whisper-largev2-mls-fr / trainer_state.json
sgangireddy's picture
End of training
07e96a8
raw
history blame
20.7 kB
{
"best_metric": 4.561620226935377,
"best_model_checkpoint": "./checkpoint-4000",
"epoch": 1.0,
"global_step": 4000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.01,
"learning_rate": 4.800000000000001e-07,
"loss": 0.6521,
"step": 25
},
{
"epoch": 0.01,
"learning_rate": 9.800000000000001e-07,
"loss": 0.4515,
"step": 50
},
{
"epoch": 0.02,
"learning_rate": 1.46e-06,
"loss": 0.2873,
"step": 75
},
{
"epoch": 0.03,
"learning_rate": 1.9600000000000003e-06,
"loss": 0.2034,
"step": 100
},
{
"epoch": 0.03,
"learning_rate": 2.46e-06,
"loss": 0.2061,
"step": 125
},
{
"epoch": 0.04,
"learning_rate": 2.96e-06,
"loss": 0.2546,
"step": 150
},
{
"epoch": 0.04,
"learning_rate": 3.46e-06,
"loss": 0.325,
"step": 175
},
{
"epoch": 0.05,
"learning_rate": 3.96e-06,
"loss": 0.1184,
"step": 200
},
{
"epoch": 0.06,
"learning_rate": 4.4600000000000005e-06,
"loss": 0.1599,
"step": 225
},
{
"epoch": 0.06,
"learning_rate": 4.960000000000001e-06,
"loss": 0.1254,
"step": 250
},
{
"epoch": 0.07,
"learning_rate": 5.460000000000001e-06,
"loss": 0.361,
"step": 275
},
{
"epoch": 0.07,
"learning_rate": 5.9600000000000005e-06,
"loss": 0.1701,
"step": 300
},
{
"epoch": 0.08,
"learning_rate": 6.460000000000001e-06,
"loss": 0.1532,
"step": 325
},
{
"epoch": 0.09,
"learning_rate": 6.96e-06,
"loss": 0.1086,
"step": 350
},
{
"epoch": 0.09,
"learning_rate": 7.4600000000000006e-06,
"loss": 0.0763,
"step": 375
},
{
"epoch": 0.1,
"learning_rate": 7.960000000000002e-06,
"loss": 0.1134,
"step": 400
},
{
"epoch": 0.11,
"learning_rate": 8.46e-06,
"loss": 0.1014,
"step": 425
},
{
"epoch": 0.11,
"learning_rate": 8.94e-06,
"loss": 0.2977,
"step": 450
},
{
"epoch": 0.12,
"learning_rate": 9.440000000000001e-06,
"loss": 0.3106,
"step": 475
},
{
"epoch": 0.12,
"learning_rate": 9.940000000000001e-06,
"loss": 0.1249,
"step": 500
},
{
"epoch": 0.13,
"learning_rate": 9.937142857142858e-06,
"loss": 0.1412,
"step": 525
},
{
"epoch": 0.14,
"learning_rate": 9.865714285714285e-06,
"loss": 0.1077,
"step": 550
},
{
"epoch": 0.14,
"learning_rate": 9.794285714285714e-06,
"loss": 0.0803,
"step": 575
},
{
"epoch": 0.15,
"learning_rate": 9.722857142857143e-06,
"loss": 0.1111,
"step": 600
},
{
"epoch": 0.16,
"learning_rate": 9.651428571428572e-06,
"loss": 0.1667,
"step": 625
},
{
"epoch": 0.16,
"learning_rate": 9.58e-06,
"loss": 0.131,
"step": 650
},
{
"epoch": 0.17,
"learning_rate": 9.508571428571429e-06,
"loss": 0.274,
"step": 675
},
{
"epoch": 0.17,
"learning_rate": 9.437142857142858e-06,
"loss": 0.1309,
"step": 700
},
{
"epoch": 0.18,
"learning_rate": 9.365714285714287e-06,
"loss": 0.1207,
"step": 725
},
{
"epoch": 0.19,
"learning_rate": 9.294285714285714e-06,
"loss": 0.0882,
"step": 750
},
{
"epoch": 0.19,
"learning_rate": 9.222857142857143e-06,
"loss": 0.1223,
"step": 775
},
{
"epoch": 0.2,
"learning_rate": 9.151428571428572e-06,
"loss": 0.1185,
"step": 800
},
{
"epoch": 0.21,
"learning_rate": 9.080000000000001e-06,
"loss": 0.4117,
"step": 825
},
{
"epoch": 0.21,
