convlab2.dst.sumbt.multiwoz package

Submodules

convlab2.dst.sumbt.multiwoz.convert_to_glue_format module

convlab2.dst.sumbt.multiwoz.convert_to_glue_format.convert_to_glue_format(data_dir, sumbt_dir)

convlab2.dst.sumbt.multiwoz.sumbt module

class convlab2.dst.sumbt.multiwoz.sumbt.SUMBTTracker(data_dir='/home/travis/build/thu-coai/ConvLab-2/data/multiwoz', model_file='https://convlab.blob.core.windows.net/convlab-2/sumbt.tar.gz', eval_slots=['attraction-area', 'attraction-name', 'attraction-type', 'hotel-day', 'hotel-people', 'hotel-stay', 'hotel-area', 'hotel-internet', 'hotel-name', 'hotel-parking', 'hotel-pricerange', 'hotel-stars', 'hotel-type', 'restaurant-day', 'restaurant-people', 'restaurant-time', 'restaurant-area', 'restaurant-food', 'restaurant-name', 'restaurant-pricerange', 'taxi-arriveby', 'taxi-departure', 'taxi-destination', 'taxi-leaveat', 'train-people', 'train-arriveby', 'train-day', 'train-departure', 'train-destination', 'train-leaveat'])

Bases: convlab2.dst.dst.DST

Transferable multi-domain dialogue state tracker, adopted from https://github.com/SKTBrain/SUMBT

construct_query(context)

Construct query from context

detect_requestable_slots(observation)
download_model()
init_session()

Init the class variables for a new session.

load_weights(model_path=None)
predict(query)
test(mode='dev', model_path=None)

Testing funciton of TRADE (to be added)

train(load_model=False, model_path=None)

Model training entry point

update(user_act=None)

Update the dialogue state with the generated tokens from TRADE

convlab2.dst.sumbt.multiwoz.sumbt.get_label_embedding(labels, max_seq_length, tokenizer, device)
convlab2.dst.sumbt.multiwoz.sumbt.test_update()
convlab2.dst.sumbt.multiwoz.sumbt.test_update_bak()

convlab2.dst.sumbt.multiwoz.sumbt_config module

class convlab2.dst.sumbt.multiwoz.sumbt_config.DotMap

Bases: object

convlab2.dst.sumbt.multiwoz.sumbt_utils module

class convlab2.dst.sumbt.multiwoz.sumbt_utils.DataProcessor

Bases: object

Base class for data converters for sequence classification data sets.

get_dev_examples(data_dir)

Gets a collection of `InputExample`s for the dev set.

get_labels()

Gets the list of labels for this data set.

get_train_examples(data_dir)

Gets a collection of `InputExample`s for the train set.

class convlab2.dst.sumbt.multiwoz.sumbt_utils.InputExample(guid, text_a, text_b=None, label=None)

Bases: object

A single training/test example for simple sequence classification.

class convlab2.dst.sumbt.multiwoz.sumbt_utils.InputFeatures(input_ids, input_len, label_id)

Bases: object

A single set of features of data.

class convlab2.dst.sumbt.multiwoz.sumbt_utils.Processor(config)

Bases: convlab2.dst.sumbt.multiwoz.sumbt_utils.DataProcessor

Processor for the belief tracking dataset (GLUE version).

get_dev_examples(data_dir, accumulation=False)

See base class.

get_labels()

See base class.

get_test_examples(data_dir, accumulation=False)

See base class.

get_train_examples(data_dir, accumulation=False)

See base class.

convlab2.dst.sumbt.multiwoz.sumbt_utils.convert_examples_to_features(examples, label_list, max_seq_length, tokenizer, max_turn_length)

Loads a data file into a list of `InputBatch`s.

convlab2.dst.sumbt.multiwoz.sumbt_utils.eval_all_accs(pred_slot, labels, accuracies)
convlab2.dst.sumbt.multiwoz.sumbt_utils.get_label_embedding(labels, max_seq_length, tokenizer, device)
convlab2.dst.sumbt.multiwoz.sumbt_utils.insertSpace(token, text)
convlab2.dst.sumbt.multiwoz.sumbt_utils.normalize_text(text)
convlab2.dst.sumbt.multiwoz.sumbt_utils.warmup_linear(x, warmup=0.002)

Module contents