Woot Succubus

세부 정보

모델 설명

This LoRA generates the goat girl/succubus which comes from this single source image: https://doodleordie.com/c/uH576FLjb#4

Works best around strengths 0.6-0.8, but it does okay up to 1.0 and below 0.6 it makes similar characters.

Trained model is Toonify but it works fine with many models.

Trigger word: ws goat woman

Assisting words: ws skin, ws breasts, ws horns, ws hair, ws eyes, ws goat ears, ws hoop earring, ws lipstick, ws tail, ws star face tattoo

You don't need to include "ws" on each of these, though the effect may be stronger if you include it. Captioning it this way associates these captions more strongly with the trigger word and so they're softly on by default. Adding them will make them much stronger.

ws skin - red skin (without mentioning the color red)

ws breasts - ridiculously huge breasts, especially stratospheric with "huge breasts" or hyperfusion

ws horns - red horns with the specific shape (without mentioning red)

ws hair - the light green aqua hair of the character (without mentioning light green aqua)

ws eyes - yellow eyes with a purple or red ring (without mentioning this coloring)

ws goat ears - the distinctive goat ears of the character, differs slightly from normal demon ears

ws hoop earring - a simple golden hoop earring

ws lipstick - black/blue lipstick (without mentioning these colors)

ws tail - a demon tail (without mentioning demon)

ws star face tattoo - a black star tattoo on the face, works best close up

You can pick and choose which elements you use.

Other relevant captions: cleavage, choker, black shirt, arms bent, hands on breasts, breasts on table, smile, speech bubble, text on shirt, simple background, close up, back of head

The entire dataset uses "simple background" so if you want a background it's highly suggested that you include it as a negative prompt.

This model is intended to be used with high res fix at a scale of 2x. It has 2 native resolution pairs: 520x392 -> 1040x784 and 512x512 -> 1024x1024. That said, it seems to work fine over a wide range of resolutions.

이 모델로 만든 이미지

이미지를 찾을 수 없습니다.