Image Classification on ImageNet 1k (val) (Specialized Top-1 Accuracy Variants)
80.2Top-1 Acc (1% labels)DPT
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| DPTArchitecture=ViT-Huge2023.02 | 80.2 | — | — | — | |
| Semi-ViTArchitecture=ViT-Huge2023.02 | 80 | — | — | — | |
| Semi-ViT (baseline)Architecture=ViT-Huge2023.02 | 79.4 | — | — | — | |
| SimCLRv2Architecture=ResNet-1522023.02 | 76.6 | — | — | — | |
| MSNArchitecture=ViT-B/42023.02 | 75.7 | 54.3 | 64.6 | 72.4 | |
| MSNArchitecture=ViT-L/72023.02 | 75.1 | 57.1 | 66.4 | 72.1 | |
| BYOLArchitecture=ResNet-2002023.02 | 71.2 | — | — | — | |
| PAWSArchitecture=ResNet-502023.02 | 66.5 | — | — | — | |
| EMANArchitecture=ResNet-502023.02 | 63 | — | — | — | |
| DINOArchitecture=ViT-B/82023.02 | — | 45.8 | 55.9 | 64.6 | |
| DPTArchitecture=ViT-B/42023.02 | — | 58.6 | 69.5 | 74.4 | |
| DPTArchitecture=ViT-L/72023.02 | — | 58.9 | 69.2 | 73.4 | |
| iBOTArchitecture=ViT-B/162023.02 | — | 46.1 | 56.2 | 64.7 | |
| MAEArchitecture=ViT-H/142023.02 | — | 11.6 | 18.6 | 32.8 | |
| MSN (baseline)Architecture=ViT-B/42023.02 | — | 52.9 | 64.9 | 72.4 | |
| MSN (baseline)Architecture=ViT-L/72023.02 | — | 56.2 | 66.5 | 72 |