license: cc-by-nc-sa-4.0
language:
- en
size_categories:
- 1K<n<10K
configs:
- config_name: default
data_files:
- split: train
path: data/train/*
Gutenberg 8K
This is a subset of the files available at 15000 Gutenberg Books
About Dataset
The Gutenberg dataset represents a corpus of over 15,000 book texts, their authors and titles. The data has been scraped from the Project Gutenberg website using a custom script to parse all bookshelves. The text download links for the books have been stored in the gutenberg_metadata.csv file, alongside their respective book's title, author and bookshelf (category). The download links have been scraped with respect to the category the Gutenberg website associated them with. Some books from the website are not part of the dataset, as they have not yet been categorized.
The text data itself can be downloaded using the gutenberg_download.py script, which will parse the metadata file, download the text data for each book and save the results as a csv file. The final csv file containing the book texts, the authors, the titles and the categories will have a size of around 5 GB.
Please note that some books are audiobooks. We have decided to keep these in the dataset in case someone wishes to work with the audio data instead of text.
Acknowledgement
All credits to the original 15000 Gutenberg Books project.