Windows GPU python closing down

[Tinbum1]

I'm finding with this that when you start multiple instances of DeepStack all looks OK until you process the first image. When this happens all instances of Python close down.

I can see an error;

‘Key already registered with the same priority: GroupSpatialSoftmax’

[Tinbum1]

I'm still having real problems with this on 2 computers.

I'm finding that in most cases all instances of Python are closing down but in other cases where they don't then DeepStack will return error 500. After a few more tries it will then work.

If i roll back to version GPU 2020.12 beta it works ok.

[johnolafenwa]

Hello @Tinbum1 , thanks for sharing this. To clarify, this happens only when running multiple instances of DeepStack?

[Tinbum1]

I've just tried and it seems ok with one instance and even 3 but after that it gets temperamental. I've been using 5 or 6 and once they are running it's been great. The GPU is showing when running 6 well;

[Tinbum1]

I'm wondering that if when multiple instances of DeepStack are started there are that many initial instances of python.exe running it runs out of memory before they close down and just leave the 2 instances needed running as shown in the post above. ?

[FortWorthMadScientist]

I am using a TESLA K40m and have seen the exact same thing but it only takes 1 instance for Python to drop.

[Tinbum1]

My cards are a GTX 750 Ti and a GTX 1050 Ti

[FortWorthMadScientist]

Any new word on this issue?

[Tinbum1]

Any update on this issue? Thanks.

[Tinbum1]

I'm still having problems running multiple instances. Any update on this.