{ "epoch": 0.999244142101285, "eval_beta_dpo/beta": 0.09165588766336441, "eval_beta_dpo/beta_margin_grad_mean": -0.38951602578163147, "eval_beta_dpo/beta_margin_grad_std": 0.142334446310997, "eval_beta_dpo/beta_margin_mean": 2.0180046558380127, "eval_beta_dpo/beta_margin_std": 2.601975917816162, "eval_beta_dpo/beta_used": 0.09165588766336441, "eval_beta_dpo/beta_used_raw": -0.021205438300967216, "eval_beta_dpo/gap_mean": 18.827608108520508, "eval_beta_dpo/gap_std": 28.881866455078125, "eval_beta_dpo/loss_margin_mean": 16.80752182006836, "eval_beta_dpo/mask_keep_frac": 1.0, "eval_logits/chosen": 0.4999350905418396, "eval_logits/rejected": 0.4466552138328552, "eval_loss": 0.6952157020568848, "eval_runtime": 38.7686, "eval_samples": 2303, "eval_samples_per_second": 59.404, "eval_steps_per_second": 1.857 }