When vectorizing a pdf to later talk to it. Can we also store and retrieve an image?
For example my pdf has this line:
I used this model:‘sentence-transformers/multi-qa-mpnet-base-dot-v1’ and retrieve the line from the pdf with this result:•To re-engage Super Cruise, press ].
So in this case the image was replaced by ‘]’. Which is the same symbol I get if I try to copy paste the image from my pdf.
So my question is. Is there a way or model which will allow me to retrieve the text an image?