This isn't a part of speech (POS) tagging model. From the examples it is quite clear that it has been trained on the named entity recognition (NER) tags instead. But it looks like it might be annotating with the POS tags instead of the NER tags.
· Sign up or log in to comment
