Diffusion Transformers (DiT) excel at image and video generation but face
computational challenges due to self-attention's quadratic complexity. We
propose DiTFastAttn, a novel post-training compression method to alleviate
DiT's computational bottleneck. We identify three key redundancies in the
attention computation during DiT inference: 1. spatial redundancy, where many
attention heads focus on local information; 2. temporal redundancy, with high
similarity between neighboring steps' attention outputs; 3. conditional
redundancy, where conditional and unconditional inferences exhibit significant
similarity. To tackle these redundancies, we propose three techniques: 1.
Window Attention with Residual Caching to reduce spatial redundancy; 2.
Temporal Similarity Reduction to exploit the similarity between steps; 3.
Conditional Redundancy Elimination to skip redundant computations during
conditional generation. To demonstrate the effectiveness of DiTFastAttn, we
apply it to DiT, PixArt-Sigma for image generation tasks, and OpenSora for
video generation tasks. Evaluation results show that for image generation, our
method reduces up to 88\% of the FLOPs and achieves up to 1.6x speedup at high
resolution generation.

Diffusion Transformers（DiT）在图像和视频生成方面表现出色，但由于自注意力的二次复杂度而面临计算挑战。我们提出 DiTFastAttn，这是一种新颖的后训练压缩方法，用于减轻 DiT 的计算瓶颈。我们确定了 DiT 推理过程中注意力计算中的三个关键冗余：1. 空间冗余，即许多注意力头部关注局部信息；2. 时间冗余，即相邻步骤的注意力输出之间存在高相似性；3. 条件冗余，即条件和无条件推理之间存在显著相似性。为了解决这些冗余，我们提出了三种技术：1. 带残留缓存的窗口注意力，以减少空间冗余；2. 时间相似性减少，以利用步骤之间的相似性；3. 条件冗余消除，以在条件生成过程中跳过冗余计算。为了证明 DiTFastAttn 的有效性，我们将其应用于 DiT、PixArt-Sigma 进行图像生成任务，以及 OpenSora 进行视频生成任务。评估结果显示，对于图像生成，我们的方法减少了高分辨率生成中高达 88％的 FLOPs，实现了高达 1.6 倍的加速。