First lora release.
This is experimental version, a bit overtrained, follow sample image for better use.
Keep trigger word at the first place in your prompt, and lora load at the end:
kkw-comicP, prompt, <lora:kkw-comicP:0.6>
Trained with kohya_ss.
I'll work on it for future update, maybe better result but posting here for see how it work on different checkpoint.
Keep prompt simple, not so much study.