Graph Classification on ogbg-molpcba (test)
32.43APGPTrans-L
Evaluation Results
| Method | Links | |
|---|---|---|
| GPTrans-LParams=86.0M, Pre-trained on PCQM4M-v2=true2023.12 | 32.43 | |
| GraphGPT-B (24 layers)Params=227.3M, Pre-trained on PCQM4M-v2=true, Layers=242023.12 | 31.81 | |
| GRPE-LargeParams=118.3M, Pre-trained on PCQM4M-v2=true2023.12 | 31.5 | |
| GatedGCNvirtual node=true2024.06 | 31.41 | |
| Graphormer-FLAG#Param=119.5M, Pre-training=true2021.08 | 31.4 | |
| Graphormer-LParams=119.5M, Pre-trained on PCQM4M-v2=true2023.12 | 31.4 | |
| GatedGCNhierarchical support graph=true2024.06 | 31.29 | |
| GraphGPT-B (12 layers)Params=113.6M, Pre-trained on PCQM4M-v2=true, Layers=122023.12 | 31.28 | |
| NeuralWalker2024.06 | 30.86 | |
| GatedGCN2024.06 | 30.66 | |
| PDFNumber of parameters=3.8m2023.05 | 30.31 | |
| PDIFFParams=3.84M, Pre-trained on PCQM4M-v2=false2023.12 | 30.31 | |
| GraphGPT-MParams=37.7M, Pre-trained on PCQM4M-v2=true2023.12 | 30.13 | |
| Nested GINVirtual node=true, Ensemble=true2021.10 | 30.07 | |
| NGIN-VNParams=44.19M, Pre-trained on PCQM4M-v2=false2023.12 | 30.07 | |
| CRaWl2022.05 | 29.86 | |
| CRAWLparameter budget=500k2021.02 | 29.86 | |
| CRAWL2024.06 | 29.86 | |
| SigGate-GT2026.04 | 29.84 | |
| GatedGCN+enhanced=true2025.02 | 29.81 | |
| GINE-APPNPNumber of Parameters=6.15m2021.12 | 29.79 | |
| GINE-APNumber of parameters=6.2m2023.05 | 29.79 | |
| Specformer2023.10 | 29.72 | |
| Specformer2025.02 | 29.72 | |
| Spec-GNNumber of Parameters=1.74m2021.12 | 29.65 | |
| EGT (30 layers)#Param=110.8M, Pre-training=true2021.08 | 29.61 | |
| EGTpre-trained=true, additional pretraining datasets=true2025.02 | 29.61 | |
| GECO2025.02 | 29.61 | |
| EGT-LargerParams=110.8M, Pre-trained on PCQM4M-v2=true2023.12 | 29.61 | |
| Norm-GNNumber of Parameters=1.74m2021.12 | 29.51 | |
| PHC-GNNVirtual node=false2021.10 | 29.47 | |
| PHC-GNN2021.10 | 29.47 | |
| PHC-GNN#Param=1.69M2021.08 | 29.47 | |
| PHC-GNNNumber of Parameters=1.69m2021.12 | 29.47 | |
| PHC-GNNumber of parameters=1.7m2023.05 | 29.47 | |
| Hodge1Lapbackbone=GPS2023.10 | 29.37 | |
| EdgeRWSEbackbone=GPS2023.10 | 29.34 | |
| GIN-AK+2022.05 | 29.3 | |
| GIN-AK+2021.02 | 29.3 | |
| GIN-AK+2023.10 | 29.3 | |
| GNNAK+Number of runs=42023.09 | 29.3 | |
| Exphormer2026.04 | 29.2 | |
| GINEVirtual node=true2021.10 | 29.17 | |
| GINE-VN#Param=6.15M2021.08 | 29.17 | |
| GINE-vnNumber of Parameters=6.15m2021.12 | 29.17 | |
| GINE+2021.02 | 29.17 | |
| GINE-VNParams=6.1M, Pre-trained on PCQM4M-v2=false2023.12 | 29.17 | |
| TransformerComplexity=O(N^2), Positional Encoding=Baseline2025.09 | 29.1 | |
| GPS2022.05 | 29.07 | |
| GPSNumber of parameters=9.7m2023.05 | 29.07 | |
