thiagobarbosa's picture
End of training
890ab67 verified
{
"best_metric": 25.502585649644473,
"best_model_checkpoint": "./whisper-base-common-voice-16-pt-v8/checkpoint-4000",
"epoch": 5.181347150259067,
"eval_steps": 1000,
"global_step": 7000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.02,
"learning_rate": 8.8e-07,
"loss": 3.14,
"step": 25
},
{
"epoch": 0.04,
"learning_rate": 1.8800000000000002e-06,
"loss": 2.574,
"step": 50
},
{
"epoch": 0.06,
"learning_rate": 2.88e-06,
"loss": 1.9629,
"step": 75
},
{
"epoch": 0.07,
"learning_rate": 3.88e-06,
"loss": 1.1213,
"step": 100
},
{
"epoch": 0.09,
"learning_rate": 4.880000000000001e-06,
"loss": 0.8777,
"step": 125
},
{
"epoch": 0.11,
"learning_rate": 5.8800000000000005e-06,
"loss": 0.767,
"step": 150
},
{
"epoch": 0.13,
"learning_rate": 6.88e-06,
"loss": 0.7085,
"step": 175
},
{
"epoch": 0.15,
"learning_rate": 7.88e-06,
"loss": 0.7318,
"step": 200
},
{
"epoch": 0.17,
"learning_rate": 8.880000000000001e-06,
"loss": 0.676,
"step": 225
},
{
"epoch": 0.19,
"learning_rate": 9.88e-06,
"loss": 0.6427,
"step": 250
},
{
"epoch": 0.2,
"learning_rate": 1.0880000000000001e-05,
"loss": 0.6226,
"step": 275
},
{
"epoch": 0.22,
"learning_rate": 1.188e-05,
"loss": 0.6164,
"step": 300
},
{
"epoch": 0.24,
"learning_rate": 1.2880000000000002e-05,
"loss": 0.6673,
"step": 325
},
{
"epoch": 0.26,
"learning_rate": 1.3880000000000001e-05,
"loss": 0.577,
"step": 350
},
{
"epoch": 0.28,
"learning_rate": 1.4880000000000002e-05,
"loss": 0.5764,
"step": 375
},
{
"epoch": 0.3,
"learning_rate": 1.588e-05,
"loss": 0.5949,
"step": 400
},
{
"epoch": 0.31,
"learning_rate": 1.688e-05,
"loss": 0.5552,
"step": 425
},
{
"epoch": 0.33,
"learning_rate": 1.788e-05,
"loss": 0.5624,
"step": 450
},
{
"epoch": 0.35,
"learning_rate": 1.8880000000000002e-05,
"loss": 0.5672,
"step": 475
},
{
"epoch": 0.37,
"learning_rate": 1.9880000000000003e-05,
"loss": 0.5403,
"step": 500
},
{
"epoch": 0.39,
"learning_rate": 1.9932307692307693e-05,
"loss": 0.5491,
"step": 525
},
{
"epoch": 0.41,
"learning_rate": 1.9855384615384618e-05,
"loss": 0.5894,
"step": 550
},
{
"epoch": 0.43,
"learning_rate": 1.977846153846154e-05,
"loss": 0.5285,
"step": 575
},
{
"epoch": 0.44,
"learning_rate": 1.9701538461538463e-05,
"loss": 0.5467,
"step": 600
},
{
"epoch": 0.46,
"learning_rate": 1.9624615384615387e-05,
"loss": 0.6033,
"step": 625
},
{
"epoch": 0.48,
"learning_rate": 1.9547692307692308e-05,
"loss": 0.5171,
"step": 650
},
{
"epoch": 0.5,
"learning_rate": 1.9470769230769233e-05,
"loss": 0.5449,
"step": 675
},
{
"epoch": 0.52,
"learning_rate": 1.9393846153846157e-05,
"loss": 0.4971,
"step": 700
},
{
"epoch": 0.54,
"learning_rate": 1.9316923076923078e-05,
"loss": 0.5514,
"step": 725
},
{
"epoch": 0.56,
"learning_rate": 1.9240000000000002e-05,
"loss": 0.4778,
"step": 750
},
{
"epoch": 0.57,
"learning_rate": 1.9163076923076923e-05,
"loss": 0.5503,
"step": 775
},
{
"epoch": 0.59,
"learning_rate": 1.9086153846153847e-05,
"loss": 0.4861,
"step": 800
},
{
"epoch": 0.61,
"learning_rate": 1.9009230769230772e-05,
"loss": 0.4682,
"step": 825
},
{
"epoch": 0.63,
"learning_rate": 1.8932307692307693e-05,
"loss": 0.4802,
"step": 850
},
{
"epoch": 0.65,
"learning_rate": 1.8855384615384617e-05,
"loss": 0.4811,
"step": 875
},
{
"epoch": 0.67,
"learning_rate": 1.877846153846154e-05,
"loss": 0.4963,
"step": 900
},
{
"epoch": 0.68,
"learning_rate": 1.8701538461538462e-05,
