whisper-tiny-it-6 / checkpoint-3000 /trainer_state.json
GIanlucaRub's picture
Copied from Google Drive
5ab1cc8
raw
history blame
15.3 kB
{
"best_metric": 47.76760953965613,
"best_model_checkpoint": "./whisper-tiny-it-6/checkpoint-3000",
"epoch": 2.8625954198473282,
"global_step": 3000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.02,
"learning_rate": 3.8e-06,
"loss": 10.1581,
"step": 25
},
{
"epoch": 0.05,
"learning_rate": 8.599999999999999e-06,
"loss": 8.1216,
"step": 50
},
{
"epoch": 0.07,
"learning_rate": 1.3600000000000002e-05,
"loss": 6.9275,
"step": 75
},
{
"epoch": 0.1,
"learning_rate": 1.86e-05,
"loss": 5.6547,
"step": 100
},
{
"epoch": 0.12,
"learning_rate": 2.36e-05,
"loss": 5.2734,
"step": 125
},
{
"epoch": 0.14,
"learning_rate": 2.86e-05,
"loss": 5.5818,
"step": 150
},
{
"epoch": 0.17,
"learning_rate": 3.3600000000000004e-05,
"loss": 4.9292,
"step": 175
},
{
"epoch": 0.19,
"learning_rate": 3.86e-05,
"loss": 4.8341,
"step": 200
},
{
"epoch": 0.21,
"learning_rate": 4.36e-05,
"loss": 4.1192,
"step": 225
},
{
"epoch": 0.24,
"learning_rate": 4.86e-05,
"loss": 4.2109,
"step": 250
},
{
"epoch": 0.26,
"learning_rate": 5.360000000000001e-05,
"loss": 3.9181,
"step": 275
},
{
"epoch": 0.29,
"learning_rate": 5.86e-05,
"loss": 3.1145,
"step": 300
},
{
"epoch": 0.31,
"learning_rate": 6.36e-05,
"loss": 3.4049,
"step": 325
},
{
"epoch": 0.33,
"learning_rate": 6.860000000000001e-05,
"loss": 3.458,
"step": 350
},
{
"epoch": 0.36,
"learning_rate": 7.36e-05,
"loss": 3.6412,
"step": 375
},
{
"epoch": 0.38,
"learning_rate": 7.860000000000001e-05,
"loss": 3.0229,
"step": 400
},
{
"epoch": 0.41,
"learning_rate": 8.36e-05,
"loss": 2.9048,
"step": 425
},
{
"epoch": 0.43,
"learning_rate": 8.86e-05,
"loss": 2.8555,
"step": 450
},
{
"epoch": 0.45,
"learning_rate": 9.360000000000001e-05,
"loss": 2.4833,
"step": 475
},
{
"epoch": 0.48,
"learning_rate": 9.86e-05,
"loss": 2.7599,
"step": 500
},
{
"epoch": 0.5,
"learning_rate": 9.948571428571429e-05,
"loss": 2.8813,
"step": 525
},
{
"epoch": 0.52,
"learning_rate": 9.877142857142858e-05,
"loss": 2.6932,
"step": 550
},
{
"epoch": 0.55,
"learning_rate": 9.805714285714286e-05,
"loss": 2.7549,
"step": 575
},
{
"epoch": 0.57,
"learning_rate": 9.734285714285714e-05,
"loss": 2.7286,
"step": 600
},
{
"epoch": 0.6,
"learning_rate": 9.662857142857144e-05,
"loss": 2.1775,
"step": 625
},
{
"epoch": 0.62,
"learning_rate": 9.591428571428572e-05,
"loss": 2.4533,
"step": 650
},
{
"epoch": 0.64,
"learning_rate": 9.52e-05,
"loss": 3.0003,
"step": 675
},
{
"epoch": 0.67,
"learning_rate": 9.448571428571429e-05,
"loss": 2.6656,
"step": 700
},
{
"epoch": 0.69,
"learning_rate": 9.377142857142857e-05,
"loss": 2.4105,
"step": 725
},
{
"epoch": 0.72,
"learning_rate": 9.305714285714286e-05,
"loss": 2.2225,
"step": 750
},
{
"epoch": 0.74,
"learning_rate": 9.234285714285714e-05,
