Github user dbtsai commented on a diff in the pull request:
https://github.com/apache/spark/pull/353#discussion_r11404515
--- Diff:
mllib/src/main/scala/org/apache/spark/mllib/optimization/LBFGS.scala ---
@@ -0,0 +1,251 @@
+/*
+ * Licensed to the Apache Software Foundation (ASF) under one or more
+ * contributor license agreements. See the NOTICE file distributed with
+ * this work for additional information regarding copyright ownership.
+ * The ASF licenses this file to You under the Apache License, Version 2.0
+ * (the "License"); you may not use this file except in compliance with
+ * the License. You may obtain a copy of the License at
+ *
+ * http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package org.apache.spark.mllib.optimization
+
+import scala.Array
+import scala.collection.mutable.ArrayBuffer
+
+import breeze.linalg.{DenseVector => BDV}
+import breeze.optimize.{CachedDiffFunction, DiffFunction}
+
+import org.apache.spark.Logging
+import org.apache.spark.rdd.RDD
+import org.apache.spark.mllib.linalg.{Vectors, Vector}
+
+/**
+ * Class used to solve an optimization problem using Limited-memory BFGS.
+ * @param gradient Gradient function to be used.
+ * @param updater Updater to be used to update weights after every
iteration.
+ */
+class LBFGS(var gradient: Gradient, var updater: Updater)
+ extends Optimizer with Logging
+{
+ private var numCorrections: Int = 10
+ private var lineSearchTolerance: Double = 0.9
+ private var convTolerance: Double = 1E-4
+ private var maxNumIterations: Int = 100
+ private var regParam: Double = 0.0
+ private var miniBatchFraction: Double = 1.0
+
+ /**
+ * Set the number of corrections used in the LBFGS update. Default 10.
+ * Values of m less than 3 are not recommended; large values of m
+ * will result in excessive computing time. 3 < m < 10 is recommended.
+ * Restriction: m > 0
+ */
+ def setNumCorrections(corrections: Int): this.type = {
+ assert(corrections > 0)
+ this.numCorrections = corrections
+ this
+ }
+
+ /**
+ * Set the tolerance to control the accuracy of the line search in
mcsrch step. Default 0.9.
+ * If the function and gradient evaluations are inexpensive with respect
to the cost of
+ * the iteration (which is sometimes the case when solving very large
problems) it may
+ * be advantageous to set to a small value. A typical small value is 0.1.
+ * Restriction: should be greater than 1e-4.
+ */
+ def setLineSearchTolerance(tolerance: Double): this.type = {
+ this.lineSearchTolerance = tolerance
+ this
+ }
+
+ /**
+ * Set fraction of data to be used for each L-BFGS iteration. Default
1.0.
+ */
+ def setMiniBatchFraction(fraction: Double): this.type = {
+ this.miniBatchFraction = fraction
+ this
+ }
+
+ /**
+ * Set the convergence tolerance of iterations for L-BFGS. Default 1E-4.
+ * Smaller value will lead to higher accuracy with the cost of more
iterations.
+ */
+ def setConvTolerance(tolerance: Int): this.type = {
+ this.convTolerance = tolerance
+ this
+ }
+
+ /**
+ * Set the maximal number of iterations for L-BFGS. Default 100.
+ */
+ def setMaxNumIterations(iters: Int): this.type = {
+ this.maxNumIterations = iters
+ this
+ }
+
+ /**
+ * Set the regularization parameter. Default 0.0.
+ */
+ def setRegParam(regParam: Double): this.type = {
+ this.regParam = regParam
+ this
+ }
+
+ /**
+ * Set the gradient function (of the loss function of one single data
example)
+ * to be used for L-BFGS.
+ */
+ def setGradient(gradient: Gradient): this.type = {
+ this.gradient = gradient
+ this
+ }
+
+ /**
+ * Set the updater function to actually perform a gradient step in a
given direction.
+ * The updater is responsible to perform the update from the
regularization term as well,
+ * and therefore determines what kind or regularization is used, if any.
+ */
+ def setUpdater(updater: Updater): this.type = {
+ this.updater = updater
+ this
+ }
+
+ def optimize(data: RDD[(Double, Vector)], initialWeights: Vector):
Vector = {
+ val (weights, _) = LBFGS.runMiniBatchLBFGS(
+ data,
+ gradient,
+ updater,
+ numCorrections,
+ lineSearchTolerance,
+ convTolerance,
+ maxNumIterations,
+ regParam,
+ miniBatchFraction,
+ initialWeights)
+ weights
+ }
+
+}
+
+// Top-level method to run LBFGS.
+object LBFGS extends Logging {
+ /**
+ * Run Limited-memory BFGS (L-BFGS) in parallel using mini batches.
+ * In each iteration, we sample a subset (fraction miniBatchFraction) of
the total data
+ * in order to compute a gradient estimate.
+ * Sampling, and averaging the subgradients over this subset is
performed using one standard
+ * spark map-reduce in each iteration.
+ *
+ * @param data - Input data for L-BFGS. RDD of the set of data examples,
each of
+ * the form (label, [feature values]).
+ * @param gradient - Gradient object (used to compute the gradient of
the loss function of
+ * one single data example)
+ * @param updater - Updater function to actually perform a gradient step
in a given direction.
+ * @param numCorrections - The number of corrections used in the L-BFGS
update.
+ * @param lineSearchTolerance - The tolerance to control the accuracy of
the line search.
+ * @param convTolerance - The convergence tolerance of iterations for
L-BFGS
+ * @param maxNumIterations - Maximal number of iterations that L-BFGS
can be run.
+ * @param regParam - Regularization parameter
+ * @param miniBatchFraction - Fraction of the input data set that should
be used for
+ * one iteration of L-BFGS. Default value 1.0.
+ *
+ * @return A tuple containing two elements. The first element is a
column matrix containing
+ * weights for every feature, and the second element is an array
containing the loss
+ * computed for every iteration.
+ */
+ def runMiniBatchLBFGS(
+ data: RDD[(Double, Vector)],
+ gradient: Gradient,
+ updater: Updater,
+ numCorrections: Int,
+ lineSearchTolerance: Double,
+ convTolerance: Double,
+ maxNumIterations: Int,
+ regParam: Double,
+ miniBatchFraction: Double,
+ initialWeights: Vector): (Vector, Array[Double]) = {
+
+ val lossHistory = new ArrayBuffer[Double](maxNumIterations)
+
+ val nexamples: Long = data.count()
+ val miniBatchSize = nexamples * miniBatchFraction
+ var i = 0
+
+ val costFun = new DiffFunction[BDV[Double]] {
--- End diff --
For cost function, I intend to do it in this way because in the code of
cost function, I want to access and modify variables outside the cost function,
for example, "i", "lossHistory", and if I create a private class for this, it
will be extra effort to achieve this without changing breeze DiffFunction
signature.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
