Semantic Segmentation on PASCAL-Context (Clean)
73.79mIoUFTP
Evaluation results
| Method | Links | ||
|---|---|---|---|
| FTPBackbone=Swin-Tiny, Pre-training=ImageNet-21K, Decoder=Segformer, Inference Time (s/it)=0.4012023.10 | 73.79 | 11.76 | |
| L2-SPBackbone=Swin-Tiny, Pre-training=ImageNet-21K, Decoder=Segformer, Inference Time (s/it)=0.3472023.10 | 73.47 | 11.27 | |
| TPGMBackbone=Swin-Tiny, Pre-training=ImageNet-21K, Decoder=Segformer, Inference Time (s/it)=0.6112023.10 | 72.56 | 9.89 | |
| LLRDBackbone=Swin-Tiny, Pre-training=ImageNet-21K, Decoder=Segformer, Inference Time (s/it)=0.2892023.10 | 72.09 | 9.18 | |
| AdapterBackbone=Swin-Tiny, Pre-training=ImageNet-21K, Decoder=Segformer, Inference Time (s/it)=0.2332023.10 | 71.85 | 8.82 | |
| BitFitBackbone=Swin-Tiny, Pre-training=ImageNet-21K, Decoder=Segformer, Inference Time (s/it)=0.2482023.10 | 70.31 | 6.49 | |
| MARS-SPBackbone=Swin-Tiny, Pre-training=ImageNet-21K, Decoder=Segformer, Inference Time (s/it)=0.3182023.10 | 66.24 | 0.32 | |
| Vanilla FTBackbone=Swin-Tiny, Pre-training=ImageNet-21K, Decoder=Segformer, Inference Time (s/it)=0.2882023.10 | 66.03 | 0 |