We attempted to enable Spark NLP pre-trained NER model (based out of BERT) at indexing pipeline stage and we are running into errors. How does this functionality work?
Fusion Server 4.2.2, Fusion AI, Index Pipeline Stages
We imported pre-trained NER and POS models from John Snow available at this link:
Only the pre-trained NER model is supported. If choosing an NER model, download the NerDLModel instead of the NerCRFModel or NERDL Model with BERT. More info is found on our documentation site here:
NerDLModel (NER with BERT) ner_dl_bert was downloaded which is not currently supported with 4.2.2