正文 · AI 翻译
IT Home reported on October 30 that Baidu officially released and open-sourced its ERNIE Unlimited OCR 2 model on October 27. The model pushes the document parsing limit from dozens of pages to “unlimited,” and can parse all 7 Harry Potter books, totaling more than 3,600 pages, in a single pass.
Unlimited OCR 2 natively supports the segmentation and recognition of layout elements, as well as merging text across pages and recognizing and merging tables. With the same computing resources, it processes more, faster, and at lower cost.
Baidu said ERNIE Unlimited OCR 2 scored 84.6 on MPDocBench, a benchmark for parsing multi-page documents, and 77.1 on MarathonDocBench, which specifically evaluates the ability to parse very long documents of around 100 pages. It achieved state-of-the-art end-to-end performance on both.
Reference
https://ai.baidu.com/support/news?action=detail&id=3290
Advertising disclosure: External links in this article (including but not limited to hyperlinks, QR codes, and access codes) are provided to convey more information and save time in selecting sources. Results are for reference only. All IT Home articles containing external links include this disclosure.