Read the following two resources to realize the inherent complexity in GANs and tuning them. No need to get into intimate detail, just skim at this point. This information will come in handy when you're looking at papers to see which techniques they've implemented.

What I've found works best is if you checkpoint your model when you're starting to get some "knowable" results (ie. your GAN is showing promise). From there, you spin up separate experiments and run each for a number of epochs.

Something like this is good:

  1. Run for 100 epochs with the same params
  2. Run for 100 epochs but edit one param in an "upwards" direction
  3. Run for 100 epochs but edit one param in a "downwards" direction
  4. Compare and tune your baseline (as applicable)
  5. Repeat.
Covers: implementation of GAN Tuning
Questions this item addresses:
  • What are the common ways to tune GANs?
Author(s) / creator(s) / reference(s)
Jason Brownlee
0 comment

Quick tips to tune your GANs

Total time needed: ~20 minutes
Help you get started in developing and tuning GANs which generate high-resolution artwork
Potential Use Cases
image generation of high-resolution artwork
Who is this for ?
INTERMEDIATEPython developers with some experience with Deep Neural Nets and want to try their hands at generating artwork
Click on each of the following annotated items to see details.
WRITEUP 1. GAN - Ways to improve GAN performance
  • What are the common ways to tune GANs?
10 minutes
WRITEUP 2. A couple of good ways to deal with limited datasets....
  • What happens when I don't have enough data?
20 minutes
REPO 3. Use DeepSpeed
  • What are the ways that I can parallelize and speed up my GANs' training?
10 minutes
WRITEUP 4. Use Dilations on your Discriminator to increase the receptive field size of your convolutional filters.
  • What does increasing the dilation of your convolutional filters do?
10 minutes

Concepts Covered

0 comment