learning_cnn.R

#https://towardsdatascience.com/how-to-implement-deep-learning-in-r-using-keras-and-tensorflow-82d135ae4889

devtools::install_github("rstudio/tensorflow")
library(tensorflow)
install_tensorflow(method = "conda", version="nightly")
#install_tensorflow() #<- apparently this works now?!??
library(keras)
use_condaenv('r-tensorflow')

? dataset_cifar10 #to see the help file for details of dataset
cifar <- dataset_cifar10()

train_x <- cifar$train$x / 255
#convert a vector class to binary class matrix
#converting the target variable to once hot encoded vectors using 
#keras inbuilt function 'to_categorical()
train_y <- to_categorical(cifar$train$y, num_classes = 10)
#TEST DATA
test_x <- cifar$test$x / 255
test_y <- to_categorical(cifar$test$y, num_classes = 10)
#checking the dimentions
dim(train_x)
cat("No of training samples\t",
    dim(train_x)[[1]],
    "\tNo of test samples\t",
    dim(test_x)[[1]])

#a linear stack of layers
model <- keras_model_sequential()
#configuring the Model
model %>%
  #defining a 2-D convolution layer
  layer_conv_2d(
    filter = 32,
    kernel_size = c(3, 3),
    padding = "same",
    input_shape = c(32, 32, 3)
  ) %>%
  layer_activation("relu") %>%
  #another 2-D convolution layer
  
  layer_conv_2d(filter = 32 , kernel_size = c(3, 3))  %>%  layer_activation("relu") %>%
  #Defining a Pooling layer which reduces the dimentions of the features map 
  #and reduces the computational 
        #complexity of the model
  layer_max_pooling_2d(pool_size = c(2, 2)) %>%
  #dropout layer to avoid overfitting
  layer_dropout(0.25) %>%
  layer_conv_2d(filter = 32 ,
                kernel_size = c(3, 3),
                padding = "same") %>% layer_activation("relu") %>%  layer_conv_2d(filter =
                32, kernel_size = c(3, 3)) %>%  layer_activation("relu") %>%
  layer_max_pooling_2d(pool_size = c(2, 2)) %>%
  layer_dropout(0.25) %>%
  #flatten the input
  layer_flatten() %>%
  layer_dense(512) %>%
  layer_activation("relu") %>%
  layer_dropout(0.5) %>%
  #output layer-10 classes-10 units
  layer_dense(10) %>%
  #applying softmax nonlinear activation function to the output layer 
  #to calculate cross-entropy
  layer_activation("softmax")
  #for computing Probabilities of classes-"logit(log probabilities)

#Model's Optimizer
#defining the type of optimizer-ADAM-Adaptive Momentum Estimation
opt<-optimizer_adam( lr= 0.0001 , decay = 1e-6 )
#lr-learning rate , decay - learning rate decay over each update

model %>%
  compile(loss="categorical_crossentropy",
          optimizer=opt,metrics = "accuracy")
#Summary of the Model and its Architecture
summary(model)

#TRAINING PROCESS OF THE MODEL
data_augmentation <- TRUE  
if(!data_augmentation) {  
  model %>% fit( train_x,train_y ,batch_size=32,
                 epochs=80,validation_data = list(test_x, test_y),
                 shuffle=TRUE)
} else {  
  #Generating images
  
  gen_images <- image_data_generator(featurewise_center = TRUE,
                                     featurewise_std_normalization = TRUE,
                                     rotation_range = 20,
                                     width_shift_range = 0.30,
                                     height_shift_range = 0.30,
                                     horizontal_flip = TRUE  )
  #Fit image data generator internal statistics to some sample data
  gen_images %>% fit_image_data_generator(train_x)
  #Generates batches of augmented/normalized data from image data and 
  #labels to visually see the generated images by the Model
  model %>% fit_generator(
    flow_images_from_data(train_x, train_y,gen_images,
                          batch_size=32,save_to_dir="F:/PROJECTS/CNNcifarimages/"),
    steps_per_epoch=as.integer(50000/32),epochs = 80,
    validation_data = list(test_x, test_y) )
}
#use save_to_dir argument to specify the directory to save the 
#images generated by the Model and to visually check the Model's 
#output and ability to classify images.