Pinned Loading

1. HiRED HiRED Public

   [AAAI 2025] HiRED strategically drops visual tokens in the image encoding stage to improve inference efficiency for High-Resolution Vision-Language Models (e.g., LLaVA-Next) under a fixed token bud…

   Python 16 4

2. llava-hallunication-fix llava-hallunication-fix Public

   The goal of this project is to mitigate the in context hallucination of LLaVA by tweaking its attention mask during token generation

   Python 1

3. Ray-Tracing- Ray-Tracing- Public

   C++ 2

4. LLaVA LLaVA Public

   Forked from haotian-liu/LLaVA

   [NeurIPS'23 Oral] Visual Instruction Tuning (LLaVA) built towards GPT-4V level capabilities and beyond.

   Python

5. Lines-Of-Action Lines-Of-Action Public

   Java 2