### Create and Install Token via CLI (Self-Service)

Source: https://docs.degirum.com/pysdk/token-installation-and-management

Initiates a self-service enrollment process for developer-managed fleets. The CLI guides the user through authentication and installs the token locally.

```bash
degirum token create
```

--------------------------------

### Install OpenVINO Runtime via CLI

Source: https://docs.degirum.com/pysdk/release-notes

Install the OpenVINO runtime on Windows using the `degirum install-runtime` command. Administrator privileges may be required for installation in shared directories.

```bash
degirum install-runtime openvino
```

--------------------------------

### Query Documentation API Example

Source: https://docs.degirum.com/pysdk/quickstart

To get additional information not directly on the page, make an HTTP GET request to the page URL with the 'ask' query parameter. The question should be specific and in natural language.

```http
GET https://docs.degirum.com/pysdk/quickstart.md?ask=<question>
```

--------------------------------

### Install Token via CLI

Source: https://docs.degirum.com/pysdk/token-installation-and-management

Use this command to install a workspace token on a device. Installing a new token will replace any existing one.

```bash
degirum token install <YOUR_TOKEN>
```

--------------------------------

### Install HailoRT PCIe Driver

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/hailort

Install the HailoRT PCIe driver by running this command after downloading the driver package. A system restart is required after installation.

```bash
sudo dpkg --install hailort-pcie-driver_4.23.0_all.deb
```

--------------------------------

### Install Akida PCIe Driver

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/akida

Installs the Akida PCIe driver by cloning the repository and running the provided installation script. This is necessary if no Brainchip device is detected.

```bash
git clone https://github.com/Brainchip-Inc/akida_dw_edma
cd akida_dw_edma
sudo ./install.sh
```

--------------------------------

### Install AI Hub Token using Degirum CLI

Source: https://docs.degirum.com/pysdk/installation

Install your AI Hub token using the `degirum` CLI. Replace `<TOKEN>` with your actual AI Hub token string.

```bash
degirum token install <TOKEN>
```

--------------------------------

### Verify HailoRT Installation

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/hailort

After installing HailoRT and its driver, and restarting your PC, use this command to verify that the installation was successful and the hardware is recognized.

```bash
hailortcli fw-control identify
```

--------------------------------

### Install RKNN Runtime

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/rknn

Install the RKNN runtime by cloning the official repository and copying the shared library to the system's library directory. Ensure you have sudo privileges for the copy operation.

```bash
git clone https://github.com/airockchip/rknn-toolkit2.git
cd rknn-toolkit2
sudo cp ./rknpu2/runtime/Linux/librknn_api/aarch64/librknnrt.so /usr/lib/
```

--------------------------------

### Start AI Server with Custom Zoo

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/command-line-interface

Starts the AI server, serving models from a specified local directory. By default, it uses the 'asio' protocol and the current directory as the model zoo.

```bash
degirum server start --zoo ./my-zoo
```

--------------------------------

### Install Device-Specific Token during Provisioning

Source: https://docs.degirum.com/pysdk/token-installation-and-management

This command is used during the provisioning process to install a unique token for each device, recommended for IT-managed fleets.

```bash
degirum token install <DEVICE_SPECIFIC_TOKEN>
```

--------------------------------

### Install Workspace Token

Source: https://docs.degirum.com/pysdk/token-installation-and-management

Install the correct Workspace token if one is missing or if you need to switch to a different Workspace. Replace `<TOKEN>` with your actual token string.

```bash
degirum token install --token <TOKEN>
```

--------------------------------

### Install HailoRT Package

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/hailort

After downloading the appropriate HailoRT .deb package for your system's architecture, use this command to install it. Replace `<architecture>` with your system's architecture (e.g., amd64, arm64).

```bash
sudo dpkg --install hailort_4.23.0_<architecture>.deb
```

--------------------------------

### Example Inference Output

Source: https://docs.degirum.com/pysdk/quickstart

This is an example of the output format you can expect after performing inference with the DeGirum PySDK.

```text
- bbox: [241.37208024326583, 113.8805459421775, 897.3911454167619, 704.7711576639625]
  category_id: 15
  label: cat
  score: 0.8781571984291077
```

--------------------------------

### Install Latest Runtime

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/command-line-interface

Installs the latest available version of a specified AI accelerator runtime library. Ensure the runtime name is valid.

```bash
degirum install-runtime onnx
```

--------------------------------

### Verify PySDK Installation (Linux)

Source: https://docs.degirum.com/pysdk/installation

Checks the PySDK installation by running the 'degirum sys-info' command. A list of detected devices indicates a successful installation.

```bash
degirum sys-info
```

