Field Detail

• DEFAULT_ROC_THRESHOLD_STEPS

  public static final int DEFAULT_ROC_THRESHOLD_STEPS

  See Also:

  Constant Field Values

• DEFAULT_EVAL_SCORE_BATCH_SIZE

  public static final int DEFAULT_EVAL_SCORE_BATCH_SIZE

  See Also:

  Constant Field Values

• DEFAULT_EVAL_WORKERS

  public static final int DEFAULT_EVAL_WORKERS

  See Also:

  Constant Field Values

Constructor Detail

• SparkComputationGraph

  public SparkComputationGraph(org.apache.spark.SparkContext sparkContext,
                               ComputationGraph network,
                               TrainingMaster trainingMaster)

  Instantiate a ComputationGraph instance with the given context and network.

  Parameters:

  sparkContext - the spark context to use

  network - the network to use

• SparkComputationGraph

  public SparkComputationGraph(org.apache.spark.api.java.JavaSparkContext javaSparkContext,
                               ComputationGraph network,
                               TrainingMaster trainingMaster)

• SparkComputationGraph

  public SparkComputationGraph(org.apache.spark.SparkContext sparkContext,
                               ComputationGraphConfiguration conf,
                               TrainingMaster trainingMaster)

• SparkComputationGraph

  public SparkComputationGraph(org.apache.spark.api.java.JavaSparkContext sparkContext,
                               ComputationGraphConfiguration conf,
                               TrainingMaster trainingMaster)

Method Detail

• getSparkContext

  public org.apache.spark.api.java.JavaSparkContext getSparkContext()

• setCollectTrainingStats

  public void setCollectTrainingStats(boolean collectTrainingStats)

• getSparkTrainingStats

  public SparkTrainingStats getSparkTrainingStats()

• getNetwork

  public ComputationGraph getNetwork()

  Returns:

  The trained ComputationGraph

• getTrainingMaster

  public TrainingMaster getTrainingMaster()

  Returns:

  The TrainingMaster for this network

• setNetwork

  public void setNetwork(ComputationGraph network)

• getDefaultEvaluationWorkers

  public int getDefaultEvaluationWorkers()

  Returns the currently set default number of evaluation workers/threads. Note that when the number of workers is provided explicitly in an evaluation method, the default value is not used.
  In many cases, we may want this to be smaller than the number of Spark threads, to reduce memory requirements. For example, with 32 Spark threads and a large network, we don't want to spin up 32 instances of the network to perform evaluation. Better (for memory requirements, and reduced cache thrashing) to use say 4 workers.
  If it is not set explicitly, DEFAULT_EVAL_WORKERS will be used

  Returns:

  Default number of evaluation workers (threads).

• setDefaultEvaluationWorkers

  public void setDefaultEvaluationWorkers(int workers)

  Set the default number of evaluation workers/threads. Note that when the number of workers is provided explicitly in an evaluation method, the default value is not used.
  In many cases, we may want this to be smaller than the number of Spark threads, to reduce memory requirements. For example, with 32 Spark threads and a large network, we don't want to spin up 32 instances of the network to perform evaluation. Better (for memory requirements, and reduced cache thrashing) to use say 4 workers.
  If it is not set explicitly, DEFAULT_EVAL_WORKERS will be used

• fit

  public ComputationGraph fit(org.apache.spark.rdd.RDD<DataSet> rdd)

  Fit the ComputationGraph with the given data set

  Parameters:

  rdd - Data to train on

  Returns:

  Trained network

• fit

  public ComputationGraph fit(org.apache.spark.api.java.JavaRDD<DataSet> rdd)

  Fit the ComputationGraph with the given data set

  Parameters:

  rdd - Data to train on

  Returns:

  Trained network

• fit

  public ComputationGraph fit(java.lang.String path)

  Fit the SparkComputationGraph network using a directory of serialized DataSet objects The assumption here is that the directory contains a number of DataSet objects, each serialized using DataSet.save(OutputStream)

  Parameters:

  path - Path to the directory containing the serialized DataSet objcets

  Returns:

  The MultiLayerNetwork after training

• fit

  @Deprecated
  public ComputationGraph fit(java.lang.String path,
                                          int minPartitions)

  Deprecated. Use fit(String)

• fitPaths

  public ComputationGraph fitPaths(org.apache.spark.api.java.JavaRDD<java.lang.String> paths)

  Fit the network using a list of paths for serialized DataSet objects.

