If you make use of this code in your own work, please cite our paper: Jiawei E，Yinglong Zhang, Shangying Yang, Hong Wang, Xuewwen Xia, Xing Xu. GraphSAGE++: Weighted Multi-scale GNN for Graph Representation Learning. Neural Processing Letters, accepted

PyTorch Geometric Datasets Loading Guide

PyTorch Geometric (PyG) is a library built on PyTorch that offers a wide range of tools for dealing with graph data. This document outlines how to obtain some common datasets - Amazon, Cora, Pubmed, Citeseer, and PPI - using PyG.

Installation of PyTorch Geometric

Before accessing the datasets, make sure that you have PyTorch Geometric installed. You can install it using the following commands: pip install torch-scatter torch-sparse torch-cluster torch-spline-conv torch-geometric

Ensure that these installation commands are compatible with your PyTorch and CUDA versions.

Datasets Acquisition

Cora, Pubmed, Citeseer

These datasets are part of the Planetoid datasets and can be loaded as follows: from torch_geometric.datasets import Planetoid

Select the dataset

datasets = ["Cora", "Pubmed", "Citeseer"]

for dataset_name in datasets: dataset = Planetoid(root=f'./data/{dataset_name}', name=dataset_name) print(f'{dataset_name} dataset loaded. Contains {len(dataset)} graphs.')

Amazon

The Amazon datasets can be loaded as follows: from torch_geometric.datasets import Amazon

Select the dataset, e.g., 'Computers' or 'Photo'

dataset_name = 'Computers' # or 'Photo' dataset = Amazon(root=f'./data/Amazon/{dataset_name}', name=dataset_name) print(f'Amazon {dataset_name} dataset loaded. Contains {len(dataset)} graphs.')

PPI

The PPI (Protein-Protein Interaction) dataset can be loaded as follows: from torch_geometric.datasets import PPI dataset = PPI(root='./data/PPI') print(f'PPI dataset loaded. Contains {len(dataset)} graphs.')