跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.20704 · v1

QASK-Edit: Adaptive Sparse Attention Across Queries and Key Sources for Efficient Video Editing

首次提交 2026-11-16 16:12
版本更新 2026-11-16 16:12
雷达观测 2026-11-17 13:07
arXiv ↗PDF ↗

摘要

Diffusion Transformers (DiTs) have achieved strong performance in generation based video editing, but suffer from high inference latency due to the long token sequences formed by conditioning inputs and generated videos. Although training-free sparse attention reduces this cost, existing methods typically apply a fixed retention ratio or a single threshold shared across all tokens. In this work, we view video editing as conditional generation and identify a key insight: attention demand varies significantly both across different key sources (e.g., source videos, reference images, audio, text, and generated videos) and across queries within each source. Motivated by this, we propose QASK-Edit (Query-Adaptive Sparsity over Key Sources for Video Editing), a training-free sparse attention framework that allocates computation independently for each query and key source. Specifically, for each query, QASK-Edit considers both the attention mass and approximation risk of each block. It then evaluates candidate routing thresholds against each key source's risk budget, choosing exact computation or approximations to dynamically adjust sparsity for each query. Extensive experiments on three video editing models demonstrate that QASK-Edit achieves up to 1.46×1.46\times end-to-end speedup over dense attention pipeline while preserving both editing quality and unedited region fidelity. Furthermore, QASK-Edit generalizes directly to Reference-to-Video generation, delivering up to 1.36×1.36\times speedup while maintaining generation quality.

AI 论文解读

基于摘要 · 非全文审阅

此版本尚未生成 AI 解读。解读由后台任务生成并按论文版本保存,登录后共享已有结果。

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-16 16:12 · 最新观测版本

QASK-Edit: Adaptive Sparse Attention Across Queries and Key Sources for Efficient Video Editing

Diffusion Transformers (DiTs) have achieved strong performance in generation based video editing, but suffer from high inference latency due to the long token sequences formed by conditioning inputs and generated videos. Although training-free sparse attention reduces this cost, existing methods typically apply a fixed retention ratio or a single threshold shared across all tokens. In this work, we view video editing as conditional generation and identify a key insight: attention demand varies significantly both across different key sources (e.g., source videos, reference images, audio, text, and generated videos) and across queries within each source. Motivated by this, we propose QASK-Edit (Query-Adaptive Sparsity over Key Sources for Video Editing), a training-free sparse attention framework that allocates computation independently for each query and key source. Specifically, for each query, QASK-Edit considers both the attention mass and approximation risk of each block. It then evaluates candidate routing thresholds against each key source's risk budget, choosing exact computation or approximations to dynamically adjust sparsity for each query. Extensive experiments on three video editing models demonstrate that QASK-Edit achieves up to 1.46×1.46\times end-to-end speedup over dense attention pipeline while preserving both editing quality and unedited region fidelity. Furthermore, QASK-Edit generalizes directly to Reference-to-Video generation, delivering up to 1.36×1.36\times speedup while maintaining generation quality.

查看此版本解读 →查看此版本原文 ↗