/// <summary>
/// The train.
/// </summary>
/// <param name="scenarioId">
/// The scenario id.
/// </param>
/// <returns>
/// The <see cref="ScenarioTrainings"/>.
/// </returns>
public ScenarioTrainings Train(string scenarioId)
{
using (var dbContext = new OpenAIEntities1())
{
Scenario scenario = dbContext.Scenarios.Find(scenarioId);
if (scenario == null)
{
return(null);
}
DataTable table = this.ParseCsv(scenario.Contents);
// Get hypothesis for each feature / output
var trainings = new ScenarioTrainings
{
ScenarioId = scenarioId,
TrainingByFeatureId = new Dictionary <string, TrainerHelper>()
};
trainingByScenario[scenarioId] = trainings;
foreach (DataColumn column in table.Columns)
{
TrainerHelper thing = trainer.Train(table, column.ColumnName);
trainings.TrainingByFeatureId[column.ColumnName] = thing;
}
return(trainings);
}
}
private static IPredictor TrainCore(IHostEnvironment env, IChannel ch, RoleMappedData data, ITrainer trainer, RoleMappedData validData,
IComponentFactory <ICalibratorTrainer> calibrator, int maxCalibrationExamples, bool?cacheData, IPredictor inputPredictor = null, RoleMappedData testData = null)
{
Contracts.CheckValue(env, nameof(env));
env.CheckValue(ch, nameof(ch));
ch.CheckValue(data, nameof(data));
ch.CheckValue(trainer, nameof(trainer));
ch.CheckValueOrNull(validData);
ch.CheckValueOrNull(inputPredictor);
AddCacheIfWanted(env, ch, trainer, ref data, cacheData);
ch.Trace("Training");
if (validData != null)
{
AddCacheIfWanted(env, ch, trainer, ref validData, cacheData);
}
if (inputPredictor != null && !trainer.Info.SupportsIncrementalTraining)
{
ch.Warning("Ignoring " + nameof(TrainCommand.Arguments.InputModelFile) +
": Trainer does not support incremental training.");
inputPredictor = null;
}
ch.Assert(validData == null || trainer.Info.SupportsValidation);
var predictor = trainer.Train(new TrainContext(data, validData, testData, inputPredictor));
var caliTrainer = calibrator?.CreateComponent(env);
return(CalibratorUtils.TrainCalibratorIfNeeded(env, ch, caliTrainer, maxCalibrationExamples, trainer, predictor, data));
}
public static long GetTrainingLength(ITrainer trainer, List <TrainingElement> trainingData)
{
var stopWatch = new Stopwatch();
stopWatch.Start();
trainer.Train(trainingData);
stopWatch.Stop();
return(stopWatch.ElapsedMilliseconds);
}
public IActionResult Train([FromQuery] int iterations, int approximationRank, double learningRate)
{
try
{
return(Ok(_trainer.Train(iterations, approximationRank, learningRate)));
}
catch (Exception e)
{
_logger.LogError(e, e.Message);
return(StatusCode(500, e.Message));
}
}
protected override IPredictor Train(TrainContext ctx)
{
var data = ctx.TrainingSet;
Contracts.CheckValue(data, "data");
Contracts.CheckValue(_trainer, "_trainer");
IDataView view = data.Data;
// Preprocess only for training.
if (_args.preTrainType != null)
{
using (var ch2 = Host.Start("PreProcessTraining"))
{
ch2.Info("Applies a preprocess only for training: {0}", _args.preTrainType);
var trSett = ScikitSubComponent <IDataTransform, SignatureDataTransform> .AsSubComponent(_args.preTrainType);
_preTrainProcess = trSett.CreateInstance(Host, view);
}
view = _preTrainProcess;
}
// Preprocess.
