metadata

language:
  - en
dataset_info:
  features:
    - name: text
      dtype: string
    - name: metadata
      struct:
        - name: pile_set_name
          sequence: string
    - name: id
      dtype: int64
  splits:
    - name: train
      num_bytes: 641011558
      num_examples: 403027
  download_size: 397624536
  dataset_size: 641011558
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

Description

This dataset is a sampled subset of the Pile dataset.

The subset sample distribution is:

{
   'Pile-CC': 198245,
   'OpenWebText2': 122382,
   'FreeLaw': 37517,
   'USPTO Backgrounds': 10195,
   'Wikipedia (en)': 8072,
   'PubMed Central': 5849,
   'PubMed Abstracts': 4965,
   'Gutenberg (PG-19)': 2712,
   'BookCorpus2': 2550,
   'Books3': 2432,
   'StackExchange': 1753,
   'PhilPapers': 1560,
   'YoutubeSubtitles': 1187,
   'OpenSubtitles': 1015,
   'ArXiv': 610,
   'NIH ExPorter': 476,
   'Enron Emails': 439,
   'EuroParl': 419,
   'Github': 390,
   'HackerNews': 259
}

The dataset contains ~100M words of text. This can be checked with:

from datasets import load_dataset

ds = load_dataset("PatrickHaller/pile-100M-words")

count = 0
for row in ds["train"]:
    count += len(row["text"].split(" "))

print(count)

# Out: 99999861