## ----setup, include=FALSE-----------------------------------------------------
library(keras)
knitr::opts_chunk$set(eval = FALSE)

## ----eval= FALSE--------------------------------------------------------------
# # Replicates `model` on 8 GPUs.
# # This assumes that your machine has 8 available GPUs.
# parallel_model <- multi_gpu_model(model, gpus=8)
# parallel_model %>% compile(
#   loss = "categorical_crossentropy",
#   optimizer = "rmsprop"
# )
# 
# # This `fit` call will be distributed on 8 GPUs.
# # Since the batch size is 256, each GPU will process 32 samples.
# parallel_model %>% fit(x, y, epochs = 20, batch_size = 256)

## ----eval = FALSE-------------------------------------------------------------
# # Model where a shared LSTM is used to encode two different sequences in parallel
# input_a <- layer_input(shape = c(140, 256))
# input_b <- layer_input(shape = c(140, 256))
# 
# shared_lstm <- layer_lstm(units = 64)
# 
# # Process the first sequence on one GPU
# library(tensorflow)
# with(tf$device_scope("/gpu:0", {
#   encoded_a <- shared_lstm(tweet_a)
# }):
# 
# # Process the next sequence on another GPU
# with(tf$device_scope("/gpu:1", {
#   encoded_b <- shared_lstm(tweet_b)
# }):
# 
# # Concatenate results on CPU
# with(tf$device_scope("/cpu:0", {
#   merged_vector <- layer_concatenate(list(encoded_a, encoded_b))
# }):

## -----------------------------------------------------------------------------
# model %>%
#   layer_dense(units = 32, activation = 'relu', input_shape = c(784)) %>%
#   layer_dense(units = 10, activation = 'softmax')

## -----------------------------------------------------------------------------
# model <- model %>%
#   layer_dense(units = 32, activation = 'relu', input_shape = c(784)) %>%
#   layer_dense(units = 10, activation = 'softmax')

## -----------------------------------------------------------------------------
# save_model_hdf5(model, 'my_model.h5')
# model <- load_model_hdf5('my_model.h5')

## -----------------------------------------------------------------------------
# json_string <- model_to_json(model)
# yaml_string <- model_to_yaml(model)

## -----------------------------------------------------------------------------
# model <- model_from_json(json_string)
# model <- model_from_yaml(yaml_string)

## -----------------------------------------------------------------------------
# save_model_weights_hdf5('my_model_weights.h5')

## -----------------------------------------------------------------------------
# model %>% load_model_weights_hdf5('my_model_weights.h5')

## -----------------------------------------------------------------------------
# model %>% load_model_weights_hdf5('my_model_weights.h5', by_name = TRUE)

## -----------------------------------------------------------------------------
# # assuming the original model looks like this:
# #   model <- keras_model_sequential()
# #   model %>%
# #     layer_dense(units = 2, input_dim = 3, name = "dense 1") %>%
# #     layer_dense(units = 3, name = "dense_3") %>%
# #     ...
# #   save_model_weights(model, fname)
# 
# # new model
# model <- keras_model_sequential()
# model %>%
#   layer_dense(units = 2, input_dim = 3, name = "dense 1") %>%  # will be loaded
#   layer_dense(units = 3, name = "dense_3")                     # will not be loaded
# 
# # load weights from first model; will only affect the first layer, dense_1.
# load_model_weights(fname, by_name = TRUE)

## -----------------------------------------------------------------------------
# model <- ...  # create the original model
# 
# layer_name <- 'my_layer'
# intermediate_layer_model <- keras_model(inputs = model$input,
#                                         outputs = get_layer(model, layer_name)$output)
# intermediate_output <- predict(intermediate_layer_model, data)

## -----------------------------------------------------------------------------
# sampling_generator <- function(X_data, Y_data, batch_size) {
#   function() {
#     rows <- sample(1:nrow(X_data), batch_size, replace = TRUE)
#     list(X_data[rows,], Y_data[rows,])
#   }
# }
# 
# model %>%
#   fit_generator(sampling_generator(X_train, Y_train, batch_size = 128),
#                 steps_per_epoch = nrow(X_train) / 128, epochs = 10)
# 

