TY  - RPRT
TI  - Variation-aware Vision Token Dropping for Faster Large Vision-Language Models
AU  - Junjie Chen
AU  - Xuyang Liu
AU  - Zichen Wen
AU  - Yiyu Wang
AU  - Siteng Huang
AU  - Honggang Chen
PY  - 2026
UR  - https://arxiv.org/abs/2509.01552
ID  - 2509.01552
ER  -