mskov's picture
Upload 9 files
4fed837
raw
history blame
20.1 kB
{
"best_metric": 87.8,
"best_model_checkpoint": "mskov/whisper_esc50/checkpoint-1000",
"epoch": 63.492063492063494,
"global_step": 4000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.4,
"learning_rate": 5.000000000000001e-07,
"loss": 6.7514,
"step": 25
},
{
"epoch": 0.79,
"learning_rate": 1.0000000000000002e-06,
"loss": 5.6562,
"step": 50
},
{
"epoch": 1.19,
"learning_rate": 1.5e-06,
"loss": 4.0209,
"step": 75
},
{
"epoch": 1.59,
"learning_rate": 2.0000000000000003e-06,
"loss": 1.5856,
"step": 100
},
{
"epoch": 1.98,
"learning_rate": 2.5e-06,
"loss": 0.5934,
"step": 125
},
{
"epoch": 2.38,
"learning_rate": 3e-06,
"loss": 0.2997,
"step": 150
},
{
"epoch": 2.78,
"learning_rate": 3.5e-06,
"loss": 0.2029,
"step": 175
},
{
"epoch": 3.17,
"learning_rate": 4.000000000000001e-06,
"loss": 0.1297,
"step": 200
},
{
"epoch": 3.57,
"learning_rate": 4.5e-06,
"loss": 0.0937,
"step": 225
},
{
"epoch": 3.97,
"learning_rate": 5e-06,
"loss": 0.0809,
"step": 250
},
{
"epoch": 4.37,
"learning_rate": 5.500000000000001e-06,
"loss": 0.0521,
"step": 275
},
{
"epoch": 4.76,
"learning_rate": 6e-06,
"loss": 0.0477,
"step": 300
},
{
"epoch": 5.16,
"learning_rate": 6.5000000000000004e-06,
"loss": 0.0271,
"step": 325
},
{
"epoch": 5.56,
"learning_rate": 7e-06,
"loss": 0.0308,
"step": 350
},
{
"epoch": 5.95,
"learning_rate": 7.500000000000001e-06,
"loss": 0.0385,
"step": 375
},
{
"epoch": 6.35,
"learning_rate": 8.000000000000001e-06,
"loss": 0.0473,
"step": 400
},
{
"epoch": 6.75,
"learning_rate": 8.5e-06,
"loss": 0.0313,
"step": 425
},
{
"epoch": 7.14,
"learning_rate": 9e-06,
"loss": 0.0246,
"step": 450
},
{
"epoch": 7.54,
"learning_rate": 9.5e-06,
"loss": 0.0227,
"step": 475
},
{
"epoch": 7.94,
"learning_rate": 1e-05,
"loss": 0.0236,
"step": 500
},
{
"epoch": 8.33,
"learning_rate": 9.92857142857143e-06,
"loss": 0.014,
"step": 525
},
{
"epoch": 8.73,
"learning_rate": 9.857142857142859e-06,
"loss": 0.0144,
"step": 550
},
{
"epoch": 9.13,
"learning_rate": 9.785714285714286e-06,
"loss": 0.0086,
"step": 575
},
{
"epoch": 9.52,
"learning_rate": 9.714285714285715e-06,
"loss": 0.0119,
"step": 600
},
{
"epoch": 9.92,
"learning_rate": 9.642857142857144e-06,
"loss": 0.0044,
"step": 625
},
{
"epoch": 10.32,
"learning_rate": 9.571428571428573e-06,
"loss": 0.0052,
"step": 650
},
{
"epoch": 10.71,
"learning_rate": 9.5e-06,
"loss": 0.0017,
"step": 675
},
{
"epoch": 11.11,
"learning_rate": 9.42857142857143e-06,
"loss": 0.0034,
"step": 700
},
{
"epoch": 11.51,
"learning_rate": 9.357142857142859e-06,
"loss": 0.0055,
"step": 725
},
{
"epoch": 11.9,
"learning_rate": 9.285714285714288e-06,
"loss": 0.0021,
"step": 750
},
{
"epoch": 12.3,
"learning_rate": 9.214285714285715e-06,
"loss": 0.0028,
"step": 775
},
{
"epoch": 12.7,
"learning_rate": 9.142857142857144e-06,
"loss": 0.0007,
"step": 800
},
{
"epoch": 13.1,
"learning_rate": 9.071428571428573e-06,
"loss": 0.0027,
