it does not have a database and with each refresh it generates a complete new face
Generator is responsible to generate from the given data and discriminator has to say if it is fake or not
the practice and learn continues till the discriminator is unable to tell if generated data is fake or real
a really fascinating video about GAN with an interesting example of building a GAN from scratch
the code of the example in this video: https://github.com/luisguiserrano/gans/blob/master/GANs_in_Slanted_Land.ipynb?short_path=28407f0
GitHub
gans/GANs_in_Slanted_Land.ipynb at master · luisguiserrano/gans
GANs in slanted land. Contribute to luisguiserrano/gans development by creating an account on GitHub.