Merge remote-tracking branch 'upstream/master'

VERSION

@@ -1 +1 @@
-0.4-rc3.9-SNAPSHOT
+0.6.1-SNAPSHOT

deeplearning4j-core/pom.xml

@@ -30,7 +30,6 @@
                 <plugin>
                     <groupId>org.apache.maven.plugins</groupId>
                     <artifactId>maven-surefire-plugin</artifactId>
-                    <version>2.18.1</version>
                     <configuration>
                         <argLine>-Ddtype=double</argLine>
                     </configuration>
@@ -62,11 +61,6 @@
                 <version>${nd4j.version}</version>
                 <scope>test</scope>
             </dependency>
-            <dependency>
-                <groupId>org.apache.commons</groupId>
-                <artifactId>commons-math3</artifactId>
-                <version>${commonsmath.version}</version>
-            </dependency>
         </dependencies>
     </dependencyManagement>
 
@@ -81,7 +75,6 @@
         <dependency>
             <groupId>ch.qos.logback</groupId>
             <artifactId>logback-classic</artifactId>
-            <version>${logback.version}</version>
             <scope>test</scope>
         </dependency>
 
@@ -95,7 +88,6 @@
         <dependency>
             <groupId>org.apache.commons</groupId>
             <artifactId>commons-math3</artifactId>
-            <version>${commonsmath.version}</version>
         </dependency>
         <dependency>
             <groupId>commons-io</groupId>

deeplearning4j-core/src/main/java/org/deeplearning4j/datasets/mnist/MnistManager.java

@@ -92,7 +92,7 @@ public static void writeImageToPpm(int[][] image, String ppmFileName) throws IOE
     public MnistManager(String imagesFile, String labelsFile, boolean train) throws IOException {
         if (imagesFile != null) {
             images = new MnistImageFile(imagesFile, "r");
-            if(train) imagesArr = new MnistImageFile(imagesFile, "r").readImagesUnsafe(MnistDataFetcher.NUM_EXAMPLES);
+            if(train) imagesArr = images.readImagesUnsafe(MnistDataFetcher.NUM_EXAMPLES);
             else imagesArr = images.readImagesUnsafe(MnistDataFetcher.NUM_EXAMPLES_TEST);
         }
         if (labelsFile != null) {

deeplearning4j-core/src/test/java/org/deeplearning4j/datasets/iterator/MultipleEpochsIteratorTest.java

@@ -104,7 +104,7 @@ public void testCifarDataSetIteratorReset() {
 
         MultiLayerNetwork net = new MultiLayerNetwork(conf);
         net.init();
-        net.setListeners(Arrays.asList((IterationListener) new ScoreIterationListener(1)));
+        net.setListeners(new ScoreIterationListener(1));
 
         MultipleEpochsIterator ds = new MultipleEpochsIterator(epochs, new CifarDataSetIterator(10,20, new int[]{20,20,1}));
         net.fit(ds);

deeplearning4j-core/src/test/java/org/deeplearning4j/eval/EvalTest.java

@@ -652,14 +652,16 @@ public void testEvaluationWithMetaData() throws Exception {
 
         List<Prediction> errors = e.getPredictionErrors();          //*** New - get list of prediction errors from evaluation ***
         List<RecordMetaData> metaForErrors = new ArrayList<>();
-        for(Prediction p : errors) metaForErrors.add(p.getRecordMetaData());
+        for(Prediction p : errors){
+            metaForErrors.add((RecordMetaData)p.getRecordMetaData());
+        }
         DataSet ds = rrdsi.loadFromMetaData(metaForErrors);         //*** New - dynamically load a subset of the data, just for prediction errors ***
         INDArray output = net.output(ds.getFeatures());
 
         int count = 0;
         for(Prediction t : errors){
             System.out.println(t
-                    + "\t\tRaw Data: " + csv.loadFromMetaData(t.getRecordMetaData()).getRecord()    //*** New - load subset of data from MetaData object (usually batched for efficiency) ***
+                    + "\t\tRaw Data: " + csv.loadFromMetaData((RecordMetaData)t.getRecordMetaData()).getRecord()    //*** New - load subset of data from MetaData object (usually batched for efficiency) ***
                     + "\tNormalized: " + ds.getFeatureMatrix().getRow(count) + "\tLabels: " + ds.getLabels().getRow(count)
                     + "\tNetwork predictions: " + output.getRow(count));
             count++;

deeplearning4j-core/src/test/java/org/deeplearning4j/gradientcheck/CNNGradientCheckTest.java

