whisper-small-miso / checkpoint-4000 /trainer_state.json
mskov's picture
Upload 5 files
d941328
{
"best_metric": 111.76470588235294,
"best_model_checkpoint": "mskov/whisper_miso/checkpoint-1000",
"epoch": 1333.3333333333333,
"global_step": 4000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 8.33,
"learning_rate": 5.000000000000001e-07,
"loss": 4.8805,
"step": 25
},
{
"epoch": 16.67,
"learning_rate": 1.0000000000000002e-06,
"loss": 3.3934,
"step": 50
},
{
"epoch": 25.0,
"learning_rate": 1.5e-06,
"loss": 1.6092,
"step": 75
},
{
"epoch": 33.33,
"learning_rate": 2.0000000000000003e-06,
"loss": 0.1399,
"step": 100
},
{
"epoch": 41.67,
"learning_rate": 2.5e-06,
"loss": 0.0008,
"step": 125
},
{
"epoch": 50.0,
"learning_rate": 3e-06,
"loss": 0.0002,
"step": 150
},
{
"epoch": 58.33,
"learning_rate": 3.5e-06,
"loss": 0.0001,
"step": 175
},
{
"epoch": 66.67,
"learning_rate": 4.000000000000001e-06,
"loss": 0.0001,
"step": 200
},
{
"epoch": 75.0,
"learning_rate": 4.5e-06,
"loss": 0.0001,
"step": 225
},
{
"epoch": 83.33,
"learning_rate": 5e-06,
"loss": 0.0001,
"step": 250
},
{
"epoch": 91.67,
"learning_rate": 5.500000000000001e-06,
"loss": 0.0,
"step": 275
},
{
"epoch": 100.0,
"learning_rate": 6e-06,
"loss": 0.0,
"step": 300
},
{
"epoch": 108.33,
"learning_rate": 6.5000000000000004e-06,
"loss": 0.0,
"step": 325
},
{
"epoch": 116.67,
"learning_rate": 7e-06,
"loss": 0.0,
"step": 350
},
{
"epoch": 125.0,
"learning_rate": 7.500000000000001e-06,
"loss": 0.0,
"step": 375
},
{
"epoch": 133.33,
"learning_rate": 8.000000000000001e-06,
"loss": 0.0,
"step": 400
},
{
"epoch": 141.67,
"learning_rate": 8.5e-06,
"loss": 0.0,
"step": 425
},
{
"epoch": 150.0,
"learning_rate": 9e-06,
"loss": 0.0,
"step": 450
},
{
"epoch": 158.33,
"learning_rate": 9.5e-06,
"loss": 0.0,
"step": 475
},
{
"epoch": 166.67,
"learning_rate": 1e-05,
"loss": 0.0,
"step": 500
},
{
"epoch": 175.0,
"learning_rate": 9.92857142857143e-06,
"loss": 0.0,
"step": 525
},
{
"epoch": 183.33,
"learning_rate": 9.857142857142859e-06,
"loss": 0.0,
"step": 550
},
{
"epoch": 191.67,
"learning_rate": 9.785714285714286e-06,
"loss": 0.0,
"step": 575
},
{
"epoch": 200.0,
"learning_rate": 9.714285714285715e-06,
"loss": 0.0,
"step": 600
},
{
"epoch": 208.33,
"learning_rate": 9.642857142857144e-06,
"loss": 0.0,
"step": 625
},
{
"epoch": 216.67,
"learning_rate": 9.571428571428573e-06,
"loss": 0.0,
"step": 650
},
{
"epoch": 225.0,
"learning_rate": 9.5e-06,
"loss": 0.0,
"step": 675
},
{
"epoch": 233.33,
"learning_rate": 9.42857142857143e-06,
"loss": 0.0,
"step": 700
},
{
"epoch": 241.67,
"learning_rate": 9.357142857142859e-06,
"loss": 0.0,
"step": 725
},
{
"epoch": 250.0,
"learning_rate": 9.285714285714288e-06,
"loss": 0.0,
"step": 750
},
{
"epoch": 258.33,
"learning_rate": 9.214285714285715e-06,
"loss": 0.0,
"step": 775
},
{
"epoch": 266.67,
"learning_rate": 9.142857142857144e-06,
"loss": 0.0,
"step": 800
},
{
"epoch": 275.0,
"learning_rate": 9.071428571428573e-06,
"loss": 0.0,
"step": 825
},
{
