---
language:
  - en
tags:
  - text-to-speech
  - speech
  - emotional-speech
  - instruction-following
  - emotion-trajectory
  - emotion-blending
base_model: FunAudioLLM/Fun-CosyVoice3-0.5B-2512
---

# HybridEmo

HybridEmo is an instruction-following multi-emotion text-to-speech model developed by post-training [CosyVoice 3](https://huggingface.co/FunAudioLLM/Fun-CosyVoice3-0.5B-2512). It supports sequential emotion trajectories and simultaneous emotion blending within a single utterance.

## Resources

- [Paper](https://arxiv.org/abs/2608.30325)
- [GitHub Repository](https://github.com/ictnlp/HybridEmo)
- [Audio Demo](https://zhouyan19.github.io/HybridEmo-demo/)
- [Hugging Face Collection](https://huggingface.co/collections/ICTNLP/hybridemo)

## Usage

Please refer to the [GitHub repository](https://github.com/ictnlp/HybridEmo) for environment setup and inference instructions.

## Citation

If you use HybridEmo in your research, please cite our paper:

```bibtex
@misc{zhou2026sequentialtrajectoriessimultaneousblending,
  title={Sequential Trajectories and Simultaneous Blending: Multi-Emotion Modeling for Instruction-Following TTS},
  author={Yan Zhou and Yun Hong and Yang Feng},
  year={2026},
  eprint={2608.30325},
  archivePrefix={arXiv},
  primaryClass={cs.CL},
  url={https://arxiv.org/abs/2608.30325},
}
```