I don’t see this in the documentation, but is there a way to get predictions from a bambi model to run in parallel? It seems that there is no way to pass a number of cores into the function like with fit() at the moment, so is it possible to break up the new data and give each chunk to a different core to process?
.predict() calls a method that mostly uses numpy arrays to compute the precitions (i.e. it does not call any PyMC related function). In my experience this results in faster computations. Do you have a use-case where this is slow and you’re thinking about going parallel?
I’m not sure about the following, but doesn’t NumPy automatically perform some parallel stuff when computing dot products? (
.predict() uses dot products basically)