Fixed X64Fact tests

[mstfbl]

Fixed X64Fact tests with x86 specific baselines added

Currently obtaining baselines from x86 builds from CI

[mstfbl]

There are 4 more X64Fact tests not fully enabled in TestPredictors.cs with relation to Issue #1216

machinelearning/test/Microsoft.ML.Predictor.Tests/TestPredictors.cs

Lines 758 to 838 in adad3e4

	[X64Fact("x86 fails. Associated GitHubIssue: https://github.com/dotnet/machinelearning/issues/1216")]
	public void TestEnsembleCombiner()
	{
	var dataPath = GetDataPath("breast-cancer.txt");
	var dataView = ML.Data.LoadFromTextFile(dataPath);
	var predictors = new PredictorModel[]
	{
	FastTree.TrainBinary(ML, new FastTreeBinaryTrainer.Options
	{
	FeatureColumnName = "Features",
	NumberOfTrees = 5,
	NumberOfLeaves = 4,
	LabelColumnName = DefaultColumnNames.Label,
	TrainingData = dataView
	}).PredictorModel,
	AveragedPerceptronTrainer.TrainBinary(ML, new AveragedPerceptronTrainer.Options()
	{
	FeatureColumnName = "Features",
	LabelColumnName = DefaultColumnNames.Label,
	NumberOfIterations = 2,
	TrainingData = dataView,
	NormalizeFeatures = NormalizeOption.No
	}).PredictorModel,
	LbfgsLogisticRegressionBinaryTrainer.TrainBinary(ML, new LbfgsLogisticRegressionBinaryTrainer.Options()
	{
	FeatureColumnName = "Features",
	LabelColumnName = DefaultColumnNames.Label,
	OptimizationTolerance = 10e-4F,
	TrainingData = dataView,
	NormalizeFeatures = NormalizeOption.No
	}).PredictorModel,
	LbfgsLogisticRegressionBinaryTrainer.TrainBinary(ML, new LbfgsLogisticRegressionBinaryTrainer.Options()
	{
	FeatureColumnName = "Features",
	LabelColumnName = DefaultColumnNames.Label,
	OptimizationTolerance = 10e-3F,
	TrainingData = dataView,
	NormalizeFeatures = NormalizeOption.No
	}).PredictorModel
	};
	CombineAndTestEnsembles(dataView, "pe", "oc=average", PredictionKind.BinaryClassification, predictors);
	Done();
	}
	[X64Fact("x86 fails. Associated GitHubIssue: https://github.com/dotnet/machinelearning/issues/1216")]
	public void TestMulticlassEnsembleCombiner()
	{
	var dataPath = GetDataPath("breast-cancer.txt");
	var dataView = ML.Data.LoadFromTextFile(dataPath);
	var predictors = new PredictorModel[]
	{
	LightGbm.TrainMulticlass(Env, new LightGbmMulticlassTrainer.Options
	{
	FeatureColumnName = "Features",
	NumberOfIterations = 5,
	NumberOfLeaves = 4,
	LabelColumnName = DefaultColumnNames.Label,
	TrainingData = dataView
	}).PredictorModel,
	LbfgsMaximumEntropyMulticlassTrainer.TrainMulticlass(Env, new LbfgsMaximumEntropyMulticlassTrainer.Options()
	{
	FeatureColumnName = "Features",
	LabelColumnName = DefaultColumnNames.Label,
	OptimizationTolerance = 10e-4F,
	TrainingData = dataView,
	NormalizeFeatures = NormalizeOption.No
	}).PredictorModel,
	LbfgsMaximumEntropyMulticlassTrainer.TrainMulticlass(Env, new LbfgsMaximumEntropyMulticlassTrainer.Options()
	{
	FeatureColumnName = "Features",
	LabelColumnName = DefaultColumnNames.Label,
	OptimizationTolerance = 10e-3F,
	TrainingData = dataView,
	NormalizeFeatures = NormalizeOption.No
	}).PredictorModel
	};
	CombineAndTestEnsembles(dataView, "weightedensemblemulticlass", "oc=multiaverage", PredictionKind.MulticlassClassification, predictors);
	}

These tests cause crashes during runs, should be explored more in the near future, and re-enabled under a different PR.

[harishsk]

Yes, I noticed that too when I was debugging earlier.

Can you please do me a favor?

Now that we have gone towards different baselines, I think it would be good to be able to easily track which tests are using configuration specific baselines. Can you please add some code in BaseTestBaseline.cs to log whenever we use a configuration specific baseline? I just realized that this is something we should do.

---

In reply to: 399621483 [](ancestors = 399621483)

[harishsk]

You can do it as a separate PR if you wish.

[frank-dong-ms-zz]

[mstfbl]

I will make a separate PR for logging when a configuration specific benchmark is utilized.