Using more than one GPU at same time

Hey guys maybe someone can help me but the moment i run the docker container it seems to just use the one Gpu even though i have two install and select the gpus=all tag… do i have to run a second docker and select the second gpu???
I have also tried nvidia-docker but does the same …
does use the gpu when running api but doesnt split on two.

@OlafenwaMoses and @gouthamravee do you guys maybe have a idea?? should one run two instances then?? one on one card and one on other?

I think the gpus=all means any, not load load balance the requests. I don’t think deepstack provides load balancing between the GPUs. Perhaps, you can run a load balancer like nginx, choose 2 different ports and run two deepstack instances, one for each gpu, specifying the gpu for each instance and then sending the requests to the load balancer.

@skoenman Please create an issue on the DeepStack repository as a feature request for this.