Yaming Wang, Supermicro
mwc 2019
In this talk we will go through several edge server and Nvidia GPU cards configurations, running GoogleNet and ResNet-50 with different batch sizes to compare the results for best edge server inference setup.