train_images<-train_images / 255 test_images <- test_images / 255 train_labels <- to_categorical(train_labels, 10) test_labels <- to_categorical(test_labels, 10) # defining the model and layers model <- k 'softmax') summary(model) # compile (define loss and optimizer) model %>% compile( loss = 'categorical_crossentropy', optimizer = optimizer_rmsprop(), metrics = c('accuracy') ) # train (fit)