ptrendx opened a new pull request #14627: Speed up SequenceReverse URL: https://github.com/apache/incubator-mxnet/pull/14627 ## Description ## This PR makes the SequenceReverse operator use more parallelism in its execution, resulting in significant speedup. I tested it on the GNMT en-vi example from GluonNLP (https://gluon-nlp.mxnet.io/model_zoo/machine_translation/index.html#google-neural-machine-translation) where it resulted in end to end speed increase from 33k wps to 44-45k wps on a single V100. FYI @eric-haibin-lin @szha @fhieber @sbodenstein ## Checklist ## ### Essentials ### Please feel free to remove inapplicable items for your PR. - [x] Changes are complete (i.e. I finished coding on this PR) - [x] All changes have test coverage: - [x] To the my best knowledge, examples are either not affected by this change, or have been fixed to be compatible with this change
---------------------------------------------------------------- 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: [email protected] With regards, Apache Git Services
