So this is my attempt at a NoobAI version of my Onibi Series loras. This time I used a better GPU option on google colab so it took just 3 hours rather than the usual 9+.
I'm not that good/experienced with NoobAI so I can't be certain this turned out well, but it's functional. It seems to have had trouble learning colours and hairstyles, so you have to use some descriptive tags along with the main triggers.
It was trained on the EPS version, so it works better on EPS-pred checkpoints, but it looks like it works decently with V-pred checkpoints too.
Like the previous versions of this lora, any form of commercial use is forbidden.
Description
You might want to put "colored inner hair, red streaks, low ponytail, print kimono" in the negative prompt because it has a tendency to add those features for no reason.