Skip to content
Efficient implementation of linear algebra for Haskell with very concise Einstein syntax for matrices and tensors.
Branch: master
Clone or download
Latest commit c4d9e6b Jan 8, 2019
Permalink
Type Name Latest commit message Commit time
Failed to load latest commit information.
benchmark Benchmark build Dec 7, 2018
cbits Untested build Dec 2, 2018
src Build 19008.68055 Jan 8, 2019
test/all Build 19008.67957 Jan 8, 2019
.gitignore Tested build 19004.56614 Jan 4, 2019
.travis.yml merge with master vol2 Nov 14, 2018
ChangeLog.md Build 18329.47431 Nov 25, 2018
LICENSE Update LICENSE Nov 1, 2018
README.md Update README.md Nov 23, 2018
appveyor.yml delete deprecated lts-9 AppVeyor build Nov 2, 2018
benchmark.sh
build.sh Temp 19005.48818 Jan 5, 2019
deploy.sh Benchmark build Dec 3, 2018
id-build.sh
loc.sh Build 19004.67851 Jan 4, 2019
package.yaml Temp 19005.51880 Jan 5, 2019
stack.yaml port to LTS-12.16 resolver Nov 1, 2018
test.sh Build 19004.68511 Jan 4, 2019
vcs.sh Build 19005.48327 Jan 5, 2019

README.md

Multilinear Hackage Build Status BuildStatus

Summary

Multilinear is general - purpose linear algebra and multi-dimensional array library for Haskell. It provides generic and efficient implementation of linear algebra operations on vectors, linear functionals, matrices and its higher - rank analoges: tensors. It can also be used as simply a miltidimensional arrays for everyone.

AS FOR NOW, THE LIBRARY IS IN BETA PHASE - DO NOT USE IT FOR PRODUCTION!!

Scripting

Multilinear is optimized to being used from GHCi. It uses easy and concise notation of Einstein summation convention to calculate complex tasks. Using this, you are able to write for example a deep learnin neural network from scratch in just a few lines of interpreter code. If you want to know more about Einstein convention, see the Wikipedia: https://en.wikipedia.org/wiki/Einstein_notation

Machine learning

Multi-dimensional algebra is especially useful to quickly write machine learning algorithms (eg. neural networks) from scratch. When library will be stable, some examples will be available.

Contribution

If you want to contribute to this library, contact with me.

Who do I talk to?

All copyrights to Artur M. Brodzki. Contact mail: artur@brodzki.org

You can’t perform that action at this time.