Drug discovery classification on ClinTox
94ROC-AUCUni-Mol
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Uni-MolInput Modality=Molecular graph, uses ground-truth 3D conformers=true2026.05 | 94 | — | |
| MolFormerPooling method=Max, Layer=122026.06 | 91 | — | |
| SAEPooling method=Max, Layer=62026.06 | 91 | — | |
| ECFP02026.06 | 90 | — | |
| MolFormerPooling method=Mean, Layer=92026.06 | 90 | — | |
| SAEPooling method=Mean, Layer=62026.06 | 90 | — | |
| MOLINTERACTInput Modality=Molecular graph2026.05 | 88.4 | — | |
| MOLEBLENDInput Modality=Molecular graph2026.05 | 87.6 | — | |
| MoleculeSDEInput Modality=Molecular graph2026.05 | 87 | — | |
| ECFP22026.06 | 87 | — | |
| MolSightInput Modality=2D image, Training Curriculum=S62026.05 | 86.8 | — | |
| ECFP42026.06 | 82 | — | |
| 3D InfoMaxInput Modality=Molecular graph2026.05 | 79.9 | — | |
| GraphMVPInput Modality=Molecular graph2026.05 | 79 | — | |
| PcDs2026.06 | 77 | — | |
| BLIP-2 (LoRA)Category=VLM, requires LLM/VLM at inference=true, protocol=LoRA2026.05 | — | 89 | |
| GPT-4o (ICL)Category=LLM, requires LLM/VLM at inference=true, protocol=ICL2026.05 | — | 59 | |
| Janus-Pro 7B (ICL)Category=VLM, requires LLM/VLM at inference=true, protocol=ICL2026.05 | — | 83 | |
| MolSight (SigLIP2, S6)Category=Image-only, Configuration=single-model results2026.05 | — | 94 | |
| MolSight-BestCategory=Image-only, Configuration=best curriculum-trained architecture per task2026.05 | — | 94 | |
| Uni-MolCategory=3D GNN, requires 3D conformer generation=true2026.05 | — | 94 |