README.md

metadata

language:
  - en
dataset_info:
  features:
    - name: id
      dtype: string
    - name: title
      dtype: string
    - name: context
      dtype: string
    - name: question
      dtype: string
    - name: answers
      sequence:
        - name: text
          dtype: string
        - name: answer_start
          dtype: int32
  splits:
    - name: train
      num_bytes: 24489106
      num_examples: 27045
    - name: valid
      num_bytes: 3337865
      num_examples: 3688
    - name: test
      num_bytes: 10472984
      num_examples: 10570
  download_size: 19442674
  dataset_size: 38299955
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: valid
        path: data/valid-*
      - split: test
        path: data/test-*