Few-shot classification on ImageNet All categories (test)
58.16Top-5 Acc (N=1)Cosine Classifier & Att. Weight Gen
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Cosine Classifier & Att. Weight GenBackbone=ResNet-102018.04 | 58.16 | 65.21 | 72.72 | 76.5 | 78.74 | |
| Cosine Classifier & Avg. Weight GenBackbone=ResNet-102018.04 | 57.65 | 64.69 | 72.35 | 76.18 | 78.46 | |
| Prototype Matching Nets w/ HBackbone=ResNet-10, Hallucination (w/ H)=true2018.04 | 57.6 | 64.7 | 71.9 | 75.2 | 77.5 | |
| Prototype Matching NetsBackbone=ResNet-102018.04 | 55.8 | 63.1 | 71.1 | 75 | 77.1 | |
| Matching NetworksBackbone=ResNet-102018.04 | 54.4 | 61 | 69 | 73.7 | 76.5 | |
| SGM w/ HBackbone=ResNet-10, Hallucination (w/ H)=true2018.04 | 54.3 | 62.1 | 71.3 | 75.8 | 78.1 | |
| Logistic regression w/ HBackbone=ResNet-10, Hallucination (w/ H)=true2018.04 | 52.2 | 59.4 | 67.6 | 72.8 | 76.9 | |
| Prototypical-NetsBackbone=ResNet-102018.04 | 49.5 | 61 | 69.7 | 72.9 | 74.6 | |
| Batch SGMBackbone=ResNet-102018.04 | 49.3 | 60.5 | 71.4 | 75.8 | 78.5 | |
| Logistic regressionBackbone=ResNet-102018.04 | 40.8 | 49.9 | 64.2 | 71.9 | 76.9 |