The model used for document information extraction however depends a lot on how to ask questions. So I wondered if there was a way to turn it into an OCR model or if there was a prompt structure to extract the best information..?
· Sign up or log in to comment