Emotional Photography
Focuses on capturing human emotions as the core, using lighting, color, setting, and composition to intentionally create an atmosphere that amplifies or mirrors the subject's feelings, aiming to convey stories and moods through visual language.
情绪摄影
以人物情感状态为创作起点，通过光线、色彩、场景、构图等视觉元素的协同作用，重点营造与情绪高度契合的环境氛围，最终实现“用画面传递情绪与故事”。
IMPORTANT USAGE TIPS:
First, you'll need to install this plugin for your ComfyUI: https://github.com/ClownsharkBatwing/RES4LYF
Once it's installed, you'll get a bunch of new, high-quality samplers and schedulers in your ComfyUI.
I recommend using Sampler: res_2s and Scheduler: bong_tangent.
I also recommend using it with this distilled LoRA for faster generation:
https://huggingface.co/lightx2v/Qwen-Image-Lightning/resolve/main/Qwen-Image-fp8-e4m3fn-Lightning-4steps-V1.0-fp32.safetensors?download=true
All demo images were generated using:
Steps: 4
CFG: 1
Sampler: res_2s
Scheduler: bong_tangent
Emotional Photography lora strength 1.0
Qwen-Image-Lightning-4steps lora strength 0.4~0.5
PS: If you plan on merging this, please give credit or link back to this page. Thanks
Description
FAQ
Comments (11)
Is this biased towards Japanese people, or is that just a coincidence?
not Japanese, they are Chinese
faralson OK fair enough, looking again you are right. My question was more - does the lora induce a particular ethnicity, or is it easy to get the aesthetic style effects without the lora making everyone Chinese
But I see the author has uploaded another example of a caucasian person, so will download and check it out - the style looks really good!
These show great emotional body language! Fantastic idea! I would love to see a more western version in the future. But for now this is inspired! great job!
大佬！模型很棒！请问fp8和bf16是不是差距比较大阿，加了加速完全没有例图这么好看
fp8我没测试，不敢说影响有多大，但模型是在fp8量化下训练的。另外如果你说加速是指蒸馏加速的lora（8步），那么在加速后效果会有衰退，需要调整权重和更换采样器，推荐deis系列采样器。使用加速lora后对真实人像的质感影响都比较大，目前我没有找到两全其美的办法。
颗粒感是直出就有的吗😊
颗粒主要来自节点，图片都带有comfyui工作流，我在工作流中有说明
fok3827 大佬牛逼~~~模型质量都好高
Lora's quality is good. Thank you for your hard work and care.
I usually use flux, but I think qwen is one level higher in terms of body anatomy. However, it lacks aesthetics. For example, the subject and the background are separated too clearly. It's a phenomenon where the texture of the skin is uniform
For qwen-image to be successful, we need to have a lot of lora and fine tuning that people like you make. The reason we use generative ai is not just for nsfw :)
To add one more thing, if you set the weight of lora to 1.0, the saturation tends to go up with the concentration of color grading, so I set it to 0.5
大佬有没有工作流，我下了那个节点老是报错。。。
Details
Files
Available On (1 platform)
Same model published on other platforms. May have additional downloads or version variants.