首先,我们来梳理一下本次Xlabs-AI团队放出的七款Lora模型,分别是
写实、Furry(但是目前该模型还不支持ComfyUI)、MJV6、动漫、迪斯尼、风景和艺术风格。每种模型都经过精心训练,能够为 Flux 模型带来卓越的风格转换能力。
XLabs-AI/flux-lora-collection at main (huggingface.co)
Kijai/flux-loras-comfyui at main (hf-mirror.com)
flux-lora-collection · 模型库 (modelscope.cn)
下载地址目前不建议从huggingface上面下载,建议从modelscope上面下载。这里提醒一下小伙伴,只使用ComfyUI的小伙伴,可以只下载带comfy的文件,下载后放入ComfyUI对应的文件夹即可。
- mjv6_lora - Midjourney 风格
- realism_lora - 写实风格
- anime_lora - 动漫风格
- disney_lora - 迪斯尼风格
- scenery_lora - 风景风格
- art_lora - 艺术风格
- furry
Flux+LoRA在comfyUI中的使用
上一篇文章我们已经讲了如何在comfyUI中使用Flux。下面我们来看在comfyUI中的使用,先创建一个简单的工作流:
Anime LoRA:精致的日式动漫风格
Anime LoRA 专为捕捉日本动漫的精髓而设计。这个模型能够精确再现动漫角色的大眼睛、夸张的发型和鲜艳的色彩,擅长表现丰富的角色表情和动态姿势,同时能生成与动漫美学相符的背景场景,让您的作品充满生动的画面感。
触发词:anime
提示词:A cute corgi lives in a house made out of sushi, anime
提示词:a girl with orange hair, standing in a room with a window, looking out at a cityscape, anime
Art LoRA:插图艺术
Art LoRA 模型致力于将普通场景转化为具有独特艺术气息的作品。无论是插画还是概念设计,这个模型都能赋予作品独特的艺术魅力,使每一幅画面都散发出创意和美感。
触发词:art
提示词:
a classic fairy tale character. She is wearing her iconic red cloak with a hood, which is beautifully detailed with intricate patterns. Little Red Riding Hood is standing in a lush, enchanted forest with tall trees and vibrant flowers. She is holding a woven basket filled with goodies, and her expression is a mix of curiosity and caution as she ventures deeper into the forest. The background features soft, dappled sunlight filtering through the trees, creating a magical and whimsical atmosphere. The scene may include subtle hints of the Big Bad Wolf in the distance, adding a touch of intrigue and tension to the setting, art
提示词:castle in the middle of forest at night, art
Disney LoRA:呈现迪士尼动画的魔法
Disney LoRA 模型完美再现了迪士尼动画的独特魅力。它不仅能够创造出圆润可爱的角色设计,还能呈现出迪士尼特有的温暖色调和柔和光影效果,非常适合创作充满童话色彩的场景和富有感染力的角色形象,带领观众进入一个充满魔法的世界。
提示词:disney mermaid princesst, disney style
提示词:Disney Mulan Liu Yifei, disney style
提示词:disney animals collection, disney style
Furry LoRA:打造拟人化卡通角色
Furry LoRA 专注于创造逼真的拟人化卡通角色。这个模型能够精确呈现各种动物特征,如毛发纹理、耳朵和尾巴的细节,同时保持人类化的体型和表情。它尤其擅长创作具有独特个性的拟人化卡通角色,使每一个角色都栩栩如生。
提示词: Female with text 'Happy Birthday', furry
提示词:Female furry Pixie with text 'hello world'
提示词:Female furry Pixie
提示词:Male furry Lycanthrope with fur-covered body in ancient ruins, howling at the full moon, surrounded by eerie mist, werewolf transformation, elder scrolls, eslweyr, glitch aesthetic, anime-inspired, digital illustration, artstation, furry
MJV6 LoRA:引领 Midjourney V6 风格
MJV6 LoRA 模型擅长模仿 Midjourney V6 版本的风格。它能够生成精致的面部细节、时尚的服装设计,以及充满未来科技感的背景元素,为您的作品增添现代和前卫的气息。
提示词:A girl in a suit covered with bold tattoos and holding a vest pistol, beautiful woman, 25 years old, cool, future fantasy, turquoise & light orange ping curl hair
提示词:a woman She carries a long sword, wears a flowing soft silk Hanfu, and her long hair moves in the wind, facing the viewer in a full-body shot against a dark gray background, an epic fantasy scene with dark clouds in the sky and smoke effects, in the style of a Chinese ink painting depicting the martial arts world with cloud mist effects, like movie footage.
