kumar-shridhar / PyTorch-BayesianCNN

Bayesian Convolutional Neural Network with Variational Inference based on Bayes by Backprop in PyTorch.
MIT License
1.42k stars 323 forks source link

testset is used in training function #2

Closed Lobotomy92 closed 5 years ago

Lobotomy92 commented 5 years ago

In line 173 the m is computed as m = math.ceil(len(testset) / cf.batch_size) However, it is in the training function, should it be len(trainset)? Because usually the test set should not be seen in the training phase.

kumar-shridhar commented 5 years ago

Hi @Lobotomy92 ,

It was a typo in the latest commit. Thanks for raising the issue. Noted and changed.