whisper-small-it / trainer_state.json
EdoAbati's picture
End of training
4327de1
raw
history blame
20.6 kB
{
"best_metric": 9.26934935147778,
"best_model_checkpoint": "./checkpoint-4000",
"epoch": 1.40375,
"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": 4.4e-07,
"loss": 1.0133,
"step": 25
},
{
"epoch": 0.01,
"learning_rate": 9.400000000000001e-07,
"loss": 0.8159,
"step": 50
},
{
"epoch": 0.02,
"learning_rate": 1.44e-06,
"loss": 0.5038,
"step": 75
},
{
"epoch": 0.03,
"learning_rate": 1.94e-06,
"loss": 0.4105,
"step": 100
},
{
"epoch": 0.03,
"learning_rate": 2.4400000000000004e-06,
"loss": 0.3996,
"step": 125
},
{
"epoch": 0.04,
"learning_rate": 2.9400000000000002e-06,
"loss": 0.347,
"step": 150
},
{
"epoch": 0.04,
"learning_rate": 3.44e-06,
"loss": 0.4348,
"step": 175
},
{
"epoch": 0.05,
"learning_rate": 3.94e-06,
"loss": 0.3964,
"step": 200
},
{
"epoch": 0.06,
"learning_rate": 4.440000000000001e-06,
"loss": 0.3034,
"step": 225
},
{
"epoch": 0.06,
"learning_rate": 4.94e-06,
"loss": 0.2421,
"step": 250
},
{
"epoch": 0.07,
"learning_rate": 5.4400000000000004e-06,
"loss": 0.2746,
"step": 275
},
{
"epoch": 0.07,
"learning_rate": 5.94e-06,
"loss": 0.2483,
"step": 300
},
{
"epoch": 0.08,
"learning_rate": 6.440000000000001e-06,
"loss": 0.2342,
"step": 325
},
{
"epoch": 0.09,
"learning_rate": 6.9400000000000005e-06,
"loss": 0.2251,
"step": 350
},
{
"epoch": 0.09,
"learning_rate": 7.440000000000001e-06,
"loss": 0.2035,
"step": 375
},
{
"epoch": 0.1,
"learning_rate": 7.94e-06,
"loss": 0.1945,
"step": 400
},
{
"epoch": 0.11,
"learning_rate": 8.44e-06,
"loss": 0.1967,
"step": 425
},
{
"epoch": 0.11,
"learning_rate": 8.94e-06,
"loss": 0.2041,
"step": 450
},
{
"epoch": 0.12,
"learning_rate": 9.440000000000001e-06,
"loss": 0.2311,
"step": 475
},
{
"epoch": 0.12,
"learning_rate": 9.940000000000001e-06,
"loss": 0.1878,
"step": 500
},
{
"epoch": 0.13,
"learning_rate": 9.937142857142858e-06,
"loss": 0.1996,
"step": 525
},
{
"epoch": 0.14,
"learning_rate": 9.865714285714285e-06,
"loss": 0.2163,
"step": 550
},
{
"epoch": 0.14,
"learning_rate": 9.794285714285714e-06,
"loss": 0.2071,
"step": 575
},
{
"epoch": 0.15,
"learning_rate": 9.722857142857143e-06,
"loss": 0.2486,
"step": 600
},
{
"epoch": 0.16,
"learning_rate": 9.651428571428572e-06,
"loss": 0.2038,
"step": 625
},
{
"epoch": 0.16,
"learning_rate": 9.58e-06,
"loss": 0.2007,
"step": 650
},
{
"epoch": 0.17,
"learning_rate": 9.508571428571429e-06,
"loss": 0.216,
"step": 675
},
{
"epoch": 0.17,
"learning_rate": 9.437142857142858e-06,
"loss": 0.2132,
"step": 700
},
{
"epoch": 0.18,
"learning_rate": 9.365714285714287e-06,
"loss": 0.1804,
"step": 725
},
{
"epoch": 0.19,
"learning_rate": 9.294285714285714e-06,
"loss": 0.1888,
"step": 750
},
{
"epoch": 0.19,
"learning_rate": 9.222857142857143e-06,
"loss": 0.1907,
"step": 775
},
{
"epoch": 0.2,
"learning_rate": 9.151428571428572e-06,
"loss": 0.2283,
"step": 800
},
{
"epoch": 0.21,
"learning_rate": 9.080000000000001e-06,
"loss": 0.1758,
"step": 825
},
{
"epoch": 0.21,
"learning_rate": 9.00857142857143e-06,
