Object recognition on N-ImageNet 1.0 (test)
75.2Top-1 AccuracyGEP
Evaluation results
| Method | Links | ||
|---|---|---|---|
| GEPBackbone=ViT-B/16, Pre-training Dataset=Event-1.8M, Ep.=242026.03 | 75.2 | 92.9 | |
| GEP (Sup. Align)Backbone=ViT-S/16, Pre-training Dataset=N-ImageNet, Ep.=62026.03 | 69.41 | 90.2 | |
| STPBackbone=Swin-T, Pre-training Dataset=N-ImageNet, Ep.=N/A2026.03 | 68.87 | 89.65 | |
| GEPBackbone=ViT-S/16, Pre-training Dataset=Event-1.8M, Ep.=242026.03 | 65.11 | 87.36 | |
| ECDPBackbone=ViT-S/16, Pre-training Dataset=N-ImageNet, Ep.=3002026.03 | 64.83 | 86.3 | |
| DINOv2Backbone=ViT-S/16, Pre-training Dataset=LVD-142M, Ep.=N/A2026.03 | 60.8 | 83.97 | |
| MEMBackbone=ViT-S/16, Pre-training Dataset=N-ImageNet, Ep.=752026.03 | 57.89 | — | |
| EventBindBackbone=ViT-B/16, Pre-training Dataset=N-ImageNet, Ep.=N/A2026.03 | 51.4 | — | |
| MAEBackbone=ViT-B/16, Pre-training Dataset=ImageNet-1K, Ep.=8002026.03 | 51.25 | 72.64 | |
| ESTBackbone=N/A, Pre-training Dataset=N/A, Ep.=N/A2026.03 | 48.93 | — | |
| BeiTBackbone=ViT-B/16, Pre-training Dataset=ImageNet-1K, Ep.=8002026.03 | 47.15 | 69.27 | |
| ViTBackbone=ViT-S/16, Pre-training Dataset=N-ImageNet, Ep.=3002026.03 | 46.7 | 69.89 | |
| MoCo-v3Backbone=ViT-S/16, Pre-training Dataset=ImageNet-1K, Ep.=3002026.03 | 45.77 | 68.89 |