Graph Regression On Pcqm4Mv2 Lsc

평가 지표

Test MAE

Validation MAE

평가 결과

이 벤치마크에서 각 모델의 성능 결과

모델 이름	Test MAE	Validation MAE	Paper Title	Repository
EGT	0.0862	0.0857	Global Self-Attention as a Replacement for Graph Convolution
GPTrans-L	0.0821	0.0809	Graph Propagation Transformer for Graph Representation Learning
GPS	0.0862	0.0852	Recipe for a General, Powerful, Scalable Graph Transformer
TIGT	-	0.0826	Topology-Informed Graph Transformer
Graphormer + GFSA	-	0.0860	Graph Convolutions Enrich the Self-Attention in Transformers!
Graphormer	-	0.0864	Do Transformers Really Perform Bad for Graph Representation?
EGT+SSA+Self-ensemble	-	0.0865	The Information Pathways Hypothesis: Transformers are Dynamic Self-Ensembles
GCN	0.1398	0.1379	Semi-Supervised Classification with Graph Convolutional Networks
ESA (Edge set attention, no positional encodings)	N/A	0.0235	An end-to-end attention-based approach for learning on graphs	-
MLP-Fingerprint	0.1760	0.1753	OGB-LSC: A Large-Scale Challenge for Machine Learning on Graphs
GRIT	-	0.0859	Graph Inductive Biases in Transformers without Message Passing
Uni-Mol+	0.0705	0.0693	Highly Accurate Quantum Chemical Property Prediction with Uni-Mol+
GRPE-Large	0.0876	0.0867	GRPE: Relative Positional Encoding for Graph Transformer
TokenGT	0.0919	0.0910	Pure Transformers are Powerful Graph Learners
EGT + Triangular Attention	0.0683	0.0671	Global Self-Attention as a Replacement for Graph Convolution
TGT-At	0.0683	0.0671	Triplet Interaction Improves Graph Transformers: Accurate Molecular Graph Learning with Triplet Graph Transformers
GIN	0.1218	0.1195	How Powerful are Graph Neural Networks?
GPTrans-T	0.0842	0.0833	Graph Propagation Transformer for Graph Representation Learning
Transformer-M	0.0782	0.0772	One Transformer Can Understand Both 2D & 3D Molecular Data
EGT+SSA	-	0.0876	The Information Pathways Hypothesis: Transformers are Dynamic Self-Ensembles

0 of 20 row(s) selected.