Focal-L (HTC++, multi-scale)

Reported on 6 benchmarks across 1 task · 1 paper

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

Computer Vision6 results

Instance SegmentationonCOCO minival
mask AP· 2021-07-01
50.9
best: 56.6 (Co-DETR)
Focal Self-attention for Local-Global Interactions in Vision Transformers arXiv:2107.00641
Instance SegmentationonCOCO test-dev
AP50· 2021-07-01
75.4
best: 80.8 (InternImage-H)
Focal Self-attention for Local-Global Interactions in Vision Transformers arXiv:2107.00641
Instance SegmentationonCOCO test-dev
AP75· 2021-07-01
56.5
best: 63.4 (Co-DETR)
Focal Self-attention for Local-Global Interactions in Vision Transformers arXiv:2107.00641
Instance SegmentationonCOCO test-dev
APL· 2021-07-01
64.2
best: 72.4 (EVA)
Focal Self-attention for Local-Global Interactions in Vision Transformers arXiv:2107.00641
Instance SegmentationonCOCO test-dev
APS· 2021-07-01
35.6
best: 41.6 (Co-DETR)
Focal Self-attention for Local-Global Interactions in Vision Transformers arXiv:2107.00641
Instance SegmentationonCOCO test-dev
mask AP· 2021-07-01
51.3
best: 57.1 (Co-DETR)
Focal Self-attention for Local-Global Interactions in Vision Transformers arXiv:2107.00641