Lung Segmentation on CXR Lung Segmentation (test)
95.63DSCSAM Ada-LoRA + D-QAT
Evaluation Results
| Method | Links | |
|---|---|---|
| SAM Ada-LoRA + D-QATTotal (M)=89.7, Trainable (M)=5.4, Reduction=16.6×2026.04 | 95.63 | |
| SAM Ada-LoRATotal (M)=89.7, Trainable (M)=5.4, Reduction=16.6×2026.04 | 95.6 | |
| SAM Ada-LoRA + Full QATTotal (M)=89.7, Trainable (M)=5.4, Reduction=16.6×2026.04 | 95.59 | |
| SAM Decoder FTTotal (M)=89.7, Trainable (M)=3.8, Reduction=23.6×2026.04 | 95.55 | |
| DeepLabV3+Total (M)=41.0, Trainable (M)=41.02026.04 | 95.36 | |
| nnU-NetTotal (M)=30.0, Trainable (M)=30.02026.04 | 94.85 | |
| SegFormerTotal (M)=84.6, Trainable (M)=84.62026.04 | 92.28 |