README.md

metadata

dataset_info:
  - config_name: coqa-instruct-7s
    features:
      - name: interactions
        list:
          - name: answer
            dtype: string
          - name: query
            dtype: string
    splits:
      - name: train
        num_bytes: 12736689
        num_examples: 6522
    download_size: 8242530
    dataset_size: 12736689
  - config_name: coqa-mix-7s
    features:
      - name: interactions
        list:
          - name: answer
            dtype: string
          - name: query
            dtype: string
          - name: think
            dtype: string
    splits:
      - name: train
        num_bytes: 35903600
        num_examples: 5343
    download_size: 18358303
    dataset_size: 35903600
  - config_name: coqa-reasoning-7s
    features:
      - name: interactions
        list:
          - name: answer
            dtype: string
          - name: query
            dtype: string
          - name: think
            dtype: string
    splits:
      - name: train
        num_bytes: 48032882
        num_examples: 4063
    download_size: 22755125
    dataset_size: 48032882
  - config_name: narrativeqa-instruct-7s
    features:
      - name: interactions
        list:
          - name: answer
            dtype: string
          - name: query
            dtype: string
    splits:
      - name: train
        num_bytes: 15963018
        num_examples: 4029
    download_size: 4367955
    dataset_size: 15963018
  - config_name: narrativeqa-mix-7s
    features:
      - name: interactions
        list:
          - name: answer
            dtype: string
          - name: query
            dtype: string
          - name: think
            dtype: string
    splits:
      - name: train
        num_bytes: 28037225
        num_examples: 2540
    download_size: 13989587
    dataset_size: 28037225
  - config_name: narrativeqa-reasoning-7s
    features:
      - name: interactions
        list:
          - name: answer
            dtype: string
          - name: query
            dtype: string
          - name: think
            dtype: string
    splits:
      - name: train
        num_bytes: 30253631
        num_examples: 1750
    download_size: 12125278
    dataset_size: 30253631
  - config_name: passkey-multitopic-instruct-7s
    features:
      - name: interactions
        list:
          - name: answer
            dtype: string
          - name: query
            dtype: string
    splits:
      - name: train
        num_bytes: 64804413
        num_examples: 1100
    download_size: 34425675
    dataset_size: 64804413
  - config_name: passkey-multitopic-mix-7s
    features:
      - name: interactions
        list:
          - name: answer
            dtype: string
          - name: query
            dtype: string
          - name: think
            dtype: string
    splits:
      - name: train
        num_bytes: 88633909
        num_examples: 1409
    download_size: 46718123
    dataset_size: 88633909
  - config_name: passkey-multitopic-reasoning-7s
    features:
      - name: interactions
        list:
          - name: answer
            dtype: string
          - name: query
            dtype: string
          - name: think
            dtype: string
    splits:
      - name: train
        num_bytes: 73781125
        num_examples: 1100
    download_size: 38472887
    dataset_size: 73781125
  - config_name: quac-instruct-7s
    features:
      - name: interactions
        list:
          - name: answer
            dtype: string
          - name: query
            dtype: string
    splits:
      - name: train
        num_bytes: 27169873
        num_examples: 7092
    download_size: 16163197
    dataset_size: 27169873
  - config_name: quac-mix-7s
    features:
      - name: interactions
        list:
          - name: answer
            dtype: string
          - name: query
            dtype: string
          - name: think
            dtype: string
    splits:
      - name: train
        num_bytes: 69188896
        num_examples: 7351
    download_size: 36696327
    dataset_size: 69188896
  - config_name: quac-reasoning-7s
    features:
      - name: interactions
        list:
          - name: answer
            dtype: string
          - name: query
            dtype: string
          - name: think
            dtype: string
    splits:
      - name: train
        num_bytes: 91213041
        num_examples: 5650
    download_size: 43397939
    dataset_size: 91213041
configs:
  - config_name: coqa-instruct-7s
    data_files:
      - split: train
        path: coqa-instruct-7s/train-*
  - config_name: coqa-mix-7s
    data_files:
      - split: train
        path: coqa-mix-7s/train-*
  - config_name: coqa-reasoning-7s
    data_files:
      - split: train
        path: coqa-reasoning-7s/train-*
  - config_name: narrativeqa-instruct-7s
    data_files:
      - split: train
        path: narrativeqa-instruct-7s/train-*
  - config_name: narrativeqa-mix-7s
    data_files:
      - split: train
        path: narrativeqa-mix-7s/train-*
  - config_name: narrativeqa-reasoning-7s
    data_files:
      - split: train
        path: narrativeqa-reasoning-7s/train-*
  - config_name: passkey-multitopic-instruct-7s
    data_files:
      - split: train
        path: passkey-multitopic-instruct-7s/train-*
  - config_name: passkey-multitopic-mix-7s
    data_files:
      - split: train
        path: passkey-multitopic-mix-7s/train-*
  - config_name: passkey-multitopic-reasoning-7s
    data_files:
      - split: train
        path: passkey-multitopic-reasoning-7s/train-*
  - config_name: quac-instruct-7s
    data_files:
      - split: train
        path: quac-instruct-7s/train-*
  - config_name: quac-mix-7s
    data_files:
      - split: train
        path: quac-mix-7s/train-*
  - config_name: quac-reasoning-7s
    data_files:
      - split: train
        path: quac-reasoning-7s/train-*
license: apache-2.0

ReactiveAI / ConverstationalRetrieval Dataset for Supervised Memory-Aware Training (SMAT)

Description in progress