This is a fork of Multicore t-SNE cython wrapper. This code has similar speed with Multicore t-SNE. In addition, support partial fitting function to continuously add points into tsne map.
Cython wrappers are available.
- Python3
- Numpy (>=1.18.0)
- Cython (>=0.28.2)
- cysignals
- cmake
- OpenMP 2(slow if not install)
pip install numpy cython cysignals
mkdir build
cd build
cmake -DCMAKE_BUILD_TYPE=RELEASE --DPYTHON_EXECUTABLE=$(which python) ..
make
make install
pip uninstall PyFastTsne
Tested with 3.6 (conda) and Ubuntu 16.04.
You can use it as a near drop-in replacement for sklearn.manifold.TSNE.
from PyFastTsne import PyTsne
x_dim = 728
y_dim = 2
tsne = PyTsne(x_dim, y_dim)
Y = tsne.fit_transform(X)
## continuously add extra points
tsne = tsne.partial_fit(extra_X, ret_Y, n_iter=300)
Please refer to sklearn TSNE manual for parameters explanation.
This implementation n_components=2, which is the most common case (use Barnes-Hut t-SNE or sklearn otherwise). Also note that some parameters are there just for the sake of compatibility with sklearn and are otherwise ignored. See MulticoreTSNE class docstring for more info.
You can test it on MNIST dataset with the following command:
cd build/PyFastTsne
python test.py
Inherited from original repo's license.
- Allow other types than double
- Improve step 2 performance (possible)