all_results.json

{
    "epoch": 1.0,
    "eval_dpo_losses": 0.6679562926292419,
    "eval_logits/chosen": -2.784212112426758,
    "eval_logits/rejected": -2.7470052242279053,
    "eval_logps/chosen": -269.742919921875,
    "eval_logps/rejected": -249.1260223388672,
    "eval_loss": 0.688105046749115,
    "eval_positive_losses": 0.09528277069330215,
    "eval_rewards/accuracies": 0.6919999718666077,
    "eval_rewards/chosen": 0.14850492775440216,
    "eval_rewards/margins": 0.05397578701376915,
    "eval_rewards/margins_max": 0.21397222578525543,
    "eval_rewards/margins_min": -0.09459882229566574,
    "eval_rewards/margins_std": 0.10261555016040802,
    "eval_rewards/rejected": 0.09452912956476212,
    "eval_runtime": 428.7571,
    "eval_samples": 2000,
    "eval_samples_per_second": 4.665,
    "eval_steps_per_second": 0.292,
    "train_loss": 0.6874516553791823,
    "train_runtime": 3895.4317,
    "train_samples": 5263,
    "train_samples_per_second": 1.351,
    "train_steps_per_second": 0.084
}