"epoch": 283.33,
"learning_rate": 9e-06,
"loss": 0.0,
"step": 850
},
{
"epoch": 291.67,
"learning_rate": 8.92857142857143e-06,
"loss": 0.0,
"step": 875
},
{
"epoch": 300.0,
"learning_rate": 8.857142857142858e-06,
"loss": 0.0,
"step": 900
},
{
"epoch": 308.33,
"learning_rate": 8.785714285714286e-06,
"loss": 0.0,
"step": 925
},
{
"epoch": 316.67,
"learning_rate": 8.714285714285715e-06,
"loss": 0.0,
"step": 950
},
{
"epoch": 325.0,
"learning_rate": 8.642857142857144e-06,
"loss": 0.0,
"step": 975
},
{
"epoch": 333.33,
"learning_rate": 8.571428571428571e-06,
"loss": 0.0,
"step": 1000
},
{
"epoch": 333.33,
"eval_loss": 0.3114524483680725,
"eval_runtime": 13.4346,
"eval_samples_per_second": 1.265,
"eval_steps_per_second": 0.223,
"eval_wer": 111.76470588235294,
"step": 1000
},
{
"epoch": 341.67,
"learning_rate": 8.5e-06,
"loss": 0.0,
"step": 1025
},
{
"epoch": 350.0,
"learning_rate": 8.428571428571429e-06,
"loss": 0.0,
"step": 1050
},
{
"epoch": 358.33,
"learning_rate": 8.357142857142858e-06,
"loss": 0.0,
"step": 1075
},
{
"epoch": 366.67,
"learning_rate": 8.285714285714287e-06,
"loss": 0.0,
"step": 1100
},
{
"epoch": 375.0,
"learning_rate": 8.214285714285714e-06,
"loss": 0.0,
"step": 1125
},
{
"epoch": 383.33,
"learning_rate": 8.142857142857143e-06,
"loss": 0.0,
"step": 1150
},
{
"epoch": 391.67,
"learning_rate": 8.071428571428572e-06,
"loss": 0.0,
"step": 1175
},
{
"epoch": 400.0,
"learning_rate": 8.000000000000001e-06,
"loss": 0.0,
"step": 1200
},
{
"epoch": 408.33,
"learning_rate": 7.928571428571429e-06,
"loss": 0.0,
"step": 1225
},
{
"epoch": 416.67,
"learning_rate": 7.857142857142858e-06,
"loss": 0.0,
"step": 1250
},
{
"epoch": 425.0,
"learning_rate": 7.785714285714287e-06,
"loss": 0.0,
"step": 1275
},
{
"epoch": 433.33,
"learning_rate": 7.714285714285716e-06,
"loss": 0.0,
"step": 1300
},
{
"epoch": 441.67,
"learning_rate": 7.642857142857143e-06,
"loss": 0.0,
"step": 1325
},
{
"epoch": 450.0,
"learning_rate": 7.571428571428572e-06,
"loss": 0.0,
"step": 1350
},
{
"epoch": 458.33,
"learning_rate": 7.500000000000001e-06,
"loss": 0.0,
"step": 1375
},
{
"epoch": 466.67,
"learning_rate": 7.428571428571429e-06,
"loss": 0.0,
"step": 1400
},
{
"epoch": 475.0,
"learning_rate": 7.357142857142858e-06,
"loss": 0.0,
"step": 1425
},
{
"epoch": 483.33,
"learning_rate": 7.285714285714286e-06,
"loss": 0.0,
"step": 1450
},
{
"epoch": 491.67,
"learning_rate": 7.2142857142857145e-06,
"loss": 0.0,
"step": 1475
},
{
"epoch": 500.0,
"learning_rate": 7.1428571428571436e-06,
"loss": 0.0,
"step": 1500
},
{
"epoch": 508.33,
"learning_rate": 7.0714285714285726e-06,
"loss": 0.0,
"step": 1525
},
{
"epoch": 516.67,
"learning_rate": 7e-06,
"loss": 0.0,
"step": 1550
},
{
"epoch": 525.0,
"learning_rate": 6.928571428571429e-06,
"loss": 0.0,
"step": 1575
},
{
"epoch": 533.33,
"learning_rate": 6.857142857142858e-06,
"loss": 0.0,
"step": 1600
},
{
"epoch": 541.67,
"learning_rate": 6.785714285714287e-06,
"loss": 0.0,
"step": 1625
},
{
"epoch": 550.0,
"learning_rate": 6.714285714285714e-06,
"loss": 0.0,
"step": 1650
},
{
"epoch": 558.33,
"learning_rate": 6.642857142857143e-06,
"loss": 0.0,
"step": 1675
},
{
"epoch": 566.67,
"learning_rate": 6.571428571428572e-06,
"loss": 0.0,
"step": 1700
},
{
"epoch": 575.0,
"learning_rate": 6.5000000000000004e-06,
"loss": 0.0,
"step": 1725
},
{
"epoch": 583.33,
"learning_rate": 6.4285714285714295e-06,
"loss": 0.0,
"step": 1750
},
{
"epoch": 591.67,
"learning_rate": 6.357142857142858e-06,
"loss": 0.0,
"step": 1775
},
{
"epoch": 600.0,
"learning_rate": 6.285714285714286e-06,
"loss": 0.0,
"step": 1800
},
{
"epoch": 608.33,
"learning_rate": 6.214285714285715e-06,
"loss": 0.0,
"step": 1825
},
{
"epoch": 616.67,
"learning_rate": 6.142857142857144e-06,
"loss": 0.0,
"step": 1850
},
{
"epoch": 625.0,
"learning_rate": 6.071428571428571e-06,
"loss": 0.0,
"step": 1875
},
{
"epoch": 633.33,
"learning_rate": 6e-06,
"loss": 0.0,
"step": 1900
},
{
"epoch": 641.67,
"learning_rate": 5.928571428571429e-06,
"loss": 0.0,
"step": 1925
},
{
"epoch": 650.0,
"learning_rate": 5.857142857142858e-06,
"loss": 0.0,
"step": 1950
},
{
"epoch": 658.33,
"learning_rate": 5.785714285714286e-06,
"loss": 0.0,
"step": 1975
},
{
"epoch": 666.67,
"learning_rate": 5.7142857142857145e-06,
"loss": 0.0,
"step": 2000
},
{
"epoch": 666.67,
"eval_loss": 0.3412761986255646,
"eval_runtime": 12.3441,
"eval_samples_per_second": 1.377,
"eval_steps_per_second": 0.243,
"eval_wer": 111.76470588235294,
"step": 2000
},
{
"epoch": 675.0,
"learning_rate": 5.6428571428571435e-06,
"loss": 0.0,
"step": 2025
},
{
"epoch": 683.33,
"learning_rate": 5.571428571428572e-06,
"loss": 0.0,
"step": 2050
},
{
"epoch": 691.67,
"learning_rate": 5.500000000000001e-06,
"loss": 0.0,
"step": 2075
},
{
"epoch": 700.0,
"learning_rate": 5.428571428571429e-06,
"loss": 0.0,
"step": 2100
},
{
"epoch": 708.33,
"learning_rate": 5.357142857142857e-06,
"loss": 0.0,
"step": 2125
},
{
"epoch": 716.67,
"learning_rate": 5.285714285714286e-06,
"loss": 0.0,
"step": 2150
},
{
"epoch": 725.0,
"learning_rate": 5.214285714285715e-06,
"loss": 0.0,
"step": 2175
},
{
"epoch": 733.33,
"learning_rate": 5.142857142857142e-06,
"loss": 0.0,
"step": 2200
},
{
"epoch": 741.67,
"learning_rate": 5.071428571428571e-06,
"loss": 0.0,
"step": 2225
},
{
"epoch": 750.0,
"learning_rate": 5e-06,
"loss": 0.0,
"step": 2250
},
{
"epoch": 758.33,
"learning_rate": 4.928571428571429e-06,
"loss": 0.0,
"step": 2275
},
{
"epoch": 766.67,
"learning_rate": 4.857142857142858e-06,
"loss": 0.0,
"step": 2300
},
{
"epoch": 775.0,
"learning_rate": 4.785714285714287e-06,
"loss": 0.0,
"step": 2325
},
{
"epoch": 783.33,
"learning_rate": 4.714285714285715e-06,
"loss": 0.0,
"step": 2350
},
{
"epoch": 791.67,
"learning_rate": 4.642857142857144e-06,
"loss": 0.0,
"step": 2375
},
{
"epoch": 800.0,
"learning_rate": 4.571428571428572e-06,
"loss": 0.0,
"step": 2400
},
{
"epoch": 808.33,
"learning_rate": 4.5e-06,
"loss": 0.0,
"step": 2425
},
{
"epoch": 816.67,
"learning_rate": 4.428571428571429e-06,
