Update README.md

README.md

@@ -69,6 +69,8 @@ model-index:
 
 [![Hugging Face Models](https://img.shields.io/badge/%F0%9F%A4%97%20Hugging%20Face-ViCA2-blue)](https://huggingface.co/nkkbr/ViCA2)
 
+[![arXiv](https://img.shields.io/badge/arXiv-2505.12312-B31B1B?logo=arxiv&link=https://arxiv.org/abs/2505.12312)](https://arxiv.org/abs/2505.12312)
+
 ## Overview
 
 **ViCA-7B** is a vision-language model specifically fine-tuned for *visuospatial reasoning* in indoor video environments. Built upon the LLaVA-Video-7B-Qwen2 architecture, it is trained using our newly proposed **ViCA-322K dataset**, which emphasizes both structured spatial annotations and complex instruction-based reasoning tasks.