Image Classification on WHU-RS19
99.8AccuracyCLAdapter
Evaluation Results
| Method | Links | |
|---|---|---|
| CLAdapterbackbone=ConvNeXt-B2025.12 | 99.8 | |
| SOTASetting=Transductive, Transductive Inference (TransCLIP)=false2025.06 | 99.3 | |
| CLAdapterbackbone=ViT-B2025.12 | 99.2 | |
| MS2AP2025.12 | 98.88 | |
| T-GeoRSCLIPSetting=Transductive, Transductive Inference (TransCLIP)=true2025.06 | 98.7 | |
| GeoMELTBackbone=BEiT-3, zero-shot=true2025.12 | 98.54 | |
| GeoVLM-R1Backbone=Qwen2.5VL-3B, zero-shot=true2025.12 | 97.91 | |
| T-RemoteCLIPSetting=Transductive, Transductive Inference (TransCLIP)=true2025.06 | 97.9 | |
| SF-MSFormer2025.12 | 97.8 | |
| T-SkyCLIP50Setting=Transductive, Transductive Inference (TransCLIP)=true2025.06 | 97.8 | |
| EMTCAL*2025.12 | 97.6 | |
| RSFJR2025.12 | 97.48 | |
| S-CLIPModel=ViT-B/162023.05 | 97.3 | |
| S-CLIPModel=ViT-B/322023.05 | 97.2 | |
| Swin2025.12 | 97.12 | |
| CLIPModel=ViT-B/16, Training Mode=fine-tune2023.05 | 96.7 | |
| ViT2025.12 | 96.42 | |
| EarthDialBackbone=Phi-3, zero-shot=true2025.12 | 96.21 | |
| RemoteCLIPBackbone=ViT-B-32, zero-shot=true2023.06 | 96.12 | |
| RemoteCLIPBackbone=ViT-B-32, zero-shot=true2023.06 | 96.12 | |
| Earth-Agent2025.09 | 96.12 | |
| S-CLIPModel=ResNet-502023.05 | 95.8 | |
| CLIPModel=ViT-B/32, Training Mode=fine-tune2023.05 | 95.8 | |
| VHM2025.09 | 95.8 | |
| RemoteCLIPARCH=ViT-B/32, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 95.3 | |
| RemoteCLIPSetting=Transductive, Transductive Inference (TransCLIP)=false2025.06 | 95.2 | |
| RemoteCLIPBackbone=ResNet-50, zero-shot=true2023.06 | 95.15 | |
| RemoteCLIPBackbone=ResNet-50, zero-shot=true2023.06 | 95.15 | |
| RemoteCLIPArchitecture=ViT-B/32, Setting=Zero-shot, Template=DST2026.06 | 95.1 | |
| RemoteCLIPBackbone=ViT-L-14, zero-shot=true2023.06 | 94.66 | |
| RemoteCLIPBackbone=ViT-L-14, zero-shot=true2023.06 | 94.66 | |
| RemoteCLIPBackbone=ViT-B, zero-shot=true2025.12 | 94.17 | |
| GeoRSCLIPArchitecture=ViT-L/14, Setting=Zero-shot, Template=DST2026.06 | 94.1 | |
| CLIPModel=ResNet-50, Training Mode=fine-tune2023.05 | 93.9 | |
| DCA-Fusion2025.12 | 93.56 | |
| LHRS-Bot2025.09 | 93.17 | |
| SkySenseGPTParams.=7B2026.05 | 93.16 | |
| SigLIP2Architecture=ViT-L/16, Setting=Zero-shot, Template=DST2026.06 | 93 | |
| GLM162025.12 | 92.99 | |
| RemoteCLIPARCH=RN-50, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 92.9 | |
| RemoteCLIPArchitecture=RN-50, Setting=Zero-shot, Template=DST2026.06 | 92.8 | |
| RemoteCLIPBackbone=ViT-L, zero-shot=true2025.12 | 92.72 | |
| VHMParams.=7B2026.05 | 91.84 | |
| SigLIP2ARCH=ViT-L/16, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 91.7 | |
| GeoRSCLIPArchitecture=ViT-B/32, Setting=Zero-shot, Template=DST2026.06 | 91.3 | |
| RemoteCLIPARCH=ViT-L/14, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 91.2 | |
| SkyCLIP50Setting=Transductive, Transductive Inference (TransCLIP)=false2025.06 | 91 | |
| MetaCLIP-2.5BARCH=ViT-L/14, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 90.9 | |
| SkyCLIPARCH=ViT-L/14, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 90.9 | |
| SigLIP2Architecture=ViT-B/16, Setting=Zero-shot, Template=DST2026.06 | 90.8 | |
| MetaCLIP-2.5BArchitecture=ViT-B/16, Setting=Zero-shot, Template=DST2026.06 | 90.7 | |
| SigLIPArchitecture=ViT-L/16, Setting=Zero-shot, Template=DST2026.06 | 90.7 | |
