{
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 39.95179987797437,
  "eval_steps": 100.0,
  "global_step": 32760,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.6101281269066504,
      "grad_norm": 109.53021240234375,
      "learning_rate": 1.188e-06,
      "loss": 53.9873,
      "step": 500
    },
    {
      "epoch": 1.0,
      "eval_cer": 1.1283783783783783,
      "eval_loss": 19.713674545288086,
      "eval_runtime": 142.3027,
      "eval_samples_per_second": 48.067,
      "eval_steps_per_second": 6.008,
      "eval_wer": 1.0,
      "step": 820
    },
    {
      "epoch": 1.2196461256863942,
      "grad_norm": 96.87247467041016,
      "learning_rate": 2.3880000000000003e-06,
      "loss": 20.2826,
      "step": 1000
    },
    {
      "epoch": 1.8297742525930445,
      "grad_norm": 75.96553802490234,
      "learning_rate": 3.588e-06,
      "loss": 16.2788,
      "step": 1500
    },
    {
      "epoch": 2.0,
      "eval_cer": 1.1284080132764343,
      "eval_loss": 12.832098960876465,
      "eval_runtime": 145.6125,
      "eval_samples_per_second": 46.974,
      "eval_steps_per_second": 5.872,
      "eval_wer": 1.0,
      "step": 1640
    },
    {
      "epoch": 2.4392922513727884,
      "grad_norm": 39.1208381652832,
      "learning_rate": 4.788e-06,
      "loss": 11.6832,
      "step": 2000
    },
    {
      "epoch": 3.0,
      "eval_cer": 1.1284080132764343,
      "eval_loss": 5.618701934814453,
      "eval_runtime": 142.6056,
      "eval_samples_per_second": 47.964,
      "eval_steps_per_second": 5.996,
      "eval_wer": 1.0,
      "step": 2460
    },
    {
      "epoch": 3.048810250152532,
      "grad_norm": 27.440048217773438,
      "learning_rate": 5.988e-06,
      "loss": 7.1726,
      "step": 2500
    },
    {
      "epoch": 3.6589383770591826,
      "grad_norm": 5.054373264312744,
      "learning_rate": 7.1880000000000005e-06,
      "loss": 4.3059,
      "step": 3000
    },
    {
      "epoch": 4.0,
      "eval_cer": 1.1284080132764343,
      "eval_loss": 3.3938002586364746,
      "eval_runtime": 143.6745,
      "eval_samples_per_second": 47.608,
      "eval_steps_per_second": 5.951,
      "eval_wer": 1.0,
      "step": 3280
    },
    {
      "epoch": 4.268456375838926,
      "grad_norm": 1.1531486511230469,
      "learning_rate": 8.388e-06,
      "loss": 3.4228,
      "step": 3500
    },
    {
      "epoch": 4.878584502745577,
      "grad_norm": 1.089985966682434,
      "learning_rate": 9.588e-06,
      "loss": 3.1294,
      "step": 4000
    },
    {
      "epoch": 5.0,
      "eval_cer": 1.1284080132764343,
      "eval_loss": 2.8853955268859863,
      "eval_runtime": 142.5876,
      "eval_samples_per_second": 47.971,
      "eval_steps_per_second": 5.996,
      "eval_wer": 1.0,
      "step": 4100
    },
    {
      "epoch": 5.48810250152532,
      "grad_norm": 2.860553503036499,
      "learning_rate": 1.0787999999999999e-05,
      "loss": 2.6448,
      "step": 4500
    },
    {
      "epoch": 6.0,
      "eval_cer": 1.1018551446183025,
      "eval_loss": 1.4072996377944946,
      "eval_runtime": 143.5621,
      "eval_samples_per_second": 47.645,
      "eval_steps_per_second": 5.956,
      "eval_wer": 1.0,
      "step": 4920
    },
    {
      "epoch": 6.097620500305064,
      "grad_norm": 4.616542816162109,