Realism LoRA:照片级真实感的极致
Realism LoRA 模型旨在提升 Flux 模型的照片级真实感效果。无论是细腻的皮肤纹理还是复杂的光影变化,这个模型都能逼真地再现现实世界的细节,让每一幅作品都如同真实照片般栩栩如生。(真实感跟MJV6 LoRA相比,个人觉得还是MJV6 LoRA强一点)
提示词:A handsome man in a suit, 25 years old, cool, futuristic
提示词:beautiful Chinese girl wearing red Hanfu
Scenery LoRA:描绘壮丽自然与城市景观
Scenery LoRA 专为创造壮观的景观而生。这个模型能够精确捕捉自然元素的纹理和光影效果,如山川、河流、森林,同时也能创作出复杂的城市建筑和未来都市景观。它擅长创作大尺度的全景图和富有氛围感的环境设计,让您的作品充满震撼力和感染力。
提示词:A fantasy cityscape with multiple buildings and skyscrapers all of which are covered in snow and ice, scenery style
提示词:A large ornate building with multiple levels and arches surrounded by trees and greenery. In front of it there are several statues and sculptures on pedestals with fire burning brightly in front of them, scenery style
如果觉得文章不错就点个赞,点点关注,一起学习AI~~
Anime LoRA:精致的日式动漫风格
Anime LoRA 专为捕捉日本动漫的精髓而设计。这个模型能够精确再现动漫角色的大眼睛、夸张的发型和鲜艳的色彩,擅长表现丰富的角色表情和动态姿势,同时能生成与动漫美学相符的背景场景,让您的作品充满生动的画面感。
提示词:A cute corgi lives in a house made out of sushi, anime
提示词:a girl with orange hair, standing in a room with a window, looking out at a cityscape, anime
Art LoRA:插图艺术
Art LoRA 模型致力于将普通场景转化为具有独特艺术气息的作品。无论是插画还是概念设计,这个模型都能赋予作品独特的艺术魅力,使每一幅画面都散发出创意和美感。
提示词:
a classic fairy tale character. She is wearing her iconic red cloak with a hood, which is beautifully detailed with intricate patterns. Little Red Riding Hood is standing in a lush, enchanted forest with tall trees and vibrant flowers. She is holding a woven basket filled with goodies, and her expression is a mix of curiosity and caution as she ventures deeper into the forest. The background features soft, dappled sunlight filtering through the trees, creating a magical and whimsical atmosphere. The scene may include subtle hints of the Big Bad Wolf in the distance, adding a touch of intrigue and tension to the setting, art
提示词:castle in the middle of forest at night, art
Disney LoRA:呈现迪士尼动画的魔法
Disney LoRA 模型完美再现了迪士尼动画的独特魅力。它不仅能够创造出圆润可爱的角色设计,还能呈现出迪士尼特有的温暖色调和柔和光影效果,非常适合创作充满童话色彩的场景和富有感染力的角色形象,带领观众进入一个充满魔法的世界。
提示词:disney mermaid princesst, disney style
提示词:Disney Mulan Liu Yifei, disney style
提示词:disney animals collection, disney style
Furry LoRA:打造拟人化卡通角色
Furry LoRA 专注于创造逼真的拟人化卡通角色。这个模型能够精确呈现各种动物特征,如毛发纹理、耳朵和尾巴的细节,同时保持人类化的体型和表情。它尤其擅长创作具有独特个性的拟人化卡通角色,使每一个角色都栩栩如生。
提示词:Female furry Pixie with text 'hello world'
提示词:Female furry Pixie
提示词:Male furry Lycanthrope with fur-covered body in ancient ruins, howling at the full moon, surrounded by eerie mist, werewolf transformation, elder scrolls, eslweyr, glitch aesthetic, anime-inspired, digital illustration, artstation, furry
MJV6 LoRA:引领 Midjourney V6 风格
MJV6 LoRA 模型擅长模仿 Midjourney V6 版本的风格。它能够生成精致的面部细节、时尚的服装设计,以及充满未来科技感的背景元素,为您的作品增添现代和前卫的气息。
提示词:A girl in a suit covered with bold tattoos and holding a vest pistol, beautiful woman, 25 years old, cool, future fantasy, turquoise & light orange ping curl hair
提示词:a woman She carries a long sword, wears a flowing soft silk Hanfu, and her long hair moves in the wind, facing the viewer in a full-body shot against a dark gray background, an epic fantasy scene with dark clouds in the sky and smoke effects, in the style of a Chinese ink painting depicting the martial arts world with cloud mist effects, like movie footage.
