liang-hou / adcgan Goto Github PK

View Code? Open in Web Editor NEW

25.0 25.0 5.0 2.55 MB

[ICML 2022] Conditional GANs with Auxiliary Discriminative Classifier

Home Page: https://arxiv.org/abs/2107.10060

Python 95.12% Shell 4.88%

cgans icml-2022 pytorch

adcgan's Introduction

Hi there 👋

adcgan's People

Contributors

Stargazers

Watchers

Forkers

houliangcs peterouzh zeta1999 orgxin

adcgan's Issues

About aux_loss in adcgan

Thank you for a great job. I'm trying to reproduce your results using TensorFlow, and I'd like to know why the "+1" operation is performed on the label in D_adc_loss_fake and G_adc_loss_neg.
in train_fns.py
Line 63
D_adc_loss_fake = losses.classifier_loss_dis(D_adc_fake, y_[:config['batch_size']] * 2 + 1, config['hinge'])
and
Line 106
G_adc_loss_neg = losses.classifier_loss_gen(D_adc_fake, y_ * 2 + 1, config['hinge'])

Looking forward to your answer, thank you!

Pre-trained Model

Great Work! Are there any pre-trained models available? If so, could you please share some of them, especially on CIFAR and ImageNet?

About PD-GAN

Hi, I just want to make sure how you deal with PD-GAN in your paper, because I can not find anything about PD-GAN in this version of codes. I think it is a good project and it helps me a lot. Thank you so much!

Recommend Projects