Final fix

deeplearning4j · Nov 30, 2017 · e68cdd5 · e68cdd5
1 parent e94bc45
commit e68cdd5
Show file tree

Hide file tree

Showing 2 changed files with 124 additions and 5 deletions.
diff --git a/deeplearning4j-core/src/test/java/org/deeplearning4j/Debug.java b/deeplearning4j-core/src/test/java/org/deeplearning4j/Debug.java
@@ -0,0 +1,116 @@
+package org.deeplearning4j;
+
+import org.deeplearning4j.nn.conf.ConvolutionMode;
+import org.deeplearning4j.nn.conf.MultiLayerConfiguration;
+import org.deeplearning4j.nn.conf.NeuralNetConfiguration;
+import org.deeplearning4j.nn.conf.WorkspaceMode;
+import org.deeplearning4j.nn.conf.layers.ConvolutionLayer;
+import org.deeplearning4j.nn.conf.layers.OutputLayer;
+import org.deeplearning4j.nn.conf.layers.PoolingType;
+import org.deeplearning4j.nn.multilayer.MultiLayerNetwork;
+import org.deeplearning4j.nn.weights.WeightInit;
+import org.junit.Test;
+import org.nd4j.linalg.activations.Activation;
+import org.nd4j.linalg.api.ndarray.INDArray;
+import org.nd4j.linalg.api.ops.executioner.OpExecutioner;
+import org.nd4j.linalg.factory.Nd4j;
+import org.nd4j.linalg.lossfunctions.LossFunctions;
+
+import java.util.Arrays;
+
+import static org.junit.Assert.assertEquals;
+
+public class Debug {
+
+    @Test
+    public void debug6() {
+        Nd4j.getRandom().setSeed(12345);
+
+        Nd4j.getExecutioner().setProfilingMode(OpExecutioner.ProfilingMode.SCOPE_PANIC);
+        int depthIn = 2;
+        int depthOut = 2;
+        int nOut = 2;
+        int width = 3;
+
+        PoolingType pt = PoolingType.SUM;
+        MultiLayerConfiguration conf = new NeuralNetConfiguration.Builder().weightInit(WeightInit.XAVIER)
+                .convolutionMode(ConvolutionMode.Same).seed(12345L).list()
+                .layer(0, new ConvolutionLayer.Builder().nIn(depthIn).nOut(depthOut).kernelSize(2, width)
+                        .hasBias(false)
+                        .stride(1, width).activation(Activation.TANH).build())
+                .layer(1, new org.deeplearning4j.nn.conf.layers.GlobalPoolingLayer.Builder().poolingType(pt)
+                        .build())
+                .layer(2, new OutputLayer.Builder(LossFunctions.LossFunction.MCXENT)
+                        .activation(Activation.SOFTMAX).nIn(depthOut).nOut(nOut).build())
+                .pretrain(false).backprop(true).build();
+
+        MultiLayerNetwork net = new MultiLayerNetwork(conf.clone());
+        net.init();
+        net.getLayerWiseConfigurations().setTrainingWorkspaceMode(WorkspaceMode.SEPARATE);
+        net.getLayerWiseConfigurations().setInferenceWorkspaceMode(WorkspaceMode.SEPARATE);
+
+//        //No workspace: passes
+//        net.getLayerWiseConfigurations().setTrainingWorkspaceMode(WorkspaceMode.NONE);
+//        net.getLayerWiseConfigurations().setInferenceWorkspaceMode(WorkspaceMode.NONE);
+
+//        //Single workspace: passes
+//        net.getLayerWiseConfigurations().setTrainingWorkspaceMode(WorkspaceMode.SINGLE);
+//        net.getLayerWiseConfigurations().setInferenceWorkspaceMode(WorkspaceMode.SINGLE);
+
+        INDArray in = Nd4j.rand(new int[]{1, 2, 5, 3});
+        INDArray out1 = net.output(in);
+        INDArray out2 = net.output(in);
+
+        System.out.println(Arrays.toString(out1.dup().data().asFloat()));
+        System.out.println(Arrays.toString(out2.dup().data().asFloat()));
+
+        assertEquals(out1, out2);
+
+    }
+
+    @Test
+    public void debug7() {
+        Nd4j.getRandom().setSeed(12345);
+
+        Nd4j.getExecutioner().setProfilingMode(OpExecutioner.ProfilingMode.SCOPE_PANIC);
+        int depthIn = 2;
+        int depthOut = 2;
