一作ACL 2026Poster· Natural Language Processing
Reducing Token Redundancy in LVLM s: A Systematic Review of Token Pruning Methods
Hanzhang Yuan · Mengxuan Hu · Wenhao Zhang · Tianlong Wang · Zhongliang Zhou · Jiasen Lu · Sheng Li
摘要
Large Vision-Language Models (LVLMs) excel at visual understanding but face severe computational bottlenecks when processing high-resolution images and long videos due to massive visual token counts. Token pruning mitigates this by selectively removing less informative tokens while maintaining pe...