      "learning_rate": 1.1988000000000001e-05,
      "loss": 1.7348,
      "step": 5000
    },
    {
      "epoch": 6.707748627211714,
      "grad_norm": 4.588995933532715,
      "learning_rate": 1.3188e-05,
      "loss": 1.022,
      "step": 5500
    },
    {
      "epoch": 7.0,
      "eval_cer": 1.033576339497392,
      "eval_loss": 0.6322615146636963,
      "eval_runtime": 146.5646,
      "eval_samples_per_second": 46.669,
      "eval_steps_per_second": 5.834,
      "eval_wer": 1.0,
      "step": 5740
    },
    {
      "epoch": 7.317266625991458,
      "grad_norm": 3.685751438140869,
      "learning_rate": 1.4388000000000002e-05,
      "loss": 0.6784,
      "step": 6000
    },
    {
      "epoch": 7.927394752898109,
      "grad_norm": 5.702773094177246,
      "learning_rate": 1.5588e-05,
      "loss": 0.4449,
      "step": 6500
    },
    {
      "epoch": 8.0,
      "eval_cer": 1.0403330962541488,
      "eval_loss": 0.3109678328037262,
      "eval_runtime": 145.6018,
      "eval_samples_per_second": 46.977,
      "eval_steps_per_second": 5.872,
      "eval_wer": 0.9988304093567252,
      "step": 6560
    },
    {
      "epoch": 8.536912751677852,
      "grad_norm": 7.984670162200928,
      "learning_rate": 1.6788e-05,
      "loss": 0.305,
      "step": 7000
    },
    {
      "epoch": 9.0,
      "eval_cer": 1.0484826932195352,
      "eval_loss": 0.23479318618774414,
      "eval_runtime": 140.3606,
      "eval_samples_per_second": 48.732,
      "eval_steps_per_second": 6.091,
      "eval_wer": 0.9991228070175439,
      "step": 7380
    },
    {
      "epoch": 9.146430750457595,
      "grad_norm": 2.9451138973236084,
      "learning_rate": 1.7988e-05,
      "loss": 0.2291,
      "step": 7500
    },
    {
      "epoch": 9.756558877364247,
      "grad_norm": 8.89615535736084,
      "learning_rate": 1.9188e-05,
      "loss": 0.1895,
      "step": 8000
    },
    {
      "epoch": 10.0,
      "eval_cer": 1.0229670459933617,
      "eval_loss": 0.11090246587991714,
      "eval_runtime": 134.2378,
      "eval_samples_per_second": 50.954,
      "eval_steps_per_second": 6.369,
      "eval_wer": 0.9988304093567252,
      "step": 8200
    },
    {
      "epoch": 10.36607687614399,
      "grad_norm": 5.290111541748047,
      "learning_rate": 2.0388e-05,
      "loss": 0.1634,
      "step": 8500
    },
    {
      "epoch": 10.97620500305064,
      "grad_norm": 6.373618125915527,
      "learning_rate": 2.1588e-05,
      "loss": 0.1437,
      "step": 9000
    },
    {
      "epoch": 11.0,
      "eval_cer": 1.0221076339497392,
      "eval_loss": 0.09310559928417206,
      "eval_runtime": 142.4966,
      "eval_samples_per_second": 48.001,
      "eval_steps_per_second": 6.0,
      "eval_wer": 0.9989766081871345,
      "step": 9020
    },
    {
      "epoch": 11.585723001830385,
      "grad_norm": 7.480563640594482,
      "learning_rate": 2.2788000000000003e-05,
      "loss": 0.1258,
      "step": 9500
    },
    {
      "epoch": 12.0,
      "eval_cer": 1.0258119962067331,
      "eval_loss": 0.0827784314751625,
      "eval_runtime": 143.9999,
      "eval_samples_per_second": 47.5,
      "eval_steps_per_second": 5.938,
      "eval_wer": 0.9988304093567252,
      "step": 9840
    },
    {
      "epoch": 12.195241000610128,
