
    {
      "epoch": 0.041908534537703976,
      "grad_norm": 1.1077616214752197,
      "learning_rate": 1e-05,
      "loss": 0.2533,
      "step": 12230
    },
    {
      "epoch": 0.041942801532420006,
      "grad_norm": 1.0907562971115112,
      "learning_rate": 1e-05,
      "loss": 0.2443,
      "step": 12240
    },
    {
      "epoch": 0.041977068527136036,
      "grad_norm": 1.3562718629837036,
      "learning_rate": 1e-05,
      "loss": 0.2523,
      "step": 12250
    },
    {
      "epoch": 0.042011335521852065,
      "grad_norm": 1.0229142904281616,
      "learning_rate": 1e-05,
      "loss": 0.2328,
      "step": 12260
    },
    {
      "epoch": 0.042045602516568095,
      "grad_norm": 1.0843278169631958,
      "learning_rate": 1e-05,
      "loss": 0.2328,
      "step": 12270
    },
    {
      "epoch": 0.04207986951128412,
      "grad_norm": 1.2029650211334229,
      "learning_rate": 1e-05,
      "loss": 0.2458,
      "step": 12280
    },
    {
      "epoch": 0.04211413650600015,
      "grad_norm": 0.9493764638900757,
      "learning_rate": 1e-05,
      "loss": 0.2232,
      "step": 12290
    },
    {
      "epoch": 0.04214840350071618,
      "grad_norm": 1.2031728029251099,
      "learning_rate": 1e-05,
      "loss": 0.2473,
      "step": 12300
    },
    {
      "epoch": 0.04218267049543221,
      "grad_norm": 1.1091227531433105,
      "learning_rate": 1e-05,
      "loss": 0.2429,
      "step": 12310
    },
    {
      "epoch": 0.04221693749014824,
      "grad_norm": 1.1729086637496948,
      "learning_rate": 1e-05,
      "loss": 0.2594,
      "step": 12320
    },
    {
      "epoch": 0.04225120448486427,
      "grad_norm": 1.0592730045318604,
      "learning_rate": 1e-05,
      "loss": 0.2506,
      "step": 12330
    },
    {
      "epoch": 0.0422854714795803,
      "grad_norm": 1.2366282939910889,
      "learning_rate": 1e-05,
      "loss": 0.2457,
      "step": 12340
    },
    {
      "epoch": 0.04231973847429633,
      "grad_norm": 1.12427818775177,
      "learning_rate": 1e-05,
      "loss": 0.246,
      "step": 12350
    },
    {
      "epoch": 0.04235400546901236,
      "grad_norm": 1.1663504838943481,
      "learning_rate": 1e-05,
      "loss": 0.2491,
      "step": 12360
    },
    {
      "epoch": 0.04238827246372839,
      "grad_norm": 1.2383378744125366,
      "learning_rate": 1e-05,
      "loss": 0.2492,
      "step": 12370
    },
    {
      "epoch": 0.04242253945844442,
      "grad_norm": 1.184813380241394,
      "learning_rate": 1e-05,
      "loss": 0.2498,
      "step": 12380
    },
    {
      "epoch": 0.042456806453160446,
      "grad_norm": 1.035650610923767,
      "learning_rate": 1e-05,
      "loss": 0.2317,
      "step": 12390
    },
    {
      "epoch": 0.042491073447876476,
      "grad_norm": 1.0495967864990234,
      "learning_rate": 1e-05,
      "loss": 0.2467,
      "step": 12400
    },
    {
      "epoch": 0.042525340442592506,
      "grad_norm": 1.0791754722595215,
      "learning_rate": 1e-05,
      "loss": 0.245,
      "step": 12410
    },
    {
      "epoch": 0.042559607437308536,
      "grad_norm": 1.1513383388519287,
      "learning_rate": 1e-05,
      "loss": 0.2599,
      "step": 12420
    },
    {
      "epoch": 0.042593874432024566,
      "grad_norm": 1.1093658208847046,
      "learning_rate": 1e-05,
      "loss": 0.2584,
      "step": 12430
    },
    {
      "epoch": 0.04262814142674059,
      "grad_norm": 1.0504255294799805,
      "learning_rate": 1e-05,
      "loss": 0.2355,
      "step": 12440
    },
    {
      "epoch": 0.04266240842145662,
      "grad_norm": 1.121837854385376,
      "learning_rate": 1e-05,
      "loss": 0.2309,
      "step": 12450
    },
    {
      "epoch": 0.04269667541617265,
      "grad_norm": 1.2266592979431152,
      "learning_rate": 1e-05,
      "loss": 0.2384,
      "step": 12460
    },
    {
      "epoch": 0.04273094241088868,
      "grad_norm": 1.0912328958511353,
      "learning_rate": 1e-05,
      "loss": 0.2392,
      "step": 12470
    },
    {
      "epoch": 0.04276520940560471,
      "grad_norm": 1.1504424810409546,
      "learning_rate": 1e-05,
      "loss": 0.2716,
      "step": 12480
    },
    {
      "epoch": 0.04279947640032074,
      "grad_norm": 1.011088490486145,
      "learning_rate": 1e-05,
      "loss": 0.2501,
      "step": 12490
    },
    {
      "epoch": 0.04283374339503677,
      "grad_norm": 1.2423217296600342,
      "learning_rate": 1e-05,
      "loss": 0.2562,
      "step": 12500
    },
    {
      "epoch": 0.04283374339503677,
      "eval_cer": 12.92115011465867,
      "eval_loss": 0.24911069869995117,
      "eval_normalized_cer": 9.452438049560353,
      "eval_runtime": 228.1651,
      "eval_samples_per_second": 2.244,
      "eval_steps_per_second": 0.035,
      "step": 12500
    },
    {
      "epoch": 0.0428680103897528,
      "grad_norm": 1.4834926128387451,
      "learning_rate": 1e-05,
      "loss": 0.2354,
      "step": 12510
    },
    {
      "epoch": 0.04290227738446883,
      "grad_norm": 1.26629638671875,
      "learning_rate": 1e-05,
      "loss": 0.2569,
      "step": 12520
    },
    {
      "epoch": 0.04293654437918486,
      "grad_norm": 1.204516053199768,
      "learning_rate": 1e-05,
      "loss": 0.2502,
      "step": 12530
    },
    {
      "epoch": 0.04297081137390089,
      "grad_norm": 1.0527433156967163,
      "learning_rate": 1e-05,
      "loss": 0.2503,
      "step": 12540
    },
    {
      "epoch": 0.04300507836861692,
      "grad_norm": 1.0310479402542114,
      "learning_rate": 1e-05,
      "loss": 0.2289,
      "step": 12550
    },
    {
      "epoch": 0.04303934536333295,
      "grad_norm": 1.2252111434936523,
      "learning_rate": 1e-05,
      "loss": 0.2485,
      "step": 12560
    },
    {
      "epoch": 0.04307361235804898,
      "grad_norm": 1.0729095935821533,
      "learning_rate": 1e-05,
      "loss": 0.2302,
      "step": 12570
    },
    {
      "epoch": 0.043107879352765006,
      "grad_norm": 1.000106930732727,
      "learning_rate": 1e-05,
      "loss": 0.2192,
      "step": 12580
    },
    {
      "epoch": 0.043142146347481036,
      "grad_norm": 1.0674782991409302,
      "learning_rate": 1e-05,
      "loss": 0.2334,
      "step": 12590
    },
    {
      "epoch": 0.04317641334219706,
      "grad_norm": 1.1148403882980347,
      "learning_rate": 1e-05,
      "loss": 0.2402,
      "step": 12600
    },
    {
      "epoch": 0.04321068033691309,
      "grad_norm": 1.0144375562667847,
      "learning_rate": 1e-05,
      "loss": 0.2492,
      "step": 12610
    },
    {
      "epoch": 0.04324494733162912,
      "grad_norm": 1.1123058795928955,
      "learning_rate": 1e-05,
      "loss": 0.2334,
      "step": 12620
    },
    {
      "epoch": 0.04327921432634515,
      "grad_norm": 1.1008777618408203,
      "learning_rate": 1e-05,
      "loss": 0.2319,
      "step": 12630
    },
    {
      "epoch": 0.04331348132106118,
      "grad_norm": 1.1487098932266235,
      "learning_rate": 1e-05,
      "loss": 0.2451,
      "step": 12640
    },
    {
      "epoch": 0.04334774831577721,
      "grad_norm": 1.1339664459228516,
      "learning_rate": 1e-05,
      "loss": 0.2267,
      "step": 12650
    },
    {
      "epoch": 0.04338201531049324,
      "grad_norm": 1.198195219039917,
      "learning_rate": 1e-05,
      "loss": 0.2472,
      "step": 12660
    },
    {
      "epoch": 0.04341628230520927,
      "grad_norm": 0.9989431500434875,
      "learning_rate": 1e-05,
      "loss": 0.2381,
      "step": 12670
    },
    {
      "epoch": 0.0434505492999253,
      "grad_norm": 1.4252516031265259,
      "learning_rate": 1e-05,
      "loss": 0.2513,
      "step": 12680
    },
    {
      "epoch": 0.04348481629464133,
      "grad_norm": 1.1313762664794922,
      "learning_rate": 1e-05,
      "loss": 0.2591,
      "step": 12690
    },
    {
      "epoch": 0.04351908328935736,
      "grad_norm": 1.0512256622314453,
      "learning_rate": 1e-05,
      "loss": 0.238,
      "step": 12700
    },
    {
      "epoch": 0.04355335028407339,
      "grad_norm": 1.158078670501709,
      "learning_rate": 1e-05,
      "loss": 0.2446,
      "step": 12710
    },
    {
      "epoch": 0.04358761727878942,
      "grad_norm": 1.0620396137237549,
      "learning_rate": 1e-05,
      "loss": 0.2403,
      "step": 12720
    },
    {
      "epoch": 0.04362188427350545,
      "grad_norm": 1.0640372037887573,
      "learning_rate": 1e-05,
      "loss": 0.2378,
      "step": 12730
    },
    {
      "epoch": 0.04365615126822148,
      "grad_norm": 1.113105058670044,
      "learning_rate": 1e-05,
      "loss": 0.236,
      "step": 12740
    },
    {
      "epoch": 0.04369041826293751,
      "grad_norm": 1.0416456460952759,
      "learning_rate": 1e-05,
      "loss": 0.2424,
      "step": 12750
    },
    {
      "epoch": 0.04372468525765354,
      "grad_norm": 1.2312722206115723,
      "learning_rate": 1e-05,
      "loss": 0.2397,
      "step": 12760
    },
    {
      "epoch": 0.04375895225236956,
      "grad_norm": 1.228950023651123,
      "learning_rate": 1e-05,
      "loss": 0.2462,
      "step": 12770
    },
    {
      "epoch": 0.04379321924708559,
      "grad_norm": 1.0736054182052612,
      "learning_rate": 1e-05,
      "loss": 0.2299,
      "step": 12780
    },
    {
      "epoch": 0.04382748624180162,
      "grad_norm": 1.4141355752944946,
      "learning_rate": 1e-05,
      "loss": 0.2377,
      "step": 12790
    },
    {
      "epoch": 0.04386175323651765,
      "grad_norm": 1.2224112749099731,
      "learning_rate": 1e-05,
      "loss": 0.2334,
      "step": 12800
    },
    {
      "epoch": 0.04389602023123368,
      "grad_norm": 1.3090282678604126,
      "learning_rate": 1e-05,
      "loss": 0.2548,
      "step": 12810
    },
    {
      "epoch": 0.04393028722594971,
      "grad_norm": 1.0864715576171875,
      "learning_rate": 1e-05,
      "loss": 0.2711,
      "step": 12820
    },
    {
      "epoch": 0.04396455422066574,
      "grad_norm": 1.0953795909881592,
      "learning_rate": 1e-05,
      "loss": 0.2467,
      "step": 12830
    },
    {
      "epoch": 0.04399882121538177,
      "grad_norm": 0.9681864976882935,
      "learning_rate": 1e-05,
      "loss": 0.2217,
      "step": 12840
    },
    {
      "epoch": 0.0440330882100978,
      "grad_norm": 0.9268914461135864,
      "learning_rate": 1e-05,
      "loss": 0.2492,
      "step": 12850
    },
    {
      "epoch": 0.04406735520481383,
      "grad_norm": 1.0900733470916748,
      "learning_rate": 1e-05,
      "loss": 0.2421,
      "step": 12860
    },
    {
      "epoch": 0.04410162219952986,
      "grad_norm": 1.1551947593688965,
      "learning_rate": 1e-05,
      "loss": 0.2549,
      "step": 12870
    },
    {
      "epoch": 0.04413588919424589,
      "grad_norm": 1.0035364627838135,
      "learning_rate": 1e-05,
      "loss": 0.248,
      "step": 12880
    },
    {
      "epoch": 0.04417015618896192,
      "grad_norm": 1.2478151321411133,
      "learning_rate": 1e-05,
      "loss": 0.253,
      "step": 12890
    },
    {
      "epoch": 0.04420442318367795,
      "grad_norm": 1.1512874364852905,
      "learning_rate": 1e-05,
      "loss": 0.2247,
      "step": 12900
    },
    {
      "epoch": 0.04423869017839398,
      "grad_norm": 1.2012622356414795,
      "learning_rate": 1e-05,
      "loss": 0.2635,
      "step": 12910
    },
    {
      "epoch": 0.04427295717311001,
      "grad_norm": 1.1266357898712158,
      "learning_rate": 1e-05,
      "loss": 0.2464,
      "step": 12920
    },
    {
      "epoch": 0.04430722416782603,
      "grad_norm": 1.11850905418396,
      "learning_rate": 1e-05,
      "loss": 0.2499,
      "step": 12930
    },
    {
      "epoch": 0.04434149116254206,
      "grad_norm": 1.1375716924667358,
      "learning_rate": 1e-05,
      "loss": 0.2474,
      "step": 12940
    },
    {
      "epoch": 0.04437575815725809,
      "grad_norm": 1.3423253297805786,
      "learning_rate": 1e-05,
      "loss": 0.2525,
      "step": 12950
    },
    {
      "epoch": 0.04441002515197412,
      "grad_norm": 1.0608446598052979,
      "learning_rate": 1e-05,
      "loss": 0.2445,
      "step": 12960
    },
    {
      "epoch": 0.04444429214669015,
      "grad_norm": 1.059899091720581,
      "learning_rate": 1e-05,
      "loss": 0.2444,
      "step": 12970
    },
    {
      "epoch": 0.04447855914140618,
      "grad_norm": 1.117346167564392,
      "learning_rate": 1e-05,
      "loss": 0.2529,
      "step": 12980
    },
    {
      "epoch": 0.04451282613612221,
      "grad_norm": 1.2896045446395874,
      "learning_rate": 1e-05,
      "loss": 0.2473,
      "step": 12990
    },
    {
      "epoch": 0.04454709313083824,
      "grad_norm": 0.9721153974533081,
      "learning_rate": 1e-05,
      "loss": 0.2461,
      "step": 13000
    },
    {
      "epoch": 0.04454709313083824,
      "eval_cer": 12.84177103545599,
      "eval_loss": 0.2521709203720093,
      "eval_normalized_cer": 9.152677857713828,
      "eval_runtime": 226.778,
      "eval_samples_per_second": 2.258,
      "eval_steps_per_second": 0.035,
      "step": 13000
    },
    {
      "epoch": 0.04458136012555427,
      "grad_norm": 1.232352375984192,
      "learning_rate": 1e-05,
      "loss": 0.2387,
      "step": 13010
    },
    {
      "epoch": 0.0446156271202703,
      "grad_norm": 1.2386256456375122,
      "learning_rate": 1e-05,
      "loss": 0.255,
      "step": 13020
    },
    {
      "epoch": 0.04464989411498633,
      "grad_norm": 1.2183597087860107,
      "learning_rate": 1e-05,
      "loss": 0.2635,
      "step": 13030
    },
    {
      "epoch": 0.04468416110970236,
      "grad_norm": 1.166823387145996,
      "learning_rate": 1e-05,
      "loss": 0.2563,
      "step": 13040
    },
    {
      "epoch": 0.04471842810441839,
      "grad_norm": 1.1574853658676147,
      "learning_rate": 1e-05,
      "loss": 0.2557,
      "step": 13050
    },
    {
      "epoch": 0.04475269509913442,
      "grad_norm": 1.1207836866378784,
      "learning_rate": 1e-05,
      "loss": 0.2578,
      "step": 13060
    },
    {
      "epoch": 0.04478696209385045,
      "grad_norm": 1.2590343952178955,
      "learning_rate": 1e-05,
      "loss": 0.2551,
      "step": 13070
    },
    {
      "epoch": 0.04482122908856648,
      "grad_norm": 1.0984435081481934,
      "learning_rate": 1e-05,
      "loss": 0.2816,
      "step": 13080
    },
    {
      "epoch": 0.0448554960832825,
      "grad_norm": 1.1435647010803223,
      "learning_rate": 1e-05,
      "loss": 0.2481,
      "step": 13090
    },
    {
      "epoch": 0.04488976307799853,
      "grad_norm": 1.1446672677993774,
      "learning_rate": 1e-05,
      "loss": 0.2561,
      "step": 13100
    },
    {
      "epoch": 0.04492403007271456,
      "grad_norm": 1.0957670211791992,
      "learning_rate": 1e-05,
      "loss": 0.2458,
      "step": 13110
    },
    {
      "epoch": 0.04495829706743059,
      "grad_norm": 1.1321167945861816,
      "learning_rate": 1e-05,
      "loss": 0.2621,
      "step": 13120
    },
    {
      "epoch": 0.04499256406214662,
      "grad_norm": 1.140914797782898,
      "learning_rate": 1e-05,
      "loss": 0.2376,
      "step": 13130
    },
    {
      "epoch": 0.04502683105686265,
      "grad_norm": 1.1879481077194214,
      "learning_rate": 1e-05,
      "loss": 0.2457,
      "step": 13140
    },
    {
      "epoch": 0.04506109805157868,
      "grad_norm": 1.240084171295166,
      "learning_rate": 1e-05,
      "loss": 0.2486,
      "step": 13150
    },
    {
      "epoch": 0.04509536504629471,
      "grad_norm": 1.1524683237075806,
      "learning_rate": 1e-05,
      "loss": 0.2533,
      "step": 13160
    },
    {
      "epoch": 0.04512963204101074,
      "grad_norm": 1.1614208221435547,
      "learning_rate": 1e-05,
      "loss": 0.2495,
      "step": 13170
    },
    {
      "epoch": 0.04516389903572677,
      "grad_norm": 1.1307048797607422,
      "learning_rate": 1e-05,
      "loss": 0.2549,
      "step": 13180
    },
    {
      "epoch": 0.0451981660304428,
      "grad_norm": 1.0327478647232056,
      "learning_rate": 1e-05,
      "loss": 0.2451,
      "step": 13190
    },
    {
      "epoch": 0.04523243302515883,
      "grad_norm": 1.2401607036590576,
      "learning_rate": 1e-05,
      "loss": 0.2492,
      "step": 13200
    },
    {
      "epoch": 0.04526670001987486,
      "grad_norm": 1.095413088798523,
      "learning_rate": 1e-05,
      "loss": 0.2487,
      "step": 13210
    },
    {
      "epoch": 0.04530096701459089,
      "grad_norm": 1.2537821531295776,
      "learning_rate": 1e-05,
      "loss": 0.246,
      "step": 13220
    },
    {
      "epoch": 0.04533523400930692,
      "grad_norm": 1.1861079931259155,
      "learning_rate": 1e-05,
      "loss": 0.2487,
      "step": 13230
    },
    {
      "epoch": 0.04536950100402295,
      "grad_norm": 1.1059224605560303,
      "learning_rate": 1e-05,
      "loss": 0.2465,
      "step": 13240
    },
    {
      "epoch": 0.04540376799873898,
      "grad_norm": 1.159122109413147,
      "learning_rate": 1e-05,
      "loss": 0.2472,
      "step": 13250
    },
    {
      "epoch": 0.045438034993455,
      "grad_norm": 1.0307060480117798,
      "learning_rate": 1e-05,
      "loss": 0.2583,
      "step": 13260
    },
    {
      "epoch": 0.04547230198817103,
      "grad_norm": 1.0377501249313354,
      "learning_rate": 1e-05,
      "loss": 0.2547,
      "step": 13270
    },
    {
      "epoch": 0.04550656898288706,
      "grad_norm": 1.124543309211731,
      "learning_rate": 1e-05,
      "loss": 0.2396,
      "step": 13280
    },
    {
      "epoch": 0.04554083597760309,
      "grad_norm": 1.0829116106033325,
      "learning_rate": 1e-05,
      "loss": 0.2417,
      "step": 13290
    },
    {
      "epoch": 0.04557510297231912,
      "grad_norm": 1.057477355003357,
      "learning_rate": 1e-05,
      "loss": 0.2527,
      "step": 13300
    },
    {
      "epoch": 0.04560936996703515,
      "grad_norm": 1.062674641609192,
      "learning_rate": 1e-05,
      "loss": 0.2423,
      "step": 13310
    },
    {
      "epoch": 0.04564363696175118,
      "grad_norm": 1.131895661354065,
      "learning_rate": 1e-05,
      "loss": 0.2517,
      "step": 13320
    },
    {
      "epoch": 0.04567790395646721,
      "grad_norm": 1.016940951347351,
      "learning_rate": 1e-05,
      "loss": 0.2419,
      "step": 13330
    },
    {
      "epoch": 0.04571217095118324,
      "grad_norm": 1.273378849029541,
      "learning_rate": 1e-05,
      "loss": 0.2465,
      "step": 13340
    },
    {
      "epoch": 0.04574643794589927,
      "grad_norm": 1.0757806301116943,
      "learning_rate": 1e-05,
      "loss": 0.2447,
      "step": 13350
    },
    {
      "epoch": 0.0457807049406153,
      "grad_norm": 1.3264166116714478,
      "learning_rate": 1e-05,
      "loss": 0.2545,
      "step": 13360
    },
    {
      "epoch": 0.04581497193533133,
      "grad_norm": 1.1011106967926025,
      "learning_rate": 1e-05,
      "loss": 0.269,
      "step": 13370
    },
    {
      "epoch": 0.04584923893004736,
      "grad_norm": 1.0483593940734863,
      "learning_rate": 1e-05,
      "loss": 0.2443,
      "step": 13380
    },
    {
      "epoch": 0.04588350592476339,
      "grad_norm": 1.2940049171447754,
      "learning_rate": 1e-05,
      "loss": 0.2459,
      "step": 13390
    },
    {
      "epoch": 0.04591777291947942,
      "grad_norm": 1.1630951166152954,
      "learning_rate": 1e-05,
      "loss": 0.2702,
      "step": 13400
    },
    {
      "epoch": 0.04595203991419545,
      "grad_norm": 1.0715082883834839,
      "learning_rate": 1e-05,
      "loss": 0.2402,
      "step": 13410
    },
    {
      "epoch": 0.04598630690891147,
      "grad_norm": 1.0946441888809204,
      "learning_rate": 1e-05,
      "loss": 0.2452,
      "step": 13420
    },
    {
      "epoch": 0.0460205739036275,
      "grad_norm": 1.0796674489974976,
      "learning_rate": 1e-05,
      "loss": 0.2449,
      "step": 13430
    },
    {
      "epoch": 0.04605484089834353,
      "grad_norm": 1.0534013509750366,
      "learning_rate": 1e-05,
      "loss": 0.2385,
      "step": 13440
    },
    {
      "epoch": 0.04608910789305956,
      "grad_norm": 1.0427377223968506,
      "learning_rate": 1e-05,
      "loss": 0.2557,
      "step": 13450
    },
    {
      "epoch": 0.04612337488777559,
      "grad_norm": 1.1708178520202637,
      "learning_rate": 1e-05,
      "loss": 0.2452,
      "step": 13460
    },
    {
      "epoch": 0.04615764188249162,
      "grad_norm": 1.0531684160232544,
      "learning_rate": 1e-05,
      "loss": 0.2594,
      "step": 13470
    },
    {
      "epoch": 0.04619190887720765,
      "grad_norm": 1.1972299814224243,
      "learning_rate": 1e-05,
      "loss": 0.2566,
      "step": 13480
    },
    {
      "epoch": 0.04622617587192368,
      "grad_norm": 1.0194915533065796,
      "learning_rate": 1e-05,
      "loss": 0.2504,
      "step": 13490
    },
    {
      "epoch": 0.04626044286663971,
      "grad_norm": 1.2437708377838135,
      "learning_rate": 1e-05,
      "loss": 0.251,
      "step": 13500
    },
    {
      "epoch": 0.04626044286663971,
      "eval_cer": 12.744752160874933,
      "eval_loss": 0.24938170611858368,
      "eval_normalized_cer": 8.952837729816148,
      "eval_runtime": 229.544,
      "eval_samples_per_second": 2.231,
      "eval_steps_per_second": 0.035,
      "step": 13500
    },
    {
      "epoch": 0.04629470986135574,
      "grad_norm": 1.2545411586761475,
      "learning_rate": 1e-05,
      "loss": 0.2714,
      "step": 13510
    },
    {
      "epoch": 0.04632897685607177,
      "grad_norm": 1.0949839353561401,
      "learning_rate": 1e-05,
      "loss": 0.2442,
      "step": 13520
    },
    {
      "epoch": 0.0463632438507878,
      "grad_norm": 1.1343241930007935,
      "learning_rate": 1e-05,
      "loss": 0.2454,
      "step": 13530
    },
    {
      "epoch": 0.04639751084550383,
      "grad_norm": 1.2311172485351562,
      "learning_rate": 1e-05,
      "loss": 0.2608,
      "step": 13540
    },
    {
      "epoch": 0.04643177784021986,
      "grad_norm": 1.0095285177230835,
      "learning_rate": 1e-05,
      "loss": 0.2558,
      "step": 13550
    },
    {
      "epoch": 0.04646604483493589,
      "grad_norm": 1.0752111673355103,
      "learning_rate": 1e-05,
      "loss": 0.237,
      "step": 13560
    },
    {
      "epoch": 0.04650031182965192,
      "grad_norm": 1.0850863456726074,
      "learning_rate": 1e-05,
      "loss": 0.2552,
      "step": 13570
    },
    {
      "epoch": 0.04653457882436794,
      "grad_norm": 1.185105323791504,
      "learning_rate": 1e-05,
      "loss": 0.2567,
      "step": 13580
    },
    {
      "epoch": 0.04656884581908397,
      "grad_norm": 1.221077799797058,
      "learning_rate": 1e-05,
      "loss": 0.2591,
      "step": 13590
    },
    {
      "epoch": 0.0466031128138,
      "grad_norm": 1.23322594165802,
      "learning_rate": 1e-05,
      "loss": 0.2485,
      "step": 13600
    },
    {
      "epoch": 0.04663737980851603,
      "grad_norm": 1.1576459407806396,
      "learning_rate": 1e-05,
      "loss": 0.2573,
      "step": 13610
    },
    {
      "epoch": 0.04667164680323206,
      "grad_norm": 1.202359914779663,
      "learning_rate": 1e-05,
      "loss": 0.2555,
      "step": 13620
    },
    {
      "epoch": 0.04670591379794809,
      "grad_norm": 1.1896847486495972,
      "learning_rate": 1e-05,
      "loss": 0.2586,
      "step": 13630
    },
    {
      "epoch": 0.04674018079266412,
      "grad_norm": 0.8355448246002197,
      "learning_rate": 1e-05,
      "loss": 0.2374,
      "step": 13640
    },
    {
      "epoch": 0.04677444778738015,
      "grad_norm": 1.1680315732955933,
      "learning_rate": 1e-05,
      "loss": 0.2365,
      "step": 13650
    },
    {
      "epoch": 0.04680871478209618,
      "grad_norm": 1.2542601823806763,
      "learning_rate": 1e-05,
      "loss": 0.2494,
      "step": 13660
    },
    {
      "epoch": 0.04684298177681221,
      "grad_norm": 1.1530771255493164,
      "learning_rate": 1e-05,
      "loss": 0.2268,
      "step": 13670
    },
    {
      "epoch": 0.04687724877152824,
      "grad_norm": 1.0363566875457764,
      "learning_rate": 1e-05,
      "loss": 0.2623,
      "step": 13680
    },
    {
      "epoch": 0.04691151576624427,
      "grad_norm": 1.0039604902267456,
      "learning_rate": 1e-05,
      "loss": 0.2524,
      "step": 13690
    },
    {
      "epoch": 0.0469457827609603,
      "grad_norm": 1.2030223608016968,
      "learning_rate": 1e-05,
      "loss": 0.2346,
      "step": 13700
    },
    {
      "epoch": 0.04698004975567633,
      "grad_norm": 1.1221191883087158,
      "learning_rate": 1e-05,
      "loss": 0.2483,
      "step": 13710
    },
    {
      "epoch": 0.04701431675039236,
      "grad_norm": 0.9777095913887024,
      "learning_rate": 1e-05,
      "loss": 0.2484,
      "step": 13720
    },
    {
      "epoch": 0.04704858374510839,
      "grad_norm": 1.1439709663391113,
      "learning_rate": 1e-05,
      "loss": 0.2447,
      "step": 13730
    },
    {
      "epoch": 0.04708285073982441,
      "grad_norm": 1.0818963050842285,
      "learning_rate": 1e-05,
      "loss": 0.2378,
      "step": 13740
    },
    {
      "epoch": 0.04711711773454044,
      "grad_norm": 1.1691282987594604,
      "learning_rate": 1e-05,
      "loss": 0.2433,
      "step": 13750
    },
    {
      "epoch": 0.04715138472925647,
      "grad_norm": 1.1172744035720825,
      "learning_rate": 1e-05,
      "loss": 0.2297,
