|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 0.0006379178361826997, |
|
"eval_steps": 3, |
|
"global_step": 10, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 6.379178361826996e-05, |
|
"grad_norm": 0.4209086000919342, |
|
"learning_rate": 2e-05, |
|
"loss": 0.2181, |
|
"step": 1 |
|
}, |
|
{ |
|
"epoch": 6.379178361826996e-05, |
|
"eval_loss": 0.3880921006202698, |
|
"eval_runtime": 146.0065, |
|
"eval_samples_per_second": 45.21, |
|
"eval_steps_per_second": 22.609, |
|
"step": 1 |
|
}, |
|
{ |
|
"epoch": 0.00012758356723653993, |
|
"grad_norm": 0.5722362995147705, |
|
"learning_rate": 4e-05, |
|
"loss": 0.2823, |
|
"step": 2 |
|
}, |
|
{ |
|
"epoch": 0.0001913753508548099, |
|
"grad_norm": 0.4928780794143677, |
|
"learning_rate": 6e-05, |
|
"loss": 0.393, |
|
"step": 3 |
|
}, |
|
{ |
|
"epoch": 0.0001913753508548099, |
|
"eval_loss": 0.3858100473880768, |
|
"eval_runtime": 143.7304, |
|
"eval_samples_per_second": 45.926, |
|
"eval_steps_per_second": 22.967, |
|
"step": 3 |
|
}, |
|
{ |
|
"epoch": 0.00025516713447307985, |
|
"grad_norm": 0.480742871761322, |
|
"learning_rate": 8e-05, |
|
"loss": 0.4093, |
|
"step": 4 |
|
}, |
|
{ |
|
"epoch": 0.00031895891809134984, |
|
"grad_norm": 0.4650613069534302, |
|
"learning_rate": 0.0001, |
|
"loss": 0.3471, |
|
"step": 5 |
|
}, |
|
{ |
|
"epoch": 0.0003827507017096198, |
|
"grad_norm": 0.4250987768173218, |
|
"learning_rate": 0.00012, |
|
"loss": 0.3982, |
|
"step": 6 |
|
}, |
|
{ |
|
"epoch": 0.0003827507017096198, |
|
"eval_loss": 0.36651504039764404, |
|
"eval_runtime": 143.664, |
|
"eval_samples_per_second": 45.947, |
|
"eval_steps_per_second": 22.977, |
|
"step": 6 |
|
}, |
|
{ |
|
"epoch": 0.00044654248532788977, |
|
"grad_norm": 0.3496827781200409, |
|
"learning_rate": 0.00014, |
|
"loss": 0.3249, |
|
"step": 7 |
|
}, |
|
{ |
|
"epoch": 0.0005103342689461597, |
|
"grad_norm": 0.34638479351997375, |
|
"learning_rate": 0.00016, |
|
"loss": 0.3686, |
|
"step": 8 |
|
}, |
|
{ |
|
"epoch": 0.0005741260525644297, |
|
"grad_norm": 0.2880222201347351, |
|
"learning_rate": 0.00018, |
|
"loss": 0.2585, |
|
"step": 9 |
|
}, |
|
{ |
|
"epoch": 0.0005741260525644297, |
|
"eval_loss": 0.3495641052722931, |
|
"eval_runtime": 145.1558, |
|
"eval_samples_per_second": 45.475, |
|
"eval_steps_per_second": 22.741, |
|
"step": 9 |
|
}, |
|
{ |
|
"epoch": 0.0006379178361826997, |
|
"grad_norm": 0.29309144616127014, |
|
"learning_rate": 0.0002, |
|
"loss": 0.3457, |
|
"step": 10 |
|
} |
|
], |
|
"logging_steps": 1, |
|
"max_steps": 10, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 1, |
|
"save_steps": 3, |
|
"stateful_callbacks": { |
|
"TrainerControl": { |
|
"args": { |
|
"should_epoch_stop": false, |
|
"should_evaluate": false, |
|
"should_log": false, |
|
"should_save": true, |
|
"should_training_stop": true |
|
}, |
|
"attributes": {} |
|
} |
|
}, |
|
"total_flos": 175850373513216.0, |
|
"train_batch_size": 2, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|