StochasticGradientDescentClassificationTrainer

Author: abgoswam

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

@@ -31,7 +31,7 @@
     "lc",
     "sasdca")]
 
-[assembly: LoadableClass(typeof(StochasticGradientDescentClassificationTrainer), typeof(StochasticGradientDescentClassificationTrainer.Arguments),
+[assembly: LoadableClass(typeof(StochasticGradientDescentClassificationTrainer), typeof(StochasticGradientDescentClassificationTrainer.Options),
     new[] { typeof(SignatureBinaryClassifierTrainer), typeof(SignatureTrainer), typeof(SignatureFeatureScorerTrainer) },
     StochasticGradientDescentClassificationTrainer.UserNameValue,
     StochasticGradientDescentClassificationTrainer.LoadNameValue,
@@ -69,6 +69,13 @@ private protected LinearTrainerBase(IHostEnvironment env, string featureColumn,
         {
         }
 
+        private protected LinearTrainerBase(IHostEnvironment env, string featureColumn, SchemaShape.Column labelColumn,
+            SchemaShape.Column weightColumn)
+            : base(Contracts.CheckRef(env, nameof(env)).Register(RegisterName), TrainerUtils.MakeR4VecFeature(featureColumn),
+                  labelColumn, weightColumn)
+        {
+        }
+
         private protected override TModel TrainModelCore(TrainContext context)
         {
             Host.CheckValue(context, nameof(context));
@@ -1595,7 +1602,7 @@ public sealed class StochasticGradientDescentClassificationTrainer :
         internal const string UserNameValue = "Hogwild SGD (binary)";
         internal const string ShortName = "HogwildSGD";
 
-        public sealed class Arguments : LearnerInputBaseWithWeight
+        public sealed class Options : LearnerInputBaseWithWeight
         {
             [Argument(ArgumentType.Multiple, HelpText = "Loss Function", ShortName = "loss", SortOrder = 50)]
             public ISupportClassificationLossFactory LossFunction = new LogLossFactory();
@@ -1670,7 +1677,7 @@ internal static class Defaults
         }
 
         private readonly IClassificationLoss _loss;
-        private readonly Arguments _args;
+        private readonly Options _args;
 
         protected override bool ShuffleData => _args.Shuffle;
 
@@ -1689,29 +1696,24 @@ internal static class Defaults
         /// <param name="initLearningRate">The initial learning rate used by SGD.</param>
         /// <param name="l2Weight">The L2 regularizer constant.</param>
         /// <param name="loss">The loss function to use.</param>
-        /// <param name="advancedSettings">A delegate to apply all the advanced arguments to the algorithm.</param>
-        public StochasticGradientDescentClassificationTrainer(IHostEnvironment env,
+        internal StochasticGradientDescentClassificationTrainer(IHostEnvironment env,
             string labelColumn = DefaultColumnNames.Label,
             string featureColumn = DefaultColumnNames.Features,
             string weightColumn = null,
-            int maxIterations = Arguments.Defaults.MaxIterations,
-            double initLearningRate = Arguments.Defaults.InitLearningRate,
-            float l2Weight = Arguments.Defaults.L2Weight,
-            ISupportClassificationLossFactory loss = null,
-            Action<Arguments> advancedSettings = null)
+            int maxIterations = Options.Defaults.MaxIterations,
+            double initLearningRate = Options.Defaults.InitLearningRate,
+            float l2Weight = Options.Defaults.L2Weight,
+            ISupportClassificationLossFactory loss = null)
             : base(env, featureColumn, TrainerUtils.MakeBoolScalarLabel(labelColumn), weightColumn)
         {
             Host.CheckNonEmpty(featureColumn, nameof(featureColumn));
             Host.CheckNonEmpty(labelColumn, nameof(labelColumn));
 
-            _args = new Arguments();
+            _args = new Options();
             _args.MaxIterations = maxIterations;
             _args.InitLearningRate = initLearningRate;
             _args.L2Weight = l2Weight;
 
-            // Apply the advanced args, if the user supplied any.
-            advancedSettings?.Invoke(_args);
-
             _args.FeatureColumn = featureColumn;
             _args.LabelColumn = labelColumn;
             _args.WeightColumn = weightColumn;
@@ -1728,8 +1730,10 @@ public StochasticGradientDescentClassificationTrainer(IHostEnvironment env,
         /// <summary>
         /// Initializes a new instance of <see cref="StochasticGradientDescentClassificationTrainer"/>
         /// </summary>
-        internal StochasticGradientDescentClassificationTrainer(IHostEnvironment env, Arguments args)
-            : base(env, args.FeatureColumn, TrainerUtils.MakeBoolScalarLabel(args.LabelColumn), args.WeightColumn)
+        /// <param name="env">The environment to use.</param>
+        /// <param name="args">Advanced arguments to the algorithm.</param>
+        internal StochasticGradientDescentClassificationTrainer(IHostEnvironment env, Options args)
+            : base(env, args.FeatureColumn, TrainerUtils.MakeBoolScalarLabel(args.LabelColumn), TrainerUtils.MakeR4ScalarWeightColumn(args.WeightColumn, args.WeightColumn.IsExplicit))
         {
             args.Check(env);
             _loss = args.LossFunction.CreateComponent(env);
@@ -1948,14 +1952,14 @@ private protected override void CheckLabel(RoleMappedData examples, out int weig
         }
 
