Files
tinyllama-1.1b-dpo-pku-safe…/eval_results.json
ModelHub XC 096673a501 初始化项目,由ModelHub XC社区提供模型
Model: AIPlans/tinyllama-1.1b-dpo-pku-saferlhf
Source: Original Platform
2026-06-11 00:38:47 +08:00

15 lines
564 B
JSON

{
"epoch": 0.9997600191984641,
"eval_logits/chosen": -2.71634578704834,
"eval_logits/rejected": -2.595935583114624,
"eval_logps/chosen": -232.62538146972656,
"eval_logps/rejected": -205.43319702148438,
"eval_loss": 0.6742354035377502,
"eval_rewards/accuracies": 0.675000011920929,
"eval_rewards/chosen": 0.050759363919496536,
"eval_rewards/margins": 0.041003111749887466,
"eval_rewards/rejected": 0.009756244719028473,
"eval_runtime": 21.1604,
"eval_samples_per_second": 47.258,
"eval_steps_per_second": 11.815
}