A Python-embedded modeling language for convex optimization problems.
Clone or download
Pull request Compare This branch is 21 commits ahead, 1169 commits behind cvxgrp:master.
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Permalink
Failed to load latest commit information.
conda-recipe
continuous_integration
cvxpy
doc
examples
.bumpversion.cfg
.coveragerc
.gitignore
.travis.yml
CHANGELOG.md
COPYING
MANIFEST
README.md
add_gpl
appveyor.yml
copyright.txt
examples_roadmap.txt
pylint_settings.py
requirements.txt
roadmap.txt
setup.py

README.md

This is a special version of CVXPY designed for matrix-free convex optimization. See these papers for further details on matrix-free convex optimization. See the CVXPY site for more information on CVXPY itself.

From the perspective of the user, matrix-free CVXPY is exactly the same as standard CVXPY except it has an additional solver "SCS_MAT_FREE", which is a matrix-free version of SCS. The solver is preliminary and not at all optimized.

To install matrix-free CVXPY, clone this Github repository and run python setup.py install in the repository directory. You must also install the matrix-free SCS solver from source (repository here).

See this repository for instructions on installing a faster version of the matrix-free SCS solver.