Update ConsumeModel.cs to enhance it's performance when being called for multiple times

src/Microsoft.ML.CodeGenerator/Templates/Console/ConsumeModel.cs

@@ -39,32 +39,39 @@ public virtual string TransformText()
 {
 	public class ConsumeModel
     {
-        // For more info on consuming ML.NET models, visit https://aka.ms/model-builder-consume
+        private static Lazy<PredictionEngine<ModelInput, ModelOutput>> PredictionEngine = new Lazy<PredictionEngine<ModelInput, ModelOutput>>(CreatePredictionEngine);
+
+        // For more info on consuming ML.NET models, visit https://aka.ms/mlnet-consume
         // Method for consuming model in your app
         public static ModelOutput Predict(ModelInput input)
         {
-
+            ModelOutput result = PredictionEngine.Value.Predict(input);
+            return result;
+        }
+
+        public static PredictionEngine<ModelInput, ModelOutput> CreatePredictionEngine()
+        {
             // Create new MLContext
             MLContext mlContext = new MLContext();
 ");
 if(HasNormalizeMapping){ 
-            this.Write(" \r\n\t\t\t// Register NormalizeMapping\r\n            mlContext.ComponentCatalog.Regist" +
-                    "erAssembly(typeof(NormalizeMapping).Assembly);\r\n");
+            this.Write(" \r\n\t\t\t// Register NormalizeMapping to calculate probabilities for each Label.\r\n  " +
+                    "          mlContext.ComponentCatalog.RegisterAssembly(typeof(NormalizeMapping).A" +
+                    "ssembly);\r\n");
 } 
 if(HasLabelMapping){ 
-            this.Write(" \r\n\t\t\t// Register LabelMapping\r\n            mlContext.ComponentCatalog.RegisterAs" +
-                    "sembly(typeof(LabelMapping).Assembly);\r\n");
+            this.Write(" \r\n\t\t\t// Register LabelMapping to map predicted Labels to their corresponding pro" +
+                    "babilities (likelihood of specified Labels)\r\n            mlContext.ComponentCata" +
+                    "log.RegisterAssembly(typeof(LabelMapping).Assembly);\r\n");
 } 
             this.Write("\r\n            // Load model & create prediction engine\r\n            string modelP" +
                     "ath = @\"");
             this.Write(this.ToStringHelper.ToStringWithCulture(MLNetModelpath));
             this.Write(@""";
             ITransformer mlModel = mlContext.Model.Load(modelPath, out var modelInputSchema);
             var predEngine = mlContext.Model.CreatePredictionEngine<ModelInput, ModelOutput>(mlModel);
-
-            // Use model to make prediction on input data
-            ModelOutput result = predEngine.Predict(input);
-            return result;
+
+            return predEngine;
         }
     }
 }

src/Microsoft.ML.CodeGenerator/Templates/Console/ConsumeModel.tt

@@ -20,30 +20,35 @@ namespace <#= Namespace #>.Model
 {
 	public class ConsumeModel
     {
-        // For more info on consuming ML.NET models, visit https://aka.ms/model-builder-consume
+        private static Lazy<PredictionEngine<ModelInput, ModelOutput>> PredictionEngine = new Lazy<PredictionEngine<ModelInput, ModelOutput>>(CreatePredictionEngine);
+
+        // For more info on consuming ML.NET models, visit https://aka.ms/mlnet-consume
         // Method for consuming model in your app
         public static ModelOutput Predict(ModelInput input)
         {
-
+            ModelOutput result = PredictionEngine.Value.Predict(input);
+            return result;
+        }
+
+        public static PredictionEngine<ModelInput, ModelOutput> CreatePredictionEngine()
+        {
             // Create new MLContext
             MLContext mlContext = new MLContext();
 <#if(HasNormalizeMapping){ #> 
-			// Register NormalizeMapping
+			// Register NormalizeMapping to calculate probabilities for each Label.
             mlContext.ComponentCatalog.RegisterAssembly(typeof(NormalizeMapping).Assembly);
 <#} #>
 <#if(HasLabelMapping){ #> 
-			// Register LabelMapping
+			// Register LabelMapping to map predicted Labels to their corresponding probabilities (likelihood of specified Labels)
             mlContext.ComponentCatalog.RegisterAssembly(typeof(LabelMapping).Assembly);
 <#} #>
 
             // Load model & create prediction engine
             string modelPath = @"<#= MLNetModelpath #>";
             ITransformer mlModel = mlContext.Model.Load(modelPath, out var modelInputSchema);
             var predEngine = mlContext.Model.CreatePredictionEngine<ModelInput, ModelOutput>(mlModel);
-
-            // Use model to make prediction on input data
-            ModelOutput result = predEngine.Predict(input);
-            return result;
+
+            return predEngine;
         }
     }
 }

test/Microsoft.ML.CodeGenerator.Tests/ApprovalTests/ConsoleCodeGeneratorTests.AzureCodeGeneratorTest.ConsumeModel.cs.approved.txt

