{
  "best_global_step": 2000,
  "best_metric": 32.74580361567799,
  "best_model_checkpoint": "./HAUSA_B/checkpoint-2000",
  "epoch": 2.706838185511171,
  "eval_steps": 200,
  "global_step": 2000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.027081922816519974,
      "grad_norm": 106.59215545654297,
      "learning_rate": 8.000000000000001e-07,
      "loss": 6.1915,
      "step": 20
    },
    {
      "epoch": 0.05416384563303995,
      "grad_norm": 30.563518524169922,
      "learning_rate": 1.8000000000000001e-06,
      "loss": 4.5886,
      "step": 40
    },
    {
      "epoch": 0.08124576844955991,
      "grad_norm": 16.528457641601562,
      "learning_rate": 2.8000000000000003e-06,
      "loss": 3.0338,
      "step": 60
    },
    {
      "epoch": 0.1083276912660799,
      "grad_norm": 13.852096557617188,
      "learning_rate": 3.8000000000000005e-06,
      "loss": 2.3188,
      "step": 80
    },
    {
      "epoch": 0.13540961408259986,
      "grad_norm": 13.284646987915039,
      "learning_rate": 4.800000000000001e-06,
      "loss": 1.9725,
      "step": 100
    },
    {
      "epoch": 0.16249153689911983,
      "grad_norm": 13.212055206298828,
      "learning_rate": 5.8e-06,
      "loss": 1.7839,
      "step": 120
    },
    {
      "epoch": 0.1895734597156398,
      "grad_norm": 12.006990432739258,
      "learning_rate": 6.800000000000001e-06,
      "loss": 1.5991,
      "step": 140
    },
    {
      "epoch": 0.2166553825321598,
      "grad_norm": 12.490514755249023,
      "learning_rate": 7.800000000000002e-06,
      "loss": 1.4411,
      "step": 160
    },
    {
      "epoch": 0.24373730534867977,
      "grad_norm": 11.587646484375,
      "learning_rate": 8.8e-06,
      "loss": 1.3487,
      "step": 180
    },
    {
      "epoch": 0.2708192281651997,
      "grad_norm": 11.574933052062988,
      "learning_rate": 9.800000000000001e-06,
      "loss": 1.2119,
      "step": 200
    },
    {
      "epoch": 0.2708192281651997,
      "eval_loss": 1.1157827377319336,
      "eval_runtime": 1601.2991,
      "eval_samples_per_second": 3.688,
      "eval_steps_per_second": 0.462,
      "eval_wer": 64.22461160550755,
      "step": 200
    },
    {
      "epoch": 0.2979011509817197,
      "grad_norm": 8.368565559387207,
      "learning_rate": 9.920556107249256e-06,
      "loss": 0.9452,
      "step": 220
    },
    {
      "epoch": 0.32498307379823965,
      "grad_norm": 7.2616682052612305,
      "learning_rate": 9.821251241310825e-06,
      "loss": 0.9298,
      "step": 240
    },
    {
      "epoch": 0.35206499661475965,
      "grad_norm": 7.266394138336182,
      "learning_rate": 9.721946375372395e-06,
      "loss": 0.8567,
      "step": 260
    },
    {
      "epoch": 0.3791469194312796,
      "grad_norm": 8.120067596435547,
      "learning_rate": 9.622641509433963e-06,
      "loss": 0.8251,
      "step": 280
    },
    {
      "epoch": 0.4062288422477996,
      "grad_norm": 7.483547687530518,
      "learning_rate": 9.523336643495532e-06,
      "loss": 0.7774,
      "step": 300
    },
    {
      "epoch": 0.4333107650643196,
      "grad_norm": 7.521896839141846,
      "learning_rate": 9.4240317775571e-06,
      "loss": 0.7266,
      "step": 320
    },
    {
      "epoch": 0.46039268788083954,
      "grad_norm": 7.073266506195068,
      "learning_rate": 9.32472691161867e-06,
      "loss": 0.7248,
      "step": 340
    },
