Hey All,
Has anyone ran into this error? I am using v4.2 with Topaz 0.2.5-2 that is intalled through SBgrid. I haven’t processed data for several versions of cryosparc, but would use topaz with SBgrid no problem in the past. I’m running 4 3090’s with CUDA 11.6. Any help would be greatly appreciated.
Hi J_Fin,
Could you try using Topaz on the same server through the commandline (not through cryosparc) and see if you get the same error?
Thanks,
-Alex
HI Alex,
Thanks for the quick response!
I ran the preprocessing fine in commandline, but am unsure where or how to obtain the .csv file? I tried exporting but only see .cs files. I thought I had done this before back in cryosparc v2 but don’t quite remember.
Best
Justin
Hi Justin,
Thanks for the info. Did the commandline Topaz run have any errors?
Best,
-Alex
HI ALex,
I haven’t ran it because I’m not sure how to generate the csv file for my particles.
Thanks
Justin
Hi Justin,
Oh I see. Could you try running topaz denoise from the commandline? You can generate a command here: Topaz GUI
Best,
-Alex
Hi ALex,
I received the same error running denoise in commandline (also when running extract in cryrosparc after training a model CPU only). Do you think it maybe an issue with CUDA 11.6? Here is my GPU set up.
Hi Justin,
Sorry for the delay. My guess is that it’s an issue with your Topaz installation or environment. I suggest sending the specific commandline commands and errors to your local IT who manages this server for you so they can fix it.
Best,
-Alex
Hi J_Fin,
Same trouble confused me for days. It seemed caused by the version conflict of cuda. I installed Topaz 0.2.5 following official recommendation with cuda-tool-kit version 9.0 and ran into same error. Using cuda-tool-kit version 9.2 could solve the problem in my computer.
Try remove your exsit topaz conda environment and reinstall by:
conda create -n topaz topaz cudatoolkit=9.2 -c tbepler -c pytorch
Hope it also work for you.
Lai
