Fine-grained classification on Oxford-IIIT Pets base classes
31Mean Class Error (MCE)CoOp
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CoOpPrompt-tuning strategy=CoOp, Calibration Method=None2026.02 | 31 | 62 | |
| TCPT (Ours)Prompt-tuning strategy=MaPLe, Calibration Method=TCPT2026.02 | 34 | 80 | |
| MaPLePrompt-tuning strategy=MaPLe, Calibration Method=None2026.02 | 60 | 119 | |
| TCPT (Ours)Prompt-tuning strategy=CoOp, Calibration Method=TCPT2026.02 | 101 | 210 | |
| ZS-NormPrompt-tuning strategy=MaPLe, Calibration Method=ZS-Norm2026.02 | 101 | 321 | |
| TCPT (Ours)Prompt-tuning strategy=KGCoOp, Calibration Method=TCPT2026.02 | 105 | 283 | |
| ZS-NormPrompt-tuning strategy=KGCoOp, Calibration Method=ZS-Norm2026.02 | 113 | 306 | |
| KGCoOpPrompt-tuning strategy=KGCoOp, Calibration Method=None2026.02 | 117 | 295 | |
| PenaltyPrompt-tuning strategy=KGCoOp, Calibration Method=Penalty2026.02 | 140 | 461 | |
| PenaltyPrompt-tuning strategy=CoOp, Calibration Method=Penalty2026.02 | 185 | 641 | |
| PenaltyPrompt-tuning strategy=MaPLe, Calibration Method=Penalty2026.02 | 185 | 629 | |
| ZS-NormPrompt-tuning strategy=CoOp, Calibration Method=ZS-Norm2026.02 | 217 | 770 |