Graph Property Prediction On Ogbg Ppa
評価指標
Ext. data
Number of params
Test Accuracy
Validation Accuracy
評価結果
このベンチマークにおける各モデルのパフォーマンス結果
モデル名 | Ext. data | Number of params | Test Accuracy | Validation Accuracy | Paper Title | Repository |
|---|---|---|---|---|---|---|
| ExpC | No | 1369397 | 0.7976 ± 0.0072 | 0.7518 ± 0.0080 | Breaking the Expressive Bottlenecks of Graph Neural Networks | |
| GCN | No | 479437 | 0.6839 ± 0.0084 | 0.6497 ± 0.0034 | Semi-Supervised Classification with Graph Convolutional Networks | |
| GatedGCN+ | No | 5547557 | 0.8258 ± 0.0055 | 0.7815 ± 0.0043 | Can Classic GNNs Be Strong Baselines for Graph-level Tasks? Simple Architectures Meet Excellence | |
| GCN+ | No | 5549605 | 0.8077 ± 0.0041 | 0.7586 ± 0.0032 | Can Classic GNNs Be Strong Baselines for Graph-level Tasks? Simple Architectures Meet Excellence | |
| GC-T+MCL(6.0) | No | 4006704 | 0.7432 ± 0.0033 | 0.6989 ± 0.0037 | - | - |
| DeeperGCN | No | 2336421 | 0.7712 ± 0.0071 | 0.7313 ± 0.0078 | DeeperGCN: All You Need to Train Deeper GCNs | |
| GIN+virtual node | No | 3288042 | 0.7037 ± 0.0107 | 0.6678 ± 0.0105 | How Powerful are Graph Neural Networks? | |
| DeeperGCN+FLAG | No | 2336421 | 0.7752 ± 0.0069 | 0.7484 ± 0.0052 | Robust Optimization as Data Augmentation for Large-scale Graphs | |
| GIN+FLAG | No | 1836942 | 0.6905 ± 0.0092 | 0.6465 ± 0.0070 | Robust Optimization as Data Augmentation for Large-scale Graphs | |
| PAS+F2GNN | No | 16346166 | 0.8201 ± 0.0019 | 0.7720 ± 0.0023 | - | - |
| GIN+ | No | 8173605 | 0.8107 ± 0.0053 | 0.7786 ± 0.0095 | Can Classic GNNs Be Strong Baselines for Graph-level Tasks? Simple Architectures Meet Excellence | |
| PAS | No | 3717160 | 0.7828 ± 0.0024 | 0.7523 ± 0.0028 | - | - |
| GCN+virtual node+FLAG | No | 1930537 | 0.6944 ± 0.0052 | 0.6638 ± 0.0055 | Robust Optimization as Data Augmentation for Large-scale Graphs | |
| GIN+virtual node+FLAG | No | 3288042 | 0.7245 ± 0.0114 | 0.6789 ± 0.0079 | Robust Optimization as Data Augmentation for Large-scale Graphs | |
| GIN | No | 1836942 | 0.6892 ± 0.0100 | 0.6562 ± 0.0107 | How Powerful are Graph Neural Networks? | |
| ExpC*+bag of tricks | No | 3758642 | 0.8140 ± 0.0028 | 0.7811 ± 0.0012 | - | - |
| GCN+virtual node | No | 1930537 | 0.6857 ± 0.0061 | 0.6511 ± 0.0048 | Semi-Supervised Classification with Graph Convolutional Networks | |
| GPS | No | 3434533 | 0.8015 | 0.7556 ± 0.0027 | Recipe for a General, Powerful, Scalable Graph Transformer |
0 of 18 row(s) selected.