whisper-medium-sw / trainer_state.json
hedronstone's picture
Upload . with huggingface_hub
2cfb154
raw
history blame
15.5 kB
{
"best_metric": 20.505582521486684,
"best_model_checkpoint": "./whisper-medium-sw/checkpoint-3000",
"epoch": 1.311,
"global_step": 3000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.01,
"learning_rate": 1.1500000000000002e-06,
"loss": 2.7757,
"step": 25
},
{
"epoch": 0.02,
"learning_rate": 2.4000000000000003e-06,
"loss": 1.9286,
"step": 50
},
{
"epoch": 0.03,
"learning_rate": 3.65e-06,
"loss": 1.3214,
"step": 75
},
{
"epoch": 0.03,
"learning_rate": 4.9000000000000005e-06,
"loss": 1.0883,
"step": 100
},
{
"epoch": 0.04,
"learning_rate": 6.15e-06,
"loss": 0.8492,
"step": 125
},
{
"epoch": 0.05,
"learning_rate": 7.4e-06,
"loss": 0.7121,
"step": 150
},
{
"epoch": 0.06,
"learning_rate": 8.65e-06,
"loss": 0.5631,
"step": 175
},
{
"epoch": 0.07,
"learning_rate": 9.9e-06,
"loss": 0.4165,
"step": 200
},
{
"epoch": 0.07,
"learning_rate": 1.1150000000000002e-05,
"loss": 0.339,
"step": 225
},
{
"epoch": 0.08,
"learning_rate": 1.2400000000000002e-05,
"loss": 0.2574,
"step": 250
},
{
"epoch": 0.09,
"learning_rate": 1.3650000000000001e-05,
"loss": 0.1914,
"step": 275
},
{
"epoch": 0.1,
"learning_rate": 1.4900000000000001e-05,
"loss": 0.2216,
"step": 300
},
{
"epoch": 0.11,
"learning_rate": 1.6150000000000003e-05,
"loss": 0.2163,
"step": 325
},
{
"epoch": 0.12,
"learning_rate": 1.7400000000000003e-05,
"loss": 0.2382,
"step": 350
},
{
"epoch": 0.12,
"learning_rate": 1.8650000000000003e-05,
"loss": 0.1823,
"step": 375
},
{
"epoch": 0.13,
"learning_rate": 1.9900000000000003e-05,
"loss": 0.1548,
"step": 400
},
{
"epoch": 0.14,
"learning_rate": 1.9823076923076924e-05,
"loss": 0.1404,
"step": 425
},
{
"epoch": 0.15,
"learning_rate": 1.9630769230769234e-05,
"loss": 0.1404,
"step": 450
},
{
"epoch": 0.16,
"learning_rate": 1.943846153846154e-05,
"loss": 0.147,
"step": 475
},
{
"epoch": 0.17,
"learning_rate": 1.924615384615385e-05,
"loss": 0.1353,
"step": 500
},
{
"epoch": 0.17,
"learning_rate": 1.9053846153846157e-05,
"loss": 0.092,
"step": 525
},
{
"epoch": 0.18,
"learning_rate": 1.886153846153846e-05,
"loss": 0.1137,
"step": 550
},
{
"epoch": 0.19,
"learning_rate": 1.866923076923077e-05,
"loss": 0.0946,
"step": 575
},
{
"epoch": 0.2,
"learning_rate": 1.847692307692308e-05,
"loss": 0.0928,
"step": 600
},
{
"epoch": 0.21,
"learning_rate": 1.8284615384615387e-05,
"loss": 0.0871,
"step": 625
},
{
"epoch": 0.22,
"learning_rate": 1.8092307692307694e-05,
"loss": 0.0995,
"step": 650
},
{
"epoch": 0.23,
"learning_rate": 1.79e-05,
"loss": 0.103,
"step": 675
},
{
"epoch": 0.23,
"learning_rate": 1.770769230769231e-05,
"loss": 0.1157,
"step": 700
},
{
"epoch": 0.24,
"learning_rate": 1.7515384615384616e-05,
"loss": 0.1345,
"step": 725
},
{
"epoch": 0.25,
"learning_rate": 1.7323076923076924e-05,
"loss": 0.1196,
"step": 750
},
{
"epoch": 0.26,
