DiNAT-L (Mask2Former, single-scale)

Reported on 4 benchmarks across 1 task · 1 paper · 2 SOTA

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

Computer Vision4 results

Instance SegmentationonADE20K val
AP· 2022-09-29
35.4
best: 44.2 (OneFormer (InternImage-H, emb_dim=1024, single-scale, 896x896, COCO-Pretrained))
SOTA
Dilated Neighborhood Attention Transformer arXiv:2209.15001
Instance SegmentationonADE20K val
APL· 2022-09-29
55.5
best: 64.3 (OneFormer (InternImage-H, emb_dim=1024, single-scale, 896x896, COCO-Pretrained))
SOTA
Dilated Neighborhood Attention Transformer arXiv:2209.15001
Instance SegmentationonADE20K val
APM· 2022-09-29
39
best: 49.9 (OneFormer (InternImage-H, emb_dim=1024, single-scale, 896x896, COCO-Pretrained))
Dilated Neighborhood Attention Transformer arXiv:2209.15001
Instance SegmentationonADE20K val
APS· 2022-09-29
16.3
best: 23.7 (OneFormer (InternImage-H, emb_dim=1024, single-scale, 896x896, COCO-Pretrained))
Dilated Neighborhood Attention Transformer arXiv:2209.15001