Resequencing frames

[Tortoise17]

I am trying to look for a solution where I could re-sequence or resort images based on text. Like if I have three images in a folder, the image matches to first sentence takes first place. Is it possible with this kind of approach in some way? If anyone can help and guide.

[mvsoom]

You could try https://github.com/mapluisch/LLaVA-CLI-with-multiple-images.