search

dflemin3 / approxposterior

A Python package for approximate Bayesian inference and optimization using Gaussian processes
https://dflemin3.github.io/approxposterior/
MIT License
41 stars 9 forks source link

Parallel approxposterior using python 3.8+ multiprocessing #61

Open dflemin3 opened 4 years ago

dflemin3 commented 4 years ago

As per the Python 3.8 release notes:

multiprocessing can now use shared memory segments to avoid pickling costs between processes

This new change can enable parallel approxposterior in terms of parallelizing various optimization tasks as now the GP no longer needs to be pickled as this task was slower than the optimization.