memory usage with party::cforest

Is there a way to shrink the size of RandomForest-class (an S4
object), so that it requires less memory during run-time and less disk
space for serialization?

On my system the data slot is about 2GB, which is causing problems,
and I'd like to see whether predict() works without it.

# example with a much smaller data set (i.e., less than 2GB)
require(party)
data(iris)
cf <- cforest(Species ~ ., data=iris)
str(cf, max.level=2)
cf at data <- NULL # this fails



Andrew