Description

If you run a training/fit with the PrivBayes plugin on a dataset with continuous columns and enough lines, you will encounter an out-of-memory system error (basically the program stops).
If you dig a bit, what happens is that the continuous variables are encoded as categorical, with many categories (as we deal with floats) and it represents too much in the computer memory, which leads to a crash.

It seems to have an error in the control logic of the _encode function in the file: src/synthcity/plugins/privacy/plugin_privbayes.py