Image Classification on Pets (Linear Accuracy)
0.809Linear AccuracyBYOL
Evaluation results
| Method | Links | |
|---|---|---|
| BYOLBackbone=ResNet-50, Width=1.0x, Params=22.56M, MACS=4.11G, Evaluation Protocol=Linear evaluation2023.03 | 0.809 | |
| US3LBackbone=ResNet-50, Width=1.0x, Params=22.56M, MACS=4.11G, Evaluation Protocol=Linear evaluation2023.03 | 0.794 | |
| US3LBackbone=ResNet-50, Width=0.75x, Params=14.77M, MACS=2.34G, Evaluation Protocol=Linear evaluation2023.03 | 0.78 | |
| US3LBackbone=ResNet-50, Width=0.5x, Params=6.92M, MACS=1.06G, Evaluation Protocol=Linear evaluation2023.03 | 0.768 | |
| BYOLBackbone=ResNet-50, Width=0.5x, Params=6.92M, MACS=1.06G, Evaluation Protocol=Linear evaluation2023.03 | 0.75 | |
| BYOLBackbone=ResNet-50, Width=0.75x, Params=14.77M, MACS=2.34G, Evaluation Protocol=Linear evaluation2023.03 | 0.742 | |
| US3LBackbone=ResNet-50, Width=0.25x, Params=1.99M, MACS=0.28G, Evaluation Protocol=Linear evaluation2023.03 | 0.74 | |
| BYOLBackbone=ResNet-50, Width=0.25x, Params=1.99M, MACS=0.28G, Evaluation Protocol=Linear evaluation2023.03 | 0.647 |