if (_args.preType != null)
{
using (var ch2 = Host.Start("PreProcess"))
{
ch2.Info("Applies a preprocess: {0}", _args.preType);
var trSett = ScikitSubComponent <IDataTransform, SignatureDataTransform> .AsSubComponent(_args.preType);
_preProcess = trSett.CreateInstance(Host, view);
}
}
else
{
_preProcess = new PassThroughTransform(Host, new PassThroughTransform.Arguments {
}, view);
}
view = _preProcess;
// New RoleDataMapping
var roles = data.Schema.GetColumnRoleNames()
.Where(kvp => kvp.Key.Value != CR.Feature.Value)
.Where(kvp => kvp.Key.Value != CR.Group.Value)
.Where(kvp => kvp.Key.Value != CR.Label.Value)
.Where(kvp => kvp.Key.Value != CR.Name.Value)
.Where(kvp => kvp.Key.Value != CR.Weight.Value);
if (data.Schema.Feature != null)
{
roles = roles.Prepend(CR.Feature.Bind(data.Schema.Feature.Value.Name));
}
if (data.Schema.Group != null)
{
roles = roles.Prepend(CR.Group.Bind(data.Schema.Group.Value.Name));
}
if (data.Schema.Label != null)
{
roles = roles.Prepend(CR.Label.Bind(data.Schema.Label.Value.Name));
}
if (data.Schema.Weight != null)
{
roles = roles.Prepend(CR.Weight.Bind(data.Schema.Weight.Value.Name));
}
var td = new RoleMappedData(view, roles);
// Train.
if (_args.predictorType != null)
{
using (var ch2 = Host.Start("Training"))
{
var sch1 = SchemaHelper.ToString(data.Schema.Schema);
var sch2 = SchemaHelper.ToString(td.Schema.Schema);
ch2.Info("Initial schema: {0}", sch1);
ch2.Info("DataViewSchema before training: {0}", sch2);
ch2.Info("Train a predictor: {0}", _args.predictorType);
_predictor = _trainer.Train(td);
}
}
// Predictor as a transform.
{
using (var ch2 = Host.Start("Predictor as Transform"))
{
ch2.Info("Creates a transfrom from a predictor");
_inputColumn = td.Schema.Feature.Value.Name;
_predictorAsTransform = new TransformFromValueMapper(Host, _predictor as IValueMapper,
view, td.Schema.Feature.Value.Name, _outputColumn);
}
view = _predictorAsTransform;
}
// Postprocess.
if (_args.postType != null)
{
using (var ch2 = Host.Start("PostProcess"))
{
ch2.Info("Applies a postprocess: {0}", _args.postType);
var postSett = ScikitSubComponent <IDataTransform, SignatureDataTransform> .AsSubComponent(_args.postType);
_postProcess = postSett.CreateInstance(Host, view);
}
}
else
{
_postProcess = null;
}
return(CreatePredictor());
}
/// <summary>
/// Convenience train extension for the case where one has only a training set with no auxiliary information.
/// Equivalent to calling <see cref="ITrainer{TPredictor}.Train(TrainContext)"/>
/// on a <see cref="TrainContext"/> constructed with <paramref name="trainData"/>.
/// </summary>
/// <param name="trainer">The trainer</param>
/// <param name="trainData">The training data.</param>
/// <returns>The trained predictor</returns>
public static TPredictor Train <TPredictor>(this ITrainer <TPredictor> trainer, RoleMappedData trainData) where TPredictor : IPredictor
=> trainer.Train(new TrainContext(trainData));
/// <summary> /// Convenience train extension for the case where one has only a training set with no auxiliary information. /// Equivalent to calling <see cref="ITrainer.Train(TrainContext)"/> /// on a <see cref="TrainContext"/> constructed with <paramref name="trainData"/>. /// </summary> /// <param name="trainer">The trainer</param> /// <param name="trainData">The training data.</param> /// <returns>The trained predictor</returns> public static IPredictor Train(this ITrainer trainer, RoleMappedData trainData) => trainer.Train(new TrainContext(trainData));
public void StochasticGradientDescentNewBenchmark()
{
_trainer.Train(_trainingData);
}
public IPredictor Train(TrainContext context)
