今天来聊一款近期备受关注的国产Flux LORA模型——新世界电商超模 FLUX.1-SuperModel 东方审美。顾名思义,这款AI绘画模型主打电商场景下的“东方审美”方向,并通过LiblibAI平台的FLUX在线训练器完成训练,堪称一次相当克制的落地实践。
先划几个重点:
- 这是一款首发免费的LORA模型,在线出图完全免费。更值得一提的是,作者直接赋予了自动商用授权,对于从事电商内容创作的用户而言,非常友好省心。
- 它的核心价值在于:有效弥补了Flux基底模型在亚洲面孔、服饰及文化元素方面的单一表现问题。简单来说,以往Flux出图虽然能生成好看的脸,但面孔类型相对单一。引入这个LORA后,更多东方特征的细节——从五官轮廓到服装纹理——都能得到更精准的呈现。
- 应用场景十分明确:虚拟模特、模特换装、图片优化、虚拟试穿、自动化内容生产、低成本解决方案、创意设计辅助。这些都是当前AIGC领域最贴合商业需求的落地方向。
说回参数:
- 推荐CFG:2.5~3.5
- 基础触发词:无(不依赖特定唤醒词)
- 推荐LORA权重:0.85
这个权重设定比较合理——0.85既能保留Flux本身出色的光影和质感,又能有效激活LORA的东方风格特征,不会喧宾夺主。
新世界电商超模 ComfyUI体验
如果你想在本地ComfyUI中运行这款模型,可以参考之前发过的Flux系列文章,特别是关于Dev版工作流搭建的那篇。下面展示的是基于该工作流、配合这个LORA模型生成的一组样片。样片涵盖多个风格方向,充分验证了模型的覆盖能力。
工作流
关于Flux模型本地ComfyUI工作流体验,可参考之前的相关内容。本文涉及的ComfyUI工作流和模型均可在LIBLIBAI上下载或在线运行体验。

样片01:写实摄影。采用85mm f/1.8中焦镜头,极简背景,模特面部细节丰富,肤色自然,完全没有明显的“AI脸”痕迹。
Photographer: Philip Halsman (Philippe Halsman)
Model features: Asian female models dressed in simple modern attire, showing confidence and elegance
Shot Description: The camera uses a 85mm f/1.8 mid-focus lens to capture details of the model against a minimalist background.

样片02:拱手作揖。偏创意的场景,模特手持小猴,动物呈冥想手势,整体传达一种人与自然和谐共处的意象。
Alexander McQueen, Model, “Capture the serene interaction between human and nature with a thoughtful composition. Using an 85mm f/1.8 mid-range lens, focus on an Asian female model dressed in contemporary simplicity, exuding confidence and grace. She holds a small monkey in her palm, the creature in a meditative pose, hands clasped, embodying a moment of mutual respect and tranquility.”

样片03/04:敦煌风。两组作品方向一致——利用24mm广角镜头,结合自然光与环境光,模特身穿传统服饰,光影层次处理得十分舒适。这类风格对LORA在中国元素还原能力上是一次极致的考验。
“Envision a captivating outdoor portrait by renowned photographer Cyril Vannini, where an Asian female model, adorned in exquisite traditional costumes, embodies a rich tapestry of cultural heritage. Utilizing a 24mm f/1.8 wide-angle lens, the scene is bathed in the soft, golden hues of natural sunlight, emphasizing the model’s graceful interaction with her surroundings.”


样片05:都市生活。偏街头抓拍风格,模特穿着日常休闲装,35mm f/2.0广角镜头,营造出浓厚的城市生活氛围。
Photographer: Elliot Erwitt
Model features: Asian female models wear daily casual wear, showing the atmosphere of urban life.
Lens description: The camera uses a wide-angle lens of 35mm f/2.0 to capture the dynamic and urban atmosphere of the model on the street.

样片06:沙滩跑步。模特身着运动装,50mm f/1.8标准镜头,在自然光下捕捉动态瞬间。
Photographer: Martin Munkácsi
Model features: Asian female models wear light sportswear, showing vitality and movement on the beach.
Lens description: The camera uses a standard lens of 50mm f/1.8 to capture the moment the model runs in natural light.

样片07:打工人。35mm f/1.8广角镜头,工作场景真实感强烈,模特展现出工人群体的坚韧与力量。
Photographer: Sebastian Salgado (Salgado Sebasti)
Model features: Asian female models wear work clothes, showing the tenacity of workers.
Lens description: The camera uses a wide-angle lens 35mm f/1.8 to capture the real moment of the model on the job site.

样片08:涂鸦墙。街头朋克风格,亚洲面孔搭配格纹套装、安全别针与锁链元素,涂鸦墙作为背景,Nikon D850 + 35mm定焦,整体张力十足。
Vivienne Westwood,
Model: An Asian woman wearing a punk-inspired tartan ensemble with safety pins and chains.
Description: The model is photographed against a graffiti-covered wall, giving a rebellious vibe. The designer used a Nikon D850 with a 35mm prime lens to capture the raw energy of the look.

样片09:自由写真。昏暗地下空间中拍摄,光影交错,35mm镜头营造出亲密感和神秘氛围。
Kohei Yoshiyuki (Jixing)
Model: An Asian woman dressed in a modern, a vant-garde outfit, incorporating bold patterns and textures.
Description: The model is photographed in a dimly lit underground space, with shadows playing across her form. Yoshiyuki used a compact camera with a 35mm lens to create a sense of intimacy and mystery.

样片10:热闹沙滩。采用24mm超广角镜头,模特在拥挤海滩中脱颖而出,背景人群丰富,色彩抢眼。
Martin Parr
Model: An Asian woman dressed in bright, patterned beachwear and holding a colorful beach towel.
Description: The model is photographed at a crowded beach, with other beachgoers in the background. Parr used a digital camera with a 24mm wide-angle lens to capture the lively scene and the model’s presence among the crowd.

从以上样片来看,这款LORA对各种风格场景的适配度都相当到位。无论是商业写真、传统敦煌风,还是街头抓拍,它都能在保留Flux大模型底子优势的同时,提升东方题材的呈现精度。如果你从事电商设计、服装展示、模特方案快速验证等工作,非常值得一试这款模型。而且免费的商用授权,用起来更省心。
