SDXL風(fēng)格插件,超越Midjourney的時刻來了


在AI繪圖圈,最著名的兩大派系是stable diffsion和Midjourney。
前者主打開源,模型、LoRA、插件層出不窮,我光下載模型就差不多2個T。而Midjourney則閉源(收費),主打一個大模型包容萬物。
我在研究模型的時候,發(fā)現(xiàn)很多嚴(yán)肅的SD模型創(chuàng)作者都逐漸有一個困惑:
模型層出不窮,卻同質(zhì)化嚴(yán)重,有沒有一種可能,可以整合所有的大模型,就靠一個模型出圖?然后用LoRA和各種插件來優(yōu)化細(xì)節(jié)?
SDXL給出了回答:可以。
自從SDXL發(fā)布后,煉丹群的群友們很少再用回SD1.5了,原因很簡單,它無所不能。
雖然很多模型創(chuàng)作者也迅速跟進(jìn),推出了很多SDXL的模型,但基礎(chǔ)模型已經(jīng)足以應(yīng)對大部分場景?;蛘哒业揭粌蓚€用得順手的模型,任意發(fā)揮自己的想象力。
和SD1.5相比,SDXL對自然語言的理解更加到位,和SD1.5的模型相比,SDXL的模型素材更多、更清晰,只需要結(jié)合幾個常用的LoRA,就能最大限度的還原想象中的場景。
WebUI的創(chuàng)作者,在7月20日發(fā)布了一款插件(8月5日進(jìn)行了第一次更新),這個插件的作用簡單又復(fù)雜:生成各種風(fēng)格(style)。

安裝這個插件很簡單,只需要在擴展--網(wǎng)址安裝--輸入網(wǎng)址即可。

https://github.com/ahgsql/StyleSelectorXL
當(dāng)然了,因為國內(nèi)訪問GitHub不太穩(wěn)定,如果安裝失敗,需要多試幾次,通常10分鐘左右肯定可以。
使用也非常方便,只需要點擊“啟用風(fēng)格”,然后選擇18種風(fēng)格種的一種即可。
我順手漢化了這個插件,文末附漢化包下載。
在漢化過程中,也對這個插件的工作原理有了一定的了解。
本質(zhì)上,作者的每個風(fēng)格,其實就是一段prompt(以及一段負(fù)面prompt),在生成圖的時候,這段prompt被追加在你原來的prompt后面。
如果有特別常用的prompt,也可以通過修改插件對應(yīng)文件的形式,追加一個或者多個按鈕。
下文的樣圖,使用的模型是dreamshaperXL,
提示詞:masterpiece,best quality,1girl,garden in dystopian postapo ruins, upper body, midnight
1、base
"prompt": "{prompt}",
"negative_prompt": ""
這是一段空白,這個按鈕還有存在的必要嗎?!

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、膠片
"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、動畫
"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、電影
"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、漫畫
"prompt": "comic {prompt} . graphic illustration, comic art, graphic novel art, vibrant, highly detailed",
"negative_prompt": "photograph, deformed, glitch, noisy, realistic, stock photo"

看起來,是美漫風(fēng)格,日漫還是要自己找LoRA(我正在測試吉卜力風(fēng)格)。
7、粘土(其實我覺得翻譯成泥人比較貼切)
"prompt": "play-doh style {prompt} . sculpture, clay art, centered composition, Claymation",
"negative_prompt": "sloppy, messy, grainy, highly detailed, ultra textured, photo"

8、數(shù)字
"prompt": "concept art {prompt} . digital artwork, illustrative, painterly, matte painting, highly detailed",
"negative_prompt": "photo, photorealistic, realism, ugly"

我想了想,數(shù)字風(fēng)格應(yīng)該就是CG風(fēng)格,也就是常見的2.5D。
9、增強
"prompt": "breathtaking {prompt} . award-winning, professional, highly detailed",
"negative_prompt": "ugly, deformed, noisy, blurry, distorted, grainy"

大致相當(dāng)于增加細(xì)節(jié)的LoRA。
10、幻想(在AI創(chuàng)作圈,fantasy是非常熱門的一種風(fēng)格,模型、LoRA有很多,XL下,一行prompt就解決了)
"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,按照字面意思翻譯了下,實際效果是畫面整整齊齊那種)
"prompt": "isometric style {prompt} . vibrant, beautiful, crisp, detailed, ultra detailed, intricate",
"negative_prompt": "deformed, mutated, ugly, disfigured, blur, blurry, noise, noisy, realistic, photographic"

12、線稿
"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,一種說不上來的略抽象的風(fēng)格,我看有的翻譯成低多邊形)
"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(直譯應(yīng)該是霓虹朋克,為了按鈕好看,翻譯成霓虹)
"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、折紙
"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(和膠片相比,可以認(rèn)為是更清晰的數(shù)碼相機效果)
"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"

魂斗羅的背景音樂響起。
18、質(zhì)感(原文是texture,直譯的話,是紋理,但我覺得從畫面效果看,質(zhì)感更合適)
"prompt": "texture {prompt} top down close-up",
"negative_prompt": "ugly, deformed, noisy, blurry"

經(jīng)過短短幾小時的試用,發(fā)現(xiàn)基本上不需要改善畫質(zhì)和風(fēng)格的LoRA了。
漢化包解壓縮后,替換掉extensions\StyleSelectorXL 即可。
漢化包下載:
https://www.123pan.com/s/ueDeVv-WO0I.html?提取碼:lgyh