whisper-small-amet / trainer_state.json
drmeeseeks's picture
End of training
1aaecd9
raw
history blame
14.8 kB
{
"best_metric": 100.0,
"best_model_checkpoint": "./whisper-small-amet/checkpoint-1100",
"epoch": 2000.0,
"global_step": 2000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 25.0,
"learning_rate": 4.2000000000000006e-07,
"loss": 2.9938,
"step": 25
},
{
"epoch": 50.0,
"learning_rate": 9.200000000000001e-07,
"loss": 2.2746,
"step": 50
},
{
"epoch": 75.0,
"learning_rate": 1.4000000000000001e-06,
"loss": 1.7055,
"step": 75
},
{
"epoch": 100.0,
"learning_rate": 1.9000000000000002e-06,
"loss": 0.9013,
"step": 100
},
{
"epoch": 100.0,
"eval_loss": 2.7051408290863037,
"eval_runtime": 8.4302,
"eval_samples_per_second": 0.593,
"eval_steps_per_second": 0.119,
"eval_wer": 276.0,
"step": 100
},
{
"epoch": 125.0,
"learning_rate": 2.4000000000000003e-06,
"loss": 0.1244,
"step": 125
},
{
"epoch": 150.0,
"learning_rate": 2.9e-06,
"loss": 0.0031,
"step": 150
},
{
"epoch": 175.0,
"learning_rate": 3.4000000000000005e-06,
"loss": 0.0004,
"step": 175
},
{
"epoch": 200.0,
"learning_rate": 3.900000000000001e-06,
"loss": 0.0002,
"step": 200
},
{
"epoch": 200.0,
"eval_loss": 3.741528034210205,
"eval_runtime": 8.6419,
"eval_samples_per_second": 0.579,
"eval_steps_per_second": 0.116,
"eval_wer": 334.6666666666667,
"step": 200
},
{
"epoch": 225.0,
"learning_rate": 4.4e-06,
"loss": 0.0002,
"step": 225
},
{
"epoch": 250.0,
"learning_rate": 4.9000000000000005e-06,
"loss": 0.0001,
"step": 250
},
{
"epoch": 275.0,
"learning_rate": 5.400000000000001e-06,
"loss": 0.0001,
"step": 275
},
{
"epoch": 300.0,
"learning_rate": 5.9e-06,
"loss": 0.0001,
"step": 300
},
{
"epoch": 300.0,
"eval_loss": 3.840170383453369,
"eval_runtime": 1.5102,
"eval_samples_per_second": 3.311,
"eval_steps_per_second": 0.662,
"eval_wer": 117.33333333333333,
"step": 300
},
{
"epoch": 325.0,
"learning_rate": 6.4000000000000006e-06,
"loss": 0.0001,
"step": 325
},
{
"epoch": 350.0,
"learning_rate": 6.9e-06,
"loss": 0.0001,
"step": 350
},
{
"epoch": 375.0,
"learning_rate": 7.4e-06,
"loss": 0.0001,
"step": 375
},
{
"epoch": 400.0,
"learning_rate": 7.9e-06,
"loss": 0.0001,
"step": 400
},
{
"epoch": 400.0,
"eval_loss": 3.8931400775909424,
"eval_runtime": 8.4556,
"eval_samples_per_second": 0.591,
"eval_steps_per_second": 0.118,
"eval_wer": 340.0,
"step": 400
},
{
"epoch": 425.0,
"learning_rate": 8.400000000000001e-06,
"loss": 0.0001,
"step": 425
},
{
"epoch": 450.0,
"learning_rate": 8.900000000000001e-06,
"loss": 0.0001,
"step": 450
},
{
"epoch": 475.0,
"learning_rate": 9.4e-06,
"loss": 0.0001,
"step": 475
},
{
"epoch": 500.0,
"learning_rate": 9.9e-06,
"loss": 0.0001,
"step": 500
},
{
"epoch": 500.0,
"eval_loss": 4.06705904006958,
"eval_runtime": 8.4836,
"eval_samples_per_second": 0.589,
