hubert_base_arabic_mdd / trainer_state.json
nrshoudi's picture
End of training
8edf60c
raw
history blame
8.04 kB
{
"best_metric": 3.2237985134124756,
"best_model_checkpoint": "nrshoudi/hubert_base_arabic_mdd/checkpoint-4911",
"epoch": 20.0,
"eval_steps": 500,
"global_step": 32740,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 1.0,
"learning_rate": 5e-05,
"loss": 6.526,
"step": 1637
},
{
"epoch": 1.0,
"eval_loss": 3.364994525909424,
"eval_per": 1.0,
"eval_runtime": 79.0541,
"eval_samples_per_second": 10.284,
"eval_steps_per_second": 5.148,
"eval_wer": 1.0,
"step": 1637
},
{
"epoch": 2.0,
"learning_rate": 0.0001,
"loss": 3.2555,
"step": 3274
},
{
"epoch": 2.0,
"eval_loss": 3.275482416152954,
"eval_per": 1.0,
"eval_runtime": 81.9459,
"eval_samples_per_second": 9.921,
"eval_steps_per_second": 4.967,
"eval_wer": 1.0,
"step": 3274
},
{
"epoch": 3.0,
"learning_rate": 9.444444444444444e-05,
"loss": 3.2548,
"step": 4911
},
{
"epoch": 3.0,
"eval_loss": 3.2237985134124756,
"eval_per": 1.0,
"eval_runtime": 79.8785,
"eval_samples_per_second": 10.178,
"eval_steps_per_second": 5.095,
"eval_wer": 1.0,
"step": 4911
},
{
"epoch": 4.0,
"learning_rate": 8.888888888888889e-05,
"loss": 3.2385,
"step": 6548
},
{
"epoch": 4.0,
"eval_loss": 3.284461259841919,
"eval_per": 1.0,
"eval_runtime": 80.0287,
"eval_samples_per_second": 10.159,
"eval_steps_per_second": 5.086,
"eval_wer": 1.0,
"step": 6548
},
{
"epoch": 5.0,
"learning_rate": 8.333333333333334e-05,
"loss": 3.2358,
"step": 8185
},
{
"epoch": 5.0,
"eval_loss": 3.227111339569092,
"eval_per": 1.0,
"eval_runtime": 79.4783,
"eval_samples_per_second": 10.229,
"eval_steps_per_second": 5.121,
"eval_wer": 1.0,
"step": 8185
},
{
"epoch": 6.0,
"learning_rate": 7.777777777777778e-05,
"loss": 3.237,
"step": 9822
},
{
"epoch": 6.0,
"eval_loss": 3.2472596168518066,
"eval_per": 1.0,
"eval_runtime": 80.029,
"eval_samples_per_second": 10.159,
"eval_steps_per_second": 5.086,
"eval_wer": 1.0,
"step": 9822
},
{
"epoch": 7.0,
"learning_rate": 7.222222222222222e-05,
"loss": 3.2622,
"step": 11459
},
{
"epoch": 7.0,
"eval_loss": 3.2289435863494873,
"eval_per": 1.0,
"eval_runtime": 81.2354,
"eval_samples_per_second": 10.008,
"eval_steps_per_second": 5.01,
"eval_wer": 1.0,
"step": 11459
},
{
"epoch": 8.0,
"learning_rate": 6.666666666666667e-05,
"loss": 3.2614,
"step": 13096
},
{
"epoch": 8.0,
"eval_loss": 3.2283053398132324,
"eval_per": 1.0,
"eval_runtime": 80.9742,
"eval_samples_per_second": 10.04,
"eval_steps_per_second": 5.026,
"eval_wer": 1.0,
"step": 13096
},
{
"epoch": 9.0,
"learning_rate": 6.111111111111112e-05,
"loss": 3.224,
"step": 14733
},
{
"epoch": 9.0,
"eval_loss": 3.224926233291626,
"eval_per": 1.0,
"eval_runtime": 79.9396,
"eval_samples_per_second": 10.17,
"eval_steps_per_second": 5.091,
"eval_wer": 1.0,
"step": 14733
},
{
"epoch": 10.0,
"learning_rate": 5.555555555555556e-05,
"loss": 3.2221,
"step": 16370
},
{
"epoch": 10.0,
"eval_loss": 3.233494520187378,
"eval_per": 1.0,
"eval_runtime": 79.7551,
"eval_samples_per_second": 10.194,
"eval_steps_per_second": 5.103,
"eval_wer": 1.0,
"step": 16370
},
{
"epoch": 11.0,
"learning_rate": 5e-05,
