Skip to main content

Questions tagged [wasserstein-gan]

For questions related to the Wasserstein GAN, introduced in "Wasserstein Generative Adversarial Networks" (2017, PMLR) by Martin Arjovsky et al.

Filter by
Sorted by
Tagged with
1 vote
0 answers
19 views

How exactly do you backpropagate the gradient penalty in WGAN-GP?

I am trying to implement WGANs from scratch. The loss function for the critic is given by : which i implement in my code as L = average(real output) - average(fake output) + lambda*GP. For ...
hidden_machine's user avatar
0 votes
1 answer
34 views

Generator of my GAN ( WGAN with spectral norm just gets worse and worse after a while and then gets stuck

I am having this issue where the generated images get better after a while, but just keep getting worse and worse. I monitored the gradient coming into the generator for the picture of "3": ...
hidden_machine's user avatar
0 votes
1 answer
39 views

Do WGAN gradients require multi-variable calculus?

The generator tries to maximise this function D(G(z)). That much I understand. But how can the critic maximise D(x) - D(G(z)). ...
zacoons's user avatar
  • 13
4 votes
0 answers
404 views

GAN : Why does a perfect discriminator mean no gradient for the generator?

In the training of a Generative Adversarial Networks (GAN) system, a perfect discriminator (D) is one which outputs 1 ("true image") for all images of the training dataset and 0 ("false ...
Soltius's user avatar
  • 281
0 votes
1 answer
384 views

Is Relativistic GAN better than WGAN-GP?

I am currently reading the ESRGAN paper and I noticed that they have used Relativistic GAN for training discriminator. So, is it because Relativistic GAN leads to better results than WGAN-GP?
SANJAY BHANDARI's user avatar
2 votes
1 answer
1k views

What is being optimized with WGAN loss? Is the generator maximizing or minimizing the critic value?

I am kind of new to the field of GANs and decided to develop a WGAN. All of the information online seems to be kind of contradicting itself. The more I read, the more I become confused, so I'm hoping ...
Gabriel Mongaras's user avatar
2 votes
1 answer
443 views

Why do we use a linear interpolation of fake and real data to penalize the gradient of discriminator in WGAN-GP

I'm trying to better frame/summarize the formulations and motivations behind Wasserstein GAN with gradient penalty, based on my understanding. For the basic GAN we are trying to optimize the following ...
James Arten's user avatar
0 votes
0 answers
307 views

Possible improvements to WGAN-GP output images

I am mapping rather complex data into what essentially amounts to a greyscale image to take better advantage of GANs for generative means. Here is an example of some real data: All real data is of ...
Zintho's user avatar
  • 1
0 votes
1 answer
824 views

How to calculate the gradient penalty proposed in "Improved Training of Wasserstein GANs"?

The research paper titled Improved Training of Wasserstein GANs proposed a gradient penalty in order to avoid undesired behavior due to weight clipping of the discriminator. We now propose an ...
hanugm's user avatar
  • 3,890
0 votes
1 answer
1k views

What to do with a GAN that trained well but got worse over time?

I am training a WGAN-GP network based on the following paper, though I am using a different dataset. Now, for the first ~ 60-70 epochs, my network trained really well, which I could see in the loss ...
Anonymous5638's user avatar
1 vote
0 answers
49 views

How can I estimate the minimum number of training samples needed to get interesting results with WGAN?

Let's say we have a WGAN where the generator and critic have 8 layers and 5 million parameters each. I know that the greater the number of training samples the better, but is there a way to know the ...
FalseSemiColon's user avatar
2 votes
1 answer
98 views

Classifying generated samples with Wasserstein-GAN as real or fake

I'm quite new to GANs and I am trying to use a Wasserstein GAN as an augmentation technique. I found this article https://www.sciencedirect.com/science/article/pii/S2095809918301127, and would like to ...
Ebba's user avatar
  • 21
2 votes
1 answer
338 views

Aren't scores in the Wasserstein GAN probabilities?

I am quite new to GAN and I am reading about WGAN vs DCGAN. Relating to the Wasserstein GAN (WGAN), I read here Instead of using a discriminator to classify or predict the probability of generated ...
Stefano Barone's user avatar
5 votes
1 answer
2k views

Wasserstein GAN: Implemention of Critic Loss Correct?

The WGAN paper concretely proposes Algorithm 1 (cf. page 8). Now, they also state what their loss for the critic and the generator is. When implementing the critic loss (so lines 5 and 6 of Algorithm ...
Anonymous5638's user avatar
1 vote
0 answers
188 views

WGAN-GP Loss formalization

I have to write the formalization of the loss function of my network, built following the WGAN-GP model. The discriminator takes 3 consecutive images as input (such as 3 consecutive frames of a video) ...
Gibser's user avatar
  • 111
4 votes
1 answer
583 views

What is the reason for mode collapse in GAN as opposed to WGAN?

In this article I am reading: $D_{KL}$ gives us inifity when two distributions are disjoint. The value of $D_{JS}$ has sudden jump, not differentiable at $\theta=0$. Only Wasserstein metric provides ...
craft's user avatar
  • 141
1 vote
0 answers
41 views

Under what conditions can one find the optimal critic in WGAN?

The Kantorovich-Rubinstein duality for the optimal transport problem implies that the Wasserstein distance between two distributions $\mu_1$ and $\mu_2$ can be computed as (equation 2 in section 3 in ...
Subho's user avatar
  • 41
1 vote
0 answers
35 views

Wasserstein GAN with non-negative weights in the critic

I want to train a WGAN where the convolution layers in the critic are only allowed to have non-negative weights (for a technical reason). The biases, nonetheless, can take both +/- values. There is no ...
Subho's user avatar
  • 41