Classification

[sctrueew]

Hi, thanks for your work. Is it possible to support classification?

[dme-compunet]

Yes, here is a sample code:

var image = "path/to/image";

using var predictor = new YoloV8("path/to/model");

var result = predictor.Classify(image);

[sctrueew]

@dme-compunet Hi, I exported my trained model with this script

from ultralytics import YOLO model = YOLO(r"..\runs\classify\train19\weights\best.pt") model.export(format='onnx', opset=15, imgsz = 224)
and then I used this code for inference:

` void ClassifyDemo(string image1, string model)
{
Console.WriteLine();
Console.WriteLine("================ CLASSIFICATION DEMO ================");
Console.WriteLine();
Console.WriteLine("Loading model...");
using var predictor = new YoloV8(model);

  Console.WriteLine("Working...");
  var result = predictor.Classify(image);

  Console.WriteLine();

  Console.WriteLine($"Result: {result}");
  Console.WriteLine($"Speed: {result.Speed}");

  Console.WriteLine();

  foreach (var item in result.Probabilities)
  {
      Console.WriteLine($"{item.Class} {item.Confidence}");
  }

}
`

but the results are not correct.

[dme-compunet]

I'm missing details, in the pre-trained model on ImageNet it works great, can you try to understand what the problem is?

[sctrueew]

I tested with the Yolo command:
yolo classify predict model=runs\classify\train32\weights\best.onnx source="..\n02099601_1421.JPEG" imgsz=224

The result is correct, but when I run it with C# code, the output is different.

[xumix]

I tested with the Yolo command: yolo classify predict model=runs\classify\train32\weights\best.onnx source="..\n02099601_1421.JPEG" imgsz=224

The result is correct, but when I run it with C# code, the output is different.

Try sorting by item.Confidence