All Downloads are FREE. Search and download functionalities are using the official Maven repository.
Please wait. This can take some minutes ...
Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance.
Project price only 1 $
You can buy this project and download/modify it how often you want.
org.nd4j.linalg.api.ops.random.impl.Range Maven / Gradle / Ivy
/*******************************************************************************
* Copyright (c) 2015-2018 Skymind, Inc.
*
* 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.
*
* 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.random.impl;
import lombok.val;
import onnx.OnnxProto3;
import org.nd4j.autodiff.samediff.SDVariable;
import org.nd4j.autodiff.samediff.SameDiff;
import org.nd4j.imports.graphmapper.tf.TFGraphMapper;
import org.nd4j.linalg.api.ops.DynamicCustomOp;
import org.nd4j.linalg.api.ops.Op;
import org.nd4j.linalg.factory.Nd4j;
import org.tensorflow.framework.AttrValue;
import org.tensorflow.framework.GraphDef;
import org.tensorflow.framework.NodeDef;
import java.util.Arrays;
import java.util.Collections;
import java.util.List;
import java.util.Map;
/**
* Range Op implementation, generates from..to distribution within Z
*
* @author [email protected]
*/
public class Range extends DynamicCustomOp {
private Double from;
private Double to;
private Double delta;
//used for initWithArrays when there are place holder
//values that need to be resolved
// private String fromVertexId,toVertexId,deltaVertexId;
public Range() {
// no-op
}
public Range(SameDiff sd, double from, double to, double step){
super(null, sd, new SDVariable[0]);
addTArgument(from, to, step);
this.from = from;
this.to = to;
this.delta = step;
}
@Override
public int opNum() {
return 4;
}
@Override
public String opName() {
return "range";
}
@Override
public String onnxName() {
return "Range";
}
@Override
public String tensorflowName() {
return "Range";
}
@Override
public void initFromTensorFlow(NodeDef nodeDef, SameDiff initWith, Map attributesForNode, GraphDef graph) {
super.initFromTensorFlow(nodeDef, initWith, attributesForNode, graph);
NodeDef startNode = null,endNode = null,deltaNode = null;
for(val node : graph.getNodeList()) {
if(node.getName().equals(nodeDef.getInput(0))) {
startNode = node;
}
if(node.getName().equals(nodeDef.getInput(1))) {
endNode = node;
}
if(node.getName().equals(nodeDef.getInput(2))) {
deltaNode = node;
}
if(startNode != null && endNode != null && deltaNode != null)
break;
}
val start = TFGraphMapper.getInstance().getNDArrayFromTensor("value",startNode,graph);
val end = TFGraphMapper.getInstance().getNDArrayFromTensor("value",endNode,graph);
val delta = TFGraphMapper.getInstance().getNDArrayFromTensor("value",deltaNode,graph);
if(start != null && end != null && delta != null) {
if(endNode.getName() != null && endNode.getName().equalsIgnoreCase("Rank")){
//Edge case: TF seems to produce rank 0 arrays for range op sometimes, in spite of docs - properties are [range/start, Rank, range/delta]
this.from = start.getDouble(0);
this.to = from+1;
this.delta = 1.0;
} else {
this.from = start.getDouble(0);
this.to = end.getDouble(0);
this.delta = delta.getDouble(0);
}
val outputVars = outputVariables();
addTArgument(this.from,this.to,this.delta);
val outputVertexId = outputVars[0].getVarName();
if(sameDiff.getArrForVarName(outputVertexId) == null) {
if(outputVars[0].getShape() == null) {
val calcShape = calculateOutputShape();
sameDiff.putShapeForVarName(outputVars[0].getVarName(),calcShape.get(0));
}
long[] shape = outputVars[0].getShape();
val arr = Nd4j.create(shape);
initWith.putArrayForVarName(outputVertexId, arr);
addOutputArgument(arr);
}
}
val fromVar = initWith.getVariable(TFGraphMapper.getInstance().getNodeName(startNode.getName()));
val toVar = initWith.getVariable(TFGraphMapper.getInstance().getNodeName(endNode.getName()));
val deltaVar = initWith.getVariable(TFGraphMapper.getInstance().getNodeName(deltaNode.getName()));
// this.fromVertexId = fromVar.getVarName();
// this.toVertexId = toVar.getVarName();
// this.deltaVertexId = deltaVar.getVarName();
}
@Override
public void initFromOnnx(OnnxProto3.NodeProto node, SameDiff initWith, Map attributesForNode, OnnxProto3.GraphProto graph) {
super.initFromOnnx(node, initWith, attributesForNode, graph);
}
@Override
public List calculateOutputShape() {
val iArgs = iArgs();
val tArgs = tArgs();
val inputArgs = inputArguments();
int cnt = 0;
if (iArgs.length > 0) {
int start = (int) iArgs[0];
int stop = (int) iArgs[1];
int step = (int) iArgs[2];
double e = (double) start;
if (start > stop) {
while (e > (double) stop) {
cnt++;
e = (double) step > 0.0 ? e - step : e + step;
}
} else {
while (e < (double) stop) {
cnt++;
e += step;
}
}
return Arrays.asList(new long[]{cnt});
}
else if (tArgs.length > 0) {
double start = tArgs[0];
double stop = tArgs[1];
double step = tArgs[2];
double e = start;
if (start > stop) {
while (e > stop) {
cnt++;
e = step > 0.0 ? e - step : e + step;
}
} else {
while (e < stop) {
cnt++;
e += step;
}
}
return Arrays.asList(new long[]{cnt});
}
else if(inputArgs.length > 0) {
double start = inputArgs[0].getDouble(0);
double stop = inputArgs[1].getDouble(0);
double step = inputArgs[2].getDouble(0);
double e = start;
if (start > stop) {
while (e > stop) {
cnt++;
e = step > 0.0 ? e - step : e + step;
}
} else {
while (e < stop) {
cnt++;
e += step;
}
}
return Arrays.asList(new long[]{cnt});
}
return Collections.emptyList();
}
@Override
public List doDiff(List f1) {
return Collections.emptyList();
}
@Override
public Op.Type opType() {
return Op.Type.CUSTOM;
}
}