PFLlib/system/flcore/servers/serverfomo.py

# PFLlib: Personalized Federated Learning Algorithm Library
# Copyright (C) 2021  Jianqing Zhang

# This program is free software; you can redistribute it and/or modify
# it under the terms of the GNU General Public License as published by
# the Free Software Foundation; either version 2 of the License, or
# (at your option) any later version.

# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
# GNU General Public License for more details.

# You should have received a copy of the GNU General Public License along
# with this program; if not, write to the Free Software Foundation, Inc.,
# 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA.

import torch
import time
import copy
import random
import numpy as np
from flcore.clients.clientfomo import clientFomo
from flcore.servers.serverbase import Server
from threading import Thread
from utils.dlg import DLG


class FedFomo(Server):
    def __init__(self, args, times):
        super().__init__(args, times)

        # select slow clients
        self.set_slow_clients()
        self.set_clients(clientFomo)

        self.P = torch.diag(torch.ones(self.num_clients, device=self.device))
        self.uploaded_ids = []
        self.M = min(args.M, self.num_join_clients)
        self.client_models = [copy.deepcopy(self.global_model) for _ in range(self.num_clients)]
            
        print(f"\nJoin ratio / total clients: {self.join_ratio} / {self.num_clients}")
        print("Finished creating server and clients.")
        self.Budget = []


    def train(self):
        for i in range(self.global_rounds+1):
            s_t = time.time()
            self.selected_clients = self.select_clients()
            self.send_models()

            if i%self.eval_gap == 0:
                print(f"\n-------------Round number: {i}-------------")
                print("\nEvaluate personalized models")
                self.evaluate()

            for client in self.selected_clients:
                client.train()

            # threads = [Thread(target=client.train)
            #            for client in self.selected_clients]
            # [t.start() for t in threads]
            # [t.join() for t in threads]

            if self.dlg_eval and i%self.dlg_gap == 0:
                self.call_dlg(i)
            self.receive_models()

            self.Budget.append(time.time() - s_t)
            print('-'*25, 'time cost', '-'*25, self.Budget[-1])

            if self.auto_break and self.check_done(acc_lss=[self.rs_test_acc], top_cnt=self.top_cnt):
                break

        print("\nBest accuracy.")
        # self.print_(max(self.rs_test_acc), max(
        #     self.rs_train_acc), min(self.rs_train_loss))
        print(max(self.rs_test_acc))
        print("\nAverage time cost per round.")
        print(sum(self.Budget[1:])/len(self.Budget[1:]))

        self.save_results()


    def send_models(self):
        assert (len(self.selected_clients) > 0)
        for client in self.clients:
            start_time = time.time()

            if client.send_slow:
                time.sleep(0.1 * np.abs(np.random.rand()))

            M_ = min(self.M, len(self.uploaded_ids)) # if clients dropped
            indices = torch.topk(self.P[client.id], M_).indices.tolist()

            send_ids = []
            send_models = []
            for i in indices:
                send_ids.append(i)
                send_models.append(self.client_models[i])

            client.receive_models(send_ids, send_models)

            client.send_time_cost['num_rounds'] += 1
            client.send_time_cost['total_cost'] += 2 * (time.time() - start_time)

    def receive_models(self):
        assert (len(self.selected_clients) > 0)

        active_clients = random.sample(
            self.selected_clients, int((1-self.client_drop_rate) * self.current_num_join_clients))

        self.uploaded_ids = []
        self.uploaded_weights = []
        tot_samples = 0
        for client in active_clients:
            try:
                client_time_cost = client.train_time_cost['total_cost'] / client.train_time_cost['num_rounds'] + \
                        client.send_time_cost['total_cost'] / client.send_time_cost['num_rounds']
            except ZeroDivisionError:
                client_time_cost = 0
            if client_time_cost <= self.time_threthold:
                tot_samples += client.train_samples
                self.uploaded_ids.append(client.id)
                self.uploaded_weights.append(client.train_samples)
                self.client_models[client.id] = copy.deepcopy(client.model)
                self.P[client.id] += client.weight_vector
        for i, w in enumerate(self.uploaded_weights):
            self.uploaded_weights[i] = w / tot_samples

    def call_dlg(self, R):
        # items = []
        cnt = 0
        psnr_val = 0
        for cid, client_model_server in zip(range(self.num_clients), self.client_models):
            client_model = self.clients[cid].model
            client_model.eval()
            origin_grad = []
            for gp, pp in zip(client_model_server.parameters(), client_model.parameters()):
                origin_grad.append(gp.data - pp.data)

            target_inputs = []
            trainloader, _ = self.clients[cid].load_train_data()
            with torch.no_grad():
                for i, (x, y) in enumerate(trainloader):
                    if i >= self.batch_num_per_client:
                        break

                    if type(x) == type([]):
                        x[0] = x[0].to(self.device)
                    else:
                        x = x.to(self.device)
                    y = y.to(self.device)
                    output = client_model(x)
                    target_inputs.append((x, output))

            d = DLG(client_model, origin_grad, target_inputs)
            if d is not None:
                psnr_val += d
                cnt += 1
            
            # items.append((client_model, origin_grad, target_inputs))
                
        if cnt > 0:
            print('PSNR value is {:.2f} dB'.format(psnr_val / cnt))
        else:
            print('PSNR error')

        # self.save_item(items, f'DLG_{R}')