About G/D loss function

[zicair]

Thanks for you code, In Paper, the loss function of 'Generator' formulated as

And the loss function of 'Discriminator' formulated as

they both use BCE loss; I found out that you didn't do it that way,you used log(output) in 'Discriminator' and -log(output) in 'Generator' and I was confused about that! Can you explain it for me.Thanks a lot.

[lyq893]

@zicair
hello!
Glad to find you following this code.
I want to know how does this train the model? How does it work in the first place? Can you answer my question?
Looking forward to your answer! thanks.