         [TlcModule.EntryPoint(Name = "Trainers.StochasticGradientDescentBinaryClassifier", Desc = "Train an Hogwild SGD binary model.", UserName = UserNameValue, ShortName = ShortName)]
-        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("TrainHogwildSGD");
             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 StochasticGradientDescentClassificationTrainer(host, input),
                 () => LearnerEntryPointsUtils.FindColumn(host, input.TrainingData.Schema, input.LabelColumn),
                 () => LearnerEntryPointsUtils.FindColumn(host, input.TrainingData.Schema, input.WeightColumn),

src/Microsoft.ML.StandardLearners/StandardLearnersCatalog.cs

@@ -13,7 +13,7 @@
 namespace Microsoft.ML
 {
     using LRArguments = LogisticRegression.Arguments;
-    using SgdArguments = StochasticGradientDescentClassificationTrainer.Arguments;
+    using SgdOptions = StochasticGradientDescentClassificationTrainer.Options;
 
     /// <summary>
     /// TrainerEstimator extension methods.
@@ -31,20 +31,32 @@ public static class StandardLearnersCatalog
         /// <param name="initLearningRate">The initial learning rate used by SGD.</param>
         /// <param name="l2Weight">The L2 regularization constant.</param>
         /// <param name="loss">The loss function to use.</param>
-        /// <param name="advancedSettings">A delegate to apply all the advanced arguments to the algorithm.</param>
         public static StochasticGradientDescentClassificationTrainer StochasticGradientDescent(this BinaryClassificationContext.BinaryClassificationTrainers ctx,
             string labelColumn = DefaultColumnNames.Label,
             string featureColumn = DefaultColumnNames.Features,
             string weights = null,
-            int maxIterations = SgdArguments.Defaults.MaxIterations,
-            double initLearningRate = SgdArguments.Defaults.InitLearningRate,
-            float l2Weight = SgdArguments.Defaults.L2Weight,
-            ISupportClassificationLossFactory loss = null,
-            Action<SgdArguments> advancedSettings = null)
+            int maxIterations = SgdOptions.Defaults.MaxIterations,
+            double initLearningRate = SgdOptions.Defaults.InitLearningRate,
+            float l2Weight = SgdOptions.Defaults.L2Weight,
+            ISupportClassificationLossFactory loss = null)
+        {
+            Contracts.CheckValue(ctx, nameof(ctx));
+            var env = CatalogUtils.GetEnvironment(ctx);
+            return new StochasticGradientDescentClassificationTrainer(env, labelColumn, featureColumn, weights, maxIterations, initLearningRate, l2Weight, loss);
+        }
+
+        /// <summary>
+        ///  Predict a target using a linear binary classification model trained with the <see cref="StochasticGradientDescentClassificationTrainer"/> trainer.
+        /// </summary>
+        /// <param name="ctx">The binary classificaiton context trainer object.</param>
+        /// <param name="advancedSettings">Advanced arguments to the algorithm.</param>
+        public static StochasticGradientDescentClassificationTrainer StochasticGradientDescent(this BinaryClassificationContext.BinaryClassificationTrainers ctx,
+            SgdOptions advancedSettings)
         {
             Contracts.CheckValue(ctx, nameof(ctx));
             var env = CatalogUtils.GetEnvironment(ctx);
-            return new StochasticGradientDescentClassificationTrainer(env, labelColumn, featureColumn, weights, maxIterations, initLearningRate, l2Weight, loss, advancedSettings);
+
+            return new StochasticGradientDescentClassificationTrainer(env, advancedSettings);
         }
 
         /// <summary>

src/Microsoft.ML.StaticPipe/SgdStatic.cs

@@ -9,7 +9,7 @@
 
 namespace Microsoft.ML.StaticPipe
 {
-    using Arguments = StochasticGradientDescentClassificationTrainer.Arguments;
+    using Options = StochasticGradientDescentClassificationTrainer.Options;
 
