[jira] [Commented] (SINGA-414) TfVgg16 OperationalError for Rafiki

["Ngin Yun Chuan (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SINGA-414?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16873733#comment-16873733 ] 

Ngin Yun Chuan commented on SINGA-414:
--------------------------------------

Hi [~Yisen], this should be resolved for the latest `v0.1.0` release at `master`. 

> TfVgg16 OperationalError for Rafiki
> -----------------------------------
>
>                 Key: SINGA-414
>                 URL: https://issues.apache.org/jira/browse/SINGA-414
>             Project: Singa
>          Issue Type: Bug
>            Reporter: Ng Yi Sen
>            Priority: Major
>         Attachments: Screenshot 2018-11-29 at 4.55.45 PM.png
>
>
> While attempting to run the example TfVgg16 script located in [examples/models/image_classification/|http://examples/models/image_classification/,], I encountered this error after the training and evaluation steps:
> {{OperationalError: out of memory}}{{DETAIL: }}{{Cannot enlarge string buffer containing 0 bytes by 1076732998 more bytes.}}
>  
> It seemed to arise from the PostgreSQL database when the client tried to send in a query larger than 1gb (1076732998 bytes in the screenshot). I am not exactly how to resolve this. Previously, there were also other dimensional input errors but I modified the scripts slightly with additional reshape functions so the training can proceed.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)