      "step": 13760
    },
    {
      "epoch": 0.0471856517239725,
      "grad_norm": 1.170318841934204,
      "learning_rate": 1e-05,
      "loss": 0.2451,
      "step": 13770
    },
    {
      "epoch": 0.04721991871868853,
      "grad_norm": 1.152566909790039,
      "learning_rate": 1e-05,
      "loss": 0.2497,
      "step": 13780
    },
    {
      "epoch": 0.04725418571340456,
      "grad_norm": 1.023813247680664,
      "learning_rate": 1e-05,
      "loss": 0.2654,
      "step": 13790
    },
    {
      "epoch": 0.04728845270812059,
      "grad_norm": 1.159853458404541,
      "learning_rate": 1e-05,
      "loss": 0.2537,
      "step": 13800
    },
    {
      "epoch": 0.04732271970283662,
      "grad_norm": 1.0579999685287476,
      "learning_rate": 1e-05,
      "loss": 0.2524,
      "step": 13810
    },
    {
      "epoch": 0.04735698669755265,
      "grad_norm": 1.1741176843643188,
      "learning_rate": 1e-05,
      "loss": 0.2574,
      "step": 13820
    },
    {
      "epoch": 0.04739125369226868,
      "grad_norm": 1.1931475400924683,
      "learning_rate": 1e-05,
      "loss": 0.2504,
      "step": 13830
    },
    {
      "epoch": 0.04742552068698471,
      "grad_norm": 0.9977951645851135,
      "learning_rate": 1e-05,
      "loss": 0.2515,
      "step": 13840
    },
    {
      "epoch": 0.04745978768170074,
      "grad_norm": 1.177424669265747,
      "learning_rate": 1e-05,
      "loss": 0.2399,
      "step": 13850
    },
    {
      "epoch": 0.04749405467641677,
      "grad_norm": 1.0663032531738281,
      "learning_rate": 1e-05,
      "loss": 0.2464,
      "step": 13860
    },
    {
      "epoch": 0.0475283216711328,
      "grad_norm": 0.9985357522964478,
      "learning_rate": 1e-05,
      "loss": 0.2518,
      "step": 13870
    },
    {
      "epoch": 0.04756258866584883,
      "grad_norm": 1.1378836631774902,
      "learning_rate": 1e-05,
      "loss": 0.2366,
      "step": 13880
    },
    {
      "epoch": 0.04759685566056486,
      "grad_norm": 1.130223035812378,
      "learning_rate": 1e-05,
      "loss": 0.2521,
      "step": 13890
    },
    {
      "epoch": 0.04763112265528089,
      "grad_norm": 1.1601965427398682,
      "learning_rate": 1e-05,
      "loss": 0.2635,
      "step": 13900
    },
    {
      "epoch": 0.04766538964999691,
      "grad_norm": 1.0559511184692383,
      "learning_rate": 1e-05,
      "loss": 0.2496,
      "step": 13910
    },
    {
      "epoch": 0.04769965664471294,
      "grad_norm": 1.0258302688598633,
      "learning_rate": 1e-05,
      "loss": 0.2472,
      "step": 13920
    },
    {
      "epoch": 0.04773392363942897,
      "grad_norm": 1.1403284072875977,
      "learning_rate": 1e-05,
      "loss": 0.2544,
      "step": 13930
    },
    {
      "epoch": 0.047768190634145,
      "grad_norm": 1.1353199481964111,
      "learning_rate": 1e-05,
      "loss": 0.2502,
      "step": 13940
    },
    {
      "epoch": 0.04780245762886103,
      "grad_norm": 1.2859739065170288,
      "learning_rate": 1e-05,
      "loss": 0.2314,
      "step": 13950
    },
    {
      "epoch": 0.04783672462357706,
      "grad_norm": 1.1392145156860352,
      "learning_rate": 1e-05,
      "loss": 0.2493,
      "step": 13960
    },
    {
      "epoch": 0.04787099161829309,
      "grad_norm": 1.055909276008606,
      "learning_rate": 1e-05,
      "loss": 0.2406,
      "step": 13970
    },
    {
      "epoch": 0.04790525861300912,
      "grad_norm": 1.0544830560684204,
      "learning_rate": 1e-05,
      "loss": 0.248,
      "step": 13980
    },
    {
      "epoch": 0.04793952560772515,
      "grad_norm": 1.0616220235824585,
      "learning_rate": 1e-05,
      "loss": 0.2489,
      "step": 13990
    },
    {
      "epoch": 0.04797379260244118,
      "grad_norm": 1.0163713693618774,
      "learning_rate": 1e-05,
      "loss": 0.2359,
      "step": 14000
    },
    {
      "epoch": 0.04797379260244118,
      "eval_cer": 12.683012877050626,
      "eval_loss": 0.24888557195663452,
      "eval_normalized_cer": 9.082733812949641,
      "eval_runtime": 228.6884,
      "eval_samples_per_second": 2.239,
      "eval_steps_per_second": 0.035,
      "step": 14000
    },
    {
      "epoch": 0.04800805959715721,
      "grad_norm": 1.0705435276031494,
      "learning_rate": 1e-05,
      "loss": 0.2407,
      "step": 14010
    },
    {
      "epoch": 0.04804232659187324,
      "grad_norm": 0.9681057333946228,
      "learning_rate": 1e-05,
      "loss": 0.2506,
      "step": 14020
    },
    {
      "epoch": 0.04807659358658927,
      "grad_norm": 1.0451055765151978,
      "learning_rate": 1e-05,
      "loss": 0.2523,
      "step": 14030
    },
    {
      "epoch": 0.0481108605813053,
      "grad_norm": 1.0783635377883911,
      "learning_rate": 1e-05,
      "loss": 0.242,
      "step": 14040
    },
    {
      "epoch": 0.04814512757602133,
      "grad_norm": 1.1066149473190308,
      "learning_rate": 1e-05,
      "loss": 0.2514,
      "step": 14050
    },
    {
      "epoch": 0.04817939457073736,
      "grad_norm": 1.163319706916809,
      "learning_rate": 1e-05,
      "loss": 0.2336,
      "step": 14060
    },
    {
      "epoch": 0.04821366156545338,
      "grad_norm": 1.1571089029312134,
      "learning_rate": 1e-05,
      "loss": 0.2626,
      "step": 14070
    },
    {
      "epoch": 0.04824792856016941,
      "grad_norm": 1.153372049331665,
      "learning_rate": 1e-05,
      "loss": 0.2499,
      "step": 14080
    },
    {
      "epoch": 0.04828219555488544,
      "grad_norm": 1.1359412670135498,
      "learning_rate": 1e-05,
      "loss": 0.245,
      "step": 14090
    },
    {
      "epoch": 0.04831646254960147,
      "grad_norm": 1.1467552185058594,
      "learning_rate": 1e-05,
      "loss": 0.2185,
      "step": 14100
    },
    {
      "epoch": 0.0483507295443175,
      "grad_norm": 1.1838459968566895,
      "learning_rate": 1e-05,
      "loss": 0.2236,
      "step": 14110
    },
    {
      "epoch": 0.04838499653903353,
      "grad_norm": 1.1658817529678345,
      "learning_rate": 1e-05,
      "loss": 0.2527,
      "step": 14120
    },
    {
      "epoch": 0.04841926353374956,
      "grad_norm": 1.0678468942642212,
      "learning_rate": 1e-05,
      "loss": 0.2299,
      "step": 14130
    },
    {
      "epoch": 0.04845353052846559,
      "grad_norm": 1.169755220413208,
      "learning_rate": 1e-05,
      "loss": 0.256,
      "step": 14140
    },
    {
      "epoch": 0.04848779752318162,
      "grad_norm": 1.1786571741104126,
      "learning_rate": 1e-05,
      "loss": 0.2294,
      "step": 14150
    },
    {
      "epoch": 0.04852206451789765,
      "grad_norm": 1.0959957838058472,
      "learning_rate": 1e-05,
      "loss": 0.2415,
      "step": 14160
    },
    {
      "epoch": 0.04855633151261368,
      "grad_norm": 1.3755067586898804,
      "learning_rate": 1e-05,
      "loss": 0.2402,
      "step": 14170
    },
    {
      "epoch": 0.04859059850732971,
      "grad_norm": 1.0811392068862915,
      "learning_rate": 1e-05,
      "loss": 0.2565,
      "step": 14180
    },
    {
      "epoch": 0.04862486550204574,
      "grad_norm": 1.0909180641174316,
      "learning_rate": 1e-05,
      "loss": 0.2338,
      "step": 14190
    },
    {
      "epoch": 0.04865913249676177,
      "grad_norm": 1.168319821357727,
      "learning_rate": 1e-05,
      "loss": 0.2422,
      "step": 14200
    },
    {
      "epoch": 0.0486933994914778,
      "grad_norm": 1.0762922763824463,
      "learning_rate": 1e-05,
      "loss": 0.2419,
      "step": 14210
    },
    {
      "epoch": 0.04872766648619383,
      "grad_norm": 0.9447901248931885,
      "learning_rate": 1e-05,
      "loss": 0.2257,
      "step": 14220
    },
    {
      "epoch": 0.048761933480909854,
      "grad_norm": 1.0396913290023804,
      "learning_rate": 1e-05,
      "loss": 0.2351,
      "step": 14230
    },
    {
      "epoch": 0.048796200475625884,
      "grad_norm": 1.0487229824066162,
      "learning_rate": 1e-05,
      "loss": 0.229,
      "step": 14240
    },
    {
      "epoch": 0.048830467470341914,
      "grad_norm": 1.1879597902297974,
      "learning_rate": 1e-05,
      "loss": 0.251,
      "step": 14250
    },
    {
      "epoch": 0.048864734465057943,
      "grad_norm": 1.2501291036605835,
      "learning_rate": 1e-05,
      "loss": 0.2421,
      "step": 14260
    },
    {
      "epoch": 0.04889900145977397,
      "grad_norm": 1.119624137878418,
      "learning_rate": 1e-05,
      "loss": 0.242,
      "step": 14270
    },
    {
      "epoch": 0.04893326845449,
      "grad_norm": 1.0920354127883911,
      "learning_rate": 1e-05,
      "loss": 0.2352,
      "step": 14280
    },
    {
      "epoch": 0.04896753544920603,
      "grad_norm": 1.0519214868545532,
      "learning_rate": 1e-05,
      "loss": 0.2466,
      "step": 14290
    },
    {
      "epoch": 0.04900180244392206,
      "grad_norm": 1.1111136674880981,
      "learning_rate": 1e-05,
      "loss": 0.2438,
      "step": 14300
    },
    {
      "epoch": 0.04903606943863809,
      "grad_norm": 1.2424713373184204,
      "learning_rate": 1e-05,
      "loss": 0.2358,
      "step": 14310
    },
    {
      "epoch": 0.04907033643335412,
      "grad_norm": 1.0222002267837524,
      "learning_rate": 1e-05,
      "loss": 0.2302,
      "step": 14320
    },
    {
      "epoch": 0.04910460342807015,
      "grad_norm": 0.9916470050811768,
      "learning_rate": 1e-05,
      "loss": 0.2068,
      "step": 14330
    },
    {
      "epoch": 0.04913887042278618,
      "grad_norm": 1.31315016746521,
      "learning_rate": 1e-05,
      "loss": 0.2489,
      "step": 14340
    },
    {
      "epoch": 0.04917313741750221,
      "grad_norm": 1.1765952110290527,
      "learning_rate": 1e-05,
      "loss": 0.2504,
      "step": 14350
    },
    {
      "epoch": 0.04920740441221824,
      "grad_norm": 1.1690514087677002,
      "learning_rate": 1e-05,
      "loss": 0.2441,
      "step": 14360
    },
    {
      "epoch": 0.04924167140693427,
      "grad_norm": 1.1756752729415894,
      "learning_rate": 1e-05,
      "loss": 0.2669,
      "step": 14370
    },
    {
      "epoch": 0.0492759384016503,
      "grad_norm": 1.1072384119033813,
      "learning_rate": 1e-05,
      "loss": 0.2426,
      "step": 14380
    },
    {
      "epoch": 0.049310205396366324,
      "grad_norm": 1.2008529901504517,
      "learning_rate": 1e-05,
      "loss": 0.2545,
      "step": 14390
    },
    {
      "epoch": 0.049344472391082354,
      "grad_norm": 1.1737167835235596,
      "learning_rate": 1e-05,
      "loss": 0.25,
      "step": 14400
    },
    {
      "epoch": 0.049378739385798384,
      "grad_norm": 1.0450342893600464,
      "learning_rate": 1e-05,
      "loss": 0.2539,
      "step": 14410
    },
    {
      "epoch": 0.049413006380514414,
      "grad_norm": 1.0435712337493896,
      "learning_rate": 1e-05,
      "loss": 0.2404,
      "step": 14420
    },
    {
      "epoch": 0.049447273375230444,
      "grad_norm": 1.2220741510391235,
      "learning_rate": 1e-05,
      "loss": 0.2544,
      "step": 14430
    },
    {
      "epoch": 0.049481540369946474,
      "grad_norm": 1.285495400428772,
      "learning_rate": 1e-05,
      "loss": 0.2563,
      "step": 14440
    },
    {
      "epoch": 0.049515807364662504,
      "grad_norm": 1.2037091255187988,
      "learning_rate": 1e-05,
      "loss": 0.269,
      "step": 14450
    },
    {
      "epoch": 0.04955007435937853,
      "grad_norm": 1.1641725301742554,
      "learning_rate": 1e-05,
      "loss": 0.2582,
      "step": 14460
    },
    {
      "epoch": 0.04958434135409456,
      "grad_norm": 1.30322265625,
      "learning_rate": 1e-05,
      "loss": 0.2637,
      "step": 14470
    },
    {
      "epoch": 0.04961860834881059,
      "grad_norm": 1.0870246887207031,
      "learning_rate": 1e-05,
      "loss": 0.2319,
      "step": 14480
    },
    {
      "epoch": 0.04965287534352662,
      "grad_norm": 1.1934154033660889,
      "learning_rate": 1e-05,
      "loss": 0.2654,
      "step": 14490
    },
    {
      "epoch": 0.04968714233824265,
      "grad_norm": 1.0471813678741455,
      "learning_rate": 1e-05,
      "loss": 0.2672,
      "step": 14500
    },
    {
      "epoch": 0.04968714233824265,
      "eval_cer": 12.859410830834362,
      "eval_loss": 0.2491358071565628,
      "eval_normalized_cer": 9.292565947242206,
      "eval_runtime": 235.6794,
      "eval_samples_per_second": 2.172,
      "eval_steps_per_second": 0.034,
      "step": 14500
    },
    {
      "epoch": 0.04972140933295868,
      "grad_norm": 1.1028441190719604,
      "learning_rate": 1e-05,
      "loss": 0.2506,
      "step": 14510
    },
    {
      "epoch": 0.04975567632767471,
      "grad_norm": 1.0829942226409912,
      "learning_rate": 1e-05,
      "loss": 0.2624,
      "step": 14520
    },
    {
      "epoch": 0.04978994332239074,
      "grad_norm": 1.094115138053894,
      "learning_rate": 1e-05,
      "loss": 0.2393,
      "step": 14530
    },
    {
      "epoch": 0.04982421031710677,
      "grad_norm": 1.1440691947937012,
      "learning_rate": 1e-05,
      "loss": 0.2454,
      "step": 14540
    },
    {
      "epoch": 0.0498584773118228,
      "grad_norm": 1.171846866607666,
      "learning_rate": 1e-05,
      "loss": 0.2528,
      "step": 14550
    },
    {
      "epoch": 0.049892744306538825,
      "grad_norm": 1.1416808366775513,
      "learning_rate": 1e-05,
      "loss": 0.2545,
      "step": 14560
    },
    {
      "epoch": 0.049927011301254855,
      "grad_norm": 1.2240614891052246,
      "learning_rate": 1e-05,
      "loss": 0.2511,
      "step": 14570
    },
    {
      "epoch": 0.049961278295970885,
      "grad_norm": 1.224147081375122,
      "learning_rate": 1e-05,
      "loss": 0.2662,
      "step": 14580
    },
    {
      "epoch": 0.049995545290686914,
      "grad_norm": 1.0746549367904663,
      "learning_rate": 1e-05,
      "loss": 0.2466,
      "step": 14590
    },
    {
      "epoch": 0.050029812285402944,
      "grad_norm": 1.0464617013931274,
      "learning_rate": 1e-05,
      "loss": 0.2483,
      "step": 14600
    },
    {
      "epoch": 0.050064079280118974,
      "grad_norm": 1.0312144756317139,
      "learning_rate": 1e-05,
      "loss": 0.2496,
      "step": 14610
    },
    {
      "epoch": 0.050098346274835004,
      "grad_norm": 1.109796166419983,
      "learning_rate": 1e-05,
      "loss": 0.2459,
      "step": 14620
    },
    {
      "epoch": 0.050132613269551034,
      "grad_norm": 1.100741982460022,
      "learning_rate": 1e-05,
      "loss": 0.2442,
      "step": 14630
    },
    {
      "epoch": 0.050166880264267064,
      "grad_norm": 1.187683343887329,
      "learning_rate": 1e-05,
      "loss": 0.2434,
      "step": 14640
    },
    {
      "epoch": 0.05020114725898309,
      "grad_norm": 1.1926337480545044,
      "learning_rate": 1e-05,
      "loss": 0.2554,
      "step": 14650
    },
    {
      "epoch": 0.05023541425369912,
      "grad_norm": 0.9210227727890015,
      "learning_rate": 1e-05,
      "loss": 0.2497,
      "step": 14660
    },
    {
      "epoch": 0.05026968124841515,
      "grad_norm": 1.010360836982727,
      "learning_rate": 1e-05,
      "loss": 0.2505,
      "step": 14670
    },
    {
      "epoch": 0.05030394824313118,
      "grad_norm": 1.0252940654754639,
      "learning_rate": 1e-05,
      "loss": 0.2461,
      "step": 14680
    },
    {
      "epoch": 0.05033821523784721,
      "grad_norm": 1.297200322151184,
      "learning_rate": 1e-05,
      "loss": 0.2526,
      "step": 14690
    },
    {
      "epoch": 0.05037248223256324,
      "grad_norm": 1.2773388624191284,
      "learning_rate": 1e-05,
      "loss": 0.2484,
      "step": 14700
    },
    {
      "epoch": 0.05040674922727927,
      "grad_norm": 1.110484004020691,
      "learning_rate": 1e-05,
      "loss": 0.2542,
      "step": 14710
    },
    {
      "epoch": 0.050441016221995295,
      "grad_norm": 1.0664464235305786,
      "learning_rate": 1e-05,
      "loss": 0.2528,
      "step": 14720
    },
    {
      "epoch": 0.050475283216711325,
      "grad_norm": 1.1421204805374146,
      "learning_rate": 1e-05,
      "loss": 0.2464,
      "step": 14730
    },
    {
      "epoch": 0.050509550211427355,
      "grad_norm": 1.115225911140442,
      "learning_rate": 1e-05,
      "loss": 0.2584,
      "step": 14740
    },
    {
      "epoch": 0.050543817206143385,
      "grad_norm": 1.1027865409851074,
      "learning_rate": 1e-05,
      "loss": 0.2606,
      "step": 14750
    },
    {
      "epoch": 0.050578084200859415,
      "grad_norm": 1.1306570768356323,
      "learning_rate": 1e-05,
      "loss": 0.2561,
      "step": 14760
    },
    {
      "epoch": 0.050612351195575445,
      "grad_norm": 1.0982325077056885,
      "learning_rate": 1e-05,
      "loss": 0.2688,
      "step": 14770
    },
    {
      "epoch": 0.050646618190291474,
      "grad_norm": 1.004801630973816,
      "learning_rate": 1e-05,
      "loss": 0.2596,
      "step": 14780
    },
    {
      "epoch": 0.050680885185007504,
      "grad_norm": 1.2542450428009033,
      "learning_rate": 1e-05,
      "loss": 0.2521,
      "step": 14790
    },
    {
      "epoch": 0.050715152179723534,
      "grad_norm": 1.311880111694336,
      "learning_rate": 1e-05,
      "loss": 0.2426,
      "step": 14800
    },
    {
      "epoch": 0.050749419174439564,
      "grad_norm": 1.119271159172058,
      "learning_rate": 1e-05,
      "loss": 0.2345,
      "step": 14810
    },
    {
      "epoch": 0.050783686169155594,
      "grad_norm": 1.1003872156143188,
      "learning_rate": 1e-05,
      "loss": 0.2518,
      "step": 14820
    },
    {
      "epoch": 0.050817953163871624,
      "grad_norm": 1.17613685131073,
      "learning_rate": 1e-05,
      "loss": 0.2546,
      "step": 14830
    },
    {
      "epoch": 0.050852220158587653,
      "grad_norm": 1.188706398010254,
      "learning_rate": 1e-05,
      "loss": 0.247,
      "step": 14840
    },
    {
      "epoch": 0.05088648715330368,
      "grad_norm": 1.1993244886398315,
      "learning_rate": 1e-05,
      "loss": 0.2619,
      "step": 14850
    },
    {
      "epoch": 0.05092075414801971,
      "grad_norm": 1.0679277181625366,
      "learning_rate": 1e-05,
      "loss": 0.2452,
      "step": 14860
    },
    {
      "epoch": 0.05095502114273574,
      "grad_norm": 1.1368016004562378,
      "learning_rate": 1e-05,
      "loss": 0.2453,
      "step": 14870
    },
    {
      "epoch": 0.050989288137451766,
      "grad_norm": 1.0620750188827515,
      "learning_rate": 1e-05,
      "loss": 0.2365,
      "step": 14880
    },
    {
      "epoch": 0.051023555132167796,
      "grad_norm": 1.1344637870788574,
      "learning_rate": 1e-05,
      "loss": 0.2479,
      "step": 14890
    },
    {
      "epoch": 0.051057822126883826,
      "grad_norm": 1.1324440240859985,
      "learning_rate": 1e-05,
      "loss": 0.2501,
      "step": 14900
    },
    {
      "epoch": 0.051092089121599855,
      "grad_norm": 1.139269232749939,
      "learning_rate": 1e-05,
      "loss": 0.2467,
      "step": 14910
    },
    {
      "epoch": 0.051126356116315885,
      "grad_norm": 1.1169798374176025,
      "learning_rate": 1e-05,
      "loss": 0.2616,
      "step": 14920
    },
    {
      "epoch": 0.051160623111031915,
      "grad_norm": 1.057564377784729,
      "learning_rate": 1e-05,
      "loss": 0.2554,
      "step": 14930
    },
    {
      "epoch": 0.051194890105747945,
      "grad_norm": 1.084874153137207,
      "learning_rate": 1e-05,
      "loss": 0.2399,
      "step": 14940
    },
    {
      "epoch": 0.051229157100463975,
      "grad_norm": 1.1470558643341064,
      "learning_rate": 1e-05,
      "loss": 0.252,
      "step": 14950
    },
    {
      "epoch": 0.051263424095180005,
      "grad_norm": 1.0080534219741821,
      "learning_rate": 1e-05,
      "loss": 0.256,
      "step": 14960
    },
    {
      "epoch": 0.051297691089896034,
      "grad_norm": 1.071164608001709,
      "learning_rate": 1e-05,
      "loss": 0.2371,
      "step": 14970
    },
    {
      "epoch": 0.051331958084612064,
      "grad_norm": 1.1828765869140625,
      "learning_rate": 1e-05,
      "loss": 0.2346,
      "step": 14980
    },
    {
      "epoch": 0.051366225079328094,
      "grad_norm": 1.2067548036575317,
      "learning_rate": 1e-05,
      "loss": 0.2414,
      "step": 14990
    },
    {
      "epoch": 0.051400492074044124,
      "grad_norm": 1.0730314254760742,
      "learning_rate": 1e-05,
      "loss": 0.2563,
      "step": 15000
    },
    {
      "epoch": 0.051400492074044124,
      "eval_cer": 12.568354207091199,
      "eval_loss": 0.2477913200855255,
      "eval_normalized_cer": 8.832933653077538,
      "eval_runtime": 229.5255,
      "eval_samples_per_second": 2.231,
      "eval_steps_per_second": 0.035,
      "step": 15000
    },
    {
      "epoch": 0.051434759068760154,
      "grad_norm": 1.0130051374435425,
      "learning_rate": 1e-05,
      "loss": 0.2571,
      "step": 15010
    },
    {
      "epoch": 0.051469026063476184,
      "grad_norm": 1.1843127012252808,
      "learning_rate": 1e-05,
      "loss": 0.2501,
      "step": 15020
    },
    {
      "epoch": 0.051503293058192214,
      "grad_norm": 1.1182798147201538,
      "learning_rate": 1e-05,
      "loss": 0.2559,
      "step": 15030
    },
    {
      "epoch": 0.051537560052908236,
      "grad_norm": 1.1553055047988892,
      "learning_rate": 1e-05,
      "loss": 0.2396,
      "step": 15040
    },
    {
      "epoch": 0.051571827047624266,
      "grad_norm": 1.030463695526123,
      "learning_rate": 1e-05,
      "loss": 0.2368,
      "step": 15050
    },
    {
      "epoch": 0.051606094042340296,
      "grad_norm": 1.2701278924942017,
      "learning_rate": 1e-05,
      "loss": 0.2495,
      "step": 15060
    },
    {
      "epoch": 0.051640361037056326,
      "grad_norm": 1.1329874992370605,
      "learning_rate": 1e-05,
      "loss": 0.2389,
      "step": 15070
    },
    {
      "epoch": 0.051674628031772356,
      "grad_norm": 1.132430911064148,
      "learning_rate": 1e-05,
      "loss": 0.2479,
      "step": 15080
    },
    {
      "epoch": 0.051708895026488386,
      "grad_norm": 1.1582975387573242,
      "learning_rate": 1e-05,
      "loss": 0.2511,
      "step": 15090
    },
    {
      "epoch": 0.051743162021204415,
      "grad_norm": 1.0918657779693604,
      "learning_rate": 1e-05,
      "loss": 0.2335,
      "step": 15100
    },
    {
      "epoch": 0.051777429015920445,
      "grad_norm": 1.1993087530136108,
      "learning_rate": 1e-05,
      "loss": 0.2468,
      "step": 15110
    },
    {
      "epoch": 0.051811696010636475,
      "grad_norm": 1.1809076070785522,
      "learning_rate": 1e-05,
      "loss": 0.2652,
      "step": 15120
    },
    {
      "epoch": 0.051845963005352505,
      "grad_norm": 1.2104005813598633,
      "learning_rate": 1e-05,
      "loss": 0.2444,
      "step": 15130
    },
    {
      "epoch": 0.051880230000068535,
      "grad_norm": 1.0615415573120117,
      "learning_rate": 1e-05,
      "loss": 0.2558,
      "step": 15140
    },
    {
      "epoch": 0.051914496994784565,
      "grad_norm": 1.1397675275802612,
      "learning_rate": 1e-05,
      "loss": 0.2543,
      "step": 15150
    },
    {
      "epoch": 0.051948763989500595,
      "grad_norm": 1.1353163719177246,
      "learning_rate": 1e-05,
      "loss": 0.2445,
      "step": 15160
    },
    {
      "epoch": 0.051983030984216624,
      "grad_norm": 1.1440542936325073,
      "learning_rate": 1e-05,
      "loss": 0.2396,
      "step": 15170
    },
    {
      "epoch": 0.052017297978932654,
      "grad_norm": 1.2371265888214111,
      "learning_rate": 1e-05,
      "loss": 0.2318,
      "step": 15180
    },
    {
      "epoch": 0.052051564973648684,
      "grad_norm": 1.03389310836792,
      "learning_rate": 1e-05,
      "loss": 0.2312,
      "step": 15190
    },
    {
      "epoch": 0.052085831968364714,
      "grad_norm": 1.3557147979736328,
      "learning_rate": 1e-05,
      "loss": 0.2387,
      "step": 15200
    },
    {
      "epoch": 0.05212009896308074,
      "grad_norm": 1.1793631315231323,
      "learning_rate": 1e-05,
      "loss": 0.2528,
      "step": 15210
    },
    {
      "epoch": 0.05215436595779677,
      "grad_norm": 1.1020557880401611,
      "learning_rate": 1e-05,
      "loss": 0.2438,
      "step": 15220
    },
    {
      "epoch": 0.052188632952512796,
      "grad_norm": 1.0566291809082031,
      "learning_rate": 1e-05,
      "loss": 0.2493,
      "step": 15230
    },
    {
      "epoch": 0.052222899947228826,
      "grad_norm": 1.0652698278427124,
      "learning_rate": 1e-05,
      "loss": 0.2385,
      "step": 15240
    },
    {
      "epoch": 0.052257166941944856,
      "grad_norm": 1.337236762046814,
      "learning_rate": 1e-05,
      "loss": 0.2532,
      "step": 15250
    },
    {
      "epoch": 0.052291433936660886,
      "grad_norm": 1.1121892929077148,
      "learning_rate": 1e-05,
      "loss": 0.2428,
      "step": 15260
    },
    {
      "epoch": 0.052325700931376916,
      "grad_norm": 1.1201363801956177,
      "learning_rate": 1e-05,
      "loss": 0.2477,
      "step": 15270
    },
    {
      "epoch": 0.052359967926092946,
      "grad_norm": 1.1055474281311035,
      "learning_rate": 1e-05,
      "loss": 0.243,
      "step": 15280
    },
    {
      "epoch": 0.052394234920808976,
      "grad_norm": 1.1349745988845825,
      "learning_rate": 1e-05,
      "loss": 0.2441,
      "step": 15290
    },
    {
      "epoch": 0.052428501915525005,
