from paddle_fl.core.scheduler.agent_master import FLScheduler
worker_num = 4
server_num = 1
#Define number of worker/server and the port for scheduler
scheduler = FLScheduler(worker_num, server_num, port=9091)
scheduler.set_sample_worker_num(4)
scheduler.init_env()
print("init env done.")
scheduler.start_fl_training()
print("submit mpi job done.")
# start scheduler and receive the ip of allocated endpoints
context = zmq.Context()
zmq_socket = context.socket(zmq.REP)
zmq_socket.bind("tcp://{}:{}".format(current_ip, random_port))
print("binding tcp://{}:{}".format(current_ip, random_port))
all_ips_ready = False
ip_list = []
scheduler = FLScheduler(int(default_dict["worker_nodes"]),
int(default_dict["server_nodes"]),
port=random_port, socket=zmq_socket)
scheduler.set_sample_worker_num(int(default_dict["worker_nodes"]))
print("going to wait all ips ready")
while not all_ips_ready:
message = zmq_socket.recv()
group = message.split("\t")
if group[0] == "ENDPOINT":
ip_list.append(group[1])
zmq_socket.send("ACCEPT\t{}".format(group[1]))
else:
zmq_socket.send("WAIT\t0")
if len(ip_list) == \
from paddle_fl.core.scheduler.agent_master import FLScheduler
worker_num = 5
server_num = 1
scheduler = FLScheduler(worker_num, server_num)
scheduler.set_sample_worker_num(5)
scheduler.init_env()
print("init env done.")
scheduler.start_fl_training()