For this lora I once again went with as many steps as I could without causing too much overtraining. It was a sacrifice I deemed necessary to make the lora as accurate to Hanai's style as possible. Of course, that does mean that you will occasionally experience some overtrained aspects (text, speech bubbles, certain clothes, censors, etc.), but if you use your negatives and prompt well then it shouldn't be a big problem.Occasionally prompting lips, shiny skin, long eyelashes, eyeshadow (glitter eyeshadow, black eyeshadow, purple eyeshadow, etc.), or other more descriptive details in either positives or negatives will let you generate different variations of Hanai's style found in the manhwa.I won't speak much about the content of the manhwa given how controversial it may be for some...but I will say that the art is absolutely gorgeous in some panels. I also haven't 'read' the manhwa either, actually. I skimmed through the chapters looking for the best images to use as training data, so please don't post any spoilers for those who may be interested in the series, of course.Recommended weight = 1, but as always, feel free to play around with the weight.

Files with SHA256: adc33c12876548da5d8914c3dd712bc47f54f1dc56c118f2e5ced8b8bf04817b