USO

<p align="center"> <img src="assets/uso.webp" width="100"/> <p> <h3 align="center"> Unified Style and Subject-Driven Generation via Disentangled and Reward Learning </h3>

Paper: USO: Unified Style and Subject-Driven Generation via Disentangled and Reward Learning

<p align="center"> <a href="https://github.com/bytedance/USO"><img alt="Build" src="https://img.shields.io/github/stars/bytedance/USO"></a> <a href="https://bytedance.github.io/USO/"><img alt="Build" src="https://img.shields.io/badge/Project%20Page-USO-blue"></a> <a href="https://arxiv.org/abs/2508.18966"><img alt="Build" src="https://img.shields.io/badge/Tech%20Report-USO-b31b1b.svg"></a> <a href="https://huggingface.co/bytedance-research/USO"><img src="https://img.shields.io/static/v1?label=%F0%9F%A4%97%20Hugging%20Face&message=Model&color=green"></a> </p>

Abstract

Existing literature typically treats style-driven and subject-driven generation as two disjoint tasks: the former prioritizes stylistic similarity, whereas the latter insists on subject consistency, resulting in an apparent antagonism. We argue that both objectives can be unified under a single framework because they ultimately concern the disentanglement and re-composition of content and style, a long-standing theme in style-driven research. To this end, we present USO, a Unified Style-Subject Optimized customization model. First, we construct a large-scale triplet dataset consisting of content images, style images, and their corresponding stylized content images. Second, we introduce a disentangled learning scheme that simultaneously aligns style features and disentangles content from style through two complementary objectives, style-alignment training and content-style disentanglement training. Third, we incorporate a style reward-learning paradigm denoted as SRL to further enhance the model's performance. Finally, we release USO-Bench, the first benchmark that jointly evaluates style similarity and subject fidelity across multiple metrics. Extensive experiments demonstrate that USO achieves state-of-the-art performance among open-source models along both dimensions of subject consistency and style similarity. Code and model: this https URL

⚡️ Quick Start

🔧 Requirements and Installation

Install the requirements

## create a virtual environment with python >= 3.10 <= 3.12, like
python -m venv uso_env
source uso_env/bin/activate
## or
conda create -n uso_env python=3.10 -y
conda activate uso_env
## then install the requirements by you need
pip install -r requirements.txt # legacy installation command

Then download checkpoints in one of the following ways:

• Suppose you already have some of the checkpoints

# 1. download USO official checkpoints
pip install huggingface_hub
huggingface-cli download bytedance-research/USO --local-dir <YOU_SAVE_DIR> --local-dir-use-symlinks False

# 2. Then set the environment variable for FLUX.1 base model
export AE="YOUR_AE_PATH"
export FLUX_DEV="YOUR_FLUX_DEV_PATH"
export T5="YOUR_T5_PATH"
export CLIP="YOUR_CLIP_PATH"
# or export HF_HOME="YOUR_HF_HOME"

# 3. Then set the environment variable for USO
export LORA="<YOU_SAVE_DIR>/uso_flux_v1.0/dit_lora.safetensors"
export PROJECTION_MODEL="<YOU_SAVE_DIR>/uso_flux_v1.0/projector.safetensors"

• Directly run the inference scripts, the checkpoints will be downloaded automatically by the hf_hub_download function in the code.

✍️ Inference

Start from the examples below to explore and spark your creativity. ✨

# the first image is a content reference, and the rest are style references.

# for subject-driven generation
python inference.py --prompt "The man in flower shops carefully match bouquets, conveying beautiful emotions and blessings with flowers. " --image_paths "assets/gradio_examples/identity1.jpg" --width 1024 --height 1024
# for style-driven generation
# please keep the first image path empty
python inference.py --prompt "A cat sleeping on a chair." --image_paths "" "assets/gradio_examples/style1.webp" --width 1024 --height 1024
# for ip-style generation
python inference.py --prompt "The woman gave an impassioned speech on the podium." --image_paths "assets/gradio_examples/identity2.webp" "assets/gradio_examples/style2.webp" --width 1024 --height 1024
# for multi-style generation
# please keep the first image path empty
python inference.py --prompt "A handsome man." --image_paths "" "assets/gradio_examples/style3.webp" "assets/gradio_examples/style4.webp" --width 1024 --height 1024

📄 Disclaimer

<p> We open-source this project for academic research. The vast majority of images used in this project are either generated or from open-source datasets. If you have any concerns, please contact us, and we will promptly remove any inappropriate content. Our project is released under the Apache 2.0 License. If you apply to other base models, please ensure that you comply with the original licensing terms. <br><br>This research aims to advance the field of generative AI. Users are free to create images using this tool, provided they comply with local laws and exercise responsible usage. The developers are not liable for any misuse of the tool by users.</p>

Citation

We also appreciate it if you could give a star ⭐ to our Github repository. Thanks a lot!

If you find this project useful for your research, please consider citing our paper:

@article{wu2025uso,
    title={USO: Unified Style and Subject-Driven Generation via Disentangled and Reward Learning},
    author={Shaojin Wu and Mengqi Huang and Yufeng Cheng and Wenxu Wu and Jiahe Tian and Yiming Luo and Fei Ding and Qian He},
    year={2025},
    eprint={2508.18966},
    archivePrefix={arXiv},
    primaryClass={cs.CV},
}