"loss": 0.5029,
"step": 925
},
{
"epoch": 0.7,
"learning_rate": 1.8624615384615387e-05,
"loss": 0.4717,
"step": 950
},
{
"epoch": 0.72,
"learning_rate": 1.8547692307692308e-05,
"loss": 0.5072,
"step": 975
},
{
"epoch": 0.74,
"learning_rate": 1.8470769230769232e-05,
"loss": 0.4883,
"step": 1000
},
{
"epoch": 0.74,
"eval_WER": 28.031674208144796,
"eval_WER_NORMALIZED": 21.832715365724468,
"eval_loss": 0.3803258538246155,
"eval_runtime": 1308.0449,
"eval_samples_per_second": 7.186,
"eval_steps_per_second": 0.45,
"step": 1000
},
{
"epoch": 0.76,
"learning_rate": 1.8393846153846156e-05,
"loss": 0.4572,
"step": 1025
},
{
"epoch": 0.78,
"learning_rate": 1.8316923076923077e-05,
"loss": 0.4627,
"step": 1050
},
{
"epoch": 0.8,
"learning_rate": 1.824e-05,
"loss": 0.434,
"step": 1075
},
{
"epoch": 0.81,
"learning_rate": 1.8163076923076926e-05,
"loss": 0.4043,
"step": 1100
},
{
"epoch": 0.83,
"learning_rate": 1.8086153846153847e-05,
"loss": 0.464,
"step": 1125
},
{
"epoch": 0.85,
"learning_rate": 1.800923076923077e-05,
"loss": 0.4155,
"step": 1150
},
{
"epoch": 0.87,
"learning_rate": 1.7932307692307692e-05,
"loss": 0.4151,
"step": 1175
},
{
"epoch": 0.89,
"learning_rate": 1.7855384615384616e-05,
"loss": 0.4333,
"step": 1200
},
{
"epoch": 0.91,
"learning_rate": 1.777846153846154e-05,
"loss": 0.4288,
"step": 1225
},
{
"epoch": 0.93,
"learning_rate": 1.7701538461538462e-05,
"loss": 0.4837,
"step": 1250
},
{
"epoch": 0.94,
"learning_rate": 1.7624615384615386e-05,
"loss": 0.4438,
"step": 1275
},
{
"epoch": 0.96,
"learning_rate": 1.754769230769231e-05,
"loss": 0.4399,
"step": 1300
},
{
"epoch": 0.98,
"learning_rate": 1.747076923076923e-05,
"loss": 0.4127,
"step": 1325
},
{
"epoch": 1.0,
"learning_rate": 1.7393846153846156e-05,
"loss": 0.4059,
"step": 1350
},
{
"epoch": 1.02,
"learning_rate": 1.7316923076923077e-05,
"loss": 0.2737,
"step": 1375
},
{
"epoch": 1.04,
"learning_rate": 1.724e-05,
"loss": 0.2745,
"step": 1400
},
{
"epoch": 1.05,
"learning_rate": 1.7163076923076925e-05,
"loss": 0.2944,
"step": 1425
},
{
"epoch": 1.07,
"learning_rate": 1.7086153846153846e-05,
"loss": 0.2879,
"step": 1450
},
{
"epoch": 1.09,
"learning_rate": 1.700923076923077e-05,
"loss": 0.2987,
"step": 1475
},
{
"epoch": 1.11,
"learning_rate": 1.6932307692307695e-05,
"loss": 0.2546,
"step": 1500
},
{
"epoch": 1.13,
"learning_rate": 1.6855384615384616e-05,
"loss": 0.2703,
"step": 1525
},
{
"epoch": 1.15,
"learning_rate": 1.677846153846154e-05,
"loss": 0.2698,
"step": 1550
},
{
"epoch": 1.17,
"learning_rate": 1.670153846153846e-05,
"loss": 0.2657,
"step": 1575
},
{
"epoch": 1.18,
"learning_rate": 1.6624615384615385e-05,
"loss": 0.2583,
"step": 1600
},
{
"epoch": 1.2,
"learning_rate": 1.654769230769231e-05,
"loss": 0.284,
"step": 1625
},
{
"epoch": 1.22,
"learning_rate": 1.647076923076923e-05,
"loss": 0.2543,
"step": 1650
},
{
"epoch": 1.24,
"learning_rate": 1.6393846153846155e-05,
"loss": 0.2786,
"step": 1675
},
{
"epoch": 1.26,
"learning_rate": 1.631692307692308e-05,
"loss": 0.2472,
"step": 1700
},
{
"epoch": 1.28,
"learning_rate": 1.6240000000000004e-05,
"loss": 0.2536,
"step": 1725
},
{
"epoch": 1.3,
"learning_rate": 1.6163076923076925e-05,
"loss": 0.2562,
"step": 1750
},
{
"epoch": 1.31,
"learning_rate": 1.6086153846153846e-05,
"loss": 0.2944,
"step": 1775
},
{
"epoch": 1.33,
"learning_rate": 1.600923076923077e-05,
"loss": 0.2398,
"step": 1800
