ML Research Wiki / Benchmarks / Formation Energy / QM9

QM9

Formation Energy Benchmark

Performance Over Time

📊 Showing 11 results | 📏 Metric: MAE

Top Performing Models

Rank	Model	Paper	MAE	Date	Code
1	TensorNet	TensorNet: Cartesian Tensor Representations for Efficient Learning of Molecular Potentials	0.09	2023-06-10	📦 torchmd/torchmd-net 📦 materialsvirtuallab/matgl 📦 jla-gardner/graph-pes
2	PAMNet	A Universal Framework for Accurate and Efficient Geometric Deep Learning of Molecular Systems	0.14	2023-11-19	📦 XieResearchGroup/Physics-aware-Multiplex-GNN
3	MXMNet	Molecular Mechanics-Driven Graph Neural Network with Multiplex Graph for Molecular Structures	0.14	2020-11-15	📦 zetayue/MXMNet 📦 smiles724/protmd 📦 Lin-Group-at-UMass/FBGNN-MBE
4	HMGNN	Heterogeneous Molecular Graph Neural Networks for Predicting Molecule Properties	0.14	2020-09-26	📦 shuix007/HMGNN
5	xGPR -- Gaussian process, graph convolution kernel	Linear-scaling kernels for protein sequences and small molecules outperform deep learning while providing uncertainty quantitation and improved interpretability	0.17	2023-02-07	📦 jlparki/xgpr 📦 jlparkI/benchmarking_xGPR
6	DimeNet	Directional Message Passing for Molecular Graphs	0.19	2020-03-06	📦 dmlc/dgl 📦 klicperajo/dimenet 📦 intellabs/matsciml 📦 akirasosa/pytorch-dimenet
7	SchNet-edge-update	Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials	0.24	2018-06-08	📦 nrel/m2p 📦 toshi-k/kaggle-champs-scalar-coupling 📦 peterbjorgensen/msgnet 📦 toshi-k/kaggle-trends-assessment-prediction 📦 tisabe/jraph_mpeu
8	HIP-NN	Hierarchical modeling of molecular energies using a deep neural network	0.26	2017-09-29	-
9	SchNet	Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials	0.31	2018-06-08	📦 nrel/m2p 📦 toshi-k/kaggle-champs-scalar-coupling 📦 peterbjorgensen/msgnet 📦 toshi-k/kaggle-trends-assessment-prediction 📦 tisabe/jraph_mpeu
10	MPNN	Neural Message Passing for Quantum Chemistry	0.49	2017-04-04	📦 Microsoft/gated-graph-neural-network-samples 📦 cts198859/deeprl_network 📦 cts198859/deeprl_dist

All Papers (11)

TensorNet: Cartesian Tensor Representations for Efficient Learning of Molecular Potentials

2023

TensorNet

torchmd/torchmd-net materialsvirtuallab/matgl jla-gardner/graph-pes

A Universal Framework for Accurate and Efficient Geometric Deep Learning of Molecular Systems

2023

PAMNet

XieResearchGroup/Physics-aware-Multiplex-GNN

Molecular Mechanics-Driven Graph Neural Network with Multiplex Graph for Molecular Structures

2020

MXMNet

zetayue/MXMNet smiles724/protmd Lin-Group-at-UMass/FBGNN-MBE

Heterogeneous Molecular Graph Neural Networks for Predicting Molecule Properties

2020

HMGNN

shuix007/HMGNN

Linear-scaling kernels for protein sequences and small molecules outperform deep learning while providing uncertainty quantitation and improved interpretability

2023

xGPR -- Gaussian process, graph convolution kernel

jlparki/xgpr jlparkI/benchmarking_xGPR

Directional Message Passing for Molecular Graphs

2020

DimeNet

dmlc/dgl klicperajo/dimenet

Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials

2018

SchNet-edge-update

nrel/m2p toshi-k/kaggle-champs-scalar-coupling

Hierarchical modeling of molecular energies using a deep neural network

2017

HIP-NN

Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials

2018

SchNet

nrel/m2p toshi-k/kaggle-champs-scalar-coupling

Neural Message Passing for Quantum Chemistry

2017

MPNN

Microsoft/gated-graph-neural-network-samples cts198859/deeprl_network

Wigner kernels: body-ordered equivariant machine learning without a basis

2023

Wigner Kernels

QM9

Performance Over Time

Edit Benchmark Results

Edit Result

Top Performing Models

All Papers (11)

TensorNet: Cartesian Tensor Representations for Efficient Learning of Molecular Potentials

A Universal Framework for Accurate and Efficient Geometric Deep Learning of Molecular Systems

Molecular Mechanics-Driven Graph Neural Network with Multiplex Graph for Molecular Structures

Heterogeneous Molecular Graph Neural Networks for Predicting Molecule Properties

Linear-scaling kernels for protein sequences and small molecules outperform deep learning while providing uncertainty quantitation and improved interpretability

Directional Message Passing for Molecular Graphs

Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials

Hierarchical modeling of molecular energies using a deep neural network

Neural Message Passing with Edge Updates for Predicting Properties of Molecules and Materials

Neural Message Passing for Quantum Chemistry

Wigner kernels: body-ordered equivariant machine learning without a basis

Model	Paper	MAE	Date
TensorNet	TensorNet: Cartesian Tensor Representations for E…	0.09	2023-06-10
PAMNet	A Universal Framework for Accurate and Efficient …	0.14	2023-11-19
MXMNet	Molecular Mechanics-Driven Graph Neural Network w…	0.14	2020-11-15
HMGNN	Heterogeneous Molecular Graph Neural Networks for…	0.14	2020-09-26
xGPR -- Gaussian process, graph convolution kernel	Linear-scaling kernels for protein sequences and …	0.17	2023-02-07
DimeNet	Directional Message Passing for Molecular Graphs	0.19	2020-03-06
SchNet-edge-update	Neural Message Passing with Edge Updates for Pred…	0.24	2018-06-08
HIP-NN	Hierarchical modeling of molecular energies using…	0.26	2017-09-29
SchNet	Neural Message Passing with Edge Updates for Pred…	0.31	2018-06-08
MPNN	Neural Message Passing for Quantum Chemistry	0.49	2017-04-04
Wigner Kernels	Wigner kernels: body-ordered equivariant machine …		2023-03-07