search

ClimbsRocks / auto_ml

[UNMAINTAINED] Automated machine learning for analytics & production
http://auto-ml.readthedocs.io
MIT License
1.64k stars 310 forks source link

look into optimizing with numba #123

Open ClimbsRocks opened 8 years ago

ClimbsRocks commented 8 years ago

particularly jit, and http://numba.pydata.org/numba-doc/dev/user/jitclass.html

eventually, might be able to do some gpu optimization.

vishnumani2009 commented 8 years ago

what kind of optimisation are we looking here?

vishnumani2009 commented 8 years ago

If its wrt speed of model training then I guess we cant do much as most of important scikit API's either are wrapper around C library's like LIBLINEAR , LIBSVM etc which will again be compiled code.

On the other hand if my understanding is correct critical time consuming submodules with in fully pythonic code are written in C/C++ and compiled using Cython(pyrex) which doesn't make it eligible for jit compilation.

vishnumani2009 commented 8 years ago

But yea numpy array/data frame looping operations etc can be very optimised for speed.