{
  "best_metric": 22.41163981516173,
  "best_model_checkpoint": "/scratch/andoni.sudupe/checkpoints/sweep_new/tiny/lr1e-5/checkpoint-9000",
  "epoch": 6.605019815059445,
  "eval_steps": 500,
  "global_step": 10000,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.02,
      "grad_norm": 34.077606201171875,
      "learning_rate": 4.4e-07,
      "loss": 4.2904,
      "step": 25
    },
    {
      "epoch": 0.03,
      "grad_norm": 23.43511199951172,
      "learning_rate": 9.400000000000001e-07,
      "loss": 3.9844,
      "step": 50
    },
    {
      "epoch": 0.05,
      "grad_norm": 16.568195343017578,
      "learning_rate": 1.44e-06,
      "loss": 3.4183,
      "step": 75
    },
    {
      "epoch": 0.07,
      "grad_norm": 14.12092113494873,
      "learning_rate": 1.94e-06,
      "loss": 2.821,
      "step": 100
    },
    {
      "epoch": 0.08,
      "grad_norm": 7.320805549621582,
      "learning_rate": 2.4400000000000004e-06,
      "loss": 2.2669,
      "step": 125
    },
    {
      "epoch": 0.1,
      "grad_norm": 6.777274131774902,
      "learning_rate": 2.9400000000000002e-06,
      "loss": 1.9647,
      "step": 150
    },
    {
      "epoch": 0.12,
      "grad_norm": 6.5238447189331055,
      "learning_rate": 3.44e-06,
      "loss": 1.7396,
      "step": 175
    },
    {
      "epoch": 0.13,
      "grad_norm": 6.274177074432373,
      "learning_rate": 3.94e-06,
      "loss": 1.5854,
      "step": 200
    },
    {
      "epoch": 0.15,
      "grad_norm": 6.0430803298950195,
      "learning_rate": 4.440000000000001e-06,
      "loss": 1.45,
      "step": 225
    },
    {
      "epoch": 0.17,
      "grad_norm": 5.32307243347168,
      "learning_rate": 4.94e-06,
      "loss": 1.3254,
      "step": 250
    },
    {
      "epoch": 0.18,
      "grad_norm": 3.9543581008911133,
      "learning_rate": 5.4400000000000004e-06,
      "loss": 1.2132,
      "step": 275
    },
    {
      "epoch": 0.2,
      "grad_norm": 3.564781427383423,
      "learning_rate": 5.94e-06,
      "loss": 1.1088,
      "step": 300
    },
    {
      "epoch": 0.21,
      "grad_norm": 3.724240303039551,
      "learning_rate": 6.440000000000001e-06,
      "loss": 0.9806,
      "step": 325
    },
    {
      "epoch": 0.23,
      "grad_norm": 5.0616254806518555,
      "learning_rate": 6.9400000000000005e-06,
      "loss": 0.7913,
      "step": 350
    },
    {
      "epoch": 0.25,
      "grad_norm": 5.042689800262451,
      "learning_rate": 7.440000000000001e-06,
      "loss": 0.6871,
      "step": 375
    },
    {
      "epoch": 0.26,
      "grad_norm": 3.1566641330718994,
      "learning_rate": 7.94e-06,
      "loss": 0.6291,
      "step": 400
    },
    {
      "epoch": 0.28,
      "grad_norm": 3.285292148590088,
      "learning_rate": 8.44e-06,
      "loss": 0.5892,
      "step": 425
    },
    {
      "epoch": 0.3,
      "grad_norm": 3.3190643787384033,
      "learning_rate": 8.94e-06,
      "loss": 0.567,
      "step": 450
    },
    {
      "epoch": 0.31,
      "grad_norm": 3.9897689819335938,
      "learning_rate": 9.440000000000001e-06,
      "loss": 0.5383,
      "step": 475
    },
    {
      "epoch": 0.33,
      "grad_norm": 2.80560040473938,
      "learning_rate": 9.940000000000001e-06,
      "loss": 0.5114,
      "step": 500
    },
    {
      "epoch": 0.33,
      "eval_loss": 0.6930010914802551,
      "eval_runtime": 21.1291,
      "eval_samples_per_second": 80.032,
      "eval_steps_per_second": 0.663,
      "eval_wer": 52.52279255651305,
      "step": 500
    },
    {
      "epoch": 0.35,
      "grad_norm": 2.685612678527832,
      "learning_rate": 9.976842105263158e-06,
      "loss": 0.4912,
      "step": 525
    },
    {
      "epoch": 0.36,
      "grad_norm": 2.799072742462158,
      "learning_rate": 9.950526315789475e-06,
      "loss": 0.4752,
      "step": 550
    },
    {
      "epoch": 0.38,
      "grad_norm": 2.7478973865509033,
      "learning_rate": 9.92421052631579e-06,
      "loss": 0.4556,
      "step": 575
    },
    {
      "epoch": 0.4,
      "grad_norm": 2.917057991027832,
      "learning_rate": 9.897894736842107e-06,
      "loss": 0.4415,
      "step": 600
    },
    {
      "epoch": 0.41,
      "grad_norm": 4.174392223358154,
      "learning_rate": 9.871578947368422e-06,
      "loss": 0.4228,
      "step": 625
    },
    {
      "epoch": 0.43,
      "grad_norm": 2.487973213195801,
      "learning_rate": 9.845263157894738e-06,
      "loss": 0.4103,
      "step": 650
    },
    {
      "epoch": 0.45,
      "grad_norm": 2.432943820953369,
      "learning_rate": 9.818947368421053e-06,
      "loss": 0.3984,
      "step": 675
    },
    {
      "epoch": 0.46,
      "grad_norm": 2.607616901397705,
      "learning_rate": 9.79263157894737e-06,
      "loss": 0.3877,
      "step": 700
    },
    {
      "epoch": 0.48,
      "grad_norm": 2.3920302391052246,
      "learning_rate": 9.766315789473685e-06,
      "loss": 0.3788,
      "step": 725
    },
    {
      "epoch": 0.5,
      "grad_norm": 2.441880702972412,
      "learning_rate": 9.74e-06,
      "loss": 0.3758,
      "step": 750
    },
    {
      "epoch": 0.51,
      "grad_norm": 2.3188793659210205,
      "learning_rate": 9.713684210526317e-06,
      "loss": 0.3662,
      "step": 775
    },
    {
      "epoch": 0.53,
      "grad_norm": 2.532235622406006,
      "learning_rate": 9.687368421052632e-06,
      "loss": 0.3582,
      "step": 800
    },
    {
      "epoch": 0.54,
      "grad_norm": 2.390369176864624,
      "learning_rate": 9.661052631578948e-06,
      "loss": 0.3468,
      "step": 825
    },
    {
      "epoch": 0.56,
      "grad_norm": 2.3737587928771973,
      "learning_rate": 9.634736842105265e-06,
      "loss": 0.3445,
      "step": 850
    },
    {
      "epoch": 0.58,
      "grad_norm": 2.4866981506347656,
      "learning_rate": 9.60842105263158e-06,
      "loss": 0.3341,
      "step": 875
    },
    {
      "epoch": 0.59,
      "grad_norm": 2.2460451126098633,
      "learning_rate": 9.582105263157897e-06,
      "loss": 0.3253,
      "step": 900
    },
    {
      "epoch": 0.61,
      "grad_norm": 2.3266069889068604,
      "learning_rate": 9.555789473684211e-06,
      "loss": 0.3278,
      "step": 925
    },
    {
      "epoch": 0.63,
      "grad_norm": 2.203608274459839,
      "learning_rate": 9.529473684210528e-06,
      "loss": 0.3218,
      "step": 950
    },
    {
      "epoch": 0.64,
      "grad_norm": 2.13688063621521,
      "learning_rate": 9.503157894736843e-06,
      "loss": 0.3183,
      "step": 975
    },
    {
      "epoch": 0.66,
      "grad_norm": 2.313721179962158,
      "learning_rate": 9.476842105263158e-06,
      "loss": 0.3117,
      "step": 1000
    },
    {
      "epoch": 0.66,
      "eval_loss": 0.5081741809844971,
      "eval_runtime": 21.6341,
      "eval_samples_per_second": 78.164,
      "eval_steps_per_second": 0.647,
      "eval_wer": 37.14874484825778,
      "step": 1000
    },
    {
      "epoch": 0.68,
      "grad_norm": 2.4097204208374023,
      "learning_rate": 9.450526315789475e-06,
      "loss": 0.308,
      "step": 1025
    },
    {
      "epoch": 0.69,
      "grad_norm": 2.286888599395752,
      "learning_rate": 9.42421052631579e-06,
      "loss": 0.3042,
      "step": 1050
    },
    {
      "epoch": 0.71,
      "grad_norm": 2.2226040363311768,
      "learning_rate": 9.397894736842106e-06,
      "loss": 0.3014,
      "step": 1075
    },
    {
      "epoch": 0.73,
      "grad_norm": 2.1482009887695312,
      "learning_rate": 9.371578947368421e-06,
      "loss": 0.295,
      "step": 1100
    },
    {
      "epoch": 0.74,
      "grad_norm": 2.2194724082946777,
      "learning_rate": 9.345263157894738e-06,
      "loss": 0.2911,
      "step": 1125
    },
    {
      "epoch": 0.76,
      "grad_norm": 2.2383620738983154,
      "learning_rate": 9.318947368421053e-06,
      "loss": 0.2925,
      "step": 1150
    },
    {
      "epoch": 0.78,
      "grad_norm": 2.307551145553589,
      "learning_rate": 9.292631578947368e-06,
      "loss": 0.2834,
      "step": 1175
    },
    {
      "epoch": 0.79,
      "grad_norm": 2.0051376819610596,
      "learning_rate": 9.266315789473685e-06,
      "loss": 0.2778,
      "step": 1200
    },
    {
      "epoch": 0.81,
      "grad_norm": 2.2903966903686523,
      "learning_rate": 9.240000000000001e-06,
      "loss": 0.2828,
      "step": 1225
    },
    {
      "epoch": 0.83,
      "grad_norm": 2.056866407394409,
      "learning_rate": 9.213684210526316e-06,
      "loss": 0.2757,
      "step": 1250
    },
    {
      "epoch": 0.84,
      "grad_norm": 2.2591381072998047,
      "learning_rate": 9.187368421052633e-06,
      "loss": 0.2742,
