Metraipolis › archive › dc_97a9ae02ffd2
Unlimited OCR Works 2 — arXiv:2610.39308
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.39308
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-31T06:16:59.979Z
- SHA-256 of the archived text
228297e27e74789c9bb8957ae3ec6e75e554d5e7a990c8e15dbb7e4564e2b099- Stable URL of this page
https://metraipolis.com/documento/dc_97a9ae02ffd2
How to cite
Unlimited OCR Works 2 — arXiv:2610.39308 — arXiv (en), https://arxiv.org/abs/2610.39308 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_97a9ae02ffd2 (Metraipolis archive dc_97a9ae02ffd2, sha256 228297e27e74789c9bb8957ae3ec6e75e554d5e7a990c8e15dbb7e4564e2b099)
BibTeX
@misc{dc97a9ae02ffd2,
title = {Unlimited OCR Works 2 — arXiv:2610.39308},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.39308},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_97a9ae02ffd2; sha256 228297e27e74789c9bb8957ae3ec6e75e554d5e7a990c8e15dbb7e4564e2b099},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.39308. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Unlimited OCR ushers the era of long-horizon parsing, substantially mitigating the growth of working memory as document lengthens. However, as document length scales to hundreds of pages, the visual context grows substantially, sharply increasing the computational cost of each decoding step and limiting the scalability of ultra-long document parsing. Inspired by human book transcription via localized focus without re-scanning preceding pages, we introduce Unlimited OCR 2, a document parsing model that looks only at what it transcribes. Building on Unlimited OCR, we upgrade all the decoder attention modules to Sparse Reference Sliding Window Attention (SR-SWA), bounding active reference tokens to a fixed capacity and thus sustaining constant computational costs across arbitrarily long documents. By combining SR-SWA with Unlimited-Horizon Decoding Paradigm, Unlimited OCR 2 enables document parsing over virtually unlimited lengths without prohibitive memory overhead. At the hundred-page scale, Unlimited OCR 2 achieves 7.0$\times$ and 5.6$\times$ per-step speedups over Full Attention and R-SWA, respectively. Furthermore, it delivers state-of-the-art end-to-end performance on both MPDocBench and MarathonDocBench, our newly proposed benchmark targeting hundred-page document parsing. Ultimately, the capacity for uninterrupted, one-shot document parsing signifies a fundamental paradigm shift into the OCR 3.0 era. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.CL, nuovi lavori del giorno). Categorie arXiv: Computation and Language (cs.CL) ; Computer Vision and Pattern Recognition (cs.CV). Pagina del lavoro: https://arxiv.org/abs/2610.39308 — 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 2026-10-31T06:07:23.355Z dalla pagina https://arxiv.org/list/cs.CL/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).