I have a flask application that calls Bayesian models written in PyMC3. Whenever more than one request is made simultaneously, the application processes each request separately keeping the other ones in queue.
I have tried using 'threaded=True' and running through a Gunicorn server with gevent asynchronous workers as mentioned in How to run Flask with Gunicorn in multithreaded mode , but still the issue persists.
Is this a specific issue related to pyMC3?
