OpenNER

Preliminary data releases

OpenNER is distributed in two forms:

1. Standardized: all datasets containing the originally annotated types, but with names standardized (e.g. in all datasets that annotate persons, the type is PER)
2. Core types: all datasets but only containing location (LOC), organization (ORG), and person (PER) annotations.

The first release of OpenNER is available via:

• GitHub releases: BIO formatted files
• Hugging Face releases: standardized and core types

More releases coming soon!

We are in the process of releasing the code to accompany the releases.