Not a problem but a suggestion - Refinement server option

[Yonny24]

Hi,

How are you doing? Your tool has been invaluable - thanks for all the hard work done btw.

In regards to the AI server options - there is a Refinement server option for when you have more than 1 deepstack server running.

What would be very useful is to have 2 options under Refinement server.
the 2nd one being - Use this server ONLY is the first server does NOT find a relevant object.
I have a custom model for 1 specific object that the main finds it hard to discover at night.
My custom model is very sensitive and both run for every image sent.

If the main server finds the object I don't need to run the custom model - if it does not find the object then I'd like the option of running the custom model only.

Hope that is clear.

Do you think this would be possible in the foreseeable future as an option?

Thanks
Chris

[GmorkDK]

In the lines of this request; Maybe add a failover/backup server? In case the primary is unresponsive for x amount of seconds for some reason.