@@ -11,25 +11,30 @@ namespace Test.Model
 {
     public class ConsumeModel
     {
-        // For more info on consuming ML.NET models, visit https://aka.ms/model-builder-consume
+        private static Lazy<PredictionEngine<ModelInput, ModelOutput>> PredictionEngine = new Lazy<PredictionEngine<ModelInput, ModelOutput>>(CreatePredictionEngine);
+
+        // For more info on consuming ML.NET models, visit https://aka.ms/mlnet-consume
         // Method for consuming model in your app
         public static ModelOutput Predict(ModelInput input)
         {
+            ModelOutput result = PredictionEngine.Value.Predict(input);
+            return result;
+        }
 
+        public static PredictionEngine<ModelInput, ModelOutput> CreatePredictionEngine()
+        {
             // Create new MLContext
             MLContext mlContext = new MLContext();
 
-            // Register LabelMapping
+            // Register LabelMapping to map predicted Labels to their corresponding probabilities (likelihood of specified Labels)
             mlContext.ComponentCatalog.RegisterAssembly(typeof(LabelMapping).Assembly);
 
             // Load model & create prediction engine
             string modelPath = @"\path\to\model";
             ITransformer mlModel = mlContext.Model.Load(modelPath, out var modelInputSchema);
             var predEngine = mlContext.Model.CreatePredictionEngine<ModelInput, ModelOutput>(mlModel);
 
-            // Use model to make prediction on input data
-            ModelOutput result = predEngine.Predict(input);
-            return result;
+            return predEngine;
         }
     }
 }

test/Microsoft.ML.CodeGenerator.Tests/ApprovalTests/ConsoleCodeGeneratorTests.AzureImageCodeGeneratorTest.ConsumeModel.cs.approved.txt

@@ -11,28 +11,33 @@ namespace CodeGenTest.Model
 {
     public class ConsumeModel
     {
-        // For more info on consuming ML.NET models, visit https://aka.ms/model-builder-consume
+        private static Lazy<PredictionEngine<ModelInput, ModelOutput>> PredictionEngine = new Lazy<PredictionEngine<ModelInput, ModelOutput>>(CreatePredictionEngine);
+
+        // For more info on consuming ML.NET models, visit https://aka.ms/mlnet-consume
         // Method for consuming model in your app
         public static ModelOutput Predict(ModelInput input)
         {
+            ModelOutput result = PredictionEngine.Value.Predict(input);
+            return result;
+        }
 
+        public static PredictionEngine<ModelInput, ModelOutput> CreatePredictionEngine()
+        {
             // Create new MLContext
             MLContext mlContext = new MLContext();
 
-            // Register NormalizeMapping
+            // Register NormalizeMapping to calculate probabilities for each Label.
             mlContext.ComponentCatalog.RegisterAssembly(typeof(NormalizeMapping).Assembly);
 
-            // Register LabelMapping
+            // Register LabelMapping to map predicted Labels to their corresponding probabilities (likelihood of specified Labels)
             mlContext.ComponentCatalog.RegisterAssembly(typeof(LabelMapping).Assembly);
 
             // Load model & create prediction engine
             string modelPath = @"/path/to/model";
             ITransformer mlModel = mlContext.Model.Load(modelPath, out var modelInputSchema);
             var predEngine = mlContext.Model.CreatePredictionEngine<ModelInput, ModelOutput>(mlModel);
 
-            // Use model to make prediction on input data
-            ModelOutput result = predEngine.Predict(input);
-            return result;
+            return predEngine;
         }
     }
 }

test/Microsoft.ML.CodeGenerator.Tests/ApprovalTests/ConsoleCodeGeneratorTests.ConsumeModelContentTest.approved.txt

@@ -15,11 +15,18 @@ namespace TestNamespace.Model
 {
     public class ConsumeModel
     {
-        // For more info on consuming ML.NET models, visit https://aka.ms/model-builder-consume
+        private static Lazy<PredictionEngine<ModelInput, ModelOutput>> PredictionEngine = new Lazy<PredictionEngine<ModelInput, ModelOutput>>(CreatePredictionEngine);
+
+        // For more info on consuming ML.NET models, visit https://aka.ms/mlnet-consume
         // Method for consuming model in your app
         public static ModelOutput Predict(ModelInput input)
         {
+            ModelOutput result = PredictionEngine.Value.Predict(input);
+            return result;
+        }
 
