public void TestSimpleExperiment()
{
var dataPath = GetDataPath(@"adult.tiny.with-schema.txt");
using (var env = new TlcEnvironment())
{
var experiment = env.CreateExperiment();
var importInput = new ML.Data.TextLoader(dataPath);
var importOutput = experiment.Add(importInput);
var normalizeInput = new ML.Transforms.MinMaxNormalizer
{
Data = importOutput.Data
};
normalizeInput.AddColumn("NumericFeatures");
var normalizeOutput = experiment.Add(normalizeInput);
experiment.Compile();
experiment.SetInput(importInput.InputFile, new SimpleFileHandle(env, dataPath, false, false));
experiment.Run();
var data = experiment.GetOutput(normalizeOutput.OutputData);
var schema = data.Schema;
Assert.Equal(5, schema.ColumnCount);
var expected = new[] { "Label", "Workclass", "Categories", "NumericFeatures", "NumericFeatures" };
for (int i = 0; i < schema.ColumnCount; i++)
{
Assert.Equal(expected[i], schema.GetColumnName(i));
}
}
}
public ILearningPipelineStep ApplyStep(ILearningPipelineStep previousStep, Experiment experiment)
{
Contracts.Assert(previousStep == null);
ImportTextInput = new Data.TextLoader();
ImportTextInput.CustomSchema = CustomSchema;
var importOutput = experiment.Add(ImportTextInput);
return(new TextLoaderPipelineStep(importOutput.Data));
}
public void TestSimpleTrainExperiment()
{
var dataPath = GetDataPath(@"adult.tiny.with-schema.txt");
using (var env = new TlcEnvironment())
{
var experiment = env.CreateExperiment();
var importInput = new ML.Data.TextLoader(dataPath);
var importOutput = experiment.Add(importInput);
var catInput = new ML.Transforms.CategoricalOneHotVectorizer
{
Data = importOutput.Data
};
catInput.AddColumn("Categories");
var catOutput = experiment.Add(catInput);
var concatInput = new ML.Transforms.ColumnConcatenator
{
Data = catOutput.OutputData
};
concatInput.AddColumn("Features", "Categories", "NumericFeatures");
var concatOutput = experiment.Add(concatInput);
var sdcaInput = new ML.Trainers.StochasticDualCoordinateAscentBinaryClassifier
{
TrainingData = concatOutput.OutputData,
LossFunction = new HingeLossSDCAClassificationLossFunction()
{
Margin = 1.1f
},
NumThreads = 1,
Shuffle = false
};
var sdcaOutput = experiment.Add(sdcaInput);
var scoreInput = new ML.Transforms.DatasetScorer
{
Data = concatOutput.OutputData,
PredictorModel = sdcaOutput.PredictorModel
};
var scoreOutput = experiment.Add(scoreInput);
var evalInput = new ML.Models.BinaryClassificationEvaluator
{
Data = scoreOutput.ScoredData
};
var evalOutput = experiment.Add(evalInput);
experiment.Compile();
experiment.SetInput(importInput.InputFile, new SimpleFileHandle(env, dataPath, false, false));
experiment.Run();
var data = experiment.GetOutput(evalOutput.OverallMetrics);
var schema = data.Schema;
var b = schema.TryGetColumnIndex("AUC", out int aucCol);
Assert.True(b);
using (var cursor = data.GetRowCursor(col => col == aucCol))
{
var getter = cursor.GetGetter <double>(aucCol);
b = cursor.MoveNext();
Assert.True(b);
double auc = 0;
getter(ref auc);
Assert.Equal(0.93, auc, 2);
b = cursor.MoveNext();
Assert.False(b);
}
}
}
public void TestCrossValidationMacroWithStratification()
{
var dataPath = GetDataPath(@"breast-cancer.txt");
using (var env = new TlcEnvironment())
{
var subGraph = env.CreateExperiment();
var nop = new ML.Transforms.NoOperation();
var nopOutput = subGraph.Add(nop);
var learnerInput = new ML.Trainers.StochasticDualCoordinateAscentBinaryClassifier
{
TrainingData = nopOutput.OutputData,
NumThreads = 1
};
var learnerOutput = subGraph.Add(learnerInput);
