Installation
Follow comprehensive installation guides for PySDK, covering virtual environments, Docker images, and verification.
Supported Environments
PySDK currently supports these operating systems, CPU architectures, and Python versions.
Ubuntu Linux 20.04, 22.04, 24.04
x86-64
3.9 ... 3.12
Ubuntu Linux 20.04, 22.04, 24.04
ARM AArch64
3.9 ... 3.12
Raspberry Pi OS (64 bit)
ARM AArch64
3.9 ... 3.12
Windows 10/11
x86-64
3.9 ... 3.12
macOS 12
ARM AArch64
3.9 ... 3.12
Supported Hardware
Below is a summary of what hardware we support, including the runtime, devices, and model binary formats.
Hailo
HAILORT
HAILO8, HAILO8L
.hef
TFLITE
EDGETPU
.tflite
Intel
OPENVINO
CPU, GPU, NPU
.onnx, .bin & .xml
BrainChip
AKIDA
NSoC_v2, AKD1500
.bin
DeGirum
N2X
ORCA1
.n2x
Rockchip
RKNN
RK3588, RK3568, RK3566
.rknn
MemryX
MEMRYX
MX3
.dfp
DeepX
DEEPX
M1A
.onnx
AMD
ONNX
VITIS_NPU
.onnx
NVIDIA
TENSORRT
DLA, GPU
.onnx
Installation Options
Using pip install
in a Virtual Environment
pip install
in a Virtual EnvironmentWe recommend creating a Python virtual environment for PySDK and other DeGirum Python packages.
Open a command prompt or terminal, then follow the steps for your system below.
Windows
If Python is not installed, disable the Windows Python App Execution Alias before installing Python and PySDK. Ensure the Visual C++ Redistributable for Visual Studio 2015 or later is installed.
Installing PySDK in a Python Virtual Environment on Windows
Ensure a supported Python version is installed, then watch the video or follow the steps below:
Create a Python Virtual Environment
Use this command to create a virtual environment:
python -m venv <virtual-environment-name>
If you have Python Launcher and multiple Python versions installed, create the virtual environment with py
and specify the version:
py <your-python-version> -m venv <virtual-environment-name>
Activate the Python Virtual Environment
After creating the environment with venv
, navigate to the Scripts
folder in your virtual environment.
cd <virtual-environment-name>\Scripts
Once in that folder, run the activation command.
On cmd.exe:
activate
On pwsh.exe:
.\Activate.ps1
After activation, the virtual environment name appears before the terminal prompt.

degirum-windows
virtual environmentInstall PySDK with pip install
With the environment active, use pip
to install PySDK and optionally degirum-tools
. The PySDK package is degirum
, while degirum-tools
helps you build AI applications with PySDK.
To install only PySDK:
pip install degirum
To install both PySDK and degirum-tools
:
pip install degirum degirum-tools
Verify PySDK Installation
Run this command in the active virtual environment:
degirum sys-info
A list of detected devices confirms that PySDK is installed.
Linux
Ensure a Python version supported by PySDK is installed.
Create a Python Virtual Environment
Use this command to create a virtual environment:
Verify that python3
points to a supported version with python3 --version
, then create the environment:
python3 -m venv <virtual-environment-name>
If you have multiple Python versions installed, you can specify a Python release version to create the virtual environment:
python<python-version> -m venv <virtual-environment-name>

Activate the Python Virtual Environment
After creating the environment with venv
, activate it by running source <virtual-environment-folder>/bin/activate
.
source <virtual-environment-folder>/bin/activate
After activation, the virtual environment name appears before the terminal prompt.

degirum
virtual environmentInstall PySDK with pip install
With the environment active, use pip
to install PySDK and optionally degirum-tools
. The PySDK package is degirum
, while degirum-tools
helps you build AI applications with PySDK.
To install only PySDK:
pip install degirum
To install both PySDK and degirum-tools
:
pip install degirum degirum-tools
Docker Images
We also provide Docker images for the DeGirum AI server and PySDK AI client installations.
Follow this GitHub link to DeGirum Docker repo for details on how to run these images and for corresponding source Dockerfiles.
Runtimes and Drivers
We support multiple hardware options and runtime environments. PySDK comes with support for N2X and TFLite runtimes. Follow this link for more details.
Limiting Runtime Plugin Loading
PySDK searches for runtime plugins at startup. To restrict which plugins load,
set the DG_PLUGINS_ALLOWED
environment variable to a list of plugin prefixes
separated by any non alphanumeric character. For example:
export DG_PLUGINS_ALLOWED=n2x_runtime_agent;onnx_runtime_agent
Only plugins with prefixes in the list load when DG_PLUGINS_ALLOWED
is set.
If the variable is not defined, all available plugins load as before.
Supported plugin prefixes include n2x_runtime_agent
, onnx_runtime_agent
,hailort_runtime_agent
, tflite_runtime_agent
, openvino_runtime_agent
,tensorrt_runtime_agent
, rknn_runtime_agent
, memryxrt_runtime_agent
, andakida_runtime_agent
.
Verification
After installing the required runtimes and device drivers, verify that PySDK can access them by running this command:
degirum sys-info
The output lists the devices available to PySDK with relevant information.
Devices:
<other device types>
<runtime>/<device type>:
- '@Index': 0
<device dependent information, optional>
- '@Index': 1
<device dependent information, optional>
<other devices>
<other device types>
Troubleshooting
If you encounter errors such as:
ERROR: Could not find a version that satisfies the requirement degirum
ERROR: No matching distribution found for degirum
Create a new virtual environment and reinstall PySDK.
You can also try upgrading pip
with this command:
python3 -m pip install -U pip
On Windows, confirm that the Visual C++ Redistributable for Visual Studio 2015 or later is installed.
Last updated
Was this helpful?