nlp_architect.models.bist package

Subpackages

• nlp_architect.models.bist.eval package
  • Subpackages
    • nlp_architect.models.bist.eval.conllu package
      • Submodules
      • nlp_architect.models.bist.eval.conllu.conll17_ud_eval module
      • Module contents
  • Module contents

Submodules

nlp_architect.models.bist.decoder module

nlp_architect.models.bist.decoder.parse_proj(scores, gold=None)

Parse using Eisner’s algorithm.

nlp_architect.models.bist.mstlstm module

class nlp_architect.models.bist.mstlstm.MSTParserLSTM(vocab, w2i, pos, rels, options)

Bases: object

Underlying LSTM model for MSTParser used by BIST parser.

load(filename)

predict(conll_path=None, conll=None)

save(filename)

train(conll_path)

nlp_architect.models.bist.utils module

nlp_architect.models.bist.utils.get_options_dict(activation, lstm_dims, lstm_layers, pos_dims)

Generates dictionary with all parser options.

nlp_architect.models.bist.utils.is_conllu(path)

Determines if the file is in CoNLL-U format.

nlp_architect.models.bist.utils.read_conll(path)

Yields CoNLL sentences read from CoNLL formatted file..

nlp_architect.models.bist.utils.run_eval(gold, test)

Evaluates a set of predictions using the appropriate script.

nlp_architect.models.bist.utils.vocab(conll_path)

nlp_architect.models.bist.utils.write_conll(path, conll_gen)

Writes CoNLL sentences to CoNLL formatted file.

Module contents