LatinCy NER

Curated NER training data for LatinCy Latin language models.

Status

This dataset is currently in preparation and will be available shortly.

Overview

• Entity types: PERSON, LOC, NORP
• Sources: Universal Dependencies, biblical texts, Latin primers, Tesserae, and other annotated corpora
• Format: spaCy-compatible JSON singles with character-offset span annotations
• Splits: Train, dev, and held-out document-level test set

License

CC BY 4.0