"loss": 2.6048,
"step": 775
},
{
"epoch": 0.76,
"learning_rate": 9.162857142857144e-05,
"loss": 2.3541,
"step": 800
},
{
"epoch": 0.79,
"learning_rate": 9.091428571428572e-05,
"loss": 1.6534,
"step": 825
},
{
"epoch": 0.81,
"learning_rate": 9.020000000000001e-05,
"loss": 2.6717,
"step": 850
},
{
"epoch": 0.83,
"learning_rate": 8.948571428571429e-05,
"loss": 2.1232,
"step": 875
},
{
"epoch": 0.86,
"learning_rate": 8.877142857142857e-05,
"loss": 2.1955,
"step": 900
},
{
"epoch": 0.88,
"learning_rate": 8.805714285714286e-05,
"loss": 2.4756,
"step": 925
},
{
"epoch": 0.91,
"learning_rate": 8.734285714285714e-05,
"loss": 2.0444,
"step": 950
},
{
"epoch": 0.93,
"learning_rate": 8.662857142857144e-05,
"loss": 2.124,
"step": 975
},
{
"epoch": 0.95,
"learning_rate": 8.591428571428572e-05,
"loss": 1.7168,
"step": 1000
},
{
"epoch": 0.95,
"eval_loss": 1.2106691598892212,
"eval_runtime": 380.7845,
"eval_samples_per_second": 3.939,
"eval_steps_per_second": 0.494,
"eval_wer": 64.80865224625624,
"step": 1000
},
{
"epoch": 0.98,
"learning_rate": 8.52e-05,
"loss": 2.451,
"step": 1025
},
{
"epoch": 1.0,
"learning_rate": 8.448571428571429e-05,
"loss": 1.4766,
"step": 1050
},
{
"epoch": 1.03,
"learning_rate": 8.377142857142858e-05,
"loss": 1.6965,
"step": 1075
},
{
"epoch": 1.05,
"learning_rate": 8.305714285714285e-05,
"loss": 1.4733,
"step": 1100
},
{
"epoch": 1.07,
"learning_rate": 8.234285714285714e-05,
"loss": 1.9818,
"step": 1125
},
{
"epoch": 1.1,
"learning_rate": 8.162857142857144e-05,
"loss": 2.0946,
"step": 1150
},
{
"epoch": 1.12,
"learning_rate": 8.091428571428572e-05,
"loss": 2.3333,
"step": 1175
},
{
"epoch": 1.15,
"learning_rate": 8.020000000000001e-05,
"loss": 1.7366,
"step": 1200
},
{
"epoch": 1.17,
"learning_rate": 7.948571428571429e-05,
"loss": 1.8155,
"step": 1225
},
{
"epoch": 1.19,
"learning_rate": 7.877142857142857e-05,
"loss": 2.1239,
"step": 1250
},
{
"epoch": 1.22,
"learning_rate": 7.805714285714286e-05,
"loss": 1.6473,
"step": 1275
},
{
"epoch": 1.24,
"learning_rate": 7.734285714285714e-05,
"loss": 1.1093,
"step": 1300
},
{
"epoch": 1.26,
"learning_rate": 7.662857142857142e-05,
"loss": 1.5792,
"step": 1325
},
{
"epoch": 1.29,
"learning_rate": 7.591428571428572e-05,
"loss": 1.9768,
"step": 1350
},
{
"epoch": 1.31,
"learning_rate": 7.52e-05,
"loss": 1.5958,
"step": 1375
},
{
"epoch": 1.34,
"learning_rate": 7.448571428571429e-05,
"loss": 1.6682,
"step": 1400
},
{
"epoch": 1.36,
"learning_rate": 7.377142857142859e-05,
"loss": 1.6654,
"step": 1425
},
{
"epoch": 1.38,
"learning_rate": 7.305714285714285e-05,
"loss": 1.6971,
"step": 1450
},
{
"epoch": 1.41,
"learning_rate": 7.234285714285715e-05,
"loss": 1.5318,
"step": 1475
},
{
"epoch": 1.43,
"learning_rate": 7.162857142857144e-05,
"loss": 1.3164,
"step": 1500
},
{
"epoch": 1.46,
"learning_rate": 7.091428571428572e-05,
"loss": 2.14,
