系统：win10
vs版本：vs2019
libtorch：1.10.0
cuda：10.2.
在win10和vs2019配置好libtorch后，发现只能调用CPU，一旦指定GPU，就会报以下错误：
PyTorch is not linked with support for cuda devices
Exception raised from getDeviceGuardImpl at C:\w\b\windows\pytorch\c10/core/impl/DeviceGuardImplInterface.h:318 (most recent call first):
00007FF82BC09A1A00007FF82BC08AA0 c10.dll!c10::detail::LogAPIUsageFakeReturn [<unknown file> @ <unknown line number>]
00007FF82BC0957A00007FF82BC08AA0 c10.dll!c10::detail::LogAPIUsageFakeReturn [<unknown file> @ <unknown line number>]
00007FF82BC0A78100007FF82BC08AA0 c10.dll!c10::detail::LogAPIUsageFakeReturn [<unknown file> @ <unknown line number>]
00007FF82BC0A3A500007FF82BC08AA0 c10.dll!c10::detail::LogAPIUsageFakeReturn [<unknown file> @ <unknown line number>]
00007FF82BC07FAF00007FF82BC07F40 c10.dll!c10::Error::Error [<unknown file> @ <unknown line number>]
00007FF82BC067D600007FF82BC06710 c10.dll!c10::detail::torchCheckFail [<unknown file> @ <unknown line number>]
00007FFF88D8E90B00007FFF88D8E650 torch_cpu.dll!at::Context::getDeviceFromPtr [<unknown file> @ <unknown line number>]
00007FFF8967A2BE00007FFF89679AA0 torch_cpu.dll!at::native::view_dtype [<unknown file> @ <unknown line number>]
00007FFF8967965900007FFF896795C0 torch_cpu.dll!at::native::to [<unknown file> @ <unknown line number>]
00007FFF8A6FE0E600007FFF8A6EC710 torch_cpu.dll!at::compositeimplicitautograd::unflatten_dense_tensors [<unknown file> @ <unknown line number>]
00007FFF8A8E742300007FFF8A6EC710 torch_cpu.dll!at::compositeimplicitautograd::unflatten_dense_tensors [<unknown file> @ <unknown line number>]
00007FFF8A95123E00007FFF8A6EC710 torch_cpu.dll!at::compositeimplicitautograd::unflatten_dense_tensors [<unknown file> @ <unknown line number>]
00007FFF89D9341500007FFF89CDFB10 torch_cpu.dll!at::_ops::flatten_dense_tensors::redispatch [<unknown file> @ <unknown line number>]
00007FFF89D7BD3900007FFF89CDFB10 torch_cpu.dll!at::_ops::flatten_dense_tensors::redispatch [<unknown file> @ <unknown line number>]
00007FFF89D7BBBF00007FFF89CDFB10 torch_cpu.dll!at::_ops::flatten_dense_tensors::redispatch [<unknown file> @ <unknown line number>]
00007FFF89E0866100007FFF89CDFB10 torch_cpu.dll!at::_ops::flatten_dense_tensors::redispatch [<unknown file> @ <unknown line number>]
00007FFF89CC97FF00007FFF89CC9720 torch_cpu.dll!at::_ops::to_device::call [<unknown file> @ <unknown line number>]
00007FFF88DD677600007FFF88DD6700 torch_cpu.dll!at::Tensor::to [<unknown file> @ <unknown line number>]
00007FFF8EA8CC6000007FFF8EA8ADC0 torch_cpu.dll!torch::jit::Unpickler::readInstruction [<unknown file> @ <unknown line number>]
00007FFF8EA8DF6200007FFF8EA8DE10 torch_cpu.dll!torch::jit::Unpickler::run [<unknown file> @ <unknown line number>]
00007FFF8EA8967D00007FFF8EA89630 torch_cpu.dll!torch::jit::Unpickler::parse_ivalue [<unknown file> @ <unknown line number>]
00007FFF8EA265A100007FFF8EA25ED0 torch_cpu.dll!torch::jit::readArchiveAndTensors [<unknown file> @ <unknown line number>]
00007FFF8EA1DE1B00007FFF8EA1D020 torch_cpu.dll!torch::jit::load [<unknown file> @ <unknown line number>]
00007FFF8EA1DB6000007FFF8EA1D020 torch_cpu.dll!torch::jit::load [<unknown file> @ <unknown line number>]
00007FFF8EA1D1CA00007FFF8EA1D020 torch_cpu.dll!torch::jit::load [<unknown file> @ <unknown line number>]
00007FFF8EA1CE9B00007FFF8EA1CDC0 torch_cpu.dll!torch::jit::load [<unknown file> @ <unknown line number>]
00007FFF8EA1CD7100007FFF8EA1CCF0 torch_cpu.dll!torch::jit::load [<unknown file> @ <unknown line number>]
00007FF7E6598A4B00007FF7E6598840 torch_resnet50.exe!main [E:\C++\torch_resnet50\torch_resnet50\源.cp @ 22]
00007FF7E659AB2900007FF7E659AAF0 torch_resnet50.exe!invoke_main [d:\a01\_work\2\s\src\vctools\crt\vcstartup\src\startup\exe_common.inl @ 79]
00007FF7E659AA0E00007FF7E659A8E0 torch_resnet50.exe!__scrt_common_main_seh [d:\a01\_work\2\s\src\vctools\crt\vcstartup\src\startup\exe_common.inl @ 288]
00007FF7E659A8CE00007FF7E659A8C0 torch_resnet50.exe!__scrt_common_main [d:\a01\_work\2\s\src\vctools\crt\vcstartup\src\startup\exe_common.inl @ 331]
00007FF7E659ABBE00007FF7E659ABB0 torch_resnet50.exe!mainCRTStartup [d:\a01\_work\2\s\src\vctools\crt\vcstartup\src\startup\exe_main.cpp @ 17]
00007FF83EA5703400007FF83EA57020 KERNEL32.DLL!BaseThreadInitThunk [<unknown file> @ <unknown line number>]
00007FF83FB6265100007FF83FB62630 ntdll.dll!RtlUserThreadStart [<unknown file> @ <unknown line number>]
可以看到没法动态链接到CUDA上，最后在官方的issuse上找到了解决方法，只需要在链接器->命令行 输入 -INCLUDE:?warp_size@cuda@at@@YAHXZ
其他版本libntoech配置方法从这位大佬的博文可以找到：Windows下PyTorch(LibTorch)配置cuda加速