Class/Object

org.platanios.tensorflow.api.learn.layers.rnn.cell

DropoutWrapper

Related Docs: object DropoutWrapper | package cell

Permalink

class DropoutWrapper[O, OS, S, SS] extends RNNCell[O, OS, S, SS]

RNN cell that applies dropout to the provided RNN cell.

Note that currently, a different dropout mask is used for each time step in an RNN (i.e., not using the variational recurrent dropout method described in ["A Theoretically Grounded Application of Dropout in Recurrent Neural Networks"](https://arxiv.org/abs/1512.05287).

Note also that for LSTM cells, no dropout is applied to the memory tensor of the state. It is only applied to the state tensor.

Linear Supertypes
RNNCell[O, OS, S, SS], Layer[Tuple[O, S], Tuple[O, S]], AnyRef, Any
Ordering
  1. Alphabetic
  2. By Inheritance
Inherited
  1. DropoutWrapper
  2. RNNCell
  3. Layer
  4. AnyRef
  5. Any
  1. Hide All
  2. Show All
Visibility
  1. Public
  2. All

Instance Constructors

  1. new DropoutWrapper(name: String, cell: RNNCell[O, OS, S, SS], inputKeepProbability: Float = 1.0f, outputKeepProbability: Float = 1.0f, stateKeepProbability: Float = 1.0f, seed: Option[Int] = None)(implicit evO: Aux[O, OS], evS: Aux[S, SS], evODropout: Supported[O], evSDropout: Supported[S])

    Permalink

    name

    Name scope (also acting as variable scope) for this layer.

    cell

    RNN cell on which to perform dropout.

    inputKeepProbability

    Keep probability for the input of the RNN cell.

    outputKeepProbability

    Keep probability for the output of the RNN cell.

    stateKeepProbability

    Keep probability for the output state of the RNN cell.

    seed

    Optional random seed, used to generate a random seed pair for the random number generator, when combined with the graph-level seed.

Value Members

  1. final def !=(arg0: Any): Boolean

    Permalink
    Definition Classes
    AnyRef → Any

  2. final def ##(): Int

    Permalink
    Definition Classes
    AnyRef → Any

  3. def +(other: Layer[Tuple[O, S], Tuple[O, S]]): Concatenate[Tuple[O, S], Tuple[O, S]]

    Permalink
    Definition Classes
    Layer

  4. def ++(others: Seq[Layer[Tuple[O, S], Tuple[O, S]]]): Concatenate[Tuple[O, S], Tuple[O, S]]

    Permalink
    Definition Classes
    Layer

  5. final def ==(arg0: Any): Boolean

    Permalink
    Definition Classes
    AnyRef → Any

  6. def >>[S](other: Layer[Tuple[O, S], S]): Compose[Tuple[O, S], Tuple[O, S], S]

    Permalink
    Definition Classes
    Layer

  7. def apply(input: Tuple[O, S])(implicit mode: Mode): Tuple[O, S]

    Permalink
    Definition Classes
    Layer

  8. final def asInstanceOf[T0]: T0

    Permalink
    Definition Classes
    Any

  9. val cell: RNNCell[O, OS, S, SS]

    Permalink

    RNN cell on which to perform dropout.

  10. def clone(): AnyRef

    Permalink
    Attributes
    protected[java.lang]
    Definition Classes
    AnyRef
    Annotations
    @throws( ... )

  11. def compose[S](other: Layer[Tuple[O, S], S]): Compose[Tuple[O, S], Tuple[O, S], S]

    Permalink
    Definition Classes
    Layer

  12. def concatenate(others: Layer[Tuple[O, S], Tuple[O, S]]*): Concatenate[Tuple[O, S], Tuple[O, S]]

    Permalink
    Definition Classes
    Layer

  13. final def createCell(mode: Mode, inputShape: OS): ops.rnn.cell.RNNCell[O, OS, S, SS]

    Permalink
    Definition Classes
    RNNCell

  14. def createCellWithoutContext(mode: Mode, inputShape: OS): ops.rnn.cell.RNNCell[O, OS, S, SS]

    Permalink
    Definition Classes
    DropoutWrapper → RNNCell

  15. final def currentStep: ops.Output

    Permalink
    Definition Classes
    Layer

  16. final def eq(arg0: AnyRef): Boolean

    Permalink
    Definition Classes
    AnyRef

  17. def equals(arg0: Any): Boolean

    Permalink
    Definition Classes
    AnyRef → Any

  18. def finalize(): Unit

    Permalink
    Attributes
    protected[java.lang]
    Definition Classes
    AnyRef
    Annotations
    @throws( classOf[java.lang.Throwable] )

  19. def forward(input: Tuple[O, S])(implicit mode: Mode): Tuple[O, S]

    Permalink
    Definition Classes
    Layer

  20. final def forwardWithoutContext(input: Tuple[O, S])(implicit mode: Mode): Tuple[O, S]

    Permalink
    Definition Classes
    RNNCell → Layer

  21. final def getClass(): Class[_]

    Permalink
    Definition Classes
    AnyRef → Any

  22. final def getParameter(name: String, dataType: types.DataType, shape: core.Shape, initializer: Initializer = null, regularizer: Regularizer = null, trainable: Boolean = true, reuse: Reuse = ReuseOrCreateNew, collections: Set[Key[ops.variables.Variable]] = Set.empty, cachingDevice: (OpSpecification) ⇒ String = null): ops.Output

    Permalink
    Definition Classes
    Layer

  23. def hashCode(): Int

    Permalink
    Definition Classes
    AnyRef → Any

  24. val inputKeepProbability: Float

    Permalink

    Keep probability for the input of the RNN cell.

  25. final def isInstanceOf[T0]: Boolean

    Permalink
    Definition Classes
    Any

  26. val layerType: String

    Permalink
    Definition Classes
    DropoutWrapper → Layer

  27. def map[MR](mapFn: (Tuple[O, S]) ⇒ MR): Layer[Tuple[O, S], MR]

    Permalink
    Definition Classes
    Layer

  28. val name: String

    Permalink

    Name scope (also acting as variable scope) for this layer.

    Name scope (also acting as variable scope) for this layer.

    Definition Classes
    DropoutWrapper → RNNCell → Layer

  29. final def ne(arg0: AnyRef): Boolean

    Permalink
    Definition Classes
    AnyRef

  30. final def notify(): Unit

    Permalink
    Definition Classes
    AnyRef

  31. final def notifyAll(): Unit

    Permalink
    Definition Classes
    AnyRef

  32. val outputKeepProbability: Float

    Permalink

    Keep probability for the output of the RNN cell.

  33. val seed: Option[Int]

    Permalink

    Optional random seed, used to generate a random seed pair for the random number generator, when combined with the graph-level seed.

  34. val stateKeepProbability: Float

    Permalink

    Keep probability for the output state of the RNN cell.

  35. final def synchronized[T0](arg0: ⇒ T0): T0

    Permalink
    Definition Classes
    AnyRef

  36. def toString(): String

    Permalink
    Definition Classes
    Layer → AnyRef → Any

  37. final def wait(): Unit

    Permalink
    Definition Classes
    AnyRef
    Annotations
    @throws( ... )

  38. final def wait(arg0: Long, arg1: Int): Unit

    Permalink
    Definition Classes
    AnyRef
    Annotations
    @throws( ... )

  39. final def wait(arg0: Long): Unit

    Permalink
    Definition Classes
    AnyRef
    Annotations
    @throws( ... )

Inherited from RNNCell[O, OS, S, SS]

Inherited from Layer[Tuple[O, S], Tuple[O, S]]

Inherited from AnyRef

Inherited from Any

Ungrouped