},
{
"epoch": 1.35,
"learning_rate": 1.5932307692307694e-05,
"loss": 0.2641,
"step": 1825
},
{
"epoch": 1.37,
"learning_rate": 1.5855384615384615e-05,
"loss": 0.2465,
"step": 1850
},
{
"epoch": 1.39,
"learning_rate": 1.577846153846154e-05,
"loss": 0.246,
"step": 1875
},
{
"epoch": 1.41,
"learning_rate": 1.5701538461538464e-05,
"loss": 0.2694,
"step": 1900
},
{
"epoch": 1.42,
"learning_rate": 1.5624615384615388e-05,
"loss": 0.2532,
"step": 1925
},
{
"epoch": 1.44,
"learning_rate": 1.554769230769231e-05,
"loss": 0.255,
"step": 1950
},
{
"epoch": 1.46,
"learning_rate": 1.547076923076923e-05,
"loss": 0.2695,
"step": 1975
},
{
"epoch": 1.48,
"learning_rate": 1.5393846153846154e-05,
"loss": 0.2659,
"step": 2000
},
{
"epoch": 1.48,
"eval_WER": 26.368778280542987,
"eval_WER_NORMALIZED": 20.166610487793132,
"eval_loss": 0.36767107248306274,
"eval_runtime": 1317.0971,
"eval_samples_per_second": 7.136,
"eval_steps_per_second": 0.446,
"step": 2000
},
{
"epoch": 1.5,
"learning_rate": 1.531692307692308e-05,
"loss": 0.263,
"step": 2025
},
{
"epoch": 1.52,
"learning_rate": 1.5240000000000001e-05,
"loss": 0.24,
"step": 2050
},
{
"epoch": 1.54,
"learning_rate": 1.5163076923076924e-05,
"loss": 0.2687,
"step": 2075
},
{
"epoch": 1.55,
"learning_rate": 1.5089230769230772e-05,
"loss": 0.2375,
"step": 2100
},
{
"epoch": 1.57,
"learning_rate": 1.5012307692307693e-05,
"loss": 0.2445,
"step": 2125
},
{
"epoch": 1.59,
"learning_rate": 1.4935384615384616e-05,
"loss": 0.2499,
"step": 2150
},
{
"epoch": 1.61,
"learning_rate": 1.485846153846154e-05,
"loss": 0.2412,
"step": 2175
},
{
"epoch": 1.63,
"learning_rate": 1.4781538461538463e-05,
"loss": 0.2409,
"step": 2200
},
{
"epoch": 1.65,
"learning_rate": 1.4704615384615387e-05,
"loss": 0.2235,
"step": 2225
},
{
"epoch": 1.67,
"learning_rate": 1.462769230769231e-05,
"loss": 0.2683,
"step": 2250
},
{
"epoch": 1.68,
"learning_rate": 1.4550769230769232e-05,
"loss": 0.2534,
"step": 2275
},
{
"epoch": 1.7,
"learning_rate": 1.4473846153846157e-05,
"loss": 0.2374,
"step": 2300
},
{
"epoch": 1.72,
"learning_rate": 1.4396923076923077e-05,
"loss": 0.2475,
"step": 2325
},
{
"epoch": 1.74,
"learning_rate": 1.432e-05,
"loss": 0.2156,
"step": 2350
},
{
"epoch": 1.76,
"learning_rate": 1.4243076923076924e-05,
"loss": 0.2439,
"step": 2375
},
{
"epoch": 1.78,
"learning_rate": 1.4166153846153847e-05,
"loss": 0.2332,
"step": 2400
},
{
"epoch": 1.79,
"learning_rate": 1.4089230769230771e-05,
"loss": 0.2406,
"step": 2425
},
{
"epoch": 1.81,
"learning_rate": 1.4012307692307694e-05,
"loss": 0.2513,
"step": 2450
},
{
"epoch": 1.83,
"learning_rate": 1.3935384615384617e-05,
"loss": 0.258,
"step": 2475
},
{
"epoch": 1.85,
"learning_rate": 1.3858461538461541e-05,
"loss": 0.2176,
"step": 2500
},
{
"epoch": 1.87,
"learning_rate": 1.3781538461538462e-05,
"loss": 0.2398,
"step": 2525
},
{
"epoch": 1.89,
"learning_rate": 1.3704615384615385e-05,
"loss": 0.2293,
"step": 2550
},
{
"epoch": 1.91,
"learning_rate": 1.3627692307692309e-05,
"loss": 0.2062,
"step": 2575
},
{
"epoch": 1.92,
"learning_rate": 1.3550769230769232e-05,
"loss": 0.2358,
"step": 2600
},
{
"epoch": 1.94,
"learning_rate": 1.3473846153846156e-05,
"loss": 0.2234,
"step": 2625
},
{
"epoch": 1.96,
"learning_rate": 1.3396923076923079e-05,
"loss": 0.2331,
"step": 2650
},
{
"epoch": 1.98,
"learning_rate": 1.3320000000000001e-05,
"loss": 0.2156,
"step": 2675
},
{
"epoch": 2.0,