"learning_rate": 1.7130769230769235e-05,
"loss": 0.1026,
"step": 775
},
{
"epoch": 0.27,
"learning_rate": 1.693846153846154e-05,
"loss": 0.1061,
"step": 800
},
{
"epoch": 0.28,
"learning_rate": 1.6746153846153846e-05,
"loss": 0.0921,
"step": 825
},
{
"epoch": 0.28,
"learning_rate": 1.6553846153846157e-05,
"loss": 0.1368,
"step": 850
},
{
"epoch": 0.29,
"learning_rate": 1.636153846153846e-05,
"loss": 0.1264,
"step": 875
},
{
"epoch": 0.3,
"learning_rate": 1.6169230769230772e-05,
"loss": 0.0896,
"step": 900
},
{
"epoch": 0.31,
"learning_rate": 1.597692307692308e-05,
"loss": 0.057,
"step": 925
},
{
"epoch": 0.32,
"learning_rate": 1.5784615384615383e-05,
"loss": 0.0765,
"step": 950
},
{
"epoch": 0.33,
"learning_rate": 1.5592307692307694e-05,
"loss": 0.0736,
"step": 975
},
{
"epoch": 0.33,
"learning_rate": 1.54e-05,
"loss": 0.08,
"step": 1000
},
{
"epoch": 0.33,
"eval_loss": 0.38002386689186096,
"eval_runtime": 3409.0073,
"eval_samples_per_second": 3.003,
"eval_steps_per_second": 0.375,
"eval_wer": 23.59576348009685,
"step": 1000
},
{
"epoch": 0.34,
"learning_rate": 1.520769230769231e-05,
"loss": 0.0684,
"step": 1025
},
{
"epoch": 0.35,
"learning_rate": 1.5015384615384617e-05,
"loss": 0.0638,
"step": 1050
},
{
"epoch": 0.36,
"learning_rate": 1.4823076923076924e-05,
"loss": 0.0686,
"step": 1075
},
{
"epoch": 0.37,
"learning_rate": 1.4630769230769233e-05,
"loss": 0.0533,
"step": 1100
},
{
"epoch": 0.38,
"learning_rate": 1.4438461538461539e-05,
"loss": 0.0632,
"step": 1125
},
{
"epoch": 0.38,
"learning_rate": 1.4246153846153848e-05,
"loss": 0.0767,
"step": 1150
},
{
"epoch": 0.39,
"learning_rate": 1.4053846153846155e-05,
"loss": 0.0628,
"step": 1175
},
{
"epoch": 0.4,
"learning_rate": 1.3861538461538461e-05,
"loss": 0.0466,
"step": 1200
},
{
"epoch": 0.41,
"learning_rate": 1.366923076923077e-05,
"loss": 0.052,
"step": 1225
},
{
"epoch": 0.42,
"learning_rate": 1.3476923076923078e-05,
"loss": 0.0586,
"step": 1250
},
{
"epoch": 0.42,
"learning_rate": 1.3284615384615385e-05,
"loss": 0.0626,
"step": 1275
},
{
"epoch": 0.43,
"learning_rate": 1.3092307692307693e-05,
"loss": 0.0741,
"step": 1300
},
{
"epoch": 0.44,
"learning_rate": 1.2900000000000002e-05,
"loss": 0.0427,
"step": 1325
},
{
"epoch": 0.45,
"learning_rate": 1.270769230769231e-05,
"loss": 0.0436,
"step": 1350
},
{
"epoch": 0.46,
"learning_rate": 1.2515384615384617e-05,
"loss": 0.0336,
"step": 1375
},
{
"epoch": 0.47,
"learning_rate": 1.2323076923076924e-05,
"loss": 0.0354,
"step": 1400
},
{
"epoch": 0.47,
"learning_rate": 1.2130769230769233e-05,
"loss": 0.0369,
"step": 1425
},
{
"epoch": 0.48,
"learning_rate": 1.1938461538461539e-05,
"loss": 0.0573,
"step": 1450
},
{
"epoch": 0.49,
"learning_rate": 1.1746153846153847e-05,
"loss": 0.0535,
"step": 1475
},
{
"epoch": 0.5,
"learning_rate": 1.1553846153846156e-05,
"loss": 0.0621,
"step": 1500
},
{
"epoch": 0.51,
"learning_rate": 1.1361538461538461e-05,
