VisionZip: Longer is Better but Not Necessary in Vision Language Models 논문리뷰

https://arxiv.org/pdf/2412.04467