"learning_rate": 1.3243076923076926e-05,
"loss": 0.2037,
"step": 2700
},
{
"epoch": 2.02,
"learning_rate": 1.3166153846153846e-05,
"loss": 0.1322,
"step": 2725
},
{
"epoch": 2.04,
"learning_rate": 1.3089230769230769e-05,
"loss": 0.1144,
"step": 2750
},
{
"epoch": 2.05,
"learning_rate": 1.3012307692307693e-05,
"loss": 0.1247,
"step": 2775
},
{
"epoch": 2.07,
"learning_rate": 1.2935384615384616e-05,
"loss": 0.1249,
"step": 2800
},
{
"epoch": 2.09,
"learning_rate": 1.285846153846154e-05,
"loss": 0.133,
"step": 2825
},
{
"epoch": 2.11,
"learning_rate": 1.2781538461538463e-05,
"loss": 0.1377,
"step": 2850
},
{
"epoch": 2.13,
"learning_rate": 1.2704615384615386e-05,
"loss": 0.1303,
"step": 2875
},
{
"epoch": 2.15,
"learning_rate": 1.262769230769231e-05,
"loss": 0.117,
"step": 2900
},
{
"epoch": 2.17,
"learning_rate": 1.2550769230769231e-05,
"loss": 0.1267,
"step": 2925
},
{
"epoch": 2.18,
"learning_rate": 1.2473846153846154e-05,
"loss": 0.1236,
"step": 2950
},
{
"epoch": 2.2,
"learning_rate": 1.2396923076923078e-05,
"loss": 0.1195,
"step": 2975
},
{
"epoch": 2.22,
"learning_rate": 1.232e-05,
"loss": 0.1251,
"step": 3000
},
{
"epoch": 2.22,
"eval_WER": 26.37524240465417,
"eval_WER_NORMALIZED": 20.46195085151121,
"eval_loss": 0.37302544713020325,
"eval_runtime": 1305.9589,
"eval_samples_per_second": 7.197,
"eval_steps_per_second": 0.45,
"step": 3000
},
{
"epoch": 2.24,
"learning_rate": 1.2243076923076925e-05,
"loss": 0.1251,
"step": 3025
},
{
"epoch": 2.26,
"learning_rate": 1.2166153846153848e-05,
"loss": 0.1291,
"step": 3050
},
{
"epoch": 2.28,
"learning_rate": 1.208923076923077e-05,
"loss": 0.1187,
"step": 3075
},
{
"epoch": 2.29,
"learning_rate": 1.2012307692307695e-05,
"loss": 0.1097,
"step": 3100
},
{
"epoch": 2.31,
"learning_rate": 1.1935384615384616e-05,
"loss": 0.1247,
"step": 3125
},
{
"epoch": 2.33,
"learning_rate": 1.1858461538461538e-05,
"loss": 0.1243,
"step": 3150
},
{
"epoch": 2.35,
"learning_rate": 1.1781538461538463e-05,
"loss": 0.1297,
"step": 3175
},
{
"epoch": 2.37,
"learning_rate": 1.1704615384615385e-05,
"loss": 0.1061,
"step": 3200
},
{
"epoch": 2.39,
"learning_rate": 1.162769230769231e-05,
"loss": 0.1219,
"step": 3225
},
{
"epoch": 2.41,
"learning_rate": 1.1550769230769232e-05,
"loss": 0.1225,
"step": 3250
},
{
"epoch": 2.42,
"learning_rate": 1.1473846153846155e-05,
"loss": 0.1176,
"step": 3275
},
{
"epoch": 2.44,
"learning_rate": 1.1396923076923079e-05,
"loss": 0.1323,
"step": 3300
},
{
"epoch": 2.46,
"learning_rate": 1.132e-05,
"loss": 0.1109,
"step": 3325
},
{
"epoch": 2.48,
"learning_rate": 1.1243076923076923e-05,
"loss": 0.1144,
"step": 3350
},
{
"epoch": 2.5,
"learning_rate": 1.1166153846153847e-05,
"loss": 0.1159,
"step": 3375
},
{
"epoch": 2.52,
"learning_rate": 1.108923076923077e-05,
"loss": 0.1068,
"step": 3400
},
{
"epoch": 2.54,
"learning_rate": 1.1012307692307694e-05,
"loss": 0.1204,
"step": 3425
},
{
"epoch": 2.55,
"learning_rate": 1.0935384615384617e-05,
"loss": 0.1104,
"step": 3450
},
{
"epoch": 2.57,
"learning_rate": 1.085846153846154e-05,
"loss": 0.1042,
"step": 3475
},
{
"epoch": 2.59,
"learning_rate": 1.0781538461538464e-05,
"loss": 0.1075,
"step": 3500
},
{
"epoch": 2.61,
"learning_rate": 1.0704615384615385e-05,
"loss": 0.1127,
"step": 3525
},
{
"epoch": 2.63,
"learning_rate": 1.0627692307692307e-05,
"loss": 0.1169,
"step": 3550
},
{
"epoch": 2.65,
"learning_rate": 1.0550769230769232e-05,
