README.md

metadata

dataset_info:
  features:
    - name: article_id
      dtype: int64
    - name: source_row_idx
      dtype: int64
    - name: source_split
      dtype: string
    - name: article_token_count
      dtype: int32
    - name: token_ids
      list: int32
    - name: token_weights
      list: float32
    - name: article_text
      dtype: string
  splits:
    - name: train_articles
      num_bytes: 185359320
      num_examples: 50000
    - name: validation_articles
      num_bytes: 25885695
      num_examples: 7000
  download_size: 124169907
  dataset_size: 211245015
configs:
  - config_name: default
    data_files:
      - split: train_articles
        path: data/train_articles-*
      - split: validation_articles
        path: data/validation_articles-*