2X Faster Embedded Small Batch Inference with Jetpack 3.1(devblogs.nvidia.com)

https://devblogs.nvidia.com/parallelforall/jetpack-doubles-jetson-inference-perf/