"loss": 3.222,
"step": 18007
},
{
"epoch": 11.0,
"eval_loss": 3.2356951236724854,
"eval_per": 1.0,
"eval_runtime": 80.1958,
"eval_samples_per_second": 10.138,
"eval_steps_per_second": 5.075,
"eval_wer": 1.0,
"step": 18007
},
{
"epoch": 12.0,
"learning_rate": 4.4444444444444447e-05,
"loss": 3.2218,
"step": 19644
},
{
"epoch": 12.0,
"eval_loss": 3.24910569190979,
"eval_per": 1.0,
"eval_runtime": 81.0704,
"eval_samples_per_second": 10.028,
"eval_steps_per_second": 5.02,
"eval_wer": 1.0,
"step": 19644
},
{
"epoch": 13.0,
"learning_rate": 3.888888888888889e-05,
"loss": 3.2183,
"step": 21281
},
{
"epoch": 13.0,
"eval_loss": 3.2446420192718506,
"eval_per": 1.0,
"eval_runtime": 81.004,
"eval_samples_per_second": 10.037,
"eval_steps_per_second": 5.024,
"eval_wer": 1.0,
"step": 21281
},
{
"epoch": 14.0,
"learning_rate": 3.3333333333333335e-05,
"loss": 3.2181,
"step": 22918
},
{
"epoch": 14.0,
"eval_loss": 3.2415904998779297,
"eval_per": 1.0,
"eval_runtime": 83.7321,
"eval_samples_per_second": 9.71,
"eval_steps_per_second": 4.861,
"eval_wer": 1.0,
"step": 22918
},
{
"epoch": 15.0,
"learning_rate": 2.777777777777778e-05,
"loss": 3.2164,
"step": 24555
},
{
"epoch": 15.0,
"eval_loss": 3.22587513923645,
"eval_per": 1.0,
"eval_runtime": 81.1128,
"eval_samples_per_second": 10.023,
"eval_steps_per_second": 5.018,
"eval_wer": 1.0,
"step": 24555
},
{
"epoch": 16.0,
"learning_rate": 2.2222222222222223e-05,
"loss": 3.2148,
"step": 26192
},
{
"epoch": 16.0,
"eval_loss": 3.224907159805298,
"eval_per": 1.0,
"eval_runtime": 80.4106,
"eval_samples_per_second": 10.111,
"eval_steps_per_second": 5.062,
"eval_wer": 1.0,
"step": 26192
},
{
"epoch": 17.0,
"learning_rate": 1.6666666666666667e-05,
"loss": 3.2139,
"step": 27829
},
{
"epoch": 17.0,
"eval_loss": 3.232697010040283,
"eval_per": 1.0,
"eval_runtime": 80.7791,
"eval_samples_per_second": 10.064,
"eval_steps_per_second": 5.038,
"eval_wer": 1.0,
"step": 27829
},
{
"epoch": 18.0,
"learning_rate": 1.1111111111111112e-05,
"loss": 3.2133,
"step": 29466
},
{
"epoch": 18.0,
"eval_loss": 3.2251381874084473,
"eval_per": 1.0,
"eval_runtime": 81.6112,
"eval_samples_per_second": 9.962,
"eval_steps_per_second": 4.987,
"eval_wer": 1.0,
"step": 29466
},
{
"epoch": 19.0,
"learning_rate": 5.555555555555556e-06,
"loss": 3.2128,
"step": 31103
},
{
"epoch": 19.0,
"eval_loss": 3.2287981510162354,
"eval_per": 1.0,
"eval_runtime": 83.1142,
"eval_samples_per_second": 9.782,
"eval_steps_per_second": 4.897,
"eval_wer": 1.0,
"step": 31103
},
{
"epoch": 20.0,
"learning_rate": 0.0,
"loss": 3.2113,
"step": 32740
},
{
"epoch": 20.0,
"eval_loss": 3.2265169620513916,
"eval_per": 1.0,
"eval_runtime": 81.1257,
"eval_samples_per_second": 10.021,
"eval_steps_per_second": 5.017,
"eval_wer": 1.0,
"step": 32740
},
{
"epoch": 20.0,
"step": 32740,
"total_flos": 3.133881121657545e+18,
"train_loss": 3.3939919333670585,
"train_runtime": 10237.0354,
"train_samples_per_second": 6.394,
"train_steps_per_second": 3.198
}
],
"logging_steps": 500,
"max_steps": 32740,
"num_train_epochs": 20,
"save_steps": 500,
"total_flos": 3.133881121657545e+18,
"trial_name": null,
"trial_params": null
}