{
    "epoch": 1.0,
    "eval_log_odds_chosen": NaN,
    "eval_log_odds_ratio": NaN,
    "eval_logits/chosen": NaN,
    "eval_logits/rejected": NaN,
    "eval_logps/chosen": NaN,
    "eval_logps/rejected": NaN,
    "eval_loss": NaN,
    "eval_nll_loss": NaN,
    "eval_rewards/accuracies": 0.0,
    "eval_rewards/chosen": NaN,
    "eval_rewards/margins": NaN,
    "eval_rewards/rejected": NaN,
    "eval_runtime": 306.8444,
    "eval_samples": 17893,
    "eval_samples_per_second": 58.313,
    "eval_steps_per_second": 1.825,
    "total_flos": 0.0,
    "train_loss": 0.04676872755621073,
    "train_runtime": 74758.4678,
    "train_samples": 71570,
    "train_samples_per_second": 0.957,
    "train_steps_per_second": 0.03
}