Wav2Vec2-Large-XLSR-Persian-ShEMO / trainer_state.json
lnxdx's picture
End of training
730d5bf
raw
history blame
7.89 kB
{
"best_metric": 0.3285630841121495,
"best_model_checkpoint": "/content/drive/MyDrive/wav2vec2-large-xlsr-persian-asr-shemo_me7494/checkpoint-1700",
"epoch": 12.5,
"eval_steps": 100,
"global_step": 2000,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 0.62,
"learning_rate": 1.98e-06,
"loss": 1.8553,
"step": 100
},
{
"epoch": 0.62,
"eval_loss": 1.4126451015472412,
"eval_runtime": 35.7389,
"eval_samples_per_second": 7.947,
"eval_steps_per_second": 1.007,
"eval_wer": 0.48656542056074764,
"step": 100
},
{
"epoch": 1.25,
"learning_rate": 3.980000000000001e-06,
"loss": 1.4083,
"step": 200
},
{
"epoch": 1.25,
"eval_loss": 1.0428400039672852,
"eval_runtime": 35.3623,
"eval_samples_per_second": 8.031,
"eval_steps_per_second": 1.018,
"eval_wer": 0.4366238317757009,
"step": 200
},
{
"epoch": 1.88,
"learning_rate": 5.98e-06,
"loss": 1.1718,
"step": 300
},
{
"epoch": 1.88,
"eval_loss": 0.86826491355896,
"eval_runtime": 34.9958,
"eval_samples_per_second": 8.115,
"eval_steps_per_second": 1.029,
"eval_wer": 0.4126752336448598,
"step": 300
},
{
"epoch": 2.5,
"learning_rate": 7.980000000000002e-06,
"loss": 0.9919,
"step": 400
},
{
"epoch": 2.5,
"eval_loss": 0.7921169996261597,
"eval_runtime": 33.9628,
"eval_samples_per_second": 8.362,
"eval_steps_per_second": 1.06,
"eval_wer": 0.3919392523364486,
"step": 400
},
{
"epoch": 3.12,
"learning_rate": 9.980000000000001e-06,
"loss": 0.9493,
"step": 500
},
{
"epoch": 3.12,
"eval_loss": 0.7675915360450745,
"eval_runtime": 34.6559,
"eval_samples_per_second": 8.195,
"eval_steps_per_second": 1.039,
"eval_wer": 0.3744158878504673,
"step": 500
},
{
"epoch": 3.75,
"learning_rate": 9.346666666666666e-06,
"loss": 0.9414,
"step": 600
},
{
"epoch": 3.75,
"eval_loss": 0.7247262597084045,
"eval_runtime": 34.016,
"eval_samples_per_second": 8.349,
"eval_steps_per_second": 1.058,
"eval_wer": 0.3694509345794392,
"step": 600
},
{
"epoch": 4.38,
"learning_rate": 8.68e-06,
"loss": 0.8897,
"step": 700
},
{
"epoch": 4.38,
"eval_loss": 0.720245897769928,
"eval_runtime": 35.0175,
"eval_samples_per_second": 8.11,
"eval_steps_per_second": 1.028,
"eval_wer": 0.3598130841121495,
"step": 700
},
{
"epoch": 5.0,
"learning_rate": 8.013333333333333e-06,
"loss": 0.8716,
"step": 800
},
{
"epoch": 5.0,
"eval_loss": 0.7095639109611511,
"eval_runtime": 34.2057,
"eval_samples_per_second": 8.303,
"eval_steps_per_second": 1.052,
"eval_wer": 0.35455607476635514,
"step": 800
},
{
"epoch": 5.62,
"learning_rate": 7.346666666666668e-06,
"loss": 0.8467,
"step": 900
},
{
"epoch": 5.62,
"eval_loss": 0.7023197412490845,
"eval_runtime": 34.5651,
"eval_samples_per_second": 8.216,
"eval_steps_per_second": 1.042,
"eval_wer": 0.34988317757009346,
"step": 900
},
{
"epoch": 6.25,
"learning_rate": 6.680000000000001e-06,
"loss": 0.8227,
"step": 1000
},
{
"epoch": 6.25,
"eval_loss": 0.6994471549987793,
"eval_runtime": 35.0644,
"eval_samples_per_second": 8.099,
"eval_steps_per_second": 1.027,
"eval_wer": 0.3411214953271028,
"step": 1000
},
{
"epoch": 6.88,
