Graph Regression On Kit
评估指标
R2
RMSE
评测结果
各个模型在此基准测试上的表现结果
模型名称 | R2 | RMSE | Paper Title | Repository |
|---|---|---|---|---|
| GATv2 | 0.826±0.000 | 0.453±0.826 | How Attentive are Graph Attention Networks? | - |
| GINDrop | 0.835±0.000 | 0.441±0.835 | DropGNN: Random Dropouts Increase the Expressiveness of Graph Neural Networks | - |
| TokenGT | 0.800±0.000 | 0.486±0.800 | Pure Transformers are Powerful Graph Learners | - |
| GCN | 0.814±0.000 | 0.469±0.814 | Semi-Supervised Classification with Graph Convolutional Networks | - |
| Graphormer | OOM | OOM | Do Transformers Really Perform Bad for Graph Representation? | - |
| PNA | 0.843±0.000 | 0.430±0.843 | Principal Neighbourhood Aggregation for Graph Nets | - |
| GAT | 0.833±0.000 | 0.443±0.833 | Graph Attention Networks | - |
| GIN | 0.833±0.000 | 0.444±0.833 | How Powerful are Graph Neural Networks? | - |
| ESA (Edge set attention, no positional encodings) | 0.841±0.000 | 0.433±0.841 | An end-to-end attention-based approach for learning on graphs | - |
0 of 9 row(s) selected.