This is a LoRA that is an extraction from a Base Model that I created called Princesses. It works best using only a specific positive prompt of what you want to generate and the negative prompt that is shown in the workflow, with the LoRA strength, set at 1.0, and a {randomize} sampler setting.