Indra-ocr format output xml

#1
by Soufiane22 - opened

Hi Team,

Thank you for providing this powerful OCR pipeline. As an AI Engineer, I can truly appreciate the amount of work behind such a solution. It's especially impressive to achieve these results with only 1.25 million parameters, which keeps the memory footprint extremely low. Great job, and thank you to the whole team for this achievement!

I have one question: is there any possibility of obtaining the output in XML format? From what I can see, the available formats are currently text, JSON, and Markdown, but XML would be very useful for some integration scenarios.

I also ran a few tests on handwritten documents. While the results on printed text are quite good, I noticed that handwriting recognition is still challenging in some cases. For example, certain handwritten letters such as "a" may occasionally be recognized as "o". That said, handwriting OCR is a difficult problem in general, so this observation is meant as constructive feedback rather than criticism.

Overall, I'm impressed by the performance and efficiency of the solution. Please keep up the great work!

Best regards,

Sign up or log in to comment