projects / platform / upstream / pytorch.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 5f1dd9e) | patch
Porting legacy reflection_pad2d to ATen
authorShen Li <shenli@fb.com>
Thu, 10 Jan 2019 04:53:03 +0000 (20:53 -0800)
committerFacebook Github Bot <facebook-github-bot@users.noreply.github.com>
Thu, 10 Jan 2019 04:55:27 +0000 (20:55 -0800)
commit23e28efed43d84366894fb6c6c963e64097730d8
tree97c4f57030fe9a0cd99153f38e05c14a4991e2d2tree | snapshot
parent5f1dd9e7436366d2ea4580ebbaaba1a70b22b4c9commit | diff
Porting legacy reflection_pad2d to ATen

Summary:
Other changes:
1. Avoided using `THCDeviceTensor` by re-calculating the mapping from cuda (blockIdx, threadIdx) to input/output tensor index.
2. Changed Camelcase naming to underscore naming.

Differential Revision: D13546803

fbshipit-source-id: 1df54f13e64934da3d803d9b6586bd5208d42d6d
13 files changed:
aten/src/ATen/native/LegacyNNDefinitions.cpp diff | blob | history
aten/src/ATen/native/ReflectionPad.cpp diff | blob | history
aten/src/ATen/native/cuda/ReflectionPad.cu diff | blob | history
aten/src/ATen/native/native_functions.yaml diff | blob | history
aten/src/ATen/nn.yaml diff | blob | history
aten/src/THCUNN/CMakeLists.txt diff | blob | history
aten/src/THCUNN/SpatialReflectionPadding.cu [deleted file] blob | history
aten/src/THCUNN/generic/SpatialReflectionPadding.cu [deleted file] blob | history
aten/src/THCUNN/generic/THCUNN.h diff | blob | history
aten/src/THNN/generic/SpatialReflectionPadding.c [deleted file] blob | history
aten/src/THNN/generic/THNN.h diff | blob | history
aten/src/THNN/init.cpp diff | blob | history
torch/nn/_functions/thnn/auto.py diff | blob | history
Domain: Machine Learning / ML Framework;