https://github.com/tfesenko/face-generation

Face Generator Project is a part of Udacity Deep Learning Nanodegree
https://github.com/tfesenko/face-generation

cyclegan generated-faces udacity

Last synced: about 2 months ago
JSON representation

Face Generator Project is a part of Udacity Deep Learning Nanodegree

• Host: GitHub
• URL: https://github.com/tfesenko/face-generation
• Owner: tfesenko
• License: mit
• Created: 2019-11-20T15:39:09.000Z (almost 6 years ago)
• Default Branch: master
• Last Pushed: 2022-04-28T06:29:10.000Z (over 3 years ago)
• Last Synced: 2025-07-30T15:37:27.111Z (2 months ago)
• Topics: cyclegan, generated-faces, udacity
• Language: HTML
• Size: 1.64 MB
• Stars: 4
• Watchers: 1
• Forks: 7
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# Face-Generation

Face Generator Project is a part of Udacity Deep Learning Nanodegree.

# Goal

Generate new faces using Generative Adversarial Networks (GANs).  

The model is trained on the CelebFaces Attributes Dataset (CelebA):

![Image of Training Set](https://github.com/tfesenko/Face-Generation/blob/master/assets/processed_face_data.png)

It generates new human faces that look like this:  

![Image of Generated Faces](https://github.com/tfesenko/Face-Generation/blob/master/assets/Generated_faces2.png)

## Recommended next steps

* Create a deeper model and use it to generate larger (say 128x128) images of faces.

* Read existing literature to see if you can use padding and normalization techniques to generate higher-resolution images.

* Implement a learning rate that evolves over time as they did in this [CycleGAN Github repo](https://github.com/junyanz/pytorch-CycleGAN-and-pix2pix).

* See if you can extend this model and use a CycleGAN to learn to swap different kinds of faces. For example, learn a mapping between faces that have and do not have eye/lip makeup, as they did in [this paper](https://gfx.cs.princeton.edu/pubs/Chang_2018_PAS/Chang-CVPR-2018.pdf).