## -----------------------------------------------------------------------------
# data_files_generator <- function(dir) {
# 
#   files < list.files(dir)
#   next_file <- 0
# 
#   function() {
# 
#     # move to the next file (note the <<- assignment operator)
#     next_file <<- next_file + 1
# 
#     # if we've exhausted all of the files then start again at the
#     # beginning of the list (keras generators need to yield
#     # data infinitely -- termination is controlled by the epochs
#     # and steps_per_epoch arguments to fit_generator())
#     if (next_file > length(files))
#       next_file <<- 1
# 
#     # determine the file name
#     file <- files[[next_file]]
# 
#     # process and return the data in the file. note that in a
#     # real example you'd further subdivide the data within the
#     # file into appropriately sized training batches. this
#     # would make this function much more complicated so we
#     # don't demonstrated it here
#     file_to_training_data(file)
#   }
# }

## -----------------------------------------------------------------------------
# early_stopping <- callback_early_stopping(monitor = 'val_loss', patience = 2)
# model %>% fit(X, y, validation_split = 0.2, callbacks = c(early_stopping))

## -----------------------------------------------------------------------------
# hist <- model %>% fit(X, y, validation_split=0.2)
# hist$history

## -----------------------------------------------------------------------------
# frozen_layer <- layer_dense(units = 32, trainable = FALSE)

## -----------------------------------------------------------------------------
# x <- layer_input(shape = c(32))
# layer <- layer_dense(units = 32)
# layer$trainable <- FALSE
# y <- x %>% layer
# 
# frozen_model <- keras_model(x, y)
# # in the model below, the weights of `layer` will not be updated during training
# frozen_model %>% compile(optimizer = 'rmsprop', loss = 'mse')
# 
# layer$trainable <- TRUE
# trainable_model <- keras_model(x, y)
# # with this model the weights of the layer will be updated during training
# # (which will also affect the above model since it uses the same layer instance)
# trainable_model %>% compile(optimizer = 'rmsprop', loss = 'mse')
# 
# frozen_model %>% fit(data, labels)  # this does NOT update the weights of `layer`
# trainable_model %>% fit(data, labels)  # this updates the weights of `layer`

## -----------------------------------------------------------------------------
# # instantiate a VGG16 model
# conv_base <- application_vgg16(
#   weights = "imagenet",
#   include_top = FALSE,
#   input_shape = c(150, 150, 3)
# )
# 
# # freeze it's weights
# freeze_weights(conv_base)
# 
# # create a composite model that includes the base + more layers
# model <- keras_model_sequential() %>%
#   conv_base %>%
#   layer_flatten() %>%
#   layer_dense(units = 256, activation = "relu") %>%
#   layer_dense(units = 1, activation = "sigmoid")
# 
# # compile
# model %>% compile(
#   loss = "binary_crossentropy",
#   optimizer = optimizer_rmsprop(lr = 2e-5),
#   metrics = c("accuracy")
# )
# 
# # unfreeze weights from "block5_conv1" on
# unfreeze_weights(conv_base, from = "block5_conv1")
# 
# # compile again since we froze or unfroze layers
# model %>% compile(
#   loss = "binary_crossentropy",
#   optimizer = optimizer_rmsprop(lr = 2e-5),
#   metrics = c("accuracy")
# )

## -----------------------------------------------------------------------------
# model <- keras_model_sequential()
# model %>%
#   layer_dense(units = 32, activation = 'relu', input_shape = c(784)) %>%
#   layer_dense(units = 32, activation = 'relu') %>%
#   layer_dense(units = 32, activation = 'relu')
# 
# length(model$layers)     # "3"
# model %>% pop_layer()
# length(model$layers)     # "2"

## -----------------------------------------------------------------------------
# model <- application_vgg16(weights = 'imagenet', include_top = TRUE)

## -----------------------------------------------------------------------------
# library(keras)
# use_backend("plaidml")

## -----------------------------------------------------------------------------
# library(keras)
# use_condaenv("plaidml")
# use_backend("plaidml")

## -----------------------------------------------------------------------------
# # testthat utilty for skipping tests when Keras isn't available
# skip_if_no_keras <- function(version = NULL) {
#   if (!is_keras_available(version))
#     skip("Required keras version not available for testing")
# }
# 
# # use the function within a test
# test_that("keras function works correctly", {
#   skip_if_no_keras()
#   # test code here
# })

## -----------------------------------------------------------------------------
# # Keras Python module
# keras <- NULL
# 
# # Obtain a reference to the module from the keras R package
# .onLoad <- function(libname, pkgname) {
#   keras <<- keras::implementation()
# }

## -----------------------------------------------------------------------------
# library(keras)
# use_session_with_seed(42)
# 
# # ...rest of code follows...
# 

## -----------------------------------------------------------------------------
# library(keras)
# use_session_with_seed(42, disable_gpu = FALSE, disable_parallel_cpu = FALSE)