  Parameters:

  paths - List of paths

  Returns:

  trained network

• fitPaths

  public ComputationGraph fitPaths(org.apache.spark.api.java.JavaRDD<java.lang.String> paths,
                                   DataSetLoader loader)

• fitMultiDataSet

  public ComputationGraph fitMultiDataSet(org.apache.spark.rdd.RDD<MultiDataSet> rdd)

  Fit the ComputationGraph with the given data set

  Parameters:

  rdd - Data to train on

  Returns:

  Trained network

• fitMultiDataSet

  public ComputationGraph fitMultiDataSet(org.apache.spark.api.java.JavaRDD<MultiDataSet> rdd)

  Fit the ComputationGraph with the given data set

  Parameters:

  rdd - Data to train on

  Returns:

  Trained network

• fitMultiDataSet

  public ComputationGraph fitMultiDataSet(java.lang.String path)

  Fit the SparkComputationGraph network using a directory of serialized MultiDataSet objects The assumption here is that the directory contains a number of serialized MultiDataSet objects

  Parameters:

  path - Path to the directory containing the serialized MultiDataSet objcets

  Returns:

  The MultiLayerNetwork after training

• fitPathsMultiDataSet

  public ComputationGraph fitPathsMultiDataSet(org.apache.spark.api.java.JavaRDD<java.lang.String> paths)

  Fit the network using a list of paths for serialized MultiDataSet objects.

  Parameters:

  paths - List of paths

  Returns:

  trained network

• fitPaths

  public ComputationGraph fitPaths(org.apache.spark.api.java.JavaRDD<java.lang.String> paths,
                                   MultiDataSetLoader loader)

• fitMultiDataSet

  @Deprecated
  public ComputationGraph fitMultiDataSet(java.lang.String path,
                                                      int minPartitions)

  Deprecated. use fitMultiDataSet(String)

• getScore

  public double getScore()

  Gets the last (average) minibatch score from calling fit. This is the average score across all executors for the last minibatch executed in each worker

• setScore

  public void setScore(double lastScore)

• calculateScore

  public double calculateScore(org.apache.spark.api.java.JavaRDD<DataSet> data,
                               boolean average)

  Calculate the score for all examples in the provided JavaRDD<DataSet>, either by summing or averaging over the entire data set. To calculate a score for each example individually, use scoreExamples(JavaPairRDD, boolean) or one of the similar methods. Uses default minibatch size in each worker, DEFAULT_EVAL_SCORE_BATCH_SIZE

  Parameters:

  data - Data to score

  average - Whether to sum the scores, or average them

• calculateScore

  public double calculateScore(org.apache.spark.api.java.JavaRDD<DataSet> data,
                               boolean average,
                               int minibatchSize)

  Calculate the score for all examples in the provided JavaRDD<DataSet>, either by summing or averaging over the entire data set. To calculate a score for each example individually, use scoreExamples(JavaPairRDD, boolean) or one of the similar methods

  Parameters:

  data - Data to score

  average - Whether to sum the scores, or average them

  minibatchSize - The number of examples to use in each minibatch when scoring. If more examples are in a partition than this, multiple scoring operations will be done (to avoid using too much memory by doing the whole partition in one go)

• calculateScoreMultiDataSet

  public double calculateScoreMultiDataSet(org.apache.spark.api.java.JavaRDD<MultiDataSet> data,
                                           boolean average)

  Calculate the score for all examples in the provided JavaRDD<MultiDataSet>, either by summing or averaging over the entire data set. Uses default minibatch size in each worker, DEFAULT_EVAL_SCORE_BATCH_SIZE

  Parameters:

  data - Data to score

  average - Whether to sum the scores, or average them

• calculateScoreMultiDataSet

  public double calculateScoreMultiDataSet(org.apache.spark.api.java.JavaRDD<MultiDataSet> data,
                                           boolean average,
                                           int minibatchSize)

  Calculate the score for all examples in the provided JavaRDD<MultiDataSet>, either by summing or averaging over the entire data set. *

  Parameters:

  data - Data to score

  average - Whether to sum the scores, or average them

  minibatchSize - The number of examples to use in each minibatch when scoring. If more examples are in a partition than this, multiple scoring operations will be done (to avoid using too much memory by doing the whole partition in one go)

• scoreExamples

  public org.apache.spark.api.java.JavaDoubleRDD scoreExamples(org.apache.spark.api.java.JavaRDD<DataSet> data,
                                                               boolean includeRegularizationTerms)

