2 ms·

Now if we could get a Qwen-Image-Layered 2 also at 7B ... ^^ (Or does this model has those capabilities natively ?)