Image Classification on ImageNet (val) (Intermediate Feature Metrics: conv1, KNN)
9.1Conv1 Feature MetricLA
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| LABackbone=ResNet-182019.03 | 9.1 | 18.7 | 34.8 | 48.4 | 52.8 | 45 | |
| LABackbone=ResNet-502019.03 | 10.2 | 23.3 | 39.3 | 49 | 60.2 | 49.4 | |
| RandomBackbone=AlexNet2019.03 | 11.6 | 17.1 | 16.9 | 16.3 | 14.1 | 3.5 | |
| ColorBackbone=AlexNet2019.03 | 13.1 | 24.8 | 31 | 32.6 | 31.8 | — | |
| IRBackbone=VGG16, Batch Normalization=true, Reproduced by authors=true2019.03 | 13.2 | 18.7 | 27.3 | 39.8 | 50.4 | 42.1 | |
| DCBackbone=AlexNet2019.03 | 13.4 | 32.3 | 41 | 39.6 | 38.2 | — | |
| LABackbone=VGG162019.03 | 14.3 | 23.4 | 28.3 | 44.5 | 57.6 | 46.6 | |
| IRBackbone=ResNet-502019.03 | 15.3 | 18.8 | 24.9 | 40.6 | 54 | 46.5 | |
| IRBackbone=ResNet-182019.03 | 16 | 19.9 | 29.8 | 39 | 44.5 | 41 | |
| ContextBackbone=AlexNet2019.03 | 16.2 | 23.3 | 30.2 | 31.7 | 29.6 | — | |
| DCBackbone=ResNet-18, Reproduced by authors=true2019.03 | 16.4 | 17.2 | 28.7 | 44.3 | 49.1 | — | |
| IRBackbone=VGG162019.03 | 16.5 | 21.4 | 27.6 | 35.1 | 39.2 | 33.9 | |
| IRBackbone=AlexNet2019.03 | 16.8 | 26.5 | 31.8 | 34.1 | 35.6 | 31.3 | |
| SplitBrainBackbone=AlexNet2019.03 | 17.7 | 29.3 | 35.4 | 35.2 | 32.8 | 11.8 | |
| CountBackbone=AlexNet2019.03 | 18 | 30.6 | 34.3 | 32.5 | 25.7 | — | |
| DCBackbone=VGG16, Reproduced by authors=true2019.03 | 18.2 | 27.5 | 41.5 | 51.3 | 52.7 | — | |
| IRBackbone=AlexNet, Batch Normalization=true, Reproduced by authors=true2019.03 | 18.4 | 30.1 | 34.4 | 39.2 | 39.9 | 34.9 | |
| LABackbone=AlexNet2019.03 | 18.7 | 32.7 | 38.1 | 42.3 | 42.4 | 38.1 | |
| DCBackbone=ResNet-50, Reproduced by authors=true2019.03 | 18.9 | 27.3 | 36.7 | 52.4 | 44.2 | — | |
| JigsawBackbone=AlexNet2019.03 | 19.2 | 30.1 | 34.7 | 33.9 | 28.3 | — |