booksum

Warning: Manual download required. See instructions below.

• Description:

BookSum: A Collection of Datasets for Long-form Narrative Summarization

This implementation currently only supports book and chapter summaries.

GitHub: https://github.com/salesforce/booksum

• Additional Documentation: Explore on Papers With Code

• Homepage: https://github.com/salesforce/booksum

• Source code: tfds.datasets.booksum.Builder

• Versions:

  • 1.0.0 (default): Initial release.

• Download size: Unknown size

• Manual download instructions: This dataset requires you to download the source data manually into download_config.manual_dir (defaults to ~/tensorflow_datasets/downloads/manual/):
  

  1) Go to https://github.com/salesforce/booksum, and run steps 1-3. Place the whole booksum git project in the manual folder. 2) Download the chapterized books from https://storage.cloud.google.com/sfr-books-dataset-chapters-research/all_chapterized_books.zip and unzip to the manual folder.

The manual folder should contain the following directories:

- `booksum/`
- `all_chapterized_books/`

Note: Because the BookSum dataset is based on the availability of web-scraped data and may be incomplete, the _generate_examples method will automatically skip missing entries.

• Auto-cached (documentation): Yes (test, validation), Only when shuffle_files=False (train)

• Feature structure:

FeaturesDict({
    'document': Text(shape=(), dtype=string),
    'summary': Text(shape=(), dtype=string),
})

• Feature documentation:

Feature	Class	Shape	Dtype	Description
	FeaturesDict
document	Text		string
summary	Text		string

• Supervised keys (See as_supervised doc): ('document', 'summary')

• Figure (tfds.show_examples): Not supported.

• Citation:

@article{kryscinski2021booksum,
      title={BookSum: A Collection of Datasets for Long-form Narrative Summarization},
      author={Wojciech Kry{\'s}ci{\'n}ski and Nazneen Rajani and Divyansh Agarwal and Caiming Xiong and Dragomir Radev},
      year={2021},
      eprint={2105.08209},
      archivePrefix={arXiv},
      primaryClass={cs.CL}
}

booksum/book (default config)

• Config description: Book-level summarization

• Dataset size: 208.81 MiB

• Splits:

Split	Examples
'test'	46
'train'	312
'validation'	45

• Examples (tfds.as_dataframe):

booksum/chapter

• Config description: chapter-level summarization

• Dataset size: 216.71 MiB

• Splits:

Split	Examples
'test'	1,083
'train'	6,524
'validation'	891

• Examples (tfds.as_dataframe):