Gesugao (expression) [PonyXL]

Do you like evil/sadistic women looking down on you? Me too. That's why I made this LoRA. If you like gesugaos, also check out my Ebiblue Style LoRA!

If you like the LoRA, please leave a reaction, review and/or tip. Enjoy!

Usage

• Trigger word: gesugao

• Use one (or more) support tags for desired "flavor" of gesugao:

  • evil grin

  • constricted pupils

  • evil smile

  • parted lips

  • open mouth

  • disgust

  • smug

  • shaded face

• Recommend negative: monochrome, greyscale (unless you want monochrome/black and white images)

• Recommended strength 0.8 - 1.0

• Use Clip skip=1 or Clip skip=2 (other values may also work, use at own risk)

• Works best at 1024x1024

Example

See preview images (generated with pruned version).

Unpruned vs. Pruned vs. Lite

I offer three variations of the model, which only differ in the file size and quality. Pick the one that best suits your preferences. If you can not decide, simply pick the "pruned" version. Note that the quality is relative, even the lite version usually produces images which closely resemble the intended subject.

• unpruned: The originally trained model. Largest file size, best quality. Pick if you want the best possible quality, at the cost of disk space.

• pruned: Downsized model. Almost same quality as the unpruned version, with noticably smaller file size.

• lite: Drastically downsized model. Smallest file size. Still has good quality, but noticeably worse than unpruned and pruned.

Notes

The default PonyXL model is already able to create some decent gesugaos, but I felt that something was missing and the expressions were sometimes quite inconsistent. So I made this LoRA to fix that.

I learned a lot about gesugaos while making this LoRA. Mostly, that it is difficult to limit it to a specific expression and that there are many different kinds of gesugaos. I also learned that there is a severe lack of gesugao art out there. Luckily, ebiblue exists. Their art makes up around two thirds of the data set. So since I already had the data, I decided to also make a style LoRA of their art, which I will upload in the next few days.

Using training data

Feel free to contact me if you want access to the training data.