Base-to-new generalization on Average (PatternNet, RSICD, RESISC45, MLRSNet)
92.52Base ScoreAPPLeNet
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| APPLeNetTrainable Params=5M2026.01 | 92.52 | 53.67 | 67.84 | |
| TCPTrainable Params=0.3M2026.01 | 92.44 | 65.9 | 76.9 | |
| MaPLeTrainable Params=3M2026.01 | 92.37 | 62.67 | 74.51 | |
| BiMoRSTrainable Params=1M2026.01 | 91.82 | 67.62 | 77.85 | |
| CoOpTrainable Params=20482026.01 | 91.01 | 55.81 | 69.01 | |
| CoCoOpTrainable Params=0.035M2026.01 | 90.72 | 60.03 | 72.16 | |
| ProGradTrainable Params=81922026.01 | 88.03 | 64.95 | 74.7 | |
| CLIPTrainable Params=02026.01 | 67.32 | 62.5 | 64.69 |