ogbn-papers-100M

The 'ogbn-papers100M' dataset is a large-scale graph dataset used for evaluating the transferability of graph self-supervised learning methods.

Papers using ogbn-papers-100M (12)

GSTBench: A Benchmark Study on the Transferability of Graph Self-Supervised Learning2025Generalizing Graph Transformers Across Diverse Graphs and Tasks via Pre-training2024 · 1 citesSGFormer: Simplifying and Empowering Transformers for Large-Graph Representations2023 · 12 citesGraphMAE2: A Decoding-Enhanced Masked Self-Supervised Graph Learner2023 · 6 citesDink-Net: Neural Clustering on Large Graphs2023 · 6 citesDistGNN-MB: Distributed Large-Scale Graph Neural Network Training on x86 via Minibatch Sampling2022 · 1 citesFreshGNN: Reducing Memory Access via Stable Historical Embeddings for Graph Neural Network Training2023 · 1 citesFedGTA: Topology-aware Averaging for Federated Graph Learning2024Disentangled Condensation for Large-scale Graphs2024Scalable and Certifiable Graph Unlearning: Overcoming the Approximation Error Barrier2024Haste Makes Waste: A Simple Approach for Scaling Graph Neural Networks2024Graph Attention Multi-Layer Perceptron2022