static void Main(string[] args)
{
// Load Data
IEnumerable <ModelInput> images = ModelInput.LoadImagesFromDirectory(Constants.AssetsRelativePath);
IDataView trainingDataView = mlContext.Data.LoadFromEnumerable(images);
// Build training pipeline
IEstimator <ITransformer> trainingPipeline = BuildTrainingPipeline(mlContext);
// Evaluate quality of Model
Evaluate(mlContext, trainingDataView, trainingPipeline);
// Train Model
ITransformer mlModel = TrainModel(mlContext, trainingDataView, trainingPipeline);
// Save model
if (!Directory.Exists(Constants.WorkspaceRelativePath))
{
Directory.CreateDirectory(Constants.WorkspaceRelativePath);
}
SaveModel(mlContext, mlModel, Constants.ModelPath, trainingDataView.Schema);
Console.WriteLine("Finished task. Press any key to exit.");
Console.ReadKey();
}
static void Main(string[] args)
{
NumberOfLabels = Constants.Labels.Length;
var images = ModelInput.LoadImagesFromDirectory(Constants.TestImagesRelativePath, false);
WriteLine($"Total number of images for testing: {images.Count()}");
confusionMatrix = new int[NumberOfLabels, NumberOfLabels];
labels = new Dictionary <string, int>();
foreach (var label in Constants.Labels)
{
labels.Add(label, Array.IndexOf(Constants.Labels, label));
}
NumberOfImages = 0;
WriteLine("Actual label\tPredicted Label\tScore");
foreach (var image in images)
{
var prediction = ConsumeModel.Predict(image, Constants.ModelPath);
WriteLine($"{image.Label}\t{prediction.Prediction}\t{prediction.Score.Max()}");
// Format: `confusionMatrix[predictedLabel, actualLabel]`
confusionMatrix[labels[prediction.Prediction], labels[image.Label]]++;
NumberOfImages++;
}
WriteLine($"Categorized {NumberOfImages} images in {NumberOfLabels} classes.");
PrintConfusionMatrix();
PrintPrecisionRecall();
}