---
title: WSYue-TTS-eval
canonical_url: "https://www.modelscope.cn/datasets/ASLP-lab/WSYue-TTS-eval"
md_url: "https://www.modelscope.cn/datasets/ASLP-lab/WSYue-TTS-eval.md"
repository: ASLP-lab/WSYue-TTS-eval
last_updated: 2025-09-09
license: cc-by-nc-4.0
storage_size: "127 MB"
downloads: 267
stars: 0
---

# WSYue-TTS-eval

> WSYue-TTS-eval - ASLP-lab 在 ModelScope 开源的数据集。WSYue-TTS-eval: Cantonese TTS Benchmark

ASLP-lab/WSYue-TTS-eval 是 ModelScope 魔搭社区上的数据集，存储大小 127 MB，采用 cc-by-nc-4.0 许可。

- **Repository**: ASLP-lab/WSYue-TTS-eval
- **License**: cc-by-nc-4.0
- **Storage size**: 127 MB
- **Downloads**: 267
- **Stars**: 0
- **Last updated**: 2025-09-09

Source: https://www.modelscope.cn/datasets/ASLP-lab/WSYue-TTS-eval

---

# WSYue-TTS-eval: Cantonese TTS Benchmark

To address the unique linguistic characteristics of Cantonese in speech synthesis, we propose WSYue-TTS-eval, a benchmark specifically designed for zero-shot Cantonese TTS evaluation. It enables rigorous assessment of model generalization and robustness across diverse domains and challenging linguistic phenomena. 

## Subsets
- Base  
  - 1,000 prompt–text pairs sampled from CommonVoice  
  - Evaluates performance on real-world data distributions  
  - Mainly daily conversational style

- Coverage 
  - Combines manually curated and LLM-generated texts  
  - Spans multiple domains: daily life, news, entertainment, poetry 
  - Covers diverse Cantonese linguistic phenomena:
    - Polyphonic characters  
    - Tone sandhi  
    - Code-switching  
    - Proper nouns & numerals  
    - Other challenging linguistic cases  

### Highlights
- Designed for rigorous evaluation of TTS systems  
- Tests generalization and robustness across diverse and complex Cantonese scenarios