Few-Shot Classification on CUB (Balanced and Unbalanced N-Shot Accuracy)
83.1Unbalanced 1-Shot AccBAVARDAGE
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| BAVARDAGEBackbone=ResNet-122022.09 | 83.1 | 90.8 | 87.4 | 92 | |
| BAVARDAGEBackbone=ResNet-182022.09 | 82 | 90.7 | 85.6 | 91.4 | |
| alpha-TIMBackbone=ResNet-182022.09 | 75.7 | 89.8 | — | — | |
| TIMBackbone=ResNet-182022.09 | 74.8 | 86.9 | 80.3 | 90.5 | |
| BD-CSPNBackbone=ResNet-182022.09 | 74.5 | 87.1 | 77.9 | 88.9 | |
| LaplacianShotBackbone=ResNet-182022.09 | 73.7 | 87.7 | 78.9 | 88.8 | |
| Entropy-minBackbone=ResNet-182022.09 | 67.5 | 82.9 | 72.8 | 88.9 | |
| PT-MAPBackbone=ResNet-182022.09 | 65.1 | 71.3 | 85.5 | 91.3 |