      "grad_norm": 8.04295825958252,
      "learning_rate": 2.3988e-05,
      "loss": 0.1275,
      "step": 10000
    },
    {
      "epoch": 12.805369127516778,
      "grad_norm": 1.162207007408142,
      "learning_rate": 2.5188e-05,
      "loss": 0.1175,
      "step": 10500
    },
    {
      "epoch": 13.0,
      "eval_cer": 1.0232337600758654,
      "eval_loss": 0.08141058683395386,
      "eval_runtime": 145.6362,
      "eval_samples_per_second": 46.966,
      "eval_steps_per_second": 5.871,
      "eval_wer": 0.9991228070175439,
      "step": 10660
    },
    {
      "epoch": 13.414887126296522,
      "grad_norm": 3.6727077960968018,
      "learning_rate": 2.6388000000000002e-05,
      "loss": 0.1083,
      "step": 11000
    },
    {
      "epoch": 14.0,
      "eval_cer": 1.0193515884305357,
      "eval_loss": 0.04150446131825447,
      "eval_runtime": 145.6958,
      "eval_samples_per_second": 46.947,
      "eval_steps_per_second": 5.868,
      "eval_wer": 0.9988304093567252,
      "step": 11480
    },
    {
      "epoch": 14.024405125076266,
      "grad_norm": 0.8475801348686218,
      "learning_rate": 2.7585600000000002e-05,
      "loss": 0.1029,
      "step": 11500
    },
    {
      "epoch": 14.634533251982916,
      "grad_norm": 2.529123067855835,
      "learning_rate": 2.87856e-05,
      "loss": 0.0974,
      "step": 12000
    },
    {
      "epoch": 15.0,
      "eval_cer": 1.0239153627311521,
      "eval_loss": 0.0653274655342102,
      "eval_runtime": 140.8698,
      "eval_samples_per_second": 48.555,
      "eval_steps_per_second": 6.069,
      "eval_wer": 0.9989766081871345,
      "step": 12300
    },
    {
      "epoch": 15.24405125076266,
      "grad_norm": 4.318971157073975,
      "learning_rate": 2.99856e-05,
      "loss": 0.1027,
      "step": 12500
    },
    {
      "epoch": 15.854179377669311,
      "grad_norm": 5.327524662017822,
      "learning_rate": 2.9956013070043084e-05,
      "loss": 0.0967,
      "step": 13000
    },
    {
      "epoch": 16.0,
      "eval_cer": 1.0200331910858227,
      "eval_loss": 0.04946194589138031,
      "eval_runtime": 129.206,
      "eval_samples_per_second": 52.939,
      "eval_steps_per_second": 6.617,
      "eval_wer": 0.9991228070175439,
      "step": 13120
    },
    {
      "epoch": 16.463697376449055,
      "grad_norm": 3.569943904876709,
      "learning_rate": 2.9822174136311704e-05,
      "loss": 0.087,
      "step": 13500
    },
    {
      "epoch": 17.0,
      "eval_cer": 1.0223743480322427,
      "eval_loss": 0.06013474613428116,
      "eval_runtime": 137.3873,
      "eval_samples_per_second": 49.786,
      "eval_steps_per_second": 6.223,
      "eval_wer": 0.9989766081871345,
      "step": 13940
    },
    {
      "epoch": 17.0732153752288,
      "grad_norm": 6.857547283172607,
      "learning_rate": 2.9599814696946643e-05,
      "loss": 0.0885,
      "step": 14000
    },
    {
      "epoch": 17.683343502135447,
      "grad_norm": 0.8758929371833801,
      "learning_rate": 2.9289379955813937e-05,
      "loss": 0.0798,
      "step": 14500
    },
    {
      "epoch": 18.0,
      "eval_cer": 1.0217816500711236,
      "eval_loss": 0.054376766085624695,
      "eval_runtime": 139.8471,
      "eval_samples_per_second": 48.911,
      "eval_steps_per_second": 6.114,
