convlab2.nlg.sclstm.crosswoz package¶
Submodules¶
convlab2.nlg.sclstm.crosswoz.evaluate module¶
Evaluate NLG models on utterances of Multiwoz_zh test dataset Metric: dataset level BLEU-4, slot error rate Usage: python evaluate.py [usr|sys|all]
-
convlab2.nlg.sclstm.crosswoz.evaluate.get_bleu4(dialog_acts, golden_utts, gen_utts, data_key)¶
-
convlab2.nlg.sclstm.crosswoz.evaluate.get_err_slot(dialog_acts, gen_slots)¶
-
convlab2.nlg.sclstm.crosswoz.evaluate.split_delex_sentence(sen)¶
-
convlab2.nlg.sclstm.crosswoz.evaluate.value_replace(sentences, dialog_act)¶
convlab2.nlg.sclstm.crosswoz.generate_resources module¶
-
convlab2.nlg.sclstm.crosswoz.generate_resources.main()¶
-
convlab2.nlg.sclstm.crosswoz.generate_resources.read_json(filename)¶
convlab2.nlg.sclstm.crosswoz.sc_lstm module¶
-
class
convlab2.nlg.sclstm.crosswoz.sc_lstm.SCLSTM(archive_file='/home/travis/build/thu-coai/ConvLab-2/convlab2/nlg/sclstm/crosswoz/models/nlg_sclstm_crosswoz.zip', use_cuda=False, is_user=False, model_file='https://convlab.blob.core.windows.net/convlab-2/nlg_sclstm_crosswoz.zip')¶ Bases:
convlab2.nlg.nlg.NLG-
generate(meta)¶ Generate a natural language utterance conditioned on the dialog act.
- Args:
- action (list of list):
The dialog action produced by dialog policy module, which is in dialog act format.
- Returns:
- utterance (str):
A natural langauge utterance.
-
generate_delex(meta)¶ - meta = [
- [
“General”, “greet”, “none”, “none”
], [
“Request”, “景点”, “名称”, “”
], [
“Inform”, “景点”, “门票”, “免费”
]
]
-
generate_slots(meta)¶
-
-
convlab2.nlg.sclstm.crosswoz.sc_lstm.parse(is_user)¶
convlab2.nlg.sclstm.crosswoz.train module¶
-
convlab2.nlg.sclstm.crosswoz.train.evaluate(config, dataset, model, data_type, beam_search, beam_size, batch_size)¶
-
convlab2.nlg.sclstm.crosswoz.train.get_slot_error(dataset, gens, refs, sv_indexes)¶ - Args:
gens: (batch_size, beam_size) refs: (batch_size,) sv: (batch_size,)
- Returns:
count: accumulative slot error of a batch countPerGen: slot error for each sample
-
convlab2.nlg.sclstm.crosswoz.train.interact(config, args)¶
-
convlab2.nlg.sclstm.crosswoz.train.parse()¶
-
convlab2.nlg.sclstm.crosswoz.train.read(config, args, mode)¶
-
convlab2.nlg.sclstm.crosswoz.train.score(feat, gen, template)¶ feat = [‘d-a-s-v:Booking-Book-Day-1’, ‘d-a-s-v:Booking-Book-Name-1’, ‘d-a-s-v:Booking-Book-Name-2’] gen = ‘xxx slot-booking-book-name xxx slot-booking-book-time’
-
convlab2.nlg.sclstm.crosswoz.train.str2bool(v)¶
-
convlab2.nlg.sclstm.crosswoz.train.test(config, args)¶
-
convlab2.nlg.sclstm.crosswoz.train.train(config, args)¶
-
convlab2.nlg.sclstm.crosswoz.train.train_epoch(config, dataset, model)¶