"learning_rate": 9.00857142857143e-06,
"loss": 0.139,
"step": 850
},
{
"epoch": 0.22,
"learning_rate": 8.937142857142857e-06,
"loss": 0.079,
"step": 875
},
{
"epoch": 0.23,
"learning_rate": 8.865714285714287e-06,
"loss": 0.1328,
"step": 900
},
{
"epoch": 0.23,
"learning_rate": 8.794285714285716e-06,
"loss": 0.1391,
"step": 925
},
{
"epoch": 0.24,
"learning_rate": 8.722857142857145e-06,
"loss": 0.1196,
"step": 950
},
{
"epoch": 0.24,
"learning_rate": 8.651428571428572e-06,
"loss": 0.1083,
"step": 975
},
{
"epoch": 0.25,
"learning_rate": 8.580000000000001e-06,
"loss": 0.1303,
"step": 1000
},
{
"epoch": 0.25,
"eval_loss": 0.12187333405017853,
"eval_runtime": 6950.7062,
"eval_samples_per_second": 0.349,
"eval_steps_per_second": 0.044,
"eval_wer": 6.36176168796884,
"step": 1000
},
{
"epoch": 0.26,
"learning_rate": 8.511428571428571e-06,
"loss": 0.1151,
"step": 1025
},
{
"epoch": 0.26,
"learning_rate": 8.44e-06,
"loss": 0.113,
"step": 1050
},
{
"epoch": 0.27,
"learning_rate": 8.36857142857143e-06,
"loss": 0.0866,
"step": 1075
},
{
"epoch": 0.28,
"learning_rate": 8.297142857142859e-06,
"loss": 0.0873,
"step": 1100
},
{
"epoch": 0.28,
"learning_rate": 8.225714285714288e-06,
"loss": 0.1041,
"step": 1125
},
{
"epoch": 0.29,
"learning_rate": 8.154285714285715e-06,
"loss": 0.2313,
"step": 1150
},
{
"epoch": 0.29,
"learning_rate": 8.082857142857144e-06,
"loss": 0.2704,
"step": 1175
},
{
"epoch": 0.3,
"learning_rate": 8.011428571428573e-06,
"loss": 0.2215,
"step": 1200
},
{
"epoch": 0.31,
"learning_rate": 7.94e-06,
"loss": 0.2238,
"step": 1225
},
{
"epoch": 0.31,
"learning_rate": 7.86857142857143e-06,
"loss": 0.5506,
"step": 1250
},
{
"epoch": 0.32,
"learning_rate": 7.797142857142858e-06,
"loss": 0.1451,
"step": 1275
},
{
"epoch": 0.33,
"learning_rate": 7.725714285714286e-06,
"loss": 0.116,
"step": 1300
},
{
"epoch": 0.33,
"learning_rate": 7.654285714285715e-06,
"loss": 0.3363,
"step": 1325
},
{
"epoch": 0.34,
"learning_rate": 7.5828571428571444e-06,
"loss": 0.109,
"step": 1350
},
{
"epoch": 0.34,
"learning_rate": 7.511428571428572e-06,
"loss": 0.1077,
"step": 1375
},
{
"epoch": 0.35,
"learning_rate": 7.440000000000001e-06,
"loss": 0.0927,
"step": 1400
},
{
"epoch": 0.36,
"learning_rate": 7.36857142857143e-06,
"loss": 0.074,
"step": 1425
},
{
"epoch": 0.36,
"learning_rate": 7.297142857142858e-06,
"loss": 0.1369,
"step": 1450
},
{
"epoch": 0.37,
"learning_rate": 7.225714285714286e-06,
"loss": 0.1043,
"step": 1475
},
{
"epoch": 0.38,
"learning_rate": 7.154285714285715e-06,
"loss": 0.0997,
"step": 1500
},
{
"epoch": 0.38,
"learning_rate": 7.082857142857143e-06,
"loss": 0.0757,
"step": 1525
},
{
"epoch": 0.39,
"learning_rate": 7.011428571428572e-06,
"loss": 0.0928,
"step": 1550
},
{
"epoch": 0.39,
"learning_rate": 6.9400000000000005e-06,
"loss": 0.0726,
"step": 1575
},
{
"epoch": 0.4,
"learning_rate": 6.868571428571429e-06,
"loss": 0.0723,
"step": 1600
},
{
"epoch": 0.41,
"learning_rate": 6.797142857142858e-06,
"loss": 0.1479,
"step": 1625
},
{
"epoch": 0.41,
"learning_rate": 6.725714285714287e-06,
"loss": 0.0664,
"step": 1650
},
