Create Env Error
Opened this issue · 6 comments
when i try to create env. i get this error anyone know why is that
Installing pip dependencies: - Ran pip subprocess with arguments:
['C:\\Users\\alpak\\anaconda3\\envs\\halp\\envs\\c3dgs\\python.exe', '-m', 'pip', 'install', '-U', '-r', 'C:\\Users\\alpak\\OneDrive\\Desktop\\alp\\take_home_exam\\SLAMPy-Monocular-SLAM-implementation-in-Python\\Compact-3DGS\\condaenv.f9b2y9ry.requirements.txt', '--exists-action=b']
Pip subprocess output:
Collecting git+https://github.com/NVlabs/tiny-cuda-nn/#subdirectory=bindings/torch (from -r C:\Users\alpak\OneDrive\Desktop\alp\take_home_exam\SLAMPy-Monocular-SLAM-implementation-in-Python\Compact-3DGS\condaenv.f9b2y9ry.requirements.txt (line 3))
Cloning https://github.com/NVlabs/tiny-cuda-nn/ to c:\users\alpak\appdata\local\temp\pip-req-build-tzed9wl8
Resolved https://github.com/NVlabs/tiny-cuda-nn/ to commit 235d1fde956dc04966940f9d1bec66aa3bdb705a
Preparing metadata (setup.py): started
Preparing metadata (setup.py): finished with status 'error'
Pip subprocess error:
Running command git clone --filter=blob:none --quiet https://github.com/NVlabs/tiny-cuda-nn/ 'C:\Users\alpak\AppData\Local\Temp\pip-req-build-tzed9wl8'
Running command git submodule update --init --recursive -q
error: subprocess-exited-with-error
× python setup.py egg_info did not run successfully.
│ exit code: 1
╰─> [10 lines of output]
C:\Users\alpak\AppData\Local\Temp\pip-req-build-tzed9wl8\bindings/torch\setup.py:5: DeprecationWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html
from pkg_resources import parse_version
Traceback (most recent call last):
File "<string>", line 36, in <module>
File "<pip-setuptools-caller>", line 34, in <module>
File "C:\Users\alpak\AppData\Local\Temp\pip-req-build-tzed9wl8\bindings/torch\setup.py", line 51, in <module>
raise EnvironmentError("Unknown compute capability. Specify the target compute capabilities in the TCNN_CUDA_ARCHITECTURES environment variable or install PyTorch with the CUDA backend to detect it automatically.")
OSError: Unknown compute capability. Specify the target compute capabilities in the TCNN_CUDA_ARCHITECTURES environment variable or install PyTorch with the CUDA backend to detect it automatically.
No CUDA runtime is found, using CUDA_HOME='C:\Program Files\NVIDIA GPU Computing Toolkit\CUDA\v11.1'
Building PyTorch extension for tiny-cuda-nn version 1.7
[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.
failed
CondaEnvException: Pip failed
Are you trying on Windows?
It seems to have some trouble installing tiny-cudn-nn, so please refer to its requirements (https://github.com/NVlabs/tiny-cuda-nn).
Thanks!
I get this error too, Have you found a good solution?
Installing pip dependencies: \ Ran pip subprocess with arguments:
['C:\\Users\\zwj\\.conda\\envs\\c3dgs\\python.exe', '-m', 'pip', 'install', '-U', '-r', 'E:\\Code\\Python\\SZU\\Compact-3DGS\\condaenv.w3k2v8oo.requirements.txt', '--exists-action=b']
Pip subprocess output:
Collecting git+https://github.com/NVlabs/tiny-cuda-nn/#subdirectory=bindings/torch (from -r E:\Code\Python\SZU\Compact-3DGS\condaenv.w3k2v8oo.requirements.txt (line 3))
Cloning https://github.com/NVlabs/tiny-cuda-nn/ to c:\users\zwj\appdata\local\temp\pip-req-build-x7u26lvj
Resolved https://github.com/NVlabs/tiny-cuda-nn/ to commit 235d1fde956dc04966940f9d1bec66aa3bdb705a
Preparing metadata (setup.py): started
Preparing metadata (setup.py): finished with status 'error'
Pip subprocess error:
Running command git clone --filter=blob:none --quiet https://github.com/NVlabs/tiny-cuda-nn/ 'C:\Users\zwj\AppData\Local\Temp\pip-req-build-x7u26lvj'
Running command git submodule update --init --recursive -q
error: subprocess-exited-with-error
× python setup.py egg_info did not run successfully.
│ exit code: 1
╰─> [10 lines of output]
C:\Users\zwj\AppData\Local\Temp\pip-req-build-x7u26lvj\bindings/torch\setup.py:5: DeprecationWarning: pkg_resources is deprecated as an API. See https://setuptools.pypa.io/en/latest/pkg_resources.html
from pkg_resources import parse_version
Traceback (most recent call last):
File "<string>", line 36, in <module>
File "<pip-setuptools-caller>", line 34, in <module>
File "C:\Users\zwj\AppData\Local\Temp\pip-req-build-x7u26lvj\bindings/torch\setup.py", line 72, in <module>
raise RuntimeError("Could not locate a supported Microsoft Visual C++ installation")
RuntimeError: Could not locate a supported Microsoft Visual C++ installation
Building PyTorch extension for tiny-cuda-nn version 1.7
Obtained compute capability 86 from PyTorch
[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.
failed
CondaEnvException: Pip failed
I use VS Code as my code editor, but it's mentioned that it has been tested and works flawlessly in Visual Studio 2022 under the tinny-cuda repository. Additionally, when installing Cuda Toolkit 11.6, it says that it couldn't find a correctly matched version of Visual Studio. Do you think the problem might stem from this? Could it be that VS Code as a code editor is insufficient?
Because I'm not a user of either Windows or VSCode, I can not be sure, but I think VSCode is not the matter.
Have you tried the following command before creating env? (command line for Windows users in 3DGS):
SET DISTUTILS_USE_SDK=1
If you can create an env of 3DGS, based on which you can set by installing tiny-cuda-nn manually I think.
Yes i tried that command before creating env but I think the problem stems from my GPU; even though PyTorch is installed correctly, it seems it is not installed with CUDA support, and my current GPU, I believe, does not support this CUDA version, which causes issues with the tiny-CUDA-NN package. However, I still want to ask one last thing: Do you think manually installing the packages expected to be installed via pip one by one is an appropriate method?
Thanks again.
It would be nice to check that installing tinycudann is the only matter.
Once you succeed in setting 3DGS's environment, you may install tinycudann (building manually if needed) and a few other packages (e.g., dahuffman, pytorch VQ).