Daniel
Daniel

Reputation: 385

Cannot setup package in conda environment with Pytorch installed

All

After setting up the PyTorch 1.7.1 with CUDA 11.2 on a conda virtual environment, I run python setup.py install it always returns me the following error message.

Traceback (most recent call last):
  File "setup.py", line 2, in <module>
    from torch.utils.cpp_extension import BuildExtension, CUDAExtension
  File "anaconda3/envs/pytorch/lib/python3.7/site-packages/torch/__init__.py", line 189, in <module>
    _load_global_deps()
  File "anaconda3/envs/pytorch/lib/python3.7/site-packages/torch/__init__.py", line 142, in _load_global_deps
    ctypes.CDLL(lib_path, mode=ctypes.RTLD_GLOBAL)
  File "anaconda3/envs/pytorch/lib/python3.7/ctypes/__init__.py", line 364, in __init__
    self._handle = _dlopen(self._name, mode)
OSError: anaconda3/envs/pytorch/lib/python3.7/site-packages/torch/lib/../../../../libcublas.so.11: symbol free_gemm_select version libcublasLt.so.11 not defined in file libcublasLt.so.11 with link time reference

Could anyone can help with this? Thanks!

Upvotes: 4

Views: 4638

Answers (1)

Daniel
Daniel

Reputation: 385

Finally, I find the solution by just using the pip from the Pytorch official website.

pip install torch==1.7.1+cu110 torchvision==0.8.2+cu110 torchaudio===0.7.2 -f https://download.pytorch.org/whl/torch_stable.html

Upvotes: 3

Related Questions