ogb 开源项目推荐（19 个）

VITA-Group /Deep_GCN_Benchmarking

[TPAMI 2022]“训练深层图神经网络的技巧袋综合基准研究”作者：陈天龙 *、周凯雄 *、段克宇、郑文青、王培浩、夏胡、王张洋

alibaba /graph-gpt

Generative Pre-trained Graph Eulerian Transformer [ICML2025]

ytchx1999 /PyG-OGB-Tricks

Bags of Tricks in OGB (node classification) with GCNs.

yzhuoning /DeepAUC_OGB_Challenge

Top1 Solution on OGB Challenge (Graph Property Prediction on HIV dataset)

xnuohz /DeeperGCN-dgl

A DGL implementation of "DeeperGCN: All You Need to Train Deeper GCNs".

venomouscyanide /S3GRL_OGB

S3GRL is a scalable SGRL method for faster link prediction using efficient precomputations.

xnuohz /CorrectAndSmooth-dgl

A DGL implementation of "Combining Label Propagation and Simple Models Out-performs Graph Neural Networks" (ICLR 2021).

yao8839836 /ogb_report

该项目在 GitHub 上没有填写描述

ytchx1999 /ogbl-vessel

OGBL-Vessel submmision.

declansam /OverFiT

beeHIVe: AI-powered platform for HIV drug discovery and prediction.

anmo-fish /aa-dc

AA-DC is a structural correction framework for link prediction. It addresses two systematic biases that arise when relational feasibility is evaluated under coarse recognition mappings, such as Adamic–Adar (AA).

Markel15 /ogb-project

GNNs for protein graph classification on OGBG-PPA with GCN, GIN and baseline models.

pinpin-http /arxiv-gnn

Réseaux de neurones sur graphe (GCN, GraphSAGE, GAT) en PyTorch sur ogbn-arxiv : ce que les citations apportent au texte seul (55,7 % → 71,5-71,9 %), sur le protocole officiel d'OGB.

wojtke /graph-u-nets

该项目在 GitHub 上没有填写描述

Wb-az /pyg-graph-networks

This repository classifies nodes and graphs and predicts links using PyG, covering data processing, model building, and testing

Vignesh3538 /EdgeGNN

Exploration of vanilla GIN, GIN with edge heuristics in prediction head, Heuristics weighted Edge-GNN for link prediction in ogbl-ddi dataset

jawadelMorabit-smi /graphormer-molhiv

Fine-tuning Microsoft Graphormer on OGB molHIV: test ROC-AUC 0.694 using only 18% of training data (6k molecules, 3 epochs, 1xT4). Transfer learning across molecular graphs

TUM-DI-Lab-Graph-Scaling /baseline-evaluation

Baseline Evaluation for the project "Acceleration of Neural Network Training with Microsoft DeepSpeed"

nata2627 /gbd

Машинное обучение на графах: метрики, эмбеддинги вершин, предсказание свойств вершин на ogbn-arxiv (Open Graph Benchmark)