I am just reading:
Deep Learning and Scientific Computing with R torch:
https://skeydan.github.io/Deep-Learning-and-Scientific-Computing-with-R-torch/
and tried to build the example 14.2.3 with the data from titanic.
The example from the book run smoothly. When I use the titanic data R crashes with fatal error:

train <- train %>% select_if(is.numeric) ## use only numeric columns
train <- train %>% na.omit()### remove NA
 ## create a torch dataset
 ds <- tensor_dataset(
    torch_tensor(as.matrix(train[, c(3:7)])),
    torch_tensor(as.numeric(train$Survived)
                 ) $to(torch_long()))

  d_in <- 5
  ### dimensionality of hidden layer
  d_hidden <- 32
  ### output dimensionality (number of predicted features)
  d_out <- 1
  ## net module
  net <- nn_module(
  initialize = function(d_in, d_hidden, d_out) {
    self$net <- nn_sequential(
      nn_linear(d_in, d_hidden),
      nn_relu(),
      nn_linear(d_hidden, d_out)
    )
  },
  forward = function(x) {
    self$net(x)
  }
)
fitted <- net %>%
  setup(
    loss = nn_mse_loss(),
    optimizer = optim_adam) %>% set_hparams(d_in = d_in,
              d_hidden = d_hidden,
              d_out = d_out) %>%
  fit(
    ds,
    epochs = 200
  )
 

Any ideas what I am doing wrong? I reinstalled torch, all libraries are uptodate.