Image Classification on AwA2 (test)
93.11AccuracyCEM
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| CEMBackbone=ResNet2026.06 | 93.11 | 64.6 | 2.278 | |
| HybridCBMBackbone=ResNet2026.06 | 92.25 | 69.9 | 2.803 | |
| FoCA CBMBackbone=ResNet2026.06 | 92.13 | 57.1 | 2.057 | |
| LaBoBackbone=ResNet2026.06 | 91.53 | 69.9 | 2.803 | |
| SCBMBackbone=ResNet2026.06 | 90.4 | 61.5 | 2.096 | |
| Vanilla CBMBackbone=ResNet2026.06 | 90.36 | 62.8 | 2.137 | |
| ProbCBMBackbone=ResNet2026.06 | 89.8 | 66.1 | 2.373 | |
| MLPCBMBackbone=ResNet2026.06 | 89.65 | 63.7 | 2.12 | |
| CF-CBMBackbone=ResNet2026.06 | 89.19 | 69.9 | 2.803 | |
| FoCA CBM-NBackbone=ResNet2026.06 | 88.26 | 65.9 | 2.273 | |
| LFCBMBackbone=ResNet2026.06 | 83.03 | 69.9 | 2.803 | |
| Posthoc CBMBackbone=ResNet2026.06 | 81 | 77.3 | 2.674 |