Skip to content

walkerwu/GroundHog

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

GroundHog by lisa-groundhog

WARNING: Groundhog development is over. Please consider using Blocks instead. For an example of machine translation using Blocks please see Blocks-examples repository

GroundHog is a python framework on top of Theano (http://deeplearning.net/software/theano/) that aims to provide a flexible, yet efficient way of implementing complex recurrent neural network models. It supports a variety of recurrent layers, such as DT-RNN, DOT-RNN, RNN with gated hidden units and LSTM. Furthermore, it enables the flexible combination of various layers, for instance, to build a neural translation model.

This is a version forked from the original GroundHog (https://github.com/pascanur/GroundHog) developed by Razvan Pascanu, Caglar Gulcehre and Kyunghyun Cho. This fork will be the version developed and maintained by the members of the LISA Lab at the University of Montreal. The main contributors and maintainers of this fork are currently Dzmitry Bahdanau and Kyunghyun Cho.

Most of the library documentation is still work in progress, but check the files containing Tut (in tutorials) for a quick tutorial on how to use the library.

The library is under the 3-clause BSD license, so it may be used for commercial purposes.

Installation

To install Groundhog in a multi-user setting (such as the LISA lab)

python setup.py develop --user

For general installation, simply use

python setup.py develop

NOTE: This will install the development version of Theano, if Theano is not currently installed.

Neural Machine Translation

See experiments/nmt/README.md

About

Library for implementing RNNs with Theano

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 92.3%
  • Perl 5.6%
  • HTML 1.6%
  • Other 0.5%