Questions tagged [gan]
Generative Adversarial Networks (GANs) are neural networks that are trained in an adversarial manner to generate data mimicking some distribution.
226
questions
0
votes
0
answers
12
views
cGAN: Discriminator loss going to zero while Generator's going always up but the result is very good
I have a Conditional Generative Adversarial Network for Quantum State Tomography. The metrics I am monitoring during the training process are the losses and the Fidelity (the degree of similarity ...
- 15
0
votes
0
answers
8
views
GAN training: Why not wait for the G or D to catch up?
It is known, that GANs are hard to train.
... and it proved to be very easy that one of the two overpowers the other early in the training process
(Quote from here)
I was wondering, why this isn't ...
- 156
0
votes
0
answers
31
views
How to Change Architecture of DCGans?
https://pytorch.org/tutorials/beginner/dcgan_faces_tutorial.html
I was refering this notebook but default size is 64*64
I want to change architecture to 256 or 512
Can anyone help me with training
...
0
votes
0
answers
172
views
How to calculate FID for a set with a small number of images?
I need to evaluate my generative model using FID (Fréchet inception distance). However, the dataset of real images that I have only contains 2719 examples. I've read that the authors of the metric ...
- 101
0
votes
0
answers
8
views
Trying to understand CGAN loss curves
I am trying to understand the loss curves of a CGAN model better. From what I've seen so far, it seems like there is no consistent answer as to what the loss curves should look like. Some tutorials ...
0
votes
0
answers
74
views
Discriminator and Generator's losses in a GAN
I'm a little confused about the Generator's and Discriminator's losses while training a cGAN. I am aware that a stable GAN is one where the Discriminator's loss reaches 0.5. I know that we can read ...
- 15
0
votes
0
answers
15
views
How to improve the expressibility of a cGAN
I am training a cGAN on the problem of reconstructing a density matrix. My inputs to the network are matrices and expectation values. That is, I have a set $(A,x)$ where A are measurement operators, ...
- 15
0
votes
0
answers
4
views
image gap filling using GAN and ancillary datasets
I have several daily single band images (with reflectivity values). there are several gaps in these images. To fill the gaps, I collected several ancillary datasets (including surface temperate ...
- 1
0
votes
1
answer
18
views
distributions in "A Note on the Inception Score" [closed]
I'm reading the following paper A Note on the Inception Score. IS is a evaluation metric for GAN. While reading the paper I got very confused about the notation of so called distributions.
I will ...
- 103
0
votes
0
answers
13
views
Optimal discriminator equation for LSGAN
In the article, LSGan's optimal discriminator is as follow.
Where can I find the detailed explanation of how optimal discriminator equation is obtained?
- 483
1
vote
0
answers
106
views
Keras RMSProp what is the alternative to "decay" (no longer available after Keras 2.3)
Background: Hello, I'm creating a GAN with an RMSProp optimizer for both discriminator & generator. The generator model has half the learning rate of the discriminator (1e-4) and half the decay of ...
- 11
2
votes
2
answers
45
views
Implementing loss function for training GAN
I am trying to understand the implementation of the following loss function in training Discriminator part.
Loss function maximize log(D(x)) + log(1 - D(G(z)))
...
- 483
4
votes
2
answers
153
views
Using Generative Adversarial Networks for joint distribution estimation
I am trying to use GAN model to generate N-dimensional samples with joint probability distribution that looks like some training data. I am having trouble getting the probability distribution of the ...
- 55
0
votes
0
answers
19
views
How to make the generated machine learning model satisfied with specified proportion
If an image is composed of three categorical variables with fixed proportions (such as 0.2,0.3, and 0.5), how could I ensure the generated image (such as GAN model) is satisfied with this proportion?
- 61
1
vote
0
answers
97
views
Numerical data imputation: Generative Adversarial Imputation Nets (GAIN) not reproducible?
I am interested in numerical data imputation problems: how to properly estimate missing values in a tabular data set (rows and columns) with missing numerical values?
In 2018, Yoon et al. proposed the ...
- 163
3
votes
1
answer
756
views
GAN : Why does a perfect discriminator mean no gradient for the generator?
(Note : this is a cross-post from Artificial Intelligence. As I got no answer there in two weeks, I'm trying my luck on a more popuplated SE site. I know this is against SE's policy on cross-posting ...
