Image Classification on ImageNet 1k (val) (Top-1 Accuracy & Latency)
76.1Top-1 AccOriginal
Evaluation results
| Method | Links | ||
|---|---|---|---|
| OriginalNetwork=ResNet-50, Pruning=None2022.02 | 76.1 | 83.8 | |
| PRACTISENetwork=ResNet-50, Pruning=block2022.02 | 74.8 | 66.2 | |
| ADI*Network=ResNet-50, Pruning=filter2022.02 | 73.5 | 79.9 | |
| ADINetwork=ResNet-50, Pruning=filter2022.02 | 73.3 | — | |
| DINetwork=ResNet-50, Pruning=filter2022.02 | 72 | — | |
| OriginalNetwork=MobileNetV2, Pruning=None2022.02 | 71.9 | 37.6 | |
| MixMixNetwork=ResNet-50, Pruning=filter2022.02 | 69.8 | — | |
| PRACTISENetwork=MobileNetV2, Pruning=block2022.02 | 68 | 30.4 | |
| ADI*Network=MobileNetV2, Pruning=filter2022.02 | 62.8 | 30.8 | |
| MixMixNetwork=MobileNetV2, Pruning=filter2022.02 | 42.5 | — | |
| DINetwork=MobileNetV2, Pruning=filter2022.02 | 15.3 | — |