Ledger

Abstract Value Members

abstract def apply(in: Array[Int]): OutInfo

Apply this layer (forward computation)
Apply this layer (forward computation)
in
Input value
returns
Output computation information.

Definition Classes
Layer
abstract def backprop(seq: ParSeq[((Array[Int], OutInfo), DataVec)]): (ParSeq[DataVec], ParSeq[() ⇒ Unit])

Backward computation
Backward computation
seq
Sequence of entries to be used for backward computation.
returns
Error sequence, to backpropagate into previous layer.

Definition Classes
Layer
Note
For the computation, we only used denominator layout. (cf. Wikipedia Page of Matrix Computation) For the computation rules, see "Matrix Cookbook" from MIT.
abstract val outVecOf: (OutInfo) ⇒ DataVec

Output converter from OutInfo to Vector *
Output converter from OutInfo to Vector *

Definition Classes
Layer

Concrete Value Members

final def !=(arg0: Any): Boolean

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

Definition Classes
AnyRef → Any
final def ==(arg0: Any): Boolean

Definition Classes
AnyRef → Any
var NIn: Int

Size of input *
Size of input *

Definition Classes
Layer
var NOut: Int

Size of output *
Size of output *

Definition Classes
Layer
var algorithm: LedgerAlgorithm
def apply(in: RDD[(Long, Array[Int])]): RDD[(Long, DataVec)]

Apply RDD of vector.
Apply RDD of vector.
in
RDD of (ID, Value), ID is Long value.
returns
RDD of (ID, Vector), with same ID for input.

Definition Classes
Layer
final def asInstanceOf[T0]: T0

Definition Classes
Any
def backward(error: ParSeq[DataVec]): (ParSeq[DataVec], ParSeq[() ⇒ Unit])

Backward computation using propagated error.
Backward computation using propagated error.
error
Propagated error sequence.
returns
Error sequence for back propagation.

Definition Classes
Layer
var bcModel: Broadcast[LedgerModel]
def broadcast(sc: SparkContext): Unit

Broadcast resources of this layer.
Broadcast resources of this layer.
sc
Spark Context

Definition Classes
Ledger → InputLayer
var builder: LedgerBuilder
def clone(): AnyRef

Attributes
protected[java.lang]
Definition Classes
AnyRef
Annotations
@throws( ... )
var dimension: Int
final def eq(arg0: AnyRef): Boolean

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

Definition Classes
AnyRef → Any
implicit val evidenceI: ClassTag[Array[Int]]

ClassTag for input *
ClassTag for input *

Attributes
protected
Definition Classes
Ledger → Layer
def finalize(): Unit

Attributes
protected[java.lang]
Definition Classes
AnyRef
Annotations
@throws( classOf[java.lang.Throwable] )
def forward(in: ParSeq[Array[Int]]): ParSeq[DataVec]

Apply Parallel Sequence of vector.
Apply Parallel Sequence of vector.
in
Parallel Sequence of Input
returns
Parallel Sequence of Vector

Definition Classes
Layer
final def forward(in: RDD[(Long, Array[Int])]): RDD[(Long, DataVec)]

Apply RDD of vector.
Apply RDD of vector.
in
RDD of (ID, Value), ID is Long value.
returns
RDD of (ID, Vector), with same ID for input.

Definition Classes
Layer
final def forward(in: Array[Int]): DataVec

Apply this layer (forward computation)
Apply this layer (forward computation)
in
Input value
returns
Output Vector

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

Definition Classes
AnyRef → Any
def hashCode(): Int

Definition Classes
AnyRef → Any
def initiateBy(builder: WeightBuilder): Ledger.this.type

Set weight builder for this layer.
Set weight builder for this layer.
builder
Weight builder to be applied
returns
self

Definition Classes
Layer
var inoutSEQ: ParSeq[(Array[Int], OutInfo)]

Sequence for backpropagation.
Sequence for backpropagation. Stores output values. *

Attributes
protected
Definition Classes
Layer
final def isInstanceOf[T0]: Boolean

Definition Classes
Any
var isUpdatable: Boolean

True if this layer affected by backward propagation *
True if this layer affected by backward propagation *

Attributes
protected
Definition Classes
Layer
def loss: Double

Weight Loss of this layer
Weight Loss of this layer
returns
Weight loss

Definition Classes
Ledger → Layer
var model: LedgerModel
final def ne(arg0: AnyRef): Boolean

Definition Classes
AnyRef
final def notify(): Unit

Definition Classes
AnyRef
final def notifyAll(): Unit

Definition Classes
AnyRef
def pad: DataVec

Attributes
protected
var padID: Int

Attributes
protected
def read(kryo: Kryo, input: Input): Unit

Definition Classes
Ledger → Layer → KryoSerializable
def setUpdatable(bool: Boolean): Unit

Assign whether this layer updatable or not.
Assign whether this layer updatable or not. value.
bool
True if this layer used in backpropagation.

