zhengruifeng created SPARK-30641:
------------------------------------
Summary: ML algs blockify input vectors
Key: SPARK-30641
URL: https://issues.apache.org/jira/browse/SPARK-30641
Project: Spark
Issue Type: New Feature
Components: ML, PySpark
Affects Versions: 3.0.0
Reporter: zhengruifeng
Assignee: zhengruifeng
stacking input vectors into blocks will benefit ML algs:
1, less RAM to persist datasets, since the overhead of object header is reduced;
2, optimization potential for impl, since high-level BLAS can be used; Proven
in ALS/MLP;
3, maybe a way to perform efficient mini-batch sampling (To be confirmed)
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