- 1,246
0
votes
0
answers
11
views
SEQGAN model strange behaviour
I have been trying SEQGAN to increase the size of my small dataset with around 5000 observations. I used the the defaults settings so far and the model looked like underfit because the generated ...
- 11
0
votes
0
answers
31
views
How does changing GAN training method afftects the GAN?
I am training a GAN in which I am optimizing or reducing the loss of the generator and the discriminator simultaneously. However, the images generated are very noisy. What could be the hidden reason? ...
- 101
0
votes
2
answers
333
views
Generating synthetic time series data with limited data
I would like some opinions on my current situation.
I have a set of time series data that I want to forecast. The data however is not very long (around 500 rows) so I was looking into generating many ...
1
vote
1
answer
213
views
Does Discriminator in GAN Train only on Real Data or it also Trains on Fake Generated Data
I have been studying GANs and I got confused in the training phase from the discriminator. Which I think only trains on Real data, not on the generated data which then helps in distinguishing or ...
1
vote
0
answers
695
views
How can Discriminator and Generator loss both move to 0?
I have been working with GANs for years and I still can't figure this out. A common thing I see is when the loss of both G and D are near zero. I usually use MSE or cross entropy loss.
Shouldn't this ...
- 2,121
1
vote
0
answers
108
views
CycleGAN cycle loss
I was reading the paper of CycleGAN and I was trying to implement it.
However, my models does not converge to any good solution whatsoever, and since I've checked the implementation many times, I ...
- 859
0
votes
0
answers
214
views
What's GAN's Input-size Limitations?
I am interested in GAN for generating synthetic data. I am studying the input limitations for GAN starting from which GAN is no longer usable.
I have found many applications that use GANs for ...
- 101
1
vote
0
answers
32
views
Conditional GAN training
Quite a simple question about this paper:
How is the cGAN trained? I'm interested in the Pix2Pix network
In particular, given the batch approach, given a specific step, is this the correct procedure?
...
- 859
0
votes
0
answers
27
views
Pix2Pix facede dataset, prevent "gray" in dataset to be predicted
I'm trying to build from scratch the pix2pix architecture, the one on this paper.
As they did, I'm using the facade dataset, and this is one of their result:
I'm particularly interested in the last ...
- 859
1
vote
1
answer
318
views
GAN artifacts on borders
not quite a math-question, but I have a doubt.
I'm trying to build from scratch the Pix2pix network, on the facades dataset, and I think I finally got a good model (from the paper I borrowed just the ...
- 859
0
votes
1
answer
100
views
plotting the latent space of a GAN
I am working on gans and wanted to know how I can plot the latent space of gan. Like I have a latent space of shape (50,250). So it is an n-d array of length 50 and 250 points representing each one of ...
- 1
1
vote
0
answers
106
views
Can you make a simple 2D ring with a GAN?
I am trying to model simple 2d continuous distributions with GANs. Here, I focus on a 2d distribution following a ring structure. The architecture of my networks are:
...
- 163
1
vote
0
answers
63
views
Can I adjust the Wasserstein GAN loss function for my particular data?
I am working on building Generative Adversarial Networks for the purpose of generating synthetic flight data. The GAN will be trained on actual time-series flight data in the form of a (n,m,9) array ...
3
votes
2
answers
2k
views
Should I be using batchnorm and/or dropout in a VAE or GAN?
I am trying to design some generative NN models on datasets of RGB images and was debating on whether I should be using dropout and/or batch norm.
Here are my thoughts (I may be completely wrong):
...
- 207
1
vote
1
answer
691
views
Initialization of GAN discriminator
The question is pretty straightforward: how are GAN and WGAN discriminators typically initialized?
I couldn't find much info on this. E.x. for GANs, I imagine you would theoretically want the ...
- 133
1
vote
0
answers
170
views
FID as a metric to evaluate the quality of synthetic datasets (Non GAN generated) for training models for a given classification task
I am working on a problem of generating synthetic data (algorithmically by blender, not using GANs) to aid the training of some CNN for a classification ask. Ideally, I want to generate an algorithm ...
- 177
2
votes
0
answers
169
views
Hyperparameters tuning on GANs
I have seen this post talking about how to tune hyperparameters on GANs. I'm actually wondering, more generally, how does one go about tuning hyperparameters on GANs.
Obviously you cannot (I mean you ...
1
vote
0
answers
227
views
The discriminator is classifying everything as fake. What does it mean?
I am using a conditional GAN with a relativistic loss function for both generator and discriminator (https://arxiv.org/abs/1807.00734). Before I added the relativistic part, the discriminator ...
- 11
0
votes
1
answer
40
views
Is there such a thing as intra-sample modal collapse in GANs?
Mode collapse is a known issue in generative adversarial networks (GANs) whereby the generator only learns a subset of the real data distribution. In those cases, it only outputs variations of a small ...
- 169
1
vote
0
answers
199
views
Why does a GAN generate samples from a random prior?
I've been reading Goodfellow et. al.'s paper on GANs and also the conditional GAN one by Mirza et. al.
While relatively straight forward, I'm not sure I understand why the prior for the generator is ...
- 679
1
vote
0
answers
177
views
Synthetic data generation - GANs vs Simulator?
For synthetic data generation, does the GAN perform better than a simulator? If so, what are the limitations of the simulator?
If we consider Conditional GANs, we could generate data based on the ...
- 111
1
vote
1
answer
64
views
Question: Optimal D notation in Generative Adversarial Network (GANs)
I am completely new to Computer Vision and how Deep Neural Networks work on images in general. In particular, I have questions on the Discriminator component of Adversarial Generative Network (GANs). ...
- 21
1
vote
1
answer
1k
views
Question about latent space for GANs
I am currently reading about GANs and I had a question about latent space.
A site mentions:
Latent space refers to an abstract multi-dimensional space containing
feature values that we cannot ...
- 113
1
vote
0
answers
60
views
GANs training initially degrades pre-trained generator
I have an issue with the training of a GAN, which consists of a generator and two discriminators. The generator is used to generate waveforms.
1-The generator is independently pre-trained by ...
- 131
2
votes
0
answers
63
views
Which part of an encoder/decoder generative network is improved by adding a discriminator loss term?
Lets say you're doing a superresolution image task with "deep learning" constructs. You encode to a latent representation using some parameterized model (like a neural network), then decode ...
- 141
0
votes
0
answers
54
views
What is a good approach to increase the depth of Nifti file format or Dicom file series?
I have a CT scan dataset of skull fracture consisting of multiple fractures and normal cases, the CT scans are in Dicom format. I want to do multi-class classification. But not every Dicom image ...
- 33
0
votes
0
answers
305
views
I need help understanding the meaning of the loss values of a WGAN with Gradient Penalty
I am currently working on training a Auxiliary Classifier Wasserstein GAN with Gradient Penalty. I based my implementation off of https://keras.io/examples/generative/wgan_gp/ (to which I added the ...
0
votes
1
answer
114
views
Generating the next heatmap in a sequence of labelled heatmaps without assuming continuity
I have a sequence of labelled heatmaps and I want to generate a new heatmap that is the 'best guess' at what the next heatmap in the sequence will look like, without assuming that the next heatmap ...
- 1
1
vote
1
answer
43
views
The motivation of the value function of the generative adversarial network
The seminal paper on the generative adversarial network, proposes to
$$\min_G\max_D V(D,G)$$
where the value function
$$V(D,G):=\int p_{\text{data}}(x)\ln p(D=\text{data}|x)\,dx+\int p_{\text{gen}}(x)\...
- 945
1
vote
0
answers
185
views
GAN discriminator loss randomly jumps up to over 1
I'm training my GAN, but it seems the losses (especially the discriminator loss) are quite erratic. It quickly converges to 0, but will randomly jump up to 1 or even 2-3 sometimes. I'm wondering what ...
- 11
1
vote
1
answer
31
views
How to approach making a GAN where row order does not matter
I am working on a project that has the aim of generating "recipes" which are the summation of "ingredients" (1-dimentional length N tensors, where each index is a value that ...
-1
votes
1
answer
75
views
Can a single GAN learn multiple probability distributions
Assuming for example to generate a sentence, I have two distributions i.e one that represents a semantic and another a syntactic structure of a sentence. This distributions are different. Can I train ...
- 21
0
votes
0
answers
150
views
Using GAN for generating data to augment training data
I want to model an experiment data using a neural network but as the data set size is too small (25 samples), I decided to use a GAN to generate more data. The input is a 4×1 tensor representing 4 ...
1
vote
1
answer
31
views
Affects of including generated data into "real" dataset
I was thinking about what the outcome of the following idea would be. Let's say that we have a Generative Adversarial Network (GAN) that has "successfully" (i.e., Discriminator is not able ...
