metadata

dataset_info:
  features:
    - name: title
      dtype: string
    - name: text
      dtype: string
  splits:
    - name: train
      num_bytes: 346623887
      num_examples: 970
  download_size: 212801176
  dataset_size: 346623887

Gutenberg-dutch

This dataset contains all Dutch books made available in the public domain on Project Gutenberg - collected in September 2023. We use a Project publicly available scraper to download the 970 full Dutch books.

While small, this dataset can be particularly helpful for training Dutch language models, as used in the development of ChocoLLaMA (add link).

If you find this dataset useful for your research, kindly cite our paper (add bibtex).