0

I’m currently trying to implement Deepmask (Link to FAIR's Paper) using Pytorch, so far I have defined the Joint Loss Function, and the model’s learn-able parameters and the forward pass.

I was working on the training phase, and as the paper says that training must be done in an alternative back-propagation fashion across the two branches, I have written the code for the same.

But there is some problem with training, I tried to train the model with a fake data-set (a randomly generated data-set), for minibatches other than the first mini-batch the loss of the model is turning out to be nan.

What could be the reason for this nan loss?

Link to current version of my code

JVJ
  • 9
  • 5
  • I would check your loss function again, perhaps you're doing division by zero somewhere. – Vadim Jun 08 '18 at 13:10
  • I have gone through the loss function, and I can't find any variable being in denominator. Thanks for the suggestion. – JVJ Jun 08 '18 at 13:49

0 Answers0