Graph-level Tasks on QM9 (test)
10Inference Throughput (graphs/sec)FGWBARY-NN
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| FGWBARY-NNK=10, Hardware=NVIDIA V100/Intel Xeon E5-26602024.02 | 10 | — | 10 | |
| RELATIONFORMERHardware=NVIDIA V100/Intel Xeon E5-26602024.02 | 10 | — | 2 | |
| ANY2GRAPHHardware=NVIDIA V100/Intel Xeon E5-26602024.02 | 10 | — | 1 | |
| FGWBARY-ILEK=25, Hardware=NVIDIA V100/Intel Xeon E5-26602024.02 | 1 | — | — | |
| Baselinesexperimental setup=Gc-train-to-Gc-infer2024.10 | — | 0.0017 | — | |
| FIT-GNNcoarsening ratio (r)=0.3, experimental setup=Gc-train-to-Gc-infer2024.10 | — | 0.0018 | — | |
| FIT-GNNcoarsening ratio (r)=0.5, experimental setup=Gc-train-to-Gc-infer2024.10 | — | 0.0019 | — |