Datasets:

piyazon
/

multilingual_voice_ug_en_zh

dataset_info:
  features:
    - name: audio
      dtype:
        audio:
          sampling_rate: 16000
    - name: sentence
      dtype: string
    - name: sentence_latn
      dtype: string
    - name: language
      dtype: string
  splits:
    - name: train
      num_bytes: 12844208621
      num_examples: 349559
    - name: validation
      num_bytes: 129743256
      num_examples: 3531
  download_size: 12630059723
  dataset_size: 12973951877
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
task_categories:
  - automatic-speech-recognition
language:
  - ug
  - zh
  - en

train:

'Uyghur': 266672
'Chinese': 50215
'English': 32672

validation:

'Uyghur': 2712
'Chinese': 491
'English': 328