PySDK Integration

Integrate DeGirum PySDK with AI Hub to produce compact, efficient code that taps into powerful cloud-hosted inference and diverse hardware resources.

DeGirum PySDK features integrations with AI Hub that enable producing compact, easy-to-understand code. Whenever you write code with PySDK, the inference_host_address indicates where inference will be performed. Setting inference_host_address to "@cloud" connects your inference request to DeGirum’s cloud-hosted hardware, eliminating the need for local setup.

Example Code

The following code demonstrates all you need to connect to the remote device farm and run inference on a file you have locally.

import degirum as dg

# Connect to the AI Hub
inference_manager = dg.connect(
    inference_host_address = "@cloud",
    zoo_url = "degirum/degirum",
    token = "<your_token>"
)

# Load a model. In this case, an image detection model from degirum/degirum.
model = inference_manager.load_model("yolov8n_relu6_coco--640x640_quant_n2x_orca1_1")

# Run inference on an image
result = model("<your_image>")

# Print raw results
print(result)

Make sure you generate an access token. You cannot access the device farm or model zoo without an access token.

For our AI Hub's integrations with PySDK, you need to generate an access token.

1

In the navigation bar, click Tokens.

2

Click Generate New Token.

3

Provide a name for the token and optionally set an expiration date.

4

Set this token as an environment variable on your system. It'll give you access to the AI Hub whenever using @cloud.

Prevent abuse of your access token by keeping it private.