- [2025/06/28] 🧠 Building on GEMeX, we now introduce ThinkVG, which helps LVLMs think while perform visually grounding.
- [2025/06/26] 🎉 Great News! Our paper has been accepted to ICCV 2025!
- [2025/01/17] We release pre-trained LLaVA-Med-GEMeX, a simple baseline model based on GEMeX VQA data.
- [2024/12/01] We release part of data and complete training code.
git clone https://github.com/Awenbocc/GEMeX-Project.git
cd llava-med
bash fine_tune_gemex.sh # modify data location