Lunala LoRA trained on NovelAI with Kohya training scripts using the DAdaptAdam optimizer at 512x512 resolution. Very spicy, I recommend adding (dark purple skin, wings) to your prompt to improve consistency. Skin color and patterns are inconsistent. Wing pattern and shape can be unstable. Recommended to lower the LoRA strength if you're running into issues. Despite the problems, good results are possible. Should have wide compatibility with model mixes based on NovelAI, which is most of them. Trained on an all female dataset.
Description
Initial release
FAQ
Comments (3)
It's a shame this 'mon is so hard to work with when it comes to making Loras because it's a very lovely one 🦇
Thank you for doing it!
Lunala is such a cool design, and is very underrated if you ask me. Because the design is so complex and doesn't anthrofy very easily, there's no real agreement on how to draw an anthro Lunala. It was pretty hard getting semi-consistent training images because of that. I'm pretty happy with how it came out with the limited training data I had.