"loss": 0.1197,
"step": 3575
},
{
"epoch": 2.66,
"learning_rate": 1.0473846153846154e-05,
"loss": 0.1139,
"step": 3600
},
{
"epoch": 2.68,
"learning_rate": 1.0396923076923079e-05,
"loss": 0.1143,
"step": 3625
},
{
"epoch": 2.7,
"learning_rate": 1.0320000000000001e-05,
"loss": 0.1079,
"step": 3650
},
{
"epoch": 2.72,
"learning_rate": 1.0243076923076924e-05,
"loss": 0.1164,
"step": 3675
},
{
"epoch": 2.74,
"learning_rate": 1.0166153846153848e-05,
"loss": 0.1098,
"step": 3700
},
{
"epoch": 2.76,
"learning_rate": 1.008923076923077e-05,
"loss": 0.1271,
"step": 3725
},
{
"epoch": 2.78,
"learning_rate": 1.0012307692307692e-05,
"loss": 0.1115,
"step": 3750
},
{
"epoch": 2.79,
"learning_rate": 9.935384615384616e-06,
"loss": 0.1148,
"step": 3775
},
{
"epoch": 2.81,
"learning_rate": 9.858461538461539e-06,
"loss": 0.1229,
"step": 3800
},
{
"epoch": 2.83,
"learning_rate": 9.781538461538463e-06,
"loss": 0.1098,
"step": 3825
},
{
"epoch": 2.85,
"learning_rate": 9.704615384615386e-06,
"loss": 0.1015,
"step": 3850
},
{
"epoch": 2.87,
"learning_rate": 9.627692307692308e-06,
"loss": 0.1072,
"step": 3875
},
{
"epoch": 2.89,
"learning_rate": 9.550769230769231e-06,
"loss": 0.1075,
"step": 3900
},
{
"epoch": 2.91,
"learning_rate": 9.473846153846155e-06,
"loss": 0.1059,
"step": 3925
},
{
"epoch": 2.92,
"learning_rate": 9.396923076923078e-06,
"loss": 0.1138,
"step": 3950
},
{
"epoch": 2.94,
"learning_rate": 9.32e-06,
"loss": 0.1013,
"step": 3975
},
{
"epoch": 2.96,
"learning_rate": 9.243076923076923e-06,
"loss": 0.1071,
"step": 4000
},
{
"epoch": 2.96,
"eval_WER": 25.502585649644473,
"eval_WER_NORMALIZED": 19.54703776825412,
"eval_loss": 0.38669511675834656,
"eval_runtime": 1309.2157,
"eval_samples_per_second": 7.179,
"eval_steps_per_second": 0.449,
"step": 4000
},
{
"epoch": 2.98,
"learning_rate": 9.166153846153848e-06,
"loss": 0.1061,
"step": 4025
},
{
"epoch": 3.0,
"learning_rate": 9.08923076923077e-06,
"loss": 0.1171,
"step": 4050
},
{
"epoch": 3.02,
"learning_rate": 9.012307692307693e-06,
"loss": 0.0605,
"step": 4075
},
{
"epoch": 3.03,
"learning_rate": 8.935384615384615e-06,
"loss": 0.0543,
"step": 4100
},
{
"epoch": 3.05,
"learning_rate": 8.85846153846154e-06,
"loss": 0.0492,
"step": 4125
},
{
"epoch": 3.07,
"learning_rate": 8.781538461538462e-06,
"loss": 0.0476,
"step": 4150
},
{
"epoch": 3.09,
"learning_rate": 8.704615384615385e-06,
"loss": 0.0499,
"step": 4175
},
{
"epoch": 3.11,
"learning_rate": 8.627692307692308e-06,
"loss": 0.0559,
"step": 4200
},
{
"epoch": 3.13,
"learning_rate": 8.550769230769232e-06,
"loss": 0.0506,
"step": 4225
},
{
"epoch": 3.15,
"learning_rate": 8.473846153846155e-06,
"loss": 0.0548,
"step": 4250
},
{
"epoch": 3.16,
"learning_rate": 8.396923076923077e-06,
"loss": 0.0496,
"step": 4275
},
{
"epoch": 3.18,
"learning_rate": 8.32e-06,
"loss": 0.0541,
"step": 4300
},
{
"epoch": 3.2,
"learning_rate": 8.243076923076924e-06,
"loss": 0.0519,
"step": 4325
},
{
"epoch": 3.22,
"learning_rate": 8.166153846153847e-06,
"loss": 0.0562,
"step": 4350
},
{
"epoch": 3.24,
"learning_rate": 8.08923076923077e-06,
"loss": 0.0571,
"step": 4375
},
{
"epoch": 3.26,
"learning_rate": 8.012307692307692e-06,
"loss": 0.0554,
"step": 4400
},
{
"epoch": 3.28,
"learning_rate": 7.935384615384617e-06,
"loss": 0.0472,
"step": 4425
},
{
"epoch": 3.29,
"learning_rate": 7.85846153846154e-06,
"loss": 0.0529,
"step": 4450
},
{
"epoch": 3.31,
