<h2 id="heading-171">Takano Suzu style LoCon</h2>Trained on 50 images by Takano Suzu. I've only tested it on the <a target="_blank" rel="ugc" href="https://civarchive.com/models/23900/anylora-checkpoint">AnyLoRA</a> 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.

Files with SHA256: ca37c94c89b0c4d4c7e945330c4a2affb3c0438cf3852b25b975a2e1f5ecf257