Image Retrieval on Flickr30k (test)
100R@1DeBias-Attack
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| DeBias-AttackSurrogate=ALBEF, Target=ALBEF2026.06 | 100 | — | — | — | — | — | — | — | — | |
| SA-AETSurrogate=TCL, Target=TCL2026.06 | 100 | — | — | — | — | — | — | — | — | |
| DeBias-AttackSurrogate=TCL, Target=TCL2026.06 | 100 | — | — | — | — | — | — | — | — | |
| SGASurrogate=CLIPViT, Target=CLIPViT2026.06 | 100 | — | — | — | — | — | — | — | — | |
| DRASurrogate=CLIPViT, Target=CLIPViT2026.06 | 100 | — | — | — | — | — | — | — | — | |
| TMMSurrogate=CLIPViT, Target=CLIPViT2026.06 | 100 | — | — | — | — | — | — | — | — | |
| SA-AETSurrogate=CLIPViT, Target=CLIPViT2026.06 | 100 | — | — | — | — | — | — | — | — | |
| DeBias-AttackSurrogate=CLIPViT, Target=CLIPViT2026.06 | 100 | — | — | — | — | — | — | — | — | |
| DeBias-AttackSurrogate=CLIPCNN, Target=CLIPCNN2026.06 | 100 | — | — | — | — | — | — | — | — | |
| DRASurrogate=TCL, Target=TCL2026.06 | 99.98 | — | — | — | — | — | — | — | — | |
| SA-AETSurrogate=ALBEF, Target=ALBEF2026.06 | 99.97 | — | — | — | — | — | — | — | — | |
| GLEAMSurrogate=ALBEF, Target=ALBEF2026.06 | 99.97 | — | — | — | — | — | — | — | — | |
| SGASurrogate=TCL, Target=TCL2026.06 | 99.96 | — | — | — | — | — | — | — | — | |
| SGASurrogate=ALBEF, Target=ALBEF2026.06 | 99.95 | — | — | — | — | — | — | — | — | |
| SA-AETSurrogate=CLIPCNN, Target=CLIPCNN2026.06 | 99.95 | — | — | — | — | — | — | — | — | |
| GLEAMSurrogate=CLIPViT, Target=CLIPViT2026.06 | 99.94 | — | — | — | — | — | — | — | — | |
| DRASurrogate=ALBEF, Target=ALBEF2026.06 | 99.92 | — | — | — | — | — | — | — | — | |
| GLEAMSurrogate=TCL, Target=TCL2026.06 | 99.9 | — | — | — | — | — | — | — | — | |
| DRASurrogate=CLIPCNN, Target=CLIPCNN2026.06 | 99.89 | — | — | — | — | — | — | — | — | |
| SGASurrogate=CLIPCNN, Target=CLIPCNN2026.06 | 99.88 | — | — | — | — | — | — | — | — | |
| GLEAMSurrogate=CLIPCNN, Target=CLIPCNN2026.06 | 99.81 | — | — | — | — | — | — | — | — | |
| GLEAMSurrogate=TCL, Target=ALBEF2026.06 | 98.46 | — | — | — | — | — | — | — | — | |
| SA-AETSurrogate=TCL, Target=ALBEF2026.06 | 98.41 | — | — | — | — | — | — | — | — | |
| Co-AttackSurrogate=ALBEF, Target=ALBEF2026.06 | 98.12 | — | — | — | — | — | — | — | — | |
| TMMSurrogate=TCL, Target=TCL2026.06 | 97.6 | — | — | — | — | — | — | — | — | |
| TMMSurrogate=ALBEF, Target=ALBEF2026.06 | 97.51 | — | — | — | — | — | — | — | — | |
| TMMSurrogate=CLIPCNN, Target=CLIPCNN2026.06 | 97.34 | — | — | — | — | — | — | — | — | |
| DeBias-AttackSurrogate=ALBEF, Target=TCL2026.06 | 97.33 | — | — | — | — | — | — | — | — | |
| Co-AttackSurrogate=CLIPCNN, Target=CLIPCNN2026.06 | 96.21 | — | — | — | — | — | — | — | — | |
| GLEAMSurrogate=ALBEF, Target=TCL2026.06 | 95.91 | — | — | — | — | — | — | — | — | |
| SA-AETSurrogate=ALBEF, Target=TCL2026.06 | 95.79 | — | — | — | — | — | — | — | — | |
| DRASurrogate=TCL, Target=ALBEF2026.06 | 95.36 | — | — | — | — | — | — | — | — | |
| Co-AttackSurrogate=TCL, Target=TCL2026.06 | 95.1 | — | — | — | — | — | — | — | — | |
| DeBias-AttackSurrogate=TCL, Target=ALBEF2026.06 | 94.6 | — | — | — | — | — | — | — | — | |
| PGDSurrogate=ALBEF, Target=ALBEF2026.06 | 93.88 | — | — | — | — | — | — | — | — | |
| SGASurrogate=TCL, Target=ALBEF2026.06 | 93.02 | — | — | — | — | — | — | — | — | |
| PGDSurrogate=CLIPCNN, Target=CLIPCNN2026.06 | 92.87 | — | — | — | — | — | — | — | — | |
