GeoNUSAF TC-SegFormer -- random split, fold 1
Kathmandu Valley land-use segmentation, 6 classes, ignore_index=255, 512x512 @ 0.586 m/px.
| backbone | nvidia/segformer-b0-finetuned-ade-512-512 |
| split mode | random (sequence-block CV, export-order proxy) |
| fold | 1 of 3 |
| seed | 42 |
| best epoch | 101 |
| detail path | True |
| CSA reweighting | True (tau=[0.6, 0.35, 0.35, 0.6, 0.6, 0.6], w_min=0.25) |
| soft-clDice | True (mu=0.3) |
| balanced sampler | True |
Validation
mIoU 0.5650 | mF1 0.7119 | OA 0.8121 | kappa 0.6946
| class | IoU | UA (prec) | PA (rec) |
|---|---|---|---|
| Residential | 0.8437 | 0.9207 | 0.9099 |
| Road | 0.4261 | 0.5454 | 0.6607 |
| River | 0.5218 | 0.6511 | 0.7243 |
| Forest | 0.6332 | 0.8627 | 0.7042 |
| UnusedLand | 0.4211 | 0.6019 | 0.5836 |
| Agricultural | 0.5443 | 0.6637 | 0.7517 |
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support