Graph Classification On Mnist

Metrics

Accuracy

Results

Performance results of various models on this benchmark

		Paper Title
ESA (Edge set attention, no positional encodings, tuned)	98.917±0.020	An end-to-end attention-based approach for learning on graphs
NeuralWalker	98.760 ± 0.079	Learning Long Range Dependencies on Graphs via Random Walks
ESA (Edge set attention, no positional encodings)	98.753±0.041	An end-to-end attention-based approach for learning on graphs
GatedGCN+	98.712 ± 0.137	Can Classic GNNs Be Strong Baselines for Graph-level Tasks? Simple Architectures Meet Excellence
CKGCN	98.423	CKGConv: General Graph Convolution with Continuous Kernels
Exphormer	98.414±0.038	Exphormer: Sparse Transformers for Graphs
GCN+	98.382 ± 0.095	Can Classic GNNs Be Strong Baselines for Graph-level Tasks? Simple Architectures Meet Excellence
EIGENFORMER	98.362	Graph Transformers without Positional Encodings
TIGT	98.230±0.133	Topology-Informed Graph Transformer
EGT	98.173	Global Self-Attention as a Replacement for Graph Convolution
GRIT	98.108	Graph Inductive Biases in Transformers without Message Passing
GPS	98.05	Recipe for a General, Powerful, Scalable Graph Transformer
GatedGCN	97.340	Benchmarking Graph Neural Networks

0 of 13 row(s) selected.

Graph Classification On Mnist

Metrics

Accuracy

Results

Performance results of various models on this benchmark

		Paper Title
ESA (Edge set attention, no positional encodings, tuned)	98.917±0.020	An end-to-end attention-based approach for learning on graphs
NeuralWalker	98.760 ± 0.079	Learning Long Range Dependencies on Graphs via Random Walks
ESA (Edge set attention, no positional encodings)	98.753±0.041	An end-to-end attention-based approach for learning on graphs
GatedGCN+	98.712 ± 0.137	Can Classic GNNs Be Strong Baselines for Graph-level Tasks? Simple Architectures Meet Excellence
CKGCN	98.423	CKGConv: General Graph Convolution with Continuous Kernels
Exphormer	98.414±0.038	Exphormer: Sparse Transformers for Graphs
GCN+	98.382 ± 0.095	Can Classic GNNs Be Strong Baselines for Graph-level Tasks? Simple Architectures Meet Excellence
EIGENFORMER	98.362	Graph Transformers without Positional Encodings
TIGT	98.230±0.133	Topology-Informed Graph Transformer
EGT	98.173	Global Self-Attention as a Replacement for Graph Convolution
GRIT	98.108	Graph Inductive Biases in Transformers without Message Passing
GPS	98.05	Recipe for a General, Powerful, Scalable Graph Transformer
GatedGCN	97.340	Benchmarking Graph Neural Networks

0 of 13 row(s) selected.

Graph Classification On Mnist | SOTA | HyperAI