{
"epoch": 0.42,
"learning_rate": 6.654285714285716e-06,
"loss": 0.1177,
"step": 1675
},
{
"epoch": 0.42,
"learning_rate": 6.582857142857143e-06,
"loss": 0.2685,
"step": 1700
},
{
"epoch": 0.43,
"learning_rate": 6.511428571428572e-06,
"loss": 0.2067,
"step": 1725
},
{
"epoch": 0.44,
"learning_rate": 6.440000000000001e-06,
"loss": 0.0782,
"step": 1750
},
{
"epoch": 0.44,
"learning_rate": 6.368571428571429e-06,
"loss": 0.0838,
"step": 1775
},
{
"epoch": 0.45,
"learning_rate": 6.297142857142857e-06,
"loss": 0.188,
"step": 1800
},
{
"epoch": 0.46,
"learning_rate": 6.225714285714286e-06,
"loss": 0.3572,
"step": 1825
},
{
"epoch": 0.46,
"learning_rate": 6.1542857142857145e-06,
"loss": 0.1352,
"step": 1850
},
{
"epoch": 0.47,
"learning_rate": 6.0828571428571435e-06,
"loss": 0.082,
"step": 1875
},
{
"epoch": 0.47,
"learning_rate": 6.011428571428572e-06,
"loss": 0.0891,
"step": 1900
},
{
"epoch": 0.48,
"learning_rate": 5.94e-06,
"loss": 0.0739,
"step": 1925
},
{
"epoch": 0.49,
"learning_rate": 5.868571428571429e-06,
"loss": 0.0702,
"step": 1950
},
{
"epoch": 0.49,
"learning_rate": 5.797142857142858e-06,
"loss": 0.1163,
"step": 1975
},
{
"epoch": 0.5,
"learning_rate": 5.725714285714287e-06,
"loss": 0.0751,
"step": 2000
},
{
"epoch": 0.5,
"eval_loss": 0.10329055041074753,
"eval_runtime": 6831.581,
"eval_samples_per_second": 0.355,
"eval_steps_per_second": 0.044,
"eval_wer": 5.390462338490352,
"step": 2000
},
{
"epoch": 0.51,
"learning_rate": 5.654285714285714e-06,
"loss": 0.1457,
"step": 2025
},
{
"epoch": 0.51,
"learning_rate": 5.582857142857143e-06,
"loss": 0.2301,
"step": 2050
},
{
"epoch": 0.52,
"learning_rate": 5.511428571428572e-06,
"loss": 0.1268,
"step": 2075
},
{
"epoch": 0.53,
"learning_rate": 5.4400000000000004e-06,
"loss": 0.0593,
"step": 2100
},
{
"epoch": 0.53,
"learning_rate": 5.368571428571429e-06,
"loss": 0.0815,
"step": 2125
},
{
"epoch": 0.54,
"learning_rate": 5.297142857142858e-06,
"loss": 0.0908,
"step": 2150
},
{
"epoch": 0.54,
"learning_rate": 5.225714285714286e-06,
"loss": 0.1282,
"step": 2175
},
{
"epoch": 0.55,
"learning_rate": 5.154285714285715e-06,
"loss": 0.1269,
"step": 2200
},
{
"epoch": 0.56,
"learning_rate": 5.082857142857144e-06,
"loss": 0.1063,
"step": 2225
},
{
"epoch": 0.56,
"learning_rate": 5.011428571428571e-06,
"loss": 0.0867,
"step": 2250
},
{
"epoch": 0.57,
"learning_rate": 4.94e-06,
"loss": 0.0746,
"step": 2275
},
{
"epoch": 0.57,
"learning_rate": 4.868571428571429e-06,
"loss": 0.0794,
"step": 2300
},
{
"epoch": 0.58,
"learning_rate": 4.797142857142857e-06,
"loss": 0.4292,
"step": 2325
},
{
"epoch": 0.59,
"learning_rate": 4.725714285714286e-06,
"loss": 0.2346,
"step": 2350
},
{
"epoch": 0.59,
"learning_rate": 4.6542857142857145e-06,
"loss": 0.0712,
"step": 2375
},
{
"epoch": 0.6,
"learning_rate": 4.5828571428571435e-06,
"loss": 0.0567,
"step": 2400
},
{
"epoch": 0.61,
"learning_rate": 4.511428571428572e-06,
"loss": 0.0548,
"step": 2425
},
{
"epoch": 0.61,
"learning_rate": 4.440000000000001e-06,
"loss": 0.0599,
"step": 2450
},
{
"epoch": 0.62,
"learning_rate": 4.368571428571429e-06,
