Skip to main content

Installation

This guide covers the installation of NVIDIA DeepStream SDK on different platforms.

Prerequisites​

Hardware Requirements​

For NVIDIA Jetson Devices:

  • Jetson Nano, Xavier NX, Xavier AGX, Orin Nano, Orin NX, or Orin AGX
  • Minimum 4GB RAM (8GB+ recommended)
  • MicroSD card (32GB+ recommended) or NVMe SSD

For x86_64 Systems:

  • NVIDIA GPU with compute capability 6.0 or higher
  • Minimum 8GB RAM (16GB+ recommended)
  • 20GB free disk space

Software Requirements​

  • Ubuntu: 20.04 or 22.04 LTS
  • NVIDIA Driver: 470.x or higher
  • NVIDIA CUDA: 11.4 or higher (12.x recommended)
  • TensorRT: 8.5 or higher
  • GStreamer: 1.16.2 or higher

Installation Methods​

Method 1: Using NVIDIA SDK Manager (Jetson Only)​

The easiest way to install DeepStream on Jetson devices:

  1. Download and install SDK Manager
  2. Connect your Jetson device via USB
  3. Select DeepStream SDK during component selection
  4. Follow the installation wizard

Step 1: Install Dependencies​

# Update package lists
sudo apt update

# Install GStreamer and dependencies
sudo apt install -y \
    libgstreamer1.0-dev \
    libgstreamer-plugins-base1.0-dev \
    libgstreamer-plugins-bad1.0-dev \
    gstreamer1.0-plugins-base \
    gstreamer1.0-plugins-good \
    gstreamer1.0-plugins-bad \
    gstreamer1.0-plugins-ugly \
    gstreamer1.0-libav \
    gstreamer1.0-tools \
    gstreamer1.0-x \
    gstreamer1.0-alsa \
    gstreamer1.0-gl \
    gstreamer1.0-gtk3 \
    gstreamer1.0-qt5 \
    gstreamer1.0-pulseaudio

Step 2: Install CUDA and TensorRT​

For x86_64:

# Download CUDA toolkit from NVIDIA website
wget https://developer.download.nvidia.com/compute/cuda/repos/ubuntu2204/x86_64/cuda-keyring_1.1-1_all.deb
sudo dpkg -i cuda-keyring_1.1-1_all.deb
sudo apt update
sudo apt install cuda-toolkit-12-2

# Install TensorRT
sudo apt install tensorrt

For Jetson:

CUDA and TensorRT are included in JetPack. Skip this step if using SDK Manager.

Step 3: Download and Install DeepStream​

For x86_64:

# Download DeepStream 6.4 (check for latest version)
cd /tmp
wget https://developer.nvidia.com/downloads/deepstream-sdk-v6.4-x86_64-deb

# Install DeepStream
sudo apt install ./deepstream-6.4_6.4.0-1_amd64.deb

For Jetson:

# Download DeepStream 6.4 for Jetson
cd /tmp
wget https://developer.nvidia.com/downloads/deepstream-sdk-v6.4-jetson-deb

# Install DeepStream
sudo apt install ./deepstream-6.4_6.4.0-1_arm64.deb

Step 4: Install Additional Dependencies​

# Install librdkafka for message broker support
git clone https://github.com/edenhill/librdkafka.git
cd librdkafka
git checkout v1.9.2
./configure
make
sudo make install
sudo ldconfig

# Copy librdkafka to DeepStream lib directory
sudo mkdir -p /opt/nvidia/deepstream/deepstream/lib
sudo cp /usr/local/lib/librdkafka* /opt/nvidia/deepstream/deepstream/lib

Method 3: Using Docker (Cross-Platform)​

Docker provides a containerized environment for DeepStream:

Pull DeepStream Docker Image​

# For x86_64 with dGPU
docker pull nvcr.io/nvidia/deepstream:6.4-gc-triton-devel

# For Jetson
docker pull nvcr.io/nvidia/deepstream-l4t:6.4-samples

Run DeepStream Container​

# For x86_64
docker run -it --rm \
    --gpus all \
    -e DISPLAY=$DISPLAY \
    -v /tmp/.X11-unix:/tmp/.X11-unix \
    nvcr.io/nvidia/deepstream:6.4-gc-triton-devel

