Hello everyone, Today I'm uploading a requested model of Lady Ozai, based on an XXX parody of Avatar: The Last Airbender. This model won't be particularly flexible, as I only had 11 training images available. Note: Some elements may be difficult to remove. For example, the tassel earrings appear in almost every training image, so their influence on the model is quite strong. You can try using <code>(earrings, tassel earrings:2.0)</code> in the negative prompt, but success is not guarantied. Additionally, the outfits may not be entirely accurate due to the limited amount of visual reference material available. Note 2: Please ignore the version number in the filename. I created two earlier versions with different learning rates before settling on this one.<hr />Prompt: look at "Trigger Words" on the right side 👉Changelog: look at "About this Version" on the right side 👉<h1></h1>My Settings <a target="_blank" rel="ugc" href="https://github.com/Haoming02/sd-webui-forge-classic">(Forge <s>Classic </s>Neo</a>)*:*note, the example images only use high-res fix.Checkpoint: waiNSFWIllustrious_v<s>160 </s>170Sampler: Euler a, CFG Scale <s>7</s> 6Scheduler: <s>Automatic</s> KarrasSamplings Steps: 30Hires.fix: R-ERSGAN 4x+Anime6B, 20 Hires steps, Denoising Strength: <s>0.5</s> 0.3-0.45Extension:Adetailer (highly recommended)

Files with SHA256: 2b83b2e6cafae9f4eb86c661e65cafec6442cb396ecb1a44bc5051188170f5f1