Video Codec-Guided Training-Free Token Compression for Video Large Language Models

作者：Xihua Sheng, Ruijie Liu, Chang Wen Chen · 会议/期刊：arXiv preprint · 方向：cs.CV · 发布日期：2026-10-12