Code from BasicOptimizer.scala:75 executed in 1367.52 seconds (7.008 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: 1149145273390700
Reset training subject: 1149166591588700
Adding measurement 24d17d3e to history. Total: 0
LBFGS Accumulation History: 1 points
Constructing line search parameters: GD+Trust
th(0)=1.917255662381649;dx=-2.6970598823630907E-7
New Minimum: 1.917255662381649 > 1.9172548577189445
WOLFE (weak): th(2.154434690031884)=1.9172548577189445; dx=-2.6756023945329155E-7 evalInputDelta=8.046627044677734E-7
New Minimum: 1.9172548577189445 > 1.9172532558441162
WOLFE (weak): th(4.308869380063768)=1.9172532558441162; dx=-2.676205989438818E-7 evalInputDelta=2.4065375328063965E-6
New Minimum: 1.9172532558441162 > 1.9172475077211857
WOLFE (weak): th(12.926608140191302)=1.9172475077211857; dx=-2.675792521818957E-7 evalInputDelta=8.15466046333313E-6
New Minimum: 1.9172475077211857 > 1.9172229915857315
WOLFE (weak): th(51.70643256076521)=1.9172229915857315; dx=-2.676127764335358E-7 evalInputDelta=3.2670795917510986E-5
New Minimum: 1.9172229915857315 > 1.9170911237597466
WOLFE (weak): th(258.53216280382605)=1.9170911237597466; dx=-2.6757796595521043E-7 evalInputDelta=1.6453862190246582E-4
New Minimum: 1.9170911237597466 > 1.9162680432200432
WOLFE (weak): th(1551.1929768229563)=1.9162680432200432; dx=-2.674974995840877E-7 evalInputDelta=9.87619161605835E-4
New Minimum: 1.9162680432200432 > 1.9103482589125633
WOLFE (weak): th(10858.350837760694)=1.9103482589125633; dx=-2.6695500947363785E-7 evalInputDelta=0.006907403469085693
New Minimum: 1.9103482589125633 > 1.8624548092484474
WOLFE (weak): th(86866.80670208555)=1.8624548092484474; dx=-2.624584288380782E-7 evalInputDelta=0.0548008531332016
New Minimum: 1.8624548092484474 > 1.470915250480175
END: th(781801.26031877)=1.470915250480175; dx=-2.0553168741567723E-7 evalInputDelta=0.446340411901474
Fitness changed from 1.917255662381649 to 1.470915250480175
Iteration 1 complete. Error: 1.470915250480175 Total: 267.6630; Orientation: 0.0149; Line Search: 206.0169
<a id="p-3"></a>Iteration 1
<a id="p-2"></a>=-0.4699948951601982;dx=-7.153073119921534E-9
Armijo: th(1.462100482588026E7)=-0.4637882933020592; dx=3.79385365063198E-9 evalInputDelta=-0.006206601858139038
New Minimum: -0.4699948951601982 > -0.5061127170920372
END: th(7310502.41294013)=-0.5061127170920372; dx=-4.675563837332247E-10 evalInputDelta=0.03611782193183899
Fitness changed from -0.4699948951601982 to -0.5061127170920372
Iteration 14 complete. Error: -0.5061127170920372 Total: 84.4460; Orientation: 0.4050; Line Search: 63.1779
Adding measurement 79d2d302 to history. Total: 3
Rejected: LBFGS Orientation magnitude: 1.004e+03, gradient 7.057e-05, dot -0.707; [a8df715c-6590-4d6f-8773-c274c857c6ad = 1.000/1.000e+00]
Orientation rejected. Popping history element from -0.37710758298635483, -0.43283799663186073, -0.4699948951601982, -0.5061127170920372
LBFGS Accumulation History: 3 points
Removed measurement 31341562 to history. Total: 4
th(0)=-0.5061127170920372;dx=-4.968447349204996E-9
New Minimum: -0.5061127170920372 > -0.5192940309643745
WOLF (strong): th(1.5750000000000007E7)=-0.5192940309643745; dx=2.5848642263990302E-9 evalInputDelta=0.013181313872337341
New Minimum: -0.5192940309643745 > -0.5491219609975815
END: th(7875000.000000004)=-0.5491219609975815; dx=-4.5344393530587577E-10 evalInputDelta=0.04300924390554428
Fitness changed from -0.5061127170920372 to -0.5491219609975815
Iteration 15 complete. Error: -0.5491219609975815 Total: 82.5812; Orientation: 0.4026; Line Search: 61.8766
<a id="p-15"></a>Iteration 15
<a id="p-14"></a>
Final threshold in iteration 16: -0.5491219609975815 (> -Infinity) after 1367.523s (< 3600.000s)
-0.5491219609975815