Image Classification on CIFAR-100 (Acc. (%) and Acc./#ReLU)
78.23Accuracy (%)Baseline
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Baseline#ReLUs (K)=491.52, Backbone=ResNet-182025.12 | 78.23 | 0.16 | |
| DeepShare#ReLUs (K)=49.9, Backbone=ResNet-182025.12 | 77.98 | 1.56 | |
| DeepShare#ReLUs (K)=12.9, Backbone=ResNet-182025.12 | 77.29 | 5.99 | |
| HybReNet#ReLUs (K)=52, Backbone=ResNet-182025.12 | 77.14 | 1.48 | |
| DeepShare#ReLUs (K)=6, Backbone=ResNet-182025.12 | 76.13 | 12.69 | |
| Seesaw#ReLUs (K)=78, Backbone=ResNet-182025.12 | 75.89 | 0.97 | |
| Spatial sharing#ReLUs (K)=50, Backbone=ResNet-182025.12 | 75.51 | 1.51 | |
| AutoRep#ReLUs (K)=50, Backbone=ResNet-182025.12 | 75.48 | 1.51 | |
| SENet#ReLUs (K)=49.6, Backbone=ResNet-182025.12 | 75.28 | 1.51 | |
| AutoRep#ReLUs (K)=12.9, Backbone=ResNet-182025.12 | 74.92 | 5.8 | |
| AutoRep#ReLUs (K)=6, Backbone=ResNet-182025.12 | 73.79 | 12.29 | |
| SNL#ReLUs (K)=49.9, Backbone=ResNet-182025.12 | 73.75 | 1.48 | |
| Spatial sharing#ReLUs (K)=30, Backbone=ResNet-182025.12 | 73.03 | 2.43 | |
| HybReNet#ReLUs (K)=26, Backbone=ResNet-182025.12 | 72.65 | 2.79 | |
| Seesaw#ReLUs (K)=36, Backbone=ResNet-182025.12 | 72.32 | 2 | |
| SENet#ReLUs (K)=24.6, Backbone=ResNet-182025.12 | 70.59 | 2.87 | |
| Sphynx#ReLUs (K)=51.2, Backbone=ResNet-182025.12 | 69.57 | 1.36 | |
| CryptoNAS#ReLUs (K)=100, Backbone=ResNet-182025.12 | 68.5 | 0.68 | |
| SNL#ReLUs (K)=12.9, Backbone=ResNet-182025.12 | 66.53 | 5.15 |