      "step": 1275
    },
    {
      "epoch": 0.86,
      "grad_norm": 2.1059749126434326,
      "learning_rate": 9.161052631578948e-06,
      "loss": 0.2694,
      "step": 1300
    },
    {
      "epoch": 0.88,
      "grad_norm": 2.204620838165283,
      "learning_rate": 9.134736842105265e-06,
      "loss": 0.2675,
      "step": 1325
    },
    {
      "epoch": 0.89,
      "grad_norm": 2.11024808883667,
      "learning_rate": 9.10842105263158e-06,
      "loss": 0.2657,
      "step": 1350
    },
    {
      "epoch": 0.91,
      "grad_norm": 2.131837844848633,
      "learning_rate": 9.082105263157896e-06,
      "loss": 0.2635,
      "step": 1375
    },
    {
      "epoch": 0.92,
      "grad_norm": 2.171182870864868,
      "learning_rate": 9.055789473684211e-06,
      "loss": 0.2592,
      "step": 1400
    },
    {
      "epoch": 0.94,
      "grad_norm": 2.0435309410095215,
      "learning_rate": 9.029473684210526e-06,
      "loss": 0.2584,
      "step": 1425
    },
    {
      "epoch": 0.96,
      "grad_norm": 2.009622573852539,
      "learning_rate": 9.003157894736843e-06,
      "loss": 0.2615,
      "step": 1450
    },
    {
      "epoch": 0.97,
      "grad_norm": 2.5420186519622803,
      "learning_rate": 8.976842105263158e-06,
      "loss": 0.2598,
      "step": 1475
    },
    {
      "epoch": 0.99,
      "grad_norm": 2.2081310749053955,
      "learning_rate": 8.950526315789474e-06,
      "loss": 0.2552,
      "step": 1500
    },
    {
      "epoch": 0.99,
      "eval_loss": 0.4230075776576996,
      "eval_runtime": 15.907,
      "eval_samples_per_second": 106.306,
      "eval_steps_per_second": 0.88,
      "eval_wer": 30.26102160609467,
      "step": 1500
    },
    {
      "epoch": 1.01,
      "grad_norm": 2.008251905441284,
      "learning_rate": 8.92421052631579e-06,
      "loss": 0.2425,
      "step": 1525
    },
    {
      "epoch": 1.02,
      "grad_norm": 1.9225270748138428,
      "learning_rate": 8.897894736842106e-06,
      "loss": 0.2319,
      "step": 1550
    },
    {
      "epoch": 1.04,
      "grad_norm": 1.8944556713104248,
      "learning_rate": 8.871578947368423e-06,
      "loss": 0.2306,
      "step": 1575
    },
    {
      "epoch": 1.06,
      "grad_norm": 1.915433645248413,
      "learning_rate": 8.845263157894738e-06,
      "loss": 0.2331,
      "step": 1600
    },
    {
      "epoch": 1.07,
      "grad_norm": 2.067833662033081,
      "learning_rate": 8.818947368421054e-06,
      "loss": 0.2327,
      "step": 1625
    },
    {
      "epoch": 1.09,
      "grad_norm": 2.1529483795166016,
      "learning_rate": 8.79263157894737e-06,
      "loss": 0.2285,
      "step": 1650
    },
    {
      "epoch": 1.11,
      "grad_norm": 2.0223615169525146,
      "learning_rate": 8.766315789473684e-06,
      "loss": 0.2258,
      "step": 1675
    },
    {
      "epoch": 1.12,
      "grad_norm": 2.064879894256592,
      "learning_rate": 8.740000000000001e-06,
      "loss": 0.2261,
      "step": 1700
    },
    {
      "epoch": 1.14,
      "grad_norm": 2.033409595489502,
      "learning_rate": 8.713684210526316e-06,
      "loss": 0.2266,
      "step": 1725
    },
    {
      "epoch": 1.16,
      "grad_norm": 1.8101595640182495,
      "learning_rate": 8.687368421052633e-06,
      "loss": 0.2248,
      "step": 1750
    },
    {
      "epoch": 1.17,
      "grad_norm": 2.1755690574645996,
      "learning_rate": 8.661052631578948e-06,
      "loss": 0.2184,
      "step": 1775
    },
    {
      "epoch": 1.19,
      "grad_norm": 1.8470789194107056,
      "learning_rate": 8.634736842105264e-06,
      "loss": 0.2205,
      "step": 1800
    },
    {
      "epoch": 1.21,
      "grad_norm": 2.2927639484405518,
      "learning_rate": 8.60842105263158e-06,
      "loss": 0.2173,
      "step": 1825
    },
    {
      "epoch": 1.22,
      "grad_norm": 1.9602240324020386,
      "learning_rate": 8.582105263157894e-06,
      "loss": 0.2169,
      "step": 1850
    },
    {
      "epoch": 1.24,
      "grad_norm": 1.941825270652771,
      "learning_rate": 8.555789473684213e-06,
      "loss": 0.2131,
      "step": 1875
    },
    {
      "epoch": 1.25,
      "grad_norm": 2.0443484783172607,
      "learning_rate": 8.529473684210528e-06,
      "loss": 0.2151,
      "step": 1900
    },
    {
      "epoch": 1.27,
      "grad_norm": 1.8085869550704956,
      "learning_rate": 8.503157894736843e-06,
      "loss": 0.2168,
      "step": 1925
    },
    {
      "epoch": 1.29,
      "grad_norm": 1.8386337757110596,
      "learning_rate": 8.47684210526316e-06,
      "loss": 0.214,
      "step": 1950
    },
    {
      "epoch": 1.3,
      "grad_norm": 2.097592830657959,
      "learning_rate": 8.450526315789474e-06,
      "loss": 0.2106,
      "step": 1975
    },
    {
      "epoch": 1.32,
      "grad_norm": 2.083630323410034,
      "learning_rate": 8.42421052631579e-06,
      "loss": 0.2102,
      "step": 2000
    },
    {
      "epoch": 1.32,
      "eval_loss": 0.3910539448261261,
      "eval_runtime": 16.2845,
      "eval_samples_per_second": 103.841,
      "eval_steps_per_second": 0.86,
      "eval_wer": 28.606219557886845,
      "step": 2000
    },
    {
      "epoch": 1.34,
      "grad_norm": 1.8663444519042969,
      "learning_rate": 8.397894736842106e-06,
      "loss": 0.2077,
      "step": 2025
    },
    {
      "epoch": 1.35,
      "grad_norm": 1.958034873008728,
      "learning_rate": 8.371578947368422e-06,
      "loss": 0.2102,
      "step": 2050
    },
    {
      "epoch": 1.37,
      "grad_norm": 1.7769043445587158,
      "learning_rate": 8.345263157894737e-06,
      "loss": 0.2055,
      "step": 2075
    },
    {
      "epoch": 1.39,
      "grad_norm": 1.8657371997833252,
      "learning_rate": 8.318947368421052e-06,
      "loss": 0.2073,
      "step": 2100
    },
    {
      "epoch": 1.4,
      "grad_norm": 2.0003774166107178,
      "learning_rate": 8.292631578947369e-06,
      "loss": 0.2094,
      "step": 2125
    },
    {
      "epoch": 1.42,
      "grad_norm": 1.8350632190704346,
      "learning_rate": 8.266315789473684e-06,
      "loss": 0.2027,
      "step": 2150
    },
    {
      "epoch": 1.44,
      "grad_norm": 2.0300979614257812,
      "learning_rate": 8.24e-06,
      "loss": 0.2082,
      "step": 2175
    },
    {
      "epoch": 1.45,
      "grad_norm": 2.034996747970581,
      "learning_rate": 8.213684210526316e-06,
      "loss": 0.206,
      "step": 2200
    },
    {
      "epoch": 1.47,
      "grad_norm": 1.8040868043899536,
      "learning_rate": 8.187368421052632e-06,
      "loss": 0.1994,
      "step": 2225
    },
    {
      "epoch": 1.49,
      "grad_norm": 1.9792579412460327,
      "learning_rate": 8.161052631578949e-06,
      "loss": 0.1985,
      "step": 2250
    },
    {
      "epoch": 1.5,
      "grad_norm": 2.0966882705688477,
      "learning_rate": 8.134736842105264e-06,
      "loss": 0.1989,
      "step": 2275
    },
    {
      "epoch": 1.52,
      "grad_norm": 1.958456039428711,
      "learning_rate": 8.10842105263158e-06,
      "loss": 0.204,
      "step": 2300
    },
    {
      "epoch": 1.54,
      "grad_norm": 1.7953109741210938,
      "learning_rate": 8.082105263157896e-06,
      "loss": 0.1953,
      "step": 2325
    },
    {
      "epoch": 1.55,
      "grad_norm": 1.8202308416366577,
      "learning_rate": 8.05578947368421e-06,
      "loss": 0.2001,
      "step": 2350
    },
    {
      "epoch": 1.57,
      "grad_norm": 1.690443754196167,
      "learning_rate": 8.029473684210527e-06,
      "loss": 0.1964,
      "step": 2375
    },
    {
      "epoch": 1.59,
      "grad_norm": 1.8226059675216675,
      "learning_rate": 8.003157894736842e-06,
      "loss": 0.1949,
      "step": 2400
    },
    {
      "epoch": 1.6,
      "grad_norm": 1.7331619262695312,
      "learning_rate": 7.976842105263159e-06,
      "loss": 0.1946,
      "step": 2425
    },
    {
      "epoch": 1.62,
      "grad_norm": 1.9106403589248657,
      "learning_rate": 7.950526315789474e-06,
      "loss": 0.191,
      "step": 2450
    },
    {
      "epoch": 1.63,
      "grad_norm": 1.8373783826828003,
      "learning_rate": 7.92421052631579e-06,
      "loss": 0.1903,
      "step": 2475
    },
    {
      "epoch": 1.65,
      "grad_norm": 1.9064444303512573,
      "learning_rate": 7.897894736842106e-06,
      "loss": 0.1925,
      "step": 2500
    },
    {
      "epoch": 1.65,
      "eval_loss": 0.37334463000297546,
      "eval_runtime": 14.6238,
      "eval_samples_per_second": 115.634,
      "eval_steps_per_second": 0.957,
      "eval_wer": 26.745347820656924,
      "step": 2500
    },
    {
      "epoch": 1.67,
      "grad_norm": 1.888606071472168,
      "learning_rate": 7.87157894736842e-06,
