all_results.json

{
    "epoch": 3.0,
    "eval_logits/chosen": -12.6875,
    "eval_logits/rejected": -12.1875,
    "eval_logps/chosen": -656.0,
    "eval_logps/rejected": -672.0,
    "eval_loss": 1.014468789100647,
    "eval_rewards/accuracies": 0.50390625,
    "eval_rewards/chosen": -3.375,
    "eval_rewards/margins": 0.45703125,
    "eval_rewards/rejected": -3.84375,
    "eval_runtime": 46.5719,
    "eval_samples": 2000,
    "eval_samples_per_second": 42.944,
    "eval_steps_per_second": 0.687,
    "total_flos": 0.0,
    "train_loss": 0.023411974788198075,
    "train_runtime": 12588.5672,
    "train_samples": 61119,
    "train_samples_per_second": 14.565,
    "train_steps_per_second": 0.228
}