# Ride the Wave: Precision-Allocated Sparse Attention for Smooth Video Generation Source: [Ride the Wave: Precision-Allocated Sparse Attention for Smooth Video Generation]() Collection index: [llms.txt]() Data updated: 2026-04-16T05:00:04.684Z Authors: Wentai Zhang, Ronghui Xi, Shiyao Peng, Jiayu Huang, Haoran Luo, Zichen Tang, Haihong E Published: 2026-04-14T02:51:52.000Z ## Abstract Video Diffusion Transformers have revolutionized high-fidelity video generation but suffer from the massive computational burden of self-attention. While sparse attention provides a promising acceleration solution, existing methods frequently provoke severe visual flickering caused by static sparsity patterns and deterministic block routing. To resolve these limitations, we propose Precision-Allocated Sparse Attention (PASA), a training-free framework designed for highly efficient and temporally smooth video generation. First, we implement a curvature-aware dynamic budgeting mechanism. By profiling the generation trajectory acceleration across timesteps, we elastically allocate the exact-computation budget to secure high-precision processing strictly during critical semantic transitions. Second, we replace global homogenizing estimations with hardware-aligned grouped approximations, successfully capturing fine-grained local variations while maintaining peak compute throughput. Finally, we incorporate a stochastic selection bias into the attention routing mechanism. This probabilistic approach softens rigid selection boundaries and eliminates selection oscillation, effectively eradicating the localized computational starvation that drives temporal flickering. Extensive evaluations on leading video diffusion models demonstrate that PASA achieves substantial inference acceleration while consistently producing remarkably fluid and structurally stable video sequences. ## Recorded benchmark results A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. - [Text-to-Video Generation on Wan 720P 2.1-14B](): Text-to-Video Generation; Wan Text-to-Video 720P 2.1-14B; T.F.: 98.48. - [Text-to-Video Generation on Hunyuan-13B (720P)](): Text-to-Video Generation; Hunyuan-13B Text-to-Video 720P; Temporal Fidelity (T.F.): 99.32. - [Text-to-Video Generation on Wan 1.3B Text-to-Video 480P 2.1](): Text-to-Video Generation; Wan 1.3B Text-to-Video 480P 2.1; Temporal Fidelity: 97.06. ## Original sources and implementation [Original paper]() [PDF]()