SOTA over time
For each benchmark, plots how the SOTA developed over time.
Graph Learning Leaderboards
Quickly identify the SOTA for common graph-learning benchmarks. Identify core baselines to test against.
Core Features
Explicit lower bounds
GCN/GIN and MLP performance is specifically highlighted to see lower bounds.
Baselines included
Benchmark list also contains baseline results from the papers.
Papers in context
Paper pages present reported results in context of current SOTA and selected simple baselines.
Recently active benchmark groups
OGB
Open Graph Benchmark — large-scale standardised graph ML benchmarks covering graph-level, node-level, and link-level prediction across diverse domains.
ogbn-products, ogbg-molhiv, ogbn-papers100M and 13 more
MoleculeNet
MoleculeNet — benchmark collection for molecular machine learning.
Heterogeneous Graph Benchmarks
Heterogeneous graph datasets from the MAGNN and HGB benchmark papers.
LRGB
Long Range Graph Benchmark — five graph ML tasks designed to test long-range dependency modelling. Introduced by Dwivedi et al. (2022).
Peptides-struct, Peptides-func, PascalVOC-SP and 2 more
Heterophily Benchmark
Five replacement datasets proposed for more reliable evaluation under graph heterophily.
Roman-empire, Amazon-ratings, Minesweeper and 2 more
GNNBenchmark
Benchmarking Graph Neural Networks — artificially and semi-artificially generated graph datasets.