Trained on the style of Alessandro Barbucci. This version is a more stable than the Anima version and more creative than the Krea2 version, striking a good compromise. Moss prompt adherence is better than Anima, but not quite as good as Krea2. Trigger: AlessBarbuRecommended: American graphic novel style illustration with thick lines and flat color,Negatives (if not turbo and using recommended style guide prompts): speech bubbleI notice that if you prompt 'comic' or 'graphic novel' for style, Moss is likely add a speech bubble. I didn't include speech bubbles in training images, so that is all Moss. Negative for 'speech bubble' mostly alleviates that issue. If you're using a turbo model, those generally can't do negative prompts. You can try to guide it with natural language like 'no speech bubbles' but that could backfire to. It'd probably be best then to to remove 'comic' or 'graphic novel' from the style prompts and hope for the best.

Files with SHA256: d76e1cd2ba40d91357e52796ede7f8d51f76fec777a0fe0d787a3f4f297a357b