Preparing the Data

The MNIST dataset is included with Keras and can be accessed using the dataset_mnist() function. Here we load the dataset then create variables for our test and training data:

library(keras)
mnist <- dataset_mnist()
x_train <- mnist$train$x
y_train <- mnist$train$y
x_test <- mnist$test$x
y_test <- mnist$test$y

The x data is a 3-d array (images,width,height) of grayscale values. To prepare the data for training we convert the 3-d arrays into matrices by reshaping width and height into a single dimension (28x28 images are flattened into length 784 vectors). Then, we convert the grayscale values from integers ranging between 0 to 255 into floating point values ranging between 0 and 1:

# reshape
dim(x_train) <- c(nrow(x_train), 784)
dim(x_test) <- c(nrow(x_test), 784)
# rescale
x_train <- x_train / 255
x_test <- x_test / 255

The y data is an integer vector with values ranging from 0 to 9. To prepare this data for training we one-hot encode the vectors into binary class matrices using the Keras to_categorical() function:

y_train <- to_categorical(y_train, 10)
y_test <- to_categorical(y_test, 10)

Defining the Model

The core data structure of Keras is a model, a way to organize layers. The simplest type of model is the sequential model, a linear stack of layers.

We begin by creating a sequential model and then adding layers using the pipe (%>%) operator:

model <- keras_model_sequential() 
model %>% 
  layer_dense(units = 256, activation = "relu", input_shape = c(784)) %>% 
  layer_dropout(rate = 0.4) %>% 
  layer_dense(units = 128, activation = "relu") %>%
  layer_dropout(rate = 0.3) %>%
  layer_dense(units = 10, activation = "softmax")

The input_shape argument to the first layer specifies the shape of the input data (a length 784 numeric vector representing a grayscale image). The final layer outputs a length 10 numeric vector (probabilities for each digit) using a softmax activation function.

Use the summary() function to print the details of the model:

summary(model)
Model
________________________________________________________________________________
Layer (type)                        Output Shape                    Param #     
================================================================================
dense_1 (Dense)                     (None, 256)                     200960      
________________________________________________________________________________
dropout_1 (Dropout)                 (None, 256)                     0           
________________________________________________________________________________
dense_2 (Dense)                     (None, 128)                     32896       
________________________________________________________________________________
dropout_2 (Dropout)                 (None, 128)                     0           
________________________________________________________________________________
dense_3 (Dense)                     (None, 10)                      1290        
================================================================================
Total params: 235,146
Trainable params: 235,146
Non-trainable params: 0
________________________________________________________________________________

Next, compile the model with appropriate loss function, optimizer, and metrics:

model %>% compile(
  loss = "categorical_crossentropy",
  optimizer = optimizer_rmsprop(),
  metrics = c("accuracy")
)

Training and Evaluation

Use the fit() function to train the model for 30 epochs using batches of 128 images:

history <- model %>% fit(
  x_train, y_train, 
  epochs = 30, batch_size = 128, 
  validation_split = 0.2
)

The history object returned by fit() includes loss and accuracy metrics which we can plot:

plot(history)

Evaluate the model’s performance on the test data:

model %>% evaluate(x_test, y_test,verbose = 0)
$loss
[1] 0.1149

$acc
[1] 0.9807

Generate predictions on new data:

model %>% predict_classes(x_test)
  [1] 7 2 1 0 4 1 4 9 5 9 0 6 9 0 1 5 9 7 3 4 9 6 6 5 4 0 7 4 0 1 3 1 3 4 7 2 7 1 2
 [40] 1 1 7 4 2 3 5 1 2 4 4 6 3 5 5 6 0 4 1 9 5 7 8 9 3 7 4 6 4 3 0 7 0 2 9 1 7 3 2
 [79] 9 7 7 6 2 7 8 4 7 3 6 1 3 6 9 3 1 4 1 7 6 9
 [ reached getOption("max.print") -- omitted 9900 entries ]

Keras provides a vocabulary for building deep learning models that is simple, elegant, and intuitive. Building a question answering system, an image classification model, a neural Turing machine, or any other model is just as straightforward.

The Guide to the Sequential Model article describes the basics of Keras sequential models in more depth.