Nan Chen† ·Yiyang Cai† · Rongchang Xie · Junwen Pan · Cheng Chen · Weinan Jia · Zhuowei Chen ·
Wen Zhou‡ · Zhenbang Sun · Wenhan Luo✉
†Equal Contribution. ‡Project Leader ✉Corresponding Author
TL; DR: We propose DomainShuttle, an open-domain subject-driven text-to-video method that flexibly handles both in-domain fidelity and cross-domain editability by decoupling reference and video features, modeling domain attributes, and learning intrinsic subject representations.
📖 Jun 23, 2026: We release the technical report.
🔥 Jun 23, 2026: We release the unofficial DomainShuttle implementation weights, inference code, and project page.
- [✅] Inference code
- [✅] checkpoint
- [✅] Technical report
conda create -n DomainShuttle python=3.10
conda activate DomainShuttle
pip install torch==2.5.1 torchvision==0.20.1 torchaudio==2.5.1 --index-url https://download.pytorch.org/whl/cu124
bash build_env_conda.sh
| Models | Download Link | Notes |
|---|---|---|
| Wan2.2-T2V-A14B | 🤗 Huggingface | Base model |
| DomainShuttle | 🤗 Huggingface | Our 14B weights |
Download models using huggingface-cli:
mkdir models/Diffusion_Transformers
hf download CNcreator0331/DomainShuttle_weight --local-dir ./models/Diffusion_Transformers/Wan2.2-DomainShuttle-A14B
hf download Wan-AI/Wan2.2-T2V-A14B --local-dir ./checkpoints/Wan2.2-T2V-A14B
mv ./checkpoints/Wan2.2-T2V-A14B/google ./models/Diffusion_Transformers/Wan2.2-DomainShuttle-A14B/
mv ./checkpoints/Wan2.2-T2V-A14B/Wan2.1_VAE.pth ./models/Diffusion_Transformers/Wan2.2-DomainShuttle-A14B/
mv ./checkpoints/Wan2.2-T2V-A14B/configuration.json ./models/Diffusion_Transformers/Wan2.2-DomainShuttle-A14B/
mv ./checkpoints/Wan2.2-T2V-A14B/models_t5_umt5-xxl-enc-bf16.pth ./models/Diffusion_Transformers/Wan2.2-DomainShuttle-A14B/
The directory should be organized as follows.
models/Diffusion_Transformers/
└── Wan2.2-DomainShuttle-A14B/
├── google/
├── high_noise_model/
├── low_noise_model/
├── configuration.json
└── models_t5_umt5-xxl-enc-bf16.pth
└── Wan2.1_VAE.pth
This unofficial inference script can run 480p/720p inference on GPU.
bash run_wan22_domainshuttle.shWe provide some test cases in test_case folder. You can also try our model with your own data. You can change the reference image in the json file.
If you find our work useful in your research, please consider citing.
The models in this repository are licensed under the Apache 2.0 License. We claim no rights over the your generated contents, granting you the freedom to use them while ensuring that your usage complies with the provisions of this license. You are fully accountable for your use of the models, which must not involve sharing any content that violates applicable laws, causes harm to individuals or groups, disseminates personal information intended for harm, spreads misinformation, or targets vulnerable populations.