      "eval_wer": 0.9989766081871345,
      "step": 14760
    },
    {
      "epoch": 18.29286150091519,
      "grad_norm": 3.2238011360168457,
      "learning_rate": 2.8893968733503523e-05,
      "loss": 0.0765,
      "step": 15000
    },
    {
      "epoch": 18.902989627821842,
      "grad_norm": 5.363631725311279,
      "learning_rate": 2.84143727148899e-05,
      "loss": 0.0719,
      "step": 15500
    },
    {
      "epoch": 19.0,
      "eval_cer": 1.0190848743480323,
      "eval_loss": 0.04263332858681679,
      "eval_runtime": 125.432,
      "eval_samples_per_second": 54.532,
      "eval_steps_per_second": 6.816,
      "eval_wer": 0.9989766081871345,
      "step": 15580
    },
    {
      "epoch": 19.512507626601586,
      "grad_norm": 0.8454777002334595,
      "learning_rate": 2.785418066112353e-05,
      "loss": 0.0731,
      "step": 16000
    },
    {
      "epoch": 20.0,
      "eval_cer": 1.0208333333333333,
      "eval_loss": 0.058695003390312195,
      "eval_runtime": 139.9224,
      "eval_samples_per_second": 48.884,
      "eval_steps_per_second": 6.111,
      "eval_wer": 0.9991228070175439,
      "step": 16400
    },
    {
      "epoch": 20.12202562538133,
      "grad_norm": 1.441945195198059,
      "learning_rate": 2.7216758309791792e-05,
      "loss": 0.0686,
      "step": 16500
    },
    {
      "epoch": 20.73215375228798,
      "grad_norm": 1.6329078674316406,
      "learning_rate": 2.6505935412410244e-05,
      "loss": 0.0693,
      "step": 17000
    },
    {
      "epoch": 21.0,
      "eval_cer": 1.0221965386439071,
      "eval_loss": 0.060316070914268494,
      "eval_runtime": 126.6576,
      "eval_samples_per_second": 54.004,
      "eval_steps_per_second": 6.75,
      "eval_wer": 0.9988304093567252,
      "step": 17220
    },
    {
      "epoch": 21.341671751067725,
      "grad_norm": 2.643927574157715,
      "learning_rate": 2.5725982724566367e-05,
      "loss": 0.0665,
      "step": 17500
    },
    {
      "epoch": 21.951799877974373,
      "grad_norm": 6.064180850982666,
      "learning_rate": 2.4881586346429215e-05,
      "loss": 0.0614,
      "step": 18000
    },
    {
      "epoch": 22.0,
      "eval_cer": 1.0191145092460883,
      "eval_loss": 0.03614737093448639,
      "eval_runtime": 143.2008,
      "eval_samples_per_second": 47.765,
      "eval_steps_per_second": 5.971,
      "eval_wer": 0.9988304093567252,
      "step": 18040
    },
    {
      "epoch": 22.561317876754117,
      "grad_norm": 0.2802174389362335,
      "learning_rate": 2.3977819567791885e-05,
      "loss": 0.0582,
      "step": 18500
    },
    {
      "epoch": 23.0,
      "eval_cer": 1.0173364153627311,
      "eval_loss": 0.03316638618707657,
      "eval_runtime": 137.4557,
      "eval_samples_per_second": 49.761,
      "eval_steps_per_second": 6.22,
      "eval_wer": 0.9988304093567252,
      "step": 18860
    },
    {
      "epoch": 23.17083587553386,
      "grad_norm": 4.600666522979736,
      "learning_rate": 2.302011238680703e-05,
      "loss": 0.0614,
      "step": 19000
    },
    {
      "epoch": 23.780964002440513,
      "grad_norm": 3.028637647628784,
      "learning_rate": 2.2014218885552525e-05,
      "loss": 0.0535,
      "step": 19500
    },
    {
      "epoch": 24.0,
      "eval_cer": 1.0172475106685632,
