Few-shot Image Classification on Average 11 datasets (test)
87.41Average Accuracy (Few-shot)CAPT
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| CAPTShot=16, Backbone=ViT-B/16 CLIP2026.03 | 87.41 | — | — | — | — | — | |
| PromptKDShot=16, Backbone=ViT-B/16 CLIP2026.03 | 86.96 | — | — | — | — | — | |
| 2SFSShot=16, Backbone=ViT-B/16 CLIP2026.03 | 85.5 | — | — | — | — | — | |
| CAPTShot=8, Backbone=ViT-B/16 CLIP2026.03 | 81.91 | — | — | — | — | — | |
| SigLIP2Classifier=LDA, Embedding Projection=PCA<-, Backbone=ViT-B2026.03 | 81 | — | — | — | — | — | |
| 2SFSShot=8, Backbone=ViT-B/16 CLIP2026.03 | 80.7 | — | — | — | — | — | |
| SigLIP2Classifier=LDA, Embedding Projection=Original, Backbone=ViT-B2026.03 | 80.5 | — | — | — | — | — | |
| PromptKDShot=8, Backbone=ViT-B/16 CLIP2026.03 | 79.56 | — | — | — | — | — | |
| SigLIPClassifier=LDA, Embedding Projection=PCA<-, Backbone=ViT-B2026.03 | 79.4 | — | — | — | — | — | |
| SigLIP2Classifier=Prototype, Embedding Projection=PCA<-, Backbone=ViT-B2026.03 | 79.4 | — | — | — | — | — | |
| CoOpShot=16, Backbone=ViT-B/16 CLIP2026.03 | 79.01 | — | — | — | — | — | |
| SigLIPClassifier=LDA, Embedding Projection=Original, Backbone=ViT-B2026.03 | 79 | — | — | — | — | — | |
| CAPTShot=4, Backbone=ViT-B/16 CLIP2026.03 | 78.69 | — | — | — | — | — | |
| MaPLeShot=16, Backbone=ViT-B/16 CLIP2026.03 | 78.55 | — | — | — | — | — | |
| 2SFSShot=4, Backbone=ViT-B/16 CLIP2026.03 | 78.5 | — | — | — | — | — | |
| SigLIPClassifier=Prototype, Embedding Projection=PCA<-, Backbone=ViT-B2026.03 | 78.5 | — | — | — | — | — | |
| PromptKDShot=4, Backbone=ViT-B/16 CLIP2026.03 | 78.46 | — | — | — | — | — | |
| SeMoBridge-TParam.=0.77 M, Avg. Time=27 s, Backbone=ViT-B/162025.09 | 78.15 | — | — | — | — | — | |
| SeMoBridge-T w/o CSBParam.=0.26 M, Avg. Time=22 s, Backbone=ViT-B/162025.09 | 78.14 | — | — | — | — | — | |
| PromptSRCParam.=0.05 M, Avg. Time=1 h 42 min, Backbone=ViT-B/162025.09 | 77.9 | — | — | — | — | — | |
| APE-TParam.=0.51 M, Avg. Time=3 min 30 s, Backbone=ViT-B/162025.09 | 77.18 | — | — | — | — | — | |
| LDCParam.=69 M, Avg. Time=2 min, Backbone=ViT-B/162025.09 | 77.17 | — | — | — | — | — | |
| SigLIP2Classifier=Prototype, Embedding Projection=Original, Backbone=ViT-B2026.03 | 77 | — | — | — | — | — | |
| CoOpShot=8, Backbone=ViT-B/16 CLIP2026.03 | 76.45 | — | — | — | — | — | |
| SigLIPClassifier=Prototype, Embedding Projection=Original, Backbone=ViT-B2026.03 | 76.3 | — | — | — | — | — | |
| Tip-Adapter-FParam.=16.3 M, Avg. Time=4 min, Backbone=ViT-B/162025.09 | 75.9 | — | — | — | — | — | |
| CAPTShot=2, Backbone=ViT-B/16 CLIP2026.03 | 75.83 | — | — | — | — | — | |