    {
      "epoch": 0.48747461069735953,
      "grad_norm": 6.335423469543457,
      "learning_rate": 9.22542204568024e-06,
      "loss": 0.7151,
      "step": 360
    },
    {
      "epoch": 0.5145565335138795,
      "grad_norm": 6.936922550201416,
      "learning_rate": 9.126117179741808e-06,
      "loss": 0.6881,
      "step": 380
    },
    {
      "epoch": 0.5416384563303994,
      "grad_norm": 7.596807479858398,
      "learning_rate": 9.026812313803377e-06,
      "loss": 0.6995,
      "step": 400
    },
    {
      "epoch": 0.5416384563303994,
      "eval_loss": 0.6352065801620483,
      "eval_runtime": 1597.3635,
      "eval_samples_per_second": 3.697,
      "eval_steps_per_second": 0.463,
      "eval_wer": 51.63653601672089,
      "step": 400
    },
    {
      "epoch": 0.5687203791469194,
      "grad_norm": 6.981812477111816,
      "learning_rate": 8.927507447864945e-06,
      "loss": 0.6795,
      "step": 420
    },
    {
      "epoch": 0.5958023019634394,
      "grad_norm": 6.481506824493408,
      "learning_rate": 8.828202581926516e-06,
      "loss": 0.6643,
      "step": 440
    },
    {
      "epoch": 0.6228842247799594,
      "grad_norm": 6.537086009979248,
      "learning_rate": 8.728897715988084e-06,
      "loss": 0.6459,
      "step": 460
    },
    {
      "epoch": 0.6499661475964793,
      "grad_norm": 6.739567756652832,
      "learning_rate": 8.629592850049653e-06,
      "loss": 0.6736,
      "step": 480
    },
    {
      "epoch": 0.6770480704129993,
      "grad_norm": 7.422546863555908,
      "learning_rate": 8.530287984111221e-06,
      "loss": 0.6591,
      "step": 500
    },
    {
      "epoch": 0.7041299932295193,
      "grad_norm": 5.7051215171813965,
      "learning_rate": 8.430983118172792e-06,
      "loss": 0.6442,
      "step": 520
    },
    {
      "epoch": 0.7312119160460393,
      "grad_norm": 7.166143417358398,
      "learning_rate": 8.33167825223436e-06,
      "loss": 0.6326,
      "step": 540
    },
    {
      "epoch": 0.7582938388625592,
      "grad_norm": 7.759460926055908,
      "learning_rate": 8.232373386295929e-06,
      "loss": 0.6328,
      "step": 560
    },
    {
      "epoch": 0.7853757616790792,
      "grad_norm": 5.876537799835205,
      "learning_rate": 8.133068520357497e-06,
      "loss": 0.5945,
      "step": 580
    },
    {
      "epoch": 0.8124576844955992,
      "grad_norm": 6.475106716156006,
      "learning_rate": 8.033763654419066e-06,
      "loss": 0.6038,
      "step": 600
    },
    {
      "epoch": 0.8124576844955992,
      "eval_loss": 0.5318673849105835,
      "eval_runtime": 1614.0202,
      "eval_samples_per_second": 3.659,
      "eval_steps_per_second": 0.458,
      "eval_wer": 44.61419121291129,
      "step": 600
    },
    {
      "epoch": 0.8395396073121192,
      "grad_norm": 6.8348259925842285,
      "learning_rate": 7.934458788480636e-06,
      "loss": 0.6046,
      "step": 620
    },
    {
      "epoch": 0.8666215301286392,
      "grad_norm": 6.060784339904785,
      "learning_rate": 7.835153922542206e-06,
      "loss": 0.6055,
      "step": 640
    },
    {
      "epoch": 0.8937034529451591,
      "grad_norm": 6.01262903213501,
      "learning_rate": 7.735849056603775e-06,
      "loss": 0.6139,
      "step": 660
    },
    {
      "epoch": 0.9207853757616791,
      "grad_norm": 6.706854820251465,
      "learning_rate": 7.636544190665344e-06,