var modelCombine = new ML.Transforms.ManyHeterogeneousModelCombiner
{
TransformModels = new ArrayVar <ITransformModel>(nopOutput.Model),
PredictorModel = learnerOutput.PredictorModel
};
var modelCombineOutput = subGraph.Add(modelCombine);
var experiment = env.CreateExperiment();
var importInput = new ML.Data.TextLoader(dataPath);
importInput.Arguments.Column = new ML.Data.TextLoaderColumn[]
{
new ML.Data.TextLoaderColumn {
Name = "Label", Source = new[] { new ML.Data.TextLoaderRange(0) }
},
new ML.Data.TextLoaderColumn {
Name = "Strat", Source = new[] { new ML.Data.TextLoaderRange(1) }
},
new ML.Data.TextLoaderColumn {
Name = "Features", Source = new[] { new ML.Data.TextLoaderRange(2, 9) }
}
};
var importOutput = experiment.Add(importInput);
var crossValidate = new ML.Models.CrossValidator
{
Data = importOutput.Data,
Nodes = subGraph,
TransformModel = null,
StratificationColumn = "Strat"
};
crossValidate.Inputs.Data = nop.Data;
crossValidate.Outputs.Model = modelCombineOutput.PredictorModel;
var crossValidateOutput = experiment.Add(crossValidate);
experiment.Compile();
experiment.SetInput(importInput.InputFile, new SimpleFileHandle(env, dataPath, false, false));
experiment.Run();
var data = experiment.GetOutput(crossValidateOutput.OverallMetrics[0]);
var schema = data.Schema;
var b = schema.TryGetColumnIndex("AUC", out int metricCol);
Assert.True(b);
using (var cursor = data.GetRowCursor(col => col == metricCol))
{
var getter = cursor.GetGetter <double>(metricCol);
b = cursor.MoveNext();
Assert.True(b);
double val = 0;
getter(ref val);
Assert.Equal(0.99, val, 2);
b = cursor.MoveNext();
Assert.False(b);
}
}
}
public void TestCrossValidationMacro()
{
var dataPath = GetDataPath(TestDatasets.winequality.trainFilename);
using (var env = new TlcEnvironment())
{
var subGraph = env.CreateExperiment();
var nop = new ML.Transforms.NoOperation();
var nopOutput = subGraph.Add(nop);
var learnerInput = new ML.Trainers.StochasticDualCoordinateAscentRegressor
{
TrainingData = nopOutput.OutputData,
NumThreads = 1
};
var learnerOutput = subGraph.Add(learnerInput);
var modelCombine = new ML.Transforms.ManyHeterogeneousModelCombiner
{
TransformModels = new ArrayVar <ITransformModel>(nopOutput.Model),
PredictorModel = learnerOutput.PredictorModel
};
var modelCombineOutput = subGraph.Add(modelCombine);
var experiment = env.CreateExperiment();
var importInput = new ML.Data.TextLoader(dataPath)
{
Arguments = new TextLoaderArguments
{
Separator = new[] { ';' },
HasHeader = true,
Column = new[]
{
new TextLoaderColumn()
{
Name = "Label",
Source = new [] { new TextLoaderRange(11) },
Type = DataKind.Num
},
new TextLoaderColumn()
{
Name = "Features",
Source = new [] { new TextLoaderRange(0, 10) },
Type = DataKind.Num
}
}
}
};
var importOutput = experiment.Add(importInput);
var crossValidate = new ML.Models.CrossValidator
{
Data = importOutput.Data,
Nodes = subGraph,
Kind = ML.Models.MacroUtilsTrainerKinds.SignatureRegressorTrainer,
TransformModel = null
};
crossValidate.Inputs.Data = nop.Data;
crossValidate.Outputs.Model = modelCombineOutput.PredictorModel;
var crossValidateOutput = experiment.Add(crossValidate);
experiment.Compile();
experiment.SetInput(importInput.InputFile, new SimpleFileHandle(env, dataPath, false, false));
experiment.Run();
var data = experiment.GetOutput(crossValidateOutput.OverallMetrics[0]);
var schema = data.Schema;
var b = schema.TryGetColumnIndex("L1(avg)", out int metricCol);
Assert.True(b);
using (var cursor = data.GetRowCursor(col => col == metricCol))
{
var getter = cursor.GetGetter <double>(metricCol);
b = cursor.MoveNext();
Assert.True(b);
double val = 0;
getter(ref val);
Assert.Equal(0.58, val, 1);
b = cursor.MoveNext();
Assert.False(b);
}
}
}
public void TestCrossValidationBinaryMacro()
{
var dataPath = GetDataPath(@"adult.tiny.with-schema.txt");
using (var env = new TlcEnvironment())
{
var subGraph = env.CreateExperiment();
var catInput = new ML.Transforms.CategoricalOneHotVectorizer();
catInput.AddColumn("Categories");
var catOutput = subGraph.Add(catInput);
var concatInput = new ML.Transforms.ColumnConcatenator
{
Data = catOutput.OutputData
};
concatInput.AddColumn("Features", "Categories", "NumericFeatures");
var concatOutput = subGraph.Add(concatInput);
var lrInput = new ML.Trainers.LogisticRegressionBinaryClassifier
{
TrainingData = concatOutput.OutputData,
NumThreads = 1
};
var lrOutput = subGraph.Add(lrInput);
var modelCombine = new ML.Transforms.ManyHeterogeneousModelCombiner
{
TransformModels = new ArrayVar <ITransformModel>(catOutput.Model, concatOutput.Model),
PredictorModel = lrOutput.PredictorModel
};
var modelCombineOutput = subGraph.Add(modelCombine);
var experiment = env.CreateExperiment();
var importInput = new ML.Data.TextLoader(dataPath);
var importOutput = experiment.Add(importInput);
var crossValidateBinary = new ML.Models.BinaryCrossValidator
{
Data = importOutput.Data,
Nodes = subGraph
};
crossValidateBinary.Inputs.Data = catInput.Data;
crossValidateBinary.Outputs.Model = modelCombineOutput.PredictorModel;
var crossValidateOutput = experiment.Add(crossValidateBinary);
experiment.Compile();
experiment.SetInput(importInput.InputFile, new SimpleFileHandle(env, dataPath, false, false));
experiment.Run();
var data = experiment.GetOutput(crossValidateOutput.OverallMetrics[0]);
var schema = data.Schema;
var b = schema.TryGetColumnIndex("AUC", out int aucCol);
Assert.True(b);
using (var cursor = data.GetRowCursor(col => col == aucCol))
{
var getter = cursor.GetGetter <double>(aucCol);
b = cursor.MoveNext();
Assert.True(b);
double auc = 0;
getter(ref auc);
Assert.Equal(0.87, auc, 1);
b = cursor.MoveNext();
Assert.False(b);
}
}
}
// Note that we don't filter out rows with parsing issues since it's not acceptable to
// produce a different set of rows when subsetting columns. Any parsing errors need to be
// translated to NaN, not result in skipping the row. We should produce some diagnostics
// to alert the user to the issues.
private Cursor(TextLoader parent, ParseStats stats, bool[] active, LineReader reader, int srcNeeded, int cthd)
: base(parent._host)
{
Ch.Assert(active == null || active.Length == parent._bindings.OutputSchema.Count);
Ch.AssertValue(reader);
Ch.AssertValue(stats);
Ch.Assert(srcNeeded >= 0);
Ch.Assert(cthd > 0);
_total = -1;
_batch = -1;
_bindings = parent._bindings;
_parser = parent._parser;
_active = active;
_reader = reader;
_stats = stats;
_srcNeeded = srcNeeded;
ParallelState state = null;
if (cthd > 1)
{
state = new ParallelState(this, out _rows, cthd);
}
else
{
_rows = _parser.CreateRowSet(_stats, 1, _active);
}
try
{
_getters = new Delegate[_bindings.Infos.Length];
for (int i = 0; i < _getters.Length; i++)
{
if (_active != null && !_active[i])
{
continue;
}
ColumnPipe v = _rows.Pipes[i];
Ch.Assert(v != null);
_getters[i] = v.GetGetter();
Ch.Assert(_getters[i] != null);
}
if (state != null)
{
_ator = ParseParallel(state).GetEnumerator();
state = null;
}
else
{
_ator = ParseSequential().GetEnumerator();
}
}
finally
{
if (state != null)
{
state.Dispose();
}
}
}