Learning Efficient Object Detection Models with Knowledge Distillation

Learning Efficient Object Detection Models with Knowledge Distillation Guobin Chen1,2 Wongun Choi1 Xiang Yu1 Tony Han2 Manmohan Chandraker1,3 1 NEC Labs America 2 University of Missouri 3 University of California, San Diego Abstract Despite significant accuracy improvement in convolutional neural networks (CNN) based object detectors, they often require prohibitive runtimes to process an image for

[knok]

Faster-RCNNのモデルをコンパクト化