[
https://issues.apache.org/jira/browse/FLINK-1992?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14548246#comment-14548246
]
ASF GitHub Bot commented on FLINK-1992:
---------------------------------------
GitHub user thvasilo opened a pull request:
https://github.com/apache/flink/pull/692
[WIP] [FLINK-1992] [ml] Add convergence criterion to SGD optimizer
This a WIP PR in order to add convergence critertion to SGD. This will also
allow us to replace the custom SGD used in Multiple Linear Regression
(FLINK-1993)
You can merge this pull request into a Git repository by running:
$ git pull https://github.com/thvasilo/flink sgd-convegrence
Alternatively you can review and apply these changes as the patch at:
https://github.com/apache/flink/pull/692.patch
To close this pull request, make a commit to your master/trunk branch
with (at least) the following in the commit message:
This closes #692
----
commit e0eb199009cdf602bf4c96fc6c37f70560594d81
Author: Theodore Vasiloudis <t...@sics.se>
Date: 2015-05-13T15:29:31Z
Some refactoring for better generalization
commit c224cf9ecb35147175f3de4b188c03168f03fa1f
Author: Theodore Vasiloudis <t...@sics.se>
Date: 2015-05-15T20:11:48Z
Optimization refactoring
Removed runParameters, to avoid having to maintain two parameter maps.
Added apply methods.
Some renaming to avoid confusion between parameters and actual types.
commit 25ae46bf7e0789d9f13479ef984c15d75526223a
Author: Theodore Vasiloudis <t...@sics.se>
Date: 2015-05-18T15:27:16Z
Removed redundant arguments from LossFunction.lossAndGradient
commit 1b81d8f16c1778c8950e72d8de0026eed6e085fd
Author: Theodore Vasiloudis <t...@sics.se>
Date: 2015-05-18T16:30:54Z
Frist draft version of convergence
----
> Add convergence criterion to SGD optimizer
> ------------------------------------------
>
> Key: FLINK-1992
> URL: https://issues.apache.org/jira/browse/FLINK-1992
> Project: Flink
> Issue Type: Improvement
> Components: Machine Learning Library
> Reporter: Till Rohrmann
> Assignee: Theodore Vasiloudis
> Labels: ML
>
> Currently, Flink's SGD optimizer runs for a fixed number of iterations. It
> would be good to support a dynamic convergence criterion, too.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
