Opening this issue following the advice of @pieper :
I installed the automatic version of pytorch using the extension. Installation completed successfully, and then Slicer crashed.
Now if I open Slicer and try to open the PyTorch module, Slicer crashes.
In the Python console, I can import torch, and it says cuda is available. If I import torchvision, Slicer crashes.
In the commandline PythonSlicer, I can import both torch and torchvsion, there seems to be no problem.
Python 3.12.10 (main, Jul 15 2026, 15:05:10) [GCC 14.2.1 20250110 (Red Hat 14.2.1-7)] on linux
Type "help", "copyright", "credits" or "license" for more information.
>>> import torchvision
>>> torchvision.__version__
'0.28.0+cu130'
>>> torchvision.version.cuda
13000
>>> torchvision.version.git_version
'8fb87713a24951e639c494b0f2a8a81b5f8e33a6'
The issue may be related to #19 and dalcalab/SlicerMultiverSeg#34 -- at least, I do see the same wheel versions,
$ cat Slicer-5.12.2-linux-amd64/lib/Python/lib/python3.12/site-packages/torchvision-0.28.0+cu130.dist-info/WHEEL
Wheel-Version: 1.0
Generator: setuptools (72.1.0)
Root-Is-Purelib: false
Tag: cp312-cp312-manylinux_2_28_x86_64
$ cat Slicer-5.12.2-linux-amd64/lib/Python/lib/python3.12/site-packages/torch-2.13.0+cu130.dist-info/WHEEL
Wheel-Version: 1.0
Generator: setuptools (81.0.0)
Root-Is-Purelib: false
Tag: cp312-cp312-manylinux_2_28_x86_64
Ubuntu 22.04, and as you can see from the above, Slicer 5.12.2, cuda 13.0
Thu Jul 23 11:52:04 2026
+-----------------------------------------------------------------------------------------+
| NVIDIA-SMI 580.173.02 Driver Version: 580.173.02 CUDA Version: 13.0 |
+-----------------------------------------+------------------------+----------------------+
| 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 GeForce RTX 3080 ... Off | 00000000:01:00.0 Off | N/A |
| N/A 42C P8 12W / 115W | 18MiB / 16384MiB | 0% Default |
| | | N/A |
+-----------------------------------------+------------------------+----------------------+
Opening this issue following the advice of @pieper :
I installed the automatic version of pytorch using the extension. Installation completed successfully, and then Slicer crashed.
Now if I open Slicer and try to open the PyTorch module, Slicer crashes.
In the Python console, I can import
torch, and it says cuda is available. If I importtorchvision, Slicer crashes.In the commandline
PythonSlicer, I can import bothtorchandtorchvsion, there seems to be no problem.The issue may be related to #19 and dalcalab/SlicerMultiverSeg#34 -- at least, I do see the same wheel versions,
Ubuntu 22.04, and as you can see from the above, Slicer 5.12.2, cuda 13.0