torchvision c++ api build from source, Build error about C++14 and C++11

[weijian-sun]

torchvision c++ api build from source, in master branch, Build error about C++14 and C++11

Environment

PyTorch Version: 1.5.1 build from source
OS (e.g., Linux): Ubuntu 16
How you installed PyTorch (conda, pip, source): source
Build command you used (if compiling from source):
#!/bin/bash
WORKSPACE=$(cd $(dirname $0);pwd)
export PATH=/opt/protobuf/bin:$PATH
mkdir build
cd build
/opt/cmake/bin/cmake -DCMAKE_INSTALL_PREFIX=${WORKSPACE}/output
-DCMAKE_PREFIX_PATH=/opt/had_deps/libtorch1.5 -DWITH_CUDA=ON ..
make -jgetconf _NPROCESSORS_CONF install
Python version: 3.5
CUDA/cuDNN version: 10.0

error message:

• 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
  -- The CUDA compiler identification is NVIDIA 10.0.130
  -- Check for working CUDA compiler: /usr/local/cuda/bin/nvcc
  -- Check for working CUDA compiler: /usr/local/cuda/bin/nvcc -- works
  -- Detecting CUDA compiler ABI info
  -- Detecting CUDA compiler ABI info - done
  -- Found Python3: /usr/lib/python3.5/config-3.5m-x86_64-linux-gnu/libpython3.5m.so (found version "3.5.2") found components: Development
  -- Looking for pthread.h
  -- Looking for pthread.h - found
  -- Looking for pthread_create
  -- Looking for pthread_create - not found
  -- Looking for pthread_create in pthreads
  -- Looking for pthread_create in pthreads - not found
  -- Looking for pthread_create in pthread
  -- Looking for pthread_create in pthread - found
  -- Found Threads: TRUE
  CMake Warning (dev) at /opt/had_deps/lib/cmake/protobuf/protobuf-config-version.cmake:50 (if):
  Policy CMP0054 is not set: Only interpret if() arguments as variables or
  keywords when unquoted. Run "cmake --help-policy CMP0054" for policy
  details. Use the cmake_policy command to set the policy and suppress this
  warning.

  Quoted variables like "" will no longer be dereferenced when the policy is
  set to NEW. Since the policy is not set the OLD behavior will be used.
  Call Stack (most recent call first):
  /opt/had_deps/libtorch1.5/share/cmake/Caffe2/public/protobuf.cmake:4 (find_package)
  /opt/had_deps/libtorch1.5/share/cmake/Caffe2/Caffe2Config.cmake:56 (include)
  /opt/had_deps/libtorch1.5/share/cmake/Torch/TorchConfig.cmake:40 (find_package)
  CMakeLists.txt:16 (find_package)
  This warning is for project developers. Use -Wno-dev to suppress it.

-- Found ZLIB: /usr/lib/x86_64-linux-gnu/libz.so (found version "1.2.8")
-- Caffe2: Found protobuf with new-style protobuf targets.
-- Caffe2: Protobuf version 3.3.0
-- Found CUDA: /usr/local/cuda (found version "10.0")
-- Caffe2: CUDA detected: 10.0
-- Caffe2: CUDA nvcc is: /usr/local/cuda/bin/nvcc
-- Caffe2: CUDA toolkit directory: /usr/local/cuda
-- Caffe2: Header version is: 10.0
-- Found CUDNN: /usr/lib/x86_64-linux-gnu/libcudnn.so
-- Found cuDNN: v7.5.0 (include: /usr/include, library: /usr/lib/x86_64-linux-gnu/libcudnn.so)
-- Autodetected CUDA architecture(s): 7.5
-- Added CUDA NVCC flags for: -gencode;arch=compute_75,code=sm_75
-- Found torch: /opt/had_deps/libtorch1.5/lib/libtorch.so
-- Found PNG: /usr/lib/x86_64-linux-gnu/libpng.so (found version "1.2.54")
-- Found JPEG: /usr/lib/x86_64-linux-gnu/libjpeg.so (found version "80")
-- Configuring done
-- Generating done
-- Build files have been written to: /home/administrator/workspace/vision/build
Scanning dependencies of target torchvision
[ 3%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/models/alexnet.cpp.o
[ 7%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/models/densenet.cpp.o
[ 10%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/models/googlenet.cpp.o
[ 14%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/models/inception.cpp.o
[ 17%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/models/mnasnet.cpp.o
[ 21%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/models/mobilenet.cpp.o
[ 25%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/models/resnet.cpp.o
[ 28%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/models/shufflenetv2.cpp.o
[ 32%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/models/squeezenet.cpp.o
[ 35%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/models/vgg.cpp.o
[ 39%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/cpu/PSROIAlign_cpu.cpp.o
[ 42%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/cpu/DeformConv_cpu.cpp.o
[ 46%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/cpu/PSROIPool_cpu.cpp.o
[ 50%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/cpu/ROIAlign_cpu.cpp.o
[ 53%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/cpu/ROIPool_cpu.cpp.o
[ 57%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/cpu/image/image.cpp.o
[ 60%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/cpu/image/readjpeg_cpu.cpp.o
[ 64%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/cpu/image/readpng_cpu.cpp.o
[ 67%] Building CXX object CMakeFiles/torchvision.dir/torchvision/csrc/cpu/nms_cpu.cpp.o
[ 71%] Building CUDA object CMakeFiles/torchvision.dir/torchvision/csrc/cuda/DeformConv_cuda.cu.o
In file included from /usr/include/c++/5/initializer_list:36:0,
from /opt/had_deps/libtorch1.5/include/c10/util/SmallVector.h:31,
from /opt/had_deps/libtorch1.5/include/c10/util/ArrayRef.h:18,
from /opt/had_deps/libtorch1.5/include/c10/core/MemoryFormat.h:5,
from /opt/had_deps/libtorch1.5/include/ATen/core/TensorBody.h:5,
from /opt/had_deps/libtorch1.5/include/ATen/Tensor.h:11,
from /opt/had_deps/libtorch1.5/include/ATen/Context.h:4,
from /opt/had_deps/libtorch1.5/include/ATen/ATen.h:5,
from /home/administrator/workspace/vision/torchvision/csrc/cuda/DeformConv_cuda.cu:69:
/usr/include/c++/5/bits/c++0x_warning.h:32:2: error: #error This file requires compiler and library support for the ISO C++ 2011 standard. This support must be enabled with the -std=c++11 or -std=gnu++11 compiler options.
#error This file requires compiler and library support
^
In file included from /opt/had_deps/libtorch1.5/include/c10/util/ArrayRef.h:19:0,
from /opt/had_deps/libtorch1.5/include/c10/core/MemoryFormat.h:5,
from /opt/had_deps/libtorch1.5/include/ATen/core/TensorBody.h:5,
from /opt/had_deps/libtorch1.5/include/ATen/Tensor.h:11,
from /opt/had_deps/libtorch1.5/include/ATen/Context.h:4,
from /opt/had_deps/libtorch1.5/include/ATen/ATen.h:5,
from /home/administrator/workspace/vision/torchvision/csrc/cuda/DeformConv_cuda.cu:69:
/opt/had_deps/libtorch1.5/include/c10/util/C++17.h:24:2: error: #error You need C++14 to compile PyTorch
#error You need C++14 to compile PyTorch

[pmeier]

Hi @weijian-sun,

I'm out of my depth for the build process, we'll have to wait for someone else. Still, I noticed one thing while skimming the output: In Environment you stated you have CUDA 10.2, but the build -- Found CUDA: /usr/local/cuda (found version "10.0").

[weijian-sun]

Hi @weijian-sun,

I'm out of my depth for the build process, we'll have to wait for someone else. Still, I noticed one thing while skimming the output: In Environment you stated you have CUDA 10.2, but the build -- Found CUDA: /usr/local/cuda (found version "10.0").

oh, sorry for the typo, it's cuda 10.0, and the driver version is 440.33.01

[deweyamer]

did u solved this problem？

[weijian-sun]

switch to a clean docker image, proglem solved, maybe some environment related issue

[deweyamer]

sorry, do u means use docker to download the vision?

[fmassa]

@Dewey1994 if you are compiling torchvision from source, you need a compiler that supports C++14