大江東去浪淘盡，千古風流人物===============================================Version1.1更新：<ul><li>部分解决文官女性化问题，仍需要添加负面tag：earrings, hair flower, female clothes, female ornament, female</li><li>一个简单的脸谱化方法：直接把人名加prompt里，参见样例。</li></ul>===============================================Version1.0更新：<ul><li>新加入了三国志战略版武将</li><li>使用新的底模：<a target="_blank" rel="ugc" href="https://civarchive.com/models/4468/counterfeit-v25">CounterfeitV25_25.safetensors </a>注意要点选V2.5，作者现在更新到3.0了</li><li>由于新底模机制不同，不要添加类似realistic, photo-realistic的prompt</li><li>素材分辨率提升至最高768x1024</li><li>删去了阴险脸（如李儒）和异民族，新删去黄巾贼</li><li>样例使用Eular a + Latent(Nearest-exact)，Denoising Strength=0.5</li><li>看板娘照例使用了其他Lora和Controlnet调脸，理论上男武将也可以这样操作</li></ul>已知问题：由于女性角色数量激增，男性文官生成中可能出现女性化，暂时添加负面prompt来workaround：earrings, hair flower, female clothes, female ornament===============================================Version0.999更新：<ul><li>新加入了近200名武将</li><li>删去了阴险脸（如李儒）和异民族服饰</li><li>Denoising Strength=0.5</li></ul>===============================================三国人物实验lora尝鲜，过几天有空继续更新<ul><li>示例使用guofeng3_v32或其Light版本，512 x 768</li><li>weight = 1</li><li>不需要特别的触发词，可以在前面加【chinese male, chinese female, chinese】强调一下，其他的可以参考示例的tag</li></ul>其他：由于long hair可能会被识别成女人，所以葛亮和宫颈……我直接用1girl===============================================Sanguo (The Three Kingdoms) character experimental lora. I will continue to update later.<ul><li>Examples using guofeng3_v32 or Light version, 512 x 768</li><li>set weight = 1</li><li>No trigger words are required, you can add 【chinese male, chinese female, chinese】to emphasize. and refer to the tags of the examples as well</li></ul>

三国 The Three Kingdoms

Files with SHA256: 88d19efd1ddab63e8d3f88620713ea7979393f2ebc263d10171507114cb35f98