README.md

ArcheoBERTje

A Dutch BERT model for the Archaeology domain

This model is based on the Dutch BERTje model by wietsedv (https://github.com/wietsedv/bertje).

We further finetuned BERTje with a corpus of roughly 60k Dutch excavation reports (~650 million tokens) from the DANS data archive (https://easy.dans.knaw.nl/ui/home).