"eval_steps_per_second": 0.118,
"eval_wer": 397.3333333333333,
"step": 500
},
{
"epoch": 525.0,
"learning_rate": 9.866666666666668e-06,
"loss": 0.0001,
"step": 525
},
{
"epoch": 550.0,
"learning_rate": 9.7e-06,
"loss": 0.0001,
"step": 550
},
{
"epoch": 575.0,
"learning_rate": 9.533333333333334e-06,
"loss": 0.0001,
"step": 575
},
{
"epoch": 600.0,
"learning_rate": 9.366666666666668e-06,
"loss": 0.0001,
"step": 600
},
{
"epoch": 600.0,
"eval_loss": 4.284416675567627,
"eval_runtime": 1.5895,
"eval_samples_per_second": 3.146,
"eval_steps_per_second": 0.629,
"eval_wer": 137.33333333333334,
"step": 600
},
{
"epoch": 625.0,
"learning_rate": 9.200000000000002e-06,
"loss": 0.0001,
"step": 625
},
{
"epoch": 650.0,
"learning_rate": 9.033333333333334e-06,
"loss": 0.0,
"step": 650
},
{
"epoch": 675.0,
"learning_rate": 8.866666666666668e-06,
"loss": 0.0,
"step": 675
},
{
"epoch": 700.0,
"learning_rate": 8.700000000000001e-06,
"loss": 0.0,
"step": 700
},
{
"epoch": 700.0,
"eval_loss": 4.469689846038818,
"eval_runtime": 4.8142,
"eval_samples_per_second": 1.039,
"eval_steps_per_second": 0.208,
"eval_wer": 289.33333333333337,
"step": 700
},
{
"epoch": 725.0,
"learning_rate": 8.533333333333335e-06,
"loss": 0.0,
"step": 725
},
{
"epoch": 750.0,
"learning_rate": 8.366666666666667e-06,
"loss": 0.0,
"step": 750
},
{
"epoch": 775.0,
"learning_rate": 8.2e-06,
"loss": 0.0,
"step": 775
},
{
"epoch": 800.0,
"learning_rate": 8.033333333333335e-06,
"loss": 0.0,
"step": 800
},
{
"epoch": 800.0,
"eval_loss": 4.627803325653076,
"eval_runtime": 8.4484,
"eval_samples_per_second": 0.592,
"eval_steps_per_second": 0.118,
"eval_wer": 449.3333333333333,
"step": 800
},
{
"epoch": 825.0,
"learning_rate": 7.866666666666667e-06,
"loss": 0.0,
"step": 825
},
{
"epoch": 850.0,
"learning_rate": 7.7e-06,
"loss": 0.0,
"step": 850
},
{
"epoch": 875.0,
"learning_rate": 7.533333333333334e-06,
"loss": 0.0,
"step": 875
},
{
"epoch": 900.0,
"learning_rate": 7.3666666666666676e-06,
"loss": 0.0,
"step": 900
},
{
"epoch": 900.0,
"eval_loss": 4.779428005218506,
"eval_runtime": 8.432,
"eval_samples_per_second": 0.593,
"eval_steps_per_second": 0.119,
"eval_wer": 678.6666666666667,
"step": 900
},
{
"epoch": 925.0,
"learning_rate": 7.2000000000000005e-06,
"loss": 0.0,
"step": 925
},
{
"epoch": 950.0,
"learning_rate": 7.033333333333334e-06,
"loss": 0.0,
"step": 950
},
{
"epoch": 975.0,
"learning_rate": 6.886666666666667e-06,
"loss": 0.0753,
"step": 975
},
{
"epoch": 1000.0,
"learning_rate": 6.720000000000001e-06,
"loss": 0.0405,
"step": 1000
},
{
"epoch": 1000.0,
"eval_loss": 4.676939487457275,
"eval_runtime": 8.4726,
"eval_samples_per_second": 0.59,
"eval_steps_per_second": 0.118,
"eval_wer": 261.3333333333333,
"step": 1000
},
{
"epoch": 1025.0,
"learning_rate": 6.553333333333334e-06,
"loss": 0.0057,
"step": 1025
