|
{ |
|
"best_metric": 0.7846666666666666, |
|
"best_model_checkpoint": "Bert_tinybert-distilled/run-13/checkpoint-344", |
|
"epoch": 2.0, |
|
"eval_steps": 500, |
|
"global_step": 344, |
|
"is_hyper_param_search": true, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 1.0, |
|
"grad_norm": 1.6519217491149902, |
|
"learning_rate": 0.0003489409909619128, |
|
"loss": 0.6241, |
|
"step": 172 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_AUC": 0.8201937777777778, |
|
"eval_BAC": 0.738, |
|
"eval_MCC": 0.4761528166708038, |
|
"eval_PR_AUC": 0.7788333782261532, |
|
"eval_accuracy": 0.738, |
|
"eval_loss": 0.5376982688903809, |
|
"eval_runtime": 6.2979, |
|
"eval_samples_per_second": 238.174, |
|
"eval_steps_per_second": 7.463, |
|
"step": 172 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"grad_norm": 2.3186373710632324, |
|
"learning_rate": 0.0003101697697439225, |
|
"loss": 0.5634, |
|
"step": 344 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_AUC": 0.8520746666666666, |
|
"eval_BAC": 0.7846666666666666, |
|
"eval_MCC": 0.5843621161716221, |
|
"eval_PR_AUC": 0.8014637687014566, |
|
"eval_accuracy": 0.7846666666666666, |
|
"eval_loss": 0.4982443153858185, |
|
"eval_runtime": 6.2909, |
|
"eval_samples_per_second": 238.439, |
|
"eval_steps_per_second": 7.471, |
|
"step": 344 |
|
} |
|
], |
|
"logging_steps": 500, |
|
"max_steps": 1720, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 10, |
|
"save_steps": 500, |
|
"stateful_callbacks": { |
|
"TrainerControl": { |
|
"args": { |
|
"should_epoch_stop": false, |
|
"should_evaluate": false, |
|
"should_log": false, |
|
"should_save": true, |
|
"should_training_stop": false |
|
}, |
|
"attributes": {} |
|
} |
|
}, |
|
"total_flos": 900755460000.0, |
|
"train_batch_size": 32, |
|
"trial_name": null, |
|
"trial_params": { |
|
"alpha": 0.9793694957428477, |
|
"learning_rate": 0.0003877122121799031, |
|
"num_train_epochs": 10, |
|
"temperature": 8 |
|
} |
|
} |
|
|