def main():
words, word_to_vec_map = read_glove_vecs('../data/glove.6B.50d.txt')
print(len(word_to_vec_map))
if DO_TEST: test(word_to_vec_map)
# on the below, we unbiase some words
g = word_to_vec_map['woman'] - word_to_vec_map['man']
if DO_TEST: show_bias(word_to_vec_map)
if DO_TEST: test_unbiase(word_to_vec_map, g)
def test_cosine_similarity_sklearn():
words, word_to_vec_map = read_glove_vecs('data/glove.6B.50d.txt')
ball = word_to_vec_map["ball"]
crocodile = word_to_vec_map["crocodile"]
france = word_to_vec_map["france"]
italy = word_to_vec_map["italy"]
paris = word_to_vec_map["paris"]
rome = word_to_vec_map["rome"]
assert np.isclose(cosine_similarity_sklearn(ball, crocodile),
0.274392462614)
assert np.isclose(cosine_similarity_sklearn(france - paris, rome - italy),
-0.675147930817)
import numpy as np
import w2v_utils
if __name__ == '__main__':
words, word_to_vec_map = w2v_utils.read_glove_vecs('data/glove.6B.50d.txt')
print(word_to_vec_map['hello'])
def setUpClass(cls):
cls.words, cls.word_to_vec_map = read_glove_vecs(
'data/glove.6B.50d.txt')
cls.model_service = ModelService()
import numpy as np
import sklearn
from numpy.linalg import norm
from sklearn.metrics.pairwise import cosine_similarity
from w2v_utils import read_glove_vecs
def cosine_similarity_simple(u, v):
return np.dot(u, v) / (norm(u) * norm(v))
def cosine_similarity_sklearn(u, v):
return cosine_similarity(u.reshape(1, -1), v.reshape(1, -1)).ravel()[0]
if __name__ == '__main__':
words, word_to_vec_map = read_glove_vecs('data/glove.6B.50d.txt')