python3 setup.py bdist_wheel error

[Nirvana93]

Hello,thanks for your code,I have following problem when running python3 setup.py bdist_wheel error:

[hpc-7-005 pytorch binding]$ python3 setup.py bdist_wheel running bdist_wheel runnina build running build deps The C compiler identification is GNU 5.4.0 The CXX compiler identification is GNU 5.4.0 Check for working c compiler: /opt/compiler/acc-5.4-os7/bin/gcc Check for working c compiler: /opt/compiler/gcc-5.4-os7/bin/gcc --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: /opt/compiler/qcc-5.4-os7/bin/g++ Check for working CXX compiler: /opt/compiler/acc-5.4-0s7/bin/a++ -- works Detecting CXX compiler ABI into Detecting CXX compiler ABI info - done Detectina CXX compile features Detecting CXX compile features -done Found CUDA: /opt/lib/cuda-9.1.1 (found suitable version "9.1", minimum required is "6.5"1- cuda found TRUE CMake Warning at CMakeLists.txt:52 (FIND PACKAGE): By not providing "FindTorch.cmake" in CMAKE MODULE PATH this project hasasked CMake to find a package configuration file provided by "Torch", butCMake did not find one.

Could not find a package configuration file provided by "Torch" with any ofthe following names:

TorchConfig.cmake torch-config.cmake

Add the installation prefix of "Torch" to CMAKE PREFIX PATH or set"Torch DIR" to a directory containing one of the above files. If "Torch"provides a separate development package or SDK, be sure it has beeninstalled.

Torch found Torch DIR-NOTFOUND Building shared Library with GPU support Configuring done Generatina done Build files have been written to: /ps2/cv6/shuaishao3/0CR/warp-ctc-travis-ci/pytorch binding/build/lib

[16%] Building NVCC (Device) obiect CMakeFiles/warpctc.dir/src/warpctc generated_reduce.cu.o [33%] Building NVCC (Device) obiect CMakeFiles/warpctc.dir/src/warpctc generated_ctc_entrypoint.cu. Scanning dependencies of target warpctc Linking CXX shared Library Libwarpctc.so [33%] Built target warpctc Scanning dependencies of target test cpu [50%] Buildina CXX object CMakeFiles/test cpu.dir/tests/test cpu.cpp.o [66%] Building CXX object CMakeFiles/test cpu.dir/tests/random.cpp.o Linking cxx executabLe test cpu [66%] Built target test cpu [83%] Building NVCC (Device) obiect CMakeFiles/test gpu.dir/tests/test gpu generated test gpu.cu.ocanning dependencies of target test gpu [100%] Building Cxx obiect CMakeFiles/test qpu.dir/tests/random.cpp.o Linking CxX executabLe test gpu [100%] Built target test qpu running build py reatina build/lib.linux-x86 64-3.6 reating build/lib.linux-x86 64-3.6/warpcto copying warpctc/ init .py -> build/lib. Linux-x86 64-3.6/warpctc reating build/lib.Linux-x86 64-3.6/warpctc/lib copying warpctc/lib/libwarpctc.so -> build/lib.Linux-x86 64-3.6/warpctc/lib runnina build ext building 'warpctc. warpctc' extension reating build/temp.Linux-x86 64-3.6 creating build/temp.linux-x86 64-3.6/sro gcc -Wsign-compare-DNDEBUG -g-fwrapv -03 -WalL-Wstrict-prototypes -fPIC-I../include -I/home/cv6/huaishao3/anaconda3/lib/python3.6/site-packages/torch/lib/include -I/home/cv6/shuaishao3/anaconda3/lib/python3.6/site-packages/torch/lib/include/TH -I/home/cv6/shuaishao3/anaconda3/lib/python3.6/site-packages/torch/lib/incLude/THC -I/home/cv6/shuaishao3/anaconda3/include/python3.6m -c src/ warpctcpp -o build/temp.linux-x86 64-3.6/src/ warpctc.o-DTORCH EXTENSION NAME- warpctc-D GLIBCXX USE CXX11 ABI-O-std=C++11 cclplus: warning: command line option '-Wstrict-prototypes' is valid for C/ObiC but not for C++ src/ _warpctc.cpp:33:79: error: macro "AT_ASSERT" passed 2 arguments, but takes just 1 AT ASSERT(! input_Lengths.type().is_cuda(), "input_Lengths must be on CPU"); src/ _warpctc.cpp:39:65: error: macro "AT_ASSERT" passed 2 arguments, but takes just 1 AT ASSERT(! input_Lengths.type().is_cuda(), "input_Lengths must be on CPU"); src/ _warpctc.cpp:45:79: error: macro "AT_ASSERT" passed 2 arguments, but takes just 1 ASSERT(! Label_Lengths .type().is_cuda(), "label_Lengths must be on CPU"); src/ warpctc.cpp: In function 'std::tuple<at::Tensor, at::Tensor> ctc(at::Tensor, at::Tensor, at::Tensor, at::Tensor, int, bool)': src/ warpctc.cpp:33:5: error: 'AT ASSERT' was not declared in this scope AT ASSERT(! input Lengths .type().is cuda(), "input Lengths must be on CPU"); error: command 'gcc' failed with exit status 1