Visual Question Answering on FloodNet (existence, counting, total metrics)
96.15Existence AccuracyNeFo
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| NeFoBase Model=Qwen3-VL, Evaluation Mode=Zero-shot transfer2026.06 | 96.15 | 40.1 | 63.74 | |
| TLMBase Model=Qwen3-VL, Evaluation Mode=Zero-shot transfer2026.06 | 95.69 | 38.09 | 62.58 | |
| SARBase Model=Qwen3-VL, Evaluation Mode=Zero-shot transfer2026.06 | 95.47 | 38.76 | 62.87 | |
| NeFoBase Model=RS-LLaVA, Evaluation Mode=Zero-shot transfer2026.06 | 95.24 | 22.32 | 53.33 | |
| Qwen3-VLEvaluation Mode=Zero-shot transfer2026.06 | 94.78 | 39.26 | 62.87 | |
| TLMBase Model=RS-LLaVA, Evaluation Mode=Zero-shot transfer2026.06 | 94.33 | 21.48 | 52.46 | |
| RS-LLaVAEvaluation Mode=Zero-shot transfer2026.06 | 94.1 | 20.13 | 52.36 | |
| TentBase Model=Qwen3-VL, Evaluation Mode=Zero-shot transfer2026.06 | 93.88 | 34.06 | 59.5 | |
| SARBase Model=RS-LLaVA, Evaluation Mode=Zero-shot transfer2026.06 | 93.65 | 20.3 | 51.49 | |
| TentBase Model=RS-LLaVA, Evaluation Mode=Zero-shot transfer2026.06 | 92.52 | 19.3 | 50.43 | |
| NeFoBase Model=Qwen2.5-VL, Evaluation Mode=Zero-shot transfer2026.06 | 82.54 | 38.93 | 57.47 | |
| NeFoBase Model=GeoReason, Evaluation Mode=Zero-shot transfer2026.06 | 79.6 | 34.4 | 53.62 | |
| SARBase Model=GeoReason, Evaluation Mode=Zero-shot transfer2026.06 | 78.46 | 28.19 | 49.57 | |
| TLMBase Model=GeoReason, Evaluation Mode=Zero-shot transfer2026.06 | 78.46 | 28.02 | 49.47 | |
| GeoReasonEvaluation Mode=Zero-shot transfer2026.06 | 77.1 | 30.54 | 50.34 | |
| TentBase Model=GeoReason, Evaluation Mode=Zero-shot transfer2026.06 | 70.75 | 23.49 | 43.59 | |
| Qwen2.5-VLEvaluation Mode=Zero-shot transfer2026.06 | 40.14 | 35.23 | 37.32 | |
| TentBase Model=Qwen2.5-VL, Evaluation Mode=Zero-shot transfer2026.06 | 38.1 | 35.74 | 36.74 | |
| TLMBase Model=Qwen2.5-VL, Evaluation Mode=Zero-shot transfer2026.06 | 35.37 | 34.06 | 34.62 | |
| SARBase Model=Qwen2.5-VL, Evaluation Mode=Zero-shot transfer2026.06 | 31.97 | 36.24 | 34.43 |