Metraipolis › archive › dc_4b5dcb08abac

Elastic Token Compression for Pixel-Space Diffusion Transformers — arXiv:2608.29281

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.29281
License (as archived)
CC0-1.0
Archived
2026-10-17T09:58:54.154Z
SHA-256 of the archived text
9b2fb468b1fe0d47e13797ffce8bd91f837c691041813723394d98216d8e2a11
Stable URL of this page
https://metraipolis.com/documento/dc_4b5dcb08abac

How to cite

Elastic Token Compression for Pixel-Space Diffusion Transformers — arXiv:2608.29281 — arXiv (en), https://arxiv.org/abs/2608.29281 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4b5dcb08abac (Metraipolis archive dc_4b5dcb08abac, sha256 9b2fb468b1fe0d47e13797ffce8bd91f837c691041813723394d98216d8e2a11)
BibTeX
@misc{dc4b5dcb08abac,
  title = {Elastic Token Compression for Pixel-Space Diffusion Transformers — arXiv:2608.29281},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.29281},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4b5dcb08abac; sha256 9b2fb468b1fe0d47e13797ffce8bd91f837c691041813723394d98216d8e2a11},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.29281. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Natural images concentrate their detail in a small fraction of the frame, yet diffusion models spend a full token on every patch, in every layer and at every timestep. The waste is largest in pixel-space models, with no autoencoder to absorb low-level redundancy first. Probing a pretrained pixel text-to-image transformer, we find its middle-block tokens redundant wherever the image is flat. The redundancy occupies connected, content-shaped regions, and exploiting it requires tokens with the same geometry. Cutting a Hilbert ordering of the patches provides them. Consecutive positions are always image neighbours, so any contiguous run is a connected region whose size and shape follow the content, and grouping in two dimensions becomes a cut in one. Existing reductions each lose part of this. Similarity merging scatters its groups, latent bottlenecks discard position, and skipping deletes what it should summarize. We cut where the model's features change most and pool each run into one region token. Our Region Token Interface (\method{}) adapts a diffusion model to these tokens, with the region count drawn at random during fine-tuning so one checkpoint serves every budget. \method{} leads prior reduction methods at matched budgets, matches dense quality at $2.0\times$ the speed, and stays close at $2.6\times$. The code and models are open-sourced at this https URL

Tema degli esploratori: statistica e scienza dei dati (retro-catalogo arXiv: annunci del 2026-09-01 in cs.CV).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV).
Pagina del lavoro: https://arxiv.org/abs/2608.29281 — 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 2026-10-17T09:28:48.019Z dalla pagina https://arxiv.org/catchup/cs.CV/2026-09-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).