      "loss": 0.1907,
      "step": 2525
    },
    {
      "epoch": 1.68,
      "grad_norm": 1.7304494380950928,
      "learning_rate": 7.845263157894739e-06,
      "loss": 0.1906,
      "step": 2550
    },
    {
      "epoch": 1.7,
      "grad_norm": 1.856978416442871,
      "learning_rate": 7.818947368421054e-06,
      "loss": 0.1866,
      "step": 2575
    },
    {
      "epoch": 1.72,
      "grad_norm": 1.7459311485290527,
      "learning_rate": 7.792631578947369e-06,
      "loss": 0.1904,
      "step": 2600
    },
    {
      "epoch": 1.73,
      "grad_norm": 1.777932047843933,
      "learning_rate": 7.766315789473685e-06,
      "loss": 0.1895,
      "step": 2625
    },
    {
      "epoch": 1.75,
      "grad_norm": 1.8538662195205688,
      "learning_rate": 7.74e-06,
      "loss": 0.186,
      "step": 2650
    },
    {
      "epoch": 1.77,
      "grad_norm": 1.790150761604309,
      "learning_rate": 7.713684210526317e-06,
      "loss": 0.1858,
      "step": 2675
    },
    {
      "epoch": 1.78,
      "grad_norm": 1.7638112306594849,
      "learning_rate": 7.687368421052632e-06,
      "loss": 0.1885,
      "step": 2700
    },
    {
      "epoch": 1.8,
      "grad_norm": 1.9235881567001343,
      "learning_rate": 7.661052631578949e-06,
      "loss": 0.1869,
      "step": 2725
    },
    {
      "epoch": 1.82,
      "grad_norm": 1.7366169691085815,
      "learning_rate": 7.634736842105264e-06,
      "loss": 0.1838,
      "step": 2750
    },
    {
      "epoch": 1.83,
      "grad_norm": 1.7510082721710205,
      "learning_rate": 7.6084210526315795e-06,
      "loss": 0.1856,
      "step": 2775
    },
    {
      "epoch": 1.85,
      "grad_norm": 1.8957219123840332,
      "learning_rate": 7.582105263157895e-06,
      "loss": 0.1866,
      "step": 2800
    },
    {
      "epoch": 1.87,
      "grad_norm": 1.7200273275375366,
      "learning_rate": 7.555789473684211e-06,
      "loss": 0.1831,
      "step": 2825
    },
    {
      "epoch": 1.88,
      "grad_norm": 1.9370934963226318,
      "learning_rate": 7.529473684210526e-06,
      "loss": 0.1831,
      "step": 2850
    },
    {
      "epoch": 1.9,
      "grad_norm": 1.6578595638275146,
      "learning_rate": 7.503157894736842e-06,
      "loss": 0.1826,
      "step": 2875
    },
    {
      "epoch": 1.92,
      "grad_norm": 1.7850892543792725,
      "learning_rate": 7.476842105263159e-06,
      "loss": 0.1787,
      "step": 2900
    },
    {
      "epoch": 1.93,
      "grad_norm": 1.7334098815917969,
      "learning_rate": 7.4505263157894744e-06,
      "loss": 0.18,
      "step": 2925
    },
    {
      "epoch": 1.95,
      "grad_norm": 1.7443405389785767,
      "learning_rate": 7.42421052631579e-06,
      "loss": 0.1776,
      "step": 2950
    },
    {
      "epoch": 1.96,
      "grad_norm": 1.7523411512374878,
      "learning_rate": 7.397894736842106e-06,
      "loss": 0.1811,
      "step": 2975
    },
    {
      "epoch": 1.98,
      "grad_norm": 1.8293405771255493,
      "learning_rate": 7.371578947368422e-06,
      "loss": 0.1768,
      "step": 3000
    },
    {
      "epoch": 1.98,
      "eval_loss": 0.35854896903038025,
      "eval_runtime": 14.7913,
      "eval_samples_per_second": 114.324,
      "eval_steps_per_second": 0.947,
      "eval_wer": 24.96565505182965,
      "step": 3000
    },
    {
      "epoch": 2.0,
      "grad_norm": 1.6644916534423828,
      "learning_rate": 7.345263157894738e-06,
      "loss": 0.1776,
      "step": 3025
    },
    {
      "epoch": 2.01,
      "grad_norm": 1.7578173875808716,
      "learning_rate": 7.3189473684210535e-06,
      "loss": 0.1698,
      "step": 3050
    },
    {
      "epoch": 2.03,
      "grad_norm": 1.7905900478363037,
      "learning_rate": 7.292631578947369e-06,
      "loss": 0.1691,
      "step": 3075
    },
    {
      "epoch": 2.05,
      "grad_norm": 1.583471655845642,
      "learning_rate": 7.266315789473684e-06,
      "loss": 0.1668,
      "step": 3100
    },
    {
      "epoch": 2.06,
      "grad_norm": 1.7078475952148438,
      "learning_rate": 7.24e-06,
      "loss": 0.1624,
      "step": 3125
    },
    {
      "epoch": 2.08,
      "grad_norm": 1.8642972707748413,
      "learning_rate": 7.213684210526316e-06,
      "loss": 0.1645,
      "step": 3150
    },
    {
      "epoch": 2.1,
      "grad_norm": 1.739005446434021,
      "learning_rate": 7.187368421052632e-06,
      "loss": 0.162,
      "step": 3175
    },
    {
      "epoch": 2.11,
      "grad_norm": 1.7232236862182617,
      "learning_rate": 7.161052631578948e-06,
      "loss": 0.1649,
      "step": 3200
    },
    {
      "epoch": 2.13,
      "grad_norm": 1.9032922983169556,
      "learning_rate": 7.134736842105264e-06,
      "loss": 0.1584,
      "step": 3225
    },
    {
      "epoch": 2.15,
      "grad_norm": 1.8413559198379517,
      "learning_rate": 7.10842105263158e-06,
      "loss": 0.1582,
      "step": 3250
    },
    {
      "epoch": 2.16,
      "grad_norm": 1.672963261604309,
      "learning_rate": 7.082105263157896e-06,
      "loss": 0.1648,
      "step": 3275
    },
    {
      "epoch": 2.18,
      "grad_norm": 1.9063982963562012,
      "learning_rate": 7.055789473684212e-06,
      "loss": 0.1621,
      "step": 3300
    },
    {
      "epoch": 2.2,
      "grad_norm": 1.7335928678512573,
      "learning_rate": 7.0294736842105275e-06,
      "loss": 0.1593,
      "step": 3325
    },
    {
      "epoch": 2.21,
      "grad_norm": 1.6883697509765625,
      "learning_rate": 7.0031578947368425e-06,
      "loss": 0.1613,
      "step": 3350
    },
    {
      "epoch": 2.23,
      "grad_norm": 1.6064486503601074,
      "learning_rate": 6.976842105263158e-06,
      "loss": 0.1602,
      "step": 3375
    },
    {
      "epoch": 2.25,
      "grad_norm": 1.62557852268219,
      "learning_rate": 6.950526315789474e-06,
      "loss": 0.1574,
      "step": 3400
    },
    {
      "epoch": 2.26,
      "grad_norm": 1.7775442600250244,
      "learning_rate": 6.92421052631579e-06,
      "loss": 0.1563,
      "step": 3425
    },
    {
      "epoch": 2.28,
      "grad_norm": 1.6719475984573364,
      "learning_rate": 6.897894736842106e-06,
      "loss": 0.162,
      "step": 3450
    },
    {
      "epoch": 2.3,
      "grad_norm": 1.8289706707000732,
      "learning_rate": 6.871578947368422e-06,
      "loss": 0.1653,
      "step": 3475
    },
    {
      "epoch": 2.31,
      "grad_norm": 1.6735001802444458,
      "learning_rate": 6.845263157894737e-06,
      "loss": 0.1572,
      "step": 3500
    },
    {
      "epoch": 2.31,
      "eval_loss": 0.35393840074539185,
      "eval_runtime": 15.9677,
      "eval_samples_per_second": 105.901,
      "eval_steps_per_second": 0.877,
      "eval_wer": 25.346571749718994,
      "step": 3500
    },
    {
      "epoch": 2.33,
      "grad_norm": 1.6916478872299194,
      "learning_rate": 6.818947368421052e-06,
      "loss": 0.1514,
      "step": 3525
    },
    {
      "epoch": 2.34,
      "grad_norm": 1.6720538139343262,
      "learning_rate": 6.792631578947368e-06,
      "loss": 0.1579,
      "step": 3550
    },
    {
      "epoch": 2.36,
      "grad_norm": 1.7501777410507202,
      "learning_rate": 6.766315789473685e-06,
      "loss": 0.1566,
      "step": 3575
    },
    {
      "epoch": 2.38,
      "grad_norm": 1.8388835191726685,
      "learning_rate": 6.740000000000001e-06,
      "loss": 0.1591,
      "step": 3600
    },
    {
      "epoch": 2.39,
      "grad_norm": 1.8639742136001587,
      "learning_rate": 6.7136842105263165e-06,
      "loss": 0.156,
      "step": 3625
    },
    {
      "epoch": 2.41,
      "grad_norm": 1.8133294582366943,
      "learning_rate": 6.687368421052632e-06,
      "loss": 0.16,
      "step": 3650
    },
    {
      "epoch": 2.43,
      "grad_norm": 1.734179139137268,
      "learning_rate": 6.661052631578948e-06,
      "loss": 0.1546,
      "step": 3675
    },
    {
      "epoch": 2.44,
      "grad_norm": 1.736973524093628,
      "learning_rate": 6.634736842105264e-06,
      "loss": 0.1554,
      "step": 3700
    },
    {
      "epoch": 2.46,
      "grad_norm": 1.5530760288238525,
      "learning_rate": 6.60842105263158e-06,
      "loss": 0.1547,
      "step": 3725
    },
    {
      "epoch": 2.48,
      "grad_norm": 1.634820580482483,
      "learning_rate": 6.582105263157896e-06,
      "loss": 0.1542,
      "step": 3750
    },
    {
      "epoch": 2.49,
      "grad_norm": 1.5735784769058228,
      "learning_rate": 6.5557894736842106e-06,
      "loss": 0.1579,
      "step": 3775
    },
    {
      "epoch": 2.51,
      "grad_norm": 1.7229827642440796,
      "learning_rate": 6.529473684210526e-06,
      "loss": 0.1536,
