研究人员开发了SQuad，一个子二次注意力蒸馏框架，旨在提高视频扩散Transformer（DiTs）的效率。该新方法降低了自注意力机制的计算成本，该机制通常随token数量呈二次方增长。SQuad实现了O(n√n)的复杂度，显著降低了FLOPs和延迟，同时保持了可比的视频生成质量。 AI
影响 这项研究通过降低扩散模型的计算需求，可能实现更高分辨率和更长时间的视频生成。
排序理由 该集群描述了一篇详细介绍一种新颖的提高AI模型效率的方法的新研究论文。
在 Hugging Face Daily Papers 阅读 →
AI 生成摘要 · Google Gemini · 来自 3 个来源。 我们如何撰写摘要 →
完整方法见我们的编辑标准。
arXiv:2608.18484v1 Announce Type: cross Abstract: Training-free block-sparse attention can accelerate video transformers, but row-wise attention concentration does not by itself specify an executable sparse operator. Queries sharing a block route may have poorly overlapping suppo…
Video Diffusion Transformers (DiTs) spend most of their compute inside the Self-Attention operation, whose cost grows quadratically, $\mathcal{O}(n^2)$, with the number of latent tokens $n$. For the task of video generation, the token count is large, so this term dominates runtim…
arXiv:2608.16585v1 Announce Type: new Abstract: Video Diffusion Transformers (DiTs) spend most of their compute inside the Self-Attention operation, whose cost grows quadratically, $\mathcal{O}(n^2)$, with the number of latent tokens $n$. For the task of video generation, the tok…