RIM from Kamitsubaki studioMain trigger word is rimI still don't know optimal settings for SDXL, but following settings produce pretty good results.<ul><li>Positive Prompts: anime,(masterpiece, top quality, best quality, official art, beautiful and aesthetic:1.2), cinematic lighting, upper body, extreme detailed, studio anime, key visual, (bioluminescent details:1.1), (subsurface scattering:1.1), colorful, rim, virtual youtuber, 1girl, solo, blue hair, multicolored hair, red hair, short hair, bangs, multicolored eyes, hood, &lt;lora:rim_sdxl_v1:1.0&gt; looking at viewer, dynamic pose, dynamic angle, dynamic lighting</li><li>Negative Prompts: (worst quality:1.2, low quality:1.2, bad anatomy:1.2, cropped hands, extra digit, fewer digit), text, error, jpeg artifacts, signature, watermark, username, artist name, <a rel="ugc" href="https://civarchive.com/models/119032/unaestheticxl-or-negative-ti">unaestheticXLv1</a></li><li>Sampler: DPM++ 2M Karras</li><li>Step: 30</li><li>CFG Scale: 8~10</li></ul><a rel="ugc" href="https://drive.google.com/file/d/1h7PaY50COBR6G6JSCxatDbr9IEXmFRCo/view?usp=drive_link">Tags</a> - This csv file contains all of the prompts that were used when I trained this modelI am not the artist of the character nor RIM herself nor related to thinkr incoporation.All right reserved to <a target="_blank" rel="ugc" href="https://www.youtube.com/@RIM_virtual">RIM</a> and <a target="_blank" rel="ugc" href="https://kamitsubaki.jp/">KAMITSUBAKI Studio</a>(<a target="_blank" rel="ugc" href="https://thinkr.jp/">THINKR inc</a>)

Files with SHA256: f04f7feeca633f36292fcfd6a45ff5b321c3344b8ef856f953f7a5383677ee1f