      "loss": 0.5789,
      "step": 680
    },
    {
      "epoch": 0.9478672985781991,
      "grad_norm": 6.464681625366211,
      "learning_rate": 7.537239324726913e-06,
      "loss": 0.5857,
      "step": 700
    },
    {
      "epoch": 0.9749492213947191,
      "grad_norm": 7.8035478591918945,
      "learning_rate": 7.437934458788482e-06,
      "loss": 0.5739,
      "step": 720
    },
    {
      "epoch": 1.001354096140826,
      "grad_norm": 5.574532985687256,
      "learning_rate": 7.33862959285005e-06,
      "loss": 0.5443,
      "step": 740
    },
    {
      "epoch": 1.028436018957346,
      "grad_norm": 5.275153160095215,
      "learning_rate": 7.23932472691162e-06,
      "loss": 0.4705,
      "step": 760
    },
    {
      "epoch": 1.055517941773866,
      "grad_norm": 5.722527027130127,
      "learning_rate": 7.140019860973188e-06,
      "loss": 0.4692,
      "step": 780
    },
    {
      "epoch": 1.0825998645903858,
      "grad_norm": 5.501762866973877,
      "learning_rate": 7.040714995034758e-06,
      "loss": 0.4791,
      "step": 800
    },
    {
      "epoch": 1.0825998645903858,
      "eval_loss": 0.46895861625671387,
      "eval_runtime": 1599.6376,
      "eval_samples_per_second": 3.691,
      "eval_steps_per_second": 0.462,
      "eval_wer": 40.9726561658299,
      "step": 800
    },
    {
      "epoch": 1.1096817874069058,
      "grad_norm": 5.207645893096924,
      "learning_rate": 6.941410129096326e-06,
      "loss": 0.4569,
      "step": 820
    },
    {
      "epoch": 1.1367637102234258,
      "grad_norm": 5.584475517272949,
      "learning_rate": 6.842105263157896e-06,
      "loss": 0.4262,
      "step": 840
    },
    {
      "epoch": 1.1638456330399458,
      "grad_norm": 5.881894588470459,
      "learning_rate": 6.742800397219464e-06,
      "loss": 0.4533,
      "step": 860
    },
    {
      "epoch": 1.1909275558564658,
      "grad_norm": 6.939334869384766,
      "learning_rate": 6.643495531281034e-06,
      "loss": 0.4675,
      "step": 880
    },
    {
      "epoch": 1.2180094786729858,
      "grad_norm": 5.780360698699951,
      "learning_rate": 6.544190665342602e-06,
      "loss": 0.4238,
      "step": 900
    },
    {
      "epoch": 1.2450914014895058,
      "grad_norm": 5.867160797119141,
      "learning_rate": 6.444885799404172e-06,
      "loss": 0.444,
      "step": 920
    },
    {
      "epoch": 1.2721733243060258,
      "grad_norm": 6.121824741363525,
      "learning_rate": 6.34558093346574e-06,
      "loss": 0.4333,
      "step": 940
    },
    {
      "epoch": 1.2992552471225456,
      "grad_norm": 5.132157802581787,
      "learning_rate": 6.24627606752731e-06,
      "loss": 0.4544,
      "step": 960
    },
    {
      "epoch": 1.3263371699390656,
      "grad_norm": 6.384315013885498,
      "learning_rate": 6.146971201588878e-06,
      "loss": 0.4332,
      "step": 980
    },
    {
      "epoch": 1.3534190927555856,
      "grad_norm": 4.54695987701416,
      "learning_rate": 6.047666335650447e-06,
      "loss": 0.4416,
      "step": 1000
    },
    {
      "epoch": 1.3534190927555856,
      "eval_loss": 0.43043428659439087,
      "eval_runtime": 1614.3899,
      "eval_samples_per_second": 3.658,
      "eval_steps_per_second": 0.458,
      "eval_wer": 38.00771402098731,
      "step": 1000
    },
    {
      "epoch": 1.3805010155721056,
      "grad_norm": 4.973900318145752,
