SeesawFaceNets: sparse and robust face verification model for mobile platform

Benchmark Model Rank Results
lightweight-face-recognition-on-agedb-30Seesaw-shuffleFaceNet(mobi)#2Accuracy: 0.9648MParams: 2.8
lightweight-face-recognition-on-cfp-fpSeesaw-shuffleFaceNet(mobi)#4Accuracy: 0.9307MParams: 2.8
lightweight-face-recognition-on-lfwSeesaw-shuffleFaceNet(mobi)#4Accuracy: 0.9965MParams: 2.8