Code from BasicOptimizer.scala:75 executed in 2459.03 seconds (21.331 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: 1075709410744500
Reset training subject: 1075729323780600
Adding measurement 157e0f6b to history. Total: 0
LBFGS Accumulation History: 1 points
Constructing line search parameters: GD+Trust
th(0)=2.2474588644801203;dx=-3.01751255237909E-7
New Minimum: 2.2474588644801203 > 2.247458521753413
WOLFE (weak): th(2.154434690031884)=2.247458521753413; dx=-2.991182352538564E-7 evalInputDelta=3.427267074584961E-7
New Minimum: 2.247458521753413 > 2.247457746893031
WOLFE (weak): th(4.308869380063768)=2.247457746893031; dx=-2.9911753103535445E-7 evalInputDelta=1.1175870895385742E-6
New Minimum: 2.247457746893031 > 2.2474549604062113
WOLFE (weak): th(12.926608140191302)=2.2474549604062113; dx=-2.9911655029858883E-7 evalInputDelta=3.90407390904457E-6
New Minimum: 2.2474549604062113 > 2.247442711762897
WOLFE (weak): th(51.70643256076521)=2.247442711762897; dx=-2.991132539074324E-7 evalInputDelta=1.6152717223327784E-5
New Minimum: 2.247442711762897 > 2.247378652216751
WOLFE (weak): th(258.53216280382605)=2.247378652216751; dx=-2.991015060979827E-7 evalInputDelta=8.021226336918374E-5
New Minimum: 2.247378652216751 > 2.2469759816351424
WOLFE (weak): th(1551.1929768229563)=2.2469759816351424; dx=-2.9905653281313733E-7 evalInputDelta=4.8288284497788325E-4
New Minimum: 2.2469759816351424 > 2.2440804130797583
WOLFE (weak): th(10858.350837760694)=2.2440804130797583; dx=-2.98597329905654E-7 evalInputDelta=0.0033784514003620814
New Minimum: 2.2440804130797583 > 2.2205927211278356
WOLFE (weak): th(86866.80670208555)=2.2205927211278356; dx=-2.949522126279808E-7 evalInputDelta=0.026866143352284766
New Minimum: 2.2205927211278356 > 2.0205797189481003
END: th(781801.26031877)=2.0205797189481003; dx=-2.582502406045801E-7 evalInputDelta=0.22687914553202004
Fitness changed from 2.2474588644801203 to 2.0205797189481003
Iteration 1 complete. Error: 2.0205797189481003 Total: 251.1734; Orientation: 0.0341; Line Search: 192.3532
<a id="p-3"></a>Iteration 1
<a id="p-2"></a>=-0.20382016544495074;dx=-5.345188956893634E-9
New Minimum: -0.20382016544495074 > -0.2142269512909445
END: th(7834471.893154294)=-0.2142269512909445; dx=-1.7864364084256995E-9 evalInputDelta=0.010406785845993749
Fitness changed from -0.20382016544495074 to -0.2142269512909445
Iteration 29 complete. Error: -0.2142269512909445 Total: 59.3060; Orientation: 1.0519; Line Search: 38.6706
Adding measurement 5ea24702 to history. Total: 3
Rejected: LBFGS Orientation magnitude: 1.645e+03, gradient 7.696e-05, dot -0.677; [57988dfa-89e4-4984-95a6-79f9f6508d70 = 1.000/1.000e+00]
Orientation rejected. Popping history element from -0.1822457727365947, -0.18733474163806818, -0.20382016544495074, -0.2142269512909445
LBFGS Accumulation History: 3 points
Removed measurement 56a8b19f to history. Total: 4
th(0)=-0.2142269512909445;dx=-5.9235169171414575E-9
Armijo: th(1.6878858024691377E7)=-0.20065140791937885; dx=3.0939492529576957E-10 evalInputDelta=-0.013575543371565635
New Minimum: -0.2142269512909445 > -0.22049521625506863
END: th(8439429.012345688)=-0.22049521625506863; dx=-8.570512892504599E-10 evalInputDelta=0.006268264964124137
Fitness changed from -0.2142269512909445 to -0.22049521625506863
Iteration 30 complete. Error: -0.22049521625506863 Total: 78.3191; Orientation: 1.0407; Line Search: 57.3682
<a id="p-21"></a>Iteration 30
<a id="p-20"></a>
Final threshold in iteration 31: -0.22049521625506863 (> -Infinity) after 2459.028s (< 3600.000s)
-0.22049521625506863