     /// <summary>
     /// Binary Classification trainer estimators.
@@ -27,7 +27,6 @@ public static class SgdStaticExtensions
         /// <param name="initLearningRate">The initial learning rate used by SGD.</param>
         /// <param name="l2Weight">The L2 regularization constant.</param>
         /// <param name="loss">The loss function to use.</param>
-        /// <param name="advancedSettings">A delegate to apply all the advanced arguments to the algorithm.</param>
         /// <param name="onFit">A delegate that is called every time the
         /// <see cref="Estimator{TTupleInShape, TTupleOutShape, TTransformer}.Fit(DataView{TTupleInShape})"/> method is called on the
         /// <see cref="Estimator{TTupleInShape, TTupleOutShape, TTransformer}"/> instance created out of this. This delegate will receive
@@ -38,17 +37,51 @@ public static (Scalar<float> score, Scalar<float> probability, Scalar<bool> pred
             Scalar<bool> label,
             Vector<float> features,
             Scalar<float> weights = null,
-            int maxIterations = Arguments.Defaults.MaxIterations,
-            double initLearningRate = Arguments.Defaults.InitLearningRate,
-            float l2Weight = Arguments.Defaults.L2Weight,
+            int maxIterations = Options.Defaults.MaxIterations,
+            double initLearningRate = Options.Defaults.InitLearningRate,
+            float l2Weight = Options.Defaults.L2Weight,
             ISupportClassificationLossFactory loss = null,
-            Action<Arguments> advancedSettings = null,
             Action<IPredictorWithFeatureWeights<float>> onFit = null)
         {
             var rec = new TrainerEstimatorReconciler.BinaryClassifier(
                 (env, labelName, featuresName, weightsName) =>
                 {
-                    var trainer = new StochasticGradientDescentClassificationTrainer(env, labelName, featuresName, weightsName, maxIterations, initLearningRate, l2Weight, loss, advancedSettings);
+                    var trainer = new StochasticGradientDescentClassificationTrainer(env, labelName, featuresName, weightsName, maxIterations, initLearningRate, l2Weight, loss);
+
+                    if (onFit != null)
+                        return trainer.WithOnFitDelegate(trans => onFit(trans.Model));
+                    return trainer;
+
+                }, label, features, weights);
+
+            return rec.Output;
+        }
+
+        /// <summary>
+        ///  Predict a target using a linear binary classification model trained with the <see cref="Microsoft.ML.Trainers.StochasticGradientDescentClassificationTrainer"/> trainer.
+        /// </summary>
+        /// <param name="ctx">The binary classificaiton context trainer object.</param>
+        /// <param name="label">The name of the label column.</param>
+        /// <param name="features">The name of the feature column.</param>
+        /// <param name="weights">The name for the example weight column.</param>
+        /// <param name="advancedSettings">Advanced arguments to the algorithm.</param>
+        /// <param name="onFit">A delegate that is called every time the
+        /// <see cref="Estimator{TTupleInShape, TTupleOutShape, TTransformer}.Fit(DataView{TTupleInShape})"/> method is called on the
+        /// <see cref="Estimator{TTupleInShape, TTupleOutShape, TTransformer}"/> instance created out of this. This delegate will receive
+        /// the linear model that was trained.  Note that this action cannot change the result in any way; it is only a way for the caller to
+        /// be informed about what was learnt.</param>
+        /// <returns>The predicted output.</returns>
+        public static (Scalar<float> score, Scalar<float> probability, Scalar<bool> predictedLabel) StochasticGradientDescentClassificationTrainer(this BinaryClassificationContext.BinaryClassificationTrainers ctx,
+            Scalar<bool> label,
+            Vector<float> features,
+            Scalar<float> weights,
+            Options advancedSettings,
+            Action<IPredictorWithFeatureWeights<float>> onFit = null)
+        {
+            var rec = new TrainerEstimatorReconciler.BinaryClassifier(
+                (env, labelName, featuresName, weightsName) =>
+                {
+                    var trainer = new StochasticGradientDescentClassificationTrainer(env, advancedSettings);
 
                     if (onFit != null)
                         return trainer.WithOnFitDelegate(trans => onFit(trans.Model));