},
{
"epoch": 1050.0,
"learning_rate": 6.386666666666668e-06,
"loss": 0.0001,
"step": 1050
},
{
"epoch": 1075.0,
"learning_rate": 6.220000000000001e-06,
"loss": 0.0001,
"step": 1075
},
{
"epoch": 1100.0,
"learning_rate": 6.0533333333333335e-06,
"loss": 0.0002,
"step": 1100
},
{
"epoch": 1100.0,
"eval_loss": 5.499487400054932,
"eval_runtime": 8.5264,
"eval_samples_per_second": 0.586,
"eval_steps_per_second": 0.117,
"eval_wer": 100.0,
"step": 1100
},
{
"epoch": 1125.0,
"learning_rate": 5.886666666666667e-06,
"loss": 0.0002,
"step": 1125
},
{
"epoch": 1150.0,
"learning_rate": 5.72e-06,
"loss": 0.0002,
"step": 1150
},
{
"epoch": 1175.0,
"learning_rate": 5.553333333333334e-06,
"loss": 0.0002,
"step": 1175
},
{
"epoch": 1200.0,
"learning_rate": 5.386666666666667e-06,
"loss": 0.0002,
"step": 1200
},
{
"epoch": 1200.0,
"eval_loss": 6.003348350524902,
"eval_runtime": 8.4985,
"eval_samples_per_second": 0.588,
"eval_steps_per_second": 0.118,
"eval_wer": 100.0,
"step": 1200
},
{
"epoch": 1225.0,
"learning_rate": 5.220000000000001e-06,
"loss": 0.0002,
"step": 1225
},
{
"epoch": 1250.0,
"learning_rate": 5.053333333333334e-06,
"loss": 0.0002,
"step": 1250
},
{
"epoch": 1275.0,
"learning_rate": 4.886666666666668e-06,
"loss": 0.0002,
"step": 1275
},
{
"epoch": 1300.0,
"learning_rate": 4.7200000000000005e-06,
"loss": 0.0002,
"step": 1300
},
{
"epoch": 1300.0,
"eval_loss": 6.288400650024414,
"eval_runtime": 8.612,
"eval_samples_per_second": 0.581,
"eval_steps_per_second": 0.116,
"eval_wer": 100.0,
"step": 1300
},
{
"epoch": 1325.0,
"learning_rate": 4.5533333333333335e-06,
"loss": 0.0002,
"step": 1325
},
{
"epoch": 1350.0,
"learning_rate": 4.3866666666666665e-06,
"loss": 0.0002,
"step": 1350
},
{
"epoch": 1375.0,
"learning_rate": 4.22e-06,
"loss": 0.0002,
"step": 1375
},
{
"epoch": 1400.0,
"learning_rate": 4.053333333333333e-06,
"loss": 0.0002,
"step": 1400
},
{
"epoch": 1400.0,
"eval_loss": 6.474369049072266,
"eval_runtime": 8.436,
"eval_samples_per_second": 0.593,
"eval_steps_per_second": 0.119,
"eval_wer": 100.0,
"step": 1400
},
{
"epoch": 1425.0,
"learning_rate": 3.886666666666667e-06,
"loss": 0.0002,
"step": 1425
},
{
"epoch": 1450.0,
"learning_rate": 3.7200000000000004e-06,
"loss": 0.0002,
"step": 1450
},
{
"epoch": 1475.0,
"learning_rate": 3.5533333333333338e-06,
"loss": 0.0002,
"step": 1475
},
{
"epoch": 1500.0,
"learning_rate": 3.386666666666667e-06,
"loss": 0.0002,
"step": 1500
},
{
"epoch": 1500.0,
"eval_loss": 6.596408843994141,
"eval_runtime": 8.5317,
"eval_samples_per_second": 0.586,
"eval_steps_per_second": 0.117,
"eval_wer": 100.0,
"step": 1500
},
{
"epoch": 1525.0,
"learning_rate": 3.2200000000000005e-06,
"loss": 0.0002,
"step": 1525
},
{
"epoch": 1550.0,
"learning_rate": 3.053333333333334e-06,
"loss": 0.0001,
"step": 1550
},
{
"epoch": 1575.0,
