com.simiacryptus.mindseye.layers.java.MaxPoolingLayer Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of mindseye-java Show documentation
Show all versions of mindseye-java Show documentation
Pure Java Neural Networks Components
/*
* Copyright (c) 2019 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.layers.java;
import com.google.gson.JsonObject;
import com.simiacryptus.lang.Tuple2;
import com.simiacryptus.mindseye.lang.*;
import com.simiacryptus.util.JsonUtil;
import com.simiacryptus.util.Util;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;
import javax.annotation.Nonnull;
import javax.annotation.Nullable;
import java.util.Arrays;
import java.util.List;
import java.util.Map;
import java.util.UUID;
import java.util.function.Function;
import java.util.function.IntToDoubleFunction;
import java.util.stream.Collectors;
import java.util.stream.IntStream;
/**
* The type Max subsample key.
*/
@SuppressWarnings("serial")
public class MaxPoolingLayer extends LayerBase {
private static final Function>> calcRegionsCache = Util.cache(MaxPoolingLayer::calcRegions);
@SuppressWarnings("unused")
private static final Logger log = LoggerFactory.getLogger(MaxPoolingLayer.class);
private int[] kernelDims;
/**
* Instantiates a new Max subsample key.
*/
protected MaxPoolingLayer() {
super();
}
/**
* Instantiates a new Max subsample key.
*
* @param kernelDims the kernel dims
*/
public MaxPoolingLayer(@Nonnull final int... kernelDims) {
this.kernelDims = Arrays.copyOf(kernelDims, kernelDims.length);
}
/**
* Instantiates a new Max subsample key.
*
* @param id the id
* @param kernelDims the kernel dims
*/
protected MaxPoolingLayer(@Nonnull final JsonObject id, @Nonnull final int... kernelDims) {
super(id);
this.kernelDims = Arrays.copyOf(kernelDims, kernelDims.length);
}
private static List> calcRegions(@Nonnull final MaxPoolingLayer.CalcRegionsParameter p) {
@Nonnull final Tensor input = new Tensor(p.inputDims);
final int[] newDims = IntStream.range(0, p.inputDims.length).map(i -> {
//assert 0 == p.inputDims[i] % p.kernelDims[i];
return (int) Math.ceil(p.inputDims[i] * 1.0 / p.kernelDims[i]);
}).toArray();
@Nonnull final Tensor output = new Tensor(newDims);
List> tuple2s = output.coordStream(true).map(o -> {
Tensor tensor = new Tensor(p.kernelDims);
final int[] inCoords = tensor.coordStream(true).mapToInt(kernelCoord -> {
@Nonnull final int[] result = new int[o.getCoords().length];
for (int index = 0; index < o.getCoords().length; index++) {
final int outputCoordinate = o.getCoords()[index];
final int kernelSize = p.kernelDims[index];
final int baseCoordinate = Math.min(outputCoordinate * kernelSize, p.inputDims[index] - kernelSize);
final int kernelCoordinate = kernelCoord.getCoords()[index];
result[index] = baseCoordinate + kernelCoordinate;
}
return input.index(result);
}).toArray();
tensor.freeRef();
return new Tuple2<>(o.getIndex(), inCoords);
}).collect(Collectors.toList());
input.freeRef();
output.freeRef();
return tuple2s;
}
/**
* From json max subsample key.
