Code from BasicOptimizer.scala:75 executed in 132.29 seconds (1.552 gc):
val lineSearchInstance: LineSearchStrategy = lineSearchFactory
IterativeTrainer.wrap(trainable)
.setOrientation(orientation())
.setMonitor(new TrainingMonitor() {
override def clear(): Unit = trainingMonitor.clear()
override def log(msg: String): Unit = trainingMonitor.log(msg)
override def onStepFail(currentPoint: Step): Boolean = {
BasicOptimizer.this.onStepFail(trainable, currentPoint)
}
override def onStepComplete(currentPoint: Step): Unit = {
if (0 < logEvery && (0 == currentPoint.iteration % logEvery || currentPoint.iteration < logEvery)) {
val image = currentImage
timelineAnimation += image
val caption = "Iteration " + currentPoint.iteration
out.p(caption + "\n" + out.jpg(image, caption))
}
BasicOptimizer.this.onStepComplete(trainable, currentPoint)
trainingMonitor.onStepComplete(currentPoint)
super.onStepComplete(currentPoint)
}
})
.setTimeout(trainingMinutes, TimeUnit.MINUTES)
.setMaxIterations(trainingIterations)
.setLineSearchFactory((_: CharSequence) => lineSearchInstance)
.setTerminateThreshold(java.lang.Double.NEGATIVE_INFINITY)
.runAndFree
.asInstanceOf[lang.Double]
Reset training subject: 899764863633400
Reset training subject: 899766055309500
Adding measurement 6870cfac to history. Total: 0
LBFGS Accumulation History: 1 points
Constructing line search parameters: GD+Trust
th(0)=0.8700454384088516;dx=-6.019906636806742E-8
New Minimum: 0.8700454384088516 > 0.8700453191995621
WOLFE (weak): th(2.154434690031884)=0.8700453191995621; dx=-6.019903762733195E-8 evalInputDelta=1.1920928955078125E-7
New Minimum: 0.8700453191995621 > 0.8700453042984009
WOLFE (weak): th(4.308869380063768)=0.8700453042984009; dx=-6.019901499366456E-8 evalInputDelta=1.341104507446289E-7
New Minimum: 0.8700453042984009 > 0.8700447976589203
WOLFE (weak): th(12.926608140191302)=0.8700447976589203; dx=-6.019889261889508E-8 evalInputDelta=6.407499313354492E-7
New Minimum: 0.8700447976589203 > 0.8700428903102875
WOLFE (weak): th(51.70643256076521)=0.8700428903102875; dx=-6.019842803508405E-8 evalInputDelta=2.5480985641479492E-6
New Minimum: 0.8700428903102875 > 0.8700327575206757
WOLFE (weak): th(258.53216280382605)=0.8700327575206757; dx=-6.019658789819888E-8 evalInputDelta=1.2680888175964355E-5
New Minimum: 0.8700327575206757 > 0.8699692487716675
WOLFE (weak): th(1551.1929768229563)=0.8699692487716675; dx=-6.01802951204271E-8 evalInputDelta=7.618963718414307E-5
New Minimum: 0.8699692487716675 > 0.8695130050182343
WOLFE (weak): th(10858.350837760694)=0.8695130050182343; dx=-6.007579528052929E-8 evalInputDelta=5.324333906173706E-4
New Minimum: 0.8695130050182343 > 0.8658049702644348
WOLFE (weak): th(86866.80670208555)=0.8658049702644348; dx=-5.945462087777701E-8 evalInputDelta=0.004240468144416809
New Minimum: 0.8658049702644348 > 0.8335308283567429
END: th(781801.26031877)=0.8335308283567429; dx=-5.364366654536032E-8 evalInputDelta=0.036514610052108765
Fitness changed from 0.8700454384088516 to 0.8335308283567429
Iteration 1 complete. Error: 0.8335308283567429 Total: 15.7025; Orientation: 0.0151; Line Search: 12.1337
<a id="p-3"></a>Iteration 1
<a id="p-2"></a>=-0.4580078348517418;dx=-1.1664070160477168E-10
New Minimum: -0.4580078348517418 > -0.4598141759634018
END: th(3.645833333333337E7)=-0.4598141759634018; dx=-7.133251578245851E-11 evalInputDelta=0.0018063411116600037
Fitness changed from -0.4580078348517418 to -0.4598141759634018
Iteration 24 complete. Error: -0.4598141759634018 Total: 3.8559; Orientation: 0.4230; Line Search: 2.3157
Adding measurement 23f60b7d to history. Total: 3
Rejected: LBFGS Orientation magnitude: 2.825e+03, gradient 1.019e-05, dot -0.849; [ee815948-f439-4925-941d-6d378ec1cdcc = 1.000/1.000e+00]
Orientation rejected. Popping history element from -0.45426400750875473, -0.4563176780939102, -0.4580078348517418, -0.4598141759634018
LBFGS Accumulation History: 3 points
Removed measurement 570b2057 to history. Total: 4
th(0)=-0.4598141759634018;dx=-1.038107277361544E-10
New Minimum: -0.4598141759634018 > -0.4636022299528122
END: th(7.854709807407919E7)=-0.4636022299528122; dx=-4.8349557453206916E-11 evalInputDelta=0.0037880539894104004
Fitness changed from -0.4598141759634018 to -0.4636022299528122
Iteration 25 complete. Error: -0.4636022299528122 Total: 3.9351; Orientation: 0.4375; Line Search: 2.3263
<a id="p-19"></a>Iteration 25
<a id="p-18"></a>
Final threshold in iteration 26: -0.4636022299528122 (> -Infinity) after 132.293s (< 3600.000s)
-0.4636022299528122