      "grad_norm": 1.0891187191009521,
      "learning_rate": 1e-05,
      "loss": 0.2544,
      "step": 15300
    },
    {
      "epoch": 0.052462768910241035,
      "grad_norm": 1.0537917613983154,
      "learning_rate": 1e-05,
      "loss": 0.2492,
      "step": 15310
    },
    {
      "epoch": 0.052497035904957065,
      "grad_norm": 1.1704713106155396,
      "learning_rate": 1e-05,
      "loss": 0.267,
      "step": 15320
    },
    {
      "epoch": 0.052531302899673095,
      "grad_norm": 0.9952285289764404,
      "learning_rate": 1e-05,
      "loss": 0.2291,
      "step": 15330
    },
    {
      "epoch": 0.052565569894389125,
      "grad_norm": 0.9887141585350037,
      "learning_rate": 1e-05,
      "loss": 0.2507,
      "step": 15340
    },
    {
      "epoch": 0.052599836889105155,
      "grad_norm": 1.20647394657135,
      "learning_rate": 1e-05,
      "loss": 0.2512,
      "step": 15350
    },
    {
      "epoch": 0.052634103883821184,
      "grad_norm": 1.0504127740859985,
      "learning_rate": 1e-05,
      "loss": 0.2559,
      "step": 15360
    },
    {
      "epoch": 0.05266837087853721,
      "grad_norm": 1.010195255279541,
      "learning_rate": 1e-05,
      "loss": 0.2292,
      "step": 15370
    },
    {
      "epoch": 0.05270263787325324,
      "grad_norm": 1.0885406732559204,
      "learning_rate": 1e-05,
      "loss": 0.2244,
      "step": 15380
    },
    {
      "epoch": 0.05273690486796927,
      "grad_norm": 0.9946883916854858,
      "learning_rate": 1e-05,
      "loss": 0.2488,
      "step": 15390
    },
    {
      "epoch": 0.0527711718626853,
      "grad_norm": 1.163482666015625,
      "learning_rate": 1e-05,
      "loss": 0.2401,
      "step": 15400
    },
    {
      "epoch": 0.05280543885740133,
      "grad_norm": 1.0451022386550903,
      "learning_rate": 1e-05,
      "loss": 0.2324,
      "step": 15410
    },
    {
      "epoch": 0.05283970585211736,
      "grad_norm": 1.0925875902175903,
      "learning_rate": 1e-05,
      "loss": 0.241,
      "step": 15420
    },
    {
      "epoch": 0.052873972846833386,
      "grad_norm": 0.9957507252693176,
      "learning_rate": 1e-05,
      "loss": 0.2377,
      "step": 15430
    },
    {
      "epoch": 0.052908239841549416,
      "grad_norm": 1.1372512578964233,
      "learning_rate": 1e-05,
      "loss": 0.2565,
      "step": 15440
    },
    {
      "epoch": 0.052942506836265446,
      "grad_norm": 1.2347620725631714,
      "learning_rate": 1e-05,
      "loss": 0.2392,
      "step": 15450
    },
    {
      "epoch": 0.052976773830981476,
      "grad_norm": 1.0665175914764404,
      "learning_rate": 1e-05,
      "loss": 0.2433,
      "step": 15460
    },
    {
      "epoch": 0.053011040825697506,
      "grad_norm": 1.2257585525512695,
      "learning_rate": 1e-05,
      "loss": 0.2371,
      "step": 15470
    },
    {
      "epoch": 0.053045307820413536,
      "grad_norm": 1.1391757726669312,
      "learning_rate": 1e-05,
      "loss": 0.2368,
      "step": 15480
    },
    {
      "epoch": 0.053079574815129565,
      "grad_norm": 1.0963542461395264,
      "learning_rate": 1e-05,
      "loss": 0.2318,
      "step": 15490
    },
    {
      "epoch": 0.053113841809845595,
      "grad_norm": 1.0512635707855225,
      "learning_rate": 1e-05,
      "loss": 0.2445,
      "step": 15500
    },
    {
      "epoch": 0.053113841809845595,
      "eval_cer": 12.947609807726229,
      "eval_loss": 0.2476710081100464,
      "eval_normalized_cer": 9.192645883293366,
      "eval_runtime": 227.5054,
      "eval_samples_per_second": 2.25,
      "eval_steps_per_second": 0.035,
      "step": 15500
    },
    {
      "epoch": 0.053148108804561625,
      "grad_norm": 1.1827884912490845,
      "learning_rate": 1e-05,
      "loss": 0.2591,
      "step": 15510
    },
    {
      "epoch": 0.053182375799277655,
      "grad_norm": 1.1053601503372192,
      "learning_rate": 1e-05,
      "loss": 0.2313,
      "step": 15520
    },
    {
      "epoch": 0.05321664279399368,
      "grad_norm": 0.9013387560844421,
      "learning_rate": 1e-05,
      "loss": 0.2418,
      "step": 15530
    },
    {
      "epoch": 0.05325090978870971,
      "grad_norm": 1.2677943706512451,
      "learning_rate": 1e-05,
      "loss": 0.2374,
      "step": 15540
    },
    {
      "epoch": 0.05328517678342574,
      "grad_norm": 1.2304824590682983,
      "learning_rate": 1e-05,
      "loss": 0.2598,
      "step": 15550
    },
    {
      "epoch": 0.05331944377814177,
      "grad_norm": 1.0899723768234253,
      "learning_rate": 1e-05,
      "loss": 0.2278,
      "step": 15560
    },
    {
      "epoch": 0.0533537107728578,
      "grad_norm": 1.1334359645843506,
      "learning_rate": 1e-05,
      "loss": 0.2428,
      "step": 15570
    },
    {
      "epoch": 0.05338797776757383,
      "grad_norm": 0.9581900835037231,
      "learning_rate": 1e-05,
      "loss": 0.2353,
      "step": 15580
    },
    {
      "epoch": 0.05342224476228986,
      "grad_norm": 1.1402366161346436,
      "learning_rate": 1e-05,
      "loss": 0.2224,
      "step": 15590
    },
    {
      "epoch": 0.05345651175700589,
      "grad_norm": 1.0247036218643188,
      "learning_rate": 1e-05,
      "loss": 0.2364,
      "step": 15600
    },
    {
      "epoch": 0.05349077875172192,
      "grad_norm": 1.0679134130477905,
      "learning_rate": 1e-05,
      "loss": 0.2338,
      "step": 15610
    },
    {
      "epoch": 0.053525045746437946,
      "grad_norm": 1.0253273248672485,
      "learning_rate": 1e-05,
      "loss": 0.2338,
      "step": 15620
    },
    {
      "epoch": 0.053559312741153976,
      "grad_norm": 1.066657543182373,
      "learning_rate": 1e-05,
      "loss": 0.2321,
      "step": 15630
    },
    {
      "epoch": 0.053593579735870006,
      "grad_norm": 1.1637462377548218,
      "learning_rate": 1e-05,
      "loss": 0.2452,
      "step": 15640
    },
    {
      "epoch": 0.053627846730586036,
      "grad_norm": 1.0810452699661255,
      "learning_rate": 1e-05,
      "loss": 0.2385,
      "step": 15650
    },
    {
      "epoch": 0.053662113725302066,
      "grad_norm": 1.1184097528457642,
      "learning_rate": 1e-05,
      "loss": 0.2448,
      "step": 15660
    },
    {
      "epoch": 0.053696380720018096,
      "grad_norm": 0.9914514422416687,
      "learning_rate": 1e-05,
      "loss": 0.2246,
      "step": 15670
    },
    {
      "epoch": 0.053730647714734125,
      "grad_norm": 1.1371618509292603,
      "learning_rate": 1e-05,
      "loss": 0.2434,
      "step": 15680
    },
    {
      "epoch": 0.05376491470945015,
      "grad_norm": 1.0953712463378906,
      "learning_rate": 1e-05,
      "loss": 0.2414,
      "step": 15690
    },
    {
      "epoch": 0.05379918170416618,
      "grad_norm": 1.05868399143219,
      "learning_rate": 1e-05,
      "loss": 0.2533,
      "step": 15700
    },
    {
      "epoch": 0.05383344869888221,
      "grad_norm": 1.2226771116256714,
      "learning_rate": 1e-05,
      "loss": 0.2342,
      "step": 15710
    },
    {
      "epoch": 0.05386771569359824,
      "grad_norm": 1.1739540100097656,
      "learning_rate": 1e-05,
      "loss": 0.2335,
      "step": 15720
    },
    {
      "epoch": 0.05390198268831427,
      "grad_norm": 1.1677112579345703,
      "learning_rate": 1e-05,
      "loss": 0.2373,
      "step": 15730
    },
    {
      "epoch": 0.0539362496830303,
      "grad_norm": 1.040004849433899,
      "learning_rate": 1e-05,
      "loss": 0.2385,
      "step": 15740
    },
    {
      "epoch": 0.05397051667774633,
      "grad_norm": 1.0462521314620972,
      "learning_rate": 1e-05,
      "loss": 0.2213,
      "step": 15750
    },
    {
      "epoch": 0.05400478367246236,
      "grad_norm": 1.1457821130752563,
      "learning_rate": 1e-05,
      "loss": 0.234,
      "step": 15760
    },
    {
      "epoch": 0.05403905066717839,
      "grad_norm": 1.1003904342651367,
      "learning_rate": 1e-05,
      "loss": 0.2378,
      "step": 15770
    },
    {
      "epoch": 0.05407331766189442,
      "grad_norm": 0.9855090975761414,
      "learning_rate": 1e-05,
      "loss": 0.2456,
      "step": 15780
    },
    {
      "epoch": 0.05410758465661045,
      "grad_norm": 0.9081568717956543,
      "learning_rate": 1e-05,
      "loss": 0.2283,
      "step": 15790
    },
    {
      "epoch": 0.05414185165132648,
      "grad_norm": 1.0517041683197021,
      "learning_rate": 1e-05,
      "loss": 0.241,
      "step": 15800
    },
    {
      "epoch": 0.054176118646042506,
      "grad_norm": 1.098231315612793,
      "learning_rate": 1e-05,
      "loss": 0.2364,
      "step": 15810
    },
    {
      "epoch": 0.054210385640758536,
      "grad_norm": 1.2353124618530273,
      "learning_rate": 1e-05,
      "loss": 0.223,
      "step": 15820
    },
    {
      "epoch": 0.054244652635474566,
      "grad_norm": 1.2910332679748535,
      "learning_rate": 1e-05,
      "loss": 0.2391,
      "step": 15830
    },
    {
      "epoch": 0.054278919630190596,
      "grad_norm": 0.9715086221694946,
      "learning_rate": 1e-05,
      "loss": 0.232,
      "step": 15840
    },
    {
      "epoch": 0.054313186624906626,
      "grad_norm": 0.9929107427597046,
      "learning_rate": 1e-05,
      "loss": 0.2425,
      "step": 15850
    },
    {
      "epoch": 0.05434745361962265,
      "grad_norm": 1.104021430015564,
      "learning_rate": 1e-05,
      "loss": 0.2242,
      "step": 15860
    },
    {
      "epoch": 0.05438172061433868,
      "grad_norm": 1.1469321250915527,
      "learning_rate": 1e-05,
      "loss": 0.2403,
      "step": 15870
    },
    {
      "epoch": 0.05441598760905471,
      "grad_norm": 1.158736228942871,
      "learning_rate": 1e-05,
      "loss": 0.2543,
      "step": 15880
    },
    {
      "epoch": 0.05445025460377074,
      "grad_norm": 1.034866213798523,
      "learning_rate": 1e-05,
      "loss": 0.2524,
      "step": 15890
    },
    {
      "epoch": 0.05448452159848677,
      "grad_norm": 1.0600367784500122,
      "learning_rate": 1e-05,
      "loss": 0.2434,
      "step": 15900
    },
    {
      "epoch": 0.0545187885932028,
      "grad_norm": 1.2206172943115234,
      "learning_rate": 1e-05,
      "loss": 0.2419,
      "step": 15910
    },
    {
      "epoch": 0.05455305558791883,
      "grad_norm": 1.3504046201705933,
      "learning_rate": 1e-05,
      "loss": 0.2663,
      "step": 15920
    },
    {
      "epoch": 0.05458732258263486,
      "grad_norm": 0.9881328344345093,
      "learning_rate": 1e-05,
      "loss": 0.2507,
      "step": 15930
    },
    {
      "epoch": 0.05462158957735089,
      "grad_norm": 1.104619026184082,
      "learning_rate": 1e-05,
      "loss": 0.2552,
      "step": 15940
    },
    {
      "epoch": 0.05465585657206692,
      "grad_norm": 0.9734206795692444,
      "learning_rate": 1e-05,
      "loss": 0.2552,
      "step": 15950
    },
    {
      "epoch": 0.05469012356678295,
      "grad_norm": 1.0191655158996582,
      "learning_rate": 1e-05,
      "loss": 0.2564,
      "step": 15960
    },
    {
      "epoch": 0.05472439056149898,
      "grad_norm": 1.0736887454986572,
      "learning_rate": 1e-05,
      "loss": 0.235,
      "step": 15970
    },
    {
      "epoch": 0.05475865755621501,
      "grad_norm": 0.9910275936126709,
      "learning_rate": 1e-05,
      "loss": 0.2476,
      "step": 15980
    },
    {
      "epoch": 0.05479292455093104,
      "grad_norm": 1.2393155097961426,
      "learning_rate": 1e-05,
      "loss": 0.2613,
      "step": 15990
    },
    {
      "epoch": 0.05482719154564707,
      "grad_norm": 1.1756526231765747,
      "learning_rate": 1e-05,
      "loss": 0.2568,
      "step": 16000
    },
    {
      "epoch": 0.05482719154564707,
      "eval_cer": 13.300405715293703,
      "eval_loss": 0.2448866218328476,
      "eval_normalized_cer": 9.672262190247801,
      "eval_runtime": 228.492,
      "eval_samples_per_second": 2.241,
      "eval_steps_per_second": 0.035,
      "step": 16000
    },
    {
      "epoch": 0.054861458540363096,
      "grad_norm": 1.2805075645446777,
      "learning_rate": 1e-05,
      "loss": 0.2613,
      "step": 16010
    },
    {
      "epoch": 0.05489572553507912,
      "grad_norm": 1.01006019115448,
      "learning_rate": 1e-05,
      "loss": 0.2767,
      "step": 16020
    },
    {
      "epoch": 0.05492999252979515,
      "grad_norm": 1.2456789016723633,
      "learning_rate": 1e-05,
      "loss": 0.2519,
      "step": 16030
    },
    {
      "epoch": 0.05496425952451118,
      "grad_norm": 1.2959610223770142,
      "learning_rate": 1e-05,
      "loss": 0.2583,
      "step": 16040
    },
    {
      "epoch": 0.05499852651922721,
      "grad_norm": 1.1990777254104614,
      "learning_rate": 1e-05,
      "loss": 0.2465,
      "step": 16050
    },
    {
      "epoch": 0.05503279351394324,
      "grad_norm": 1.0876649618148804,
      "learning_rate": 1e-05,
      "loss": 0.2602,
      "step": 16060
    },
    {
      "epoch": 0.05506706050865927,
      "grad_norm": 0.9991684556007385,
      "learning_rate": 1e-05,
      "loss": 0.2363,
      "step": 16070
    },
    {
      "epoch": 0.0551013275033753,
      "grad_norm": 1.1685731410980225,
      "learning_rate": 1e-05,
      "loss": 0.2489,
      "step": 16080
    },
    {
      "epoch": 0.05513559449809133,
      "grad_norm": 1.086849331855774,
      "learning_rate": 1e-05,
      "loss": 0.2525,
      "step": 16090
    },
    {
      "epoch": 0.05516986149280736,
      "grad_norm": 1.0985273122787476,
      "learning_rate": 1e-05,
      "loss": 0.2482,
      "step": 16100
    },
    {
      "epoch": 0.05520412848752339,
      "grad_norm": 1.112008810043335,
      "learning_rate": 1e-05,
      "loss": 0.244,
      "step": 16110
    },
    {
      "epoch": 0.05523839548223942,
      "grad_norm": 1.068474531173706,
      "learning_rate": 1e-05,
      "loss": 0.2349,
      "step": 16120
    },
    {
      "epoch": 0.05527266247695545,
      "grad_norm": 1.001871943473816,
      "learning_rate": 1e-05,
      "loss": 0.242,
      "step": 16130
    },
    {
      "epoch": 0.05530692947167148,
      "grad_norm": 1.1889883279800415,
      "learning_rate": 1e-05,
      "loss": 0.2293,
      "step": 16140
    },
    {
      "epoch": 0.05534119646638751,
      "grad_norm": 1.1890829801559448,
      "learning_rate": 1e-05,
      "loss": 0.2399,
      "step": 16150
    },
    {
      "epoch": 0.05537546346110354,
      "grad_norm": 1.0574150085449219,
      "learning_rate": 1e-05,
      "loss": 0.2379,
      "step": 16160
    },
    {
      "epoch": 0.05540973045581957,
      "grad_norm": 1.1458338499069214,
      "learning_rate": 1e-05,
      "loss": 0.2501,
      "step": 16170
    },
    {
      "epoch": 0.05544399745053559,
      "grad_norm": 1.0738166570663452,
      "learning_rate": 1e-05,
      "loss": 0.2292,
      "step": 16180
    },
    {
      "epoch": 0.05547826444525162,
      "grad_norm": 1.1803652048110962,
      "learning_rate": 1e-05,
      "loss": 0.2286,
      "step": 16190
    },
    {
      "epoch": 0.05551253143996765,
      "grad_norm": 1.02376127243042,
      "learning_rate": 1e-05,
      "loss": 0.2495,
      "step": 16200
    },
    {
      "epoch": 0.05554679843468368,
      "grad_norm": 0.9988088011741638,
      "learning_rate": 1e-05,
      "loss": 0.2191,
      "step": 16210
    },
    {
      "epoch": 0.05558106542939971,
      "grad_norm": 1.2596487998962402,
      "learning_rate": 1e-05,
      "loss": 0.2373,
      "step": 16220
    },
    {
      "epoch": 0.05561533242411574,
      "grad_norm": 1.1729799509048462,
      "learning_rate": 1e-05,
      "loss": 0.2355,
      "step": 16230
    },
    {
      "epoch": 0.05564959941883177,
      "grad_norm": 1.0790057182312012,
      "learning_rate": 1e-05,
      "loss": 0.2211,
      "step": 16240
    },
    {
      "epoch": 0.0556838664135478,
      "grad_norm": 1.044366717338562,
      "learning_rate": 1e-05,
      "loss": 0.226,
      "step": 16250
    },
    {
      "epoch": 0.05571813340826383,
      "grad_norm": 1.1206332445144653,
      "learning_rate": 1e-05,
      "loss": 0.2121,
      "step": 16260
    },
    {
      "epoch": 0.05575240040297986,
      "grad_norm": 1.0622124671936035,
      "learning_rate": 1e-05,
      "loss": 0.2179,
      "step": 16270
    },
    {
      "epoch": 0.05578666739769589,
      "grad_norm": 1.1448779106140137,
      "learning_rate": 1e-05,
      "loss": 0.2266,
      "step": 16280
    },
    {
      "epoch": 0.05582093439241192,
      "grad_norm": 0.9783304929733276,
      "learning_rate": 1e-05,
      "loss": 0.2254,
      "step": 16290
    },
    {
      "epoch": 0.05585520138712795,
      "grad_norm": 1.2645550966262817,
      "learning_rate": 1e-05,
      "loss": 0.2235,
      "step": 16300
    },
    {
      "epoch": 0.05588946838184398,
      "grad_norm": 1.1020660400390625,
      "learning_rate": 1e-05,
      "loss": 0.2363,
      "step": 16310
    },
    {
      "epoch": 0.05592373537656001,
      "grad_norm": 1.1464896202087402,
      "learning_rate": 1e-05,
      "loss": 0.2407,
      "step": 16320
    },
    {
      "epoch": 0.05595800237127604,
      "grad_norm": 1.2940075397491455,
      "learning_rate": 1e-05,
      "loss": 0.2315,
      "step": 16330
    },
    {
      "epoch": 0.05599226936599206,
      "grad_norm": 1.0869808197021484,
      "learning_rate": 1e-05,
      "loss": 0.2176,
      "step": 16340
    },
    {
      "epoch": 0.05602653636070809,
      "grad_norm": 1.025872826576233,
      "learning_rate": 1e-05,
      "loss": 0.2138,
      "step": 16350
    },
    {
      "epoch": 0.05606080335542412,
      "grad_norm": 1.0538456439971924,
      "learning_rate": 1e-05,
      "loss": 0.2204,
      "step": 16360
    },
    {
      "epoch": 0.05609507035014015,
      "grad_norm": 1.0765165090560913,
      "learning_rate": 1e-05,
      "loss": 0.2327,
      "step": 16370
    },
    {
      "epoch": 0.05612933734485618,
      "grad_norm": 1.1909908056259155,
      "learning_rate": 1e-05,
      "loss": 0.239,
      "step": 16380
    },
    {
      "epoch": 0.05616360433957221,
      "grad_norm": 1.249849557876587,
      "learning_rate": 1e-05,
      "loss": 0.2318,
      "step": 16390
    },
    {
      "epoch": 0.05619787133428824,
      "grad_norm": 1.0325443744659424,
      "learning_rate": 1e-05,
      "loss": 0.2372,
      "step": 16400
    },
    {
      "epoch": 0.05623213832900427,
      "grad_norm": 2.3129172325134277,
      "learning_rate": 1e-05,
      "loss": 0.2334,
      "step": 16410
    },
    {
      "epoch": 0.0562664053237203,
      "grad_norm": 1.1196421384811401,
      "learning_rate": 1e-05,
      "loss": 0.2389,
      "step": 16420
    },
    {
      "epoch": 0.05630067231843633,
      "grad_norm": 1.170823335647583,
      "learning_rate": 1e-05,
      "loss": 0.2435,
      "step": 16430
    },
    {
      "epoch": 0.05633493931315236,
      "grad_norm": 1.2110240459442139,
      "learning_rate": 1e-05,
      "loss": 0.2499,
      "step": 16440
    },
    {
      "epoch": 0.05636920630786839,
      "grad_norm": 1.2487872838974,
      "learning_rate": 1e-05,
      "loss": 0.2556,
      "step": 16450
    },
    {
      "epoch": 0.05640347330258442,
      "grad_norm": 1.0886249542236328,
      "learning_rate": 1e-05,
      "loss": 0.2297,
      "step": 16460
    },
    {
      "epoch": 0.05643774029730045,
      "grad_norm": 1.274640679359436,
      "learning_rate": 1e-05,
      "loss": 0.236,
      "step": 16470
    },
    {
      "epoch": 0.05647200729201648,
      "grad_norm": 1.1046870946884155,
      "learning_rate": 1e-05,
      "loss": 0.2497,
      "step": 16480
    },
    {
      "epoch": 0.05650627428673251,
      "grad_norm": 1.0462591648101807,
      "learning_rate": 1e-05,
      "loss": 0.2298,
      "step": 16490
    },
    {
      "epoch": 0.05654054128144854,
      "grad_norm": 1.1058859825134277,
      "learning_rate": 1e-05,
      "loss": 0.2367,
      "step": 16500
    },
    {
      "epoch": 0.05654054128144854,
      "eval_cer": 13.247486329158582,
      "eval_loss": 0.24819624423980713,
      "eval_normalized_cer": 9.602318145483613,
      "eval_runtime": 228.0591,
      "eval_samples_per_second": 2.245,
      "eval_steps_per_second": 0.035,
      "step": 16500
    },
    {
      "epoch": 0.05657480827616456,
      "grad_norm": 1.1613426208496094,
      "learning_rate": 1e-05,
      "loss": 0.2318,
      "step": 16510
    },
    {
      "epoch": 0.05660907527088059,
      "grad_norm": 1.1319721937179565,
      "learning_rate": 1e-05,
      "loss": 0.2461,
      "step": 16520
    },
    {
      "epoch": 0.05664334226559662,
      "grad_norm": 1.2690738439559937,
      "learning_rate": 1e-05,
      "loss": 0.2392,
      "step": 16530
    },
    {
      "epoch": 0.05667760926031265,
      "grad_norm": 1.119575023651123,
      "learning_rate": 1e-05,
      "loss": 0.2467,
      "step": 16540
    },
    {
      "epoch": 0.05671187625502868,
      "grad_norm": 1.119841456413269,
      "learning_rate": 1e-05,
      "loss": 0.223,
      "step": 16550
    },
    {
      "epoch": 0.05674614324974471,
      "grad_norm": 1.0248748064041138,
      "learning_rate": 1e-05,
      "loss": 0.2392,
      "step": 16560
    },
    {
      "epoch": 0.05678041024446074,
      "grad_norm": 1.0252872705459595,
      "learning_rate": 1e-05,
      "loss": 0.2426,
      "step": 16570
    },
    {
      "epoch": 0.05681467723917677,
      "grad_norm": 1.1694291830062866,
      "learning_rate": 1e-05,
      "loss": 0.2078,
      "step": 16580
    },
    {
      "epoch": 0.0568489442338928,
      "grad_norm": 1.152016520500183,
      "learning_rate": 1e-05,
      "loss": 0.2423,
      "step": 16590
    },
    {
      "epoch": 0.05688321122860883,
      "grad_norm": 0.9418520927429199,
      "learning_rate": 1e-05,
      "loss": 0.2214,
      "step": 16600
    },
    {
      "epoch": 0.05691747822332486,
      "grad_norm": 1.1249022483825684,
      "learning_rate": 1e-05,
      "loss": 0.2222,
      "step": 16610
    },
    {
      "epoch": 0.05695174521804089,
      "grad_norm": 1.0502556562423706,
      "learning_rate": 1e-05,
      "loss": 0.2366,
      "step": 16620
    },
    {
      "epoch": 0.05698601221275692,
      "grad_norm": 1.0879663228988647,
      "learning_rate": 1e-05,
      "loss": 0.2253,
      "step": 16630
    },
    {
      "epoch": 0.05702027920747295,
      "grad_norm": 1.0701017379760742,
      "learning_rate": 1e-05,
      "loss": 0.2369,
      "step": 16640
    },
    {
      "epoch": 0.05705454620218898,
      "grad_norm": 1.0117312669754028,
      "learning_rate": 1e-05,
      "loss": 0.2419,
      "step": 16650
    },
    {
      "epoch": 0.05708881319690501,
      "grad_norm": 1.1599793434143066,
      "learning_rate": 1e-05,
      "loss": 0.2444,
      "step": 16660
    },
    {
      "epoch": 0.05712308019162103,
      "grad_norm": 1.1197978258132935,
      "learning_rate": 1e-05,
      "loss": 0.2388,
      "step": 16670
    },
    {
      "epoch": 0.05715734718633706,
      "grad_norm": 1.1201450824737549,
      "learning_rate": 1e-05,
      "loss": 0.2401,
      "step": 16680
    },
    {
      "epoch": 0.05719161418105309,
      "grad_norm": 1.0935217142105103,
      "learning_rate": 1e-05,
      "loss": 0.2191,
      "step": 16690
    },
    {
      "epoch": 0.05722588117576912,
      "grad_norm": 1.237023949623108,
      "learning_rate": 1e-05,
      "loss": 0.2509,
      "step": 16700
    },
    {
      "epoch": 0.05726014817048515,
      "grad_norm": 1.0313085317611694,
      "learning_rate": 1e-05,
      "loss": 0.2271,
      "step": 16710
    },
    {
      "epoch": 0.05729441516520118,
      "grad_norm": 1.2026563882827759,
      "learning_rate": 1e-05,
      "loss": 0.2344,
      "step": 16720
    },
    {
      "epoch": 0.05732868215991721,
      "grad_norm": 1.2139136791229248,
      "learning_rate": 1e-05,
      "loss": 0.2425,
      "step": 16730
    },
    {
      "epoch": 0.05736294915463324,
      "grad_norm": 1.06145179271698,
      "learning_rate": 1e-05,
      "loss": 0.2547,
      "step": 16740
    },
    {
      "epoch": 0.05739721614934927,
      "grad_norm": 1.1620399951934814,
      "learning_rate": 1e-05,
      "loss": 0.228,
      "step": 16750
    },
    {
      "epoch": 0.0574314831440653,
      "grad_norm": 1.0586286783218384,
      "learning_rate": 1e-05,
      "loss": 0.2202,
      "step": 16760
    },
    {
      "epoch": 0.05746575013878133,
      "grad_norm": 1.0937540531158447,
      "learning_rate": 1e-05,
      "loss": 0.2415,
      "step": 16770
    },
    {
      "epoch": 0.05750001713349736,
      "grad_norm": 1.0289047956466675,
      "learning_rate": 1e-05,
      "loss": 0.2327,
      "step": 16780
    },
    {
      "epoch": 0.05753428412821339,
      "grad_norm": 1.0515446662902832,
      "learning_rate": 1e-05,
      "loss": 0.2306,
      "step": 16790
    },
    {
      "epoch": 0.05756855112292942,
      "grad_norm": 0.9734529852867126,
      "learning_rate": 1e-05,
      "loss": 0.2196,
      "step": 16800
    },
    {
      "epoch": 0.05760281811764545,
      "grad_norm": 1.0374795198440552,
      "learning_rate": 1e-05,
      "loss": 0.2263,
      "step": 16810
    },
    {
      "epoch": 0.05763708511236148,
      "grad_norm": 1.0560572147369385,
      "learning_rate": 1e-05,
      "loss": 0.2466,
      "step": 16820
    },
    {
      "epoch": 0.0576713521070775,
      "grad_norm": 0.9350127577781677,
      "learning_rate": 1e-05,
      "loss": 0.229,
      "step": 16830
    },
    {