"loss": 0.068,
"step": 2475
},
{
"epoch": 0.62,
"learning_rate": 4.297142857142858e-06,
"loss": 0.4456,
"step": 2500
},
{
"epoch": 0.63,
"learning_rate": 4.225714285714286e-06,
"loss": 0.1059,
"step": 2525
},
{
"epoch": 0.64,
"learning_rate": 4.154285714285714e-06,
"loss": 0.1816,
"step": 2550
},
{
"epoch": 0.64,
"learning_rate": 4.082857142857143e-06,
"loss": 0.1162,
"step": 2575
},
{
"epoch": 0.65,
"learning_rate": 4.011428571428571e-06,
"loss": 0.1061,
"step": 2600
},
{
"epoch": 0.66,
"learning_rate": 3.94e-06,
"loss": 0.1751,
"step": 2625
},
{
"epoch": 0.66,
"learning_rate": 3.8685714285714286e-06,
"loss": 0.259,
"step": 2650
},
{
"epoch": 0.67,
"learning_rate": 3.7971428571428576e-06,
"loss": 0.1915,
"step": 2675
},
{
"epoch": 0.68,
"learning_rate": 3.7257142857142857e-06,
"loss": 0.0778,
"step": 2700
},
{
"epoch": 0.68,
"learning_rate": 3.6542857142857148e-06,
"loss": 0.0664,
"step": 2725
},
{
"epoch": 0.69,
"learning_rate": 3.582857142857143e-06,
"loss": 0.12,
"step": 2750
},
{
"epoch": 0.69,
"learning_rate": 3.511428571428572e-06,
"loss": 0.0841,
"step": 2775
},
{
"epoch": 0.7,
"learning_rate": 3.44e-06,
"loss": 0.1226,
"step": 2800
},
{
"epoch": 0.71,
"learning_rate": 3.3685714285714287e-06,
"loss": 0.1037,
"step": 2825
},
{
"epoch": 0.71,
"learning_rate": 3.2971428571428577e-06,
"loss": 0.068,
"step": 2850
},
{
"epoch": 0.72,
"learning_rate": 3.225714285714286e-06,
"loss": 0.0879,
"step": 2875
},
{
"epoch": 0.72,
"learning_rate": 3.154285714285715e-06,
"loss": 0.0941,
"step": 2900
},
{
"epoch": 0.73,
"learning_rate": 3.082857142857143e-06,
"loss": 0.0577,
"step": 2925
},
{
"epoch": 0.74,
"learning_rate": 3.0114285714285716e-06,
"loss": 0.0587,
"step": 2950
},
{
"epoch": 0.74,
"learning_rate": 2.9400000000000002e-06,
"loss": 0.0486,
"step": 2975
},
{
"epoch": 0.75,
"learning_rate": 2.868571428571429e-06,
"loss": 0.0613,
"step": 3000
},
{
"epoch": 0.75,
"eval_loss": 0.09700169414281845,
"eval_runtime": 6788.178,
"eval_samples_per_second": 0.357,
"eval_steps_per_second": 0.045,
"eval_wer": 4.91925762843566,
"step": 3000
},
{
"epoch": 0.76,
"learning_rate": 2.797142857142857e-06,
"loss": 0.0781,
"step": 3025
},
{
"epoch": 0.76,
"learning_rate": 2.725714285714286e-06,
"loss": 0.0807,
"step": 3050
},
{
"epoch": 0.77,
"learning_rate": 2.654285714285714e-06,
"loss": 0.1322,
"step": 3075
},
{
"epoch": 0.78,
"learning_rate": 2.582857142857143e-06,
"loss": 0.3381,
"step": 3100
},
{
"epoch": 0.78,
"learning_rate": 2.5114285714285718e-06,
"loss": 0.1083,
"step": 3125
},
{
"epoch": 0.79,
"learning_rate": 2.4400000000000004e-06,
"loss": 0.0692,
"step": 3150
},
{
"epoch": 0.79,
"learning_rate": 2.3685714285714285e-06,
"loss": 0.237,
"step": 3175
},
{
"epoch": 0.8,
"learning_rate": 2.297142857142857e-06,
"loss": 0.178,
"step": 3200
},
{
"epoch": 0.81,
"learning_rate": 2.2257142857142857e-06,
"loss": 0.0748,
"step": 3225
},
{
"epoch": 0.81,
"learning_rate": 2.1542857142857147e-06,
"loss": 0.0661,
"step": 3250
},
{
"epoch": 0.82,
"learning_rate": 2.0828571428571433e-06,
"loss": 0.0694,
"step": 3275
},
{
"epoch": 0.82,
