YuxueYang1204
commented
3 months ago It seems like you might have installed the CPU version of PyTorch. It looks like the error you're encountering is due to libcudart.so.10.1: cannot open shared object file: No such file or directory. This issue is likely related to the configuration of PyTorch3D. I recommend checking the PyTorch3D documentation or their support channels for a solution. Additionally, here is the partial information about my environment, which I hope might be helpful to you:
pytorch 2.1.0 py3.8_cuda12.1_cudnn8.9.2_0 pytorch
pytorch-cuda 12.1 ha16c6d3_5 pytorch
pytorch-mutex 1.0 cuda pytorch
pytorch-scatter 2.1.2 py38_torch_2.1.0_cu121 pyg
pytorch3d 0.7.5 py38_cu121_pyt210 pytorch3d
Abyssaledge
When I ran python scripts/run_dtu.py, no error occurred, but it did not work either. I could see no activity in nvidia-smi.
Previously, I encountered the following errors:
File "/hdd/jhee/3D/TrimGS/Trim3DGS/scene/dataset_readers.py", line 24, in <module> from scene.gaussian_model import BasicPointCloud File "/hdd/jhee/3D/TrimGS/Trim3DGS/scene/gaussian_model.py", line 19, in <module> from pytorch3d.ops import knn_points File "/home/jhee/miniconda3/envs/trimgs/lib/python3.8/site-packages/pytorch3d/ops/__init__.py", line 5, in <module> from .graph_conv import GraphConv File "/home/jhee/miniconda3/envs/trimgs/lib/python3.8/site-packages/pytorch3d/ops/graph_conv.py", line 8, in <module> from pytorch3d import _C ImportError: libcudart.so.10.1: cannot open shared object file: No such file or directory`(trimgs) jhee@vsclab04:/hdd/jhee/3D/TrimGS/Trim3DGS$ python scripts/run_dtu.py Starting job on GPU 0 with scene 24OMP_NUM_THREADS=4 CUDA_VISIBLE_DEVICES=0 python train.py -s data/dtu_dataset/DTU/scan24 -m output/DTU_3DGS/scan24 -r 2
Traceback (most recent call last):
File "train.py", line 16, in
from gaussian_renderer import render, network_gui
File "/hdd/jhee/3D/TrimGS/Trim3DGS/gaussian_renderer/init.py", line 14, in
from diff_gaussian_rasterization import GaussianRasterizationSettings, GaussianRasterizer
File "/home/jhee/miniconda3/envs/trimgs/lib/python3.8/site-packages/diff_gaussian_rasterization/init.py", line 15, in
from . import _C
ImportError: /home/jhee/miniconda3/envs/trimgs/lib/python3.8/site-packages/torch/lib/libtorch_cuda_cpp.so: undefined symbol: _ZTIN4c10d12ProcessGroup4WorkE
Job (0, (24, 2)) has finished., rellasing GPU 0
Starting job on GPU 0 with scene 37
OMP_NUM_THREADS=4 CUDA_VISIBLE_DEVICES=0 python train.py -s data/dtu_dataset/DTU/scan37 -m output/DTU_3DGS/scan37 -r 2
dispatch_jobs(jobs, executor)
File "scripts/run_dtu.py", line 74, in dispatch_jobs
time.sleep(5)
KeyboardInterrupt`
Traceback (most recent call last):
File "train.py", line 16, in
from gaussian_renderer import render, network_gui
File "/hdd/jhee/3D/TrimGS/Trim3DGS/gaussian_renderer/init.py", line 14, in
from diff_gaussian_rasterization import GaussianRasterizationSettings, GaussianRasterizer
File "/home/jhee/miniconda3/envs/trimgs/lib/python3.8/site-packages/diff_gaussian_rasterization/init.py", line 15, in
from . import _C
ImportError: /home/jhee/miniconda3/envs/trimgs/lib/python3.8/site-packages/torch/lib/libtorch_cuda_cpp.so: undefined symbol: _ZTIN4c10d12ProcessGroup4WorkE
Job (0, (37, 2)) has finished., rellasing GPU 0
^CTraceback (most recent call last):
File "scripts/run_dtu.py", line 81, in