FLUX-customID: Realistically Customize Your Personal ID to Perfection

This repository is the official implementation of FLUX-customID. It is capable of generating images based on your face image at a level equivalent to real photographic quality. Our base model is FLUX.dev, which ensures the generation of high-quality images.

News

• 🌟2024-11-13: Released the code and weights for FLUX-customID.

Gallery

Here are some example samples generated by our method.

Deployment Options

Option 1: Local Setup

1. Setup Repository and Environment

conda create -n customID python=3.10 -y
conda activate customID
conda install pytorch==2.4.0 torchvision==0.19.0 pytorch-cuda=11.8 -c pytorch -c nvidia  -y
pip install -i https://mirrors.cloud.tencent.com/pypi/simple  diffusers==0.31.0 transformers onnxruntime-gpu insightface sentencepiece matplotlib imageio tqdm numpy einops accelerate peft

2. Prepare Pretrained Checkpoints

git clone https://github.com/damo-cv/FLUX-customID.git
cd FLUX-customID

mkdir pretrained_ckpt
cd pretrained_ckpt

#Download CLIP
export HF_ENDPOINT=https://hf-mirror.com
pip install -U "huggingface_hub[cli]"

huggingface-cli download \
--resume-download "laion/CLIP-ViT-H-14-laion2B-s32B-b79K" \
--cache-dir your_dir/

ln -s your_dir/models--laion--CLIP-ViT-H-14-laion2B-s32B-b79K/snapshots/de081ac0a0ca8dc9d1533eed1ae884bb8ae1404b pretrained_ckpt/openclip-vit-h-14

#Download FLUX.1-dev
huggingface-cli download \
--resume-download "black-forest-labs/FLUX.1-dev" \
--cache-dir your_dir/

ln -s your_dir/models--black-forest-labs--FLUX.1-dev/snapshots/303875135fff3f05b6aa893d544f28833a237d58 pretrained_ckpt/flux.1-dev

#Download FLUX-customID
Download our trained checkpoint from https://huggingface.co/Damo-vision/FLUX-customID and place FLUX-customID.pt in the floder pretrained_ckpt/

3. Quick Inference

run infer_customID.ipynb

Option 2: RunPod Serverless Deployment

For production deployments or scaling to multiple users, we provide a RunPod serverless implementation.

Key Features

• Automatic scaling based on demand
• GPU acceleration with NVIDIA A5000 or better
• Persistent model storage
• RESTful API endpoint
• Detailed logging and monitoring

Quick Setup

1. Follow the detailed instructions in RUNPOD_DEPLOYMENT.md
2. Key requirements:
   • RunPod account
   • 80GB storage volume for models
   • NVIDIA GPU (16GB+ VRAM)

API Usage

import requests
import base64

def encode_image(image_path):
    with open(image_path, "rb") as f:
        return base64.b64encode(f.read()).decode('utf-8')

# Replace with your endpoint URL
endpoint = "https://your-endpoint.runpod.net"
image = encode_image("path/to/your/image.jpg")

response = requests.post(endpoint, json={
    "input": {
        "image": image,
        "prompt": "professional ID photo",
        "num_samples": 1,
        "height": 1024,
        "width": 1024
    }
})

# Response includes base64-encoded generated images
result = response.json()

For detailed deployment instructions and troubleshooting, see RUNPOD_DEPLOYMENT.md

Preview for CustomAnyID

We would like to announce that we are currently working on a related project, CustomAnyID. Below are some preliminary experimental results:

Preview for Controlnet

We would like to announce our Controlnet model. Below are some preliminary experimental results:

System Requirements

Local Setup

• Python 3.10
• CUDA 11.8
• 16GB+ GPU VRAM
• 32GB+ System RAM

RunPod Deployment

• NVIDIA A5000 or better GPU
• 80GB storage volume
• 20GB container disk

Contact Us

Dongyang Li: [email protected]

Acknowledgements

The partial code is implemented based on IP-Adapter and PhotoMaker.