TorchCraft
build cherrypi on windows : python setup.py build command fail
-- Caffe2: Header version is: 9.0 -- Found CUDNN: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v9.0/include
-- Found cuDNN: v7.4.2 (include: C:/Program Files/NVIDIA GPU Computing Toolkit/ CUDA/v9.0/include, library: C:/Program Files/NVIDIA GPU Computing Toolkit/CUDA/v 9.0/lib/x64/cudnn.lib) CMake Error at cmake/public/cuda.cmake:278 (message): CUDA support not available with 32-bit windows. Did you forget to set Win64 in the generator target? Call Stack (most recent call first): cmake/Dependencies.cmake:626 (include) CMakeLists.txt:200 (include)
-- Configuring incomplete, errors occurred! See also "E:/Game/BWAPI/TorchCraftAI/3rdparty/pytorch/build/CMakeFiles/CMakeOutp ut.log". See also "E:/Game/BWAPI/TorchCraftAI/3rdparty/pytorch/build/CMakeFiles/CMakeErro r.log".
E:\Game\BWAPI\TorchCraftAI\3rdparty\pytorch\build>IF ERRORLEVEL 1 exit 1 Failed to run 'tools\build_pytorch_libs.bat --use-cuda --use-fbgemm --use-nnpack --use-mkldnn --use-qnnpack caffe2'
It sounds like you forgot to run
set CMAKE_GENERATOR=Visual Studio 15 2017 Win64
before you started building, as mentioned in Initialize Environment
My compilation environment is windows7,CUDA Toolkit 9.0, python 37-32
5>Building NVCC (Device) object caffe2/CMakeFiles/caffe2_gpu.dir//aten/src/THC/Release/caffe2_gpu_generated_THCReduceApplyUtils.cu.obj 5>nvcc fatal : Unsupported gpu architecture 'compute_20' 5>CMake Error at caffe2_gpu_generated_THCReduceApplyUtils.cu.obj.Release.cmake:219 (message): 5> Error generating 5> E:/Game/BWAPI/TorchCraftAI/3rdparty/pytorch/build/caffe2/CMakeFiles/caffe2_gpu.dir//aten/src/THC/Release/caffe2_gpu_generated_THCReduceApplyUtils.cu.obj 5> 5> 5>D:\Program Files\Microsoft Visual Studio\2017\Community\Common7\IDE\VC\VCTargets\Microsoft.CppCommon.targets(209,5): error MSB6006: “cmd.exe”已退出，代码为 1。
Which GPU are you using? From a Google search, it's seems that compute capability 2.0 was removed in cuda 9, and if you're using a sufficiently old GPU you might need CUDA 8: https://github.com/kaldi-asr/kaldi/issues/1918
I see how this works
Delete "- gencode arch = compute_20, code = sm_20"
But I don't know in what file to delete this statement. Is this statement configured in this project or where?
I'm not sure, we rely on PyTorch to build their system correctly, you might find more information asking in their github repository. https://github.com/pytorch/pytorch/