arXiv cs.AI· Peiliang Cai, Evelyn Zhang, Jiacheng Liu, Hao Lin, Ruiqi Zhang, Weile Mo, Yue Ma, Shikang Zheng, Jiehang Huang, Dongrui Liu, Linfeng Zhang·· 14 小时前AI 评分31
Focused Forcing：面向高效自回归视频扩散的内容感知逐帧 KV 选择
Focused Forcing: Content-Aware Per-Frame KV Selection for Efficient Autoregressive Video Diffusion
AI 导读
研究者提出免训练的 KV 选择方法 Focused Forcing，在自回归视频扩散中按生成帧和注意力头两个维度聚焦缓存历史，实现最高 1.48x 端到端加速，同时提升视觉质量与文本对齐。该方法结合注意力分数与历史帧多样性分数为每个生成帧保留最相关且独特的历史帧，并按估计重要性为注意力头分配更大预算。代码已开源。
正文 · 原文
Authors:Peiliang Cai, Evelyn Zhang, Jiacheng Liu, Hao Lin, Ruiqi Zhang, Weile Mo, Yue Ma, Shikang Zheng, Jiehang Huang, Dongrui Liu, Linfeng Zhang
Abstract:Recent advances in autoregressive video diffusion have enabled sequential and streaming video generation. However, long-horizon generation requires increasingly large KV caches, making efficient compression without sacrificing quality challenging. Existing methods mostly select historical frames based on attention scores, but their context decisions remain coarse. When multiple frames are generated in the same chunk, these methods often apply a shared history selection to the whole chunk, score historical frames solely by attention, and assign head-wise budgets either uniformly or by attention-pattern heuristics rather than explicit head-importance estimation. We show that frames within the same generated chunk can depend on distinct historical frames, that the same historical frame can receive different attention scores as its relative temporal distance to the current frames changes, and that masking different heads induces unequal generation degradation. Motivated by these findings, we propose Focused Forcing, a training-free KV selection method that focuses cached history along both generated-frame and head dimensions. For each generated frame, Focused Forcing preserves the most relevant and distinctive historical frames by combining attention scores with diversity scores of historical frames, while assigning larger budgets to heads with higher estimated importance. Across multiple autoregressive generation paradigms, Focused Forcing achieves up to 1.48x end-to-end acceleration without training, while improving visual quality and text alignment. Our code is available at this https URL.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2605.18346 [cs.CV] |
| (or arXiv:2605.18346v2 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2605.18346 arXiv-issued DOI via DataCite |
Submission history
From: Peiliang Cai [view email]
[v1]
Mon, 18 May 2026 12:58:13 UTC (9,137 KB)
[v2]
Thu, 29 Oct 2026 11:32:02 UTC (9,130 KB)
来源：arXiv cs.AI · arxiv.org