Aggregating Italian news articles and summaries translated from the Spanish MLSum corpus using the Helsinki-NLP/opus-mt-es-it model. It features two columns, source and target, representing the original news text from BBC/mundo and its corresponding abstractive summary.
Use Cases
- Train sequence-to-sequence models for abstractive summarization using the source and target columns.
- Benchmark Italian language model performance on news-specific text processing using the source feature.
- Develop text simplification or compression algorithms for Italian media using the target summaries as reference.
Strengths
- Translated from Spanish to Italian using the Helsinki-NLP/opus-mt-es-it model.
- Contains two primary features: source (input article) and target (summary).
- Derived from the BBC/mundo portion of the multilingual MLSum dataset.
- Supports abstractive-summarization and summarization tasks.