steinert-a / JetsonInference

Jetson Nano Developer Kit 2GB Inference with C++, Qt, OpenCV and TensorRT

Jetson Inference

The Jetson Inference project is used to deploy a pre-trained model with the Jetson TensorFlow Run Time (TensorRT).

Development

This project does use some complex sofware technics. In order to simplify the understanding some general hints are given.

CMake variables

variable	description
CMAKE_SOURCE_DIR	Directory of the top-level CMakeLists.txt file.
PROJECT_SOURCE_DIR	Directory of the most recent project() call.

C++ Style Guide

The style guide used here is based on the Google C++ Style Guide.

• MyClassName.cpp / MyClassName.h
• class MyClass / myClassMember_ / myClassMemberFunction
• struct MyStruct
• GlobalFunction / parameter
• #define MY_MACRO 123
• enum class EnumTypeName {EnumValue = 0, ... };

TensorRT terminology

term	description
Deep Learning Accelerator (DLA)	DLA also called AI Accelerator is a class of hardware acceleration for machine learning algorithms.
Open Neural Network Exchange (ONNX)	ONNX is a standard developed to exchange models between frameworks.

Dependencies

For the build of the Jetson Inference project different dependencies are needed. This section does describe how to make the requirements available.

Qt

Qt library is used for basic C++ tasks under the LGPL license, like parameter line parsing or logging.

• Qt
• Qt Offline Installer

Qt library can be installed manually by compiling the Qt source on the target system.

wget https://download.qt.io/official_releases/qt/5.15/5.15.2/single/qt-everywhere-src-5.15.2.tar.xz
tar -xpf qt-everywhere-src-5.15.2.tar.xz
cd qt-everywhere-src-5.15.2/
make
make install

The build takes several hours/days. For this it is helpful to run the build in the bakground with the & symbol at the end of the command. This wil run the build in a sub-terminal and can be detached from the current terminal.

cd qt-everywhere-src-5.15.2/
make -j4 > build_stream.log &
# [1] <your-PID>
disown -h <your-PID>
# view log file
tail -f ./build_stream.log

OpenCV

Useful links with information about OpenCV:

• OpenCV Official Homepage
• OpenCV Tutorials Point
• OpenCV GitHub

OpenCV installation instruction:

git clone https://github.com/opencv/opencv.git
cd opencv
mkdir build
cd build
cmake -D CMAKE_BUILD_TYPE=RELEASE -D BUILD_EXAMPLES=OFF -D BUILD_TESTS=OFF -D BUILD_PERF_TESTS=OFF -D BUILD_EXAMPLES=OFF ..
make
make install

TensorRT

Useful links with information about TensorRT:

• https://docs.nvidia.com/deeplearning/tensorrt/api/c_api/namespacenvinfer1.html

Build

Building the Jetson Inference project can be done by executing the known commands.

mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE=Debug ..
make