"loss": 0.0694,
"step": 1525
},
{
"epoch": 0.52,
"learning_rate": 1.116923076923077e-05,
"loss": 0.0574,
"step": 1550
},
{
"epoch": 0.53,
"learning_rate": 1.0976923076923078e-05,
"loss": 0.0739,
"step": 1575
},
{
"epoch": 0.53,
"learning_rate": 1.0784615384615384e-05,
"loss": 0.0451,
"step": 1600
},
{
"epoch": 0.54,
"learning_rate": 1.0592307692307693e-05,
"loss": 0.0566,
"step": 1625
},
{
"epoch": 0.55,
"learning_rate": 1.04e-05,
"loss": 0.0589,
"step": 1650
},
{
"epoch": 0.56,
"learning_rate": 1.020769230769231e-05,
"loss": 0.0666,
"step": 1675
},
{
"epoch": 0.57,
"learning_rate": 1.0015384615384615e-05,
"loss": 0.0449,
"step": 1700
},
{
"epoch": 0.57,
"learning_rate": 9.823076923076924e-06,
"loss": 0.0349,
"step": 1725
},
{
"epoch": 0.58,
"learning_rate": 9.630769230769232e-06,
"loss": 0.0427,
"step": 1750
},
{
"epoch": 0.59,
"learning_rate": 9.43846153846154e-06,
"loss": 0.0482,
"step": 1775
},
{
"epoch": 0.6,
"learning_rate": 9.246153846153847e-06,
"loss": 0.0714,
"step": 1800
},
{
"epoch": 0.61,
"learning_rate": 9.053846153846154e-06,
"loss": 0.0503,
"step": 1825
},
{
"epoch": 0.62,
"learning_rate": 8.861538461538463e-06,
"loss": 0.0536,
"step": 1850
},
{
"epoch": 0.62,
"learning_rate": 8.669230769230769e-06,
"loss": 0.0488,
"step": 1875
},
{
"epoch": 0.63,
"learning_rate": 8.476923076923078e-06,
"loss": 0.04,
"step": 1900
},
{
"epoch": 0.64,
"learning_rate": 8.284615384615386e-06,
"loss": 0.0477,
"step": 1925
},
{
"epoch": 0.65,
"learning_rate": 8.092307692307693e-06,
"loss": 0.0318,
"step": 1950
},
{
"epoch": 0.66,
"learning_rate": 7.9e-06,
"loss": 0.0323,
"step": 1975
},
{
"epoch": 0.67,
"learning_rate": 7.707692307692308e-06,
"loss": 0.0295,
"step": 2000
},
{
"epoch": 0.67,
"eval_loss": 0.3322092294692993,
"eval_runtime": 3500.3035,
"eval_samples_per_second": 2.925,
"eval_steps_per_second": 0.366,
"eval_wer": 20.510172467210573,
"step": 2000
},
{
"epoch": 0.68,
"learning_rate": 7.5153846153846155e-06,
"loss": 0.0306,
"step": 2025
},
{
"epoch": 0.68,
"learning_rate": 7.323076923076924e-06,
"loss": 0.0426,
"step": 2050
},
{
"epoch": 1.0,
"learning_rate": 7.130769230769231e-06,
"loss": 0.0344,
"step": 2075
},
{
"epoch": 1.01,
"learning_rate": 6.9384615384615395e-06,
"loss": 0.0341,
"step": 2100
},
{
"epoch": 1.02,
"learning_rate": 6.746153846153847e-06,
"loss": 0.0497,
"step": 2125
},
{
"epoch": 1.03,
"learning_rate": 6.553846153846154e-06,
"loss": 0.0802,
"step": 2150
},
{
"epoch": 1.04,
"learning_rate": 6.361538461538463e-06,
"loss": 0.0641,
"step": 2175
},
{
"epoch": 1.04,
"learning_rate": 6.169230769230769e-06,
"loss": 0.0439,
"step": 2200
},
{
"epoch": 1.05,
"learning_rate": 5.976923076923078e-06,
"loss": 0.03,
"step": 2225
},
{
"epoch": 1.06,
"learning_rate": 5.784615384615385e-06,
"loss": 0.0279,
"step": 2250
},
{
"epoch": 1.07,
"learning_rate": 5.5923076923076925e-06,
"loss": 0.0321,
"step": 2275
},
{
"epoch": 1.08,
