Version 0.4

Version 0.4 of the AjMC NE corpus. Compared to the previous version (v. 0.3, used in the shared task HIPE-2022) it has improved data quality and includes an additional annotation layer of bibliographic references.

Version 0.3

Version 0.3 of the AjMC NE corpus, which was released as part of the HIPE-2022-data version 2.1 – N.B. some of the masked/unmasked test files are contained only in this private repo, and will be pushed to HIPE-2022-data at due time.

Version 0.2

Version 0.2 of the AjMC NE corpus, which was released as part of the HIPE-2022-data version 2.0.

New in this release:

• full train + dev sets for fr, en, de documents
• added mappings [OCR-gold transcript] for noisy entities

Version 0.1

This release contains the sample data (mini-reference corpus) which was included in the HIPE-2022-data release v. 1.0.