French Public Domain Newspapers: A Large Historical Text Collection

A large collection aggregating French newspapers and periodicals in the public domain, originally compiled by Pierre-Carl Langlais. The corpus is based on work by Benoît de Courson and Benjamin Azoulay for the Gallicagram project, in cooperation with OpenLLMFrance. The dataset page was last updated on 2024-03-19.

Use Cases

• Historical language modeling based on the aggregated corpus of French periodicals.
• Cultural trend analysis using the n-gram data referenced from the Gallicagram project.
• Training or fine-tuning large language models on public domain French text.
• Studying the evolution of language and discourse in French media over time.

Strengths

• Described as a 'large collection' aiming to aggregate all relevant public domain sources.
• Built upon a corpus curated for the established Gallicagram cultural analytics project.
• Compiled in cooperation with OpenLLMFrance, suggesting a focus on AI/ML applications.

Limitations

• Row count, file formats, and column-level documentation are unknown.
• The description metadata is limited; actual data structure and quality require manual inspection after download.
• License information is unspecified, which may restrict certain uses.

Provenance

Source

Hugging Face user 'PleIAs', based on work by Pierre-Carl Langlais, Benoît de Courson, and Benjamin Azoulay.

Collection Method

Aggregated from a large corpus curated for the Gallicagram project.

Freshness

Last updated 2024-03-19 15:19:31; freshness should be verified.

Geography

France (implied by language and source focus)