Xerneas LoRA trained on NovelAI with Kohya training scripts using the DAdaptAdam optimizer at 512x512 resolution. Horns can come out wonky a lot of the time because they have a lot of overlap with each other, which Stable Diffusion has a hard time handling. Body can come out blue, but should be black. Adding (black body) to your prompt will fix it. Depending on your base model, may require adding (hooves) to your prompt to get correct feet. Should have wide compatibility with model mixes based on NovelAI, which is most of them. Trained on an all female dataset.