Code from BasicOptimizer.scala:75 executed in 33.45 seconds (0.560 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: 905579280029500
Reset training subject: 905579792423300
Adding measurement 72015156 to history. Total: 0
LBFGS Accumulation History: 1 points
Constructing line search parameters: GD+Trust
th(0)=1.7948784530162811;dx=-2.6153322468395754E-7
New Minimum: 1.7948784530162811 > 1.7948780953884125
WOLFE (weak): th(2.154434690031884)=1.7948780953884125; dx=-2.6153323500100455E-7 evalInputDelta=3.5762786865234375E-7
New Minimum: 1.7948780953884125 > 1.794877365231514
WOLFE (weak): th(4.308869380063768)=1.794877365231514; dx=-2.6153327302690357E-7 evalInputDelta=1.087784767150879E-6
New Minimum: 1.794877365231514 > 1.7948750853538513
WOLFE (weak): th(12.926608140191302)=1.7948750853538513; dx=-2.615328730144612E-7 evalInputDelta=3.3676624298095703E-6
New Minimum: 1.7948750853538513 > 1.794864982366562
WOLFE (weak): th(51.70643256076521)=1.794864982366562; dx=-2.615329979993116E-7 evalInputDelta=1.3470649719238281E-5
New Minimum: 1.794864982366562 > 1.794810727238655
WOLFE (weak): th(258.53216280382605)=1.794810727238655; dx=-2.615028035574283E-7 evalInputDelta=6.77257776260376E-5
New Minimum: 1.794810727238655 > 1.7944728136062622
WOLFE (weak): th(1551.1929768229563)=1.7944728136062622; dx=-2.61514799049201E-7 evalInputDelta=4.056394100189209E-4
New Minimum: 1.7944728136062622 > 1.7920367121696472
WOLFE (weak): th(10858.350837760694)=1.7920367121696472; dx=-2.618935594331886E-7 evalInputDelta=0.002841740846633911
New Minimum: 1.7920367121696472 > 1.772078961133957
WOLFE (weak): th(86866.80670208555)=1.772078961133957; dx=-2.6211873997640013E-7 evalInputDelta=0.02279949188232422
New Minimum: 1.772078961133957 > 1.6033691465854645
END: th(781801.26031877)=1.6033691465854645; dx=-2.2066293190830404E-7 evalInputDelta=0.19150930643081665
Fitness changed from 1.7948784530162811 to 1.6033691465854645
Iteration 1 complete. Error: 1.6033691465854645 Total: 6.4899; Orientation: 0.0035; Line Search: 4.9718
<a id="p-3"></a>Iteration 1
<a id="p-2"></a>=-0.34228915721178055;dx=-4.122129110203499E-10
New Minimum: -0.34228915721178055 > -0.34502118825912476
END: th(9747336.550586838)=-0.34502118825912476; dx=-1.5169109444267596E-10 evalInputDelta=0.0027320310473442078
Fitness changed from -0.34228915721178055 to -0.34502118825912476
Iteration 14 complete. Error: -0.34502118825912476 Total: 1.6621; Orientation: 0.1048; Line Search: 1.0595
Adding measurement 611a91df to history. Total: 3
Rejected: LBFGS Orientation magnitude: 3.059e+03, gradient 1.301e-05, dot -0.954; [c182d263-e3f3-42a9-af16-884359de5fbe = 1.000/1.000e+00]
Orientation rejected. Popping history element from -0.32841235399246216, -0.33924517035484314, -0.34228915721178055, -0.34502118825912476
LBFGS Accumulation History: 3 points
Removed measurement 1db9c2cf to history. Total: 4
th(0)=-0.34502118825912476;dx=-1.6935293294962526E-10
New Minimum: -0.34502118825912476 > -0.34832555055618286
END: th(2.1000000000000007E7)=-0.34832555055618286; dx=-1.522001137839909E-10 evalInputDelta=0.0033043622970581055
Fitness changed from -0.34502118825912476 to -0.34832555055618286
Iteration 15 complete. Error: -0.34832555055618286 Total: 1.5714; Orientation: 0.1044; Line Search: 0.9689
<a id="p-15"></a>Iteration 15
<a id="p-14"></a>
Final threshold in iteration 16: -0.34832555055618286 (> -Infinity) after 33.450s (< 1800.000s)
-0.34832555055618286