List of usage examples for org.apache.hadoop.io WritableUtils writeString
public static void writeString(DataOutput out, String s) throws IOException
From source file:org.apache.hama.ml.ann.SmallLayeredNeuralNetwork.java
License:Apache License
@Override public void write(DataOutput output) throws IOException { super.write(output); // write squashing functions output.writeInt(this.squashingFunctionList.size()); for (DoubleFunction aSquashingFunctionList : this.squashingFunctionList) { WritableUtils.writeString(output, aSquashingFunctionList.getFunctionName()); }/*ww w. j a va2 s.com*/ // write weight matrices output.writeInt(this.weightMatrixList.size()); for (DoubleMatrix aWeightMatrixList : this.weightMatrixList) { MatrixWritable.write(aWeightMatrixList, output); } // DO NOT WRITE WEIGHT UPDATE }
From source file:org.apache.hama.ml.perception.SmallMultiLayerPerceptron.java
License:Apache License
@Override public void write(DataOutput output) throws IOException { WritableUtils.writeString(output, MLPType); output.writeDouble(learningRate);/*from w w w . ja v a 2s. com*/ output.writeDouble(regularization); output.writeDouble(momentum); output.writeInt(numberOfLayers); WritableUtils.writeString(output, squashingFunctionName); WritableUtils.writeString(output, costFunctionName); // write the number of neurons for each layer for (int i = 0; i < this.numberOfLayers; ++i) { output.writeInt(this.layerSizeArray[i]); } for (int i = 0; i < numberOfLayers - 1; ++i) { MatrixWritable matrixWritable = new MatrixWritable(this.weightMatrice[i]); matrixWritable.write(output); } // serialize the feature transformer Class<? extends FeatureTransformer> featureTransformerCls = this.featureTransformer.getClass(); byte[] featureTransformerBytes = SerializationUtils.serialize(featureTransformerCls); output.writeInt(featureTransformerBytes.length); output.write(featureTransformerBytes); }
From source file:org.apache.hama.ml.perception.TestSmallMultiLayerPerceptron.java
License:Apache License
/** * Test the output of an example MLP.//w w w. ja v a 2s.c om */ @Test public void testOutput() { // write the MLP meta-data manually String modelPath = "/tmp/sampleModel-testOutput.data"; Configuration conf = new Configuration(); try { FileSystem fs = FileSystem.get(conf); FSDataOutputStream output = fs.create(new Path(modelPath), true); String MLPType = SmallMultiLayerPerceptron.class.getName(); double learningRate = 0.5; double regularization = 0.0; double momentum = 0.1; String squashingFunctionName = "Sigmoid"; String costFunctionName = "SquaredError"; int[] layerSizeArray = new int[] { 3, 2, 3, 3 }; int numberOfLayers = layerSizeArray.length; WritableUtils.writeString(output, MLPType); output.writeDouble(learningRate); output.writeDouble(regularization); output.writeDouble(momentum); output.writeInt(numberOfLayers); WritableUtils.writeString(output, squashingFunctionName); WritableUtils.writeString(output, costFunctionName); // write the number of neurons for each layer for (int i = 0; i < numberOfLayers; ++i) { output.writeInt(layerSizeArray[i]); } double[][] matrix01 = { // 4 by 2 { 0.5, 0.2 }, { 0.1, 0.1 }, { 0.2, 0.5 }, { 0.1, 0.5 } }; double[][] matrix12 = { // 3 by 3 { 0.1, 0.2, 0.5 }, { 0.2, 0.5, 0.2 }, { 0.5, 0.5, 0.1 } }; double[][] matrix23 = { // 4 by 3 { 0.2, 0.5, 0.2 }, { 0.5, 0.1, 0.5 }, { 0.1, 0.2, 0.1 }, { 0.1, 0.2, 0.5 } }; DoubleMatrix[] matrices = { new DenseDoubleMatrix(matrix01), new DenseDoubleMatrix(matrix12), new DenseDoubleMatrix(matrix23) }; for (DoubleMatrix mat : matrices) { MatrixWritable.write(mat, output); } // serialize the feature transformer FeatureTransformer transformer = new DefaultFeatureTransformer(); Class<? extends FeatureTransformer> featureTransformerCls = transformer.getClass(); byte[] featureTransformerBytes = SerializationUtils.serialize(featureTransformerCls); output.writeInt(featureTransformerBytes.length); output.write(featureTransformerBytes); output.close(); } catch (IOException e) { e.printStackTrace(); } // initial the mlp with existing model meta-data and get the output MultiLayerPerceptron mlp = new SmallMultiLayerPerceptron(modelPath); DoubleVector input = new DenseDoubleVector(new double[] { 1, 2, 3 }); try { DoubleVector result = mlp.output(input); assertArrayEquals(new double[] { 0.6636557, 0.7009963, 0.7213835 }, result.toArray(), 0.0001); } catch (Exception e1) { e1.printStackTrace(); } // delete meta-data try { FileSystem fs = FileSystem.get(conf); fs.delete(new Path(modelPath), true); } catch (IOException e) { e.printStackTrace(); } }
From source file:org.apache.hcatalog.mapreduce.HCatSplit.java
License:Apache License
@Override public void write(DataOutput output) throws IOException { String partitionInfoString = HCatUtil.serialize(partitionInfo); // write partitionInfo into output WritableUtils.writeString(output, partitionInfoString); WritableUtils.writeString(output, baseMapRedSplit.getClass().getName()); Writable baseSplitWritable = (Writable) baseMapRedSplit; //write baseSplit into output baseSplitWritable.write(output);/*from w w w . j a v a2 s .c o m*/ //write the table schema into output String tableSchemaString = HCatUtil.serialize(tableSchema); WritableUtils.writeString(output, tableSchemaString); }
