all_results.json

{
    "epoch": 2.0,
    "eval_logits/chosen": 0.6623885631561279,
    "eval_logits/rejected": 0.7668179869651794,
    "eval_logps/chosen": -271.2955017089844,
    "eval_logps/rejected": -249.1475830078125,
    "eval_loss": 2156.2255859375,
    "eval_rewards/accuracies": 0.6940000057220459,
    "eval_rewards/chosen": -0.11051338165998459,
    "eval_rewards/margins": 0.06660113483667374,
    "eval_rewards/rejected": -0.17711451649665833,
    "eval_runtime": 325.4702,
    "eval_samples": 2000,
    "eval_samples_per_second": 6.145,
    "eval_steps_per_second": 0.384,
    "train_loss": 2246.599348344143,
    "train_runtime": 18130.6033,
    "train_samples": 30567,
    "train_samples_per_second": 3.372,
    "train_steps_per_second": 0.053
}