      "step": 3800
    },
    {
      "epoch": 2.53,
      "grad_norm": 1.6279429197311401,
      "learning_rate": 6.503157894736842e-06,
      "loss": 0.1518,
      "step": 3825
    },
    {
      "epoch": 2.54,
      "grad_norm": 1.5785133838653564,
      "learning_rate": 6.476842105263158e-06,
      "loss": 0.1488,
      "step": 3850
    },
    {
      "epoch": 2.56,
      "grad_norm": 1.494593620300293,
      "learning_rate": 6.450526315789474e-06,
      "loss": 0.1525,
      "step": 3875
    },
    {
      "epoch": 2.58,
      "grad_norm": 1.7033801078796387,
      "learning_rate": 6.42421052631579e-06,
      "loss": 0.1539,
      "step": 3900
    },
    {
      "epoch": 2.59,
      "grad_norm": 1.574308156967163,
      "learning_rate": 6.397894736842106e-06,
      "loss": 0.1488,
      "step": 3925
    },
    {
      "epoch": 2.61,
      "grad_norm": 1.597985863685608,
      "learning_rate": 6.371578947368422e-06,
      "loss": 0.1528,
      "step": 3950
    },
    {
      "epoch": 2.63,
      "grad_norm": 1.7028511762619019,
      "learning_rate": 6.345263157894738e-06,
      "loss": 0.1549,
      "step": 3975
    },
    {
      "epoch": 2.64,
      "grad_norm": 1.7545559406280518,
      "learning_rate": 6.318947368421054e-06,
      "loss": 0.1503,
      "step": 4000
    },
    {
      "epoch": 2.64,
      "eval_loss": 0.3429431915283203,
      "eval_runtime": 15.347,
      "eval_samples_per_second": 110.185,
      "eval_steps_per_second": 0.912,
      "eval_wer": 24.19133258398901,
      "step": 4000
    },
    {
      "epoch": 2.66,
      "grad_norm": 1.7337292432785034,
      "learning_rate": 6.292631578947369e-06,
      "loss": 0.1538,
      "step": 4025
    },
    {
      "epoch": 2.68,
      "grad_norm": 1.7260130643844604,
      "learning_rate": 6.266315789473685e-06,
      "loss": 0.1501,
      "step": 4050
    },
    {
      "epoch": 2.69,
      "grad_norm": 1.6987520456314087,
      "learning_rate": 6.24e-06,
      "loss": 0.1542,
      "step": 4075
    },
    {
      "epoch": 2.71,
      "grad_norm": 1.781073808670044,
      "learning_rate": 6.213684210526316e-06,
      "loss": 0.1534,
      "step": 4100
    },
    {
      "epoch": 2.72,
      "grad_norm": 1.6806445121765137,
      "learning_rate": 6.187368421052632e-06,
      "loss": 0.147,
      "step": 4125
    },
    {
      "epoch": 2.74,
      "grad_norm": 1.4836201667785645,
      "learning_rate": 6.161052631578948e-06,
      "loss": 0.1509,
      "step": 4150
    },
    {
      "epoch": 2.76,
      "grad_norm": 1.7774155139923096,
      "learning_rate": 6.134736842105264e-06,
      "loss": 0.1519,
      "step": 4175
    },
    {
      "epoch": 2.77,
      "grad_norm": 1.613662600517273,
      "learning_rate": 6.108421052631579e-06,
      "loss": 0.15,
      "step": 4200
    },
    {
      "epoch": 2.79,
      "grad_norm": 1.713094711303711,
      "learning_rate": 6.0821052631578945e-06,
      "loss": 0.1481,
      "step": 4225
    },
    {
      "epoch": 2.81,
      "grad_norm": 1.7286796569824219,
      "learning_rate": 6.055789473684212e-06,
      "loss": 0.15,
      "step": 4250
    },
    {
      "epoch": 2.82,
      "grad_norm": 1.566652774810791,
      "learning_rate": 6.029473684210527e-06,
      "loss": 0.147,
      "step": 4275
    },
    {
      "epoch": 2.84,
      "grad_norm": 1.7226457595825195,
      "learning_rate": 6.003157894736843e-06,
      "loss": 0.1474,
      "step": 4300
    },
    {
      "epoch": 2.86,
      "grad_norm": 1.82681405544281,
      "learning_rate": 5.976842105263159e-06,
      "loss": 0.1518,
      "step": 4325
    },
    {
      "epoch": 2.87,
      "grad_norm": 1.8655240535736084,
      "learning_rate": 5.950526315789474e-06,
      "loss": 0.1449,
      "step": 4350
    },
    {
      "epoch": 2.89,
      "grad_norm": 1.6956390142440796,
      "learning_rate": 5.92421052631579e-06,
      "loss": 0.149,
      "step": 4375
    },
    {
      "epoch": 2.91,
      "grad_norm": 1.6391839981079102,
      "learning_rate": 5.897894736842106e-06,
      "loss": 0.1494,
      "step": 4400
    },
    {
      "epoch": 2.92,
      "grad_norm": 1.7154439687728882,
      "learning_rate": 5.871578947368422e-06,
      "loss": 0.1464,
      "step": 4425
    },
    {
      "epoch": 2.94,
      "grad_norm": 1.708503246307373,
      "learning_rate": 5.845263157894737e-06,
      "loss": 0.1473,
      "step": 4450
    },
    {
      "epoch": 2.96,
      "grad_norm": 1.6649551391601562,
      "learning_rate": 5.818947368421053e-06,
      "loss": 0.1467,
      "step": 4475
    },
    {
      "epoch": 2.97,
      "grad_norm": 1.6294652223587036,
      "learning_rate": 5.7926315789473685e-06,
      "loss": 0.1485,
      "step": 4500
    },
    {
      "epoch": 2.97,
      "eval_loss": 0.3344377875328064,
      "eval_runtime": 14.5388,
      "eval_samples_per_second": 116.31,
      "eval_steps_per_second": 0.963,
      "eval_wer": 22.904958161608594,
      "step": 4500
    },
    {
      "epoch": 2.99,
      "grad_norm": 1.691629409790039,
      "learning_rate": 5.766315789473684e-06,
      "loss": 0.1448,
      "step": 4525
    },
    {
      "epoch": 3.01,
      "grad_norm": 1.4993510246276855,
      "learning_rate": 5.74e-06,
      "loss": 0.1406,
      "step": 4550
    },
    {
      "epoch": 3.02,
      "grad_norm": 1.5252876281738281,
      "learning_rate": 5.713684210526316e-06,
      "loss": 0.1366,
      "step": 4575
    },
    {
      "epoch": 3.04,
      "grad_norm": 1.4705634117126465,
      "learning_rate": 5.687368421052633e-06,
      "loss": 0.1355,
      "step": 4600
    },
    {
      "epoch": 3.05,
      "grad_norm": 1.661219596862793,
      "learning_rate": 5.661052631578948e-06,
      "loss": 0.1354,
      "step": 4625
    },
    {
      "epoch": 3.07,
      "grad_norm": 1.7826298475265503,
      "learning_rate": 5.634736842105264e-06,
      "loss": 0.1369,
      "step": 4650
    },
    {
      "epoch": 3.09,
      "grad_norm": 1.7129111289978027,
      "learning_rate": 5.60842105263158e-06,
      "loss": 0.1359,
      "step": 4675
    },
    {
      "epoch": 3.1,
      "grad_norm": 1.5526037216186523,
      "learning_rate": 5.582105263157895e-06,
      "loss": 0.1355,
      "step": 4700
    },
    {
      "epoch": 3.12,
      "grad_norm": 1.690327763557434,
      "learning_rate": 5.555789473684211e-06,
      "loss": 0.1346,
      "step": 4725
    },
    {
      "epoch": 3.14,
      "grad_norm": 1.5185409784317017,
      "learning_rate": 5.529473684210527e-06,
      "loss": 0.1323,
      "step": 4750
    },
    {
      "epoch": 3.15,
      "grad_norm": 1.617923617362976,
      "learning_rate": 5.5031578947368425e-06,
      "loss": 0.1366,
      "step": 4775
    },
    {
      "epoch": 3.17,
      "grad_norm": 1.6980522871017456,
      "learning_rate": 5.476842105263158e-06,
      "loss": 0.136,
      "step": 4800
    },
    {
      "epoch": 3.19,
      "grad_norm": 1.5043766498565674,
      "learning_rate": 5.450526315789474e-06,
      "loss": 0.1339,
      "step": 4825
    },
    {
      "epoch": 3.2,
      "grad_norm": 1.8342564105987549,
      "learning_rate": 5.42421052631579e-06,
      "loss": 0.135,
      "step": 4850
    },
    {
      "epoch": 3.22,
      "grad_norm": 1.4562426805496216,
      "learning_rate": 5.397894736842105e-06,
      "loss": 0.1359,
      "step": 4875
    },
    {
      "epoch": 3.24,
      "grad_norm": 1.6603587865829468,
      "learning_rate": 5.371578947368421e-06,
      "loss": 0.13,
      "step": 4900
    },
    {
      "epoch": 3.25,
      "grad_norm": 1.5824819803237915,
      "learning_rate": 5.345263157894738e-06,
      "loss": 0.1332,
      "step": 4925
    },
    {
      "epoch": 3.27,
      "grad_norm": 1.6331573724746704,
      "learning_rate": 5.318947368421053e-06,
      "loss": 0.132,
      "step": 4950
    },
    {
      "epoch": 3.29,
      "grad_norm": 1.6373783349990845,
      "learning_rate": 5.292631578947369e-06,
      "loss": 0.1376,
      "step": 4975
    },
    {
      "epoch": 3.3,
      "grad_norm": 1.617693305015564,
      "learning_rate": 5.266315789473685e-06,
      "loss": 0.131,
      "step": 5000
    },
    {
      "epoch": 3.3,
      "eval_loss": 0.3334658145904541,
      "eval_runtime": 14.2594,
      "eval_samples_per_second": 118.588,
      "eval_steps_per_second": 0.982,
      "eval_wer": 22.979892593980267,
      "step": 5000
    },
    {
      "epoch": 3.32,
      "grad_norm": 1.6260693073272705,
      "learning_rate": 5.240000000000001e-06,
      "loss": 0.1321,
      "step": 5025
    },
    {
      "epoch": 3.34,
      "grad_norm": 1.679488182067871,
      "learning_rate": 5.2136842105263165e-06,