      "eval_loss": 0.03474853187799454,
      "eval_runtime": 142.7947,
      "eval_samples_per_second": 47.901,
      "eval_steps_per_second": 5.988,
      "eval_wer": 0.9988304093567252,
      "step": 19680
    },
    {
      "epoch": 24.390482001220256,
      "grad_norm": 0.652704656124115,
      "learning_rate": 2.096618265844089e-05,
      "loss": 0.0509,
      "step": 20000
    },
    {
      "epoch": 25.0,
      "grad_norm": 1.3256402015686035,
      "learning_rate": 1.988230050118496e-05,
      "loss": 0.0467,
      "step": 20500
    },
    {
      "epoch": 25.0,
      "eval_cer": 1.0179587482219061,
      "eval_loss": 0.033395446836948395,
      "eval_runtime": 143.6182,
      "eval_samples_per_second": 47.626,
      "eval_steps_per_second": 5.953,
      "eval_wer": 0.9988304093567252,
      "step": 20500
    },
    {
      "epoch": 25.610128126906652,
      "grad_norm": 2.4054596424102783,
      "learning_rate": 1.876908457848333e-05,
      "loss": 0.0456,
      "step": 21000
    },
    {
      "epoch": 26.0,
      "eval_cer": 1.0163880986249407,
      "eval_loss": 0.028318189084529877,
      "eval_runtime": 136.1666,
      "eval_samples_per_second": 50.233,
      "eval_steps_per_second": 6.279,
      "eval_wer": 0.9988304093567252,
      "step": 21320
    },
    {
      "epoch": 26.219646125686396,
      "grad_norm": 2.3439478874206543,
      "learning_rate": 1.7635513102937044e-05,
      "loss": 0.041,
      "step": 21500
    },
    {
      "epoch": 26.829774252593044,
      "grad_norm": 4.5279974937438965,
      "learning_rate": 1.6486170220352805e-05,
      "loss": 0.0389,
      "step": 22000
    },
    {
      "epoch": 27.0,
      "eval_cer": 1.0172178757705073,
      "eval_loss": 0.03608373552560806,
      "eval_runtime": 133.2593,
      "eval_samples_per_second": 51.329,
      "eval_steps_per_second": 6.416,
      "eval_wer": 0.9988304093567252,
      "step": 22140
    },
    {
      "epoch": 27.439292251372787,
      "grad_norm": 0.17273186147212982,
      "learning_rate": 1.5325608410059234e-05,
      "loss": 0.04,
      "step": 22500
    },
    {
      "epoch": 28.0,
      "eval_cer": 1.016714082503556,
      "eval_loss": 0.025806378573179245,
      "eval_runtime": 139.0344,
      "eval_samples_per_second": 49.196,
      "eval_steps_per_second": 6.15,
      "eval_wer": 0.9988304093567252,
      "step": 22960
    },
    {
      "epoch": 28.04881025015253,
      "grad_norm": 0.3772958219051361,
      "learning_rate": 1.4163090284146517e-05,
      "loss": 0.0383,
      "step": 23000
    },
    {
      "epoch": 28.658938377059183,
      "grad_norm": 2.2773993015289307,
      "learning_rate": 1.3005600466773616e-05,
      "loss": 0.0348,
      "step": 23500
    },
    {
      "epoch": 29.0,
      "eval_cer": 1.0176031294452348,
      "eval_loss": 0.03283878415822983,
      "eval_runtime": 134.9976,
      "eval_samples_per_second": 50.668,
      "eval_steps_per_second": 6.333,
      "eval_wer": 0.9989766081871345,
      "step": 23780
    },
    {
      "epoch": 29.268456375838927,
      "grad_norm": 0.017471132799983025,
      "learning_rate": 1.186009337109073e-05,
      "loss": 0.0356,
      "step": 24000
    },
    {
      "epoch": 29.878584502745575,
      "grad_norm": 0.18886829912662506,
