Variational-Ladder-Autoencoder icon indicating copy to clipboard operation
Variational-Ladder-Autoencoder copied to clipboard

Published 20 hours ago • verified publisher icon ermongroup

is_training not set in one of the layer of vladder_small

Open l3robot opened this issue 5 years ago • 1 comments

Hi, very interesting paper btw! I'm replicating the code in PyTorch, but noticed that you've not used the is_training parameter for one of the layer below. I was wandering if that was intended, since some papers are keeping the training behaviour of batch normalization in evaluation mode. https://github.com/ermongroup/Variational-Ladder-Autoencoder/blob/00d517dd3c7baa341c83298a58c898b3a8267db7/vladder_small.py#L66 Thanks

l3robot avatar Feb 24 '19 17:02 l3robot

@l3robot Have you had success replicating the code in Pytorch? If so, is there any chance it could be made public?

DanRuta avatar Nov 27 '19 11:11 DanRuta