Modify API for advanced settings (several learners)

src/Microsoft.ML.StandardLearners/Properties/AssemblyInfo.cs

@@ -5,6 +5,7 @@
 using System.Runtime.CompilerServices;
 using Microsoft.ML;
 
+[assembly: InternalsVisibleTo(assemblyName: "Microsoft.ML.Ensemble" + PublicKey.Value)]
 [assembly: InternalsVisibleTo(assemblyName: "Microsoft.ML.StaticPipe" + PublicKey.Value)]
 
 [assembly: InternalsVisibleTo(assemblyName: "RunTests" + InternalPublicKey.Value)]

src/Microsoft.ML.StandardLearners/Standard/LogisticRegression/LbfgsPredictorBase.cs

@@ -156,7 +156,6 @@ internal LbfgsTrainerBase(IHostEnvironment env,
             string featureColumn,
             SchemaShape.Column labelColumn,
             string weightColumn,
-            Action<TArgs> advancedSettings,
             float l1Weight,
             float l2Weight,
             float optimizationTolerance,
@@ -173,7 +172,7 @@ internal LbfgsTrainerBase(IHostEnvironment env,
                             MemorySize = memorySize,
                             EnforceNonNegativity = enforceNoNegativity
                         },
-                  labelColumn, advancedSettings)
+                  labelColumn)
         {
         }
 

src/Microsoft.ML.StandardLearners/Standard/LogisticRegression/LogisticRegression.cs

@@ -16,7 +16,7 @@
 using Microsoft.ML.Numeric;
 using Microsoft.ML.Training;
 
-[assembly: LoadableClass(LogisticRegression.Summary, typeof(LogisticRegression), typeof(LogisticRegression.Arguments),
+[assembly: LoadableClass(LogisticRegression.Summary, typeof(LogisticRegression), typeof(LogisticRegression.Options),
     new[] { typeof(SignatureBinaryClassifierTrainer), typeof(SignatureTrainer), typeof(SignatureFeatureScorerTrainer) },
     LogisticRegression.UserNameValue,
     LogisticRegression.LoadNameValue,
@@ -30,15 +30,15 @@ namespace Microsoft.ML.Learners
 
     /// <include file='doc.xml' path='doc/members/member[@name="LBFGS"]/*' />
     /// <include file='doc.xml' path='docs/members/example[@name="LogisticRegressionBinaryClassifier"]/*' />
-    public sealed partial class LogisticRegression : LbfgsTrainerBase<LogisticRegression.Arguments, BinaryPredictionTransformer<ParameterMixingCalibratedPredictor>, ParameterMixingCalibratedPredictor>
+    public sealed partial class LogisticRegression : LbfgsTrainerBase<LogisticRegression.Options, BinaryPredictionTransformer<ParameterMixingCalibratedPredictor>, ParameterMixingCalibratedPredictor>
     {
         public const string LoadNameValue = "LogisticRegression";
         internal const string UserNameValue = "Logistic Regression";
         internal const string ShortName = "lr";
         internal const string Summary = "Logistic Regression is a method in statistics used to predict the probability of occurrence of an event and can "
             + "be used as a classification algorithm. The algorithm predicts the probability of occurrence of an event by fitting data to a logistical function.";
 