      "loss": 0.1345,
      "step": 5050
    },
    {
      "epoch": 3.35,
      "grad_norm": 1.550234079360962,
      "learning_rate": 5.187368421052632e-06,
      "loss": 0.1299,
      "step": 5075
    },
    {
      "epoch": 3.37,
      "grad_norm": 1.5804076194763184,
      "learning_rate": 5.161052631578948e-06,
      "loss": 0.1348,
      "step": 5100
    },
    {
      "epoch": 3.39,
      "grad_norm": 1.7430262565612793,
      "learning_rate": 5.134736842105263e-06,
      "loss": 0.1329,
      "step": 5125
    },
    {
      "epoch": 3.4,
      "grad_norm": 1.6812238693237305,
      "learning_rate": 5.108421052631579e-06,
      "loss": 0.1343,
      "step": 5150
    },
    {
      "epoch": 3.42,
      "grad_norm": 1.6852658987045288,
      "learning_rate": 5.082105263157895e-06,
      "loss": 0.1321,
      "step": 5175
    },
    {
      "epoch": 3.43,
      "grad_norm": 1.7165955305099487,
      "learning_rate": 5.0557894736842105e-06,
      "loss": 0.133,
      "step": 5200
    },
    {
      "epoch": 3.45,
      "grad_norm": 1.722699522972107,
      "learning_rate": 5.029473684210526e-06,
      "loss": 0.129,
      "step": 5225
    },
    {
      "epoch": 3.47,
      "grad_norm": 1.7014281749725342,
      "learning_rate": 5.003157894736842e-06,
      "loss": 0.1325,
      "step": 5250
    },
    {
      "epoch": 3.48,
      "grad_norm": 1.8094806671142578,
      "learning_rate": 4.976842105263158e-06,
      "loss": 0.1312,
      "step": 5275
    },
    {
      "epoch": 3.5,
      "grad_norm": 1.600211262702942,
      "learning_rate": 4.950526315789474e-06,
      "loss": 0.1284,
      "step": 5300
    },
    {
      "epoch": 3.52,
      "grad_norm": 1.5237622261047363,
      "learning_rate": 4.92421052631579e-06,
      "loss": 0.1297,
      "step": 5325
    },
    {
      "epoch": 3.53,
      "grad_norm": 1.542418122291565,
      "learning_rate": 4.897894736842106e-06,
      "loss": 0.1317,
      "step": 5350
    },
    {
      "epoch": 3.55,
      "grad_norm": 1.639620065689087,
      "learning_rate": 4.871578947368421e-06,
      "loss": 0.132,
      "step": 5375
    },
    {
      "epoch": 3.57,
      "grad_norm": 1.6733545064926147,
      "learning_rate": 4.845263157894737e-06,
      "loss": 0.1308,
      "step": 5400
    },
    {
      "epoch": 3.58,
      "grad_norm": 1.700808048248291,
      "learning_rate": 4.818947368421053e-06,
      "loss": 0.1305,
      "step": 5425
    },
    {
      "epoch": 3.6,
      "grad_norm": 1.4252574443817139,
      "learning_rate": 4.792631578947369e-06,
      "loss": 0.1296,
      "step": 5450
    },
    {
      "epoch": 3.62,
      "grad_norm": 1.7240992784500122,
      "learning_rate": 4.7663157894736845e-06,
      "loss": 0.1339,
      "step": 5475
    },
    {
      "epoch": 3.63,
      "grad_norm": 1.6297061443328857,
      "learning_rate": 4.74e-06,
      "loss": 0.1294,
      "step": 5500
    },
    {
      "epoch": 3.63,
      "eval_loss": 0.3302935063838959,
      "eval_runtime": 17.4619,
      "eval_samples_per_second": 96.839,
      "eval_steps_per_second": 0.802,
      "eval_wer": 23.179717746971402,
      "step": 5500
    },
    {
      "epoch": 3.65,
      "grad_norm": 1.5596210956573486,
      "learning_rate": 4.713684210526316e-06,
      "loss": 0.1296,
      "step": 5525
    },
    {
      "epoch": 3.67,
      "grad_norm": 1.6876928806304932,
      "learning_rate": 4.687368421052632e-06,
      "loss": 0.1268,
      "step": 5550
    },
    {
      "epoch": 3.68,
      "grad_norm": 1.5455940961837769,
      "learning_rate": 4.661052631578948e-06,
      "loss": 0.129,
      "step": 5575
    },
    {
      "epoch": 3.7,
      "grad_norm": 1.594913125038147,
      "learning_rate": 4.634736842105264e-06,
      "loss": 0.1274,
      "step": 5600
    },
    {
      "epoch": 3.72,
      "grad_norm": 1.5149903297424316,
      "learning_rate": 4.6084210526315794e-06,
      "loss": 0.1306,
      "step": 5625
    },
    {
      "epoch": 3.73,
      "grad_norm": 1.661870002746582,
      "learning_rate": 4.582105263157895e-06,
      "loss": 0.1334,
      "step": 5650
    },
    {
      "epoch": 3.75,
      "grad_norm": 1.7134344577789307,
      "learning_rate": 4.55578947368421e-06,
      "loss": 0.1264,
      "step": 5675
    },
    {
      "epoch": 3.76,
      "grad_norm": 1.5931401252746582,
      "learning_rate": 4.529473684210527e-06,
      "loss": 0.1291,
      "step": 5700
    },
    {
      "epoch": 3.78,
      "grad_norm": 1.6768360137939453,
      "learning_rate": 4.503157894736843e-06,
      "loss": 0.128,
      "step": 5725
    },
    {
      "epoch": 3.8,
      "grad_norm": 1.5798521041870117,
      "learning_rate": 4.4768421052631585e-06,
      "loss": 0.1255,
      "step": 5750
    },
    {
      "epoch": 3.81,
      "grad_norm": 1.5002096891403198,
      "learning_rate": 4.450526315789474e-06,
      "loss": 0.1275,
      "step": 5775
    },
    {
      "epoch": 3.83,
      "grad_norm": 1.6101495027542114,
      "learning_rate": 4.424210526315789e-06,
      "loss": 0.1279,
      "step": 5800
    },
    {
      "epoch": 3.85,
      "grad_norm": 1.5645090341567993,
      "learning_rate": 4.397894736842105e-06,
      "loss": 0.1337,
      "step": 5825
    },
    {
      "epoch": 3.86,
      "grad_norm": 1.6023365259170532,
      "learning_rate": 4.371578947368421e-06,
      "loss": 0.1285,
      "step": 5850
    },
    {
      "epoch": 3.88,
      "grad_norm": 1.4570146799087524,
      "learning_rate": 4.345263157894738e-06,
      "loss": 0.1279,
      "step": 5875
    },
    {
      "epoch": 3.9,
      "grad_norm": 1.6708632707595825,
      "learning_rate": 4.3189473684210535e-06,
      "loss": 0.1276,
      "step": 5900
    },
    {
      "epoch": 3.91,
      "grad_norm": 1.553534746170044,
      "learning_rate": 4.2926315789473684e-06,
      "loss": 0.1311,
      "step": 5925
    },
    {
      "epoch": 3.93,
      "grad_norm": 1.6338897943496704,
      "learning_rate": 4.266315789473684e-06,
      "loss": 0.1258,
      "step": 5950
    },
    {
      "epoch": 3.95,
      "grad_norm": 1.5521620512008667,
      "learning_rate": 4.24e-06,
      "loss": 0.1278,
      "step": 5975
    },
    {
      "epoch": 3.96,
      "grad_norm": 1.508652925491333,
      "learning_rate": 4.213684210526316e-06,
      "loss": 0.1287,
      "step": 6000
    },
    {
      "epoch": 3.96,
      "eval_loss": 0.3290025293827057,
      "eval_runtime": 15.8008,
      "eval_samples_per_second": 107.02,
      "eval_steps_per_second": 0.886,
      "eval_wer": 23.073560634444863,
      "step": 6000
    },
    {
      "epoch": 3.98,
      "grad_norm": 1.6393396854400635,
      "learning_rate": 4.1873684210526325e-06,
      "loss": 0.1276,
      "step": 6025
    },
    {
      "epoch": 4.0,
      "grad_norm": 1.571065902709961,
      "learning_rate": 4.1610526315789475e-06,
      "loss": 0.1261,
      "step": 6050
    },
    {
      "epoch": 4.01,
      "grad_norm": 1.5454843044281006,
      "learning_rate": 4.134736842105263e-06,
      "loss": 0.12,
      "step": 6075
    },
    {
      "epoch": 4.03,
      "grad_norm": 1.5475822687149048,
      "learning_rate": 4.108421052631579e-06,
      "loss": 0.1226,
      "step": 6100
    },
    {
      "epoch": 4.05,
      "grad_norm": 1.4355708360671997,
      "learning_rate": 4.082105263157895e-06,
      "loss": 0.1216,
      "step": 6125
    },
    {
      "epoch": 4.06,
      "grad_norm": 1.546095848083496,
      "learning_rate": 4.055789473684211e-06,
      "loss": 0.1202,
      "step": 6150
    },
    {
      "epoch": 4.08,
      "grad_norm": 1.5846285820007324,
      "learning_rate": 4.029473684210527e-06,
      "loss": 0.1174,
      "step": 6175
    },
    {
      "epoch": 4.1,
      "grad_norm": 1.6201484203338623,
      "learning_rate": 4.0031578947368424e-06,
      "loss": 0.1202,
      "step": 6200
    },
    {
      "epoch": 4.11,
      "grad_norm": 1.5635933876037598,
      "learning_rate": 3.976842105263158e-06,
      "loss": 0.1179,
      "step": 6225
    },
    {
      "epoch": 4.13,
      "grad_norm": 1.5160921812057495,
      "learning_rate": 3.950526315789474e-06,
      "loss": 0.1184,
      "step": 6250
    },
    {
      "epoch": 4.14,
      "grad_norm": 1.8259953260421753,
      "learning_rate": 3.92421052631579e-06,
      "loss": 0.1167,
      "step": 6275
    },
    {
      "epoch": 4.16,
      "grad_norm": 1.5983636379241943,
      "learning_rate": 3.897894736842106e-06,
      "loss": 0.1214,
      "step": 6300
    },
    {
      "epoch": 4.18,
      "grad_norm": 1.5877853631973267,
      "learning_rate": 3.8715789473684215e-06,
      "loss": 0.1188,
      "step": 6325
