Nature Communications

Table 2 The benchmark results on OC20 IS2RE task

From: Data-driven quantum chemical property prediction leveraging 3D conformations with Uni-Mol+

Results on validation set
	Energy MAE (eV) ↓					EwT (%) ↑
Model	ID	OOD Ads.	OOD Cat.	OOD Both	AVG.	ID	OOD Ads.	OOD Cat.	OOD Both	AVG.
SchNet²⁶	0.6465	0.7074	0.6475	0.6626	0.6660	2.96	2.22	3.03	2.38	2.65
DimeNet++⁴⁰	0.5636	0.7127	0.5612	0.6492	0.6217	4.25	2.48	4.40	2.56	3.42
GemNet-T⁴¹	0.5561	0.7342	0.5659	0.6964	0.6382	4.51	2.24	4.37	2.38	3.38
SphereNet⁴²	0.5632	0.6682	0.5590	0.6190	0.6024	4.56	2.70	4.59	2.70	3.64
Graphormer-3D⁵	0.4329	0.5850	0.4441	0.5299	0.4980	-	-	-	-	-
GNS⁴³	0.54	0.65	0.55	0.59	0.5825	-	-	-	-	-
GNS+NN⁴³	0.47	0.51	0.48	0.46	0.4800	-	-	-	-	-
EquiFormer⁴⁴	0.4222	0.5420	0.4231	0.4754	0.4657	7.23	3.77	7.13	4.10	5.56
EquiFormer+NN⁴⁴	0.4156	0.4976	0.4165	0.4344	0.4410	7.47	4.64	7.19	4.84	6.04
DRFormer⁴⁵	0.4187	0.4863	0.4321	0.4332	0.4425	8.39	5.42	8.12	5.44	6.84
Uni-Mol+	0.3787	0.4519	0.4009	0.4048	0.4119	11.02	6.61	10.00	6.38	8.60
	± 0.0007	± 0.0049	± 0.0001	± 0.0037	± 0.0036
Results on test set
SchNet²⁶	0.639	0.734	0.662	0.704	0.6848	2.96	2.33	2.94	2.21	2.61
DimeNet++⁴⁰	0.562	0.725	0.576	0.661	0.631	4.25	2.07	4.1	2.41	3.21
SphereNet⁴²	0.563	0.703	0.571	0.638	0.6188	4.47	2.29	4.09	2.41	3.32
Graphormer-3D⁵	0.3976	0.5719	0.4166	0.5029	0.4722	8.97	3.45	8.18	3.79	6.1
GNS+NN⁴³	0.4219	0.5678	0.4366	0.4651	0.4728	9.12	4.25	8.01	4.64	6.5
EquiFormer⁴⁴	0.5037	0.6881	0.5213	0.6301	0.5858	5.14	2.41	4.67	2.69	3.73
EquiFormer+NN⁴⁴	0.4171	0.5479	0.4248	0.4741	0.4660	7.71	3.70	7.15	4.07	5.66
DRFormer⁴⁵	0.3865	0.5435	0.4060	0.4677	0.4509	9.18	4.01	8.39	4.33	6.48
Uni-Mol+	0.3745	0.4760	0.3980	0.4086	0.4143	11.29	6.05	9.53	6.06	8.23

NN refers to Noisy Nodes⁴³. We highlight the best results in bold. Source data are provided as a Source Data file.

Back to article page

Search

Advanced search

Quick links