Image Classification on ImageNet (val) (Accuracy, Params, and FLOPs)
76.31Top-1 AccDECORE-8
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| DECORE-8Backbone=ResNet-50, lambda=82021.06 | 76.31 | 93.02 | — | — | |
| ResNet 50Backbone=ResNet-50, Role=Baseline2021.06 | 76.15 | 92.87 | — | — | |
| DECORE-6Backbone=ResNet-50, lambda=62021.06 | 74.58 | 92.18 | — | — | |
| SSS-32Backbone=ResNet-502021.06 | 74.18 | 91.91 | — | — | |
| He et al.Backbone=ResNet-502021.06 | 72.3 | 90.8 | — | — | |
| DECORE-5Backbone=ResNet-50, lambda=52021.06 | 72.06 | 90.82 | — | — | |
| ThiNet-70Backbone=ResNet-502021.06 | 72.04 | 90.67 | — | — | |
| HRank-1Backbone=ResNet-502021.06 | 71.98 | 91.01 | — | — | |
| GAL-0.5Backbone=ResNet-502021.06 | 71.95 | 90.94 | — | — | |
| SSS-26Backbone=ResNet-502021.06 | 71.82 | 90.79 | — | — | |
| GAL-0.5-jointBackbone=ResNet-50, joint=true2021.06 | 71.8 | 90.82 | — | — | |
| GDP-0.6Backbone=ResNet-502021.06 | 71.19 | 90.71 | — | — | |
| ThiNet-50Backbone=ResNet-502021.06 | 70.01 | 90.02 | — | — | |
| GAL-1Backbone=ResNet-502021.06 | 69.88 | 89.75 | — | — | |
| DECORE-4Backbone=ResNet-50, lambda=42021.06 | 69.71 | 89.37 | — | — | |
| GDP-0.5Backbone=ResNet-502021.06 | 69.58 | 90.14 | — | — | |
| GAL-1-jointBackbone=ResNet-50, joint=true2021.06 | 69.31 | 89.12 | — | — | |
| HRank-2Backbone=ResNet-502021.06 | 69.1 | 89.58 | — | — | |
| ThiNet-30Backbone=ResNet-502021.06 | 68.42 | 88.3 | — | — |