test/BaselineOutput/Common/EntryPoints/core_ep-list.tsv

@@ -68,7 +68,7 @@ Trainers.PoissonRegressor	Train an Poisson regression model.	Microsoft.ML.Traine
 Trainers.StochasticDualCoordinateAscentBinaryClassifier	Train an SDCA binary model.	Microsoft.ML.Trainers.Sdca	TrainBinary	Microsoft.ML.Trainers.SdcaBinaryTrainer+Arguments	Microsoft.ML.EntryPoints.CommonOutputs+BinaryClassificationOutput
 Trainers.StochasticDualCoordinateAscentClassifier	The SDCA linear multi-class classification trainer.	Microsoft.ML.Trainers.Sdca	TrainMultiClass	Microsoft.ML.Trainers.SdcaMultiClassTrainer+Arguments	Microsoft.ML.EntryPoints.CommonOutputs+MulticlassClassificationOutput
 Trainers.StochasticDualCoordinateAscentRegressor	The SDCA linear regression trainer.	Microsoft.ML.Trainers.Sdca	TrainRegression	Microsoft.ML.Trainers.SdcaRegressionTrainer+Arguments	Microsoft.ML.EntryPoints.CommonOutputs+RegressionOutput
-Trainers.StochasticGradientDescentBinaryClassifier	Train an Hogwild SGD binary model.	Microsoft.ML.Trainers.StochasticGradientDescentClassificationTrainer	TrainBinary	Microsoft.ML.Trainers.StochasticGradientDescentClassificationTrainer+Arguments	Microsoft.ML.EntryPoints.CommonOutputs+BinaryClassificationOutput
+Trainers.StochasticGradientDescentBinaryClassifier	Train an Hogwild SGD binary model.	Microsoft.ML.Trainers.StochasticGradientDescentClassificationTrainer	TrainBinary	Microsoft.ML.Trainers.StochasticGradientDescentClassificationTrainer+Options	Microsoft.ML.EntryPoints.CommonOutputs+BinaryClassificationOutput
 Trainers.SymSgdBinaryClassifier	Train a symbolic SGD.	Microsoft.ML.Trainers.SymSgd.SymSgdClassificationTrainer	TrainSymSgd	Microsoft.ML.Trainers.SymSgd.SymSgdClassificationTrainer+Arguments	Microsoft.ML.EntryPoints.CommonOutputs+BinaryClassificationOutput
 Transforms.ApproximateBootstrapSampler	Approximate bootstrap sampling.	Microsoft.ML.Transforms.BootstrapSample	GetSample	Microsoft.ML.Transforms.BootstrapSamplingTransformer+Arguments	Microsoft.ML.EntryPoints.CommonOutputs+TransformOutput
 Transforms.BinaryPredictionScoreColumnsRenamer	For binary prediction, it renames the PredictedLabel and Score columns to include the name of the positive class.	Microsoft.ML.EntryPoints.ScoreModel	RenameBinaryPredictionScoreColumns	Microsoft.ML.EntryPoints.ScoreModel+RenameBinaryPredictionScoreColumnsInput	Microsoft.ML.EntryPoints.CommonOutputs+TransformOutput

test/Microsoft.ML.StaticPipelineTesting/Training.cs

@@ -940,8 +940,7 @@ public void HogwildSGDBinaryClassification()
             var est = reader.MakeNewEstimator()
                 .Append(r => (r.label, preds: ctx.Trainers.StochasticGradientDescentClassificationTrainer(r.label, r.features,
                     l2Weight: 0,
-                    onFit: (p) => { pred = p; },
-                    advancedSettings: s => s.NumThreads = 1)));
+                    onFit: (p) => { pred = p; })));
 
             var pipe = reader.Append(est);
 

test/Microsoft.ML.Tests/FeatureContributionTests.cs

@@ -9,6 +9,7 @@
 using Microsoft.ML.Internal.Internallearn;
 using Microsoft.ML.Internal.Utilities;
 using Microsoft.ML.RunTests;
+using Microsoft.ML.Trainers;
 using Microsoft.ML.Training;
 using Microsoft.ML.Transforms;
 using Xunit;
@@ -152,7 +153,9 @@ public void TestSDCABinary()
         [Fact]
         public void TestSGDBinary()
         {
-            TestFeatureContribution(ML.BinaryClassification.Trainers.StochasticGradientDescent(advancedSettings: args => { args.NumThreads = 1; }), GetSparseDataset(TaskType.BinaryClassification, 100), "SGDBinary");
+            TestFeatureContribution(ML.BinaryClassification.Trainers.StochasticGradientDescent(
+                new StochasticGradientDescentClassificationTrainer.Options { NumThreads = 1}), 
+                GetSparseDataset(TaskType.BinaryClassification, 100), "SGDBinary");
         }
 
         [Fact]

test/Microsoft.ML.Tests/TrainerEstimators/TrainerEstimators.cs

@@ -86,7 +86,7 @@ public void KMeansEstimator()
         public void TestEstimatorHogwildSGD()
         {
             (IEstimator<ITransformer> pipe, IDataView dataView) = GetBinaryClassificationPipeline();
-            var trainer = new StochasticGradientDescentClassificationTrainer(Env, "Label", "Features");
+            var trainer = ML.BinaryClassification.Trainers.StochasticGradientDescent();
             var pipeWithTrainer = pipe.Append(trainer);
             TestEstimatorCore(pipeWithTrainer, dataView);