Datasetpublic

Varho/gutenberg-fi

Finnish Project Gutenberg Books Dataset Description A collection of 3,505 Finnish-language books from Project Gutenberg, extracted from a January 2026 ZIM archive and converted to Markdown. Statistics Total books 3,505 Unique authors 1,101 Books with translator 1,451 Total text ~0.9 GB Median book length 178k characters Mean book length 258k characters Min book length ~12k characters Max book length ~2.5M characters… See the full description on the dataset page: https://huggingface.co/datasets/Varho/gutenberg-fi.

sourceHugging Facecc0-1.0updated 8mo agoView on Hugging Face

0likes9downloads

Dataset Card

Finnish Project Gutenberg Books

Dataset Description

A collection of 3,505 Finnish-language books from Project Gutenberg, extracted from a January 2026 ZIM archive and converted to Markdown.

Statistics


Total books	3,505
Unique authors	1,101
Books with translator	1,451
Total text	~0.9 GB
Median book length	178k characters
Mean book length	258k characters
Min book length	~12k characters
Max book length	~2.5M characters

Data Format

Each item has the following fields:

Field	Type	Description
`id`	`int`	Project Gutenberg book ID
`title`	`string`	Book title
`author`	`string \	null`	Author name(s), semicolon-separated if multiple
`translator`	`string \	null`	Translator name(s), if the work is a translation
`language`	`string`	ISO 639-1 language code (all entries: `fi`)
`subjects`	`list[string]`	Library of Congress subject headings
`rights`	`string`	Rights statement (all entries: "Public domain in the USA.")
`created`	`string`	Date first posted to Project Gutenberg (ISO date, e.g. `2024-04-24`)
`url`	`string`	Link to the book on gutenberg.org
`text`	`string`	Full Markdown text of the book

Example

json

{
  "id": 73459,
  "title": "\"... Mutta -- naivat tummaverisiä\"",
  "author": "Loos, Anita, 1893-1981",
  "translator": "Hedman, Valfrid, 1872-1939",
  "language": "fi",
  "subjects": ["Humorous stories", "Young women -- Fiction", "Diary fiction"],
  "rights": "Public domain in the USA.",
  "created": "2024-04-24",
  "url": "https://www.gutenberg.org/ebooks/73459",
  "text": "\"… MUTTA — NAIVAT TUMMAVERISIÄ\"\n\nKirj.\n\nAnita Loos\n..."
}

Source and Processing

The dataset was extracted from gutenberg_fi_all_2026-01.zim (the Kiwix offline archive of Finnish Gutenberg books) in the following way:

1.Each HTML book entry is parsed to extract metadata
2.HTML is converted to Markdown using markdownify, preserving headings, emphasis, blockquotes, tables, and hr. Other elements are stripped.
3.Non-book and non-Finnish entries were filtered out.

Intended Use

This dataset is intended for training and evaluating Finnish language models, particularly for:

—Pretraining or fine-tuning text generation models
—Finnish NLP research

License

All texts carry the rights statement "Public domain in the USA." as recorded in each book's dc.rights metadata.

The books should not contain references to Project Gutenberg which would require following: https://www.gutenberg.org/policy/license.html