|
{ |
|
"best_metric": 0.5458308458328247, |
|
"best_model_checkpoint": "bert_base_lda_50_v1_qnli/checkpoint-1230", |
|
"epoch": 8.0, |
|
"eval_steps": 500, |
|
"global_step": 3280, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 1.0, |
|
"grad_norm": 2.3502628803253174, |
|
"learning_rate": 4.9e-05, |
|
"loss": 0.6503, |
|
"step": 410 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_accuracy": 0.6505583012996522, |
|
"eval_loss": 0.6147003173828125, |
|
"eval_runtime": 3.5547, |
|
"eval_samples_per_second": 1536.826, |
|
"eval_steps_per_second": 6.189, |
|
"step": 410 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"grad_norm": 3.2205982208251953, |
|
"learning_rate": 4.8e-05, |
|
"loss": 0.5535, |
|
"step": 820 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_accuracy": 0.7155409115870401, |
|
"eval_loss": 0.5496997237205505, |
|
"eval_runtime": 3.5498, |
|
"eval_samples_per_second": 1538.962, |
|
"eval_steps_per_second": 6.198, |
|
"step": 820 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"grad_norm": 3.7460732460021973, |
|
"learning_rate": 4.7e-05, |
|
"loss": 0.4425, |
|
"step": 1230 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_accuracy": 0.7367746659344683, |
|
"eval_loss": 0.5458308458328247, |
|
"eval_runtime": 3.5648, |
|
"eval_samples_per_second": 1532.48, |
|
"eval_steps_per_second": 6.171, |
|
"step": 1230 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"grad_norm": 6.930105209350586, |
|
"learning_rate": 4.600000000000001e-05, |
|
"loss": 0.3413, |
|
"step": 1640 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_accuracy": 0.733479773018488, |
|
"eval_loss": 0.6162947416305542, |
|
"eval_runtime": 3.5841, |
|
"eval_samples_per_second": 1524.219, |
|
"eval_steps_per_second": 6.138, |
|
"step": 1640 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"grad_norm": 7.669921398162842, |
|
"learning_rate": 4.5e-05, |
|
"loss": 0.2508, |
|
"step": 2050 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_accuracy": 0.7384221123924584, |
|
"eval_loss": 0.6988028883934021, |
|
"eval_runtime": 3.562, |
|
"eval_samples_per_second": 1533.698, |
|
"eval_steps_per_second": 6.176, |
|
"step": 2050 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"grad_norm": 8.73188304901123, |
|
"learning_rate": 4.4000000000000006e-05, |
|
"loss": 0.1792, |
|
"step": 2460 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_accuracy": 0.7338458722313747, |
|
"eval_loss": 0.8530865907669067, |
|
"eval_runtime": 3.5879, |
|
"eval_samples_per_second": 1522.601, |
|
"eval_steps_per_second": 6.132, |
|
"step": 2460 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"grad_norm": 6.224167346954346, |
|
"learning_rate": 4.3e-05, |
|
"loss": 0.1282, |
|
"step": 2870 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"eval_accuracy": 0.7285374336445176, |
|
"eval_loss": 0.9750887155532837, |
|
"eval_runtime": 3.585, |
|
"eval_samples_per_second": 1523.83, |
|
"eval_steps_per_second": 6.137, |
|
"step": 2870 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"grad_norm": 6.511787414550781, |
|
"learning_rate": 4.2e-05, |
|
"loss": 0.0942, |
|
"step": 3280 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"eval_accuracy": 0.727622185612301, |
|
"eval_loss": 1.022977352142334, |
|
"eval_runtime": 3.522, |
|
"eval_samples_per_second": 1551.101, |
|
"eval_steps_per_second": 6.246, |
|
"step": 3280 |
|
}, |
|
{ |
|
"epoch": 8.0, |
|
"step": 3280, |
|
"total_flos": 1.1023616508628992e+17, |
|
"train_loss": 0.33002244205009645, |
|
"train_runtime": 1412.0901, |
|
"train_samples_per_second": 3708.793, |
|
"train_steps_per_second": 14.517 |
|
} |
|
], |
|
"logging_steps": 1, |
|
"max_steps": 20500, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 50, |
|
"save_steps": 500, |
|
"stateful_callbacks": { |
|
"EarlyStoppingCallback": { |
|
"args": { |
|
"early_stopping_patience": 5, |
|
"early_stopping_threshold": 0.0 |
|
}, |
|
"attributes": { |
|
"early_stopping_patience_counter": 5 |
|
} |
|
}, |
|
"TrainerControl": { |
|
"args": { |
|
"should_epoch_stop": false, |
|
"should_evaluate": false, |
|
"should_log": false, |
|
"should_save": true, |
|
"should_training_stop": true |
|
}, |
|
"attributes": {} |
|
} |
|
}, |
|
"total_flos": 1.1023616508628992e+17, |
|
"train_batch_size": 256, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|