CustomVisionOnnx won't work properly after upgrading to ML. NET 2.0

[BigHeadDev]

I can confirm that the problem is with the nuget package Microsoft.ML.ImageAnalytics Due to some changes in its dependency during the process from 1. x to 2.0, it no longer relies on System.Drawing.Common

The following code snippet shows the key errors：

public class WineInput {// C# Input Mapping
        [ImageType(ImageSettings.imageWidth, ImageSettings.imageHeight)]
        public Bitmap Image { get; set; }
    }
    public struct ImageSettings {
        public const int imageHeight = 416;
        public const int imageWidth = 416;
    }

MLContext context = new MLContext();
            var emptyData = new List<WineInput>();
            var data = context.Data.LoadFromEnumerable(emptyData);// Error!

How should I modify the code to use Microsoft.ML.ImageAnalytics 2.0

Thank you very much for the video posted on Youtube and learning a lot~I hope you can help me solve this problem 😁

[jwood803]

Sorry for the late reply. I think 2.0 removed some image packages in favor of some that are more cross platform. I do have a plan to make a video that updates the Custom Vision Object Detection scenario where you build an object detection model in Custom Vision and use it in ML.NET. Hope that will help!