Metrics Phase 1

tensorflow-framework/src/main/java/org/tensorflow/framework/metrics/BinaryCrossentropy.java

@@ -41,7 +41,7 @@ public class BinaryCrossentropy<U extends TNumber, T extends TNumber>
    *
    * @param tf the TensorFlow Ops
    * @param name the name of this metric, if null then metric name is {@link Class#getSimpleName()}.
-   * @param fromLogits Whether to interpret predictions as a tensor of logit values or not.
+   * @param fromLogits Whether to interpret predictions as a tensor of logit values as opposed to a probability distribution.
    * @param labelSmoothing value used to smooth labels, When 0, no smoothing occurs. When &gt; 0,
    *     compute the loss between the predicted labels and a smoothed version of the true labels,
    *     where the smoothing squeezes the labels towards 0.5. Larger values of label_smoothing

tensorflow-framework/src/main/java/org/tensorflow/framework/metrics/CategoricalCrossentropy.java

@@ -48,7 +48,7 @@ public class CategoricalCrossentropy<U extends TNumber, T extends TNumber>
    *
    * @param tf the TensorFlow Ops
    * @param name the name of this metric, if null then metric name is {@link Class#getSimpleName()}.
-   * @param fromLogits Whether to interpret predictions as a tensor of logit values or not.
+   * @param fromLogits Whether to interpret predictions as a tensor of logit values oras opposed to a probability distribution.
    * @param labelSmoothing value used to smooth labels, When &gt; 0, label values are smoothed,
    *     meaning the confidence on label values are relaxed. e.g. <code>labelSmoothing=0.2</code>
    *     means that we will use a value of <code>0.1</code> for label <code>0</code> and <code>0.9
@@ -68,7 +68,7 @@ public CategoricalCrossentropy(
    *
    * @param tf the TensorFlow Ops
    * @param name the name of this metric, if null then metric name is {@link Class#getSimpleName()}.
-   * @param fromLogits Whether to interpret predictions as a tensor of logit values or not.
+   * @param fromLogits Whether to interpret predictions as a tensor of logit values as opposed to a probability distribution.
    * @param labelSmoothing value used to smooth labels, When &gt; 0, label values are smoothed,
    *     meaning the confidence on label values are relaxed. e.g. <code>labelSmoothing=0.2</code>
    *     means that we will use a value of <code>0.1</code> for label <code>0</code> and <code>0.9

tensorflow-framework/src/main/java/org/tensorflow/framework/metrics/Metrics.java

@@ -64,19 +64,19 @@ public static <T extends TNumber, U extends TNumber> Operand<T> topKCategoricalA
    * @param tf the TensorFlow Ops
    * @param labels The ground truth values.
    * @param predictions The prediction values.
-   * @param axis The dimension along which the cosine similarity is computed.
+   * @param axes The dimensions along which the cosine similarity is computed.
    * @param <U> the data type for the labels
    * @param <T> the data type for the predictions and result
    * @return Cosine similarity value.
    */
   public static <T extends TNumber, U extends TNumber> Operand<T> cosineProximity(
-      Ops tf, Operand<U> labels, Operand<T> predictions, int[] axis) {
+      Ops tf, Operand<U> labels, Operand<T> predictions, int[] axes) {
     Operand<T> labelsNorm = CastHelper.cast(tf, labels, predictions.type());
-    labelsNorm = l2Normalize(tf, labelsNorm, axis);
+    labelsNorm = l2Normalize(tf, labelsNorm, axes);
 
-    Operand<T> predictionsNorm = l2Normalize(tf, predictions, axis);
+    Operand<T> predictionsNorm = l2Normalize(tf, predictions, axes);
     Operand<T> mathMul = tf.math.mul(labelsNorm, predictionsNorm);
-    return tf.reduceSum(mathMul, tf.constant(axis), ReduceSum.keepDims(Boolean.FALSE));
+    return tf.reduceSum(mathMul, tf.constant(axes), ReduceSum.keepDims(Boolean.FALSE));
   }
 
   /**

tensorflow-framework/src/main/java/org/tensorflow/framework/metrics/SparseCategoricalCrossentropy.java

@@ -32,30 +32,30 @@ public class SparseCategoricalCrossentropy<U extends TNumber, T extends TNumber>
     extends MeanMetricWrapper<U, T> implements LossMetric<T> {
 
   private final boolean fromLogits;
-  private final int axes;
+  private final int axis;
 
   /**
    * Creates a SparseCategoricalCrossentropy metric
    *
    * @param tf the TensorFlow Ops
    * @param name the name of this metric, if null then metric name is {@link Class#getSimpleName()}.
-   * @param fromLogits Whether to interpret predictions as a tensor of logit values or not.
-   * @param axes The dimension along which the entropy is computed.
+   * @param fromLogits Whether to interpret predictions as a tensor of logit values as opposed to a probability distribution.
+   * @param axis The dimension along which the entropy is computed.
    * @param seed the seed for random number generation. An initializer created with a given seed
    *     will always produce the same random tensor for a given shape and data type.
    * @param type the type for the variables and result
    */
   public SparseCategoricalCrossentropy(
-      Ops tf, String name, boolean fromLogits, int axes, long seed, Class<T> type) {
+      Ops tf, String name, boolean fromLogits, int axis, long seed, Class<T> type) {
     super(tf, name, seed, type);
     setLoss(this);
     this.fromLogits = fromLogits;
-    this.axes = axes;
+    this.axis = axis;
   }
 
   /** {@inheritDoc} */
   @Override
   public <V extends TNumber> Operand<T> call(Operand<V> labels, Operand<T> predictions) {
-    return Losses.sparseCategoricalCrossentropy(getTF(), labels, predictions, fromLogits, axes);
+    return Losses.sparseCategoricalCrossentropy(getTF(), labels, predictions, fromLogits, axis);
   }
 }

tensorflow-framework/src/main/java/org/tensorflow/framework/metrics/impl/MeanMetricWrapper.java

@@ -85,7 +85,7 @@ protected void setLoss(LossMetric<T> loss) {
    *     [batch_size, d0, .. dN-1] (or can be broadcasted to this shape), then each loss element of
    *     predictions is scaled by the corresponding value of sampleWeights. (Note on dN-1: all loss
    *     functions reduce by 1 dimension, usually axis=-1.)
-   * @param <V> the datatype of the predictions
+   * @param <V> the datatype of the labels
    * @param <S> the data type for sampleWeights
    * @return a List of control operations that updates the Mean state variables.
    */