Search in sources :

Example 6 with MaxEpochsTerminationCondition

use of org.deeplearning4j.earlystopping.termination.MaxEpochsTerminationCondition in project deeplearning4j by deeplearning4j.

the class TestEarlyStoppingCompGraph method testListeners.

@Test
public void testListeners() {
    ComputationGraphConfiguration conf = new NeuralNetConfiguration.Builder().optimizationAlgo(OptimizationAlgorithm.STOCHASTIC_GRADIENT_DESCENT).iterations(1).updater(Updater.SGD).weightInit(WeightInit.XAVIER).graphBuilder().addInputs("in").addLayer("0", new OutputLayer.Builder().nIn(4).nOut(3).lossFunction(LossFunctions.LossFunction.MCXENT).build(), "in").setOutputs("0").pretrain(false).backprop(true).build();
    ComputationGraph net = new ComputationGraph(conf);
    net.setListeners(new ScoreIterationListener(1));
    DataSetIterator irisIter = new IrisDataSetIterator(150, 150);
    EarlyStoppingModelSaver<ComputationGraph> saver = new InMemoryModelSaver<>();
    EarlyStoppingConfiguration<ComputationGraph> esConf = new EarlyStoppingConfiguration.Builder<ComputationGraph>().epochTerminationConditions(new MaxEpochsTerminationCondition(5)).iterationTerminationConditions(new MaxTimeIterationTerminationCondition(1, TimeUnit.MINUTES)).scoreCalculator(new DataSetLossCalculatorCG(irisIter, true)).modelSaver(saver).build();
    LoggingEarlyStoppingListener listener = new LoggingEarlyStoppingListener();
    IEarlyStoppingTrainer trainer = new EarlyStoppingGraphTrainer(esConf, net, irisIter, listener);
    trainer.fit();
    assertEquals(1, listener.onStartCallCount);
    assertEquals(5, listener.onEpochCallCount);
    assertEquals(1, listener.onCompletionCallCount);
}
Also used : OutputLayer(org.deeplearning4j.nn.conf.layers.OutputLayer) InMemoryModelSaver(org.deeplearning4j.earlystopping.saver.InMemoryModelSaver) MaxEpochsTerminationCondition(org.deeplearning4j.earlystopping.termination.MaxEpochsTerminationCondition) IrisDataSetIterator(org.deeplearning4j.datasets.iterator.impl.IrisDataSetIterator) NeuralNetConfiguration(org.deeplearning4j.nn.conf.NeuralNetConfiguration) DataSetLossCalculatorCG(org.deeplearning4j.earlystopping.scorecalc.DataSetLossCalculatorCG) EarlyStoppingGraphTrainer(org.deeplearning4j.earlystopping.trainer.EarlyStoppingGraphTrainer) IEarlyStoppingTrainer(org.deeplearning4j.earlystopping.trainer.IEarlyStoppingTrainer) ComputationGraphConfiguration(org.deeplearning4j.nn.conf.ComputationGraphConfiguration) ComputationGraph(org.deeplearning4j.nn.graph.ComputationGraph) ScoreIterationListener(org.deeplearning4j.optimize.listeners.ScoreIterationListener) IrisDataSetIterator(org.deeplearning4j.datasets.iterator.impl.IrisDataSetIterator) DataSetIterator(org.nd4j.linalg.dataset.api.iterator.DataSetIterator) MaxTimeIterationTerminationCondition(org.deeplearning4j.earlystopping.termination.MaxTimeIterationTerminationCondition) Test(org.junit.Test)

Example 7 with MaxEpochsTerminationCondition

use of org.deeplearning4j.earlystopping.termination.MaxEpochsTerminationCondition in project deeplearning4j by deeplearning4j.

the class TestEarlyStoppingCompGraph method testTimeTermination.

