LMescheder GAN_stability issues

LMescheder / GAN_stability

Code for paper "Which Training Methods for GANs do actually Converge? (ICML 2018)"

MIT License

917 stars 114 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

Why RMSProp but not Adam?

#22 LynnHo opened 3 years ago
0
2D-data distributions possible to recreate with code?

#21 SimonVerlinden opened 3 years ago
0
Question about Eq. 1 in the article.

#20 yoon28 opened 4 years ago
0
Why use different architectures for ImageNet, celeba and church?

#19 MiaoyunZhao opened 4 years ago
0
It only support Standard GAN loss and WGAN-GP ??

#18 Johnson-yue opened 5 years ago
0
Fix suggested reference

#17 kylematoba closed 5 years ago
1
can you release your pretrain model on celebA-HQ?

#16 qivigor closed 5 years ago
1
Instance noise used in all experiments?

#15 zplizzi closed 5 years ago
4
Questions with batch normalization

#14 LuChengTHU opened 5 years ago
1
question about config

#13 wmpauli opened 5 years ago
1
Parameters for ImageNet GAN

#12 ryanwebster90 closed 5 years ago
2
Why multiply by 0.1 in the residual block?

#11 zplizzi opened 5 years ago
5
Gradient Penalty with Auxiliary classifier

#10 dougsouza opened 5 years ago
3
typo.. toogle->toggle

#9 caffeinism closed 5 years ago
0
discriminator don't need embedding layer.. so

#8 caffeinism closed 5 years ago
0
May I ask where is the implementation of instance noise in this code?

#7 caffeinism closed 5 years ago
0
train own data

#6 niuniu111 opened 5 years ago
0
Weight average

#5 Blade6570 opened 5 years ago
1
Minimal Dirac-GAN

#4 see-- closed 5 years ago
3
Spectral Normalization

#3 oleksandrlazariev opened 5 years ago
3
What's the reason for LeakyReLU activation in Descriminator at the last linear layer?

#2 Arsey closed 6 years ago
1
cifar10 config file

#1 aditya-grover closed 6 years ago
1