发表机构
KlingAI Research（克林人工智能研究公司）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究针对自回归视频扩散模型注意力成本高的问题，提出HeadCast框架，基于预训练模型注意力头行为进行一次性分类，重组KV缓存，保留全局头，在不同分辨率下有效加速推理且保持质量，减少闪烁。
AI 中文摘要
自回归（AR）视频扩散模型已成为长视频和流式视频合成的一种有前景的范式，但不断增长的键值（KV）缓存使注意力成为主要推理成本，尤其是在高分辨率下。现有补救措施要么采用粗略启发式方法清除缓存导致帧间闪烁，要么需要重新训练模型。我们提出了HeadCast，这是一个无需训练、即插即用的加速框架。基于预训练AR模型的注意力头表现出稳定、异质行为这一观察结果，HeadCast在最大噪声步骤进行一次性分类，将每个头分为四种原型之一，并将整体KV缓存重组为特定于头的路径。它保留了全局头以保持长程时间一致性。在跨最先进的AR模型中，HeadCast在720P时将推理加速高达1.62倍，在1080P时加速1.95倍，同时保持VBench质量与全注意力相当且基本无闪烁。
英文摘要
Autoregressive (AR) video diffusion models have become a promising paradigm for long and streaming video synthesis, but the continuously growing Key-Value (KV) cache makes attention the dominant inference cost, especially at high resolution where each frame contributes many tokens. Existing remedies either evict the cache with coarse heuristics that cause inter-frame flickering, or require model re-training. We propose HeadCast, a training-free, plug-and-play acceleration framework built on the observation that a pre-trained AR model's attention heads exhibit stable, heterogeneous behaviors. After a short warm-up, HeadCast performs a one-time classification at the maximum-noise step that sorts every head into one of four archetypes: Sink, Dummy, Spatial, and Global, and restructures the monolithic KV cache into head-specific pathways. Crucially, it retains the Global heads that preserve the long-range temporal consistency aggressive eviction destroys. Because the Spatial pathway operates on a fixed-size grid, its savings grow with resolution: across state-of-the-art AR models, HeadCast accelerates inference by up to 1.62x at 720P and 1.95x at 1080P, while keeping VBench quality on par with full attention and largely flicker-free. Code is available at https://github.com/sjlgaga/HeadCast .