  DataSet version of scoreExamples(JavaRDD, boolean)

• scoreExamples

  public org.apache.spark.api.java.JavaDoubleRDD scoreExamples(org.apache.spark.api.java.JavaRDD<DataSet> data,
                                                               boolean includeRegularizationTerms,
                                                               int batchSize)

  DataSet version of scoreExamples(JavaPairRDD, boolean, int)

• scoreExamples

  public <K> org.apache.spark.api.java.JavaPairRDD<K,java.lang.Double> scoreExamples(org.apache.spark.api.java.JavaPairRDD<K,DataSet> data,
                                                                                     boolean includeRegularizationTerms)

  DataSet version of scoreExamples(JavaPairRDD, boolean)

• scoreExamples

  public <K> org.apache.spark.api.java.JavaPairRDD<K,java.lang.Double> scoreExamples(org.apache.spark.api.java.JavaPairRDD<K,DataSet> data,
                                                                                     boolean includeRegularizationTerms,
                                                                                     int batchSize)

  DataSet version of scoreExamples(JavaPairRDD, boolean, int)

• scoreExamplesMultiDataSet

  public org.apache.spark.api.java.JavaDoubleRDD scoreExamplesMultiDataSet(org.apache.spark.api.java.JavaRDD<MultiDataSet> data,
                                                                           boolean includeRegularizationTerms)

  Score the examples individually, using the default batch size DEFAULT_EVAL_SCORE_BATCH_SIZE. Unlike calculateScore(JavaRDD, boolean), this method returns a score for each example separately. If scoring is needed for specific examples use either scoreExamples(JavaPairRDD, boolean) or scoreExamples(JavaPairRDD, boolean, int) which can have a key for each example.

  Parameters:

  data - Data to score

  includeRegularizationTerms - If true: include the l1/l2 regularization terms with the score (if any)

  Returns:

  A JavaDoubleRDD containing the scores of each example

  See Also:

  ComputationGraph.scoreExamples(MultiDataSet, boolean)

• scoreExamplesMultiDataSet

  public org.apache.spark.api.java.JavaDoubleRDD scoreExamplesMultiDataSet(org.apache.spark.api.java.JavaRDD<MultiDataSet> data,
                                                                           boolean includeRegularizationTerms,
                                                                           int batchSize)

  Score the examples individually, using a specified batch size. Unlike calculateScore(JavaRDD, boolean), this method returns a score for each example separately. If scoring is needed for specific examples use either scoreExamples(JavaPairRDD, boolean) or scoreExamples(JavaPairRDD, boolean, int) which can have a key for each example.

  Parameters:

  data - Data to score

  includeRegularizationTerms - If true: include the l1/l2 regularization terms with the score (if any)

  batchSize - Batch size to use when doing scoring

  Returns:

  A JavaDoubleRDD containing the scores of each example

  See Also:

  ComputationGraph.scoreExamples(MultiDataSet, boolean)

• scoreExamplesMultiDataSet

  public <K> org.apache.spark.api.java.JavaPairRDD<K,java.lang.Double> scoreExamplesMultiDataSet(org.apache.spark.api.java.JavaPairRDD<K,MultiDataSet> data,
                                                                                                 boolean includeRegularizationTerms)

  Score the examples individually, using the default batch size DEFAULT_EVAL_SCORE_BATCH_SIZE. Unlike calculateScore(JavaRDD, boolean), this method returns a score for each example separately
  Note: The provided JavaPairRDD has a key that is associated with each example and returned score.
  Note: The DataSet objects passed in must have exactly one example in them (otherwise: can't have a 1:1 association between keys and data sets to score)

  Type Parameters:

  K - Key type

  Parameters:

  data - Data to score

  includeRegularizationTerms - If true: include the l1/l2 regularization terms with the score (if any)

  Returns:

  A JavaPairRDD<K,Double> containing the scores of each example

  See Also:

  MultiLayerNetwork.scoreExamples(DataSet, boolean)

• feedForwardWithKeySingle

  public <K> org.apache.spark.api.java.JavaPairRDD<K,INDArray> feedForwardWithKeySingle(org.apache.spark.api.java.JavaPairRDD<K,INDArray> featuresData,
                                                                                        int batchSize)

  Feed-forward the specified data, with the given keys. i.e., get the network output/predictions for the specified data

  Type Parameters:

  K - Type of data for key - may be anything

  Parameters:

  featuresData - Features data to feed through the network

  batchSize - Batch size to use when doing feed forward operations

  Returns:

