You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@mxnet.apache.org by GitBox <gi...@apache.org> on 2019/10/31 21:24:44 UTC
[GitHub] [incubator-mxnet] ptrendx edited a comment on issue #16542: Faster
GPU NMS operator
ptrendx edited a comment on issue #16542: Faster GPU NMS operator
URL: https://github.com/apache/incubator-mxnet/pull/16542#issuecomment-548574934
@zhreshold Do you have any other comments?
For the record - I think there are a few outstanding items to fully mark GPU implementation of NMS as "done", namely
- handle the output required for backward in the new implementation
- change the way sorting is done (currently sorting happens `batch size` times, and it turns out that for large topk it is not beneficial, the previous implementation that would always do 2 (larger) sorts was better in that regard.
That said, I will not be able to finish those 2 outstanding items before 1.6 code freeze and I believe performance improvements are enough to merge this PR and address those 2 points as part of the future PR.
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
users@infra.apache.org
With regards,
Apache Git Services