tiny-MOAT-2 (IN-1K pretraining, single-scale)

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 minival
box AP· 2022-10-04
53
best: 66 (PE_spatial (DETA))
MOAT: Alternating Mobile Convolution and Attention Brings Strong Vision Models arXiv:2210.01820
2D ClassificationonCOCO minival
box AP· 2022-10-04
53
best: 66 (PE_spatial (DETA))
MOAT: Alternating Mobile Convolution and Attention Brings Strong Vision Models arXiv:2210.01820
2D Object DetectiononCOCO minival
box AP· 2022-10-04
53
best: 66 (PE_spatial (DETA))
MOAT: Alternating Mobile Convolution and Attention Brings Strong Vision Models arXiv:2210.01820
16konCOCO minival
box AP· 2022-10-04
53
best: 66 (PE_spatial (DETA))
MOAT: Alternating Mobile Convolution and Attention Brings Strong Vision Models arXiv:2210.01820

Computer Vision2 results

Object DetectiononCOCO minival
box AP· 2022-10-04
53
best: 66 (PE_spatial (DETA))
MOAT: Alternating Mobile Convolution and Attention Brings Strong Vision Models arXiv:2210.01820
Instance SegmentationonCOCO minival
mask AP· 2022-10-04
45
best: 56.6 (Co-DETR)
MOAT: Alternating Mobile Convolution and Attention Brings Strong Vision Models arXiv:2210.01820