      "epoch": 0.05770561910179353,
      "grad_norm": 1.180124282836914,
      "learning_rate": 1e-05,
      "loss": 0.2395,
      "step": 16840
    },
    {
      "epoch": 0.05773988609650956,
      "grad_norm": 1.17545747756958,
      "learning_rate": 1e-05,
      "loss": 0.2427,
      "step": 16850
    },
    {
      "epoch": 0.05777415309122559,
      "grad_norm": 1.1822388172149658,
      "learning_rate": 1e-05,
      "loss": 0.2525,
      "step": 16860
    },
    {
      "epoch": 0.05780842008594162,
      "grad_norm": 1.0005474090576172,
      "learning_rate": 1e-05,
      "loss": 0.2215,
      "step": 16870
    },
    {
      "epoch": 0.05784268708065765,
      "grad_norm": 1.2070783376693726,
      "learning_rate": 1e-05,
      "loss": 0.2476,
      "step": 16880
    },
    {
      "epoch": 0.05787695407537368,
      "grad_norm": 1.3960411548614502,
      "learning_rate": 1e-05,
      "loss": 0.2535,
      "step": 16890
    },
    {
      "epoch": 0.05791122107008971,
      "grad_norm": 1.1246318817138672,
      "learning_rate": 1e-05,
      "loss": 0.2532,
      "step": 16900
    },
    {
      "epoch": 0.05794548806480574,
      "grad_norm": 1.042833685874939,
      "learning_rate": 1e-05,
      "loss": 0.2453,
      "step": 16910
    },
    {
      "epoch": 0.05797975505952177,
      "grad_norm": 1.1697344779968262,
      "learning_rate": 1e-05,
      "loss": 0.2396,
      "step": 16920
    },
    {
      "epoch": 0.0580140220542378,
      "grad_norm": 1.1674904823303223,
      "learning_rate": 1e-05,
      "loss": 0.2413,
      "step": 16930
    },
    {
      "epoch": 0.05804828904895383,
      "grad_norm": 1.2486639022827148,
      "learning_rate": 1e-05,
      "loss": 0.2397,
      "step": 16940
    },
    {
      "epoch": 0.05808255604366986,
      "grad_norm": 1.0194487571716309,
      "learning_rate": 1e-05,
      "loss": 0.2323,
      "step": 16950
    },
    {
      "epoch": 0.05811682303838589,
      "grad_norm": 1.1312390565872192,
      "learning_rate": 1e-05,
      "loss": 0.238,
      "step": 16960
    },
    {
      "epoch": 0.05815109003310192,
      "grad_norm": 1.0825895071029663,
      "learning_rate": 1e-05,
      "loss": 0.2523,
      "step": 16970
    },
    {
      "epoch": 0.05818535702781795,
      "grad_norm": 1.1746137142181396,
      "learning_rate": 1e-05,
      "loss": 0.2389,
      "step": 16980
    },
    {
      "epoch": 0.05821962402253397,
      "grad_norm": 1.1029731035232544,
      "learning_rate": 1e-05,
      "loss": 0.2347,
      "step": 16990
    },
    {
      "epoch": 0.05825389101725,
      "grad_norm": 1.1937546730041504,
      "learning_rate": 1e-05,
      "loss": 0.2492,
      "step": 17000
    },
    {
      "epoch": 0.05825389101725,
      "eval_cer": 12.621273593226318,
      "eval_loss": 0.24555271863937378,
      "eval_normalized_cer": 9.272581934452438,
      "eval_runtime": 227.9675,
      "eval_samples_per_second": 2.246,
      "eval_steps_per_second": 0.035,
      "step": 17000
    },
    {
      "epoch": 0.05828815801196603,
      "grad_norm": 1.1056500673294067,
      "learning_rate": 1e-05,
      "loss": 0.2267,
      "step": 17010
    },
    {
      "epoch": 0.05832242500668206,
      "grad_norm": 1.1719837188720703,
      "learning_rate": 1e-05,
      "loss": 0.2567,
      "step": 17020
    },
    {
      "epoch": 0.05835669200139809,
      "grad_norm": 1.0683894157409668,
      "learning_rate": 1e-05,
      "loss": 0.2185,
      "step": 17030
    },
    {
      "epoch": 0.05839095899611412,
      "grad_norm": 1.0233598947525024,
      "learning_rate": 1e-05,
      "loss": 0.2233,
      "step": 17040
    },
    {
      "epoch": 0.05842522599083015,
      "grad_norm": 0.9998723864555359,
      "learning_rate": 1e-05,
      "loss": 0.236,
      "step": 17050
    },
    {
      "epoch": 0.05845949298554618,
      "grad_norm": 1.1496772766113281,
      "learning_rate": 1e-05,
      "loss": 0.2409,
      "step": 17060
    },
    {
      "epoch": 0.05849375998026221,
      "grad_norm": 1.112373948097229,
      "learning_rate": 1e-05,
      "loss": 0.2435,
      "step": 17070
    },
    {
      "epoch": 0.05852802697497824,
      "grad_norm": 0.9641957879066467,
      "learning_rate": 1e-05,
      "loss": 0.2398,
      "step": 17080
    },
    {
      "epoch": 0.05856229396969427,
      "grad_norm": 1.1222591400146484,
      "learning_rate": 1e-05,
      "loss": 0.2295,
      "step": 17090
    },
    {
      "epoch": 0.0585965609644103,
      "grad_norm": 1.1070595979690552,
      "learning_rate": 1e-05,
      "loss": 0.2393,
      "step": 17100
    },
    {
      "epoch": 0.05863082795912633,
      "grad_norm": 1.2548716068267822,
      "learning_rate": 1e-05,
      "loss": 0.2446,
      "step": 17110
    },
    {
      "epoch": 0.05866509495384236,
      "grad_norm": 1.2774968147277832,
      "learning_rate": 1e-05,
      "loss": 0.2496,
      "step": 17120
    },
    {
      "epoch": 0.05869936194855839,
      "grad_norm": 1.0819348096847534,
      "learning_rate": 1e-05,
      "loss": 0.2413,
      "step": 17130
    },
    {
      "epoch": 0.05873362894327442,
      "grad_norm": 1.1262927055358887,
      "learning_rate": 1e-05,
      "loss": 0.2231,
      "step": 17140
    },
    {
      "epoch": 0.05876789593799045,
      "grad_norm": 1.0115349292755127,
      "learning_rate": 1e-05,
      "loss": 0.2367,
      "step": 17150
    },
    {
      "epoch": 0.05880216293270647,
      "grad_norm": 1.1211316585540771,
      "learning_rate": 1e-05,
      "loss": 0.2455,
      "step": 17160
    },
    {
      "epoch": 0.0588364299274225,
      "grad_norm": 1.213244080543518,
      "learning_rate": 1e-05,
      "loss": 0.2067,
      "step": 17170
    },
    {
      "epoch": 0.05887069692213853,
      "grad_norm": 1.1806918382644653,
      "learning_rate": 1e-05,
      "loss": 0.2359,
      "step": 17180
    },
    {
      "epoch": 0.05890496391685456,
      "grad_norm": 1.0859711170196533,
      "learning_rate": 1e-05,
      "loss": 0.2327,
      "step": 17190
    },
    {
      "epoch": 0.05893923091157059,
      "grad_norm": 1.084893822669983,
      "learning_rate": 1e-05,
      "loss": 0.2515,
      "step": 17200
    },
    {
      "epoch": 0.05897349790628662,
      "grad_norm": 1.1239533424377441,
      "learning_rate": 1e-05,
      "loss": 0.219,
      "step": 17210
    },
    {
      "epoch": 0.05900776490100265,
      "grad_norm": 0.9819833040237427,
      "learning_rate": 1e-05,
      "loss": 0.2283,
      "step": 17220
    },
    {
      "epoch": 0.05904203189571868,
      "grad_norm": 1.075325608253479,
      "learning_rate": 1e-05,
      "loss": 0.219,
      "step": 17230
    },
    {
      "epoch": 0.05907629889043471,
      "grad_norm": 1.141569972038269,
      "learning_rate": 1e-05,
      "loss": 0.2436,
      "step": 17240
    },
    {
      "epoch": 0.05911056588515074,
      "grad_norm": 1.0299127101898193,
      "learning_rate": 1e-05,
      "loss": 0.2275,
      "step": 17250
    },
    {
      "epoch": 0.05914483287986677,
      "grad_norm": 1.084275245666504,
      "learning_rate": 1e-05,
      "loss": 0.2119,
      "step": 17260
    },
    {
      "epoch": 0.0591790998745828,
      "grad_norm": 1.1073039770126343,
      "learning_rate": 1e-05,
      "loss": 0.2272,
      "step": 17270
    },
    {
      "epoch": 0.05921336686929883,
      "grad_norm": 1.2507880926132202,
      "learning_rate": 1e-05,
      "loss": 0.2328,
      "step": 17280
    },
    {
      "epoch": 0.05924763386401486,
      "grad_norm": 0.9464664459228516,
      "learning_rate": 1e-05,
      "loss": 0.2265,
      "step": 17290
    },
    {
      "epoch": 0.05928190085873089,
      "grad_norm": 1.1672245264053345,
      "learning_rate": 1e-05,
      "loss": 0.2426,
      "step": 17300
    },
    {
      "epoch": 0.05931616785344692,
      "grad_norm": 1.1788960695266724,
      "learning_rate": 1e-05,
      "loss": 0.2235,
      "step": 17310
    },
    {
      "epoch": 0.05935043484816294,
      "grad_norm": 1.2637144327163696,
      "learning_rate": 1e-05,
      "loss": 0.2304,
      "step": 17320
    },
    {
      "epoch": 0.05938470184287897,
      "grad_norm": 1.0590788125991821,
      "learning_rate": 1e-05,
      "loss": 0.2486,
      "step": 17330
    },
    {
      "epoch": 0.059418968837595,
      "grad_norm": 0.9958922266960144,
      "learning_rate": 1e-05,
      "loss": 0.2322,
      "step": 17340
    },
    {
      "epoch": 0.05945323583231103,
      "grad_norm": 1.1164788007736206,
      "learning_rate": 1e-05,
      "loss": 0.2338,
      "step": 17350
    },
    {
      "epoch": 0.05948750282702706,
      "grad_norm": 1.0397776365280151,
      "learning_rate": 1e-05,
      "loss": 0.2293,
      "step": 17360
    },
    {
      "epoch": 0.05952176982174309,
      "grad_norm": 1.0913232564926147,
      "learning_rate": 1e-05,
      "loss": 0.2352,
      "step": 17370
    },
    {
      "epoch": 0.05955603681645912,
      "grad_norm": 1.1954469680786133,
      "learning_rate": 1e-05,
      "loss": 0.2473,
      "step": 17380
    },
    {
      "epoch": 0.05959030381117515,
      "grad_norm": 0.9494290947914124,
      "learning_rate": 1e-05,
      "loss": 0.2181,
      "step": 17390
    },
    {
      "epoch": 0.05962457080589118,
      "grad_norm": 0.9757588505744934,
      "learning_rate": 1e-05,
      "loss": 0.214,
      "step": 17400
    },
    {
      "epoch": 0.05965883780060721,
      "grad_norm": 1.0311897993087769,
      "learning_rate": 1e-05,
      "loss": 0.2359,
      "step": 17410
    },
    {
      "epoch": 0.05969310479532324,
      "grad_norm": 1.087958574295044,
      "learning_rate": 1e-05,
      "loss": 0.2268,
      "step": 17420
    },
    {
      "epoch": 0.05972737179003927,
      "grad_norm": 1.053572654724121,
      "learning_rate": 1e-05,
      "loss": 0.2295,
      "step": 17430
    },
    {
      "epoch": 0.0597616387847553,
      "grad_norm": 1.2911629676818848,
      "learning_rate": 1e-05,
      "loss": 0.2391,
      "step": 17440
    },
    {
      "epoch": 0.05979590577947133,
      "grad_norm": 1.1952745914459229,
      "learning_rate": 1e-05,
      "loss": 0.2498,
      "step": 17450
    },
    {
      "epoch": 0.05983017277418736,
      "grad_norm": 1.3044365644454956,
      "learning_rate": 1e-05,
      "loss": 0.2454,
      "step": 17460
    },
    {
      "epoch": 0.05986443976890339,
      "grad_norm": 0.9195178151130676,
      "learning_rate": 1e-05,
      "loss": 0.2312,
      "step": 17470
    },
    {
      "epoch": 0.059898706763619414,
      "grad_norm": 1.1249533891677856,
      "learning_rate": 1e-05,
      "loss": 0.2393,
      "step": 17480
    },
    {
      "epoch": 0.059932973758335444,
      "grad_norm": 1.0563664436340332,
      "learning_rate": 1e-05,
      "loss": 0.2279,
      "step": 17490
    },
    {
      "epoch": 0.05996724075305147,
      "grad_norm": 1.069501519203186,
      "learning_rate": 1e-05,
      "loss": 0.2615,
      "step": 17500
    },
    {
      "epoch": 0.05996724075305147,
      "eval_cer": 12.744752160874933,
      "eval_loss": 0.24637901782989502,
      "eval_normalized_cer": 9.102717825739408,
      "eval_runtime": 227.2598,
      "eval_samples_per_second": 2.253,
      "eval_steps_per_second": 0.035,
      "step": 17500
    },
    {
      "epoch": 0.0600015077477675,
      "grad_norm": 1.0828664302825928,
      "learning_rate": 1e-05,
      "loss": 0.2585,
      "step": 17510
    },
    {
      "epoch": 0.06003577474248353,
      "grad_norm": 1.144832968711853,
      "learning_rate": 1e-05,
      "loss": 0.2707,
      "step": 17520
    },
    {
      "epoch": 0.06007004173719956,
      "grad_norm": 1.1772241592407227,
      "learning_rate": 1e-05,
      "loss": 0.2484,
      "step": 17530
    },
    {
      "epoch": 0.06010430873191559,
      "grad_norm": 1.0854761600494385,
      "learning_rate": 1e-05,
      "loss": 0.2543,
      "step": 17540
    },
    {
      "epoch": 0.06013857572663162,
      "grad_norm": 0.9568967223167419,
      "learning_rate": 1e-05,
      "loss": 0.2328,
      "step": 17550
    },
    {
      "epoch": 0.06017284272134765,
      "grad_norm": 1.1131020784378052,
      "learning_rate": 1e-05,
      "loss": 0.2502,
      "step": 17560
    },
    {
      "epoch": 0.06020710971606368,
      "grad_norm": 1.1434942483901978,
      "learning_rate": 1e-05,
      "loss": 0.2459,
      "step": 17570
    },
    {
      "epoch": 0.06024137671077971,
      "grad_norm": 1.0522421598434448,
      "learning_rate": 1e-05,
      "loss": 0.2276,
      "step": 17580
    },
    {
      "epoch": 0.06027564370549574,
      "grad_norm": 0.9989955425262451,
      "learning_rate": 1e-05,
      "loss": 0.237,
      "step": 17590
    },
    {
      "epoch": 0.06030991070021177,
      "grad_norm": 1.0393595695495605,
      "learning_rate": 1e-05,
      "loss": 0.2534,
      "step": 17600
    },
    {
      "epoch": 0.0603441776949278,
      "grad_norm": 1.1289571523666382,
      "learning_rate": 1e-05,
      "loss": 0.2408,
      "step": 17610
    },
    {
      "epoch": 0.06037844468964383,
      "grad_norm": 1.4361021518707275,
      "learning_rate": 1e-05,
      "loss": 0.2517,
      "step": 17620
    },
    {
      "epoch": 0.06041271168435986,
      "grad_norm": 1.1056772470474243,
      "learning_rate": 1e-05,
      "loss": 0.2496,
      "step": 17630
    },
    {
      "epoch": 0.060446978679075884,
      "grad_norm": 1.1014364957809448,
      "learning_rate": 1e-05,
      "loss": 0.2345,
      "step": 17640
    },
    {
      "epoch": 0.060481245673791914,
      "grad_norm": 1.040041208267212,
      "learning_rate": 1e-05,
      "loss": 0.2478,
      "step": 17650
    },
    {
      "epoch": 0.060515512668507944,
      "grad_norm": 1.0661914348602295,
      "learning_rate": 1e-05,
      "loss": 0.243,
      "step": 17660
    },
    {
      "epoch": 0.060549779663223974,
      "grad_norm": 1.0410475730895996,
      "learning_rate": 1e-05,
      "loss": 0.2423,
      "step": 17670
    },
    {
      "epoch": 0.060584046657940004,
      "grad_norm": 1.140838861465454,
      "learning_rate": 1e-05,
      "loss": 0.2258,
      "step": 17680
    },
    {
      "epoch": 0.06061831365265603,
      "grad_norm": 1.060164451599121,
      "learning_rate": 1e-05,
      "loss": 0.2318,
      "step": 17690
    },
    {
      "epoch": 0.06065258064737206,
      "grad_norm": 0.982888400554657,
      "learning_rate": 1e-05,
      "loss": 0.227,
      "step": 17700
    },
    {
      "epoch": 0.06068684764208809,
      "grad_norm": 1.038228988647461,
      "learning_rate": 1e-05,
      "loss": 0.2193,
      "step": 17710
    },
    {
      "epoch": 0.06072111463680412,
      "grad_norm": 1.1862796545028687,
      "learning_rate": 1e-05,
      "loss": 0.229,
      "step": 17720
    },
    {
      "epoch": 0.06075538163152015,
      "grad_norm": 1.1852322816848755,
      "learning_rate": 1e-05,
      "loss": 0.2319,
      "step": 17730
    },
    {
      "epoch": 0.06078964862623618,
      "grad_norm": 1.2296260595321655,
      "learning_rate": 1e-05,
      "loss": 0.228,
      "step": 17740
    },
    {
      "epoch": 0.06082391562095221,
      "grad_norm": 0.94486004114151,
      "learning_rate": 1e-05,
      "loss": 0.2152,
      "step": 17750
    },
    {
      "epoch": 0.06085818261566824,
      "grad_norm": 1.045202612876892,
      "learning_rate": 1e-05,
      "loss": 0.2366,
      "step": 17760
    },
    {
      "epoch": 0.06089244961038427,
      "grad_norm": 1.0080078840255737,
      "learning_rate": 1e-05,
      "loss": 0.2116,
      "step": 17770
    },
    {
      "epoch": 0.0609267166051003,
      "grad_norm": 1.0263158082962036,
      "learning_rate": 1e-05,
      "loss": 0.2189,
      "step": 17780
    },
    {
      "epoch": 0.06096098359981633,
      "grad_norm": 1.137341022491455,
      "learning_rate": 1e-05,
      "loss": 0.2388,
      "step": 17790
    },
    {
      "epoch": 0.06099525059453236,
      "grad_norm": 1.1267235279083252,
      "learning_rate": 1e-05,
      "loss": 0.2309,
      "step": 17800
    },
    {
      "epoch": 0.061029517589248385,
      "grad_norm": 1.1590142250061035,
      "learning_rate": 1e-05,
      "loss": 0.215,
      "step": 17810
    },
    {
      "epoch": 0.061063784583964414,
      "grad_norm": 1.0661910772323608,
      "learning_rate": 1e-05,
      "loss": 0.2269,
      "step": 17820
    },
    {
      "epoch": 0.061098051578680444,
      "grad_norm": 1.178369402885437,
      "learning_rate": 1e-05,
      "loss": 0.238,
      "step": 17830
    },
    {
      "epoch": 0.061132318573396474,
      "grad_norm": 1.1601570844650269,
      "learning_rate": 1e-05,
      "loss": 0.2183,
      "step": 17840
    },
    {
      "epoch": 0.061166585568112504,
      "grad_norm": 1.098443627357483,
      "learning_rate": 1e-05,
      "loss": 0.2478,
      "step": 17850
    },
    {
      "epoch": 0.061200852562828534,
      "grad_norm": 1.3805477619171143,
      "learning_rate": 1e-05,
      "loss": 0.2497,
      "step": 17860
    },
    {
      "epoch": 0.061235119557544564,
      "grad_norm": 1.1133577823638916,
      "learning_rate": 1e-05,
      "loss": 0.2335,
      "step": 17870
    },
    {
      "epoch": 0.061269386552260593,
      "grad_norm": 1.1253570318222046,
      "learning_rate": 1e-05,
      "loss": 0.2199,
      "step": 17880
    },
    {
      "epoch": 0.06130365354697662,
      "grad_norm": 1.1934068202972412,
      "learning_rate": 1e-05,
      "loss": 0.2304,
      "step": 17890
    },
    {
      "epoch": 0.06133792054169265,
      "grad_norm": 1.1207107305526733,
      "learning_rate": 1e-05,
      "loss": 0.2419,
      "step": 17900
    },
    {
      "epoch": 0.06137218753640868,
      "grad_norm": 0.9266690015792847,
      "learning_rate": 1e-05,
      "loss": 0.234,
      "step": 17910
    },
    {
      "epoch": 0.06140645453112471,
      "grad_norm": 1.296251893043518,
      "learning_rate": 1e-05,
      "loss": 0.2322,
      "step": 17920
    },
    {
      "epoch": 0.06144072152584074,
      "grad_norm": 1.2122479677200317,
      "learning_rate": 1e-05,
      "loss": 0.2467,
      "step": 17930
    },
    {
      "epoch": 0.06147498852055677,
      "grad_norm": 1.2585302591323853,
      "learning_rate": 1e-05,
      "loss": 0.2653,
      "step": 17940
    },
    {
      "epoch": 0.0615092555152728,
      "grad_norm": 1.0809286832809448,
      "learning_rate": 1e-05,
      "loss": 0.231,
      "step": 17950
    },
    {
      "epoch": 0.06154352250998883,
      "grad_norm": 1.1828172206878662,
      "learning_rate": 1e-05,
      "loss": 0.245,
      "step": 17960
    },
    {
      "epoch": 0.061577789504704855,
      "grad_norm": 1.2036045789718628,
      "learning_rate": 1e-05,
      "loss": 0.2543,
      "step": 17970
    },
    {
      "epoch": 0.061612056499420885,
      "grad_norm": 1.0810980796813965,
      "learning_rate": 1e-05,
      "loss": 0.2474,
      "step": 17980
    },
    {
      "epoch": 0.061646323494136915,
      "grad_norm": 1.0696234703063965,
      "learning_rate": 1e-05,
      "loss": 0.2277,
      "step": 17990
    },
    {
      "epoch": 0.061680590488852945,
      "grad_norm": 1.0389044284820557,
      "learning_rate": 1e-05,
      "loss": 0.2471,
      "step": 18000
    },
    {
      "epoch": 0.061680590488852945,
      "eval_cer": 13.08872817075322,
      "eval_loss": 0.24683812260627747,
      "eval_normalized_cer": 9.30255795363709,
      "eval_runtime": 227.6579,
      "eval_samples_per_second": 2.249,
      "eval_steps_per_second": 0.035,
      "step": 18000
    },
    {
      "epoch": 0.061714857483568974,
      "grad_norm": 1.071568250656128,
      "learning_rate": 1e-05,
      "loss": 0.2496,
      "step": 18010
    },
    {
      "epoch": 0.061749124478285004,
      "grad_norm": 1.141601800918579,
      "learning_rate": 1e-05,
      "loss": 0.2607,
      "step": 18020
    },
    {
      "epoch": 0.061783391473001034,
      "grad_norm": 1.1443320512771606,
      "learning_rate": 1e-05,
      "loss": 0.2493,
      "step": 18030
    },
    {
      "epoch": 0.061817658467717064,
      "grad_norm": 1.1623589992523193,
      "learning_rate": 1e-05,
      "loss": 0.2597,
      "step": 18040
    },
    {
      "epoch": 0.061851925462433094,
      "grad_norm": 1.1392161846160889,
      "learning_rate": 1e-05,
      "loss": 0.2445,
      "step": 18050
    },
    {
      "epoch": 0.061886192457149124,
      "grad_norm": 1.1709184646606445,
      "learning_rate": 1e-05,
      "loss": 0.2671,
      "step": 18060
    },
    {
      "epoch": 0.061920459451865154,
      "grad_norm": 1.2600747346878052,
      "learning_rate": 1e-05,
      "loss": 0.2624,
      "step": 18070
    },
    {
      "epoch": 0.06195472644658118,
      "grad_norm": 0.9713681936264038,
      "learning_rate": 1e-05,
      "loss": 0.2399,
      "step": 18080
    },
    {
      "epoch": 0.06198899344129721,
      "grad_norm": 1.1438798904418945,
      "learning_rate": 1e-05,
      "loss": 0.2584,
      "step": 18090
    },
    {
      "epoch": 0.06202326043601324,
      "grad_norm": 0.9961438775062561,
      "learning_rate": 1e-05,
      "loss": 0.251,
      "step": 18100
    },
    {
      "epoch": 0.06205752743072927,
      "grad_norm": 1.0935770273208618,
      "learning_rate": 1e-05,
      "loss": 0.2478,
      "step": 18110
    },
    {
      "epoch": 0.0620917944254453,
      "grad_norm": 1.1075562238693237,
      "learning_rate": 1e-05,
      "loss": 0.2538,
      "step": 18120
    },
    {
      "epoch": 0.062126061420161326,
      "grad_norm": 1.1211317777633667,
      "learning_rate": 1e-05,
      "loss": 0.2524,
      "step": 18130
    },
    {
      "epoch": 0.062160328414877355,
      "grad_norm": 1.1476017236709595,
      "learning_rate": 1e-05,
      "loss": 0.242,
      "step": 18140
    },
    {
      "epoch": 0.062194595409593385,
      "grad_norm": 1.217791199684143,
      "learning_rate": 1e-05,
      "loss": 0.258,
      "step": 18150
    },
    {
      "epoch": 0.062228862404309415,
      "grad_norm": 1.0843958854675293,
      "learning_rate": 1e-05,
      "loss": 0.2562,
      "step": 18160
    },
    {
      "epoch": 0.062263129399025445,
      "grad_norm": 0.9931436777114868,
      "learning_rate": 1e-05,
      "loss": 0.2478,
      "step": 18170
    },
    {
      "epoch": 0.062297396393741475,
      "grad_norm": 1.0133415460586548,
      "learning_rate": 1e-05,
      "loss": 0.2526,
      "step": 18180
    },
    {
      "epoch": 0.062331663388457505,
      "grad_norm": 1.2083938121795654,
      "learning_rate": 1e-05,
      "loss": 0.2604,
      "step": 18190
    },
    {
      "epoch": 0.062365930383173535,
      "grad_norm": 1.0791544914245605,
      "learning_rate": 1e-05,
      "loss": 0.2543,
      "step": 18200
    },
    {
      "epoch": 0.062400197377889564,
      "grad_norm": 1.117967128753662,
      "learning_rate": 1e-05,
      "loss": 0.2551,
      "step": 18210
    },
    {
      "epoch": 0.062434464372605594,
      "grad_norm": 1.1042882204055786,
      "learning_rate": 1e-05,
      "loss": 0.2492,
      "step": 18220
    },
    {
      "epoch": 0.062468731367321624,
      "grad_norm": 1.3039363622665405,
      "learning_rate": 1e-05,
      "loss": 0.2631,
      "step": 18230
    },
    {
      "epoch": 0.06250299836203765,
      "grad_norm": 1.123572587966919,
      "learning_rate": 1e-05,
      "loss": 0.2432,
      "step": 18240
    },
    {
      "epoch": 0.06253726535675368,
      "grad_norm": 1.1475626230239868,
      "learning_rate": 1e-05,
      "loss": 0.2355,
      "step": 18250
    },
    {
      "epoch": 0.06257153235146971,
      "grad_norm": 1.0993095636367798,
      "learning_rate": 1e-05,
      "loss": 0.257,
      "step": 18260
    },
    {
      "epoch": 0.06260579934618574,
      "grad_norm": 1.0722532272338867,
      "learning_rate": 1e-05,
      "loss": 0.2385,
      "step": 18270
    },
    {
      "epoch": 0.06264006634090177,
      "grad_norm": 1.051962971687317,
      "learning_rate": 1e-05,
      "loss": 0.2436,
      "step": 18280
    },
    {
      "epoch": 0.0626743333356178,
      "grad_norm": 1.1819945573806763,
      "learning_rate": 1e-05,
      "loss": 0.2556,
      "step": 18290
    },
    {
      "epoch": 0.06270860033033383,
      "grad_norm": 1.0978039503097534,
      "learning_rate": 1e-05,
      "loss": 0.2352,
      "step": 18300
    },
    {
      "epoch": 0.06274286732504986,
      "grad_norm": 1.1181819438934326,
      "learning_rate": 1e-05,
      "loss": 0.2395,
      "step": 18310
    },
    {
      "epoch": 0.06277713431976589,
      "grad_norm": 1.1631795167922974,
      "learning_rate": 1e-05,
      "loss": 0.2475,
      "step": 18320
    },
    {
      "epoch": 0.06281140131448192,
      "grad_norm": 1.160852313041687,
      "learning_rate": 1e-05,
      "loss": 0.2481,
      "step": 18330
    },
    {
      "epoch": 0.06284566830919795,
      "grad_norm": 0.9235048890113831,
      "learning_rate": 1e-05,
      "loss": 0.2289,
      "step": 18340
    },
    {
      "epoch": 0.06287993530391398,
      "grad_norm": 1.215062141418457,
      "learning_rate": 1e-05,
      "loss": 0.2366,
      "step": 18350
    },
    {
      "epoch": 0.06291420229863001,
      "grad_norm": 1.2124155759811401,
      "learning_rate": 1e-05,
      "loss": 0.2343,
      "step": 18360
    },
    {
      "epoch": 0.06294846929334603,
      "grad_norm": 1.0954176187515259,
      "learning_rate": 1e-05,
      "loss": 0.2468,
