Metraipolis › archive › dc_2cb64b8dbb1d

SpectralCache: Accelerating Diffusion-Based World Models via Spectral Feature Caching — arXiv:2610.02660

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.02660
License (as archived)
CC0-1.0
Archived
2026-10-06T09:07:04.293Z
SHA-256 of the archived text
d648084aea26b4d6f21a3b7f5e60e5133943f1b39d247ec47f9f06ebbb08b521
Stable URL of this page
https://metraipolis.com/documento/dc_2cb64b8dbb1d

How to cite

SpectralCache: Accelerating Diffusion-Based World Models via Spectral Feature Caching — arXiv:2610.02660 — arXiv (en), https://arxiv.org/abs/2610.02660 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2cb64b8dbb1d (Metraipolis archive dc_2cb64b8dbb1d, sha256 d648084aea26b4d6f21a3b7f5e60e5133943f1b39d247ec47f9f06ebbb08b521)
BibTeX
@misc{dc2cb64b8dbb1d,
  title = {SpectralCache: Accelerating Diffusion-Based World Models via Spectral Feature Caching — arXiv:2610.02660},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.02660},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2cb64b8dbb1d; sha256 d648084aea26b4d6f21a3b7f5e60e5133943f1b39d247ec47f9f06ebbb08b521},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.02660. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Diffusion-based world models enable high-quality interactive environment generation but suffer from substantial inference overhead due to repeated Transformer evaluations during denoising. Existing caching methods mainly exploit temporal redundancy at the feature or token level, leaving the underlying mathematical structure of diffusion features largely unexplored. In this work, we reveal that world-model features exhibit highly stable singular subspaces across nearby denoising steps, while their singular values follow predictable evolution patterns. Building on this observation, we propose SpectralCache, a training-free spectral caching framework that reuses stable singular subspaces and estimates only low-dimensional singular values through linear extrapolation. We further exploit the spectral consistency between neighboring full-computation features to skip selected expensive backbone evaluations via singular value scaling. Extensive experiments on representative world models demonstrate that SpectralCache consistently improves inference efficiency while preserving generation quality. On HunyuanWorld-Voyager-13B, SpectralCache achieves 5.22x acceleration while maintaining a WorldScore of 65.90 for static scenes, substantially outperforming existing training-free caching methods in inference efficiency.

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-10-05 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2610.02660 — 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 2026-10-06T08:21:44.163Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-10-05?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).