Question about Conv2d + ConvTranspose2d

[absentio]

Dear @tlatkowski, I was looking at your code about CIFAR-10 conditional DCGAN and It is the best I've found, even though all your code is written using TensorFlow and I only have experience with PyTorch. Anyway, I found quite strange to see the usage of both convolution and transposed convolution layers in the Generator model. Can you explain me what led you to this choice?
Thanks again for your great work.

[tlatkowski]

Hi @absentio ,

thanks for your interest in my repo, could you provide the code reference?

Regards,
Tomasz