      "learning_rate": 5.948361469712016e-06,
      "loss": 0.4401,
      "step": 1020
    },
    {
      "epoch": 1.4075829383886256,
      "grad_norm": 5.952788352966309,
      "learning_rate": 5.849056603773585e-06,
      "loss": 0.4865,
      "step": 1040
    },
    {
      "epoch": 1.4346648612051456,
      "grad_norm": 6.601942539215088,
      "learning_rate": 5.749751737835154e-06,
      "loss": 0.4435,
      "step": 1060
    },
    {
      "epoch": 1.4617467840216656,
      "grad_norm": 6.17143440246582,
      "learning_rate": 5.650446871896723e-06,
      "loss": 0.451,
      "step": 1080
    },
    {
      "epoch": 1.4888287068381856,
      "grad_norm": 5.782886981964111,
      "learning_rate": 5.551142005958292e-06,
      "loss": 0.4311,
      "step": 1100
    },
    {
      "epoch": 1.5159106296547056,
      "grad_norm": 5.734127998352051,
      "learning_rate": 5.451837140019861e-06,
      "loss": 0.4632,
      "step": 1120
    },
    {
      "epoch": 1.5429925524712256,
      "grad_norm": 5.601761341094971,
      "learning_rate": 5.35253227408143e-06,
      "loss": 0.424,
      "step": 1140
    },
    {
      "epoch": 1.5700744752877456,
      "grad_norm": 5.866110801696777,
      "learning_rate": 5.253227408142999e-06,
      "loss": 0.4273,
      "step": 1160
    },
    {
      "epoch": 1.5971563981042654,
      "grad_norm": 5.120361328125,
      "learning_rate": 5.153922542204568e-06,
      "loss": 0.4297,
      "step": 1180
    },
    {
      "epoch": 1.6242383209207854,
      "grad_norm": 5.093082427978516,
      "learning_rate": 5.054617676266137e-06,
      "loss": 0.4321,
      "step": 1200
    },
    {
      "epoch": 1.6242383209207854,
      "eval_loss": 0.3993258476257324,
      "eval_runtime": 1619.3754,
      "eval_samples_per_second": 3.646,
      "eval_steps_per_second": 0.456,
      "eval_wer": 37.091943373052636,
      "step": 1200
    },
    {
      "epoch": 1.6513202437373053,
      "grad_norm": 5.076949119567871,
      "learning_rate": 4.955312810327706e-06,
      "loss": 0.4446,
      "step": 1220
    },
    {
      "epoch": 1.6784021665538253,
      "grad_norm": 5.729442596435547,
      "learning_rate": 4.856007944389276e-06,
      "loss": 0.4375,
      "step": 1240
    },
    {
      "epoch": 1.7054840893703453,
      "grad_norm": 5.660022258758545,
      "learning_rate": 4.756703078450844e-06,
      "loss": 0.4318,
      "step": 1260
    },
    {
      "epoch": 1.7325660121868651,
      "grad_norm": 5.320847511291504,
      "learning_rate": 4.657398212512414e-06,
      "loss": 0.4255,
      "step": 1280
    },
    {
      "epoch": 1.7596479350033851,
      "grad_norm": 4.652419567108154,
      "learning_rate": 4.558093346573982e-06,
      "loss": 0.4231,
      "step": 1300
    },
    {
      "epoch": 1.7867298578199051,
      "grad_norm": 6.301181793212891,
      "learning_rate": 4.458788480635551e-06,
      "loss": 0.4426,
      "step": 1320
    },
    {
      "epoch": 1.8138117806364251,
      "grad_norm": 5.540495872497559,
      "learning_rate": 4.35948361469712e-06,
      "loss": 0.4361,
      "step": 1340
    },
    {
      "epoch": 1.8408937034529451,
      "grad_norm": 5.618797302246094,
      "learning_rate": 4.260178748758689e-06,
      "loss": 0.4081,
      "step": 1360
    },
    {
      "epoch": 1.8679756262694651,
      "grad_norm": 5.519278526306152,
      "learning_rate": 4.160873882820258e-06,