*
* @param json the json
* @param rs the rs
* @return the max subsample key
*/
public static MaxPoolingLayer fromJson(@Nonnull final JsonObject json, Map rs) {
return new MaxPoolingLayer(json,
JsonUtil.getIntArray(json.getAsJsonArray("heapCopy")));
}
@Nonnull
@Override
public Result eval(@Nonnull final Result... inObj) {
Arrays.stream(inObj).forEach(nnResult -> nnResult.addRef());
final Result in = inObj[0];
in.getData().length();
@Nonnull final int[] inputDims = in.getData().getDimensions();
final List> regions = MaxPoolingLayer.calcRegionsCache.apply(new MaxPoolingLayer.CalcRegionsParameter(inputDims, kernelDims));
final Tensor[] outputA = IntStream.range(0, in.getData().length()).mapToObj(dataIndex -> {
final int[] newDims = IntStream.range(0, inputDims.length).map(i -> {
return (int) Math.ceil(inputDims[i] * 1.0 / kernelDims[i]);
}).toArray();
@Nonnull final Tensor output = new Tensor(newDims);
return output;
}).toArray(i -> new Tensor[i]);
Arrays.stream(outputA).mapToInt(x -> x.length()).sum();
@Nonnull final int[][] gradientMapA = new int[in.getData().length()][];
IntStream.range(0, in.getData().length()).forEach(dataIndex -> {
@Nullable final Tensor input = in.getData().get(dataIndex);
final Tensor output = outputA[dataIndex];
@Nonnull final IntToDoubleFunction keyExtractor = inputCoords -> input.get(inputCoords);
@Nonnull final int[] gradientMap = new int[input.length()];
regions.parallelStream().forEach(tuple -> {
final Integer from = tuple.getFirst();
final int[] toList = tuple.getSecond();
int toMax = -1;
double bestValue = Double.NEGATIVE_INFINITY;
for (final int c : toList) {
final double value = keyExtractor.applyAsDouble(c);
if (-1 == toMax || bestValue < value) {
bestValue = value;
toMax = c;
}
}
gradientMap[from] = toMax;
output.set(from, input.get(toMax));
});
input.freeRef();
gradientMapA[dataIndex] = gradientMap;
});
return new Result(TensorArray.wrap(outputA), (@Nonnull final DeltaSet buffer, @Nonnull final TensorList data) -> {
if (in.isAlive()) {
@Nonnull TensorArray tensorArray = TensorArray.wrap(IntStream.range(0, in.getData().length()).parallel().mapToObj(dataIndex -> {
@Nonnull final Tensor backSignal = new Tensor(inputDims);
final int[] ints = gradientMapA[dataIndex];
@Nullable final Tensor datum = data.get(dataIndex);
for (int i = 0; i < datum.length(); i++) {
backSignal.add(ints[i], datum.get(i));
}
datum.freeRef();
return backSignal;
}).toArray(i -> new Tensor[i]));
in.accumulate(buffer, tensorArray);
}
}) {
@Override
protected void _free() {
Arrays.stream(inObj).forEach(nnResult -> nnResult.freeRef());
}
@Override
public boolean isAlive() {
return in.isAlive();
}
};
}
@Nonnull
@Override
public JsonObject getJson(Map resources, DataSerializer dataSerializer) {
@Nonnull final JsonObject json = super.getJsonStub();
json.add("heapCopy", JsonUtil.getJson(kernelDims));
return json;
}
@Nonnull
@Override
public List state() {
return Arrays.asList();
}
/**
* The type Calc regions parameter.
*/
public static class CalcRegionsParameter {
/**
* The Input dims.
*/
public int[] inputDims;
/**
* The Kernel dims.
*/
public int[] kernelDims;
/**
* Instantiates a new Calc regions parameter.
*
* @param inputDims the input dims
* @param kernelDims the kernel dims
*/
public CalcRegionsParameter(final int[] inputDims, final int[] kernelDims) {
this.inputDims = inputDims;
this.kernelDims = kernelDims;
}
@Override
public boolean equals(@Nullable final Object obj) {
if (this == obj) {
return true;
}
if (obj == null) {
return false;
}
if (getClass() != obj.getClass()) {
return false;
}
@Nonnull final MaxPoolingLayer.CalcRegionsParameter other = (MaxPoolingLayer.CalcRegionsParameter) obj;
if (!Arrays.equals(inputDims, other.inputDims)) {
return false;
}
return Arrays.equals(kernelDims, other.kernelDims);
}
@Override
public int hashCode() {
final int prime = 31;
int result = 1;
result = prime * result + Arrays.hashCode(inputDims);
result = prime * result + Arrays.hashCode(kernelDims);
return result;
}
}
}