|
{ |
|
"best_metric": 0.926605504587156, |
|
"best_model_checkpoint": "electra-distilled-sst\\run-6\\checkpoint-2108", |
|
"epoch": 6.0, |
|
"eval_steps": 500, |
|
"global_step": 3162, |
|
"is_hyper_param_search": true, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 1.0, |
|
"grad_norm": 17.92670440673828, |
|
"learning_rate": 2.010330074098376e-05, |
|
"loss": 1.6952, |
|
"step": 527 |
|
}, |
|
{ |
|
"epoch": 1.0, |
|
"eval_accuracy": 0.8910550458715596, |
|
"eval_loss": 0.9234904050827026, |
|
"eval_runtime": 1.446, |
|
"eval_samples_per_second": 603.052, |
|
"eval_steps_per_second": 4.841, |
|
"step": 527 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"grad_norm": 43.87211990356445, |
|
"learning_rate": 1.759575125325375e-05, |
|
"loss": 0.8731, |
|
"step": 1054 |
|
}, |
|
{ |
|
"epoch": 2.0, |
|
"eval_accuracy": 0.9059633027522935, |
|
"eval_loss": 0.770187258720398, |
|
"eval_runtime": 1.4489, |
|
"eval_samples_per_second": 601.851, |
|
"eval_steps_per_second": 4.831, |
|
"step": 1054 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"grad_norm": 13.310898780822754, |
|
"learning_rate": 1.508343456117444e-05, |
|
"loss": 0.6622, |
|
"step": 1581 |
|
}, |
|
{ |
|
"epoch": 3.0, |
|
"eval_accuracy": 0.9139908256880734, |
|
"eval_loss": 0.7737818956375122, |
|
"eval_runtime": 1.4497, |
|
"eval_samples_per_second": 601.516, |
|
"eval_steps_per_second": 4.829, |
|
"step": 1581 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"grad_norm": 12.57313060760498, |
|
"learning_rate": 1.2571117869095134e-05, |
|
"loss": 0.5443, |
|
"step": 2108 |
|
}, |
|
{ |
|
"epoch": 4.0, |
|
"eval_accuracy": 0.926605504587156, |
|
"eval_loss": 0.7234447598457336, |
|
"eval_runtime": 1.451, |
|
"eval_samples_per_second": 600.977, |
|
"eval_steps_per_second": 4.824, |
|
"step": 2108 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"grad_norm": 28.64960289001465, |
|
"learning_rate": 1.0058801177015824e-05, |
|
"loss": 0.48, |
|
"step": 2635 |
|
}, |
|
{ |
|
"epoch": 5.0, |
|
"eval_accuracy": 0.9174311926605505, |
|
"eval_loss": 0.7743226289749146, |
|
"eval_runtime": 1.4478, |
|
"eval_samples_per_second": 602.308, |
|
"eval_steps_per_second": 4.835, |
|
"step": 2635 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"grad_norm": 10.863966941833496, |
|
"learning_rate": 7.546484484936517e-06, |
|
"loss": 0.4283, |
|
"step": 3162 |
|
}, |
|
{ |
|
"epoch": 6.0, |
|
"eval_accuracy": 0.9220183486238532, |
|
"eval_loss": 0.7562010288238525, |
|
"eval_runtime": 1.4492, |
|
"eval_samples_per_second": 601.691, |
|
"eval_steps_per_second": 4.83, |
|
"step": 3162 |
|
} |
|
], |
|
"logging_steps": 500, |
|
"max_steps": 4743, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 9, |
|
"save_steps": 500, |
|
"total_flos": 1123721737396596.0, |
|
"train_batch_size": 128, |
|
"trial_name": null, |
|
"trial_params": { |
|
"alpha": 0.7277023096419001, |
|
"learning_rate": 2.261085022871377e-05, |
|
"num_train_epochs": 9, |
|
"temperature": 29 |
|
} |
|
} |
|
|