cs.CVNov 18, 2026

Spark-H3: Adaptive Block Sparse Attention for Video Diffusion Models

Authors: Zecheng Tang, Jiaye Fu, Youtian Lin, Xinyao Rao, Liuxuan Sun, Jian Zhang, Yao Yao, Wei Li

Organizations: Peking University · Nanjing University

Abstract

Long video sequences make attention a substantial cost in video diffusion models. Block sparse attention (BSA) reduces this cost by computing selected blocks exactly while compressing the remaining blocks. However, fixed token grouping can place tokens with dissimilar attention preferences in the same group, while uniform key averaging can underestimate the attention mass of compressed blocks. We introduce Spark-Attn, a training-free BSA method comprising Spark-Reblock and Spark-Reweight. Spark-Reblock recursively groups tokens by attention preference while preserving regular block shapes for efficient GPU execution. Spark-Reweight uses the global mean query as an anchor to construct attention-weighted key-value representations and logit biases for the compressed blocks, calibrating their attention mass. We instantiate Spark-Attn in MiniMax-H3 as Spark-H3. At 10% exact-block density, Spark-H3 delivers a 2.49×\times DiT speedup and reaches 24.81 dB PSNR against the corresponding dense outputs on MiniMax-H3 (20 denoising steps, 345 frames, 1344×\times768). Increasing the density to 20% further improves PSNR to 26.76 dB while still achieving a 2.26×\times DiT speedup. These results demonstrate that Spark-H3 achieves a state-of-the-art quality-efficiency trade-off for training-free BSA on video diffusion models.

Figures & tables

Explore similar work

CardsList
  1. AB-Attention: Accurate Block Attention for Accelerating Video Diffusion Transformers

    Oct 29, 2026Bin Xu, Mo Zhang, Zhengfeiyang Zhang +6Diffusion Model Inference AccelerationBlock-Sparse Attention

  2. SPADE: An Input-Adaptive Sparse Attention Engine for Fast Video Diffusion Models Inference

    Aug 4, 2026Shanghao Liu, Renze Chen, Size Zheng +4Video Diffusion ModelsDiffusion Transformer