Deep Learning with R Cookbook

There's more...

While working with large datasets, we often run out of memory while training a deep learning model. The keras library in R provides various generator utility functions, which generate batches of training data on the fly during the training process. It also provides a utility function for creating batches of temporal data. The following code creates a supervised form of data similar to what we created in the How to do it.. section of this recipe, using a generator utility:

# importing required libraries
library(reticulate)
library(keras)

# generating dummy data
data =  seq(from = 1,to = 10)

# timseries generator
gen = timeseries_generator(data = data,targets = data,length = 1,batch_size = 5)

# Print first batch
iter_next(as_iterator ...