allennlp.modules.matrix_attention.cosine_matrix_attention#

CosineMatrixAttention#

CosineMatrixAttention(self)

Computes attention between every entry in matrix_1 with every entry in matrix_2 using cosine similarity.