segmentation fault

[pconway11]

I have installed detectron2 and torch 1.10 with pip

I am getting a segmentation fault as detectron 2 seems to look for torch in a non-existent directory.
Is there an environment variable or something I need to set so it looks in the correct place

see trace below
/opt/python/ does not exist

Thread 1 "python3" received signal SIGSEGV, Segmentation fault. 0x00007fffd5d51c94 in c10::detail::infer_schema::make_function_schema(std::string&&, std::string&&, c10::ArrayRef<c10::detail::infer_schema::ArgumentDef>, c10::ArrayRef<c10::detail::infer_schema::ArgumentDef>) () from /home/ubuntu/.local/lib/python3.8/site-packages/torch/lib/libtorch_cpu.so (gdb) bt #0 0x00007fffd5d51c94 in c10::detail::infer_schema::make_function_schema(std::string&&, std::string&&, c10::ArrayRef<c10::detail::infer_schema::ArgumentDef>, c10::ArrayRef<c10::detail::infer_schema::ArgumentDef>) () from /home/ubuntu/.local/lib/python3.8/site-packages/torch/lib/libtorch_cpu.so #1 0x00007fffd5d5248a in c10::detail::infer_schema::make_function_schema(c10::ArrayRef<c10::detail::infer_schema::ArgumentDef>, c10::ArrayRef<c10::detail::infer_schema::ArgumentDef>) () from /home/ubuntu/.local/lib/python3.8/site-packages/torch/lib/libtorch_cpu.so #2 0x00007fffc6808467 in c10::detail::infer_schema::createFunctionSchemaFromTraitsFlattenedReturns<c10::guts::function_traits<at::Tensor (at::Tensor const&, at::Tensor const&, double)> >() () at /opt/python/cp38-cp38/lib/python3.8/site-packages/torch/include/ATen/core/op_registration/infer_schema.h:125 #3 c10::inferFunctionSchemaFlattenedReturns<at::Tensor (at::Tensor const&, at::Tensor const&, double)>() () at /opt/python/cp38-cp38/lib/python3.8/site-packages/torch/include/ATen/core/op_registration/infer_schema.h:149 #4 c10::detail::inferFunctionSchemaFromFunctor<at::Tensor (*)(at::Tensor const&, at::Tensor const&, double)> () at /opt/python/cp38-cp38/lib/python3.8/site-packages/torch/include/ATen/core/op_registration/op_registration.h:30 #5 0x00007fffc67fc869 in torch::CppFunction::CppFunction<at::Tensor (at::Tensor const&, at::Tensor const&, double)>(at::Tensor (*)(at::Tensor const&, at::Tensor const&, double), std::enable_if<c10::guts::is_function_type<at::Tensor (at::Tensor const&, at::Tensor const&, double)>::value, decltype(nullptr)>::type) (f=0x7fffc6803960 <detectron2::nms_rotated(at::Tensor const&, at::Tensor const&, double)>, this=0x7fffffff5b20) at /opt/python/cp38-cp38/lib/python3.8/site-packages/torch/include/torch/library.h:111 #6 torch::Library::def<char const (&) [12], at::Tensor (*)(at::Tensor const&, at::Tensor const&, double)>(char const (&) [12], at::Tensor (*&&)(at::Tensor const&, at::Tensor const&, double)) & (raw_name_or_schema=..., raw_f=<optimized out>, this=0x7fffc6a4c9e0 <detectron2::TORCH_LIBRARY_static_init_detectron2>) at /opt/python/cp38-cp38/lib/python3.8/site-packages/torch/include/torch/library.h:542 #7 detectron2::TORCH_LIBRARY_init_detectron2 (m=...) at /detectron2/detectron2/layers/csrc/vision.cpp:126 #8 torch::detail::TorchLibraryInit::TorchLibraryInit (line=125, file=0x7fffc6840530 "/detectron2/detectron2/layers/csrc/vision.cpp", k=..., ns=0x7fffc683f098 "detectron2", fn=<optimized out>, kind=torch::Library::DEF, this=0x7fffc6a4c9e0 <detectron2::TORCH_LIBRARY_static_init_detectron2>) at /opt/python/cp38-cp38/lib/python3.8/site-packages/torch/include/torch/library.h:712 #9 __static_initialization_and_destruction_0 (__initialize_p=1, __priority=65535) at /detectron2/detectron2/layers/csrc/vision.cpp:125 #10 _GLOBAL__sub_I_vision.cpp(void) () at /detectron2/detectron2/layers/csrc/vision.cpp:129 #11 0x00007ffff7fe0b8a in ?? () from /lib64/ld-linux-x86-64.so.2 #12 0x00007ffff7fe0c91 in ?? () from /lib64/ld-linux-x86-64.so.2 #13 0x00007ffff7f2c895 in __GI__dl_catch_exception (exception=<optimized out>, operate=<optimized out>, args=<optimized out>) at dl-error-skeleton.c:182 #14 0x00007ffff7fe50bf in ?? () from /lib64/ld-linux-x86-64.so.2 #15 0x00007ffff7f2c838 in __GI__dl_catch_exception (exception=<optimized out>, operate=<optimized out>, args=<optimized out>) at dl-error-skeleton.c:208 #16 0x00007ffff7fe45fa in ?? () from /lib64/ld-linux-x86-64.so.2 #17 0x00007ffff7da434c in dlopen_doit (a=a@entry=0x7fffffff61b0) at dlopen.c:66 #18 0x00007ffff7f2c838 in __GI__dl_catch_exception (exception=exception@entry=0x7fffffff6150, operate=<optimized out>, args=<optimized out>) at dl-error-skeleton.c:208 #19 0x00007ffff7f2c903 in __GI__dl_catch_error (objname=0xa15ef0, errstring=0xa15ef8, mallocedp=0xa15ee8, operate=<optimized out>, args=<optimized out>) at dl-error-skeleton.c:227 #20 0x00007ffff7da4b59 in _dlerror_run (operate=operate@entry=0x7ffff7da42f0 <dlopen_doit>, args=args@entry=0x7fffffff61b0) at dlerror.c:170 #21 0x00007ffff7da43da in __dlopen (file=<optimized out>, mode=<optimized out>) at dlopen.c:87

[pconway11]

installing from source fixed the issue..
error happens with prebuilt CPU version..

python -m pip install 'git+https://github.com/facebookresearch/detectron2.git'