| { |
| "best_global_step": null, |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 4.0, |
| "eval_steps": 500, |
| "global_step": 1140, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 1.0, |
| "eval_accuracy": 0.9513888888888888, |
| "eval_loss": 0.25660452246665955, |
| "eval_runtime": 9.1525, |
| "eval_samples_per_second": 31.467, |
| "eval_steps_per_second": 3.933, |
| "step": 285 |
| }, |
| { |
| "epoch": 1.7543859649122808, |
| "grad_norm": 0.02743104100227356, |
| "learning_rate": 3.249122807017544e-05, |
| "loss": 0.0946, |
| "step": 500 |
| }, |
| { |
| "epoch": 2.0, |
| "eval_accuracy": 0.9826388888888888, |
| "eval_loss": 0.10773849487304688, |
| "eval_runtime": 9.0685, |
| "eval_samples_per_second": 31.758, |
| "eval_steps_per_second": 3.97, |
| "step": 570 |
| }, |
| { |
| "epoch": 3.0, |
| "eval_accuracy": 0.9861111111111112, |
| "eval_loss": 0.04608719050884247, |
| "eval_runtime": 8.9558, |
| "eval_samples_per_second": 32.158, |
| "eval_steps_per_second": 4.02, |
| "step": 855 |
| }, |
| { |
| "epoch": 3.5087719298245617, |
| "grad_norm": 0.015116340480744839, |
| "learning_rate": 1.4947368421052632e-05, |
| "loss": 0.0468, |
| "step": 1000 |
| }, |
| { |
| "epoch": 4.0, |
| "eval_accuracy": 0.9826388888888888, |
| "eval_loss": 0.08199767768383026, |
| "eval_runtime": 9.0401, |
| "eval_samples_per_second": 31.858, |
| "eval_steps_per_second": 3.982, |
| "step": 1140 |
| } |
| ], |
| "logging_steps": 500, |
| "max_steps": 1425, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 5, |
| "save_steps": 500, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": false |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 2397467936440320.0, |
| "train_batch_size": 8, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|