projects / platform / upstream / pytorch.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 1e0eab5) | patch
Removes THCNumerics usages in RNN.cu (#15085)
authorSyed Tousif Ahmed <syed.ahmed.emails@gmail.com>
Thu, 13 Dec 2018 08:19:13 +0000 (00:19 -0800)
committerFacebook Github Bot <facebook-github-bot@users.noreply.github.com>
Thu, 13 Dec 2018 08:24:17 +0000 (00:24 -0800)
commitaa022313cb694546cca7fcb0123a818bf2d668b9
tree38506ebec051504cf68954e33578de38990f62e1tree | snapshot
parent1e0eab5df839ae4124fa558c9c82586c9993eaabcommit | diff
Removes THCNumerics usages in RNN.cu (#15085)

Summary:
We don't need THCNumerics here since at::Half can be implicitly converted to float and the cuda math dispatches are handled by `/usr/local/cuda/include/crt/math_functions.hpp` and `cmath`. ATen should be free of THCNumerics after this and when porting kernels from THC, one should not use THCNumerics.

Should close: https://github.com/pytorch/pytorch/issues/11878
Pull Request resolved: https://github.com/pytorch/pytorch/pull/15085

Differential Revision: D13447558

Pulled By: soumith

fbshipit-source-id: 4ff5cbf838edcd01e2d1397e4d7f4f920e9e9fc3
aten/src/ATen/native/cuda/RNN.cu diff | blob | history
Domain: Machine Learning / ML Framework;