README.md

metadata

title: PrismAudio
emoji: 🎵
colorFrom: purple
colorTo: blue
sdk: gradio
sdk_version: 5.42.0
python_version: '3.10'
app_file: app.py
pinned: false

PrismAudio

       

---

PrismAudio is the first framework to integrate reinforcement learning into video-to-audio (V2A) generation, equipped with a dedicated Chain-of-Thought (CoT) planning mechanism. Building on the pioneering CoT-based V2A framework of ThinkSound, PrismAudio further decomposes single-step reasoning into four specialized CoT modules — semantic, temporal, aesthetic, and spatial — each with targeted reward functions, enabling multi-dimensional RL optimization that simultaneously improves reasoning across all perceptual dimensions.

Quick Start

For full training and inference details, please refer to the ThinkSound prismaudio branch.

git clone -b prismaudio https://github.com/liuhuadai/ThinkSound.git
cd ThinkSound

conda create -n prismaudio python=3.10
conda activate prismaudio
chmod +x scripts/PrismAudio/setup/build_env.sh
./scripts/PrismAudio/setup/build_env.sh

# Download pretrained weights to ckpts/
# From Hugging Face:  https://huggingface.co/FunAudioLLM/PrismAudio
# From ModelScope:    https://www.modelscope.cn/models/iic/PrismAudio
git lfs install
git clone https://huggingface.co/FunAudioLLM/PrismAudio ckpts

---

License

This project is released under the MIT License.

Note: The code, model weights, and datasets are intended for research and educational purposes only. Commercial use is not permitted without explicit authorization from the authors.

---

Citation

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

@misc{liu2025thinksoundchainofthoughtreasoningmultimodal,
      title={ThinkSound: Chain-of-Thought Reasoning in Multimodal Large Language Models for Audio Generation and Editing}, 
      author={Huadai Liu and Jialei Wang and Kaicheng Luo and Wen Wang and Qian Chen and Zhou Zhao and Wei Xue},
      year={2025},
      eprint={2506.21448},
      archivePrefix={arXiv},
      primaryClass={eess.AS},
      url={https://arxiv.org/abs/2506.21448}, 
}

@misc{liu2025prismaudiodecomposedchainofthoughtsmultidimensional,
        title={PrismAudio: Decomposed Chain-of-Thoughts and Multi-dimensional Rewards for Video-to-Audio Generation}, 
        author={Huadai Liu and Kaicheng Luo and Wen Wang and Qian Chen and Peiwen Sun and Rongjie Huang and Xiangang Li and Jieping Ye and Wei Xue},
        year={2025},
        eprint={2511.18833},
        archivePrefix={arXiv},
        primaryClass={cs.SD},
        url={https://arxiv.org/abs/2511.18833}, 
}

---

Contact

If you have any questions or suggestions, feel free to open an issue