{
  "best_global_step": null,
  "best_metric": null,
  "best_model_checkpoint": null,
  "epoch": 3.0,
  "eval_steps": 500,
  "global_step": 30,
  "is_hyper_param_search": false,
  "is_local_process_zero": true,
  "is_world_process_zero": true,
  "log_history": [
    {
      "epoch": 0.5333333333333333,
      "grad_norm": 0.9012682437896729,
      "learning_rate": 9.966191788709714e-07,
      "log_odds_chosen": 0.9927773475646973,
      "log_odds_ratio": -0.47375020384788513,
      "logits/chosen": 4.043400764465332,
      "logits/rejected": 4.5044355392456055,
      "logps/chosen": -1.5020328760147095,
      "logps/rejected": -2.389401912689209,
      "loss": 2.3713,
      "nll_loss": 2.3238818645477295,
      "rewards/accuracies": 0.800000011920929,
      "rewards/chosen": -0.15020330250263214,
      "rewards/margins": 0.088736891746521,
      "rewards/rejected": -0.23894019424915314,
      "step": 5
    },
    {
      "epoch": 1.0,
      "grad_norm": 0.9028968214988708,
      "learning_rate": 8.83022221559489e-07,
      "log_odds_chosen": 0.9902886748313904,
      "log_odds_ratio": -0.5574756264686584,
      "logits/chosen": 4.209258079528809,
      "logits/rejected": 4.346748352050781,
      "logps/chosen": -1.4047656059265137,
      "logps/rejected": -2.313467264175415,
      "loss": 2.3148,
      "nll_loss": 2.2753841876983643,
      "rewards/accuracies": 0.6285714507102966,
      "rewards/chosen": -0.14047658443450928,
      "rewards/margins": 0.09087015688419342,
      "rewards/rejected": -0.2313467264175415,
      "step": 10
    },
    {
      "epoch": 1.5333333333333332,
      "grad_norm": 0.8105126619338989,
      "learning_rate": 6.434016163555451e-07,
      "log_odds_chosen": 0.8671220541000366,
      "log_odds_ratio": -0.544124960899353,
      "logits/chosen": 4.157012939453125,
      "logits/rejected": 4.290781021118164,
      "logps/chosen": -1.4531787633895874,
      "logps/rejected": -2.2471916675567627,
      "loss": 2.3567,
      "nll_loss": 2.30228853225708,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.14531788229942322,
      "rewards/margins": 0.07940130680799484,
      "rewards/rejected": -0.22471916675567627,
      "step": 15
    },
    {
      "epoch": 2.0,
      "grad_norm": 0.8526118397712708,
      "learning_rate": 3.56598383644455e-07,
      "log_odds_chosen": 1.1391531229019165,
      "log_odds_ratio": -0.4789137840270996,
      "logits/chosen": 4.081291675567627,
      "logits/rejected": 4.590013027191162,
      "logps/chosen": -1.4622834920883179,
      "logps/rejected": -2.482956886291504,
      "loss": 2.3068,
      "nll_loss": 2.2980434894561768,
      "rewards/accuracies": 0.7714285850524902,
      "rewards/chosen": -0.1462283581495285,
      "rewards/margins": 0.10206736624240875,
      "rewards/rejected": -0.24829572439193726,
      "step": 20
    },
    {
      "epoch": 2.533333333333333,
      "grad_norm": 0.7531288862228394,
      "learning_rate": 1.1697777844051104e-07,
      "log_odds_chosen": 1.1075918674468994,
      "log_odds_ratio": -0.5166323184967041,
      "logits/chosen": 4.147093772888184,
      "logits/rejected": 4.431370735168457,
      "logps/chosen": -1.4894083738327026,
      "logps/rejected": -2.5067427158355713,
      "loss": 2.3673,
      "nll_loss": 2.315680980682373,
      "rewards/accuracies": 0.699999988079071,
      "rewards/chosen": -0.14894084632396698,
      "rewards/margins": 0.10173343122005463,
      "rewards/rejected": -0.250674307346344,
      "step": 25
    },
    {
      "epoch": 3.0,
      "grad_norm": 0.8306752443313599,
      "learning_rate": 3.380821129028488e-09,
      "log_odds_chosen": 0.8599945902824402,
      "log_odds_ratio": -0.507698655128479,
      "logits/chosen": 4.097240447998047,
      "logits/rejected": 4.431210994720459,
      "logps/chosen": -1.415029764175415,
      "logps/rejected": -2.174788475036621,
      "loss": 2.3028,
      "nll_loss": 2.276947498321533,
      "rewards/accuracies": 0.7714285850524902,
      "rewards/chosen": -0.1415029764175415,
      "rewards/margins": 0.07597588747739792,
      "rewards/rejected": -0.21747885644435883,
      "step": 30
    }
  ],
  "logging_steps": 5,
  "max_steps": 30,
  "num_input_tokens_seen": 0,
  "num_train_epochs": 3,
  "save_steps": 50,
  "stateful_callbacks": {
    "TrainerControl": {
      "args": {
        "should_epoch_stop": false,
        "should_evaluate": false,
        "should_log": false,
        "should_save": true,
        "should_training_stop": true
      },
      "attributes": {}
    }
  },
  "total_flos": 0.0,
  "train_batch_size": 1,
  "trial_name": null,
  "trial_params": null
}