All problems
Speaker Cosine Similarity
Problem
Return the cosine similarity between the mean embeddings (centroids) of two speaker embedding sets.
Return 0.0 when either set is empty or a centroid has zero norm.
Hint
Average each embedding set first, then take the cosine of the two centroids.
Test cases
case 1
in: [[[1,0],[1,0]],[[1,0]]]
expected: 1
case 2
in: [[[1,0]],[[0,1]]]
expected: 0
case 3
in: [[[1,2],[3,4]],[[2,3]]]
expected: 1.0000000000000002
case 4
in: [[],[[1]]]
expected: 0
case 5
in: [[[1,0]],[[-1,0]]]
expected: -1
Starter codePython
def speaker_cosine(embeddings_a, embeddings_b):
# Your code here
pass5 test cases · run your code against them in the editor