{
return(_trainer.Train(context));
}
public void StochasticGradientDescentBenchmark()
{
_trainer.Train(_trainingData.ToList());
}
/// <summary>
/// Enterance method
/// </summary>
/// <param name="args">Comand line arguments</param>
private static void Main(string[] args)
{
Console.WriteLine(
new System.Text.RegularExpressions.Regex("^\\s*(?:\\+?(\\d{1,3}))?[-. (]*(\\d{3})[-. )]*(\\d{3})[-. ]*(\\d{4})(?: *x(\\d+))?\\s*$")
.Matches(Console.ReadLine()).Cast <System.Text.RegularExpressions.Capture>().Select(x => x.Value).Aggregate((i, j) => i + " \n" + j)
);
Console.ReadLine();
TrainerManager tm = new TrainerManager();
//ICalculatableNetwork nn2 =
// tm.GetTrainer(LearningSubject.AndGate, LearningObject.Network).Train<ICalculatableNetwork>();
//ICalculatableNeuron n1 =
// tm.GetTrainer(LearningSubject.AndGate, LearningObject.Neuron).Train<ICalculatableNeuron>();
//ICalculatableNeuron n2 =
// tm.GetTrainer(LearningSubject.OrGate, LearningObject.Neuron).Train<ICalculatableNeuron>();
//ICalculatableNetwork nn1 =
// tm.GetTrainer(LearningSubject.XorGate, LearningObject.Network).Train<ICalculatableNetwork>();
ITrainer imageTrainer =
tm.GetTrainer(LearningSubject.Image, LearningObject.Network);
var imageNetwork = imageTrainer.Train <ICalculatableNetwork>();
var d = imageTrainer.GetTestData();
double succesRate = 0;
foreach (var pair in d)
{
var di = (new DigitImage(pair.Key, (byte)pair.Value.ToList().IndexOf(1)));
var data = (imageNetwork.Calculate(pair.Key).ToList());
if (di.Label == data.IndexOf(data.Max()))
{
succesRate += 1.0 / d.Count;
}
}
Console.WriteLine(succesRate);
//var m = tm.GetTrainer(LearningSubject.XorGate, (ILearnable) nn);
//var m2 = tm.GetTrainer(LearningSubject.XorGate, LearningObject.Network).Train<ICalculatableNetwork>();
//var m3 = m2.Calculate(1,1).First();
//ITrainer andTrainer = new AndGateNeuronTrainer((ILearnableNeuron) n1);
//ITrainer orTrainer = new OrGateNeuronTrainer((ILearnableNeuron) n2);
//ITrainer xorNetworkTrainer = new XorGateNetworkTrainer((ILearnableNetwork) nn);
//andTrainer.Train();
//orTrainer.Train();
//// Not posible -> Use multilayer
//// xorNeuronTrainer.Train();
//xorNetworkTrainer.Train();
//Console.WriteLine("----------AND Gate----------");
//Console.WriteLine($"(1,1) => {n1.Calculate(1, 1)}");
//Console.WriteLine($"(1,0) => {n1.Calculate(1, 0)}");
//Console.WriteLine($"(0,1) => {n1.Calculate(0, 1)}");
//Console.WriteLine($"(0,0) => {n1.Calculate(0, 0)}");
//Console.WriteLine("----------OR Gate----------");
//Console.WriteLine($"(1,1) => {n2.Calculate(1, 1)}");
//Console.WriteLine($"(1,0) => {n2.Calculate(1, 0)}");
//Console.WriteLine($"(0,1) => {n2.Calculate(0, 1)}");
//Console.WriteLine($"(0,0) => {n2.Calculate(0, 0)}");
//Console.WriteLine("----------XOR Gate----------");
//Console.WriteLine($"(1,1) => {nn1.Calculate(1, 1).First()}");
//Console.WriteLine($"(1,0) => {nn1.Calculate(1, 0).First()}");
//Console.WriteLine($"(0,1) => {nn1.Calculate(0, 1).First()}");
//Console.WriteLine($"(0,0) => {nn1.Calculate(0, 0).First()}");
//Console.WriteLine("----------AND Gate----------");
//Console.WriteLine($"(1,1) => {nn2.Calculate(1, 1).First()}");
//Console.WriteLine($"(1,0) => {nn2.Calculate(1, 0).First()}");
//Console.WriteLine($"(0,1) => {nn2.Calculate(0, 1).First()}");
//Console.WriteLine($"(0,0) => {nn2.Calculate(0, 0).First()}");
//DigitImageLoader loader = new DigitImageLoader("train-images.idx3-ubyte", "train-labels.idx1-ubyte");
//var data = loader.Load().OrderBy(x=>x.Label);
//foreach (DigitImage image in data)
//{
// Console.WriteLine(image);
// Console.ReadLine();
//}
Console.ReadLine();
}