This hypernetwork can generate images in a cartoon, anime, or realistic styles. It's not great, but hey, it works! The original cartoon style can sometimes be tricky.
The hypernetwork is trained on model Any v4.5, but other models, especially in the anime style, are fine too.
Activation tags - FionnaTheHuman, fionna_the_human_girl
For bunny ears don't forget to add - aimal_ears, bunny_ears, animal_hood, bunny_hood, white hood
Also don't forget to play with hypernet strength I prefer somewhere in the 0.6-1
There is room for improvement, and I might do a version 2 if this hypernetwork becomes popular.
Description
FAQ
Comments (2)
this looks fun to mess around with, love fiona! what made you wanna do a hypernetwork tho out of curiosity
man I wonder how training would even begin to try to under stand Jake or Cake lol
Well I just have been watching a lot of adventure time laitly. And when I found that there is no hypernet or LoRA with Fionna then I decided that I must made one! Train your hypernet is not so hard acctually. The hardest part is to collect good dataset and create good tags for it.