      "step": 18370
    },
    {
      "epoch": 0.06298273628806206,
      "grad_norm": 1.1772414445877075,
      "learning_rate": 1e-05,
      "loss": 0.2383,
      "step": 18380
    },
    {
      "epoch": 0.0630170032827781,
      "grad_norm": 1.0725330114364624,
      "learning_rate": 1e-05,
      "loss": 0.2434,
      "step": 18390
    },
    {
      "epoch": 0.06305127027749412,
      "grad_norm": 1.178328037261963,
      "learning_rate": 1e-05,
      "loss": 0.2377,
      "step": 18400
    },
    {
      "epoch": 0.06308553727221015,
      "grad_norm": 1.1152503490447998,
      "learning_rate": 1e-05,
      "loss": 0.2386,
      "step": 18410
    },
    {
      "epoch": 0.06311980426692618,
      "grad_norm": 1.273404836654663,
      "learning_rate": 1e-05,
      "loss": 0.2674,
      "step": 18420
    },
    {
      "epoch": 0.06315407126164221,
      "grad_norm": 1.1588289737701416,
      "learning_rate": 1e-05,
      "loss": 0.2365,
      "step": 18430
    },
    {
      "epoch": 0.06318833825635824,
      "grad_norm": 1.245696783065796,
      "learning_rate": 1e-05,
      "loss": 0.2304,
      "step": 18440
    },
    {
      "epoch": 0.06322260525107427,
      "grad_norm": 1.0686434507369995,
      "learning_rate": 1e-05,
      "loss": 0.2459,
      "step": 18450
    },
    {
      "epoch": 0.0632568722457903,
      "grad_norm": 1.135810375213623,
      "learning_rate": 1e-05,
      "loss": 0.237,
      "step": 18460
    },
    {
      "epoch": 0.06329113924050633,
      "grad_norm": 0.9825658202171326,
      "learning_rate": 1e-05,
      "loss": 0.2323,
      "step": 18470
    },
    {
      "epoch": 0.06332540623522236,
      "grad_norm": 0.9794711470603943,
      "learning_rate": 1e-05,
      "loss": 0.2259,
      "step": 18480
    },
    {
      "epoch": 0.06335967322993839,
      "grad_norm": 1.1655373573303223,
      "learning_rate": 1e-05,
      "loss": 0.24,
      "step": 18490
    },
    {
      "epoch": 0.06339394022465442,
      "grad_norm": 1.1974196434020996,
      "learning_rate": 1e-05,
      "loss": 0.2414,
      "step": 18500
    },
    {
      "epoch": 0.06339394022465442,
      "eval_cer": 12.885870523901923,
      "eval_loss": 0.24566347897052765,
      "eval_normalized_cer": 9.05275779376499,
      "eval_runtime": 227.2597,
      "eval_samples_per_second": 2.253,
      "eval_steps_per_second": 0.035,
      "step": 18500
    },
    {
      "epoch": 0.06342820721937045,
      "grad_norm": 1.1439276933670044,
      "learning_rate": 1e-05,
      "loss": 0.2357,
      "step": 18510
    },
    {
      "epoch": 0.06346247421408648,
      "grad_norm": 1.0722652673721313,
      "learning_rate": 1e-05,
      "loss": 0.2413,
      "step": 18520
    },
    {
      "epoch": 0.06349674120880251,
      "grad_norm": 1.071835994720459,
      "learning_rate": 1e-05,
      "loss": 0.215,
      "step": 18530
    },
    {
      "epoch": 0.06353100820351854,
      "grad_norm": 1.0905380249023438,
      "learning_rate": 1e-05,
      "loss": 0.2262,
      "step": 18540
    },
    {
      "epoch": 0.06356527519823456,
      "grad_norm": 0.9569874405860901,
      "learning_rate": 1e-05,
      "loss": 0.2111,
      "step": 18550
    },
    {
      "epoch": 0.0635995421929506,
      "grad_norm": 1.1387230157852173,
      "learning_rate": 1e-05,
      "loss": 0.2278,
      "step": 18560
    },
    {
      "epoch": 0.06363380918766662,
      "grad_norm": 1.1088684797286987,
      "learning_rate": 1e-05,
      "loss": 0.236,
      "step": 18570
    },
    {
      "epoch": 0.06366807618238265,
      "grad_norm": 1.1542891263961792,
      "learning_rate": 1e-05,
      "loss": 0.2248,
      "step": 18580
    },
    {
      "epoch": 0.06370234317709868,
      "grad_norm": 1.0670695304870605,
      "learning_rate": 1e-05,
      "loss": 0.2258,
      "step": 18590
    },
    {
      "epoch": 0.06373661017181471,
      "grad_norm": 1.1448674201965332,
      "learning_rate": 1e-05,
      "loss": 0.2445,
      "step": 18600
    },
    {
      "epoch": 0.06377087716653074,
      "grad_norm": 1.283347487449646,
      "learning_rate": 1e-05,
      "loss": 0.2318,
      "step": 18610
    },
    {
      "epoch": 0.06380514416124677,
      "grad_norm": 1.2149467468261719,
      "learning_rate": 1e-05,
      "loss": 0.2383,
      "step": 18620
    },
    {
      "epoch": 0.0638394111559628,
      "grad_norm": 1.0885416269302368,
      "learning_rate": 1e-05,
      "loss": 0.2213,
      "step": 18630
    },
    {
      "epoch": 0.06387367815067883,
      "grad_norm": 1.1276935338974,
      "learning_rate": 1e-05,
      "loss": 0.2265,
      "step": 18640
    },
    {
      "epoch": 0.06390794514539486,
      "grad_norm": 1.15458345413208,
      "learning_rate": 1e-05,
      "loss": 0.2366,
      "step": 18650
    },
    {
      "epoch": 0.0639422121401109,
      "grad_norm": 1.1190539598464966,
      "learning_rate": 1e-05,
      "loss": 0.2243,
      "step": 18660
    },
    {
      "epoch": 0.06397647913482692,
      "grad_norm": 1.1128675937652588,
      "learning_rate": 1e-05,
      "loss": 0.2351,
      "step": 18670
    },
    {
      "epoch": 0.06401074612954295,
      "grad_norm": 1.0355507135391235,
      "learning_rate": 1e-05,
      "loss": 0.2245,
      "step": 18680
    },
    {
      "epoch": 0.06404501312425898,
      "grad_norm": 1.2043906450271606,
      "learning_rate": 1e-05,
      "loss": 0.2343,
      "step": 18690
    },
    {
      "epoch": 0.064079280118975,
      "grad_norm": 1.089877724647522,
      "learning_rate": 1e-05,
      "loss": 0.2363,
      "step": 18700
    },
    {
      "epoch": 0.06411354711369104,
      "grad_norm": 1.1575262546539307,
      "learning_rate": 1e-05,
      "loss": 0.2482,
      "step": 18710
    },
    {
      "epoch": 0.06414781410840706,
      "grad_norm": 1.0474443435668945,
      "learning_rate": 1e-05,
      "loss": 0.2422,
      "step": 18720
    },
    {
      "epoch": 0.0641820811031231,
      "grad_norm": 0.9865725040435791,
      "learning_rate": 1e-05,
      "loss": 0.2245,
      "step": 18730
    },
    {
      "epoch": 0.06421634809783912,
      "grad_norm": 1.126033067703247,
      "learning_rate": 1e-05,
      "loss": 0.2336,
      "step": 18740
    },
    {
      "epoch": 0.06425061509255516,
      "grad_norm": 1.0604937076568604,
      "learning_rate": 1e-05,
      "loss": 0.2341,
      "step": 18750
    },
    {
      "epoch": 0.06428488208727118,
      "grad_norm": 1.0804321765899658,
      "learning_rate": 1e-05,
      "loss": 0.258,
      "step": 18760
    },
    {
      "epoch": 0.06431914908198721,
      "grad_norm": 1.0379873514175415,
      "learning_rate": 1e-05,
      "loss": 0.2503,
      "step": 18770
    },
    {
      "epoch": 0.06435341607670324,
      "grad_norm": 1.169416069984436,
      "learning_rate": 1e-05,
      "loss": 0.2207,
      "step": 18780
    },
    {
      "epoch": 0.06438768307141927,
      "grad_norm": 1.037367820739746,
      "learning_rate": 1e-05,
      "loss": 0.2359,
      "step": 18790
    },
    {
      "epoch": 0.0644219500661353,
      "grad_norm": 1.0716742277145386,
      "learning_rate": 1e-05,
      "loss": 0.2641,
      "step": 18800
    },
    {
      "epoch": 0.06445621706085133,
      "grad_norm": 1.097924828529358,
      "learning_rate": 1e-05,
      "loss": 0.255,
      "step": 18810
    },
    {
      "epoch": 0.06449048405556736,
      "grad_norm": 1.2333221435546875,
      "learning_rate": 1e-05,
      "loss": 0.2481,
      "step": 18820
    },
    {
      "epoch": 0.0645247510502834,
      "grad_norm": 1.1249961853027344,
      "learning_rate": 1e-05,
      "loss": 0.2337,
      "step": 18830
    },
    {
      "epoch": 0.06455901804499942,
      "grad_norm": 1.2257179021835327,
      "learning_rate": 1e-05,
      "loss": 0.2496,
      "step": 18840
    },
    {
      "epoch": 0.06459328503971545,
      "grad_norm": 1.141400694847107,
      "learning_rate": 1e-05,
      "loss": 0.2447,
      "step": 18850
    },
    {
      "epoch": 0.06462755203443148,
      "grad_norm": 0.9263083338737488,
      "learning_rate": 1e-05,
      "loss": 0.2446,
      "step": 18860
    },
    {
      "epoch": 0.0646618190291475,
      "grad_norm": 1.0573432445526123,
      "learning_rate": 1e-05,
      "loss": 0.2482,
      "step": 18870
    },
    {
      "epoch": 0.06469608602386354,
      "grad_norm": 1.2113983631134033,
      "learning_rate": 1e-05,
      "loss": 0.2459,
      "step": 18880
    },
    {
      "epoch": 0.06473035301857956,
      "grad_norm": 1.2218704223632812,
      "learning_rate": 1e-05,
      "loss": 0.2285,
      "step": 18890
    },
    {
      "epoch": 0.0647646200132956,
      "grad_norm": 1.0855586528778076,
      "learning_rate": 1e-05,
      "loss": 0.2393,
      "step": 18900
    },
    {
      "epoch": 0.06479888700801162,
      "grad_norm": 1.1320980787277222,
      "learning_rate": 1e-05,
      "loss": 0.2438,
      "step": 18910
    },
    {
      "epoch": 0.06483315400272766,
      "grad_norm": 1.2386951446533203,
      "learning_rate": 1e-05,
      "loss": 0.2554,
      "step": 18920
    },
    {
      "epoch": 0.06486742099744368,
      "grad_norm": 1.1133781671524048,
      "learning_rate": 1e-05,
      "loss": 0.2331,
      "step": 18930
    },
    {
      "epoch": 0.06490168799215972,
      "grad_norm": 1.149234652519226,
      "learning_rate": 1e-05,
      "loss": 0.2708,
      "step": 18940
    },
    {
      "epoch": 0.06493595498687574,
      "grad_norm": 1.2425532341003418,
      "learning_rate": 1e-05,
      "loss": 0.2443,
      "step": 18950
    },
    {
      "epoch": 0.06497022198159177,
      "grad_norm": 1.070039987564087,
      "learning_rate": 1e-05,
      "loss": 0.2586,
      "step": 18960
    },
    {
      "epoch": 0.0650044889763078,
      "grad_norm": 1.2911726236343384,
      "learning_rate": 1e-05,
      "loss": 0.2472,
      "step": 18970
    },
    {
      "epoch": 0.06503875597102383,
      "grad_norm": 1.2151503562927246,
      "learning_rate": 1e-05,
      "loss": 0.2594,
      "step": 18980
    },
    {
      "epoch": 0.06507302296573986,
      "grad_norm": 1.1766403913497925,
      "learning_rate": 1e-05,
      "loss": 0.2394,
      "step": 18990
    },
    {
      "epoch": 0.0651072899604559,
      "grad_norm": 1.0176496505737305,
      "learning_rate": 1e-05,
      "loss": 0.2448,
      "step": 19000
    },
    {
      "epoch": 0.0651072899604559,
      "eval_cer": 12.691832774739813,
      "eval_loss": 0.24545185267925262,
      "eval_normalized_cer": 9.082733812949641,
      "eval_runtime": 228.5313,
      "eval_samples_per_second": 2.24,
      "eval_steps_per_second": 0.035,
      "step": 19000
    },
    {
      "epoch": 0.06514155695517192,
      "grad_norm": 1.2762819528579712,
      "learning_rate": 1e-05,
      "loss": 0.2723,
      "step": 19010
    },
    {
      "epoch": 0.06517582394988795,
      "grad_norm": 1.2979320287704468,
      "learning_rate": 1e-05,
      "loss": 0.2491,
      "step": 19020
    },
    {
      "epoch": 0.06521009094460398,
      "grad_norm": 1.2459253072738647,
      "learning_rate": 1e-05,
      "loss": 0.242,
      "step": 19030
    },
    {
      "epoch": 0.06524435793932,
      "grad_norm": 1.130518913269043,
      "learning_rate": 1e-05,
      "loss": 0.2644,
      "step": 19040
    },
    {
      "epoch": 0.06527862493403604,
      "grad_norm": 1.0955489873886108,
      "learning_rate": 1e-05,
      "loss": 0.255,
      "step": 19050
    },
    {
      "epoch": 0.06531289192875206,
      "grad_norm": 1.0841457843780518,
      "learning_rate": 1e-05,
      "loss": 0.2397,
      "step": 19060
    },
    {
      "epoch": 0.0653471589234681,
      "grad_norm": 1.0940138101577759,
      "learning_rate": 1e-05,
      "loss": 0.2474,
      "step": 19070
    },
    {
      "epoch": 0.06538142591818412,
      "grad_norm": 1.1498796939849854,
      "learning_rate": 1e-05,
      "loss": 0.2495,
      "step": 19080
    },
    {
      "epoch": 0.06541569291290016,
      "grad_norm": 1.0884509086608887,
      "learning_rate": 1e-05,
      "loss": 0.2384,
      "step": 19090
    },
    {
      "epoch": 0.06544995990761618,
      "grad_norm": 1.119750738143921,
      "learning_rate": 1e-05,
      "loss": 0.2509,
      "step": 19100
    },
    {
      "epoch": 0.06548422690233222,
      "grad_norm": 0.973027765750885,
      "learning_rate": 1e-05,
      "loss": 0.2466,
      "step": 19110
    },
    {
      "epoch": 0.06551849389704824,
      "grad_norm": 1.1478537321090698,
      "learning_rate": 1e-05,
      "loss": 0.2464,
      "step": 19120
    },
    {
      "epoch": 0.06555276089176428,
      "grad_norm": 1.2103520631790161,
      "learning_rate": 1e-05,
      "loss": 0.2445,
      "step": 19130
    },
    {
      "epoch": 0.0655870278864803,
      "grad_norm": 1.000458002090454,
      "learning_rate": 1e-05,
      "loss": 0.2498,
      "step": 19140
    },
    {
      "epoch": 0.06562129488119633,
      "grad_norm": 1.1135921478271484,
      "learning_rate": 1e-05,
      "loss": 0.2369,
      "step": 19150
    },
    {
      "epoch": 0.06565556187591236,
      "grad_norm": 1.0207141637802124,
      "learning_rate": 1e-05,
      "loss": 0.243,
      "step": 19160
    },
    {
      "epoch": 0.0656898288706284,
      "grad_norm": 1.1605093479156494,
      "learning_rate": 1e-05,
      "loss": 0.2366,
      "step": 19170
    },
    {
      "epoch": 0.06572409586534442,
      "grad_norm": 1.242141604423523,
      "learning_rate": 1e-05,
      "loss": 0.2397,
      "step": 19180
    },
    {
      "epoch": 0.06575836286006044,
      "grad_norm": 0.9129644632339478,
      "learning_rate": 1e-05,
      "loss": 0.2433,
      "step": 19190
    },
    {
      "epoch": 0.06579262985477648,
      "grad_norm": 0.9504500031471252,
      "learning_rate": 1e-05,
      "loss": 0.2429,
      "step": 19200
    },
    {
      "epoch": 0.0658268968494925,
      "grad_norm": 1.2368412017822266,
      "learning_rate": 1e-05,
      "loss": 0.2406,
      "step": 19210
    },
    {
      "epoch": 0.06586116384420854,
      "grad_norm": 1.2066251039505005,
      "learning_rate": 1e-05,
      "loss": 0.2491,
      "step": 19220
    },
    {
      "epoch": 0.06589543083892456,
      "grad_norm": 1.1131211519241333,
      "learning_rate": 1e-05,
      "loss": 0.2605,
      "step": 19230
    },
    {
      "epoch": 0.0659296978336406,
      "grad_norm": 1.211876392364502,
      "learning_rate": 1e-05,
      "loss": 0.2627,
      "step": 19240
    },
    {
      "epoch": 0.06596396482835662,
      "grad_norm": 1.0012997388839722,
      "learning_rate": 1e-05,
      "loss": 0.2563,
      "step": 19250
    },
    {
      "epoch": 0.06599823182307266,
      "grad_norm": 1.2079874277114868,
      "learning_rate": 1e-05,
      "loss": 0.2423,
      "step": 19260
    },
    {
      "epoch": 0.06603249881778868,
      "grad_norm": 1.0952426195144653,
      "learning_rate": 1e-05,
      "loss": 0.2573,
      "step": 19270
    },
    {
      "epoch": 0.06606676581250472,
      "grad_norm": 0.9400359392166138,
      "learning_rate": 1e-05,
      "loss": 0.2516,
      "step": 19280
    },
    {
      "epoch": 0.06610103280722074,
      "grad_norm": 1.026219367980957,
      "learning_rate": 1e-05,
      "loss": 0.2436,
      "step": 19290
    },
    {
      "epoch": 0.06613529980193678,
      "grad_norm": 1.0502564907073975,
      "learning_rate": 1e-05,
      "loss": 0.2471,
      "step": 19300
    },
    {
      "epoch": 0.0661695667966528,
      "grad_norm": 1.2147547006607056,
      "learning_rate": 1e-05,
      "loss": 0.2358,
      "step": 19310
    },
    {
      "epoch": 0.06620383379136884,
      "grad_norm": 1.2542741298675537,
      "learning_rate": 1e-05,
      "loss": 0.2489,
      "step": 19320
    },
    {
      "epoch": 0.06623810078608486,
      "grad_norm": 1.031640887260437,
      "learning_rate": 1e-05,
      "loss": 0.2391,
      "step": 19330
    },
    {
      "epoch": 0.0662723677808009,
      "grad_norm": 1.2736186981201172,
      "learning_rate": 1e-05,
      "loss": 0.2391,
      "step": 19340
    },
    {
      "epoch": 0.06630663477551692,
      "grad_norm": 1.091873288154602,
      "learning_rate": 1e-05,
      "loss": 0.2407,
      "step": 19350
    },
    {
      "epoch": 0.06634090177023294,
      "grad_norm": 1.1229465007781982,
      "learning_rate": 1e-05,
      "loss": 0.2414,
      "step": 19360
    },
    {
      "epoch": 0.06637516876494898,
      "grad_norm": 1.2272335290908813,
      "learning_rate": 1e-05,
      "loss": 0.247,
      "step": 19370
    },
    {
      "epoch": 0.066409435759665,
      "grad_norm": 1.063962697982788,
      "learning_rate": 1e-05,
      "loss": 0.241,
      "step": 19380
    },
    {
      "epoch": 0.06644370275438104,
      "grad_norm": 1.2155866622924805,
      "learning_rate": 1e-05,
      "loss": 0.252,
      "step": 19390
    },
    {
      "epoch": 0.06647796974909706,
      "grad_norm": 0.9977822303771973,
      "learning_rate": 1e-05,
      "loss": 0.2334,
      "step": 19400
    },
    {
      "epoch": 0.0665122367438131,
      "grad_norm": 1.1466742753982544,
      "learning_rate": 1e-05,
      "loss": 0.2434,
      "step": 19410
    },
    {
      "epoch": 0.06654650373852912,
      "grad_norm": 0.9723268747329712,
      "learning_rate": 1e-05,
      "loss": 0.245,
      "step": 19420
    },
    {
      "epoch": 0.06658077073324516,
      "grad_norm": 0.9507194757461548,
      "learning_rate": 1e-05,
      "loss": 0.2268,
      "step": 19430
    },
    {
      "epoch": 0.06661503772796118,
      "grad_norm": 1.1839369535446167,
      "learning_rate": 1e-05,
      "loss": 0.24,
      "step": 19440
    },
    {
      "epoch": 0.06664930472267722,
      "grad_norm": 1.1277563571929932,
      "learning_rate": 1e-05,
      "loss": 0.2448,
      "step": 19450
    },
    {
      "epoch": 0.06668357171739324,
      "grad_norm": 0.9276403188705444,
      "learning_rate": 1e-05,
      "loss": 0.2489,
      "step": 19460
    },
    {
      "epoch": 0.06671783871210928,
      "grad_norm": 0.9769749045372009,
      "learning_rate": 1e-05,
      "loss": 0.2361,
      "step": 19470
    },
    {
      "epoch": 0.0667521057068253,
      "grad_norm": 1.1307224035263062,
      "learning_rate": 1e-05,
      "loss": 0.2413,
      "step": 19480
    },
    {
      "epoch": 0.06678637270154134,
      "grad_norm": 1.0125936269760132,
      "learning_rate": 1e-05,
      "loss": 0.2496,
      "step": 19490
    },
    {
      "epoch": 0.06682063969625736,
      "grad_norm": 1.3682901859283447,
      "learning_rate": 1e-05,
      "loss": 0.2573,
      "step": 19500
    },
    {
      "epoch": 0.06682063969625736,
      "eval_cer": 12.436055741753396,
      "eval_loss": 0.24245715141296387,
      "eval_normalized_cer": 8.81294964028777,
      "eval_runtime": 227.5796,
      "eval_samples_per_second": 2.25,
      "eval_steps_per_second": 0.035,
      "step": 19500
    },
    {
      "epoch": 0.06685490669097338,
      "grad_norm": 1.3106615543365479,
      "learning_rate": 1e-05,
      "loss": 0.2473,
      "step": 19510
    },
    {
      "epoch": 0.06688917368568942,
      "grad_norm": 1.1861129999160767,
      "learning_rate": 1e-05,
      "loss": 0.2478,
      "step": 19520
    },
    {
      "epoch": 0.06692344068040544,
      "grad_norm": 1.0658881664276123,
      "learning_rate": 1e-05,
      "loss": 0.2629,
      "step": 19530
    },
    {
      "epoch": 0.06695770767512148,
      "grad_norm": 1.3097602128982544,
      "learning_rate": 1e-05,
      "loss": 0.2332,
      "step": 19540
    },
    {
      "epoch": 0.0669919746698375,
      "grad_norm": 1.1135023832321167,
      "learning_rate": 1e-05,
      "loss": 0.2613,
      "step": 19550
    },
    {
      "epoch": 0.06702624166455354,
      "grad_norm": 1.2106831073760986,
      "learning_rate": 1e-05,
      "loss": 0.2518,
      "step": 19560
    },
    {
      "epoch": 0.06706050865926956,
      "grad_norm": 1.0989844799041748,
      "learning_rate": 1e-05,
      "loss": 0.2602,
      "step": 19570
    },
    {
      "epoch": 0.0670947756539856,
      "grad_norm": 1.0839687585830688,
      "learning_rate": 1e-05,
      "loss": 0.245,
      "step": 19580
    },
    {
      "epoch": 0.06712904264870162,
      "grad_norm": 1.249485731124878,
      "learning_rate": 1e-05,
      "loss": 0.251,
      "step": 19590
    },
    {
      "epoch": 0.06716330964341766,
      "grad_norm": 1.0540473461151123,
      "learning_rate": 1e-05,
      "loss": 0.2682,
      "step": 19600
    },
    {
      "epoch": 0.06719757663813368,
      "grad_norm": 1.1540120840072632,
      "learning_rate": 1e-05,
      "loss": 0.264,
      "step": 19610
    },
    {
      "epoch": 0.06723184363284972,
      "grad_norm": 1.2563894987106323,
      "learning_rate": 1e-05,
      "loss": 0.2627,
      "step": 19620
    },
    {
      "epoch": 0.06726611062756574,
      "grad_norm": 1.1335647106170654,
      "learning_rate": 1e-05,
      "loss": 0.2662,
      "step": 19630
    },
    {
      "epoch": 0.06730037762228178,
      "grad_norm": 1.0809390544891357,
      "learning_rate": 1e-05,
      "loss": 0.2477,
      "step": 19640
    },
    {
      "epoch": 0.0673346446169978,
      "grad_norm": 1.0609475374221802,
      "learning_rate": 1e-05,
      "loss": 0.2491,
      "step": 19650
    },
    {
      "epoch": 0.06736891161171384,
      "grad_norm": 1.2045155763626099,
      "learning_rate": 1e-05,
      "loss": 0.2586,
      "step": 19660
    },
    {
      "epoch": 0.06740317860642986,
      "grad_norm": 1.0101896524429321,
      "learning_rate": 1e-05,
      "loss": 0.2514,
      "step": 19670
    },
    {
      "epoch": 0.06743744560114588,
      "grad_norm": 1.070760726928711,
      "learning_rate": 1e-05,
      "loss": 0.2677,
      "step": 19680
    },
    {
      "epoch": 0.06747171259586192,
      "grad_norm": 1.1094374656677246,
      "learning_rate": 1e-05,
      "loss": 0.2601,
      "step": 19690
    },
    {
      "epoch": 0.06750597959057794,
      "grad_norm": 1.2062901258468628,
      "learning_rate": 1e-05,
      "loss": 0.2351,
      "step": 19700
    },
    {
      "epoch": 0.06754024658529398,
      "grad_norm": 1.2172263860702515,
      "learning_rate": 1e-05,
      "loss": 0.248,
      "step": 19710
    },
    {
      "epoch": 0.06757451358001,
      "grad_norm": 0.9713698625564575,
      "learning_rate": 1e-05,
      "loss": 0.2366,
      "step": 19720
    },
    {
      "epoch": 0.06760878057472604,
      "grad_norm": 1.046995997428894,
      "learning_rate": 1e-05,
      "loss": 0.2225,
      "step": 19730
    },
    {
      "epoch": 0.06764304756944206,
      "grad_norm": 1.002658724784851,
      "learning_rate": 1e-05,
      "loss": 0.2286,
      "step": 19740
    },
    {
      "epoch": 0.0676773145641581,
      "grad_norm": 1.072398066520691,
      "learning_rate": 1e-05,
      "loss": 0.232,
      "step": 19750
    },
    {
      "epoch": 0.06771158155887412,
      "grad_norm": 1.1962021589279175,
      "learning_rate": 1e-05,
      "loss": 0.2416,
      "step": 19760
    },
    {
      "epoch": 0.06774584855359016,
      "grad_norm": 1.19411039352417,
      "learning_rate": 1e-05,
      "loss": 0.2448,
      "step": 19770
    },
    {
      "epoch": 0.06778011554830618,
      "grad_norm": 1.0695290565490723,
      "learning_rate": 1e-05,
      "loss": 0.2273,
      "step": 19780
    },
    {
      "epoch": 0.06781438254302222,
      "grad_norm": 1.0701625347137451,
      "learning_rate": 1e-05,
      "loss": 0.2355,
      "step": 19790
    },
    {
      "epoch": 0.06784864953773824,
      "grad_norm": 1.1018661260604858,
      "learning_rate": 1e-05,
      "loss": 0.2301,
      "step": 19800
    },
    {
      "epoch": 0.06788291653245428,
      "grad_norm": 1.2072187662124634,
      "learning_rate": 1e-05,
      "loss": 0.2311,
      "step": 19810
    },
    {
      "epoch": 0.0679171835271703,
      "grad_norm": 1.2599499225616455,
      "learning_rate": 1e-05,
      "loss": 0.2327,
      "step": 19820
    },
    {
      "epoch": 0.06795145052188634,
      "grad_norm": 1.0052101612091064,
      "learning_rate": 1e-05,
      "loss": 0.2303,
      "step": 19830
    },
    {
      "epoch": 0.06798571751660236,
      "grad_norm": 1.1168869733810425,
      "learning_rate": 1e-05,
      "loss": 0.239,
      "step": 19840
    },
    {
      "epoch": 0.06801998451131838,
      "grad_norm": 1.166634440422058,
      "learning_rate": 1e-05,
      "loss": 0.2255,
      "step": 19850
    },
    {
      "epoch": 0.06805425150603442,
      "grad_norm": 1.0218929052352905,
      "learning_rate": 1e-05,
      "loss": 0.2244,
      "step": 19860
    },
    {
      "epoch": 0.06808851850075044,
      "grad_norm": 1.1705201864242554,
      "learning_rate": 1e-05,
      "loss": 0.2177,
      "step": 19870
    },
    {
      "epoch": 0.06812278549546648,
      "grad_norm": 0.9937231540679932,
      "learning_rate": 1e-05,
      "loss": 0.2384,
      "step": 19880
    },
    {
      "epoch": 0.0681570524901825,
      "grad_norm": 1.045133352279663,
      "learning_rate": 1e-05,
      "loss": 0.2208,
      "step": 19890
    },
    {
      "epoch": 0.06819131948489854,
      "grad_norm": 1.1548084020614624,
      "learning_rate": 1e-05,
      "loss": 0.2262,
      "step": 19900
    },
    {
      "epoch": 0.06822558647961456,
      "grad_norm": 1.2487162351608276,
      "learning_rate": 1e-05,
      "loss": 0.234,
      "step": 19910
    },
    {
      "epoch": 0.0682598534743306,
      "grad_norm": 1.0547343492507935,
      "learning_rate": 1e-05,
      "loss": 0.2337,
      "step": 19920
    },
    {
      "epoch": 0.06829412046904662,
      "grad_norm": 0.8701609373092651,
      "learning_rate": 1e-05,
      "loss": 0.2132,
      "step": 19930
    },
    {
      "epoch": 0.06832838746376266,
      "grad_norm": 1.0996358394622803,
      "learning_rate": 1e-05,
      "loss": 0.2201,
      "step": 19940
    },
    {
      "epoch": 0.06836265445847868,
      "grad_norm": 1.0274715423583984,
      "learning_rate": 1e-05,
      "loss": 0.221,
      "step": 19950
    },
    {
      "epoch": 0.06839692145319472,
      "grad_norm": 0.9839652180671692,
      "learning_rate": 1e-05,
      "loss": 0.2142,
      "step": 19960
    },
    {
      "epoch": 0.06843118844791074,
      "grad_norm": 1.121183156967163,
      "learning_rate": 1e-05,
      "loss": 0.2204,
      "step": 19970
    },
    {
      "epoch": 0.06846545544262678,
      "grad_norm": 1.0347726345062256,
      "learning_rate": 1e-05,
      "loss": 0.2388,
      "step": 19980
    },
    {
      "epoch": 0.0684997224373428,
      "grad_norm": 1.0449944734573364,
      "learning_rate": 1e-05,
      "loss": 0.2242,
      "step": 19990
    },
    {
      "epoch": 0.06853398943205882,
      "grad_norm": 1.0335705280303955,
      "learning_rate": 1e-05,
      "loss": 0.2323,
      "step": 20000
    },
    {
      "epoch": 0.06853398943205882,
      "eval_cer": 13.097548068442405,
      "eval_loss": 0.2434115707874298,
      "eval_normalized_cer": 9.482414068745005,
      "eval_runtime": 242.1003,
      "eval_samples_per_second": 2.115,
      "eval_steps_per_second": 0.033,
      "step": 20000
    },
    {
      "epoch": 0.06856825642677486,
      "grad_norm": 1.1393816471099854,
      "learning_rate": 1e-05,
      "loss": 0.2257,
      "step": 20010
    },
    {
      "epoch": 0.06860252342149088,
      "grad_norm": 1.2440309524536133,
      "learning_rate": 1e-05,
      "loss": 0.2045,
      "step": 20020
    },
    {
      "epoch": 0.06863679041620692,
      "grad_norm": 1.2834079265594482,
      "learning_rate": 1e-05,
      "loss": 0.2312,
      "step": 20030
    },
    {
      "epoch": 0.06867105741092294,
      "grad_norm": 1.0295487642288208,
      "learning_rate": 1e-05,
      "loss": 0.2288,
      "step": 20040
    },
    {
      "epoch": 0.06870532440563898,
      "grad_norm": 1.0935391187667847,
      "learning_rate": 1e-05,
      "loss": 0.2191,
      "step": 20050
    },
    {
      "epoch": 0.068739591400355,
      "grad_norm": 1.2545008659362793,
      "learning_rate": 1e-05,
      "loss": 0.2204,
      "step": 20060
    },
    {
      "epoch": 0.06877385839507104,
      "grad_norm": 1.0660568475723267,
      "learning_rate": 1e-05,
      "loss": 0.2214,
      "step": 20070
    },
    {
      "epoch": 0.06880812538978706,
      "grad_norm": 1.0966823101043701,
      "learning_rate": 1e-05,
      "loss": 0.2148,
      "step": 20080
    },
    {
      "epoch": 0.0688423923845031,
      "grad_norm": 1.1360703706741333,
      "learning_rate": 1e-05,
      "loss": 0.2145,
      "step": 20090
    },
    {
      "epoch": 0.06887665937921912,
      "grad_norm": 1.0804933309555054,
      "learning_rate": 1e-05,
      "loss": 0.2256,
      "step": 20100
    },
    {
      "epoch": 0.06891092637393516,
      "grad_norm": 1.2052994966506958,
      "learning_rate": 1e-05,
      "loss": 0.2188,
      "step": 20110
    },
    {
      "epoch": 0.06894519336865118,
      "grad_norm": 1.2262095212936401,
      "learning_rate": 1e-05,
      "loss": 0.2283,
      "step": 20120
    },
    {
      "epoch": 0.06897946036336722,
      "grad_norm": 1.2067344188690186,
      "learning_rate": 1e-05,
      "loss": 0.2107,
      "step": 20130
    },
    {
      "epoch": 0.06901372735808324,
      "grad_norm": 1.1655231714248657,
      "learning_rate": 1e-05,
      "loss": 0.2097,
      "step": 20140
    },
    {
      "epoch": 0.06904799435279928,
      "grad_norm": 1.1139166355133057,
      "learning_rate": 1e-05,
      "loss": 0.2079,
      "step": 20150
    },
    {
      "epoch": 0.0690822613475153,
      "grad_norm": 1.0625183582305908,
      "learning_rate": 1e-05,
      "loss": 0.2097,
      "step": 20160
    },
    {
      "epoch": 0.06911652834223132,
      "grad_norm": 1.0871198177337646,
      "learning_rate": 1e-05,
      "loss": 0.2041,
      "step": 20170
    },
    {
      "epoch": 0.06915079533694736,
      "grad_norm": 1.1072641611099243,
      "learning_rate": 1e-05,
      "loss": 0.2178,
      "step": 20180
    },
    {
      "epoch": 0.06918506233166338,
      "grad_norm": 1.1643656492233276,
      "learning_rate": 1e-05,
      "loss": 0.2264,
      "step": 20190
    },
    {
      "epoch": 0.06921932932637942,
      "grad_norm": 1.0108036994934082,
      "learning_rate": 1e-05,
      "loss": 0.2069,
      "step": 20200
    },
    {
      "epoch": 0.06925359632109544,
      "grad_norm": 1.1025501489639282,
      "learning_rate": 1e-05,
      "loss": 0.2146,
      "step": 20210
    },
    {
      "epoch": 0.06928786331581148,
      "grad_norm": 1.2304129600524902,
      "learning_rate": 1e-05,
      "loss": 0.2333,
      "step": 20220
    },
    {
      "epoch": 0.0693221303105275,
      "grad_norm": 1.2792093753814697,
      "learning_rate": 1e-05,
      "loss": 0.2453,
      "step": 20230
    },
    {
      "epoch": 0.06935639730524354,
      "grad_norm": 0.9987597465515137,
      "learning_rate": 1e-05,
      "loss": 0.2484,
      "step": 20240
    },
    {
      "epoch": 0.06939066429995956,
      "grad_norm": 1.2693513631820679,
      "learning_rate": 1e-05,
      "loss": 0.2439,
      "step": 20250
    },
    {
      "epoch": 0.0694249312946756,
      "grad_norm": 1.2504794597625732,
      "learning_rate": 1e-05,
      "loss": 0.2478,
      "step": 20260
    },
    {
      "epoch": 0.06945919828939162,
      "grad_norm": 1.1483858823776245,
      "learning_rate": 1e-05,
      "loss": 0.2615,
      "step": 20270
    },
    {
      "epoch": 0.06949346528410766,
      "grad_norm": 1.1267510652542114,
      "learning_rate": 1e-05,
      "loss": 0.2492,
      "step": 20280
    },
    {
      "epoch": 0.06952773227882368,
      "grad_norm": 1.0131007432937622,
      "learning_rate": 1e-05,
      "loss": 0.2426,
      "step": 20290
    },
    {
      "epoch": 0.06956199927353972,
      "grad_norm": 1.1513702869415283,
      "learning_rate": 1e-05,
      "loss": 0.2473,
      "step": 20300
    },
    {
      "epoch": 0.06959626626825574,
      "grad_norm": 1.0932912826538086,
      "learning_rate": 1e-05,
      "loss": 0.2444,
      "step": 20310
    },
    {
      "epoch": 0.06963053326297178,
      "grad_norm": 1.1496869325637817,
      "learning_rate": 1e-05,
      "loss": 0.2543,
      "step": 20320
    },
    {
      "epoch": 0.0696648002576878,
      "grad_norm": 1.0396265983581543,
      "learning_rate": 1e-05,
      "loss": 0.2528,
      "step": 20330
    },
    {
      "epoch": 0.06969906725240382,
      "grad_norm": 1.2144852876663208,
      "learning_rate": 1e-05,
      "loss": 0.263,
      "step": 20340
    },
    {
      "epoch": 0.06973333424711986,
      "grad_norm": 1.0001530647277832,
      "learning_rate": 1e-05,
      "loss": 0.2411,
      "step": 20350
    },
    {
      "epoch": 0.06976760124183588,
      "grad_norm": 1.1693116426467896,
      "learning_rate": 1e-05,
      "loss": 0.2643,
      "step": 20360
    },
    {
      "epoch": 0.06980186823655192,
      "grad_norm": 1.1341391801834106,
      "learning_rate": 1e-05,
      "loss": 0.2432,
      "step": 20370
    },
    {
      "epoch": 0.06983613523126794,
      "grad_norm": 1.0701422691345215,
      "learning_rate": 1e-05,
      "loss": 0.2526,
      "step": 20380
    },
    {
      "epoch": 0.06987040222598398,
      "grad_norm": 1.1302570104599,
      "learning_rate": 1e-05,
      "loss": 0.2557,
      "step": 20390
    },
    {
      "epoch": 0.0699046692207,
      "grad_norm": 0.940927267074585,
      "learning_rate": 1e-05,
      "loss": 0.2379,
      "step": 20400
    },
    {
      "epoch": 0.06993893621541604,
      "grad_norm": 1.210973858833313,
      "learning_rate": 1e-05,
      "loss": 0.2565,
      "step": 20410
    },
    {
      "epoch": 0.06997320321013206,
      "grad_norm": 1.1883857250213623,
      "learning_rate": 1e-05,
      "loss": 0.2688,
      "step": 20420
    },
    {
      "epoch": 0.0700074702048481,
      "grad_norm": 1.0678610801696777,
      "learning_rate": 1e-05,
      "loss": 0.2414,
      "step": 20430
    },
    {
      "epoch": 0.07004173719956412,
      "grad_norm": 1.0714175701141357,
      "learning_rate": 1e-05,
      "loss": 0.2607,
      "step": 20440
    },
    {
      "epoch": 0.07007600419428016,
      "grad_norm": 1.171006441116333,
      "learning_rate": 1e-05,
      "loss": 0.2542,
      "step": 20450
    },
    {
      "epoch": 0.07011027118899618,
      "grad_norm": 1.1522200107574463,
      "learning_rate": 1e-05,
      "loss": 0.2412,
      "step": 20460
    },
    {
      "epoch": 0.07014453818371222,
      "grad_norm": 1.2820340394973755,
      "learning_rate": 1e-05,
      "loss": 0.2648,
      "step": 20470
    },
    {
      "epoch": 0.07017880517842824,
      "grad_norm": 1.191225290298462,
      "learning_rate": 1e-05,
      "loss": 0.2448,
      "step": 20480
    },
    {
      "epoch": 0.07021307217314426,
      "grad_norm": 1.1028372049331665,
      "learning_rate": 1e-05,
      "loss": 0.2448,
      "step": 20490
    },
    {
      "epoch": 0.0702473391678603,
      "grad_norm": 1.0639230012893677,
      "learning_rate": 1e-05,
      "loss": 0.2435,
      "step": 20500
    },
    {
      "epoch": 0.0702473391678603,
      "eval_cer": 12.136179220321045,
      "eval_loss": 0.2431151568889618,
      "eval_normalized_cer": 8.593125499600319,
      "eval_runtime": 226.8501,
      "eval_samples_per_second": 2.257,
      "eval_steps_per_second": 0.035,
      "step": 20500
    },
    {
      "epoch": 0.07028160616257632,
      "grad_norm": 1.1122690439224243,
      "learning_rate": 1e-05,
      "loss": 0.2462,
      "step": 20510
    },
    {
      "epoch": 0.07031587315729236,
      "grad_norm": 1.1983146667480469,
      "learning_rate": 1e-05,
      "loss": 0.2684,
      "step": 20520
    },
    {
      "epoch": 0.07035014015200838,
      "grad_norm": 1.0748060941696167,
      "learning_rate": 1e-05,
      "loss": 0.2446,
      "step": 20530
    },
    {
      "epoch": 0.07038440714672442,
      "grad_norm": 1.0608043670654297,
      "learning_rate": 1e-05,
      "loss": 0.2461,
      "step": 20540
    },
    {
      "epoch": 0.07041867414144044,
      "grad_norm": 1.0185233354568481,
      "learning_rate": 1e-05,
      "loss": 0.2558,
      "step": 20550
    },
    {
      "epoch": 0.07045294113615648,
      "grad_norm": 1.0136014223098755,
      "learning_rate": 1e-05,
      "loss": 0.237,
      "step": 20560
    },
    {
      "epoch": 0.0704872081308725,
      "grad_norm": 1.0096406936645508,
      "learning_rate": 1e-05,
      "loss": 0.257,
      "step": 20570
    },
    {
      "epoch": 0.07052147512558854,
      "grad_norm": 1.0398333072662354,
      "learning_rate": 1e-05,
      "loss": 0.2586,
      "step": 20580
    },
    {
      "epoch": 0.07055574212030456,
      "grad_norm": 1.0201914310455322,
      "learning_rate": 1e-05,
      "loss": 0.2369,
      "step": 20590
    },
    {
      "epoch": 0.0705900091150206,
      "grad_norm": 1.2178388833999634,
      "learning_rate": 1e-05,
      "loss": 0.2593,
      "step": 20600
    },
    {
      "epoch": 0.07062427610973662,
      "grad_norm": 1.120365858078003,
      "learning_rate": 1e-05,
      "loss": 0.2616,
      "step": 20610
    },
    {
      "epoch": 0.07065854310445266,
      "grad_norm": 1.3457388877868652,
      "learning_rate": 1e-05,
      "loss": 0.2561,
      "step": 20620
    },
    {
      "epoch": 0.07069281009916868,
      "grad_norm": 1.004683256149292,
      "learning_rate": 1e-05,
      "loss": 0.2421,
      "step": 20630
    },
    {
      "epoch": 0.07072707709388472,
      "grad_norm": 1.122775912284851,
      "learning_rate": 1e-05,
      "loss": 0.236,
      "step": 20640
    },
    {
      "epoch": 0.07076134408860074,
      "grad_norm": 1.1064289808273315,
      "learning_rate": 1e-05,
      "loss": 0.2383,
      "step": 20650
    },
    {
      "epoch": 0.07079561108331676,
      "grad_norm": 1.0714755058288574,
      "learning_rate": 1e-05,
      "loss": 0.2376,
      "step": 20660
    },
    {
      "epoch": 0.0708298780780328,
      "grad_norm": 1.0543241500854492,
      "learning_rate": 1e-05,
      "loss": 0.2575,
      "step": 20670
    },
    {
      "epoch": 0.07086414507274882,
      "grad_norm": 1.0677913427352905,
      "learning_rate": 1e-05,
      "loss": 0.2505,
      "step": 20680
    },
    {
      "epoch": 0.07089841206746486,
      "grad_norm": 0.9517323970794678,
      "learning_rate": 1e-05,
      "loss": 0.2273,
      "step": 20690
    },
    {
      "epoch": 0.07093267906218088,
      "grad_norm": 1.2108800411224365,
      "learning_rate": 1e-05,
      "loss": 0.2436,
      "step": 20700
    },
    {
      "epoch": 0.07096694605689692,
      "grad_norm": 1.1150835752487183,
      "learning_rate": 1e-05,
      "loss": 0.2434,
      "step": 20710
    },
    {
      "epoch": 0.07100121305161294,
      "grad_norm": 1.1059845685958862,
      "learning_rate": 1e-05,
      "loss": 0.2603,
      "step": 20720
    },
    {
      "epoch": 0.07103548004632898,
      "grad_norm": 1.2442775964736938,
      "learning_rate": 1e-05,
      "loss": 0.2608,
      "step": 20730
    },
    {
      "epoch": 0.071069747041045,
      "grad_norm": 1.2519625425338745,
      "learning_rate": 1e-05,
      "loss": 0.2354,
      "step": 20740
    },
    {
      "epoch": 0.07110401403576104,
      "grad_norm": 1.049765706062317,
      "learning_rate": 1e-05,
      "loss": 0.2397,
      "step": 20750
    },
    {
      "epoch": 0.07113828103047706,
      "grad_norm": 1.1388698816299438,
      "learning_rate": 1e-05,
      "loss": 0.235,
      "step": 20760
    },
    {
      "epoch": 0.0711725480251931,
      "grad_norm": 1.265163779258728,
      "learning_rate": 1e-05,
      "loss": 0.2259,
      "step": 20770
    },
    {
      "epoch": 0.07120681501990912,
      "grad_norm": 1.1405965089797974,
      "learning_rate": 1e-05,
      "loss": 0.2256,
      "step": 20780
    },
    {
      "epoch": 0.07124108201462516,
      "grad_norm": 1.0761326551437378,
      "learning_rate": 1e-05,
      "loss": 0.2207,
      "step": 20790
    },
    {
      "epoch": 0.07127534900934118,
      "grad_norm": 0.958968460559845,
      "learning_rate": 1e-05,
      "loss": 0.2206,
      "step": 20800
    },
    {
      "epoch": 0.0713096160040572,
      "grad_norm": 1.0726120471954346,
      "learning_rate": 1e-05,
      "loss": 0.2268,
      "step": 20810
    },
    {
      "epoch": 0.07134388299877324,
      "grad_norm": 1.0687642097473145,
      "learning_rate": 1e-05,
      "loss": 0.2205,
      "step": 20820
    },
    {
      "epoch": 0.07137814999348926,
      "grad_norm": 1.1116247177124023,
      "learning_rate": 1e-05,
      "loss": 0.2343,
      "step": 20830
    },
    {
      "epoch": 0.0714124169882053,
      "grad_norm": 1.1039427518844604,
      "learning_rate": 1e-05,
      "loss": 0.2401,
      "step": 20840
    },
    {
      "epoch": 0.07144668398292132,
      "grad_norm": 1.1759549379348755,
      "learning_rate": 1e-05,
      "loss": 0.2346,
      "step": 20850
    },
    {
      "epoch": 0.07148095097763736,
      "grad_norm": 1.1680679321289062,
      "learning_rate": 1e-05,
      "loss": 0.2225,
      "step": 20860
    },
    {
      "epoch": 0.07151521797235338,
      "grad_norm": 1.0508878231048584,
      "learning_rate": 1e-05,
      "loss": 0.2326,
      "step": 20870
    },
    {
      "epoch": 0.07154948496706942,
      "grad_norm": 1.0902260541915894,
      "learning_rate": 1e-05,
      "loss": 0.2208,
      "step": 20880
    },
    {
      "epoch": 0.07158375196178544,
      "grad_norm": 1.2788329124450684,
      "learning_rate": 1e-05,
      "loss": 0.2138,
      "step": 20890
    },
    {
      "epoch": 0.07161801895650148,
      "grad_norm": 1.062601089477539,
      "learning_rate": 1e-05,
      "loss": 0.223,
      "step": 20900
    },
    {
      "epoch": 0.0716522859512175,
      "grad_norm": 1.0100822448730469,
      "learning_rate": 1e-05,
      "loss": 0.2252,
      "step": 20910
    },
    {
      "epoch": 0.07168655294593354,
      "grad_norm": 1.2648316621780396,
      "learning_rate": 1e-05,
      "loss": 0.2331,
      "step": 20920
    },
    {
      "epoch": 0.07172081994064956,
      "grad_norm": 1.2350354194641113,
      "learning_rate": 1e-05,
      "loss": 0.2405,
      "step": 20930
    },
    {
      "epoch": 0.0717550869353656,
      "grad_norm": 1.0435272455215454,
      "learning_rate": 1e-05,
      "loss": 0.2181,
      "step": 20940
    },
    {
      "epoch": 0.07178935393008162,
      "grad_norm": 0.9889156818389893,
      "learning_rate": 1e-05,
      "loss": 0.229,
      "step": 20950
    },
    {
      "epoch": 0.07182362092479766,
      "grad_norm": 1.131627082824707,
      "learning_rate": 1e-05,
      "loss": 0.2493,
      "step": 20960
    },
    {
      "epoch": 0.07185788791951368,
      "grad_norm": 1.1382976770401,
      "learning_rate": 1e-05,
      "loss": 0.2149,
      "step": 20970
    },
    {
      "epoch": 0.0718921549142297,
      "grad_norm": 1.1330755949020386,
      "learning_rate": 1e-05,
      "loss": 0.2281,
      "step": 20980
    },
    {
      "epoch": 0.07192642190894574,
      "grad_norm": 1.165480375289917,
      "learning_rate": 1e-05,
      "loss": 0.2376,
      "step": 20990
    },
    {
      "epoch": 0.07196068890366177,
      "grad_norm": 1.2060060501098633,
      "learning_rate": 1e-05,
      "loss": 0.2231,
      "step": 21000
    },
    {
      "epoch": 0.07196068890366177,
      "eval_cer": 12.956429705415417,
      "eval_loss": 0.2448483556509018,
      "eval_normalized_cer": 9.502398081534773,
      "eval_runtime": 227.7694,
      "eval_samples_per_second": 2.248,
      "eval_steps_per_second": 0.035,
      "step": 21000
    },
    {
      "epoch": 0.0719949558983778,
      "grad_norm": 1.4561773538589478,
      "learning_rate": 1e-05,
      "loss": 0.2405,
      "step": 21010
    },
    {
      "epoch": 0.07202922289309382,
      "grad_norm": 1.0566471815109253,
      "learning_rate": 1e-05,
      "loss": 0.2223,
      "step": 21020
    },
    {
      "epoch": 0.07206348988780986,
      "grad_norm": 1.1329096555709839,
      "learning_rate": 1e-05,
      "loss": 0.245,
      "step": 21030
    },
    {
      "epoch": 0.07209775688252588,
      "grad_norm": 1.1300888061523438,
      "learning_rate": 1e-05,
      "loss": 0.2312,
      "step": 21040
    },
    {
      "epoch": 0.07213202387724192,
      "grad_norm": 0.9974931478500366,
      "learning_rate": 1e-05,
      "loss": 0.2148,
      "step": 21050
    },
    {
      "epoch": 0.07216629087195794,
      "grad_norm": 1.1847643852233887,
      "learning_rate": 1e-05,
      "loss": 0.2357,
      "step": 21060
    },
    {
      "epoch": 0.07220055786667398,
      "grad_norm": 1.0534117221832275,
      "learning_rate": 1e-05,
      "loss": 0.2469,
      "step": 21070
    },
    {
      "epoch": 0.07223482486139,
      "grad_norm": 1.1248983144760132,
      "learning_rate": 1e-05,
      "loss": 0.2222,
      "step": 21080
    },
    {
      "epoch": 0.07226909185610604,
      "grad_norm": 0.9976629614830017,
      "learning_rate": 1e-05,
      "loss": 0.2374,
      "step": 21090
    },
    {
      "epoch": 0.07230335885082206,
      "grad_norm": 1.0231010913848877,
      "learning_rate": 1e-05,
      "loss": 0.2341,
      "step": 21100
    },
    {
      "epoch": 0.0723376258455381,
      "grad_norm": 1.1117150783538818,
      "learning_rate": 1e-05,
      "loss": 0.2159,
      "step": 21110
    },
    {
      "epoch": 0.07237189284025412,
      "grad_norm": 0.9355179071426392,
      "learning_rate": 1e-05,
      "loss": 0.2196,
      "step": 21120
    },
    {
      "epoch": 0.07240615983497016,
      "grad_norm": 1.1800999641418457,
      "learning_rate": 1e-05,
      "loss": 0.224,
      "step": 21130
    },
    {
      "epoch": 0.07244042682968618,
      "grad_norm": 1.061677098274231,
      "learning_rate": 1e-05,
      "loss": 0.2348,
      "step": 21140
    },
    {
      "epoch": 0.0724746938244022,
      "grad_norm": 1.0895639657974243,
      "learning_rate": 1e-05,
      "loss": 0.2364,
      "step": 21150
    },
    {
      "epoch": 0.07250896081911824,
      "grad_norm": 1.0691415071487427,
      "learning_rate": 1e-05,
      "loss": 0.2219,
      "step": 21160
    },
    {
      "epoch": 0.07254322781383427,
      "grad_norm": 1.1175389289855957,
      "learning_rate": 1e-05,
      "loss": 0.2223,
      "step": 21170
    },
    {
      "epoch": 0.0725774948085503,
      "grad_norm": 1.064145565032959,
      "learning_rate": 1e-05,
      "loss": 0.226,
      "step": 21180
    },
    {
      "epoch": 0.07261176180326633,
      "grad_norm": 1.0937738418579102,
      "learning_rate": 1e-05,
      "loss": 0.2292,
      "step": 21190
    },
    {
      "epoch": 0.07264602879798236,
      "grad_norm": 1.106714129447937,
      "learning_rate": 1e-05,
      "loss": 0.2164,
      "step": 21200
    },
    {
      "epoch": 0.07268029579269839,
      "grad_norm": 0.9919803738594055,
      "learning_rate": 1e-05,
      "loss": 0.2361,
      "step": 21210
    },
    {
      "epoch": 0.07271456278741442,
      "grad_norm": 1.055652141571045,
      "learning_rate": 1e-05,
      "loss": 0.2194,
      "step": 21220
    },
    {
      "epoch": 0.07274882978213044,
      "grad_norm": 1.098995327949524,
      "learning_rate": 1e-05,
      "loss": 0.2339,
      "step": 21230
    },
    {
      "epoch": 0.07278309677684648,
      "grad_norm": 1.0142837762832642,
      "learning_rate": 1e-05,
      "loss": 0.2222,
      "step": 21240
    },
    {
      "epoch": 0.0728173637715625,
      "grad_norm": 1.1160062551498413,
      "learning_rate": 1e-05,
      "loss": 0.2358,
      "step": 21250
    },
    {
      "epoch": 0.07285163076627854,
      "grad_norm": 1.0952587127685547,
      "learning_rate": 1e-05,
      "loss": 0.2329,
      "step": 21260
    },
    {
      "epoch": 0.07288589776099456,
      "grad_norm": Infinity,
      "learning_rate": 1e-05,
      "loss": 0.2513,
      "step": 21270
    },
    {
      "epoch": 0.0729201647557106,
      "grad_norm": 1.2418158054351807,
      "learning_rate": 1e-05,
      "loss": 0.2421,
      "step": 21280
    },
    {
      "epoch": 0.07295443175042662,
      "grad_norm": 1.184396505355835,
      "learning_rate": 1e-05,
      "loss": 0.2604,
      "step": 21290
    },
    {
      "epoch": 0.07298869874514265,
      "grad_norm": 1.0499835014343262,
      "learning_rate": 1e-05,
      "loss": 0.2575,
      "step": 21300
    },
    {
      "epoch": 0.07302296573985868,
      "grad_norm": 1.371702790260315,
      "learning_rate": 1e-05,
      "loss": 0.2452,
      "step": 21310
    },
    {
      "epoch": 0.0730572327345747,
      "grad_norm": 1.0391615629196167,
      "learning_rate": 1e-05,
      "loss": 0.2492,
      "step": 21320
    },
    {
      "epoch": 0.07309149972929074,
      "grad_norm": 1.1651289463043213,
      "learning_rate": 1e-05,
      "loss": 0.2413,
      "step": 21330
    },
    {
      "epoch": 0.07312576672400677,
      "grad_norm": 1.1316466331481934,
      "learning_rate": 1e-05,
      "loss": 0.2438,
      "step": 21340
    },
    {
      "epoch": 0.0731600337187228,
      "grad_norm": 1.0570766925811768,
      "learning_rate": 1e-05,
      "loss": 0.2494,
      "step": 21350
    },
    {
      "epoch": 0.07319430071343883,
      "grad_norm": 1.0197632312774658,
      "learning_rate": 1e-05,
      "loss": 0.2429,
      "step": 21360
    },
    {
      "epoch": 0.07322856770815486,
      "grad_norm": 1.2528513669967651,
      "learning_rate": 1e-05,
      "loss": 0.2415,
      "step": 21370
    },
    {
      "epoch": 0.07326283470287089,
      "grad_norm": 1.3180269002914429,
      "learning_rate": 1e-05,
      "loss": 0.2618,
      "step": 21380
    },
    {
      "epoch": 0.07329710169758692,
      "grad_norm": 1.0602507591247559,
      "learning_rate": 1e-05,
      "loss": 0.2586,
      "step": 21390
    },
    {
      "epoch": 0.07333136869230295,
      "grad_norm": 0.9718620181083679,
      "learning_rate": 1e-05,
      "loss": 0.2519,
      "step": 21400
    },
    {
      "epoch": 0.07336563568701898,
      "grad_norm": 1.1401866674423218,
      "learning_rate": 1e-05,
      "loss": 0.2645,
      "step": 21410
    },
    {
      "epoch": 0.073399902681735,
      "grad_norm": 1.2167593240737915,
      "learning_rate": 1e-05,
      "loss": 0.2361,
      "step": 21420
    },
    {
      "epoch": 0.07343416967645104,
      "grad_norm": 1.2634971141815186,
      "learning_rate": 1e-05,
      "loss": 0.2551,
      "step": 21430
    },
    {
      "epoch": 0.07346843667116706,
      "grad_norm": 1.0504337549209595,
      "learning_rate": 1e-05,
      "loss": 0.2408,
      "step": 21440
    },
    {
      "epoch": 0.0735027036658831,
      "grad_norm": 1.1132304668426514,
      "learning_rate": 1e-05,
      "loss": 0.2534,
      "step": 21450
    },
    {
