README.md

metadata

dataset_info:
  features:
    - name: messages
      list:
        - name: role
          dtype: string
        - name: content
          dtype: string
    - name: task
      dtype: string
    - name: source
      dtype: string
  splits:
    - name: train
      num_bytes: 26670620
      num_examples: 26238
    - name: validation
      num_bytes: 2558296
      num_examples: 2459
  download_size: 9435021
  dataset_size: 29228916
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*