| DINOv2Classifier=LDA, Embedding Projection=Original, Backbone=ViT-B2026.03 | 75.3 | — | — | — | — | — | |
| PromptKDShot=2, Backbone=ViT-B/16 CLIP2026.03 | 75.19 | — | — | — | — | — | |
| MaPLeShot=8, Backbone=ViT-B/16 CLIP2026.03 | 73.82 | — | — | — | — | — | |
| CoOpShot=4, Backbone=ViT-B/16 CLIP2026.03 | 73.59 | — | — | — | — | — | |
| 2SFSShot=2, Backbone=ViT-B/16 CLIP2026.03 | 73.28 | — | — | — | — | — | |
| CAPTShot=1, Backbone=ViT-B/16 CLIP2026.03 | 73.11 | — | — | — | — | — | |
| 2SFSShot=1, Backbone=ViT-B/16 CLIP2026.03 | 72.55 | — | — | — | — | — | |
| PromptKDShot=1, Backbone=ViT-B/16 CLIP2026.03 | 72.47 | — | — | — | — | — | |
| DINOv2Classifier=Prototype, Embedding Projection=Original, Backbone=ViT-B2026.03 | 71.8 | — | — | — | — | — | |
| CLIPClassifier=LDA, Embedding Projection=PCA<-, Backbone=ViT-B2026.03 | 70.8 | — | — | — | — | — | |
| MaPLeShot=4, Backbone=ViT-B/16 CLIP2026.03 | 70.66 | — | — | — | — | — | |
| CoOpShot=2, Backbone=ViT-B/16 CLIP2026.03 | 70.13 | — | — | — | — | — | |
| CLIPClassifier=LDA, Embedding Projection=Original, Backbone=ViT-B2026.03 | 69.8 | — | — | — | — | — | |
| CLIPClassifier=Prototype, Embedding Projection=PCA<-, Backbone=ViT-B2026.03 | 69.7 | — | — | — | — | — | |
| CLIP-AdapterParam.=0.52 M, Avg. Time=32 min, Backbone=ViT-B/162025.09 | 69.45 | — | — | — | — | — | |
| CoOpShot=1, Backbone=ViT-B/16 CLIP2026.03 | 68.09 | — | — | — | — | — | |
| MaPLeShot=2, Backbone=ViT-B/16 CLIP2026.03 | 65.28 | — | — | — | — | — | |
| CoOpParam.=0.01 M, Avg. Time=10 h 0 min, Backbone=ViT-B/162025.09 | 63.9 | — | — | — | — | — | |
| CLIPClassifier=Prototype, Embedding Projection=Original, Backbone=ViT-B2026.03 | 63.8 | — | — | — | — | — | |
| MaPLeShot=1, Backbone=ViT-B/16 CLIP2026.03 | 61.79 | — | — | — | — | — | |
| CoCoOp2025.03 | — | 66.79 | 67.65 | 71.21 | 72.96 | 74.9 | |
| CoCoOp2025.05 | — | 66.79 | 67.65 | 71.21 | 72.96 | 74.9 | |
| CoOp2025.03 | — | 67.56 | 70.65 | 74.02 | 76.98 | 79.89 | |
| CoOp2025.05 | — | 67.56 | 70.65 | 74.02 | 76.98 | 79.89 | |
| Linear probe CLIP2025.03 | — | 45.83 | 57.98 | 68.01 | 74.47 | 78.79 | |
| Linear probe CLIP2025.05 | — | 45.83 | 57.98 | 68.01 | 74.47 | 78.79 | |
| MaPLe2025.03 | — | 69.27 | 72.58 | 75.37 | 78.89 | 81.79 | |
| MaPLe2025.05 | — | 69.27 | 72.58 | 75.37 | 78.89 | 81.79 | |
| MMA2025.03 | — | 69.28 | 72.08 | 76.38 | 79.57 | 82.76 | |
| MMA2025.05 | — | 69.28 | 72.08 | 76.38 | 79.57 | 82.76 | |
| MMRL2025.03 | — | 72.67 | 75.9 | 79.2 | 81.47 | 84.34 | |
| MMRL2025.05 | — | 72.67 | 75.9 | 79.2 | 81.47 | 84.34 | |
| MMRL++2025.05 | — | 72.66 | 75.25 | 79 | 81.33 | 84.05 | |
| PromptSRC2025.03 | — | 72.32 | 75.29 | 78.35 | 80.69 | 82.87 | |
| PromptSRC2025.05 | — | 72.32 | 75.29 | 78.35 | 80.69 | 82.87 |