CriteriaPO/qwen2.5-3b-dpo-finegrained
1
0
Fork 0
Code Issues Pull Requests Actions Projects Releases Wiki Activity

初始化项目，由ModelHub XC社区提供模型

Browse Source 
Model: CriteriaPO/qwen2.5-3b-dpo-finegrained
Source: Original Platform
This commit is contained in:
ModelHub XC
2026-08-12 18:09:31 +08:00
commit 3b1dbd380c
138 changed files with 675275 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

16
eval_results.json Normal file
View File

@@ -0,0 +1,16 @@
{
"epoch": 1.0,
"eval_logits/chosen": -2.9375,
"eval_logits/rejected": -2.625,
"eval_logps/chosen": -772.0,
"eval_logps/rejected": -1168.0,
"eval_loss": 0.21195410192012787,
"eval_rewards/accuracies": 0.9136363863945007,
"eval_rewards/chosen": -6.0,
"eval_rewards/margins": 4.1875,
"eval_rewards/rejected": -10.1875,
"eval_runtime": 3472.8109,
"eval_samples": 99217,
"eval_samples_per_second": 28.364,
"eval_steps_per_second": 0.443
}