You are viewing a plain text version of this content. The canonical link for it is here.

Posted to submarine-dev@hadoop.apache.org by "Zac Zhou (JIRA)" <ji...@apache.org> on 2019/08/13 09:13:00 UTC

[jira] [Updated] (SUBMARINE-125) Mini-submarine failed to run mnist job due to lack of access to data url

     [ https://issues.apache.org/jira/browse/SUBMARINE-125?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Zac Zhou updated SUBMARINE-125:
-------------------------------
    Summary: Mini-submarine failed to run mnist job due to lack of access to data url  (was: Mini-submarine failed to run mnist test due to lack of access to data url)

> Mini-submarine failed to run mnist job due to lack of access to data url
> ------------------------------------------------------------------------
>
>                 Key: SUBMARINE-125
>                 URL: https://issues.apache.org/jira/browse/SUBMARINE-125
>             Project: Hadoop Submarine
>          Issue Type: Task
>            Reporter: Zac Zhou
>            Assignee: Zac Zhou
>            Priority: Major
>
> The default mnist data url for tensorflow is [https://storage.googleapis.com/cvdf-datasets/mnist/.|https://storage.googleapis.com/cvdf-datasets/mnist/]
> Mnist job needs to get data from that url before the training. If the network is unavailable, the job would fail in Mini-submarine.
> We can add a parameter for users to specify the mirrors of the data, so that submarine job can be tested more easily.



--
This message was sent by Atlassian JIRA
(v7.6.14#76016)