Repository · GitHub

opendatalab/MinerU-Diffusion

[ECCV 2026] A diffusion-based framework for document OCR that replaces autoregressive decoding with block-level parallel diffusion decoding.

View on GitHubReport an issue

Owner	opendatalab
License (as declared)	MIT
Stars	620
Forks	42
Main language	Python
Created	13 March 2026
Last updated at source	18 June 2026

pythonai4sciencediffusiondlmdocument-analysisextract-datalayout-analysislladaocrparserpdfpdf-converterpdf-extractor-llmpdf-extractor-pretrainpdf-extractor-ragpdf-parser

Embed this entry

<iframe src="https://silytics.si/embed/item/0c180cbe-08ba-48c8-b9ea-3fd0c444b11f" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Free, no tracking. More about feeds and embeds.

Stars, last 3 days

620 now (+1 since 2026-10-16). Counts as reported by the source, snapshotted daily by Silytics.

More from opendatalab

MinerURepository
MinerU-HTMLRepository

Silytics shows public metadata from GitHub and links to the original. Information may be incomplete or out of date. Last refreshed 18 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.