Image Classification on ImageNet-1k (Acc@1 and FLOPs)
75.46Top-1 AccuracyViT-Tiny
Evaluation results
| Method | Links | ||
|---|---|---|---|
| ViT-TinyBackbone=ViT-Tiny2026.07 | 75.46 | 0 | |
| Fusion (BALANCED)Backbone=ViT-Tiny, Profile=BALANCED2026.07 | 71.51 | 31.3 | |
| DynamicViT-TinyBackbone=ViT-Tiny2026.07 | 71.15 | 37 | |
| Fusion (POWER_SAVE)Backbone=ViT-Tiny, Profile=POWER_SAVE2026.07 | 70.27 | 41.2 | |
| Fusion parallel (interference ref.)Backbone=ViT-Tiny, Variant=parallel-composition2026.07 | 51.44 | 46.9 |