There were only 5 images in the image sample with 125 steps, so it will be refined in the future.
Trained on abyss orange mix 2 NSFW, but works with less power on other models as well.
It is better to set the Lora force 0.5-0.9
Future improvements for the model:
1. Add more images and edit promt
2.
get rid of the very occasional oblong texture line artifacts3. Detail the body
Description
trained on 5 portrait images
FAQ
Comments (3)
This lora was so good! Do you plan to make a ponyxl version of it?
Yes, I think I will soon make a lara on this topic too ^-^
Released for PonyXL, you can read it here
https://civitai.com/models/416456/lerapi-style-pony-xl?modelVersionId=464001