    },
    {
      "epoch": 4.19,
      "grad_norm": 1.5386959314346313,
      "learning_rate": 3.845263157894737e-06,
      "loss": 0.1196,
      "step": 6350
    },
    {
      "epoch": 4.21,
      "grad_norm": 1.8094242811203003,
      "learning_rate": 3.818947368421053e-06,
      "loss": 0.1199,
      "step": 6375
    },
    {
      "epoch": 4.23,
      "grad_norm": 1.5226629972457886,
      "learning_rate": 3.792631578947369e-06,
      "loss": 0.1176,
      "step": 6400
    },
    {
      "epoch": 4.24,
      "grad_norm": 1.605747938156128,
      "learning_rate": 3.766315789473685e-06,
      "loss": 0.1166,
      "step": 6425
    },
    {
      "epoch": 4.26,
      "grad_norm": 1.4122657775878906,
      "learning_rate": 3.74e-06,
      "loss": 0.1193,
      "step": 6450
    },
    {
      "epoch": 4.28,
      "grad_norm": 1.538041114807129,
      "learning_rate": 3.713684210526316e-06,
      "loss": 0.1162,
      "step": 6475
    },
    {
      "epoch": 4.29,
      "grad_norm": 1.5662447214126587,
      "learning_rate": 3.687368421052632e-06,
      "loss": 0.1153,
      "step": 6500
    },
    {
      "epoch": 4.29,
      "eval_loss": 0.3267483413219452,
      "eval_runtime": 14.5761,
      "eval_samples_per_second": 116.012,
      "eval_steps_per_second": 0.96,
      "eval_wer": 22.061945797427253,
      "step": 6500
    },
    {
      "epoch": 4.31,
      "grad_norm": 1.511885643005371,
      "learning_rate": 3.6610526315789472e-06,
      "loss": 0.1197,
      "step": 6525
    },
    {
      "epoch": 4.33,
      "grad_norm": 1.5087144374847412,
      "learning_rate": 3.6347368421052635e-06,
      "loss": 0.1214,
      "step": 6550
    },
    {
      "epoch": 4.34,
      "grad_norm": 1.5032167434692383,
      "learning_rate": 3.6084210526315793e-06,
      "loss": 0.1177,
      "step": 6575
    },
    {
      "epoch": 4.36,
      "grad_norm": 1.4949249029159546,
      "learning_rate": 3.582105263157895e-06,
      "loss": 0.1209,
      "step": 6600
    },
    {
      "epoch": 4.38,
      "grad_norm": 1.6159617900848389,
      "learning_rate": 3.555789473684211e-06,
      "loss": 0.1204,
      "step": 6625
    },
    {
      "epoch": 4.39,
      "grad_norm": 1.5025402307510376,
      "learning_rate": 3.5294736842105263e-06,
      "loss": 0.1143,
      "step": 6650
    },
    {
      "epoch": 4.41,
      "grad_norm": 1.4929492473602295,
      "learning_rate": 3.503157894736842e-06,
      "loss": 0.12,
      "step": 6675
    },
    {
      "epoch": 4.43,
      "grad_norm": 1.6627602577209473,
      "learning_rate": 3.476842105263158e-06,
      "loss": 0.1176,
      "step": 6700
    },
    {
      "epoch": 4.44,
      "grad_norm": 1.5767234563827515,
      "learning_rate": 3.450526315789474e-06,
      "loss": 0.1158,
      "step": 6725
    },
    {
      "epoch": 4.46,
      "grad_norm": 1.4062097072601318,
      "learning_rate": 3.42421052631579e-06,
      "loss": 0.118,
      "step": 6750
    },
    {
      "epoch": 4.47,
      "grad_norm": 1.5933609008789062,
      "learning_rate": 3.3978947368421054e-06,
      "loss": 0.1178,
      "step": 6775
    },
    {
      "epoch": 4.49,
      "grad_norm": 1.4828532934188843,
      "learning_rate": 3.3715789473684212e-06,
      "loss": 0.1163,
      "step": 6800
    },
    {
      "epoch": 4.51,
      "grad_norm": 1.4716224670410156,
      "learning_rate": 3.345263157894737e-06,
      "loss": 0.1154,
      "step": 6825
    },
    {
      "epoch": 4.52,
      "grad_norm": 1.5611412525177002,
      "learning_rate": 3.318947368421053e-06,
      "loss": 0.1181,
      "step": 6850
    },
    {
      "epoch": 4.54,
      "grad_norm": 1.626551866531372,
      "learning_rate": 3.292631578947369e-06,
      "loss": 0.1189,
      "step": 6875
    },
    {
      "epoch": 4.56,
      "grad_norm": 1.6640559434890747,
      "learning_rate": 3.2663157894736845e-06,
      "loss": 0.12,
      "step": 6900
    },
    {
      "epoch": 4.57,
      "grad_norm": 1.5578954219818115,
      "learning_rate": 3.2400000000000003e-06,
      "loss": 0.1151,
      "step": 6925
    },
    {
      "epoch": 4.59,
      "grad_norm": 1.4524513483047485,
      "learning_rate": 3.213684210526316e-06,
      "loss": 0.1124,
      "step": 6950
    },
    {
      "epoch": 4.61,
      "grad_norm": 1.5898208618164062,
      "learning_rate": 3.187368421052632e-06,
      "loss": 0.1202,
      "step": 6975
    },
    {
      "epoch": 4.62,
      "grad_norm": 1.52357816696167,
      "learning_rate": 3.1610526315789474e-06,
      "loss": 0.1176,
      "step": 7000
    },
    {
      "epoch": 4.62,
      "eval_loss": 0.3240759074687958,
      "eval_runtime": 18.5252,
      "eval_samples_per_second": 91.281,
      "eval_steps_per_second": 0.756,
      "eval_wer": 24.116398151617336,
      "step": 7000
    },
    {
      "epoch": 4.64,
      "grad_norm": 1.4377552270889282,
      "learning_rate": 3.134736842105263e-06,
      "loss": 0.1189,
      "step": 7025
    },
    {
      "epoch": 4.66,
      "grad_norm": 1.679721713066101,
      "learning_rate": 3.1084210526315794e-06,
      "loss": 0.1165,
      "step": 7050
    },
    {
      "epoch": 4.67,
      "grad_norm": 1.7221696376800537,
      "learning_rate": 3.0821052631578952e-06,
      "loss": 0.1182,
      "step": 7075
    },
    {
      "epoch": 4.69,
      "grad_norm": 1.5595306158065796,
      "learning_rate": 3.055789473684211e-06,
      "loss": 0.1154,
      "step": 7100
    },
    {
      "epoch": 4.71,
      "grad_norm": 1.5186105966567993,
      "learning_rate": 3.0294736842105264e-06,
      "loss": 0.1167,
      "step": 7125
    },
    {
      "epoch": 4.72,
      "grad_norm": 1.6027384996414185,
      "learning_rate": 3.0031578947368423e-06,
      "loss": 0.1142,
      "step": 7150
    },
    {
      "epoch": 4.74,
      "grad_norm": 1.4588899612426758,
      "learning_rate": 2.976842105263158e-06,
      "loss": 0.1149,
      "step": 7175
    },
    {
      "epoch": 4.76,
      "grad_norm": 1.6083778142929077,
      "learning_rate": 2.9505263157894735e-06,
      "loss": 0.1152,
      "step": 7200
    },
    {
      "epoch": 4.77,
      "grad_norm": 1.3121345043182373,
      "learning_rate": 2.92421052631579e-06,
      "loss": 0.114,
      "step": 7225
    },
    {
      "epoch": 4.79,
      "grad_norm": 1.5029678344726562,
      "learning_rate": 2.8978947368421055e-06,
      "loss": 0.1175,
      "step": 7250
    },
    {
      "epoch": 4.81,
      "grad_norm": 1.5415253639221191,
      "learning_rate": 2.8715789473684214e-06,
      "loss": 0.1158,
      "step": 7275
    },
    {
      "epoch": 4.82,
      "grad_norm": 1.5756951570510864,
      "learning_rate": 2.845263157894737e-06,
      "loss": 0.1188,
      "step": 7300
    },
    {
      "epoch": 4.84,
      "grad_norm": 1.5500506162643433,
      "learning_rate": 2.8189473684210526e-06,
      "loss": 0.1138,
      "step": 7325
    },
    {
      "epoch": 4.85,
      "grad_norm": 1.4593719244003296,
      "learning_rate": 2.7926315789473684e-06,
      "loss": 0.1147,
      "step": 7350
    },
    {
      "epoch": 4.87,
      "grad_norm": 1.521674633026123,
      "learning_rate": 2.766315789473684e-06,
      "loss": 0.1166,
      "step": 7375
    },
    {
      "epoch": 4.89,
      "grad_norm": 1.594332218170166,
      "learning_rate": 2.7400000000000004e-06,
      "loss": 0.1156,
      "step": 7400
    },
    {
      "epoch": 4.9,
      "grad_norm": 1.5146125555038452,
      "learning_rate": 2.7136842105263163e-06,
      "loss": 0.1174,
      "step": 7425
    },
    {
      "epoch": 4.92,
      "grad_norm": 1.5220681428909302,
      "learning_rate": 2.6873684210526317e-06,
      "loss": 0.1151,
      "step": 7450
    },
    {
      "epoch": 4.94,
      "grad_norm": 1.5523508787155151,
      "learning_rate": 2.6610526315789475e-06,
      "loss": 0.1192,
      "step": 7475
    },
    {
      "epoch": 4.95,
      "grad_norm": 1.439876914024353,
      "learning_rate": 2.6347368421052633e-06,
      "loss": 0.1153,
      "step": 7500
    },
    {
      "epoch": 4.95,
      "eval_loss": 0.32601335644721985,
      "eval_runtime": 20.8661,
      "eval_samples_per_second": 81.041,
      "eval_steps_per_second": 0.671,
      "eval_wer": 23.960284750843012,
      "step": 7500
    },
    {
      "epoch": 4.97,
      "grad_norm": 1.5585688352584839,
      "learning_rate": 2.608421052631579e-06,
      "loss": 0.1204,
      "step": 7525
    },
    {
      "epoch": 4.99,
      "grad_norm": 1.5496189594268799,
      "learning_rate": 2.5821052631578954e-06,
      "loss": 0.1171,
      "step": 7550
    },
    {
      "epoch": 5.0,
      "grad_norm": 1.4494433403015137,
