README.md

GRLC: Graph Representation Learning with Constraints

This repository contains the reference code for the paper Graph Representation Learning with Constraints (TNNLS submission).

Contents

0. Installation
1. Preparation
2. Basics
3. [GPU Setting](#GPU Setting)

Installation

pip install -r requirements.txt

Preparation

Dataset (--dataset-class, --dataset-name,--Custom-key)

Dataset class	Dataset name	Custom key
Planetoid	Cora	classification
Planetoid	CiteSeer	classification
Planetoid	PubMed	classification
WikiCS	WikiCS	classification
MyAmazon	Photo	classification
MyCitationFull	CoraFull	classification
MyCitationFull	DBLP	classification
Crocodile	Crocodile	classification
PygNodePropPredDataset	ogbn-arxiv	classification
PygNodePropPredDataset	ogbn-mag	classification
PygNodePropPredDataset	ogbn-products	classification

Important args:

• --usepretraining Test checkpoints
• --dataset-class Planetoid, MyAmazon, WikiCS, MyCitationFull, Crocodile, PygNodePropPredDataset
• --dataset-name Cora, CiteSeer, PubMed, Photo, WikiCS, CoraFull, DBLP, Crocodile, ogbn-arxiv, ogbn-mag, ogbn-products
• --custom_key classification, link, clu

Basics

• The main train/test code is in Code_GRLC/train.py.
• If you want to see the UGRL layer in PyTorch Geometric MessagePassing grammar, refer to Code_GRLC/layers.

GPU Setting

There are three arguments for GPU settings (--num-gpus-total, --num-gpus-to-use, --black-list). Default values are from the author's machine, so we recommend you modify these values from GRLC/args.yaml or by the command line.

• --num-gpus-total (default 4): The total number of GPUs in your machine.
• --num-gpus-to-use (default 1): The number of GPUs you want to use.
• --black-list (default: [1, 2, 3]): The ids of GPUs you want to not use.