+        int nOut = 2;
+        int width = 3;
+
+        PoolingType pt = PoolingType.SUM;
+        MultiLayerConfiguration conf = new NeuralNetConfiguration.Builder().weightInit(WeightInit.XAVIER)
+                .convolutionMode(ConvolutionMode.Same).seed(12345L).list()
+//                .layer(0, new ConvolutionLayer.Builder().nIn(depthIn).nOut(depthOut).kernelSize(2, width)
+//                        .stride(1, width).activation(Activation.TANH).build())
+                .layer(0, new org.deeplearning4j.nn.conf.layers.GlobalPoolingLayer.Builder().poolingType(pt)
+                        .build())
+                .layer(1, new OutputLayer.Builder(LossFunctions.LossFunction.MCXENT)
+                        .activation(Activation.SOFTMAX).nIn(depthOut).nOut(nOut).build())
+                .pretrain(false).backprop(true).build();
+
+        MultiLayerNetwork net = new MultiLayerNetwork(conf.clone());
+        net.init();
+        net.getLayerWiseConfigurations().setTrainingWorkspaceMode(WorkspaceMode.SEPARATE);
+        net.getLayerWiseConfigurations().setInferenceWorkspaceMode(WorkspaceMode.SEPARATE);
+
+//        //No workspace: passes
+//        net.getLayerWiseConfigurations().setTrainingWorkspaceMode(WorkspaceMode.NONE);
+//        net.getLayerWiseConfigurations().setInferenceWorkspaceMode(WorkspaceMode.NONE);
+
+//        //Single workspace: passes
+//        net.getLayerWiseConfigurations().setTrainingWorkspaceMode(WorkspaceMode.SINGLE);
+//        net.getLayerWiseConfigurations().setInferenceWorkspaceMode(WorkspaceMode.SINGLE);
+
+        INDArray in = Nd4j.rand(new int[]{1, 2, 5, 1});
+        INDArray out1 = net.output(in);
+        INDArray out2 = net.output(in);
+
+        System.out.println(Arrays.toString(out1.dup().data().asFloat()));
+        System.out.println(Arrays.toString(out2.dup().data().asFloat()));
+
+        assertEquals(out1, out2);
+
+    }
+
+}
diff --git a/deeplearning4j-nn/src/main/java/org/deeplearning4j/nn/multilayer/MultiLayerNetwork.java b/deeplearning4j-nn/src/main/java/org/deeplearning4j/nn/multilayer/MultiLayerNetwork.java
@@ -1927,17 +1927,20 @@ public INDArray output(INDArray input, boolean train) {
         WorkspaceMode cMode = layerWiseConfigurations.getTrainingWorkspaceMode();
         layerWiseConfigurations.setTrainingWorkspaceMode(layerWiseConfigurations.getInferenceWorkspaceMode());
         MemoryWorkspace workspace =
-                        layerWiseConfigurations.getTrainingWorkspaceMode() == WorkspaceMode.NONE ? new DummyWorkspace()
-                                        : Nd4j.getWorkspaceManager().getWorkspaceForCurrentThread(
-                                                        workspaceConfigurationExternal, workspaceExternal);
+                layerWiseConfigurations.getTrainingWorkspaceMode() == WorkspaceMode.NONE ? new DummyWorkspace()
+                        : Nd4j.getWorkspaceManager().getWorkspaceForCurrentThread(
+                        workspaceConfigurationExternal, workspaceExternal);
 
+        INDArray ret = null;
         try (MemoryWorkspace wsE = workspace.notifyScopeEntered()) {
-            INDArray ret = silentOutput(input, train).detach();
+            ret = silentOutput(input, train).detach();
 
             layerWiseConfigurations.setTrainingWorkspaceMode(cMode);
             clearLayersStates();    //Ensure INDArrays in layer input fields don't leak out of workspace (via .input() etc)
-            return ret;
         }
+
+        layerWiseConfigurations.setTrainingWorkspaceMode(cMode);
+        return ret;
     }
 
     protected INDArray silentOutput(INDArray input, boolean train) {