curious to understand why model size differs drastically on training while i am just making minor changes in nlu and core data?
every time I am using rasa train command to train the model
in my case model size is either 60mb or 120mb randomly.
pls see snap-shot, data is sorted by date.
