Object Classification on PASCAL VOC 2007 (mAP k=1 to k=16)
54.3mAP@1Supervised
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| Supervisedarchitecture=ResNet-50, pretrained_dataset=ImageNet-1M, pretrained_epochs=2002020.05 | 54.3 | 67.8 | 73.9 | 79.6 | 82.3 | |
| PCL v2architecture=ResNet-50-MLP, pretrained_dataset=ImageNet-1M, pretrained_epochs=2002020.05 | 47.9 | 59.6 | 66.2 | 74.5 | 78.3 | |
| PCLarchitecture=ResNet-50, pretrained_dataset=ImageNet-1M, pretrained_epochs=2002020.05 | 46.9 | 56.4 | 62.8 | 70.2 | 74.3 | |
| MoCo v2architecture=ResNet-50-MLP, pretrained_dataset=ImageNet-1M, pretrained_epochs=2002020.05 | 46.3 | 58.3 | 64.9 | 72.5 | 76.1 | |
| SimCLRarchitecture=ResNet-50-MLP, pretrained_dataset=ImageNet-1M, pretrained_epochs=2002020.05 | 32.7 | 43.1 | 52.5 | 61 | 67.1 | |
| MoCoarchitecture=ResNet-50, pretrained_dataset=ImageNet-1M, pretrained_epochs=2002020.05 | 31.4 | 42 | 49.5 | 60 | 65.9 | |
| Jigsawarchitecture=ResNet-50, pretrained_dataset=ImageNet-14M2020.05 | 26.5 | 31.1 | 40 | 46.7 | 51.8 | |
| Randomarchitecture=ResNet-50, pretrained_dataset=ImageNet-1M, pretrained_epochs=2002020.05 | 8 | 8.2 | 8.2 | 8.2 | 8.5 |