      "learning_rate": 1.0733451415837331e-05,
      "loss": 0.0343,
      "step": 24500
    },
    {
      "epoch": 30.0,
      "eval_cer": 1.016239924134661,
      "eval_loss": 0.027611197903752327,
      "eval_runtime": 136.0309,
      "eval_samples_per_second": 50.283,
      "eval_steps_per_second": 6.285,
      "eval_wer": 0.9988304093567252,
      "step": 24600
    },
    {
      "epoch": 30.48810250152532,
      "grad_norm": 0.37121689319610596,
      "learning_rate": 9.634615680568962e-06,
      "loss": 0.0323,
      "step": 25000
    },
    {
      "epoch": 31.0,
      "eval_cer": 1.0165066382171646,
      "eval_loss": 0.029725994914770126,
      "eval_runtime": 136.5772,
      "eval_samples_per_second": 50.082,
      "eval_steps_per_second": 6.26,
      "eval_wer": 0.9988304093567252,
      "step": 25420
    },
    {
      "epoch": 31.097620500305062,
      "grad_norm": 0.04804076626896858,
      "learning_rate": 8.56578623342252e-06,
      "loss": 0.0298,
      "step": 25500
    },
    {
      "epoch": 31.707748627211714,
      "grad_norm": 3.0204851627349854,
      "learning_rate": 7.535614733981355e-06,
      "loss": 0.0283,
      "step": 26000
    },
    {
      "epoch": 32.0,
      "eval_cer": 1.0165362731152205,
      "eval_loss": 0.029090913012623787,
      "eval_runtime": 146.0361,
      "eval_samples_per_second": 46.838,
      "eval_steps_per_second": 5.855,
      "eval_wer": 0.9988304093567252,
      "step": 26240
    },
    {
      "epoch": 32.31726662599146,
      "grad_norm": 0.10753169655799866,
      "learning_rate": 6.550290643366546e-06,
      "loss": 0.0297,
      "step": 26500
    },
    {
      "epoch": 32.92739475289811,
      "grad_norm": 0.41164013743400574,
      "learning_rate": 5.615733971162722e-06,
      "loss": 0.0275,
      "step": 27000
    },
    {
      "epoch": 33.0,
      "eval_cer": 1.0160917496443813,
      "eval_loss": 0.0251940805464983,
      "eval_runtime": 140.9294,
      "eval_samples_per_second": 48.535,
      "eval_steps_per_second": 6.067,
      "eval_wer": 0.9988304093567252,
      "step": 27060
    },
    {
      "epoch": 33.53691275167785,
      "grad_norm": 0.8919934630393982,
      "learning_rate": 4.737559706904321e-06,
      "loss": 0.0256,
      "step": 27500
    },
    {
      "epoch": 34.0,
      "eval_cer": 1.0164473684210527,
      "eval_loss": 0.024480827152729034,
      "eval_runtime": 145.8102,
      "eval_samples_per_second": 46.91,
      "eval_steps_per_second": 5.864,
      "eval_wer": 0.9988304093567252,
      "step": 27880
    },
    {
      "epoch": 34.1464307504576,
      "grad_norm": 0.016134992241859436,
      "learning_rate": 3.921044084178765e-06,
      "loss": 0.0265,
      "step": 28000
    },
    {
      "epoch": 34.75655887736425,
      "grad_norm": 0.048243045806884766,
      "learning_rate": 3.171092880038068e-06,
      "loss": 0.0241,
      "step": 28500
    },
    {
      "epoch": 35.0,
      "eval_cer": 1.0159139402560455,
      "eval_loss": 0.023953931406140327,
      "eval_runtime": 141.1272,
      "eval_samples_per_second": 48.467,
      "eval_steps_per_second": 6.058,
      "eval_wer": 0.9988304093567252,
      "step": 28700
    },
    {
      "epoch": 35.36607687614399,
      "grad_norm": 0.09765101969242096,
