You are viewing a plain text version of this content. The canonical link for it is here.

Posted to commits@tvm.apache.org by GitBox <gi...@apache.org> on 2020/07/26 05:15:45 UTC

[GitHub] [incubator-tvm] Laurawly commented on pull request #6108: Fix CUDA Compute Function For `get_valid_counts` and `nms`

Laurawly commented on pull request #6108:
URL: https://github.com/apache/incubator-tvm/pull/6108#issuecomment-663936994


   I'm wondering what's the purpose of this PR. Currently, there's no correctness issue with `get_valid_counts` and `nms` in end-to-end performance of object detection related models using them. And based on the benchmarks shown in this [PR](https://github.com/apache/incubator-tvm/pull/5339), there's a lot of performance gain on recent changes. And as a summarization of that PR, we know data rearrangement on the GPU is super slow.


----------------------------------------------------------------
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