AI绘图圈,最著名的两大派系是stable diffsion和Midjourney。
前者主打开源,模型、LoRA、插件层出不穷,我光下载模型就差不多2个T。而Midjourney则闭源(收费),主打一个大模型包容万物。
我在研究模型的时候,发现很多严肃的SD模型创作者都逐渐有一个困惑:
模型层出不穷,却同质化严重,有没有一种可能,可以整合所有的大模型,就靠一个模型出图?然后用LoRA和各种插件来优化细节?
SDXL给出了回答:可以。
自从SDXL发布后,炼丹群的群友们很少再用回SD1.5了,原因很简单,它无所不能。
虽然很多模型创作者也迅速跟进,推出了很多SDXL的模型,但基础模型已经足以应对大部分场景。或者找到一两个用得顺手的模型,任意发挥自己的想象力。
和SD1.5相比,SDXL对自然语言的理解更加到位,和SD1.5的模型相比,SDXL的模型素材更多、更清晰,只需要结合几个常用的LoRA,就能最大限度的还原想象中的场景。
WebUI的创作者,在7月20日发布了一款插件(8月5日进行了第一次更新),这个插件的作用简单又复杂:生成各种风格(style)。
安装这个插件很简单,只需要在扩展--网址安装--输入网址即可。
https://github.com/ahgsql/StyleSelectorXL
当然了,因为国内访问GitHub不太稳定,如果安装失败,需要多试几次,通常10分钟左右肯定可以。
使用也非常方便,只需要点击“启用风格”,然后选择18种风格种的一种即可。
我顺手汉化了这个插件,文末附汉化包下载。
在汉化过程中,也对这个插件的工作原理有了一定的了解。
本质上,作者的每个风格,其实就是一段prompt(以及一段负面prompt),在生成图的时候,这段prompt被追加在你原来的prompt后面。
如果有特别常用的prompt,也可以通过修改插件对应文件的形式,追加一个或者多个按钮。
下文的样图,使用的模型是dreamshaperXL,
提示词:masterpiece,best quality,1girl,garden in dystopian postapo ruins, upper body, midnight
1、base
"prompt": "{prompt}",
"negative_prompt": ""
这是一段空白,这个按钮还有存在的必要吗?!
2、3D模型
"prompt": "professional 3d model {prompt} . octane render, highly detailed, volumetric, dramatic lighting",
"negative_prompt": "ugly, deformed, noisy, low poly, blurry, painting"
3、胶片
"prompt": "analog film photo {prompt} . faded film, desaturated, 35mm photo, grainy, vignette, vintage, Kodachrome, Lomography, stained, highly detailed, found footage",
"negative_prompt": "painting, drawing, illustration, glitch, deformed, mutated, cross-eyed, ugly, disfigured"
4、动画
"prompt": "anime artwork {prompt} . anime style, key visual, vibrant, studio anime, highly detailed",
"negative_prompt": "photo, deformed, black and white, realism, disfigured, low contrast"
5、电影
"prompt": "cinematic film still {prompt} . shallow depth of field, vignette, highly detailed, high budget, bokeh, cinemascope, moody, epic, gorgeous, film grain, grainy",
"negative_prompt": "anime, cartoon, graphic, text, painting, crayon, graphite, abstract, glitch, deformed, mutated, ugly, disfigured"
6、漫画
"prompt": "comic {prompt} . graphic illustration, comic art, graphic novel art, vibrant, highly detailed",
"negative_prompt": "photograph, deformed, glitch, noisy, realistic, stock photo"
看起来,是美漫风格,日漫还是要自己找LoRA(我正在测试吉卜力风格)。
7、粘土(其实我觉得翻译成泥人比较贴切)
"prompt": "play-doh style {prompt} . sculpture, clay art, centered composition, Claymation",
"negative_prompt": "sloppy, messy, grainy, highly detailed, ultra textured, photo"
8、数字
"prompt": "concept art {prompt} . digital artwork, illustrative, painterly, matte painting, highly detailed",
"negative_prompt": "photo, photorealistic, realism, ugly"
我想了想,数字风格应该就是CG风格,也就是常见的2.5D。
9、增强
"prompt": "breathtaking {prompt} . award-winning, professional, highly detailed",
"negative_prompt": "ugly, deformed, noisy, blurry, distorted, grainy"
