A Generative Adversarial Network is trained to generate realistic human faces from celebrity images.
In this project, a Generative Adversarial Network is first used to generate new characters using the MNIST dataset. Then, on photographs of celebrities, a fresh network is trained to generate new faces.
face_generation.ipynb
– Project notebook.
face_generation.html
– HTML Export of the project notebook.
helper.py
– Helper functions.
problem_unittests.py
– Unit tests.