|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 5.0, |
|
"eval_steps": 500, |
|
"global_step": 4470, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 1.0, |
|
"grad_norm": 2.3870694637298584, |
|
"learning_rate": 0.0008, |
|
"loss": 0.781, |
|
"step": 894 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_gen_len": 1.0, |
|
"eval_loss": 0.5142351388931274, |
|
"eval_rouge1": 0.6277, |
|
"eval_rouge2": 0.0, |
|
"eval_rougeL": 0.6309, |
|
"eval_rougeLsum": 0.628, |
|
"eval_runtime": 35.7617, |
|
"eval_samples_per_second": 20.888, |
|
"eval_steps_per_second": 0.671, |
|
"step": 894 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"grad_norm": 0.8088672757148743, |
|
"learning_rate": 0.0006, |
|
"loss": 0.6103, |
|
"step": 1788 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_gen_len": 1.0, |
|
"eval_loss": 0.48342767357826233, |
|
"eval_rouge1": 0.6516, |
|
"eval_rouge2": 0.0, |
|
"eval_rougeL": 0.6528, |
|
"eval_rougeLsum": 0.6502, |
|
"eval_runtime": 36.2393, |
|
"eval_samples_per_second": 20.613, |
|
"eval_steps_per_second": 0.662, |
|
"step": 1788 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"grad_norm": 1.094125747680664, |
|
"learning_rate": 0.0004, |
|
"loss": 0.5612, |
|
"step": 2682 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_gen_len": 1.0, |
|
"eval_loss": 0.47685039043426514, |
|
"eval_rouge1": 0.6537, |
|
"eval_rouge2": 0.0, |
|
"eval_rougeL": 0.654, |
|
"eval_rougeLsum": 0.6571, |
|
"eval_runtime": 35.8197, |
|
"eval_samples_per_second": 20.854, |
|
"eval_steps_per_second": 0.67, |
|
"step": 2682 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"grad_norm": 0.5754472613334656, |
|
"learning_rate": 0.0002, |
|
"loss": 0.5242, |
|
"step": 3576 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_gen_len": 1.0, |
|
"eval_loss": 0.4704252779483795, |
|
"eval_rouge1": 0.6285, |
|
"eval_rouge2": 0.0, |
|
"eval_rougeL": 0.6294, |
|
"eval_rougeLsum": 0.6312, |
|
"eval_runtime": 36.0416, |
|
"eval_samples_per_second": 20.726, |
|
"eval_steps_per_second": 0.666, |
|
"step": 3576 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"grad_norm": 0.8965361714363098, |
|
"learning_rate": 0.0, |
|
"loss": 0.4933, |
|
"step": 4470 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_gen_len": 1.0, |
|
"eval_loss": 0.4665641784667969, |
|
"eval_rouge1": 0.6446, |
|
"eval_rouge2": 0.0, |
|
"eval_rougeL": 0.6437, |
|
"eval_rougeLsum": 0.6455, |
|
"eval_runtime": 38.3287, |
|
"eval_samples_per_second": 19.489, |
|
"eval_steps_per_second": 0.626, |
|
"step": 4470 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"step": 4470, |
|
"total_flos": 4.9318600605696e+16, |
|
"train_loss": 0.5939977931762969, |
|
"train_runtime": 3736.9827, |
|
"train_samples_per_second": 19.12, |
|
"train_steps_per_second": 1.196 |
|
} |
|
], |
|
"logging_steps": 500, |
|
"max_steps": 4470, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 5, |
|
"save_steps": 500, |
|
"total_flos": 4.9318600605696e+16, |
|
"train_batch_size": 16, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|