com.simiacryptus.mindseye.util.TFConverter Maven / Gradle / Ivy
/*
* 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.util;
import com.google.common.collect.Streams;
import com.simiacryptus.mindseye.lang.Coordinate;
import com.simiacryptus.mindseye.lang.Layer;
import com.simiacryptus.mindseye.lang.Tensor;
import com.simiacryptus.mindseye.layers.cudnn.*;
import com.simiacryptus.mindseye.layers.cudnn.conv.SimpleConvolutionLayer;
import com.simiacryptus.mindseye.layers.java.FullyConnectedLayer;
import com.simiacryptus.mindseye.layers.tensorflow.MatMulLayer;
import com.simiacryptus.mindseye.layers.tensorflow.TFLayer;
import com.simiacryptus.mindseye.layers.tensorflow.TFLayerBase;
import com.simiacryptus.mindseye.network.DAGNode;
import com.simiacryptus.mindseye.network.InnerNode;
import com.simiacryptus.mindseye.network.PipelineNetwork;
import com.simiacryptus.ref.lang.RefUtil;
import com.simiacryptus.ref.wrappers.*;
import com.simiacryptus.tensorflow.GraphModel;
import com.simiacryptus.tensorflow.ImageNetworkPipeline;
import org.jetbrains.annotations.NotNull;
import org.tensorflow.framework.AttrValue;
import org.tensorflow.framework.GraphDef;
import javax.annotation.Nonnull;
import javax.annotation.Nullable;
import java.util.List;
import java.util.Map;
import java.util.function.Consumer;
public class TFConverter {
public static RefList getLayers(@Nonnull ImageNetworkPipeline pipeline) {
return RefIntStream.range(0, pipeline.graphDefs.size()).mapToObj(i -> getLayer(pipeline, i))
.collect(RefCollectors.toList());
}
@Nonnull
public static TFLayer getLayer(@Nonnull ImageNetworkPipeline pipeline, int i) {
GraphDef graphDef = pipeline.graphDefs.get(i);
String output = pipeline.nodeIds().get(i);
String input = i == 0 ? "input" : pipeline.nodeIds().get(i - 1);
TFLayer temp_05_0004 = new TFLayer(graphDef.toByteArray(), new RefHashMap<>(), output, input);
temp_05_0004.setFloat(true);
TFLayer temp_05_0003 = temp_05_0004.addRef();
temp_05_0004.freeRef();
return temp_05_0003;
}
@Nonnull
public FullyConnectedLayer getFCLayer(@Nonnull MatMulLayer matMulLayer) {
RefMap temp_05_0010 = matMulLayer.getWeights();
assert temp_05_0010 != null;
Tensor weights = temp_05_0010.get("weights");
temp_05_0010.freeRef();
int[] intputDims = matMulLayer.getIntputDims();
int[] outputDims = matMulLayer.getOutputDims();
matMulLayer.freeRef();
int[] tfView = Streams
.concat(RefArrays.stream(outputDims),
RefIntStream.range(0, intputDims.length).map(i -> intputDims.length - 1 - i).map(i -> intputDims[i]))
.toArray();
int[] tfPermute = Streams
.concat(RefIntStream.range(0, intputDims.length).map(i -> outputDims.length + intputDims.length - 1 - i),
RefIntStream.range(0, outputDims.length))
.toArray();
assert weights != null;
Tensor temp_05_0011 = weights.reshapeCast(tfView);
Tensor rearranged = temp_05_0011.permuteDimensions(tfPermute);
temp_05_0011.freeRef();
weights.freeRef();
FullyConnectedLayer fullyConnectedLayer = new FullyConnectedLayer(intputDims, outputDims);
Tensor temp_05_0012 = fullyConnectedLayer.getWeights();
assert temp_05_0012 != null;
temp_05_0012.set(rearranged.addRef());
temp_05_0012.freeRef();
rearranged.freeRef();
return fullyConnectedLayer;
}
@Nonnull
public PipelineNetwork convert(@Nonnull TFLayerBase tfLayer) {
