new metrics

- all the metric functions gain the ability to return a `keras$metrics$Metric`
   instance if called without `y_true` and `y_pred`

- Each metric function is now documented separately, with a common
  `?Metric` topic demonstrating example usage.

- New built in metrics:
        metric_true_negatives
        metric_true_positives
        metric_false_negatives
        metric_false_positives
        metric_specificity_at_sensitivity
        metric_sensitivity_at_specificity
        metric_precision
        metric_precision_at_recall
        metric_sum
        metric_recall
        metric_recall_at_precision
        metric_root_mean_squared_error
        metric_sparse_categorical_accuracy
        metric_mean_tensor
        metric_mean_wrapper
        metric_mean_iou
        metric_mean_relative_error
        metric_logcosh_error
        metric_mean
        metric_cosine_similarity
        metric_categorical_hinge
        metric_accuracy
        metric_auc

Author: t-kalinowski

NAMESPACE

@@ -446,22 +446,45 @@ export(loss_poisson)
 export(loss_sparse_categorical_crossentropy)
 export(loss_squared_hinge)
 export(make_sampling_table)
+export(metric_accuracy)
+export(metric_auc)
 export(metric_binary_accuracy)
 export(metric_binary_crossentropy)
 export(metric_categorical_accuracy)
 export(metric_categorical_crossentropy)
+export(metric_categorical_hinge)
 export(metric_cosine_proximity)
+export(metric_cosine_similarity)
+export(metric_false_negatives)
+export(metric_false_positives)
 export(metric_hinge)
 export(metric_kullback_leibler_divergence)
+export(metric_logcosh_error)
+export(metric_mean)
 export(metric_mean_absolute_error)
 export(metric_mean_absolute_percentage_error)
+export(metric_mean_iou)
+export(metric_mean_relative_error)
 export(metric_mean_squared_error)
 export(metric_mean_squared_logarithmic_error)
+export(metric_mean_tensor)
+export(metric_mean_wrapper)
 export(metric_poisson)
+export(metric_precision)
+export(metric_precision_at_recall)
+export(metric_recall)
+export(metric_recall_at_precision)
+export(metric_root_mean_squared_error)
+export(metric_sensitivity_at_specificity)
+export(metric_sparse_categorical_accuracy)
 export(metric_sparse_categorical_crossentropy)
 export(metric_sparse_top_k_categorical_accuracy)
+export(metric_specificity_at_sensitivity)
 export(metric_squared_hinge)
+export(metric_sum)
 export(metric_top_k_categorical_accuracy)
+export(metric_true_negatives)
+export(metric_true_positives)
 export(mobilenet_decode_predictions)
 export(mobilenet_load_model_hdf5)
 export(mobilenet_preprocess_input)

NEWS.md

@@ -63,6 +63,34 @@ New Features:
     (a `keras$losses$Loss` instance) if `y_true` and `y_pred` arguments are missing.
   - new builtin loss functions: loss_huber(), loss_kl_divergence()
 
+- metric functions:
+  - all the metric functions gain the ability to return a `keras$metrics$Metric` instance if called without `y_true` and `y_pred`
+  - Each metric function is now documented separately, with a common `?Metric` topic demonstrating example usage.
+  - New built in metrics:
+        metric_true_negatives
+        metric_true_positives
+        metric_false_negatives
+        metric_false_positives
+        metric_specificity_at_sensitivity
+        metric_sensitivity_at_specificity
+        metric_precision
+        metric_precision_at_recall
+        metric_sum
+        metric_recall
+        metric_recall_at_precision
+        metric_root_mean_squared_error
+        metric_sparse_categorical_accuracy
+        metric_mean_tensor
+        metric_mean_wrapper
+        metric_mean_iou
+        metric_mean_relative_error
+        metric_logcosh_error
+        metric_mean
+        metric_cosine_similarity
+        metric_categorical_hinge
+        metric_accuracy
+        metric_auc
+
 - keras_model_sequential() gains the ability to accept arguments that
   define input layer like `input_shape` and `dtype`.
   See `?keras_model_sequential` for details and examples.

R/losses.R

@@ -370,4 +370,3 @@ loss_cosine_proximity <- function(...) {
   loss_cosine_similarity(...)
 }
 attr(loss_cosine_proximity, "py_function_name") <- "cosine_proximity"
-