A corpus of 11,551,703 pages of historical German newspapers in the public domain, harvested from the Deutsche Digitale Bibliothek (DDB) and its Zeitungsportal.
It covers 1,607,744 issues from 796 newspapers published between 1638 and 1964, totalling 25.8 billion whitespace tokens (157.9 billion characters) of OCR fulltext. Every page carries an explicit per-page license (Public Domain Mark or CC0) and links back to its full-resolution scan (IIIF) and its… See the full description on the dataset page:
https://huggingface.co/datasets/histde/ddb-newspaper-corpus.
