README.md

metadata

dataset_info:
  - config_name: c4
    features:
      - name: text
        dtype: string
      - name: id
        dtype: string
    splits:
      - name: train
        num_examples: 8596372
  - config_name: msmarco
    features:
      - name: text
        dtype: string
      - name: id
        dtype: string
    splits:
      - name: train
        num_examples: 8596372
configs:
  - config_name: c4
    data_files:
      - split: train
        path: c4/*.jsonl.tar.gz
  - config_name: msmarco
    data_files:
      - split: train
        path: msmarco/*.jsonl.tar.gz

This dataset contains two configurations: c4 and msmarco, both stored in gz format