[
https://issues.apache.org/jira/browse/MADLIB-1418?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Frank McQuillan closed MADLIB-1418.
-----------------------------------
Resolution: Fixed
https://github.com/apache/madlib/pull/490
> DL: performance improvements to fit transition function
> -------------------------------------------------------
>
> Key: MADLIB-1418
> URL: https://issues.apache.org/jira/browse/MADLIB-1418
> Project: Apache MADlib
> Issue Type: Improvement
> Components: Deep Learning
> Reporter: Nikhil Kak
> Assignee: Nikhil Kak
> Priority: Major
> Fix For: v1.17
>
>
>
> {{We can speed up our per row/buffer time by not serializing the model
> weights as part of state except for the last row. For all rows except the
> last row, all we need from the model state is just the agg_image_count, so we
> can just set the state to agg_image_count for all rows except the last row.
> For the last row, the behavior remains the same as before}}
> {{}}
> {{With places10 20 msts gpdb6 (with 1 epoch for all msts), per iteration time
> went down from around 1700-1800 secs to 1600-1700 secs. }}
> {{This will be more pronounced when training with a bigger dataset like
> places365/imagenet.}}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
