Image Classification on ImageNet-C Severity Level 5, Mixed Shifts v1 (test)
70.9AccuracyDualTTA
Evaluation results
| Method | Links | |
|---|---|---|
| DualTTABackbone=ViTBase-LN2026.04 | 70.9 | |
| DeYOBackbone=ViTBase-LN2026.04 | 69.73 | |
| DeYOBackbone=VitBase-LN2024.03 | 59.4 | |
| EATABackbone=ViTBase-LN2026.04 | 58.64 | |
| SARBackbone=VitBase-LN2024.03 | 57.1 | |
| EATABackbone=VitBase-LN2024.03 | 56.4 | |
| No adaptBackbone=ViTBase-LN2026.04 | 54.37 | |
| SARBackbone=ViTBase-LN2026.04 | 54.33 | |
| TentBackbone=ViTBase-LN2026.04 | 52.82 | |
| DeYOBackbone=ResNet-GN2026.04 | 39.26 | |
| MEMOBackbone=VitBase-LN2024.03 | 39.1 | |
| DualTTABackbone=ResNet-GN2026.04 | 38.92 | |
| DeYOBackbone=ResNet-50-GN2024.03 | 38.6 | |
| EATABackbone=ResNet-GN2026.04 | 38.5 | |
| SARBackbone=ResNet-50-GN2024.03 | 38.3 | |
| EATABackbone=ResNet-50-GN2024.03 | 38.2 | |
| SARBackbone=ResNet-GN2026.04 | 35.09 | |
| TentBackbone=ResNet-50-GN2024.03 | 34.2 | |
| No adaptBackbone=ResNet-GN2026.04 | 31.44 | |
| MEMOBackbone=ResNet-50-GN2024.03 | 31.2 | |
| ResNet-50-GNBackbone=ResNet-50-GN2024.03 | 30.6 | |
| VitBase-LNBackbone=VitBase-LN2024.03 | 29.9 | |
| DualTTABackbone=ResNet-BN2026.04 | 27.51 | |
| SARBackbone=ResNet-BN2026.04 | 26.1 | |
| TentBackbone=VitBase-LN2024.03 | 24.1 | |
| EATABackbone=ResNet-BN2026.04 | 19.62 | |
| TentBackbone=ResNet-GN2026.04 | 13.16 | |
| DeYOBackbone=ResNet-BN2026.04 | 11.63 | |
| TentBackbone=ResNet-BN2026.04 | 2.35 | |
| No adaptBackbone=ResNet-BN2026.04 | 0.14 |