QuillworksIllustrious

At its core, Quillworks18 represents a return to form— V15 merged back into Hassaku v2.2 as a foundation for its well-balanced structure and strong baseline fidelity. However, this iteration takes things further with a newly trained LoRA focused on digital art and Western cartoon sensibilities built directly into the model. This targeted fine-tuning helps guide the model’s output toward more deliberate forms, tighter linework, and stronger shape language without becoming locked into any single style.

What's New in Quillworks18?

• Improved Output Reliability This version tones down some of the wilder stylistic fluctuations seen in earlier versions. The goal was to create a model that delivers more predictable and prompt-faithful results, especially with character-focused art and single-subject compositions.
• Streamlined Style Handling While styles like sketch and heavily textured painting have been dialed back to reduce inconsistency, Quillworks18 still retains excellent support for stylistic exploration. Watercolor, flat cartoon, anime, and graphic ink remain fully usable and responsive through prompt control, giving creators flexibility without chaos.
• Western Cartoon Influence This version leans slightly into Western character design, borrowing from clean vector aesthetics, expressive faces, and simplified forms without losing the flexibility that Quillworks users appreciate. It plays well with both digital realism and more stylized approaches. It can still do some sketch styles but requires a lot heavier prompting to get the effects.

The model works bes* ***ng DPM++ Karras, I like using CFG 4.5-6.0 but mostly 5.0.

I like to run it between 20-30 steps.

Its still an illustrious based model and will struggle with hands, faces, and eyes at times  but this should give a bit more consistency. Full body portraits, especially with a glow effect will often cause issues with the eyes. Use face fixers for that.

Prompting Guide: "masterwork quality,  [styles like cartoon anime watercolor], [main subject], [background], [extra details]"

Please explore style prompting, it leans towards digital art now but its still capable of producing many many styles when prompted with care and creativity.