Skip to content

Latest commit

 

History

History
78 lines (53 loc) · 2.02 KB

README.md

File metadata and controls

78 lines (53 loc) · 2.02 KB
Raw

PGGAN-tensorflow

The tensorflow implementation of PROGRESSIVE GROWING OF GANS FOR IMPROVED QUALITY, STABILITY, AND VARIATION.

The generative process of PG-GAN

Differences with the original paper.

  • Recently, just generate 64x64 and 128x128 pixels samples.

Setup

Prerequisites

  • TensorFlow >= 1.4
  • python 2.7 or 3

Getting Started

  • Clone this repo:
git clone https://github.com/zhangqianhui/progressive_growing_of_gans_tensorflow.git
cd progressive_growing_of_gans_tensorflow
  • Download the CelebA dataset

You can download the CelebA dataset and unzip CelebA into a directory. Noted that this directory don't contain the sub-directory.

  • The method for creating CelebA-HQ can be found on Method

  • Train the model on CelebA dataset

python main.py --path=your celeba data-path --celeba=True
  • Train the model on CelebA-HQ dataset
python main.py --path=your celeba-hq data-path --celeba=False

Results on celebA dataset

Here is the generated 64x64 results(Left: generated; Right: Real):

Here is the generated 128x128 results(Left: generated; Right: Real):

Results on CelebA-HQ dataset

Here is the generated 64x64 results(Left: Real; Right: Generated):

Here is the generated 128x128 results(Left: Real; Right: Generated):

Issue

If you find some bugs, Thanks for your issue to propose it.

Reference code

PGGAN Theano

PGGAN Pytorch