Molecular Property Prediction On Esol
Métriques
R2
RMSE
Résultats
Résultats de performance de divers modèles sur ce benchmark
Tableau comparatif
| Nom du modèle | R2 | RMSE |
|---|---|---|
| graph-attention-networks | 0.930±0.007 | 0.540±0.027 |
| chemrl-gem-geometry-enhanced-molecular | - | 0.798 |
| do-transformers-really-perform-bad-for-graph | 0.908±0.021 | 0.618±0.068 |
| moleculenet-a-benchmark-for-molecular-machine | - | 0.99 |
| how-powerful-are-graph-neural-networks | 0.938±0.011 | 0.509±0.044 |
| are-learned-molecular-representations-ready | - | 1.050 |
| moleculenet-a-benchmark-for-molecular-machine | - | 0.58 |
| semi-supervised-classification-with-graph | 0.936±0.006 | 0.520±0.024 |
| how-attentive-are-graph-attention-networks | 0.928±0.005 | 0.549±0.020 |
| pre-training-graph-neural-networks-on | - | 0.816±0.019 |
| principal-neighbourhood-aggregation-for-graph | 0.942±0.006 | 0.493±0.026 |
| a-bayesian-flow-network-framework-for | - | 0.884 |
| dropgnn-random-dropouts-increase-the | 0.935±0.012 | 0.520±0.048 |
| molecular-structure-property-co-trained | - | 0.810 |
| uni-mol-a-universal-3d-molecular | - | 0.788 |
| self-guided-masked-autoencoders-for-domain | - | 0.623 |
| chemberta-2-towards-chemical-foundation | - | 0.889 |
| pure-transformers-are-powerful-graph-learners | 0.892±0.036 | 0.667±0.103 |
| recipe-for-a-general-powerful-scalable-graph | 0.911±0.003 | 0.613±0.010 |
| masked-attention-is-all-you-need-for-graphs | 0.944±0.002 | 0.485±0.009 |