From Fedora Project Wiki
(adding warning that the SIG is inactive)
 
(37 intermediate revisions by 10 users not shown)
Line 1: Line 1:
{{admon/warning|The Machine Learning SIG is inactive and the information on this page is mostly out of date. The [[SIGs/AI-ML|AI/ML SIG]] has replaced this SIG and will be the better reference going forward.}}
= Machine Learning SIG =
= Machine Learning SIG =
The Machine Learning SIG, aims to make Fedora the best platform for all things related to [http://en.wikipedia.org/wiki/Machine_learning Machine Learning].
The Machine Learning SIG's goal is to make Fedora the best platform for all things related to [http://en.wikipedia.org/wiki/Machine_learning Machine Learning].  We aim to act a hub in the gap between the Astronomy, Bigdata, Fedora Medical and Science and Technology SIGs and also between all interesting projects related to machine learning in Fedora.


== Members ==
== Members ==
Line 6: Line 9:
[[User:besser82|Björn Esser (besser82)]] <[mailto:besser82@fedoraproject.org besser82@fedoraproject.org]>
[[User:besser82|Björn Esser (besser82)]] <[mailto:besser82@fedoraproject.org besser82@fedoraproject.org]>


= Machine Learning Packages =
[[User:kushal124|Kushal Khandelwal (kushal124)]] <[mailto:kushalkhandelwal10@gmail.com kushalkhandelwal10@gmail.com]>
 
[[User:lupinix|Christian Dersch (lupinix)]] <[mailto:lupinix@fedoraproject.org lupinix@fedoraproject.org]>
 
[[User:dhanesh95|Dhanesh B. Sabane (dhanesh95)]] <[mailto:dhanesh95@fedoraproject.org dhanesh95@fedoraproject.org]>
 
[[User:lbalhar|Lumír Balhar (lbalhar)]] <[mailto:lbalhar@redhat.com lbalhar@redhat.com]>
 
[[User:canderson9|Christiano Anderson (canderson9)]] <[mailto:canderson9@fedoraproject.org canderson9@fedoraproject.org]>
 
[[User:dkirwan|David Kirwan (dkirwan)]] <[mailto:dkirwan@redhat.com dkirwan@redhat.com]>
 
[[User:willo|Graham Williamson (willo)]] <[mailto:willo@fedoraproject.org willo@fedoraproject.org]>
 
[[User:trix|Tom Rix (trix)]] <[mailto:trix@redhat.com trix@redhat.com]>
 
[[User:fed500|Benson Muite (fed500)]] <[mailto:fed500@fedoraproject.org fed500@fedoraproject.org]>
 
= Machine Learning Projects =
 
== Thoth  ==
 
[https://thoth-station.ninja/ Thoth] is a promising project which uses artificial intelligence to analyze and recommend software stack for artificial intelligence applications. It's based on an assumption that the proper combination of tools and libraries can have a significant impact on the performance of your project.
 
Thoth is now in the alpha state but it should be soon ready for beta testers.
 
Github organization: [https://github.com/thoth-station github.com/thoth-station]
 
== Optimized TensorFlow builds ==
 
[https://tensorflow.pypi.thoth-station.ninja/ The Artificial Intelligence Center of Excellence (AICoE)] provides a [https://tensorflow.pypi.thoth-station.ninja/index/ simple Python package index] with builds of Tensorflow optimized for Fedora/Centos/RHEL.
 
== ML-Pack ==
 
[https://mlpack.org/ ML-Pack] is a lightweight C++ machine learning library that has Python and Julia bindings. It is available in Fedora repositories.
 
== Shogun ==
 
[https://shogun-toolbox.org/ Shogun] is a machine learning toolbox with interfaces to a number of other languages including Python, Ruby and R.
 
== Dlib ==
 
[http://dlib.net Dlib] is an open source C++ machine learning library that can also be called from Python.
 
== FANN ==


== Interesting software waiting for being packaged ==
[http://leenissen.dk/fann/wp/ FANN(Fast Artificial Neural Network Library)] is an open source C++ machine learning library with a wide variety of language bindings. It is available in Fedora repositories.


