Datasets:

Languages:
English
Multilinguality:
monolingual
Size Categories:
100K<n<1M
Language Creators:
found
Source Datasets:
extended|other
ArXiv:
License:
dfki-nlp commited on
Commit
abc0652
1 Parent(s): d13bb8a

fixed class label feature set for re-tacred

Browse files
Files changed (1) hide show
  1. tacred.py +1 -1
tacred.py CHANGED
@@ -285,7 +285,7 @@ class Tacred(datasets.GeneratorBasedBuilder):
285
  "stanford_ner": datasets.Sequence(datasets.Value("string")),
286
  "stanford_deprel": datasets.Sequence(datasets.Value("string")),
287
  "stanford_head": datasets.Sequence(datasets.Value("int32")),
288
- "relation": datasets.ClassLabel(names=_CLASS_LABELS),
289
  }
290
  )
291
 
 
285
  "stanford_ner": datasets.Sequence(datasets.Value("string")),
286
  "stanford_deprel": datasets.Sequence(datasets.Value("string")),
287
  "stanford_head": datasets.Sequence(datasets.Value("int32")),
288
+ "relation": datasets.ClassLabel(names=_RETACRED_CLASS_LABELS if self.config.name == 're-tacred' else _CLASS_LABELS),
289
  }
290
  )
291