SPACETomo model fail to load

[xtalfire]

Hi Fabian,

We have a G4 Titan with Falcon4i, and SerialEM runs on both microscope PC and support PC. I installed SPACEtomo[gui] on MPC and SPACEtomo[gui, predict] on SPC. I downloaded the 2 models and added and activated them to config.py successfully. But when i run SPACEtomo_run.py or SPACEtomo_manMap.py, both MPC and SPC instances gave me error message"Lamella segmentation model could not be found! Please import it using the SPACEmodel command!"
SPACEmodel list on the terminal shows 2 models were active, so I wonder what could have I missed? Thanks a lot

Best,
Tat

[eisfabian]

Hi Tat,

Sorry for the late reply.
This sounds like it's an issue with the config file itself.
Could you try starting from a fresh config file? Alternatively, could you send me your config file then I can have a look at it.

Thank you and best wishes,
Fabian

[xtalfire]

Hi Fabian,

I realized I didn't set up PathToPython to point to the venv location, so SerialEM doesn't know about the environment. It was my bad and now it is fixed. And today finally the scope is free for some testing. Even though SPACE is not running smoothly yet, it is really great!!

I have set up SPACE on both MPC [gui] and SPC [gui, predict], somehow the MPC [gui] gives me some trouble , e.g. in the utils_sem.py there is a sem.exit() instead of sem.Exit(), which gives error.

In the meantime SPACE runs better on SPC. Level 1 can be completed successfully. But there is some trouble in doing level2. The IM is set at 580x, and typically LD view is 8700x. During the first step of level2, it failed in the Realign step between IM and LD view.
If I switch view to 4800x, the Realign step still fails, but from whatever stage position it started collecting MM and finished. The lamella is off centered.

We would really like to use 8700x to set up MM, what would advise us to do?

Thanks a lot and best wishes

Tat

[eisfabian]

Thanks, Tat! Good to hear!

I fixed the sem.exit() in utils_sempy 4 months ago. You might want to update from the Github repo.

Regarding the alignment between mags: Yes, that's the most common issue. I would recommend playing with IM mag and also do the "Mag IS Offsets" calibrations again to make sure the mags are not too far off from each other.

As a temporary workaround, there is an option in the Targets GUI to reacquire MM maps and you can recenter it.

Let me know if that works for you!

Thanks and best wishes,
Fabian

[xtalfire]

Hi Fabain,

Sorry for delayed reply. We had long scope downtime because of series of events and autoloader issues.
Thanks for the advice to update. It works much better after the upgrade, like CLEM, BIGSMALL etc.
I am not sure I did the update properly though, the whl I got from piwheels was released 27.03.2025, so I decided to make a whl from the cloned github and reinstall in on the SPC and it worked.

One thing I had trouble with is that there is a SetupPolygonMontage step after between IM and MMM. It works fine when I set the LDview at 4800x, and serialEM fails when LDview is at 8700x (see error message). I can manually set up a MMM at that stage position with 8700x that covers the whole lamella, but it is puzzling to me what makes the SetupPolygonMontage fails.

Tat