"learning_rate": 7.781538461538462e-06,
"loss": 0.0496,
"step": 4475
},
{
"epoch": 3.33,
"learning_rate": 7.704615384615384e-06,
"loss": 0.0503,
"step": 4500
},
{
"epoch": 3.35,
"learning_rate": 7.627692307692308e-06,
"loss": 0.0539,
"step": 4525
},
{
"epoch": 3.37,
"learning_rate": 7.5507692307692315e-06,
"loss": 0.0523,
"step": 4550
},
{
"epoch": 3.39,
"learning_rate": 7.473846153846155e-06,
"loss": 0.0479,
"step": 4575
},
{
"epoch": 3.4,
"learning_rate": 7.396923076923077e-06,
"loss": 0.0514,
"step": 4600
},
{
"epoch": 3.42,
"learning_rate": 7.32e-06,
"loss": 0.0502,
"step": 4625
},
{
"epoch": 3.44,
"learning_rate": 7.243076923076924e-06,
"loss": 0.0528,
"step": 4650
},
{
"epoch": 3.46,
"learning_rate": 7.166153846153847e-06,
"loss": 0.0557,
"step": 4675
},
{
"epoch": 3.48,
"learning_rate": 7.089230769230769e-06,
"loss": 0.0479,
"step": 4700
},
{
"epoch": 3.5,
"learning_rate": 7.0123076923076925e-06,
"loss": 0.0521,
"step": 4725
},
{
"epoch": 3.52,
"learning_rate": 6.935384615384616e-06,
"loss": 0.053,
"step": 4750
},
{
"epoch": 3.53,
"learning_rate": 6.8584615384615395e-06,
"loss": 0.049,
"step": 4775
},
{
"epoch": 3.55,
"learning_rate": 6.781538461538461e-06,
"loss": 0.0517,
"step": 4800
},
{
"epoch": 3.57,
"learning_rate": 6.704615384615385e-06,
"loss": 0.0495,
"step": 4825
},
{
"epoch": 3.59,
"learning_rate": 6.627692307692308e-06,
"loss": 0.0459,
"step": 4850
},
{
"epoch": 3.61,
"learning_rate": 6.550769230769232e-06,
"loss": 0.0546,
"step": 4875
},
{
"epoch": 3.63,
"learning_rate": 6.4738461538461535e-06,
"loss": 0.0499,
"step": 4900
},
{
"epoch": 3.65,
"learning_rate": 6.396923076923077e-06,
"loss": 0.0502,
"step": 4925
},
{
"epoch": 3.66,
"learning_rate": 6.3200000000000005e-06,
"loss": 0.0466,
"step": 4950
},
{
"epoch": 3.68,
"learning_rate": 6.243076923076924e-06,
"loss": 0.0519,
"step": 4975
},
{
"epoch": 3.7,
"learning_rate": 6.166153846153846e-06,
"loss": 0.0523,
"step": 5000
},
{
"epoch": 3.7,
"eval_WER": 25.709437621202326,
"eval_WER_NORMALIZED": 19.685077286078876,
"eval_loss": 0.41479501128196716,
"eval_runtime": 1279.8151,
"eval_samples_per_second": 7.344,
"eval_steps_per_second": 0.459,
"step": 5000
},
{
"epoch": 3.72,
"learning_rate": 6.089230769230769e-06,
"loss": 0.0485,
"step": 5025
},
{
"epoch": 3.74,
"learning_rate": 6.012307692307693e-06,
"loss": 0.0468,
"step": 5050
},
{
"epoch": 3.76,
"learning_rate": 5.935384615384616e-06,
"loss": 0.0495,
"step": 5075
},
{
"epoch": 3.77,
"learning_rate": 5.858461538461538e-06,
"loss": 0.0471,
"step": 5100
},
{
"epoch": 3.79,
"learning_rate": 5.7815384615384615e-06,
"loss": 0.0458,
"step": 5125
},
{
"epoch": 3.81,
"learning_rate": 5.704615384615385e-06,
"loss": 0.0465,
"step": 5150
},
{
"epoch": 3.83,
"learning_rate": 5.6276923076923085e-06,
"loss": 0.0533,
"step": 5175
},
{
"epoch": 3.85,
"learning_rate": 5.55076923076923e-06,
"loss": 0.0504,
"step": 5200
},
{
"epoch": 3.87,
"learning_rate": 5.473846153846154e-06,
"loss": 0.0559,
"step": 5225
},
{
"epoch": 3.89,
"learning_rate": 5.396923076923077e-06,
"loss": 0.0514,
"step": 5250
},
{
"epoch": 3.9,
"learning_rate": 5.320000000000001e-06,
"loss": 0.0534,
"step": 5275
},
{
"epoch": 3.92,
"learning_rate": 5.243076923076923e-06,
"loss": 0.0443,
"step": 5300
},
{
"epoch": 3.94,
"learning_rate": 5.166153846153846e-06,
"loss": 0.0506,
"step": 5325
},
{
"epoch": 3.96,
"learning_rate": 5.0892307692307695e-06,
