naver-clova-ix

/

donut-base

like 181

Follow

NAVER CLOVA INFORMATION EXTRACTION 43

I am trying to export intermediate result of DONUT base model and use it similar to OCR just to extract entire text of a given document. Don't want to do classification, VDQ, parsing. Is there any way to do so? Basically, given a doc, extract as much text as you can.

Thanks