-        public sealed class Arguments : ArgumentsBase
+        public sealed class Options : ArgumentsBase
         {
             /// <summary>
             /// If set to <value>true</value>training statistics will be generated at the end of training.
@@ -53,7 +53,7 @@ public sealed class Arguments : ArgumentsBase
             /// <summary>
             /// The instance of <see cref="ComputeLRTrainingStd"/> that computes the std of the training statistics, at the end of training.
             /// The calculations are not part of Microsoft.ML package, due to the size of MKL.
-            /// If you need these calculations, add the Microsoft.ML.HalLearners package, and initialize <see cref="LogisticRegression.Arguments.StdComputer"/>.
+            /// If you need these calculations, add the Microsoft.ML.HalLearners package, and initialize <see cref="LogisticRegression.Options.StdComputer"/>.
             /// to the <see cref="ComputeLRTrainingStd"/> implementation in the Microsoft.ML.HalLearners package.
             /// </summary>
             public ComputeLRTrainingStd StdComputer;
@@ -74,18 +74,16 @@ public sealed class Arguments : ArgumentsBase
         /// <param name="l2Weight">Weight of L2 regularizer term.</param>
         /// <param name="memorySize">Memory size for <see cref="LogisticRegression"/>. Low=faster, less accurate.</param>
         /// <param name="optimizationTolerance">Threshold for optimizer convergence.</param>
-        /// <param name="advancedSettings">A delegate to apply all the advanced arguments to the algorithm.</param>
-        public LogisticRegression(IHostEnvironment env,
+        internal LogisticRegression(IHostEnvironment env,
             string labelColumn = DefaultColumnNames.Label,
             string featureColumn = DefaultColumnNames.Features,
             string weights = null,
-            float l1Weight = Arguments.Defaults.L1Weight,
-            float l2Weight = Arguments.Defaults.L2Weight,
-            float optimizationTolerance = Arguments.Defaults.OptTol,
-            int memorySize = Arguments.Defaults.MemorySize,
-            bool enforceNoNegativity = Arguments.Defaults.EnforceNonNegativity,
-            Action<Arguments> advancedSettings = null)
-            : base(env, featureColumn, TrainerUtils.MakeBoolScalarLabel(labelColumn), weights, advancedSettings,
+            float l1Weight = Options.Defaults.L1Weight,
+            float l2Weight = Options.Defaults.L2Weight,
+            float optimizationTolerance = Options.Defaults.OptTol,
+            int memorySize = Options.Defaults.MemorySize,
+            bool enforceNoNegativity = Options.Defaults.EnforceNonNegativity)
+            : base(env, featureColumn, TrainerUtils.MakeBoolScalarLabel(labelColumn), weights,
                   l1Weight, l2Weight, optimizationTolerance, memorySize, enforceNoNegativity)
         {
             Host.CheckNonEmpty(featureColumn, nameof(featureColumn));
@@ -98,8 +96,8 @@ public LogisticRegression(IHostEnvironment env,
         /// <summary>
         /// Initializes a new instance of <see cref="LogisticRegression"/>
         /// </summary>
-        internal LogisticRegression(IHostEnvironment env, Arguments args)
-            : base(env, args, TrainerUtils.MakeBoolScalarLabel(args.LabelColumn))
+        internal LogisticRegression(IHostEnvironment env, Options options)
+            : base(env, options, TrainerUtils.MakeBoolScalarLabel(options.LabelColumn))
         {
             _posWeight = 0;
             ShowTrainingStats = Args.ShowTrainingStats;
@@ -410,14 +408,14 @@ protected override ParameterMixingCalibratedPredictor CreatePredictor()
             XmlInclude = new[] { @"<include file='../Microsoft.ML.StandardLearners/Standard/LogisticRegression/doc.xml' path='doc/members/member[@name=""LBFGS""]/*' />",
                                  @"<include file='../Microsoft.ML.StandardLearners/Standard/LogisticRegression/doc.xml' path='doc/members/example[@name=""LogisticRegressionBinaryClassifier""]/*' />"})]
 
-        public static CommonOutputs.BinaryClassificationOutput TrainBinary(IHostEnvironment env, Arguments input)
+        public static CommonOutputs.BinaryClassificationOutput TrainBinary(IHostEnvironment env, Options input)
         {
             Contracts.CheckValue(env, nameof(env));
             var host = env.Register("TrainLRBinary");
             host.CheckValue(input, nameof(input));
             EntryPointUtils.CheckInputArgs(host, input);
 
-            return LearnerEntryPointsUtils.Train<Arguments, CommonOutputs.BinaryClassificationOutput>(host, input,
+            return LearnerEntryPointsUtils.Train<Options, CommonOutputs.BinaryClassificationOutput>(host, input,
                 () => new LogisticRegression(host, input),
                 () => LearnerEntryPointsUtils.FindColumn(host, input.TrainingData.Schema, input.LabelColumn),
                 () => LearnerEntryPointsUtils.FindColumn(host, input.TrainingData.Schema, input.WeightColumn));
@@ -437,7 +435,7 @@ public abstract class ComputeLRTrainingStd
         /// Computes the standard deviation matrix of each of the non-zero training weights, needed to calculate further the standard deviation,
         /// p-value and z-Score.
         /// The calculations are not part of Microsoft.ML package, due to the size of MKL.
-        /// If you need these calculations, add the Microsoft.ML.HalLearners package, and initialize <see cref="LogisticRegression.Arguments.StdComputer"/>
+        /// If you need these calculations, add the Microsoft.ML.HalLearners package, and initialize <see cref="LogisticRegression.Options.StdComputer"/>
         /// to the <see cref="ComputeLRTrainingStd"/> implementation in the Microsoft.ML.HalLearners package.
         /// Due to the existence of regularization, an approximation is used to compute the variances of the trained linear coefficients.
         /// </summary>

src/Microsoft.ML.StandardLearners/Standard/LogisticRegression/MulticlassLogisticRegression.cs

