|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 10.0, |
|
"global_step": 220, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 4.55, |
|
"learning_rate": 1.8181818181818182e-05, |
|
"loss": 4.5806, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 9.09, |
|
"learning_rate": 3.0303030303030305e-06, |
|
"loss": 2.5073, |
|
"step": 200 |
|
}, |
|
{ |
|
"epoch": 10.0, |
|
"step": 220, |
|
"total_flos": 396870987244656.0, |
|
"train_loss": 3.405086222561923, |
|
"train_runtime": 100.589, |
|
"train_samples_per_second": 25.45, |
|
"train_steps_per_second": 2.187 |
|
} |
|
], |
|
"max_steps": 220, |
|
"num_train_epochs": 10, |
|
"total_flos": 396870987244656.0, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|