Fastai Wide ResNet implementation

Hi everyone,

I was going through fastai implementation of Wide ResNet and I’m wondering what the 0.2 multiply at the end of the forward pass of the BasicBlock is doing. As far as I can tell, it’s not mentioned in the original Wide ResNet paper and isn’t present in vanilla ResNets either.

Any idea?

Does this help


Thank you very much! I somewhat missed that answer when I searched the forums :slight_smile: