Use better AutiFit overload and ignore Multiclass (dotnet#64)

Author: vinodshanbhag

src/mlnet/Commands/NewCommand.cs

@@ -58,15 +58,7 @@ private static (Pipeline, ITransformer) ExploreModels(
 
             if (options.MlTask == TaskKind.BinaryClassification)
             {
-                var result = context.BinaryClassification.AutoFit(trainData, label, validationData, settings:
-                    new AutoFitSettings()
-                    {
-                        StoppingCriteria = new ExperimentStoppingCriteria()
-                        {
-                            //default need to have a way to override
-                            TimeOutInMinutes = 10
-                        }
-                    });
+                var result = context.BinaryClassification.AutoFit(trainData, label, validationData, 10);
                 result = result.OrderByDescending(t => t.Metrics.Accuracy);
                 var bestIteration = result.FirstOrDefault();
                 pipelineToDeconstruct = bestIteration.Pipeline;
@@ -75,22 +67,14 @@ private static (Pipeline, ITransformer) ExploreModels(
 
             if (options.MlTask == TaskKind.Regression)
             {
-                var result = context.Regression.AutoFit(trainData, label, validationData, settings:
-                    new AutoFitSettings()
-                    {
-                        StoppingCriteria = new ExperimentStoppingCriteria()
-                        {
-                            //default need to have a way to override
-                            TimeOutInMinutes = 10
-                        }
-                    });
+                var result = context.Regression.AutoFit(trainData, label, validationData, 10);
                 result = result.OrderByDescending(t => t.Metrics.RSquared);
                 var bestIteration = result.FirstOrDefault();
                 pipelineToDeconstruct = bestIteration.Pipeline;
                 model = bestIteration.Model;
             }
 
-            if (options.MlTask == TaskKind.Regression)
+            if (options.MlTask == TaskKind.MulticlassClassification)
             {
                 throw new NotImplementedException();
             }