|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 0.001383253412025083, |
|
"eval_steps": 8, |
|
"global_step": 30, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 4.610844706750277e-05, |
|
"eval_loss": NaN, |
|
"eval_runtime": 652.4716, |
|
"eval_samples_per_second": 13.996, |
|
"eval_steps_per_second": 6.998, |
|
"step": 1 |
|
}, |
|
{ |
|
"epoch": 0.0001383253412025083, |
|
"grad_norm": NaN, |
|
"learning_rate": 6e-05, |
|
"loss": 0.0, |
|
"step": 3 |
|
}, |
|
{ |
|
"epoch": 0.0002766506824050166, |
|
"grad_norm": NaN, |
|
"learning_rate": 0.00012, |
|
"loss": 0.0, |
|
"step": 6 |
|
}, |
|
{ |
|
"epoch": 0.00036886757654002215, |
|
"eval_loss": NaN, |
|
"eval_runtime": 652.4014, |
|
"eval_samples_per_second": 13.998, |
|
"eval_steps_per_second": 6.999, |
|
"step": 8 |
|
}, |
|
{ |
|
"epoch": 0.0004149760236075249, |
|
"grad_norm": NaN, |
|
"learning_rate": 0.00018, |
|
"loss": 0.0, |
|
"step": 9 |
|
}, |
|
{ |
|
"epoch": 0.0005533013648100332, |
|
"grad_norm": NaN, |
|
"learning_rate": 0.00019510565162951537, |
|
"loss": 0.0, |
|
"step": 12 |
|
}, |
|
{ |
|
"epoch": 0.0006916267060125415, |
|
"grad_norm": NaN, |
|
"learning_rate": 0.00017071067811865476, |
|
"loss": 0.0, |
|
"step": 15 |
|
}, |
|
{ |
|
"epoch": 0.0007377351530800443, |
|
"eval_loss": NaN, |
|
"eval_runtime": 652.574, |
|
"eval_samples_per_second": 13.994, |
|
"eval_steps_per_second": 6.997, |
|
"step": 16 |
|
}, |
|
{ |
|
"epoch": 0.0008299520472150498, |
|
"grad_norm": NaN, |
|
"learning_rate": 0.00013090169943749476, |
|
"loss": 0.0, |
|
"step": 18 |
|
}, |
|
{ |
|
"epoch": 0.0009682773884175581, |
|
"grad_norm": NaN, |
|
"learning_rate": 8.435655349597689e-05, |
|
"loss": 0.0, |
|
"step": 21 |
|
}, |
|
{ |
|
"epoch": 0.0011066027296200663, |
|
"grad_norm": NaN, |
|
"learning_rate": 4.12214747707527e-05, |
|
"loss": 0.0, |
|
"step": 24 |
|
}, |
|
{ |
|
"epoch": 0.0011066027296200663, |
|
"eval_loss": NaN, |
|
"eval_runtime": 652.3612, |
|
"eval_samples_per_second": 13.998, |
|
"eval_steps_per_second": 6.999, |
|
"step": 24 |
|
}, |
|
{ |
|
"epoch": 0.0012449280708225748, |
|
"grad_norm": NaN, |
|
"learning_rate": 1.0899347581163221e-05, |
|
"loss": 0.0, |
|
"step": 27 |
|
}, |
|
{ |
|
"epoch": 0.001383253412025083, |
|
"grad_norm": NaN, |
|
"learning_rate": 0.0, |
|
"loss": 0.0, |
|
"step": 30 |
|
} |
|
], |
|
"logging_steps": 3, |
|
"max_steps": 30, |
|
"num_input_tokens_seen": 0, |
|
"num_train_epochs": 1, |
|
"save_steps": 10, |
|
"stateful_callbacks": { |
|
"TrainerControl": { |
|
"args": { |
|
"should_epoch_stop": false, |
|
"should_evaluate": false, |
|
"should_log": false, |
|
"should_save": true, |
|
"should_training_stop": true |
|
}, |
|
"attributes": {} |
|
} |
|
}, |
|
"total_flos": 1.0572960236568576e+16, |
|
"train_batch_size": 2, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|