Image Classification on ImageNet (val) (Top-1 Accuracy and FLOPs)
0.1Training FLOPsSET
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| SETBackbone=ResNet-50, sparsity=0.92021.02 | 0.1 | — | — | 0.1 | |
| SETBackbone=ResNet-50, sparsity=0.82021.02 | 0.23 | — | — | 0.23 | |
| StaticBackbone=ResNet-50, sparsity=0.92021.02 | 0.24 | — | 0.1 | 0.24 | |
| RigLBackbone=ResNet-50, sparsity=0.92021.02 | 0.25 | — | — | 0.24 | |
| RigL-ITOPBackbone=ResNet-50, sparsity=0.92021.02 | 0.25 | — | 0.83 | 0.24 | |
| StaticBackbone=ResNet-50, sparsity=0.82021.02 | 0.42 | — | 0.2 | 0.42 | |
| RigLBackbone=ResNet-50, sparsity=0.82021.02 | 0.42 | — | — | 0.42 | |
| RigL-ITOPBackbone=ResNet-50, sparsity=0.82021.02 | 0.42 | — | 0.93 | 0.42 | |
| SNFSBackbone=ResNet-50, sparsity=0.92021.02 | 0.5 | — | — | 0.24 | |
| RigL-ITOP_2xBackbone=ResNet-50, sparsity=0.9, Training Time=Extended 2x2021.02 | 0.5 | — | 0.89 | 0.24 | |
| GMPBackbone=ResNet-50, sparsity=0.82021.02 | 0.51 | 75.6 | — | 0.1 | |
| GMPBackbone=ResNet-50, sparsity=0.92021.02 | 0.56 | 73.9 | — | 0.23 | |
| SNFSBackbone=ResNet-50, sparsity=0.82021.02 | 0.61 | — | — | 0.42 | |
| RigL-ITOP_2xBackbone=ResNet-50, sparsity=0.8, Training Time=Extended 2x2021.02 | 0.84 | — | 0.97 | 0.42 | |
| DenseBackbone=ResNet-50, sparsity=0.02021.02 | 13,200,000,000,000,000,000 | — | 1 | 18,200,000,000 |