| { |
| "best_metric": null, |
| "best_model_checkpoint": null, |
| "epoch": 4.0, |
| "eval_steps": 500, |
| "global_step": 1644, |
| "is_hyper_param_search": false, |
| "is_local_process_zero": true, |
| "is_world_process_zero": true, |
| "log_history": [ |
| { |
| "epoch": 1.0, |
| "eval_accuracy": 0.5871421461904744, |
| "eval_loss": 2.005277633666992, |
| "eval_runtime": 29.0983, |
| "eval_samples_per_second": 113.134, |
| "eval_steps_per_second": 3.54, |
| "step": 411 |
| }, |
| { |
| "epoch": 1.22, |
| "learning_rate": 4.9996641797696206e-05, |
| "loss": 2.0885, |
| "step": 500 |
| }, |
| { |
| "epoch": 2.0, |
| "eval_accuracy": 0.5971017152277686, |
| "eval_loss": 1.9286963939666748, |
| "eval_runtime": 28.862, |
| "eval_samples_per_second": 114.06, |
| "eval_steps_per_second": 3.569, |
| "step": 822 |
| }, |
| { |
| "epoch": 2.43, |
| "learning_rate": 2.9684532864643122e-05, |
| "loss": 1.8728, |
| "step": 1000 |
| }, |
| { |
| "epoch": 3.0, |
| "eval_accuracy": 0.603926221807302, |
| "eval_loss": 1.8916035890579224, |
| "eval_runtime": 28.7797, |
| "eval_samples_per_second": 114.386, |
| "eval_steps_per_second": 3.579, |
| "step": 1233 |
| }, |
| { |
| "epoch": 3.65, |
| "learning_rate": 1.9095509616124385e-06, |
| "loss": 1.7214, |
| "step": 1500 |
| }, |
| { |
| "epoch": 4.0, |
| "eval_accuracy": 0.6049892568138169, |
| "eval_loss": 1.891752004623413, |
| "eval_runtime": 28.8553, |
| "eval_samples_per_second": 114.086, |
| "eval_steps_per_second": 3.57, |
| "step": 1644 |
| }, |
| { |
| "epoch": 4.0, |
| "step": 1644, |
| "total_flos": 2.475153948672e+16, |
| "train_loss": 1.875104602525994, |
| "train_runtime": 621.1187, |
| "train_samples_per_second": 84.686, |
| "train_steps_per_second": 2.647 |
| }, |
| { |
| "epoch": 4.0, |
| "eval_accuracy": 0.6049892568138169, |
| "eval_loss": 1.891752004623413, |
| "eval_runtime": 28.9822, |
| "eval_samples_per_second": 113.587, |
| "eval_steps_per_second": 3.554, |
| "step": 1644 |
| } |
| ], |
| "logging_steps": 500, |
| "max_steps": 1644, |
| "num_input_tokens_seen": 0, |
| "num_train_epochs": 4, |
| "save_steps": 500, |
| "total_flos": 2.475153948672e+16, |
| "train_batch_size": 32, |
| "trial_name": null, |
| "trial_params": null |
| } |
|
|
