Makes images in a monochrome/greyscale style with hatching textures. LyCORIS variant of previously released <a target="_blank" rel="ugc" href="https://civarchive.com/models/102664">LoRA</a>. This version is trained on a larger dataset with the monster content removed for, hopefully, more reliable humanoid outputs.How to use Trained on both "1boy" and "1girl" images this LyCORIS should handle both. I highly recommend using "greyscale" and "monochrome" tags to capture the style best. Other tags to try is "traditional media", "sepia", "depth of field", "hatching \(texture\)", "drawing of", "sketch" and other drawing style related tags. The style was caught enough to also do general imagery. For specific prompt tips see the sample images or the uploaded zip of training prompts. Recommended strength is anything from 0.5 to 1.0 depending on what you prompt. For samples I recommend DIMM or DPM++ 2M Karras since those best recreate the hand drawn style. Model recommendations is anything with good support for drawing styles. Putting "3D" and/or "realistic" in negatives can help if it's trained on a lot of photorealism. For upscaling I recommend one with good attention to detail like 4x_foolhardy_Remacri.MISC Trained on NAI at 512px with a dataset of ~350 images. All images are raw txt2img upscaled without any inpainting or helper embeddings, e.g things should look even better if you use them. 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 artist <a target="_blank" rel="ugc" href="https://www.pixiv.net/en/artworks/91455208">Epeulu</a>. Go support them if you want the good stuff.

Files with SHA256: 43b6b4cbd56219b4a5844a32cc5cb6b6c0e67c4a9273ef7894b96a9404deeabe