com.simiacryptus.mindseye.art.examples.OperatorSurvey.scala Maven / Gradle / Ivy
The newest version!
/*
* Copyright (c) 2020 by Andrew Charneski.
*
* The author licenses this file to you under the
* Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance
* with the License. You may obtain a copy
* of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing,
* software distributed under the License is distributed on an
* "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
* KIND, either express or implied. See the License for the
* specific language governing permissions and limitations
* under the License.
*/
package com.simiacryptus.mindseye.art.examples
import java.awt.image.BufferedImage
import java.awt.{Font, Graphics2D}
import java.net.URI
import com.simiacryptus.mindseye.art.models.VGG16
import com.simiacryptus.mindseye.art.ops._
import com.simiacryptus.mindseye.art.util.ArtSetup.{ec2client, s3client}
import com.simiacryptus.mindseye.art.util.{BasicOptimizer, _}
import com.simiacryptus.mindseye.eval.Trainable
import com.simiacryptus.mindseye.lang.Tensor
import com.simiacryptus.mindseye.opt.Step
import com.simiacryptus.notebook.NotebookOutput
import com.simiacryptus.ref.wrappers.RefAtomicReference
import com.simiacryptus.sparkbook.NotebookRunner._
import com.simiacryptus.sparkbook._
import com.simiacryptus.sparkbook.util.Java8Util._
import com.simiacryptus.sparkbook.util.LocalRunner
import scala.collection.mutable.ArrayBuffer
import scala.util.Try
object OperatorSurvey extends OperatorSurvey with LocalRunner[Object] with NotebookRunner[Object]
class OperatorSurvey extends ArtSetup[Object] {
val styleUrl = "upload:Style"
val initUrl: String = "50 + noise * 0.5"
val s3bucket: String = "test.deepartist.org"
val resolution = 800
val animationDelay = 1000
val magnification = Array(4.0)
override def indexStr = "102"
override def description =
Paints a texture using a variety of signal operators but each having:
- A single input image to define style
- Random noise initialization
- Standard VGG16 layers to define the style
- A single resolution
It demonstrates the variety of effects that can be obtained using varied operators.
.toString.trim
override def inputTimeoutSeconds = 3600
override def postConfigure(log: NotebookOutput) = log.eval { () =>
() => {
implicit val l = log
// First, basic configuration so we publish to our s3 site
if (Option(s3bucket).filter(!_.isEmpty).isDefined)
log.setArchiveHome(URI.create(s"s3://$s3bucket/$className/${log.getId}/"))
log.onComplete(() => upload(log): Unit)
// Fetch input images (user upload prompts) and display rescaled copies
log.out(log.jpg(ImageArtUtil.loadImage(log, styleUrl, (resolution * Math.sqrt(magnification.head)).toInt), "Input Style"))
val renderedCanvases = new ArrayBuffer[() => BufferedImage]
// Execute the main process while registered with the site index
val registration = registerWithIndexGIF(renderedCanvases.map(_ ()).toList, delay = animationDelay)
withMonitoredGif(() => renderedCanvases.map(_ ()).toList, delay = animationDelay) {
try {
val operatorMap = Map(
"GramMatrixEnhancer" -> new GramMatrixEnhancer(),
"GramMatrixEnhancer (100)" -> new GramMatrixEnhancer().setMinMax(-100, 100),
"ChannelMeanMatcher" -> new ChannelMeanMatcher(),
"MomentMatcher" -> new MomentMatcher(),
"MomentMatcher (no-cov)" -> {
val momentMatcher = new MomentMatcher()
momentMatcher.setCovCoeff(0.0)
momentMatcher
},
"MomentMatcher (no-pos)" -> {
val momentMatcher = new MomentMatcher()
momentMatcher.setPosCoeff(0.0)
momentMatcher
},
"GramMatrixMatcher" -> new GramMatrixMatcher(),
"ChannelPowerEnhancer" -> new ChannelPowerEnhancer()
)
for (modifiers <- oneAtATime(operatorMap)) {
val canvas = new RefAtomicReference[Tensor](null)
val keys = modifiers.keys.toList.sorted
log.h2(keys.mkString(" + "))
renderedCanvases += (() => {
val image = canvas.get().toImage
if (null == image) image else {
val graphics = image.getGraphics.asInstanceOf[Graphics2D]
graphics.setFont(new Font("Calibri", Font.BOLD, 24))
var y = 25
for (modifier <- keys) {
graphics.drawString(modifier, 10, y)
y = y + 50
}
image
}
})
withMonitoredJpg(() => Option(canvas.get()).map(_.toRgbImage).orNull) {
var steps = 0
Try {
log.subreport("Painting", (sub: NotebookOutput) => {
paint(
contentUrl = styleUrl,
initUrl = initUrl,
canvas = canvas,
network = new VisualStyleNetwork(
styleLayers = List(
// We select all the lower-level layers to achieve a good balance between speed and accuracy.
VGG16.VGG16_0b,
VGG16.VGG16_1a,
VGG16.VGG16_1b1,
VGG16.VGG16_1b2,
VGG16.VGG16_1c1,
VGG16.VGG16_1c2,
VGG16.VGG16_1c3
),
styleModifiers = modifiers.values.toList,
styleUrl = List(styleUrl),
magnification = magnification
),
optimizer = new BasicOptimizer {
override val trainingMinutes: Int = 60
override val trainingIterations: Int = 20
override val maxRate = 1e9
override def onStepComplete(trainable: Trainable, currentPoint: Step): Boolean = {
steps = steps + 1
super.onStepComplete(trainable, currentPoint)
}
},
aspect = None,
resolutions = new GeometricSequence {
override val min: Double = resolution
override val max: Double = resolution
override val steps = 1
}.toStream.map(_.round.toDouble))(sub)
null
})
}
if (steps < 3 && !renderedCanvases.isEmpty) {
renderedCanvases.remove(renderedCanvases.size - 1)
}
uploadAsync(log)
}(log)
}
null
} finally {
Try {
registration.foreach(_.stop()(s3client, ec2client))
}
}
}
}
}()
def oneAtATime[T](map: Map[String, T]): List[Map[String, T]] = {
map.toList.map(Map(_))
}
/**
* Full expansion of sets containing one or more element of the source set
*
* @param map Input collection
* @tparam T collection type
* @return All valid subsets of the collection
*/
def submaps[T](map: Map[String, T]): List[Map[String, T]] = {
subsets(map.toList).sortBy(_.toList.sortBy(_._1).mkString("")).map(_.toMap)
}
def subsets[T](map: Seq[T]) = {
map.map(x => List(List.empty, List(x)))
.reduce((a, b) => for (a <- a; b <- b) yield a ++ b)
.filterNot(_.isEmpty).map(_.toSet).distinct
}
}
© 2015 - 2025 Weber Informatics LLC | Privacy Policy