MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis
Generative Adversarial Network (GAN) which is widely used for Image synthesis via generative modelling suffers peculiarly from training instability. One of the known reasons for this instability is the passage of uninformative gradients from the Discriminator to the Generator due to learning imbalance between them during training. In this work, we propose Multi-Scale Gradients Generative Adversarial Network (MSG-GAN), a simplistic but effective technique for addressing this problem; by allowing the flow of gradients from the Discriminator to the Generator at multiple scales. This results in the Generator acquiring the ability to synthesize synchronized images at multiple resolutions simultaneously. We also highlight a suite of techniques that together buttress the stability of training without excessive hyperparameter tuning. Our MSG-GAN technique is a generic mathematical framework which has multiple instantiations. We present an intuitive form of this technique which uses the concatenation operation in the Discriminator computations and empirically validate it through experiments on the CelebA-HQ, CIFAR10 and Oxford102 flowers datasets and by comparing it with some of the current state-of-the-art techniques.
NurtureToken New!

Token crowdsale for this paper ends in

Buy Nurture Tokens

Authors

Are you an author of this paper? Check the Twitter handle we have for you is correct.

Animesh Karnewar (edit)
Raghu Sesha Iyengar (add twitter)
Ask The Authors

Ask the authors of this paper a question or leave a comment.

Read it. Rate it.
#1. Which part of the paper did you read?

#2. The paper contains new data or analyses that is openly accessible?
#3. The conclusion is supported by the data and analyses?
#4. The conclusion is of scientific interest?
#5. The result is likely to lead to future research?

Github
Stargazers:
202
Forks:
30
Open Issues:
3
Network:
30
Subscribers:
12
Language:
Python
ProGAN package implemented as an extension of PyTorch nn.Module
Youtube
Link:
None (add)
Views:
0
Likes:
0
Dislikes:
0
Favorites:
0
Comments:
0
Other
Sample Sizes (N=):
Inserted:
Words Total:
Words Unique:
Source:
Abstract:
None
03/14/19 06:01PM
5,375
1,813
Tweets
ArxivVision: #arxiv #ComputerVision [https://t.co/xLSoLEW1zO] Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis. (arXiv:1903.06048 [https://t.co/xLSoLEW1zO] [https://t.co/D87vUclEmM]) https://t.co/T8aGvUBEmM While Generative Adversarial Networks (GANs)...
AnimeshKarnewar: Anybody can GAN with MSG! our paper titled "MSG-GAN: Multi-Scale Gradient GAN for Stable Image Synthesis" is available at https://t.co/BoirCFeobr. Code at: https://t.co/iAXm44PDTv. Requires little to no hyperparameter tuning. Special thanks to @jm_alexia. #GAN #MSG_GAN #AI #DL
arxiv_cscv: MSG-GAN: Multi-Scale Gradient GAN for Stable Image Synthesis https://t.co/9vZusPj2aG
arxiv_cscv: MSG-GAN: Multi-Scale Gradient GAN for Stable Image Synthesis https://t.co/9vZusPj2aG
udmrzn: RT @arxiv_cscv: MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis https://t.co/9vZusPj2aG
arxiv_cscv: MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis https://t.co/9vZusPj2aG
arxiv_cscv: MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis https://t.co/9vZusPj2aG
arxiv_cscv: MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis https://t.co/9vZusPj2aG
arxiv_cscv: MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis https://t.co/9vZusPj2aG
udmrzn: RT @arxiv_cscv: MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis https://t.co/9vZusPj2aG
arxiv_cscv: MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis https://t.co/9vZusPj2aG
udmrzn: RT @arxiv_cscv: MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis https://t.co/9vZusPAD2e
arxiv_cscv: MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis https://t.co/9vZusPAD2e
arxivml: "MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis", Animesh Karnewar… https://t.co/rqGb3Niz9T
ibu_hoshina: RT @StatsPapers: MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis. https://t.co/EgZ8uVp7gV
arxiv_cs_LG: MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis. Animesh Karnewar and Raghu Sesha Iyengar https://t.co/LSBma3tcGy
StatsPapers: MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis. https://t.co/EgZ8uVp7gV
arxiv_cscv: MSG-GAN: Multi-Scale Gradients GAN for more stable and synchronized multi-scale image synthesis https://t.co/9vZusPj2aG
Images
Related