When training models, it’s essential to describe the content of your dataset. Creating files that detail image content becomes crucial. Now, regarding taggers or interrogators: I primarily use WD 1.4 variants, but for more complex scenes, I’m interested in tools that generate full sentences. I recall finding such a tagger previously but lost it during a reinstall. So, what yields the best results for tagging your datasets? CLIP, BLIP, or perhaps something else? And finally, what’s your favorite tool or UI for tagging?