Open breznak opened 5 years ago
Thanks to PR #769 local inhibition (only for wrapAround=true
) is now 2x faster!
local inhibition is very slow! 70x (1s vs 74s global/local inhibition)
which equals to improvement and it's now only 10x slower than global inh!
CC CUDA SP #50, SP faster LocalInho #86 , #92
From https://github.com/htm-community/nupic.cpp/pull/119#pullrequestreview-178811874 :
I'm evaluation the performance impact of ND vs 1D, where algorithmically the 1D is really simpler.
Global Inh
Local, 1D (to be removed):
local, ND (proposed default, resp. only choice):
TL;DR