大致相当于增加细节的LoRA。
10、幻想(在AI创作圈,fantasy是非常热门的一种风格,模型、LoRA有很多,XL下,一行prompt就解决了)
"prompt": "ethereal fantasy concept art of {prompt} . magnificent, celestial, ethereal, painterly, epic, majestic, magical, fantasy art, cover art, dreamy",
"negative_prompt": "photographic, realistic, realism, 35mm film, dslr, cropped, frame, text, deformed, glitch, noise, noisy, off-center, deformed, cross-eyed, closed eyes, bad anatomy, ugly, disfigured, sloppy, duplicate, mutated, black and white"
11、等距(原文是isometric style,按照字面意思翻译了下,实际效果是画面整整齐齐那种)
"prompt": "isometric style {prompt} . vibrant, beautiful, crisp, detailed, ultra detailed, intricate",
"negative_prompt": "deformed, mutated, ugly, disfigured, blur, blurry, noise, noisy, realistic, photographic"
12、线稿
"prompt": "line art drawing {prompt} . professional, sleek, modern, minimalist, graphic, line art, vector graphics",
"negative_prompt": "anime, photorealistic, 35mm film, deformed, glitch, blurry, noisy, off-center, deformed, cross-eyed, closed eyes, bad anatomy, ugly, disfigured, mutated, realism, realistic, impressionism, expressionism, oil, acrylic"
13、低聚(原文是low poly,一种说不上来的略抽象的风格,我看有的翻译成低多边形)
"prompt": "low-poly style {prompt} . low-poly game art, polygon mesh, jagged, blocky, wireframe edges, centered composition",
"negative_prompt": "noisy, sloppy, messy, grainy, highly detailed, ultra textured, photo"
14、霓虹(直译应该是霓虹朋克,为了按钮好看,翻译成霓虹)
"prompt": "neonpunk style {prompt} . cyberpunk, vaporwave, neon, vibes, vibrant, stunningly beautiful, crisp, detailed, sleek, ultramodern, magenta highlights, dark purple shadows, high contrast, cinematic, ultra detailed, intricate, professional",
"negative_prompt": "painting, drawing, illustration, glitch, deformed, mutated, cross-eyed, ugly, disfigured"
15、折纸
"prompt": "origami style {prompt} . paper art, pleated paper, folded, origami art, pleats, cut and fold, centered composition",
"negative_prompt": "noisy, sloppy, messy, grainy, highly detailed, ultra textured, photo"
16、相片(和胶片相比,可以认为是更清晰的数码相机效果)
"prompt": "cinematic photo {prompt} . 35mm photograph, film, bokeh, professional, 4k, highly detailed",
"negative_prompt": "drawing, painting, crayon, sketch, graphite, impressionist, noisy, blurry, soft, deformed, ugly"
17、像素
"prompt": "pixel-art {prompt} . low-res, blocky, pixel art style, 8-bit graphics",
"negative_prompt": "sloppy, messy, blurry, noisy, highly detailed, ultra textured, photo, realistic"
魂斗罗的背景音乐响起。
18、质感(原文是texture,直译的话,是纹理,但我觉得从画面效果看,质感更合适)
"prompt": "texture {prompt} top down close-up",
"negative_prompt": "ugly, deformed, noisy, blurry"
经过短短几小时的试用,发现基本上不需要改善画质和风格的LoRA了。
汉化包下载:
汉化包解压缩后,替换掉extensionsStyleSelectorXL 即可。
https://www.123pan.com/s/ueDeVv-WO0I.html 提取码:lgyh