"step": 825
},
{
"epoch": 13.49,
"learning_rate": 9e-06,
"loss": 0.0003,
"step": 850
},
{
"epoch": 13.89,
"learning_rate": 8.92857142857143e-06,
"loss": 0.0001,
"step": 875
},
{
"epoch": 14.29,
"learning_rate": 8.857142857142858e-06,
"loss": 0.0001,
"step": 900
},
{
"epoch": 14.68,
"learning_rate": 8.785714285714286e-06,
"loss": 0.0001,
"step": 925
},
{
"epoch": 15.08,
"learning_rate": 8.714285714285715e-06,
"loss": 0.0001,
"step": 950
},
{
"epoch": 15.48,
"learning_rate": 8.642857142857144e-06,
"loss": 0.0,
"step": 975
},
{
"epoch": 15.87,
"learning_rate": 8.571428571428571e-06,
"loss": 0.0,
"step": 1000
},
{
"epoch": 15.87,
"eval_loss": 0.1191895380616188,
"eval_runtime": 460.7316,
"eval_samples_per_second": 2.17,
"eval_steps_per_second": 0.271,
"eval_wer": 87.8,
"step": 1000
},
{
"epoch": 16.27,
"learning_rate": 8.5e-06,
"loss": 0.0,
"step": 1025
},
{
"epoch": 16.67,
"learning_rate": 8.428571428571429e-06,
"loss": 0.0,
"step": 1050
},
{
"epoch": 17.06,
"learning_rate": 8.357142857142858e-06,
"loss": 0.0,
"step": 1075
},
{
"epoch": 17.46,
"learning_rate": 8.285714285714287e-06,
"loss": 0.0,
"step": 1100
},
{
"epoch": 17.86,
"learning_rate": 8.214285714285714e-06,
"loss": 0.0,
"step": 1125
},
{
"epoch": 18.25,
"learning_rate": 8.142857142857143e-06,
"loss": 0.0,
"step": 1150
},
{
"epoch": 18.65,
"learning_rate": 8.071428571428572e-06,
"loss": 0.0,
"step": 1175
},
{
"epoch": 19.05,
"learning_rate": 8.000000000000001e-06,
"loss": 0.0,
"step": 1200
},
{
"epoch": 19.44,
"learning_rate": 7.928571428571429e-06,
"loss": 0.0,
"step": 1225
},
{
"epoch": 19.84,
"learning_rate": 7.857142857142858e-06,
"loss": 0.0,
"step": 1250
},
{
"epoch": 20.24,
"learning_rate": 7.785714285714287e-06,
"loss": 0.0,
"step": 1275
},
{
"epoch": 20.63,
"learning_rate": 7.714285714285716e-06,
"loss": 0.0,
"step": 1300
},
{
"epoch": 21.03,
"learning_rate": 7.642857142857143e-06,
"loss": 0.0,
"step": 1325
},
{
"epoch": 21.43,
"learning_rate": 7.571428571428572e-06,
"loss": 0.0,
"step": 1350
},
{
"epoch": 21.83,
"learning_rate": 7.500000000000001e-06,
"loss": 0.0,
"step": 1375
},
{
"epoch": 22.22,
"learning_rate": 7.428571428571429e-06,
"loss": 0.0,
"step": 1400
},
{
"epoch": 22.62,
"learning_rate": 7.357142857142858e-06,
"loss": 0.0,
"step": 1425
},
{
"epoch": 23.02,
"learning_rate": 7.285714285714286e-06,
"loss": 0.0,
"step": 1450
},
{
"epoch": 23.41,
"learning_rate": 7.2142857142857145e-06,
"loss": 0.0,
"step": 1475
},
{
"epoch": 23.81,
"learning_rate": 7.1428571428571436e-06,
"loss": 0.0,
"step": 1500
},
{
"epoch": 24.21,
"learning_rate": 7.0714285714285726e-06,
"loss": 0.0,
"step": 1525
},
{
"epoch": 24.6,
"learning_rate": 7e-06,
"loss": 0.0,
"step": 1550
},
{
"epoch": 25.0,
"learning_rate": 6.928571428571429e-06,
"loss": 0.0,
"step": 1575
},
{
"epoch": 25.4,
"learning_rate": 6.857142857142858e-06,
"loss": 0.0,
"step": 1600
},
{
"epoch": 25.79,
"learning_rate": 6.785714285714287e-06,
"loss": 0.0,
"step": 1625
},
{
