Takano Suzu style LoCon

Trained on 50 images by Takano Suzu. I've only tested it on the AnyLoRA checkpoint and i would recommend using it. Use a weight around 0.8 and a CFG Scale of 7 for best results. It suffers a little bit from being trained on mosaic images and i probably should have put a combination of black bar and mosaic images but it is what it is, I may redo it in the future. It works fine with character loras.

I recommend using comfyui but I don't really know if it makes any real difference. My workflow is in the metadata of the images if you really want it but it's just something i've thrown together.