Implementing a network with triplet loss in Neural Network toolbox

I needed some guidance on how to implement triplet loss like in this paper(FaceNet: A Unified Embedding for Face Recognition and Clustering) in Neural Network Toolbox?

[Greg Heath]

Thanks for the vague reference. However, I would like to see your interpretation of triplet loss.

Most importantly, how does it differ from the two explanations given in the NEWSGROUP and ANSWERS when I search using the search words

 triplet loss

Thanks,

Greg

Answer 1

[Roy Talman]

Hi I implementing this, its only preliminary and need some improvements, but it works.

https://www.mathworks.com/matlabcentral/fileexchange/65927-cnn-triplet-loss