I wanted to see if I could train a Jessica Rabbit LoRA that would keep her original cel shaded styling from the movie Who Framed Roger Rabbit. Trained with AI-Toolkit, 3500 steps, 50 images. You may need to tweak the strength between 0.8 to 1.1, depending on your checkpoint. Trained with <a rel="ugc" href="https://civarchive.com/models/1595884/hyphoria-illu-and-nai">Hyphoria [Illu &amp; NAI] </a> - you may see varied results on other checkpoints (eg SmoothMix will heavily adapt with its style) Please post images against the model - will tip those I particularly enjoy!

Files with SHA256: fdebedfb13cd04888e423f703735026e1cc80e2ce4337d54de046901070091f7