Hi,
I’ve adopted the documentation example for jitting the logistic regression equation with parallel=True, and miserably fell flat on my face at it ― Elapsed running time of parallel jit for the logistic regression documentation example is dozens of seconds · Issue #9395 · numba/numba · GitHub.
However in the meanwhile, I also am not certain from the discussion on that page what form of parallelization should have taken place, meaning, fiddling with pen and paper is the function of logistic regression really one that turns from a recursive definition to a regular equivalent function which can be then computed modularly in a parallel calculation graph? or does numba parallelize computing in isolation or by blocks for the indices of w, the weights vector, or in some totally different obvious way or aspect?
Anyway, thanks for your support,
Matan