Image Classification on ImageNet-1K (Top-1 Acc, Throughput, FLOPs)
85.7Top-1 AccBaseline
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| BaselineModel=ViT-L/224 (MAE)2025.12 | 85.7 | — | — | |
| HyNAPModel=ViT-L/224 (MAE)2025.12 | 85.7 | — | — | |
| DiffRateModel=ViT-L/224 (MAE)2025.12 | 85.6 | — | — | |
| ToMeModel=ViT-L/224 (MAE)2025.12 | 85.4 | — | — | |
| EViTModel=ViT-L/224 (MAE)2025.12 | 85.1 | — | — | |
| BaselineModel=ViT-B/224 (MAE)2025.12 | 83.6 | — | — | |
| BaselineModel=DeiT-B/2242025.12 | 81.8 | — | — | |
| HyNAPModel=ViT-B/224 (MAE)2025.12 | 80.1 | — | — | |
| DiffRateModel=ViT-B/224 (MAE)2025.12 | 79.9 | — | — | |
| BaselineModel=DeiT-S/2242025.12 | 79.8 | — | — | |
| DiffRateModel=DeiT-S/2242025.12 | 79.6 | — | — | |
| HyNAPModel=DeiT-S/2242025.12 | 79.6 | — | — | |
| ATSModel=DeiT-B/2242025.12 | 79.6 | — | — | |
| Zero-TPModel=DeiT-S/2242025.12 | 79.4 | — | — | |
| HyNAPModel=DeiT-B/2242025.12 | 79.3 | — | — | |
| ATSModel=DeiT-S/2242025.12 | 79.2 | — | — | |
| GTP-ViTModel=DeiT-S/2242025.12 | 79.1 | — | — | |
| DiffRateModel=DeiT-B/2242025.12 | 79 | — | — | |
| ToMeModel=ViT-B/224 (MAE)2025.12 | 78.9 | — | — | |
| LoTMModel=DeiT-S/2242025.12 | 78.8 | — | — | |
| ToMeModel=DeiT-S/2242025.12 | 78.7 | — | — | |
| EViTModel=DeiT-S/2242025.12 | 78.5 | — | — | |
| GTP-ViTModel=DeiT-B/2242025.12 | 78.3 | — | — | |
| ToMeModel=DeiT-B/2242025.12 | 77.5 | — | — | |
| Evo-ViTModel=DeiT-S/2242025.12 | 77.4 | — | — | |
| EViTModel=DeiT-B/2242025.12 | 75.1 | — | — | |
| EViTModel=ViT-B/224 (MAE)2025.12 | 75.1 | — | — | |
| DyViTModel=DeiT-S/2242025.12 | 67.4 | — | — | |
| Evo-ViTModel=DeiT-B/2242025.12 | 60.6 | — | — | |
| DyViTModel=DeiT-B/2242025.12 | 50.2 | — | — |