Using DCGAN on a (very small) dataset of art


I am developing a DCGAN using the this tutorial in PyCharm. As my usage of this tutorial suggests, I am quite new to DCGANs as I've previously only had a few experiences with machine learning algorithms on classifying problems. My goal is to feed my DCGAN a dataset of paintings of a specific painter, and get 'new' paintings in return. Needless to say, a painter does not paint thousands of paintings in his life, leaving me with a dataset of around 60 paintings. One of the smallest datasets I have ever worked with. I have two, related, questions:

  1. Is it realistic to properly train a DCGAN on this type of dataset? If not, would there be any alternative you would suggest?
  2. What would be a good set of parameters to start of from to properly train this DCGAN?

Thanks in advance!


Posted 2020-04-22T12:05:46.760

Reputation: 11

No answers