--------------------------------

### Start AI Server with Both Protocols and Custom Ports

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/command-line-interface

Starts the AI server with both 'asio' and 'http' protocols enabled, binding to specified ports. The first port is for 'asio', and the second is for 'http'.

```bash
degirum server start --zoo ./my-zoo --port 12345 --protocol both
```

--------------------------------

### Install PySDK with pip on Windows

Source: https://docs.degirum.com/pysdk/installation

Install PySDK and optionally degirum-tools using pip, specifying the custom package index.

```powershell
pip install -i https://pkg.degirum.com degirum
```

```powershell
pip install -i https://pkg.degirum.com degirum degirum-tools
```

--------------------------------

### Install Akida Runtime

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/akida

Installs the specified version of the Akida Runtime package. Ensure your pip is up to date before running.

```bash
pip install --upgrade pip
pip install akida==2.11.0
```

--------------------------------

### List Available Runtimes

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/command-line-interface

Use this command to list all available AI accelerator runtime libraries and their versions that can be installed.

```bash
degirum install-runtime --list
```

--------------------------------

### Install Specific Runtime Version

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/command-line-interface

Installs a specific version of an AI accelerator runtime library. You can also install multiple versions by providing them as a comma-separated list.

```bash
degirum install-runtime openvino 2025.3.0
```

--------------------------------

### Query Documentation via HTTP GET

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/tensorrt

To get additional information not explicitly present on the page, perform an HTTP GET request with an 'ask' query parameter. The question should be specific and self-contained.

```http
GET https://docs.degirum.com/pysdk/runtimes-and-drivers/tensorrt.md?ask=<question>
```

--------------------------------

### Minimum Code Example for Loading a Model

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/loading-an-ai-model

This streamlined example shows the essential parameters required to load a model directly using degirum.load_model(). It includes the model name, inference host address, zoo URL, and token, with a comment indicating where optional parameters can be added.

```python
import degirum as dg

# Loading a model
model = dg.load_model(
    model_name = "yolov8n_relu6_coco--640x640_quant_rknn_rk3588_1",
    inference_host_address = "@cloud",
    zoo_url = "degirum/public",
    token = your_token 
    # optional parameters, such as overlay_show_probabilities = True
)
print(model)
```

--------------------------------

### Install DeGirum PySDK using Pip

Source: https://docs.degirum.com/pysdk/installation

Use this command to install the DeGirum PySDK from DeGirum's package index. It is recommended to do this within a virtual environment.

```bash
pip install -i https://pkg.degirum.com degirum
```

--------------------------------

### Verify Akida Installation

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/akida

Checks if the Akida Runtime is installed correctly by querying its version.

```bash
akida version
```

--------------------------------

### Query Documentation via HTTP GET

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/archive

To get information not directly on the page, make an HTTP GET request to the current page URL with an 'ask' query parameter. The question should be specific and in natural language. The response includes the answer and relevant documentation excerpts.

```http
GET https://docs.degirum.com/pysdk/user-guide-pysdk/archive.md?ask=<question>
```

--------------------------------

### Query Documentation via HTTP GET

Source: https://docs.degirum.com/pysdk/eula

Perform an HTTP GET request on the current page URL with the `ask` query parameter to dynamically query the documentation. The question should be specific and self-contained. Use this when the answer is not explicitly present, for clarification, or to retrieve related documentation.

```HTTP
GET https://docs.degirum.com/pysdk/eula.md?ask=<question>
```

--------------------------------

### Query Documentation via GET Request

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/api-ref/misc

To get additional information not directly present on the page, perform an HTTP GET request to the current page URL with the 'ask' query parameter. The question should be specific and self-contained.

```http
GET https://docs.degirum.com/pysdk/user-guide-pysdk/api-ref/misc.md?ask=<question>
```

--------------------------------

### Query Documentation via HTTP GET

Source: https://docs.degirum.com/pysdk/token-installation-and-management

Perform an HTTP GET request to the documentation URL with the `ask` query parameter to dynamically query for information not explicitly present on the page. The question should be specific and self-contained.

```http
GET https://docs.degirum.com/pysdk/token-installation-and-management.md?ask=<question>
```

--------------------------------

### Install Axelera Runtime and Driver Packages

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/axelera

Installs the specified versions of the Axelera runtime, device driver, Metis DKMS, and RISC-V toolchain.

```bash
sudo apt install -y axelera-runtime-1.4.0 axelera-device-1.4.0 metis-dkms=1.2.2 axelera-riscv-gnu-newlib-toolchain-409b951ba662-7
```

--------------------------------

### Example Output for Model Input Data Type

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/running-ai-model-inference

Shows an example output when querying a model's input types, indicating that the model expects image data.

```text
['Image']
```

--------------------------------

### Query Axelera Documentation

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/axelera

Demonstrates how to query the documentation dynamically using an HTTP GET request with an 'ask' query parameter.

```http
GET https://docs.degirum.com/pysdk/runtimes-and-drivers/axelera.md?ask=<question>
```

--------------------------------

### Querying Documentation Dynamically

Source: https://docs.degirum.com/pysdk/user-guide-pysdk

To get information not explicitly on the page, perform an HTTP GET request to the current page URL with the `ask` query parameter. The question should be specific and self-contained. The response includes the answer and relevant excerpts.

```http
GET https://docs.degirum.com/pysdk/user-guide-pysdk.md?ask=<question>
```

--------------------------------

### Start degirum Linux Service

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/setting-up-an-ai-server

Start the degirum AI server as a Linux system service using systemctl. This command requires sudo privileges.

```bash
sudo systemctl start degirum.service
```

--------------------------------

### Query Documentation via HTTP GET

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/setting-up-an-ai-server

Perform an HTTP GET request to the current page URL with the `ask` query parameter to dynamically query documentation. The question should be specific and in natural language.

```http
GET https://docs.degirum.com/pysdk/user-guide-pysdk/setting-up-an-ai-server.md?ask=<question>
```

--------------------------------

### Query Documentation Dynamically

Source: https://docs.degirum.com/pysdk/installation

To get additional information not directly on the page, perform an HTTP GET request with the 'ask' query parameter. The question should be specific and in natural language.

```http
GET https://docs.degirum.com/pysdk/installation.md?ask=<question>
```

--------------------------------

### Perform HTTP GET Request with `ask` Parameter

Source: https://docs.degirum.com/pysdk/release-notes/retired-versions

Use this method to query documentation dynamically when information is not directly available. The question should be specific and self-contained.

```http
GET https://docs.degirum.com/pysdk/release-notes/retired-versions.md?ask=<question>
```

--------------------------------

### Install an AI Hub Token

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/command-line-interface

Use this command to save a provided token to local storage. The token will then be automatically used by PySDK.

```bash
degirum token install --token <your_token>
```

--------------------------------

### Ask a Question via GET Request

Source: https://docs.degirum.com/pysdk

Use this method to perform an HTTP GET request on the documentation index with the 'ask' parameter to find specific information. The question should be specific, self-contained, and written in natural language.

```http
GET https://docs.degirum.com/readme.md?ask=<question>
```

--------------------------------

### Query Documentation Dynamically

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers

To get additional information not directly present on the page, perform an HTTP GET request with the 'ask' query parameter. The question should be specific and in natural language. Use this for clarifications or to retrieve related documentation.

```http
GET https://docs.degirum.com/pysdk/runtimes-and-drivers.md?ask=<question>
```

--------------------------------

### Query Documentation Dynamically

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/running-ai-model-inference

To get additional information not present on the page, perform an HTTP GET request to the current page URL with the `ask` query parameter. The question should be specific and in natural language.

```http
GET https://docs.degirum.com/pysdk/user-guide-pysdk/running-ai-model-inference.md?ask=<question>
```

--------------------------------

### Landmark Object Detection Results Example

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/api-ref/postprocessor

Illustrates the structure for landmark detection results, which may include keypoint coordinates and connectivity information.

```python
[
    {
        'landmarks': [

```

--------------------------------

### Get Supported Devices

Source: https://docs.degirum.com/pysdk/release-notes/retired-versions

Obtain a list of supported runtime/device combinations for a specified inference engine. This helps in identifying compatible hardware and software configurations.

```python
supported_device_types = degirum.get_supported_devices("localhost")
```

--------------------------------

### Example Output of Loaded Model

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/loading-an-ai-model

This shows the typical output when a model is successfully loaded, indicating the object type and its memory address.

```text
<degirum.model._CloudServerModel object at 0x000002C079FEE270>
```

--------------------------------

### Configure OpenVINO Environment on Linux

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/openvino

Run this command in a new terminal after installing OpenVINO via a .tgz archive to configure the environment. Ensure your interpreter is started from a terminal where this command has been executed.

```bash
source <OpenVINO Directory>/setupvars.sh
```