+        public static PredictionEngine<ModelInput, ModelOutput> CreatePredictionEngine()
+        {
             // Create new MLContext
             MLContext mlContext = new MLContext();
 
@@ -28,9 +35,7 @@ namespace TestNamespace.Model
             ITransformer mlModel = mlContext.Model.Load(modelPath, out var modelInputSchema);
             var predEngine = mlContext.Model.CreatePredictionEngine<ModelInput, ModelOutput>(mlModel);
 
-            // Use model to make prediction on input data
-            ModelOutput result = predEngine.Predict(input);
-            return result;
+            return predEngine;
         }
     }
 }

test/Microsoft.ML.CodeGenerator.Tests/ApprovalTests/ConsoleCodeGeneratorTests.Recommendation_GenerateModelProjectContents_VerifyConsumeModel.approved.txt

@@ -15,11 +15,18 @@ namespace TestNamespace.Model
 {
     public class ConsumeModel
     {
-        // For more info on consuming ML.NET models, visit https://aka.ms/model-builder-consume
+        private static Lazy<PredictionEngine<ModelInput, ModelOutput>> PredictionEngine = new Lazy<PredictionEngine<ModelInput, ModelOutput>>(CreatePredictionEngine);
+
+        // For more info on consuming ML.NET models, visit https://aka.ms/mlnet-consume
         // Method for consuming model in your app
         public static ModelOutput Predict(ModelInput input)
         {
+            ModelOutput result = PredictionEngine.Value.Predict(input);
+            return result;
+        }
 
+        public static PredictionEngine<ModelInput, ModelOutput> CreatePredictionEngine()
+        {
             // Create new MLContext
             MLContext mlContext = new MLContext();
 
@@ -28,9 +35,7 @@ namespace TestNamespace.Model
             ITransformer mlModel = mlContext.Model.Load(modelPath, out var modelInputSchema);
             var predEngine = mlContext.Model.CreatePredictionEngine<ModelInput, ModelOutput>(mlModel);
 
-            // Use model to make prediction on input data
-            ModelOutput result = predEngine.Predict(input);
-            return result;
+            return predEngine;
         }
     }
 }

test/Microsoft.ML.CodeGenerator.Tests/ApprovalTests/TemplateTest.TestConsumeModel.approved.txt

@@ -0,0 +1,43 @@
+// This file was auto-generated by ML.NET Model Builder. 
+
+using System;
+using System.Collections.Generic;
+using System.Linq;
+using System.Text;
+using Microsoft.ML;
+using Namespace.Model;
+
+namespace Namespace.Model
+{
+	public class ConsumeModel
+    {
+        private static Lazy<PredictionEngine<ModelInput, ModelOutput>> PredictionEngine = new Lazy<PredictionEngine<ModelInput, ModelOutput>>(CreatePredictionEngine);
+
+        // For more info on consuming ML.NET models, visit https://aka.ms/mlnet-consume
+        // Method for consuming model in your app
+        public static ModelOutput Predict(ModelInput input)
+        {
+            ModelOutput result = PredictionEngine.Value.Predict(input);
+            return result;
+        }
+
+        public static PredictionEngine<ModelInput, ModelOutput> CreatePredictionEngine()
+        {
+            // Create new MLContext
+            MLContext mlContext = new MLContext();
+
+			// Register NormalizeMapping to calculate probabilities for each Label.
+            mlContext.ComponentCatalog.RegisterAssembly(typeof(NormalizeMapping).Assembly);
+
+			// Register LabelMapping to map predicted Labels to their corresponding probabilities (likelihood of specified Labels)
+            mlContext.ComponentCatalog.RegisterAssembly(typeof(LabelMapping).Assembly);
+
+            // Load model & create prediction engine
+            string modelPath = @"/path/to/model";
+            ITransformer mlModel = mlContext.Model.Load(modelPath, out var modelInputSchema);
+            var predEngine = mlContext.Model.CreatePredictionEngine<ModelInput, ModelOutput>(mlModel);
+
+            return predEngine;
+        }
+    }
+}

test/Microsoft.ML.CodeGenerator.Tests/ApprovalTests/TemplateTest.cs

@@ -39,5 +39,21 @@ public void TestPredictProgram_WithSampleData()
             };
             Approvals.Verify(predictProgram.TransformText());
         }
+
+        [Fact]
+        [UseReporter(typeof(DiffReporter))]
+        [MethodImpl(MethodImplOptions.NoInlining)]
+        public void TestConsumeModel()
+        {
+            var consumeModel = new ConsumeModel()
+            {
+                Namespace = "Namespace",
+                HasNormalizeMapping = true,
+                HasLabelMapping = true,
+                MLNetModelpath = @"/path/to/model",
+            };
+
+            Approvals.Verify(consumeModel.TransformText());
+        }
     }
 }