|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 0.003514629645901063, |
|
"eval_steps": 10, |
|
"global_step": 50, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 7.029259291802126e-05, |
|
"eval_loss": NaN, |
|
"eval_runtime": 224.3798, |
|
"eval_samples_per_second": 26.696, |
|
"eval_steps_per_second": 13.348, |
|
"step": 1 |
|
}, |
|
{ |
|
"epoch": 0.00035146296459010633, |
|
"grad_norm": NaN, |
|
"learning_rate": 5e-05, |
|
"loss": 0.0, |
|
"step": 5 |
|
}, |
|
{ |
|
"epoch": 0.0007029259291802127, |
|
"grad_norm": NaN, |
|
"learning_rate": 0.0001, |
|
"loss": 0.0, |
|
"step": 10 |
|
}, |
|
{ |
|
"epoch": 0.0007029259291802127, |
|
"eval_loss": NaN, |
|
"eval_runtime": 222.0838, |
|
"eval_samples_per_second": 26.972, |
|
"eval_steps_per_second": 13.486, |
|
"step": 10 |
|
}, |
|
{ |
|
"epoch": 0.001054388893770319, |
|
"grad_norm": NaN, |
|
"learning_rate": 9.619397662556435e-05, |
|
"loss": 0.0, |
|
"step": 15 |
|
}, |
|
{ |
|
"epoch": 0.0014058518583604253, |
|
"grad_norm": NaN, |
|
"learning_rate": 8.535533905932738e-05, |
|
"loss": 0.0, |
|
"step": 20 |
|
}, |
|
{ |
|
"epoch": 0.0014058518583604253, |
|
"eval_loss": NaN, |
|
"eval_runtime": 223.2606, |
|
"eval_samples_per_second": 26.83, |
|
"eval_steps_per_second": 13.415, |
|
"step": 20 |
|
}, |
|
{ |
|
"epoch": 0.0017573148229505316, |
|
"grad_norm": NaN, |
|
"learning_rate": 6.91341716182545e-05, |
|
"loss": 0.0, |
|
"step": 25 |
|
}, |
|
{ |
|
"epoch": 0.002108777787540638, |
|
"grad_norm": NaN, |
|
"learning_rate": 5e-05, |
|
"loss": 0.0, |
|
"step": 30 |
|
}, |
|
{ |
|
"epoch": 0.002108777787540638, |
|
"eval_loss": NaN, |
|
"eval_runtime": 222.7897, |
|
"eval_samples_per_second": 26.886, |
|
"eval_steps_per_second": 13.443, |
|
"step": 30 |
|
}, |
|
{ |
|
"epoch": 0.0024602407521307442, |
|
"grad_norm": NaN, |
|
"learning_rate": 3.086582838174551e-05, |
|
"loss": 0.0, |
|
"step": 35 |
|
}, |
|
{ |
|
"epoch": 0.0028117037167208507, |
|
"grad_norm": NaN, |
|
"learning_rate": 1.4644660940672627e-05, |
|
"loss": 0.0, |
|
"step": 40 |
|
}, |
|
{ |
|
"epoch": 0.0028117037167208507, |
|
"eval_loss": NaN, |
|
"eval_runtime": 221.815, |
|
"eval_samples_per_second": 27.004, |
|
"eval_steps_per_second": 13.502, |
|
"step": 40 |
|
}, |
|
{ |
|
"epoch": 0.0031631666813109567, |
|
"grad_norm": NaN, |
|
"learning_rate": 3.8060233744356633e-06, |
|
"loss": 0.0, |
|
"step": 45 |
|
}, |
|
{ |
|
"epoch": 0.003514629645901063, |
|
"grad_norm": NaN, |
|
"learning_rate": 0.0, |
|
"loss": 0.0, |
|
"step": 50 |
|
}, |
|
{ |
|
"epoch": 0.003514629645901063, |
|
"eval_loss": NaN, |
|
"eval_runtime": 223.0492, |
|
"eval_samples_per_second": 26.855, |
|
"eval_steps_per_second": 13.428, |
|
"step": 50 |
|
} |
|
], |
|
"logging_steps": 5, |
|
"max_steps": 50, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 1, |
|
"save_steps": 13, |
|
"stateful_callbacks": { |
|
"TrainerControl": { |
|
"args": { |
|
"should_epoch_stop": false, |
|
"should_evaluate": false, |
|
"should_log": false, |
|
"should_save": true, |
|
"should_training_stop": true |
|
}, |
|
"attributes": {} |
|
} |
|
}, |
|
"total_flos": 752372484145152.0, |
|
"train_batch_size": 2, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|