Hey everybody,
If you are looking for a very fast Hadamard, check out the code here (which was developed in 2014 as part of my master thesis at City University of Hong Kong and Carnegie Mellon):
https://arxiv.org/pdf/1702.08159
https://github.com/curto2/mckernel
If you are looking for a very fast Hadamard, check out the code here (which was developed in 2014 as part of my master thesis at City University of Hong Kong and Carnegie Mellon):
https://arxiv.org/pdf/1702.08159
https://github.com/curto2/mckernel
More information here also, where emphasis is given to McKernel (a library that presents a framework to use kernel approximates in the mini-batch setting with Stochastic Gradient Descent)
https://paperswithcode.com/method/mckernel
Usage on how to use only the HADAMARD routines here:
https://github.com/curto2/mckernel/tree/master/standard
If you want to use FWH, add #include "hpp/McKernel.hpp" in your test file and compile with flags -msse2 (-mavx) -O3 -std=c++11 FWH. Here is an example of how to compile the code: $ g++-4.8 -o FWH_Test FWH_Test.cpp -msse2 -mavx -O3 -std=c++11 $ ./FWH_Test *Note: if -mavx is not compatible with your computer, please remove the flag.
Best regards,
De Curtò i DíAz.