Skip to content

MXNet implementation of Graph Convolutional Neural Networks

License

Notifications You must be signed in to change notification settings

saravsak/graph-cnn.mxnet

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

18 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

MXNet implementation of Graph Convolutional Neural Networks detailed in Semi-Supervised Classification with Graph Convolutional Networks

Tensorflow Implementation (Original) Author's PyTorch Implementation Pytorch Implementation by Bumsoo Kim

Graph Convolutional Networks

Graph Convolutions are best explained in this amazing blogpost by Thomas Kipf the author of the paper Semi-Supervised Classification with Graph Convolutional Networks

Requirements

  • Python
  • MXnet
  • Numpy

Usage

python train.py

Notes/Observations:

  1. Unable to reproduce results from the paper with dropout of 0.5. Dropout=0 gives results similar to paper

TODO:

  1. NetworkX integration for loading data
  2. Add options to pick loss function
  3. Support datasets other than CORA
  4. Option to save model
  5. Add visualization support of loss parameters