"epoch": 26.19,
"learning_rate": 6.714285714285714e-06,
"loss": 0.0,
"step": 1650
},
{
"epoch": 26.59,
"learning_rate": 6.642857142857143e-06,
"loss": 0.0,
"step": 1675
},
{
"epoch": 26.98,
"learning_rate": 6.571428571428572e-06,
"loss": 0.0,
"step": 1700
},
{
"epoch": 27.38,
"learning_rate": 6.5000000000000004e-06,
"loss": 0.0,
"step": 1725
},
{
"epoch": 27.78,
"learning_rate": 6.4285714285714295e-06,
"loss": 0.0,
"step": 1750
},
{
"epoch": 28.17,
"learning_rate": 6.357142857142858e-06,
"loss": 0.0,
"step": 1775
},
{
"epoch": 28.57,
"learning_rate": 6.285714285714286e-06,
"loss": 0.0,
"step": 1800
},
{
"epoch": 28.97,
"learning_rate": 6.214285714285715e-06,
"loss": 0.0,
"step": 1825
},
{
"epoch": 29.37,
"learning_rate": 6.142857142857144e-06,
"loss": 0.0,
"step": 1850
},
{
"epoch": 29.76,
"learning_rate": 6.071428571428571e-06,
"loss": 0.0,
"step": 1875
},
{
"epoch": 30.16,
"learning_rate": 6e-06,
"loss": 0.0,
"step": 1900
},
{
"epoch": 30.56,
"learning_rate": 5.928571428571429e-06,
"loss": 0.0,
"step": 1925
},
{
"epoch": 30.95,
"learning_rate": 5.857142857142858e-06,
"loss": 0.0,
"step": 1950
},
{
"epoch": 31.35,
"learning_rate": 5.785714285714286e-06,
"loss": 0.0,
"step": 1975
},
{
"epoch": 31.75,
"learning_rate": 5.7142857142857145e-06,
"loss": 0.0,
"step": 2000
},
{
"epoch": 31.75,
"eval_loss": 0.12031258642673492,
"eval_runtime": 468.2579,
"eval_samples_per_second": 2.136,
"eval_steps_per_second": 0.267,
"eval_wer": 89.9,
"step": 2000
},
{
"epoch": 32.14,
"learning_rate": 5.6428571428571435e-06,
"loss": 0.0,
"step": 2025
},
{
"epoch": 32.54,
"learning_rate": 5.571428571428572e-06,
"loss": 0.0,
"step": 2050
},
{
"epoch": 32.94,
"learning_rate": 5.500000000000001e-06,
"loss": 0.0,
"step": 2075
},
{
"epoch": 33.33,
"learning_rate": 5.428571428571429e-06,
"loss": 0.0,
"step": 2100
},
{
"epoch": 33.73,
"learning_rate": 5.357142857142857e-06,
"loss": 0.0,
"step": 2125
},
{
"epoch": 34.13,
"learning_rate": 5.285714285714286e-06,
"loss": 0.0,
"step": 2150
},
{
"epoch": 34.52,
"learning_rate": 5.214285714285715e-06,
"loss": 0.0,
"step": 2175
},
{
"epoch": 34.92,
"learning_rate": 5.142857142857142e-06,
"loss": 0.0,
"step": 2200
},
{
"epoch": 35.32,
"learning_rate": 5.071428571428571e-06,
"loss": 0.0,
"step": 2225
},
{
"epoch": 35.71,
"learning_rate": 5e-06,
"loss": 0.0,
"step": 2250
},
{
"epoch": 36.11,
"learning_rate": 4.928571428571429e-06,
"loss": 0.0,
"step": 2275
},
{
"epoch": 36.51,
"learning_rate": 4.857142857142858e-06,
"loss": 0.0,
"step": 2300
},
{
"epoch": 36.9,
"learning_rate": 4.785714285714287e-06,
"loss": 0.0,
"step": 2325
},
{
"epoch": 37.3,
"learning_rate": 4.714285714285715e-06,
"loss": 0.0,
"step": 2350
},
{
"epoch": 37.7,
"learning_rate": 4.642857142857144e-06,
"loss": 0.0,
"step": 2375
},
{
"epoch": 38.1,
"learning_rate": 4.571428571428572e-06,
"loss": 0.0,
"step": 2400
},
{
"epoch": 38.49,
"learning_rate": 4.5e-06,
"loss": 0.0,
"step": 2425
},
{
"epoch": 38.89,