final PipelineNetwork converted = new PipelineNetwork(1);
RefConcurrentHashMap nodes = new RefConcurrentHashMap<>();
RefUtil.freeRef(getNode(tfLayer.getOutputNode(), converted.addRef(),
new GraphModel(tfLayer.constGraph().toByteArray()), RefUtil.addRef(nodes)));
tfLayer.freeRef();
nodes.freeRef();
return converted;
}
@Nullable
protected DAGNode getNode(@Nonnull String id, @Nonnull PipelineNetwork network, @Nonnull GraphModel tfModel,
@Nonnull RefMap map) {
try {
if (!map.containsKey(id)) {
DAGNode result = getDagNode(id, network, tfModel, map.addRef());
if (!map.containsKey(id)) {
RefUtil.freeRef(map.put(id, result));
} else {
result.freeRef();
}
} else {
network.freeRef();
}
DAGNode node = map.get(id);
map.freeRef();
return node;
} catch (Throwable e) {
throw new RuntimeException("Error converting " + id, e);
}
}
@Nonnull
protected PoolingLayer getPoolingLayer(@Nonnull GraphModel.GraphNode graphNode) {
PoolingLayer temp_05_0005 = new PoolingLayer();
temp_05_0005.setMode(PoolingLayer.PoolingMode.Max);
PoolingLayer poolingLayer = temp_05_0005.addRef();
temp_05_0005.freeRef();
Map attrMap = graphNode.getNodeDef().getAttrMap();
assert "SAME".equals(attrMap.get("padding").getS().toStringUtf8());
AttrValue _ksize = attrMap.get("ksize");
if (null != _ksize) {
List ksize = _ksize.getList().getIList();
poolingLayer.setWindowX(Math.toIntExact(ksize.get(1)));
poolingLayer.setWindowY(Math.toIntExact(ksize.get(2)));
}
AttrValue _strides = attrMap.get("strides");
if (null != _strides) {
List strides = _strides.getList().getIList();
poolingLayer.setStrideX(Math.toIntExact(strides.get(1)));
poolingLayer.setStrideY(Math.toIntExact(strides.get(2)));
}
return poolingLayer;
}
@Nonnull
protected ImgBandBiasLayer getBiasAdd(@Nonnull GraphModel.GraphNode graphNode) {
GraphModel.GraphNode dataNode = graphNode.getInputs().get(1);
assert dataNode.getOp().equals("Const");
double[] data = dataNode.getData();
assert data != null;
Tensor tensor = new Tensor(data, data.length);
ImgBandBiasLayer temp_05_0006 = new ImgBandBiasLayer(data.length);
temp_05_0006.set(tensor.addRef());
ImgBandBiasLayer temp_05_0001 = temp_05_0006.addRef();
temp_05_0006.freeRef();
tensor.freeRef();
return temp_05_0001;
}
@Nonnull
protected Layer getConv2D(@Nonnull GraphModel.GraphNode graphNode) {
GraphModel.GraphNode dataNode = graphNode.getInputs().get(1);
assert dataNode.getOp().equals("Const");
int[] kernelDims = RefArrays.stream(dataNode.getShape()).mapToInt(x -> (int) x).toArray();
double[] data = dataNode.getData();
if (kernelDims.length == 0) {
assert data != null;
kernelDims = new int[]{data.length};
}
Tensor temp_05_0007 = new Tensor(data, kernelDims[3], kernelDims[2], kernelDims[1], kernelDims[0]);
Tensor sourceKernel = temp_05_0007.invertDimensions();
temp_05_0007.freeRef();
int[] sourceKernelDimensions = sourceKernel.getDimensions();
// ConvolutionLayer convolutionLayer = new ConvolutionLayer(sourceKernelDimensions[0], sourceKernelDimensions[1], sourceKernelDimensions[2], sourceKernelDimensions[3]);
SimpleConvolutionLayer convolutionLayer = new SimpleConvolutionLayer(sourceKernelDimensions[0],
sourceKernelDimensions[1], sourceKernelDimensions[2] * sourceKernelDimensions[3]);
Tensor targetKernel = new Tensor(sourceKernelDimensions[0], sourceKernelDimensions[1], sourceKernelDimensions[2],
