<h1 id="heading-190">XPortrait, the First SDXL documentary portrait model</h1><h2 id="heading-191">Let the AI image convey true feelings, a new level of virtual humanistic photography</h2>This is my first attempt at SDXL Lora training, based on 100 portraits of people trained, committed to creating real light and shadow, restore high definition facial texture, let the characters vivid and lively.With the advantage of SDXL, you can use natural language instead of short sentences to describe the content of the picture in detail, trying to make the picture full of "story"; Try adding the photographer's name and camera model to achieve different effects;At the same time, SDXL also optimizes a lot of graphics logic, and crashes are greatly reducedThe recommended weight is 0.6~0.8Try adding the following prefix words：portrait, 1990 s, photographed by Steve McCurry, Leica camera, (face focus), blurred background, emotional face, Clear face,<h3 id="heading-192">XPortrait，首个SDXL人文肖像模型</h3>让影像传递真情实感，虚拟人文摄影新台阶这是我第一次尝试SDXL Lora 训练，基于100张人物肖像图训练而成，青龙圣者底模，致力于打造真实光影，还原高清晰度面部纹理，让人物生动传神基于SDXL的优势，您可以使用自然语言而非短句，详细描述画面内容，试着让画面充满“故事感”；试着增加摄影师名字以及相机型号，实现不同画面效果；同时，SDXL也优化了许多图像逻辑，崩坏大大降低建议权重0.6~0.8

Files with SHA256: 97665a9515fd4bfa45a06ecfd16309bdc668c8ed794f09db170a3302a1628f7a