InfoMerge: Information-aware Token Compression for Efficient Video Large Language Models 事件

InfoMerge: Information-aware Token Compression for Efficient Video Large Language Models arXiv:2606.02161v1 Announce Type: new Abstract: Video Large Language Models (Video-LLMs) achieve strong performance in video understanding, but their excessive visual tokens bring substantial computational overhead. Existing training-free compression methods improve inference efficiency by reducing visual tokens, yet they often rely on local adjacent-frame similarity for temporal redundancy estimation or al