#6520 Fix setLearningRate(double) for the no updater state (SGD, etc)…

… case

deeplearning4j/deeplearning4j-core/src/test/java/org/deeplearning4j/nn/misc/TestLrChanges.java

@@ -33,6 +33,7 @@
 import org.nd4j.linalg.learning.config.Adam;
 import org.nd4j.linalg.learning.config.NoOp;
 import org.nd4j.linalg.learning.config.RmsProp;
+import org.nd4j.linalg.learning.config.Sgd;
 import org.nd4j.linalg.lossfunctions.LossFunctions;
 import org.nd4j.linalg.schedule.ExponentialSchedule;
 import org.nd4j.linalg.schedule.ScheduleType;
@@ -140,6 +141,35 @@ public void testChangeLrMLN(){
         assertEquals(net.getUpdater().getStateViewArray(), net3.getUpdater().getStateViewArray());
     }
 
+    @Test
+    public void testChangeLSGD() {
+        //Simple test for no updater nets
+        MultiLayerConfiguration conf = new NeuralNetConfiguration.Builder()
+                .activation(Activation.TANH)
+                .seed(12345)
+                .updater(new Sgd(0.1))
+                .list()
+                .layer(new DenseLayer.Builder().nIn(10).nOut(10).build())
+                .layer(new DenseLayer.Builder().nIn(10).nOut(10).build())
+                .layer(new OutputLayer.Builder().nIn(10).nOut(10).lossFunction(LossFunctions.LossFunction.MSE).build())
+                .build();
+
+        MultiLayerNetwork net = new MultiLayerNetwork(conf);
+        net.init();
+        net.setLearningRate(1.0);
+        net.setLearningRate(1, 0.5);
+        assertEquals(1.0, net.getLearningRate(0), 0.0);
+        assertEquals(0.5, net.getLearningRate(1), 0.0);
+
+
+        ComputationGraph cg = net.toComputationGraph();
+        cg.setLearningRate(2.0);
+        cg.setLearningRate("1", 2.5);
+        assertEquals(2.0, cg.getLearningRate("0"), 0.0);
+        assertEquals(2.5, cg.getLearningRate("1"), 0.0);
+
+    }
+
     @Test
     public void testChangeLrMLNSchedule(){
         //First: Set LR for a *single* layer and compare vs. equivalent net config

deeplearning4j/deeplearning4j-nn/src/main/java/org/deeplearning4j/nn/conf/NeuralNetConfiguration.java

@@ -1140,10 +1140,10 @@ private void copyConfigToLayer(String layerName, Layer layer) {
 
                 //Configure updaters:
                 if(iUpdater != null && bLayer.getIUpdater() == null){
-                    bLayer.setIUpdater(iUpdater);
+                    bLayer.setIUpdater(iUpdater.clone());   //Clone the updater to avoid shared instances - in case of setLearningRate calls later
                 }
                 if(biasUpdater != null && bLayer.getBiasUpdater() == null){
-                    bLayer.setBiasUpdater(biasUpdater);
+                    bLayer.setBiasUpdater(biasUpdater.clone());     //Clone the updater to avoid shared instances - in case of setLearningRate calls later
                 }
 
                 if(bLayer.getIUpdater() == null && iUpdater == null && bLayer.initializer().numParams(bLayer) > 0){

deeplearning4j/deeplearning4j-nn/src/main/java/org/deeplearning4j/nn/graph/ComputationGraph.java

@@ -306,7 +306,8 @@ public Layer[] getLayers() {
      * Get a given layer by name.
      */
     public Layer getLayer(String name) {
-        return verticesMap.get(name).getLayer(); //TODO checks
+        Preconditions.checkState(verticesMap.containsKey(name), "Layer with name %s does not exist in the network", name);
+        return verticesMap.get(name).getLayer();
     }
 
     /**

deeplearning4j/deeplearning4j-nn/src/main/java/org/deeplearning4j/nn/multilayer/MultiLayerNetwork.java

@@ -3752,7 +3752,7 @@ public void setLearningRate(int layerNumber, ISchedule newLr){
      * @return Learning rate for the specified layer, or null
      */
     public Double getLearningRate(int layerNumber){
-        return NetworkUtils.getLearningRate(this, layerIndex);
+        return NetworkUtils.getLearningRate(this, layerNumber);
     }
 
     /**

deeplearning4j/deeplearning4j-nn/src/main/java/org/deeplearning4j/nn/updater/BaseMultiLayerUpdater.java

@@ -211,6 +211,13 @@ public BaseMultiLayerUpdater(T network, INDArray updaterState) {
      * @param viewArray The new updater state
      */
     public void setStateViewArray(INDArray viewArray) {
+        if(this.updaterStateViewArray == null){
+            if(viewArray == null)
+                return; //No op - for example, SGD and NoOp updater - i.e., no stored state
+            else {
+                throw new IllegalStateException("Attempting to set updater state view array with null value");
+            }
+        }
         if (this.updaterStateViewArray.length() != viewArray.length())
             throw new IllegalStateException("Invalid input: view arrays differ in length. " + "Expected length "
                             + this.updaterStateViewArray.length() + ", got length " + viewArray.length());

nd4j/nd4j-backends/nd4j-api-parent/nd4j-api/src/main/java/org/nd4j/linalg/api/ndarray/BaseNDArray.java

@@ -1405,7 +1405,6 @@ public INDArray assign(final INDArray arr) {
                         " with x.shape=%ndShape and y.shape=%ndShape", this, arr );
         Nd4j.getExecutioner().exec(new org.nd4j.linalg.api.ops.impl.transforms.Set(this, arr, this, length()));
         return this;
-
     }
 
     @Override

nd4j/nd4j-backends/nd4j-api-parent/nd4j-api/src/main/java/org/nd4j/linalg/factory/Nd4j.java

@@ -959,11 +959,12 @@ public static INDArray gemm(INDArray a,
      * depending on setting of arguments transposeA and transposeB.<br>
      * Note that matrix c MUST be fortran order, have zero offset and have c.data().length == c.length().
      * i.e., the result array must not be a view. An exception will be thrown otherwise.<br>
+     * (Note: some views are allowed, if and only if they have f order and are contiguous in the buffer other than an
+     * offset. Put another way, they must be f order and have strides identical to a non-view/default array of the same shape)<br>
      * Don't use this unless you know about level 3 blas and NDArray storage orders.
      * @param a First matrix
      * @param b Second matrix
-     * @param c result matrix. Used in calculation (assuming beta != 0) and result is stored in this. f order,
-     *          zero offset and length == data.length only
+     * @param c result matrix. Used in calculation (assuming beta != 0) and result is stored in this. f order, and not a view only
      * @param transposeA if true: transpose matrix a before mmul
      * @param transposeB if true: transpose matrix b before mmul
      * @return result, i.e., matrix c is returned for convenience
@@ -975,7 +976,7 @@ public static INDArray gemm(INDArray a,
                                 boolean transposeB,
                                 double alpha,
                                 double beta) {
-        Preconditions.checkState(c.length() == 1 || (c.ordering() == 'f' && Shape.hasDefaultStridesForShape(c) && !c.isView()),
+        Preconditions.checkState(c.length() == 1 || (c.ordering() == 'f' && Shape.hasDefaultStridesForShape(c)),    //Note: some views
                 "C (result) array is not F order or is a view. Nd4j.gemm requires the result array to be F order " +
                         "and not a view. C (result) array: [%ndSInfo]", c);
         getBlasWrapper().level3().gemm(a, b, c, transposeA, transposeB, alpha, beta);