{
    "epoch": 1.0,
    "eval_logits/chosen": -2.159156560897827,
    "eval_logits/rejected": -1.9642761945724487,
    "eval_logps/chosen": -304.89056396484375,
    "eval_logps/rejected": -175.10516357421875,
    "eval_loss": 0.04293975234031677,
    "eval_rewards/accuracies": 0.9741379022598267,
    "eval_rewards/chosen": -0.5987315773963928,
    "eval_rewards/margins": 9.656463623046875,
    "eval_rewards/rejected": -10.255194664001465,
    "eval_runtime": 497.484,
    "eval_samples": 1843,
    "eval_samples_per_second": 3.705,
    "eval_steps_per_second": 0.117,
    "train_loss": 0.06308119606848633,
    "train_runtime": 23531.0319,
    "train_samples": 55762,
    "train_samples_per_second": 2.37,
    "train_steps_per_second": 0.074
}