Zoroark LoRA trained on NovelAI with Kohya training scripts using the DAdaptAdam optimizer at 512x512 resolution. Depending on your base model, can try to give Zoroark a tail, which is wrong. I recommend adding (tail) to your negative and (tailless) to your positive if that's an issue you run into. Zoroark's claws don't play nice with any embeddings that try to shape hands, so I recommend removing those from your prompt when using this LoRA. Should have wide compatibility with model mixes based on NovelAI, which is most of them. Trained on an all female dataset.