"learning_rate": 2.8866666666666673e-06,
"loss": 0.0001,
"step": 1575
},
{
"epoch": 1600.0,
"learning_rate": 2.7200000000000002e-06,
"loss": 0.0001,
"step": 1600
},
{
"epoch": 1600.0,
"eval_loss": 6.679154872894287,
"eval_runtime": 8.5665,
"eval_samples_per_second": 0.584,
"eval_steps_per_second": 0.117,
"eval_wer": 100.0,
"step": 1600
},
{
"epoch": 1625.0,
"learning_rate": 2.5533333333333336e-06,
"loss": 0.0001,
"step": 1625
},
{
"epoch": 1650.0,
"learning_rate": 2.386666666666667e-06,
"loss": 0.0001,
"step": 1650
},
{
"epoch": 1675.0,
"learning_rate": 2.2200000000000003e-06,
"loss": 0.0001,
"step": 1675
},
{
"epoch": 1700.0,
"learning_rate": 2.0533333333333337e-06,
"loss": 0.0001,
"step": 1700
},
{
"epoch": 1700.0,
"eval_loss": 6.7370285987854,
"eval_runtime": 8.4426,
"eval_samples_per_second": 0.592,
"eval_steps_per_second": 0.118,
"eval_wer": 100.0,
"step": 1700
},
{
"epoch": 1725.0,
"learning_rate": 1.8866666666666669e-06,
"loss": 0.0001,
"step": 1725
},
{
"epoch": 1750.0,
"learning_rate": 1.72e-06,
"loss": 0.0001,
"step": 1750
},
{
"epoch": 1775.0,
"learning_rate": 1.5533333333333334e-06,
"loss": 0.0001,
"step": 1775
},
{
"epoch": 1800.0,
"learning_rate": 1.3866666666666668e-06,
"loss": 0.0001,
"step": 1800
},
{
"epoch": 1800.0,
"eval_loss": 6.773484230041504,
"eval_runtime": 8.4858,
"eval_samples_per_second": 0.589,
"eval_steps_per_second": 0.118,
"eval_wer": 100.0,
"step": 1800
},
{
"epoch": 1825.0,
"learning_rate": 1.2200000000000002e-06,
"loss": 0.0001,
"step": 1825
},
{
"epoch": 1850.0,
"learning_rate": 1.0533333333333333e-06,
"loss": 0.0001,
"step": 1850
},
{
"epoch": 1875.0,
"learning_rate": 8.866666666666668e-07,
"loss": 0.0001,
"step": 1875
},
{
"epoch": 1900.0,
"learning_rate": 7.2e-07,
"loss": 0.0001,
"step": 1900
},
{
"epoch": 1900.0,
"eval_loss": 6.795783042907715,
"eval_runtime": 8.5107,
"eval_samples_per_second": 0.587,
"eval_steps_per_second": 0.117,
"eval_wer": 100.0,
"step": 1900
},
{
"epoch": 1925.0,
"learning_rate": 5.533333333333334e-07,
"loss": 0.0001,
"step": 1925
},
{
"epoch": 1950.0,
"learning_rate": 3.8666666666666674e-07,
"loss": 0.0001,
"step": 1950
},
{
"epoch": 1975.0,
"learning_rate": 2.2e-07,
"loss": 0.0001,
"step": 1975
},
{
"epoch": 2000.0,
"learning_rate": 5.3333333333333334e-08,
"loss": 0.0001,
"step": 2000
},
{
"epoch": 2000.0,
"eval_loss": 6.801175117492676,
"eval_runtime": 8.5138,
"eval_samples_per_second": 0.587,
"eval_steps_per_second": 0.117,
"eval_wer": 100.0,
"step": 2000
},
{
"epoch": 2000.0,
"step": 2000,
"total_flos": 5.7717080064e+17,
"train_loss": 0.10164999849759625,
"train_runtime": 1221.0979,
"train_samples_per_second": 104.824,
"train_steps_per_second": 1.638
}
],
"max_steps": 2000,
"num_train_epochs": 2000,
"total_flos": 5.7717080064e+17,
"trial_name": null,
"trial_params": null
}