Using Probabilistic U-net on Matlab

Hello,

I am trying to implement a "Probablistic U-net" by Simon(2018) to cope with label ambiguity in semantic segmentation.

And I found some kind open source programs written in Torch:

https://github.com/stefanknegt/Probabilistic-Unet-Pytorch

As I barely know anything about torch(and python), I have hard time applying the aforementioned program in matlab(specifically, the type of dataset used for the training in Torch is quite different from the matlab)

I would appreciate any suggestion or tips to implement the above code in matlab.

Thank you very much in advance!!!!!