Segmentation fault (core dump)
phamquandung opened this issue · 7 comments
Could you please help me with this issue? @ToniRV @jrpowers
When I try to run the example: python ./examples/slam_demo.py --dataset_dir=./datasets/Replica/office0 --dataset_name=nerf --buffer=100 --slam --parallel_run --img_stride=2 --fusion='nerf' --gui
It turns out that: "Module: VioSLAM - Spinning. Segmentation fault (core dumped)" as
below figure
After that, the code continues running as following figure.
Finally, the result shows nothing.
Hi, I have the same problem as you.
I follow the instruction from @jrpowers and it really helps. In the beginning, everything goes fine.
When it comes to the Run part, things get worse.
It turns out that: "Module: VioSLAM - Spinning. Segmentation fault (core dumped).
Have you solved this problem yet?
Hi, I have the same problem as you. I follow the instruction from @jrpowers and it really helps. In the beginning, everything goes fine. When it comes to the Run part, things get worse. It turns out that: "Module: VioSLAM - Spinning. Segmentation fault (core dumped). Have you solved this problem yet?
I then use 2 GPUs and the problem is solved.
Hi, I have the same problem as you. I follow the instruction from @jrpowers and it really helps. In the beginning, everything goes fine. When it comes to the Run part, things get worse. It turns out that: "Module: VioSLAM - Spinning. Segmentation fault (core dumped). Have you solved this problem yet?
I then use 2 GPUs and the problem is solved.
Hello, I had the same problem(core dumped). too.
I tried with 2 GPUs, but I faced same error.
How can I fix this error? Is there anyone who solved this problem?!
No. My problem still exists, and I have to give it up.
I solved the problem with reinstall package named 'torch-scatter' after upgrading pip.
pip install --upgrade pip setuptools wheel
pip uninstall torch-scatter
pip install torch-scatter
I hope this comment helps someone.
Hello @apeach64, I've try uninstall and reinstall torch-scatter, but the problem remains. Can you please provide which version of torch-scatter you use?
Besides, my device has only one GPU. I'm curious if you can still run the demo using one GPU.
Hi @EdwardjkFeng, my torch-scatter version is 2.1.0
Also I still using NeRF-SLAM with one GPU.
I'll provide some information about my environment in case it might be helpful.
My environment spec is
GeForce RTX 3090
ubuntu 18.04
cuda 11.7 + cudnn 8.7.0
I hope you can solve the problem.