Tag (images -> for now)

Use two AI models to generate tags for images. The tags will be saved in a json with the same name as the inputs file and script works recursively through an image folder. One of the models need keywords as input. I put some examples, but i do not think to much keywords is a good idea, so i only use the custom- one.

Whats next

Actually make interchangeable objects that gives tags per file
handle videos: (idea: split using ffmpeg - tag those and somehow combine into taglist)
ui
progress (collect files and print progress)