README.md

---
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: dev
    path: data/dev-*
  - split: test
    path: data/test-*
dataset_info:
  features:
  - name: correct_audio
    dtype:
      audio:
        sampling_rate: 16000
  - name: correct_transcription
    dtype: string
  - name: correct_file
    dtype: string
  - name: wrong_audio
    dtype:
      audio:
        sampling_rate: 16000
  - name: wrong_transcription
    dtype: string
  - name: wrong_file
    dtype: string
  splits:
  - name: train
    num_bytes: 7561544424.98
    num_examples: 23549
  - name: dev
    num_bytes: 2665949331.86
    num_examples: 8505
  - name: test
    num_bytes: 929488114.48
    num_examples: 3176
  download_size: 10860817060
  dataset_size: 11156981871.32
license: mit
language:
- zh
- en
---

This dataset contains the Mandarin-English track of the benchmark from ICASSP 2024: Zero Resource Code-Switched Speech Benchmark Using Speech Utterance Pairs for Multiple Spoken Languages.    
Though the benchmark is originally designed to assess the semantic and syntactic abilities of the speech foundation models, you can also use this dataset for code-switching ASR.


If you find this dataset helpful, please consider to cite the following paper:
```
@INPROCEEDINGS{10446737,
  author={Huang, Kuan-Po and Yang, Chih-Kai and Fu, Yu-Kuan and Dunbar, Ewan and Lee, Hung-Yi},
  booktitle={ICASSP 2024 - 2024 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)}, 
  title={Zero Resource Code-Switched Speech Benchmark Using Speech Utterance Pairs for Multiple Spoken Languages}, 
  year={2024},
  volume={},
  number={},
  pages={10006-10010},
  keywords={Speech coding;Benchmark testing;Signal processing;Linguistics;Acoustics;Speech processing;Task analysis;Code-switch;Multilingual;Discrete unit;Zero resource;Self-supervised},
  doi={10.1109/ICASSP48485.2024.10446737}}
```