      "learning_rate": 2.5557894736842108e-06,
      "loss": 0.1143,
      "step": 7575
    },
    {
      "epoch": 5.02,
      "grad_norm": 1.5850895643234253,
      "learning_rate": 2.5294736842105266e-06,
      "loss": 0.1077,
      "step": 7600
    },
    {
      "epoch": 5.04,
      "grad_norm": 1.3688222169876099,
      "learning_rate": 2.5031578947368424e-06,
      "loss": 0.1095,
      "step": 7625
    },
    {
      "epoch": 5.05,
      "grad_norm": 1.4580186605453491,
      "learning_rate": 2.476842105263158e-06,
      "loss": 0.1111,
      "step": 7650
    },
    {
      "epoch": 5.07,
      "grad_norm": 1.3927934169769287,
      "learning_rate": 2.4505263157894736e-06,
      "loss": 0.1077,
      "step": 7675
    },
    {
      "epoch": 5.09,
      "grad_norm": 1.497953176498413,
      "learning_rate": 2.42421052631579e-06,
      "loss": 0.1105,
      "step": 7700
    },
    {
      "epoch": 5.1,
      "grad_norm": 1.5126893520355225,
      "learning_rate": 2.3978947368421052e-06,
      "loss": 0.1065,
      "step": 7725
    },
    {
      "epoch": 5.12,
      "grad_norm": 1.4988917112350464,
      "learning_rate": 2.371578947368421e-06,
      "loss": 0.1132,
      "step": 7750
    },
    {
      "epoch": 5.14,
      "grad_norm": 1.4799151420593262,
      "learning_rate": 2.3452631578947373e-06,
      "loss": 0.1078,
      "step": 7775
    },
    {
      "epoch": 5.15,
      "grad_norm": 1.4531598091125488,
      "learning_rate": 2.3189473684210527e-06,
      "loss": 0.1086,
      "step": 7800
    },
    {
      "epoch": 5.17,
      "grad_norm": 1.437127709388733,
      "learning_rate": 2.2926315789473685e-06,
      "loss": 0.1109,
      "step": 7825
    },
    {
      "epoch": 5.18,
      "grad_norm": 1.5070149898529053,
      "learning_rate": 2.2663157894736843e-06,
      "loss": 0.1086,
      "step": 7850
    },
    {
      "epoch": 5.2,
      "grad_norm": 1.4844095706939697,
      "learning_rate": 2.24e-06,
      "loss": 0.1103,
      "step": 7875
    },
    {
      "epoch": 5.22,
      "grad_norm": 1.5023523569107056,
      "learning_rate": 2.213684210526316e-06,
      "loss": 0.1103,
      "step": 7900
    },
    {
      "epoch": 5.23,
      "grad_norm": 1.4412981271743774,
      "learning_rate": 2.187368421052632e-06,
      "loss": 0.1084,
      "step": 7925
    },
    {
      "epoch": 5.25,
      "grad_norm": 1.438542127609253,
      "learning_rate": 2.1610526315789476e-06,
      "loss": 0.1091,
      "step": 7950
    },
    {
      "epoch": 5.27,
      "grad_norm": 1.4971038103103638,
      "learning_rate": 2.1347368421052634e-06,
      "loss": 0.1113,
      "step": 7975
    },
    {
      "epoch": 5.28,
      "grad_norm": 1.4342187643051147,
      "learning_rate": 2.1084210526315792e-06,
      "loss": 0.1093,
      "step": 8000
    },
    {
      "epoch": 5.28,
      "eval_loss": 0.32474491000175476,
      "eval_runtime": 18.2289,
      "eval_samples_per_second": 92.765,
      "eval_steps_per_second": 0.768,
      "eval_wer": 23.854127638316474,
      "step": 8000
    },
    {
      "epoch": 5.3,
      "grad_norm": 1.5157432556152344,
      "learning_rate": 2.082105263157895e-06,
      "loss": 0.11,
      "step": 8025
    },
    {
      "epoch": 5.32,
      "grad_norm": 1.5796202421188354,
      "learning_rate": 2.055789473684211e-06,
      "loss": 0.1101,
      "step": 8050
    },
    {
      "epoch": 5.33,
      "grad_norm": 1.3845866918563843,
      "learning_rate": 2.0294736842105263e-06,
      "loss": 0.1076,
      "step": 8075
    },
    {
      "epoch": 5.35,
      "grad_norm": 1.5526548624038696,
      "learning_rate": 2.003157894736842e-06,
      "loss": 0.1107,
      "step": 8100
    },
    {
      "epoch": 5.37,
      "grad_norm": 1.3970515727996826,
      "learning_rate": 1.976842105263158e-06,
      "loss": 0.1083,
      "step": 8125
    },
    {
      "epoch": 5.38,
      "grad_norm": 1.5313230752944946,
      "learning_rate": 1.9505263157894737e-06,
      "loss": 0.1096,
      "step": 8150
    },
    {
      "epoch": 5.4,
      "grad_norm": 1.4472230672836304,
      "learning_rate": 1.9242105263157896e-06,
      "loss": 0.1109,
      "step": 8175
    },
    {
      "epoch": 5.42,
      "grad_norm": 1.5823153257369995,
      "learning_rate": 1.8978947368421056e-06,
      "loss": 0.1094,
      "step": 8200
    },
    {
      "epoch": 5.43,
      "grad_norm": 1.5765074491500854,
      "learning_rate": 1.8715789473684212e-06,
      "loss": 0.1103,
      "step": 8225
    },
    {
      "epoch": 5.45,
      "grad_norm": 1.442234754562378,
      "learning_rate": 1.845263157894737e-06,
      "loss": 0.1084,
      "step": 8250
    },
    {
      "epoch": 5.47,
      "grad_norm": 1.40919828414917,
      "learning_rate": 1.8189473684210528e-06,
      "loss": 0.1068,
      "step": 8275
    },
    {
      "epoch": 5.48,
      "grad_norm": 1.5205286741256714,
      "learning_rate": 1.7926315789473686e-06,
      "loss": 0.1069,
      "step": 8300
    },
    {
      "epoch": 5.5,
      "grad_norm": 1.5483301877975464,
      "learning_rate": 1.7663157894736843e-06,
      "loss": 0.1123,
      "step": 8325
    },
    {
      "epoch": 5.52,
      "grad_norm": 1.5091420412063599,
      "learning_rate": 1.74e-06,
      "loss": 0.1071,
      "step": 8350
    },
    {
      "epoch": 5.53,
      "grad_norm": 1.4874048233032227,
      "learning_rate": 1.713684210526316e-06,
      "loss": 0.1113,
      "step": 8375
    },
    {
      "epoch": 5.55,
      "grad_norm": 1.3622137308120728,
      "learning_rate": 1.6873684210526317e-06,
      "loss": 0.1085,
      "step": 8400
    },
    {
      "epoch": 5.56,
      "grad_norm": 1.591680884361267,
      "learning_rate": 1.6610526315789473e-06,
      "loss": 0.1104,
      "step": 8425
    },
    {
      "epoch": 5.58,
      "grad_norm": 1.479125738143921,
      "learning_rate": 1.6347368421052633e-06,
      "loss": 0.1097,
      "step": 8450
    },
    {
      "epoch": 5.6,
      "grad_norm": 1.5606906414031982,
      "learning_rate": 1.6084210526315792e-06,
      "loss": 0.1082,
      "step": 8475
    },
    {
      "epoch": 5.61,
      "grad_norm": 1.5121188163757324,
      "learning_rate": 1.5821052631578948e-06,
      "loss": 0.1129,
      "step": 8500
    },
    {
      "epoch": 5.61,
      "eval_loss": 0.3241681754589081,
      "eval_runtime": 17.5037,
      "eval_samples_per_second": 96.608,
      "eval_steps_per_second": 0.8,
      "eval_wer": 22.848757337329836,
      "step": 8500
    },
    {
      "epoch": 5.63,
      "grad_norm": 1.3083462715148926,
      "learning_rate": 1.5557894736842106e-06,
      "loss": 0.1073,
      "step": 8525
    },
    {
      "epoch": 5.65,
      "grad_norm": 1.4445217847824097,
      "learning_rate": 1.5294736842105264e-06,
      "loss": 0.108,
      "step": 8550
    },
    {
      "epoch": 5.66,
      "grad_norm": 1.4660565853118896,
      "learning_rate": 1.5031578947368422e-06,
      "loss": 0.1087,
      "step": 8575
    },
    {
      "epoch": 5.68,
      "grad_norm": 1.5583314895629883,
      "learning_rate": 1.4768421052631578e-06,
      "loss": 0.1044,
      "step": 8600
    },
    {
      "epoch": 5.7,
      "grad_norm": 1.4255391359329224,
      "learning_rate": 1.4505263157894739e-06,
      "loss": 0.1071,
      "step": 8625
    },
    {
      "epoch": 5.71,
      "grad_norm": 1.5511982440948486,
      "learning_rate": 1.4242105263157897e-06,
      "loss": 0.108,
      "step": 8650
    },
    {
      "epoch": 5.73,
      "grad_norm": 1.5022433996200562,
      "learning_rate": 1.3978947368421053e-06,
      "loss": 0.1092,
      "step": 8675
    },
    {
      "epoch": 5.75,
      "grad_norm": 1.5003283023834229,
      "learning_rate": 1.3715789473684213e-06,
      "loss": 0.1066,
      "step": 8700
    },
    {
      "epoch": 5.76,
      "grad_norm": 1.5088740587234497,
      "learning_rate": 1.345263157894737e-06,
      "loss": 0.1055,
      "step": 8725
    },
    {
      "epoch": 5.78,
      "grad_norm": 1.563700556755066,
      "learning_rate": 1.3189473684210527e-06,
      "loss": 0.1069,
      "step": 8750
    },
    {
      "epoch": 5.8,
      "grad_norm": 1.4417468309402466,
      "learning_rate": 1.2926315789473683e-06,
      "loss": 0.1061,
      "step": 8775
    },
    {
      "epoch": 5.81,
      "grad_norm": 1.412596583366394,
      "learning_rate": 1.2663157894736844e-06,
      "loss": 0.1095,
      "step": 8800
    },
    {
      "epoch": 5.83,
      "grad_norm": 1.4061535596847534,
      "learning_rate": 1.2400000000000002e-06,
      "loss": 0.1099,
      "step": 8825
    },
    {
      "epoch": 5.85,
      "grad_norm": 1.4590774774551392,
