Fixed from Unit test results

Author: JimClarke5

tensorflow-core/tensorflow-core-api/src/gen/annotations/org/tensorflow/op/NnOps.java

@@ -1863,14 +1863,13 @@ public <T extends TNumber> Softmax<T> softmax(Operand<T> logits) {
    * @param logits Per-label activations, typically a linear output. These activation energies are
    *      interpreted as unnormalized log probabilities.
    * @param axis The class dimension. -1 is the last dimension.
-   * @param <U> the data type of the <code>logits</code>
    * @param <T> the number type of the operands
    * @return the softmax cross entropy loss. Its type is the same as <code>logits</code> and its
    *      shape is the same as <code>labels</code> except that it does not have the last dimension of
    *      <code>labels</code>.
    */
-  public <U extends TType, T extends TNumber> Operand<T> softmaxCrossEntropyWithLogits(
-      Operand<T> labels, Operand<U> logits, int axis) {
+  public <T extends TNumber, U extends TNumber> Operand<T> softmaxCrossEntropyWithLogits(
+      Operand<U> labels, Operand<T> logits, int axis) {
     return SoftmaxCrossEntropyWithLogits.softmaxCrossEntropyWithLogits(scope, labels, logits, axis);
   }
 

tensorflow-framework/src/main/java/org/tensorflow/framework/optimizers/Optimizer.java

@@ -71,6 +71,14 @@ protected Optimizer(Graph graph, String name) {
     this.globals = new ArrayList<>();
   }
 
+  /**
+   * Gets the Optimizer's Ops instance
+   * @return the Optimizer's Ops instance
+   */
+  public final Ops getTF() {
+    return tf;
+  }
+
   /**
    * Creates a name by combining a variable name and a slot name
    *

tensorflow-framework/src/main/java/org/tensorflow/framework/optimizers/Optimizers.java

@@ -15,7 +15,7 @@ public enum Optimizers {
   NADAM(Nadam::new),
   RMSPROP(RMSProp::new),
   MOMENTUM(Momentum::new),
-  GRADIENT_DESCENT(Momentum::new);
+  GRADIENT_DESCENT(GradientDescent::new);
 
   private final Function<Graph, Optimizer> creator;
 

tensorflow-framework/src/test/java/org/tensorflow/framework/optimizers/AdaGradDATest.java

@@ -55,9 +55,13 @@ public void testBasic() {
     float[] grads0Init = {0.1F, 0.2F};
     float[] grads1Init = {0.01F, 0.02F};
     try (TestSession session = TestSession.createTestSession(tfMode)) {
-      Ops tf = session.getTF();
       Graph graph = session.getGraph();
 
+      float learningRate = 3.0F;
+
+      AdaGradDA instance = new AdaGradDA(graph, learningRate);
+      Ops tf = instance.getTF();
+
       Shape shape0 = Shape.of(var0Init.length);
       Shape shape1 = Shape.of(var1Init.length);
       Variable<TFloat32> var0 = tf.withName("var0").variable(shape0, TFloat32.DTYPE);
@@ -70,13 +74,10 @@ public void testBasic() {
       Constant<TFloat32> grads1 = tf.constant(grads1Init);
 
       /* initialize the local variables */
-      /* initialize the local variables */
+
       session.run(var0Initializer);
       session.run(var1Initializer);
 
-      float learningRate = 3.0F;
-
-      AdaGrad instance = new AdaGrad(graph, learningRate);
 
       /* build the GradsAnvVars */
       List<Optimizer.GradAndVar<? extends TType>> gradsAndVars = new ArrayList<>();

tensorflow-framework/src/test/java/org/tensorflow/framework/optimizers/AdaGradTest.java

@@ -71,9 +71,12 @@ public void testBasic() {
     FloatNdArray accum1Np = NdArrays.vectorOf(accum1);
 
     try (TestSession session = TestSession.createTestSession(tfMode)) {
-      Ops tf = session.getTF();
       Graph graph = session.getGraph();
 
+      float learningRate = 3.0F;
+      AdaGrad instance = new AdaGrad(graph, learningRate, 0.1f);
+      Ops tf = instance.getTF();
+
       Shape shape0 = Shape.of(var0Init.length);
       Shape shape1 = Shape.of(var1Init.length);
       Variable<TFloat32> var0 = tf.withName("var0").variable(shape0, TFloat32.DTYPE);
@@ -85,9 +88,7 @@ public void testBasic() {
       Constant<TFloat32> grads0 = tf.constant(grads0Init);
       Constant<TFloat32> grads1 = tf.constant(grads1Init);
 
-      float learningRate = 3.0F;
 
