News articles across five main categories—France, world, politics, automotive, and society—scraped from the Orange Actu website. It covers nearly a decade of French-language content from February 2011 to September 2020, providing professionally written titles and abstracts for each entry.
Use Cases
- Train French-language abstractive summarization models using the abstract field as the target.
- Develop headline generation systems by mapping article content to the single-sentence title.
- Perform hierarchical text classification using the main category and society subcategory labels.
Strengths
- Spans a 9-year period of news coverage from February 2011 to September 2020.
- Includes five primary categories: France, world, politics, automotive, and society.
- Contains 8 specific subcategories for society news, including health, environment, and high-tech.
- Provides two distinct summarization targets: a single-sentence title and a brief abstract.