Skip to content
Example code for Weight Normalization, from "Weight Normalization: A Simple Reparameterization to Accelerate Training of Deep Neural Networks"
Python
Branch: master
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
keras
lasagne
tensorflow
LICENSE.md
README.md

README.md

Status: Archive (code is provided as-is, no updates expected)

Weight Normalization

This repo contains example code for Weight Normalization, as described in the following paper:

Weight Normalization: A Simple Reparameterization to Accelerate Training of Deep Neural Networks, by Tim Salimans, and Diederik P. Kingma.

  • The folder 'lasagne' contains code using the Lasagne package for Theano. This code was used to run the CIFAR-10 experiments in the paper.
  • The folder 'tensorflow' contains a single nn.py file with a direct implementation copied from our PixelCNN++ repository.
  • The folder 'keras' contains example code for use with the Keras package.

Citation

If you find this code useful please cite us in your work:

@inproceedings{Salimans2016WeightNorm,
  title={Weight Normalization: A Simple Reparameterization to Accelerate Training of Deep Neural Networks},
  author={Tim Salimans and Diederik P. Kingma},
  booktitle={Neural Information Processing Systems 2016},
  year={2016}
}
You can’t perform that action at this time.