@@ -22,7 +22,7 @@
 using Microsoft.ML.Training;
 using Newtonsoft.Json.Linq;
 
-[assembly: LoadableClass(typeof(MulticlassLogisticRegression), typeof(MulticlassLogisticRegression.Arguments),
+[assembly: LoadableClass(typeof(MulticlassLogisticRegression), typeof(MulticlassLogisticRegression.Options),
     new[] { typeof(SignatureMultiClassClassifierTrainer), typeof(SignatureTrainer) },
     MulticlassLogisticRegression.UserNameValue,
     MulticlassLogisticRegression.LoadNameValue,
@@ -38,14 +38,14 @@ namespace Microsoft.ML.Learners
 {
     /// <include file = 'doc.xml' path='doc/members/member[@name="LBFGS"]/*' />
     /// <include file = 'doc.xml' path='docs/members/example[@name="LogisticRegressionClassifier"]/*' />
-    public sealed class MulticlassLogisticRegression : LbfgsTrainerBase<MulticlassLogisticRegression.Arguments,
+    public sealed class MulticlassLogisticRegression : LbfgsTrainerBase<MulticlassLogisticRegression.Options,
         MulticlassPredictionTransformer<MulticlassLogisticRegressionModelParameters>, MulticlassLogisticRegressionModelParameters>
     {
         public const string LoadNameValue = "MultiClassLogisticRegression";
         internal const string UserNameValue = "Multi-class Logistic Regression";
         internal const string ShortName = "mlr";
 
-        public sealed class Arguments : ArgumentsBase
+        public sealed class Options : ArgumentsBase
         {
             [Argument(ArgumentType.AtMostOnce, HelpText = "Show statistics of training examples.", ShortName = "stat", SortOrder = 50)]
             public bool ShowTrainingStats = false;
@@ -82,19 +82,16 @@ public sealed class Arguments : ArgumentsBase
         /// <param name="l2Weight">Weight of L2 regularizer term.</param>
         /// <param name="memorySize">Memory size for <see cref="LogisticRegression"/>. Low=faster, less accurate.</param>
         /// <param name="optimizationTolerance">Threshold for optimizer convergence.</param>
-        /// <param name="advancedSettings">A delegate to apply all the advanced arguments to the algorithm.</param>
-        public MulticlassLogisticRegression(IHostEnvironment env,
+        internal MulticlassLogisticRegression(IHostEnvironment env,
             string labelColumn = DefaultColumnNames.Label,
             string featureColumn = DefaultColumnNames.Features,
             string weights = null,
-            float l1Weight = Arguments.Defaults.L1Weight,
-            float l2Weight = Arguments.Defaults.L2Weight,
-            float optimizationTolerance = Arguments.Defaults.OptTol,
-            int memorySize = Arguments.Defaults.MemorySize,
-            bool enforceNoNegativity = Arguments.Defaults.EnforceNonNegativity,
-            Action<Arguments> advancedSettings = null)
-            : base(env, featureColumn, TrainerUtils.MakeU4ScalarColumn(labelColumn), weights, advancedSettings,
-                  l1Weight, l2Weight, optimizationTolerance, memorySize, enforceNoNegativity)
+            float l1Weight = Options.Defaults.L1Weight,
+            float l2Weight = Options.Defaults.L2Weight,
+            float optimizationTolerance = Options.Defaults.OptTol,
+            int memorySize = Options.Defaults.MemorySize,
+            bool enforceNoNegativity = Options.Defaults.EnforceNonNegativity)
+            : base(env, featureColumn, TrainerUtils.MakeU4ScalarColumn(labelColumn), weights, l1Weight, l2Weight, optimizationTolerance, memorySize, enforceNoNegativity)
         {
             Host.CheckNonEmpty(featureColumn, nameof(featureColumn));
             Host.CheckNonEmpty(labelColumn, nameof(labelColumn));
@@ -105,8 +102,8 @@ public MulticlassLogisticRegression(IHostEnvironment env,
         /// <summary>
         /// Initializes a new instance of <see cref="MulticlassLogisticRegression"/>
         /// </summary>
-        internal MulticlassLogisticRegression(IHostEnvironment env, Arguments args)
-            : base(env, args, TrainerUtils.MakeU4ScalarColumn(args.LabelColumn))
+        internal MulticlassLogisticRegression(IHostEnvironment env, Options options)
+            : base(env, options, TrainerUtils.MakeU4ScalarColumn(options.LabelColumn))
         {
             ShowTrainingStats = Args.ShowTrainingStats;
         }
@@ -1007,14 +1004,14 @@ public partial class LogisticRegression
             ShortName = MulticlassLogisticRegression.ShortName,
             XmlInclude = new[] { @"<include file='../Microsoft.ML.StandardLearners/Standard/LogisticRegression/doc.xml' path='doc/members/member[@name=""LBFGS""]/*' />",
                                  @"<include file='../Microsoft.ML.StandardLearners/Standard/LogisticRegression/doc.xml' path='doc/members/example[@name=""LogisticRegressionClassifier""]/*' />" })]
-        public static CommonOutputs.MulticlassClassificationOutput TrainMultiClass(IHostEnvironment env, MulticlassLogisticRegression.Arguments input)
+        public static CommonOutputs.MulticlassClassificationOutput TrainMultiClass(IHostEnvironment env, MulticlassLogisticRegression.Options input)
         {
             Contracts.CheckValue(env, nameof(env));
             var host = env.Register("TrainLRMultiClass");
             host.CheckValue(input, nameof(input));
             EntryPointUtils.CheckInputArgs(host, input);
 
