3 ms·

Any vision model is better than commercial OCR software.

• szundi 1y ago
  [dead]
• Etheryte 1y ago
  I'm not really sure if that's an accurate summary of the state of the art, [0] is a better overview. In short, SOTA multi-modal LLMs are the best option for handwriting, nearly anything is good at printed text, for printed media, specialty models from hyperscalers are slightly better than multi-modal LLMs. [0] https://research.aimultiple.com/ocr-accuracy/ https://research.aimultiple.com/ocr-accuracy/
  • ozim 1y ago
    I see it confirms what I wrote state of art is “not using tessaract anymore” and I think bunch of commercial solutions are stuck with tessaract.
    • ares623 1y ago
      I assume Tesseract has the advantage of being able to give a confidence score?
• dragonwriter 1y ago
  Since “commercial OCR software” includes VLM-based commercial offerings, that's clearly not correct.