|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 0.036034412864285395, |
|
"global_step": 1000, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.0, |
|
"learning_rate": 9.755282581475769e-05, |
|
"loss": 10.6517, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 0.01, |
|
"learning_rate": 9.045084971874738e-05, |
|
"loss": 2.1456, |
|
"step": 200 |
|
}, |
|
{ |
|
"epoch": 0.01, |
|
"learning_rate": 7.938926261462366e-05, |
|
"loss": 1.3578, |
|
"step": 300 |
|
}, |
|
{ |
|
"epoch": 0.01, |
|
"learning_rate": 6.545084971874738e-05, |
|
"loss": 1.0559, |
|
"step": 400 |
|
}, |
|
{ |
|
"epoch": 0.02, |
|
"learning_rate": 5e-05, |
|
"loss": 0.9472, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 0.02, |
|
"learning_rate": 3.4549150281252636e-05, |
|
"loss": 0.8765, |
|
"step": 600 |
|
}, |
|
{ |
|
"epoch": 0.03, |
|
"learning_rate": 2.061073738537635e-05, |
|
"loss": 0.8402, |
|
"step": 700 |
|
}, |
|
{ |
|
"epoch": 0.03, |
|
"learning_rate": 9.549150281252633e-06, |
|
"loss": 0.8968, |
|
"step": 800 |
|
}, |
|
{ |
|
"epoch": 0.03, |
|
"learning_rate": 2.4471741852423237e-06, |
|
"loss": 0.8694, |
|
"step": 900 |
|
}, |
|
{ |
|
"epoch": 0.04, |
|
"learning_rate": 0.0, |
|
"loss": 0.8573, |
|
"step": 1000 |
|
} |
|
], |
|
"max_steps": 1000, |
|
"num_train_epochs": 1, |
|
"total_flos": 5478058819584000.0, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|