Hi, I have a question related to the distributed learning scenario.
While I am doing distributed learning and specified only the number of workers.
Based on the description here:
Each worker node will also specify a server process.
However, what is the role of the node launch the distributed learning job?
It is a or a server or schdueler?
Also, if we have more than one server process, does it mean we are also using model parallesim ?
Best Regards,
Johnny