@Test
public void testTimeTermination() {
    //test termination after max time
    Nd4j.getRandom().setSeed(12345);
    ComputationGraphConfiguration conf = new NeuralNetConfiguration.Builder().seed(12345).optimizationAlgo(OptimizationAlgorithm.STOCHASTIC_GRADIENT_DESCENT).iterations(1).updater(Updater.SGD).learningRate(1e-6).weightInit(WeightInit.XAVIER).graphBuilder().addInputs("in").addLayer("0", new OutputLayer.Builder().nIn(4).nOut(3).lossFunction(LossFunctions.LossFunction.MCXENT).build(), "in").setOutputs("0").pretrain(false).backprop(true).build();
    ComputationGraph net = new ComputationGraph(conf);
    net.setListeners(new ScoreIterationListener(1));
    DataSetIterator irisIter = new IrisDataSetIterator(150, 150);
    EarlyStoppingModelSaver<ComputationGraph> saver = new InMemoryModelSaver<>();
    EarlyStoppingConfiguration<ComputationGraph> esConf = new EarlyStoppingConfiguration.Builder<ComputationGraph>().epochTerminationConditions(new MaxEpochsTerminationCondition(10000)).iterationTerminationConditions(new MaxTimeIterationTerminationCondition(3, TimeUnit.SECONDS), //Initial score is ~2.5
    new MaxScoreIterationTerminationCondition(7.5)).modelSaver(saver).build();
    IEarlyStoppingTrainer trainer = new EarlyStoppingGraphTrainer(esConf, net, irisIter);
    long startTime = System.currentTimeMillis();
    EarlyStoppingResult result = trainer.fit();
    long endTime = System.currentTimeMillis();
    int durationSeconds = (int) (endTime - startTime) / 1000;
    assertTrue(durationSeconds >= 3);
    assertTrue(durationSeconds <= 9);
    assertEquals(EarlyStoppingResult.TerminationReason.IterationTerminationCondition, result.getTerminationReason());
    String expDetails = new MaxTimeIterationTerminationCondition(3, TimeUnit.SECONDS).toString();
    assertEquals(expDetails, result.getTerminationDetails());
}
Also used : OutputLayer(org.deeplearning4j.nn.conf.layers.OutputLayer) InMemoryModelSaver(org.deeplearning4j.earlystopping.saver.InMemoryModelSaver) MaxEpochsTerminationCondition(org.deeplearning4j.earlystopping.termination.MaxEpochsTerminationCondition) IrisDataSetIterator(org.deeplearning4j.datasets.iterator.impl.IrisDataSetIterator) NeuralNetConfiguration(org.deeplearning4j.nn.conf.NeuralNetConfiguration) EarlyStoppingGraphTrainer(org.deeplearning4j.earlystopping.trainer.EarlyStoppingGraphTrainer) IEarlyStoppingTrainer(org.deeplearning4j.earlystopping.trainer.IEarlyStoppingTrainer) ComputationGraphConfiguration(org.deeplearning4j.nn.conf.ComputationGraphConfiguration) ComputationGraph(org.deeplearning4j.nn.graph.ComputationGraph) MaxScoreIterationTerminationCondition(org.deeplearning4j.earlystopping.termination.MaxScoreIterationTerminationCondition) ScoreIterationListener(org.deeplearning4j.optimize.listeners.ScoreIterationListener) IrisDataSetIterator(org.deeplearning4j.datasets.iterator.impl.IrisDataSetIterator) DataSetIterator(org.nd4j.linalg.dataset.api.iterator.DataSetIterator) MaxTimeIterationTerminationCondition(org.deeplearning4j.earlystopping.termination.MaxTimeIterationTerminationCondition) Test(org.junit.Test)

Example 8 with MaxEpochsTerminationCondition

use of org.deeplearning4j.earlystopping.termination.MaxEpochsTerminationCondition in project deeplearning4j by deeplearning4j.

the class TestEarlyStoppingSpark method testEarlyStoppingIris.

@Test
public void testEarlyStoppingIris() {
    MultiLayerConfiguration conf = new NeuralNetConfiguration.Builder().optimizationAlgo(OptimizationAlgorithm.STOCHASTIC_GRADIENT_DESCENT).iterations(1).updater(Updater.SGD).weightInit(WeightInit.XAVIER).list().layer(0, new OutputLayer.Builder().nIn(4).nOut(3).lossFunction(LossFunctions.LossFunction.MCXENT).build()).pretrain(false).backprop(true).build();
    MultiLayerNetwork net = new MultiLayerNetwork(conf);
    net.setListeners(new ScoreIterationListener(1));
    JavaRDD<DataSet> irisData = getIris();
    EarlyStoppingModelSaver<MultiLayerNetwork> saver = new InMemoryModelSaver<>();
    EarlyStoppingConfiguration<MultiLayerNetwork> esConf = new EarlyStoppingConfiguration.Builder<MultiLayerNetwork>().epochTerminationConditions(new MaxEpochsTerminationCondition(5)).iterationTerminationConditions(new MaxTimeIterationTerminationCondition(1, TimeUnit.MINUTES)).scoreCalculator(new SparkDataSetLossCalculator(irisData, true, sc.sc())).modelSaver(saver).build();
    IEarlyStoppingTrainer<MultiLayerNetwork> trainer = new SparkEarlyStoppingTrainer(getContext().sc(), new ParameterAveragingTrainingMaster.Builder(irisBatchSize()).saveUpdater(true).averagingFrequency(1).build(), esConf, net, irisData);
    EarlyStoppingResult<MultiLayerNetwork> result = trainer.fit();
    System.out.println(result);
    assertEquals(5, result.getTotalEpochs());
    assertEquals(EarlyStoppingResult.TerminationReason.EpochTerminationCondition, result.getTerminationReason());
    Map<Integer, Double> scoreVsIter = result.getScoreVsEpoch();
    assertEquals(5, scoreVsIter.size());
    String expDetails = esConf.getEpochTerminationConditions().get(0).toString();
    assertEquals(expDetails, result.getTerminationDetails());
    MultiLayerNetwork out = result.getBestModel();
    assertNotNull(out);
    //Check that best score actually matches (returned model vs. manually calculated score)
    MultiLayerNetwork bestNetwork = result.getBestModel();
    double score = bestNetwork.score(new IrisDataSetIterator(150, 150).next());
    double bestModelScore = result.getBestModelScore();
    assertEquals(bestModelScore, score, 1e-3);
}
Also used : IrisDataSetIterator(org.deeplearning4j.datasets.iterator.impl.IrisDataSetIterator) DataSet(org.nd4j.linalg.dataset.DataSet) SparkEarlyStoppingTrainer(org.deeplearning4j.spark.earlystopping.SparkEarlyStoppingTrainer) SparkDataSetLossCalculator(org.deeplearning4j.spark.earlystopping.SparkDataSetLossCalculator) EarlyStoppingConfiguration(org.deeplearning4j.earlystopping.EarlyStoppingConfiguration) MultiLayerConfiguration(org.deeplearning4j.nn.conf.MultiLayerConfiguration) MultiLayerNetwork(org.deeplearning4j.nn.multilayer.MultiLayerNetwork) ScoreIterationListener(org.deeplearning4j.optimize.listeners.ScoreIterationListener) InMemoryModelSaver(org.deeplearning4j.earlystopping.saver.InMemoryModelSaver) MaxEpochsTerminationCondition(org.deeplearning4j.earlystopping.termination.MaxEpochsTerminationCondition) NeuralNetConfiguration(org.deeplearning4j.nn.conf.NeuralNetConfiguration) MaxTimeIterationTerminationCondition(org.deeplearning4j.earlystopping.termination.MaxTimeIterationTerminationCondition) Test(org.junit.Test)

Example 9 with MaxEpochsTerminationCondition

use of org.deeplearning4j.earlystopping.termination.MaxEpochsTerminationCondition in project deeplearning4j by deeplearning4j.

the class TestEarlyStoppingSpark method testBadTuning.

