[
https://issues.apache.org/jira/browse/MXNET-7?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16354508#comment-16354508
]
Chris Olivier commented on MXNET-7:
-----------------------------------
Are the perf bottlenecks you're seeing in the python code up to and including
the entry into the C API call, or the API function itself that is ultimately
running the indexing/slicing op? I am assuming here that the indexing and
slicing is some NNVM-registerred op.
> NDArray cython refactoring
> --------------------------
>
> Key: MXNET-7
> URL: https://issues.apache.org/jira/browse/MXNET-7
> Project: Apache MXNet
> Issue Type: Improvement
> Reporter: Chris Olivier
> Priority: Major
> Labels: cython, performance
>
> Refactor some NDArray python code to cython
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
