Tom Jubb

I'll try and answer this one after spending some time thinking through the code. First, the masking must be done throughout every layer (a regular convolution would pick up "future"...

I also believe based on some checks that the "blind spot" issue originally raised in the gated pixelCNN paper is actually present in this model (despite the use of a...

I've also run into this problem and tried almost everything suggested. I have Ubuntu 18, which doesn't support CUDA version 8, which is needed for tensorflow 1.2.1. So swapping to...

> I've also run into this problem and tried almost everything suggested. > > I have Ubuntu 18, which doesn't support CUDA version 8, which is needed for tensorflow 1.2.1....