No trigger words! Just use the LoRA, suggested weight is 0.5-0.8:

After the latest ComfyUI update, I cannot recreate ANY of the sample images. Looks like weight 1.0 is WAY too much for the LoRA. 0.7 gets very close to the sample images.

Disclaimer: This is an experimental LoRA! Just trying to see what Z-Image-Turbo is capable; still waiting for the full base model to be released. The base model can barely do NSFW and Nudity and the LoRA Dataset is 100% SFW (its just official artwork!), don't expect the LoRA to do NSFW!

Start your prompt with: "A detailed comic book illustration by Clay Mann..." or "An digital illustration by Clay Mann featuring..." or "A hand-drawn illustration by Clay Mann...". Just make sure to mention the media (illustration, comic book, digital, drawing, paper, etc) and the artist name; the base model is great with Natural Language and long prompts; try to avoid tag based prompting.

ComfyUI Metadata is included in the sample images! Click the little icon in the bottom right to see it.

Please check out my Ko-Fi page if you are interested in supporting me. LoRA commissions for SD1.5, SDXL(Pony, Illustrious and Photoreal) and FLUX open!

It was trained on 107 1024x1024 images scraped from the internet and extracted from scanned comic books.