Files
llama-3-8b-base-new-dpo-hh-…/eval_results.json
ModelHub XC cf60694ff3 初始化项目,由ModelHub XC社区提供模型
Model: jackf857/llama-3-8b-base-new-dpo-hh-harmless-4xh200-batch-64-q_t-0.5-s_star-1.0
Source: Original Platform
2026-06-03 22:08:02 +08:00

17 lines
641 B
JSON

{
"epoch": 0.999244142101285,
"eval_fcm_dpo/beta": 0.567912757396698,
"eval_logits/chosen": 0.20129863917827606,
"eval_logits/rejected": 0.16339144110679626,
"eval_logps/chosen": -77.83040618896484,
"eval_logps/ref_chosen": -74.85946655273438,
"eval_logps/ref_rejected": -79.54898834228516,
"eval_logps/rejected": -83.91458892822266,
"eval_loss": 0.5609715580940247,
"eval_margin_dpo/margin_mean": 1.3946577310562134,
"eval_margin_dpo/margin_std": 2.489074230194092,
"eval_runtime": 37.9328,
"eval_samples": 2303,
"eval_samples_per_second": 60.713,
"eval_steps_per_second": 1.898
}