KiToke通过全局冗余检测，实现视频大模型的高效压缩。
KiToke: Kernel-based Interval-aware Token Compression for Video Large Language Models
- 用核方法全局衡量视觉令牌冗余，自适应选择关键信息。
- 在仅保留1%令牌时仍保持性能，显著优于现有无训练压缩方法。
- 适合资源受限场景下的视频大模型推理加速，尤其适合极端压缩需求。
视频大语言模型（Video LLMs）在视频理解任务中表现强劲，但因视觉令牌数量庞大导致推理成本高。本文提出KiToke，一种无需训练、与查询无关的令牌压缩方法，可减少时空冗余并保留关键视觉信息。该方法利用基于核的冗余度量全局估计令牌多样性，实现内容自适应选择，即使在极低令牌预算下依然有效；同时引入轻量级时间区间构建与区间感知的令牌合并机制，以维持时间连贯性。与依赖局部或片段级启发式方法不同，KiToke显式捕捉整段视频的全局冗余，提升令牌利用效率。在多个视频理解基准和视频大模型主干上进行的大量实验表明，KiToke持续优于现有无训练压缩方法，尤其在激进保留率（低至1%）下表现出显著优势。
原文摘要 · Abstract (English)
Video Large Language Models (Video LLMs) achieve strong performance on video understanding tasks but suffer from high inference costs due to the large number of visual tokens. We propose KiToke, a training-free, query-agnostic token compression approach that reduces spatiotemporal redundancy while preserving critical visual information. Our method estimates token diversity globally using a kernel-based redundancy measure, enabling content-adaptive selection that remains effective under extreme token budgets, and further introduces a lightweight temporal interval construction with interval-aware token merging to maintain temporal coherence. Unlike prior methods that rely on local or segment-level heuristics, KiToke explicitly captures global redundancy across an entire video, leading to more efficient token utilization. Extensive experiments on multiple video understanding benchmarks and Video LLM backbones demonstrate that KiToke consistently outperforms existing training-free compression methods, with particularly large gains at aggressive retention ratios down to 1%.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。