      "learning_rate": 1.213684210526316e-06,
      "loss": 0.1089,
      "step": 8850
    },
    {
      "epoch": 5.86,
      "grad_norm": 1.417046070098877,
      "learning_rate": 1.1873684210526316e-06,
      "loss": 0.1097,
      "step": 8875
    },
    {
      "epoch": 5.88,
      "grad_norm": 1.5805013179779053,
      "learning_rate": 1.1610526315789474e-06,
      "loss": 0.1122,
      "step": 8900
    },
    {
      "epoch": 5.89,
      "grad_norm": 1.562596321105957,
      "learning_rate": 1.1347368421052633e-06,
      "loss": 0.1085,
      "step": 8925
    },
    {
      "epoch": 5.91,
      "grad_norm": 1.5319128036499023,
      "learning_rate": 1.108421052631579e-06,
      "loss": 0.112,
      "step": 8950
    },
    {
      "epoch": 5.93,
      "grad_norm": 1.468923568725586,
      "learning_rate": 1.082105263157895e-06,
      "loss": 0.1098,
      "step": 8975
    },
    {
      "epoch": 5.94,
      "grad_norm": 1.4811102151870728,
      "learning_rate": 1.0557894736842105e-06,
      "loss": 0.1069,
      "step": 9000
    },
    {
      "epoch": 5.94,
      "eval_loss": 0.3231496810913086,
      "eval_runtime": 16.5945,
      "eval_samples_per_second": 101.901,
      "eval_steps_per_second": 0.844,
      "eval_wer": 22.41163981516173,
      "step": 9000
    },
    {
      "epoch": 5.96,
      "grad_norm": 1.4732502698898315,
      "learning_rate": 1.0294736842105265e-06,
      "loss": 0.1115,
      "step": 9025
    },
    {
      "epoch": 5.98,
      "grad_norm": 1.5112581253051758,
      "learning_rate": 1.0031578947368421e-06,
      "loss": 0.1067,
      "step": 9050
    },
    {
      "epoch": 5.99,
      "grad_norm": 1.4593307971954346,
      "learning_rate": 9.76842105263158e-07,
      "loss": 0.1092,
      "step": 9075
    },
    {
      "epoch": 6.01,
      "grad_norm": 1.3361061811447144,
      "learning_rate": 9.505263157894738e-07,
      "loss": 0.1077,
      "step": 9100
    },
    {
      "epoch": 6.03,
      "grad_norm": 1.5821435451507568,
      "learning_rate": 9.242105263157895e-07,
      "loss": 0.1023,
      "step": 9125
    },
    {
      "epoch": 6.04,
      "grad_norm": 1.3687916994094849,
      "learning_rate": 8.978947368421054e-07,
      "loss": 0.1012,
      "step": 9150
    },
    {
      "epoch": 6.06,
      "grad_norm": 1.5094650983810425,
      "learning_rate": 8.715789473684211e-07,
      "loss": 0.1,
      "step": 9175
    },
    {
      "epoch": 6.08,
      "grad_norm": 1.3472874164581299,
      "learning_rate": 8.452631578947369e-07,
      "loss": 0.1047,
      "step": 9200
    },
    {
      "epoch": 6.09,
      "grad_norm": 1.446553111076355,
      "learning_rate": 8.189473684210527e-07,
      "loss": 0.1041,
      "step": 9225
    },
    {
      "epoch": 6.11,
      "grad_norm": 1.5492606163024902,
      "learning_rate": 7.926315789473685e-07,
      "loss": 0.1081,
      "step": 9250
    },
    {
      "epoch": 6.13,
      "grad_norm": 1.4887338876724243,
      "learning_rate": 7.663157894736843e-07,
      "loss": 0.1071,
      "step": 9275
    },
    {
      "epoch": 6.14,
      "grad_norm": 1.290645718574524,
      "learning_rate": 7.4e-07,
      "loss": 0.1025,
      "step": 9300
    },
    {
      "epoch": 6.16,
      "grad_norm": 1.4855327606201172,
      "learning_rate": 7.136842105263158e-07,
      "loss": 0.1032,
      "step": 9325
    },
    {
      "epoch": 6.18,
      "grad_norm": 1.4177087545394897,
      "learning_rate": 6.873684210526315e-07,
      "loss": 0.105,
      "step": 9350
    },
    {
      "epoch": 6.19,
      "grad_norm": 1.3693170547485352,
      "learning_rate": 6.610526315789475e-07,
      "loss": 0.1064,
      "step": 9375
    },
    {
      "epoch": 6.21,
      "grad_norm": 1.5773143768310547,
      "learning_rate": 6.347368421052633e-07,
      "loss": 0.1036,
      "step": 9400
    },
    {
      "epoch": 6.23,
      "grad_norm": 1.4657046794891357,
      "learning_rate": 6.08421052631579e-07,
      "loss": 0.1046,
      "step": 9425
    },
    {
      "epoch": 6.24,
      "grad_norm": 1.4977328777313232,
      "learning_rate": 5.821052631578948e-07,
      "loss": 0.1021,
      "step": 9450
    },
    {
      "epoch": 6.26,
      "grad_norm": 1.3914306163787842,
      "learning_rate": 5.557894736842106e-07,
      "loss": 0.0998,
      "step": 9475
    },
    {
      "epoch": 6.27,
      "grad_norm": 1.487586498260498,
      "learning_rate": 5.294736842105263e-07,
      "loss": 0.1029,
      "step": 9500
    },
    {
      "epoch": 6.27,
      "eval_loss": 0.3232876658439636,
      "eval_runtime": 16.4967,
      "eval_samples_per_second": 102.505,
      "eval_steps_per_second": 0.849,
      "eval_wer": 22.567753215936058,
      "step": 9500
    },
    {
      "epoch": 6.29,
      "grad_norm": 1.5918724536895752,
      "learning_rate": 5.031578947368422e-07,
      "loss": 0.1054,
      "step": 9525
    },
    {
      "epoch": 6.31,
      "grad_norm": 1.422078251838684,
      "learning_rate": 4.768421052631579e-07,
      "loss": 0.1055,
      "step": 9550
    },
    {
      "epoch": 6.32,
      "grad_norm": 1.4294945001602173,
      "learning_rate": 4.505263157894737e-07,
      "loss": 0.1043,
      "step": 9575
    },
    {
      "epoch": 6.34,
      "grad_norm": 1.469376802444458,
      "learning_rate": 4.242105263157895e-07,
      "loss": 0.1032,
      "step": 9600
    },
    {
      "epoch": 6.36,
      "grad_norm": 1.555346965789795,
      "learning_rate": 3.9789473684210533e-07,
      "loss": 0.1055,
      "step": 9625
    },
    {
      "epoch": 6.37,
      "grad_norm": 1.4565141201019287,
      "learning_rate": 3.715789473684211e-07,
      "loss": 0.1079,
      "step": 9650
    },
    {
      "epoch": 6.39,
      "grad_norm": 1.4277021884918213,
      "learning_rate": 3.4526315789473686e-07,
      "loss": 0.1034,
      "step": 9675
    },
    {
      "epoch": 6.41,
      "grad_norm": 1.5641932487487793,
      "learning_rate": 3.1894736842105263e-07,
      "loss": 0.1065,
      "step": 9700
    },
    {
      "epoch": 6.42,
      "grad_norm": 1.4544775485992432,
      "learning_rate": 2.9263157894736845e-07,
      "loss": 0.1026,
      "step": 9725
    },
    {
      "epoch": 6.44,
      "grad_norm": 1.44139564037323,
      "learning_rate": 2.663157894736842e-07,
      "loss": 0.1059,
      "step": 9750
    },
    {
      "epoch": 6.46,
      "grad_norm": 1.3898285627365112,
      "learning_rate": 2.4000000000000003e-07,
      "loss": 0.104,
      "step": 9775
    },
    {
      "epoch": 6.47,
      "grad_norm": 1.4317445755004883,
      "learning_rate": 2.136842105263158e-07,
      "loss": 0.1051,
      "step": 9800
    },
    {
      "epoch": 6.49,
      "grad_norm": 1.672619342803955,
      "learning_rate": 1.8736842105263162e-07,
      "loss": 0.1057,
      "step": 9825
    },
    {
      "epoch": 6.51,
      "grad_norm": 1.528280258178711,
      "learning_rate": 1.6105263157894738e-07,
      "loss": 0.1048,
      "step": 9850
    },
    {
      "epoch": 6.52,
      "grad_norm": 1.4008196592330933,
      "learning_rate": 1.3473684210526317e-07,
      "loss": 0.1048,
      "step": 9875
    },
    {
      "epoch": 6.54,
      "grad_norm": 1.4791390895843506,
      "learning_rate": 1.0842105263157895e-07,
      "loss": 0.1039,
      "step": 9900
    },
    {
      "epoch": 6.56,
      "grad_norm": 1.4586882591247559,
      "learning_rate": 8.210526315789474e-08,
      "loss": 0.1057,
      "step": 9925
    },
    {
      "epoch": 6.57,
      "grad_norm": 1.4133316278457642,
      "learning_rate": 5.578947368421053e-08,
      "loss": 0.106,
      "step": 9950
    },
    {
      "epoch": 6.59,
      "grad_norm": 1.4374464750289917,
      "learning_rate": 2.9473684210526318e-08,
      "loss": 0.1055,
      "step": 9975
    },
    {
      "epoch": 6.61,
      "grad_norm": 1.4011142253875732,
      "learning_rate": 3.1578947368421054e-09,
      "loss": 0.1056,
      "step": 10000
    },
    {
      "epoch": 6.61,
      "eval_loss": 0.3232288956642151,
      "eval_runtime": 15.5989,
      "eval_samples_per_second": 108.405,
      "eval_steps_per_second": 0.897,
      "eval_wer": 22.42412888722368,
      "step": 10000
    },
    {
      "epoch": 6.61,
      "step": 10000,
      "total_flos": 6.300100720410624e+19,
      "train_loss": 0.2321115765094757,
      "train_runtime": 12814.2682,
      "train_samples_per_second": 199.777,
      "train_steps_per_second": 0.78
    }
  ],
  "logging_steps": 25,
  "max_steps": 10000,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 7,
  "save_steps": 1000,
  "total_flos": 6.300100720410624e+19,
  "train_batch_size": 256,
  "trial_name": null,
  "trial_params": null
}