Are you looking for a model that's both 'Realistic'? Try Lemix Realistic Asians!全能的亚洲真人、真物、真景写实模型，喜欢可以试一下！★ Please upload at least one photo made using the Lemix Realistic Asians model uploaded to the gallery.把你生成的满意作品，上传到Civitai，谢谢！To achieve the best image quality, please follow these parameter settings: 下面是调试时使用的参数：<ul><li>Sampling method &amp; type: DPM++ 2M Karras</li><li>Sampling steps: 25 ~ 50</li><li>CFG Scale: 5 ~ 7 (7 is best)</li><li>Resolution: 512x768 建议使用这个尺寸生成，满意的作品再放大。</li><li>Textual Inversion: <a target="_blank" rel="ugc" href="https://civarchive.com/models/4629/deep-negative-v1x">ng_deepnegative_v1_75t </a>and <a target="_blank" rel="ugc" href="https://civarchive.com/models/7808/easynegative">easynegative</a> ,don't use Bad-Hand! Bad-Hand不好用，推荐使用<a target="_blank" rel="ugc" href="https://civarchive.com/models/4629/deep-negative-v1x">ng_deepnegative_v1_75t</a>和<a target="_blank" rel="ugc" href="https://civarchive.com/models/7808/easynegative">easynegative</a>。</li></ul>To achieve more realistic character details, it is recommended to use <a target="_blank" rel="ugc" href="https://github.com/Bing-su/adetailer">Adetailer</a> with the following parameters:建议安装<a target="_blank" rel="ugc" href="https://github.com/Bing-su/adetailer">Adetailer</a>来对人物进行优化，脸部、眼睛的效果非常显著，手部效果一般。<ul><li>Face model: face_yolov8n (mediapipe_face_mesh is my favorite)</li><li>Eyes model: mediapipe_face_mesh_eyes_only (recommend)</li><li>Hand model: hand_yolov8n (There are no better hand re-painting models. None of them work well, so I don't recommend them.)</li><li>Redraw amplitude:<ul><li>For close-up portraits: 0.25-0.3</li><li>For upper body portraits: 0.3-0.35</li><li>For full body portraits: 0.35-0.4</li></ul></li></ul>If you like my model, you can encourage me with "buzz".准备训练提升模型能力的Lora，有"buzz"的朋友赞助点，谢谢！I only publish my models on <a target="_blank" rel="ugc" href="https://civarchive.com/models/564358">Civitai</a>, the rest are all stolen.

Lemix Realistic Asians

v2 is a bridge version that combines the models we want and has had some training on character features.v2是一个过渡版本，重点是融合想要的模型，少量的训练了人物特征。

Files with SHA256: 29ed1f6e809717322cdf7aa2fd8bc8ae0673d51c822e506cef87031ccdbfed8b