| Benchmark | Model | Rank | Results |
|---|---|---|---|
| cross-modal-retrieval-on-rsicd | HarMA (w/ GeoRSCLIP) | #1 | Mean Recall: 38.95%Image-to-text R@1: 20.52%text-to-image R@1: 15.84% |
| cross-modal-retrieval-on-rsitmd | HarMA (w/ GeoRSCLIP) | #1 | Image-to-text R@1: 32.74%Mean Recall: 52.27%text-to-imageR@1: 25.62% |