It would really help with speed for the package on 1M+ rows if we could turn off the top factors calculation. The overhead of doing both model trainings is really rough when models take hours to train.
Without knowing the architecture of the package completely, perhaps an input flag streamline == True could turn off some of the more extraneous and computationally expensive things?
Aylr
commented
6 years ago
It would really help with speed for the package on 1M+ rows if we could turn off the top factors calculation. The overhead of doing both model trainings is really rough when models take hours to train.
Without knowing the architecture of the package completely, perhaps an input flag
streamline == Truecould turn off some of the more extraneous and computationally expensive things?