|
{ |
|
"best_metric": 2.672280788421631, |
|
"best_model_checkpoint": "output/bring-me-the-horizon/checkpoint-52", |
|
"epoch": 2.0, |
|
"global_step": 52, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.19, |
|
"learning_rate": 0.00012505669320030482, |
|
"loss": 2.9107, |
|
"step": 5 |
|
}, |
|
{ |
|
"epoch": 0.38, |
|
"learning_rate": 9.292589525111794e-05, |
|
"loss": 2.6407, |
|
"step": 10 |
|
}, |
|
{ |
|
"epoch": 0.58, |
|
"learning_rate": 5.218294542987356e-05, |
|
"loss": 2.5181, |
|
"step": 15 |
|
}, |
|
{ |
|
"epoch": 0.77, |
|
"learning_rate": 1.725216267546246e-05, |
|
"loss": 2.4529, |
|
"step": 20 |
|
}, |
|
{ |
|
"epoch": 0.96, |
|
"learning_rate": 5.001712368734975e-07, |
|
"loss": 2.466, |
|
"step": 25 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_loss": 2.72078013420105, |
|
"eval_runtime": 0.543, |
|
"eval_samples_per_second": 75.509, |
|
"eval_steps_per_second": 11.05, |
|
"step": 26 |
|
}, |
|
{ |
|
"epoch": 1.15, |
|
"learning_rate": 7.857716640189785e-06, |
|
"loss": 2.5427, |
|
"step": 30 |
|
}, |
|
{ |
|
"epoch": 1.35, |
|
"learning_rate": 3.671999039779749e-05, |
|
"loss": 2.3173, |
|
"step": 35 |
|
}, |
|
{ |
|
"epoch": 1.54, |
|
"learning_rate": 7.686881626551516e-05, |
|
"loss": 2.2888, |
|
"step": 40 |
|
}, |
|
{ |
|
"epoch": 1.73, |
|
"learning_rate": 0.00011409021435531856, |
|
"loss": 2.3614, |
|
"step": 45 |
|
}, |
|
{ |
|
"epoch": 1.92, |
|
"learning_rate": 0.00013520660867542716, |
|
"loss": 2.1993, |
|
"step": 50 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_loss": 2.672280788421631, |
|
"eval_runtime": 0.5476, |
|
"eval_samples_per_second": 74.866, |
|
"eval_steps_per_second": 10.956, |
|
"step": 52 |
|
} |
|
], |
|
"max_steps": 1300, |
|
"num_train_epochs": 50, |
|
"total_flos": 52780990464000.0, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|