Skip to content

This repository contains PyTorch implementation of the following paper: "Order Matters: Probabilistic Modeling of Node Sequence for Graph Generation"

License

Notifications You must be signed in to change notification settings

tufts-ml/graph-generation-vi

Repository files navigation

Order Matters: Probabilistic Modeling of Node Sequencefor Graph Generation

This repository contains PyTorch implementation of the following paper: "Order Matters: Probabilistic Modeling of Node Sequence for Graph Generation"

0. Environment Setup

enviroment setup: "run conda install -f enviroment.yml"

installation of Graph automorphism library: https://web.cs.dal.ca/~peter/software/pynauty/html/install.html

1. Experiment

#DGMG sh experiments/DGMG_caveman_small.sh sh experiments/DGMG_ENZYMES.sh #GraphRNN sh experiments/GraphRNN_caveman_small.sh sh experiments/GraphRNN_Lung.sh #Graphgen sh experiments/Graphgen_citeseer_small.sh sh experiments/Graphgen_ENZYMES.sh 

2. Training

To list the arguments, run the following command:

python main.py -h 

To train the given model on Lung dataset, run the following:

python main.py \ --graph_tyep Lung \ --note <GraphRNN, DGMG, Graphgen> \ --sample_size 16 \ --gcn_type <gat, gcn, appnp> \ --max_cr_iteration 5 \ --enable_gcn 

To train the given model on ENZYMES dataset, run the following:

python main.py \ --graph_tyep ENZYMES \ --note <GraphRNN, DGMG, Graphgen> \ --sample_size 16 \ --gcn_type <gat, gcn, appnp> \ --max_cr_iteration 5 \ --enable_gcn 

To train the given model on caveman_small dataset, run the following:

python main.py \ --graph_tyep caveman_small \ --note <GraphRNN, DGMG, Graphgen> \ --sample_size 16 \ --gcn_type <gat, gcn, appnp> \ --max_cr_iteration 5 \ --enable_gcn 

To train the given model on citeseer_small dataset, run the following:

python main.py \ --graph_tyep citeseer_small \ --note <GraphRNN, DGMG, Graphgen> \ --sample_size 16 \ --gcn_type <gat, gcn, appnp> \ --max_cr_iteration 5 \ --enable_gcn 

About

This repository contains PyTorch implementation of the following paper: "Order Matters: Probabilistic Modeling of Node Sequence for Graph Generation"

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 2

  •  
  •