I’m getting warnings when trying to run Total Segmentator.
The code for the first warning is as follows:
Traceback (most recent call last):
File “D:\3D Slicer\Slicer 5.2.2\bin\Python\slicer\util.py”, line 2967, in tryWithErrorDisplay
yield
File “D:/3D Slicer/Slicer 5.2.2/NA-MIC/Extensions-31382/TotalSegmentator/lib/Slicer-5.2/qt-scripted-modules/TotalSegmentator.py”, line 279, in onPackageInfoUpdate
self.ui.packageInfoTextBrowser.plainText = self.logic.installedTotalSegmentatorPythonPackageInfo().rstrip()
File “D:/3D Slicer/Slicer 5.2.2/NA-MIC/Extensions-31382/TotalSegmentator/lib/Slicer-5.2/qt-scripted-modules/TotalSegmentator.py”, line 568, in installedTotalSegmentatorPythonPackageInfo
versionInfo = subprocess.check_output([shutil.which(‘PythonSlicer’), “-m”, “pip”, “show”, “TotalSegmentator”]).decode()
File “D:\3D Slicer\Slicer 5.2.2\lib\Python\Lib\subprocess.py”, line 424, in check_output
return run(*popenargs, stdout=PIPE, timeout=timeout, check=True,
File “D:\3D Slicer\Slicer 5.2.2\lib\Python\Lib\subprocess.py”, line 528, in run
raise CalledProcessError(retcode, process.args,
subprocess.CalledProcessError: Command ‘[‘D:/3D Slicer/Slicer 5.2.2/bin/…/bin\PythonSlicer.EXE’, ‘-m’, ‘pip’, ‘show’, ‘TotalSegmentator’]’ returned non-zero exit status 1.
The other issue is that the GPU cannot be found error. I downloaded CUDA. However, the same error persists.
TotalSegmentator v2 is available as of the Slicer 5.6.0 release. Please download the latest Slicer stable (https://download.slicer.org/) and install the Total Segmentator extension to receive support related to Total Segmentator.
Please consult here in the Total Segmentator readme regarding information about the licenses associated with the various sub tasks.
I downloaded and installed Slicer 5.6.1. But total segmentator cannot detect my GPU again. I have two graphic cards: Intel and Nvidia. I tried this solution of problem like this topic (Total Segmenter cannot detect my GPU (1080ti x 2)). But it’s not working.
Try to follow the instructions linked below regarding using the PyTorch Util module to check the installed pytorch version and to install/uninstall versions as part of debugging.
Also can you provide details about which Nvidia GPU your system has installed?
Yes you’re going to have to remain on the CPU version of pytorch as the GeForce GT 720M “Fermi” based microarchitecture is not supported by the CUDA compute capability of latest pytorch. However, based on the age of your mobile GPU, I would expect that your CPU and memory are also going to struggle running inference. All these new ML tools generally require fairly recent hardware to run well.
You can probably run the segmentation on the CPU, just make sure you force using the CPU. Low-resolution segmentation may be completed in a couple of minutes, but full-resolution segmentation on CPU will take tens of minutes.
Thank you very much for your detailed answer. Yes, my computer is 11 years old. However, I can easily run graphics-based software in terms of CPU, RAM and SSD. I have no problems with manual segmentation either. However, I encountered such a problem in automatic segmentation. All segmentations are completed in approximately 30 minutes (long version). Thank you again!
Full resolution segmentation via CPU takes approximately 30 minutes. Low-resolution segmentation is completed in a few minutes, but it is not of much use to me.
Father of Total Segmentator, thank you very much for your efforts!
Thanks for the update, I’m happy that things work for you.
Note that I developed and I am supporting the Slicer module for TotalSegmentator, but the main person who developed the TotalSegmentator model is Jakob Wasserthal.