sourceKernelDimensions[3]);
sourceKernel.coordStream(false).forEach(RefUtil.wrapInterface((Consumer) c -> {
int[] coord = c.getCoords();
targetKernel.set(sourceKernelDimensions[0] - 1 - coord[0], sourceKernelDimensions[1] - 1 - coord[1], coord[2],
coord[3], sourceKernel.get(c));
}, targetKernel.addRef(), sourceKernel.addRef()));
sourceKernel.freeRef();
convolutionLayer.set(targetKernel.addRef());
targetKernel.freeRef();
AttrValue stridesArr = graphNode.getNodeDef().getAttrMap().get("strides");
if (null != stridesArr) {
int[] strides = stridesArr.getList().getIList().stream().mapToInt(x -> Math.toIntExact(x)).toArray();
int strideX = strides[1];
int strideY = strides[2];
if (strideX > 1 || strideY > 1) {
convolutionLayer.setStrideX(strideX);
convolutionLayer.setStrideY(strideY);
return convolutionLayer;
} else {
return convolutionLayer;
}
} else {
return convolutionLayer;
}
}
@NotNull
private DAGNode getDagNode(@Nonnull String id, @Nonnull PipelineNetwork network, @Nonnull GraphModel tfModel, @Nonnull RefMap map) {
GraphModel.GraphNode graphNode = tfModel.getChild(id);
assert null != graphNode;
try {
if (graphNode.getOp().equals("Conv2D")) {
return network.add(getConv2D(graphNode), getNode(graphNode.getInputKeys().get(0),
network.addRef(), tfModel, map));
} else if (graphNode.getOp().equals("BiasAdd")) {
return network.add(getBiasAdd(graphNode), getNode(graphNode.getInputKeys().get(0),
network.addRef(), tfModel, map));
} else if (graphNode.getOp().equals("Relu")) {
return network.add(new ActivationLayer(ActivationLayer.Mode.RELU), getNode(graphNode.getInputKeys().get(0),
network.addRef(), tfModel, map));
} else if (graphNode.getOp().equals("LRN")) {
return network.add(getLRNLayer(graphNode), getNode(graphNode.getInputKeys().get(0),
network.addRef(), tfModel, map));
} else if (graphNode.getOp().equals("MaxPool")) {
return network.add(getPoolingLayer(graphNode), getNode(graphNode.getInputKeys().get(0),
network.addRef(), tfModel, map));
} else if (graphNode.getOp().equals("Concat")) {
List inputKeys = graphNode.getInputKeys();
InnerNode innerNode = network.add(new ImgConcatLayer(),
inputKeys.stream().skip(1)
.map(inputKey -> getNode(inputKey,
network.addRef(), tfModel, RefUtil.addRef(map)))
.toArray(i -> new DAGNode[i]));
map.freeRef();
return innerNode;
} else if (graphNode.getOp().equals("Placeholder")) {
map.freeRef();
return network.getInput(0);
} else {
map.freeRef();
throw new IllegalArgumentException(graphNode.getOp());
}
} finally {
network.freeRef();
}
}
@Nonnull
private LRNLayer getLRNLayer(@Nonnull GraphModel.GraphNode graphNode) {
Map attrMap = graphNode.getNodeDef().getAttrMap();
long depth_radius = attrMap.get("depth_radius").getI();
float alpha = attrMap.get("alpha").getF();
float bias = attrMap.get("bias").getF();
float beta = attrMap.get("beta").getF();
long width = depth_radius * 2 + 1;
LRNLayer temp_05_0009 = new LRNLayer((int) width);
temp_05_0009.setAlpha(alpha * width);
LRNLayer temp_05_0014 = temp_05_0009.addRef();
temp_05_0014.setBeta(beta);
LRNLayer temp_05_0015 = temp_05_0014.addRef();
temp_05_0015.setK(bias);
LRNLayer temp_05_0008 = temp_05_0015.addRef();
temp_05_0015.freeRef();
temp_05_0014.freeRef();
temp_05_0009.freeRef();
return temp_05_0008;
}
}
© 2015 - 2025 Weber Informatics LLC | Privacy Policy