How to setup workflow (Jetson Nano + x86 Ubuntu PC)

Autonomous Machines Jetson & Embedded Systems Jetson Nano
1

Hi,
Can someone explaine me what is a proper way to setup workflow for Jetson app development?
For a past 2 days I can’t compile jetson-inference[1] for a x86 Ubuntu machine, each time there is some conflicts or missing libs | headers | whatever stuff.
There is totally no docs how one can develope app on host PC. Jetpack is only for Jetsons, Nvidia SDK Manager lacks TensorRT, manual installation… don’t want even talk about it.

Jetson family are ARM64, so there is no IDE.
Any help will be really apreciated. Realy!

[1] GitHub - dusty-nv/jetson-inference: Hello AI World guide to deploying deep-learning inference networks and deep vision primitives with TensorRT and NVIDIA Jetson.

2

Hi ckryvomaz, please refer to this GitHub Issue filed about the same topic: https://github.com/dusty-nv/jetson-inference/issues/462

Here is a JetsonHacks article about using Visual Studio Code on Jetson Nano: https://www.jetsonhacks.com/2019/10/01/jetson-nano-visual-studio-code-python/

3

Thanks.
I know this github issue, its me created it.

4

Hey, ckryvomaz

Not sure if you are using Python or C++ but I’ve been using NetBeans (g++ compiler) and C++ to develop apps on my Linux laptop which I then copy and compile using the same IDE on my Jetson Nano. I’ve been mostly using libraries such as Boost, OpenCV, CUDA and TensorRT.

Even (Tensor, OpenCV…) C++ code created in Visual Studio (Win) can be easily ported to Linux(Jetson Nano) as long as it doesn’t use libraries that are only Win specific. I’ve recently started programming on Linux (been programming on Win my whole life) and it turned out to be a much better experience than what I had anticipated :)

5

Fwiw, Jetbrains products work fine as well. I’ve used both CLion and Pycharm Pro for remote development and debug.