https://github.com/titu1994/tf_gon
Tensorflow 2.x implementation of Gradient Origin Networks
https://github.com/titu1994/tf_gon
Last synced: 6 months ago
JSON representation
Tensorflow 2.x implementation of Gradient Origin Networks
- Host: GitHub
- URL: https://github.com/titu1994/tf_gon
- Owner: titu1994
- License: mit
- Created: 2020-07-13T18:33:05.000Z (over 5 years ago)
- Default Branch: master
- Last Pushed: 2020-07-13T18:47:51.000Z (over 5 years ago)
- Last Synced: 2025-03-25T05:34:07.155Z (7 months ago)
- Language: Python
- Size: 19.5 MB
- Stars: 12
- Watchers: 4
- Forks: 3
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project
README
# Gradient Origin Network for Tensorflow 2.0
Tensorflow 2.0 port of [Gradient Origin Networs](https://arxiv.org/abs/2007.02798) paper from the original author's implementation found at - https://github.com/cwkx/GON.
# Usage
## MNIST
MNIST models (4.3k parameter models) can be trained in under 500 steps to good quality, so they dont have weights restoration. They take just a few minutes to train, even on CPU.
Simply run `mnist_example.py`, and you it will create a directory called `mnist/images` which stores 3 kinds of images - reconstruction, sampled images and spherical interpolation between images.
## CIFAR 10
CIFAR 10 models are much larger - 260k parameters, with a larger latent space (256 dim), and need to be trained for a lot longer (100k train steps). Therefore weights have been provided along with log of training.
Simply run `cifar10_train.py` which takes around 12 - 14 hours on a good GPU to train.
To evaluate results with pre-trained model, run `cifar10_eval.py`. It will create images under `cifar10/images/*_eval.png`
# Requirements
`pip install -r requirements.txt`.
# Citation
Please cite the original authors for their work -
```
@article{bondtaylor2020gradient,
title={Gradient Origin Networks},
author={Sam Bond-Taylor and Chris G. Willcocks},
journal={arXiv preprint arXiv:2007.02798},
year={2020}
}
```