While working with cosine similarity, I am facing issue calculating the inner product of two vectors.
Code:
from gensim.similarities import (
WordEmbeddingSimilarityIndex,
SparseTermSimilarityMatrix
)
w2v_model = api.load("glove-wiki-gigaword-50")
similarity_index = WordEmbeddingSimilarityIndex(w2v_model)
similarity_matrix = SparseTermSimilarityMatrix(similarity_index, dictionary)
score = similarity_matrix.inner_product(
X = [
(0, 1), (1, 1), (2, 1), (3, 2), (4, 1),
(5, 1), (6, 1), (7, 1), (8, 1), (9, 1),
(10, 1), (11, 1), (12, 1), (13, 1), (14, 1),
(15, 1), (16, 3)
],
Y = [(221, 1), (648, 1), (8238, 1)],
normalized = True
)
Error:
TypeError Traceback (most recent call last)
Input In [77], in <cell line: 1>()
----> 1 similarity_matrix.inner_product(
2 [(0, 1), (1, 1), (2, 1), (3, 2), (4, 1), (5, 1), (6, 1), (7, 1),
3 (8, 1), (9, 1), (10, 1), (11, 1), (12, 1), (13, 1), (14, 1), (15, 1), (16, 3)],
4 [(221, 1), (648, 1), (8238, 1)], normalized=True)
File ~\Anaconda3\lib\site-packages\gensim\similarities\termsim.py:558, in SparseTermSimilarityMatrix.inner_product(self, X, Y, normalized)
555 if not X or not Y:
556 return self.matrix.dtype.type(0.0)
--> 558 normalized_X, normalized_Y = normalized
559 valid_normalized_values = (True, False, 'maintain')
561 if normalized_X not in valid_normalized_values:
TypeError: cannot unpack non-iterable bool object
I am not sure why it says bool objects when both X and Y are list.
The normalized parameter should be a 2-tuple which declares for both X and Y separately (as in the docs).
Therefore, the call should look like this:
score = similarity_matrix.inner_product(
X = [
(0, 1), (1, 1), (2, 1), (3, 2), (4, 1),
(5, 1), (6, 1), (7, 1), (8, 1), (9, 1),
(10, 1), (11, 1), (12, 1), (13, 1), (14, 1),
(15, 1), (16, 3)
],
Y = [(221, 1), (648, 1), (8238, 1)],
normalized = (True, True)
)