Image-Text Retrieval on Flickr30K (R@1)
100R@1MDCS-SGA
Evaluation Results
| Method | Links | |
|---|---|---|
| MDCS-SGASource Model=ALBEF, Target Model=ALBEF, Attack Setting=White-box2026.02 | 100 | |
| MDCS-SAAETSource Model=TCL, Target Model=TCL, Attack Setting=White-box2026.02 | 100 | |
| MDCS-SAAETSource Model=CLIPViT, Target Model=CLIPViT, Attack Setting=White-box2026.02 | 100 | |
| MDCS-SAAETSource Model=CLIPCNN, Target Model=CLIPCNN, Attack Setting=White-box2026.02 | 100 | |
| MDCS-SAAETSource Model=ALBEF, Target Model=ALBEF, Attack Setting=White-box2026.02 | 100 | |
| SGASource Model=CLIPViT, Target Model=CLIPViT, Attack Setting=White-box2026.02 | 100 | |
| MDCS-DRASource Model=ALBEF, Target Model=ALBEF, Attack Setting=White-box2026.02 | 99.9 | |
| MDCS-SAAETSource Model=ALBEF, Target Model=ALBEF, Attack Setting=White-box2026.02 | 99.9 | |
| MDCS-SAAETSource Model=TCL, Target Model=ALBEF, Attack Setting=Black-box2026.02 | 99.17 | |
| MDCS-SAAETSource Model=ALBEF, Target Model=TCL, Attack Setting=Black-box2026.02 | 96.71 | |
| MDCS-SAAETSource Model=ALBEF, Target Model=TCL, Attack Setting=Black-box2026.02 | 96.52 | |
| MDCS-SGASource Model=ALBEF, Target Model=TCL, Attack Setting=Black-box2026.02 | 91.78 | |
| SETOKIMSize=7B, Visual Features=semantic-equivalent (SE), Visual Tokens=continuous (C), Observed during training=false2024.06 | 86.9 | |
| NEXT-GPTSize=7B, Visual Features=patch-level features (P), Visual Tokens=continuous (C), Observed during training=false2024.06 | 84.5 | |
| LaVITSize=7B, Visual Features=patch-level features (P), Visual Tokens=continuous (C), Observed during training=false2024.06 | 83 | |
| HQ-CLIPDataset (Data Size)=DataComp-DFN (130M), Samples Seen@Resolution=3.20B@224, GPU-hours=≈ 2675, Zero-shot=true2026.04 | 77.9 | |
| EmuSize=7B, Visual Features=learnable queries (Q), Visual Tokens=continuous (C), Observed during training=false2024.06 | 77.4 | |
| MetaCLIP-400MDataset (Data Size)=(400M), Samples Seen@Resolution=12.8B@224, GPU-hours=≈ 10700, Zero-shot=true2026.04 | 76.7 | |
| ChameleonSize=34B, Visual Features=patch-level features (P), Visual Tokens=discrete (D), Observed during training=false2024.06 | 74.7 | |
| OpenCLIPDataset (Data Size)=LAION-400M (400M), Samples Seen@Resolution=12.8B@224, GPU-hours=10736, Zero-shot=true2026.04 | 74.6 | |
| DynamiCSDataset (Data Size)=LAION-400M (298M), Samples Seen@Resolution=2.56B@112 + 128M@224, GPU-hours=299, Zero-shot=true2026.04 | 73.9 | |
| DynamiCSDataset (Data Size)=LAION-400M (298M), Samples Seen@Resolution=1.28B@112 + 128M@224, GPU-hours=163, Zero-shot=true2026.04 | 72.7 | |
| RECLIP*Dataset (Data Size)=LAION-400M (298M), Samples Seen@Resolution=2.56B@112 + 128M@224, GPU-hours=280, Zero-shot=true2026.04 | 72.4 | |
| OpenAI-WITDataset (Data Size)=-, Samples Seen@Resolution=12.8B@224, GPU-hours=10700, Zero-shot=true2026.04 | 72.2 | |
| CLIPA*Dataset (Data Size)=LAION-400M (298M), Samples Seen@Resolution=2.56B@112 + 128M@224, GPU-hours=269, Zero-shot=true2026.04 | 71.7 | |
| DynamiCSDataset (Data Size)=DataComp-DFN (130M), Samples Seen@Resolution=2.56B@112 + 128M@224, GPU-hours=299, Zero-shot=true2026.04 | 71.3 | |
| RandomDataset (Data Size)=LAION-400M (298M), Samples Seen@Resolution=1.28B@112 + 128M@224, GPU-hours=151, Zero-shot=true2026.04 | 69.4 | |
| DynamiCSDataset (Data Size)=DataComp-DFN (130M), Samples Seen@Resolution=1.28B@112 + 128M@224, GPU-hours=163, Zero-shot=true2026.04 | 69.3 | |
| DFN*Dataset (Data Size)=DataComp-DFN (130M), Samples Seen@Resolution=1.28B@112 + 128M@224, GPU-hours=151, Zero-shot=true2026.04 | 68.1 | |
| MDCS-SAAETSource Model=ALBEF, Target Model=CLIPCNN, Attack Setting=Black-box2026.02 | 67.89 | |
| MDCS-SAAETSource Model=ALBEF, Target Model=CLIPViT, Attack Setting=Black-box2026.02 | 67.01 | |
| DynamiCSDataset (Data Size)=DataComp-DFN (130M), Samples Seen@Resolution=0.64B@112 + 128M@224, GPU-hours=95, Zero-shot=true2026.04 | 66.6 | |
| DataCompDataset (Data Size)=DataComp-Large (1.28B), Samples Seen@Resolution=1.28B@224, GPU-hours=≈ 1070, Zero-shot=true2026.04 | 63.6 | |
| RandomDataset (Data Size)=DataComp-DFN (130M), Samples Seen@Resolution=0.64B@112 + 128M@224, GPU-hours=90, Zero-shot=true2026.04 | 62.3 | |
| MDCS-SAAETSource Model=ALBEF, Target Model=CLIPCNN, Attack Setting=Black-box2026.02 | 60.54 | |
| MDCS-SAAETSource Model=ALBEF, Target Model=CLIPViT, Attack Setting=Black-box2026.02 | 60.25 | |
| PGDSource Model=ALBEF, Target Model=ALBEF, Attack Setting=White-box2026.02 | 52.45 | |
| PGDSource Model=ALBEF, Target Model=CLIPCNN, Attack Setting=Black-box2026.02 | 10.34 | |
| PGDSource Model=ALBEF, Target Model=CLIPViT, Attack Setting=Black-box2026.02 | 8.96 | |
| PGDSource Model=ALBEF, Target Model=TCL, Attack Setting=Black-box2026.02 | 3.06 |