<p><span style="color:rgb(193, 194, 197)">Sceptile LoRA trained on NovelAI with Kohya training scripts using the DAdaptAdam optimizer at 512x512 resolution. Doesn't have a good grasp of what Sceptile's tail should look like. There was a lot of variance in the tail in the training data, which is reflected in the final model. I tried adding (frond tail) to get a better tail, and it worked to very limited effect. Sceptile's yellow 'orbs' only appear occasionally, I haven't figured out a way to prompt for them. May require adding (red jaw) to your prompt to get the right jaw color. Otherwise great at reproducing the body shape of Sceptile. Should have wide compatibility with model mixes based on NovelAI, which is most of them. Trained on an all female dataset.</span></p>

Files with SHA256: d418302610b615dadecfdb46368326031731baae22c4d6ab2ff2defee0ea7c62