[mlpack-git] [mlpack/mlpack] How to build mlpack with visual studio 2013 (#659)

[Lau]

@zoq Here is the way I build armadillo: 
1. I build OpenBLAS according to [How to use OpenBLAS in Microsoft Visual Studio](https://github.com/xianyi/OpenBLAS/wiki/How-to-use-OpenBLAS-in-Microsoft-Visual-Studio) with MinGW64 (using MSYS2). And download LAPACK and BLAS (.lib files) from [LAPACK for Windows](http://icl.cs.utk.edu/lapack-for-windows/lapack/).
2. Configure armadillo with CMake by following settings:
[CMake_Settings.png](http://s33.postimg.org/kjpl8yh0v/cmake_settings.png).
3. Open armadillo.sln with vs2013 and build ALL_BUILD as well as INSTALL both in Debug and Release mode.

And there is weired thing I just found. In my last post I said I use three 32-D data vectors to train the guassian distribution. But if I transpose the data matrix(turn 32x3 matrix to 3x32 matrix, which representing 32 three-dimension vectors), the function
```
mlpack::GaussianDistribution normalDistribution;
normalDistribution.Train(matrix);
```
works fine, there is no exception showing up. But it computes the distribution of 32 3d data items rather than three 32-Dimension data items I want, right? Is GaussianDistribution unable to train too few data items?

Of course I'd like to use NuGet. Please show me some instructions to do so OK?


---
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
https://github.com/mlpack/mlpack/issues/659#issuecomment-223276511
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mailman.cc.gatech.edu/pipermail/mlpack-git/attachments/20160602/070a7fe6/attachment.html>