Files
genv3pair1NoGT_1.5B_cdpo_eb…/eval_results.json
ModelHub XC 742c3e40b8 初始化项目,由ModelHub XC社区提供模型
Model: YuchenLi01/genv3pair1NoGT_1.5B_cdpo_ebs32_lr1e-06_beta0.1_epoch16.0_42
Source: Original Platform
2026-05-25 23:10:23 +08:00

16 lines
556 B
JSON

{
"epoch": 16.0,
"eval_logits/chosen": -3.25758695602417,
"eval_logits/rejected": -3.3809757232666016,
"eval_logps/chosen": -30.838027954101562,
"eval_logps/rejected": -56.25493621826172,
"eval_loss": 1.0559927225112915,
"eval_rewards/accuracies": 0.550000011920929,
"eval_rewards/chosen": -0.07679860293865204,
"eval_rewards/margins": -0.07679860293865204,
"eval_rewards/rejected": 0.0,
"eval_runtime": 45.9469,
"eval_samples": 310,
"eval_samples_per_second": 6.747,
"eval_steps_per_second": 0.218
}