NER extractor with BERT

(Anubhavpnp) #1

Has anyone tried to modified the ner_crf extractor of RASA NLU to take into account a wordpiece tokenizer that we have for google bert.So insead of a whitespace tokenzer , we have a wordpiece tokeniser. Or does anyone in general have tried to integrate BERT’s NER extraction mechanism with RASA?

(G) #2

I am interested in this as well. I feel like maybe integration with flair might be the most straightforward first step? It can integrate with BERT, allows custom model training, and does NER, though i’ve not yet experimented with using BERT for NER specifically, and then integration with Rasa on top of that.