Molecular property prediction on BACE (test)
89.4ROC-AUCBioT5
Evaluation Results
| Method | Links | |
|---|---|---|
| BioT5Model Category=Single-task Specialist Models2024.02 | 89.4 | |
| MolCLRModel Category=Single-task Specialist Models2024.02 | 89 | |
| MolXPTModel Category=Single-task Specialist Models2024.02 | 88.4 | |
| ProtoW-L2Contrastive regularization=true, Distance metric=L22020.06 | 87.3 | |
| ProtoW-DotContrastive regularization=true, Distance metric=Dot2020.06 | 87.1 | |
| Fingerprint+MLP2020.06 | 87 | |
| ProtoW-L2Contrastive regularization=false, Distance metric=L22020.06 | 87 | |
| ProtoW-DotContrastive regularization=false, Distance metric=Dot2020.06 | 86.7 | |
| D-MPNNPooling method=summation2020.06 | 86.5 | |
| ProtoS-L2Contrastive regularization=true, Distance metric=L22020.06 | 86.5 | |
| BioT5+Model Category=LLM-based Generalist Models2024.02 | 86.2 | |
| GINPooling method=summation2020.06 | 86.1 | |
| GATPooling method=summation2020.06 | 86 | |
| D-MPNN+TopK PoolPooling method=TopK Pool2020.06 | 86 | |
| InstructMol-G-6.9BModel Category=LLM-based Generalist Models2024.02 | 85.9 | |
| GPF-plusPre-training Strategy=ContextPred2022.09 | 85.81 | |
| GPF-plusPre-training Strategy=AttrMasking2022.09 | 85.76 | |
| Uni-MolModel Category=Single-task Specialist Models2024.02 | 85.7 | |
| GEMModel Category=Single-task Specialist Models2024.02 | 85.6 | |
| D-MPNN+SAG PoolPooling method=SAG Pool2020.06 | 85.5 | |
| GPFPre-training Strategy=ContextPred2022.09 | 85.32 | |
| SUPTsoftPre-training Strategy=ContextPred2024.02 | 85.27 | |
| SUPTsoftPre-training Strategy=AttrMasking2024.02 | 85.22 | |
| SUPThardPre-training Strategy=ContextPred2024.02 | 85.21 | |
| GPFPre-training Strategy=ContextPred2024.02 | 85.03 | |
| GPF-plusPre-training Strategy=AttrMasking2024.02 | 85.01 | |
| SUPThardPre-training Strategy=AttrMasking2024.02 | 84.98 | |
| GPFPre-training Strategy=AttrMasking2024.02 | 84.77 | |
| FTPre-training Strategy=ContextPred2022.09 | 84.66 | |
| FTPre-training Strategy=ContextPred2024.02 | 84.66 | |
| GPF-plusPre-training Strategy=ContextPred2024.02 | 84.5 | |
| GPFPre-training Strategy=AttrMasking2022.09 | 84.33 | |
| SUPThardPre-training Strategy=Infomax2024.02 | 84.05 | |
| SUPTsoftPre-training Strategy=Infomax2024.02 | 84.01 | |
| MolFMModel Category=Single-task Specialist Models2024.02 | 83.9 | |
| GPF-plusPre-training Strategy=Infomax2022.09 | 83.67 | |
| GPFPre-training Strategy=Infomax2022.09 | 83.6 | |
| GPFPre-training Strategy=Infomax2024.02 | 83.24 | |
| GPF-plusPre-training Strategy=Infomax2024.02 | 82.96 | |
| InstructMol-GS-6.9BModel Category=LLM-based Generalist Models2024.02 | 82.3 | |
| GPF-plusPre-training Strategy=EdgePred2022.09 | 81.75 | |
| SUPThardPre-training Strategy=EdgePred2024.02 | 81.72 | |
| GPFPre-training Strategy=EdgePred2022.09 | 81.57 | |
| SUPTsoftPre-training Strategy=EdgePred2024.02 | 81.44 | |
| FTPre-training Strategy=Infomax2022.09 | 81.32 | |
