# Backend-Agnostic Sparse Attention for Fast High-Resolution Visual Generation

arXiv: https://arxiv.org/abs/2610.08772 · HTML: https://mlsanity.com/paper/2610.08772/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Video generation, one of 14 directions in Diffusion models.
- Builds on 13 of its 39 references that are in the map.
- First posted on arXiv 6 Oct 2026.

## Placed in

- Diffusion models / [Video generation](https://mlsanity.com/diffusion/video-generation/)

## Builds on (most cited first, 13 of 13)

- [Wan: Open and Advanced Large-Scale Video Generative Models](https://mlsanity.com/paper/2503.20314/) (2025)
- [HunyuanVideo: A Systematic Framework For Large Video Generative Models](https://mlsanity.com/paper/2412.03603/) (2024)
- [Sparse VideoGen: Accelerating Video Diffusion Transformers with Spatial-Temporal Sparsity](https://mlsanity.com/paper/2502.01776/) (2025)
- [Sparse VideoGen2: Accelerate Video Generation with Sparse Attention via Semantic-Aware Permutation](https://mlsanity.com/paper/2505.18875/) (2025)
- [VSA: Faster Video Diffusion with Trainable Sparse Attention](https://mlsanity.com/paper/2505.13389/) (2025)
- [HiDiffusion: Unlocking Higher-Resolution Creativity and Efficiency in Pretrained Diffusion Models](https://mlsanity.com/paper/2311.17528/) (2023)
- [CLEAR: Conv-Like Linearization Revs Pre-Trained Diffusion Transformers Up](https://mlsanity.com/paper/2412.16112/) (2024)
- [I-Max: Maximize the Resolution Potential of Pre-trained Rectified Flow Transformers with Projected Flow](https://mlsanity.com/paper/2410.07536/) (2024)
- [Generalized Neighborhood Attention: Multi-dimensional Sparse Attention at the Speed of Light](https://mlsanity.com/paper/2504.16922/) (2025)
- [UltraImage: Rethinking Resolution Extrapolation in Image Diffusion Transformers](https://mlsanity.com/paper/2512.04504/) (2025)
- [UltraImageGen: Efficient Ultra-High-Resolution Image Generation with Hierarchical Local Attention](https://mlsanity.com/paper/2510.16325/) (2025)
- [CineScale: Tuning-Free High-Resolution Video Generation](https://mlsanity.com/paper/2508.15774/) (2025)
- [EDITBRIDGE: Towards Faithful and Efficient Ultra-High-Resolution Image Editing](https://mlsanity.com/paper/2608.18063/) (2026)

## Abstract (arXiv)

Diffusion Transformers (DiTs) have achieved strong performance in image and video generation, but the quadratic complexity of full attention makes high-resolution generation computationally expensive. Window attention offers an efficient alternative, yet existing methods face a practical trade-off: partitioned window attention typically achieves computational efficiency consistent with its theoretical complexity. However, isolated windows block cross-window interaction, often introducing visible grid-like artifacts in the generated results. Fine-grained sliding-window attention effectively restores interactions across neighboring windows and improves visual quality. However, its irregular computation patterns create a substantial gap between theoretical and practical speedups and require specialized kernels tailored to each hardware backend. To tackle these challenges, we propose BASA, a backend-agnostic sparse attention, which brings the best of both worlds: visual quality and practical acceleration. Specifically, BASA replaces visual self-attention with shifted local-window attention. By introducing a structured window-shifting scheme across DiT blocks, we allow tokens divided by window boundaries in one layer to communicate in the following layers, thereby achieving global information exchange and eliminating window-induced visual artifacts. Notably, our design introduces no additional irregular operators or customized kernels, making it readily deployable on existing attention backends and closing the gap between theoretical sparsity and practical acceleration. Experiments demonstrate that BASA achieves measured speedups exceeding 90% of the theoretical estimates on FLUX and delivers a 4.52× attention speedup on Wan while maintaining competitive generation quality. Codes are publicly available at: https://github.com/lama0110/BASA.