SoulX-FlashHead-1_3B

Soul-AILab/SoulX-FlashHead-1_3B

Tan Yu*, Qian Qiao*✉, Le Shen*, Ke Zhou, Jincheng Hu, Dian Sheng, Bo Hu, Haoming Qin, Jun Gao, Changhai Zhou, Shunshun Yin, Siyuan Liu ✉

*Equal Contribution ✉Corresponding Author

⚡ Highlights

• Model_Lite Released get 96 FPS, or 3-concurrent real-time(25+ FPS) streaming on single RTX4090.
• Model_Pro Released can generate high-quality videos with 10.8 FPS on single RTX4090, or real-time(25+ FPS) on two RTX5090.
• Model_Pretrained is coming soon, providing high-performance weights and experimental foundations for community research.

🔥 News

• 2026.02.12 - The online demo is now available via the Soul App. Download it today to try it out. Download it today to try it out.
• 2026.02.12 - We have released the inference code, and the model weights.
• 2026.02.12 - We released Project page on SoulX-FlashHead.
• 2026.02.07 - We released Dataset.
• 2026.02.07 - We released SoulX-FlashHead Technical Report on Arxiv and GitHub repository.

📑 Todo List

• Technical report
• Project Page
• Inference code
• Distilled Checkpoint of Pro-Model & Lite-Model release
• Pretrained Checkpoint release

🌰 Examples

More examples are available in the project.

📖 Quickstart

🔧 Installation

1. Create a Conda environment

conda create -n flashhead python=3.10
conda activate flashhead

2. Install PyTorch on CUDA

pip install torch==2.7.1 torchvision==0.22.1 --index-url https://download.pytorch.org/whl/cu128

3. Install other dependencies

pip install -r requirements.txt

4. FlashAttention installation:

pip install ninja
pip install flash_attn==2.8.0.post2 --no-build-isolation

-- If it takes a long time, we recommend the way below.

1. download wheel file from here
2. pip install xxx.whl

5. SageAttention installation (Optional)

pip install sageattention==2.2.0 --no-build-isolation

5. FFmpeg installation

# Ubuntu / Debian
apt-get install ffmpeg
# CentOS / RHEL
yum install ffmpeg ffmpeg-devel

or

# Conda (no root required) 
conda install -c conda-forge ffmpeg==7

🤗 Model download

# If you are in china mainland, run this first: export HF_ENDPOINT=https://hf-mirror.com
pip install "huggingface_hub[cli]"
huggingface-cli download Soul-AILab/SoulX-FlashHead-1_3B --local-dir ./models/SoulX-FlashHead-1_3B
huggingface-cli download facebook/wav2vec2-base-960h --local-dir ./models/wav2vec2-base-960h

🚀 Inference

# Infer with [Pro-Model] on single GPU
bash inference_script_single_gpu_pro.sh


# Infer with [Pro-Model] on multy GPUs
bash inference_script_multi_gpu_pro.sh
# Real-time inference speed of Pro-Model can only be supported on two RTX-5090 with SageAttention.


# Infer with [Lite-Model] on single GPU
bash inference_script_single_gpu_lite.sh
# Real-time inference speed can be supported on single RTX-4090 (up to 3 concurrent).

👋 Online Experience

For a real-time interactive experience, scan the QR code to enter the event link. [2026.2.12~2026.3.11]

📧 Contact Us

If you are interested in leaving a message to our work, feel free to email yutan@soulapp.cn or qiaoqian@soulapp.cn or le.shen@mail.dhu.edu.cn or zhouke@soulapp.cn or liusiyuan@soulapp.cn

We have opened a WeChat group. Additionally, we represent SoulApp and warmly welcome everyone to download the app and join our Soul group for further technical discussions and updates!

Component Licenses:

1. SoulX‑FlashHead‑1.3B transformer DiT weights: Apache‑2.0
2. VAE_WAN (Wan‑2.1 VAE): Apache‑2.0
3. VAE_LTX (LTX‑Video‑0.9 VAE): Open‑RAIL‑M
4. wav2vec2‑base‑960h: Apache‑2.0

Important note: When you redistribute this model archive, you MUST preserve the license of VAE_LTX as Open‑RAIL‑M, cannot re‑license it to Apache‑2.0. If you only use Pro variant (VAE_WAN), the whole workflow is under Apache‑2.0. Lite variant requires complying to Open‑RAIL‑M terms for the LTX‑0.9 VAE component.

📚 Citation

If you find our work useful in your research, please consider citing:

@misc{yu2026soulxflashheadoracleguidedgenerationinfinite,
      title={SoulX-FlashHead: Oracle-guided Generation of Infinite Real-time Streaming Talking Heads}, 
      author={Tan Yu and Qian Qiao and Le Shen and Ke Zhou and Jincheng Hu and Dian Sheng and Bo Hu and Haoming Qin and Jun Gao and Changhai Zhou and Shunshun Yin and Siyuan Liu},
      year={2026},
      eprint={2602.07449},
      archivePrefix={arXiv},
      primaryClass={cs.CV},
      url={https://arxiv.org/abs/2602.07449}, 
}

🙇 Acknowledgement

• Wan: the base model we built upon.
• LTX-Video: the VAE of our Lite-Model.
• Self forcing: the codebase we built upon.
• DMD and Self forcing++: the key distillation technique used by our method.
• SoulX-FlashTalk is another model developed by our team, featuring 14B parameters and real-time capabilities.

[!TIP] If you find our work useful, please also consider starring the original repositories of these foundational methods.

💡 Star History