"loss": 0.1676,
"step": 850
},
{
"epoch": 0.22,
"learning_rate": 8.937142857142857e-06,
"loss": 0.1707,
"step": 875
},
{
"epoch": 0.23,
"learning_rate": 8.865714285714287e-06,
"loss": 0.1903,
"step": 900
},
{
"epoch": 0.23,
"learning_rate": 8.794285714285716e-06,
"loss": 0.1788,
"step": 925
},
{
"epoch": 0.24,
"learning_rate": 8.722857142857145e-06,
"loss": 0.235,
"step": 950
},
{
"epoch": 0.24,
"learning_rate": 8.651428571428572e-06,
"loss": 0.3076,
"step": 975
},
{
"epoch": 0.25,
"learning_rate": 8.580000000000001e-06,
"loss": 0.2851,
"step": 1000
},
{
"epoch": 0.25,
"eval_loss": 0.26039254665374756,
"eval_runtime": 4150.3489,
"eval_samples_per_second": 3.615,
"eval_steps_per_second": 0.113,
"eval_wer": 11.974404635339145,
"step": 1000
},
{
"epoch": 0.26,
"learning_rate": 8.50857142857143e-06,
"loss": 0.2049,
"step": 1025
},
{
"epoch": 0.26,
"learning_rate": 8.437142857142859e-06,
"loss": 0.1866,
"step": 1050
},
{
"epoch": 0.27,
"learning_rate": 8.365714285714286e-06,
"loss": 0.1712,
"step": 1075
},
{
"epoch": 0.28,
"learning_rate": 8.294285714285715e-06,
"loss": 0.2156,
"step": 1100
},
{
"epoch": 0.28,
"learning_rate": 8.222857142857144e-06,
"loss": 0.2191,
"step": 1125
},
{
"epoch": 0.29,
"learning_rate": 8.151428571428572e-06,
"loss": 0.2273,
"step": 1150
},
{
"epoch": 0.29,
"learning_rate": 8.08e-06,
"loss": 0.2345,
"step": 1175
},
{
"epoch": 0.3,
"learning_rate": 8.00857142857143e-06,
"loss": 0.1985,
"step": 1200
},
{
"epoch": 0.31,
"learning_rate": 7.937142857142857e-06,
"loss": 0.232,
"step": 1225
},
{
"epoch": 0.31,
"learning_rate": 7.865714285714286e-06,
"loss": 0.1998,
"step": 1250
},
{
"epoch": 0.32,
"learning_rate": 7.794285714285715e-06,
"loss": 0.2125,
"step": 1275
},
{
"epoch": 0.33,
"learning_rate": 7.722857142857142e-06,
"loss": 0.2036,
"step": 1300
},
{
"epoch": 0.33,
"learning_rate": 7.651428571428571e-06,
"loss": 0.1937,
"step": 1325
},
{
"epoch": 0.34,
"learning_rate": 7.58e-06,
"loss": 0.2159,
"step": 1350
},
{
"epoch": 0.34,
"learning_rate": 7.508571428571429e-06,
"loss": 0.1933,
"step": 1375
},
{
"epoch": 0.35,
"learning_rate": 7.4371428571428575e-06,
"loss": 0.1921,
"step": 1400
},
{
"epoch": 0.36,
"learning_rate": 7.365714285714286e-06,
"loss": 0.2039,
"step": 1425
},
{
"epoch": 0.36,
"learning_rate": 7.294285714285715e-06,
"loss": 0.2126,
"step": 1450
},
{
"epoch": 0.37,
"learning_rate": 7.222857142857144e-06,
"loss": 0.1878,
"step": 1475
},
{
"epoch": 0.38,
"learning_rate": 7.151428571428573e-06,
"loss": 0.2299,
"step": 1500
},
{
"epoch": 0.38,
"learning_rate": 7.08e-06,
"loss": 0.214,
"step": 1525
},
{
"epoch": 0.39,
"learning_rate": 7.008571428571429e-06,
"loss": 0.2033,
"step": 1550
},
{
"epoch": 0.39,
"learning_rate": 6.937142857142858e-06,
"loss": 0.2081,
"step": 1575
},
{
"epoch": 0.4,
"learning_rate": 6.865714285714287e-06,
"loss": 0.1898,
"step": 1600
},
{
"epoch": 0.41,
"learning_rate": 6.794285714285714e-06,
"loss": 0.1874,
"step": 1625
},
{
"epoch": 0.41,
"learning_rate": 6.722857142857143e-06,
"loss": 0.1899,
"step": 1650
},
{
"epoch": 0.42,
