Fine-grained Image Classification on UCF101 (test)
64.4AccuracyTAPT
Evaluation results
| Method | Links | ||
|---|---|---|---|
| TAPTCategory=Tuning-based, Backbone=CLIP-ViT-B/32, Attack Protocol=PGD-100 (ϵ = 4.0)2026.06 | 64.4 | 6.4 | |
| R-TPTCategory=Tuning-based, Backbone=CLIP-ViT-B/32, Attack Protocol=PGD-100 (ϵ = 4.0)2026.06 | 62.8 | 42.4 | |
| MTACategory=Tuning-free, Backbone=CLIP-ViT-B/32, Attack Protocol=PGD-100 (ϵ = 4.0)2026.06 | 62.8 | 32.8 | |
| MACCategory=Tuning-free, Backbone=CLIP-ViT-B/32, Attack Protocol=PGD-100 (ϵ = 4.0)2026.06 | 62.3 | 48.2 | |
| TTCCategory=Tuning-free, Backbone=CLIP-ViT-B/32, Attack Protocol=PGD-100 (ϵ = 4.0)2026.06 | 62 | 6.9 | |
| CLIPCategory=Baseline, Backbone=CLIP-ViT-B/32, Attack Protocol=PGD-100 (ϵ = 4.0)2026.06 | 61.6 | 0 | |
| TPTBackbone=CLIP-ResNet50, Epsilon=1.0, Augmentation views=152026.06 | 59.5 | 0.3 | |
| CLIPBackbone=CLIP-ResNet50, Epsilon=1.02026.06 | 58.9 | 0.1 | |
| SS-TPTBackbone=CLIP-ResNet50, Epsilon=1.0, Augmentation views=152026.06 | 58.1 | 49.1 | |
| R-TPTBackbone=CLIP-ResNet50, Epsilon=1.0, Augmentation views=152026.06 | 57.1 | 40.5 | |
| EnsembleBackbone=CLIP-ResNet50, Epsilon=1.0, Augmentation views=152026.06 | 54 | 35 | |
| DOCBackbone=CLIP-ResNet50, Epsilon=1.0, Effective batch size=12026.06 | 53.2 | 21 | |
| TTCBackbone=CLIP-ResNet50, Epsilon=1.0, Effective batch size=12026.06 | 52.7 | 21.6 |