@Test
public void testBadTuning() {
    //Test poor tuning (high LR): should terminate on MaxScoreIterationTerminationCondition
    Nd4j.getRandom().setSeed(12345);
    MultiLayerConfiguration conf = new NeuralNetConfiguration.Builder().seed(12345).optimizationAlgo(OptimizationAlgorithm.STOCHASTIC_GRADIENT_DESCENT).iterations(1).updater(Updater.SGD).learningRate(//Intentionally huge LR
    10.0).weightInit(WeightInit.XAVIER).list().layer(0, new OutputLayer.Builder().nIn(4).nOut(3).activation(Activation.IDENTITY).lossFunction(LossFunctions.LossFunction.MSE).build()).pretrain(false).backprop(true).build();
    MultiLayerNetwork net = new MultiLayerNetwork(conf);
    net.setListeners(new ScoreIterationListener(1));
    JavaRDD<DataSet> irisData = getIris();
    EarlyStoppingModelSaver<MultiLayerNetwork> saver = new InMemoryModelSaver<>();
    EarlyStoppingConfiguration<MultiLayerNetwork> esConf = new EarlyStoppingConfiguration.Builder<MultiLayerNetwork>().epochTerminationConditions(new MaxEpochsTerminationCondition(5000)).iterationTerminationConditions(new MaxTimeIterationTerminationCondition(1, TimeUnit.MINUTES), //Initial score is ~2.5
    new MaxScoreIterationTerminationCondition(7.5)).scoreCalculator(new SparkDataSetLossCalculator(irisData, true, sc.sc())).modelSaver(saver).build();
    IEarlyStoppingTrainer<MultiLayerNetwork> trainer = new SparkEarlyStoppingTrainer(getContext().sc(), new ParameterAveragingTrainingMaster(true, 4, 1, 150 / 4, 1, 0), esConf, net, irisData);
    EarlyStoppingResult result = trainer.fit();
    assertTrue(result.getTotalEpochs() < 5);
    assertEquals(EarlyStoppingResult.TerminationReason.IterationTerminationCondition, result.getTerminationReason());
    String expDetails = new MaxScoreIterationTerminationCondition(7.5).toString();
    assertEquals(expDetails, result.getTerminationDetails());
}
Also used : OutputLayer(org.deeplearning4j.nn.conf.layers.OutputLayer) InMemoryModelSaver(org.deeplearning4j.earlystopping.saver.InMemoryModelSaver) MaxEpochsTerminationCondition(org.deeplearning4j.earlystopping.termination.MaxEpochsTerminationCondition) DataSet(org.nd4j.linalg.dataset.DataSet) SparkEarlyStoppingTrainer(org.deeplearning4j.spark.earlystopping.SparkEarlyStoppingTrainer) SparkDataSetLossCalculator(org.deeplearning4j.spark.earlystopping.SparkDataSetLossCalculator) NeuralNetConfiguration(org.deeplearning4j.nn.conf.NeuralNetConfiguration) ParameterAveragingTrainingMaster(org.deeplearning4j.spark.impl.paramavg.ParameterAveragingTrainingMaster) EarlyStoppingResult(org.deeplearning4j.earlystopping.EarlyStoppingResult) EarlyStoppingConfiguration(org.deeplearning4j.earlystopping.EarlyStoppingConfiguration) MultiLayerConfiguration(org.deeplearning4j.nn.conf.MultiLayerConfiguration) MaxScoreIterationTerminationCondition(org.deeplearning4j.earlystopping.termination.MaxScoreIterationTerminationCondition) MultiLayerNetwork(org.deeplearning4j.nn.multilayer.MultiLayerNetwork) ScoreIterationListener(org.deeplearning4j.optimize.listeners.ScoreIterationListener) MaxTimeIterationTerminationCondition(org.deeplearning4j.earlystopping.termination.MaxTimeIterationTerminationCondition) Test(org.junit.Test)

Example 10 with MaxEpochsTerminationCondition

use of org.deeplearning4j.earlystopping.termination.MaxEpochsTerminationCondition in project deeplearning4j by deeplearning4j.

the class TestEarlyStoppingSpark method testNoImprovementNEpochsTermination.

