|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 5.0, |
|
"eval_steps": 500, |
|
"global_step": 4460, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 1.0, |
|
"grad_norm": 0.6163806319236755, |
|
"learning_rate": 0.0008, |
|
"loss": 0.8212, |
|
"step": 892 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_gen_len": 1.0, |
|
"eval_loss": 0.5616050362586975, |
|
"eval_rouge1": 0.77, |
|
"eval_rouge2": 0.0, |
|
"eval_rougeL": 0.7703, |
|
"eval_rougeLsum": 0.7685, |
|
"eval_runtime": 38.5385, |
|
"eval_samples_per_second": 19.435, |
|
"eval_steps_per_second": 0.623, |
|
"step": 892 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"grad_norm": 0.7340933680534363, |
|
"learning_rate": 0.0006, |
|
"loss": 0.6185, |
|
"step": 1784 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_gen_len": 1.0, |
|
"eval_loss": 0.5592873096466064, |
|
"eval_rouge1": 0.7457, |
|
"eval_rouge2": 0.0, |
|
"eval_rougeL": 0.7436, |
|
"eval_rougeLsum": 0.7441, |
|
"eval_runtime": 38.0734, |
|
"eval_samples_per_second": 19.673, |
|
"eval_steps_per_second": 0.63, |
|
"step": 1784 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"grad_norm": 0.5377846360206604, |
|
"learning_rate": 0.0004, |
|
"loss": 0.5776, |
|
"step": 2676 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_gen_len": 1.0, |
|
"eval_loss": 0.5282897353172302, |
|
"eval_rouge1": 0.7664, |
|
"eval_rouge2": 0.0, |
|
"eval_rougeL": 0.7649, |
|
"eval_rougeLsum": 0.7639, |
|
"eval_runtime": 38.0432, |
|
"eval_samples_per_second": 19.688, |
|
"eval_steps_per_second": 0.631, |
|
"step": 2676 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"grad_norm": 0.627534806728363, |
|
"learning_rate": 0.0002, |
|
"loss": 0.5535, |
|
"step": 3568 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_gen_len": 1.0, |
|
"eval_loss": 0.5193020105361938, |
|
"eval_rouge1": 0.7475, |
|
"eval_rouge2": 0.0, |
|
"eval_rougeL": 0.747, |
|
"eval_rougeLsum": 0.745, |
|
"eval_runtime": 37.7969, |
|
"eval_samples_per_second": 19.816, |
|
"eval_steps_per_second": 0.635, |
|
"step": 3568 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"grad_norm": 0.7363429069519043, |
|
"learning_rate": 0.0, |
|
"loss": 0.5357, |
|
"step": 4460 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_gen_len": 1.0, |
|
"eval_loss": 0.5139681696891785, |
|
"eval_rouge1": 0.7665, |
|
"eval_rouge2": 0.0, |
|
"eval_rougeL": 0.7672, |
|
"eval_rougeLsum": 0.7666, |
|
"eval_runtime": 38.1607, |
|
"eval_samples_per_second": 19.628, |
|
"eval_steps_per_second": 0.629, |
|
"step": 4460 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"step": 4460, |
|
"total_flos": 4.90272982695936e+16, |
|
"train_loss": 0.6212816452231642, |
|
"train_runtime": 3932.3471, |
|
"train_samples_per_second": 18.135, |
|
"train_steps_per_second": 1.134 |
|
} |
|
], |
|
"logging_steps": 500, |
|
"max_steps": 4460, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 5, |
|
"save_steps": 500, |
|
"total_flos": 4.90272982695936e+16, |
|
"train_batch_size": 16, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|