[
https://issues.apache.org/jira/browse/SYSTEMML-1466?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15979562#comment-15979562
]
Mike Dusenberry edited comment on SYSTEMML-1466 at 4/21/17 11:31 PM:
---------------------------------------------------------------------
Attaching a distributed convnet file, a python script for running the convnet
using dummy data of the same size as the breast cancer data, and an output of
the statistics w/ and w/o forced remote parfor loops.
I ran this on a Yarn cluster with 19 90GB executors. I experimented with the
driver size to see if I could force it to use a remote spark parfor loop, but
that didn't work.
To run, grab a copy of the {{nn}} library and place it next to the attached
files, then execute {{run_convnet_distrib_sgd.py}}.
was (Author: mwdus...@us.ibm.com):
Attaching a distributed convnet file, a python script for running the convnet
using dummy data of the same size as the breast cancer data, and an output of
the statistics w/ and w/o forced remote parfor loops.
To run, grab a copy of the {{nn}} library and place it next to the attached
files, then execute {{run_convnet_distrib_sgd.py}}.
> Update `convnet.dml` to use distributed SGD.
> --------------------------------------------
>
> Key: SYSTEMML-1466
> URL: https://issues.apache.org/jira/browse/SYSTEMML-1466
> Project: SystemML
> Issue Type: Sub-task
> Reporter: Mike Dusenberry
> Assignee: Mike Dusenberry
> Attachments: convnet_distrib_sgd.dml, run_convnet_distrib_sgd.py,
> run_convnet_distrib_sgd-stats.txt
>
>
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
