---
license: cc0-1.0
task_categories:
- text-generation
language:
- de
tags:
- newspapers
- ocr
- public domain
pretty_name: Public Domain Newspapers (German)
size_categories:
- 10B<n<100B
---
# Dataset Card for Public Domain Newspapers (German)

<!-- Provide a quick summary of the dataset. -->

This dataset contains 13 billion words of OCR text extracted from German historical newspapers. 


## Dataset Details

### Dataset Description

<!-- Provide a longer summary of what this dataset is. -->


- **Curated by:** [Sebastian Majstorovic](https://www.storytracer.org)
- **Language(s) (NLP):** German
- **License:** Dataset: CC0, Texts: Public Domain

### Dataset Sources [optional]

<!-- Provide the basic links for the dataset. -->

- **Repository:** https://www.deutsche-digitale-bibliothek.de/newspaper

### Copyright & License

The newspapers texts have been determined to be in the Public Domain by the institutions who provided them to the newspaper portal of the German Digital National Library. The dataset itself, excluding the texts, is licensed under the [CC0 license](https://creativecommons.org/public-domain/cc0/).