Training a DCGAN on a dataset of faces to get the generator network to generate new images that look like realistic faces. The model was trained on a subset of the celebA dataset.
pyhton3 train.py
pyhton3 test.py
Plot of the training losses for the generator and discriminator:
