### R/Keras Examples for FFNs & CNNs
### Based on “Deep Learning wth R” by Chollet & Allaire; 
###   https://www.manning.com/books/deep-learning-with-r

##install Keras:
install.packages("keras")              
library(keras)
#CPU-based:                         
install_keras()


##Loading the MNIST dataset in Keras:
library(keras)
mnist <- dataset_mnist()
train_images <- mnist$train$x
train_labels <- mnist$train$y
test_images <- mnist$test$x
test_labels <- mnist$test$y

str(train_images)
 # int [1:60000, 1:28, 1:28] 0 0 0 0 0 0 0 0 0 0 ...
str(test_labels)
 #  int [1:10000(1d)] 7 2 1 0 4 1 4 9 5 9 ...

train_images <- array_reshape(train_images, c(60000, 28 * 28))
train_images <- train_images / 255
test_images <- array_reshape(test_images, c(10000, 28 * 28))
test_images <- test_images / 255

train_labels <- to_categorical(train_labels)
test_labels <- to_categorical(test_labels)


###FFN:
network <- keras_model_sequential() %>% 
  layer_dense(units = 512, activation = "relu", input_shape = c(28 * 28)) %>% 
  layer_dense(units = 10, activation = "softmax")

network %>% compile(
  optimizer = "rmsprop",
  loss = "categorical_crossentropy",
  metrics = c("accuracy")
)

network %>% fit(train_images, train_labels, epochs = 5, batch_size = 128)
#Epoch 1/5    469/469 [==============================] - 2s 4ms/step - loss: 0.2575 - accuracy: 0.9254
#Epoch 2/5    469/469 [==============================] - 2s 3ms/step - loss: 0.1031 - accuracy: 0.9695
#Epoch 3/5    469/469 [==============================] - 2s 3ms/step - loss: 0.0688 - accuracy: 0.9792
#Epoch 4/5    469/469 [==============================] - 2s 3ms/step - loss: 0.0489 - accuracy: 0.9857
#Epoch 5/5    469/469 [==============================] - 2s 3ms/step - loss: 0.0380 - accuracy: 0.9887

metrics <- network %>% evaluate(test_images, test_labels, verbose = 0)
metrics
#  loss   accuracy 
# 0.06546536 0.98079997 


###CNN:
library(keras)

model <- keras_model_sequential() %>%
  layer_conv_2d(filters = 32, kernel_size = c(3, 3), activation = "relu",
                input_shape = c(28, 28, 1)) %>%
  layer_max_pooling_2d(pool_size = c(2, 2)) %>%
  layer_conv_2d(filters = 64, kernel_size = c(3, 3), activation = "relu") %>%
  layer_max_pooling_2d(pool_size = c(2, 2)) %>%
  layer_conv_2d(filters = 64, kernel_size = c(3, 3), activation = "relu") %>%

  layer_flatten() %>%
  layer_dense(units = 64, activation = "relu") %>%
  layer_dense(units = 10, activation = "softmax")

mnist <- dataset_mnist()
c(c(train_images, train_labels), c(test_images, test_labels)) %<-% mnist
train_images <- array_reshape(train_images, c(60000, 28, 28, 1))
train_images <- train_images / 255
test_images <- array_reshape(test_images, c(10000, 28, 28, 1))
test_images <- test_images / 255
train_labels <- to_categorical(train_labels)
test_labels <- to_categorical(test_labels)
model %>% compile(
  optimizer = "rmsprop",
  loss = "categorical_crossentropy",
  metrics = c("accuracy")
)
model %>% fit(
  train_images, train_labels,
  epochs = 5, batch_size=64
)

results <- model %>% evaluate(test_images, test_labels)
results
#$loss
#[1] 0.02563557
#$acc
#[1] 0.993