Torchtext - work with images

Hi all. I’m new to nlp & torchtext. I’ve been working on an image retrieval project for a while and found this lib only very recently. Before, I was using a rather primitive dataloader written by my own to batch (text,image) pairs. I want to migrate to torchtext however haven’t found examples doing this in torchtext.

Any info related to this would be helpful. Thanks in advance!