"step": 1525
},
{
"epoch": 1.48,
"learning_rate": 7.02e-05,
"loss": 1.9155,
"step": 1550
},
{
"epoch": 1.5,
"learning_rate": 6.948571428571429e-05,
"loss": 1.4583,
"step": 1575
},
{
"epoch": 1.53,
"learning_rate": 6.877142857142857e-05,
"loss": 1.5538,
"step": 1600
},
{
"epoch": 1.55,
"learning_rate": 6.805714285714287e-05,
"loss": 1.1043,
"step": 1625
},
{
"epoch": 1.57,
"learning_rate": 6.734285714285715e-05,
"loss": 1.8925,
"step": 1650
},
{
"epoch": 1.6,
"learning_rate": 6.662857142857143e-05,
"loss": 1.7625,
"step": 1675
},
{
"epoch": 1.62,
"learning_rate": 6.591428571428572e-05,
"loss": 1.082,
"step": 1700
},
{
"epoch": 1.65,
"learning_rate": 6.52e-05,
"loss": 1.3789,
"step": 1725
},
{
"epoch": 1.67,
"learning_rate": 6.448571428571429e-05,
"loss": 1.568,
"step": 1750
},
{
"epoch": 1.69,
"learning_rate": 6.377142857142857e-05,
"loss": 1.3194,
"step": 1775
},
{
"epoch": 1.72,
"learning_rate": 6.305714285714285e-05,
"loss": 1.3629,
"step": 1800
},
{
"epoch": 1.74,
"learning_rate": 6.234285714285715e-05,
"loss": 1.6437,
"step": 1825
},
{
"epoch": 1.77,
"learning_rate": 6.162857142857144e-05,
"loss": 1.0288,
"step": 1850
},
{
"epoch": 1.79,
"learning_rate": 6.091428571428571e-05,
"loss": 1.5278,
"step": 1875
},
{
"epoch": 1.81,
"learning_rate": 6.02e-05,
"loss": 1.2791,
"step": 1900
},
{
"epoch": 1.84,
"learning_rate": 5.9485714285714286e-05,
"loss": 1.6202,
"step": 1925
},
{
"epoch": 1.86,
"learning_rate": 5.877142857142858e-05,
"loss": 1.3762,
"step": 1950
},
{
"epoch": 1.88,
"learning_rate": 5.805714285714285e-05,
"loss": 1.8212,
"step": 1975
},
{
"epoch": 1.91,
"learning_rate": 5.734285714285714e-05,
"loss": 1.1073,
"step": 2000
},
{
"epoch": 1.91,
"eval_loss": 0.9890535473823547,
"eval_runtime": 356.3677,
"eval_samples_per_second": 4.209,
"eval_steps_per_second": 0.528,
"eval_wer": 53.001941209095946,
"step": 2000
},
{
"epoch": 1.93,
"learning_rate": 5.662857142857143e-05,
"loss": 1.5077,
"step": 2025
},
{
"epoch": 1.96,
"learning_rate": 5.591428571428572e-05,
"loss": 1.1343,
"step": 2050
},
{
"epoch": 1.98,
"learning_rate": 5.520000000000001e-05,
"loss": 1.1924,
"step": 2075
},
{
"epoch": 2.0,
"learning_rate": 5.448571428571429e-05,
"loss": 1.2554,
"step": 2100
},
{
"epoch": 2.03,
"learning_rate": 5.3771428571428574e-05,
"loss": 1.5992,
"step": 2125
},
{
"epoch": 2.05,
"learning_rate": 5.305714285714286e-05,
"loss": 0.9659,
"step": 2150
},
{
"epoch": 2.08,
"learning_rate": 5.234285714285715e-05,
"loss": 1.2035,
"step": 2175
},
{
"epoch": 2.1,
"learning_rate": 5.162857142857143e-05,
"loss": 1.3945,
"step": 2200
},
{
"epoch": 2.12,
"learning_rate": 5.0914285714285714e-05,
"loss": 1.0145,
"step": 2225
},
{
"epoch": 2.15,
"learning_rate": 5.02e-05,
"loss": 1.0097,
"step": 2250
},
{
"epoch": 2.17,
"learning_rate": 4.9485714285714294e-05,
"loss": 0.9375,
"step": 2275
},
{
"epoch": 2.19,
