Code from BasicOptimizer.scala:75 executed in 115.81 seconds (1.833 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: 900054130309900
Reset training subject: 900055123212500
Adding measurement 452ec287 to history. Total: 0
LBFGS Accumulation History: 1 points
Constructing line search parameters: GD+Trust
th(0)=1.4620341956615448;dx=-3.302801116576718E-7
New Minimum: 1.4620341956615448 > 1.4620337039232254
WOLFE (weak): th(2.154434690031884)=1.4620337039232254; dx=-3.3028131151048744E-7 evalInputDelta=4.917383193969727E-7
New Minimum: 1.4620337039232254 > 1.4620329588651657
WOLFE (weak): th(4.308869380063768)=1.4620329588651657; dx=-3.302810962838615E-7 evalInputDelta=1.2367963790893555E-6
New Minimum: 1.4620329588651657 > 1.4620305001735687
WOLFE (weak): th(12.926608140191302)=1.4620305001735687; dx=-3.302810772901591E-7 evalInputDelta=3.6954879760742188E-6
New Minimum: 1.4620305001735687 > 1.4620198458433151
WOLFE (weak): th(51.70643256076521)=1.4620198458433151; dx=-3.302760928148587E-7 evalInputDelta=1.4349818229675293E-5
New Minimum: 1.4620198458433151 > 1.461962416768074
WOLFE (weak): th(258.53216280382605)=1.461962416768074; dx=-3.302578571064784E-7 evalInputDelta=7.177889347076416E-5
New Minimum: 1.461962416768074 > 1.4616042077541351
WOLFE (weak): th(1551.1929768229563)=1.4616042077541351; dx=-3.306390732095289E-7 evalInputDelta=4.2998790740966797E-4
New Minimum: 1.4616042077541351 > 1.459037721157074
WOLFE (weak): th(10858.350837760694)=1.459037721157074; dx=-3.2911300688992205E-7 evalInputDelta=0.002996474504470825
New Minimum: 1.459037721157074 > 1.438677579164505
WOLFE (weak): th(86866.80670208555)=1.438677579164505; dx=-3.199228655429477E-7 evalInputDelta=0.023356616497039795
New Minimum: 1.438677579164505 > 1.3061667382717133
END: th(781801.26031877)=1.3061667382717133; dx=-2.1357542815952788E-7 evalInputDelta=0.15586745738983154
Fitness changed from 1.4620341956615448 to 1.3061667382717133
Iteration 1 complete. Error: 1.3061667382717133 Total: 13.2788; Orientation: 0.0150; Line Search: 10.2748
<a id="p-3"></a>Iteration 1
<a id="p-2"></a>=-0.23934021592140198;dx=-1.879439130098054E-9
Armijo: th(8.203125000000007E7)=-0.20492437481880188; dx=2.6187891149998234E-11 evalInputDelta=-0.0344158411026001
New Minimum: -0.23934021592140198 > -0.24319112300872803
END: th(4.101562500000004E7)=-0.24319112300872803; dx=-4.785687666298924E-10 evalInputDelta=0.00385090708732605
Fitness changed from -0.23934021592140198 to -0.24319112300872803
Iteration 24 complete. Error: -0.24319112300872803 Total: 4.3587; Orientation: 0.4095; Line Search: 2.9818
Adding measurement 3ee68377 to history. Total: 3
Rejected: LBFGS Orientation magnitude: 2.500e+03, gradient 4.476e-05, dot -0.654; [a13f0abe-45b5-46cb-a186-a1d80936ea8c = 1.000/1.000e+00]
Orientation rejected. Popping history element from -0.22959448397159576, -0.2352561503648758, -0.23934021592140198, -0.24319112300872803
LBFGS Accumulation History: 3 points
Removed measurement 1f12d5e0 to history. Total: 4
th(0)=-0.24319112300872803;dx=-2.0036202972524417E-9
Armijo: th(8.836548533333907E7)=-0.19426393508911133; dx=2.1599074871407665E-10 evalInputDelta=-0.0489271879196167
New Minimum: -0.24319112300872803 > -0.2452404499053955
END: th(4.418274266666953E7)=-0.2452404499053955; dx=-3.4052622221177666E-10 evalInputDelta=0.0020493268966674805
Fitness changed from -0.24319112300872803 to -0.2452404499053955
Iteration 25 complete. Error: -0.2452404499053955 Total: 4.4287; Orientation: 0.4369; Line Search: 3.0162
<a id="p-19"></a>Iteration 25
<a id="p-18"></a>
Final threshold in iteration 26: -0.2452404499053955 (> -Infinity) after 115.811s (< 3600.000s)
-0.2452404499053955