metadata

license: cc0-1.0
language:
  - ru
  - uk
  - be
  - en
configs:
  - config_name: default
    data_files:
      - split: train
        path:
          - xaa.jsonl.zst
          - xab.jsonl.zst
          - xac.jsonl.zst
          - xad.jsonl.zst
          - xae.jsonl.zst
          - xaf.jsonl.zst
          - xag.jsonl.zst
          - xah.jsonl.zst
          - xai.jsonl.zst
          - xaj.jsonl.zst
task_categories:
  - text-generation
tags:
  - chemistry
  - biology
  - finance
  - legal
  - music
  - art
  - code
  - medical
  - climate
pretty_name: Вики Чтение
size_categories:
  - 10M<n<100M

Dataset Card for Wikireading

This is a dataset of book chapters scraped from a Russian website called Wikireading.

Dataset Details

Dataset Description

Wikireading is a collection of non-fiction educational books in various domains: Biology, Art, History, Religion and much more. The books are highly educational and provide vast knowledge in different domains, making this dataset a good choice for pretraining.

The resulting dataset contains ~26M rows, which in total contain ~7B tokens (~28B characters) of mostly Russian text, with some books written in other slavic languages.

Curated by: its5Q

Dataset Sources

Scraped from Wikireading

Dataset Structure

Each row in the dataset represents a single chapter of a book. Each row contains the book title (book_title), the author of that book (book_author), the HTML of the book returned by Wikireading, and the book's text extracted using Trafilatura. Additionally, there is a column named litres_preview that indicates whether or not this book is a preview provided by Litres and not a full book. Rows marked as a preview may contain incomplete chapters, but they should still provide some value.

More information

All the works provided on Wikireading, and, inherently, in this dataset, are protected by copyright law, and I do not claim ownership of, or any rights to, the content included in this dataset. The inclusion of text data from these books does not imply any affiliation, endorsement, or sponsorship by the original authors or publishers. This dataset is provided solely for research, educational, and informational purposes. If you are a copyright holder and have any concerns regarding the inclusion of your work in this dataset, please contact me, and appropriate action will be taken to address your concerns.

Dataset Card Contact

In case of any questions or concerns, I can be contacted on Telegram.