Metraipolis › archive › dc_d55dd540e17f

Unlocking Any-Order Generation in Pretrained Autoregressive Image Models — arXiv:2610.39293

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.39293
License (as archived)
CC0-1.0
Archived
2026-11-01T02:27:08.314Z
SHA-256 of the archived text
bdbd74e2bd0571ef67ae76ba6d4324a494d60e7d1e3b7e73c68fff89e7d6e022
Stable URL of this page
https://metraipolis.com/documento/dc_d55dd540e17f

How to cite

Unlocking Any-Order Generation in Pretrained Autoregressive Image Models — arXiv:2610.39293 — arXiv (en), https://arxiv.org/abs/2610.39293 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d55dd540e17f (Metraipolis archive dc_d55dd540e17f, sha256 bdbd74e2bd0571ef67ae76ba6d4324a494d60e7d1e3b7e73c68fff89e7d6e022)
BibTeX
@misc{dcd55dd540e17f,
  title = {Unlocking Any-Order Generation in Pretrained Autoregressive Image Models — arXiv:2610.39293},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.39293},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d55dd540e17f; sha256 bdbd74e2bd0571ef67ae76ba6d4324a494d60e7d1e3b7e73c68fff89e7d6e022},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.39293. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Autoregressive (AR) image generators are powerful in principle but inflexible in practice: training with a fixed raster-scan order hardcodes a single factorization of the image distribution into the model, leaving it unable to inpaint, outpaint, or edit at inference time. Existing solutions, including the modern discrete-diffusion family, train from scratch on a mixture of token orderings, paying roughly 3x the compute of standard raster-scan training; discrete diffusion further sacrifices compatibility with the AR inference ecosystem (KV caching, vLLM) by construction. We argue that this cost is unnecessary. A pretrained raster-scan AR model has already learned the joint distribution over image tokens; what it lacks is not knowledge but the ability to express that knowledge through non-sequential conditional pathways. We propose a recipe for adapting pretrained raster-scan models into any-order generators using two components: future-aware positional embeddings that resolve a target-position identifiability problem at negligible overhead, and an ordering curriculum that traces a stable optimization path from the pretrained solution to a good any-order basin. Applied to four base models across two architectures and two scales (LlamaGen-L/XL, RAR-L/XL), the recipe substantially outperforms MaskGIT on ImageNet inpainting (FID 6.21 vs. 10.44) and outpainting, is competitive with MAR-L, trades a 1.2-1.9 increase in class-conditional FID for this flexibility, and costs 13-33% additional compute beyond pretraining, while leaving the standard AR inference stack untouched. Inpainting, outpainting, and editing, our results suggest, need not be added to autoregressive image models. They can be unlocked from the ones that exist. Code is available at this https URL .

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.LG, nuovi lavori del giorno).
Categorie arXiv: Computer Vision and Pattern Recognition (cs.CV) ; Machine Learning (cs.LG).
Pagina del lavoro: https://arxiv.org/abs/2610.39293 — 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 2026-11-01T02:12:44.294Z dalla pagina https://arxiv.org/list/cs.LG/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).