Question on how to reconstruct images about glow-pytorch HOT 6 CLOSED

First you need to gather zs from blocks. Then you can reverse using these zs, but you will need to skip sampling (reparameterization) using prior. You can may refer to this: 4beed5d.

from glow-pytorch.

Hi I just tried it out and it works perfectly now! Thank you very much your help!

from glow-pytorch.

for reconstruction do we need to normalize zs (obtained during the forward pass)?

Currently, I'm training glow on mnist with affine coupling, during the forward pass zs blows up to 100k sometimes, due to which reconstructions are Nan always. Let me if normalization is required or if I'm missing any other step during reconstruction.

Thank you!

from glow-pytorch.

@koriavinash1 No, you don't need a normalization. It should be enough to give zs from forward to reverse with reconstruct=True.

from glow-pytorch.

@rosinality thanks for the quick response, do you have any idea on how to resolve the NaN issue?

from glow-pytorch.

@koriavinash1 You can try to reduce --n_flow and --n_block to make the model more stable.

from glow-pytorch.