The idea for this LoRA came from something my wife came across, a single image in a style that caught my imagination. And because it was a single image, I didn't have a lot to work from to create a dataset. The image contained three faces and some other scene elements. I carefully extracted various elements and produced enough images for a dataset. For some of the faces, I removed the background. Others I left intact. The result is this LoRA. It's quaint, sweet, a little bit storybook style. I hope you like it.

The SDXL version is least like the original images, but it's still a nice style.