Showerthought: I spent 2 months putting the whole prompt in the "style" box and my mashups were coming out mush
I've been messing with AI mashup images since March and could not figure out why my stuff always looked like one thing smeared over another, like a dog-pizza blob with no clear halves. Then I watched a guy on YouTube type "pizza cat" into the subject box and "claymation photo" into the style box, and it hit me, I was cramming everything into one field so the model just blended it all together. Split it up and my next batch actually looked like a cat wearing a pepperoni slice. Anyone have a setup that works better for keeping two things readable, or is one subject plus one style the whole trick?