Installing DeepStream and CUDA for the Jetson¶
NVIDIA® DeepStream Software Development Kit (SDK) is an accelerated AI framework to build intelligent video analytics (IVA) pipelines. DeepStream runs on NVIDIA® Jetson NX™, NVIDIA® Jetson Orin™ NX, NVIDIA® Jetson Orin™ Nano in conjunction with the EchoPilot AI.
The CUDA Toolkit provides a development environment for creating high-performance GPU-accelerated applications.
The instructions below show how to install both. These instructions were developed using a Jetson Orin NX with a 256 GB SSD, running L4T 35.4.1. In most cases, you will not be able to install this software without significant available storage space. E.g., an Xavier NX with only a 16 GB eMMC will not have enough storage space. We recommend adding a NVMe SSD before proceeding.
Note
The instructions below assume that the EchoPilot AI has internet access and you are logged in to the console. Since EchoPilotAI Jetson hardware is provided with a static IP address, it is often simpler to enable DHCP and let the Jetson get internet through your LAN's router. To do this, you can use the nmcli commands to change the static-eth0 connection profile to auto/hdcp.
sudo nmcli con mod static-eth0 ipv4.method auto
sudo nmcli con down static-eth0
sudo nmcli con up static-eth0
sudo nmcli con mod static-eth0 ipv4.addresses "10.223.44.55/16" # For example
sudo nmcli con down static-eth0
sudo nmcli con up static-eth0
Install DeepStream¶
Start by doing an apt update.
Install dependencies¶
sudo apt install \
build_essential \
libssl1.1 \
libgstreamer1.0-0 \
gstreamer1.0-tools \
gstreamer1.0-plugins-good \
gstreamer1.0-plugins-bad \
gstreamer1.0-plugins-ugly \
gstreamer1.0-libav \
libgstreamer-plugins-base1.0-dev \
libgstrtspserver-1.0-0 \
libjansson4 \
libyaml-cpp-dev
Install librdkafka¶
- Clone the repo
- Configure, build and install
- Copy the generated files to the deepstream directory.
Get and install the Deepstream SDK¶
wget --content-disposition 'https://api.ngc.nvidia.com/v2/resources/org/nvidia/deepstream/6.3/files?redirect=true&path=deepstream-6.3_6.3.0-1_arm64.deb' -O deepstream-6.3_6.3.0-1_arm64.deb
sudo apt-get install ./deepstream-6.3_6.3.0-1_arm64.deb
Install CUDA¶
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2004/sbsa/cuda-keyring_1.1-1_all.deb
sudo dpkg -i cuda-keyring_1.1-1_all.deb
sudo apt-get update
sudo apt-get -y install cuda-toolkit-12-3
Install CUDA driver¶
Boost the clocks¶
Now Reboot the system.Verify functionality¶
At this point, the deepstream-app
should run without errors:
Explore Sample Code¶
Browse and run precompiled sample applications in sources/apps/sample_apps
.
Follow the directory’s README file to run the application.