Image Classification on DTD (Base, Novel, HM)
85.67Base AccuracyMMRL
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MMRL2026.05 | 85.67 | 65 | 73.82 | |
| MMRL + NeRPRefining Method=NeRP2026.05 | 85.67 | 65.7 | 74.37 | |
| 2SFSBackbone=ViT-B/16, Shots=162025.03 | 84.6 | 65.01 | 73.52 | |
| CLIP-LoRABackbone=ViT-B/16, Shots=162025.03 | 83.95 | 62.84 | 71.39 | |
| ProVP2026.03 | 83.95 | 59.06 | 69.34 | |
| 2SFSVisual Backbone=ViT-L/14, Shots (k)=82025.03 | 83.91 | 70.01 | 76.33 | |
| SkipT2026.05 | 83.77 | 67.23 | 74.59 | |
| PromptSRCSetting=Base-to-Novel Generalization2025.12 | 83.37 | 62.97 | 71.75 | |
| RMAdapterSetting=Base-to-Novel Generalization2025.12 | 83.37 | 66.73 | 74.09 | |
| PromptSRC2026.03 | 83.37 | 62.97 | 71.75 | |
| PromptSRC2026.05 | 83.37 | 62.97 | 71.75 | |
| PromptSRC + NeRPRefining Method=NeRP2026.05 | 83.37 | 63.37 | 72.01 | |
| MMABackbone=ViT-B/16, Shots=162025.03 | 83.2 | 65.63 | 73.38 | |
| MMASetting=Base-to-Novel Generalization2025.12 | 83.2 | 65.63 | 73.38 | |
| MMA2026.03 | 83.2 | 65.63 | 73.38 | |
| MMA2026.05 | 83.2 | 65.63 | 73.38 | |
| MMA + NeRPRefining Method=NeRP2026.05 | 83.2 | 66.2 | 73.73 | |
| CoPromptSetting=Base-to-Novel Generalization2025.12 | 83.13 | 64.73 | 72.79 | |
| CoPrompt2026.05 | 83.13 | 64.73 | 72.79 | |
| MetaPrompt2026.03 | 83.1 | 58.05 | 68.35 | |
| EvoPrompt2026.03 | 83.1 | 64.1 | 72.37 | |
| TCPSetting=Base-to-Novel Generalization2025.12 | 82.77 | 58.07 | 68.25 | |
| TCP2026.03 | 82.77 | 58.07 | 68.25 | |
| ProDA2026.03 | 80.67 | 56.48 | 66.44 | |
| MaPLeBackbone=ViT-B/16, Shots=162025.03 | 80.36 | 59.18 | 68.16 | |
| MaPLeSetting=Base-to-Novel Generalization2025.12 | 80.36 | 59.18 | 68.16 | |
| MaPLe2026.03 | 80.36 | 59.18 | 68.16 | |
| MaPLe2026.05 | 80.36 | 59.18 | 68.16 | |
| 2SFSBackbone=ViT-B/16, Shots=8, Evaluation Protocol=base-to-novel generalization2025.03 | 80.17 | 64.45 | 71.46 | |
| CoOpBackbone=ViT-B/16, Shots=162025.03 | 79.44 | 41.18 | 54.24 | |
| CoOpSetting=Base-to-Novel Generalization2025.12 | 79.44 | 41.18 | 54.24 | |
| CoOp2026.03 | 79.44 | 41.18 | 54.24 | |
| CoOp2026.05 | 79.44 | 41.18 | 54.24 | |
| MMAVisual Backbone=ViT-L/14, Shots (k)=82025.03 | 78.13 | 69.9 | 73.79 | |
| KgCoOpBackbone=ViT-B/16, Shots=162025.03 | 77.55 | 54.99 | 64.35 | |
| KgCoOpSetting=Base-to-Novel Generalization2025.12 | 77.55 | 54.99 | 64.35 | |
| KgCoOp2026.03 | 77.55 | 54.99 | 64.35 | |
| ProGradBackbone=ViT-B/16, Shots=162025.03 | 77.35 | 52.35 | 62.45 | |
| MMABackbone=ViT-B/16, Shots=8, Evaluation Protocol=base-to-novel generalization2025.03 | 77.07 | 64.47 | 70.21 | |
| CoCoOpBackbone=ViT-B/16, Shots=162025.03 | 77.01 | 56 | 64.85 | |
| Co-CoOpSetting=Base-to-Novel Generalization2025.12 | 77.01 | 56 | 64.85 | |
| CoCoOp2026.03 | 77.01 | 56 | 64.85 | |
| CoCoOp2026.05 | 77.01 | 56 | 64.85 | |
| CoCoOp + NeRPRefining Method=NeRP2026.05 | 77.01 | 56.67 | 65.29 | |
| 2SFSBackbone=ViT-B/32, Shots=82025.03 | 75.85 | 55.23 | 63.92 | |
| MMABackbone=ViT-B/32, Shots=82025.03 | 73.7 | 56.07 | 63.69 | |
| CLIPVisual Backbone=ViT-L/14, Shots (k)=82025.03 | 59.14 | 67.87 | 63.21 | |
| CLIPBackbone=ViT-B/32, Shots=82025.03 | 54.17 | 58.21 | 56.12 | |
| CLIPBackbone=ViT-B/16, zero-shot=true2025.03 | 53.24 | 59.9 | 56.37 | |
| CLIPBackbone=ViT-B/16, Shots=8, Evaluation Protocol=base-to-novel generalization2025.03 | 53.24 | 59.9 | 56.37 | |
| CLIPSetting=Base-to-Novel Generalization2025.12 | 53.24 | 59.9 | 56.37 | |
| CLIP2026.03 | 53.24 | 59.9 | 56.37 |