This is a LoRA that is designed to make your images look like an older graphic novels. It should have visible outlines, strong colors, and simplified characters. I recommend running it around 0.8 strength and using clip skip 2. Its not a fully developed dataset and definitely favors modern aesthetics over fantasy or sci fi. As with many pony models its likely to put the people in dynamic poses and should respond well to typical pony prompting concepts. the dataset was tagged using simplified nouns and adjectives, not natural language, so prompt with that in mind.I recommend always prompting your backgrounds otherwise it will be pretty random but leans towards solid colors. It has a tendency to make men rather muscular and to make women's faces be somewhat standardized, this can be fixed with simple prompts though.

Graphic Novel Style

First version, simplified dataset.

Files with SHA256: 5545aaeeea23f4da6bc079bd6bcb32288956a1ddc49c98049afa665b551967f7