Skip to content
Heiko Strathmann edited this page Apr 29, 2016 · 26 revisions

Releases

This document contains TODO lists for the next release.

4.2 [Takechi Hanpeita]

Purpose:

  • Pre GSoC release, to make students get involved
  • Offers BSD license option
  • Last Shogun that does not require C++11 (which will be used for GSoC)

###TODO:

  • Make sure tests are green (include no warnings in libshogun and python modular)
  • BSD/GPL repository split
  • Fix remaining problems with meta example integration
  • Put new website online
  • See the milestone

Release text draft:

Dear all,

we are pleased to announce the 4.2 release of the Shogun Machine Learning Toolbox, in preparation of the coming Google Summer of Code 2016.

This release's biggest change is the new BSD license option. You can now use (slightly slimmer version of) Shogun under a non-viral BSD3-style license. All of Shogun remains available under the GPL license. For details, see here

The release also will be the last one which has C++11 as an optional requirement. From the next release on, this will be changed to a hard requirement.

As usual, there are a number of improvements, mostly coming from our GSoC applicants. A full list and the download can be found at https://github.com/shogun-toolbox/shogun/releases/tag/shogun_4.2.0

Heiko Strathmann, on behalf of the Shogun team

About Shogun: http://www.shogun-toolbox.org

Source code: https://github.com/shogun-toolbox

4.1 [DONE]

###Purpose:

  • Mostly contains Wu's new inference algorithms
  • Otherwise clean-up (drop migration, bug fixes)
  • Get back to speed.
  • Needed for pre GSoC 2016

###TODO:

  • Make sure tests are green
  • Make sure there are no fixable documentation warnings (Sergey, Wu)
  • Integrate new cookbook into main branch (not sure yet) (Sergey)
  • Add new algorithms to NEWS (Heiko)
  • Add clean-up patches to NEWS (Heiko)
  • Put new website online? (Kevin)
  • Announce move to stack overflow for questions (Heiko)
  • Write Release text (Heiko)

Release text draft:

Dear all,

we are pleased to announce the 4.1.0 release of the Shogun Machine Learning Toolbox.

This release features the work of the new Shogun-team member Wu Lin, who put a lot of work into Shogun's framework for large-scale variational inference for Gaussian Processes. A (very large) notebook explaining a lot of the methods can be found at http://www.shogun-toolbox.org/static/notebook/current/variational_classifier.html

The release also features a computer-science feature by Esben Sörig, who designed a meta-language to write Shogun examples in, which are then automatically translated to all of Shogun's interface languages. Documentation can be found at https://github.com/shogun-toolbox/shogun/wiki/Example_Generation

We also worked towards a smoother build process and did a number of internal efficiency improvements and clean-ups.

A full list and the download can be found at https://github.com/shogun-toolbox/shogun/releases/tag/shogun_4.1.0

On other news:

  • We plan to start using Stack Overflow to reply to Shogun questions, rather than the mailing list. This will create a more documented set of Q&A. Github issues are used as before to report/solve problems. See http://stackoverflow.com/questions/tagged/shogun
  • The Shogun foundation was recognised as a non-profit organisation by German tax law. This is an important step towards more sustainabl finances of the project.

Heiko Strathmann, on behalf of the Shogun team

About Shogun: http://www.shogun-toolbox.org

Source code: https://github.com/shogun-toolbox

Welcome to the Shogun wiki!

Clone this wiki locally