Class LinearSGDOptions

All Implemented Interfaces:

public class LinearSGDOptions extends Object implements
CLI options for training a linear classifier.
  • Field Details

    • sgoOptions

      public GradientOptimiserOptions sgoOptions
      The gradient descent optimiser options.
    • sgdEpochs

      @Option(longName="sgd-epochs", usage="Number of SGD epochs.") public int sgdEpochs
      Number of SGD epochs.
    • sgdObjective

      @Option(longName="sgd-objective", usage="Loss function.") public LinearSGDOptions.LossEnum sgdObjective
      Loss function.
    • sgdLoggingInterval

      @Option(longName="sgd-logging-interval", usage="Log the objective after <int> examples.") public int sgdLoggingInterval
      Log the objective after n examples.
    • sgdMinibatchSize

      @Option(longName="sgd-minibatch-size", usage="Minibatch size.") public int sgdMinibatchSize
      Minibatch size.
    • sgdSeed

      @Option(longName="sgd-seed", usage="Sets the random seed for the LinearSGDTrainer.") public long sgdSeed
      Sets the random seed for the LinearSGDTrainer.
  • Constructor Details

    • LinearSGDOptions

      public LinearSGDOptions()
  • Method Details

    • getLoss

      public MultiLabelObjective getLoss()
      Returns the loss function specified in the arguments.
      The loss function.
    • getTrainer

      public LinearSGDTrainer getTrainer()
      Gets the LinearSGDTrainer specified by the options in this object.
      The configured trainer.