@@ -5,6 +5,7 @@
 import org.deeplearning4j.nn.conf.MultiLayerConfiguration;
 import org.deeplearning4j.nn.conf.NeuralNetConfiguration;
 import org.deeplearning4j.nn.conf.Updater;
+import org.deeplearning4j.nn.conf.distribution.NormalDistribution;
 import org.deeplearning4j.nn.conf.inputs.InputType;
 import org.deeplearning4j.nn.conf.layers.ConvolutionLayer;
 import org.deeplearning4j.nn.conf.layers.OutputLayer;
@@ -239,6 +240,7 @@ public void testCnnWithSubsampling(){
                             .regularization(false)
                             .learningRate(1.0)
                             .updater(Updater.SGD)
+                            .weightInit(WeightInit.DISTRIBUTION).dist(new NormalDistribution(0,1))
                             .list()
                             .layer(0, new ConvolutionLayer.Builder(kernel, stride, padding)
                                     .nIn(inputDepth).nOut(3)
@@ -299,6 +301,7 @@ public void testCnnWithSubsamplingV2(){
                             .regularization(false)
                             .learningRate(1.0)
                             .updater(Updater.SGD)
+                            .weightInit(WeightInit.DISTRIBUTION).dist(new NormalDistribution(0,1))
                             .list()
                             .layer(0, new ConvolutionLayer.Builder(kernel, stride, padding)
                                     .nIn(inputDepth).nOut(3)

deeplearning4j-core/src/test/java/org/deeplearning4j/gradientcheck/GradientCheckTests.java

@@ -616,13 +616,13 @@ public void testGradientCnnFfRnn() {
         MultiLayerConfiguration conf = new NeuralNetConfiguration.Builder()
                 .updater(Updater.NONE)
                 .seed(12345)
+                .weightInit(WeightInit.DISTRIBUTION).dist(new NormalDistribution(0,1))
                 .list()
                 .layer(0, new ConvolutionLayer.Builder(5, 5)
                         .nIn(3)
                         .nOut(5)
                         .stride(1, 1)
                         .activation("tanh")
-                        .weightInit(WeightInit.XAVIER)
                         .build())    //Out: (10-5)/1+1 = 6 -> 6x6x5
                 .layer(1, new SubsamplingLayer.Builder(SubsamplingLayer.PoolingType.MAX)
                         .kernelSize(2, 2)
@@ -631,13 +631,11 @@ public void testGradientCnnFfRnn() {
                 .layer(2, new DenseLayer.Builder()
                         .nIn(5 * 5 * 5)
                         .nOut(4)
-                        .weightInit(WeightInit.XAVIER)
                         .activation("tanh")
                         .build())
                 .layer(3, new GravesLSTM.Builder()
                         .nIn(4)
                         .nOut(3)
-                        .weightInit(WeightInit.XAVIER)
                         .activation("tanh")
                         .build())
                 .layer(4, new RnnOutputLayer.Builder()
@@ -797,7 +795,7 @@ public void testAutoEncoder() {
                                 .l2(l2).l1(l1)
                                 .optimizationAlgo(OptimizationAlgorithm.CONJUGATE_GRADIENT)
                                 .seed(12345L)
-                                .weightInit(WeightInit.XAVIER)
+                                .weightInit(WeightInit.DISTRIBUTION).dist(new NormalDistribution(0,1))
                                 .updater(Updater.SGD)
                                 .list()
                                 .layer(0, new AutoEncoder.Builder()

deeplearning4j-core/src/test/java/org/deeplearning4j/nn/graph/TestComputationGraphNetwork.java

