News for hackers. Stuff that matters.

A comment on Basic Neural Network on Python. Parent comment

Re:Basic Neural Network on Python by lightcatcher on Friday July 05, 2013 @11:31AM (#5995283)
A relatively small lookup table for the sigmoid function can also work well. Here are the various sigmoid approximations that Theano (a library used for deep learning research among other things) offers: http://deeplearning.net/software/theano/library/tensor/nnet/...
[ Parent | on HN ]
Re:Basic Neural Network on Python by gamegoblin on Friday July 05, 2013 @01:44PM (#5995884)
I usually use an array with a few thousand entries. In C this gives me a 2.5x speedup over the exact function with no important decrease in accuracy.
[ Parent | on HN ]
Re:Basic Neural Network on Python by mistercow on Friday July 05, 2013 @03:02PM (#5996255)
I wonder though if you might actually do better overall with a smaller lookup table and interpolation (or even just a polynomial approximation, which can be evaluated without branching), since large lookup tables can lead to bad cache behavior.
[ Parent | on HN ]
Re:Basic Neural Network on Python by dfrodriguez143 on Friday July 05, 2013 @02:54PM (#5996218)
Very interesting and simple improvement. Definitely will take a look at that.
[ Parent | on HN ]