README.md

metadata

dataset_info:
  features:
    - name: speaker_id
      dtype: string
    - name: gender
      dtype: string
    - name: transcript
      dtype: string
    - name: interval
      sequence: float64
    - name: untranscripted
      dtype: bool
    - name: topic
      dtype: string
    - name: audio
      dtype: audio
  splits:
    - name: train
      num_bytes: 234737425.48921812
      num_examples: 3078
    - name: test
      num_bytes: 12430884.226781858
      num_examples: 163
  download_size: 242347299
  dataset_size: 247168309.71599996
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*

Data set source: https://magichub.com/datasets/egyptian-arabic-conversational-speech-corpus/