| GPS2023.10 | 29.07 | |
| GPS2024.06 | 29.07 | |
| GPS2024.06 | 29.07 | |
| GraphGPS2025.02 | 29.07 | |
| GraphGPS2026.04 | 29.07 | |
| GIN-VN#Param=3.4M, Pre-training=true2021.08 | 29.02 | |
| DGN2021.10 | 28.85 | |
| DGN#Param=6.73M2021.08 | 28.85 | |
| DGNNumber of Parameters=6.73m2021.12 | 28.85 | |
| DGN2022.05 | 28.85 | |
| DGNNumber of parameters=6.7m2023.05 | 28.85 | |
| DGNNumber of runs=42023.09 | 28.85 | |
| GDeRRemaining Ratio %=70%, Backbone=PNA, Pruning Type=Dynamic2024.10 | 28.5 | |
| Exphormer2025.02 | 28.49 | |
| DeeperGCN + FLAGBackbone=DeeperGCN, Virtual Nodes=false2020.10 | 28.42 | |
| DeeperGCN-VN-FLAG#Param=6.55M2021.08 | 28.42 | |
| DeeperGNumber of Parameters=5.55m2021.12 | 28.42 | |
| DeeperGNumber of parameters=5.6m2023.05 | 28.42 | |
| PNA-LSPE2021.10 | 28.4 | |
| PNA2021.10 | 28.38 | |
| PNA#Param=6.55M2021.08 | 28.38 | |
| PNANumber of Parameters=6.55m2021.12 | 28.38 | |
| PNA2022.05 | 28.38 | |
| PNANumber of parameters=6.6m2023.05 | 28.38 | |
| PNA2021.02 | 28.38 | |
| PNA2023.10 | 28.38 | |
| PNANumber of runs=42023.09 | 28.38 | |
| PNA2026.04 | 28.38 | |
| GIN-Virtual + FLAGBackbone=GIN, Virtual Nodes=true2020.10 | 28.34 | |
| Nested GINVirtual node=true2021.10 | 28.32 | |
| Nested GIN2021.02 | 28.32 | |
| Nested GINNumber of runs=4, Virtual node=true2023.09 | 28.32 | |
| MarginRemaining Ratio %=70%, Backbone=PNA, Pruning Type=Static2024.10 | 28.3 | |
| InfoBatchRemaining Ratio %=70%, Backbone=PNA, Pruning Type=Dynamic2024.10 | 28.3 | |
| UCBRemaining Ratio %=70%, Backbone=PNA, Pruning Type=Dynamic2024.10 | 28.1 | |
| Whole DatasetRemaining Ratio %=100%, Backbone=PNA2024.10 | 28.1 | |
| Hard RandomRemaining Ratio %=70%, Backbone=PNA, Pruning Type=Static2024.10 | 28 | |
| GraNd-4Remaining Ratio %=70%, Backbone=PNA, Pruning Type=Static2024.10 | 28 | |
| GDeRRemaining Ratio %=50%, Backbone=PNA, Pruning Type=Dynamic2024.10 | 28 | |
| ForgettingRemaining Ratio %=70%, Backbone=PNA, Pruning Type=Static2024.10 | 27.9 | |
| GlisterRemaining Ratio %=70%, Backbone=PNA, Pruning Type=Static2024.10 | 27.9 | |
| DeeperGCNvirtual_node=true2021.10 | 27.81 | |
| DeeperGCNBackbone=DeeperGCN, Virtual Nodes=false2020.10 | 27.81 | |
| DeeperGCN2022.05 | 27.81 | |
| DeeperGCN2026.04 | 27.81 | |
| CDRemaining Ratio %=70%, Backbone=PNA, Pruning Type=Static2024.10 | 27.8 | |
| Least ConfidenceRemaining Ratio %=70%, Backbone=PNA, Pruning Type=Static2024.10 | 27.8 | |
| GraNd-20Remaining Ratio %=70%, Backbone=PNA, Pruning Type=Static2024.10 | 27.8 | |
| CraigRemaining Ratio %=70%, Backbone=PNA, Pruning Type=Static2024.10 | 27.8 | |
| Soft RandomRemaining Ratio %=70%, Backbone=PNA, Pruning Type=Dynamic2024.10 | 27.8 |