SeMask (SeMask Swin-L Mask2Former)

Reported on 6 benchmarks across 2 tasks · 1 paper

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Medical3 results

Semantic SegmentationonCityscapes val
mIoU· 2021-12-23
84.98
best: 90.3 (EfficientPS (Cityscapes-fine))
SeMask: Semantically Masked Transformers for Semantic Segmentation arXiv:2112.12782
Semantic SegmentationonADE20K val
mIoU· 2021-12-23
57.5
best: 62.8 (BEiT-3)
SeMask: Semantically Masked Transformers for Semantic Segmentation arXiv:2112.12782
Semantic SegmentationonADE20K
Validation mIoU· 2021-12-23
57.5
best: 63.6 (ViT-P (InternImage-H))
SeMask: Semantically Masked Transformers for Semantic Segmentation arXiv:2112.12782

Audio3 results

10-shot image generationonCityscapes val
mIoU· 2021-12-23
84.98
best: 90.3 (EfficientPS (Cityscapes-fine))
SeMask: Semantically Masked Transformers for Semantic Segmentation arXiv:2112.12782
10-shot image generationonADE20K val
mIoU· 2021-12-23
57.5
best: 62.8 (BEiT-3)
SeMask: Semantically Masked Transformers for Semantic Segmentation arXiv:2112.12782
10-shot image generationonADE20K
Validation mIoU· 2021-12-23
57.5
best: 63.6 (ViT-P (InternImage-H))
SeMask: Semantically Masked Transformers for Semantic Segmentation arXiv:2112.12782