wav2vec2-large-xlsr-lithuanian / trainer_state.json
m3hrdadfi's picture
HP-tuning and optimize the model for a better performance
7536716
{
"best_metric": null,
"best_model_checkpoint": null,
"epoch": 59.994152046783626,
"global_step": 5100,
"is_hyper_param_search": false,
"is_local_process_zero": true,
"is_world_process_zero": true,
"log_history": [
{
"epoch": 3.53,
"learning_rate": 0.0003546,
"loss": 5.9795,
"step": 300
},
{
"epoch": 3.53,
"eval_loss": 2.985990524291992,
"eval_runtime": 34.848,
"eval_samples_per_second": 13.372,
"eval_wer": 1.0,
"step": 300
},
{
"epoch": 7.06,
"learning_rate": 0.0005781521739130435,
"loss": 0.8761,
"step": 600
},
{
"epoch": 7.06,
"eval_loss": 0.40344536304473877,
"eval_runtime": 34.0361,
"eval_samples_per_second": 13.691,
"eval_wer": 0.5147461234417756,
"step": 600
},
{
"epoch": 10.58,
"learning_rate": 0.0005396086956521739,
"loss": 0.1959,
"step": 900
},
{
"epoch": 10.58,
"eval_loss": 0.4274047613143921,
"eval_runtime": 34.4474,
"eval_samples_per_second": 13.528,
"eval_wer": 0.43356643356643354,
"step": 900
},
{
"epoch": 14.12,
"learning_rate": 0.0005010652173913044,
"loss": 0.1058,
"step": 1200
},
{
"epoch": 14.12,
"eval_loss": 0.4762655198574066,
"eval_runtime": 34.4865,
"eval_samples_per_second": 13.513,
"eval_wer": 0.4150197628458498,
"step": 1200
},
{
"epoch": 17.64,
"learning_rate": 0.00046252173913043483,
"loss": 0.079,
"step": 1500
},
{
"epoch": 17.64,
"eval_loss": 0.4416888356208801,
"eval_runtime": 34.5923,
"eval_samples_per_second": 13.471,
"eval_wer": 0.4159318941927638,
"step": 1500
},
{
"epoch": 21.18,
"learning_rate": 0.00042397826086956525,
"loss": 0.0568,
"step": 1800
},
{
"epoch": 21.18,
"eval_loss": 0.4843335449695587,
"eval_runtime": 33.8847,
"eval_samples_per_second": 13.753,
"eval_wer": 0.39464882943143814,
"step": 1800
},
{
"epoch": 24.7,
"learning_rate": 0.0003854347826086957,
"loss": 0.0462,
"step": 2100
},
{
"epoch": 24.7,
"eval_loss": 0.503822386264801,
"eval_runtime": 34.2122,
"eval_samples_per_second": 13.621,
"eval_wer": 0.40103374885983584,
"step": 2100
},
{
"epoch": 28.23,
"learning_rate": 0.0003468913043478261,
"loss": 0.0406,
"step": 2400
},
{
"epoch": 28.23,
"eval_loss": 0.5020607709884644,
"eval_runtime": 34.0407,
"eval_samples_per_second": 13.689,
"eval_wer": 0.39252052295530554,
"step": 2400
},
{
"epoch": 31.76,
"learning_rate": 0.00030834782608695654,
"loss": 0.0336,
"step": 2700
},
{
"epoch": 31.76,
"eval_loss": 0.5211769342422485,
"eval_runtime": 34.3633,
"eval_samples_per_second": 13.561,
"eval_wer": 0.3870477348738218,
"step": 2700
},
{
"epoch": 35.29,
"learning_rate": 0.00026980434782608695,
"loss": 0.0295,
"step": 3000
},
{
"epoch": 35.29,
"eval_loss": 0.5115832090377808,
"eval_runtime": 34.3852,
"eval_samples_per_second": 13.552,
"eval_wer": 0.38735177865612647,
"step": 3000
},
{
"epoch": 38.82,
"learning_rate": 0.00023126086956521742,
"loss": 0.0255,
"step": 3300
},
{
"epoch": 38.82,
"eval_loss": 0.555125892162323,
"eval_runtime": 34.6652,
"eval_samples_per_second": 13.443,
"eval_wer": 0.3931286105199149,
"step": 3300
},
{
"epoch": 42.35,
"learning_rate": 0.00019271739130434786,
"loss": 0.0223,
"step": 3600
},
{
"epoch": 42.35,
"eval_loss": 0.5234040021896362,
"eval_runtime": 34.224,
"eval_samples_per_second": 13.616,
"eval_wer": 0.37731833384007296,
"step": 3600
},
{
"epoch": 45.88,
"learning_rate": 0.00015417391304347827,
"loss": 0.0196,
"step": 3900
},
{
"epoch": 45.88,
"eval_loss": 0.5482075810432434,
"eval_runtime": 34.3543,
"eval_samples_per_second": 13.565,
"eval_wer": 0.38066281544542413,
"step": 3900
},
{
"epoch": 49.41,
"learning_rate": 0.00011563043478260871,
"loss": 0.0167,
"step": 4200
},
{
"epoch": 49.41,
"eval_loss": 0.5395901203155518,
"eval_runtime": 34.2391,
"eval_samples_per_second": 13.61,
"eval_wer": 0.3757981149285497,
"step": 4200
},
{
"epoch": 52.94,
"learning_rate": 7.708695652173913e-05,
"loss": 0.0158,
"step": 4500
},
{
"epoch": 52.94,
"eval_loss": 0.537598729133606,
"eval_runtime": 60.6582,
"eval_samples_per_second": 7.682,
"eval_wer": 0.3612040133779264,
"step": 4500
},
{
"epoch": 56.47,
"learning_rate": 3.854347826086957e-05,
"loss": 0.0135,
"step": 4800
},
{
"epoch": 56.47,
"eval_loss": 0.5266188979148865,
"eval_runtime": 34.8561,
"eval_samples_per_second": 13.369,
"eval_wer": 0.35177865612648224,
"step": 4800
},
{
"epoch": 59.99,
"learning_rate": 0.0,
"loss": 0.0114,
"step": 5100
},
{
"epoch": 59.99,
"eval_loss": 0.5329256653785706,
"eval_runtime": 34.148,
"eval_samples_per_second": 13.646,
"eval_wer": 0.3505624809972636,
"step": 5100
},
{
"epoch": 59.99,
"step": 5100,
"total_flos": 4.584946221172683e+19,
"train_runtime": 31139.6077,
"train_samples_per_second": 0.164
}
],
"max_steps": 5100,
"num_train_epochs": 60,
"total_flos": 4.584946221172683e+19,
"trial_name": null,
"trial_params": null
}