TY  - RPRT
TI  - Accelerating Multimodal Large Language Models via Dynamic Visual-Token Exit and the Empirical Findings
AU  - Qiong Wu
AU  - Wenhao Lin
AU  - Yiyi Zhou
AU  - Weihao Ye
AU  - Zhanpeng Zen
AU  - Xiaoshuai Sun
AU  - Rongrong Ji
PY  - 2025
UR  - https://arxiv.org/abs/2411.19628
ID  - 2411.19628
ER  -