A German BERT model for the Archaeology domain

This model is based on the bert-base-german-cased model (https://huggingface.co/bert-base-german-cased).

We further finetuned bert-base-german-cased with a corpus of roughly 21k documents (~107 million tokens) from the Heidelberg university library (https://digi.ub.uni-heidelberg.de/en/sammlungen/archaeologie.html).