  Network output given the input, by key

• feedForwardWithKey

  public <K> org.apache.spark.api.java.JavaPairRDD<K,INDArray[]> feedForwardWithKey(org.apache.spark.api.java.JavaPairRDD<K,INDArray[]> featuresData,
                                                                                    int batchSize)

  Feed-forward the specified data, with the given keys. i.e., get the network output/predictions for the specified data

  Type Parameters:

  K - Type of data for key - may be anything

  Parameters:

  featuresData - Features data to feed through the network

  batchSize - Batch size to use when doing feed forward operations

  Returns:

  Network output given the input, by key

• scoreExamplesMultiDataSet

  public <K> org.apache.spark.api.java.JavaPairRDD<K,java.lang.Double> scoreExamplesMultiDataSet(org.apache.spark.api.java.JavaPairRDD<K,MultiDataSet> data,
                                                                                                 boolean includeRegularizationTerms,
                                                                                                 int batchSize)

  Score the examples individually, using a specified batch size. Unlike calculateScore(JavaRDD, boolean), this method returns a score for each example separately
  Note: The provided JavaPairRDD has a key that is associated with each example and returned score.
  Note: The DataSet objects passed in must have exactly one example in them (otherwise: can't have a 1:1 association between keys and data sets to score)

  Type Parameters:

  K - Key type

  Parameters:

  data - Data to score

  includeRegularizationTerms - If true: include the l1/l2 regularization terms with the score (if any)

  Returns:

  A JavaPairRDD<K,Double> containing the scores of each example

  See Also:

  MultiLayerNetwork.scoreExamples(DataSet, boolean)

• evaluate

  public Evaluation evaluate(org.apache.spark.rdd.RDD<DataSet> data)

  RDD<DataSet> overload of evaluate(JavaRDD)

• evaluate

  public Evaluation evaluate(org.apache.spark.api.java.JavaRDD<DataSet> data)

  Evaluate the network (classification performance) in a distributed manner on the provided data

  Parameters:

  data - Data to evaluate on

  Returns:

  Evaluation object; results of evaluation on all examples in the data set

• evaluate

  public Evaluation evaluate(org.apache.spark.rdd.RDD<DataSet> data,
                             java.util.List<java.lang.String> labelsList)

  RDD<DataSet> overload of evaluate(JavaRDD, List)

• evaluateRegression

  public RegressionEvaluation evaluateRegression(org.apache.spark.api.java.JavaRDD<DataSet> data)

  Evaluate the network (regression performance) in a distributed manner on the provided data

  Parameters:

  data - Data to evaluate

  Returns:

  RegressionEvaluation instance with regression performance

• evaluateRegression

  public RegressionEvaluation evaluateRegression(org.apache.spark.api.java.JavaRDD<DataSet> data,
                                                 int minibatchSize)

  Evaluate the network (regression performance) in a distributed manner on the provided data

  Parameters:

  data - Data to evaluate

  minibatchSize - Minibatch size to use when doing performing evaluation

  Returns:

  RegressionEvaluation instance with regression performance

• evaluate

  public Evaluation evaluate(org.apache.spark.api.java.JavaRDD<DataSet> data,
                             java.util.List<java.lang.String> labelsList)

  Evaluate the network (classification performance) in a distributed manner, using default batch size and a provided list of labels

  Parameters:

  data - Data to evaluate on

  labelsList - List of labels used for evaluation

  Returns:

  Evaluation object; results of evaluation on all examples in the data set

• evaluateROC

  public ROC evaluateROC(org.apache.spark.api.java.JavaRDD<DataSet> data)

  Perform ROC analysis/evaluation on the given DataSet in a distributed manner, using the default number of threshold steps (DEFAULT_ROC_THRESHOLD_STEPS) and the default minibatch size (DEFAULT_EVAL_SCORE_BATCH_SIZE)

  Parameters:

  data - Test set data (to evaluate on)

  Returns:

  ROC for the entire data set

• evaluateROC

  public ROC evaluateROC(org.apache.spark.api.java.JavaRDD<DataSet> data,
                         int thresholdSteps,
                         int evaluationMinibatchSize)

  Perform ROC analysis/evaluation on the given DataSet in a distributed manner

  Parameters:

  data - Test set data (to evaluate on)

  thresholdSteps - Number of threshold steps for ROC - see ROC

  evaluationMinibatchSize - Minibatch size to use when performing ROC evaluation

  Returns:

