Makes images similar in style to the artist <a target="_blank" rel="ugc" href="https://danbooru.donmai.us/posts?tags=sukabu">Sukabu</a>. For those unfamiliar with the artist he likes to draw a futuristic world where cyborg, android and normal girls just go about their daily life having fun despite (or thanks to?) their mechanical bits and parts. With some random mechas and kaijuu sprinkled on top.How to use Prompt as normal. Recommended tags to use are "mechanical arms", "mechanical legs", "android", "robot", "humanoid robot", "robot joints", "doll joints" and similar. Anything related to machinery, cables, mechanical parts and similar themes usually work well too. The colorful, drawn style comes in more strongly at lower Pony XL quality tags where the higher ones usually gives a more "clean" output.MISC Trained on Pony XL with a dataset of 110 images at 1024 resolution. All sample images are raw txt2img except one or two using <a target="_blank" rel="ugc" href="https://github.com/wkpark/uddetailer"> μ Detection Detailer</a> faces but no helper embeddings/negatives used. E.g. - things should look even better if you do. The training was done on a small subset of art, as such the output of this LoRA is not an accurate representation neither in quality or style of the source material. Go support the artist <a rel="ugc" href="https://twitter.com/sukabu89">Sukabu</a><a target="_blank" rel="ugc" href="https://twitter.com/_zunta"> </a>if you want the good stuff.

Files with SHA256: a3bcc9b63dfe8b1dd0f38f382ed987f69aa52d4015dbb761a9cdd1f33da95d98