|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 10.0, |
|
"global_step": 580, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 8.62, |
|
"learning_rate": 6.896551724137932e-06, |
|
"loss": 2.0521, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"step": 580, |
|
"total_flos": 2134337526558720.0, |
|
"train_loss": 2.016565941120016, |
|
"train_runtime": 632.1099, |
|
"train_samples_per_second": 3.623, |
|
"train_steps_per_second": 0.918 |
|
} |
|
], |
|
"max_steps": 580, |
|
"num_train_epochs": 10, |
|
"total_flos": 2134337526558720.0, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|