all_results.json

{
    "epoch": 3.0,
    "eval_logits/chosen": 6.03125,
    "eval_logits/rejected": 8.0625,
    "eval_logps/chosen": -1640.0,
    "eval_logps/rejected": -1760.0,
    "eval_loss": 1.570968747138977,
    "eval_rewards/accuracies": 0.58984375,
    "eval_rewards/chosen": -13.1875,
    "eval_rewards/margins": 1.5,
    "eval_rewards/rejected": -14.6875,
    "eval_runtime": 46.4721,
    "eval_samples": 2000,
    "eval_samples_per_second": 43.037,
    "eval_steps_per_second": 0.689,
    "total_flos": 0.0,
    "train_loss": 0.2725709742172434,
    "train_runtime": 12848.3898,
    "train_samples": 61119,
    "train_samples_per_second": 14.271,
    "train_steps_per_second": 0.223
}