"learning_rate": 2.0114285714285715e-06,
"loss": 0.0856,
"step": 3300
},
{
"epoch": 0.83,
"learning_rate": 1.94e-06,
"loss": 0.1095,
"step": 3325
},
{
"epoch": 0.84,
"learning_rate": 1.8685714285714289e-06,
"loss": 0.0858,
"step": 3350
},
{
"epoch": 0.84,
"learning_rate": 1.8000000000000001e-06,
"loss": 0.0745,
"step": 3375
},
{
"epoch": 0.85,
"learning_rate": 1.7285714285714287e-06,
"loss": 0.0733,
"step": 3400
},
{
"epoch": 0.86,
"learning_rate": 1.657142857142857e-06,
"loss": 0.0549,
"step": 3425
},
{
"epoch": 0.86,
"learning_rate": 1.5857142857142857e-06,
"loss": 0.2172,
"step": 3450
},
{
"epoch": 0.87,
"learning_rate": 1.5142857142857145e-06,
"loss": 0.2124,
"step": 3475
},
{
"epoch": 0.88,
"learning_rate": 1.442857142857143e-06,
"loss": 0.1138,
"step": 3500
},
{
"epoch": 0.88,
"learning_rate": 1.3714285714285717e-06,
"loss": 0.081,
"step": 3525
},
{
"epoch": 0.89,
"learning_rate": 1.3e-06,
"loss": 0.0686,
"step": 3550
},
{
"epoch": 0.89,
"learning_rate": 1.2285714285714286e-06,
"loss": 0.0637,
"step": 3575
},
{
"epoch": 0.9,
"learning_rate": 1.1571428571428572e-06,
"loss": 0.0654,
"step": 3600
},
{
"epoch": 0.91,
"learning_rate": 1.0857142857142858e-06,
"loss": 0.1011,
"step": 3625
},
{
"epoch": 0.91,
"learning_rate": 1.0142857142857144e-06,
"loss": 0.2169,
"step": 3650
},
{
"epoch": 0.92,
"learning_rate": 9.42857142857143e-07,
"loss": 0.1365,
"step": 3675
},
{
"epoch": 0.93,
"learning_rate": 8.714285714285716e-07,
"loss": 0.2004,
"step": 3700
},
{
"epoch": 0.93,
"learning_rate": 8.000000000000001e-07,
"loss": 0.0842,
"step": 3725
},
{
"epoch": 0.94,
"learning_rate": 7.285714285714287e-07,
"loss": 0.0529,
"step": 3750
},
{
"epoch": 0.94,
"learning_rate": 6.571428571428571e-07,
"loss": 0.1639,
"step": 3775
},
{
"epoch": 0.95,
"learning_rate": 5.857142857142857e-07,
"loss": 0.1324,
"step": 3800
},
{
"epoch": 0.96,
"learning_rate": 5.142857142857143e-07,
"loss": 0.0956,
"step": 3825
},
{
"epoch": 0.96,
"learning_rate": 4.4285714285714286e-07,
"loss": 0.1126,
"step": 3850
},
{
"epoch": 0.97,
"learning_rate": 3.7142857142857145e-07,
"loss": 0.0956,
"step": 3875
},
{
"epoch": 0.97,
"learning_rate": 3.0000000000000004e-07,
"loss": 0.1001,
"step": 3900
},
{
"epoch": 0.98,
"learning_rate": 2.285714285714286e-07,
"loss": 0.4695,
"step": 3925
},
{
"epoch": 0.99,
"learning_rate": 1.5714285714285717e-07,
"loss": 0.1176,
"step": 3950
},
{
"epoch": 0.99,
"learning_rate": 8.571428571428573e-08,
"loss": 0.0856,
"step": 3975
},
{
"epoch": 1.0,
"learning_rate": 1.4285714285714288e-08,
"loss": 0.0796,
"step": 4000
},
{
"epoch": 1.0,
"eval_loss": 0.09033442288637161,
"eval_runtime": 6797.8014,
"eval_samples_per_second": 0.357,
"eval_steps_per_second": 0.045,
"eval_wer": 4.561620226935377,
"step": 4000
},
{
"epoch": 1.0,
"step": 4000,
"total_flos": 5.435317813248e+20,
"train_loss": 0.14483159190416336,
"train_runtime": 109034.6521,
"train_samples_per_second": 2.348,
"train_steps_per_second": 0.037
}
],
"max_steps": 4000,
"num_train_epochs": 9223372036854775807,
"total_flos": 5.435317813248e+20,
"trial_name": null,
"trial_params": null
}