Closed xiayizhan2017 closed 5 years ago
I'm getting something similar when running python generate.py configs/demo.yaml
ImportError: dlopen(/Users/{my username}/anaconda3/envs/mesh_funcspace/lib/python3.6/site-packages/torchvision/_C.cpython-36m-darwin.so, 2): Symbol not found: __ZN2at19NonVariableTypeMode10is_enabledEv
Referenced from: /Users/{my username}/anaconda3/envs/mesh_funcspace/lib/python3.6/site-packages/torchvision/_C.cpython-36m-darwin.so
Expected in: flat namespace
in /Users/{my username}/anaconda3/envs/mesh_funcspace/lib/python3.6/site-packages/torchvision/_C.cpython-36m-darwin.so
How do I fix this?
This is usually caused by environment version changes(e.g. python/pytorch version) after library compilation. You may try to delete the compiled binaries and run python setup.py build_ext --inplace
again.
@JiamingSuen Thank you for your answer, but I don't understand the environmental version change.
could you say what version of PyTorch and CUDA this code works with? I've tried by deleting all build and cache files and compiling Pytorch (and torchvision) from source, and using both clang/clang++ and gcc/g++ on MacOS 10.13.6 with CUDA 10.0 but always get the above 'symbol not found' error when trying to run generate.py
I run python3 generate.py configs/demo.yaml I get import error : /im2mesh/dmc/ops/_cuda_ext.cpython-35m-x86_64-linux-gnu.so: undefined symbol: _ZN2atErrorCLENS_14SourceLocationERKSs