supervisely deploy-agent
so you don't have to do anything. But, if you don't have a GPU device on your machine with Supervisely, you may want to deploy an additional agent on AWS or some computer with a videocard.SERVER_ADDRESS
variable in your .env
file. Agent should be able to make outbound connections to that IP or Host. As a quick solution you can find your docker0
ip address that act as locahost
for Docker: run ip addr list docker0
and use it in SERVER_ADDRESS
. It should look like 172.17.0.1
or something. Run supervisely deploy-agent
to apply your changes. Main Node at Cluster page should change status to "Running"