<h1 id="heading-10846">TL;DR:</h1>This model is a research-use experimental model, needs to cherry-pick and proper prompting to get similar effort. about 15% image should be good enough except title text. Title text is considered to be limited by token lenth, the longer it was, the less chance it will show up.I didn't try it combine with other character LoRA. but I tried use it on BArtstyleXL, it works fine.Inspired by <a target="_blank" rel="ugc" href="https://civarchive.com/models/262786/nikke-goddess-of-victory-profile-or-concept-lora-xl">NIKKE: Goddess of Victory Profile</a>, trying to figure out how will the artist tag and natural language works on captioning.This is a pre-alpha testing for BArtstyleXL V2.<h1 id="heading-10847">How-to:</h1>You need to use natural language to use this model.<h3 id="heading-10848">Trigger sentence: </h3><code>anime character poster with 4 different expressions of them, a english text title says **INSERT YOUR TITLE CHARA NAME HERE** in uppercase</code><h3 id="heading-10849">Suggest resolution: </h3>1:1.6-1.8, vertical only. e.g. 800*1280 to 800*1416<h3 id="heading-10850">Suggest Workflow:</h3><ol><li>prompting priority suggestion: character name, trigger sentence, quality tag, character identity, character clothing</li><li>turn off hi-res fix and adetailer</li><li>generate until the proper text/perspective pop out.</li><li>re-use the seed, set a small variant rate.</li><li>turn on hi-res fix and adetailer, infinite generate until you satisfied for results.</li></ol><h3 id="heading-10851">Known Issue:</h3>Alongside the title text issue, the expressions didn't recognize as a single row, sometimes it will pop out on both sides or even nothing shows up, the number of expressions is pretty random for no reason. so do those expressions themself, sometimes it may generate 6 same expressions, maybe undertrained,This model doesn't have a suggest weight due to lack of data. 0.6-0.9 works fine.<h3 id="heading-578">Disclaimer:</h3><a target="_blank" rel="ugc" href="https://civarchive.com/models/260267/">Animagine XL V3</a>'s license asked that merges of this model should be released as same permissions, which is allowed user to use generation service or sell generated images of the model.But according to Blue Archive's Fanart Guidelines ( <a target="_blank" rel="ugc" href="https://bluearchive.jp/fankit/guidelines">JP</a> | <a target="_blank" rel="ugc" href="https://www.bilibili.com/read/cv24806541/">CN </a>), this model shall not be used for any kind of commercial use, including but not limited to selling this model or merges of this model, selling image generated by this model or merges of this model, or 'paid member exclusive' for monetization platforms like Patreon/Fanbox/etc.The license change of this model is permitted by <a target="_blank" rel="ugc" href="https://civarchive.com/user/CagliostroLab">Cagliostro Research Lab</a>, the author of Animagine XL. Once again, I sincerely thank them for their understanding and dedication for open-source community. <a target="_blank" rel="ugc" href="https://huggingface.co/cagliostrolab/animagine-xl-3.0/discussions/8">(Permission Source</a>)<h3 id="heading-578">免责声明:</h3>根据 <a target="_blank" rel="ugc" href="https://civarchive.com/models/260267/">Animagine XL V3</a> 的许可证，基于该模型的衍生模型应以相同授权方式公开，即允许用户使用线上生成服务，或售卖该模型生成的图片。但根据《&lt;蔚蓝档案&gt;同人创作指引》 ( <a target="_blank" rel="ugc" href="https://bluearchive.jp/fankit/guidelines">日文</a> | <a target="_blank" rel="ugc" href="https://www.bilibili.com/read/cv24806541/">中文 </a>)，此模型不得用于任何形式的商业用途，包括但不限于售卖该模型及其衍生模型自身或其生成的图片，或Patreon/Fanbox等众筹平台的“订阅会员限定”服务。该许可证授权范围变更得到 Animagine XL 作者 <a target="_blank" rel="ugc" href="https://civarchive.com/user/CagliostroLab">Cagliostro Research Lab </a>的许可。在这里再次感谢他们的理解，以及他们对于开源社区的贡献。（<a target="_blank" rel="ugc" href="https://huggingface.co/cagliostrolab/animagine-xl-3.0/discussions/8">授权来源</a>）

Files with SHA256: 6bfe24efe030c14769da68d0437b94934cabd76ad082cff85eed60d12c5afed6