I am training a quantile forest regression using the grf package in R https://grf-labs.github.io/grf/REFERENCE.html#prediction
My dataset has 12 million records and i am using a 192gb 48 core cluster. The model trains pretty fast ( 3 mins for 50 trees), but the issue is always when using predict.
predict(q.forest, X.distinct)
The process is extremely slow and even when I predict a handful of values, it takes a long time to run.
Is there anyway to speed this process up or does it work differently