Class AdaDelta

java.lang.Object
org.tensorflow.framework.optimizers.Optimizer
org.tensorflow.framework.optimizers.AdaDelta
public class AdaDelta extends Optimizer
Optimizer that implements the Adadelta algorithm.

Adadelta optimization is a stochastic gradient descent method that is based on adaptive learning rate per dimension to address two drawbacks:

  • the continual decay of learning rates throughout training
  • the need for a manually selected global learning rate

Adadelta is a more robust extension of Adagrad that adapts learning rates based on a moving window of gradient updates, instead of accumulating all past gradients. This way, Adadelta continues learning even when many updates have been done. Compared to Adagrad, in the original version of Adadelta you don't have to set an initial learning rate. In this version, initial learning rate can be set, as in most other optimizers.

According to section 4.3 ("Effective Learning rates"), near the end of training step sizes converge to 1 which is effectively a high learning rate which would cause divergence. This occurs only near the end of the training as gradients and step sizes are small, and the epsilon constant in the numerator and denominator dominate past gradients and parameter updates which converge the learning rate to 1.

According to section 4.4("Speech Data"),where a large neural network with 4 hidden layers was trained on a corpus of US English data, ADADELTA was used with 100 network replicas.The epsilon used is 1e-6 with rho=0.95 which converged faster than ADAGRAD, by the following construction: new AdaDelta(graph, 1.0f, 0.95f, 1e-6f);

See Also:

  • Field Details

  • Constructor Details

    • AdaDelta

      public AdaDelta(Graph graph)

    • AdaDelta

      public AdaDelta(Graph graph, float learningRate)
      Creates an AdaDelta Optimizer
      Parameters:
      graph - the TensorFlow Graph
      learningRate - the learning rate

    • AdaDelta

      public AdaDelta(Graph graph, float learningRate, float rho, float epsilon)
      Creates an AdaDelta Optimizer
      Parameters:
      graph - the TensorFlow Graph
      learningRate - the learning rate
      rho - The decay factor
      epsilon - A constant epsilon used to better conditioning the grad update

    • AdaDelta

      public AdaDelta(Graph graph, String name, float learningRate)
      Creates an AdaDelta Optimizer
      Parameters:
      graph - the TensorFlow Graph
      name - the name for this Optimizer (defaults to 'Adadelta')
      learningRate - the learning rate

    • AdaDelta

      public AdaDelta(Graph graph, String name, float learningRate, float rho, float epsilon)
      Creates an AdaDelta Optimizer
      Parameters:
      graph - the TensorFlow Graph
      name - the name for this Optimizer (defaults to 'Adadelta')
      learningRate - the learning rate
      rho - The decay factor
      epsilon - A constant epsilon used to better conditioning the grad update

  • Method Details

    • createSlots

      protected void createSlots(List<Output<? extends TType>> variables)
      Performs a No-op slot creation method.
      Overrides:
      createSlots in class Optimizer
      Parameters:
      variables - The variables to create slots for.

    • applyDense

      protected <T extends TType> Op applyDense(Ops deps, Output<T> gradient, Output<T> variable)
      Generates the gradient update operations for the specific variable and gradient.
      Specified by:
      applyDense in class Optimizer
      Type Parameters:
      T - The type of the variable.
      Parameters:
      gradient - The gradient to use.
      variable - The variable to update.
      Returns:
      An operand which applies the desired optimizer update to the variable.

    • toString

      public String toString()
      Overrides:
      toString in class Object

    • getOptimizerName

      public String getOptimizerName()
      Get the Name of the optimizer.
      Specified by:
      getOptimizerName in class Optimizer
      Returns:
      The optimizer name.