models/SameFace Fix [Flux Lora] - v0.1

SameFace Fix [Flux Lora] - v0.1

7/9/2025

5:01:56 AM

Related Keywords & Tags

character,diversify women's faces,flux.1 d,lora,lora model,negative weight lora,sameface fix [flux lora],sameface lora,stable diffusion,v0.1,vadimfedenko,woman

Portrait of a woman with blue eyes and black hair, wearing a sleeveless shirt, illuminated by blue light in a minimalistic indoor setting.

A large, serene dragon with textured scales blowing dandelion seeds into the air as children gather around on a rocky hill at sunset.

Realistic CGI image of a woman with long dark braids wearing a black dress with white collar, standing on a dirt path surrounded by green foliage under daylight, with a visible hand near the hem of her dress.

A fantasy-style young woman wearing a red hooded cloak sitting beside a large, muscular wolf in a dark, misty forest illuminated by soft ethereal light and glowing mushrooms.

Realistic CGI image of a young woman wearing a blue turtleneck sweater and golden ochre skirt, standing in an autumn forest with tall trees and wearing a black cat ear headband.

A Black woman in avant-garde attire stands silhouetted indoors against a glowing amber-toned surreal installation of translucent fabric strips suspended from the ceiling.

Hyper-realistic portrait of a fair-skinned woman with silver-white hair and platinum headdress adorned with blue gemstones, wearing a celestial embroidered cloak.

A goth girl with black hair and black lipstick stands inside an ornate ancient mirror in a ruined dungeon, wearing a sheer white gown, with smoke and mist in the background.

Cinematic portrait of a woman wearing glasses and a beige coat with artistic bokeh lighting and 35mm film photography style

Close-up portrait of an Aurora Borealis Cyborg with a teal metallic face embedded with cosmic glowing pathways and celestial light, set against a starry background.

Close-up portrait of a stone cyborg with rock-like cracked skin, glowing golden veins, intricate runic carvings, mechanical details, set against misty mountains.

Tips

Use this Lora with a NEGATIVE weight to act as a negative prompt without explicitly needing one.

This Lora diversifies women's appearances by specifying what should NOT appear in the generation, rather than adding features.

Applying high guidance with this Lora maintains image quality while avoiding typical identical faces.

The Lora is highly experimental; feedback is appreciated to improve quality and explore new variations.

Version Highlights

First Version!

Important: This Lora should be used with a NEGATIVE weight.

Also: The model is aimed at diversifying the look of women only. I plan to do a separate model for men later (and men with beards especially)

Also cool fact: the size of the Lora is only 4mb (it is limited to 2 blocks with a high rank)

So, What's a "Sameface"?

It's a common problem in generative AI, in which the model generates people with identical appearances. Unlike with SD1.5/SDXL, this is not just a result of the diverse dataset converging into a single token like "woman". in Flux, Dev/Schnell were distilled from the Pro version, resulting in a loss of diversity.

Why use Negative weight?

Lora was trained on the most generic photos of women that Flux generates.

If you then apply this Lora with a NEGATIVE weight, it will work as a sort of negative prompt, but without negative prompt!

By using this clever trick, Lora will basically tell Flux to generate anything except a typical Flux face. And as result, we'll get beautiful unique faces every time!

Why Lora and not tricks like lower Guidance?

People want diverse faces with high guidance in order to maintain high prompt adherence and image quality, that's why interventions at the latent space level are optimal.

How does it differ from other character Loras?

While other Loras introduce their own "samefaces", this Lora, only tells Flux what should NOT be in the generation, not what to add. As a result, this Lora diversifies people's appearance, rather than narrowing it even further.

How was the Lora actually made?

At first I was disappointed because nothing worked out. I was about to give up this project, but then tried to train Lora with just 2 layers, and suddenly during testing, I was met with beautiful faces, and with authentic, distant from Flux, appearance.

What's next?

I have a lot to experiment with. Different learning rates, layers and ranks have to be tested in order to improve quality. Merging different Lora variants might fix the problems with artifacts at higher values. I'll make sure to make a better version, If this Lora has enough interest

I'm really forward to your feedback, as this model is highly experimental. Any feedback is appreciated and will hopefully help me to improve the model.

special thanks to Ostris for AI-Toolkit