This Programm trains & uses a GAN to generate Logos. The Data was provided here: Dataset-Logo-2k.
This project was setup with Anaconda python. `environment.yml` contains all used packages.
Basically a GAN consists of 2 Neural Nets:
- Generator The Generator generates artifical images
- Discriminator The Discriminator classifies an given Image as Real or Artifical
I couldn’t yield any good results as you can see in the given examples:
I didn’t investigate any further why this is the case.






