Field Segmentation on AI4SmallFarms Vietnam
42mIoUMAgSeg
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| MAgSegModel size=12B, Evaluation Protocol=Zero-shot2026.05 | 42 | 35 | 3.52 | 3.86 | |
| MAgSegEvaluation Mode=Zero-shot, Training Dataset=India, Model Scale=4B2026.05 | 41 | 37 | 10.33 | 25.63 | |
| GRESEvaluation Protocol=Zero-shot2026.05 | 40 | 31 | 3.67 | 10.6 | |
| MAgSegModel size=4B, Evaluation Protocol=Zero-shot2026.05 | 40 | 33 | 3.99 | 11.47 | |
| LISAtEvaluation Mode=Zero-shot, Training Dataset=India2026.05 | 36 | 30 | 9.05 | 48.41 | |
| GRESEvaluation Mode=Zero-shot, Training Dataset=India2026.05 | 33 | 22 | 12.57 | 47.19 | |
| LISAEvaluation Mode=Zero-shot, Training Dataset=India2026.05 | 30 | 19 | 24.1 | 50.48 | |
| GSVAEvaluation Mode=Zero-shot, Training Dataset=India2026.05 | 29 | 19 | 26.98 | 46.42 | |
| LISAtEvaluation Protocol=Zero-shot2026.05 | 24 | 13 | 5.51 | 34.77 | |
| GSVAEvaluation Protocol=Zero-shot2026.05 | 24 | 12 | 8.9 | 31.51 | |
| LISAEvaluation Protocol=Zero-shot2026.05 | 22 | 10 | 10.01 | 31.84 | |
| LAVTEvaluation Protocol=Zero-shot2026.05 | 19 | 7 | 9.3 | 6.28 | |
| LAVTEvaluation Mode=Zero-shot, Training Dataset=India2026.05 | 13 | 5 | 19.98 | 30.02 | |
| Text4SegVariant=R, Evaluation Protocol=Zero-shot2026.05 | 9 | 3 | 11.45 | 39.69 | |
| Text4SegVariant=C, Evaluation Protocol=Zero-shot2026.05 | 8 | 3 | 11.02 | 21.75 |