* [https://pypi.python.org/pypi/bob bob - free signal-processing and machine learning toolbox]
== h2o ==
* [https://pypi.python.org/pypi/copper copper - Fast, easy and intuitive machine learning prototyping]
* [https://pypi.python.org/pypi/ease ease - Machine learning based automated text classification library]
* [https://pypi.python.org/pypi/hyperspy hyperspy - Hyperspectral data analysis toolbox]
* [https://pypi.python.org/pypi/infer infer - machine learning toolkit for classification and assisted experimentation]
* [http://java-ml.sourceforge.net/ Java-ML]
* [https://pypi.python.org/pypi/milk milk - Machine Learning Toolkit]
* [https://pypi.python.org/pypi/MLizard MLizard - Machine Learning workflow automatization]
* [http://mlpy.sourceforge.net/ mlpy - Machine Learning Python]
* [https://pypi.python.org/pypi/mmlf Maja Machine Learning Framework]
* [https://pypi.python.org/pypi/Monte Monte - machine learning in pure Python]
* [https://pypi.python.org/pypi/nolearn nolearn - Miscellaneous utilities for machine learning]
* [https://pypi.python.org/pypi/pcSVM pcSVM]
* [https://pypi.python.org/pypi/Peach Peach - Python library for computational intelligence and machine learning]
* [http://pybrain.org/ PyBrain]
* [http://pyml.sourceforge.net/ PyML]
* [https://pypi.python.org/pypi/ramp Rapid Machine Learning Prototyping]
* [https://pypi.python.org/pypi/Reinforcement-Learning-Toolkit Reinforcement-Learning-Toolkit]
* [https://pypi.python.org/pypi/tradingmachine tadingmachine - backtester for financial algorithms]
* [http://sourceforge.net/projects/weka/ Weka - Machine Learning Software in Java]
* [https://pypi.python.org/pypi/zipline zipline - backtester for financial algorithms]


== Work in progress ==
[https://github.com/h2oai/h2o-3 h2o] is an open source Java machine learning library with Python and R bindings.


* [http://mahout.apache.org/ Apache Mahout]
= Machine Learning Packages =
* [https://pypi.python.org/pypi/astroML astroML]
* [http://orange.biolab.si/ Orange]
* [http://www.clips.ua.ac.be/pages/pattern Pattern - Web mining module for Python]
* [http://scikit-learn.org/ scikit-learn - Machine Learning in Python]
* [http://shogun-toolbox.org/ SHOGUN]


== New packages ==
== New packages ==
Line 62: Line 84:
You can find them on the [[rhbug:ML-SIG|ML-SIG review-tracker]].
You can find them on the [[rhbug:ML-SIG|ML-SIG review-tracker]].


We would be glad, if you would take one or a few.  :)
We would be glad if you would take one or a few.  :)


== Existing packages ==
== Existing packages ==


You can find the existing [https://admin.fedoraproject.org/pkgdb/users/packages/ml-sig ml-related packages on the PkgDB].
We have a new FAS group [https://src.fedoraproject.org/group/machine-learning-sig machine-learning-sig]! Do you need any help maintaining your package? Let us know on the mailing list and give the group permissions for your package.
 
== Categories ==
 
more to come soon.


= What are we going to do? =
= What are we going to do? =


more to come soon.
* We're gonna update this wiki page so it can serve as a hub of interesting projects, links to important content, RPM packages etc. Do you know about anything we should have here? Let us know!
* We'll help ML developers as much as we can. Do you need some help or some new RPM package in Fedora? Let us know!
* We'll try to make Fedora the best distribution for AI/ML developers and users. Do you know how? Let us know!


= Participation =
= Participation =
Line 85: Line 105:


* Join: {{fplist|ml}}
* Join: {{fplist|ml}}
* Archives: [http://lists.fedoraproject.org/pipermail/ml/ read]
* Archives: [https://lists.fedoraproject.org/archives/list/ml@lists.fedoraproject.org/ read]


== IRC ==
== IRC ==
Line 92: Line 112:


Haven't used IRC for communication before? More information on how to use IRC is available [[IRC|here.]]
Haven't used IRC for communication before? More information on how to use IRC is available [[IRC|here.]]
== Telegram ==
We also have a Telegram group since Flock 2019. You can find it at [https://t.me/FedoraMLSIG t.me/FedoraMLSIG]
There is also a bridge between the Telegram group and the IRC channel so you can use only one platform to stay in touch.


== Meetings ==
== Meetings ==

Latest revision as of 16:03, 15 August 2023

The Machine Learning SIG is inactive and the information on this page is mostly out of date. The AI/ML SIG has replaced this SIG and will be the better reference going forward.


Machine Learning SIG

The Machine Learning SIG's goal is to make Fedora the best platform for all things related to Machine Learning. We aim to act a hub in the gap between the Astronomy, Bigdata, Fedora Medical and Science and Technology SIGs and also between all interesting projects related to machine learning in Fedora.

Members

Björn Esser (besser82) <besser82@fedoraproject.org>

Kushal Khandelwal (kushal124) <kushalkhandelwal10@gmail.com>

Christian Dersch (lupinix) <lupinix@fedoraproject.org>

Dhanesh B. Sabane (dhanesh95) <dhanesh95@fedoraproject.org>

Lumír Balhar (lbalhar) <lbalhar@redhat.com>

Christiano Anderson (canderson9) <canderson9@fedoraproject.org>

David Kirwan (dkirwan) <dkirwan@redhat.com>

Graham Williamson (willo) <willo@fedoraproject.org>

Tom Rix (trix) <trix@redhat.com>

Benson Muite (fed500) <fed500@fedoraproject.org>

Machine Learning Projects

Thoth

Thoth is a promising project which uses artificial intelligence to analyze and recommend software stack for artificial intelligence applications. It's based on an assumption that the proper combination of tools and libraries can have a significant impact on the performance of your project.

Thoth is now in the alpha state but it should be soon ready for beta testers.

Github organization: github.com/thoth-station

Optimized TensorFlow builds

The Artificial Intelligence Center of Excellence (AICoE) provides a simple Python package index with builds of Tensorflow optimized for Fedora/Centos/RHEL.

ML-Pack

ML-Pack is a lightweight C++ machine learning library that has Python and Julia bindings. It is available in Fedora repositories.

Shogun

Shogun is a machine learning toolbox with interfaces to a number of other languages including Python, Ruby and R.

Dlib

Dlib is an open source C++ machine learning library that can also be called from Python.

FANN

FANN(Fast Artificial Neural Network Library) is an open source C++ machine learning library with a wide variety of language bindings. It is available in Fedora repositories.

h2o

h2o is an open source Java machine learning library with Python and R bindings.

Machine Learning Packages

New packages

When submitting a new ml-related package for review, please add "Blocks: ML-SIG" to your review-request. After the review has been granted don't forget to remove it, when filing the SCM-request, please.

When you are filing your SCM-admin-request, you should make sure to request InitialCC for "ml-sig".

Example:

New Package SCM Request
=======================
Package Name: pkgname
Short Description: summary of package
Owners: foo bar
Branches: f18 f19 f20 el5 el6
InitialCC: ml-sig

Packages waiting for your review

You can find them on the ML-SIG review-tracker.

We would be glad if you would take one or a few.  :)

Existing packages

We have a new FAS group machine-learning-sig! Do you need any help maintaining your package? Let us know on the mailing list and give the group permissions for your package.

What are we going to do?

  • We're gonna update this wiki page so it can serve as a hub of interesting projects, links to important content, RPM packages etc. Do you know about anything we should have here? Let us know!
  • We'll help ML developers as much as we can. Do you need some help or some new RPM package in Fedora? Let us know!
  • We'll try to make Fedora the best distribution for AI/ML developers and users. Do you know how? Let us know!

Participation

There is no formal process for participating; joining the mailing list, hanging out on IRC, or participating in meetings are all fantastic ways to get involved.

A little self-introduction on the mailing list would be nice, too. And, if you want to, add yourself to our members-section above.

Mailing list

IRC

We will likely hang out on irc.freenode.net at #fedora-ml. German members may want to come into #fedora-ml-de, too.

Haven't used IRC for communication before? More information on how to use IRC is available here.

Telegram

We also have a Telegram group since Flock 2019. You can find it at t.me/FedoraMLSIG

There is also a bridge between the Telegram group and the IRC channel so you can use only one platform to stay in touch.

Meetings

We shall have them, and see how it goes.

more to come soon.