Realism LoRA:照片级真实感的极致
Realism LoRA 模型旨在提升 Flux 模型的照片级真实感效果。无论是细腻的皮肤纹理还是复杂的光影变化,这个模型都能逼真地再现现实世界的细节,让每一幅作品都如同真实照片般栩栩如生。(真实感跟MJV6 LoRA相比,个人觉得还是MJV6 LoRA强一点)
提示词:A handsome man in a suit, 25 years old, cool, futuristic
提示词:beautiful Chinese girl wearing red Hanfu
Scenery LoRA:描绘壮丽自然与城市景观
Scenery LoRA 专为创造壮观的景观而生。这个模型能够精确捕捉自然元素的纹理和光影效果,如山川、河流、森林,同时也能创作出复杂的城市建筑和未来都市景观。它擅长创作大尺度的全景图和富有氛围感的环境设计,让您的作品充满震撼力和感染力。
提示词:A fantasy cityscape with multiple buildings and skyscrapers all of which are covered in snow and ice, scenery style
提示词:A large ornate building with multiple levels and arches surrounded by trees and greenery. In front of it there are several statues and sculptures on pedestals with fire burning brightly in front of them, scenery style
Flux_插画
一款手绘插画风格的 lora,图像效果和我们常见的插画海报很接近,适合生成节日、活动插画素材。
① 推荐设置:触发词:illustration
② 参考提示词:The picture depicts a girl, with glasses and long hair, working in front of a computer. There is a yellow lamp and a potted cactus on the table. The background is a pale pink wall. The overall style of illustration is cute, commercial flat illustration
Flux-RealismLora
最近一类 “一个人在 TED 上发表演讲” 的视频片段在互联网上广为流传,原因是它完全是 AI 制作的,但是却非常逼真,无论是人物动态还是画面质感都与现实一般无二,效果非常惊人。
这种视频的制作方式是先用 AI 生成图像,再用 Runway、Kling 等工具将图像变为视频。之所以人物效果这么逼真,是因为它使用的是最新的 Flux 模型,并且叠加了一个 flux-Realism LoRA,能生成质感自然的人像,实现以假乱真。
目前 Huggingface 上有一个免费 Demo 可以让我们体验这个 lora 的生成效果,如果你对 AI 视频足够熟悉的话,还可以尝试用 AI 为视频配音、配乐,自己做一个 “真实的” 演讲视频。感兴趣的小伙伴可以了解我的 《AI 音视频系统入门指南》,帮你快速掌握目前最流行的相关工具。
① 在线体验: https://huggingface.co/spaces/multimodalart/flux-lora-the-explorer
② 推荐设置:lora 权重:1 / 触发词:无
③ 参考提示词:A 40-year-old female CEO is captured mid-speech. She had curly brown hair, holds a black microphone in his right hand, speaking passionately.Wearing a beige suit with multiple badges and logos hanging around his neck, prominently featuring the "Power of Reflection" and a company logo. Behind him, there is a blurred background with a white banner containing logos and text, indicating a professional or conference setting. The overall scene is vibrant and dynamic, capturing the energy of a live presentation.
Flux1.dev-AsianFemale
由于训练模型的原因,Flux 默认生成的人物都是西方面孔,即使添加 Chinese 等关键词,形象也与我们熟悉的样子有明显区别。Flux1.dev-AsianFemale 这个 lora 模型则专门针对女性形象做了优化,可以让生成的图像更符合东亚的审美偏好。
① 权重:0.4 -0.65 / 触发词:a young woman
② 参考提示词:A young woman in a blue blazer and striped tie,smiles against a blue sky.
Half_illustration
Half_illustration 是一款非常酷炫的 flux lora 模型,可以将插画风格融入现实照片;图像风格大胆时尚,非常适合作为海报素材使用。
① 权重:1 / 触发词: in the style of TOK ;最好将触发词放在第一位,并描述场景中的说明性元素,例如服装或表达元素。
② 参考提示词:In the style of TOK, a photo editorial avant-garde dramatic action pose of a woman short blue hair wearing 70s round wacky sunglasses pulling glasses down looking forward, in Tokyo with large marble structures and bonsai trees at sunset with a vibrant illustrated jacket surrounded by illustrations of flowers, smoke, flames, ice cream, sparkles, rock and roll
提示词: In the style of TOK, a photo editorial avant-garde dramatic action pose of a person wearing 90s round wacky sunglasses pulling glasses down looking forward, in Tokyo with large marble structures and bonsai trees at sunset with a vibrant illustrated jacket surrounded by illustrations of flowers, smoke, flames, ice cream, sparkles, rock and roll
酷!!!太有范儿了!Flux Dev【Lora】 极具摄影和插图风格的FLUX LoRA-half_illustration-CSDN博客
下载地址:
https://pan.baidu.com/share/init?surl=VcFJVOlNIRFJHN0pDTVosg&pwd=ggev
文生图开源模型 FLUX.1 XLabs-AI 全新发布 6 个 Lora - 知乎 (zhihu.com)
好看又好用!5 款高质量的 Flux Lora 模型推荐(内附LORA安装包) (weibo.com)
FLUX.1 社区大繁荣,盘点当前支持FLUX.1的Lora模型(附使用方式:Code+ComfyUI) - 知乎 (zhihu.com)