trains-agent daemon --gpus 0 --queue default & trains-agent daemon --gpus 1 --queue default &

yap, that should do the trick

Hi JitteryCoyote63 ,

Sure you can, you have many https://allegro.ai/docs/use_cases/trains_agent_use_case_examples/ , just pick to one you need 🙂

Very cool! Run two train-agent daemons, one per GPU on the same machine, with default Nvidia/CUDA Docker This is close to my use case, I just would like to run these two daemons not with docker, would that be possible? I should just remove the --docker nvidia/cuda param right?