Definition Classes
Layer
final def synchronized[T0](arg0: ⇒ T0): T0

Definition Classes
AnyRef
def toString(): String

Definition Classes
AnyRef → Any
def unbroadcast(): Unit

Unpersist broadcast of this layer.
Unpersist broadcast of this layer.

Definition Classes
Ledger → InputLayer
def updateWord(word: Int, dx: DataVec): Unit

Attributes
protected
def vectorOf(str: Int): DataVec

Attributes
protected
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( ... )
def withModel(model: LedgerModel, builder: LedgerBuilder): Ledger.this.type
def write(kryo: Kryo, output: Output): Unit

Definition Classes
Ledger → Layer → KryoSerializable

Deprecated Value Members

def withInput(in: Int): Ledger.this.type

Set input size
Set input size
in
Size of input
returns
self

Definition Classes
Ledger → Layer
Annotations
@deprecated
Deprecated
def withOutput(out: Int): Ledger.this.type

Set output size
Set output size
out
Size of output
returns
self

Definition Classes
Ledger → Layer
Annotations
@deprecated
Deprecated

Related Doc: package layer

trait Ledger[OutInfo] extends InputLayer[Array[Int], OutInfo]

Abstract Value Members

abstract def apply(in: Array[Int]): OutInfo

abstract def backprop(seq: ParSeq[((Array[Int], OutInfo), DataVec)]): (ParSeq[DataVec], ParSeq[() ⇒ Unit])

abstract val outVecOf: (OutInfo) ⇒ DataVec

Concrete Value Members

final def !=(arg0: Any): Boolean

final def ##(): Int

final def ==(arg0: Any): Boolean

var NIn: Int

var NOut: Int

var algorithm: LedgerAlgorithm

def apply(in: RDD[(Long, Array[Int])]): RDD[(Long, DataVec)]

final def asInstanceOf[T0]: T0

def backward(error: ParSeq[DataVec]): (ParSeq[DataVec], ParSeq[() ⇒ Unit])

var bcModel: Broadcast[LedgerModel]

def broadcast(sc: SparkContext): Unit

var builder: LedgerBuilder

def clone(): AnyRef

var dimension: Int

final def eq(arg0: AnyRef): Boolean

def equals(arg0: Any): Boolean

implicit val evidenceI: ClassTag[Array[Int]]

def finalize(): Unit

def forward(in: ParSeq[Array[Int]]): ParSeq[DataVec]

final def forward(in: RDD[(Long, Array[Int])]): RDD[(Long, DataVec)]

final def forward(in: Array[Int]): DataVec

final def getClass(): Class[_]

def hashCode(): Int

def initiateBy(builder: WeightBuilder): Ledger.this.type

var inoutSEQ: ParSeq[(Array[Int], OutInfo)]

final def isInstanceOf[T0]: Boolean

var isUpdatable: Boolean

def loss: Double

var model: LedgerModel

final def ne(arg0: AnyRef): Boolean

final def notify(): Unit

final def notifyAll(): Unit

def pad: DataVec

var padID: Int

def read(kryo: Kryo, input: Input): Unit

def setUpdatable(bool: Boolean): Unit

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

def toString(): String

def unbroadcast(): Unit

def updateWord(word: Int, dx: DataVec): Unit

def vectorOf(str: Int): DataVec

final def wait(): Unit

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

final def wait(arg0: Long): Unit

def withModel(model: LedgerModel, builder: LedgerBuilder): Ledger.this.type

def write(kryo: Kryo, output: Output): Unit

Deprecated Value Members

def withInput(in: Int): Ledger.this.type

def withOutput(out: Int): Ledger.this.type

Inherited from InputLayer[Array[Int], OutInfo]

Inherited from Layer[Array[Int], OutInfo]

Inherited from KryoSerializable

Inherited from Serializable

Inherited from Serializable

Inherited from AnyRef

Inherited from Any

Ungrouped