| { |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 2.1818181818181817, |
| "eval_steps": 500, |
| "global_step": 6, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 0.7272727272727273, |
| "eval_loss": 1.1343961954116821, |
| "eval_runtime": 3.2144, |
| "eval_samples_per_second": 23.021, |
| "eval_steps_per_second": 0.622, |
| "step": 2 |
| }, |
| { |
| "epoch": 1.8181818181818183, |
| "eval_loss": 1.0645030736923218, |
| "eval_runtime": 3.119, |
| "eval_samples_per_second": 23.725, |
| "eval_steps_per_second": 0.641, |
| "step": 5 |
| }, |
| { |
| "epoch": 2.1818181818181817, |
| "eval_loss": 1.054938554763794, |
| "eval_runtime": 3.2846, |
| "eval_samples_per_second": 22.529, |
| "eval_steps_per_second": 0.609, |
| "step": 6 |
| }, |
| { |
| "epoch": 2.1818181818181817, |
| "step": 6, |
| "total_flos": 9840843816960.0, |
| "train_loss": 1.1885654131571453, |
| "train_runtime": 412.5871, |
| "train_samples_per_second": 10.107, |
| "train_steps_per_second": 0.015 |
| } |
| ], |
| "logging_steps": 10, |
| "max_steps": 6, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 3, |
| "save_steps": 500, |
| "stateful_callbacks": { |
| "TrainerControl": { |
| "args": { |
| "should_epoch_stop": false, |
| "should_evaluate": false, |
| "should_log": false, |
| "should_save": true, |
| "should_training_stop": true |
| }, |
| "attributes": {} |
| } |
| }, |
| "total_flos": 9840843816960.0, |
| "train_batch_size": 8, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
