Base-to-novel generalization on StanfordCars (test)
81.33Base AccuracyMMRL++
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MMRL++2025.05 | 81.33 | 75.27 | 78.18 | |
| MMRL2025.05 | 81.3 | 75.07 | 78.06 | |
| PromptSRCMode=Prompt learning2023.06 | 78.27 | 74.97 | 76.58 | |
| CoOpMode=Prompt learning2023.06 | 78.12 | 60.4 | 68.13 | |
| CoOp2025.05 | 78.12 | 60.4 | 68.13 | |
| ProGradMode=Prompt learning2023.06 | 77.68 | 68.63 | 72.88 | |
| CoPromptMode=Collective prompt learning2023.06 | 76.97 | 74.4 | 75.66 | |
| MaPLeMode=Multi-modal prompt learning2023.06 | 72.94 | 74 | 73.47 | |
| KgCoOpMode=Knowledge-guided prompt learning2023.06 | 71.76 | 75.04 | 73.36 | |
| Co-CoOpMode=Conditional prompt learning2023.06 | 70.49 | 73.59 | 72.01 | |
| CLIPFine-tuned=false2023.06 | 63.37 | 74.89 | 68.65 | |
| CLIP2025.05 | 63.37 | 74.89 | 68.65 |