Crystal Property Prediction on Materials Project
0.015Formation Energy MAECrysLDNet
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| CrysLDNetEvaluation Protocol=Pretrain-Finetune2026.05 | 0.015 | 0.184 | 0.032 | 0.059 | — | — | |
| PDDFormerEvaluation Protocol=Train-from-scratch2026.05 | 0.016 | 0.189 | 0.034 | 0.062 | — | — | |
| eComFormerEvaluation Protocol=Train-from-scratch2026.05 | 0.018 | 0.202 | 0.042 | 0.073 | — | — | |
| iComFormerEvaluation Protocol=Train-from-scratch2026.05 | 0.018 | 0.193 | 0.038 | 0.064 | — | — | |
| PotNetEvaluation Protocol=Train-from-scratch2026.05 | 0.019 | 0.204 | 0.04 | 0.065 | — | — | |
| DPFEvaluation Protocol=Pretrain-Finetune2026.05 | 0.02 | 0.203 | 0.042 | 0.073 | — | — | |
| MatformerEvaluation Protocol=Train-from-scratch2026.05 | 0.021 | 0.211 | 0.043 | 0.073 | — | — | |
| ALIGNNEvaluation Protocol=Train-from-scratch2026.05 | 0.022 | 0.218 | 0.051 | 0.078 | — | — | |
| MEGNetEvaluation Protocol=Train-from-scratch2026.05 | 0.03 | 0.307 | 0.06 | 0.099 | — | — | |
| CGCNNEvaluation Protocol=Train-from-scratch2026.05 | 0.031 | 0.292 | 0.047 | 0.077 | — | — | |
| SchNetEvaluation Protocol=Train-from-scratch2026.05 | 0.033 | 0.345 | 0.066 | 0.099 | — | — | |
| GATGNNEvaluation Protocol=Train-from-scratch2026.05 | 0.033 | 0.28 | 0.045 | 0.075 | — | — | |
| CrysGNNEvaluation Protocol=Pretrain-Finetune2026.05 | 0.033 | 0.266 | 0.043 | 0.076 | — | — | |
| CrysXPPEvaluation Protocol=Pretrain-Finetune2026.05 | 0.034 | 0.269 | 0.055 | 0.084 | — | — | |
| Crystal TwinsEvaluation Protocol=Pretrain-Finetune2026.05 | 0.034 | 0.269 | 0.051 | 0.082 | — | — | |
| DimeNet++Evaluation Protocol=Train-from-scratch2026.05 | 0.049 | 0.392 | 0.041 | 0.068 | — | — | |
| CoGNEvaluation Protocol=Train-from-scratch2026.05 | 0.05 | 0.204 | 0.046 | 0.07 | — | — | |
| ALM CoreConfiguration=canonical understanding r=128+IT2026.06 | 0.07 | 0.3 | — | — | 0.052 | 0.212 | |
| LLM-PropInput Format=CIF2026.06 | 0.07 | 0.317 | — | — | 0.103 | 0.156 | |
| MatBERTInput Format=CIF2026.06 | 0.091 | 0.348 | — | — | 0.059 | 0.207 | |
| CGCNNInput Format=CIF2026.06 | 0.123 | 0.366 | — | — | 0.058 | 0.246 | |
| EquiformerEvaluation Protocol=Train-from-scratch2026.05 | 0.405 | 0.565 | 0.055 | 0.075 | — | — |