Selective Classification on ImageNet-S
14AURCEnergy
Evaluation results
| Method | Links | ||
|---|---|---|---|
| EnergyBackbone=EVA, Protocol=Supervised learning, Coverage=Full2025.05 | 14 | 0.641 | |
| MaxLogitBackbone=EVA, Protocol=Supervised learning, Coverage=Full2025.05 | 10.3 | 0.437 | |
| SIRCBackbone=EVA, Protocol=Supervised learning, Coverage=Full2025.05 | 8.71 | 0.35 | |
| MSPBackbone=EVA, Protocol=Supervised learning, Coverage=Full2025.05 | 8.15 | 0.319 | |
| MDSBackbone=EVA, Protocol=Supervised learning, Coverage=Full2025.05 | 7.26 | 0.271 | |
| KNNBackbone=EVA, Protocol=Supervised learning, Coverage=Full2025.05 | 7.15 | 0.265 | |
| Δ-KNNBackbone=EVA, Protocol=Supervised learning, Coverage=Full2025.05 | 5.91 | 0.197 | |
| RLogBackbone=EVA, Protocol=Supervised learning, Coverage=Full2025.05 | 5.9 | 0.197 | |
| Δ-MDSBackbone=EVA, Protocol=Supervised learning, Coverage=Full2025.05 | 5.76 | 0.189 | |
| Δ-KNN-RLogBackbone=EVA, Protocol=Supervised learning, Coverage=Full2025.05 | 5.63 | 0.182 | |
| Δ-MDS-RLogBackbone=EVA, Protocol=Supervised learning, Coverage=Full2025.05 | 5.45 | 0.172 |