TrainingParam

Criteria: When to stop training

This case class defines when to stop training. Training stops if one of the following condition is satisfied.

- #Iteration ≥ maxIter

#Iteration ≥ current patience value, which is calculated by max(patience, bestIteration * patienceStep)
Amount of loss < lossThreshold

Validation is done for each validationFreq iterations, and whenever current/best loss ratio below improveThreshold, that iteration is marked as best iteration.

maxIter: maximum iteration count (default 100,000)
waitAfterUpdate: multiplier for calculating patience (default 1 := Wait lastupdate# * 1 after update)
improveThreshold: threshold that iteration is marked as "improved" (default 99.5% = 0.995)
lossThreshold: maximum-tolerant loss value. (default 0.0001)
validationFreq: multiplier used for count for validation. (default 1.0) Validation checked whenever (validationFreq) * (#epoch for 1 training batch). where #epoch for 1 iteration = round(1 / miniBatchFraction).
miniBatch: size of mini-batch. (default 10)
reuseSaveData: True if this trainer reuse previous temp data in disk. (default false)
dataOnLocal: True if want to collect all the data into the driver. (default false)
storageLevel: Persist level (default MEMORY_ONLY)

Linear Supertypes

Product, Equals, Serializable, Serializable, AnyRef, Any

Instance Constructors

new TrainingParam(maxIter: Int = 100, waitAfterUpdate: Int = 1, improveThreshold: Double = 0.999999, lossThreshold: Double = 0.0001, validationFreq: Double = 1.0, miniBatch: Int = 10, reuseSaveData: Boolean = false, dataOnLocal: Boolean = false, storageLevel: StorageLevel = StorageLevel.MEMORY_ONLY)

maxIter
maximum iteration count (default 100,000)
waitAfterUpdate
multiplier for calculating patience (default 1 := Wait lastupdate# * 1 after update)
improveThreshold
threshold that iteration is marked as "improved" (default 99.5% = 0.995)
lossThreshold
maximum-tolerant loss value. (default 0.0001)
validationFreq
multiplier used for count for validation. (default 1.0) Validation checked whenever (validationFreq) * (#epoch for 1 training batch). where #epoch for 1 iteration = round(1 / miniBatchFraction).
miniBatch
size of mini-batch. (default 10)
reuseSaveData
True if this trainer reuse previous temp data in disk. (default false)
dataOnLocal
True if want to collect all the data into the driver. (default false)
storageLevel
Persist level (default MEMORY_ONLY)

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
final def asInstanceOf[T0]: T0

Definition Classes
Any
def clone(): AnyRef

Attributes
protected[java.lang]
Definition Classes
AnyRef
Annotations
@throws( ... )
val dataOnLocal: Boolean

True if want to collect all the data into the driver.
True if want to collect all the data into the driver. (default false)
final def eq(arg0: AnyRef): Boolean

Definition Classes
AnyRef
def finalize(): Unit

Attributes
protected[java.lang]
Definition Classes
AnyRef
Annotations
@throws( classOf[java.lang.Throwable] )
final def getClass(): Class[_]

Definition Classes
AnyRef → Any
val improveThreshold: Double

threshold that iteration is marked as "improved" (default 99.5% = 0.995)
final def isInstanceOf[T0]: Boolean

Definition Classes
Any
val lossThreshold: Double

maximum-tolerant loss value.
maximum-tolerant loss value. (default 0.0001)
val maxIter: Int

maximum iteration count (default 100,000)
val miniBatch: Int

size of mini-batch.
size of mini-batch. (default 10)
final def ne(arg0: AnyRef): Boolean

Definition Classes
AnyRef
final def notify(): Unit

Definition Classes
AnyRef
final def notifyAll(): Unit

Definition Classes
AnyRef
val reuseSaveData: Boolean

True if this trainer reuse previous temp data in disk.
True if this trainer reuse previous temp data in disk. (default false)
val storageLevel: StorageLevel

Persist level (default MEMORY_ONLY)
final def synchronized[T0](arg0: ⇒ T0): T0

Definition Classes
AnyRef
val validationFreq: Double

multiplier used for count for validation.
multiplier used for count for validation. (default 1.0) Validation checked whenever (validationFreq) * (#epoch for 1 training batch). where #epoch for 1 iteration = round(1 / miniBatchFraction).
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( ... )
val waitAfterUpdate: Int

multiplier for calculating patience (default 1 := Wait lastupdate# * 1 after update)

Related Doc: package train

Instance Constructors

new TrainingParam(maxIter: Int = 100, waitAfterUpdate: Int = 1, improveThreshold: Double = 0.999999, lossThreshold: Double = 0.0001, validationFreq: Double = 1.0, miniBatch: Int = 10, reuseSaveData: Boolean = false, dataOnLocal: Boolean = false, storageLevel: StorageLevel = StorageLevel.MEMORY_ONLY)

Value Members

final def !=(arg0: Any): Boolean

final def ##(): Int

final def ==(arg0: Any): Boolean

final def asInstanceOf[T0]: T0

def clone(): AnyRef

val dataOnLocal: Boolean

final def eq(arg0: AnyRef): Boolean

def finalize(): Unit

final def getClass(): Class[_]

val improveThreshold: Double

final def isInstanceOf[T0]: Boolean

val lossThreshold: Double

val maxIter: Int

val miniBatch: Int

final def ne(arg0: AnyRef): Boolean

final def notify(): Unit

final def notifyAll(): Unit

val reuseSaveData: Boolean

val storageLevel: StorageLevel

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

val validationFreq: Double

final def wait(): Unit

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

final def wait(arg0: Long): Unit

val waitAfterUpdate: Int

Inherited from Product

Inherited from Equals

Inherited from Serializable

Inherited from Serializable

Inherited from AnyRef

Inherited from Any

Ungrouped