      "loss": 0.4311,
      "step": 1380
    },
    {
      "epoch": 1.8950575490859851,
      "grad_norm": 5.24718713760376,
      "learning_rate": 4.061569016881828e-06,
      "loss": 0.441,
      "step": 1400
    },
    {
      "epoch": 1.8950575490859851,
      "eval_loss": 0.37384942173957825,
      "eval_runtime": 1616.0813,
      "eval_samples_per_second": 3.654,
      "eval_steps_per_second": 0.457,
      "eval_wer": 35.20006895214291,
      "step": 1400
    },
    {
      "epoch": 1.9221394719025051,
      "grad_norm": 5.9886884689331055,
      "learning_rate": 3.962264150943396e-06,
      "loss": 0.4125,
      "step": 1420
    },
    {
      "epoch": 1.9492213947190251,
      "grad_norm": 5.233187198638916,
      "learning_rate": 3.862959285004966e-06,
      "loss": 0.4179,
      "step": 1440
    },
    {
      "epoch": 1.9763033175355451,
      "grad_norm": 5.533965587615967,
      "learning_rate": 3.763654419066535e-06,
      "loss": 0.4272,
      "step": 1460
    },
    {
      "epoch": 2.002708192281652,
      "grad_norm": 5.179864883422852,
      "learning_rate": 3.664349553128104e-06,
      "loss": 0.4082,
      "step": 1480
    },
    {
      "epoch": 2.029790115098172,
      "grad_norm": 4.748936653137207,
      "learning_rate": 3.565044687189673e-06,
      "loss": 0.3085,
      "step": 1500
    },
    {
      "epoch": 2.056872037914692,
      "grad_norm": 5.267838001251221,
      "learning_rate": 3.4657398212512415e-06,
      "loss": 0.312,
      "step": 1520
    },
    {
      "epoch": 2.083953960731212,
      "grad_norm": 6.315608501434326,
      "learning_rate": 3.3664349553128105e-06,
      "loss": 0.3512,
      "step": 1540
    },
    {
      "epoch": 2.111035883547732,
      "grad_norm": 5.232484340667725,
      "learning_rate": 3.2671300893743795e-06,
      "loss": 0.3184,
      "step": 1560
    },
    {
      "epoch": 2.138117806364252,
      "grad_norm": 4.316751956939697,
      "learning_rate": 3.1678252234359485e-06,
      "loss": 0.3129,
      "step": 1580
    },
    {
      "epoch": 2.1651997291807716,
      "grad_norm": 4.040640354156494,
      "learning_rate": 3.0685203574975175e-06,
      "loss": 0.3073,
      "step": 1600
    },
    {
      "epoch": 2.1651997291807716,
      "eval_loss": 0.3516213595867157,
      "eval_runtime": 1617.0984,
      "eval_samples_per_second": 3.652,
      "eval_steps_per_second": 0.457,
      "eval_wer": 34.23258419703075,
      "step": 1600
    },
    {
      "epoch": 2.1922816519972916,
      "grad_norm": 4.437410831451416,
      "learning_rate": 2.9692154915590865e-06,
      "loss": 0.3293,
      "step": 1620
    },
    {
      "epoch": 2.2193635748138116,
      "grad_norm": 4.284883499145508,
      "learning_rate": 2.8699106256206555e-06,
      "loss": 0.3237,
      "step": 1640
    },
    {
      "epoch": 2.2464454976303316,
      "grad_norm": 4.959640979766846,
      "learning_rate": 2.7706057596822245e-06,
      "loss": 0.3374,
      "step": 1660
    },
    {
      "epoch": 2.2735274204468516,
      "grad_norm": 5.366809368133545,
      "learning_rate": 2.6713008937437935e-06,
      "loss": 0.3086,
      "step": 1680
    },
    {
      "epoch": 2.3006093432633716,
      "grad_norm": 4.363138198852539,
      "learning_rate": 2.571996027805363e-06,
      "loss": 0.331,
      "step": 1700
    },
    {
      "epoch": 2.3276912660798916,
