You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@mxnet.apache.org by "Denis Davydenko (JIRA)" <ji...@apache.org> on 2019/01/14 21:11:00 UTC
[jira] [Commented] (MXNET-1292) Benchmark Between Image API and
Java Image processing library
[ https://issues.apache.org/jira/browse/MXNET-1292?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16742503#comment-16742503 ]
Denis Davydenko commented on MXNET-1292:
----------------------------------------
There is an effort going on with building preprocessing into model itself (by [~skm@apache.org]). That project will introduce an opportunity to run preprocessing on GPU. Once that is available - we will evaluate performance results of such preprocessing as benchmarking of how fast MXNet can run conversions vs OpenCV.
> Benchmark Between Image API and Java Image processing library
> -------------------------------------------------------------
>
> Key: MXNET-1292
> URL: https://issues.apache.org/jira/browse/MXNET-1292
> Project: Apache MXNet
> Issue Type: Improvement
> Components: Apache MXNet Java API, Apache MXNet Scala API
> Reporter: Qing Lan
> Priority: Minor
>
> Currently we have Image API from OpenCV and Java Image processing library that can both convert an image from byte array to NDArray. In this case, we need to do a benchmark to compare the performance of them.
> # Sequential inputs
> # Batch inputs
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@mxnet.apache.org
For additional commands, e-mail: issues-help@mxnet.apache.org