"learning_rate": 5.400000000000001e-06,
"loss": 0.0314,
"step": 2300
},
{
"epoch": 1.09,
"learning_rate": 5.207692307692308e-06,
"loss": 0.019,
"step": 2325
},
{
"epoch": 1.09,
"learning_rate": 5.015384615384616e-06,
"loss": 0.0117,
"step": 2350
},
{
"epoch": 1.1,
"learning_rate": 4.823076923076924e-06,
"loss": 0.0265,
"step": 2375
},
{
"epoch": 1.11,
"learning_rate": 4.630769230769231e-06,
"loss": 0.0271,
"step": 2400
},
{
"epoch": 1.12,
"learning_rate": 4.438461538461539e-06,
"loss": 0.029,
"step": 2425
},
{
"epoch": 1.13,
"learning_rate": 4.246153846153846e-06,
"loss": 0.0154,
"step": 2450
},
{
"epoch": 1.14,
"learning_rate": 4.053846153846155e-06,
"loss": 0.0119,
"step": 2475
},
{
"epoch": 1.14,
"learning_rate": 3.861538461538462e-06,
"loss": 0.0192,
"step": 2500
},
{
"epoch": 1.15,
"learning_rate": 3.6692307692307695e-06,
"loss": 0.0175,
"step": 2525
},
{
"epoch": 1.16,
"learning_rate": 3.476923076923077e-06,
"loss": 0.0231,
"step": 2550
},
{
"epoch": 1.17,
"learning_rate": 3.284615384615385e-06,
"loss": 0.0222,
"step": 2575
},
{
"epoch": 1.18,
"learning_rate": 3.0923076923076927e-06,
"loss": 0.0145,
"step": 2600
},
{
"epoch": 1.19,
"learning_rate": 2.9e-06,
"loss": 0.0257,
"step": 2625
},
{
"epoch": 1.19,
"learning_rate": 2.7076923076923076e-06,
"loss": 0.0147,
"step": 2650
},
{
"epoch": 1.2,
"learning_rate": 2.5153846153846155e-06,
"loss": 0.0121,
"step": 2675
},
{
"epoch": 1.21,
"learning_rate": 2.3230769230769234e-06,
"loss": 0.0167,
"step": 2700
},
{
"epoch": 1.22,
"learning_rate": 2.130769230769231e-06,
"loss": 0.02,
"step": 2725
},
{
"epoch": 1.23,
"learning_rate": 1.9384615384615387e-06,
"loss": 0.0254,
"step": 2750
},
{
"epoch": 1.24,
"learning_rate": 1.7461538461538464e-06,
"loss": 0.0273,
"step": 2775
},
{
"epoch": 1.24,
"learning_rate": 1.5538461538461538e-06,
"loss": 0.0451,
"step": 2800
},
{
"epoch": 1.25,
"learning_rate": 1.3615384615384617e-06,
"loss": 0.0266,
"step": 2825
},
{
"epoch": 1.26,
"learning_rate": 1.1692307692307693e-06,
"loss": 0.0193,
"step": 2850
},
{
"epoch": 1.27,
"learning_rate": 9.76923076923077e-07,
"loss": 0.0257,
"step": 2875
},
{
"epoch": 1.28,
"learning_rate": 7.846153846153847e-07,
"loss": 0.0252,
"step": 2900
},
{
"epoch": 1.29,
"learning_rate": 5.923076923076924e-07,
"loss": 0.0301,
"step": 2925
},
{
"epoch": 1.29,
"learning_rate": 4.0000000000000003e-07,
"loss": 0.0317,
"step": 2950
},
{
"epoch": 1.3,
"learning_rate": 2.076923076923077e-07,
"loss": 0.0165,
"step": 2975
},
{
"epoch": 1.31,
"learning_rate": 1.5384615384615385e-08,
"loss": 0.0078,
"step": 3000
},
{
"epoch": 1.31,
"eval_loss": 0.3208019435405731,
"eval_runtime": 3585.2008,
"eval_samples_per_second": 2.856,
"eval_steps_per_second": 0.357,
"eval_wer": 20.505582521486684,
"step": 3000
}
],
"max_steps": 3000,
"num_train_epochs": 9223372036854775807,
"total_flos": 5.540579959283712e+19,
"trial_name": null,
"trial_params": null
}