tessdata
This repository provides a range of trained data sets for Tesseract.js, designed to improve OCR accuracy. Data sets include various versions optimized for LSTM and Legacy OEMs. They are accessible via NPM packages, CDNs such as JSDelivr and Unpkg, or can be used locally. The repository facilitates a shift from deprecated resources like GitHub Pages to ensure better compliance with current size constraints. Developers can select from versions such as 'Tessdata Best,' 'Tessdata Fast,' or Tesseract v3's historic files, aligning with their language processing needs.
