all_results.json

{
    "epoch": 3.0,
    "eval_logits/chosen": -15.5625,
    "eval_logits/rejected": -15.25,
    "eval_logps/chosen": -724.0,
    "eval_logps/rejected": -740.0,
    "eval_loss": 1.1510624885559082,
    "eval_rewards/accuracies": 0.49609375,
    "eval_rewards/chosen": -4.03125,
    "eval_rewards/margins": 0.482421875,
    "eval_rewards/rejected": -4.53125,
    "eval_runtime": 46.5353,
    "eval_samples": 2000,
    "eval_samples_per_second": 42.978,
    "eval_steps_per_second": 0.688,
    "total_flos": 0.0,
    "train_loss": 0.02151861389916891,
    "train_runtime": 13053.1122,
    "train_samples": 61119,
    "train_samples_per_second": 14.047,
    "train_steps_per_second": 0.219
}