Update README.md

README.md

@@ -16,6 +16,9 @@ tags:
   
   🤗<a href="https://huggingface.co/collections/long-xing1/caprl-68d64ac32ded31596c36e189">CapRL Collection</a> | 🤗<a href="https://huggingface.co/papers/2509.22647">Daily Paper</a>
 
+Based on the same recipe as CapRL-3B, we used InternVL3.5-8B as the policy model and obtained **CapRL-InternVL3.5-8B** through CapRL. **Its performance significantly surpasses that of Qwen2.5-VL-72B**.
+
+We are working on even stronger base models and upgrading our training recipe — stay tuned!
 
 
 ## Introduction
@@ -34,10 +37,10 @@ By employing CapRL training framework, initializing with the Qwen2.5-VL-3B model
 filtered 75K QA dataset as the training set, we obtained a highly capable captioner, CapRL-3B.
 
 <p align="center">
-  <img src="./assets/teaser.png" alt="Main Results on GPT2" width="750"/>
+  <img src="./assets/teaser.png"  width="750"/>
 </p>
 <p align="center">
-  <img src="./assets/performance_update.png" alt="Main Results on GPT2" width="750"/>
+  <img src="./assets/performance_update.png" width="750"/>
 </p>
 
 ## Key Features