This model was trained on art by Lisa Frank (and possibly some fan-art). It aims to add some cartoony colorfulness to stable diffusion. There are two models uploaded, 1400 steps and 4000 steps. You can judge the difference in style by looking at the image grid.