Instance Constructors
-
new
LBFGS(numIterations: Double = 1000, maxIterations: Int = 1000, tolerance: Double = 1.0E-4, gradientTolerance: Double = 0.001, eps: Double = 1.0E-5, rankOfApproximation: Int = 4, initialStepSize: Double = 1.0)
Value Members
-
final
def
!=(arg0: AnyRef): Boolean
-
final
def
!=(arg0: Any): Boolean
-
final
def
##(): Int
-
final
def
==(arg0: AnyRef): Boolean
-
final
def
==(arg0: Any): Boolean
-
var
alpha: Array[Double]
-
final
def
asInstanceOf[T0]: T0
-
def
clone(): AnyRef
-
-
val
eps: Double
-
final
def
eq(arg0: AnyRef): Boolean
-
def
equals(arg0: Any): Boolean
-
def
finalize(): Unit
-
def
finalizeWeights(weights: WeightsSet): Unit
-
-
final
def
getClass(): Class[_]
-
var
gradientTolerance: Double
-
def
hashCode(): Int
-
val
initialStepSize: Double
-
def
initializeWeights(weights: WeightsSet): Unit
-
def
isConverged: Boolean
-
final
def
isInstanceOf[T0]: Boolean
-
var
iterations: Int
-
-
-
var
maxIterations: Int
-
final
def
ne(arg0: AnyRef): Boolean
-
final
def
notify(): Unit
-
final
def
notifyAll(): Unit
-
var
numIterations: Double
-
-
var
oldValue: Double
-
-
-
def
postIteration(iter: Int): Unit
-
def
pushDbl(l: ArrayBuffer[Double], toadd: Double): Unit
-
def
pushTensor(l: ArrayBuffer[WeightsMap], toadd: WeightsMap): Unit
-
val
rankOfApproximation: Int
-
def
reset(): Unit
-
var
rho: ArrayBuffer[Double]
-
-
def
step(weights: WeightsSet, gradient: WeightsMap, value: Double): Unit
-
var
step: Double
-
final
def
synchronized[T0](arg0: ⇒ T0): T0
-
def
toString(): String
-
var
tolerance: Double
-
final
def
wait(): Unit
-
final
def
wait(arg0: Long, arg1: Int): Unit
-
final
def
wait(arg0: Long): Unit
-
Maximize using Limited-memory BFGS, as described in Byrd, Nocedal, and Schnabel, "Representations of Quasi-Newton Matrices and Their Use in Limited Memory Methods"