Using other language models with Stanford Core NLP

Question

I want to use a Dutch model for named entity recognition with Core NLP.

I have found a pre-trained model from OpenNLP, but it doesn't seem to be interoperable with CoreNLP.

Why is that? Can we still use Core NLP with other languages than English, Chinese and Spanish?

Answer 1

CoreNLP currently does not support Dutch. There are some components which work for German and Arabic, but the pipeline is currently only for English, Chinese and Spanish. You can retrain our NER model on the same conllx data, but we have not done this.

The OpenNLP models are not compatible with CoreNLP.