Gengar LoRA trained on NovelAI with Kohya training scripts using the DAdaptAdam optimizer at 512x512 resolution. Depending on the base model used, may try and give Gengar ears, which is incorrect. This is easily fixed by adding (ears) to your negative prompt. She's going to be grinning in 90% of the images, and that expression is hard to change. Comes with the territory of Gengar, I guess. Should have wide compatibility with model mixes based on NovelAI, which is most of them. Trained on an all female dataset.