@@ -470,17 +470,17 @@ public void testPreTraining(){
                         .nIn(4).nOut(3)
                         .weightInit(WeightInit.DISTRIBUTION).dist(new UniformDistribution(0, 1))
                         .activation("tanh")
-                        .lossFunction(LossFunctions.LossFunction.RMSE_XENT).build(), "in")
+                        .lossFunction(LossFunctions.LossFunction.KL_DIVERGENCE).build(), "in")
                 .addLayer("layer1", new RBM.Builder(RBM.HiddenUnit.GAUSSIAN, RBM.VisibleUnit.GAUSSIAN)
                         .nIn(4).nOut(3)
                         .weightInit(WeightInit.DISTRIBUTION).dist(new UniformDistribution(0, 1))
                         .activation("tanh")
-                        .lossFunction(LossFunctions.LossFunction.RMSE_XENT).build(), "in")
+                        .lossFunction(LossFunctions.LossFunction.KL_DIVERGENCE).build(), "in")
                 .addLayer("layer2", new RBM.Builder(RBM.HiddenUnit.GAUSSIAN, RBM.VisibleUnit.GAUSSIAN)
                         .nIn(3).nOut(3)
                         .weightInit(WeightInit.DISTRIBUTION).dist(new UniformDistribution(0, 1))
                         .activation("tanh")
-                        .lossFunction(LossFunctions.LossFunction.RMSE_XENT).build(),"layer1")
+                        .lossFunction(LossFunctions.LossFunction.KL_DIVERGENCE).build(),"layer1")
                 .addLayer("out", new org.deeplearning4j.nn.conf.layers.OutputLayer.Builder(LossFunctions.LossFunction.MCXENT)
                         .nIn(3+3).nOut(3)
                         .weightInit(WeightInit.DISTRIBUTION).dist(new UniformDistribution(0, 1))

deeplearning4j-core/src/test/java/org/deeplearning4j/nn/layers/feedforward/rbm/RBMTests.java

@@ -80,7 +80,7 @@ public void testLfw() throws Exception {
                 .layer(new org.deeplearning4j.nn.conf.layers.RBM.Builder(org.deeplearning4j.nn.conf.layers.RBM.HiddenUnit.RECTIFIED, org.deeplearning4j.nn.conf.layers.RBM.VisibleUnit.GAUSSIAN)
                         .nIn(d.numInputs()).nOut(nOut)
                         .weightInit(WeightInit.VI)
-                        .lossFunction(LossFunctions.LossFunction.RMSE_XENT)
+                        .lossFunction(LossFunctions.LossFunction.KL_DIVERGENCE)
                         .build())
                 .optimizationAlgo(OptimizationAlgorithm.STOCHASTIC_GRADIENT_DESCENT)
                 .learningRate(1e-3f)
@@ -105,7 +105,7 @@ public void testIrisGaussianHidden() {
                 .layer(new org.deeplearning4j.nn.conf.layers.RBM.Builder(
                         org.deeplearning4j.nn.conf.layers.RBM.HiddenUnit.GAUSSIAN, org.deeplearning4j.nn.conf.layers.RBM.VisibleUnit.GAUSSIAN)
                         .nIn(d.numInputs()).nOut(3)
-                        .lossFunction(LossFunctions.LossFunction.RMSE_XENT).build())
+                        .lossFunction(LossFunctions.LossFunction.KL_DIVERGENCE).build())
                 .build();
 
         int numParams = conf.getLayer().initializer().numParams(conf,true);
@@ -127,7 +127,7 @@ public void testIris() {
                 .learningRate(1e-1f)
                 .layer(new org.deeplearning4j.nn.conf.layers.RBM.Builder(org.deeplearning4j.nn.conf.layers.RBM.HiddenUnit.RECTIFIED, org.deeplearning4j.nn.conf.layers.RBM.VisibleUnit.GAUSSIAN)
                         .nIn(d.numInputs()).nOut(3)
-                        .lossFunction(LossFunctions.LossFunction.RMSE_XENT).build())
+                        .lossFunction(LossFunctions.LossFunction.KL_DIVERGENCE).build())
                 .build();
 
