README.md

metadata

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: audio
      dtype:
        audio:
          sampling_rate: 16000
    - name: transcription
      dtype: string
    - name: translation
      dtype: string
    - name: file
      dtype: string
  splits:
    - name: train
      num_bytes: 6392221146.655
      num_examples: 17779
    - name: dev
      num_bytes: 786905707.92
      num_examples: 2997
    - name: test
      num_bytes: 4054213966.96
      num_examples: 14916
  download_size: 8220600841
  dataset_size: 11233340821.535
  license: mit

Dataset Card for "ML2021_ASR_ST"

This dataset contains the audio recordings, the transcriptions, and the English translation of the transcriptions of the Machine Learning Course in 2021 at National Taiwan Univeristy. This can be used for domain-specific and code-switching ASR/Speech-to-text translation.

If you find this dataset useful, please consider to cite the following paper:

@article{yang2023investigating,
  title={Investigating Zero-Shot Generalizability on Mandarin-English Code-Switched ASR and Speech-to-text Translation of Recent Foundation Models with Self-Supervision and Weak Supervision},
  author={Yang, Chih-Kai and Huang, Kuan-Po and Lu, Ke-Han and Kuan, Chun-Yi and Hsiao, Chi-Yuan and Lee, Hung-yi},
  journal={arXiv preprint arXiv:2401.00273},
  year={2023}
}