https://github.com/yuj-umd/graphRNN

Codes for the paper "Learning Graph-Level Representations with Gated Recurrent Neural Networks"
https://github.com/yuj-umd/graphRNN

Last synced: about 2 months ago
JSON representation

Codes for the paper "Learning Graph-Level Representations with Gated Recurrent Neural Networks"

• Host: GitHub
• URL: https://github.com/yuj-umd/graphRNN
• Owner: yuj-umd
• Created: 2018-05-19T23:07:42.000Z (almost 7 years ago)
• Default Branch: master
• Last Pushed: 2019-02-11T04:53:20.000Z (over 6 years ago)
• Last Synced: 2024-10-30T14:43:36.330Z (7 months ago)
• Language: Python
• Homepage:
• Size: 12.9 MB
• Stars: 29
• Watchers: 3
• Forks: 5
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-graph-classification - [Python Reference

README

        
# GraphRNN

Codes for the paper "Learning Graph-Level Representations with Gated Recurrent Neural Networks". (https://arxiv.org/pdf/1805.07683.pdf)

# 1. Setup

Get the source code,

    git clone https://github.com/yuj-umd/graphRNN.git

Install pytorch from https://pytorch.org/

# 2. Usage

Run

    python main.py \

            -seed 1 \

		-data $data \

		-learning_rate $learning_rate \

		-num_epochs $num_epochs \

		-hidden $hidden \

		-fold $fold \

		-embedding_dim $embedding_dim \

		-rnn_hidden_dim $rnn_hidden_dim 

Paramaters are defined as

    data: MUTAG, NCI1, NCI109, DD, ENZYMES

    feat_dim: Number of node labels

    embedding_dim: Dimension of node embedding

    num_class: Number of graph classes

    rnn_hidden_dim: Hidden unit size of RNN

    learning_rate: initial learning_rate

# 3. Reference

    @article{jin2018learning,

        title={Learning Graph-Level Representations with Gated Recurrent Neural Networks},

        author={Jin, Yu and JaJa, Joseph F},

        journal={arXiv preprint arXiv:1805.07683},

        year={2018}

     }