[mlpack-git] [mlpack] implement CNN and LeNet1 (#405)

[Shangtong Zhang]

I implement convolution with fft and block.
When input and kernel scale is large, fft performs much better than simple loop convolution.
But in my experiment simple loop convolution performs better than block convolution.
As my convolution type is different from http://lists.r-forge.r-project.org/pipermail/rcpp-devel/2012-January/003304.html, so the block convolution is also different.
My configuration:
cmake -D DEBUG=OFF -G "Xcode"
In Xcode in use Release Mode to do comparasion

---
Reply to this email directly or view it on GitHub:
https://github.com/mlpack/mlpack/pull/405#issuecomment-80835294
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20150314/e9e56a41/attachment.html>