https://github.com/2wins/BEGAN-tensorlayer

A TensorFlow Implementation of "BEGAN: Boundary Equilibrium Generative Adversarial Networks" - (EASY to READ)
https://github.com/2wins/BEGAN-tensorlayer

gan generative-adversarial-network generative-adversarial-networks tensorflow tensorlayer

Last synced: 2 months ago
JSON representation

A TensorFlow Implementation of "BEGAN: Boundary Equilibrium Generative Adversarial Networks" - (EASY to READ)

• Host: GitHub
• URL: https://github.com/2wins/BEGAN-tensorlayer
• Owner: 2wins
• Created: 2018-05-04T15:58:44.000Z (over 6 years ago)
• Default Branch: master
• Last Pushed: 2018-05-22T01:53:01.000Z (over 6 years ago)
• Last Synced: 2024-08-02T20:44:31.226Z (6 months ago)
• Topics: gan, generative-adversarial-network, generative-adversarial-networks, tensorflow, tensorlayer
• Language: Python
• Size: 14.4 MB
• Stars: 4
• Watchers: 2
• Forks: 0
• Open Issues: 0
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-tensorlayer - BEGAN

README

        
# BEGAN in TensorFlow / TensorLayer

TensorFlow / TensorLayer implementation of [BEGAN: Boundary Equilibrium Generative Adversarial Networks](http://arxiv.org/abs/1703.10717)

## Prerequisites

- Python 2.7 or Python 3.3+

- [TensorFlow==1.0+](https://www.tensorflow.org/)

- [TensorLayer==1.4+](https://github.com/tensorlayer/tensorlayer)

## Usage

First, download images to `data/celebA`:

    $ python download.py celebA		[202599 face images]

Second, train the GAN:

    $ python main.py --point "25 58"

Third, generate faces with the trained generator:

    $ python generate.py --num_imgs 1000

## Result on CelebA

From scratch to 60k (frames captured every 500 iter.)

`gamma=0.5`