Open wcoath opened 2 years ago
Hi Will,
Thanks for the message! I'm at a conference right now until Thursday and then off on vacation for a week. I'll try to have a look at it before then, but if not I'll check when ai get back in early July. Don't hesitate to email me to remind me if I don't get back to you soon!
Best
Thomas
On Tue, Jun 21, 2022, 15:23 Will Coath @.**@.>> wrote:
Hello,
I'm using APPIAN to run an SUVR pipeline with T1, freesurfer segmentation and static florbetapir PET image that is one single frame of 50-60mins post-injection. I'm using the latest release downloaded from dockerhub on 19th May 2022, running it through singularity on a HPC cluster (as I can't use docker on the HPC).
This is the command that is being run: python3 /opt/APPIAN/Launcher.py -s /appian_test/data -t /appian_test/cerebellum_suvr_10015124_baseline --sub 10015124 --ses baseline --no-group-level --quant-method suvr --start-time 3000 --end-time 3600 --quant-label-img fs-segm --quant-label 7 8 46 47 --quant-labels-ones-only --quant-label-space t1 --no-mri-mask --normalization-type nl --results-label-img fs-segm --results-label-space t1 --threads 4
It fails consistently on the 'brain_mask_space_pet' step after completing normalisation step. I've attached the full log file. Any help working out what is going wrong would be much appreciated! Many thanks - Will
appian_log.txthttps://github.com/APPIAN-PET/APPIAN/files/8949722/appian_log.txt
— Reply to this email directly, view it on GitHubhttps://github.com/APPIAN-PET/APPIAN/issues/38, or unsubscribehttps://github.com/notifications/unsubscribe-auth/ABUJ4PEVOUXQ4UIIUPW73C3VQHF5LANCNFSM5ZMRJAXQ. You are receiving this because you are subscribed to this thread.Message ID: @.***>
Neugierige sind herzlich willkommen am Sonntag, den 21. August 2022, von 10:00 bis 17:00 Uhr. Mehr unter: https://www.tagderneugier.de
Hi Thomas, Hope you had a good conference and vacation! I'm just wondering if you've had a chance to look at this yet? Cheers! Will
Hi Will, I did and just pushed the change last night. Can you run it again and let me know how it goes?
Thomas
On Fri, Jul 15, 2022, 05:12 Will Coath @.***> wrote:
Hi Thomas, Hope you had a good conference and vacation! I'm just wondering if you've had a chance to look at this yet? Cheers! Will
— Reply to this email directly, view it on GitHub https://github.com/APPIAN-PET/APPIAN/issues/38#issuecomment-1185345443, or unsubscribe https://github.com/notifications/unsubscribe-auth/ABUJ4PD3LZCN2BUYFBPLVRLVUETQNANCNFSM5ZMRJAXQ . You are receiving this because you commented.
Message ID: @.***>
Hello,
I'm using APPIAN to run an SUVR pipeline with T1, freesurfer segmentation and static florbetapir PET image that is one single frame of 50-60mins post-injection. I'm using the latest release downloaded from dockerhub on 19th May 2022, running it through singularity on a HPC cluster (as I can't use docker on the HPC).
This is the command that is being run: python3 /opt/APPIAN/Launcher.py -s /appian_test/data -t /appian_test/cerebellum_suvr_10015124_baseline --sub 10015124 --ses baseline --no-group-level --quant-method suvr --start-time 3000 --end-time 3600 --quant-label-img fs-segm --quant-label 7 8 46 47 --quant-labels-ones-only --quant-label-space t1 --no-mri-mask --normalization-type nl --results-label-img fs-segm --results-label-space t1 --threads 4
It fails consistently on the 'brain_mask_space_pet' step after completing normalisation step. I've attached the full log file. Any help working out what is going wrong would be much appreciated! Many thanks - Will
appian_log.txt