Mesh extraction segmentation fault

[cdcseacave]

similar error for any dataset I try:

python mesh_extract.py -s data/bicycle/ -m outputs/bicycle -r 2
Loaded gaussians from outputs/bicycle/point_cloud/iteration_7000/point_cloud.ply
Reading camera 194/194
cameras extent: 4.991958522796631
Segmentation fault (core dumped)

GPU info:

nvidia-smi
Wed Jun 26 12:33:11 2024
+---------------------------------------------------------------------------------------+
| NVIDIA-SMI 535.129.03             Driver Version: 535.129.03   CUDA Version: 12.2     |
|-----------------------------------------+----------------------+----------------------+
| GPU  Name                 Persistence-M | Bus-Id        Disp.A | Volatile Uncorr. ECC |
| Fan  Temp   Perf          Pwr:Usage/Cap |         Memory-Usage | GPU-Util  Compute M. |
|                                         |                      |               MIG M. |
|=========================================+======================+======================|
|   0  NVIDIA A100-SXM4-40GB          On  | 00000000:06:00.0 Off |                    0 |
| N/A   32C    P0              45W / 400W |      4MiB / 40960MiB |      0%      Default |
|                                         |                      |             Disabled |
+-----------------------------------------+----------------------+----------------------+

+---------------------------------------------------------------------------------------+
| Processes:                                                                            |
|  GPU   GI   CI        PID   Type   Process name                            GPU Memory |
|        ID   ID                                                             Usage      |
|=======================================================================================|
|  No running processes found                                                           |
+---------------------------------------------------------------------------------------+

[BaowenZ]

Please refer to #5.

[cdcseacave]

thank you, reducing a lot the resolution helped (even though I have 40GB GPU memory)

is there a way to extract the mesh only for the ROI so that I can preserve details?

[CanCanZeng]

Hello @cdcseacave , I have also encountered the same problem. However, I suspect that this crash issue is not related to the GPU memory size, as it seems that the algorithm is running on the CPU

RaDe-GS/mesh_extract.py

Line 60 in 2d4bc08

o3d_device = o3d.core.Device("CPU:0")

and no sudden increase in GPU memory has been observed. And this issue may not be related to the size of the CPU memory, as my computer has 128GB of CPU memory and I did not observe a sudden increase in memory consumption during program execution.