      "epoch": 0.07353697066059912,
      "grad_norm": 1.2953345775604248,
      "learning_rate": 1e-05,
      "loss": 0.241,
      "step": 21460
    },
    {
      "epoch": 0.07357123765531515,
      "grad_norm": 0.9960566759109497,
      "learning_rate": 1e-05,
      "loss": 0.2418,
      "step": 21470
    },
    {
      "epoch": 0.07360550465003118,
      "grad_norm": 1.1548551321029663,
      "learning_rate": 1e-05,
      "loss": 0.2709,
      "step": 21480
    },
    {
      "epoch": 0.0736397716447472,
      "grad_norm": 1.068936824798584,
      "learning_rate": 1e-05,
      "loss": 0.242,
      "step": 21490
    },
    {
      "epoch": 0.07367403863946324,
      "grad_norm": 1.0421861410140991,
      "learning_rate": 1e-05,
      "loss": 0.249,
      "step": 21500
    },
    {
      "epoch": 0.07367403863946324,
      "eval_cer": 12.524254718645263,
      "eval_loss": 0.2399865686893463,
      "eval_normalized_cer": 8.962829736211031,
      "eval_runtime": 227.9979,
      "eval_samples_per_second": 2.246,
      "eval_steps_per_second": 0.035,
      "step": 21500
    },
    {
      "epoch": 0.07370830563417927,
      "grad_norm": 1.3439817428588867,
      "learning_rate": 1e-05,
      "loss": 0.2592,
      "step": 21510
    },
    {
      "epoch": 0.0737425726288953,
      "grad_norm": 1.2546987533569336,
      "learning_rate": 1e-05,
      "loss": 0.2451,
      "step": 21520
    },
    {
      "epoch": 0.07377683962361133,
      "grad_norm": 0.9786490797996521,
      "learning_rate": 1e-05,
      "loss": 0.2273,
      "step": 21530
    },
    {
      "epoch": 0.07381110661832736,
      "grad_norm": 1.1565337181091309,
      "learning_rate": 1e-05,
      "loss": 0.2455,
      "step": 21540
    },
    {
      "epoch": 0.07384537361304339,
      "grad_norm": 1.2526615858078003,
      "learning_rate": 1e-05,
      "loss": 0.2445,
      "step": 21550
    },
    {
      "epoch": 0.07387964060775942,
      "grad_norm": 1.2073777914047241,
      "learning_rate": 1e-05,
      "loss": 0.2536,
      "step": 21560
    },
    {
      "epoch": 0.07391390760247545,
      "grad_norm": 1.0333545207977295,
      "learning_rate": 1e-05,
      "loss": 0.2433,
      "step": 21570
    },
    {
      "epoch": 0.07394817459719148,
      "grad_norm": 0.9934136867523193,
      "learning_rate": 1e-05,
      "loss": 0.2608,
      "step": 21580
    },
    {
      "epoch": 0.0739824415919075,
      "grad_norm": 1.152581810951233,
      "learning_rate": 1e-05,
      "loss": 0.2449,
      "step": 21590
    },
    {
      "epoch": 0.07401670858662354,
      "grad_norm": 1.0302467346191406,
      "learning_rate": 1e-05,
      "loss": 0.2542,
      "step": 21600
    },
    {
      "epoch": 0.07405097558133956,
      "grad_norm": 1.1744670867919922,
      "learning_rate": 1e-05,
      "loss": 0.2503,
      "step": 21610
    },
    {
      "epoch": 0.0740852425760556,
      "grad_norm": 1.1475284099578857,
      "learning_rate": 1e-05,
      "loss": 0.2571,
      "step": 21620
    },
    {
      "epoch": 0.07411950957077162,
      "grad_norm": 1.157768964767456,
      "learning_rate": 1e-05,
      "loss": 0.2336,
      "step": 21630
    },
    {
      "epoch": 0.07415377656548765,
      "grad_norm": 1.1291224956512451,
      "learning_rate": 1e-05,
      "loss": 0.2516,
      "step": 21640
    },
    {
      "epoch": 0.07418804356020368,
      "grad_norm": 1.1889537572860718,
      "learning_rate": 1e-05,
      "loss": 0.2428,
      "step": 21650
    },
    {
      "epoch": 0.07422231055491971,
      "grad_norm": 1.0830391645431519,
      "learning_rate": 1e-05,
      "loss": 0.2469,
      "step": 21660
    },
    {
      "epoch": 0.07425657754963574,
      "grad_norm": 1.0361846685409546,
      "learning_rate": 1e-05,
      "loss": 0.2397,
      "step": 21670
    },
    {
      "epoch": 0.07429084454435177,
      "grad_norm": 1.008002519607544,
      "learning_rate": 1e-05,
      "loss": 0.2401,
      "step": 21680
    },
    {
      "epoch": 0.0743251115390678,
      "grad_norm": 1.2176759243011475,
      "learning_rate": 1e-05,
      "loss": 0.2562,
      "step": 21690
    },
    {
      "epoch": 0.07435937853378383,
      "grad_norm": 1.2259795665740967,
      "learning_rate": 1e-05,
      "loss": 0.2623,
      "step": 21700
    },
    {
      "epoch": 0.07439364552849986,
      "grad_norm": 1.1422061920166016,
      "learning_rate": 1e-05,
      "loss": 0.2732,
      "step": 21710
    },
    {
      "epoch": 0.07442791252321589,
      "grad_norm": 1.1725839376449585,
      "learning_rate": 1e-05,
      "loss": 0.2431,
      "step": 21720
    },
    {
      "epoch": 0.07446217951793192,
      "grad_norm": 1.026131510734558,
      "learning_rate": 1e-05,
      "loss": 0.244,
      "step": 21730
    },
    {
      "epoch": 0.07449644651264795,
      "grad_norm": 1.1673903465270996,
      "learning_rate": 1e-05,
      "loss": 0.263,
      "step": 21740
    },
    {
      "epoch": 0.07453071350736398,
      "grad_norm": 1.0802525281906128,
      "learning_rate": 1e-05,
      "loss": 0.2496,
      "step": 21750
    },
    {
      "epoch": 0.07456498050208,
      "grad_norm": 1.0692561864852905,
      "learning_rate": 1e-05,
      "loss": 0.2452,
      "step": 21760
    },
    {
      "epoch": 0.07459924749679604,
      "grad_norm": 0.9135819673538208,
      "learning_rate": 1e-05,
      "loss": 0.2359,
      "step": 21770
    },
    {
      "epoch": 0.07463351449151207,
      "grad_norm": 0.9052550792694092,
      "learning_rate": 1e-05,
      "loss": 0.2331,
      "step": 21780
    },
    {
      "epoch": 0.07466778148622809,
      "grad_norm": 1.1125128269195557,
      "learning_rate": 1e-05,
      "loss": 0.2355,
      "step": 21790
    },
    {
      "epoch": 0.07470204848094412,
      "grad_norm": 1.2584770917892456,
      "learning_rate": 1e-05,
      "loss": 0.2487,
      "step": 21800
    },
    {
      "epoch": 0.07473631547566015,
      "grad_norm": 1.0602110624313354,
      "learning_rate": 1e-05,
      "loss": 0.2262,
      "step": 21810
    },
    {
      "epoch": 0.07477058247037618,
      "grad_norm": 1.0684539079666138,
      "learning_rate": 1e-05,
      "loss": 0.2142,
      "step": 21820
    },
    {
      "epoch": 0.07480484946509221,
      "grad_norm": 1.0720919370651245,
      "learning_rate": 1e-05,
      "loss": 0.2363,
      "step": 21830
    },
    {
      "epoch": 0.07483911645980824,
      "grad_norm": 0.9669879078865051,
      "learning_rate": 1e-05,
      "loss": 0.2338,
      "step": 21840
    },
    {
      "epoch": 0.07487338345452427,
      "grad_norm": 1.19837486743927,
      "learning_rate": 1e-05,
      "loss": 0.2487,
      "step": 21850
    },
    {
      "epoch": 0.0749076504492403,
      "grad_norm": 0.9803608059883118,
      "learning_rate": 1e-05,
      "loss": 0.2447,
      "step": 21860
    },
    {
      "epoch": 0.07494191744395633,
      "grad_norm": 1.0210012197494507,
      "learning_rate": 1e-05,
      "loss": 0.2314,
      "step": 21870
    },
    {
      "epoch": 0.07497618443867236,
      "grad_norm": 1.0393648147583008,
      "learning_rate": 1e-05,
      "loss": 0.2362,
      "step": 21880
    },
    {
      "epoch": 0.07501045143338839,
      "grad_norm": 1.1650710105895996,
      "learning_rate": 1e-05,
      "loss": 0.2245,
      "step": 21890
    },
    {
      "epoch": 0.07504471842810442,
      "grad_norm": 1.2479090690612793,
      "learning_rate": 1e-05,
      "loss": 0.2394,
      "step": 21900
    },
    {
      "epoch": 0.07507898542282045,
      "grad_norm": 1.105844497680664,
      "learning_rate": 1e-05,
      "loss": 0.2489,
      "step": 21910
    },
    {
      "epoch": 0.07511325241753648,
      "grad_norm": 1.1276116371154785,
      "learning_rate": 1e-05,
      "loss": 0.2398,
      "step": 21920
    },
    {
      "epoch": 0.0751475194122525,
      "grad_norm": 0.9943282604217529,
      "learning_rate": 1e-05,
      "loss": 0.2308,
      "step": 21930
    },
    {
      "epoch": 0.07518178640696854,
      "grad_norm": 0.9971685409545898,
      "learning_rate": 1e-05,
      "loss": 0.2191,
      "step": 21940
    },
    {
      "epoch": 0.07521605340168457,
      "grad_norm": 1.0626705884933472,
      "learning_rate": 1e-05,
      "loss": 0.2355,
      "step": 21950
    },
    {
      "epoch": 0.07525032039640059,
      "grad_norm": 1.1923019886016846,
      "learning_rate": 1e-05,
      "loss": 0.2165,
      "step": 21960
    },
    {
      "epoch": 0.07528458739111663,
      "grad_norm": 1.1665087938308716,
      "learning_rate": 1e-05,
      "loss": 0.2422,
      "step": 21970
    },
    {
      "epoch": 0.07531885438583265,
      "grad_norm": 1.1135221719741821,
      "learning_rate": 1e-05,
      "loss": 0.229,
      "step": 21980
    },
    {
      "epoch": 0.07535312138054868,
      "grad_norm": 1.035752296447754,
      "learning_rate": 1e-05,
      "loss": 0.2399,
      "step": 21990
    },
    {
      "epoch": 0.07538738837526471,
      "grad_norm": 1.0547956228256226,
      "learning_rate": 1e-05,
      "loss": 0.2174,
      "step": 22000
    },
    {
      "epoch": 0.07538738837526471,
      "eval_cer": 12.762391956253309,
      "eval_loss": 0.24175649881362915,
      "eval_normalized_cer": 9.012789768185451,
      "eval_runtime": 226.9242,
      "eval_samples_per_second": 2.256,
      "eval_steps_per_second": 0.035,
      "step": 22000
    },
    {
      "epoch": 0.07542165536998074,
      "grad_norm": 1.011830449104309,
      "learning_rate": 1e-05,
      "loss": 0.2245,
      "step": 22010
    },
    {
      "epoch": 0.07545592236469677,
      "grad_norm": 1.245556354522705,
      "learning_rate": 1e-05,
      "loss": 0.2369,
      "step": 22020
    },
    {
      "epoch": 0.0754901893594128,
      "grad_norm": 1.0645912885665894,
      "learning_rate": 1e-05,
      "loss": 0.2611,
      "step": 22030
    },
    {
      "epoch": 0.07552445635412883,
      "grad_norm": 1.173473596572876,
      "learning_rate": 1e-05,
      "loss": 0.2485,
      "step": 22040
    },
    {
      "epoch": 0.07555872334884486,
      "grad_norm": 1.1239567995071411,
      "learning_rate": 1e-05,
      "loss": 0.2523,
      "step": 22050
    },
    {
      "epoch": 0.07559299034356089,
      "grad_norm": 1.2006217241287231,
      "learning_rate": 1e-05,
      "loss": 0.2411,
      "step": 22060
    },
    {
      "epoch": 0.07562725733827692,
      "grad_norm": 1.0894744396209717,
      "learning_rate": 1e-05,
      "loss": 0.2497,
      "step": 22070
    },
    {
      "epoch": 0.07566152433299295,
      "grad_norm": 1.3402432203292847,
      "learning_rate": 1e-05,
      "loss": 0.2547,
      "step": 22080
    },
    {
      "epoch": 0.07569579132770898,
      "grad_norm": 1.247170090675354,
      "learning_rate": 1e-05,
      "loss": 0.2631,
      "step": 22090
    },
    {
      "epoch": 0.075730058322425,
      "grad_norm": 1.1156800985336304,
      "learning_rate": 1e-05,
      "loss": 0.2396,
      "step": 22100
    },
    {
      "epoch": 0.07576432531714104,
      "grad_norm": 1.0527523756027222,
      "learning_rate": 1e-05,
      "loss": 0.2524,
      "step": 22110
    },
    {
      "epoch": 0.07579859231185707,
      "grad_norm": 0.9769160151481628,
      "learning_rate": 1e-05,
      "loss": 0.2461,
      "step": 22120
    },
    {
      "epoch": 0.07583285930657309,
      "grad_norm": 1.066941261291504,
      "learning_rate": 1e-05,
      "loss": 0.2575,
      "step": 22130
    },
    {
      "epoch": 0.07586712630128913,
      "grad_norm": 1.226346492767334,
      "learning_rate": 1e-05,
      "loss": 0.2542,
      "step": 22140
    },
    {
      "epoch": 0.07590139329600515,
      "grad_norm": 0.9813686609268188,
      "learning_rate": 1e-05,
      "loss": 0.2478,
      "step": 22150
    },
    {
      "epoch": 0.07593566029072119,
      "grad_norm": 1.127979040145874,
      "learning_rate": 1e-05,
      "loss": 0.2589,
      "step": 22160
    },
    {
      "epoch": 0.07596992728543721,
      "grad_norm": 1.195841908454895,
      "learning_rate": 1e-05,
      "loss": 0.2605,
      "step": 22170
    },
    {
      "epoch": 0.07600419428015324,
      "grad_norm": 1.137445092201233,
      "learning_rate": 1e-05,
      "loss": 0.2423,
      "step": 22180
    },
    {
      "epoch": 0.07603846127486927,
      "grad_norm": 0.9853975176811218,
      "learning_rate": 1e-05,
      "loss": 0.246,
      "step": 22190
    },
    {
      "epoch": 0.0760727282695853,
      "grad_norm": 1.0872706174850464,
      "learning_rate": 1e-05,
      "loss": 0.2588,
      "step": 22200
    },
    {
      "epoch": 0.07610699526430133,
      "grad_norm": 1.1870428323745728,
      "learning_rate": 1e-05,
      "loss": 0.2482,
      "step": 22210
    },
    {
      "epoch": 0.07614126225901736,
      "grad_norm": 1.1714868545532227,
      "learning_rate": 1e-05,
      "loss": 0.2361,
      "step": 22220
    },
    {
      "epoch": 0.07617552925373339,
      "grad_norm": 1.1878902912139893,
      "learning_rate": 1e-05,
      "loss": 0.2532,
      "step": 22230
    },
    {
      "epoch": 0.07620979624844942,
      "grad_norm": 1.0626158714294434,
      "learning_rate": 1e-05,
      "loss": 0.2499,
      "step": 22240
    },
    {
      "epoch": 0.07624406324316545,
      "grad_norm": 0.9548189640045166,
      "learning_rate": 1e-05,
      "loss": 0.2515,
      "step": 22250
    },
    {
      "epoch": 0.07627833023788148,
      "grad_norm": 1.018364429473877,
      "learning_rate": 1e-05,
      "loss": 0.2523,
      "step": 22260
    },
    {
      "epoch": 0.0763125972325975,
      "grad_norm": 1.1226998567581177,
      "learning_rate": 1e-05,
      "loss": 0.2328,
      "step": 22270
    },
    {
      "epoch": 0.07634686422731353,
      "grad_norm": 1.0202254056930542,
      "learning_rate": 1e-05,
      "loss": 0.2356,
      "step": 22280
    },
    {
      "epoch": 0.07638113122202957,
      "grad_norm": 1.1206306219100952,
      "learning_rate": 1e-05,
      "loss": 0.231,
      "step": 22290
    },
    {
      "epoch": 0.07641539821674559,
      "grad_norm": 1.131699562072754,
      "learning_rate": 1e-05,
      "loss": 0.2415,
      "step": 22300
    },
    {
      "epoch": 0.07644966521146163,
      "grad_norm": 1.0941540002822876,
      "learning_rate": 1e-05,
      "loss": 0.2312,
      "step": 22310
    },
    {
      "epoch": 0.07648393220617765,
      "grad_norm": 1.0358905792236328,
      "learning_rate": 1e-05,
      "loss": 0.2464,
      "step": 22320
    },
    {
      "epoch": 0.07651819920089369,
      "grad_norm": 1.0331768989562988,
      "learning_rate": 1e-05,
      "loss": 0.239,
      "step": 22330
    },
    {
      "epoch": 0.07655246619560971,
      "grad_norm": 1.0920426845550537,
      "learning_rate": 1e-05,
      "loss": 0.2489,
      "step": 22340
    },
    {
      "epoch": 0.07658673319032575,
      "grad_norm": 1.1783214807510376,
      "learning_rate": 1e-05,
      "loss": 0.2335,
      "step": 22350
    },
    {
      "epoch": 0.07662100018504177,
      "grad_norm": 1.2217092514038086,
      "learning_rate": 1e-05,
      "loss": 0.2247,
      "step": 22360
    },
    {
      "epoch": 0.0766552671797578,
      "grad_norm": 1.1125669479370117,
      "learning_rate": 1e-05,
      "loss": 0.2493,
      "step": 22370
    },
    {
      "epoch": 0.07668953417447383,
      "grad_norm": 0.9087139964103699,
      "learning_rate": 1e-05,
      "loss": 0.2345,
      "step": 22380
    },
    {
      "epoch": 0.07672380116918986,
      "grad_norm": 1.099518060684204,
      "learning_rate": 1e-05,
      "loss": 0.2524,
      "step": 22390
    },
    {
      "epoch": 0.07675806816390589,
      "grad_norm": 1.0696626901626587,
      "learning_rate": 1e-05,
      "loss": 0.2172,
      "step": 22400
    },
    {
      "epoch": 0.07679233515862192,
      "grad_norm": 1.045519471168518,
      "learning_rate": 1e-05,
      "loss": 0.2295,
      "step": 22410
    },
    {
      "epoch": 0.07682660215333795,
      "grad_norm": 1.1601028442382812,
      "learning_rate": 1e-05,
      "loss": 0.253,
      "step": 22420
    },
    {
      "epoch": 0.07686086914805398,
      "grad_norm": 1.1112728118896484,
      "learning_rate": 1e-05,
      "loss": 0.2328,
      "step": 22430
    },
    {
      "epoch": 0.07689513614277001,
      "grad_norm": 1.0132625102996826,
      "learning_rate": 1e-05,
      "loss": 0.2302,
      "step": 22440
    },
    {
      "epoch": 0.07692940313748603,
      "grad_norm": 0.9999580383300781,
      "learning_rate": 1e-05,
      "loss": 0.2466,
      "step": 22450
    },
    {
      "epoch": 0.07696367013220207,
      "grad_norm": 1.157254934310913,
      "learning_rate": 1e-05,
      "loss": 0.2282,
      "step": 22460
    },
    {
      "epoch": 0.07699793712691809,
      "grad_norm": 1.195129632949829,
      "learning_rate": 1e-05,
      "loss": 0.2334,
      "step": 22470
    },
    {
      "epoch": 0.07703220412163413,
      "grad_norm": 1.1293895244598389,
      "learning_rate": 1e-05,
      "loss": 0.2447,
      "step": 22480
    },
    {
      "epoch": 0.07706647111635015,
      "grad_norm": 1.0318418741226196,
      "learning_rate": 1e-05,
      "loss": 0.2348,
      "step": 22490
    },
    {
      "epoch": 0.07710073811106619,
      "grad_norm": 1.0450618267059326,
      "learning_rate": 1e-05,
      "loss": 0.2459,
      "step": 22500
    },
    {
      "epoch": 0.07710073811106619,
      "eval_cer": 12.638913388604692,
      "eval_loss": 0.2446923851966858,
      "eval_normalized_cer": 9.082733812949641,
      "eval_runtime": 228.122,
      "eval_samples_per_second": 2.244,
      "eval_steps_per_second": 0.035,
      "step": 22500
    },
    {
      "epoch": 0.07713500510578221,
      "grad_norm": 1.0523881912231445,
      "learning_rate": 1e-05,
      "loss": 0.2564,
      "step": 22510
    },
    {
      "epoch": 0.07716927210049825,
      "grad_norm": 1.209961175918579,
      "learning_rate": 1e-05,
      "loss": 0.2204,
      "step": 22520
    },
    {
      "epoch": 0.07720353909521427,
      "grad_norm": 1.1583831310272217,
      "learning_rate": 1e-05,
      "loss": 0.2551,
      "step": 22530
    },
    {
      "epoch": 0.0772378060899303,
      "grad_norm": 1.4498618841171265,
      "learning_rate": 1e-05,
      "loss": 0.2528,
      "step": 22540
    },
    {
      "epoch": 0.07727207308464633,
      "grad_norm": 1.215000867843628,
      "learning_rate": 1e-05,
      "loss": 0.2464,
      "step": 22550
    },
    {
      "epoch": 0.07730634007936237,
      "grad_norm": 1.0598031282424927,
      "learning_rate": 1e-05,
      "loss": 0.2394,
      "step": 22560
    },
    {
      "epoch": 0.07734060707407839,
      "grad_norm": 1.1213382482528687,
      "learning_rate": 1e-05,
      "loss": 0.2566,
      "step": 22570
    },
    {
      "epoch": 0.07737487406879442,
      "grad_norm": 1.1188169717788696,
      "learning_rate": 1e-05,
      "loss": 0.2395,
      "step": 22580
    },
    {
      "epoch": 0.07740914106351045,
      "grad_norm": 1.0692533254623413,
      "learning_rate": 1e-05,
      "loss": 0.2376,
      "step": 22590
    },
    {
      "epoch": 0.07744340805822647,
      "grad_norm": 1.131622076034546,
      "learning_rate": 1e-05,
      "loss": 0.2414,
      "step": 22600
    },
    {
      "epoch": 0.07747767505294251,
      "grad_norm": 1.1728034019470215,
      "learning_rate": 1e-05,
      "loss": 0.2375,
      "step": 22610
    },
    {
      "epoch": 0.07751194204765853,
      "grad_norm": 0.9095053672790527,
      "learning_rate": 1e-05,
      "loss": 0.2347,
      "step": 22620
    },
    {
      "epoch": 0.07754620904237457,
      "grad_norm": 1.0425827503204346,
      "learning_rate": 1e-05,
      "loss": 0.2456,
      "step": 22630
    },
    {
      "epoch": 0.07758047603709059,
      "grad_norm": 1.189103364944458,
      "learning_rate": 1e-05,
      "loss": 0.2496,
      "step": 22640
    },
    {
      "epoch": 0.07761474303180663,
      "grad_norm": 1.1523323059082031,
      "learning_rate": 1e-05,
      "loss": 0.2463,
      "step": 22650
    },
    {
      "epoch": 0.07764901002652265,
      "grad_norm": 1.1360780000686646,
      "learning_rate": 1e-05,
      "loss": 0.2367,
      "step": 22660
    },
    {
      "epoch": 0.07768327702123869,
      "grad_norm": 1.0910000801086426,
      "learning_rate": 1e-05,
      "loss": 0.2428,
      "step": 22670
    },
    {
      "epoch": 0.07771754401595471,
      "grad_norm": 1.10030198097229,
      "learning_rate": 1e-05,
      "loss": 0.2441,
      "step": 22680
    },
    {
      "epoch": 0.07775181101067075,
      "grad_norm": 1.3179526329040527,
      "learning_rate": 1e-05,
      "loss": 0.2558,
      "step": 22690
    },
    {
      "epoch": 0.07778607800538677,
      "grad_norm": 1.0743062496185303,
      "learning_rate": 1e-05,
      "loss": 0.2352,
      "step": 22700
    },
    {
      "epoch": 0.0778203450001028,
      "grad_norm": 0.9239439368247986,
      "learning_rate": 1e-05,
      "loss": 0.2409,
      "step": 22710
    },
    {
      "epoch": 0.07785461199481883,
      "grad_norm": 1.1279993057250977,
      "learning_rate": 1e-05,
      "loss": 0.2573,
      "step": 22720
    },
    {
      "epoch": 0.07788887898953487,
      "grad_norm": 1.0824002027511597,
      "learning_rate": 1e-05,
      "loss": 0.2463,
      "step": 22730
    },
    {
      "epoch": 0.07792314598425089,
      "grad_norm": 1.1521611213684082,
      "learning_rate": 1e-05,
      "loss": 0.2311,
      "step": 22740
    },
    {
      "epoch": 0.07795741297896693,
      "grad_norm": 1.1287387609481812,
      "learning_rate": 1e-05,
      "loss": 0.2437,
      "step": 22750
    },
    {
      "epoch": 0.07799167997368295,
      "grad_norm": 1.250704050064087,
      "learning_rate": 1e-05,
      "loss": 0.2394,
      "step": 22760
    },
    {
      "epoch": 0.07802594696839897,
      "grad_norm": 1.2648214101791382,
      "learning_rate": 1e-05,
      "loss": 0.2433,
      "step": 22770
    },
    {
      "epoch": 0.07806021396311501,
      "grad_norm": 1.1684668064117432,
      "learning_rate": 1e-05,
      "loss": 0.2375,
      "step": 22780
    },
    {
      "epoch": 0.07809448095783103,
      "grad_norm": 1.0133503675460815,
      "learning_rate": 1e-05,
      "loss": 0.2426,
      "step": 22790
    },
    {
      "epoch": 0.07812874795254707,
      "grad_norm": 1.0540106296539307,
      "learning_rate": 1e-05,
      "loss": 0.2172,
      "step": 22800
    },
    {
      "epoch": 0.07816301494726309,
      "grad_norm": 0.9576539993286133,
      "learning_rate": 1e-05,
      "loss": 0.2228,
      "step": 22810
    },
    {
      "epoch": 0.07819728194197913,
      "grad_norm": 1.2314975261688232,
      "learning_rate": 1e-05,
      "loss": 0.2274,
      "step": 22820
    },
    {
      "epoch": 0.07823154893669515,
      "grad_norm": 1.1600431203842163,
      "learning_rate": 1e-05,
      "loss": 0.2212,
      "step": 22830
    },
    {
      "epoch": 0.07826581593141119,
      "grad_norm": 1.0336724519729614,
      "learning_rate": 1e-05,
      "loss": 0.2157,
      "step": 22840
    },
    {
      "epoch": 0.07830008292612721,
      "grad_norm": 1.209795355796814,
      "learning_rate": 1e-05,
      "loss": 0.2208,
      "step": 22850
    },
    {
      "epoch": 0.07833434992084325,
      "grad_norm": 1.052100419998169,
      "learning_rate": 1e-05,
      "loss": 0.2151,
      "step": 22860
    },
    {
      "epoch": 0.07836861691555927,
      "grad_norm": 1.0897157192230225,
      "learning_rate": 1e-05,
      "loss": 0.2235,
      "step": 22870
    },
    {
      "epoch": 0.0784028839102753,
      "grad_norm": 1.0175033807754517,
      "learning_rate": 1e-05,
      "loss": 0.2226,
      "step": 22880
    },
    {
      "epoch": 0.07843715090499133,
      "grad_norm": 1.3162877559661865,
      "learning_rate": 1e-05,
      "loss": 0.2323,
      "step": 22890
    },
    {
      "epoch": 0.07847141789970737,
      "grad_norm": 1.1104750633239746,
      "learning_rate": 1e-05,
      "loss": 0.2202,
      "step": 22900
    },
    {
      "epoch": 0.07850568489442339,
      "grad_norm": 1.1699038743972778,
      "learning_rate": 1e-05,
      "loss": 0.2262,
      "step": 22910
    },
    {
      "epoch": 0.07853995188913943,
      "grad_norm": 1.2070201635360718,
      "learning_rate": 1e-05,
      "loss": 0.2369,
      "step": 22920
    },
    {
      "epoch": 0.07857421888385545,
      "grad_norm": 1.0363352298736572,
      "learning_rate": 1e-05,
      "loss": 0.208,
      "step": 22930
    },
    {
      "epoch": 0.07860848587857147,
      "grad_norm": 1.0242831707000732,
      "learning_rate": 1e-05,
      "loss": 0.1995,
      "step": 22940
    },
    {
      "epoch": 0.07864275287328751,
      "grad_norm": 0.9983187317848206,
      "learning_rate": 1e-05,
      "loss": 0.2241,
      "step": 22950
    },
    {
      "epoch": 0.07867701986800353,
      "grad_norm": 1.2646509408950806,
      "learning_rate": 1e-05,
      "loss": 0.22,
      "step": 22960
    },
    {
      "epoch": 0.07871128686271957,
      "grad_norm": 1.0720429420471191,
      "learning_rate": 1e-05,
      "loss": 0.2157,
      "step": 22970
    },
    {
      "epoch": 0.07874555385743559,
      "grad_norm": 1.047745943069458,
      "learning_rate": 1e-05,
      "loss": 0.2166,
      "step": 22980
    },
    {
      "epoch": 0.07877982085215163,
      "grad_norm": 0.9465416073799133,
      "learning_rate": 1e-05,
      "loss": 0.2211,
      "step": 22990
    },
    {
