mohdgadi
(Mohdgadi)
August 20, 2018, 5:38am
1
Currently my pipeline looks like this
> pipeline:
- name: "tokenizer_whitespace"
- name: "ner_duckling_http"
- name: "ner_synonyms"
- name: "intent_featurizer_count_vectors"
- name: "intent_classifier_tensorflow_embedding"
I have trained my model with a total of 12 intents and 240 data points. Everything seems to be working fine, but I want to detect number entities eg ‘ten’,‘10’ or ‘24 years’. How do I extract these entities using duckling, Do i need to train with them. When i input them to interpret, I receive an empty entity value.Preformatted text
souvikg10
(Souvik Ghosh)
August 20, 2018, 9:36am
2
When you use duckling http, it means you have somewhere a duckling server running. There is already a topic covered here similar to your question
The duckling you installed is part of the ner_duckling_http pipeline and not ner_duckling, ner_duckling is simply the python wrapper of the Clojure version of duckling. Facebook has deprecated that. You pipeline should look like this
pipeline:
- name: "tokenizer_whitespace"
- name: "intent_entity_featurizer_regex"
- name: "ner_crf"
- name: "ner_synonyms"
- name: "intent_featurizer_count_vectors"
- name: "intent_classifier_tensorflow_embedding"
intent_tokenization_flag: true
intent_split_sy…
libindavis
(Libindavis)
December 12, 2018, 1:40am
3
Hi Souvik,
just want to confirm, do I need to install the duckling separately if I want to use the component ‘ner_duckling_http’ in pipeline?