Accuracy Prediction on NAS-Bench-201 (5% samples, train)
0.874Kendall's TauNAR-Former V2
Evaluation results
| Method | Links | |
|---|---|---|
| NAR-Former V2Backbone=Transformer, Self-evolution strategy (SE)=false2023.06 | 0.874 | |
| NAR-FormerBackbone=Transformer, Self-evolution strategy (SE)=false2023.06 | 0.849 | |
| TNASP + SEBackbone=Transformer, Self-evolution strategy (SE)=true2023.06 | 0.69 | |
| TNASPBackbone=Transformer, Self-evolution strategy (SE)=false2023.06 | 0.689 | |
| NP + SEBackbone=GNN, Self-evolution strategy (SE)=true2023.06 | 0.652 | |
| NPBackbone=GNN, Self-evolution strategy (SE)=false2023.06 | 0.634 | |
| NAO + SEBackbone=LSTM, Self-evolution strategy (SE)=true2023.06 | 0.529 | |
| NAOBackbone=LSTM, Self-evolution strategy (SE)=false2023.06 | 0.522 |