all_results.json

{
    "epoch": 4.0,
    "eval_logits/chosen": -2.236751079559326,
    "eval_logits/rejected": -2.2318832874298096,
    "eval_logps/chosen": -34.04561233520508,
    "eval_logps/rejected": -37.546016693115234,
    "eval_loss": 0.5836836099624634,
    "eval_rewards/accuracies": 0.5427741408348083,
    "eval_rewards/chosen": -0.007742464076727629,
    "eval_rewards/margins": 0.01283606793731451,
    "eval_rewards/rejected": -0.02057853527367115,
    "eval_runtime": 145.5744,
    "eval_samples": 343,
    "eval_samples_per_second": 2.356,
    "eval_steps_per_second": 0.295,
    "train_loss": 0.20226844080856868,
    "train_runtime": 10768.956,
    "train_samples": 3079,
    "train_samples_per_second": 1.144,
    "train_steps_per_second": 0.143
}