LoRA trained on images from MidJourney's Niji style, specifically the jelly look.

I recommend using the LoRA at around 0.8 emphasis for best results.