Fix init model

This commit is contained in:
Matthew Honnibal 2018-07-03 22:16:44 +02:00
parent 2543f8c93a
commit 019d09e3c3
1 changed files with 1 additions and 6 deletions

View File

@ -69,12 +69,7 @@ def init_model(lang, output_dir, freqs_loc=None, clusters_loc=None, jsonl_loc=No
vectors_loc = ensure_path(vectors_loc) vectors_loc = ensure_path(vectors_loc)
if vectors_loc and vectors_loc.parts[-1].endswith('.npz'): if vectors_loc and vectors_loc.parts[-1].endswith('.npz'):
vector_data = numpy.load(vectors_loc.open('rb')) vector_data = numpy.load(vectors_loc.open('rb'))
nlp.vocab.vectors = Vectors(data=vector_data) vectors_keys = [lex['orth'] for lex in lex_attrs if 'id' in lex]
vectors_keys = []
for word in nlp.vocab:
if word.rank:
nlp.vocab.vectors.add(word.orth, row=word.rank)
vectors_keys.append(word.orth_)
else: else:
vectors_data, vector_keys = read_vectors(vectors_loc) if vectors_loc else (None, None) vectors_data, vector_keys = read_vectors(vectors_loc) if vectors_loc else (None, None)
nlp = create_model(lang, lex_attrs, vectors_data, vector_keys, prune_vectors) nlp = create_model(lang, lex_attrs, vectors_data, vector_keys, prune_vectors)