Browser Inference

Leverage DeGirum’s Browser Inference to run models directly in your web browser – upload inputs, view real-time results, and explore model details effortlessly.

As soon as you log into AI Hub, you can perform inferences on various hardware configurations and models, enabling rapid experimentation without any setup required.

The Browser Inference GUI

In the Browser Inference GUI, you may upload an input file, run an inference on that input file, view the source code for this inference, view the model JSON, view labels, read the Model Readme, and more.

When you view specific models in the AI Hub, you will be presented with a GUI to run an inference in your browser.

Using the Browser Inference

1

Upload an Input Image

Click the Input File button on the left side to upload your image. When you upload an input file with this image, the left side will show your input image.

2

Running Inference

After uploading your image, click Run Inference on the right side to run inference on the input image.

3

Results Display

On the right side, the car is highlighted with a yellow bounding box, and a confidence score (0.92) is displayed. We observe an inference duration of 14.51 milliseconds.

In our example, we use a DeGirum Orca1 with an N2X runtime to verify that the detected object is indeed a car.