Image Classification on ImageNet (val) (P@1 and R@5)
77.21P@1FRN layer
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FRN layerBackbone=ResnetV2-50, Batch Size=2562019.11 | 77.21 | 93.57 | |
| BatchnormBackbone=ResnetV2-50, Batch Size=2562019.11 | 76.21 | 92.98 | |
| BatchRenormBackbone=ResnetV2-50, Batch Size=2562019.11 | 75.85 | 92.9 | |
| GroupnormBackbone=ResnetV2-50, Batch Size=2562019.11 | 75.67 | 92.7 | |
| LayernormBackbone=ResnetV2-50, Batch Size=2562019.11 | 72.75 | 91.19 | |
| InstancenormBackbone=ResnetV2-50, Batch Size=2562019.11 | 71.63 | 90.46 |