|
{ |
|
"best_metric": 0.3382669687271118, |
|
"best_model_checkpoint": "../../saves/LLaMA3-70B-qlora-bnb/lora/sft/checkpoint-200", |
|
"epoch": 2.9698996655518393, |
|
"eval_steps": 100, |
|
"global_step": 222, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.13377926421404682, |
|
"grad_norm": 34.1706428527832, |
|
"learning_rate": 4.9977474155117045e-05, |
|
"loss": 12.2081, |
|
"step": 10 |
|
}, |
|
{ |
|
"epoch": 0.26755852842809363, |
|
"grad_norm": NaN, |
|
"learning_rate": 4.9640397758692715e-05, |
|
"loss": 9.9449, |
|
"step": 20 |
|
}, |
|
{ |
|
"epoch": 0.4013377926421405, |
|
"grad_norm": 28.862974166870117, |
|
"learning_rate": 4.8798182638228166e-05, |
|
"loss": 6.0546, |
|
"step": 30 |
|
}, |
|
{ |
|
"epoch": 0.5351170568561873, |
|
"grad_norm": 25.968364715576172, |
|
"learning_rate": 4.748018037239592e-05, |
|
"loss": 1.4872, |
|
"step": 40 |
|
}, |
|
{ |
|
"epoch": 0.6688963210702341, |
|
"grad_norm": 18.545757293701172, |
|
"learning_rate": 4.571274123109606e-05, |
|
"loss": 0.378, |
|
"step": 50 |
|
}, |
|
{ |
|
"epoch": 0.802675585284281, |
|
"grad_norm": 13.884383201599121, |
|
"learning_rate": 4.353120088833501e-05, |
|
"loss": 0.3846, |
|
"step": 60 |
|
}, |
|
{ |
|
"epoch": 0.9364548494983278, |
|
"grad_norm": 9.545388221740723, |
|
"learning_rate": 4.0979173970824626e-05, |
|
"loss": 0.3639, |
|
"step": 70 |
|
}, |
|
{ |
|
"epoch": 1.0702341137123745, |
|
"grad_norm": 16.583349227905273, |
|
"learning_rate": 3.8107682088930794e-05, |
|
"loss": 0.3366, |
|
"step": 80 |
|
}, |
|
{ |
|
"epoch": 1.2040133779264215, |
|
"grad_norm": 19.358070373535156, |
|
"learning_rate": 3.497413378288541e-05, |
|
"loss": 0.3343, |
|
"step": 90 |
|
}, |
|
{ |
|
"epoch": 1.3377926421404682, |
|
"grad_norm": 9.057594299316406, |
|
"learning_rate": 3.164117677777191e-05, |
|
"loss": 0.3548, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 1.3377926421404682, |
|
"eval_loss": 0.3387661278247833, |
|
"eval_runtime": 235.819, |
|
"eval_samples_per_second": 0.284, |
|
"eval_steps_per_second": 0.284, |
|
"step": 100 |
|
}, |
|
{ |
|
"epoch": 1.471571906354515, |
|
"grad_norm": 10.804256439208984, |
|
"learning_rate": 2.8175445493671972e-05, |
|
"loss": 0.4054, |
|
"step": 110 |
|
}, |
|
{ |
|
"epoch": 1.605351170568562, |
|
"grad_norm": 9.787694931030273, |
|
"learning_rate": 2.4646228851480956e-05, |
|
"loss": 0.3489, |
|
"step": 120 |
|
}, |
|
{ |
|
"epoch": 1.7391304347826086, |
|
"grad_norm": 12.635400772094727, |
|
"learning_rate": 2.1124085008395054e-05, |
|
"loss": 0.3078, |
|
"step": 130 |
|
}, |
|
{ |
|
"epoch": 1.8729096989966556, |
|
"grad_norm": 13.208198547363281, |
|
"learning_rate": 1.7679430718086243e-05, |
|
"loss": 0.3377, |
|
"step": 140 |
|
}, |
|
{ |
|
"epoch": 2.0066889632107023, |
|
"grad_norm": 7.031893730163574, |
|
"learning_rate": 1.4381133517898804e-05, |
|
"loss": 0.2855, |
|
"step": 150 |
|
}, |
|
{ |
|
"epoch": 2.140468227424749, |
|
"grad_norm": 10.516304969787598, |
|
"learning_rate": 1.1295134888882258e-05, |
|
"loss": 0.2867, |
|
"step": 160 |
|
}, |
|
{ |
|
"epoch": 2.274247491638796, |
|
"grad_norm": 6.793305397033691, |
|
"learning_rate": 8.483131915247968e-06, |
|
"loss": 0.3044, |
|
"step": 170 |
|
}, |
|
{ |
|
"epoch": 2.408026755852843, |
|
"grad_norm": 10.898367881774902, |
|
"learning_rate": 6.001343800282569e-06, |
|
"loss": 0.28, |
|
"step": 180 |
|
}, |
|
{ |
|
"epoch": 2.5418060200668897, |
|
"grad_norm": 6.905026912689209, |
|
"learning_rate": 3.8993878992512415e-06, |
|
"loss": 0.2838, |
|
"step": 190 |
|
}, |
|
{ |
|
"epoch": 2.6755852842809364, |
|
"grad_norm": 11.015267372131348, |
|
"learning_rate": 2.219287740296605e-06, |
|
"loss": 0.271, |
|
"step": 200 |
|
}, |
|
{ |
|
"epoch": 2.6755852842809364, |
|
"eval_loss": 0.3382669687271118, |
|
"eval_runtime": 235.7263, |
|
"eval_samples_per_second": 0.284, |
|
"eval_steps_per_second": 0.284, |
|
"step": 200 |
|
}, |
|
{ |
|
"epoch": 2.809364548494983, |
|
"grad_norm": 9.907255172729492, |
|
"learning_rate": 9.946328655577624e-07, |
|
"loss": 0.2943, |
|
"step": 210 |
|
}, |
|
{ |
|
"epoch": 2.94314381270903, |
|
"grad_norm": 31.69417381286621, |
|
"learning_rate": 2.499072894559057e-07, |
|
"loss": 0.2605, |
|
"step": 220 |
|
}, |
|
{ |
|
"epoch": 2.9698996655518393, |
|
"step": 222, |
|
"total_flos": 5.059970235227701e+18, |
|
"train_loss": 1.6024731107660242, |
|
"train_runtime": 15122.115, |
|
"train_samples_per_second": 0.119, |
|
"train_steps_per_second": 0.015 |
|
} |
|
], |
|
"logging_steps": 10, |
|
"max_steps": 222, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 3, |
|
"save_steps": 100, |
|
"total_flos": 5.059970235227701e+18, |
|
"train_batch_size": 1, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|