diff --git a/README.md b/README.md
index 922d81b..1cb215e 100644
--- a/README.md
+++ b/README.md
@@ -1,5 +1,7 @@
 # comfyui-Image-reward
 
+![1711282071529](https://github.com/shadowcz007/comfyui-Image-reward/assets/12645064/8a604c1d-66c2-43d2-a1a4-f2c28c02f3cb)
+![1711282130123](https://github.com/shadowcz007/comfyui-Image-reward/assets/12645064/1eb11f53-ac9c-406d-a16b-a672e0ea0787)
 
 
 [ImageReward](https://github.com/THUDM/ImageReward?tab=readme-ov-file#example-use)：文本到图像生成中的人类偏好学习。这是NeurIPS 2023的一篇论文，训练使用专业的大规模数据集ImageRewardDB：约13.7万个⽐较pairs。