andrewadare / jetson-tx2-pytorch Goto Github PK

Installing PyTorch on the Nvidia Jetson TX1/TX2

License: MIT License

jetson-tx2-pytorch's Introduction

Installing PyTorch on the NVIDIA Jetson TX1/TX2

PyTorch is a new deep learning framework that runs very well on the Jetson TX1 and TX2 boards. It is relatively simple and quick to install. Unlike TensorFlow, it requires no external swap partition to build on the TX1.

Although the TX2 has an ample 32 GB of eMMC, the TX1 has only half that, and it is easy to run out of space due to cruft from JetPack, Ubuntu packages, and installation artifacts. The cleanup section below lists ways to slim things down, and the steps here lean in the direction of minimalism.

The PyTorch developers recommend the Anaconda distribution. I was unable to find a recent Anaconda setup for ARM64, so I used the global python libraries.

Tip: On the TX2, running ~/jetson_clocks.sh throttles up the CPUs and enables two more cores. This reduces the PyTorch compilation time from 45 to 37 minutes. I didn't test on the TX1, but would expect a less dramatic speedup.

To avoid issues with system-wide installation as superuser, I appended --user to all pip3 install commands below. This puts packages in $HOME/.local/lib/python3.5/site-packages, which I added to my PYTHONPATH.

Scipy and LA libs

sudo apt install libopenblas-dev libatlas-dev liblapack-dev
sudo apt install liblapacke-dev checkinstall # For OpenCV
pip3 install numpy scipy # ~20-30 min

Build tool prerequisites

pip3 install pyyaml
pip3 install scikit-build
sudo apt install ninja-build

CMake

Check cmake --version. It looks like cmake >= 3.6 is required for the python bindings used in PyTorch. I followed http://askubuntu.com/a/865294 (and used --no-check-certificate) to get 3.7.2. The cmake executable was installed to /usr/local/bin. Make sure cmake --version reports the new version after installation.

CFFI

sudo apt install python3-dev
sudo apt install libffi-dev
pip3 install cffi

OpenCV

I followed a subset of these excellent instructions for Python 3 from the pyimagesearch blog. I skipped CUDA and OpenCL integration, and went system-wide with the Python 3 bindings (no virtualenvs). Under the OpenCV source directory, I created build/ and ran this:

cmake \
    -D ENABLE_PRECOMPILED_HEADERS=OFF \
    -D WITH_OPENCL=OFF \
    -D WITH_CUDA=OFF \
    -D WITH_CUFFT=OFF \
    -D WITH_CUBLAS=OFF \
    -D CMAKE_BUILD_TYPE=RELEASE \
    -D CMAKE_INSTALL_PREFIX=/usr/local \
    -D INSTALL_PYTHON_EXAMPLES=OFF \
    -D INSTALL_C_EXAMPLES=OFF \
    -D PYTHON_EXECUTABLE=/usr/bin/python3.5 \
    -D BUILD_EXAMPLES=OFF ..

The first option significantly reduces the storage requirements during the build, at the expense of a slightly longer build time. Before discovering this option, I encountered errors like "failed to copy PCH file" on the TX1.

With this setup, OpenCV builds quickly--19 minutes on the TX2.

CUDA and cuDNN

The Jetson does not ship ready to run deep learning models on the GPU. Unfortunately CUDA and cuDNN cannot simply be downloaded directly as for x86 systems; NVIDIA directs us to the JetPack SDK, which currently runs only on Ubuntu 14.04. In a VirtualBox instance, I downloaded the JetPack .run file. The default full installation is massive and takes hours to download, build, and flash. Moreover, it copies many GB of files to the local VM which is not what I want! I realized I could avoid the overkill and just select the CUDA and cuDNN packages, then cancel the installation once the .deb files are downloaded. Only the following 4 files are needed to install CUDA and cuDNN:

cuda-repo-l4t-8-0-local_8.0.64-1_arm64.deb
libcudnn5_5.1.10-1+cuda8.0_arm64.deb
libcudnn5-dev_5.1.10-1+cuda8.0_arm64.deb
cuda-l4t.sh

I created ~/cuda-l4t on the Jetson and copied these 4 files there.

sudo ./cuda-l4t.sh cuda-repo...arm64.deb 8.0 8-0 to install the CUDA libs. Note that this script prepends CUDA directories to PATH and LD_LIBRARY_PATH, then slaps the hard-coded redefinitions onto ~/.bashrc. If it matters to you, now is a good time to tidy this up.
To get cuDNN, do sudo dpkg -i <debfile> for files 2 and 3 above. Then do

sudo apt install -f

Check: nvcc -V
ldconfig -p | grep cu and grep dnn can be used to show the library locations. I also see /usr/include/cudnn.h

cuDNN for PyTorch

Since tools/setup_helpers/cuda.py assumes /usr/local/cuda, CUDA_HOME need not be set. But the TX1/TX2 install locations are missing from the search list in tools/setup_helpers/cudnn.py. In that script, there is a check for os.getenv(CUDNN_LIB_DIR) and another for the include dir. So I added the following to ~/.profile:

export CUDNN_LIB_DIR=/usr/lib/aarch64-linux-gnu
export CUDNN_INCLUDE_DIR=/usr/include

Note: Echoing at the prompt is not sufficient to test if environment variables are visible to the PyTorch build setup scripts. If CUDNN_LIB_DIR is properly set for the bash environment, but you also see this:

[~]$ sudo python3 -c 'import os; print(os.getenv("CUDNN_LIB_DIR"))'
None

then cuDNN will fail to be included. Here are three options:

Skip the problem and install PyTorch locally with the --user option (see next section).
use sudo -E to make user environment variables available to root.
Export from ~/.profile, then log out/in.

PyTorch source

git clone https://github.com/pytorch/pytorch.git
git checkout -b v0.1.10 v0.1.10

Now build PyTorch:

time python3 setup.py install --user

For a system-wide install,

sudo time python3 setup.py install # or sudo -E

Test out in the python3 repl (outside pytorch direcory):

import torch
torch.backends.cudnn.is_acceptable(torch.cuda.FloatTensor(1))

If True, congratulations!

TorchVision

pip3 --no-deps install torchvision

Cleanup

The postFlashTX1 repo contains some useful cleanup scripts. In addition:

sudo apt clean
sudo apt autoremove --purge
sudo rm /usr/src/*.tbz2 ## I had 6.9 GB of zip files
sudo rm /var/cuda-repo-8.0-local/*.deb
rm ~/temp # From my CMake 3.7 install

The OpenCV sources can also be removed if necessary.

jetson-tx2-pytorch's People

Contributors

Stargazers

Watchers

Forkers

amore-hdu lareinahe lukasc-ch zaie keita1 andrewssobral davis-love-ai prashanthag empireofkings zuyingwo oycq jincy7 dongzhou-1996 dsp6414 li91yang frantic0 whh14 hxdaze

jetson-tx2-pytorch's Issues

AttributeError: module 'torch' has no attribute 'version'

I was able to build pytorch using your instruction set, but when testing the installation, I ran into a missing attribute error (see below):

Nvidia TX1
Ubuntu 16
Python 3.5
Cmake 3.7.2

torch==0.1.10+ac9245a
torchvision==0.2.2.post3

(cv40py35) nvidia@tegra-ubuntu:~$ which cmake
/usr/local/bin/cmake
(cv40py35) nvidia@tegra-ubuntu:~$ cmake --version
cmake version 3.7.2
Cuda 8
Cudnn 5

(cv40py35) nvidia@tegra-ubuntu:~$ python3 -c 'import sys; print("\n".join(sys.path))'

/home/nvidia/.virtualenvs/cv40py35/lib/python35.zip
/home/nvidia/.virtualenvs/cv40py35/lib/python3.5
/home/nvidia/.virtualenvs/cv40py35/lib/python3.5/plat-aarch64-linux-gnu
/home/nvidia/.virtualenvs/cv40py35/lib/python3.5/lib-dynload
/usr/lib/python3.5
/usr/lib/python3.5/plat-aarch64-linux-gnu
/home/nvidia/.virtualenvs/cv40py35/lib/python3.5/site-packages

(cv40py35) nvidia@tegra-ubuntu:/~$ python
Python 3.5.2 (default, Nov 23 2017, 16:37:01)
[GCC 5.4.0 20160609] on linux
Type "help", "copyright", "credits" or "license" for more information.

import torch
print(torch.version)
Traceback (most recent call last):
File "", line 1, in
AttributeError: module 'torch' has no attribute 'version'

pytorch-install.log

c++: error: unrecognized command line option ‘-mavx2’

[ 15%] Building CXX object src/ATen/CMakeFiles/ATen.dir/native/cpu/ReduceOpsKernel.cpp.AVX.cpp.o
c++: error: unrecognized command line option ‘-mavx2’
src/ATen/CMakeFiles/ATen.dir/build.make:81862: recipe for target 'src/ATen/CMakeFiles/ATen.dir/native/cpu/ReduceOpsKernel.cpp.AVX2.cpp.o' failed
make[2]: *** [src/ATen/CMakeFiles/ATen.dir/native/cpu/ReduceOpsKernel.cpp.AVX2.cpp.o] Error 1
make[2]: *** Waiting for unfinished jobs....
c++: error: unrecognized command line option ‘-mavx’
src/ATen/CMakeFiles/ATen.dir/build.make:81886: recipe for target 'src/ATen/CMakeFiles/ATen.dir/native/cpu/ReduceOpsKernel.cpp.AVX.cpp.o' failed
make[2]: *** [src/ATen/CMakeFiles/ATen.dir/native/cpu/ReduceOpsKernel.cpp.AVX.cpp.o] Error 1
CMakeFiles/Makefile2:233: recipe for target 'src/ATen/CMakeFiles/ATen.dir/all' failed
make[1]: *** [src/ATen/CMakeFiles/ATen.dir/all] Error 2
Makefile:129: recipe for target 'all' failed
make: *** [all] Error

Has anyone ever met this question? How to solve it?

CMake Error at ATen_generated_THCTensorMode.cu.o.Release.cmake:279 (message): Error generating file

Hi, I met the a error as following:
[ 18%] Building NVCC (Device) object src/ATen/CMakeFiles/ATen.dir//THC/ATen_generated_THCTensorScatterGather.cu.o
[ 18%] Building NVCC (Device) object src/ATen/CMakeFiles/ATen.dir//THC/ATen_generated_THCTensorTopK.cu.o
[ 18%] Building NVCC (Device) object src/ATen/CMakeFiles/ATen.dir//THC/ATen_generated_THCTensorSort.cu.o
[ 18%] Building NVCC (Device) object src/ATen/CMakeFiles/ATen.dir//THC/ATen_generated_THCTensorTypeUtils.cu.o
[ 19%] Building NVCC (Device) object src/ATen/CMakeFiles/ATen.dir//THC/ATen_generated_THCSortUtils.cu.o
[ 19%] Building NVCC (Device) object src/ATen/CMakeFiles/ATen.dir//THC/ATen_generated_THCTensorMode.cu.o
[ 19%] Building NVCC (Device) object src/ATen/CMakeFiles/ATen.dir//THC/generated/ATen_generated_THCTensorSortByte.cu.o
[ 19%] Building NVCC (Device) object src/ATen/CMakeFiles/ATen.dir//THC/generated/ATen_generated_THCTensorMathCompareTByte.cu.o
Killed
CMake Error at ATen_generated_THCTensorMode.cu.o.Release.cmake:279 (message):
Error generating file
/home/nvidia/pytorch/aten/build/src/ATen/CMakeFiles/ATen.dir/__/THC/./ATen_generated_THCTensorMode.cu.o

src/ATen/CMakeFiles/ATen.dir/build.make:210: recipe for target 'src/ATen/CMakeFiles/ATen.dir//THC/ATen_generated_THCTensorMode.cu.o' failed
make[2]: *** [src/ATen/CMakeFiles/ATen.dir//THC/ATen_generated_THCTensorMode.cu.o] Error 1
make[2]: *** Waiting for unfinished jobs....
CMakeFiles/Makefile2:233: recipe for target 'src/ATen/CMakeFiles/ATen.dir/all' failed
make[1]: *** [src/ATen/CMakeFiles/ATen.dir/all] Error 2
Makefile:127: recipe for target 'all' failed
make: *** [all] Error 2

have u met it?waiting for ur reply! thx

Install Error: "aarch64-linux-gnu-gcc: internal compiler error"

I ran into a compiler error using cmake 3.5.1. I might try to upgrade the compiler to 3.7...

(cv40py35) nvidia@tegra-ubuntu:~/pytorch$ cmake --version
cmake version 3.5.1

Error:
"aarch64-linux-gnu-gcc: internal compiler error (program cc1plus)"

This could just be a memory issue since I tried building without swap space (usb drive)...

-- Build files have been written to: /home/nvidia/pytorch/torch/lib/build/THPP
Scanning dependencies of target THPP
[ 11%] Building CXX object CMakeFiles/THPP.dir/Traits.cpp.o
[ 22%] Building CXX object CMakeFiles/THPP.dir/storages/THCStorage.cpp.o
[ 33%] Building CXX object CMakeFiles/THPP.dir/tensors/THCTensor.cpp.o
[ 44%] Building CXX object CMakeFiles/THPP.dir/storages/THStorage.cpp.o
[ 55%] Building CXX object CMakeFiles/THPP.dir/tensors/THTensor.cpp.o
[ 66%] Building CXX object CMakeFiles/THPP.dir/tensors/THCSTensor.cpp.o
[ 77%] Building CXX object CMakeFiles/THPP.dir/tensors/THSTensor.cpp.o
[ 88%] Building CXX object CMakeFiles/THPP.dir/TraitsCuda.cpp.o
[100%] Linking CXX shared library libTHPP.so
[100%] Built target THPP
Install the project...
-- Install configuration: "Release"
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/lib/libTHPP.so.1
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/lib/libTHPP.so
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/THPP.h
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/TraitsCuda.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/Storage.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/Tensor.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/Traits.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/storages/generic/THCStorage.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/storages/generic/THStorage.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/storages/THCStorage.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/storages/THStorage.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/tensors/generic/THCSTensor.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/tensors/generic/THTensor.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/tensors/generic/THSTensor.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/tensors/generic/THCTensor.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/tensors/THCSTensor.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/tensors/THTensor.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/tensors/THSTensor.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/tensors/THCTensor.hpp
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/THPP/Type.hpp
-- The C compiler identification is GNU 5.4.0
-- The CXX compiler identification is GNU 5.4.0
-- Check for working C compiler: /usr/bin/cc
-- Check for working C compiler: /usr/bin/cc -- works
-- Detecting C compiler ABI info
-- Detecting C compiler ABI info - done
-- Detecting C compile features
-- Detecting C compile features - done
-- Check for working CXX compiler: /usr/bin/c++
-- Check for working CXX compiler: /usr/bin/c++ -- works
-- Detecting CXX compiler ABI info
-- Detecting CXX compiler ABI info - done
-- Detecting CXX compile features
-- Detecting CXX compile features - done
-- Looking for clock_gettime in rt
-- Looking for clock_gettime in rt - found
-- Configuring done
-- Generating done
CMake Warning:
Manually-specified variables were not used by the project:

CUDA_NVCC_FLAGS
NO_CUDA
THCS_LIBRARIES
THCUNN_SO_VERSION
THC_LIBRARIES
THC_SO_VERSION
THD_SO_VERSION
THNN_SO_VERSION
THPP_LIBRARIES
THS_LIBRARIES
TH_INCLUDE_PATH
TH_LIB_PATH
TH_SO_VERSION
Torch_FOUND

-- Build files have been written to: /home/nvidia/pytorch/torch/lib/build/libshm
Scanning dependencies of target torch_shm_manager
Scanning dependencies of target shm
[ 25%] Building CXX object CMakeFiles/torch_shm_manager.dir/manager.cpp.o
[ 50%] Building CXX object CMakeFiles/shm.dir/core.cpp.o
/home/nvidia/pytorch/torch/lib/libshm/manager.cpp: In function 'void print_init_message(const char*)':
/home/nvidia/pytorch/torch/lib/libshm/manager.cpp:58:37: warning: ignoring return value of 'ssize_t write(int, const void*, size_t)', declared with attribute warn_unused_result [-Wunused-result]
write(1, message, strlen(message));
^
/home/nvidia/pytorch/torch/lib/libshm/manager.cpp:59:20: warning: ignoring return value of 'ssize_t write(int, const void*, size_t)', declared with attribute warn_unused_result [-Wunused-result]
write(1, "\n", 1);
^
[ 75%] Linking CXX shared library libshm.so
[ 75%] Built target shm
[100%] Linking CXX executable torch_shm_manager
CMakeFiles/torch_shm_manager.dir/manager.cpp.o: In function main': manager.cpp:(.text.startup+0x3c): warning: the use of tmpnam' is dangerous, better use `mkstemp'
[100%] Built target torch_shm_manager
Install the project...
-- Install configuration: "Release"
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/lib/libshm.so
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/include/libshm.h
-- Installing: /home/nvidia/pytorch/torch/lib/tmp_install/bin/torch_shm_manager
running build
running build_py
-- Building version 0.1.10+ac9245a
creating build
creating build/lib.linux-aarch64-3.5
creating build/lib.linux-aarch64-3.5/torch
copying torch/_torch_docs.py -> build/lib.linux-aarch64-3.5/torch
copying torch/_utils.py -> build/lib.linux-aarch64-3.5/torch
copying torch/storage.py -> build/lib.linux-aarch64-3.5/torch
copying torch/serialization.py -> build/lib.linux-aarch64-3.5/torch
copying torch/tensor.py -> build/lib.linux-aarch64-3.5/torch
copying torch/version.py -> build/lib.linux-aarch64-3.5/torch
copying torch/functional.py -> build/lib.linux-aarch64-3.5/torch
copying torch/_tensor_str.py -> build/lib.linux-aarch64-3.5/torch
copying torch/_tensor_docs.py -> build/lib.linux-aarch64-3.5/torch
copying torch/init.py -> build/lib.linux-aarch64-3.5/torch
creating build/lib.linux-aarch64-3.5/tools
copying tools/init.py -> build/lib.linux-aarch64-3.5/tools
creating build/lib.linux-aarch64-3.5/torch/legacy
copying torch/legacy/init.py -> build/lib.linux-aarch64-3.5/torch/legacy
creating build/lib.linux-aarch64-3.5/torch/utils
copying torch/utils/model_zoo.py -> build/lib.linux-aarch64-3.5/torch/utils
copying torch/utils/hooks.py -> build/lib.linux-aarch64-3.5/torch/utils
copying torch/utils/init.py -> build/lib.linux-aarch64-3.5/torch/utils
creating build/lib.linux-aarch64-3.5/torch/backends
copying torch/backends/init.py -> build/lib.linux-aarch64-3.5/torch/backends
creating build/lib.linux-aarch64-3.5/torch/multiprocessing
copying torch/multiprocessing/pool.py -> build/lib.linux-aarch64-3.5/torch/multiprocessing
copying torch/multiprocessing/init.py -> build/lib.linux-aarch64-3.5/torch/multiprocessing
copying torch/multiprocessing/reductions.py -> build/lib.linux-aarch64-3.5/torch/multiprocessing
copying torch/multiprocessing/queue.py -> build/lib.linux-aarch64-3.5/torch/multiprocessing
creating build/lib.linux-aarch64-3.5/torch/sparse
copying torch/sparse/init.py -> build/lib.linux-aarch64-3.5/torch/sparse
creating build/lib.linux-aarch64-3.5/torch/distributed
copying torch/distributed/remote_types.py -> build/lib.linux-aarch64-3.5/torch/distributed
copying torch/distributed/init.py -> build/lib.linux-aarch64-3.5/torch/distributed
copying torch/distributed/collectives.py -> build/lib.linux-aarch64-3.5/torch/distributed
creating build/lib.linux-aarch64-3.5/torch/_thnn
copying torch/_thnn/utils.py -> build/lib.linux-aarch64-3.5/torch/_thnn
copying torch/_thnn/init.py -> build/lib.linux-aarch64-3.5/torch/_thnn
creating build/lib.linux-aarch64-3.5/torch/autograd
copying torch/autograd/gradcheck.py -> build/lib.linux-aarch64-3.5/torch/autograd
copying torch/autograd/stochastic_function.py -> build/lib.linux-aarch64-3.5/torch/autograd
copying torch/autograd/function.py -> build/lib.linux-aarch64-3.5/torch/autograd
copying torch/autograd/engine.py -> build/lib.linux-aarch64-3.5/torch/autograd
copying torch/autograd/init.py -> build/lib.linux-aarch64-3.5/torch/autograd
copying torch/autograd/variable.py -> build/lib.linux-aarch64-3.5/torch/autograd
creating build/lib.linux-aarch64-3.5/torch/nn
copying torch/nn/parameter.py -> build/lib.linux-aarch64-3.5/torch/nn
copying torch/nn/functional.py -> build/lib.linux-aarch64-3.5/torch/nn
copying torch/nn/init.py -> build/lib.linux-aarch64-3.5/torch/nn
copying torch/nn/init.py -> build/lib.linux-aarch64-3.5/torch/nn
creating build/lib.linux-aarch64-3.5/torch/cuda
copying torch/cuda/sparse.py -> build/lib.linux-aarch64-3.5/torch/cuda
copying torch/cuda/nccl.py -> build/lib.linux-aarch64-3.5/torch/cuda
copying torch/cuda/streams.py -> build/lib.linux-aarch64-3.5/torch/cuda
copying torch/cuda/random.py -> build/lib.linux-aarch64-3.5/torch/cuda
copying torch/cuda/comm.py -> build/lib.linux-aarch64-3.5/torch/cuda
copying torch/cuda/init.py -> build/lib.linux-aarch64-3.5/torch/cuda
creating build/lib.linux-aarch64-3.5/torch/optim
copying torch/optim/adam.py -> build/lib.linux-aarch64-3.5/torch/optim
copying torch/optim/adagrad.py -> build/lib.linux-aarch64-3.5/torch/optim
copying torch/optim/adadelta.py -> build/lib.linux-aarch64-3.5/torch/optim
copying torch/optim/rmsprop.py -> build/lib.linux-aarch64-3.5/torch/optim
copying torch/optim/sgd.py -> build/lib.linux-aarch64-3.5/torch/optim
copying torch/optim/asgd.py -> build/lib.linux-aarch64-3.5/torch/optim
copying torch/optim/optimizer.py -> build/lib.linux-aarch64-3.5/torch/optim
copying torch/optim/init.py -> build/lib.linux-aarch64-3.5/torch/optim
copying torch/optim/adamax.py -> build/lib.linux-aarch64-3.5/torch/optim
copying torch/optim/lbfgs.py -> build/lib.linux-aarch64-3.5/torch/optim
copying torch/optim/rprop.py -> build/lib.linux-aarch64-3.5/torch/optim
creating build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/LeakyReLU.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Sum.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Concat.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Bilinear.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/CosineDistance.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/AddConstant.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/DistKLDivCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/DotProduct.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/WeightedEuclidean.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/MarginRankingCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Add.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/L1HingeEmbeddingCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/GradientReversal.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialAdaptiveMaxPooling.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/VolumetricFullConvolution.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Cosine.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Log.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SoftShrink.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/MaskedSelect.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/CMul.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialConvolution.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/L1Cost.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Clamp.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/BatchNormalization.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/utils.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/MM.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Unsqueeze.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Sequential.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/PartialLinear.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SelectTable.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/MulConstant.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialReflectionPadding.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Reshape.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Min.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Tanh.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Criterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/HardTanh.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/VolumetricAveragePooling.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/LookupTable.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialLPPooling.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SoftMarginCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialMaxPooling.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialReplicationPadding.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Sqrt.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SplitTable.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Replicate.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/CSubTable.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialSubtractiveNormalization.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Squeeze.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Index.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Narrow.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/MultiLabelMarginCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Abs.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/ParallelTable.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/VolumetricConvolution.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialUpSamplingNearest.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialCrossMapLRN.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/DepthConcat.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/VolumetricDropout.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialDivisiveNormalization.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/MSECriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/HardShrink.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Threshold.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/CosineEmbeddingCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Module.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Identity.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/MarginCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Mean.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/WeightedMSECriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/FlattenTable.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Contiguous.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Transpose.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/CrossEntropyCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialClassNLLCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SoftMax.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/CMulTable.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialSubSampling.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Dropout.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialFractionalMaxPooling.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/CDivTable.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/TemporalConvolution.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/JoinTable.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/NarrowTable.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/VolumetricBatchNormalization.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialDilatedConvolution.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/CAddTable.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/MultiLabelSoftMarginCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialFullConvolution.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/LogSoftMax.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/MV.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/MixtureTable.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/PReLU.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialConvolutionLocal.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Container.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialAveragePooling.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/RReLU.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Linear.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Exp.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Padding.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Max.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Select.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialBatchNormalization.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/AbsCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/LogSigmoid.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/CriterionTable.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Square.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SmoothL1Criterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialDropout.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Euclidean.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/View.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/ClassNLLCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/MultiMarginCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/TemporalMaxPooling.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SoftSign.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/ReLU6.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialSoftMax.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/VolumetricMaxPooling.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Power.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialFullConvolutionMap.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Copy.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/init.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/ReLU.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Mul.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/MultiCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/VolumetricMaxUnpooling.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/PairwiseDistance.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialMaxUnpooling.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Parallel.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialConvolutionMap.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SoftPlus.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/HingeEmbeddingCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/TanhShrink.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/BCECriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Normalize.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/VolumetricReplicationPadding.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialContrastiveNormalization.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SpatialZeroPadding.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/Sigmoid.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/ConcatTable.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/TemporalSubSampling.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/L1Penalty.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/ClassSimplexCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/ParallelCriterion.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/ELU.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
copying torch/legacy/nn/SoftMin.py -> build/lib.linux-aarch64-3.5/torch/legacy/nn
creating build/lib.linux-aarch64-3.5/torch/legacy/optim
copying torch/legacy/optim/adam.py -> build/lib.linux-aarch64-3.5/torch/legacy/optim
copying torch/legacy/optim/nag.py -> build/lib.linux-aarch64-3.5/torch/legacy/optim
copying torch/legacy/optim/adagrad.py -> build/lib.linux-aarch64-3.5/torch/legacy/optim
copying torch/legacy/optim/adadelta.py -> build/lib.linux-aarch64-3.5/torch/legacy/optim
copying torch/legacy/optim/rmsprop.py -> build/lib.linux-aarch64-3.5/torch/legacy/optim
copying torch/legacy/optim/sgd.py -> build/lib.linux-aarch64-3.5/torch/legacy/optim
copying torch/legacy/optim/asgd.py -> build/lib.linux-aarch64-3.5/torch/legacy/optim
copying torch/legacy/optim/cg.py -> build/lib.linux-aarch64-3.5/torch/legacy/optim
copying torch/legacy/optim/init.py -> build/lib.linux-aarch64-3.5/torch/legacy/optim
copying torch/legacy/optim/adamax.py -> build/lib.linux-aarch64-3.5/torch/legacy/optim
copying torch/legacy/optim/lbfgs.py -> build/lib.linux-aarch64-3.5/torch/legacy/optim
copying torch/legacy/optim/rprop.py -> build/lib.linux-aarch64-3.5/torch/legacy/optim
creating build/lib.linux-aarch64-3.5/torch/utils/ffi
copying torch/utils/ffi/init.py -> build/lib.linux-aarch64-3.5/torch/utils/ffi
creating build/lib.linux-aarch64-3.5/torch/utils/trainer
copying torch/utils/trainer/trainer.py -> build/lib.linux-aarch64-3.5/torch/utils/trainer
copying torch/utils/trainer/init.py -> build/lib.linux-aarch64-3.5/torch/utils/trainer
creating build/lib.linux-aarch64-3.5/torch/utils/serialization
copying torch/utils/serialization/read_lua_file.py -> build/lib.linux-aarch64-3.5/torch/utils/serialization
copying torch/utils/serialization/init.py -> build/lib.linux-aarch64-3.5/torch/utils/serialization
creating build/lib.linux-aarch64-3.5/torch/utils/data
copying torch/utils/data/dataloader.py -> build/lib.linux-aarch64-3.5/torch/utils/data
copying torch/utils/data/dataset.py -> build/lib.linux-aarch64-3.5/torch/utils/data
copying torch/utils/data/init.py -> build/lib.linux-aarch64-3.5/torch/utils/data
copying torch/utils/data/sampler.py -> build/lib.linux-aarch64-3.5/torch/utils/data
creating build/lib.linux-aarch64-3.5/torch/utils/trainer/plugins
copying torch/utils/trainer/plugins/progress.py -> build/lib.linux-aarch64-3.5/torch/utils/trainer/plugins
copying torch/utils/trainer/plugins/accuracy.py -> build/lib.linux-aarch64-3.5/torch/utils/trainer/plugins
copying torch/utils/trainer/plugins/plugin.py -> build/lib.linux-aarch64-3.5/torch/utils/trainer/plugins
copying torch/utils/trainer/plugins/monitor.py -> build/lib.linux-aarch64-3.5/torch/utils/trainer/plugins
copying torch/utils/trainer/plugins/logger.py -> build/lib.linux-aarch64-3.5/torch/utils/trainer/plugins
copying torch/utils/trainer/plugins/loss.py -> build/lib.linux-aarch64-3.5/torch/utils/trainer/plugins
copying torch/utils/trainer/plugins/time.py -> build/lib.linux-aarch64-3.5/torch/utils/trainer/plugins
copying torch/utils/trainer/plugins/init.py -> build/lib.linux-aarch64-3.5/torch/utils/trainer/plugins
creating build/lib.linux-aarch64-3.5/torch/backends/cudnn
copying torch/backends/cudnn/rnn.py -> build/lib.linux-aarch64-3.5/torch/backends/cudnn
copying torch/backends/cudnn/init.py -> build/lib.linux-aarch64-3.5/torch/backends/cudnn
creating build/lib.linux-aarch64-3.5/torch/autograd/_functions
copying torch/autograd/_functions/stochastic.py -> build/lib.linux-aarch64-3.5/torch/autograd/_functions
copying torch/autograd/_functions/reduce.py -> build/lib.linux-aarch64-3.5/torch/autograd/_functions
copying torch/autograd/_functions/tensor.py -> build/lib.linux-aarch64-3.5/torch/autograd/_functions
copying torch/autograd/_functions/blas.py -> build/lib.linux-aarch64-3.5/torch/autograd/_functions
copying torch/autograd/_functions/basic_ops.py -> build/lib.linux-aarch64-3.5/torch/autograd/_functions
copying torch/autograd/_functions/pointwise.py -> build/lib.linux-aarch64-3.5/torch/autograd/_functions
copying torch/autograd/_functions/linalg.py -> build/lib.linux-aarch64-3.5/torch/autograd/_functions
copying torch/autograd/_functions/init.py -> build/lib.linux-aarch64-3.5/torch/autograd/_functions
copying torch/autograd/_functions/compare.py -> build/lib.linux-aarch64-3.5/torch/autograd/_functions
creating build/lib.linux-aarch64-3.5/torch/nn/utils
copying torch/nn/utils/clip_grad.py -> build/lib.linux-aarch64-3.5/torch/nn/utils
copying torch/nn/utils/rnn.py -> build/lib.linux-aarch64-3.5/torch/nn/utils
copying torch/nn/utils/init.py -> build/lib.linux-aarch64-3.5/torch/nn/utils
creating build/lib.linux-aarch64-3.5/torch/nn/backends
copying torch/nn/backends/backend.py -> build/lib.linux-aarch64-3.5/torch/nn/backends
copying torch/nn/backends/thnn.py -> build/lib.linux-aarch64-3.5/torch/nn/backends
copying torch/nn/backends/init.py -> build/lib.linux-aarch64-3.5/torch/nn/backends
creating build/lib.linux-aarch64-3.5/torch/nn/_functions
copying torch/nn/_functions/conv.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions
copying torch/nn/_functions/padding.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions
copying torch/nn/_functions/dropout.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions
copying torch/nn/_functions/loss.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions
copying torch/nn/_functions/rnn.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions
copying torch/nn/_functions/batchnorm.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions
copying torch/nn/_functions/linear.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions
copying torch/nn/_functions/init.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions
copying torch/nn/_functions/activation.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions
creating build/lib.linux-aarch64-3.5/torch/nn/parallel
copying torch/nn/parallel/data_parallel.py -> build/lib.linux-aarch64-3.5/torch/nn/parallel
copying torch/nn/parallel/replicate.py -> build/lib.linux-aarch64-3.5/torch/nn/parallel
copying torch/nn/parallel/parallel_apply.py -> build/lib.linux-aarch64-3.5/torch/nn/parallel
copying torch/nn/parallel/scatter_gather.py -> build/lib.linux-aarch64-3.5/torch/nn/parallel
copying torch/nn/parallel/_functions.py -> build/lib.linux-aarch64-3.5/torch/nn/parallel
copying torch/nn/parallel/init.py -> build/lib.linux-aarch64-3.5/torch/nn/parallel
creating build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/container.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/sparse.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/conv.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/utils.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/module.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/padding.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/upsampling.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/pooling.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/distance.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/pixelshuffle.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/dropout.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/normalization.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/loss.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/rnn.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/batchnorm.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/linear.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/init.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
copying torch/nn/modules/activation.py -> build/lib.linux-aarch64-3.5/torch/nn/modules
creating build/lib.linux-aarch64-3.5/torch/nn/_functions/thnn
copying torch/nn/_functions/thnn/sparse.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions/thnn
copying torch/nn/_functions/thnn/auto.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions/thnn
copying torch/nn/_functions/thnn/upsampling.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions/thnn
copying torch/nn/_functions/thnn/pooling.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions/thnn
copying torch/nn/_functions/thnn/normalization.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions/thnn
copying torch/nn/_functions/thnn/loss.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions/thnn
copying torch/nn/_functions/thnn/init.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions/thnn
copying torch/nn/_functions/thnn/activation.py -> build/lib.linux-aarch64-3.5/torch/nn/_functions/thnn
creating build/lib.linux-aarch64-3.5/torch/lib
copying torch/lib/libshm.so -> build/lib.linux-aarch64-3.5/torch/lib
copying torch/lib/libTHNN.so.1 -> build/lib.linux-aarch64-3.5/torch/lib
copying torch/lib/libTHCUNN.so.1 -> build/lib.linux-aarch64-3.5/torch/lib
copying torch/lib/libTHS.so.1 -> build/lib.linux-aarch64-3.5/torch/lib
copying torch/lib/libTH.so.1 -> build/lib.linux-aarch64-3.5/torch/lib
copying torch/lib/libnccl.so -> build/lib.linux-aarch64-3.5/torch/lib
copying torch/lib/libTHC.so.1 -> build/lib.linux-aarch64-3.5/torch/lib
copying torch/lib/libTHCS.so.1 -> build/lib.linux-aarch64-3.5/torch/lib
copying torch/lib/libTHPP.so.1 -> build/lib.linux-aarch64-3.5/torch/lib
copying torch/lib/torch_shm_manager -> build/lib.linux-aarch64-3.5/torch/lib
copying torch/lib/THCUNN.h -> build/lib.linux-aarch64-3.5/torch/lib
copying torch/lib/THNN.h -> build/lib.linux-aarch64-3.5/torch/lib
creating build/lib.linux-aarch64-3.5/torch/lib/include
creating build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THFilePrivate.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THRandom.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THStorage.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/TH.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THTensorMacros.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THMath.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THDiskFile.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THGeneral.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THAtomic.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THFile.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THVector.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THTensor.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THTensorApply.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THGenerateHalfType.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THGenerateAllTypes.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THHalf.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THMemoryFile.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THTensorDimApply.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THBlas.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THAllocator.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THGenerateIntTypes.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THLapack.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THGenerateFloatTypes.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
copying torch/lib/include/TH/THLogAdd.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH
creating build/lib.linux-aarch64-3.5/torch/lib/include/TH/generic
copying torch/lib/include/TH/generic/THTensorMath.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH/generic
copying torch/lib/include/TH/generic/THStorage.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH/generic
copying torch/lib/include/TH/generic/THTensorLapack.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH/generic
copying torch/lib/include/TH/generic/THTensorCopy.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH/generic
copying torch/lib/include/TH/generic/THVector.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH/generic
copying torch/lib/include/TH/generic/THTensor.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH/generic
copying torch/lib/include/TH/generic/THTensorRandom.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH/generic
copying torch/lib/include/TH/generic/THStorageCopy.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH/generic
copying torch/lib/include/TH/generic/THTensorConv.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH/generic
copying torch/lib/include/TH/generic/THBlas.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH/generic
copying torch/lib/include/TH/generic/THLapack.h -> build/lib.linux-aarch64-3.5/torch/lib/include/TH/generic
creating build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCGenerateFloatType.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCAllocator.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCSleep.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCGenerateByteType.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCGeneral.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCGenerateHalfType.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCTensorCopy.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCTensorConv.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCStream.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCCachingAllocator.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCCachingHostAllocator.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCStorage.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCGenerateShortType.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCGenerateLongType.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCHalf.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCTensorTopK.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCGenerateAllTypes.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCGenerateFloatTypes.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THC.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCThreadLocal.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCTensorRandom.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCGenerateIntType.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCStorageCopy.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCGenerateCharType.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCTensor.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCBlas.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCTensorMath.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
copying torch/lib/include/THC/THCGenerateDoubleType.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC
creating build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorMathReduce.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorMasked.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorCopy.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorMathCompareT.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorScatterGather.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCStorage.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorMathCompare.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorIndex.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorMathPointwise.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorMathBlas.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorRandom.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCStorageCopy.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorSort.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorMathScan.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorMathPairwise.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensor.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorMathMagma.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
copying torch/lib/include/THC/generic/THCTensorMath.h -> build/lib.linux-aarch64-3.5/torch/lib/include/THC/generic
running build_ext
-- Building with NumPy bindings
-- Detected cuDNN at /usr/lib/aarch64-linux-gnu, /usr/include
-- Detected CUDA at /usr/local/cuda-8.0
building 'torch._C' extension
creating build/temp.linux-aarch64-3.5
creating build/temp.linux-aarch64-3.5/torch
creating build/temp.linux-aarch64-3.5/torch/csrc
creating build/temp.linux-aarch64-3.5/torch/csrc/utils
creating build/temp.linux-aarch64-3.5/torch/csrc/autograd
creating build/temp.linux-aarch64-3.5/torch/csrc/autograd/functions
creating build/temp.linux-aarch64-3.5/torch/csrc/nn
creating build/temp.linux-aarch64-3.5/torch/csrc/cuda
creating build/temp.linux-aarch64-3.5/torch/csrc/cudnn
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/PtrWrapper.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/PtrWrapper.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/Size.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/Size.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/Storage.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/Storage.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/utils.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/utils.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/Module.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/Module.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/Exceptions.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/Exceptions.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/Tensor.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/Tensor.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/utils/object_ptr.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/utils/object_ptr.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/Generator.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/Generator.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/allocators.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/allocators.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/serialization.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/serialization.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/autograd/function.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/autograd/function.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/autograd/variable.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/autograd/variable.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/DynamicTypes.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/DynamicTypes.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/autograd/init.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/autograd/init.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/autograd/grad_buffer.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/autograd/grad_buffer.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/autograd/engine.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/autograd/engine.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/byte_order.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/byte_order.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/autograd/python_function.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/autograd/python_function.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/autograd/python_engine.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/autograd/python_engine.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/autograd/functions/batch_normalization.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/autograd/functions/batch_normalization.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/nn/THNN_generic.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/nn/THNN_generic.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/autograd/functions/init.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/autograd/functions/init.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/autograd/python_cpp_function.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/autograd/python_cpp_function.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/cuda/Stream.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/cuda/Stream.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/autograd/python_variable.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/autograd/python_variable.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/cuda/Tensor.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/cuda/Tensor.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/cuda/utils.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/cuda/utils.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/cuda/Module.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/cuda/Module.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/cuda/serialization.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/cuda/serialization.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/cuda/Storage.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/cuda/Storage.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/cudnn/BatchNorm.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/cudnn/BatchNorm.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/cudnn/Conv.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/cudnn/Conv.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++
aarch64-linux-gnu-gcc: internal compiler error: Killed (program cc1plus)
Please submit a full bug report,
with preprocessed source if appropriate.
See file:///usr/share/doc/gcc-5/README.Bugs for instructions.
aarch64-linux-gnu-gcc -pthread -DNDEBUG -g -fwrapv -O2 -Wall -Wstrict-prototypes -g -fstack-protector-strong -Wformat -Werror=format-security -Wdate-time -D_FORTIFY_SOURCE=2 -fPIC -I/home/nvidia/pytorch -I/home/nvidia/pytorch/torch/csrc -I/home/nvidia/pytorch/torch/lib/tmp_install/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/TH -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THPP -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THNN -I/usr/local/lib/python3.5/dist-packages/numpy/core/include -I/usr/local/cuda-8.0/include -I/home/nvidia/pytorch/torch/lib/tmp_install/include/THCUNN -I/usr/include -I/usr/include/python3.5m -c torch/csrc/cudnn/Handles.cpp -o build/temp.linux-aarch64-3.5/torch/csrc/cudnn/Handles.o -D_THP_CORE -std=c++11 -Wno-write-strings -DWITH_NUMPY -DWITH_CUDA -DCUDA_LIB_PATH=/usr/local/cuda-8.0/lib64 -DWITH_CUDNN
error: command 'aarch64-linux-gnu-gcc' failed with exit status 4
cc1plus: warning: command line option '-Wstrict-prototypes' is valid for C/ObjC but not for C++

parse error in template argument list

Hi,

When I run 'python setup.py install --user' ,it output an error:
[ 85%] Building NVCC (Device) object src/ATen/CMakeFiles/ATen_cuda.dir/native/cuda/ATen_cuda_generated_TensorFactories.cu.o
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:41:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SoftMaxForward<2, ::at::cuda::type , ::at::acc_type<double, (bool)1> , ::at::native::operator ::LogSoftMaxForwardEpilogue>( _ZN2at4cuda4typeIdEE *&__cuda_0,_ZN2at4cuda4typeIdEE *&__cuda_1,int &__cuda_2){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462619cunn_SoftMaxForwardILi2EddNS1_25LogSoftMaxForwardEpilogueEEEvPT0_S5_i( (_ZN2at4cuda4typeIdEE &)__cuda_0,(_ZN2at4cuda4typeIdEE &)__cuda_1,(int &)__cuda_2);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:41:17: error: template-id ‘__wrapper__device_stub_cunn_SoftMaxForward<2, at::cuda::type, at::acc_type<double, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SoftMaxForward(_ZN2at4cuda4typeIdEE&, _ZN2at4cuda4typeIdEE&, int&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:46:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SoftMaxForward<2, ::at::cuda::type , ::at::acc_type<float, (bool)1> , ::at::native::operator ::LogSoftMaxForwardEpilogue>( _ZN2at4cuda4typeIfEE *&__cuda_0,_ZN2at4cuda4typeIfEE *&__cuda_1,int &__cuda_2){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462619cunn_SoftMaxForwardILi2EffNS1_25LogSoftMaxForwardEpilogueEEEvPT0_S5_i( (_ZN2at4cuda4typeIfEE &)__cuda_0,(_ZN2at4cuda4typeIfEE &)__cuda_1,(int &)__cuda_2);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:46:17: error: template-id ‘__wrapper__device_stub_cunn_SoftMaxForward<2, at::cuda::type, at::acc_type<float, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SoftMaxForward(_ZN2at4cuda4typeIfEE&, _ZN2at4cuda4typeIfEE&, int&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:51:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SoftMaxForward<2, ::at::cuda::type< ::at::Half> , ::at::acc_type< ::__half, (bool)1> , ::at::native::operator ::LogSoftMaxForwardEpilogue>( _ZN2at4cuda4typeINS_4HalfEEE *&__cuda_0,_ZN2at4cuda4typeINS_4HalfEEE *&__cuda_1,int &__cuda_2){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462619cunn_SoftMaxForwardILi2E6__halffNS1_25LogSoftMaxForwardEpilogueEEEvPT0_S6_i( (_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_0,(_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_1,(int &)__cuda_2);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:51:17: error: template-id ‘__wrapper__device_stub_cunn_SoftMaxForward<2, at::cuda::typeat::Half, at::acc_type<__half, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SoftMaxForward(_ZN2at4cuda4typeINS_4HalfEEE&, _ZN2at4cuda4typeINS_4HalfEEE&, int&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:56:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SpatialSoftMaxForward< ::at::cuda::type , ::at::acc_type<double, (bool)1> , ::at::native::operator ::LogSoftMaxForwardEpilogue>( _ZN2at4cuda4typeIdEE *&__cuda_0,_ZN2at4cuda4typeIdEE *&__cuda_1,::uint32_t &__cuda_2,::uint32_t &__cuda_3,::uint32_t &__cuda_4){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462626cunn_SpatialSoftMaxForwardIddNS1_25LogSoftMaxForwardEpilogueEEEvPT_S5_jjj( (_ZN2at4cuda4typeIdEE &)__cuda_0,(_ZN2at4cuda4typeIdEE &)__cuda_1,(::uint32_t &)__cuda_2,(::uint32_t &)__cuda_3,(::uint32_t &)__cuda_4);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:56:17: error: template-id ‘__wrapper__device_stub_cunn_SpatialSoftMaxForward<at::cuda::type, at::acc_type<double, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SpatialSoftMaxForward(_ZN2at4cuda4typeIdEE&, _ZN2at4cuda4typeIdEE&, uint32_t&, uint32_t&, uint32_t&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:61:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SpatialSoftMaxForward< ::at::cuda::type , ::at::acc_type<float, (bool)1> , ::at::native::operator ::LogSoftMaxForwardEpilogue>( _ZN2at4cuda4typeIfEE *&__cuda_0,_ZN2at4cuda4typeIfEE *&__cuda_1,::uint32_t &__cuda_2,::uint32_t &__cuda_3,::uint32_t &__cuda_4){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462626cunn_SpatialSoftMaxForwardIffNS1_25LogSoftMaxForwardEpilogueEEEvPT_S5_jjj( (_ZN2at4cuda4typeIfEE &)__cuda_0,(_ZN2at4cuda4typeIfEE &)__cuda_1,(::uint32_t &)__cuda_2,(::uint32_t &)__cuda_3,(::uint32_t &)__cuda_4);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:61:17: error: template-id ‘__wrapper__device_stub_cunn_SpatialSoftMaxForward<at::cuda::type, at::acc_type<float, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SpatialSoftMaxForward(_ZN2at4cuda4typeIfEE&, _ZN2at4cuda4typeIfEE&, uint32_t&, uint32_t&, uint32_t&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:66:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SpatialSoftMaxForward< ::at::cuda::type< ::at::Half> , ::at::acc_type< ::__half, (bool)1> , ::at::native::operator ::LogSoftMaxForwardEpilogue>( _ZN2at4cuda4typeINS_4HalfEEE *&__cuda_0,_ZN2at4cuda4typeINS_4HalfEEE *&__cuda_1,::uint32_t &__cuda_2,::uint32_t &__cuda_3,::uint32_t &__cuda_4){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462626cunn_SpatialSoftMaxForwardI6__halffNS1_25LogSoftMaxForwardEpilogueEEEvPT_S6_jjj( (_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_0,(_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_1,(::uint32_t &)__cuda_2,(::uint32_t &)__cuda_3,(::uint32_t &)__cuda_4);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:66:17: error: template-id ‘__wrapper__device_stub_cunn_SpatialSoftMaxForward<at::cuda::typeat::Half, at::acc_type<__half, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SpatialSoftMaxForward(_ZN2at4cuda4typeINS_4HalfEEE&, _ZN2at4cuda4typeINS_4HalfEEE&, uint32_t&, uint32_t&, uint32_t&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:71:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SoftMaxBackward<2, ::at::cuda::type , ::at::acc_type<double, (bool)1> , ::at::native::operator ::LogSoftMaxBackwardEpilogue>( _ZN2at4cuda4typeIdEE *&__cuda_0,_ZN2at4cuda4typeIdEE *&__cuda_1,_ZN2at4cuda4typeIdEE *&__cuda_2,int &__cuda_3){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462620cunn_SoftMaxBackwardILi2EddNS1_26LogSoftMaxBackwardEpilogueEEEvPT0_S5_S5_i( (_ZN2at4cuda4typeIdEE &)__cuda_0,(_ZN2at4cuda4typeIdEE &)__cuda_1,(_ZN2at4cuda4typeIdEE &)__cuda_2,(int &)__cuda_3);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:71:17: error: template-id ‘__wrapper__device_stub_cunn_SoftMaxBackward<2, at::cuda::type, at::acc_type<double, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SoftMaxBackward(_ZN2at4cuda4typeIdEE&, _ZN2at4cuda4typeIdEE&, _ZN2at4cuda4typeIdEE&, int&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:76:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SoftMaxBackward<2, ::at::cuda::type , ::at::acc_type<float, (bool)1> , ::at::native::operator ::LogSoftMaxBackwardEpilogue>( _ZN2at4cuda4typeIfEE *&__cuda_0,_ZN2at4cuda4typeIfEE *&__cuda_1,_ZN2at4cuda4typeIfEE *&__cuda_2,int &__cuda_3){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462620cunn_SoftMaxBackwardILi2EffNS1_26LogSoftMaxBackwardEpilogueEEEvPT0_S5_S5_i( (_ZN2at4cuda4typeIfEE &)__cuda_0,(_ZN2at4cuda4typeIfEE &)__cuda_1,(_ZN2at4cuda4typeIfEE &)__cuda_2,(int &)__cuda_3);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:76:17: error: template-id ‘__wrapper__device_stub_cunn_SoftMaxBackward<2, at::cuda::type, at::acc_type<float, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SoftMaxBackward(_ZN2at4cuda4typeIfEE&, _ZN2at4cuda4typeIfEE&, _ZN2at4cuda4typeIfEE&, int&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:81:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SoftMaxBackward<2, ::at::cuda::type< ::at::Half> , ::at::acc_type< ::__half, (bool)1> , ::at::native::operator ::LogSoftMaxBackwardEpilogue>( _ZN2at4cuda4typeINS_4HalfEEE *&__cuda_0,_ZN2at4cuda4typeINS_4HalfEEE *&__cuda_1,_ZN2at4cuda4typeINS_4HalfEEE *&__cuda_2,int &__cuda_3){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462620cunn_SoftMaxBackwardILi2E6__halffNS1_26LogSoftMaxBackwardEpilogueEEEvPT0_S6_S6_i( (_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_0,(_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_1,(_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_2,(int &)__cuda_3);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:81:17: error: template-id ‘__wrapper__device_stub_cunn_SoftMaxBackward<2, at::cuda::typeat::Half, at::acc_type<__half, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SoftMaxBackward(_ZN2at4cuda4typeINS_4HalfEEE&, _ZN2at4cuda4typeINS_4HalfEEE&, _ZN2at4cuda4typeINS_4HalfEEE&, int&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:86:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SpatialSoftMaxBackward< ::at::cuda::type , ::at::acc_type<double, (bool)1> , ::at::native::operator ::LogSoftMaxBackwardEpilogue>( _ZN2at4cuda4typeIdEE *&__cuda_0,_ZN2at4cuda4typeIdEE *&__cuda_1,_ZN2at4cuda4typeIdEE *&__cuda_2,::uint32_t &__cuda_3,::uint32_t &__cuda_4,::uint32_t &__cuda_5){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462627cunn_SpatialSoftMaxBackwardIddNS1_26LogSoftMaxBackwardEpilogueEEEvPT_S5_S5_jjj( (_ZN2at4cuda4typeIdEE &)__cuda_0,(_ZN2at4cuda4typeIdEE &)__cuda_1,(_ZN2at4cuda4typeIdEE &)__cuda_2,(::uint32_t &)__cuda_3,(::uint32_t &)__cuda_4,(::uint32_t &)__cuda_5);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:86:17: error: template-id ‘__wrapper__device_stub_cunn_SpatialSoftMaxBackward<at::cuda::type, at::acc_type<double, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SpatialSoftMaxBackward(_ZN2at4cuda4typeIdEE&, _ZN2at4cuda4typeIdEE&, _ZN2at4cuda4typeIdEE&, uint32_t&, uint32_t&, uint32_t&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:91:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SpatialSoftMaxBackward< ::at::cuda::type , ::at::acc_type<float, (bool)1> , ::at::native::operator ::LogSoftMaxBackwardEpilogue>( _ZN2at4cuda4typeIfEE *&__cuda_0,_ZN2at4cuda4typeIfEE *&__cuda_1,_ZN2at4cuda4typeIfEE *&__cuda_2,::uint32_t &__cuda_3,::uint32_t &__cuda_4,::uint32_t &__cuda_5){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462627cunn_SpatialSoftMaxBackwardIffNS1_26LogSoftMaxBackwardEpilogueEEEvPT_S5_S5_jjj( (_ZN2at4cuda4typeIfEE &)__cuda_0,(_ZN2at4cuda4typeIfEE &)__cuda_1,(_ZN2at4cuda4typeIfEE &)__cuda_2,(::uint32_t &)__cuda_3,(::uint32_t &)__cuda_4,(::uint32_t &)__cuda_5);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:91:17: error: template-id ‘__wrapper__device_stub_cunn_SpatialSoftMaxBackward<at::cuda::type, at::acc_type<float, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SpatialSoftMaxBackward(_ZN2at4cuda4typeIfEE&, _ZN2at4cuda4typeIfEE&, _ZN2at4cuda4typeIfEE&, uint32_t&, uint32_t&, uint32_t&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:96:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SpatialSoftMaxBackward< ::at::cuda::type< ::at::Half> , ::at::acc_type< ::__half, (bool)1> , ::at::native::operator ::LogSoftMaxBackwardEpilogue>( _ZN2at4cuda4typeINS_4HalfEEE *&__cuda_0,_ZN2at4cuda4typeINS_4HalfEEE *&__cuda_1,_ZN2at4cuda4typeINS_4HalfEEE *&__cuda_2,::uint32_t &__cuda_3,::uint32_t &__cuda_4,::uint32_t &__cuda_5){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462627cunn_SpatialSoftMaxBackwardI6__halffNS1_26LogSoftMaxBackwardEpilogueEEEvPT_S6_S6_jjj( (_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_0,(_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_1,(_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_2,(::uint32_t &)__cuda_3,(::uint32_t &)__cuda_4,(::uint32_t &)__cuda_5);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:96:17: error: template-id ‘__wrapper__device_stub_cunn_SpatialSoftMaxBackward<at::cuda::typeat::Half, at::acc_type<__half, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SpatialSoftMaxBackward(_ZN2at4cuda4typeINS_4HalfEEE&, _ZN2at4cuda4typeINS_4HalfEEE&, _ZN2at4cuda4typeINS_4HalfEEE&, uint32_t&, uint32_t&, uint32_t&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:101:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SoftMaxForward<2, ::at::cuda::type , ::at::acc_type<double, (bool)1> , ::at::native::operator ::SoftMaxForwardEpilogue>( _ZN2at4cuda4typeIdEE *&__cuda_0,_ZN2at4cuda4typeIdEE *&__cuda_1,int &__cuda_2){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462619cunn_SoftMaxForwardILi2EddNS1_22SoftMaxForwardEpilogueEEEvPT0_S5_i( (_ZN2at4cuda4typeIdEE &)__cuda_0,(_ZN2at4cuda4typeIdEE &)__cuda_1,(int &)__cuda_2);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:101:17: error: template-id ‘__wrapper__device_stub_cunn_SoftMaxForward<2, at::cuda::type, at::acc_type<double, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SoftMaxForward(_ZN2at4cuda4typeIdEE&, _ZN2at4cuda4typeIdEE&, int&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:106:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SoftMaxForward<2, ::at::cuda::type , ::at::acc_type<float, (bool)1> , ::at::native::operator ::SoftMaxForwardEpilogue>( _ZN2at4cuda4typeIfEE *&__cuda_0,_ZN2at4cuda4typeIfEE *&__cuda_1,int &__cuda_2){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462619cunn_SoftMaxForwardILi2EffNS1_22SoftMaxForwardEpilogueEEEvPT0_S5_i( (_ZN2at4cuda4typeIfEE &)__cuda_0,(_ZN2at4cuda4typeIfEE &)__cuda_1,(int &)__cuda_2);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:106:17: error: template-id ‘__wrapper__device_stub_cunn_SoftMaxForward<2, at::cuda::type, at::acc_type<float, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SoftMaxForward(_ZN2at4cuda4typeIfEE&, _ZN2at4cuda4typeIfEE&, int&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:111:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SoftMaxForward<2, ::at::cuda::type< ::at::Half> , ::at::acc_type< ::__half, (bool)1> , ::at::native::operator ::SoftMaxForwardEpilogue>( _ZN2at4cuda4typeINS_4HalfEEE *&__cuda_0,_ZN2at4cuda4typeINS_4HalfEEE *&__cuda_1,int &__cuda_2){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462619cunn_SoftMaxForwardILi2E6__halffNS1_22SoftMaxForwardEpilogueEEEvPT0_S6_i( (_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_0,(_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_1,(int &)__cuda_2);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:111:17: error: template-id ‘__wrapper__device_stub_cunn_SoftMaxForward<2, at::cuda::typeat::Half, at::acc_type<__half, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SoftMaxForward(_ZN2at4cuda4typeINS_4HalfEEE&, _ZN2at4cuda4typeINS_4HalfEEE&, int&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:116:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SpatialSoftMaxForward< ::at::cuda::type , ::at::acc_type<double, (bool)1> , ::at::native::operator ::SoftMaxForwardEpilogue>( _ZN2at4cuda4typeIdEE *&__cuda_0,_ZN2at4cuda4typeIdEE *&__cuda_1,::uint32_t &__cuda_2,::uint32_t &__cuda_3,::uint32_t &__cuda_4){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462626cunn_SpatialSoftMaxForwardIddNS1_22SoftMaxForwardEpilogueEEEvPT_S5_jjj( (_ZN2at4cuda4typeIdEE &)__cuda_0,(_ZN2at4cuda4typeIdEE &)__cuda_1,(::uint32_t &)__cuda_2,(::uint32_t &)__cuda_3,(::uint32_t &)__cuda_4);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:116:17: error: template-id ‘__wrapper__device_stub_cunn_SpatialSoftMaxForward<at::cuda::type, at::acc_type<double, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SpatialSoftMaxForward(_ZN2at4cuda4typeIdEE&, _ZN2at4cuda4typeIdEE&, uint32_t&, uint32_t&, uint32_t&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:121:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SpatialSoftMaxForward< ::at::cuda::type , ::at::acc_type<float, (bool)1> , ::at::native::operator ::SoftMaxForwardEpilogue>( _ZN2at4cuda4typeIfEE *&__cuda_0,_ZN2at4cuda4typeIfEE *&__cuda_1,::uint32_t &__cuda_2,::uint32_t &__cuda_3,::uint32_t &__cuda_4){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462626cunn_SpatialSoftMaxForwardIffNS1_22SoftMaxForwardEpilogueEEEvPT_S5_jjj( (_ZN2at4cuda4typeIfEE &)__cuda_0,(_ZN2at4cuda4typeIfEE &)__cuda_1,(::uint32_t &)__cuda_2,(::uint32_t &)__cuda_3,(::uint32_t &)__cuda_4);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:121:17: error: template-id ‘__wrapper__device_stub_cunn_SpatialSoftMaxForward<at::cuda::type, at::acc_type<float, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SpatialSoftMaxForward(_ZN2at4cuda4typeIfEE&, _ZN2at4cuda4typeIfEE&, uint32_t&, uint32_t&, uint32_t&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:126:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SpatialSoftMaxForward< ::at::cuda::type< ::at::Half> , ::at::acc_type< ::__half, (bool)1> , ::at::native::operator ::SoftMaxForwardEpilogue>( _ZN2at4cuda4typeINS_4HalfEEE *&__cuda_0,_ZN2at4cuda4typeINS_4HalfEEE *&__cuda_1,::uint32_t &__cuda_2,::uint32_t &__cuda_3,::uint32_t &__cuda_4){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462626cunn_SpatialSoftMaxForwardI6__halffNS1_22SoftMaxForwardEpilogueEEEvPT_S6_jjj( (_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_0,(_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_1,(::uint32_t &)__cuda_2,(::uint32_t &)__cuda_3,(::uint32_t &)__cuda_4);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:126:17: error: template-id ‘__wrapper__device_stub_cunn_SpatialSoftMaxForward<at::cuda::typeat::Half, at::acc_type<__half, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SpatialSoftMaxForward(_ZN2at4cuda4typeINS_4HalfEEE&, _ZN2at4cuda4typeINS_4HalfEEE&, uint32_t&, uint32_t&, uint32_t&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:131:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SoftMaxBackward<2, ::at::cuda::type , ::at::acc_type<double, (bool)1> , ::at::native::operator ::SoftMaxBackwardEpilogue>( _ZN2at4cuda4typeIdEE *&__cuda_0,_ZN2at4cuda4typeIdEE *&__cuda_1,_ZN2at4cuda4typeIdEE *&__cuda_2,int &__cuda_3){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462620cunn_SoftMaxBackwardILi2EddNS1_23SoftMaxBackwardEpilogueEEEvPT0_S5_S5_i( (_ZN2at4cuda4typeIdEE &)__cuda_0,(_ZN2at4cuda4typeIdEE &)__cuda_1,(_ZN2at4cuda4typeIdEE &)__cuda_2,(int &)__cuda_3);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:131:17: error: template-id ‘__wrapper__device_stub_cunn_SoftMaxBackward<2, at::cuda::type, at::acc_type<double, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SoftMaxBackward(_ZN2at4cuda4typeIdEE&, _ZN2at4cuda4typeIdEE&, _ZN2at4cuda4typeIdEE&, int&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:136:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SoftMaxBackward<2, ::at::cuda::type , ::at::acc_type<float, (bool)1> , ::at::native::operator ::SoftMaxBackwardEpilogue>( _ZN2at4cuda4typeIfEE *&__cuda_0,_ZN2at4cuda4typeIfEE *&__cuda_1,_ZN2at4cuda4typeIfEE *&__cuda_2,int &__cuda_3){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462620cunn_SoftMaxBackwardILi2EffNS1_23SoftMaxBackwardEpilogueEEEvPT0_S5_S5_i( (_ZN2at4cuda4typeIfEE &)__cuda_0,(_ZN2at4cuda4typeIfEE &)__cuda_1,(_ZN2at4cuda4typeIfEE &)__cuda_2,(int &)__cuda_3);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:136:17: error: template-id ‘__wrapper__device_stub_cunn_SoftMaxBackward<2, at::cuda::type, at::acc_type<float, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SoftMaxBackward(_ZN2at4cuda4typeIfEE&, _ZN2at4cuda4typeIfEE&, _ZN2at4cuda4typeIfEE&, int&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:141:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SoftMaxBackward<2, ::at::cuda::type< ::at::Half> , ::at::acc_type< ::__half, (bool)1> , ::at::native::operator ::SoftMaxBackwardEpilogue>( _ZN2at4cuda4typeINS_4HalfEEE *&__cuda_0,_ZN2at4cuda4typeINS_4HalfEEE *&__cuda_1,_ZN2at4cuda4typeINS_4HalfEEE *&__cuda_2,int &__cuda_3){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462620cunn_SoftMaxBackwardILi2E6__halffNS1_23SoftMaxBackwardEpilogueEEEvPT0_S6_S6_i( (_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_0,(_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_1,(_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_2,(int &)__cuda_3);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:141:17: error: template-id ‘__wrapper__device_stub_cunn_SoftMaxBackward<2, at::cuda::typeat::Half, at::acc_type<__half, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SoftMaxBackward(_ZN2at4cuda4typeINS_4HalfEEE&, _ZN2at4cuda4typeINS_4HalfEEE&, _ZN2at4cuda4typeINS_4HalfEEE&, int&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:146:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SpatialSoftMaxBackward< ::at::cuda::type , ::at::acc_type<double, (bool)1> , ::at::native::operator ::SoftMaxBackwardEpilogue>( _ZN2at4cuda4typeIdEE *&__cuda_0,_ZN2at4cuda4typeIdEE *&__cuda_1,_ZN2at4cuda4typeIdEE *&__cuda_2,::uint32_t &__cuda_3,::uint32_t &__cuda_4,::uint32_t &__cuda_5){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462627cunn_SpatialSoftMaxBackwardIddNS1_23SoftMaxBackwardEpilogueEEEvPT_S5_S5_jjj( (_ZN2at4cuda4typeIdEE &)__cuda_0,(_ZN2at4cuda4typeIdEE &)__cuda_1,(_ZN2at4cuda4typeIdEE &)__cuda_2,(::uint32_t &)__cuda_3,(::uint32_t &)__cuda_4,(::uint32_t &)__cuda_5);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:146:17: error: template-id ‘__wrapper__device_stub_cunn_SpatialSoftMaxBackward<at::cuda::type, at::acc_type<double, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SpatialSoftMaxBackward(_ZN2at4cuda4typeIdEE&, _ZN2at4cuda4typeIdEE&, _ZN2at4cuda4typeIdEE&, uint32_t&, uint32_t&, uint32_t&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:151:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SpatialSoftMaxBackward< ::at::cuda::type , ::at::acc_type<float, (bool)1> , ::at::native::operator ::SoftMaxBackwardEpilogue>( _ZN2at4cuda4typeIfEE *&__cuda_0,_ZN2at4cuda4typeIfEE *&__cuda_1,_ZN2at4cuda4typeIfEE *&__cuda_2,::uint32_t &__cuda_3,::uint32_t &__cuda_4,::uint32_t &__cuda_5){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462627cunn_SpatialSoftMaxBackwardIffNS1_23SoftMaxBackwardEpilogueEEEvPT_S5_S5_jjj( (_ZN2at4cuda4typeIfEE &)__cuda_0,(_ZN2at4cuda4typeIfEE &)__cuda_1,(_ZN2at4cuda4typeIfEE &)__cuda_2,(::uint32_t &)__cuda_3,(::uint32_t &)__cuda_4,(::uint32_t &)__cuda_5);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:151:17: error: template-id ‘__wrapper__device_stub_cunn_SpatialSoftMaxBackward<at::cuda::type, at::acc_type<float, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SpatialSoftMaxBackward(_ZN2at4cuda4typeIfEE&, _ZN2at4cuda4typeIfEE&, _ZN2at4cuda4typeIfEE&, uint32_t&, uint32_t&, uint32_t&)’ does not match any template declaration
In file included from tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:1:0:
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:156:17: error: parse error in template argument list
template<> __specialization_static void __wrapper__device_stub_cunn_SpatialSoftMaxBackward< ::at::cuda::type< ::at::Half> , ::at::acc_type< ::__half, (bool)1> , ::at::native::operator ::SoftMaxBackwardEpilogue>( _ZN2at4cuda4typeINS_4HalfEEE *&__cuda_0,_ZN2at4cuda4typeINS_4HalfEEE *&__cuda_1,_ZN2at4cuda4typeINS_4HalfEEE *&__cuda_2,::uint32_t &__cuda_3,::uint32_t &__cuda_4,::uint32_t &__cuda_5){__device_stub__ZN2at6native66_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a462627cunn_SpatialSoftMaxBackwardI6__halffNS1_23SoftMaxBackwardEpilogueEEEvPT_S6_S6_jjj( (_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_0,(_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_1,(_ZN2at4cuda4typeINS_4HalfEEE &)__cuda_2,(::uint32_t &)__cuda_3,(::uint32_t &)__cuda_4,(::uint32_t &)__cuda_5);}}}}
^
/tmp/tmpxft_00007024_00000000-4_SoftMax.cudafe1.stub.c:156:17: error: template-id ‘__wrapper__device_stub_cunn_SpatialSoftMaxBackward<at::cuda::typeat::Half, at::acc_type<__half, true>, >’ for ‘void at::native::_GLOBAL__N__42_tmpxft_00007024_00000000_7_SoftMax_cpp1_ii_826a4626::__wrapper__device_stub_cunn_SpatialSoftMaxBackward(_ZN2at4cuda4typeINS_4HalfEEE&, _ZN2at4cuda4typeINS_4HalfEEE&, _ZN2at4cuda4typeINS_4HalfEEE&, uint32_t&, uint32_t&, uint32_t&)’ does not match any template declaration
cc1plus: warning: unrecognized command line option ‘-Wno-absolute-value’
CMake Error at ATen_cuda_generated_SoftMax.cu.o.Release.cmake:279 (message):
Error generating file
/home/ubuntu/pytorch0.3/aten/build/src/ATen/CMakeFiles/ATen_cuda.dir/native/cuda/./ATen_cuda_generated_SoftMax.cu.o

src/ATen/CMakeFiles/ATen_cuda.dir/build.make:1169: recipe for target 'src/ATen/CMakeFiles/ATen_cuda.dir/native/cuda/ATen_cuda_generated_SoftMax.cu.o' failed
make[2]: *** [src/ATen/CMakeFiles/ATen_cuda.dir/native/cuda/ATen_cuda_generated_SoftMax.cu.o] Error 1
make[2]: *** 正在等待未完成的任务....
CMakeFiles/Makefile2:209: recipe for target 'src/ATen/CMakeFiles/ATen_cuda.dir/all' failed
make[1]: *** [src/ATen/CMakeFiles/ATen_cuda.dir/all] Error 2
Makefile:129: recipe for target 'all' failed

How to fix it?

Recommend Projects

React

A declarative, efficient, and flexible JavaScript library for building user interfaces.
Vue.js

🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
Typescript

TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
TensorFlow

An Open Source Machine Learning Framework for Everyone
Django

The Web framework for perfectionists with deadlines.
Laravel

A PHP framework for web artisans
D3

Bring data to life with SVG, Canvas and HTML. 📊📈🎉

Recommend Topics

javascript

JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
web

Some thing interesting about web. New door for the world.
server

A server is a program made to process requests and deliver data to clients.
Machine learning

Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
Visualization

Some thing interesting about visualization, use data art
Game

Some thing interesting about game, make everyone happy.

Recommend Org

Facebook

We are working to build community through open source technology. NB: members must have two-factor auth.
Microsoft

Open source projects and samples from Microsoft.
Google

Google ❤️ Open Source for everyone.
Alibaba

Alibaba Open Source for everyone
D3

Data-Driven Documents codes.
Tencent

China tencent open source team.