all_results.json

{
    "epoch": 3.0,
    "eval_logits/chosen": -1.0759214162826538,
    "eval_logits/rejected": -0.9100367426872253,
    "eval_logps/chosen": -2645.154052734375,
    "eval_logps/rejected": -3544.43115234375,
    "eval_loss": 6.142495155334473,
    "eval_rewards/accuracies": 0.6259999871253967,
    "eval_rewards/chosen": -23.702728271484375,
    "eval_rewards/margins": 9.166375160217285,
    "eval_rewards/margins_max": 58.90415954589844,
    "eval_rewards/margins_min": -33.25897216796875,
    "eval_rewards/margins_std": 29.858272552490234,
    "eval_rewards/rejected": -32.86910629272461,
    "eval_runtime": 737.8634,
    "eval_samples": 2000,
    "eval_samples_per_second": 2.711,
    "eval_steps_per_second": 0.169,
    "train_loss": 0.1413031851636692,
    "train_runtime": 20921.2576,
    "train_samples": 5678,
    "train_samples_per_second": 0.814,
    "train_steps_per_second": 0.051
}