Mark deprecated functions with @keywords internal

To prevent them showing up in the package documentation index.

Author: t-kalinowski

NEWS.md

@@ -31,7 +31,9 @@
 - Fixed error raised when `k_random_binomial()` was passed a non-floating dtype.
 - Added `k_random_bernouli()` as an alias for `k_random_binomial()`.
 
-- `image_laod()` gains a `color_mode` argument.
+- `image_load()` gains a `color_mode` argument.
+
+- Deprecated functions are no longer included in the package documentation index.
 
 # keras 2.7.0
 

R/backend.R

@@ -1595,7 +1595,9 @@ k_log <- function(x) {
 }
 
 
-#' Computes log(sum(exp(elements across dimensions of a tensor))).
+#' (Deprecated) Computes log(sum(exp(elements across dimensions of a tensor))).
+#'
+#' This funciton is deprecated. Please use `tensorflow::tf$reduce_logsumexp()`.
 #'
 #' This function is more numerically stable than log(sum(exp(x))). It avoids
 #' overflows caused by taking the exp of large inputs and underflows caused by
@@ -1611,6 +1613,7 @@ k_log <- function(x) {
 #'
 #' @template roxlate-keras-backend
 #'
+#' @keywords internal
 #' @export
 k_logsumexp <- function(x, axis = NULL, keepdims = FALSE) {
 

R/callbacks.R

@@ -495,7 +495,7 @@ callback_lambda <- function(on_epoch_begin = NULL, on_epoch_end = NULL,
 #' (Deprecated) Base R6 class for Keras callbacks
 #'
 #' New custom callbacks implemented as R6 classes are encouraged to inherit from
-#' keras$callbacks$Callback directly.
+#' `keras$callbacks$Callback` directly.
 #'
 #' @docType class
 #'
@@ -526,7 +526,7 @@ callback_lambda <- function(on_epoch_begin = NULL, on_epoch_end = NULL,
 #' - `on_batch_end`: logs include `loss`, and optionally `acc` (if accuracy monitoring is enabled).
 #'
 #' @return [KerasCallback].
-#'
+#' @keywords internal
 #' @examples
 #' \dontrun{
 #' library(keras)

R/constraints.R

@@ -122,6 +122,7 @@ constraint_minmaxnorm <- function(min_value = 0.0, max_value = 1.0, rate = 1.0,
 #'
 #' @seealso [constraints]
 #'
+#' @keywords internal
 #' @export
 KerasConstraint <- R6::R6Class("KerasConstraint",
   public = list(

R/layer-custom.R

@@ -6,7 +6,7 @@
 #'
 #' @docType class
 #'
-#' @format An [R6Class] generator object #'
+#' @format An [R6Class] generator object
 #' @section Methods: \describe{ \item{\code{build(input_shape)}}{Creates the
 #'   layer weights (must be implemented by all layers that have weights)}
 #'   \item{\code{call(inputs,mask)}}{Call the layer on an input tensor.}
@@ -18,6 +18,8 @@
 #'
 #' @return [KerasLayer].
 #'
+#' @keywords internal
+#'
 #' @export
 KerasLayer <- R6Class("KerasLayer",
 

R/layers-recurrent.R

@@ -165,7 +165,7 @@ function(object,
 
 
 
-#' Fast GRU implementation backed by [CuDNN](https://developer.nvidia.com/cudnn).
+#' (Deprecated) Fast GRU implementation backed by [CuDNN](https://developer.nvidia.com/cudnn).
 #'
 #' Can only be run on GPU, with the TensorFlow backend.
 #'
@@ -184,6 +184,7 @@ function(object,
 #'   Application of Dropout in Recurrent Neural
 #'   Networks](https://arxiv.org/abs/1512.05287)
 #'
+#' @keywords internal
 #' @export
 layer_cudnn_gru <- function(object, units,
                             kernel_initializer = "glorot_uniform", recurrent_initializer = "orthogonal", bias_initializer = "zeros",
@@ -282,7 +283,7 @@ function(object,
   create_layer(keras$layers$LSTM, object, args)
 }
 
-#' Fast LSTM implementation backed by [CuDNN](https://developer.nvidia.com/cudnn).
+#' (Deprecated) Fast LSTM implementation backed by [CuDNN](https://developer.nvidia.com/cudnn).
 #'
 #' Can only be run on GPU, with the TensorFlow backend.
 #'
@@ -296,6 +297,7 @@ function(object,
 #'
 #' @family recurrent layers
 #'
+#' @keywords internal
 #' @export
 layer_cudnn_lstm <- function(object, units,
                              kernel_initializer = "glorot_uniform",  recurrent_initializer = "orthogonal",

R/losses.R

@@ -366,6 +366,7 @@ c("squared_hinge", "SquaredHinge")
 #'
 #' @param ... passed on to [loss_cosine_similarity()]
 #'
+#' @keywords internal
 #' @export
 loss_cosine_proximity <- function(...) {
   warning("loss_cosine_proximity is deprecated and will be removed in a future version.",

R/metrics.R

@@ -1324,6 +1324,7 @@ custom_metric <- function(name, metric_fn) {
 #' `tensorflow::tf$compat$v1$keras$metrics$cosine_proximity()`
 #'
 #' @inheritParams metric-or-Metric
+#' @keywords internal
 #' @export
 metric_cosine_proximity <- function(y_true, y_pred) {
   warning(

R/model-persistence.R

@@ -473,7 +473,7 @@ export_savedmodel.keras.engine.training.Model <- function(
   invisible(export_dir_base)
 }
 
-#' Export to Saved Model format
+#' (Deprecated) Export to Saved Model format
 #'
 #' @param model A Keras model to be saved. If the model is subclassed, the flag
 #'   `serving_only` must be set to `TRUE`.
@@ -493,6 +493,7 @@ export_savedmodel.keras.engine.training.Model <- function(
 #' @return Invisibly returns the `saved_model_path`.
 #' @family saved_model
 #'
+#' @keywords internal
 #' @export
 model_to_saved_model <- function(model, saved_model_path, custom_objects = NULL,
                                  as_text = FALSE, input_signature = NULL,

R/model.R

@@ -145,7 +145,7 @@ sequential_model_input_layer <- function(input_shape = NULL,
 
 
 
-#' Replicates a model on different GPUs.
+#' (Deprecated) Replicates a model on different GPUs.
 #'
 #' @param model A Keras model instance. To avoid OOM errors,
 #'   this model could have been built on CPU, for instance
@@ -247,6 +247,7 @@ sequential_model_input_layer <- function(input_shape = NULL,
 #'     classes = num_classes
 #' })
 #' ```
+#' @keywords internal
 #' @export
 multi_gpu_model <- function(model, gpus = NULL, cpu_merge = TRUE, cpu_relocation = FALSE) {
 
@@ -933,6 +934,7 @@ predict.keras.engine.training.Model <- function(object, x, batch_size=NULL, verb
 #'
 #' @family model functions
 #'
+#' @keywords internal
 #' @export
 predict_proba <- function(object, x, batch_size = NULL, verbose = 0, steps = NULL) {
   warning("`predict_proba()` is deprecated and was removed from tensorflow in version 2.6, ",
@@ -957,6 +959,7 @@ predict_proba <- function(object, x, batch_size = NULL, verbose = 0, steps = NUL
 }
 
 #' @rdname predict_proba
+#' @keywords internal
 #' @export
 predict_classes <- function(object, x, batch_size = NULL, verbose = 0, steps = NULL) {
   warning(
@@ -1048,7 +1051,7 @@ test_on_batch <- function(object, x, y, sample_weight = NULL) {
 
 
 
-#' Fits the model on data yielded batch-by-batch by a generator.
+#' (Deprecated) Fits the model on data yielded batch-by-batch by a generator.
 #'
 #' The generator is run in parallel to the model, for efficiency. For instance,
 #' this allows you to do real-time data augmentation on images on CPU in
@@ -1111,6 +1114,7 @@ test_on_batch <- function(object, x, y, sample_weight = NULL) {
 #'
 #' @family model functions
 #'
+#' @keywords internal
 #' @export
 fit_generator <- function(object, generator, steps_per_epoch, epochs = 1,
                           verbose=getOption("keras.fit_verbose", default = 1), callbacks = NULL,
@@ -1155,7 +1159,7 @@ fit_generator <- function(object, generator, steps_per_epoch, epochs = 1,
   do.call(fit, args)
 }
 
-#' Evaluates the model on a data generator.
+#' (Deprecated) Evaluates the model on a data generator.
 #'
 #' The generator should return the same kind of data as accepted by
 #' `test_on_batch()`.
@@ -1173,6 +1177,7 @@ fit_generator <- function(object, generator, steps_per_epoch, epochs = 1,
 #'
 #' @family model functions
 #'
+#' @keywords internal
 #' @export
 evaluate_generator <- function(object, generator, steps, max_queue_size = 10, workers = 1,
                                callbacks = NULL) {
@@ -1197,7 +1202,7 @@ evaluate_generator <- function(object, generator, steps, max_queue_size = 10, wo
 }
 
 
-#' Generates predictions for the input samples from a data generator.
+#' (Deprecated) Generates predictions for the input samples from a data generator.
 #'
 #' The generator should return the same kind of data as accepted by
 #' `predict_on_batch()`.
@@ -1218,6 +1223,7 @@ evaluate_generator <- function(object, generator, steps, max_queue_size = 10, wo
 #'
 #' @family model functions
 #'
+#' @keywords internal
 #' @export
 predict_generator <- function(object, generator, steps, max_queue_size = 10, workers = 1, verbose = 0,
                               callbacks = NULL) {