R/layer-wrappers.R

In dfalbel/keras: R Interface to 'Keras'

#' Apply a layer to every temporal slice of an input.
#' 
#' The input should be at least 3D, and the dimension of index one will be
#' considered to be the temporal dimension. 
#' 
#' Consider a batch of 32 samples,  where each sample is a sequence of 10 vectors of 16 dimensions. The batch
#' input shape of the layer is then `(32, 10, 16)`, and the `input_shape`, not
#' including the samples dimension, is `(10, 16)`. You can then use
#' `time_distributed` to apply a `layer_dense` to each of the 10 timesteps,
#' independently.
#' 
#' @inheritParams layer_dense
#' 
#' @param layer A layer instance.
#' 
#' @family layer wrappers
#'   
#' @export
time_distributed <- function(object, layer, input_shape = NULL,
                             batch_input_shape = NULL, batch_size = NULL, dtype = NULL, 
                             name = NULL, trainable = NULL, weights = NULL) {

  create_layer(keras$layers$TimeDistributed, object, list(
    layer = layer,
    input_shape = normalize_shape(input_shape),
    batch_input_shape = normalize_shape(batch_input_shape),
    batch_size = as_nullable_integer(batch_size),
    dtype = dtype,
    name = name,
    trainable = trainable,
    weights = weights
  ))
  
}


#' Bidirectional wrapper for RNNs.
#' 
#' @inheritParams layer_dense
#' 
#' @param layer Recurrent instance.
#' @param merge_mode Mode by which outputs of the forward and backward RNNs will
#'   be combined. One of 'sum', 'mul', 'concat', 'ave', NULL. If NULL, the
#'   outputs will not be combined, they will be returned as a list.
#'   
#' @family layer wrappers
#'   
#' @export
bidirectional <- function(object, layer, merge_mode = "concat", input_shape = NULL,
                          batch_input_shape = NULL, batch_size = NULL, dtype = NULL, 
                          name = NULL, trainable = NULL, weights = NULL) {
  
  create_layer(keras$layers$Bidirectional, object, list(
    layer = layer,
    merge_mode = merge_mode,
    input_shape = normalize_shape(input_shape),
    batch_input_shape = normalize_shape(batch_input_shape),
    batch_size = as_nullable_integer(batch_size),
    dtype = dtype,
    name = name,
    trainable = trainable,
    weights = weights
  ))
  
}