| { | |
| "best_metric": null, | |
| "best_model_checkpoint": null, | |
| "epoch": 0.0026171159382360636, | |
| "eval_steps": 20, | |
| "global_step": 20, | |
| "is_hyper_param_search": false, | |
| "is_local_process_zero": true, | |
| "is_world_process_zero": true, | |
| "log_history": [ | |
| { | |
| "epoch": 0.0, | |
| "grad_norm": 9.931995391845703, | |
| "learning_rate": 0.0, | |
| "log_odds_chosen": 0.0738741084933281, | |
| "log_odds_ratio": -0.6917311549186707, | |
| "logits/chosen": -2.934194564819336, | |
| "logits/rejected": -2.862790584564209, | |
| "logps/chosen": -0.9762001037597656, | |
| "logps/rejected": -1.031965970993042, | |
| "loss": 2.1232, | |
| "nll_loss": 2.054041624069214, | |
| "rewards/accuracies": 0.4625000059604645, | |
| "rewards/chosen": -0.09762001037597656, | |
| "rewards/margins": 0.0055765872821211815, | |
| "rewards/rejected": -0.10319659858942032, | |
| "step": 20 | |
| }, | |
| { | |
| "epoch": 0.0, | |
| "eval_log_odds_chosen": 0.1270444095134735, | |
| "eval_log_odds_ratio": -0.6689019203186035, | |
| "eval_logits/chosen": -2.976943254470825, | |
| "eval_logits/rejected": -2.9503843784332275, | |
| "eval_logps/chosen": -1.011623501777649, | |
| "eval_logps/rejected": -1.0984618663787842, | |
| "eval_loss": 1.6425423622131348, | |
| "eval_nll_loss": 1.5756521224975586, | |
| "eval_rewards/accuracies": 0.5705000162124634, | |
| "eval_rewards/chosen": -0.10116235911846161, | |
| "eval_rewards/margins": 0.008683841675519943, | |
| "eval_rewards/rejected": -0.10984619706869125, | |
| "eval_runtime": 2631.0618, | |
| "eval_samples_per_second": 0.76, | |
| "eval_steps_per_second": 0.38, | |
| "step": 20 | |
| } | |
| ], | |
| "logging_steps": 20, | |
| "max_steps": 20, | |
| "num_input_tokens_seen": 0, | |
| "num_train_epochs": 1, | |
| "save_steps": 10, | |
| "total_flos": 0.0, | |
| "train_batch_size": 2, | |
| "trial_name": null, | |
| "trial_params": null | |
| } | |