Merge pull request #99 from iotamudelta/master

Merge from upstream

Author: iotamudelta

.gitignore

@@ -141,10 +141,6 @@ docs/source/scripts/activation_images/
 # PyCharm files
 .idea
 
-# Visual Studio Code files
-.vscode
-.vs
-
 # OSX dir files
 .DS_Store
 
@@ -195,3 +191,11 @@ caffe2.egg-info
 
 # Atom/Watchman required file
 .watchmanconfig
+
+# BEGIN NOT-CLEAN-FILES (setup.py handles this marker. Do not change.)
+#
+# Below files are not deleted by "setup.py clean".
+
+# Visual Studio Code files
+.vscode
+.vs

README.md

@@ -148,7 +148,9 @@ If you want to compile with CUDA support, install
 If you want to disable CUDA support, export environment variable `NO_CUDA=1`.
 Other potentially useful environment variables may be found in `setup.py`.
 
-If you want to build on Windows, Visual Studio 2017 and NVTX are also needed.
+If you want to build on Windows, Visual Studio 2017 14.11 toolset and NVTX are also needed.
+Especially, for CUDA 8 build on Windows, there will be an additional requirement for VS 2015 Update 3 and a patch for it.
+The details of the patch can be found out [here](https://support.microsoft.com/en-gb/help/4020481/fix-link-exe-crashes-with-a-fatal-lnk1000-error-when-you-use-wholearch).
 
 #### Install optional dependencies
 
@@ -196,9 +198,10 @@ On Windows
 set "VS150COMNTOOLS=C:\Program Files (x86)\Microsoft Visual Studio\2017\Enterprise\VC\Auxiliary\Build"
 set CMAKE_GENERATOR=Visual Studio 15 2017 Win64
 set DISTUTILS_USE_SDK=1
-REM The following line is needed for Python 2.7, but the support for it is very experimental.
+REM The following two lines are needed for Python 2.7, but the support for it is very experimental.
 set MSSdk=1
-REM As for CUDA 8, VS2015 Update 2 or up is required to build PyTorch. Use the following two lines.
+set FORCE_PY27_BUILD=1
+REM As for CUDA 8, VS2015 Update 3 is also required to build PyTorch. Use the following two lines.
 set "PREBUILD_COMMAND=%VS140COMNTOOLS%\..\..\VC\vcvarsall.bat"
 set PREBUILD_COMMAND_ARGS=x64
 

aten/doc/Type.h

@@ -5,14 +5,14 @@
 #include <functional>
 
 #include "ATen/ATenGeneral.h"
-#include "ATen/ArrayRef.h"
+#include "ATen/Allocator.h"
 #include "ATen/Generator.h"
-#include "ATen/Half.h"
-#include "ATen/SparseTensorRef.h"
-#include "ATen/ScalarType.h"
 #include "ATen/Scalar.h"
+#include "ATen/ScalarType.h"
+#include "ATen/SparseTensorRef.h"
 #include "ATen/Tensor.h"
-#include "ATen/Allocator.h"
+#include "ATen/core/ArrayRef.h"
+#include "ATen/core/Half.h"
 
 // To solve the conflict of s_addr in inaddr.h
 #ifdef _MSC_VER

aten/src/ATen/ATen.h

@@ -1,25 +1,25 @@
 #pragma once
 
 #include "ATen/ATenGeneral.h"
-#include "ATen/CPUGeneral.h"
 #include "ATen/Allocator.h"
-#include "ATen/Scalar.h"
-#include "ATen/Type.h"
-#include "ATen/Generator.h"
+#include "ATen/CPUGeneral.h"
+#include "ATen/CUDAGuard.h"
 #include "ATen/Context.h"
+#include "ATen/Device.h"
+#include "ATen/DeviceGuard.h"
+#include "ATen/DimVector.h"
+#include "ATen/Dispatch.h"
+#include "ATen/Formatting.h"
+#include "ATen/Functions.h"
+#include "ATen/Generator.h"
+#include "ATen/Layout.h"
+#include "ATen/OptionsGuard.h"
+#include "ATen/Scalar.h"
 #include "ATen/Storage.h"
 #include "ATen/Tensor.h"
-#include "ATen/Device.h"
 #include "ATen/TensorGeometry.h"
-#include "ATen/Functions.h"
-#include "ATen/Formatting.h"
-#include "ATen/TensorOperators.h"
 #include "ATen/TensorMethods.h"
-#include "ATen/Dispatch.h"
-#include "ATen/DimVector.h"
-#include "ATen/DeviceGuard.h"
+#include "ATen/TensorOperators.h"
 #include "ATen/TensorOptions.h"
-#include "ATen/Layout.h"
-#include "ATen/OptionsGuard.h"
-#include "ATen/CUDAGuard.h"
-#include "ATen/Error.h"
+#include "ATen/Type.h"
+#include "ATen/core/Error.h"

aten/src/ATen/ATenGeneral.cpp

@@ -0,0 +1 @@
+#include <ATen/ATenGeneral.h>

aten/src/ATen/AccumulateType.h

@@ -1,6 +1,6 @@
 #pragma once
 #include "ATen/Config.h"
-#include "ATen/Half.h"
+#include "ATen/core/Half.h"
 
 // Defines the accumulation type for a scalar type.
 // Example:

aten/src/ATen/Allocator.h

@@ -3,9 +3,9 @@
 #include <memory>
 #include <stddef.h>
 
-#include <ATen/Error.h>
-#include <ATen/Retainable.h>
 #include <ATen/Device.h>
+#include <ATen/Retainable.h>
+#include <ATen/core/Error.h>
 #include <ATen/core/UniqueVoidPtr.h>
 
 namespace at {

aten/src/ATen/ArrayRef.cpp

@@ -1,7 +1,7 @@
 #pragma once
 
+#include "ATen/core/Error.h"
 #include "TH/TH.h"
-#include "ATen/Error.h"
 
 // This file creates a fake allocator that just throws exceptions if
 // it is actually used.

aten/src/ATen/CPUFixedAllocator.h

@@ -1,8 +1,8 @@
 #pragma once
 
-#include "ATen/Error.h"
 #include "ATen/Generator.h"
 #include "ATen/Utils.h"
+#include "ATen/core/Error.h"
 
 namespace at {