kwotsin

Hi @abdulfatir! My apologies for the late response! As I have recently moved it is a little challenging to access these models at the moment, but are there any specific...

Hi @houliangict, yes you are right there are differences in how the field computes FID these days. Recall that since the goal for GANs in this case is to model...

Hi @WingsleyLui , thanks for your kind comments! I also realised this from a similar issue #20 , and so added support for using custom datasets. In general, for using...

@WingsleyLui Thanks for your suggestions, I will take a look at the SN visualisation as well :+1:

Hi @WingsleyLui thank you for the comments! Indeed this is something I'm looking at, especially since there are some models (e.g. SAGAN) that seems to only work if I use...

Hi @KeisukeShimokawa, yes indeed I mostly compared to the version without sBN to keep the design simple as sBN is a general architectural improvement, but indeed it will be an...

@KeisukeShimokawa This is an interesting suggestion: indeed, I'm also not sure what the performance difference between the two is. I did a deeper check and it seems like the TensorFlow...

Hi @greeneggsandyaml, thanks for the suggestion! I will certainly look into StyleGAN v2 and other interesting models as well! :D

@Zehaos No worries about it. I think generally the YellowFin performance seems to be much better, but I still haven't got the opportunity to test it out past 6k steps...

This greatly depends on the hardware you used and seems to be a duplicated problem of #19. Have you checked if the bottleneck comes from the data feeding part rather...