README.md

metadata

dataset_info:
  features:
    - name: text
      dtype: string
    - name: meta
      struct:
        - name: pile_set_name
          dtype: string
  splits:
    - name: train
      num_bytes: 5315790
      num_examples: 400
  download_size: 2537440
  dataset_size: 5315790
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

First 100 rows of each of timaeus/pile-github, timaeus/pile-wikipedia_en, timaeus/pile-arxiv, timaeus/pile-pile-cc in that order.