Image Classification on ImageNet (Acc. and Rob.)
65.7AccuracyMTA
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MTACategory=Tuning-free, Backbone=CLIP-ViT-B/32, Attack scenario=PGD-100 (eps=4.0)2026.06 | 65.7 | 24.2 | |
| R-TPTCategory=Tuning-based, Backbone=CLIP-ViT-B/32, Attack scenario=PGD-100 (eps=4.0)2026.06 | 64.8 | 37.1 | |
| CLIPCategory=Baseline, Backbone=CLIP-ViT-B/32, Attack scenario=PGD-100 (eps=4.0)2026.06 | 62.2 | 0 | |
| MACCategory=Tuning-free, Backbone=CLIP-ViT-B/32, Attack scenario=PGD-100 (eps=4.0)2026.06 | 61.7 | 47.6 | |
| TTCCategory=Tuning-free, Backbone=CLIP-ViT-B/32, Attack scenario=PGD-100 (eps=4.0)2026.06 | 51.8 | 4.7 |