Now if we could get a Qwen-Image-Layered 2 also at 7B ... ^^

(Or does this model has those capabilities natively ?)