MPI-based t-SNE

This is an MPI-based implementation of Barnes-Hut t-SNE. It is built off of the sklearn implementation of BH t-SNE using a python wrapper on MPI.

Building and Running

The building instructions are the same as for original sklearn t-sne, which can be found in README_ORIGINAL.rst

Only difference is that you also need to install mpi before building:

conda install mpi4py openmpi

To run, execute a command as follows:

mpiexec -n 4 python tsne.py

Changing 4 to be the number of processes you would like to use.

Name		Name	Last commit message	Last commit date
Latest commit History 29,960 Commits
.binder		.binder
.circleci		.circleci
.github		.github
asv_benchmarks		asv_benchmarks
benchmarks		benchmarks
build_tools		build_tools
doc		doc
examples		examples
maint_tools		maint_tools
sklearn		sklearn
.cirrus.star		.cirrus.star
.codecov.yml		.codecov.yml
.coveragerc		.coveragerc
.git-blame-ignore-revs		.git-blame-ignore-revs
.gitignore		.gitignore
.mailmap		.mailmap
.pre-commit-config.yaml		.pre-commit-config.yaml
CODE_OF_CONDUCT.md		CODE_OF_CONDUCT.md
CONTRIBUTING.md		CONTRIBUTING.md
COPYING		COPYING
MANIFEST.in		MANIFEST.in
Makefile		Makefile
README.md		README.md
README_ORIGINAL.rst		README_ORIGINAL.rst
SECURITY.md		SECURITY.md
azure-pipelines.yml		azure-pipelines.yml
conftest.py		conftest.py
performance.png		performance.png
pyproject.toml		pyproject.toml
report.pdf		report.pdf
setup.cfg		setup.cfg
setup.py		setup.py
tsne.py		tsne.py