dandelin

/

vilt-b32-finetuned-nlvr2

like 2

Hey for my use-case I would need to adapt the model to not only handle 2, but 21 images for the Answer creation. Can somebody guid me on how to adapt the model to do this? Thanks for you help :)!