{
    "epoch": 1.0,
    "eval_logits/chosen": -2.827404737472534,
    "eval_logits/rejected": -2.818655014038086,
    "eval_logps/chosen": -283.37945556640625,
    "eval_logps/rejected": -265.9969787597656,
    "eval_loss": 0.5459502935409546,
    "eval_pred_label": 0.0,
    "eval_rewards/accuracies": 0.7139999866485596,
    "eval_rewards/chosen": 0.08778975158929825,
    "eval_rewards/margins": 0.7575166821479797,
    "eval_rewards/rejected": -0.6697269678115845,
    "eval_runtime": 478.4263,
    "eval_samples": 2000,
    "eval_samples_per_second": 4.18,
    "eval_steps_per_second": 0.261,
    "eval_use_label": 0.0,
    "train_loss": 0.5628191218950361,
    "train_runtime": 25746.1298,
    "train_samples": 61135,
    "train_samples_per_second": 2.375,
    "train_steps_per_second": 0.037
}