README.md

metadata

dataset_info:
  features:
    - name: task
      dtype: string
    - name: setting
      struct:
        - name: description
          dtype: string
        - name: name
          dtype: string
    - name: characters
      struct:
        - name: partner_name
          dtype: string
        - name: self_name
          dtype: string
        - name: self_persona
          dtype: string
    - name: dialogue
      sequence: string
  splits:
    - name: train
      num_bytes: 23424440
      num_examples: 17075
    - name: valid
      num_bytes: 1428522
      num_examples: 1000
    - name: test
      num_bytes: 2904545
      num_examples: 2000
  download_size: 9579936
  dataset_size: 27757507
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: valid
        path: data/valid-*
      - split: test
        path: data/test-*