[
https://issues.apache.org/jira/browse/SINGA-49?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14693120#comment-14693120
]
ASF subversion and git services commented on SINGA-49:
------------------------------------------------------
Commit d5b6a30d0efd207da143d01ca4079f2caa2ba5f7 in incubator-singa's branch
refs/heads/master from Wei Wang
[ https://git-wip-us.apache.org/repos/asf?p=incubator-singa.git;h=d5b6a30 ]
SINGA-49 Fix a bug in HandlePutMsg func that sets param fields to invalid values
The Param::HandlePutMsg doesn't set the weight decay multiplier and learning
rate multiplier correctly.
The previous code doesn't use the values from the received messages.
Fixed the bug using the correct multipliers.
> Fix a bug in HandlePutMsg func that sets param fields to invalid values
> -----------------------------------------------------------------------
>
> Key: SINGA-49
> URL: https://issues.apache.org/jira/browse/SINGA-49
> Project: Singa
> Issue Type: Bug
> Reporter: wangwei
>
> The Param::HandlePutMsg doesn't set the weight decay multiplier and learning
> rate multiplier correctly. The current code doesn't use the values from the
> received messages.
> This bug is found when testing the CIFAR10 example. That example uses a large
> weight decay for top layers to avoid over-fitting. The test is fine if
> parameters are updated at worker side but gets over-fitting when parameters
> are updated at server side. The reason is that the weight decay multiplier is
> not set properly.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