"learning_rate": 6.651428571428572e-06,
"loss": 0.1752,
"step": 1675
},
{
"epoch": 0.42,
"learning_rate": 6.5800000000000005e-06,
"loss": 0.1976,
"step": 1700
},
{
"epoch": 0.43,
"learning_rate": 6.5085714285714295e-06,
"loss": 0.1678,
"step": 1725
},
{
"epoch": 0.44,
"learning_rate": 6.437142857142858e-06,
"loss": 0.1935,
"step": 1750
},
{
"epoch": 0.44,
"learning_rate": 6.365714285714286e-06,
"loss": 0.216,
"step": 1775
},
{
"epoch": 0.45,
"learning_rate": 6.294285714285715e-06,
"loss": 0.2373,
"step": 1800
},
{
"epoch": 0.46,
"learning_rate": 6.222857142857144e-06,
"loss": 0.2108,
"step": 1825
},
{
"epoch": 0.46,
"learning_rate": 6.151428571428571e-06,
"loss": 0.1942,
"step": 1850
},
{
"epoch": 0.47,
"learning_rate": 6.08e-06,
"loss": 0.1926,
"step": 1875
},
{
"epoch": 0.47,
"learning_rate": 6.008571428571429e-06,
"loss": 0.212,
"step": 1900
},
{
"epoch": 0.48,
"learning_rate": 5.937142857142858e-06,
"loss": 0.1892,
"step": 1925
},
{
"epoch": 0.49,
"learning_rate": 5.865714285714286e-06,
"loss": 0.1573,
"step": 1950
},
{
"epoch": 0.49,
"learning_rate": 5.794285714285715e-06,
"loss": 0.146,
"step": 1975
},
{
"epoch": 0.5,
"learning_rate": 5.722857142857144e-06,
"loss": 0.1885,
"step": 2000
},
{
"epoch": 0.5,
"eval_loss": 0.21763817965984344,
"eval_runtime": 4204.389,
"eval_samples_per_second": 3.568,
"eval_steps_per_second": 0.112,
"eval_wer": 10.135817563257495,
"step": 2000
},
{
"epoch": 0.51,
"learning_rate": 5.651428571428572e-06,
"loss": 0.1969,
"step": 2025
},
{
"epoch": 0.51,
"learning_rate": 5.580000000000001e-06,
"loss": 0.1978,
"step": 2050
},
{
"epoch": 0.52,
"learning_rate": 5.508571428571429e-06,
"loss": 0.1716,
"step": 2075
},
{
"epoch": 0.53,
"learning_rate": 5.437142857142857e-06,
"loss": 0.1924,
"step": 2100
},
{
"epoch": 0.53,
"learning_rate": 5.365714285714286e-06,
"loss": 0.1559,
"step": 2125
},
{
"epoch": 0.54,
"learning_rate": 5.294285714285715e-06,
"loss": 0.1583,
"step": 2150
},
{
"epoch": 0.54,
"learning_rate": 5.2228571428571425e-06,
"loss": 0.1961,
"step": 2175
},
{
"epoch": 0.55,
"learning_rate": 5.1514285714285715e-06,
"loss": 0.1755,
"step": 2200
},
{
"epoch": 0.56,
"learning_rate": 5.0800000000000005e-06,
"loss": 0.1352,
"step": 2225
},
{
"epoch": 0.56,
"learning_rate": 5.0085714285714295e-06,
"loss": 0.2097,
"step": 2250
},
{
"epoch": 0.57,
"learning_rate": 4.937142857142858e-06,
"loss": 0.1651,
"step": 2275
},
{
"epoch": 0.57,
"learning_rate": 4.865714285714287e-06,
"loss": 0.2335,
"step": 2300
},
{
"epoch": 0.58,
"learning_rate": 4.794285714285715e-06,
"loss": 0.1868,
"step": 2325
},
{
"epoch": 0.59,
"learning_rate": 4.722857142857144e-06,
"loss": 0.1674,
"step": 2350
},
{
"epoch": 0.59,
"learning_rate": 4.651428571428572e-06,
"loss": 0.1619,
"step": 2375
},
{
"epoch": 1.0,
"learning_rate": 4.58e-06,
"loss": 0.1326,
"step": 2400
},
{
"epoch": 1.01,
"learning_rate": 4.508571428571429e-06,
"loss": 0.1076,
"step": 2425
},
{
"epoch": 1.02,
"learning_rate": 4.437142857142857e-06,
"loss": 0.1034,
"step": 2450
},
{
"epoch": 1.02,
"learning_rate": 4.3657142857142855e-06,
"loss": 0.1235,
