convlab2.nlu.jointBERT.crosswoz package¶
Submodules¶
convlab2.nlu.jointBERT.crosswoz.analyse module¶
-
convlab2.nlu.jointBERT.crosswoz.analyse.calculateF1(predict_golden, goal_type=None, intent=None, domain=None, slot=None)¶
-
convlab2.nlu.jointBERT.crosswoz.analyse.get_goal_type(data, mode)¶
-
convlab2.nlu.jointBERT.crosswoz.analyse.read_zipped_json(filepath, filename)¶
convlab2.nlu.jointBERT.crosswoz.nlu module¶
-
class
convlab2.nlu.jointBERT.crosswoz.nlu.BERTNLU(mode='all', config_file='crosswoz_all_context.json', model_file='https://convlab.blob.core.windows.net/convlab-2/bert_crosswoz_all_context.zip')¶ Bases:
convlab2.nlu.nlu.NLU-
predict(utterance, context=[])¶ Predict the dialog act of a natural language utterance.
- Args:
- utterance (str):
A natural language utterance.
- context (list of str):
Previous utterances.
- Returns:
- action (list of list):
The dialog act of utterance.
-
convlab2.nlu.jointBERT.crosswoz.postprocess module¶
-
convlab2.nlu.jointBERT.crosswoz.postprocess.calculateF1(predict_golden)¶
-
convlab2.nlu.jointBERT.crosswoz.postprocess.intent2das(intent_seq)¶
-
convlab2.nlu.jointBERT.crosswoz.postprocess.is_slot_da(da)¶
-
convlab2.nlu.jointBERT.crosswoz.postprocess.recover_intent(dataloader, intent_logits, tag_logits, tag_mask_tensor, ori_word_seq, new2ori)¶
-
convlab2.nlu.jointBERT.crosswoz.postprocess.tag2das(word_seq, tag_seq)¶
convlab2.nlu.jointBERT.crosswoz.preprocess module¶
-
convlab2.nlu.jointBERT.crosswoz.preprocess.preprocess(mode)¶
-
convlab2.nlu.jointBERT.crosswoz.preprocess.read_zipped_json(filepath, filename)¶