Help with deepstack gpu for windows

I just changed from DeepStack CPU to DeepStack GPU and I think it sounds more complicated than it is - there are several different versions and comments floating around...

I recently purchased a Quadro P400 v2 and I have it installed in my system with the motherboard bios set to use onboard graphics. So in Windows, the Task Manager will show two separate GPU's. Also, I used the latest version of CUDA from NVIDIA (11.3.1)

I wrote up the procedure I used - (attached text file). It 'seems' to be working fine, but if any one sees anything I missed or has some tips....

The main webpage with instructions and links can be found here: Using DeepStack with Windows 10 (CPU and GPU)


View attachment 129304


Thank you for this . I picked up a 1030 for a decent price (low profile as I needed it as well), went from ~3500ms to ~125ms. Amazing! Also love the custom models provided by @MikeLud1
 
Had the same problem as the OP and after banging my head against the wall for a bit I came across this: GPU 2022.01.1 - no objects detected.

Apparently the latest Deepstack version has some issues with certain video cards (GTX 1650 in my case) where it appears to be working, but never detects any objects. Going to the 2021.09.1 version fixed the issue for me and now Deepstack is detecting objects.
 
Has anyone tried with a GT 730? According to CUDA-Z it's 3.5 compute compatibility but I cannot get deepstack to stop timing out. I've tried multiple CUDA and deepstack versions with no success. Thinking I may just have to bite the bullet and get the 1030 low profile. I am running windows on Proxmox VM if that makes any difference.
 
  • Like
Reactions: Flintstone61
I just changed from DeepStack CPU to DeepStack GPU and I think it sounds more complicated than it is - there are several different versions and comments floating around...

I recently purchased a Quadro P400 v2 and I have it installed in my system with the motherboard bios set to use onboard graphics. So in Windows, the Task Manager will show two separate GPU's. Also, I used the latest version of CUDA from NVIDIA (11.3.1)

I wrote up the procedure I used - (attached text file). It 'seems' to be working fine, but if any one sees anything I missed or has some tips....

The main webpage with instructions and links can be found here: Using DeepStack with Windows 10 (CPU and GPU)


View attachment 129304
I have followed this to the letter but when i hit start now in BI it goes back to "start now" within a second. It would appear something is wrong. Also when i try to analyze video i get no boxes around anything when using testing & tuning. Running latest BI. I'm using 127.0.0.1 with port 82, also tried 5000. Not sure why your instructions show to use port 80 in powershell since this is used by BI. Before following your instructions i've tried all the others i've found online.... just about to give up. Open in browser does work btw.
 
Open Resource Monitor or Task Manager and see if senseAI is actually running and in memory.
 
Now that BI has deprecated DeepStack, the bigger question is how much longer will Deepstack work with BI? You may be chasing a solution to something that disappears soon.
 
  • Wow
Reactions: Flintstone61
what does that mean....Bi has deprecated Deepstack.....like shitcanned it?
 
what does that mean....Bi has deprecated Deepstack.....like shitcanned it?

That is what the latest help file says. I think it means it is there but will have no more improvements associated with it.

The question though is will the ability to use DeepStack be eliminated in a future update?
 
  • Like
Reactions: Flintstone61
That is what the latest help file says. I think it means it is there but will have no more improvements associated with it.

The question though is will the ability to use DeepStack be eliminated in a future update?

That doesn’t matter much if my limited knowledge is correct. People were using deepstack with BI before there was a native integration… BI just made it easier to use deepstack by adding AI support in the settings. Or am I completely wrong about that?