"step": 2475
},
{
"epoch": 1.03,
"learning_rate": 4.2942857142857146e-06,
"loss": 0.1015,
"step": 2500
},
{
"epoch": 1.03,
"learning_rate": 4.222857142857143e-06,
"loss": 0.1051,
"step": 2525
},
{
"epoch": 1.04,
"learning_rate": 4.151428571428572e-06,
"loss": 0.1102,
"step": 2550
},
{
"epoch": 1.05,
"learning_rate": 4.08e-06,
"loss": 0.1108,
"step": 2575
},
{
"epoch": 1.05,
"learning_rate": 4.008571428571429e-06,
"loss": 0.0878,
"step": 2600
},
{
"epoch": 1.06,
"learning_rate": 3.937142857142858e-06,
"loss": 0.0935,
"step": 2625
},
{
"epoch": 1.07,
"learning_rate": 3.865714285714286e-06,
"loss": 0.0991,
"step": 2650
},
{
"epoch": 1.07,
"learning_rate": 3.7942857142857147e-06,
"loss": 0.1089,
"step": 2675
},
{
"epoch": 1.08,
"learning_rate": 3.722857142857143e-06,
"loss": 0.0908,
"step": 2700
},
{
"epoch": 1.08,
"learning_rate": 3.651428571428572e-06,
"loss": 0.0827,
"step": 2725
},
{
"epoch": 1.09,
"learning_rate": 3.58e-06,
"loss": 0.0902,
"step": 2750
},
{
"epoch": 1.1,
"learning_rate": 3.508571428571429e-06,
"loss": 0.0993,
"step": 2775
},
{
"epoch": 1.1,
"learning_rate": 3.437142857142857e-06,
"loss": 0.0953,
"step": 2800
},
{
"epoch": 1.11,
"learning_rate": 3.3657142857142862e-06,
"loss": 0.1185,
"step": 2825
},
{
"epoch": 1.12,
"learning_rate": 3.2942857142857144e-06,
"loss": 0.1628,
"step": 2850
},
{
"epoch": 1.12,
"learning_rate": 3.222857142857143e-06,
"loss": 0.1548,
"step": 2875
},
{
"epoch": 1.13,
"learning_rate": 3.151428571428572e-06,
"loss": 0.1175,
"step": 2900
},
{
"epoch": 1.14,
"learning_rate": 3.08e-06,
"loss": 0.1062,
"step": 2925
},
{
"epoch": 1.14,
"learning_rate": 3.008571428571429e-06,
"loss": 0.0934,
"step": 2950
},
{
"epoch": 1.15,
"learning_rate": 2.9371428571428573e-06,
"loss": 0.1239,
"step": 2975
},
{
"epoch": 1.15,
"learning_rate": 2.865714285714286e-06,
"loss": 0.1176,
"step": 3000
},
{
"epoch": 1.15,
"eval_loss": 0.21106119453907013,
"eval_runtime": 4138.4328,
"eval_samples_per_second": 3.625,
"eval_steps_per_second": 0.113,
"eval_wer": 9.566367212417607,
"step": 3000
},
{
"epoch": 1.16,
"learning_rate": 2.7942857142857145e-06,
"loss": 0.1303,
"step": 3025
},
{
"epoch": 1.17,
"learning_rate": 2.722857142857143e-06,
"loss": 0.1489,
"step": 3050
},
{
"epoch": 1.17,
"learning_rate": 2.6514285714285713e-06,
"loss": 0.1402,
"step": 3075
},
{
"epoch": 1.18,
"learning_rate": 2.5800000000000003e-06,
"loss": 0.1265,
"step": 3100
},
{
"epoch": 1.19,
"learning_rate": 2.5085714285714285e-06,
"loss": 0.1174,
"step": 3125
},
{
"epoch": 1.19,
"learning_rate": 2.4371428571428575e-06,
"loss": 0.124,
"step": 3150
},
{
"epoch": 1.2,
"learning_rate": 2.365714285714286e-06,
"loss": 0.125,
"step": 3175
},
{
"epoch": 1.2,
"learning_rate": 2.2942857142857146e-06,
"loss": 0.1059,
"step": 3200
},
{
"epoch": 1.21,
"learning_rate": 2.222857142857143e-06,
"loss": 0.0909,
"step": 3225
},
{
"epoch": 1.22,
"learning_rate": 2.1514285714285714e-06,
"loss": 0.1227,
"step": 3250
},
{
"epoch": 1.22,
"learning_rate": 2.08e-06,
"loss": 0.1259,
"step": 3275
},
{
"epoch": 1.23,
"learning_rate": 2.0085714285714286e-06,
