Skip to content

Dataset · bevaya

PubMed-OCR: scientific PDF layout markup

bevaya/pubmed-ocr

PubMed Central page OCR markup with bounding boxes for layout-aware tasks.

What's inside

Corpus of scientific articles from PubMed Central Open Access PDFs: each page is rendered to an image and annotated with Google Cloud Vision OCR. Markup is in compact JSON with bounding boxes at word, line, and paragraph levels. Scale: 209.5k articles, about 1.5M pages, and 1.3B OCR tokens; English language.

What you can do with it

  • Training OCR and layout-aware models
  • Coordinate-grounded page QA
  • Extracting text from scientific PDFs
Tasks
Image → text, Image text → text
Languages
English
Size
1M–10M rows
Downloads this month
2.1K
Likes
73

How to load it

from datasets import load_dataset

ds = load_dataset("bevaya/pubmed-ocr")