RNN

Instance Constructors

new RNN(name: String, cell: RNNCell[O, OS, S, SS], initialState: () ⇒ S = null, timeMajor: Boolean = false, parallelIterations: Int = 32, swapMemory: Boolean = false, sequenceLengths: tensors.Tensor[types.DataType] = null)(implicit evO: Aux[O, OS], evS: Aux[S, SS])

name
Name scope (also acting as variable scope) for this layer.
cell
RNN cell to use.
initialState
Initial state to use for the RNN, which is a structure over tensors with shapes [batchSize, stateShape(i)(0), stateShape(i)(1), ...], where i corresponds to the index of the corresponding state. Defaults to a zero state.
timeMajor
Boolean value indicating whether the inputs are provided in time-major format (i.e., have shape [time, batch, depth]) or in batch-major format (i.e., have shape [batch, time, depth]).
parallelIterations
Number of RNN loop iterations allowed to run in parallel.
swapMemory
If true, GPU-CPU memory swapping support is enabled for the RNN loop.
sequenceLengths
Optional INT32 tensor with shape [batchSize] containing the sequence lengths for each row in the batch.

Value Members

final def !=(arg0: Any): Boolean

Definition Classes
AnyRef → Any
final def ##(): Int

Definition Classes
AnyRef → Any
def +(other: Layer[O, Tuple[O, S]]): Concatenate[O, Tuple[O, S]]

Definition Classes
Layer
def ++(others: Seq[Layer[O, Tuple[O, S]]]): Concatenate[O, Tuple[O, S]]

Definition Classes
Layer
final def ==(arg0: Any): Boolean

Definition Classes
AnyRef → Any
def >>[S](other: Layer[Tuple[O, S], S]): Compose[O, Tuple[O, S], S]

Definition Classes
Layer
def apply(input: O)(implicit mode: Mode): Tuple[O, S]

Definition Classes
Layer
final def asInstanceOf[T0]: T0

Definition Classes
Any
val cell: RNNCell[O, OS, S, SS]

RNN cell to use.
def clone(): AnyRef

Attributes
protected[java.lang]
Definition Classes
AnyRef
Annotations
@throws( ... )
def compose[S](other: Layer[Tuple[O, S], S]): Compose[O, Tuple[O, S], S]

Definition Classes
Layer
def concatenate(others: Layer[O, Tuple[O, S]]*): Concatenate[O, Tuple[O, S]]

Definition Classes
Layer
final def currentStep: ops.Output

Definition Classes
Layer
final def eq(arg0: AnyRef): Boolean

Definition Classes
AnyRef
def equals(arg0: Any): Boolean

Definition Classes
AnyRef → Any
def finalize(): Unit

Attributes
protected[java.lang]
Definition Classes
AnyRef
Annotations
@throws( classOf[java.lang.Throwable] )
def forward(input: O)(implicit mode: Mode): Tuple[O, S]

Definition Classes
Layer
def forwardWithoutContext(input: O)(implicit mode: Mode): Tuple[O, S]

Definition Classes
RNN → Layer
final def getClass(): Class[_]

Definition Classes
AnyRef → Any
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

Definition Classes
Layer
def hashCode(): Int

Definition Classes
AnyRef → Any
val initialState: () ⇒ S

Initial state to use for the RNN, which is a structure over tensors with shapes [batchSize, stateShape(i)(0), stateShape(i)(1), ...], where i corresponds to the index of the corresponding state.
Initial state to use for the RNN, which is a structure over tensors with shapes [batchSize, stateShape(i)(0), stateShape(i)(1), ...], where i corresponds to the index of the corresponding state. Defaults to a zero state.
final def isInstanceOf[T0]: Boolean

Definition Classes
Any
val layerType: String

Definition Classes
RNN → Layer
def map[MR](mapFn: (Tuple[O, S]) ⇒ MR): Layer[O, MR]

Definition Classes
Layer
val name: String

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

Definition Classes
RNN → Layer
final def ne(arg0: AnyRef): Boolean

Definition Classes
AnyRef
final def notify(): Unit

Definition Classes
AnyRef
final def notifyAll(): Unit

Definition Classes
AnyRef
val parallelIterations: Int

Number of RNN loop iterations allowed to run in parallel.
val sequenceLengths: tensors.Tensor[types.DataType]

Optional INT32 tensor with shape [batchSize] containing the sequence lengths for each row in the batch.
val swapMemory: Boolean

If true, GPU-CPU memory swapping support is enabled for the RNN loop.
final def synchronized[T0](arg0: ⇒ T0): T0

Definition Classes
AnyRef
val timeMajor: Boolean

Boolean value indicating whether the inputs are provided in time-major format (i.e., have shape [time, batch, depth]) or in batch-major format (i.e., have shape [batch, time, depth]).
def toString(): String

Definition Classes
Layer → AnyRef → Any
final def wait(): Unit

Definition Classes
AnyRef
Annotations
@throws( ... )
final def wait(arg0: Long, arg1: Int): Unit

Definition Classes
AnyRef
Annotations
@throws( ... )
final def wait(arg0: Long): Unit

Definition Classes
AnyRef
Annotations
@throws( ... )

Related Docs: object RNN | package rnn

class RNN[O, OS, S, SS] extends Layer[O, Tuple[O, S]]

Instance Constructors

new RNN(name: String, cell: RNNCell[O, OS, S, SS], initialState: () ⇒ S = null, timeMajor: Boolean = false, parallelIterations: Int = 32, swapMemory: Boolean = false, sequenceLengths: tensors.Tensor[types.DataType] = null)(implicit evO: Aux[O, OS], evS: Aux[S, SS])

Value Members

final def !=(arg0: Any): Boolean

final def ##(): Int

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

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

final def ==(arg0: Any): Boolean

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

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

final def asInstanceOf[T0]: T0

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

def clone(): AnyRef

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

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

final def currentStep: ops.Output

final def eq(arg0: AnyRef): Boolean

def equals(arg0: Any): Boolean

def finalize(): Unit

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

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

final def getClass(): Class[_]

def hashCode(): Int

val initialState: () ⇒ S

final def isInstanceOf[T0]: Boolean

val layerType: String

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

val name: String

final def ne(arg0: AnyRef): Boolean

final def notify(): Unit

final def notifyAll(): Unit

val parallelIterations: Int

val sequenceLengths: tensors.Tensor[types.DataType]

val swapMemory: Boolean

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

val timeMajor: Boolean

def toString(): String

final def wait(): Unit

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

final def wait(arg0: Long): Unit

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

Inherited from AnyRef

Inherited from Any

Ungrouped