Code from BasicOptimizer.scala:75 executed in 409.75 seconds (5.902 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: 3409108526014
Reset training subject: 3410618219819
Adding measurement 7aea7c1c to history. Total: 0
LBFGS Accumulation History: 1 points
Constructing line search parameters: GD+Trust
th(0)=0.8880820666949342;dx=-9.57505075727732E-8
New Minimum: 0.8880820666949342 > 0.8880818986468545
WOLFE (weak): th(2.154434690031884)=0.8880818986468545; dx=-9.575055731437762E-8 evalInputDelta=1.6804807967396584E-7
New Minimum: 0.8880818986468545 > 0.8880817305986909
WOLFE (weak): th(4.308869380063768)=0.8880817305986909; dx=-9.575089534978239E-8 evalInputDelta=3.3609624328079235E-7
New Minimum: 0.8880817305986909 > 0.8880810584077287
WOLFE (weak): th(12.926608140191302)=0.8880810584077287; dx=-9.575061167895022E-8 evalInputDelta=1.0082872055061642E-6
New Minimum: 0.8880810584077287 > 0.8880780335381336
WOLFE (weak): th(51.70643256076521)=0.8880780335381336; dx=-9.574806812646361E-8 evalInputDelta=4.033156800531401E-6
New Minimum: 0.8880780335381336 > 0.8880619019296023
WOLFE (weak): th(258.53216280382605)=0.8880619019296023; dx=-9.57591077768758E-8 evalInputDelta=2.0164765331842638E-5
New Minimum: 0.8880619019296023 > 0.8879610388063558
WOLFE (weak): th(1551.1929768229563)=0.8879610388063558; dx=-9.577331231035879E-8 evalInputDelta=1.2102788857837332E-4
New Minimum: 0.8879610388063558 > 0.8872324198413369
WOLFE (weak): th(10858.350837760694)=0.8872324198413369; dx=-9.602670436705727E-8 evalInputDelta=8.496468535972301E-4
New Minimum: 0.8872324198413369 > 0.8811943595133328
WOLFE (weak): th(86866.80670208555)=0.8811943595133328; dx=-9.655267259016018E-8 evalInputDelta=0.006887707181601366
New Minimum: 0.8811943595133328 > 0.8340277984479741
END: th(781801.26031877)=0.8340277984479741; dx=-6.276520426763219E-8 evalInputDelta=0.054054268246960024
Fitness changed from 0.8880820666949342 to 0.8340277984479741
Iteration 1 complete. Error: 0.8340277984479741 Total: 20.8042; Orientation: 0.0716; Line Search: 16.2147
<a id="p-3"></a>Iteration 1
<a id="p-2"></a>=0.24308586654291225;dx=-6.648485452324587E-10
New Minimum: 0.24308586654291225 > 0.24169180712626934
END: th(1.3700658447460849E7)=0.24169180712626934; dx=-3.4144242140729095E-10 evalInputDelta=0.0013940594166429177
Fitness changed from 0.24308586654291225 to 0.24169180712626934
Iteration 49 complete. Error: 0.24169180712626934 Total: 6.4333; Orientation: 1.6312; Line Search: 3.3058
Adding measurement 1ffe6348 to history. Total: 3
Rejected: LBFGS Orientation magnitude: 1.440e+03, gradient 2.747e-05, dot -0.726; [85094c29-dc3c-4daa-89da-f3d203c961d2 = 1.000/1.000e+00]
Orientation rejected. Popping history element from 0.24629877658963087, 0.24551684340867266, 0.24308586654291225, 0.24169180712626934
LBFGS Accumulation History: 3 points
Removed measurement 165c5129 to history. Total: 4
th(0)=0.24169180712626934;dx=-7.544803762645406E-10
Armijo: th(2.9517173835488025E7)=0.24546570963814926; dx=1.9860345972085248E-11 evalInputDelta=-0.003773902511879923
New Minimum: 0.24169180712626934 > 0.24078324153589056
END: th(1.4758586917744013E7)=0.24078324153589056; dx=-1.9191345387791618E-10 evalInputDelta=9.085655903787737E-4
Fitness changed from 0.24169180712626934 to 0.24078324153589056
Iteration 50 complete. Error: 0.24078324153589056 Total: 8.1162; Orientation: 1.6335; Line Search: 4.9888
<a id="p-29"></a>Iteration 50
<a id="p-28"></a>
Final threshold in iteration 51: 0.24078324153589056 (> -Infinity) after 409.750s (< 3600.000s)
0.24078324153589056