      "epoch": 0.07881408784686765,
      "grad_norm": 1.141426920890808,
      "learning_rate": 1e-05,
      "loss": 0.2442,
      "step": 23000
    },
    {
      "epoch": 0.07881408784686765,
      "eval_cer": 12.444875639442582,
      "eval_loss": 0.24720345437526703,
      "eval_normalized_cer": 8.76298960831335,
      "eval_runtime": 227.2823,
      "eval_samples_per_second": 2.253,
      "eval_steps_per_second": 0.035,
      "step": 23000
    },
    {
      "epoch": 0.07884835484158369,
      "grad_norm": 1.0129847526550293,
      "learning_rate": 1e-05,
      "loss": 0.2215,
      "step": 23010
    },
    {
      "epoch": 0.07888262183629971,
      "grad_norm": 1.0782142877578735,
      "learning_rate": 1e-05,
      "loss": 0.2159,
      "step": 23020
    },
    {
      "epoch": 0.07891688883101575,
      "grad_norm": 1.0121062994003296,
      "learning_rate": 1e-05,
      "loss": 0.2124,
      "step": 23030
    },
    {
      "epoch": 0.07895115582573177,
      "grad_norm": 1.1070152521133423,
      "learning_rate": 1e-05,
      "loss": 0.2446,
      "step": 23040
    },
    {
      "epoch": 0.0789854228204478,
      "grad_norm": 1.2437150478363037,
      "learning_rate": 1e-05,
      "loss": 0.2444,
      "step": 23050
    },
    {
      "epoch": 0.07901968981516383,
      "grad_norm": 1.1498879194259644,
      "learning_rate": 1e-05,
      "loss": 0.2443,
      "step": 23060
    },
    {
      "epoch": 0.07905395680987987,
      "grad_norm": 1.145849347114563,
      "learning_rate": 1e-05,
      "loss": 0.2538,
      "step": 23070
    },
    {
      "epoch": 0.07908822380459589,
      "grad_norm": 1.1981453895568848,
      "learning_rate": 1e-05,
      "loss": 0.2478,
      "step": 23080
    },
    {
      "epoch": 0.07912249079931191,
      "grad_norm": 1.113627552986145,
      "learning_rate": 1e-05,
      "loss": 0.2401,
      "step": 23090
    },
    {
      "epoch": 0.07915675779402795,
      "grad_norm": 1.152208685874939,
      "learning_rate": 1e-05,
      "loss": 0.242,
      "step": 23100
    },
    {
      "epoch": 0.07919102478874397,
      "grad_norm": 1.2640796899795532,
      "learning_rate": 1e-05,
      "loss": 0.2674,
      "step": 23110
    },
    {
      "epoch": 0.07922529178346001,
      "grad_norm": 1.1089024543762207,
      "learning_rate": 1e-05,
      "loss": 0.2374,
      "step": 23120
    },
    {
      "epoch": 0.07925955877817603,
      "grad_norm": 1.1184459924697876,
      "learning_rate": 1e-05,
      "loss": 0.2436,
      "step": 23130
    },
    {
      "epoch": 0.07929382577289207,
      "grad_norm": 1.1597626209259033,
      "learning_rate": 1e-05,
      "loss": 0.2502,
      "step": 23140
    },
    {
      "epoch": 0.07932809276760809,
      "grad_norm": 0.9689920544624329,
      "learning_rate": 1e-05,
      "loss": 0.2369,
      "step": 23150
    },
    {
      "epoch": 0.07936235976232413,
      "grad_norm": 1.0698407888412476,
      "learning_rate": 1e-05,
      "loss": 0.2347,
      "step": 23160
    },
    {
      "epoch": 0.07939662675704015,
      "grad_norm": 1.1528737545013428,
      "learning_rate": 1e-05,
      "loss": 0.2539,
      "step": 23170
    },
    {
      "epoch": 0.07943089375175619,
      "grad_norm": 1.134964108467102,
      "learning_rate": 1e-05,
      "loss": 0.2557,
      "step": 23180
    },
    {
      "epoch": 0.07946516074647221,
      "grad_norm": 1.1838676929473877,
      "learning_rate": 1e-05,
      "loss": 0.2489,
      "step": 23190
    },
    {
      "epoch": 0.07949942774118825,
      "grad_norm": 1.007856011390686,
      "learning_rate": 1e-05,
      "loss": 0.2426,
      "step": 23200
    },
    {
      "epoch": 0.07953369473590427,
      "grad_norm": 1.0157620906829834,
      "learning_rate": 1e-05,
      "loss": 0.2569,
      "step": 23210
    },
    {
      "epoch": 0.0795679617306203,
      "grad_norm": 1.1552293300628662,
      "learning_rate": 1e-05,
      "loss": 0.2655,
      "step": 23220
    },
    {
      "epoch": 0.07960222872533633,
      "grad_norm": 1.0644792318344116,
      "learning_rate": 1e-05,
      "loss": 0.2413,
      "step": 23230
    },
    {
      "epoch": 0.07963649572005237,
      "grad_norm": 1.0649666786193848,
      "learning_rate": 1e-05,
      "loss": 0.2397,
      "step": 23240
    },
    {
      "epoch": 0.07967076271476839,
      "grad_norm": 1.0416626930236816,
      "learning_rate": 1e-05,
      "loss": 0.2419,
      "step": 23250
    },
    {
      "epoch": 0.07970502970948441,
      "grad_norm": 1.0674875974655151,
      "learning_rate": 1e-05,
      "loss": 0.2472,
      "step": 23260
    },
    {
      "epoch": 0.07973929670420045,
      "grad_norm": 1.0134835243225098,
      "learning_rate": 1e-05,
      "loss": 0.2306,
      "step": 23270
    },
    {
      "epoch": 0.07977356369891647,
      "grad_norm": 1.0753681659698486,
      "learning_rate": 1e-05,
      "loss": 0.2363,
      "step": 23280
    },
    {
      "epoch": 0.07980783069363251,
      "grad_norm": 1.0712649822235107,
      "learning_rate": 1e-05,
      "loss": 0.2321,
      "step": 23290
    },
    {
      "epoch": 0.07984209768834853,
      "grad_norm": 0.9887093305587769,
      "learning_rate": 1e-05,
      "loss": 0.2194,
      "step": 23300
    },
    {
      "epoch": 0.07987636468306457,
      "grad_norm": 1.050628900527954,
      "learning_rate": 1e-05,
      "loss": 0.2413,
      "step": 23310
    },
    {
      "epoch": 0.07991063167778059,
      "grad_norm": 1.0947946310043335,
      "learning_rate": 1e-05,
      "loss": 0.2116,
      "step": 23320
    },
    {
      "epoch": 0.07994489867249663,
      "grad_norm": 0.9869987368583679,
      "learning_rate": 1e-05,
      "loss": 0.2355,
      "step": 23330
    },
    {
      "epoch": 0.07997916566721265,
      "grad_norm": 0.9494191408157349,
      "learning_rate": 1e-05,
      "loss": 0.2268,
      "step": 23340
    },
    {
      "epoch": 0.08001343266192869,
      "grad_norm": 1.2696855068206787,
      "learning_rate": 1e-05,
      "loss": 0.2252,
      "step": 23350
    },
    {
      "epoch": 0.08004769965664471,
      "grad_norm": 1.2440532445907593,
      "learning_rate": 1e-05,
      "loss": 0.2424,
      "step": 23360
    },
    {
      "epoch": 0.08008196665136075,
      "grad_norm": 1.074804425239563,
      "learning_rate": 1e-05,
      "loss": 0.2259,
      "step": 23370
    },
    {
      "epoch": 0.08011623364607677,
      "grad_norm": 1.2430806159973145,
      "learning_rate": 1e-05,
      "loss": 0.2393,
      "step": 23380
    },
    {
      "epoch": 0.08015050064079281,
      "grad_norm": 1.1680575609207153,
      "learning_rate": 1e-05,
      "loss": 0.2316,
      "step": 23390
    },
    {
      "epoch": 0.08018476763550883,
      "grad_norm": 1.1257153749465942,
      "learning_rate": 1e-05,
      "loss": 0.2395,
      "step": 23400
    },
    {
      "epoch": 0.08021903463022487,
      "grad_norm": 1.1165947914123535,
      "learning_rate": 1e-05,
      "loss": 0.2368,
      "step": 23410
    },
    {
      "epoch": 0.08025330162494089,
      "grad_norm": 1.2653453350067139,
      "learning_rate": 1e-05,
      "loss": 0.235,
      "step": 23420
    },
    {
      "epoch": 0.08028756861965691,
      "grad_norm": 1.064660906791687,
      "learning_rate": 1e-05,
      "loss": 0.2463,
      "step": 23430
    },
    {
      "epoch": 0.08032183561437295,
      "grad_norm": 1.0772967338562012,
      "learning_rate": 1e-05,
      "loss": 0.2351,
      "step": 23440
    },
    {
      "epoch": 0.08035610260908897,
      "grad_norm": 1.1206378936767578,
      "learning_rate": 1e-05,
      "loss": 0.2302,
      "step": 23450
    },
    {
      "epoch": 0.08039036960380501,
      "grad_norm": 0.9457486271858215,
      "learning_rate": 1e-05,
      "loss": 0.2412,
      "step": 23460
    },
    {
      "epoch": 0.08042463659852103,
      "grad_norm": 1.0531508922576904,
      "learning_rate": 1e-05,
      "loss": 0.245,
      "step": 23470
    },
    {
      "epoch": 0.08045890359323707,
      "grad_norm": 1.0849158763885498,
      "learning_rate": 1e-05,
      "loss": 0.2377,
      "step": 23480
    },
    {
      "epoch": 0.08049317058795309,
      "grad_norm": 1.1281490325927734,
      "learning_rate": 1e-05,
      "loss": 0.2565,
      "step": 23490
    },
    {
      "epoch": 0.08052743758266913,
      "grad_norm": 1.111106276512146,
      "learning_rate": 1e-05,
      "loss": 0.2303,
      "step": 23500
    },
    {
      "epoch": 0.08052743758266913,
      "eval_cer": 12.180278708766979,
      "eval_loss": 0.24189412593841553,
      "eval_normalized_cer": 8.733013589128696,
      "eval_runtime": 228.3161,
      "eval_samples_per_second": 2.243,
      "eval_steps_per_second": 0.035,
      "step": 23500
    },
    {
      "epoch": 0.08056170457738515,
      "grad_norm": 1.0803686380386353,
      "learning_rate": 1e-05,
      "loss": 0.217,
      "step": 23510
    },
    {
      "epoch": 0.08059597157210119,
      "grad_norm": 1.0238206386566162,
      "learning_rate": 1e-05,
      "loss": 0.2235,
      "step": 23520
    },
    {
      "epoch": 0.08063023856681721,
      "grad_norm": 1.2258280515670776,
      "learning_rate": 1e-05,
      "loss": 0.2355,
      "step": 23530
    },
    {
      "epoch": 0.08066450556153325,
      "grad_norm": 1.0921289920806885,
      "learning_rate": 1e-05,
      "loss": 0.2277,
      "step": 23540
    },
    {
      "epoch": 0.08069877255624927,
      "grad_norm": 0.9965620636940002,
      "learning_rate": 1e-05,
      "loss": 0.2344,
      "step": 23550
    },
    {
      "epoch": 0.08073303955096531,
      "grad_norm": 1.0104345083236694,
      "learning_rate": 1e-05,
      "loss": 0.2327,
      "step": 23560
    },
    {
      "epoch": 0.08076730654568133,
      "grad_norm": 1.0309514999389648,
      "learning_rate": 1e-05,
      "loss": 0.2498,
      "step": 23570
    },
    {
      "epoch": 0.08080157354039735,
      "grad_norm": 0.9816809892654419,
      "learning_rate": 1e-05,
      "loss": 0.236,
      "step": 23580
    },
    {
      "epoch": 0.08083584053511339,
      "grad_norm": 0.9675121903419495,
      "learning_rate": 1e-05,
      "loss": 0.2193,
      "step": 23590
    },
    {
      "epoch": 0.08087010752982941,
      "grad_norm": 0.9846246838569641,
      "learning_rate": 1e-05,
      "loss": 0.2478,
      "step": 23600
    },
    {
      "epoch": 0.08090437452454545,
      "grad_norm": 1.154410719871521,
      "learning_rate": 1e-05,
      "loss": 0.2506,
      "step": 23610
    },
    {
      "epoch": 0.08093864151926147,
      "grad_norm": 1.148653268814087,
      "learning_rate": 1e-05,
      "loss": 0.2418,
      "step": 23620
    },
    {
      "epoch": 0.08097290851397751,
      "grad_norm": 1.1022738218307495,
      "learning_rate": 1e-05,
      "loss": 0.2348,
      "step": 23630
    },
    {
      "epoch": 0.08100717550869353,
      "grad_norm": 1.2160323858261108,
      "learning_rate": 1e-05,
      "loss": 0.2563,
      "step": 23640
    },
    {
      "epoch": 0.08104144250340957,
      "grad_norm": 1.0583524703979492,
      "learning_rate": 1e-05,
      "loss": 0.2244,
      "step": 23650
    },
    {
      "epoch": 0.08107570949812559,
      "grad_norm": 1.0356570482254028,
      "learning_rate": 1e-05,
      "loss": 0.242,
      "step": 23660
    },
    {
      "epoch": 0.08110997649284163,
      "grad_norm": 1.0162861347198486,
      "learning_rate": 1e-05,
      "loss": 0.2675,
      "step": 23670
    },
    {
      "epoch": 0.08114424348755765,
      "grad_norm": 1.1000967025756836,
      "learning_rate": 1e-05,
      "loss": 0.2326,
      "step": 23680
    },
    {
      "epoch": 0.08117851048227369,
      "grad_norm": 1.1410856246948242,
      "learning_rate": 1e-05,
      "loss": 0.2464,
      "step": 23690
    },
    {
      "epoch": 0.08121277747698971,
      "grad_norm": 1.1000715494155884,
      "learning_rate": 1e-05,
      "loss": 0.256,
      "step": 23700
    },
    {
      "epoch": 0.08124704447170575,
      "grad_norm": 1.1132537126541138,
      "learning_rate": 1e-05,
      "loss": 0.2416,
      "step": 23710
    },
    {
      "epoch": 0.08128131146642177,
      "grad_norm": 1.1105347871780396,
      "learning_rate": 1e-05,
      "loss": 0.2533,
      "step": 23720
    },
    {
      "epoch": 0.08131557846113781,
      "grad_norm": 1.1548352241516113,
      "learning_rate": 1e-05,
      "loss": 0.2426,
      "step": 23730
    },
    {
      "epoch": 0.08134984545585383,
      "grad_norm": 1.0594673156738281,
      "learning_rate": 1e-05,
      "loss": 0.2435,
      "step": 23740
    },
    {
      "epoch": 0.08138411245056985,
      "grad_norm": 1.050610065460205,
      "learning_rate": 1e-05,
      "loss": 0.244,
      "step": 23750
    },
    {
      "epoch": 0.08141837944528589,
      "grad_norm": 1.1562016010284424,
      "learning_rate": 1e-05,
      "loss": 0.2439,
      "step": 23760
    },
    {
      "epoch": 0.08145264644000191,
      "grad_norm": 0.9773512482643127,
      "learning_rate": 1e-05,
      "loss": 0.2409,
      "step": 23770
    },
    {
      "epoch": 0.08148691343471795,
      "grad_norm": 1.160895586013794,
      "learning_rate": 1e-05,
      "loss": 0.256,
      "step": 23780
    },
    {
      "epoch": 0.08152118042943397,
      "grad_norm": 1.017398476600647,
      "learning_rate": 1e-05,
      "loss": 0.2429,
      "step": 23790
    },
    {
      "epoch": 0.08155544742415001,
      "grad_norm": 1.0898586511611938,
      "learning_rate": 1e-05,
      "loss": 0.2478,
      "step": 23800
    },
    {
      "epoch": 0.08158971441886603,
      "grad_norm": 1.2312469482421875,
      "learning_rate": 1e-05,
      "loss": 0.2474,
      "step": 23810
    },
    {
      "epoch": 0.08162398141358207,
      "grad_norm": 1.0122491121292114,
      "learning_rate": 1e-05,
      "loss": 0.2463,
      "step": 23820
    },
    {
      "epoch": 0.08165824840829809,
      "grad_norm": 1.2574313879013062,
      "learning_rate": 1e-05,
      "loss": 0.2382,
      "step": 23830
    },
    {
      "epoch": 0.08169251540301413,
      "grad_norm": 1.0722161531448364,
      "learning_rate": 1e-05,
      "loss": 0.2337,
      "step": 23840
    },
    {
      "epoch": 0.08172678239773015,
      "grad_norm": 1.1575268507003784,
      "learning_rate": 1e-05,
      "loss": 0.2497,
      "step": 23850
    },
    {
      "epoch": 0.08176104939244619,
      "grad_norm": 1.1297662258148193,
      "learning_rate": 1e-05,
      "loss": 0.2474,
      "step": 23860
    },
    {
      "epoch": 0.08179531638716221,
      "grad_norm": 0.9983540773391724,
      "learning_rate": 1e-05,
      "loss": 0.2477,
      "step": 23870
    },
    {
      "epoch": 0.08182958338187825,
      "grad_norm": 0.9470457434654236,
      "learning_rate": 1e-05,
      "loss": 0.2475,
      "step": 23880
    },
    {
      "epoch": 0.08186385037659427,
      "grad_norm": 1.1658211946487427,
      "learning_rate": 1e-05,
      "loss": 0.2746,
      "step": 23890
    },
    {
      "epoch": 0.0818981173713103,
      "grad_norm": 1.1910369396209717,
      "learning_rate": 1e-05,
      "loss": 0.2558,
      "step": 23900
    },
    {
      "epoch": 0.08193238436602633,
      "grad_norm": 1.0597240924835205,
      "learning_rate": 1e-05,
      "loss": 0.2325,
      "step": 23910
    },
    {
      "epoch": 0.08196665136074235,
      "grad_norm": 1.0382158756256104,
      "learning_rate": 1e-05,
      "loss": 0.2413,
      "step": 23920
    },
    {
      "epoch": 0.08200091835545839,
      "grad_norm": 1.107107400894165,
      "learning_rate": 1e-05,
      "loss": 0.2505,
      "step": 23930
    },
    {
      "epoch": 0.08203518535017441,
      "grad_norm": 1.1651002168655396,
      "learning_rate": 1e-05,
      "loss": 0.2391,
      "step": 23940
    },
    {
      "epoch": 0.08206945234489045,
      "grad_norm": 1.079775333404541,
      "learning_rate": 1e-05,
      "loss": 0.2651,
      "step": 23950
    },
    {
      "epoch": 0.08210371933960647,
      "grad_norm": 0.9570854902267456,
      "learning_rate": 1e-05,
      "loss": 0.2501,
      "step": 23960
    },
    {
      "epoch": 0.08213798633432251,
      "grad_norm": 1.0310214757919312,
      "learning_rate": 1e-05,
      "loss": 0.2486,
      "step": 23970
    },
    {
      "epoch": 0.08217225332903853,
      "grad_norm": 1.045013189315796,
      "learning_rate": 1e-05,
      "loss": 0.2494,
      "step": 23980
    },
    {
      "epoch": 0.08220652032375457,
      "grad_norm": 1.0111870765686035,
      "learning_rate": 1e-05,
      "loss": 0.2707,
      "step": 23990
    },
    {
      "epoch": 0.08224078731847059,
      "grad_norm": 1.1841882467269897,
      "learning_rate": 1e-05,
      "loss": 0.2536,
      "step": 24000
    },
    {
      "epoch": 0.08224078731847059,
      "eval_cer": 12.621273593226318,
      "eval_loss": 0.24069419503211975,
      "eval_normalized_cer": 8.93285371702638,
      "eval_runtime": 226.7251,
      "eval_samples_per_second": 2.258,
      "eval_steps_per_second": 0.035,
      "step": 24000
    },
    {
      "epoch": 0.08227505431318663,
      "grad_norm": 1.059084177017212,
      "learning_rate": 1e-05,
      "loss": 0.2416,
      "step": 24010
    },
    {
      "epoch": 0.08230932130790265,
      "grad_norm": 1.10452401638031,
      "learning_rate": 1e-05,
      "loss": 0.2486,
      "step": 24020
    },
    {
      "epoch": 0.08234358830261869,
      "grad_norm": 1.1392238140106201,
      "learning_rate": 1e-05,
      "loss": 0.2594,
      "step": 24030
    },
    {
      "epoch": 0.08237785529733471,
      "grad_norm": 1.1955091953277588,
      "learning_rate": 1e-05,
      "loss": 0.2607,
      "step": 24040
    },
    {
      "epoch": 0.08241212229205075,
      "grad_norm": 0.9922770857810974,
      "learning_rate": 1e-05,
      "loss": 0.2393,
      "step": 24050
    },
    {
      "epoch": 0.08244638928676677,
      "grad_norm": 1.0853592157363892,
      "learning_rate": 1e-05,
      "loss": 0.253,
      "step": 24060
    },
    {
      "epoch": 0.0824806562814828,
      "grad_norm": 1.1436183452606201,
      "learning_rate": 1e-05,
      "loss": 0.2404,
      "step": 24070
    },
    {
      "epoch": 0.08251492327619883,
      "grad_norm": 1.0038506984710693,
      "learning_rate": 1e-05,
      "loss": 0.2452,
      "step": 24080
    },
    {
      "epoch": 0.08254919027091485,
      "grad_norm": 1.1340346336364746,
      "learning_rate": 1e-05,
      "loss": 0.2395,
      "step": 24090
    },
    {
      "epoch": 0.08258345726563089,
      "grad_norm": 1.2182914018630981,
      "learning_rate": 1e-05,
      "loss": 0.2409,
      "step": 24100
    },
    {
      "epoch": 0.08261772426034691,
      "grad_norm": 1.1426423788070679,
      "learning_rate": 1e-05,
      "loss": 0.2406,
      "step": 24110
    },
    {
      "epoch": 0.08265199125506295,
      "grad_norm": 0.9850037097930908,
      "learning_rate": 1e-05,
      "loss": 0.2216,
      "step": 24120
    },
    {
      "epoch": 0.08268625824977897,
      "grad_norm": 1.177119255065918,
      "learning_rate": 1e-05,
      "loss": 0.2389,
      "step": 24130
    },
    {
      "epoch": 0.08272052524449501,
      "grad_norm": 1.1432584524154663,
      "learning_rate": 1e-05,
      "loss": 0.2545,
      "step": 24140
    },
    {
      "epoch": 0.08275479223921103,
      "grad_norm": 1.0530070066452026,
      "learning_rate": 1e-05,
      "loss": 0.2511,
      "step": 24150
    },
    {
      "epoch": 0.08278905923392707,
      "grad_norm": 1.0670721530914307,
      "learning_rate": 1e-05,
      "loss": 0.2393,
      "step": 24160
    },
    {
      "epoch": 0.0828233262286431,
      "grad_norm": 1.0654528141021729,
      "learning_rate": 1e-05,
      "loss": 0.2453,
      "step": 24170
    },
    {
      "epoch": 0.08285759322335913,
      "grad_norm": 1.0833876132965088,
      "learning_rate": 1e-05,
      "loss": 0.2427,
      "step": 24180
    },
    {
      "epoch": 0.08289186021807515,
      "grad_norm": 1.203636646270752,
      "learning_rate": 1e-05,
      "loss": 0.2445,
      "step": 24190
    },
    {
      "epoch": 0.08292612721279119,
      "grad_norm": 1.0569822788238525,
      "learning_rate": 1e-05,
      "loss": 0.2527,
      "step": 24200
    },
    {
      "epoch": 0.08296039420750721,
      "grad_norm": 1.0819365978240967,
      "learning_rate": 1e-05,
      "loss": 0.2491,
      "step": 24210
    },
    {
      "epoch": 0.08299466120222325,
      "grad_norm": 0.956221878528595,
      "learning_rate": 1e-05,
      "loss": 0.2411,
      "step": 24220
    },
    {
      "epoch": 0.08302892819693927,
      "grad_norm": 1.1548758745193481,
      "learning_rate": 1e-05,
      "loss": 0.2466,
      "step": 24230
    },
    {
      "epoch": 0.0830631951916553,
      "grad_norm": 0.9903671741485596,
      "learning_rate": 1e-05,
      "loss": 0.2483,
      "step": 24240
    },
    {
      "epoch": 0.08309746218637133,
      "grad_norm": 1.0387729406356812,
      "learning_rate": 1e-05,
      "loss": 0.2372,
      "step": 24250
    },
    {
      "epoch": 0.08313172918108735,
      "grad_norm": 1.15946626663208,
      "learning_rate": 1e-05,
      "loss": 0.2499,
      "step": 24260
    },
    {
      "epoch": 0.08316599617580339,
      "grad_norm": 1.1431846618652344,
      "learning_rate": 1e-05,
      "loss": 0.2379,
      "step": 24270
    },
    {
      "epoch": 0.08320026317051941,
      "grad_norm": 1.1871981620788574,
      "learning_rate": 1e-05,
      "loss": 0.2366,
      "step": 24280
    },
    {
      "epoch": 0.08323453016523545,
      "grad_norm": 1.285339117050171,
      "learning_rate": 1e-05,
      "loss": 0.2408,
      "step": 24290
    },
    {
      "epoch": 0.08326879715995147,
      "grad_norm": 0.9946316480636597,
      "learning_rate": 1e-05,
      "loss": 0.2258,
      "step": 24300
    },
    {
      "epoch": 0.08330306415466751,
      "grad_norm": 1.0830858945846558,
      "learning_rate": 1e-05,
      "loss": 0.243,
      "step": 24310
    },
    {
      "epoch": 0.08333733114938353,
      "grad_norm": 1.1845496892929077,
      "learning_rate": 1e-05,
      "loss": 0.2384,
      "step": 24320
    },
    {
      "epoch": 0.08337159814409957,
      "grad_norm": 0.9513915777206421,
      "learning_rate": 1e-05,
      "loss": 0.247,
      "step": 24330
    },
    {
      "epoch": 0.0834058651388156,
      "grad_norm": 1.101516604423523,
      "learning_rate": 1e-05,
      "loss": 0.2642,
      "step": 24340
    },
    {
      "epoch": 0.08344013213353163,
      "grad_norm": 1.2533254623413086,
      "learning_rate": 1e-05,
      "loss": 0.2462,
      "step": 24350
    },
    {
      "epoch": 0.08347439912824765,
      "grad_norm": 1.1113494634628296,
      "learning_rate": 1e-05,
      "loss": 0.2542,
      "step": 24360
    },
    {
      "epoch": 0.08350866612296369,
      "grad_norm": 1.0338751077651978,
      "learning_rate": 1e-05,
      "loss": 0.2364,
      "step": 24370
    },
    {
      "epoch": 0.08354293311767971,
      "grad_norm": 1.3761131763458252,
      "learning_rate": 1e-05,
      "loss": 0.2484,
      "step": 24380
    },
    {
      "epoch": 0.08357720011239574,
      "grad_norm": 1.1954588890075684,
      "learning_rate": 1e-05,
      "loss": 0.2618,
      "step": 24390
    },
    {
      "epoch": 0.08361146710711177,
      "grad_norm": 1.1829007863998413,
      "learning_rate": 1e-05,
      "loss": 0.2477,
      "step": 24400
    },
    {
      "epoch": 0.0836457341018278,
      "grad_norm": 1.0606454610824585,
      "learning_rate": 1e-05,
      "loss": 0.2591,
      "step": 24410
    },
    {
      "epoch": 0.08368000109654383,
      "grad_norm": 1.1648555994033813,
      "learning_rate": 1e-05,
      "loss": 0.2381,
      "step": 24420
    },
    {
      "epoch": 0.08371426809125986,
      "grad_norm": 1.1213902235031128,
      "learning_rate": 1e-05,
      "loss": 0.2335,
      "step": 24430
    },
    {
      "epoch": 0.08374853508597589,
      "grad_norm": 1.0329564809799194,
      "learning_rate": 1e-05,
      "loss": 0.2505,
      "step": 24440
    },
    {
      "epoch": 0.08378280208069191,
      "grad_norm": 1.0963842868804932,
      "learning_rate": 1e-05,
      "loss": 0.2465,
      "step": 24450
    },
    {
      "epoch": 0.08381706907540795,
      "grad_norm": 1.1725986003875732,
      "learning_rate": 1e-05,
      "loss": 0.2599,
      "step": 24460
    },
    {
      "epoch": 0.08385133607012397,
      "grad_norm": 1.1225870847702026,
      "learning_rate": 1e-05,
      "loss": 0.2319,
      "step": 24470
    },
    {
      "epoch": 0.08388560306484001,
      "grad_norm": 1.0629470348358154,
      "learning_rate": 1e-05,
      "loss": 0.2599,
      "step": 24480
    },
    {
      "epoch": 0.08391987005955603,
      "grad_norm": 1.012823462486267,
      "learning_rate": 1e-05,
      "loss": 0.2225,
      "step": 24490
    },
    {
      "epoch": 0.08395413705427207,
      "grad_norm": 1.0204377174377441,
      "learning_rate": 1e-05,
      "loss": 0.2512,
      "step": 24500
    },
    {
      "epoch": 0.08395413705427207,
      "eval_cer": 12.515434820956076,
      "eval_loss": 0.2421187460422516,
      "eval_normalized_cer": 8.792965627498003,
      "eval_runtime": 226.7851,
      "eval_samples_per_second": 2.258,
      "eval_steps_per_second": 0.035,
      "step": 24500
    }
  ],
  "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": 4.96759625220096e+21,
  "train_batch_size": 128,
  "trial_name": null,
  "trial_params": null
}