inst/examples/datasets.R

In kerasR: R Interface to the Keras Deep Learning Library

if (keras_available()) {
  boston <- load_boston_housing()
  X_train <- normalize(boston$X_train, 0)
  Y_train <- boston$Y_train
  X_test <- normalize(boston$X_test, 0)
  Y_test <- boston$Y_test
  
  mod <- Sequential()
  mod$add(Dense(units = 200, input_shape = 13))
  mod$add(Activation("relu"))
  mod$add(Dense(units = 200))
  mod$add(Activation("relu"))
  mod$add(Dense(units = 1))
  keras_compile(mod,  loss = 'mse', optimizer = SGD())
  
  keras_fit(mod, scale(X_train), Y_train,
            batch_size = 32, epochs = 20,
            verbose = 1, validation_split = 0.1)
}