booksum_cleaned_v2 / README.md
taczin's picture
Upload dataset
f5dbc1a verified
metadata
dataset_info:
  features:
    - name: book_title
      dtype: string
    - name: text
      dtype: string
    - name: summary
      sequence: string
    - name: split
      dtype: string
  splits:
    - name: train
      num_bytes: 81510598.58240491
      num_examples: 3882
    - name: validation
      num_bytes: 13690085.078755282
      num_examples: 652
    - name: test
      num_bytes: 14110026.338839801
      num_examples: 672
  download_size: 66112611
  dataset_size: 109310710
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*