Currently, the server only uses data from the master process in a distributed run. ideally, all the processes should be aggregated.
Currently, the server only uses data from the master process in a distributed run. ideally, all the processes should be aggregated.