[
https://issues.apache.org/jira/browse/SYSTEMML-2403?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16520708#comment-16520708
]
Matthias Boehm commented on SYSTEMML-2403:
------------------------------------------
I just ran this scenario several times with different numbers of workers but
was not able to reproduce this (I always get {{Val Accuracy: 0.8915}}). Maybe a
previous change fixes some side effect? Let's keep this task open for now - if
we don't encounter this issue again in the next weeks, we can close it.
> Paramserv low accuracy sometimes occurred
> -----------------------------------------
>
> Key: SYSTEMML-2403
> URL: https://issues.apache.org/jira/browse/SYSTEMML-2403
> Project: SystemML
> Issue Type: Bug
> Reporter: LI Guobao
> Assignee: LI Guobao
> Priority: Major
>
> We could observe that the low accuracy will sometimes occurred. Here is the
> scenario: _BSP, BATCH, DISJOINT_CONTIGUOUS (or DISJOINT_RANDOM)_ with _1
> epoch, 4 workers (I have 4 vcores in my machine) and batchSize 16_ using 60k
> minist.
> {code:java}
> Val Loss: 2.3006845853187783
> Val Accuracy: 0.11183333333333334
> {code}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)