[
https://issues.apache.org/jira/browse/SPARK-10009?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14698523#comment-14698523
]
Yanbo Liang commented on SPARK-10009:
-------------------------------------
[~kaisasak] I think what you means is Params which must be set as keyword
dictionary type if I understand you correctly. What I mean is Param such as [
scalingVec |
https://github.com/apache/spark/blob/master/mllib/src/main/scala/org/apache/spark/ml/feature/ElementwiseProduct.scala#L47
]
> PySpark Param of Vector type can be set with Python array or numpy.array
> ------------------------------------------------------------------------
>
> Key: SPARK-10009
> URL: https://issues.apache.org/jira/browse/SPARK-10009
> Project: Spark
> Issue Type: Improvement
> Components: ML, PySpark
> Reporter: Yanbo Liang
>
> If the type of Param in PySpark ML pipeline is Vector, we can set with Vector
> currently. We also need to support set it with Python array and numpy.array.
> It should be handled in the wrapper (_transfer_params_to_java).
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