  ROC for the entire data set

• evaluateROCMultiClass

  public ROCMultiClass evaluateROCMultiClass(org.apache.spark.api.java.JavaRDD<DataSet> data)

  Perform ROC analysis/evaluation (for the multi-class case, using ROCMultiClass on the given DataSet in a distributed manner

  Parameters:

  data - Test set data (to evaluate on)

  Returns:

  ROC for the entire data set

• evaluateROCMultiClass

  public ROCMultiClass evaluateROCMultiClass(org.apache.spark.api.java.JavaRDD<DataSet> data,
                                             int thresholdSteps,
                                             int evaluationMinibatchSize)

  Perform ROC analysis/evaluation (for the multi-class case, using ROCMultiClass on the given DataSet in a distributed manner

  Parameters:

  data - Test set data (to evaluate on)

  thresholdSteps - Number of threshold steps for ROC - see ROC

  evaluationMinibatchSize - Minibatch size to use when performing ROC evaluation

  Returns:

  ROCMultiClass for the entire data set

• evaluate

  public Evaluation evaluate(org.apache.spark.api.java.JavaRDD<DataSet> data,
                             java.util.List<java.lang.String> labelsList,
                             int evalBatchSize)

  Evaluate the network (classification performance) in a distributed manner, using specified batch size and a provided list of labels

  Parameters:

  data - Data to evaluate on

  labelsList - List of labels used for evaluation

  evalBatchSize - Batch size to use when conducting evaluations

  Returns:

  Evaluation object; results of evaluation on all examples in the data set

• evaluateMDS

  public Evaluation evaluateMDS(org.apache.spark.api.java.JavaRDD<MultiDataSet> data)

  Evaluate the network (classification performance) in a distributed manner on the provided data

• evaluateMDS

  public Evaluation evaluateMDS(org.apache.spark.api.java.JavaRDD<MultiDataSet> data,
                                int minibatchSize)

  Evaluate the network (classification performance) in a distributed manner on the provided data

• evaluateRegressionMDS

  public RegressionEvaluation evaluateRegressionMDS(org.apache.spark.api.java.JavaRDD<MultiDataSet> data)

  Evaluate the network (regression performance) in a distributed manner on the provided data

  Parameters:

  data - Data to evaluate

  Returns:

  RegressionEvaluation instance with regression performance

• evaluateRegressionMDS

  public RegressionEvaluation evaluateRegressionMDS(org.apache.spark.api.java.JavaRDD<MultiDataSet> data,
                                                    int minibatchSize)

  Evaluate the network (regression performance) in a distributed manner on the provided data

  Parameters:

  data - Data to evaluate

  minibatchSize - Minibatch size to use when doing performing evaluation

  Returns:

  RegressionEvaluation instance with regression performance

• evaluateROCMDS

  public ROC evaluateROCMDS(org.apache.spark.api.java.JavaRDD<MultiDataSet> data)

  Perform ROC analysis/evaluation on the given DataSet in a distributed manner, using the default number of threshold steps (DEFAULT_ROC_THRESHOLD_STEPS) and the default minibatch size (DEFAULT_EVAL_SCORE_BATCH_SIZE)

  Parameters:

  data - Test set data (to evaluate on)

  Returns:

  ROC for the entire data set

• evaluateROCMDS

  public ROC evaluateROCMDS(org.apache.spark.api.java.JavaRDD<MultiDataSet> data,
                            int rocThresholdNumSteps,
                            int minibatchSize)

  Perform ROC analysis/evaluation on the given DataSet in a distributed manner, using the specified number of steps and minibatch size

  Parameters:

  data - Test set data (to evaluate on)

  rocThresholdNumSteps - See ROC for details

  minibatchSize - Minibatch size for evaluation

  Returns:

  ROC for the entire data set

• doEvaluation

  public <T extends IEvaluation> T doEvaluation(org.apache.spark.api.java.JavaRDD<DataSet> data,
                                                T emptyEvaluation,
                                                int evalBatchSize)

  Perform distributed evaluation of any type of IEvaluation. For example, Evaluation, RegressionEvaluation, ROC, ROCMultiClass etc.