      "learning_rate": 2.4922119401819466e-06,
      "loss": 0.0243,
      "step": 29000
    },
    {
      "epoch": 35.97620500305064,
      "grad_norm": 1.2347666025161743,
      "learning_rate": 1.8896100834437107e-06,
      "loss": 0.0237,
      "step": 29500
    },
    {
      "epoch": 36.0,
      "eval_cer": 1.0165659080132765,
      "eval_loss": 0.0277857668697834,
      "eval_runtime": 141.5894,
      "eval_samples_per_second": 48.309,
      "eval_steps_per_second": 6.039,
      "eval_wer": 0.9988304093567252,
      "step": 29520
    },
    {
      "epoch": 36.58572300183038,
      "grad_norm": 2.3807311058044434,
      "learning_rate": 1.3654632704576153e-06,
      "loss": 0.0238,
      "step": 30000
    },
    {
      "epoch": 37.0,
      "eval_cer": 1.0162991939307728,
      "eval_loss": 0.027532441541552544,
      "eval_runtime": 145.7443,
      "eval_samples_per_second": 46.932,
      "eval_steps_per_second": 5.866,
      "eval_wer": 0.9988304093567252,
      "step": 30340
    },
    {
      "epoch": 37.195241000610125,
      "grad_norm": 1.2338793277740479,
      "learning_rate": 9.221049871134596e-07,
      "loss": 0.0237,
      "step": 30500
    },
    {
      "epoch": 37.80536912751678,
      "grad_norm": 0.02291560173034668,
      "learning_rate": 5.633293037499832e-07,
      "loss": 0.022,
      "step": 31000
    },
    {
      "epoch": 38.0,
      "eval_cer": 1.0162991939307728,
      "eval_loss": 0.024687692523002625,
      "eval_runtime": 136.7797,
      "eval_samples_per_second": 50.007,
      "eval_steps_per_second": 6.251,
      "eval_wer": 0.9988304093567252,
      "step": 31160
    },
    {
      "epoch": 38.41488712629652,
      "grad_norm": 1.2824342250823975,
      "learning_rate": 2.912918111057888e-07,
      "loss": 0.0184,
      "step": 31500
    },
    {
      "epoch": 39.0,
      "eval_cer": 1.0162695590327169,
      "eval_loss": 0.02619771659374237,
      "eval_runtime": 143.9756,
      "eval_samples_per_second": 47.508,
      "eval_steps_per_second": 5.939,
      "eval_wer": 0.9988304093567252,
      "step": 31980
    },
    {
      "epoch": 39.024405125076264,
      "grad_norm": 0.024750633165240288,
      "learning_rate": 1.0762696080869105e-07,
      "loss": 0.0228,
      "step": 32000
    },
    {
      "epoch": 39.634533251982916,
      "grad_norm": 0.007463211193680763,
      "learning_rate": 1.3438245287707985e-08,
      "loss": 0.0199,
      "step": 32500
    },
    {
      "epoch": 39.95179987797437,
      "eval_cer": 1.0160324798482694,
      "eval_loss": 0.024437865242362022,
      "eval_runtime": 142.7998,
      "eval_samples_per_second": 47.899,
      "eval_steps_per_second": 5.987,
      "eval_wer": 0.9988304093567252,
      "step": 32760
    },
    {
      "epoch": 39.95179987797437,
      "step": 32760,
      "total_flos": 3.0824437935703523e+19,
      "train_loss": 1.9907507799775026,
      "train_runtime": 53433.8738,
      "train_samples_per_second": 39.256,
      "train_steps_per_second": 0.613
    }
  ],
  "logging_steps": 500,
  "max_steps": 32760,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 40,
  "save_steps": 400,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 3.0824437935703523e+19,
  "train_batch_size": 32,
  "trial_name": null,
  "trial_params": null
}