https://github.com/danwild/face-generator

Face generation with DCGAN
https://github.com/danwild/face-generator

dcgan deep-learning face-generation gans pytorch udacity-nanodegree

Last synced: 4 days ago
JSON representation

Face generation with DCGAN

• Host: GitHub
• URL: https://github.com/danwild/face-generator
• Owner: danwild
• License: apache-2.0
• Created: 2019-07-01T00:55:03.000Z (almost 7 years ago)
• Default Branch: master
• Last Pushed: 2019-07-02T01:50:13.000Z (almost 7 years ago)
• Last Synced: 2025-01-07T21:35:02.518Z (over 1 year ago)
• Topics: dcgan, deep-learning, face-generation, gans, pytorch, udacity-nanodegree
• Language: HTML
• Size: 1.09 MB
• Stars: 0
• Watchers: 1
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# face-generator [![Binder](https://mybinder.org/badge_logo.svg)](https://mybinder.org/v2/gh/danwild/face-generator/master)

Face generation with Deep Convolutional Generative Adversarial Networks (DCGAN).

Uses the [CelebFaces Attributes Dataset (CelebA)](http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html) to train the networks, which then produces new faces (which are mostly white people due to bias in training dataset).

## Example output:

![Screenshot](/screenshots/faces-50.png?raw=true)

## Setup notes

I'm using AWS' [Deep Learning AMI (Ubuntu) v23](https://aws.amazon.com/marketplace/pp/B077GCH38C) (however to run remotely on CPU follow the binder link).

```

# enter provided env

source activate pytorch_p36

python -m pip install -r requirements.txt

# quick setup jupyter

jupyter notebook --generate-config

sed -ie "s/#c.NotebookApp.ip = 'localhost'/#c.NotebookApp.ip = '*'/g" ~/.jupyter/jupyter_notebook_config.py

jupyter notebook --ip=0.0.0.0 --no-browser

# navigate to url provided (will need to replace IP)

# may also need to select kernal corresponding to env above

```

## Images

Need to download:

- [CelebFaces Attributes Dataset (CelebA)](http://mmlab.ie.cuhk.edu.hk/projects/CelebA.html)

  - Goes in: `/assets`