Recommended strength: 0.75This lyco was trained on a 680 images of output from Envy Ephemeral and a couple of my loras. I just did this to get the hang of training SDXL, so I didn't spend any time curating the images (some of them were probably deformed, and I'm sure a lot of them had bad hands), but I think the results here are pretty encouraging. This lyco is mostly here for people to mess with while better stuff is trained. I don't recommend mixing it into anything in its current state, as I'm sure it'll make hand quality worse.All images were generated using hires fix. Images posted to this resource are not cherry picked, so what you see should be what you get. See images for prompts.

Files with SHA256: 18a817695464364e16bf576da10e880cece8b6683c90c9e0978d7ec27ba03eb5