|
{ |
|
"best_metric": null, |
|
"best_model_checkpoint": null, |
|
"epoch": 0.001510745175057597, |
|
"eval_steps": 3, |
|
"global_step": 10, |
|
"is_hyper_param_search": false, |
|
"is_local_process_zero": true, |
|
"is_world_process_zero": true, |
|
"log_history": [ |
|
{ |
|
"epoch": 0.00015107451750575972, |
|
"grad_norm": 0.29824402928352356, |
|
"learning_rate": 2e-05, |
|
"loss": 1.1896, |
|
"step": 1 |
|
}, |
|
{ |
|
"epoch": 0.00015107451750575972, |
|
"eval_loss": 1.3592280149459839, |
|
"eval_runtime": 87.6897, |
|
"eval_samples_per_second": 31.783, |
|
"eval_steps_per_second": 15.897, |
|
"step": 1 |
|
}, |
|
{ |
|
"epoch": 0.00030214903501151944, |
|
"grad_norm": 0.40944790840148926, |
|
"learning_rate": 4e-05, |
|
"loss": 1.5883, |
|
"step": 2 |
|
}, |
|
{ |
|
"epoch": 0.00045322355251727916, |
|
"grad_norm": 0.48245853185653687, |
|
"learning_rate": 6e-05, |
|
"loss": 1.5352, |
|
"step": 3 |
|
}, |
|
{ |
|
"epoch": 0.00045322355251727916, |
|
"eval_loss": 1.3583358526229858, |
|
"eval_runtime": 85.8753, |
|
"eval_samples_per_second": 32.454, |
|
"eval_steps_per_second": 16.233, |
|
"step": 3 |
|
}, |
|
{ |
|
"epoch": 0.0006042980700230389, |
|
"grad_norm": 0.29556429386138916, |
|
"learning_rate": 8e-05, |
|
"loss": 1.2385, |
|
"step": 4 |
|
}, |
|
{ |
|
"epoch": 0.0007553725875287985, |
|
"grad_norm": 0.3337860405445099, |
|
"learning_rate": 0.0001, |
|
"loss": 1.3206, |
|
"step": 5 |
|
}, |
|
{ |
|
"epoch": 0.0009064471050345583, |
|
"grad_norm": 0.4581671357154846, |
|
"learning_rate": 0.00012, |
|
"loss": 1.5301, |
|
"step": 6 |
|
}, |
|
{ |
|
"epoch": 0.0009064471050345583, |
|
"eval_loss": 1.3462966680526733, |
|
"eval_runtime": 85.8021, |
|
"eval_samples_per_second": 32.482, |
|
"eval_steps_per_second": 16.247, |
|
"step": 6 |
|
}, |
|
{ |
|
"epoch": 0.001057521622540318, |
|
"grad_norm": 0.43568184971809387, |
|
"learning_rate": 0.00014, |
|
"loss": 1.1669, |
|
"step": 7 |
|
}, |
|
{ |
|
"epoch": 0.0012085961400460777, |
|
"grad_norm": 0.6850478649139404, |
|
"learning_rate": 0.00016, |
|
"loss": 1.7176, |
|
"step": 8 |
|
}, |
|
{ |
|
"epoch": 0.0013596706575518375, |
|
"grad_norm": 0.4434090852737427, |
|
"learning_rate": 0.00018, |
|
"loss": 1.2542, |
|
"step": 9 |
|
}, |
|
{ |
|
"epoch": 0.0013596706575518375, |
|
"eval_loss": 1.299032211303711, |
|
"eval_runtime": 85.6635, |
|
"eval_samples_per_second": 32.534, |
|
"eval_steps_per_second": 16.273, |
|
"step": 9 |
|
}, |
|
{ |
|
"epoch": 0.001510745175057597, |
|
"grad_norm": 0.3884293735027313, |
|
"learning_rate": 0.0002, |
|
"loss": 1.1699, |
|
"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": 1849564248145920.0, |
|
"train_batch_size": 2, |
|
"trial_name": null, |
|
"trial_params": null |
|
} |
|
|