Failed to build SAM-2 [CUDA compatibility issue ?]

[kevinhelvig]

Hi,
I can't build the SAM-2. I have the following error when running python3.11 -m pip install -e . command

RuntimeError:
 The detected CUDA version (11.8) mismatches the version that was used to compile
  PyTorch (12.1). Please make sure to use the same CUDA versions.

Debug attempts :

• update conda
• try re-install pytorch with associated CUDA11.8 python3.11 -m pip install torch==2.3.1 torchvision==0.18.1 torchaudio==2.3.1 --index-url https://download.pytorch.org/whl/cu118
• also a try with force command

Thanks for help

Conda list of package :

_libgcc_mutex 0.1 main
_openmp_mutex 5.1 1_gnu
bzip2 1.0.8 h5eee18b_6
ca-certificates 2024.7.2 h06a4308_0
filelock 3.13.1 pypi_0 pypi
fsspec 2024.2.0 pypi_0 pypi
jinja2 3.1.3 pypi_0 pypi
ld_impl_linux-64 2.38 h1181459_1
libffi 3.4.4 h6a678d5_1
libgcc-ng 11.2.0 h1234567_1
libgomp 11.2.0 h1234567_1
libstdcxx-ng 11.2.0 h1234567_1
libuuid 1.41.5 h5eee18b_0
markupsafe 2.1.5 pypi_0 pypi
mpmath 1.3.0 pypi_0 pypi
ncurses 6.4 h6a678d5_0
networkx 3.2.1 pypi_0 pypi
numpy 1.26.3 pypi_0 pypi
nvidia-cublas-cu11 11.11.3.6 pypi_0 pypi
nvidia-cuda-cupti-cu11 11.8.87 pypi_0 pypi
nvidia-cuda-nvrtc-cu11 11.8.89 pypi_0 pypi
nvidia-cuda-runtime-cu11 11.8.89 pypi_0 pypi
nvidia-cudnn-cu11 8.7.0.84 pypi_0 pypi
nvidia-cufft-cu11 10.9.0.58 pypi_0 pypi
nvidia-curand-cu11 10.3.0.86 pypi_0 pypi
nvidia-cusolver-cu11 11.4.1.48 pypi_0 pypi
nvidia-cusparse-cu11 11.7.5.86 pypi_0 pypi
nvidia-nccl-cu11 2.20.5 pypi_0 pypi
nvidia-nvtx-cu11 11.8.86 pypi_0 pypi
openssl 3.0.14 h5eee18b_0
pillow 10.2.0 pypi_0 pypi
pip 24.0 py311h06a4308_0
python 3.11.9 h955ad1f_0
readline 8.2 h5eee18b_0
setuptools 69.5.1 py311h06a4308_0
sqlite 3.45.3 h5eee18b_0
sympy 1.12 pypi_0 pypi
tk 8.6.14 h39e8969_0
torch 2.3.1+cu118 pypi_0 pypi
torchaudio 2.3.1+cu118 pypi_0 pypi
torchvision 0.18.1+cu118 pypi_0 pypi
triton 2.3.1 pypi_0 pypi
typing-extensions 4.9.0 pypi_0 pypi
tzdata 2024a h04d1e81_0
wheel 0.43.0 py311h06a4308_0
xz 5.4.6 h5eee18b_1
zlib 1.2.13 h5eee18b_1

[ChinaStark]

The detected CUDA version (11.7) mismatches the version that was used to compile
PyTorch (12.1). Please make sure to use the same CUDA versions.

[ChinaStark]

did you solved ?

[ChinaStark]

cuda 12.1???

[kevinhelvig]

I tried re-installing torch etc ..; with both cuda 11.8 and 12.1 and it still doesn't work

[kevinhelvig]

my cuda version is 12.1 (nvidia-smi command) but even if I reinstall pytorch with cuda 12.1 I have the same error

[ChinaStark]

my cuda version is 12.1 (nvidia-smi command) but even if I reinstall pytorch with cuda 12.1 I have the same error

See this URL, #18

I will try later

[kevinhelvig]

Thanks, the hack proposed in #18 worked for me