You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Awesome work, congratulations! I have some questions regarding the discrepancies between your implementation and paper.
In your CIFAR and MNIST notebooks, what is the purpose of the latent variable z?
In your training loops, you have 2 forward passes through F2 network, and optimize 2 cross-entropy losses. Is this necessary? Am I missing something, or the paper doesn't exactly mention this implementation detail?
Thank you in advance!
Best regards,
Diyor
The text was updated successfully, but these errors were encountered:
Hi Nikita,
Awesome work, congratulations! I have some questions regarding the discrepancies between your implementation and paper.
Thank you in advance!
Best regards,
Diyor
The text was updated successfully, but these errors were encountered: