Charmander LoRA trained on NovelAI with Kohya training scripts using the DAdaptAdam optimizer at 512x512 resolution. Works great, main issues are tail placement and where the tail flame ends up. Sometimes the tail flame manifests in other parts of the image as a random flame. Lowering the LoRA strength should fix this issue. Should have wide compatibility with model mixes based on NovelAI, which is most of them. Trained on a mixed male/female dataset.