Installation of PyTorch when build from source #168
Description
Dear developers,
I have some problems with building openMM-torch from source. After successfully installing openmmtorch in my conda environment following the ccmake steps, I found it did not also install pytorch in my current environment, so I installed it manually. The installation seems to be successful, but after it, importing openmmtorch will give the following error:
>>> import openmmtorch
terminate called after throwing an instance of 'c10::Error'
what(): Tried to register multiple backend fallbacks for the same dispatch key Batched; previous registration registered at /pytorch/aten/src/ATen/core/VariableFallbackKernel.cpp:91, new registration registered at ./aten/src/ATen/BatchingRegistrations.cpp:1064
Exception raised from registerFallback at ./aten/src/ATen/core/dispatch/Dispatcher.cpp:277 (most recent call first):
frame #0: c10::Error::Error(c10::SourceLocation, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) + 0x9e (0x7f10d6d7f9ae in /lib/x86_64-linux-gnu/libc10.so.1.13)
frame #1: c10::detail::torchCheckFail(char const*, char const*, unsigned int, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&) + 0x74 (0x7f10d6d4c79e in /lib/x86_64-linux-gnu/libc10.so.1.13)
frame #2: c10::Dispatcher::registerFallback(c10::DispatchKey, c10::KernelFunction, std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> >) + 0x9b6 (0x7f10d095eaf6 in /lib/x86_64-linux-gnu/libtorch_cpu.so.1.13)
frame #3: torch::Library::_fallback(torch::CppFunction&&) & + 0x4bc (0x7f10d099e2cc in /lib/x86_64-linux-gnu/libtorch_cpu.so.1.13)
frame #4: <unknown function> + 0xd23075 (0x7f10d0723075 in /lib/x86_64-linux-gnu/libtorch_cpu.so.1.13)
frame #5: <unknown function> + 0xd29619 (0x7f10d0729619 in /lib/x86_64-linux-gnu/libtorch_cpu.so.1.13)
frame #6: <unknown function> + 0xcc77ab (0x7f10d06c77ab in /lib/x86_64-linux-gnu/libtorch_cpu.so.1.13)
frame #7: <unknown function> + 0x48be (0x7f12255498be in /lib64/ld-linux-x86-64.so.2)
frame #8: <unknown function> + 0x49a4 (0x7f12255499a4 in /lib64/ld-linux-x86-64.so.2)
frame #9: _dl_catch_exception + 0xd4 (0x7f12253af024 in /lib/x86_64-linux-gnu/libc.so.6)
frame #10: <unknown function> + 0xb09e (0x7f122555009e in /lib64/ld-linux-x86-64.so.2)
frame #11: _dl_catch_exception + 0x7a (0x7f12253aefca in /lib/x86_64-linux-gnu/libc.so.6)
frame #12: <unknown function> + 0xb438 (0x7f1225550438 in /lib64/ld-linux-x86-64.so.2)
frame #13: <unknown function> + 0x85438 (0x7f12252e5438 in /lib/x86_64-linux-gnu/libc.so.6)
frame #14: _dl_catch_exception + 0x7a (0x7f12253aefca in /lib/x86_64-linux-gnu/libc.so.6)
frame #15: _dl_catch_error + 0x2f (0x7f12253af07f in /lib/x86_64-linux-gnu/libc.so.6)
frame #16: <unknown function> + 0x84f27 (0x7f12252e4f27 in /lib/x86_64-linux-gnu/libc.so.6)
frame #17: dlopen + 0x69 (0x7f12252e54e9 in /lib/x86_64-linux-gnu/libc.so.6)
<omitting python frames>
frame #44: <unknown function> + 0x2724a (0x7f122528724a in /lib/x86_64-linux-gnu/libc.so.6)
frame #45: __libc_start_main + 0x85 (0x7f1225287305 in /lib/x86_64-linux-gnu/libc.so.6)
I suppose this is because openmmtorch is dependent on the libtorch libraries, thus installing another pytorch will have influences. But how can I get pytorch appropriately in my conda environment, if I want to build from the source?