Is an Object Detection suppose to take 3.5 seconds?

[borakeh]

System Information (please complete the following information):

• OS & Version: Windows 10
• ML.NET Version: v17.17.0
• .NET Version: .NET 8

Describe the bug I just managed to setup my first project and things seem to be working quite well, however I noticed that doing a simple object detection takes about 3.5 seconds. I'm not sure if this is normal due to some PredictEngine overhead or something but it seems quite long considering I'm using a 4090 and the images aren't big.

The model was trained using 10 images and I'm only trying to predict 1 object on 1 image.

DoorDetection.ModelInput sampleData = new DoorDetection.ModelInput()
{
    Image = image,
};

Stopwatch stopwatch = Stopwatch.StartNew();

var predictionResult = DoorDetection.Predict(sampleData);

stopwatch.Stop();
Console.WriteLine(stopwatch.ElapsedMilliseconds);

To Reproduce Steps to reproduce the behavior:

1. Clone my project https://github.com/borakeh/ObjectDetection
2. I guess make sure the images are on the desktop in the correct folder Folder Structure
   • Desktop/doors
   • Desktop/doors/doorsOut
3. Run the project and check how long it takes.

Expected behavior I would assume it would be a bit faster.

Screenshots, Code, Sample Projects No screenshots.

Additional context No additional context.