Flux.jl
Flux.jl copied to clipboard
Follow Up to the Addition of Siamese Loss function PR
Currently the loss function added takes in the the inputs as the euclidean distance vectors to calculate the siamese contrastive loss function. While adding this functionality in the previous PR, I also thought it would be easier for the user ( as an added option ) to precompute the euclidean distance along with the siamese contrastive loss function ( by just taking the flatten input vectors as parameters ). Was also discussed in the review to add this functionality. ( loss function to be added along with the euclidean distance, tests case to be written along with it, simple docstring ( documentation to be added for it ) just like the previous added function. If everyone is fine with the implementation of this function, I can start working on the PR right away.