-            return LearnerEntryPointsUtils.Train<MulticlassLogisticRegression.Arguments, CommonOutputs.MulticlassClassificationOutput>(host, input,
+            return LearnerEntryPointsUtils.Train<MulticlassLogisticRegression.Options, CommonOutputs.MulticlassClassificationOutput>(host, input,
                 () => new MulticlassLogisticRegression(host, input),
                 () => LearnerEntryPointsUtils.FindColumn(host, input.TrainingData.Schema, input.LabelColumn),
                 () => LearnerEntryPointsUtils.FindColumn(host, input.TrainingData.Schema, input.WeightColumn));

src/Microsoft.ML.StandardLearners/Standard/MultiClass/MetaMulticlassTrainer.cs

@@ -87,7 +87,7 @@ private TScalarTrainer CreateTrainer()
         {
             return Args.PredictorType != null ?
                 Args.PredictorType.CreateComponent(Host) :
-                new LinearSvmTrainer(Host, new LinearSvmTrainer.Arguments());
+                new LinearSvmTrainer(Host, new LinearSvmTrainer.Options());
         }
 
         private protected IDataView MapLabelsCore<T>(ColumnType type, InPredicate<T> equalsTarget, RoleMappedData data)

src/Microsoft.ML.StandardLearners/Standard/Online/AveragedPerceptron.cs

@@ -15,7 +15,7 @@
 using Microsoft.ML.Trainers.Online;
 using Microsoft.ML.Training;
 
-[assembly: LoadableClass(AveragedPerceptronTrainer.Summary, typeof(AveragedPerceptronTrainer), typeof(AveragedPerceptronTrainer.Arguments),
+[assembly: LoadableClass(AveragedPerceptronTrainer.Summary, typeof(AveragedPerceptronTrainer), typeof(AveragedPerceptronTrainer.Options),
     new[] { typeof(SignatureBinaryClassifierTrainer), typeof(SignatureTrainer), typeof(SignatureFeatureScorerTrainer) },
     AveragedPerceptronTrainer.UserNameValue,
     AveragedPerceptronTrainer.LoadNameValue, "avgper", AveragedPerceptronTrainer.ShortName)]
@@ -37,9 +37,9 @@ public sealed class AveragedPerceptronTrainer : AveragedLinearTrainer<BinaryPred
         internal const string ShortName = "ap";
         internal const string Summary = "Averaged Perceptron Binary Classifier.";
 
-        private readonly Arguments _args;
+        private readonly Options _args;
 
