I read that python has GIL issue causing it not to utilize the CPU processes & threads optimally.
Has someone used a C++ wrapper or wrote piece of C++ code in python - specifically that portion which requires multithreading?
I also looked at some C/C++ bindings and CFFI as well. Even Tensorflow C++ API documentation is not extensive.
The goal is to productionize a machine learning model (with or without Tensorflow 2.0), push it into production and use C++ wrapper which can exploit CPU's multithreading. If someone can point me in the right direction, it will be of great help.
thank you