all_results.json

{
    "epoch": 1.0,
    "eval_dpo_losses": 0.6521267890930176,
    "eval_logits/chosen": -2.2373995780944824,
    "eval_logits/rejected": -2.133667230606079,
    "eval_logps/chosen": -258.7702941894531,
    "eval_logps/rejected": -254.93106079101562,
    "eval_loss": 0.6839185953140259,
    "eval_positive_losses": 0.16242748498916626,
    "eval_rewards/accuracies": 0.692460298538208,
    "eval_rewards/chosen": 0.17003849148750305,
    "eval_rewards/margins": 0.09114638715982437,
    "eval_rewards/margins_max": 0.362783819437027,
    "eval_rewards/margins_min": -0.12145084142684937,
    "eval_rewards/margins_std": 0.1623499095439911,
    "eval_rewards/rejected": 0.07889209687709808,
    "eval_runtime": 385.7962,
    "eval_samples": 2000,
    "eval_samples_per_second": 5.184,
    "eval_steps_per_second": 0.163,
    "train_loss": 0.6847423079499896,
    "train_runtime": 42109.9082,
    "train_samples": 61134,
    "train_samples_per_second": 1.452,
    "train_steps_per_second": 0.091
}