This software package contains a Barnes-Hut implementation of the t-SNE algorithm. The implementation is described in this paper.
On Linux or OS X, compile the source using the following command:
g++ sptree.cpp tsne.cpp -o bh_tsne -O2
The executable will be called bh_tsne.
On Windows using Visual C++, do the following in your command line:
- Find the
vcvars64.batfile in your Visual C++ installation directory. This file may be namedvcvars64.bator something similar. For example:
// Visual Studio 12
"C:\Program Files (x86)\Microsoft Visual Studio 12.0\VC\bin\amd64\vcvars64.bat"
// Visual Studio 2013 Express:
C:\VisualStudioExp2013\VC\bin\x86_amd64\vcvarsx86_amd64.bat
-
From
cmd.exe, go to the directory containing that .bat file and run it. -
Go to
bhtsnedirectory and run:
nmake -f Makefile.win all
The executable will be called windows\bh_tsne.exe.
The code comes with wrappers for Matlab and Python. These wrappers write your data to a file called data.dat, run the bh_tsne binary, and read the result file result.dat that the binary produces. There are also external wrappers available for Torch, R, and Julia. Writing your own wrapper should be straightforward; please refer to one of the existing wrappers for the format of the data and result files.
Demonstration of usage in Matlab:
filename = websave('mnist_train.mat', 'https://github.com/awni/cs224n-pa4/blob/master/Simple_tSNE/mnist_train.mat?raw=true');
load(filename);
numDims = 2; pcaDims = 50; perplexity = 50; theta = .5; alg = 'svd';
map = fast_tsne(digits', numDims, pcaDims, perplexity, theta, alg);
gscatter(map(:,1), map(:,2), labels');