all_results.json

{
    "epoch": 1.0,
    "eval_logits/chosen": -0.14711737632751465,
    "eval_logits/rejected": 0.07057683169841766,
    "eval_logps/chosen": -336.2481384277344,
    "eval_logps/rejected": -404.3017578125,
    "eval_loss": 0.5472621321678162,
    "eval_rewards/accuracies": 0.7421875,
    "eval_rewards/chosen": -0.8609448075294495,
    "eval_rewards/margins": 0.6641340851783752,
    "eval_rewards/rejected": -1.5250788927078247,
    "eval_runtime": 99.6217,
    "eval_samples": 2000,
    "eval_samples_per_second": 20.076,
    "eval_steps_per_second": 0.321,
    "train_loss": 0.583915277301329,
    "train_runtime": 6210.8046,
    "train_samples": 61135,
    "train_samples_per_second": 9.843,
    "train_steps_per_second": 0.154
}