Metraipolis › archive › dc_1fb77dec612f
iCATS: Fast Video Generation via Interaction-Aware Sparse Attention and Timestep-Adaptive Sparsity — arXiv:2610.11302
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.11302
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-09T06:30:27.064Z
- SHA-256 of the archived text
114f94c3e4bdd9879995d2939c3e4fa0fe170fcc5781c95b237b592d530595f4- Stable URL of this page
https://metraipolis.com/documento/dc_1fb77dec612f
How to cite
iCATS: Fast Video Generation via Interaction-Aware Sparse Attention and Timestep-Adaptive Sparsity — arXiv:2610.11302 — arXiv (en), https://arxiv.org/abs/2610.11302 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_1fb77dec612f (Metraipolis archive dc_1fb77dec612f, sha256 114f94c3e4bdd9879995d2939c3e4fa0fe170fcc5781c95b237b592d530595f4)
BibTeX
@misc{dc1fb77dec612f,
title = {iCATS: Fast Video Generation via Interaction-Aware Sparse Attention and Timestep-Adaptive Sparsity — arXiv:2610.11302},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.11302},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_1fb77dec612f; sha256 114f94c3e4bdd9879995d2939c3e4fa0fe170fcc5781c95b237b592d530595f4},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.11302. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Training-free sparse attention offers a practical acceleration solution to Diffusion Transformers (DiTs) via reducing computations without fine-tuning. It typically involves estimating the importance of query-key regions and deriving sparse masks to compute only the important candidates, which inevitably introduces approximation errors that may degrade generation quality. To better balance the efficiency-quality trade-off, we propose iCATS, integrating improved importance estimation and sparse mask construction with an efficient hardware execution strategy. Specifically, for importance estimation, unlike previous works that perform independent clustering over query and key tokens based on feature similarity to estimate attention scores, iCATS demonstrates that clustering based on query-key dot-product interactions is more accurate and further reformulates this objective as a simple quadratic form for low-cost computation. For sparse mask construction, instead of using a fixed top-p rule, we observe that tolerance to sparse approximation errors varies across denoising timesteps and therefore introduce an SNR-guided sparsity schedule to adjust sparsity dynamically, leading to higher accuracy. Finally, for hardware execution, we devise a tail-merging strategy to reduce padding overhead caused by irregular cluster sizes, improving GPU kernel utilization. Extensive experiments show that iCATS achieves $2.03\times$ acceleration with 31.017 dB PSNR on HunyuanVideo-T2V-13B and $1.55\times$ acceleration with 29.301 dB PSNR on Wan2.1-T2V-14B, delivering a state-of-the-art efficiency-quality trade-off. Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.CV, nuovi lavori del giorno). Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2610.11302 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-09T05:55:07.295Z dalla pagina https://arxiv.org/list/cs.CV/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).