This sample targets for demonstrating TensorRT2.1 Plugin API

We leverage most of the functions from jetson_inference; please check it first if you need more DL samples:

https://github.com/dusty-nv/jetson-inference

Jetson TX2
JetPack-3.1

$ sudo apt-get install git cmake
$ git clone https://github.com/AastaNV/Face-Recognition
$ cd Face-Recognition
$ mkdir build
$ cd build
$ cmake ..
$ make

$ cd aarch64/bin
$ ./face-recognition

BboxMergeLayer
This plugin layer demonstrate how to implement a CPU-based Plugin layer

1. Make required tensor as output
2. Allocate unified memory: CPU pointer== GPU pointer

DataRoiLayer
This plugin layer demonstrate how to implement a GPU Plugin layer

1. Got input/output data pointer in enqueue function
2. Launch GPU kernel with same cuda stream

RecognitionLayer
This plugin layer demonstrate more complicated handling of Plugin layer

1. This class can handle two differient layers: selectBbox and summaryLabel
2. Define some shared variable to make between layers communication easier

Please rise your problem in our forum to get immediately support.
https://devtalk.nvidia.com/default/board/189/jetson-tx2/