README.md

metadata

dataset_info:
  - config_name: ar-en
    features:
      - name: client_id
        dtype: string
      - name: file
        dtype: string
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: sentence
        dtype: string
      - name: translation
        dtype: string
      - name: id
        dtype: string
    splits:
      - name: train
        num_examples: 2283
      - name: test
        num_examples: 1695
      - name: validation
        num_examples: 1758
  - config_name: en-ar
    features:
      - name: client_id
        dtype: string
      - name: file
        dtype: string
      - name: audio
        dtype:
          audio:
            sampling_rate: 16000
      - name: sentence
        dtype: string
      - name: translation
        dtype: string
      - name: id
        dtype: string
    splits:
      - name: train
        num_examples: 289430
      - name: test
        num_examples: 15531
      - name: validation
        num_examples: 15531
configs:
  - config_name: ar-en
    data_files:
      - split: train
        path: ar-en/train-*
      - split: validation
        path: ar-en/validation-*
      - split: test
        path: ar-en/test-*
  - config_name: en-ar
    data_files:
      - split: train
        path: en-ar/train-*
      - split: validation
        path: en-ar/validation-*
      - split: test
        path: en-ar/test-*
license: cc-by-nc-4.0
task_categories:
  - automatic-speech-recognition
  - text-to-speech
  - translation
language:
  - ar
  - en
size_categories:
  - 100K<n<1M

Dataset Description

CoVoST 2 is a large-scale multilingual speech translation corpus based on Common Voice, developed by FAIR. This is the English-to-Arabic portion of the dataset. The original dataset can be found here.

Data Splits (EN-AR)

lang	train	validation	test
EN-AR	289430	15531	15531
AR-EN	2283	1758	1695

Citation

@misc{wang2020covost,
    title={CoVoST 2: A Massively Multilingual Speech-to-Text Translation Corpus},
    author={Changhan Wang and Anne Wu and Juan Pino},
    year={2020},
    eprint={2007.10310},
    archivePrefix={arXiv},
    primaryClass={cs.CL}
}