List your company
ResearchFormerly WizWand

Semantic Segmentation on PartNet (val)

50.8mIoUpseudo grid
39.04842.09945.1548.201Jul 2, 2020
Updated 8mo ago

Evaluation results

MethodLinks
pseudo gridparam=18.5M, FLOP=8.5G2020.0750.8
PosPool*param=18.5M, FLOP=6.8G2020.0750.6
adapt weightsparam=18.5M, FLOP=7.2G2020.0750.1
PosPoolparam=18.5M, FLOP=6.8G2020.0750
point-wise MLPparam=25.6M, FLOP=9.1G2020.0748.1
PosPool* (S)param=1.1M, FLOP=0.5G, Model Scale=Small2020.0746.1
baseline† (MAX)param=18.5M, FLOP=6.7G, Pooling strategy=MAX, Equation=72020.0745.4
point-wise MLP (S)param=1.6M, FLOP=0.7G, Model Scale=Small2020.0745.3
PosPool (S)param=1.1M, FLOP=0.5G, Model Scale=Small2020.0744.6
baseline† (AVG)param=18.5M, FLOP=6.7G, Pooling strategy=AVG, Equation=72020.0744.2
pseudo grid (S)param=1.2M, FLOP=0.9G, Model Scale=Small2020.0744.2
adapt weights (S)param=1.2M, FLOP=0.5G, Model Scale=Small2020.0744.1
baseline*param=18.5M, FLOP=6.7G, Equation=62020.0742.5
baseline† (MAX, S)param=1.1M, FLOP=0.4G, Pooling strategy=MAX, Model Scale=Small, Equation=72020.0739.8
baseline† (AVG, S)param=1.1M, FLOP=0.4G, Pooling strategy=AVG, Model Scale=Small, Equation=72020.0739.5