(转) How to Train a GAN? Tips and tricks to make GANs work

Posted The Blog of Xiao Wang

tags:

篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了(转) How to Train a GAN? Tips and tricks to make GANs work相关的知识,希望对你有一定的参考价值。

 

How to Train a GAN? Tips and tricks to make GANs work

转自:https://github.com/soumith/ganhacks 

 

While research in Generative Adversarial Networks (GANs) continues to improve the fundamental stability of these models, we use a bunch of tricks to train them and make them stable day to day.

Here are a summary of some of the tricks.

Here‘s a link to the authors of this document

If you find a trick that is particularly useful in practice, please open a Pull Request to add it to the document. If we find it to be reasonable and verified, we will merge it in.

1. Normalize the inputs

  • normalize the images between -1 and 1
  • Tanh as the last layer of the generator output

2: A modified loss function

In GAN papers, the loss function to optimize G is min (log 1-D), but in practice folks practically use max log D

  • because the first formulation has vanishing gradients early on
  • Goodfellow et. al (2014)

In practice, works well:

  • Flip labels when training generator: real = fake, fake = real

3: Use a spherical Z

  • Dont sample from a Uniform distribution

技术分享

  • Sample from a gaussian distribution

技术分享

  • When doing interpolations, do the interpolation via a great circle, rather than a straight line from point A to point B
  • Tom White‘s Sampling Generative Networks has more details

4: BatchNorm

  • Construct different mini-batches for real and fake, i.e. each mini-batch needs to contain only all real images or all generated images.
  • when batchnorm is not an option use instance normalization (for each sample, subtract mean and divide by standard deviation).

技术分享

5: Avoid Sparse Gradients: ReLU, MaxPool

  • the stability of the GAN game suffers if you have sparse gradients
  • LeakyReLU = good (in both G and D)
  • For Downsampling, use: Average Pooling, Conv2d + stride
  • For Upsampling, use: PixelShuffle, ConvTranspose2d + stride

6: Use Soft and Noisy Labels

  • Label Smoothing, i.e. if you have two target labels: Real=1 and Fake=0, then for each incoming sample, if it is real, then replace the label with a random number between 0.7 and 1.2, and if it is a fake sample, replace it with 0.0 and 0.3 (for example).
    • Salimans et. al. 2016
  • make the labels the noisy for the discriminator: occasionally flip the labels when training the discriminator

7: DCGAN / Hybrid Models

  • Use DCGAN when you can. It works!
  • if you cant use DCGANs and no model is stable, use a hybrid model : KL + GAN or VAE + GAN

8: Use stability tricks from RL

  • Experience Replay
    • Keep a replay buffer of past generations and occassionally show them
    • Keep checkpoints from the past of G and D and occassionaly swap them out for a few iterations
  • All stability tricks that work for deep deterministic policy gradients
  • See Pfau & Vinyals (2016)

9: Use the ADAM Optimizer

  • optim.Adam rules!
    • See Radford et. al. 2015
  • Use SGD for discriminator and ADAM for generator

10: Track failures early

  • D loss goes to 0: failure mode
  • check norms of gradients: if they are over 100 things are screwing up
  • when things are working, D loss has low variance and goes down over time vs having huge variance and spiking
  • if loss of generator steadily decreases, then it‘s fooling D with garbage (says martin)

11: Dont balance loss via statistics (unless you have a good reason to)

  • Dont try to find a (number of G / number of D) schedule to uncollapse training
  • It‘s hard and we‘ve all tried it.
  • If you do try it, have a principled approach to it, rather than intuition

For example

while lossD > A:
  train D
while lossG > B:
  train G

12: If you have labels, use them

  • if you have labels available, training the discriminator to also classify the samples: auxillary GANs

13: Add noise to inputs, decay over time

14: [notsure] Train discriminator more (sometimes)

  • especially when you have noise
  • hard to find a schedule of number of D iterations vs G iterations

15: [notsure] Batch Discrimination

  • Mixed results

16: Discrete variables in Conditional GANs

  • Use an Embedding layer
  • Add as additional channels to images
  • Keep embedding dimensionality low and upsample to match image channel size

Authors

  • Soumith Chintala
  • Emily Denton
  • Martin Arjovsky
  • Michael Mathieu

以上是关于(转) How to Train a GAN? Tips and tricks to make GANs work的主要内容,如果未能解决你的问题,请参考以下文章

[转]Android How to Download and Make Volley.jar

[转]How to: Create a Custom Principal Identity

转-How to install an SSH Server in Windows Server 2008

转 How to Find Out Who Is Locking a Table in MySQL

How to Solve Vxdiag benz c6 doip multiplexer Can’t connect to WiFi

转How to Start Intel Hardware-assisted Virtualization (hypervisor) on Linux to Speed-up Intel Andro