| DRASurrogate=ALBEF, Target=TCL2026.06 | 90.96 | — | — | — | — | — | — | — | — | |
| PGDSurrogate=TCL, Target=TCL2026.06 | 89.45 | — | — | — | — | — | — | — | — | |
| SGASurrogate=ALBEF, Target=TCL2026.06 | 88.56 | — | — | — | — | — | — | — | — | |
| Co-AttackSurrogate=CLIPViT, Target=CLIPViT2026.06 | 87.93 | — | — | — | — | — | — | — | — | |
| ALBEFPre-training Images=14M, Evaluation Protocol=Finetuning2021.11 | 85.6 | 97.5 | 98.9 | — | — | — | — | — | — | |
| ALBEF(14M)Pre-training Data Scale=> 10M images, Evaluation Protocol=fine-tuned2022.11 | 85.6 | 97.5 | 98.9 | — | — | — | — | — | — | |
| ALIGNPre-training Data Scale=> 10M images, Evaluation Protocol=fine-tuned2022.11 | 84.9 | 97.4 | 98.6 | — | — | — | — | — | — | |
| SCLPre-training Data Scale=< 10M images, Evaluation Protocol=fine-tuned2022.11 | 84.56 | 97.42 | 98.94 | — | — | — | — | — | — | |
| PGDSurrogate=CLIPViT, Target=CLIPViT2026.06 | 84.32 | — | — | — | — | — | — | — | — | |
| DeBias-AttackSurrogate=CLIPViT, Target=CLIPCNN2026.06 | 84.12 | — | — | — | — | — | — | — | — | |
| TCLPre-training Data Scale=< 10M images, Evaluation Protocol=fine-tuned2022.11 | 84 | 96.7 | 98.5 | — | — | — | — | — | — | |
| UniMoCoZero-shot=true, Base Model=Qwen2-VL-7B2025.05 | 83.1 | 96.2 | 98.1 | — | — | — | — | — | — | |
| ALBEFPre-training Images=4M, Evaluation Protocol=Finetuning2021.11 | 82.8 | 96.7 | 98.4 | — | — | — | — | — | — | |
| ALBEF(4M)Pre-training Data Scale=< 10M images, Evaluation Protocol=fine-tuned2022.11 | 82.8 | 96.7 | 98.4 | — | — | — | — | — | — | |
| ALBEFSupervision Level=Supervised, Pre-training Data=Large-Scale Paired Image-Text Data2023.05 | 82.8 | 96.7 | 98.4 | — | — | — | — | — | — | |
| METER-CLIP-ViT-BasePre-training Images=4M, Backbone=CLIP-ViT-Base, Evaluation Protocol=Finetuning2021.11 | 82.22 | 96.34 | 98.36 | — | — | — | — | — | — | |
| METERPre-training Data Scale=< 10M images, Evaluation Protocol=fine-tuned2022.11 | 82.22 | 96.34 | 98.36 | — | — | — | — | — | — | |
| METER-CLIP-VITBASESupervision Level=Supervised, Pre-training Data=Large-Scale Paired Image-Text Data2023.05 | 82.2 | 96.3 | 98.3 | — | — | — | — | — | — | |
| ALBEF BaseParams=419, TFLOPS=7.412023.10 | 80.6 | 95.8 | 98 | — | — | — | — | — | — | |
| GLEAMSurrogate=CLIPViT, Target=CLIPCNN2026.06 | 80.43 | — | — | — | — | — | — | — | — | |
| VLM2VECZero-shot=true2025.05 | 80.3 | 95 | 97.4 | — | — | — | — | — | — | |
| ALBEF LoRAParams=431, TFLOPS=7.492023.10 | 80.2 | 95.6 | 97.7 | — | — | — | — | — | — | |
| PELAParams=173, TFLOPS=2.582023.10 | 79.7 | 94.8 | 97.5 | — | — | — | — | — | — | |
| Open CLIP-G/14zero-shot=true, backbone=ViT-G/142023.03 | 79.5 | 95 | 97.1 | — | — | — | — | — | — | |
| METER-Swin-BasePre-training Images=4M, Backbone=Swin-Base, Evaluation Protocol=Finetuning2021.11 | 79.02 | 95.58 | 98.04 | — | — | — | — | — | — | |
| EVA-01-CLIP-g/14+zero-shot=true, backbone=ViT-g/14+2023.03 | 78.9 | 94.5 | 96.9 | — | — | — | — | — | — | |
| EVA-02-CLIP-E/14zero-shot=true, backbone=ViT-E/142023.03 | 78.8 | 94.6 | 97 | — | — | — | — | — | — | |
| EVA-02-CLIP-E/14+zero-shot=true, backbone=ViT-E/14+2023.03 | 78.8 | 94.2 | 96.8 | — | — | — | — | — | — | |
| UNIMO-LargePre-training Images=<10M, Evaluation Protocol=Finetuning2021.11 | 78.04 | 94.24 | 97.12 | — | — | — | — | — | — | |
| EVA-02-CLIP-L/14+zero-shot=true, backbone=ViT-L/14+2023.03 | 77.9 | 94.2 | 96.8 | — | — | — | — | — | — | |
