Remote Sensing Visual Grounding on DIOR-RSVG official (test)
0.8341Acc@0.5LQVG
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| LQVGModel Category=Specialized Models2026.01 | 0.8341 | — | — | |
| LPVAModel Category=Specialized Models2026.01 | 0.8227 | — | — | |
| AirSpatialBotModel Category=Large Vision-Language Models2026.01 | 0.7741 | — | — | |
| MGVLFModel Category=Specialized Models2026.01 | 0.7678 | — | — | |
| RSGround-R1Params=3B, Protocol=GRPO-based RFT (0.4 epoch)2026.01 | 0.7181 | 0.5871 | 0.6338 | |
| Qwen2.5-VL-3B w/ GRPOParams=3B, Protocol=GRPO-based RFT (0.4 epoch)2026.01 | 0.6665 | 0.5615 | 0.6018 | |
| Qwen2.5-VL-3B w/ SFTParams=3B, Protocol=SFT (1 epoch)2026.01 | 0.626 | 0.476 | 0.5571 | |
| VHM w/ SFTParams=7B, Protocol=SFT (1 epoch)2026.01 | 0.6167 | 0.4573 | 0.5485 | |
| VHMParams=7B, Protocol=Zero-shot2026.01 | 0.5215 | 0.2563 | 0.4601 | |
| GeoChat w/ SFTParams=7B, Protocol=SFT (1 epoch)2026.01 | 0.4648 | 0.133 | 0.4305 | |
| Qwen2-VLModel Category=Large Vision-Language Models2026.01 | 0.4425 | — | — | |
| Ours2025.12 | 0.4027 | — | — | |
| Qwen2.5-VL-7BParams=7B, Protocol=Zero-shot2026.01 | 0.3925 | 0.252 | 0.3866 | |
| MoE-LoRAtraining_strategy=Mixture-of-Experts LoRA2025.12 | 0.3786 | — | — | |
| Qwen2.5-VL-3BParams=3B, Protocol=Zero-shot2026.01 | 0.375 | 0.2625 | 0.3731 | |
| Full-FTtraining_strategy=Full Fine-Tuning2025.12 | 0.3631 | — | — | |
| LoRAtraining_strategy=Low-Rank Adaptation2025.12 | 0.3315 | — | — | |
| LHRS-Bot-NovaParams=8B, Protocol=Zero-shot2026.01 | 0.3287 | 0.1237 | 0.3161 | |
| Qwen-VLModel Category=Large Vision-Language Models2026.01 | 0.3222 | — | — | |
| MiniGPT-v2Params=7B, Protocol=Zero-shot2026.01 | 0.3025 | 0.1146 | 0.3008 | |
| GeoChatParams=7B, Protocol=Zero-shot2026.01 | 0.2595 | 0.0453 | 0.2987 | |
| Qwen-vl-Chat2025.12 | 0.2505 | — | — | |
| GeoChatModel Category=Large Vision-Language Models2026.01 | 0.2405 | — | — | |
| SkySenseGPTParams=7B, Protocol=Zero-shot2026.01 | 0.2286 | 0.0552 | 0.2776 | |
| LHRS-BotModel Category=Large Vision-Language Models2026.01 | 0.1759 | — | — | |
| LLaVA-1.5Params=7B, Protocol=Zero-shot2026.01 | 0.1607 | 0.0415 | 0.2315 | |
| InternVL2-40BModel Category=Large Vision-Language Models2026.01 | 0.1506 | — | — | |
| InternVL2-8BModel Category=Large Vision-Language Models2026.01 | 0.1442 | — | — | |
| LHRS-BotParams=7B, Protocol=Zero-shot2026.01 | 0.1249 | 0.0302 | 0.1791 | |
| LLaVA-1.52025.12 | 0.0952 | — | — |