| FTPre-training Strategy=Infomax2024.02 | 81.32 | |
| GraphMVP-CModel Category=Single-task Specialist Models2024.02 | 81.2 | |
| GIT-MolInput Modality=Graph + SMILES2023.08 | 81.08 | |
| FTPre-training Strategy=AttrMasking2022.09 | 80.94 | |
| FTPre-training Strategy=AttrMasking2024.02 | 80.94 | |
| GPF-plusPre-training Strategy=EdgePred2024.02 | 80.91 | |
| FTPre-training Strategy=EdgePred2022.09 | 80.9 | |
| FTPre-training Strategy=EdgePred2024.02 | 80.9 | |
| Mole-BERT2023.08 | 80.8 | |
| GPFPre-training Strategy=EdgePred2024.02 | 79.76 | |
| MGSSL (DFS)Backbone=GIN, Evaluation Protocol=Fine-tuning2021.10 | 79.7 | |
| MGSSLModel Category=Single-task Specialist Models2024.02 | 79.7 | |
| GPF-plusPre-training Strategy=GCL2022.09 | 79.61 | |
| GroverBackbone=GIN, Evaluation Protocol=Fine-tuning2021.10 | 79.5 | |
| GraphMVP2023.08 | 79.3 | |
| SUPThardPre-training Strategy=GCL2024.02 | 79.17 | |
| MGSSL (BFS)Backbone=GIN, Evaluation Protocol=Fine-tuning2021.10 | 79.1 | |
| SUPTsoftPre-training Strategy=GCL2024.02 | 78.92 | |
| GPFPre-training Strategy=GCL2022.09 | 78.55 | |
| GPT-GNNBackbone=GIN, Evaluation Protocol=Fine-tuning2021.10 | 78.5 | |
| KV-PLM2023.08 | 78.5 | |
| Attribute maskingBackbone=GIN, Evaluation Protocol=Fine-tuning2021.10 | 78.3 | |
| GPF-plusPre-training Strategy=GCL2024.02 | 77.84 | |
| MoMuModel Category=Single-task Specialist Models2024.02 | 76.7 | |
| MoMu2023.08 | 76.7 | |
| GPFPre-training Strategy=GCL2024.02 | 76.67 | |
| FTPre-training Strategy=GCL2022.09 | 75.51 | |
| FTPre-training Strategy=GCL2024.02 | 75.51 | |
| GraphCLModel Category=Single-task Specialist Models2024.02 | 75.4 | |
| InfomaxBackbone=GIN, Evaluation Protocol=Fine-tuning2021.10 | 75 | |
| GCCBackbone=GIN, Evaluation Protocol=Fine-tuning2021.10 | 75 | |
| Llama-2-7B-chatModel Category=LLM-based Generalist Models, tuning=LoRA2024.02 | 74.8 | |
| GraphCL2023.08 | 74.6 | |
| Galactica-30BModel Category=LLM-based Generalist Models2024.02 | 72.7 | |
| KV-PLMModel Category=Single-task Specialist Models2024.02 | 71.9 | |
| GPPTPre-training Strategy=EdgePred2022.09 | 70.85 | |
| GPPTPre-training Strategy=EdgePred2024.02 | 70.85 | |
| GPPT (w/o ol)Pre-training Strategy=EdgePred2022.09 | 70.31 | |
| GPPTw/o olPre-training Strategy=EdgePred2024.02 | 70.31 | |
| No pretrainBackbone=GIN, Evaluation Protocol=Fine-tuning2021.10 | 70 | |
| GIT-MolInput Modality=SMILES2023.08 | 68.4 | |
| Vicuna-v1.3-7BModel Category=LLM-based Generalist Models, tuning=LoRA2024.02 | 68.3 | |
| GraphPromptPre-training Strategy=EdgePred2022.09 | 67.7 | |
| GraphPromptPre-training Strategy=EdgePred2024.02 | 67.7 | |
| GIT-MolInput Modality=Graph2023.08 | 65.8 | |
| Galactica-120BModel Category=LLM-based Generalist Models2024.02 | 61.7 | |
| Galactica-6.7BModel Category=LLM-based Generalist Models2024.02 | 58.4 | |
| Vicuna-v1.5-13B-16kModel Category=LLM-based Generalist Models, Shot=4-shot2024.02 | 49.2 |