scavetta / conf_tensorflow_training_day1 Goto Github PK

Hi,

I am learning your great tutor of keras in R but I can't understend some point in the 1_3_DL_Basics_Multi-class_Classification.Rmd though it is really a detailed tutor I have ever seen.

`network <- keras_model_sequential() %>%
layer_dense(units = 2^6, activation = "relu", input_shape = 13) %>%
layer_dense(units = 2^6, activation = "relu") %>%
layer_dense(units = 5, activation = "softmax")

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

history<- network %>% fit(
train_data,
train_targets,
epochs = 50,
batch_size = 16,
verbose = FALSE
)`

In these code, we seems haven't saved the parameter to our network to our 'network' model
How can use the optimize result to the model?

Thanks