zsh: segmentation fault

[giovannividetta]

Hello, everyone!

I am trying to apply TractSeg on my diffusion MRI data. In order to understand the algorithm, I am running TractSeg on the example data provided by @wasserth in his GitHub repository. Launching on terminal TractSeg -i Diffusion.nii.gz --raw_diffusion_input, a segfault error appears as following:

Creating brain mask...
Warning: An input intended to be a single 3D volume has multiple timepoints. Input will be truncated to first volume, but this functionality is deprecated and will be removed in a future release.
Creating peaks (1 of 3)...
Creating peaks (2 of 3)...
Creating peaks (3 of 3)...
Loading weights from: /Users/brainmilab/.tractseg/pretrained_weights_tract_segmentation_v3.npz
Processing direction (1 of 3)
zsh: segmentation fault TractSeg -i Diffusion.nii.gz --raw_diffusion_input

I looked for a solution on websites and forums, but I did not find any of relevant. I downloaded pretrained_weights_tract_segmentation_v3.npz manually from zenodo (https://zenodo.org/records/5052821) because I thought that file was corrupted when it was automatically downloaded by TractSeg. However, the segfault error persisted. I do not know if it would be a problem concerning memory access. I am using Python 3.9.13, FSL 6.0.6.2, pytorch 1.13.1, numpy 1.21.5, and MacOs 13.6.6.

Thank you for your support!

[wasserth]

Unfortunately a "segmentation fault" error can mean anything. Therefore, I can not really tell what the problem might be. Generally speaking TractSeg is most optimised for Linux. So if you have a Linux / Ubuntu machine available you can try to run on this again.

[giovannividetta]

Hi, @wasserth!

I apologise if I re-write to you. I followed your suggestion and I tried to run TractSeg on Windows 10 by using Ubuntu 22.04.3 LTS. However, althought "segmentation fault" error has been resolved, now the step of "Procesing direction (2 of 3)" is killed by the script. I reported the command line on the terminal:

$brainmilab@brainmilab:~/tratto_gio/examples TractSeg -i Diffusion.nii.gz --raw_diffusion_input
Creating brain mask...
Warning: An input intended to be a single 3D volume has multiple timepoints. Input will be truncated to first volume, but this functionality is deprecated and will be removed in a future release.
Creating peaks (1 of 3)...
Creating peaks (2 of 3)...
Creating peaks (3 of 3)...
Loading weights from: /home/brainmilab/.tractseg/pretrained_weights_tract_segmentation_v3.npz
Processing direction (1 of 3)
100%|█████████████████████████████████████████████████████████████████████████████████| 144/144 [00:27<00:00, 5.28it/s]Processing direction (2 of 3)
0%| | 0/144 [00:00<?, ?it/s]Killed

Is it a common error or am I making any mitake?
Thanks again for your time and support!