"loss": 0.0,
"step": 2450
},
{
"epoch": 825.0,
"learning_rate": 4.357142857142857e-06,
"loss": 0.0,
"step": 2475
},
{
"epoch": 833.33,
"learning_rate": 4.2857142857142855e-06,
"loss": 0.0,
"step": 2500
},
{
"epoch": 841.67,
"learning_rate": 4.2142857142857145e-06,
"loss": 0.0,
"step": 2525
},
{
"epoch": 850.0,
"learning_rate": 4.1428571428571435e-06,
"loss": 0.0,
"step": 2550
},
{
"epoch": 858.33,
"learning_rate": 4.071428571428572e-06,
"loss": 0.0,
"step": 2575
},
{
"epoch": 866.67,
"learning_rate": 4.000000000000001e-06,
"loss": 0.0,
"step": 2600
},
{
"epoch": 875.0,
"learning_rate": 3.928571428571429e-06,
"loss": 0.0,
"step": 2625
},
{
"epoch": 883.33,
"learning_rate": 3.857142857142858e-06,
"loss": 0.0,
"step": 2650
},
{
"epoch": 891.67,
"learning_rate": 3.785714285714286e-06,
"loss": 0.0,
"step": 2675
},
{
"epoch": 900.0,
"learning_rate": 3.7142857142857146e-06,
"loss": 0.0,
"step": 2700
},
{
"epoch": 908.33,
"learning_rate": 3.642857142857143e-06,
"loss": 0.0,
"step": 2725
},
{
"epoch": 916.67,
"learning_rate": 3.5714285714285718e-06,
"loss": 0.0,
"step": 2750
},
{
"epoch": 925.0,
"learning_rate": 3.5e-06,
"loss": 0.0,
"step": 2775
},
{
"epoch": 933.33,
"learning_rate": 3.428571428571429e-06,
"loss": 0.0,
"step": 2800
},
{
"epoch": 941.67,
"learning_rate": 3.357142857142857e-06,
"loss": 0.0,
"step": 2825
},
{
"epoch": 950.0,
"learning_rate": 3.285714285714286e-06,
"loss": 0.0,
"step": 2850
},
{
"epoch": 958.33,
"learning_rate": 3.2142857142857147e-06,
"loss": 0.0,
"step": 2875
},
{
"epoch": 966.67,
"learning_rate": 3.142857142857143e-06,
"loss": 0.0,
"step": 2900
},
{
"epoch": 975.0,
"learning_rate": 3.071428571428572e-06,
"loss": 0.0,
"step": 2925
},
{
"epoch": 983.33,
"learning_rate": 3e-06,
"loss": 0.0,
"step": 2950
},
{
"epoch": 991.67,
"learning_rate": 2.928571428571429e-06,
"loss": 0.0,
"step": 2975
},
{
"epoch": 1000.0,
"learning_rate": 2.8571428571428573e-06,
"loss": 0.0,
"step": 3000
},
{
"epoch": 1000.0,
"eval_loss": 0.36998674273490906,
"eval_runtime": 12.5254,
"eval_samples_per_second": 1.357,
"eval_steps_per_second": 0.24,
"eval_wer": 117.64705882352942,
"step": 3000
},
{
"epoch": 1008.33,
"learning_rate": 2.785714285714286e-06,
"loss": 0.0,
"step": 3025
},
{
"epoch": 1016.67,
"learning_rate": 2.7142857142857144e-06,
"loss": 0.0,
"step": 3050
},
{
"epoch": 1025.0,
"learning_rate": 2.642857142857143e-06,
"loss": 0.0,
"step": 3075
},
{
"epoch": 1033.33,
"learning_rate": 2.571428571428571e-06,
"loss": 0.0,
"step": 3100
},
{
"epoch": 1041.67,
"learning_rate": 2.5e-06,
"loss": 0.0,
"step": 3125
},
{
"epoch": 1050.0,
"learning_rate": 2.428571428571429e-06,
"loss": 0.0,
"step": 3150
},
{
"epoch": 1058.33,
"learning_rate": 2.3571428571428574e-06,
"loss": 0.0,
"step": 3175
},
{
"epoch": 1066.67,
"learning_rate": 2.285714285714286e-06,
"loss": 0.0,
"step": 3200
},
{
"epoch": 1075.0,
"learning_rate": 2.2142857142857146e-06,
"loss": 0.0,
"step": 3225
},
{
"epoch": 1083.33,
"learning_rate": 2.1428571428571427e-06,
"loss": 0.0,
