The method to measure the similarity of two vectors: cosine similarity, pytorch calculate cosine similarity: torch nn. CosineSimilarity(dim=1, eps=1e-08)

$more\; thanstringphaselikedegree$
stay NLP In my mission , Will calculate the similarity of the generated two word vectors , Cosine similarity formula is often used to calculate .

Cosine similarity uses the cosine value of the angle between two vectors in vector space to measure the difference between two individuals . The closer the cosine is to 1, It means that the closer the angle is 0 degree , That is, the more similar the two vectors are , This is called " Cosine similarity ".

We know , For two vectors , If the angle between them is smaller , So we think the more similar these two vectors are . Cosine similarity is the use of this theoretical idea . It measures the similarity between vectors by calculating the cosine of the angle between two vectors .

The formula is as follows ：

Reference resources ： A way to measure the similarity between two vectors ： Cosine similarity

be used for torch Cosine similarity function ：
pytorch Cosine similarity ：torch.nn.CosineSimilarity(dim=1, eps=1e-08)