Energy and force prediction on 3BPA 300 K (test)
2.9Energy RMSERACE-Ensemble
Evaluation results
| Method | Links | ||
|---|---|---|---|
| RACE-Ensembletrained_on=300 K2025.10 | 2.9 | 10.2 | |
| MACEtrained_on=300 K2025.10 | 3 | 8.8 | |
| BOTNettrained_on=300 K2025.10 | 3.1 | 11 | |
| NequIPtrained_on=300 K2025.10 | 3.3 | 10.8 | |
| RACEtrained_on=300 K2025.10 | 3.4 | 12.1 | |
| Allegrotrained_on=300 K2025.10 | 3.8 | 13 | |
| RACE-LAtrained_on=300 K, algorithm=Laplace Approximation2025.10 | 4.8 | 18.2 | |
| RACE-DE-JEFtrained_on=300 K, loss=NLL_JEF2025.10 | 5 | 14.8 | |
| MGNNtrained_on=300 K2025.10 | 5.5 | 15.7 | |
| ICTPsymNtrain=502024.05 | 13.43 | 37.27 | |
| MACENtrain=502024.05 | 14.54 | 37.68 | |
| ICTPfullNtrain=502024.05 | 14.98 | 37.21 | |
| ICTPsym+ltNtrain=502024.05 | 16.03 | 38.38 | |
| RACE-DE-Etrained_on=300 K, loss=Deep Ensemble Energy-only2025.10 | 17.5 | 52.7 |