         int numParams = conf.getLayer().initializer().numParams(conf,true);
@@ -157,7 +157,7 @@ public void testBasic() {
                 .learningRate(1e-1f)
                 .layer(new org.deeplearning4j.nn.conf.layers.RBM.Builder()
                         .nIn(6).nOut(4)
-                        .lossFunction(LossFunctions.LossFunction.RMSE_XENT).build())
+                        .lossFunction(LossFunctions.LossFunction.KL_DIVERGENCE).build())
                 .build();
 
         int numParams = conf.getLayer().initializer().numParams(conf,true);
@@ -208,7 +208,7 @@ public void testSetGetParams() {
                 .learningRate(1e-1f)
                 .layer(new org.deeplearning4j.nn.conf.layers.RBM.Builder()
                         .nIn(6).nOut(4)
-                        .lossFunction(LossFunctions.LossFunction.RMSE_XENT).build())
+                        .lossFunction(LossFunctions.LossFunction.KL_DIVERGENCE).build())
                 .build();
 
         int numParams = conf.getLayer().initializer().numParams(conf,true);
@@ -241,7 +241,7 @@ public void testCg() {
                 .learningRate(1e-1f)
                 .layer(new org.deeplearning4j.nn.conf.layers.RBM.Builder()
                         .nIn(6).nOut(4)
-                        .lossFunction(LossFunctions.LossFunction.RMSE_XENT).build())
+                        .lossFunction(LossFunctions.LossFunction.KL_DIVERGENCE).build())
                 .build();
 
         int numParams = conf.getLayer().initializer().numParams(conf,true);
@@ -276,7 +276,7 @@ public void testGradient() {
                 .learningRate(1e-1f)
                 .layer(new org.deeplearning4j.nn.conf.layers.RBM.Builder()
                         .nIn(6).nOut(4)
-                        .lossFunction(LossFunctions.LossFunction.RMSE_XENT).build())
+                        .lossFunction(LossFunctions.LossFunction.KL_DIVERGENCE).build())
                 .build();
 
         int numParams = conf.getLayer().initializer().numParams(conf,true);

deeplearning4j-core/src/test/java/org/deeplearning4j/nn/multilayer/MultiLayerTest.java

@@ -184,7 +184,7 @@ public void testDbn() throws Exception {
                         .nIn(4).nOut(3)
                         .weightInit(WeightInit.DISTRIBUTION).dist(new UniformDistribution(0, 1))
                         .activation("tanh")
-                        .lossFunction(LossFunctions.LossFunction.RMSE_XENT).build())
+                        .lossFunction(LossFunctions.LossFunction.KL_DIVERGENCE).build())
                 .layer(1, new org.deeplearning4j.nn.conf.layers.OutputLayer.Builder(LossFunctions.LossFunction.MCXENT)
                         .nIn(3).nOut(3)
                         .weightInit(WeightInit.DISTRIBUTION).dist(new UniformDistribution(0, 1))
@@ -655,6 +655,7 @@ public void testPredict() throws Exception{
                 .layer(0, new DenseLayer.Builder().nIn(400).nOut(50).activation("relu").build())
                 .layer(1, new OutputLayer.Builder(LossFunctions.LossFunction.MCXENT).activation("softmax").nIn(50).nOut(10).build())
                 .pretrain(false).backprop(true)
+                .setInputType(InputType.convolutional(20,20,1))
                 .build();
 
         MultiLayerNetwork net = new MultiLayerNetwork(conf);
@@ -698,6 +699,7 @@ public void testOutput() throws Exception{
                 .layer(0, new DenseLayer.Builder().nIn(400).nOut(50).activation("relu").build())
                 .layer(1, new OutputLayer.Builder(LossFunctions.LossFunction.MCXENT).activation("softmax").nIn(50).nOut(10).build())
                 .pretrain(false).backprop(true)
+                .setInputType(InputType.convolutional(20,20,1))
                 .build();
 
         MultiLayerNetwork net = new MultiLayerNetwork(conf);