| GeoRSCLIPARCH=ViT-L/14, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 90.6 | |
| T-CLIPSetting=Transductive, Transductive Inference (TransCLIP)=true2025.06 | 90.3 | |
| RemoteCLIPArchitecture=ViT-L/14, Setting=Zero-shot, Template=DST2026.06 | 90.2 | |
| Gemini-2.02026.05 | 90 | |
| SkyNativeParams.=8B2026.05 | 90 | |
| MetaCLIP-400MArchitecture=ViT-B/16, Setting=Zero-shot, Template=DST2026.06 | 90 | |
| MetaCLIP-2.5BArchitecture=ViT-L/14, Setting=Zero-shot, Template=DST2026.06 | 90 | |
| GeoRSCLIPARCH=ViT-B/32, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 89.9 | |
| SigLIPARCH=ViT-L/16, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 89.8 | |
| MetaCLIP-400MARCH=ViT-L/14, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 89.5 | |
| CLIP LAION-RSARCH=ViT-L/14, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 89.5 | |
| MetaCLIP-2.5BARCH=ViT-B/16, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 89.4 | |
| MetaCLIP-400MArchitecture=ViT-L/14, Setting=Zero-shot, Template=DST2026.06 | 89.2 | |
| SkyCLIPArchitecture=ViT-L/14, Setting=Zero-shot, Template=DST2026.06 | 89.2 | |
| CLIP LAION-RSArchitecture=ViT-L/14, Setting=Zero-shot, Template=DST2026.06 | 89 | |
| GeoRSCLIPSetting=Transductive, Transductive Inference (TransCLIP)=false2025.06 | 88.8 | |
| CLIPARCH=ViT-L/14, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 88.2 | |
| GM2025.12 | 88.16 | |
| CLIPBackbone=ViT-L, zero-shot=true2025.12 | 87.86 | |
| CLIPBackbone=ViT-L-14, zero-shot=true2023.06 | 87.5 | |
| CLIPBackbone=ViT-L-14, zero-shot=true2023.06 | 87.5 | |
| SigLIP2ARCH=ViT-B/16, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 87.4 | |
| MetaCLIP-400MARCH=ViT-B/16, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 87.2 | |
| Geochat2025.09 | 86.47 | |
| CLIPARCH=ViT-B/16, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 86.3 | |
| CLIPArchitecture=ViT-B/16, Setting=Zero-shot, Template=DST2026.06 | 85.5 | |
| CLIPArchitecture=ViT-L/14, Setting=Zero-shot, Template=DST2026.06 | 85.5 | |
| SigLIPArchitecture=ViT-B/16, Setting=Zero-shot, Template=DST2026.06 | 85 | |
| Qwen2.5-VLParams.=7B2026.05 | 84.74 | |
| SigLIPARCH=ViT-B/16, Zero-shot=true, Approach=Meta-prompting (LLaMA-3.1)2026.06 | 83.3 | |
| CLIPSetting=Transductive, Transductive Inference (TransCLIP)=false2025.06 | 81.1 | |
| CLIPBackbone=ViT-B-32, zero-shot=true2023.06 | 80.61 | |
| CLIPBackbone=ViT-B-32, zero-shot=true2023.06 | 80.61 | |
| GeoChatBackbone=LLaVA-1.5-7B, zero-shot=true2025.12 | 80.09 | |
| Cross-Scale MAEEvaluation protocol=Average KNN, Scale ratios=12.5%, 25%, 50%, 100%2024.01 | 79.8 | |
| Scale-MAEEvaluation protocol=Average KNN, Scale ratios=12.5%, 25%, 50%, 100%2024.01 | 79.5 | |
| VoCoTParams.=7B2026.05 | 76.84 | |
| Claude-42026.05 | 76.32 | |
| LLaVA-1.52025.09 | 74.52 | |
| CLIPBackbone=ViT-B, zero-shot=true2025.12 | 73.79 | |
| EVEv2Params.=7B2026.05 | 73.42 | |
| EarthDialParams.=4B2026.05 | 73.42 | |
| SatMAEEvaluation protocol=Average KNN, Scale ratios=12.5%, 25%, 50%, 100%2024.01 | 69.9 | |
| CLIPBackbone=ResNet-50, zero-shot=true2023.06 | 69.42 | |
| CLIPBackbone=ResNet-50, zero-shot=true2023.06 | 69.42 | |
| MiniGPT-v22025.09 | 64.8 | |
| Domain-Aware Reinforcement Fine-TuningBase Model=QwenVL, Fine-tuning Strategy=Reinforcement Fine-Tuning2026.01 | 44.4 | |
| QwenVLBase Model=QwenVL2026.01 | 29.3 |