"learning_rate": 6.013333333333335e-06,
"loss": 0.855,
"step": 1100
},
{
"epoch": 6.88,
"eval_loss": 0.6882991790771484,
"eval_runtime": 33.9338,
"eval_samples_per_second": 8.369,
"eval_steps_per_second": 1.061,
"eval_wer": 0.3431658878504673,
"step": 1100
},
{
"epoch": 7.5,
"learning_rate": 5.346666666666667e-06,
"loss": 0.8457,
"step": 1200
},
{
"epoch": 7.5,
"eval_loss": 0.6772996187210083,
"eval_runtime": 33.7458,
"eval_samples_per_second": 8.416,
"eval_steps_per_second": 1.067,
"eval_wer": 0.34258177570093457,
"step": 1200
},
{
"epoch": 8.12,
"learning_rate": 4.680000000000001e-06,
"loss": 0.7614,
"step": 1300
},
{
"epoch": 8.12,
"eval_loss": 0.6912708282470703,
"eval_runtime": 33.9892,
"eval_samples_per_second": 8.356,
"eval_steps_per_second": 1.059,
"eval_wer": 0.33440420560747663,
"step": 1300
},
{
"epoch": 8.75,
"learning_rate": 4.013333333333334e-06,
"loss": 0.8127,
"step": 1400
},
{
"epoch": 8.75,
"eval_loss": 0.6827006340026855,
"eval_runtime": 33.6998,
"eval_samples_per_second": 8.427,
"eval_steps_per_second": 1.068,
"eval_wer": 0.33352803738317754,
"step": 1400
},
{
"epoch": 9.38,
"learning_rate": 3.346666666666667e-06,
"loss": 0.8443,
"step": 1500
},
{
"epoch": 9.38,
"eval_loss": 0.6724790930747986,
"eval_runtime": 34.0431,
"eval_samples_per_second": 8.342,
"eval_steps_per_second": 1.057,
"eval_wer": 0.33557242990654207,
"step": 1500
},
{
"epoch": 10.0,
"learning_rate": 2.68e-06,
"loss": 0.7548,
"step": 1600
},
{
"epoch": 10.0,
"eval_loss": 0.6758835315704346,
"eval_runtime": 34.9833,
"eval_samples_per_second": 8.118,
"eval_steps_per_second": 1.029,
"eval_wer": 0.3317757009345794,
"step": 1600
},
{
"epoch": 10.62,
"learning_rate": 2.0133333333333337e-06,
"loss": 0.7839,
"step": 1700
},
{
"epoch": 10.62,
"eval_loss": 0.677266001701355,
"eval_runtime": 34.4604,
"eval_samples_per_second": 8.241,
"eval_steps_per_second": 1.045,
"eval_wer": 0.3285630841121495,
"step": 1700
},
{
"epoch": 11.25,
"learning_rate": 1.3466666666666668e-06,
"loss": 0.7912,
"step": 1800
},
{
"epoch": 11.25,
"eval_loss": 0.6747676730155945,
"eval_runtime": 35.0861,
"eval_samples_per_second": 8.094,
"eval_steps_per_second": 1.026,
"eval_wer": 0.3285630841121495,
"step": 1800
},
{
"epoch": 11.88,
"learning_rate": 6.800000000000001e-07,
"loss": 0.8238,
"step": 1900
},
{
"epoch": 11.88,
"eval_loss": 0.673457145690918,
"eval_runtime": 33.7389,
"eval_samples_per_second": 8.418,
"eval_steps_per_second": 1.067,
"eval_wer": 0.32973130841121495,
"step": 1900
},
{
"epoch": 12.5,
"learning_rate": 1.3333333333333334e-08,
"loss": 0.7618,
"step": 2000
},
{
"epoch": 12.5,
"eval_loss": 0.672848641872406,
"eval_runtime": 34.5736,
"eval_samples_per_second": 8.214,
"eval_steps_per_second": 1.041,
"eval_wer": 0.3285630841121495,
"step": 2000
},
{
"epoch": 12.5,
"step": 2000,
"total_flos": 9.605345431463184e+18,
"train_loss": 0.9391711463928223,
"train_runtime": 9697.5717,
"train_samples_per_second": 3.3,
"train_steps_per_second": 0.206
}
],
"logging_steps": 100,
"max_steps": 2000,
"num_train_epochs": 13,
"save_steps": 100,
"total_flos": 9.605345431463184e+18,
"trial_name": null,
"trial_params": null
}