-        public sealed class Arguments : AveragedLinearArguments
+        public sealed class Options : AveragedLinearArguments
         {
             [Argument(ArgumentType.Multiple, HelpText = "Loss Function", ShortName = "loss", SortOrder = 50)]
             public ISupportClassificationLossFactory LossFunction = new HingeLoss.Arguments();
@@ -83,10 +83,10 @@ public override LinearBinaryModelParameters CreatePredictor()
             }
         }
 
-        internal AveragedPerceptronTrainer(IHostEnvironment env, Arguments args)
-            : base(args, env, UserNameValue, TrainerUtils.MakeBoolScalarLabel(args.LabelColumn))
+        internal AveragedPerceptronTrainer(IHostEnvironment env, Options options)
+            : base(options, env, UserNameValue, TrainerUtils.MakeBoolScalarLabel(options.LabelColumn))
         {
-            _args = args;
+            _args = options;
             LossFunction = _args.LossFunction.CreateComponent(env);
         }
 
@@ -103,18 +103,16 @@ internal AveragedPerceptronTrainer(IHostEnvironment env, Arguments args)
         /// <param name="decreaseLearningRate">Wheather to decrease learning rate as iterations progress.</param>
         /// <param name="l2RegularizerWeight">L2 Regularization Weight.</param>
         /// <param name="numIterations">The number of training iteraitons.</param>
-        /// <param name="advancedSettings">A delegate to supply more advanced arguments to the algorithm.</param>
-        public AveragedPerceptronTrainer(IHostEnvironment env,
+        internal AveragedPerceptronTrainer(IHostEnvironment env,
             string labelColumn = DefaultColumnNames.Label,
             string featureColumn = DefaultColumnNames.Features,
             string weights = null,
             IClassificationLoss lossFunction = null,
-            float learningRate = Arguments.AveragedDefaultArgs.LearningRate,
-            bool decreaseLearningRate = Arguments.AveragedDefaultArgs.DecreaseLearningRate,
-            float l2RegularizerWeight = Arguments.AveragedDefaultArgs.L2RegularizerWeight,
-            int numIterations = Arguments.AveragedDefaultArgs.NumIterations,
-            Action<Arguments> advancedSettings = null)
-            : this(env, InvokeAdvanced(advancedSettings, new Arguments
+            float learningRate = Options.AveragedDefaultArgs.LearningRate,
+            bool decreaseLearningRate = Options.AveragedDefaultArgs.DecreaseLearningRate,
+            float l2RegularizerWeight = Options.AveragedDefaultArgs.L2RegularizerWeight,
+            int numIterations = Options.AveragedDefaultArgs.NumIterations)
+            : this(env, new Options
             {
                 LabelColumn = labelColumn,
                 FeatureColumn = featureColumn,
@@ -124,7 +122,7 @@ public AveragedPerceptronTrainer(IHostEnvironment env,
                 L2RegularizerWeight = l2RegularizerWeight,
                 NumIterations = numIterations,
                 LossFunction = new TrivialFactory(lossFunction ?? new HingeLoss())
-            }))
+            })
         {
         }
 
@@ -191,14 +189,14 @@ public BinaryPredictionTransformer<LinearBinaryModelParameters> Train(IDataView
              ShortName = ShortName,
              XmlInclude = new[] { @"<include file='../Microsoft.ML.StandardLearners/Standard/Online/doc.xml' path='doc/members/member[@name=""AP""]/*' />",
                                  @"<include file='../Microsoft.ML.StandardLearners/Standard/Online/doc.xml' path='doc/members/example[@name=""AP""]/*' />"})]
-        public static CommonOutputs.BinaryClassificationOutput TrainBinary(IHostEnvironment env, Arguments input)
+        public static CommonOutputs.BinaryClassificationOutput TrainBinary(IHostEnvironment env, Options input)
         {
             Contracts.CheckValue(env, nameof(env));
             var host = env.Register("TrainAP");
             host.CheckValue(input, nameof(input));
             EntryPointUtils.CheckInputArgs(host, input);
 
-            return LearnerEntryPointsUtils.Train<Arguments, CommonOutputs.BinaryClassificationOutput>(host, input,
+            return LearnerEntryPointsUtils.Train<Options, CommonOutputs.BinaryClassificationOutput>(host, input,
                 () => new AveragedPerceptronTrainer(host, input),
                 () => LearnerEntryPointsUtils.FindColumn(host, input.TrainingData.Schema, input.LabelColumn),
                 calibrator: input.Calibrator, maxCalibrationExamples: input.MaxCalibrationExamples);