Skip to content

In this project, we are going to define two adversarial networks, a generator and a discriminator, and train them until we can generate realistic faces.

Notifications You must be signed in to change notification settings

ahmedhasandrlnd/Face_Generation

Repository files navigation

Face Generation

Introduction

In this project, we'll use generative adversarial networks to generate new images of faces. The model is trained using celebrity images for 50 epochs to generate some realistic face images by its own. We can see from the images below that as we train the model for more epochs, it generates realistic images with more authentic features.

Images generated after epoch: 10

epoch10

Images generated after epoch: 20

epoch10

Images generated after epoch: 30

epoch10

Images generated after epoch: 40

epoch10

Images generated after epoch: 50

epoch10

Getting the project files

The project files can be found in our public GitHub repo, in the project-face-generation folder.

git clone https://github.com/udacity/deep-learning-v2-pytorch/tree/master/

Since we need GPU support in this project, so we decided to complete our project using the provided, in-classroom project notebook; this will just require that we run the project notebook, and download the complete files later on.

Structure of Project

  • dlnd_face_generation.ipynb: Notebook of the project
  • problem_unittests.py: Test file for different TODO functions in the project
  • dlnd_face_generation.html: HTML version of the notebook
  • images folder: different epochs outputs

About

In this project, we are going to define two adversarial networks, a generator and a discriminator, and train them until we can generate realistic faces.

Topics

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published