Feature request: swtich from to tag to raw view for caption

[StableLlama]

Working with models (esp. Flux) that use natural language captioning the current tag based view isn't easily readable, so I'd like to have the possibility to read the full caption there.

As long as there are no special ways to handle natural language (see also #267 ) a very simple work around would be an option that switches the top right "Image Tags" to a "raw" display, showing the text as it is and not in single lines/entries from the comma separated list as it is right now.

Note: a work around to the work around is to change in the settings the tag separator. Either to \n to split the lines or even to something that's not in the text like |||