To train our first auto-encoder, we first need to get R set up. In addition to the other packages in our checkpoint.R
file, we will add the data.table package to facilitate data management, as shown in the following code:
library(data.table)
Now we can source the checkpoint.R
file to set up the R environment for analysis, as follows:
source("checkpoint.R") options(width = 70, digits = 2)
For these first examples, we will use the Modified National Institute of Standards and Technology (MNIST) digits image data. The following code loads the necessary data, as in previous chapters, and sets up the H2O cluster for analysis. We use the first 20,000 rows of data for training and the next 10,000 rows for testing. In addition to loading the data and setting up the H2O cluster, the data need to be transferred to H2O, which is done using the as.h2o()
function:
## data and H2O setup digits.train <- read.csv("train.csv") digits.train$label <- factor(digits.train$label...