--------------------------------

### Batch Inference on Video File with OpenCV

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/running-ai-model-inference

Processes frames from a video file using batch inference and displays the results overlaid on the video frames using OpenCV. This example requires `degirum` and `cv2` to be installed.

```python
import degirum as dg
import cv2

# Declaring variables

```

--------------------------------

### Upgrade Pip for PySDK Installation

Source: https://docs.degirum.com/pysdk/installation

To resolve potential installation issues, upgrade your pip package installer to the latest version using this command.

```bash
python3 -m pip install -U pip
```

--------------------------------

### __enter__

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/api-ref/model

The context manager's entry point handler.

```APIDOC
## __enter__

`degirum.model.Model.__enter__()`

Context manager enter handler.
```

--------------------------------

### Loading an AI Model with degirum.load_model()

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/loading-an-ai-model

This example demonstrates loading an AI model using degirum.load_model(). It first retrieves supported devices and lists available models, then loads the first model from the list with specific optional parameters like overlay_show_probabilities and output_confidence_threshold.

```python
import degirum as dg

# Set your inference host address, model zoo, and token in these variables.
your_host_address = "@cloud" # Can be "@cloud", host:port, or "@local"
your_model_zoo = "degirum/public"
your_token = "<your_token>"

# Retrieve supported device types
supported_devices = dg.get_supported_devices(
    inference_host_address = your_host_address, 
    zoo_url = your_model_zoo, 
    token = your_token,
)

# List models from the AI Hub model zoo with specific filtering criteria
model_list = dg.list_models(
    inference_host_address = your_host_address, 
    zoo_url = your_model_zoo, 
    token = your_token,
    device_type=list(supported_devices)
)

# Load the first model from the list with some optional parameters
model = dg.load_model(
    model_name = list(model_list)[0], 
    inference_host_address = your_host_address, 
    zoo_url = your_model_zoo, 
    token = your_token,
    overlay_show_probabilities = True,
    output_confidence_threshold = 0.5
)
print(model)
```

--------------------------------

### Configure degirum.service for Linux

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/setting-up-an-ai-server

Create a systemd service configuration file to automatically start the AI server on boot. Ensure you replace placeholders like <your_username> and <path_to_python> with your specific details.

```ini
[Unit]
Description=DeGirum AI Service

[Service]
WorkingDirectory=/home/<your_username>/
ExecStart=<path_to_python> -m degirum.server --zoo /home/<your_username>/zoo
Restart=always
RestartSec=10
SyslogIdentifier=degirum-ai-server
User=<your_username>

[Install]
WantedBy=multi-user.target
```

--------------------------------

### Verify PySDK Installation on Windows

Source: https://docs.degirum.com/pysdk/installation

Run this command in your active virtual environment to check if PySDK is installed correctly. A list of detected devices indicates a successful installation.

```powershell
degirum sys-info
```

```powershell
(degirum-windows) PS C:\degirum\degirum-windows\Scripts> degirum sys-info
Devices:
  N2X/CPU:
  - '@Index': 0
  TFLITE/CPU:
  - '@Index': 0
  - '@Index': 1
  - '@Index': 2
  - '@Index': 3
Software Version: 1.0.0
```

--------------------------------

### Enable degirum Linux Service on Startup

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/setting-up-an-ai-server

Configure the degirum AI server service to start automatically when the system boots up using systemctl. This command requires sudo privileges.

```bash
sudo systemctl enable degirum.service
```

--------------------------------

### Verify CUDA Installation

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/tensorrt

Check if the CUDA installation was successful by running the nvcc command to display the compiler version.

```bash
nvcc --version
```

--------------------------------

### Clear Installed Token via CLI

Source: https://docs.degirum.com/pysdk/token-installation-and-management

Removes the currently installed token from the device. This is useful for local cleanup when a device is reachable.

```bash
degirum token clear
```

--------------------------------

### Query Documentation with 'ask' Parameter

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/memryxrt

Perform an HTTP GET request to query the documentation dynamically. Use this when the answer is not explicitly present, you need clarification, or want to retrieve related documentation sections.

```http
GET https://docs.degirum.com/pysdk/runtimes-and-drivers/memryxrt.md?ask=<question>
```