# For Jetson
sudo docker run -it --rm \
    --runtime nvidia \
    --network host \
    -e DISPLAY=$DISPLAY \
    -v /tmp/.X11-unix:/tmp/.X11-unix \
    nvcr.io/nvidia/deepstream-l4t:6.4-samples

Python Bindings Installation​

Prerequisites for Python Bindings​

sudo apt install -y python3-gi python3-dev python3-gst-1.0 \
    python-gi-dev git python-dev \
    python3 python3-pip python3.10-dev \
    cmake g++ build-essential \
    libglib2.0-dev libglib2.0-dev-bin \
    libgstreamer1.0-dev libtool m4 autoconf automake \
    libgirepository1.0-dev libcairo2-dev

Install Python Bindings​

# Clone DeepStream Python bindings
cd /opt/nvidia/deepstream/deepstream/sources
git clone https://github.com/NVIDIA-AI-IOT/deepstream_python_apps.git

# Navigate to bindings
cd deepstream_python_apps/bindings

# Build and install
mkdir build && cd build
cmake ..
make
pip3 install ./pyds-*.whl

Install Python Dependencies​

pip3 install \
    numpy \
    opencv-python \
    pycairo \
    PyGObject

Verification​

Verify DeepStream Installation​

# Check DeepStream version
deepstream-app --version

# Run sample application
cd /opt/nvidia/deepstream/deepstream/samples/configs/deepstream-app
deepstream-app -c source4_1080p_dec_infer-resnet_tracker_sgie_tiled_display_int8.txt

Verify Python Bindings​

import sys
sys.path.append('/opt/nvidia/deepstream/deepstream/lib')
import pyds

print(f"DeepStream Python bindings version: {pyds.__version__}")

Common Installation Issues​

Issue: Missing CUDA libraries​

Solution:

# Add CUDA to PATH and LD_LIBRARY_PATH
echo 'export PATH=/usr/local/cuda/bin:$PATH' >> ~/.bashrc
echo 'export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH' >> ~/.bashrc
source ~/.bashrc

Issue: GStreamer plugins not found​

Solution:

# Set GStreamer plugin path
export GST_PLUGIN_PATH=/opt/nvidia/deepstream/deepstream/lib/gst-plugins:$GST_PLUGIN_PATH

Issue: Python bindings import error​

Solution:

# Add DeepStream lib to Python path
export PYTHONPATH=/opt/nvidia/deepstream/deepstream/lib:$PYTHONPATH

Issue: Display not available in Docker​

Solution:

# Allow X server access
xhost +local:docker

# Run container with proper display settings
docker run -it --rm \
    --gpus all \
    -e DISPLAY=$DISPLAY \
    -v /tmp/.X11-unix:/tmp/.X11-unix \
    -v $HOME/.Xauthority:/root/.Xauthority \
    --net=host \
    nvcr.io/nvidia/deepstream:6.4-gc-triton-devel

Environment Setup​

Create a setup script for easy environment configuration:

setup_deepstream.sh
#!/bin/bash

# CUDA paths
export PATH=/usr/local/cuda/bin:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda/lib64:$LD_LIBRARY_PATH

# DeepStream paths
export DEEPSTREAM_DIR=/opt/nvidia/deepstream/deepstream
export LD_LIBRARY_PATH=$DEEPSTREAM_DIR/lib:$LD_LIBRARY_PATH
export GST_PLUGIN_PATH=$DEEPSTREAM_DIR/lib/gst-plugins:$GST_PLUGIN_PATH

# Python bindings
export PYTHONPATH=$DEEPSTREAM_DIR/lib:$PYTHONPATH

echo "DeepStream environment configured!"

Save and source the script:

chmod +x setup_deepstream.sh
source setup_deepstream.sh

# Add to .bashrc for automatic loading
echo "source ~/setup_deepstream.sh" >> ~/.bashrc

Next Steps​