      "grad_norm": 4.7536468505859375,
      "learning_rate": 2.4726911618669315e-06,
      "loss": 0.317,
      "step": 1720
    },
    {
      "epoch": 2.3547731888964116,
      "grad_norm": 4.352657794952393,
      "learning_rate": 2.3733862959285005e-06,
      "loss": 0.3008,
      "step": 1740
    },
    {
      "epoch": 2.3818551117129316,
      "grad_norm": 4.978397846221924,
      "learning_rate": 2.2740814299900695e-06,
      "loss": 0.3245,
      "step": 1760
    },
    {
      "epoch": 2.4089370345294516,
      "grad_norm": 4.570504665374756,
      "learning_rate": 2.174776564051639e-06,
      "loss": 0.3147,
      "step": 1780
    },
    {
      "epoch": 2.4360189573459716,
      "grad_norm": 5.904801845550537,
      "learning_rate": 2.075471698113208e-06,
      "loss": 0.3395,
      "step": 1800
    },
    {
      "epoch": 2.4360189573459716,
      "eval_loss": 0.3366335332393646,
      "eval_runtime": 1614.6836,
      "eval_samples_per_second": 3.657,
      "eval_steps_per_second": 0.458,
      "eval_wer": 33.028076450688445,
      "step": 1800
    },
    {
      "epoch": 2.4631008801624916,
      "grad_norm": 5.771766185760498,
      "learning_rate": 1.976166832174777e-06,
      "loss": 0.3064,
      "step": 1820
    },
    {
      "epoch": 2.4901828029790116,
      "grad_norm": 5.262455463409424,
      "learning_rate": 1.8768619662363458e-06,
      "loss": 0.2961,
      "step": 1840
    },
    {
      "epoch": 2.5172647257955316,
      "grad_norm": 5.288928031921387,
      "learning_rate": 1.7775571002979148e-06,
      "loss": 0.3285,
      "step": 1860
    },
    {
      "epoch": 2.5443466486120516,
      "grad_norm": 5.816193580627441,
      "learning_rate": 1.6782522343594838e-06,
      "loss": 0.3254,
      "step": 1880
    },
    {
      "epoch": 2.571428571428571,
      "grad_norm": 5.249015808105469,
      "learning_rate": 1.5789473684210526e-06,
      "loss": 0.3279,
      "step": 1900
    },
    {
      "epoch": 2.598510494245091,
      "grad_norm": 4.509056568145752,
      "learning_rate": 1.4796425024826216e-06,
      "loss": 0.317,
      "step": 1920
    },
    {
      "epoch": 2.625592417061611,
      "grad_norm": 4.113898754119873,
      "learning_rate": 1.3803376365441908e-06,
      "loss": 0.3135,
      "step": 1940
    },
    {
      "epoch": 2.652674339878131,
      "grad_norm": 5.507041931152344,
      "learning_rate": 1.2810327706057598e-06,
      "loss": 0.3302,
      "step": 1960
    },
    {
      "epoch": 2.679756262694651,
      "grad_norm": 4.8485846519470215,
      "learning_rate": 1.1817279046673288e-06,
      "loss": 0.3106,
      "step": 1980
    },
    {
      "epoch": 2.706838185511171,
      "grad_norm": 5.389250755310059,
      "learning_rate": 1.0824230387288978e-06,
      "loss": 0.3347,
      "step": 2000
    },
    {
      "epoch": 2.706838185511171,
      "eval_loss": 0.3267146348953247,
      "eval_runtime": 1650.7646,
      "eval_samples_per_second": 3.577,
      "eval_steps_per_second": 0.448,
      "eval_wer": 32.74580361567799,
      "step": 2000
    }
  ],
  "logging_steps": 20,
  "max_steps": 2214,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 200,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": false
      },
      "attributes": {}
    }
  },
  "total_flos": 1.845330483806208e+19,
  "train_batch_size": 16,
  "trial_name": null,
  "trial_params": null
}