Image Classification on ImageNet 1k (real)
88.6Top-1 AccuracyOurs (DINOv2)
Evaluation results
| Method | Links | |
|---|---|---|
| Ours (DINOv2)Data=INet-22k, Arch.=L/142024.06 | 88.6 | |
| ResNet-50 (A1)Training Procedure=A1, Number of epochs=600, Training resolution=224 x 224, Training time=110h, Peak memory by GPU (MB)=22,095, Numbers of GPU=42021.10 | 85.7 | |
| ResNet-50 (A2)Training Procedure=A2, Number of epochs=300, Training resolution=224 x 224, Training time=55h, Peak memory by GPU (MB)=22,095, Numbers of GPU=42021.10 | 85.4 | |
| ResNet-50 (A3)Training Procedure=A3, Number of epochs=100, Training resolution=160 x 160, Training time=15h, Peak memory by GPU (MB)=11,390, Numbers of GPU=42021.10 | 84.5 | |
| I-JEPAData=INet-1k, Arch.=H/142024.06 | 83.8 | |
| MAEData=INet-1k, Arch.=H/142024.06 | 83.7 | |
| I-JEPAData=INet-22k, Arch.=g/162024.06 | 78.6 | |
| BEITData=INet-1k, Arch.=L/162024.06 | 78.3 | |
| AIMData=DFN-2B+, Arch.=H/142024.06 | 64.4 |