[
https://issues.apache.org/jira/browse/SPARK-4844?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14246749#comment-14246749
]
Sean Owen commented on SPARK-4844:
----------------------------------
No, it definitely does not. See {{PartitionwiseSampledRDD}}, and how it uses
{{BernoulliSampler}} and {{PoissonSampler}}, which are already pluggable if you
want. They use gap-sampling iterators. If that's the only change, I would close
this. The PR reinvents some of the classes above.
> SGD should support custom sampling.
> -----------------------------------
>
> Key: SPARK-4844
> URL: https://issues.apache.org/jira/browse/SPARK-4844
> Project: Spark
> Issue Type: Improvement
> Components: MLlib
> Reporter: Guoqiang Li
> Fix For: 1.3.0
>
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