-      AdaGrad instance = new AdaGrad(graph, learningRate);
 
       /* build the GradsAnvVars */
       List<Optimizer.GradAndVar<? extends TType>> gradsAndVars = new ArrayList<>();
@@ -116,15 +117,11 @@ public void testBasic() {
 
       for (int step = 0; step < numSteps; step++) {
         session.run(adaUpdate);
-
-        accum0Np = caclulateAccum(accum0Np, grads0Np);
-        var0Np = calculate(var0Np, accum0Np, grads0Np, learningRate);
-        session.evaluate(var0Np, var0);
-
-        accum1Np = caclulateAccum(accum1Np, grads1Np);
-        var1Np = calculate(var1Np, accum1Np, grads1Np, learningRate);
-        session.evaluate(var1Np, var1);
       }
+      float[] expected0 = {-1.6026098728179932f, -0.6026098728179932f};
+      session.evaluate(expected0, var0);
+      float[] expected1 = {2.715679168701172f, 3.715679168701172f};
+      session.evaluate(expected1, var1);
     }
   }
 

tensorflow-framework/src/test/java/org/tensorflow/framework/optimizers/AdamTest.java

@@ -69,10 +69,15 @@ public void testBasic() {
     float epsilon1 = 1e-3F;
 
     try (TestSession session = TestSession.createTestSession(tfMode)) {
-      Ops tf = session.getTF();
+      float learningRate = 0.001F;
+      float beta1 = 0.9F;
+      float beta2 = 0.999F;
       Graph graph = session.getGraph();
+
       session.setEpsilon(epsilon1);
 
+      Adam instance = new Adam(graph, learningRate);
+      Ops tf = instance.getTF();
       Shape shape0 = Shape.of(var0Init.length);
       Shape shape1 = Shape.of(var1Init.length);
       Variable<TFloat32> var0 = tf.withName("var0").variable(shape0, TFloat32.DTYPE);
@@ -88,16 +93,14 @@ public void testBasic() {
       session.run(var0Initializer);
       session.run(var1Initializer);
 
-      float learningRate = 0.001F;
-      float beta1 = 0.9F;
-      float beta2 = 0.999F;
+
 
       /* build the GradsAnvVars */
       List<Optimizer.GradAndVar<? extends TType>> gradsAndVars = new ArrayList<>();
       gradsAndVars.add(new Optimizer.GradAndVar<>(grads0.asOutput(), var0.asOutput()));
       gradsAndVars.add(new Optimizer.GradAndVar<>(grads1.asOutput(), var1.asOutput()));
 
-      Adam instance = new Adam(graph, learningRate);
+
 
       Op update = instance.applyGradients(gradsAndVars, "AdamTest");
 
@@ -139,7 +142,7 @@ public void testBasic() {
             session
                 .getGraphSession()
                 .runner()
-                .fetch("beta1Power")
+                .fetch("beta1_power")
                 .run()
                 .get(0)
                 .expect(TFloat32.DTYPE)) {
@@ -149,7 +152,7 @@ public void testBasic() {
             session
                 .getGraphSession()
                 .runner()
-                .fetch("beta2Power")
+                .fetch("beta2_power")
                 .run()
                 .get(0)
                 .expect(TFloat32.DTYPE)) {

tensorflow-framework/src/test/java/org/tensorflow/framework/optimizers/AdamaxTest.java

@@ -94,9 +94,11 @@ public void testBasic() {
     float epsilon1 = 1e-3F;
 
     try (TestSession session = TestSession.createTestSession(tfMode)) {
-      Ops tf = session.getTF();
       Graph graph = session.getGraph();
 
+      Adamax instance = new Adamax(graph);
+      Ops tf = instance.getTF();
+
       Shape shape0 = Shape.of(var0Init.length);
       Shape shape1 = Shape.of(var1Init.length);
       Variable<TFloat32> var0 = tf.withName("var0").variable(shape0, TFloat32.DTYPE);
@@ -112,7 +114,7 @@ public void testBasic() {
       session.run(var0Initializer);
       session.run(var1Initializer);
 
-      Adamax instance = new Adamax(graph);
+
       /* build the GradsAnvVars */
       List<GradAndVar<? extends TType>> gradsAndVars = new ArrayList<>();
       gradsAndVars.add(new GradAndVar<>(grads0.asOutput(), var0.asOutput()));
@@ -151,7 +153,7 @@ public void testBasic() {
             session
                 .getGraphSession()
                 .runner()
-                .fetch("beta1Power")
+                .fetch("beta1_power")
                 .run()
                 .get(0)
                 .expect(TFloat32.DTYPE)) {