MogaNet-L (Cascade Mask R-CNN)

Reported on 6 benchmarks across 6 tasks · 1 paper

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

Methodology4 results

3DonCOCO 2017 val
AP· 2022-11-07
53.3
best: 72.2 (LOGO-CAP (Ours) HRNet-W48)
MogaNet: Multi-order Gated Aggregation Network arXiv:2211.03295
2D ClassificationonCOCO 2017 val
AP· 2022-11-07
53.3
best: 61.8 (Mr. DETR (Swin-L, 1x, 5cale))
MogaNet: Multi-order Gated Aggregation Network arXiv:2211.03295
2D Object DetectiononCOCO 2017 val
AP· 2022-11-07
53.3
best: 61.8 (Mr. DETR (Swin-L, 1x, 5cale))
MogaNet: Multi-order Gated Aggregation Network arXiv:2211.03295
16konCOCO 2017 val
AP· 2022-11-07
53.3
best: 61.8 (Mr. DETR (Swin-L, 1x, 5cale))
MogaNet: Multi-order Gated Aggregation Network arXiv:2211.03295

Computer Vision2 results

Object DetectiononCOCO 2017 val
AP· 2022-11-07
53.3
best: 61.8 (Mr. DETR (Swin-L, 1x, 5cale))
MogaNet: Multi-order Gated Aggregation Network arXiv:2211.03295
Instance SegmentationonCOCO test-dev
mask AP· 2022-11-07
46.1
best: 57.1 (Co-DETR)
MogaNet: Multi-order Gated Aggregation Network arXiv:2211.03295