Reproduce “OCRia-1 Technical Report” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Aman Rangapur, Antonio Jimeno Yepes, Daniel Bustamante, Nutchapol Dendumrongsup, Sebastián G. Acosta, Pierre-Loïc Doulcet, Simon Suo
OCRia-1 converts document pages into structured transcriptions, representing tables as HTML, charts as data tables, and formulas as LaTeX. The compact vision-language model is trained on curated real and synthetic pages and optimized for inference speed. On ParseBench, it scores 0.87 across tables, charts, formatting, and text, versus 0.75 for LightOnOCR-3-4B; it provides no bounding boxes.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.