Code from BasicOptimizer.scala:75 executed in 406.81 seconds (6.202 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: 2994381160063
Reset training subject: 2995811909984
Adding measurement 5c025122 to history. Total: 0
LBFGS Accumulation History: 1 points
Constructing line search parameters: GD+Trust
th(0)=0.9366366700557319;dx=-9.012874634113232E-8
New Minimum: 0.9366366700557319 > 0.9366365310438761
WOLFE (weak): th(2.154434690031884)=0.9366365310438761; dx=-9.012638737808244E-8 evalInputDelta=1.3901185580422748E-7
New Minimum: 0.9366365310438761 > 0.9366363920347074
WOLFE (weak): th(4.308869380063768)=0.9366363920347074; dx=-9.012660942376756E-8 evalInputDelta=2.7802102453566846E-7
New Minimum: 0.9366363920347074 > 0.9366358359984814
WOLFE (weak): th(12.926608140191302)=0.9366358359984814; dx=-9.012675520215603E-8 evalInputDelta=8.340572504872412E-7
New Minimum: 0.9366358359984814 > 0.9366333338510475
WOLFE (weak): th(51.70643256076521)=0.9366333338510475; dx=-9.012716245368812E-8 evalInputDelta=3.3362046844009896E-6
New Minimum: 0.9366333338510475 > 0.9366199842922445
WOLFE (weak): th(258.53216280382605)=0.9366199842922445; dx=-9.015834887826646E-8 evalInputDelta=1.6685763487456207E-5
New Minimum: 0.9366199842922445 > 0.936536425217887
WOLFE (weak): th(1551.1929768229563)=0.936536425217887; dx=-9.030391566407131E-8 evalInputDelta=1.0024483784487614E-4
New Minimum: 0.936536425217887 > 0.9359296667759847
WOLFE (weak): th(10858.350837760694)=0.9359296667759847; dx=-9.102017699812285E-8 evalInputDelta=7.070032797472647E-4
New Minimum: 0.9359296667759847 > 0.9307047794094278
WOLFE (weak): th(86866.80670208555)=0.9307047794094278; dx=-9.584365939925764E-8 evalInputDelta=0.005931890646304172
New Minimum: 0.9307047794094278 > 0.8803466167689128
END: th(781801.26031877)=0.8803466167689128; dx=-7.588710995972393E-8 evalInputDelta=0.05629005328681913
Fitness changed from 0.9366366700557319 to 0.8803466167689128
Iteration 1 complete. Error: 0.8803466167689128 Total: 20.5660; Orientation: 0.0636; Line Search: 16.1249
<a id="p-3"></a>Iteration 1
<a id="p-2"></a>=0.16152863785320556;dx=-4.902964567004407E-10
Armijo: th(2.4356726128819294E7)=0.16458478053777836; dx=1.5719517941809875E-10 evalInputDelta=-0.003056142684572799
New Minimum: 0.16152863785320556 > 0.16102926868856743
END: th(1.2178363064409647E7)=0.16102926868856743; dx=-2.5063399158123887E-11 evalInputDelta=4.99369164638136E-4
Fitness changed from 0.16152863785320556 to 0.16102926868856743
Iteration 49 complete. Error: 0.16102926868856743 Total: 8.0810; Orientation: 1.6481; Line Search: 4.9421
Adding measurement 40f65bf4 to history. Total: 3
Rejected: LBFGS Orientation magnitude: 6.616e+02, gradient 2.588e-05, dot -0.559; [65b8d454-07aa-484c-ad3a-f8dbb5803c80 = 1.000/1.000e+00]
Orientation rejected. Popping history element from 0.16330205744971485, 0.1622800929454408, 0.16152863785320556, 0.16102926868856743
LBFGS Accumulation History: 3 points
Removed measurement 622559d9 to history. Total: 4
th(0)=0.16102926868856743;dx=-6.697993333755044E-10
Armijo: th(2.623748785376714E7)=0.1673722114638111; dx=3.3003216450908937E-10 evalInputDelta=-0.006342942775243665
Armijo: th(1.311874392688357E7)=0.1612124430097911; dx=1.4076185067385871E-10 evalInputDelta=-1.8317432122366695E-4
New Minimum: 0.16102926868856743 > 0.1598110001500463
END: th(4372914.642294523)=0.1598110001500463; dx=-2.7734464951949257E-10 evalInputDelta=0.001218268538521139
Fitness changed from 0.16102926868856743 to 0.1598110001500463
Iteration 50 complete. Error: 0.1598110001500463 Total: 11.1970; Orientation: 1.6271; Line Search: 8.0868
<a id="p-29"></a>Iteration 50
<a id="p-28"></a>
Final threshold in iteration 51: 0.1598110001500463 (> -Infinity) after 406.809s (< 3600.000s)
0.1598110001500463