CKGConv: General Graph Convolution with Continuous Kernels

Benchmark Model Rank Results
node-classification-on-patternCKGCN#1Accuracy: 88.661
node-classification-on-clusterCKGCN#4Accuracy: 79.003
graph-classification-on-mnistCKGCN#5Accuracy: 98.423
graph-regression-on-zincCKGCN#6MAE: 0.059
graph-classification-on-cifar10-100kCKGCN#11Accuracy (%): 72.785
graph-classification-on-peptides-funcCKGCN#16AP: 0.6952
graph-regression-on-peptides-structCKGCN#18MAE: 0.2477
graph-regression-on-zinc-500kCKGCN#37MAE: 5.9