Graph Classification on MoleculeNet (test)
68.56Tox21 AccuracyPatchNet
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| PatchNetPre-training strategy=AM+CP2026.06 | 68.56 | 59.64 | 53.06 | 59.5 | 68.55 | 59.74 | 1.2 | |
| GraphGPSPre-training strategy=AM+CP2026.06 | 65.77 | 56.13 | 52.8 | 58.75 | 68.73 | 52.11 | 2.5 | |
| GraphormerPre-training strategy=AM+CP2026.06 | 63.04 | 57.14 | 52.54 | 55.61 | 66.24 | 57.56 | 3.2 | |
| GATPre-training strategy=AM+CP2026.06 | 61.62 | 53.86 | 51.83 | 58.41 | 68.14 | 54.01 | 3.8 | |
| GINPre-training strategy=AM+CP2026.06 | 60.17 | 54.19 | 51.44 | 53.59 | 68.1 | 50.02 | 5 | |
| GCNPre-training strategy=AM+CP2026.06 | 55.25 | 52.14 | 51.94 | 50.96 | 65.27 | 53.62 | 5.3 |