all_results.json

{
    "epoch": 1.98,
    "eval_logits/chosen": 0.7993869781494141,
    "eval_logits/rejected": 0.8175749778747559,
    "eval_logps/chosen": -91.77693176269531,
    "eval_logps/rejected": -94.28819274902344,
    "eval_loss": 0.009953780099749565,
    "eval_rewards/accuracies": 0.2619999945163727,
    "eval_rewards/chosen": -0.0004955825861543417,
    "eval_rewards/margins": 0.00041623329161666334,
    "eval_rewards/rejected": -0.0009118159650824964,
    "eval_runtime": 274.2974,
    "eval_samples": 2000,
    "eval_samples_per_second": 7.291,
    "eval_steps_per_second": 0.456,
    "train_loss": 0.00995884225458691,
    "train_runtime": 2568.9627,
    "train_samples": 6113,
    "train_samples_per_second": 4.759,
    "train_steps_per_second": 0.074
}