Skip to content

Pytorch implementation of various Graph Neural Networks (GNNs) for graph classification

License

Notifications You must be signed in to change notification settings

qbxlvnf11/graph-neural-networks-for-graph-classification

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

40 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Description

- Graph Neural Networks (GNNs) for Graph Classification

  • Implementation of various neural graph classification model (not node classification)
  • Training and test of various Graph Neural Networks (GNNs) models using graph classification datasets
  • Input graph: graph adjacency matrix, graph node features matrix
  • Graph classification model (graph aggregating)
    • Get latent graph node featrue matrix
    • GCN, GAT, GIN, ...
  • Readout: transforming each latent node feature to one dimension vector for graph classification
  • Feature modeling: fully-connected layer

How to use

- Details of parameter: references to help of argparse

python train.py --model_list GCN GAT --dataset_list ALL --readout_list ALL --n_agg_layer 2 -- agg_hidden 32

Contents

- Available Model

- Available Datasets

  • Node labels X, edge labels X: IMDB-BINARY, IMDB-MULTI
  • Node labels O, edge labels X: PROTEINS, ENZYMES, NCI1
  • Node labels O, edge labels O: MUTAG

- Available Readout

  • Basic readout: max, avg, sum

References

- Graph classification data processing

https://github.com/bknyaz/graph_nn

Author

About

Pytorch implementation of various Graph Neural Networks (GNNs) for graph classification

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages