[
https://issues.apache.org/jira/browse/SPARK-10014?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14741185#comment-14741185
]
Joseph K. Bradley commented on SPARK-10014:
-------------------------------------------
An optimization like a hashcode sounds cool, but it might be overkill. Let's
just rebroadcast on each predict/transform. If a user really needs to call
transform a bunch of times, they can handle broadcasting themselves; it
probably is a special use case. I can look at your PRs again once they are
updated. Thanks!
> ML model broadcasts should be stored in private vars
> ----------------------------------------------------
>
> Key: SPARK-10014
> URL: https://issues.apache.org/jira/browse/SPARK-10014
> Project: Spark
> Issue Type: Umbrella
> Components: ML, MLlib
> Reporter: Joseph K. Bradley
> Priority: Minor
>
> Multiple places in MLlib, we broadcast a model before prediction. Since
> prediction may be called many times, we should store the broadcast variable
> in a private var so that we broadcast at most once.
> I'll link subtasks for each problem case I find.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
