Problem with openfold installation.

[zeqri]

Hi,

I was able to install all packages required except for openfold. When trying to install it keep running into this error:

Preparing metadata (setup.py) ... error
error: subprocess-exited-with-error

× python setup.py egg_info did not run successfully.
│ exit code: 1
╰─> [8 lines of output]
Traceback (most recent call last):
File "", line 2, in
File "", line 34, in
File "/gpfs/bwfor/scratch/fr_ma453_o05i14/pip-req-build-qo0kbx1r/setup.py", line 56, in
_, bare_metal_major, _ = get_cuda_bare_metal_version(CUDA_HOME)
File "/gpfs/bwfor/scratch/fr_ma453_o05i14/pip-req-build-qo0kbx1r/setup.py", line 40, in get_cuda_bare_metal_version
raw_output = subprocess.check_output([cuda_dir + "/bin/nvcc", "-V"], universal_newlines=True)
TypeError: unsupported operand type(s) for +: 'NoneType' and 'str'
[end of output]

note: This error originates from a subprocess, and is likely not a problem with pip.
error: metadata-generation-failed

× Encountered error while generating package metadata.
╰─> See above for output.

note: This is an issue with the package mentioned above, not pip.
hint: See above for details.

If someone could give feedback on how to solve this issue I would highly appreciate it.

[tornikeo]

Most likely you need to install cudatoolkit, using conda, or micromamba, like so:

micromamba install cudatoolkit -c nvidia

You get TypeError: unsupported operand type(s) for +: 'NoneType' and 'str' since cuda_dir gives a None unless you have cudatoolkit, which contains the nvcc. Hope this helps.

[ejmeitz]

I had to install cudatoolkit-dev : conda install -c conda-forge cudatoolkit-dev -y

You can also check this specific issue is fixed by opening python and running:

from torch.utils.cpp_extension import CUDA_HOME
print(CUDA_HOME)