Makes sketch-like monochrome images similar in style to the artist bone Nigi. Pony XL version on request of an old NAI <a target="_blank" rel="ugc" href="https://civarchive.com/models/38007">LoRA</a>.How to use Since this LoRA was focused on capturing the shading the dataset is very limited. The artist also likes using dynamic poses and angles making virtually all characters have their bodies, fingers and hands in non-standard positions. This all leads to this LoRA being a bit "temperamental" using just raw prompting. At least when you don't prompt where the limbs should be. <s>The quality tags of Pony XL leaning heavily towards western art will also generally destroy the style with the exception of "score_5_up". "monochrome", "greyscale" and "hatching (texture)" may increase output quality depending on model, I didn't see any noticeable one on Pony XL though.</s><s>If you're not using other tools to increase control poses I highly recommend checking out the uploaded zip with training tags to get a feel of what works and then modify them for your needs.</s> Less of an issue in v1.10Tags that work well are anything catgirl related, tummy exposure, muscles, short hair and tight clothing (preferably black). For more prompt tips please see the sample images or above mentioned zip.Misc Trained on Pony SDXL at 1024 resolution with a dataset of 27 images. All images are untouched txt2img, no inpainting or helper embeddings used for quality transparence, e.g. - things should look even better if you do. The training was done on a small subset of the artists images, as such the output of this LoRA is not an accurate representation neither in quality or style of the source material. Go support the <a target="_blank" rel="ugc" href="https://skeb.jp/@bone_Nig">artist</a> if you want the good stuff.

bone Nigi Style [Pony XL]

Retrained with up-to-date tools. Might work better or worse depending on what model and prompts you use.

Files with SHA256: 32920028925baf650facafda41420999143b2d25e4941310661a73e5332c1e8a