Closed ngoyal2707 closed 5 years ago
@ngoyal2707,
Thank you for pointing out. Even I had never used the LSGAN with ProGAN as the paper had mentioned some additional stabilization tricks involving noise addition. Will make the required correction soon.
Best regards, @akanimax
@akanimax Sure, I was using LSGAN in ProGAN with noise trick from paper which is giving decent result but sometimes still get random mode collapses where suddenly after some epochs, Generator starts producing random noise.
@ngoyal2707, Interesting observation. I'll try it sometime too. BTW, the error is fixed in 0368a2665b52b843785ffe0b18b3154fd32539b0.
Closing the issue now. Thanks once again.
Cheers :beers:! @akanimax
The current code of LSGAN loss does following:
which is
square of mean
of error, probably should bemean of squared error
as following:If above is correct, same applies to both
generator
anddiscriminator
ofLSGAN
andLSGANSigmoid
.Note: I haven't actually run the above code, just noticed it while looking at the other losses in your code.