  Type Parameters:

  T - Type of evaluation instance to return

  Parameters:

  data - Data to evaluate on

  emptyEvaluation - Empty evaluation instance. This is the starting point (serialized/duplicated, then merged)

  evalBatchSize - Evaluation batch size

  Returns:

  IEvaluation instance

• doEvaluation

  public <T extends IEvaluation> T[] doEvaluation(org.apache.spark.api.java.JavaRDD<DataSet> data,
                                                  int evalBatchSize,
                                                  T... emptyEvaluations)

  Perform distributed evaluation on a single output ComputationGraph form DataSet objects using Spark. Can be used to perform multiple evaluations on this single output (for example, Evaluation and ROC) at the same time.
  Note that the default number of worker threads getDefaultEvaluationWorkers() will be used

  Parameters:

  data - Data to evaluatie

  evalBatchSize - Minibatch size for evaluation

  emptyEvaluations - Evaluations to perform

  Returns:

  Evaluations

• doEvaluation

  public <T extends IEvaluation> T[] doEvaluation(org.apache.spark.api.java.JavaRDD<DataSet> data,
                                                  int evalNumWorkers,
                                                  int evalBatchSize,
                                                  T... emptyEvaluations)

  Perform distributed evaluation on a single output ComputationGraph form DataSet objects using Spark. Can be used to perform multiple evaluations on this single output (for example, Evaluation and ROC) at the same time.
  

  Parameters:

  data - Data to evaluatie

  evalNumWorkers - Number of worker threads (per machine) to use for evaluation. May want tis to be less than the number of Spark threads per machine/JVM to reduce memory requirements

  evalBatchSize - Minibatch size for evaluation

  emptyEvaluations - Evaluations to perform

  Returns:

  Evaluations

• doEvaluationMDS

  public <T extends IEvaluation> T[] doEvaluationMDS(org.apache.spark.api.java.JavaRDD<MultiDataSet> data,
                                                     int evalBatchSize,
                                                     T... emptyEvaluations)

  Perform distributed evaluation on a single output ComputationGraph form MultiDataSet objects using Spark. Can be used to perform multiple evaluations on this single output (for example, Evaluation and ROC) at the same time.

  Parameters:

  data - Data to evaluatie

  evalBatchSize - Minibatch size for evaluation

  emptyEvaluations - Evaluations to perform

  Returns:

  Evaluations

• doEvaluationMDS

  public <T extends IEvaluation> T[] doEvaluationMDS(org.apache.spark.api.java.JavaRDD<MultiDataSet> data,
                                                     int evalNumWorkers,
                                                     int evalBatchSize,
                                                     T... emptyEvaluations)

• doEvaluation

  public IEvaluation[] doEvaluation(org.apache.spark.api.java.JavaRDD<java.lang.String> data,
                                    int evalNumWorkers,
                                    int evalBatchSize,
                                    DataSetLoader loader,
                                    IEvaluation... emptyEvaluations)

  Perform evaluation on serialized DataSet objects on disk, (potentially in any format), that are loaded using an DataSetLoader

  Parameters:

  data - List of paths to the data (that can be loaded as / converted to DataSets)

  evalNumWorkers - Number of workers to perform evaluation with. To reduce memory requirements and cache thrashing, it is common to set this to a lower value than the number of spark threads per JVM/executor

  evalBatchSize - Batch size to use when performing evaluation

  loader - Used to load DataSets from their paths

  emptyEvaluations - Evaluations to perform

  Returns:

  Evaluation

• doEvaluation

  public IEvaluation[] doEvaluation(org.apache.spark.api.java.JavaRDD<java.lang.String> data,
                                    int evalNumWorkers,
                                    int evalBatchSize,
                                    MultiDataSetLoader loader,
                                    IEvaluation... emptyEvaluations)

  Perform evaluation on serialized MultiDataSet objects on disk, (potentially in any format), that are loaded using an MultiDataSetLoader

  Parameters:

  data - List of paths to the data (that can be loaded as / converted to DataSets)

  evalNumWorkers - Number of workers to perform evaluation with. To reduce memory requirements and cache thrashing, it is common to set this to a lower value than the number of spark threads per JVM/executor

  evalBatchSize - Batch size to use when performing evaluation

  loader - Used to load MultiDataSets from their paths

  emptyEvaluations - Evaluations to perform

  Returns:

  Evaluation

• doEvaluation

  protected IEvaluation[] doEvaluation(org.apache.spark.api.java.JavaRDD<java.lang.String> data,
                                       int evalNumWorkers,
                                       int evalBatchSize,
                                       DataSetLoader loader,
                                       MultiDataSetLoader mdsLoader,
                                       IEvaluation... emptyEvaluations)