datalab-to / indra

1M param OCR model - runs on CPU, 15x faster than tesseract, close to 1B+ param models in accuracy

tesseract-ocr / tesseract

Tesseract Open Source OCR Engine (main repository)