"loss": 0.0446,
"step": 5350
},
{
"epoch": 3.98,
"learning_rate": 5.012307692307693e-06,
"loss": 0.0437,
"step": 5375
},
{
"epoch": 4.0,
"learning_rate": 4.935384615384616e-06,
"loss": 0.0421,
"step": 5400
},
{
"epoch": 4.02,
"learning_rate": 4.858461538461539e-06,
"loss": 0.0323,
"step": 5425
},
{
"epoch": 4.03,
"learning_rate": 4.781538461538462e-06,
"loss": 0.0234,
"step": 5450
},
{
"epoch": 4.05,
"learning_rate": 4.704615384615385e-06,
"loss": 0.023,
"step": 5475
},
{
"epoch": 4.07,
"learning_rate": 4.627692307692308e-06,
"loss": 0.0219,
"step": 5500
},
{
"epoch": 4.09,
"learning_rate": 4.550769230769231e-06,
"loss": 0.0261,
"step": 5525
},
{
"epoch": 4.11,
"learning_rate": 4.473846153846154e-06,
"loss": 0.0208,
"step": 5550
},
{
"epoch": 4.13,
"learning_rate": 4.3969230769230775e-06,
"loss": 0.0233,
"step": 5575
},
{
"epoch": 4.15,
"learning_rate": 4.32e-06,
"loss": 0.0255,
"step": 5600
},
{
"epoch": 4.16,
"learning_rate": 4.243076923076924e-06,
"loss": 0.0208,
"step": 5625
},
{
"epoch": 4.18,
"learning_rate": 4.166153846153846e-06,
"loss": 0.0228,
"step": 5650
},
{
"epoch": 4.2,
"learning_rate": 4.08923076923077e-06,
"loss": 0.0281,
"step": 5675
},
{
"epoch": 4.22,
"learning_rate": 4.012307692307692e-06,
"loss": 0.0231,
"step": 5700
},
{
"epoch": 4.24,
"learning_rate": 3.935384615384616e-06,
"loss": 0.0242,
"step": 5725
},
{
"epoch": 4.26,
"learning_rate": 3.8584615384615385e-06,
"loss": 0.0232,
"step": 5750
},
{
"epoch": 4.27,
"learning_rate": 3.781538461538462e-06,
"loss": 0.0218,
"step": 5775
},
{
"epoch": 4.29,
"learning_rate": 3.7046153846153847e-06,
"loss": 0.0267,
"step": 5800
},
{
"epoch": 4.31,
"learning_rate": 3.627692307692308e-06,
"loss": 0.0234,
"step": 5825
},
{
"epoch": 4.33,
"learning_rate": 3.5507692307692308e-06,
"loss": 0.0234,
"step": 5850
},
{
"epoch": 4.35,
"learning_rate": 3.4738461538461543e-06,
"loss": 0.0244,
"step": 5875
},
{
"epoch": 4.37,
"learning_rate": 3.396923076923077e-06,
"loss": 0.0242,
"step": 5900
},
{
"epoch": 4.39,
"learning_rate": 3.3200000000000004e-06,
"loss": 0.0225,
"step": 5925
},
{
"epoch": 4.4,
"learning_rate": 3.243076923076923e-06,
"loss": 0.0261,
"step": 5950
},
{
"epoch": 4.42,
"learning_rate": 3.1661538461538465e-06,
"loss": 0.0234,
"step": 5975
},
{
"epoch": 4.44,
"learning_rate": 3.089230769230769e-06,
"loss": 0.02,
"step": 6000
},
{
"epoch": 4.44,
"eval_WER": 25.680349062702,
"eval_WER_NORMALIZED": 19.575929760356978,
"eval_loss": 0.4490774869918823,
"eval_runtime": 1265.5179,
"eval_samples_per_second": 7.427,
"eval_steps_per_second": 0.465,
"step": 6000
},
{
"epoch": 4.46,
"learning_rate": 3.0123076923076927e-06,
"loss": 0.0225,
"step": 6025
},
{
"epoch": 4.48,
"learning_rate": 2.9353846153846153e-06,
"loss": 0.0231,
"step": 6050
},
{
"epoch": 4.5,
"learning_rate": 2.8584615384615388e-06,
"loss": 0.0214,
"step": 6075
},
{
"epoch": 4.52,
"learning_rate": 2.7815384615384614e-06,
"loss": 0.0209,
"step": 6100
},
{
"epoch": 4.53,
"learning_rate": 2.704615384615385e-06,
"loss": 0.0217,
"step": 6125
},
{
"epoch": 4.55,
"learning_rate": 2.6276923076923076e-06,
"loss": 0.0232,
"step": 6150
},
{
"epoch": 4.57,
"learning_rate": 2.550769230769231e-06,
"loss": 0.0223,
"step": 6175
},
{
"epoch": 4.59,
"learning_rate": 2.473846153846154e-06,
"loss": 0.0246,
"step": 6200
},
{
"epoch": 4.61,
"learning_rate": 2.396923076923077e-06,
"loss": 0.021,
"step": 6225
},
{
"epoch": 4.63,
