org.nd4j.linalg.api.ops.impl.reduce.Moments Maven / Gradle / Ivy
The newest version!
/*
* ******************************************************************************
* *
* *
* * This program and the accompanying materials are made available under the
* * terms of the Apache License, Version 2.0 which is available at
* * https://www.apache.org/licenses/LICENSE-2.0.
* *
* * See the NOTICE file distributed with this work for additional
* * information regarding copyright ownership.
* * 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.
* *
* * SPDX-License-Identifier: Apache-2.0
* *****************************************************************************
*/
package org.nd4j.linalg.api.ops.impl.reduce;
import lombok.NonNull;
import org.nd4j.autodiff.samediff.SDVariable;
import org.nd4j.autodiff.samediff.SameDiff;
import org.nd4j.common.base.Preconditions;
import org.nd4j.imports.graphmapper.tf.TFGraphMapper;
import org.nd4j.linalg.api.buffer.DataType;
import org.nd4j.linalg.api.ndarray.INDArray;
import org.nd4j.linalg.api.ops.DynamicCustomOp;
import org.nd4j.linalg.api.ops.impl.reduce.bp.MeanBp;
import org.nd4j.linalg.api.ops.impl.reduce.bp.VarianceBp;
import org.nd4j.linalg.factory.Nd4j;
import org.nd4j.shade.guava.primitives.Ints;
import org.tensorflow.framework.AttrValue;
import org.tensorflow.framework.GraphDef;
import org.tensorflow.framework.NodeDef;
import java.util.*;
public class Moments extends DynamicCustomOp {
private boolean keepDims;
public Moments() {
}
public Moments(@NonNull INDArray input, boolean keepDims, int... dimensions) {
super(new INDArray[]{input}, null);
this.dimensions = dimensions;
this.keepDims = keepDims;
addArgs();
}
public Moments(@NonNull INDArray input, int... dimensions) {
super(new INDArray[]{input}, null);
this.dimensions = dimensions;
addArgs();
}
public Moments(SameDiff sameDiff, SDVariable input) {
this(sameDiff, input, null);
addArgs();
}
public Moments(SameDiff sameDiff, SDVariable input, int[] axes) {
super(null, sameDiff, new SDVariable[] {input}, false);
this.dimensions = axes;
addArgs();
}
public Moments(INDArray in, INDArray outMean, INDArray outStd, int... axes) {
super(null, new INDArray[]{in}, new INDArray[]{outMean, outStd}, null, axes);
this.dimensions = axes;
addArgs();
}
public Moments(INDArray input, int[] axes, boolean keepDims) {
super(null,new INDArray[]{input},null);
this.keepDims = keepDims;
this.dimensions = axes;
addArgs();
}
public Moments(INDArray input, INDArray axes, boolean keepDims) {
super(null,new INDArray[]{input,axes},null);
this.keepDims = keepDims;
addArgs();
}
public Moments(SameDiff sd, SDVariable input, int[] axes, boolean keepDims) {
super(null,sd,new SDVariable[]{input},false);
this.keepDims = keepDims;
this.dimensions = axes;
addArgs();
}
public Moments(SameDiff sd, SDVariable input, SDVariable axes, boolean keepDims) {
super(null,sd,new SDVariable[]{input,axes},false);
this.keepDims = keepDims;
addArgs();
}
@Override
public void initFromTensorFlow(NodeDef nodeDef, SameDiff initWith, Map attributesForNode, GraphDef graph) {
TFGraphMapper.initFunctionFromProperties(nodeDef.getOp(), this, attributesForNode, nodeDef, graph);
addArgs();
}
@Override
public String opName() {
return "moments";
}
@Override
public List doDiff(List grad){
SDVariable dLdMean = grad.get(0);
SDVariable dLdVar = grad.get(1); //Note: non-bias-corrected variance
if(dimensions != null) {
SDVariable meanBp = new MeanBp(sameDiff, arg(), dLdMean, keepDims, dimensions).outputVariable();
SDVariable varBp = new VarianceBp(sameDiff, arg(), dLdVar, false, keepDims, dimensions).outputVariable();
return Collections.singletonList(meanBp.add(varBp));
} else if(numIArguments() > 0) {
int[] newDimensions = Ints.toArray(this.iArguments);
this.dimensions = newDimensions;
SDVariable meanBp = new MeanBp(sameDiff, arg(), dLdMean, keepDims, newDimensions).outputVariable();
SDVariable varBp = new VarianceBp(sameDiff, arg(), dLdVar, false, keepDims,newDimensions).outputVariable();
return Collections.singletonList(meanBp.add(varBp));
} else if(numInputArguments() > 1) {
SDVariable meanBp = new MeanBp(sameDiff, arg(), dLdMean, keepDims, arg(1)).outputVariable();
SDVariable varBp = new VarianceBp(sameDiff, arg(), dLdVar, false, keepDims, arg(1)).outputVariable();
return Collections.singletonList(meanBp.add(varBp));
} else {
SDVariable meanBp = new MeanBp(sameDiff, arg(), dLdMean, keepDims, dimensions).outputVariable();
SDVariable varBp = new VarianceBp(sameDiff, arg(), dLdVar, false, keepDims, dimensions).outputVariable();
return Collections.singletonList(meanBp.add(varBp));
}
}
@Override
public List calculateOutputDataTypes(List dataTypes) {
Preconditions.checkState(dataTypes != null && dataTypes.size() == 1, "Expected 1 datatype for %s, got %s", getClass(), dataTypes);
if(dataTypes.get(0).isFPType())
return Arrays.asList(dataTypes.get(0), dataTypes.get(0));
return Arrays.asList(Nd4j.defaultFloatingPointType(), Nd4j.defaultFloatingPointType());
}
@Override
public Map propertiesForFunction() {
Map properties = new HashMap<>();
properties.put("keepDims",keepDims);
properties.put("dimensions",dimensions);
return properties;
}
protected void addArgs() {
addBArgument(keepDims);
if(dimensions != null && dimensions.length > 0) {
if(dimensions.length != 1 || dimensions[0] != Integer.MAX_VALUE) {
//Integer.MAX_VALUE means "full array" but here no dimension args == full array
addIArgument(dimensions);
}
}
}
}