--------------------------------

### Model Zoo Directory Structure Example

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/organizing-models

Organize each model in its own subdirectory containing all necessary files. This structure is followed by AI Hub and can be adopted for local model zoos.

```bash
<model_zoo_directory>/
    ├── mobilenet_v2_imagenet--224x224_quant_n2x_orca1_1/
    │   ├── mobilenet_v2_imagenet--224x224_quant_n2x_orca1_1.json
    │   ├── mobilenet_v2_imagenet--224x224_quant_n2x_orca1_1.n2x
    │   └── labels.json  (if required)
    ├── yolov8n_coco--640x640_quant_openvino_1/
    │   ├── yolov8n_coco--640x640_quant_openvino_1.json
    │   ├── yolov8n_coco--640x640_quant_openvino_1.xml
    │   ├── yolov8n_coco--640x640_quant_openvino_1.bin
    │   └── postprocess.py  (if required)
    └── yolov8n_relu6_coco_pose--640x640_quant_tflite_edgetpu_1/
        ├── yolov8n_relu6_coco_pose--640x640_quant_tflite_edgetpu_1.json
        ├── yolov8n_relu6_coco_pose--640x640_quant_tflite_edgetpu_1.tflite
        └── labels.json  (if required)
```

--------------------------------

### Update PATH and LD_LIBRARY_PATH for CUDA

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/tensorrt

After installing CUDA on Linux, update your PATH and LD_LIBRARY_PATH environment variables. Replace <VERSION> with your installed CUDA version.

```bash
export PATH=/usr/local/cuda-<VERSION>/bin${PATH:+:${PATH}}
export LD_LIBRARY_PATH=/usr/local/cuda-<VERSION>/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
```

--------------------------------

### Install Python on Raspberry Pi OS

Source: https://docs.degirum.com/pysdk/installation

Updates package lists and upgrades installed packages on Raspberry Pi OS. Python 3 is pre-installed on 64-bit versions.

```bash
sudo apt update
sudo apt upgrade -y
```

--------------------------------

### Install Python on Ubuntu 24.04

Source: https://docs.degirum.com/pysdk/installation

Installs Python 3 and the venv module on Ubuntu 24.04. Ensure package lists are updated and existing packages are upgraded first.

```bash
sudo apt update
sudo apt upgrade -y
sudo apt install -y python3 python3-venv
```

--------------------------------

### Check RKNN Runtime Installation

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/rknn

Verify if the RKNN runtime library is installed by checking its version string. This command helps confirm the presence and version of the librknnrt.so file.

```bash
strings /usr/lib/librknnrt.so | grep librknnrt
```

```bash
librknnrt version: 2.0.0b0 (35a6907d79@2024-03-24T10:31:14)
```

--------------------------------

### Query Documentation Dynamically

Source: https://docs.degirum.com/pysdk/runtimes-and-drivers/openvino

Perform an HTTP GET request on the documentation URL with the 'ask' query parameter to retrieve specific information not explicitly present on the page. The question should be specific and self-contained.

```http
GET https://docs.degirum.com/pysdk/runtimes-and-drivers/openvino.md?ask=<question>
```

--------------------------------

### Sample JSON Configuration

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/model-json-structure

Include or omit parameters as your model requires. This is a template for model configuration.

```json
{
  "ConfigVersion": <config_version_number>,
  "Checksum": "<checksum>",
  "DEVICE": [
    {
      "DeviceType": "<device_type>",
      "RuntimeAgent": "<runtime_agent>",
      "SupportedDeviceTypes": "<supported_device_types>"
    }
  ],
  "PRE_PROCESS": [
    {
      "InputN": <input_N>,
      "InputH": <input_H>,
      "InputW": <input_W>,
      "InputC": <input_C>,
      "InputQuantEn": <boolean_for_quant_enabled>
    }
  ],
  "MODEL_PARAMETERS": [
    {
      "ModelPath": "<path_to_model>"
    }
  ],
  "POST_PROCESS": [
    {
      "OutputPostprocessType": "<postprocess_type>",
      "OutputNumClasses": <number_of_output_classes>,
      "LabelsPath": "<path_to_labels_json>"
    }
  ]
}
```

--------------------------------

### Start DeGirum AI Server

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/api-ref/server

Launches the DeGirum AI server. Specify the path to your local model zoo directory using `--zoo`. Optionally, you can set a custom TCP port with `--port`; otherwise, it defaults to 8778.

```bash
python -m degirum.server --zoo <local zoo path> [--port <server TCP port>]
```

--------------------------------

### Query Documentation Dynamically

Source: https://docs.degirum.com/pysdk/user-guide-pysdk/api-ref

Perform an HTTP GET request on the current page URL with the `ask` query parameter to dynamically query the documentation. The question should be specific, self-contained, and written in natural language. Use this mechanism when the answer is not explicitly present, you need clarification, or want to retrieve related documentation sections.

```http
GET https://docs.degirum.com/pysdk/user-guide-pysdk/api-ref.md?ask=<question>
```