Semantic Segmentation on XSeg (test)
72.83mIoUAPSAM
Evaluation results
| Method | Links | ||
|---|---|---|---|
| APSAMBackbone=ViT-L/14, Trainable Params(M)=11.912026.04 | 72.83 | 82.31 | |
| Mask2formerBackbone=Swin-L, Trainable Params(M)=144.852026.04 | 69.59 | 81.44 | |
| SAMUSBackbone=ViT-L/14, Trainable Params(M)=43.21, Evaluation Protocol=Finetune2026.04 | 68.56 | 78.46 | |
| SDANetBackbone=ResNet101, Trainable Params(M)=103.202026.04 | 68.15 | 80.46 | |
| SegMANBackbone=SegMANEncoder-b, Trainable Params(M)=56.342026.04 | 67.97 | 77.09 | |
| SAMBackbone=ViT-L/14, Trainable Params(M)=10.06, Evaluation Protocol=Finetune2026.04 | 67.87 | 77.45 | |
| MaskDINOBackbone=ResNet101, Trainable Params(M)=62.902026.04 | 65.22 | 78.95 | |
| SegformerBackbone=MiT-b5, Trainable Params(M)=82.012026.04 | 63.67 | 77.8 | |
| SegmenterBackbone=ViT-B/16, Trainable Params(M)=104.452026.04 | 63.27 | 77.5 | |
| TwinsBackbone=PCPVT-L, Trainable Params(M)=64.722026.04 | 61.21 | 75.94 | |
| DeepLabV3+Backbone=ResNet101, Trainable Params(M)=60.212026.04 | 57.29 | 72.84 | |
| UperNetBackbone=Swin-Tiny, Trainable Params(M)=58.942026.04 | 55.16 | 71.1 | |
| PSPNetBackbone=ResNet101, Trainable Params(M)=65.592026.04 | 54.24 | 70.34 | |
| SAMBackbone=ViT-L/14, Trainable Params(M)=0, Evaluation Protocol=Frozen2026.04 | 53.82 | 64.99 |