Implemenation of Transforming Auto-encoders (by Hinton et al.) using TensorFlow.
Paper: Hinton, Geoffrey E., Alex Krizhevsky, and Sida D. Wang. "Transforming auto-encoders." International Conference on Artificial Neural Networks. Springer, Berlin, Heidelberg, 2011.
These instructions describe how I setup my environment for this project. There are different ways to get it running using conda
or traditional pip
.
python3
pipenv
jupyter
cuda 9.0
(If you don't have a GPU, you can replacetensorflow-gpu
bytensorflow
)
Clone the repository and install the python dependencies in a virtual environment using pipenv:
$ git clone git@github.com:HedgehogCode/transforming-autoencoders-tf.git
$ cd transforming-autoencoders-tf
$ pipenv install -d
$ pipenv shell
$ python -m ipykernel install --user --name transforming-autoencoders --display-name "Python (transforming-autoencoders)"
Start jupyter notebook
and open a notebook.
- yingzha/Transforming-Autoencoders: Theano impemetation
- nikhil-dce/Transforming-Autoencoder-TF
- ndrplz/transforming-autoencoders
- ethanscho/TransformingAutoencoders
- ... (just search for Transfroming Autoencoders on GitHub)
- Add other datasets
- Fashion-MNIST should be easy to use
This project is licensed unter the 2-Clause BSD License - see the LICENSE.txt file for details.