FastTreeRankingTrainer expose non-advanced args(#1246)

Author: feiyun0112

src/Microsoft.ML.FastTree/BoostingFastTree.cs

@@ -21,10 +21,20 @@ protected BoostingFastTreeTrainerBase(IHostEnvironment env, TArgs args, SchemaSh
         {
         }
 
-        protected BoostingFastTreeTrainerBase(IHostEnvironment env, SchemaShape.Column label, string featureColumn,
-            string weightColumn = null, string groupIdColumn = null, Action<TArgs> advancedSettings = null)
-            : base(env, label, featureColumn, weightColumn, groupIdColumn, advancedSettings)
+        protected BoostingFastTreeTrainerBase(IHostEnvironment env,
+            SchemaShape.Column label,
+            string featureColumn,
+            string weightColumn = null,
+            string groupIdColumn = null,
+            int numLeaves = Defaults.NumLeaves,
+            int numTrees = Defaults.NumTrees,
+            int minDocumentsInLeafs = Defaults.MinDocumentsInLeafs,
+            double learningRate = Defaults.LearningRates,
+            Action<TArgs> advancedSettings = null)
+            : base(env, label, featureColumn, weightColumn, groupIdColumn, numLeaves, numTrees, minDocumentsInLeafs, advancedSettings)
         {
+            //override with the directly provided values.
+            Args.LearningRates = learningRate;
         }
 
         protected override void CheckArgs(IChannel ch)

src/Microsoft.ML.FastTree/FastTree.cs

@@ -92,8 +92,15 @@ public abstract class FastTreeTrainerBase<TArgs, TTransformer, TModel> :
         /// <summary>
         /// Constructor to use when instantiating the classes deriving from here through the API.
         /// </summary>
-        private protected FastTreeTrainerBase(IHostEnvironment env, SchemaShape.Column label, string featureColumn,
-            string weightColumn = null, string groupIdColumn = null, Action<TArgs> advancedSettings = null)
+        private protected FastTreeTrainerBase(IHostEnvironment env,
+            SchemaShape.Column label,
+            string featureColumn,
+            string weightColumn = null,
+            string groupIdColumn = null,
+            int numLeaves = Defaults.NumLeaves,
+            int numTrees = Defaults.NumTrees,
+            int minDocumentsInLeafs = Defaults.MinDocumentsInLeafs,
+            Action<TArgs> advancedSettings = null)
             : base(Contracts.CheckRef(env, nameof(env)).Register(RegisterName), TrainerUtils.MakeR4VecFeature(featureColumn), label, TrainerUtils.MakeR4ScalarWeightColumn(weightColumn))
         {
             Args = new TArgs();
@@ -113,6 +120,11 @@ private protected FastTreeTrainerBase(IHostEnvironment env, SchemaShape.Column l
             if (groupIdColumn != null)
                 Args.GroupIdColumn = groupIdColumn;
 
+            //override with the directly provided values.
+            Args.NumLeaves = numLeaves;
+            Args.NumTrees = numTrees;
+            Args.MinDocumentsInLeafs = minDocumentsInLeafs;
+
             // The discretization step renders this trainer non-parametric, and therefore it does not need normalization.
             // Also since it builds its own internal discretized columnar structures, it cannot benefit from caching.
             // Finally, even the binary classifiers, being logitboost, tend to not benefit from external calibration.

src/Microsoft.ML.FastTree/FastTreeCatalog.cs

@@ -94,7 +94,7 @@ public static FastTreeRankingTrainer FastTree(this RankingContext.RankingTrainer
         {
             Contracts.CheckValue(ctx, nameof(ctx));
             var env = CatalogUtils.GetEnvironment(ctx);
-            return new FastTreeRankingTrainer(env, label, features, groupId, weights, advancedSettings);
+            return new FastTreeRankingTrainer(env, label, features, groupId, weights, advancedSettings: advancedSettings);
         }
     }
 }

src/Microsoft.ML.FastTree/FastTreeClassification.cs

@@ -136,7 +136,7 @@ public FastTreeBinaryClassificationTrainer(IHostEnvironment env,
             int minDocumentsInLeafs = Defaults.MinDocumentsInLeafs,
             double learningRate = Defaults.LearningRates,
             Action<Arguments> advancedSettings = null)
-            : base(env, TrainerUtils.MakeBoolScalarLabel(labelColumn), featureColumn, weightColumn, null, advancedSettings)
+            : base(env, TrainerUtils.MakeBoolScalarLabel(labelColumn), featureColumn, weightColumn, null, numLeaves, numTrees, minDocumentsInLeafs, learningRate, advancedSettings)
         {
             Host.CheckNonEmpty(labelColumn, nameof(labelColumn));
             Host.CheckNonEmpty(featureColumn, nameof(featureColumn));

src/Microsoft.ML.FastTree/FastTreeRanking.cs

@@ -67,10 +67,22 @@ public sealed partial class FastTreeRankingTrainer
         /// <param name="featureColumn">The name of the feature column.</param>
         /// <param name="groupIdColumn">The name for the column containing the group ID. </param>
         /// <param name="weightColumn">The name for the column containing the initial weight.</param>
+        /// <param name="numLeaves">The max number of leaves in each regression tree.</param>
+        /// <param name="numTrees">Total number of decision trees to create in the ensemble.</param>
+        /// <param name="minDocumentsInLeafs">The minimal number of documents allowed in a leaf of a regression tree, out of the subsampled data.</param>
+        /// <param name="learningRate">The learning rate.</param>
         /// <param name="advancedSettings">A delegate to apply all the advanced arguments to the algorithm.</param>
-        public FastTreeRankingTrainer(IHostEnvironment env, string labelColumn, string featureColumn, string groupIdColumn,
-            string weightColumn = null, Action<Arguments> advancedSettings = null)
-            : base(env, TrainerUtils.MakeR4ScalarLabel(labelColumn), featureColumn, weightColumn, groupIdColumn, advancedSettings: advancedSettings)
+        public FastTreeRankingTrainer(IHostEnvironment env,
+            string labelColumn,
+            string featureColumn,
+            string groupIdColumn,
+            string weightColumn = null,
+            int numLeaves = Defaults.NumLeaves,
+            int numTrees = Defaults.NumTrees,
+            int minDocumentsInLeafs = Defaults.MinDocumentsInLeafs,
+            double learningRate = Defaults.LearningRates,
+            Action<Arguments> advancedSettings = null)
+            : base(env, TrainerUtils.MakeR4ScalarLabel(labelColumn), featureColumn, weightColumn, groupIdColumn, numLeaves, numTrees, minDocumentsInLeafs, learningRate, advancedSettings: advancedSettings)
         {
             Host.CheckNonEmpty(labelColumn, nameof(labelColumn));
             Host.CheckNonEmpty(featureColumn, nameof(featureColumn));

src/Microsoft.ML.FastTree/FastTreeRegression.cs

@@ -72,7 +72,7 @@ public FastTreeRegressionTrainer(IHostEnvironment env,
             int minDocumentsInLeafs = Defaults.MinDocumentsInLeafs,
             double learningRate = Defaults.LearningRates,
             Action<Arguments> advancedSettings = null)
-            : base(env, TrainerUtils.MakeR4ScalarLabel(labelColumn), featureColumn, weightColumn, null, advancedSettings)
+            : base(env, TrainerUtils.MakeR4ScalarLabel(labelColumn), featureColumn, weightColumn, null, numLeaves, numTrees, minDocumentsInLeafs, learningRate, advancedSettings)
         {
             Host.CheckNonEmpty(labelColumn, nameof(labelColumn));
             Host.CheckNonEmpty(featureColumn, nameof(featureColumn));

src/Microsoft.ML.FastTree/FastTreeStatic.cs

@@ -153,7 +153,7 @@ public static Scalar<float> FastTree<TVal>(this RankingContext.RankingTrainers c
             var rec = new TrainerEstimatorReconciler.Ranker<TVal>(
                (env, labelName, featuresName, groupIdName, weightsName) =>
                {
-                   var trainer = new FastTreeRankingTrainer(env, labelName, featuresName, groupIdName, weightsName, advancedSettings);
+                   var trainer = new FastTreeRankingTrainer(env, labelName, featuresName, groupIdName, weightsName, numLeaves, numTrees, minDatapointsInLeafs, learningRate, advancedSettings);
                    if (onFit != null)
                        return trainer.WithOnFitDelegate(trans => onFit(trans.Model));
                    return trainer;

src/Microsoft.ML.FastTree/FastTreeTweedie.cs

@@ -59,7 +59,7 @@ public sealed partial class FastTreeTweedieTrainer
         /// <param name="advancedSettings">A delegate to apply all the advanced arguments to the algorithm.</param>
         public FastTreeTweedieTrainer(IHostEnvironment env, string labelColumn, string featureColumn,
             string groupIdColumn = null, string weightColumn = null, Action<Arguments> advancedSettings = null)
-            : base(env, TrainerUtils.MakeR4ScalarLabel(labelColumn), featureColumn, weightColumn, groupIdColumn, advancedSettings)
+            : base(env, TrainerUtils.MakeR4ScalarLabel(labelColumn), featureColumn, weightColumn, groupIdColumn, advancedSettings: advancedSettings)
         {
             Host.CheckNonEmpty(labelColumn, nameof(labelColumn));
             Host.CheckNonEmpty(featureColumn, nameof(featureColumn));

src/Microsoft.ML.FastTree/RandomForest.cs

@@ -30,7 +30,7 @@ protected RandomForestTrainerBase(IHostEnvironment env, TArgs args, SchemaShape.
         /// </summary>
         protected RandomForestTrainerBase(IHostEnvironment env, SchemaShape.Column label, string featureColumn,
             string weightColumn = null, string groupIdColumn = null, bool quantileEnabled = false, Action<TArgs> advancedSettings = null)
-            : base(env, label, featureColumn, weightColumn, groupIdColumn, advancedSettings)
+            : base(env, label, featureColumn, weightColumn, groupIdColumn, advancedSettings: advancedSettings)
         {
             _quantileEnabled = quantileEnabled;
         }