[MXNET-1030] Cosine Embedding Loss Function

• Type: New Feature
• Status: To Do
• Priority: Major
• Resolution: Unresolved
• Component/s: Gluon
• Assignee: Unassigned
• Reporter: Gaurav Gireesh
• Created: 06/Oct/18 00:38
• Updated: 06/Oct/18 00:38

Description

For a target label 1 or -1, vectors target and prediction, the function computes the cosine distance between the vectors. This can be interpreted as how similar/dissimilar two input vectors are.
Can be used as a loss function while training models.