Before you begin

Before exploring the guides and running examples, review this page to ensure you're set up for success. It covers how to install DeGirum's PySDK and DeGirum Tools with Hailo devices.

Estimated read time: 3 minutes

These guides are written to be self-contained. Most code examples should work out of the box, with very few exceptions.

Prerequisites

Before starting, make sure you've completed the following setup:

1

Python virtual environment

Use a dedicated Python virtual environment with both PySDK and DeGirum Tools installed.

Click here for instructions on setting up a Python virtual environment and installing PySDK (degirum) and DeGirum Tools (degirum_tools).

We recommend avoiding your system-level Python installation to prevent conflicts.

2

Hailo runtime and drivers

Make sure the HailoRT runtime and drivers are installed.

Click here for Hailo runtime and driver installation instructions.

To verify the installation, run hailortcli fw-control identify.

For Hailo-8, you should see output similar to:

Executing on device: 0000:03:00.0
Identifying board
Control Protocol Version: 2
Firmware Version: 4.20.0 (release,app,extended context switch buffer)
Logger Version: 0
Board Name: Hailo-8
Device Architecture: HAILO8
Serial Number: HLLWMB0214600138
Part Number: HM218B1C2LA
Product Name: HAILO-8 AI ACCELERATOR M.2 B+M KEY MODULE

If the output is missing or incomplete, reinstall the HailoRT runtime and drivers.

3

Supported hardware

Check that your hardware is recognized by PySDK. Run degirum sys-info and look for a Hailo-8 or Hailo-8L device.

Here is an example for Hailo-8:

Devices:
  HAILORT/HAILO8:
  - '@Index': 0
    Board Name: Hailo-8
    Device Architecture: HAILO8
    Firmware Version: 4.20.0
    ID: '0000:03:00.0'
    Part Number: xxxxxxx
    Product Name: HAILO-8 AI ACCELERATOR M.2 B+M KEY MODULE
    Serial Number: "xxxxxxxxxxx"
  (OTHER_RUNTIMES):
  - '@Index': 0

Click here for full instructions on verifying your Hailo and PySDK installation.

4

Access to the DeGirum AI Hub

For access to pre-verified models and cloud-hosted hardware, install an AI Hub Workspace Token into your PySDK setup.

While not required, it simplifies model access and improves the development experience.

You can learn what a Workspace Token is and follow these instructions to install one.

To verify that your token is active, run degirum token status:

token: <token>
$schema: https://hub.degirum.com/schemas/GetTokenInfoOutputBody.json
created_at: '2025-01-01T11:11:11.111111Z'
description: <description>
value: <token>
expiration: '2025-01-02T11:11:11.111111Z'
user: <your_email>
space: <your_workspace>

Optional: improve your deployment environment

Although not required, the tools below can enhance your experience.

1

IDE support

We recommend using an integrated development environment like Visual Studio Code (VS Code), which makes it easier to run and edit PySDK and DeGirum Tools code compared to the default Python REPL or plain text files.

2

Jupyter Notebook extension

Many examples in our guides are written in Jupyter Notebook and can run directly in your IDE or Google Colab.

Click here for details on the Jupyter Notebook extension for VS Code.

Core concepts to know

To get the most out of these guides, it's helpful to be familiar with these PySDK and AI Hub concepts:

• AI inference engine: the component that runs AI models to make predictions.

• AI model: a trained model that performs tasks like object detection or face recognition.

• AI Model Zoo: a collection of ready-to-use models available to the inference engine.

Understanding these concepts will give you the foundation to navigate the examples and tools confidently.