Skip to content

Dataset · PleIAs

US Public Domain Newspapers Archive

PleIAs/US-PD-Newspapers

A massive corpus of digitized historical US newspapers in English for training language models.

What's inside

A collection of US newspaper issues digitized by the US Library of Congress from the Chronicling America project. The corpus contains between 10 and 100 million records (nearly 21 million unique materials and about 100 billion words) in English, completely free of copyright restrictions.

What you can do with it

  • Pre-train large language models (LLMs) on legally clean data
  • Research in historical linguistics and press analysis
Tasks
Text generation
Languages
English
Size
10M–100M rows
Downloads this month
33.5K
Likes
50

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("PleIAs/US-PD-Newspapers")