|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 7.372654155495979, |
|
"eval_steps": 500, |
|
"global_step": 5500, |
|
"is_hyper_param_search": true, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.6702412868632708, |
|
"grad_norm": 13.804513931274414, |
|
"learning_rate": 1.2284160000241888e-05, |
|
"loss": 0.5698, |
|
"step": 500 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_accuracy": 0.7532685399055481, |
|
"eval_loss": 0.4876743257045746, |
|
"eval_runtime": 9.5369, |
|
"eval_samples_per_second": 312.785, |
|
"eval_steps_per_second": 19.608, |
|
"step": 746 |
|
}, |
|
{ |
|
"epoch": 1.3404825737265416, |
|
"grad_norm": 23.23216438293457, |
|
"learning_rate": 1.1401677241603822e-05, |
|
"loss": 0.4247, |
|
"step": 1000 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_accuracy": 0.7777405381202698, |
|
"eval_loss": 0.5234329700469971, |
|
"eval_runtime": 9.4342, |
|
"eval_samples_per_second": 316.19, |
|
"eval_steps_per_second": 19.821, |
|
"step": 1492 |
|
}, |
|
{ |
|
"epoch": 2.0107238605898123, |
|
"grad_norm": 6.197817325592041, |
|
"learning_rate": 1.0519194482965755e-05, |
|
"loss": 0.3313, |
|
"step": 1500 |
|
}, |
|
{ |
|
"epoch": 2.680965147453083, |
|
"grad_norm": 20.681987762451172, |
|
"learning_rate": 9.63671172432769e-06, |
|
"loss": 0.1575, |
|
"step": 2000 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_accuracy": 0.7740529775619507, |
|
"eval_loss": 0.6654106378555298, |
|
"eval_runtime": 9.4207, |
|
"eval_samples_per_second": 316.642, |
|
"eval_steps_per_second": 19.85, |
|
"step": 2238 |
|
}, |
|
{ |
|
"epoch": 3.351206434316354, |
|
"grad_norm": 0.6577324271202087, |
|
"learning_rate": 8.754228965689622e-06, |
|
"loss": 0.1178, |
|
"step": 2500 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_accuracy": 0.7763996124267578, |
|
"eval_loss": 1.094183087348938, |
|
"eval_runtime": 9.4168, |
|
"eval_samples_per_second": 316.775, |
|
"eval_steps_per_second": 19.858, |
|
"step": 2984 |
|
}, |
|
{ |
|
"epoch": 4.021447721179625, |
|
"grad_norm": 5.022602558135986, |
|
"learning_rate": 7.871746207051555e-06, |
|
"loss": 0.0777, |
|
"step": 3000 |
|
}, |
|
{ |
|
"epoch": 4.6916890080428955, |
|
"grad_norm": 0.7826739549636841, |
|
"learning_rate": 6.989263448413488e-06, |
|
"loss": 0.044, |
|
"step": 3500 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_accuracy": 0.782433807849884, |
|
"eval_loss": 1.2965056896209717, |
|
"eval_runtime": 9.3869, |
|
"eval_samples_per_second": 317.783, |
|
"eval_steps_per_second": 19.921, |
|
"step": 3730 |
|
}, |
|
{ |
|
"epoch": 5.361930294906166, |
|
"grad_norm": 1.7349833250045776, |
|
"learning_rate": 6.106780689775422e-06, |
|
"loss": 0.0345, |
|
"step": 4000 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_accuracy": 0.7861213684082031, |
|
"eval_loss": 1.1750046014785767, |
|
"eval_runtime": 9.4116, |
|
"eval_samples_per_second": 316.949, |
|
"eval_steps_per_second": 19.869, |
|
"step": 4476 |
|
}, |
|
{ |
|
"epoch": 6.032171581769437, |
|
"grad_norm": 16.81600570678711, |
|
"learning_rate": 5.224297931137354e-06, |
|
"loss": 0.0227, |
|
"step": 4500 |
|
}, |
|
{ |
|
"epoch": 6.702412868632708, |
|
"grad_norm": 0.004451041109859943, |
|
"learning_rate": 4.341815172499288e-06, |
|
"loss": 0.0164, |
|
"step": 5000 |
|
}, |
|
{ |
|
"epoch": 7.0, |
|
"eval_accuracy": 0.7827690243721008, |
|
"eval_loss": 1.4659219980239868, |
|
"eval_runtime": 9.615, |
|
"eval_samples_per_second": 310.243, |
|
"eval_steps_per_second": 19.449, |
|
"step": 5222 |
|
}, |
|
{ |
|
"epoch": 7.372654155495979, |
|
"grad_norm": 4.934184551239014, |
|
"learning_rate": 3.4593324138612214e-06, |
|
"loss": 0.0136, |
|
"step": 5500 |
|
} |
|
], |
|
"logging_steps": 500, |
|
"max_steps": 7460, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 10, |
|
"save_steps": 500, |
|
"total_flos": 1.4332422519482328e+16, |
|
"train_batch_size": 16, |
|
"trial_name": null, |
|
"trial_params": { |
|
"learning_rate": 1.3166642758879955e-05, |
|
"per_device_train_batch_size": 16 |
|
} |
|
} |
|
|