N_jobs do not work in sklearn classes

Question

Does anyone use the "n_jobs" sklearn classes? I work with sklearn in Anaconda 3.4 64 bit. Spyder version is 2.3.8. My script cannot complete its execution after setting the "n_jobs" parameter of some sklearn class to a non-zero value. Why is this happening?

Answer 1

Several scikit-learn tools, such as GridSearchCV and cross_val_score, rely inside the Pythons multiprocessing module to parallelize execution on multiple Python processes, passing n_jobs> 1 as an argument.

Adapted from the Sklearn documentation:

, Python fork exec . , ( ) Accelerate/vecLib OSX, ( ) MKL, OpenMP GCC, nvidias Cuda ( , ), . fork, : , , . , , : OpenBLAS ( master 0.2.10), GCC OpenMP runtime ( ).