"loss": 0.1268,
"step": 3300
},
{
"epoch": 1.23,
"learning_rate": 1.9371428571428576e-06,
"loss": 0.1133,
"step": 3325
},
{
"epoch": 1.24,
"learning_rate": 1.865714285714286e-06,
"loss": 0.1203,
"step": 3350
},
{
"epoch": 1.25,
"learning_rate": 1.7942857142857146e-06,
"loss": 0.0974,
"step": 3375
},
{
"epoch": 1.25,
"learning_rate": 1.7228571428571432e-06,
"loss": 0.1053,
"step": 3400
},
{
"epoch": 1.26,
"learning_rate": 1.6514285714285715e-06,
"loss": 0.1304,
"step": 3425
},
{
"epoch": 1.27,
"learning_rate": 1.5800000000000001e-06,
"loss": 0.1118,
"step": 3450
},
{
"epoch": 1.27,
"learning_rate": 1.5085714285714287e-06,
"loss": 0.0886,
"step": 3475
},
{
"epoch": 1.28,
"learning_rate": 1.4371428571428573e-06,
"loss": 0.135,
"step": 3500
},
{
"epoch": 1.28,
"learning_rate": 1.3657142857142857e-06,
"loss": 0.1121,
"step": 3525
},
{
"epoch": 1.29,
"learning_rate": 1.2942857142857143e-06,
"loss": 0.152,
"step": 3550
},
{
"epoch": 1.3,
"learning_rate": 1.222857142857143e-06,
"loss": 0.1359,
"step": 3575
},
{
"epoch": 1.3,
"learning_rate": 1.1514285714285714e-06,
"loss": 0.1203,
"step": 3600
},
{
"epoch": 1.31,
"learning_rate": 1.08e-06,
"loss": 0.1128,
"step": 3625
},
{
"epoch": 1.32,
"learning_rate": 1.0085714285714286e-06,
"loss": 0.1219,
"step": 3650
},
{
"epoch": 1.32,
"learning_rate": 9.371428571428571e-07,
"loss": 0.1262,
"step": 3675
},
{
"epoch": 1.33,
"learning_rate": 8.657142857142858e-07,
"loss": 0.1079,
"step": 3700
},
{
"epoch": 1.33,
"learning_rate": 7.942857142857144e-07,
"loss": 0.1288,
"step": 3725
},
{
"epoch": 1.34,
"learning_rate": 7.228571428571429e-07,
"loss": 0.1093,
"step": 3750
},
{
"epoch": 1.35,
"learning_rate": 6.514285714285715e-07,
"loss": 0.1168,
"step": 3775
},
{
"epoch": 1.35,
"learning_rate": 5.800000000000001e-07,
"loss": 0.1115,
"step": 3800
},
{
"epoch": 1.36,
"learning_rate": 5.085714285714286e-07,
"loss": 0.1013,
"step": 3825
},
{
"epoch": 1.37,
"learning_rate": 4.371428571428572e-07,
"loss": 0.1175,
"step": 3850
},
{
"epoch": 1.37,
"learning_rate": 3.657142857142858e-07,
"loss": 0.1077,
"step": 3875
},
{
"epoch": 1.38,
"learning_rate": 2.942857142857143e-07,
"loss": 0.1152,
"step": 3900
},
{
"epoch": 1.39,
"learning_rate": 2.228571428571429e-07,
"loss": 0.1134,
"step": 3925
},
{
"epoch": 1.39,
"learning_rate": 1.5142857142857144e-07,
"loss": 0.1207,
"step": 3950
},
{
"epoch": 1.4,
"learning_rate": 8e-08,
"loss": 0.1132,
"step": 3975
},
{
"epoch": 1.4,
"learning_rate": 8.571428571428572e-09,
"loss": 0.1256,
"step": 4000
},
{
"epoch": 1.4,
"eval_loss": 0.201337069272995,
"eval_runtime": 4228.3195,
"eval_samples_per_second": 3.548,
"eval_steps_per_second": 0.111,
"eval_wer": 9.26934935147778,
"step": 4000
},
{
"epoch": 1.4,
"step": 4000,
"total_flos": 7.386891633451008e+19,
"train_loss": 0.18421337854862213,
"train_runtime": 44278.3137,
"train_samples_per_second": 5.782,
"train_steps_per_second": 0.09
}
],
"max_steps": 4000,
"num_train_epochs": 9223372036854775807,
"total_flos": 7.386891633451008e+19,
"trial_name": null,
"trial_params": null
}