Annotated-Semantic-Relationships-Datasets

Why we included this project

If you're training a supervised relation extractor, this collection saves you the usual scavenger hunt across papers and project pages. It groups the standard benchmark corpora by how they were annotated: closed-set information extraction, open information extraction, and distantly supervised data. Each dataset entry gives the number of classes, language, year, and the paper to cite, so you can quickly check fit and give proper credit. English and Portuguese are both covered, which matters if you work with Portuguese text and rarely find ready-made annotated resources. It's a reference index, not a tool you run, but for scoping a project it's exactly the kind of starting point you want.