Image Classification on Stanford Cars (top-1/top-5 error)
3.8Top-1 Error RateDAT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| DAT2020.10 | 3.8 | — | |
| EfficientNet-L2Backbone=EfficientNet-L2, SAM Optimization=true, Pre-training Dataset=ImageNet plus JFT2020.10 | 4.04 | — | |
| EfficientNet-L2Backbone=EfficientNet-L2, SAM Optimization=false, Pre-training Dataset=ImageNet plus JFT2020.10 | 4.93 | — | |
| TBMSL-NetPre-training Dataset=ImageNet-only2020.10 | 5 | — | |
| EfficientNet-b7Backbone=EfficientNet-b7, SAM Optimization=true, Pre-training Dataset=ImageNet-only2020.10 | 5.18 | — | |
| EfficientNet-b7Backbone=EfficientNet-b7, SAM Optimization=false, Pre-training Dataset=ImageNet-only2020.10 | 5.94 | — | |
| Individual NetworksBackbone=ResNet-502018.01 | 8.17 | — | |
| Individual NetworksBackbone=DenseNet-1212018.01 | 8.64 | — | |
| Individual NetworksBackbone=VGG-16 BN2018.01 | 9.41 | — | |
| PiggybackBackbone=VGG-16 BN2018.01 | 9.87 | — | |
| PiggybackBackbone=ResNet-502018.01 | 10.38 | — | |
| PiggybackBackbone=DenseNet-1212018.01 | 10.87 | — | |
| PackNetBackbone=ResNet-50, Task addition order=↓2018.01 | 13.89 | — | |
| Individual NetworksBackbone=VGG-16, # Models=4, Size=2,173 MB2017.11 | 13.97 | — | |
| PackNetBackbone=VGG-16 BN, Task addition order=↓2018.01 | 14.05 | — | |
| PackNetBackbone=DenseNet-121, Task addition order=↓2018.01 | 15.35 | — | |
| PackNetBackbone=VGG-16, Pruning Ratios=0.75, 0.75, 0.75, # Models=1, Size=595 MB2017.11 | 15.75 | — | |
| PackNetBackbone=VGG-16 BN, Task addition order=↑2018.01 | 17.6 | — | |
| PackNetBackbone=VGG-16, Pruning Ratios=0.50, 0.75, 0.75, # Models=1, Size=595 MB2017.11 | 18.08 | — | |
| PackNetBackbone=ResNet-50, Task addition order=↑2018.01 | 19.99 | — | |
| PackNetBackbone=DenseNet-121, Task addition order=↑2018.01 | 22.09 | — | |
| LwFBackbone=VGG-16, # Models=1, Size=562 MB2017.11 | 22.97 | — | |
| Classifier OnlyBackbone=DenseNet-1212018.01 | 43.19 | — | |
| Classifier OnlyBackbone=ResNet-502018.01 | 47.2 | — | |
| Classifier OnlyBackbone=VGG-16 BN2018.01 | 51.62 | — | |
| Classifier OnlyBackbone=VGG-16, # Models=1, Size=562 MB2017.11 | 56.42 | — |