"learning_rate": 4.8771428571428574e-05,
"loss": 0.9557,
"step": 2300
},
{
"epoch": 2.22,
"learning_rate": 4.805714285714286e-05,
"loss": 0.9563,
"step": 2325
},
{
"epoch": 2.24,
"learning_rate": 4.734285714285715e-05,
"loss": 1.2328,
"step": 2350
},
{
"epoch": 2.27,
"learning_rate": 4.6628571428571434e-05,
"loss": 1.4048,
"step": 2375
},
{
"epoch": 2.29,
"learning_rate": 4.5914285714285714e-05,
"loss": 0.9962,
"step": 2400
},
{
"epoch": 2.31,
"learning_rate": 4.52e-05,
"loss": 1.4589,
"step": 2425
},
{
"epoch": 2.34,
"learning_rate": 4.448571428571429e-05,
"loss": 1.2766,
"step": 2450
},
{
"epoch": 2.36,
"learning_rate": 4.3771428571428574e-05,
"loss": 1.3044,
"step": 2475
},
{
"epoch": 2.39,
"learning_rate": 4.3057142857142854e-05,
"loss": 2.5598,
"step": 2500
},
{
"epoch": 2.41,
"learning_rate": 4.234285714285715e-05,
"loss": 1.2634,
"step": 2525
},
{
"epoch": 2.43,
"learning_rate": 4.162857142857143e-05,
"loss": 1.2135,
"step": 2550
},
{
"epoch": 2.46,
"learning_rate": 4.0914285714285715e-05,
"loss": 1.0847,
"step": 2575
},
{
"epoch": 2.48,
"learning_rate": 4.02e-05,
"loss": 0.8392,
"step": 2600
},
{
"epoch": 2.5,
"learning_rate": 3.948571428571429e-05,
"loss": 1.7331,
"step": 2625
},
{
"epoch": 2.53,
"learning_rate": 3.8771428571428575e-05,
"loss": 1.468,
"step": 2650
},
{
"epoch": 2.55,
"learning_rate": 3.8057142857142855e-05,
"loss": 0.9024,
"step": 2675
},
{
"epoch": 2.58,
"learning_rate": 3.734285714285715e-05,
"loss": 0.8736,
"step": 2700
},
{
"epoch": 2.6,
"learning_rate": 3.662857142857143e-05,
"loss": 0.8067,
"step": 2725
},
{
"epoch": 2.62,
"learning_rate": 3.5914285714285715e-05,
"loss": 1.2113,
"step": 2750
},
{
"epoch": 2.65,
"learning_rate": 3.52e-05,
"loss": 1.2159,
"step": 2775
},
{
"epoch": 2.67,
"learning_rate": 3.448571428571429e-05,
"loss": 0.9207,
"step": 2800
},
{
"epoch": 2.7,
"learning_rate": 3.377142857142857e-05,
"loss": 1.2301,
"step": 2825
},
{
"epoch": 2.72,
"learning_rate": 3.305714285714286e-05,
"loss": 0.8843,
"step": 2850
},
{
"epoch": 2.74,
"learning_rate": 3.234285714285715e-05,
"loss": 0.998,
"step": 2875
},
{
"epoch": 2.77,
"learning_rate": 3.162857142857143e-05,
"loss": 1.0264,
"step": 2900
},
{
"epoch": 2.79,
"learning_rate": 3.0914285714285715e-05,
"loss": 0.8172,
"step": 2925
},
{
"epoch": 2.81,
"learning_rate": 3.02e-05,
"loss": 1.0094,
"step": 2950
},
{
"epoch": 2.84,
"learning_rate": 2.948571428571429e-05,
"loss": 0.9068,
"step": 2975
},
{
"epoch": 2.86,
"learning_rate": 2.8771428571428572e-05,
"loss": 1.341,
"step": 3000
},
{
"epoch": 2.86,
"eval_loss": 0.8741958141326904,
"eval_runtime": 363.1523,
"eval_samples_per_second": 4.13,
"eval_steps_per_second": 0.518,
"eval_wer": 47.76760953965613,
"step": 3000
}
],
"max_steps": 4000,
"num_train_epochs": 4,
"total_flos": 1.18136182136832e+18,
"trial_name": null,
"trial_params": null
}