| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0, | |
| "eval_steps": 500, | |
| "global_step": 0, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0, | |
| "eval_loss": 2.304720878601074, | |
| "eval_runtime": 8.4159, | |
| "eval_samples_per_second": 97.435, | |
| "eval_steps_per_second": 19.487, | |
| "step": 0 | |
| }, | |
| { | |
| "epoch": 0, | |
| "eval_loss": 2.3047213554382324, | |
| "eval_runtime": 8.4725, | |
| "eval_samples_per_second": 96.784, | |
| "eval_steps_per_second": 19.357, | |
| "step": 0 | |
| }, | |
| { | |
| "epoch": 0, | |
| "eval_loss": 2.3047211170196533, | |
| "eval_runtime": 8.4856, | |
| "eval_samples_per_second": 96.634, | |
| "eval_steps_per_second": 19.327, | |
| "step": 0 | |
| }, | |
| { | |
| "epoch": 0, | |
| "eval_loss": 2.304720878601074, | |
| "eval_runtime": 8.5069, | |
| "eval_samples_per_second": 96.393, | |
| "eval_steps_per_second": 19.279, | |
| "step": 0 | |
| }, | |
| { | |
| "epoch": 0, | |
| "eval_loss": 2.3047211170196533, | |
| "eval_runtime": 8.5052, | |
| "eval_samples_per_second": 96.412, | |
| "eval_steps_per_second": 19.282, | |
| "step": 0 | |
| }, | |
| { | |
| "epoch": 0, | |
| "eval_loss": 2.3047211170196533, | |
| "eval_runtime": 8.5056, | |
| "eval_samples_per_second": 96.407, | |
| "eval_steps_per_second": 19.281, | |
| "step": 0 | |
| }, | |
| { | |
| "epoch": 0, | |
| "step": 0, | |
| "total_flos": 3.3809589564604416e+16, | |
| "train_loss": 1404.587984085083, | |
| "train_runtime": 212.882, | |
| "train_samples_per_second": 23.111, | |
| "train_steps_per_second": 0.028 | |
| } | |
| ], | |
| "logging_steps": 5, | |
| "max_steps": 6, | |
| "num_train_epochs": 6, | |
| "save_steps": 500, | |
| "total_flos": 3.3809589564604416e+16, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |