GLOBE_V2_test / README.md
diffunity's picture
Save GLOBE_V2 test split as train split for cross validation
cc164ef verified
metadata
dataset_info:
  features:
    - name: audio
      dtype: audio
    - name: speaker_id
      dtype: string
    - name: transcript
      dtype: string
    - name: accent
      dtype: string
    - name: duration
      dtype: float64
    - name: age
      dtype: string
    - name: gender
      dtype: string
  splits:
    - name: train
      num_bytes: 755286998
      num_examples: 5046
  download_size: 722477750
  dataset_size: 755286998
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*