Formation Energy
On the QM9 dataset the numbers reported in the table are the mean absolute error in eV on the target variable U0 divided by U0's chemical accuracy, which is equal to 0.043.
Papers
Showing 1–10 of 54 papers
All datasetsQM9Materials ProjectJARVIS-DFT3BPAAcetylacetoneAspirinEthanolGeTeLiPSLiPS20NaphthaleneOQMD v1.2
Benchmark Results
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | HDAD+KRR | MAE | 0.58 | — | Unverified |
| 2 | MPNN | MAE | 0.49 | — | Unverified |
| 3 | SchNet | MAE | 0.31 | — | Unverified |
| 4 | ALIGNN | MAE | 0.3 | — | Unverified |
| 5 | MEGNet-simple | MAE | 0.28 | — | Unverified |
| 6 | HIP-NN | MAE | 0.26 | — | Unverified |
| 7 | SchNet-edge-update | MAE | 0.24 | — | Unverified |
| 8 | MEGNet-Full | MAE | 0.21 | — | Unverified |
| 9 | PhysNet | MAE | 0.19 | — | Unverified |
| 10 | DimeNet | MAE | 0.19 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | CGNN-128 | MAE | 35.7 | — | Unverified |
| 2 | CGNN-160 | MAE | 35.1 | — | Unverified |
| 3 | CGNN-192 | MAE | 34.6 | — | Unverified |
| 4 | CGNN Ensemble | MAE | 30.5 | — | Unverified |
| # | Model | Metric | Claimed | Verified | Status |
|---|---|---|---|---|---|
| 1 | SchNet | MAE | 0.31 | — | Unverified |
| 2 | CGNN Trio Ensemble | MAE | 0.04 | — | Unverified |
| 3 | CGNN Full Ensemble | MAE | 0.03 | — | Unverified |