search

Jumpat / SegmentAnythingin3D

Segment Anything in 3D with NeRFs (NeurIPS 2023)
Apache License 2.0
862 stars 53 forks source link

Mismatched CUDA Version? #64

Closed STF04 closed 3 months ago

STF04 commented 5 months ago

When attempting to download, and running the following line: "cd GroundingDINO/; pip install -e ."

I get this corresponding error: RuntimeError: The detected CUDA version (11.7) mismatches the version that was used to compile PyTorch (10.2). Please make sure to use the same CUDA versions.

Has anyone encountered this error and knows the solution, any help is greatly appreciated.

Jumpat commented 5 months ago

Hi, uninstall torch and try this. pip install torch==1.12.1+cu117 --extra-index-url https://download.pytorch.org/whl/cu117

zratclif commented 5 months ago

I am also having troubles with this; I get the same error. I entered the command you commented, and I get the following error: ERROR: Could not find a version that satisfies the requirement torch==1.12.1+cu117 (from versions: 1.11.0, 1.12.0, 1.12.1, 1.13.0, 1.13.0+cu117, 1.13.1, 1.13.1+cu117, 2.0.0, 2.0.0+cu117, 2.0.1, 2.0.1+cu117, 2.1.0, 2.1.1, 2.1.2, 2.2.0, 2.2.1, 2.2.2) ERROR: No matching distribution found for torch==1.12.1+cu117

Jumpat commented 5 months ago

I am also having troubles with this; I get the same error. I entered the command you commented, and I get the following error: ERROR: Could not find a version that satisfies the requirement torch==1.12.1+cu117 (from versions: 1.11.0, 1.12.0, 1.12.1, 1.13.0, 1.13.0+cu117, 1.13.1, 1.13.1+cu117, 2.0.0, 2.0.0+cu117, 2.0.1, 2.0.1+cu117, 2.1.0, 2.1.1, 2.1.2, 2.2.0, 2.2.1, 2.2.2) ERROR: No matching distribution found for torch==1.12.1+cu117

@zratclif This error is always disturbing😂. I think you can try to install another version of torch compiled with cu117. Our code does not have strict constrain on torch version.

zratclif commented 5 months ago

Thank you, that part finally worked. I have encountered another error: 'Segmentation fault (core dumped)' . This happens when I try to train NERF, which is running this command 'python run.py --config=configs/llff/fern.py --stop_at=20000 --render_video --i_weights=10000'. Do you know what could be the cause to this error? Thanks again for the help.

Jumpat commented 5 months ago

Thank you, that part finally worked. I have encountered another error: 'Segmentation fault (core dumped)' . This happens when I try to train NERF, which is running this command 'python run.py --config=configs/llff/fern.py --stop_at=20000 --render_video --i_weights=10000'. Do you know what could be the cause to this error? Thanks again for the help.

Nope. @zratclif You may have to check the cuda version, PyTorch version and so on. I think this is caused by the wrongly complied cuda extension. Segmentation fault is always the last problem we want to meet.😂

AsaKal commented 4 months ago

Hi i am getting OSError: libcudart.so.11.0: cannot open shared object file: No such file or directory. My system has cuda version 12.1

Jumpat commented 4 months ago

Hi i am getting OSError: libcudart.so.11.0: cannot open shared object file: No such file or directory. My system has cuda version 12.1

@AsaKal 12.1 is too high. You may need 11.7. You can refer to the repo of 3D-GS for more information.

AsaKal commented 4 months ago

Okay thanks @Jumpat i will check it out