Semantic Segmentation on Pascal Semantic Segmentation OOD Corrupted (test)
0.7174mIoU (Fog)Adam-SPD
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| Adam-SPDBackbone=SWIN-Tiny, Pre-trained=ImageNet21K2024.11 | 0.7174 | 0.5341 | 0.4417 | 0.7092 | 0.6006 | 0.365 | |
| FTPBackbone=SWIN-Tiny, Pre-trained=ImageNet21K2024.11 | 0.711 | 0.5263 | 0.4025 | 0.6981 | 0.5845 | 0.3283 | |
| L2-SPBackbone=SWIN-Tiny, Pre-trained=ImageNet21K2024.11 | 0.6987 | 0.492 | 0.391 | 0.6861 | 0.567 | 0.2885 | |
| TPGMBackbone=SWIN-Tiny, Pre-trained=ImageNet21K2024.11 | 0.6951 | 0.5088 | 0.3862 | 0.6882 | 0.5696 | 0.2944 | |
| AdapterBackbone=SWIN-Tiny, Pre-trained=ImageNet21K2024.11 | 0.6936 | 0.5094 | 0.3743 | 0.6826 | 0.565 | 0.284 | |
| LLRDBackbone=SWIN-Tiny, Pre-trained=ImageNet21K2024.11 | 0.6813 | 0.4618 | 0.3728 | 0.663 | 0.5447 | 0.2379 | |
| BitFitBackbone=SWIN-Tiny, Pre-trained=ImageNet21K2024.11 | 0.67 | 0.4639 | 0.3061 | 0.6622 | 0.5256 | 0.1944 | |
| MARS-SPBackbone=SWIN-Tiny, Pre-trained=ImageNet21K2024.11 | 0.5697 | 0.3729 | 0.2182 | 0.5827 | 0.4359 | -0.0094 | |
| Vanilla FTBackbone=SWIN-Tiny, Pre-trained=ImageNet21K2024.11 | 0.5672 | 0.3804 | 0.2321 | 0.5803 | 0.44 | 0 |