"learning_rate": 4.428571428571429e-06,
"loss": 0.0,
"step": 2450
},
{
"epoch": 39.29,
"learning_rate": 4.357142857142857e-06,
"loss": 0.0,
"step": 2475
},
{
"epoch": 39.68,
"learning_rate": 4.2857142857142855e-06,
"loss": 0.0,
"step": 2500
},
{
"epoch": 40.08,
"learning_rate": 4.2142857142857145e-06,
"loss": 0.0,
"step": 2525
},
{
"epoch": 40.48,
"learning_rate": 4.1428571428571435e-06,
"loss": 0.0,
"step": 2550
},
{
"epoch": 40.87,
"learning_rate": 4.071428571428572e-06,
"loss": 0.0,
"step": 2575
},
{
"epoch": 41.27,
"learning_rate": 4.000000000000001e-06,
"loss": 0.0,
"step": 2600
},
{
"epoch": 41.67,
"learning_rate": 3.928571428571429e-06,
"loss": 0.0,
"step": 2625
},
{
"epoch": 42.06,
"learning_rate": 3.857142857142858e-06,
"loss": 0.0,
"step": 2650
},
{
"epoch": 42.46,
"learning_rate": 3.785714285714286e-06,
"loss": 0.0,
"step": 2675
},
{
"epoch": 42.86,
"learning_rate": 3.7142857142857146e-06,
"loss": 0.0,
"step": 2700
},
{
"epoch": 43.25,
"learning_rate": 3.642857142857143e-06,
"loss": 0.0,
"step": 2725
},
{
"epoch": 43.65,
"learning_rate": 3.5714285714285718e-06,
"loss": 0.0,
"step": 2750
},
{
"epoch": 44.05,
"learning_rate": 3.5e-06,
"loss": 0.0,
"step": 2775
},
{
"epoch": 44.44,
"learning_rate": 3.428571428571429e-06,
"loss": 0.0,
"step": 2800
},
{
"epoch": 44.84,
"learning_rate": 3.357142857142857e-06,
"loss": 0.0,
"step": 2825
},
{
"epoch": 45.24,
"learning_rate": 3.285714285714286e-06,
"loss": 0.0,
"step": 2850
},
{
"epoch": 45.63,
"learning_rate": 3.2142857142857147e-06,
"loss": 0.0,
"step": 2875
},
{
"epoch": 46.03,
"learning_rate": 3.142857142857143e-06,
"loss": 0.0,
"step": 2900
},
{
"epoch": 46.43,
"learning_rate": 3.071428571428572e-06,
"loss": 0.0,
"step": 2925
},
{
"epoch": 46.83,
"learning_rate": 3e-06,
"loss": 0.0,
"step": 2950
},
{
"epoch": 47.22,
"learning_rate": 2.928571428571429e-06,
"loss": 0.0,
"step": 2975
},
{
"epoch": 47.62,
"learning_rate": 2.8571428571428573e-06,
"loss": 0.0,
"step": 3000
},
{
"epoch": 47.62,
"eval_loss": 0.12188500165939331,
"eval_runtime": 480.3764,
"eval_samples_per_second": 2.082,
"eval_steps_per_second": 0.26,
"eval_wer": 111.7,
"step": 3000
},
{
"epoch": 48.02,
"learning_rate": 2.785714285714286e-06,
"loss": 0.0,
"step": 3025
},
{
"epoch": 48.41,
"learning_rate": 2.7142857142857144e-06,
"loss": 0.0,
"step": 3050
},
{
"epoch": 48.81,
"learning_rate": 2.642857142857143e-06,
"loss": 0.0,
"step": 3075
},
{
"epoch": 49.21,
"learning_rate": 2.571428571428571e-06,
"loss": 0.0,
"step": 3100
},
{
"epoch": 49.6,
"learning_rate": 2.5e-06,
"loss": 0.0,
"step": 3125
},
{
"epoch": 50.0,
"learning_rate": 2.428571428571429e-06,
"loss": 0.0,
"step": 3150
},
{
"epoch": 50.4,
"learning_rate": 2.3571428571428574e-06,
"loss": 0.0,
"step": 3175
},
{
"epoch": 50.79,
"learning_rate": 2.285714285714286e-06,
"loss": 0.0,
"step": 3200
},
{
"epoch": 51.19,
"learning_rate": 2.2142857142857146e-06,
"loss": 0.0,
"step": 3225
},
{
"epoch": 51.59,
"learning_rate": 2.1428571428571427e-06,
