DSTC-11-Track-5 / README.md
vkehfdl1's picture
add faq id to doc_id column
912c4a3
|
raw
history blame
1.66 kB
metadata
license: apache-2.0
dataset_info:
  - config_name: default
    features:
      - name: log
        list:
          - name: speaker
            dtype: string
          - name: text
            dtype: string
      - name: target
        dtype: bool
      - name: knowledge
        list:
          - name: doc_id
            dtype: int64
          - name: doc_type
            dtype: string
          - name: domain
            dtype: string
          - name: entity_id
            dtype: int64
          - name: sent_id
            dtype: int64
      - name: response
        dtype: string
    splits:
      - name: train
        num_bytes: 22289817
        num_examples: 28431
      - name: test
        num_bytes: 4412204
        num_examples: 5475
      - name: validation
        num_bytes: 3371855
        num_examples: 4173
    download_size: 12543490
    dataset_size: 30073876
  - config_name: knowledge
    features:
      - name: domain
        dtype: string
      - name: entity_id
        dtype: int64
      - name: entity_name
        dtype: string
      - name: doc_type
        dtype: string
      - name: doc_id
        dtype: string
      - name: review_sent_id
        dtype: string
      - name: review_sentence
        dtype: string
      - name: review_metadata
        struct:
          - name: dishes
            sequence: string
          - name: drinks
            sequence: string
          - name: traveler_type
            dtype: string
      - name: faq_question
        dtype: string
      - name: faq_answer
        dtype: string
    splits:
      - name: train
        num_bytes: 2135411
        num_examples: 10882
    download_size: 535623
    dataset_size: 2135411
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: test
        path: data/test-*
      - split: validation
        path: data/validation-*
  - config_name: knowledge
    data_files:
      - split: train
        path: knowledge/train-*