Closed Ghostvv closed 5 years ago
@Ghostvv can you please provide steps to reproduce & the exact output?
For demo-rasa.md
dataset, with above config for ner_crf
, I get:
{'intent': {'name': 'restaurant_search', 'confidence': 0.9175125956535339}, 'entities': [{'start': 8, 'end': 15, 'value': 'CHINESE', 'entity': 'cuisine', 'confidence': 0.6072946511019204, 'extractor': 'ner_crf'}], 'intent_ranking': [{'name': 'restaurant_search', 'confidence': 0.9175125956535339}, {'name': 'goodbye', 'confidence': 0.06257633119821548}, {'name': 'address', 'confidence': 0.0}, {'name': 'greet', 'confidence': 0.0}, {'name': 'affirm', 'confidence': 0.0}], 'text': 'show me CHINESE restaurants'}
{'intent': {'name': 'restaurant_search', 'confidence': 0.9175125956535339}, 'entities': [{'start': 8, 'end': 15, 'value': 'chinese', 'entity': 'cuisine', 'confidence': 0.8070571545315702, 'extractor': 'ner_crf'}], 'intent_ranking': [{'name': 'restaurant_search', 'confidence': 0.9175125956535339}, {'name': 'goodbye', 'confidence': 0.06257633119821548}, {'name': 'address', 'confidence': 0.0}, {'name': 'greet', 'confidence': 0.0}, {'name': 'affirm', 'confidence': 0.0}], 'text': 'show me chinese restaurants'}
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.
@amn41 did you try to reproduce it?
no - please add to backlog
Took a look at this, suffix and prefix are parts of the tokens, so if tokenization is case sensitive, it depends on capitalization.
The following config for
ner_crf
:picks different entities in
B-OY 2018
andb-oy 2018