AI RESEARCH
EvoComp: Learning Visual Token Compression for Multimodal Large Language Models via Semantic-Guided Evolutionary Labeling
arXiv CS.LG
•
ArXi:2604.17087v1 Announce Type: cross Recent Multimodal Large Language Models (MLLMs) have nstrated strong performance on vision-language understanding tasks, yet their inference efficiency is often hampered by the large number of visual tokens, particularly in high-resolution or multi-image scenarios. To address this issue, we propose EvoComp, a visual token compression framework that significantly reduces token count while preserving task accuracy. EvoComp