"learning_rate": 2.3200000000000002e-06,
"loss": 0.0206,
"step": 6250
},
{
"epoch": 4.64,
"learning_rate": 2.2430769230769233e-06,
"loss": 0.022,
"step": 6275
},
{
"epoch": 4.66,
"learning_rate": 2.1661538461538464e-06,
"loss": 0.0225,
"step": 6300
},
{
"epoch": 4.68,
"learning_rate": 2.0892307692307694e-06,
"loss": 0.0219,
"step": 6325
},
{
"epoch": 4.7,
"learning_rate": 2.0123076923076925e-06,
"loss": 0.0213,
"step": 6350
},
{
"epoch": 4.72,
"learning_rate": 1.9353846153846156e-06,
"loss": 0.0205,
"step": 6375
},
{
"epoch": 4.74,
"learning_rate": 1.8584615384615386e-06,
"loss": 0.0224,
"step": 6400
},
{
"epoch": 4.76,
"learning_rate": 1.7815384615384617e-06,
"loss": 0.0226,
"step": 6425
},
{
"epoch": 4.77,
"learning_rate": 1.7046153846153847e-06,
"loss": 0.0248,
"step": 6450
},
{
"epoch": 4.79,
"learning_rate": 1.6276923076923078e-06,
"loss": 0.0225,
"step": 6475
},
{
"epoch": 4.81,
"learning_rate": 1.5507692307692309e-06,
"loss": 0.0213,
"step": 6500
},
{
"epoch": 4.83,
"learning_rate": 1.473846153846154e-06,
"loss": 0.021,
"step": 6525
},
{
"epoch": 4.85,
"learning_rate": 1.396923076923077e-06,
"loss": 0.0199,
"step": 6550
},
{
"epoch": 4.87,
"learning_rate": 1.32e-06,
"loss": 0.0215,
"step": 6575
},
{
"epoch": 4.89,
"learning_rate": 1.2430769230769231e-06,
"loss": 0.0212,
"step": 6600
},
{
"epoch": 4.9,
"learning_rate": 1.1661538461538464e-06,
"loss": 0.0208,
"step": 6625
},
{
"epoch": 4.92,
"learning_rate": 1.0892307692307695e-06,
"loss": 0.0237,
"step": 6650
},
{
"epoch": 4.94,
"learning_rate": 1.0123076923076925e-06,
"loss": 0.0221,
"step": 6675
},
{
"epoch": 4.96,
"learning_rate": 9.353846153846155e-07,
"loss": 0.0238,
"step": 6700
},
{
"epoch": 4.98,
"learning_rate": 8.584615384615386e-07,
"loss": 0.0253,
"step": 6725
},
{
"epoch": 5.0,
"learning_rate": 7.815384615384616e-07,
"loss": 0.0206,
"step": 6750
},
{
"epoch": 5.01,
"learning_rate": 7.046153846153847e-07,
"loss": 0.0165,
"step": 6775
},
{
"epoch": 5.03,
"learning_rate": 6.276923076923078e-07,
"loss": 0.0168,
"step": 6800
},
{
"epoch": 5.05,
"learning_rate": 5.507692307692308e-07,
"loss": 0.0145,
"step": 6825
},
{
"epoch": 5.07,
"learning_rate": 4.738461538461539e-07,
"loss": 0.0166,
"step": 6850
},
{
"epoch": 5.09,
"learning_rate": 3.9692307692307694e-07,
"loss": 0.0143,
"step": 6875
},
{
"epoch": 5.11,
"learning_rate": 3.2e-07,
"loss": 0.0141,
"step": 6900
},
{
"epoch": 5.13,
"learning_rate": 2.430769230769231e-07,
"loss": 0.0128,
"step": 6925
},
{
"epoch": 5.14,
"learning_rate": 1.6615384615384616e-07,
"loss": 0.0138,
"step": 6950
},
{
"epoch": 5.16,
"learning_rate": 8.923076923076923e-08,
"loss": 0.014,
"step": 6975
},
{
"epoch": 5.18,
"learning_rate": 1.230769230769231e-08,
"loss": 0.0134,
"step": 7000
},
{
"epoch": 5.18,
"eval_WER": 26.192630898513254,
"eval_WER_NORMALIZED": 20.002889199210287,
"eval_loss": 0.4573647677898407,
"eval_runtime": 1280.8418,
"eval_samples_per_second": 7.338,
"eval_steps_per_second": 0.459,
"step": 7000
},
{
"epoch": 5.18,
"step": 7000,
"total_flos": 1.45234503401472e+19,
"train_loss": 0.21270917607205256,
"train_runtime": 18504.0374,
"train_samples_per_second": 12.105,
"train_steps_per_second": 0.378
}
],
"logging_steps": 25,
"max_steps": 7000,
"num_input_tokens_seen": 0,
"num_train_epochs": 6,
"save_steps": 1000,
"total_flos": 1.45234503401472e+19,
"train_batch_size": 32,
"trial_name": null,
"trial_params": null
}