Semantic Localization on AIR-SLT
75.95R_sutGeoRSCLIP
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| GeoRSCLIPParadigm=GVLM + FT, Tuned on=RS5M, Backbone=ViT-H-142023.06 | 75.95 | 25.66 | 74.18 | 74.94 | |
| GeoRSCLIPParadigm=GVLM + FT, Tuned on=RS5M, Backbone=ViT-B-322023.06 | 75.46 | 26.1 | 71.8 | 74 | |
| GeoRSCLIPParadigm=GVLM + UniPELT, Tuned on=RS5M, Backbone=ViT-B-322023.06 | 74.89 | 25.5 | 70.21 | 73.58 | |
| GeoRSCLIPParadigm=GVLM + LORA, Tuned on=RS5M, Backbone=ViT-B-322023.06 | 74.61 | 26.42 | 66.36 | 72.18 | |
| GeoRSCLIPParadigm=GVLM + Prefixtuning, Tuned on=RS5M, Backbone=ViT-B-322023.06 | 74.4 | 25.51 | 70.13 | 73.36 | |
| GeoRSCLIPParadigm=GVLM + Pfeiffer, Tuned on=RS5M, Backbone=ViT-B-322023.06 | 74.02 | 25.41 | 69.48 | 73.08 | |
| SeLo-v2Paradigm=Supervised, Tuned on=RSITMD, Backbone=ViT-B-322023.06 | 71.99 | 29.25 | 66.58 | 70.21 | |
| CLIP-BaselineParadigm=GVLM, Tuned on=-, Backbone=ViT-B-322023.06 | 71.88 | 30.06 | 69.92 | 70.71 | |
| SeLo-v1Paradigm=Supervised, Tuned on=RSITMD, Backbone=ViT-B-322023.06 | 69.2 | 33.23 | 66.67 | 67.72 |