@Test
public void testNoImprovementNEpochsTermination() {
    //Idea: terminate training if score (test set loss) does not improve for 5 consecutive epochs
    //Simulate this by setting LR = 0.0
    Nd4j.getRandom().setSeed(12345);
    MultiLayerConfiguration conf = new NeuralNetConfiguration.Builder().seed(12345).optimizationAlgo(OptimizationAlgorithm.STOCHASTIC_GRADIENT_DESCENT).iterations(1).updater(Updater.SGD).learningRate(0.0).weightInit(WeightInit.XAVIER).list().layer(0, new OutputLayer.Builder().nIn(4).nOut(3).lossFunction(LossFunctions.LossFunction.MCXENT).build()).pretrain(false).backprop(true).build();
    MultiLayerNetwork net = new MultiLayerNetwork(conf);
    net.setListeners(new ScoreIterationListener(1));
    JavaRDD<DataSet> irisData = getIris();
    EarlyStoppingModelSaver<MultiLayerNetwork> saver = new InMemoryModelSaver<>();
    EarlyStoppingConfiguration<MultiLayerNetwork> esConf = new EarlyStoppingConfiguration.Builder<MultiLayerNetwork>().epochTerminationConditions(new MaxEpochsTerminationCondition(100), new ScoreImprovementEpochTerminationCondition(5)).iterationTerminationConditions(//Initial score is ~2.5
    new MaxScoreIterationTerminationCondition(7.5)).scoreCalculator(new SparkDataSetLossCalculator(irisData, true, sc.sc())).modelSaver(saver).build();
    IEarlyStoppingTrainer<MultiLayerNetwork> trainer = new SparkEarlyStoppingTrainer(getContext().sc(), new ParameterAveragingTrainingMaster(true, 4, 1, 150 / 10, 1, 0), esConf, net, irisData);
    EarlyStoppingResult result = trainer.fit();
    //Expect no score change due to 0 LR -> terminate after 6 total epochs
    //Normally expect 6 epochs exactly; get a little more than that here due to rounding + order of operations
    assertTrue(result.getTotalEpochs() < 12);
    assertEquals(EarlyStoppingResult.TerminationReason.EpochTerminationCondition, result.getTerminationReason());
    String expDetails = new ScoreImprovementEpochTerminationCondition(5).toString();
    assertEquals(expDetails, result.getTerminationDetails());
}
Also used : InMemoryModelSaver(org.deeplearning4j.earlystopping.saver.InMemoryModelSaver) MaxEpochsTerminationCondition(org.deeplearning4j.earlystopping.termination.MaxEpochsTerminationCondition) DataSet(org.nd4j.linalg.dataset.DataSet) ScoreImprovementEpochTerminationCondition(org.deeplearning4j.earlystopping.termination.ScoreImprovementEpochTerminationCondition) SparkEarlyStoppingTrainer(org.deeplearning4j.spark.earlystopping.SparkEarlyStoppingTrainer) SparkDataSetLossCalculator(org.deeplearning4j.spark.earlystopping.SparkDataSetLossCalculator) NeuralNetConfiguration(org.deeplearning4j.nn.conf.NeuralNetConfiguration) ParameterAveragingTrainingMaster(org.deeplearning4j.spark.impl.paramavg.ParameterAveragingTrainingMaster) EarlyStoppingResult(org.deeplearning4j.earlystopping.EarlyStoppingResult) EarlyStoppingConfiguration(org.deeplearning4j.earlystopping.EarlyStoppingConfiguration) MultiLayerConfiguration(org.deeplearning4j.nn.conf.MultiLayerConfiguration) MaxScoreIterationTerminationCondition(org.deeplearning4j.earlystopping.termination.MaxScoreIterationTerminationCondition) MultiLayerNetwork(org.deeplearning4j.nn.multilayer.MultiLayerNetwork) ScoreIterationListener(org.deeplearning4j.optimize.listeners.ScoreIterationListener) Test(org.junit.Test)

Aggregations

InMemoryModelSaver (org.deeplearning4j.earlystopping.saver.InMemoryModelSaver)27 MaxEpochsTerminationCondition (org.deeplearning4j.earlystopping.termination.MaxEpochsTerminationCondition)27 Test (org.junit.Test)27 ScoreIterationListener (org.deeplearning4j.optimize.listeners.ScoreIterationListener)26 NeuralNetConfiguration (org.deeplearning4j.nn.conf.NeuralNetConfiguration)25 MaxTimeIterationTerminationCondition (org.deeplearning4j.earlystopping.termination.MaxTimeIterationTerminationCondition)22 IrisDataSetIterator (org.deeplearning4j.datasets.iterator.impl.IrisDataSetIterator)19 MultiLayerConfiguration (org.deeplearning4j.nn.conf.MultiLayerConfiguration)17 MultiLayerNetwork (org.deeplearning4j.nn.multilayer.MultiLayerNetwork)17 DataSetIterator (org.nd4j.linalg.dataset.api.iterator.DataSetIterator)17 EarlyStoppingConfiguration (org.deeplearning4j.earlystopping.EarlyStoppingConfiguration)13 MaxScoreIterationTerminationCondition (org.deeplearning4j.earlystopping.termination.MaxScoreIterationTerminationCondition)13 IEarlyStoppingTrainer (org.deeplearning4j.earlystopping.trainer.IEarlyStoppingTrainer)13 DataSetLossCalculator (org.deeplearning4j.earlystopping.scorecalc.DataSetLossCalculator)11 OutputLayer (org.deeplearning4j.nn.conf.layers.OutputLayer)11 DataSet (org.nd4j.linalg.dataset.DataSet)11 ComputationGraphConfiguration (org.deeplearning4j.nn.conf.ComputationGraphConfiguration)10 ComputationGraph (org.deeplearning4j.nn.graph.ComputationGraph)10 ListDataSetIterator (org.deeplearning4j.datasets.iterator.impl.ListDataSetIterator)9 EarlyStoppingTrainer (org.deeplearning4j.earlystopping.trainer.EarlyStoppingTrainer)9