{
    "epoch": 1.0,
    "eval_logits/chosen": -2.8405630588531494,
    "eval_logits/rejected": -2.8439478874206543,
    "eval_logps/chosen": -283.7974548339844,
    "eval_logps/rejected": -260.2491760253906,
    "eval_loss": 0.6432895660400391,
    "eval_pred_label": 0.0,
    "eval_rewards/accuracies": 0.6740000247955322,
    "eval_rewards/chosen": 0.04599040374159813,
    "eval_rewards/margins": 0.14093737304210663,
    "eval_rewards/rejected": -0.0949469730257988,
    "eval_runtime": 468.6797,
    "eval_samples": 2000,
    "eval_samples_per_second": 4.267,
    "eval_steps_per_second": 0.267,
    "eval_use_label": 0.0,
    "train_loss": 0.660746982334796,
    "train_runtime": 25629.2924,
    "train_samples": 61135,
    "train_samples_per_second": 2.385,
    "train_steps_per_second": 0.037
}