[
https://issues.apache.org/jira/browse/SPARK-61844?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ruifeng Zheng resolved SPARK-61844.
-----------------------------------
Fix Version/s: 4.5.0
Resolution: Fixed
Issue resolved by pull request 60602
[https://github.com/apache/spark/pull/60602]
> Support configurable intermediate storage levels in LDA
> -------------------------------------------------------
>
> Key: SPARK-61844
> URL: https://issues.apache.org/jira/browse/SPARK-61844
> Project: Spark
> Issue Type: Improvement
> Components: ML, PySpark
> Affects Versions: 5.0.0, 4.5.0
> Reporter: Ruifeng Zheng
> Assignee: Ruifeng Zheng
> Priority: Major
> Labels: pull-request-available
> Fix For: 4.5.0
>
>
> LDA currently persists intermediate training instances with a hard-coded
> MEMORY_AND_DISK storage level when the online optimizer is used and the input
> dataset is not already cached.
> Expose the shared intermediateStorageLevel parameter in the Scala and PySpark
> LDA APIs, including the Python constructor, setParams, and setter. Keep
> MEMORY_AND_DISK as the default and preserve the existing cached-input guard.
> This parameter applies to the online optimizer; EM's graph caching is outside
> this change.
> Extend the existing Scala and Python test suites. New API annotations use
> 4.5.0, following the branch-4.4 cut.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]