Skip to content

jcatw/dcnn

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

32 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

DCNN

An implementation of Diffusion-Convolutional Neural Networks [1] in Theano and Lasagne.

Installation

git clone https://github.com/jcatw/dcnn.git
cd dcnn

Usage

Node Classification (Cora)

python -m client.run --model=node_classification --data=cora

Graph Classification (NCI1)

python -m client.run --model=graph_classification --data=nci1

Code Structure

client/: Client code for running from the command line.
  parser.py: Parses command line args into configuration parameters.
  run.py: Runs experiments.

data/: Example datasets.

python/: DCNN library.
  data.py: Dataset parsers.
  layers.py: Lasagne internals for DCNN layers.
  models.py: User-facing end-to-end models that provide a scikit-learn-like interface.
  params.py: Simple container for configuration parameters.
  util.py: Misc utility functions.

References

[1] Atwood, James, and Don Towsley. "Diffusion-convolutional neural networks." Advances in Neural Information Processing Systems. 2016.

About

An implementation of Diffusion-Convolutional Neural Networks in Lasagne and Theano.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages