support validation and incremental trainers (#610)

add proper fields to trainer info for FT, FM, and OnlineLearner regarding incremental training and validation datasets

Author: Ivanidzo4ka

src/Microsoft.ML.FastTree/FastTree.cs

@@ -95,7 +95,7 @@ private protected FastTreeTrainerBase(IHostEnvironment env, TArgs args)
             // 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.
-            Info = new TrainerInfo(normalization: false, caching: false, calibration: NeedCalibration);
+            Info = new TrainerInfo(normalization: false, caching: false, calibration: NeedCalibration, supportValid: true);
             int numThreads = Args.NumThreads ?? Environment.ProcessorCount;
             if (Host.ConcurrencyFactor > 0 && numThreads > Host.ConcurrencyFactor)
             {

src/Microsoft.ML.StandardLearners/FactorizationMachine/FactorizationMachineTrainer.cs

@@ -103,7 +103,7 @@ public FieldAwareFactorizationMachineTrainer(IHostEnvironment env, Arguments arg
             _shuffle = args.Shuffle;
             _verbose = args.Verbose;
             _radius = args.Radius;
-            Info = new TrainerInfo();
+            Info = new TrainerInfo(supportValid: true, supportIncrementalTrain: true);
         }
 
         private void InitializeTrainingState(int fieldCount, int featureCount, FieldAwareFactorizationMachinePredictor predictor, out float[] linearWeights,

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

@@ -84,7 +84,7 @@ protected OnlineLinearTrainer(TArguments args, IHostEnvironment env, string name
 
             Args = args;
             // REVIEW: Caching could be false for one iteration, if we got around the whole shuffling issue.
-            Info = new TrainerInfo(calibration: NeedCalibration);
+            Info = new TrainerInfo(calibration: NeedCalibration, supportIncrementalTrain: true);
         }
 
         /// <summary>