Compute graph wrong and one question about bicyclegan HOT 4 OPEN

I haven't seen the errors in 1 before and I am not sure about what happened.
For your question 2, note that

self.fake_B_random = self.netG(self.real_A_encoded, self.z_random)

fake_B_random is also conditional on real_A_encoded. The confusion might be caused by the naming. See #31 for more details.

from bicyclegan.

I was able to reproduce your error 1 now. It didn't happen for the previous PyTorch version. I fixed it with the latest commit.

from bicyclegan.

Hi @junyanz.
I believe that your latest commit introduced an error.

Now this comment is no longer valid.
Because self.backward_G_alone() also computes gradients for the encoder.
You must keep the encoder fixed when minimizing latent regression loss (according to your paper).

from bicyclegan.

Yes. You are correct. I updated the code with a new commit.
I set self.set_requires_grad([self.netE], False) before self.backward_G_alone(). I assume this will disabling computing graidents for E while we backprop the gradients from loss_z_L1.
I am also not sure why the previous code doesn't work in the recent PyTorch version ( I am using 1.5.0)
@ssnl

from bicyclegan.