什么是Stable Diffusion

Stable Diffusion是2022年发布的深度学习文本到图像生成模型，它主要用于根据文本的描述产生详细图像，尽管它也可以应用于其他任务，如内补绘制、外补绘制，以及在提示词（英语）指导下产生图生图的翻译。

简单的来说，就是我们常说的AI绘画，它的工作原理就是，想象你正在欣赏一幅世界名画，然后慢慢地眯起了眼睛，画面越来越模糊，Stable Diffusion 将我们眼前的模糊的画面扩散，然后得到了一个新的画面。这么说好像不对，没关系，并不影响。现在一起来尝试这款大模型吧。

Stable Diffusion的配置要求

操作系统: win10及其以上

内存要求：16-32G

显卡:一般推荐N卡

显存:6G及其以上

cpu：越高越高

安装Stable Diffusion

介于Stable Diffusion的本地部署有一定的难度，所以我选择了秋叶大佬的整合包。简单方便，开箱即用。

链接 提取码

4gsh

漫长的等待后，就可以使用了

魔法种子

1. 黑色礼服少女

   Night scenery, milf,mature women,Flickering green eyes, deep brown irises,sexy,slim,slander waist, water, happy, glowing objects, romantic, environmental details, lake water, starry sky, gentle breeze, forest, tree shadows, sexy, little girl, graceful, white snow, smooth, delicate, bright eyes, stars, slender eyebrows, elegant, straight nose, playful, tempting lips, innocent, confident, relaxed, mischievous, soft hair, charming, mysterious, fashionable, black dress, perfect body curves, mysterious aura, exquisite bracelet, high heels, lush flower field, fragrance of flowers, confidence and independence,((rim light,depth of field,blurry,cinematic lighting,backlight,ray-tracing)),<lora:xijie:0.2>, <lora:shuimu:0.1>, <lora:shuimo:0.1>, <hypernet:baisi:0.3>
   

2. 小天使

   (kokkoro_(princess_connect!): 1.2), (pointy_ears: 1.2), (rating:safe: 1.2), (1girl: 1.2), (elf: 1.2), (antenna_hair: 1.2), (hair_between_eyes: 1.2), (bangs: 1.2), (flower: 1.2), (hair_ornament: 1.2), (puffy_sleeves: 1.2), (silver_hair: 1.2), (smile: 1.2), (closed_mouth: 1.1), (puffy_long_sleeves: 1.1), (white_flower: 1.1), (hair_flower: 1.1), (dress: 1.1), (solo: 1.1), (blush: 1.1), (sleeveless: 1.1), (long_sleeves), (eyebrows_visible_through_hair), (bare_shoulders), detached_sleeves, sleeveless_dress, see-through, see-through_sleeves, looking_at_viewer, hand_up, red_eyes, green_dress, gradient, gradient_background, white_dress, upper_body,sea of flowers，cat ear,lori，black silk
   

声明

感谢秋叶大佬