"loss": 0.0,
"step": 3250
},
{
"epoch": 51.98,
"learning_rate": 2.0714285714285717e-06,
"loss": 0.0,
"step": 3275
},
{
"epoch": 52.38,
"learning_rate": 2.0000000000000003e-06,
"loss": 0.0,
"step": 3300
},
{
"epoch": 52.78,
"learning_rate": 1.928571428571429e-06,
"loss": 0.0,
"step": 3325
},
{
"epoch": 53.17,
"learning_rate": 1.8571428571428573e-06,
"loss": 0.0,
"step": 3350
},
{
"epoch": 53.57,
"learning_rate": 1.7857142857142859e-06,
"loss": 0.0,
"step": 3375
},
{
"epoch": 53.97,
"learning_rate": 1.7142857142857145e-06,
"loss": 0.0,
"step": 3400
},
{
"epoch": 54.37,
"learning_rate": 1.642857142857143e-06,
"loss": 0.0,
"step": 3425
},
{
"epoch": 54.76,
"learning_rate": 1.5714285714285714e-06,
"loss": 0.0,
"step": 3450
},
{
"epoch": 55.16,
"learning_rate": 1.5e-06,
"loss": 0.0,
"step": 3475
},
{
"epoch": 55.56,
"learning_rate": 1.4285714285714286e-06,
"loss": 0.0,
"step": 3500
},
{
"epoch": 55.95,
"learning_rate": 1.3571428571428572e-06,
"loss": 0.0,
"step": 3525
},
{
"epoch": 56.35,
"learning_rate": 1.2857142857142856e-06,
"loss": 0.0,
"step": 3550
},
{
"epoch": 56.75,
"learning_rate": 1.2142857142857144e-06,
"loss": 0.0,
"step": 3575
},
{
"epoch": 57.14,
"learning_rate": 1.142857142857143e-06,
"loss": 0.0,
"step": 3600
},
{
"epoch": 57.54,
"learning_rate": 1.0714285714285714e-06,
"loss": 0.0,
"step": 3625
},
{
"epoch": 57.94,
"learning_rate": 1.0000000000000002e-06,
"loss": 0.0,
"step": 3650
},
{
"epoch": 58.33,
"learning_rate": 9.285714285714287e-07,
"loss": 0.0,
"step": 3675
},
{
"epoch": 58.73,
"learning_rate": 8.571428571428572e-07,
"loss": 0.0,
"step": 3700
},
{
"epoch": 59.13,
"learning_rate": 7.857142857142857e-07,
"loss": 0.0,
"step": 3725
},
{
"epoch": 59.52,
"learning_rate": 7.142857142857143e-07,
"loss": 0.0,
"step": 3750
},
{
"epoch": 59.92,
"learning_rate": 6.428571428571428e-07,
"loss": 0.0,
"step": 3775
},
{
"epoch": 60.32,
"learning_rate": 5.714285714285715e-07,
"loss": 0.0,
"step": 3800
},
{
"epoch": 60.71,
"learning_rate": 5.000000000000001e-07,
"loss": 0.0,
"step": 3825
},
{
"epoch": 61.11,
"learning_rate": 4.285714285714286e-07,
"loss": 0.0,
"step": 3850
},
{
"epoch": 61.51,
"learning_rate": 3.5714285714285716e-07,
"loss": 0.0,
"step": 3875
},
{
"epoch": 61.9,
"learning_rate": 2.8571428571428575e-07,
"loss": 0.0,
"step": 3900
},
{
"epoch": 62.3,
"learning_rate": 2.142857142857143e-07,
"loss": 0.0,
"step": 3925
},
{
"epoch": 62.7,
"learning_rate": 1.4285714285714287e-07,
"loss": 0.0,
"step": 3950
},
{
"epoch": 63.1,
"learning_rate": 7.142857142857144e-08,
"loss": 0.0,
"step": 3975
},
{
"epoch": 63.49,
"learning_rate": 0.0,
"loss": 0.0,
"step": 4000
},
{
"epoch": 63.49,
"eval_loss": 0.12252455949783325,
"eval_runtime": 494.6401,
"eval_samples_per_second": 2.022,
"eval_steps_per_second": 0.253,
"eval_wer": 111.7,
"step": 4000
}
],
"max_steps": 4000,
"num_train_epochs": 64,
"total_flos": 1.832401857871872e+19,
"trial_name": null,
"trial_params": null
}