README.md

metadata

dataset_info:
  features:
    - name: ep_id
      dtype: string
    - name: video
      dtype: string
    - name: question
      dtype: string
    - name: answer
      dtype: string
    - name: task_id
      dtype: string
    - name: high_level_category
      dtype: string
    - name: low_level_category
      dtype: string
    - name: num_interactions
      dtype: int64
  splits:
    - name: train
      num_bytes: 107509435
      num_examples: 79218
    - name: validation
      num_bytes: 9665221
      num_examples: 5876
  download_size: 14791026
  dataset_size: 117174656
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*