From source file:org.apache.hive.hcatalog.mapreduce.HCatSplit.java
License:Apache License
@Override public void write(DataOutput output) throws IOException { String partitionInfoString = HCatUtil.serialize(partitionInfo); // write partitionInfo into output WritableUtils.writeString(output, partitionInfoString); WritableUtils.writeString(output, baseMapRedSplit.getClass().getName()); Writable baseSplitWritable = (Writable) baseMapRedSplit; //write baseSplit into output baseSplitWritable.write(output);//from w w w . ja v a 2s . c om }
From source file:org.apache.horn.core.AbstractLayeredNeuralNetwork.java
License:Apache License
@Override public void write(DataOutput output) throws IOException { super.write(output); // write regularization weight output.writeFloat(this.regularizationWeight); // write momentum weight output.writeFloat(this.momentumWeight); // write cost function WritableUtils.writeString(output, costFunction.getFunctionName()); // write layer size list output.writeInt(this.layerSizeList.size()); for (Integer aLayerSizeList : this.layerSizeList) { output.writeInt(aLayerSizeList); }/*from w ww. j a va 2 s . c o m*/ WritableUtils.writeEnum(output, this.trainingMethod); WritableUtils.writeEnum(output, this.learningStyle); }
From source file:org.apache.horn.core.AbstractNeuralNetwork.java
License:Apache License
@Override public void write(DataOutput output) throws IOException { // write model type WritableUtils.writeString(output, modelType); // write learning rate output.writeFloat(learningRate);//www . ja v a 2 s . c om // write model path if (this.modelPath != null) { WritableUtils.writeString(output, modelPath); } else { WritableUtils.writeString(output, "null"); } // serialize the class Class<? extends FloatFeatureTransformer> featureTransformerCls = this.featureTransformer.getClass(); byte[] featureTransformerBytes = SerializationUtils.serialize(featureTransformerCls); output.writeInt(featureTransformerBytes.length); output.write(featureTransformerBytes); }
From source file:org.apache.horn.core.LayeredNeuralNetwork.java
License:Apache License
@Override public void write(DataOutput output) throws IOException { super.write(output); output.writeInt(finalLayerIdx);//from ww w. jav a 2 s. com output.writeFloat(dropRate); // write neuron classes output.writeInt(this.neuronClassList.size()); for (Class<? extends Neuron> clazz : this.neuronClassList) { output.writeUTF(clazz.getName()); } // write squashing functions output.writeInt(this.squashingFunctionList.size()); for (FloatFunction aSquashingFunctionList : this.squashingFunctionList) { WritableUtils.writeString(output, aSquashingFunctionList.getFunctionName()); } // write weight matrices output.writeInt(this.weightMatrixList.size()); for (FloatMatrix aWeightMatrixList : this.weightMatrixList) { FloatMatrixWritable.write(aWeightMatrixList, output); } // DO NOT WRITE WEIGHT UPDATE }
From source file:org.apache.horn.core.RecurrentLayeredNeuralNetwork.java
License:Apache License
@Override public void write(DataOutput output) throws IOException { super.write(output); output.writeInt(finalLayerIdx);// w w w .jav a2 s .com output.writeFloat(dropRate); // write neuron classes output.writeInt(this.neuronClassList.size()); for (Class<? extends Neuron> clazz : this.neuronClassList) { output.writeUTF(clazz.getName()); } // write squashing functions output.writeInt(this.squashingFunctionList.size()); for (FloatFunction aSquashingFunctionList : this.squashingFunctionList) { WritableUtils.writeString(output, aSquashingFunctionList.getFunctionName()); } // write recurrent step size output.writeInt(this.recurrentStepSize); // write recurrent step size output.writeInt(this.numOutCells); // write recurrent layer list output.writeInt(this.recurrentLayerList.size()); for (Boolean isReccurentLayer : recurrentLayerList) { output.writeBoolean(isReccurentLayer); } // write weight matrices output.writeInt(this.getSizeOfWeightmatrix()); for (List<FloatMatrix> aWeightMatrixLists : this.weightMatrixLists) { for (FloatMatrix aWeightMatrixList : aWeightMatrixLists) { FloatMatrixWritable.write(aWeightMatrixList, output); } } // DO NOT WRITE WEIGHT UPDATE }
From source file:org.apache.mahout.classifier.chi_rw.data.Dataset.java
License:Apache License
@Override public void write(DataOutput out) throws IOException { out.writeInt(attributes.length); // nb attributes for (Attribute attr : attributes) { WritableUtils.writeString(out, attr.name()); }/*from www . ja v a 2 s . c o m*/ Chi_RWUtils.writeArray(out, ignored); // only CATEGORICAL attributes have values for (String[] vals : values) { if (vals != null) { WritableUtils.writeStringArray(out, vals); } } // only NUMERICAL attributes have values for (double[] vals : nvalues) { if (vals != null) { Chi_RWUtils.writeArray(out, vals); } } for (double[] vals : minmaxvalues) { if (vals != null) { Chi_RWUtils.writeArray(out, vals); } } out.writeInt(labelId); out.writeInt(nbInstances); }