"step": 3250
},
{
"epoch": 1091.67,
"learning_rate": 2.0714285714285717e-06,
"loss": 0.0,
"step": 3275
},
{
"epoch": 1100.0,
"learning_rate": 2.0000000000000003e-06,
"loss": 0.0,
"step": 3300
},
{
"epoch": 1108.33,
"learning_rate": 1.928571428571429e-06,
"loss": 0.0,
"step": 3325
},
{
"epoch": 1116.67,
"learning_rate": 1.8571428571428573e-06,
"loss": 0.0,
"step": 3350
},
{
"epoch": 1125.0,
"learning_rate": 1.7857142857142859e-06,
"loss": 0.0,
"step": 3375
},
{
"epoch": 1133.33,
"learning_rate": 1.7142857142857145e-06,
"loss": 0.0,
"step": 3400
},
{
"epoch": 1141.67,
"learning_rate": 1.642857142857143e-06,
"loss": 0.0,
"step": 3425
},
{
"epoch": 1150.0,
"learning_rate": 1.5714285714285714e-06,
"loss": 0.0,
"step": 3450
},
{
"epoch": 1158.33,
"learning_rate": 1.5e-06,
"loss": 0.0,
"step": 3475
},
{
"epoch": 1166.67,
"learning_rate": 1.4285714285714286e-06,
"loss": 0.0,
"step": 3500
},
{
"epoch": 1175.0,
"learning_rate": 1.3571428571428572e-06,
"loss": 0.0,
"step": 3525
},
{
"epoch": 1183.33,
"learning_rate": 1.2857142857142856e-06,
"loss": 0.0,
"step": 3550
},
{
"epoch": 1191.67,
"learning_rate": 1.2142857142857144e-06,
"loss": 0.0,
"step": 3575
},
{
"epoch": 1200.0,
"learning_rate": 1.142857142857143e-06,
"loss": 0.0,
"step": 3600
},
{
"epoch": 1208.33,
"learning_rate": 1.0714285714285714e-06,
"loss": 0.0,
"step": 3625
},
{
"epoch": 1216.67,
"learning_rate": 1.0000000000000002e-06,
"loss": 0.0,
"step": 3650
},
{
"epoch": 1225.0,
"learning_rate": 9.285714285714287e-07,
"loss": 0.0,
"step": 3675
},
{
"epoch": 1233.33,
"learning_rate": 8.571428571428572e-07,
"loss": 0.0,
"step": 3700
},
{
"epoch": 1241.67,
"learning_rate": 7.857142857142857e-07,
"loss": 0.0,
"step": 3725
},
{
"epoch": 1250.0,
"learning_rate": 7.142857142857143e-07,
"loss": 0.0,
"step": 3750
},
{
"epoch": 1258.33,
"learning_rate": 6.428571428571428e-07,
"loss": 0.0,
"step": 3775
},
{
"epoch": 1266.67,
"learning_rate": 5.714285714285715e-07,
"loss": 0.0,
"step": 3800
},
{
"epoch": 1275.0,
"learning_rate": 5.000000000000001e-07,
"loss": 0.0,
"step": 3825
},
{
"epoch": 1283.33,
"learning_rate": 4.285714285714286e-07,
"loss": 0.0,
"step": 3850
},
{
"epoch": 1291.67,
"learning_rate": 3.5714285714285716e-07,
"loss": 0.0,
"step": 3875
},
{
"epoch": 1300.0,
"learning_rate": 2.8571428571428575e-07,
"loss": 0.0,
"step": 3900
},
{
"epoch": 1308.33,
"learning_rate": 2.142857142857143e-07,
"loss": 0.0,
"step": 3925
},
{
"epoch": 1316.67,
"learning_rate": 1.4285714285714287e-07,
"loss": 0.0,
"step": 3950
},
{
"epoch": 1325.0,
"learning_rate": 7.142857142857144e-08,
"loss": 0.0,
"step": 3975
},
{
"epoch": 1333.33,
"learning_rate": 0.0,
"loss": 0.0,
"step": 4000
},
{
"epoch": 1333.33,
"eval_loss": 0.3857485353946686,
"eval_runtime": 12.6654,
"eval_samples_per_second": 1.342,
"eval_steps_per_second": 0.237,
"eval_wer": 117.64705882352942,
"step": 4000
}
],
"max_steps": 4000,
"num_train_epochs": 1334,
"total_flos": 8.46536413298688e+18,
"trial_name": null,
"trial_params": null
}