| Open CLIP-H/14zero-shot=true, backbone=ViT-H/142023.03 | 77.8 | 94.1 | 96.6 | — | — | — | — | — | — | |
| Open CLIP-g/14zero-shot=true, backbone=ViT-g/142023.03 | 77.7 | 94.1 | 96.9 | — | — | — | — | — | — | |
| EVA-02-CLIP-L/14zero-shot=true, backbone=ViT-L/142023.03 | 77.3 | 93.6 | 96.8 | — | — | — | — | — | — | |
| ERNIE-ViLEvaluation Protocol=Finetune2021.02 | 76.7 | 93.6 | 96.4 | — | — | — | — | — | — | |
| DeBias-AttackSurrogate=TCL, Target=CLIPCNN2026.06 | 76.47 | — | — | — | — | — | — | — | — | |
| VILLAEvaluation Protocol=Finetune2021.02 | 76.3 | 94.2 | 96.8 | — | — | — | — | — | — | |
| VILLAParams=1102023.10 | 76.3 | 94.2 | 96.8 | — | — | — | — | — | — | |
| VILLA-large2020.06 | 76.26 | 94.24 | 96.84 | — | — | — | — | — | — | |
| VILLA-LargePre-training Images=<10M, Evaluation Protocol=Finetuning2021.11 | 76.26 | 94.24 | 96.84 | — | — | — | — | — | — | |
| UniterEvaluation Protocol=Finetune2021.02 | 75.6 | 94.1 | 96.8 | — | — | — | — | — | — | |
| UNITERParams=110, TFLOPS=0.372023.10 | 75.6 | 94.1 | 96.8 | — | — | — | — | — | — | |
| UNITER-large2020.06 | 75.56 | 94.08 | 96.76 | — | — | — | — | — | — | |
| UNITER-LargePre-training Images=<10M, Evaluation Protocol=Finetuning2021.11 | 75.56 | 94.08 | 96.76 | — | — | — | — | — | — | |
| UNITERPre-training Data Scale=< 10M images, Evaluation Protocol=fine-tuned2022.11 | 75.56 | 94.08 | 96.76 | — | — | — | — | — | — | |
| DeBias-AttackSurrogate=CLIPCNN, Target=CLIPViT2026.06 | 75.52 | — | — | — | — | — | — | — | — | |
| Open CLIP-L/14zero-shot=true, backbone=ViT-L/142023.03 | 75 | 92.5 | 95.6 | — | — | — | — | — | — | |
| DeBias-AttackSurrogate=ALBEF, Target=CLIPCNN2026.06 | 74.96 | — | — | — | — | — | — | — | — | |
| VILLA-base2020.06 | 74.74 | 92.86 | 95.82 | — | — | — | — | — | — | |
| ERNIE-ViLParams=210M, Architecture=Two stream2021.06 | 74.44 | 92.72 | 95.94 | — | — | — | — | — | — | |
| SA-AETSurrogate=CLIPViT, Target=CLIPCNN2026.06 | 74.24 | — | — | — | — | — | — | — | — | |
| E2E-VLPParams=94M, Architecture=End2End2021.06 | 73.58 | 92.42 | 96.03 | — | — | — | — | — | — | |
| Visual ParsingPre-training Images=<10M, Evaluation Protocol=Finetuning2021.11 | 73.5 | 93.1 | 96.4 | — | — | — | — | — | — | |
| DeBias-AttackSurrogate=ALBEF, Target=CLIPViT2026.06 | 73.26 | — | — | — | — | — | — | — | — | |
| DeBias-AttackSurrogate=TCL, Target=CLIPViT2026.06 | 72.65 | — | — | — | — | — | — | — | — | |
| EVA-01-CLIP-g/14zero-shot=true, backbone=ViT-g/142023.03 | 72.6 | 91.6 | 95.1 | — | — | — | — | — | — | |
| UNITER-base2020.06 | 72.52 | 92.36 | 96.08 | — | — | — | — | — | — | |
| UNITERParams=110M, Architecture=Single stream2021.06 | 72.52 | 92.36 | 96.08 | — | — | — | — | — | — | |
| SOHOPre-training Data Scale=< 10M images, Evaluation Protocol=fine-tuned2022.11 | 72.5 | 92.7 | 96.1 | — | — | — | — | — | — | |
| UNITERSupervision Level=Supervised, Pre-training Data=Large-Scale Paired Image-Text Data2023.05 | 72.5 | 92.4 | 96.1 | — | — | — | — | — | — | |
| TMMSurrogate=TCL, Target=ALBEF2026.06 | 72.3 | — | — | — | — | — | — | — | — | |
| Unicoder-VL2020.06 | 71.5 | 90.9 | 94.9 | — | — | — | — | — | — | |
| Unicoder-VLEvaluation Protocol=Finetune2021.02 | 71.5 | 90.9 | 94.9 | — | — | — | — | — | — | |
| Unicoder-VLParams=110M, Architecture=Single stream2021.06 | 71.5 | 90.9 | 94.9 | — | — | — | — | — | — |