Style LoRA to replicate velminths (aka hessra) art style, based on PonyXL v6, though I recommend using it with AutismMix, for best image quality.

Since velminth is mainly an NSFW artist, the model does best with NSFW images, but is also able to generate SFW images. The LoRA is pretty good at generating lean, muscular women, so if that is your thing, you're in luck!

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

Usage

• No activation/trigger needed. Simply include the LoRA in the prompt.

• Recommended to use at strength 0.8 - 1.0

• Use the lite version (see below).

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

• Works best at ca. 1024x1024. Struggles to get the style right at lower resolutions.

• I recommend to use SDXL VAE for the must authentic recreation of the art style, since velminth mostly uses somewhat muted colors. If you want a more "anime" look, I recommend Flatpiece VAE.

Example

See preview images.

Lite version

The lite version is less than half the size of the original and does not cause a noticeable quality loss (though it does generate slightly different images, see example). I recommend using the lite version.

Notes

This LoRA was actually pretty easy to make. Even though this is only my second LoRA, I think I am starting to get the hang of it.

Using training data

See lite model for training data.
The training data (images and tags) has been sourced from rule34.xxx and cleaned up manually by me. Natural language captions have been added automatically using BLIP. When using the training data or modified versions of it to create models or datasets, please credit both me and the original artist.

Disclaimer

This LoRA is not affiliated with the artist it is based on. If you like the art style, please consider supporting the original artist on their Patreon: https://www.patreon.com/velminth (also check out their game "Saria Reclaimed" on itch.io!)