Datasets:

bigbio

/

n2c2_2009

like 1

Some entities in the kb schema have no offsets or text.

from train sample 64

 'entities': [{'id': '133875-entity-train-0-0-0-0',
   'type': 'r',
   'text': [],
   'offsets': [],
   'normalized': []},

this is probably b/c they are not filtered properly if they have text "na"