FedPEFD/train.py at main · Qyizos/FedPEFD · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
# -*- coding: utf-8 -*-
import sys, os
import logging
import glob
import time
import shutil
import re
from torch.utils import data
from torch.utils.data import dataset

base_path = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
sys.path.append(base_path)
import numpy as np
import random
import torch
import torch.nn as nn
import torch.optim as optim
import torch.backends.cudnn as cudnn

from federated.configs import set_configs

from federated.fl_api import *
# from federated.fl_api import FedAvgAPI, FedNucAPI, FedProxAPI
from federated.model_trainer_segmentation import ModelTrainerSegmentation

from utils.dataset import DataFolder
from utils.my_transforms import get_transforms
from utils.readVal import getVal
from torch.utils.data import DataLoader
from utils.loss import DiceLoss, entropy_loss
from nets.model import *
import test

def main(curTime=None, curmodel=ResUNet34(pretrained = True), notes=None):
    # 参数初始化
    args = set_configs()
    args.generalize = False
    args.source = ['Sconsep', 'Scrag', 'Sdpath', 'Sglas', 'Spannuke']
    args.notes = notes
    args.transform = dict()
    args.transform['train'] = {
        'random_resize': [0.8, 1.25],
        'horizontal_flip': True,
        'vertical_flip': True,
        'random_affine': 0.3,
        'random_rotation': 90,
        'random_crop': args.input_size,
        'label_encoding': 2,    # 将倒数几个输入转为单通道，对应模型有几组标签
        'to_tensor': 1          #
    }
    args.transform['test'] = {
        'to_tensor': 1
    }

    valDSet = args.source[0].split('_')[0]    # 单数据集测试

    deterministic(args.seed)
    if curTime:
        args.sonName = curTime
    else:
        args.sonName = time.strftime('%y%m%d-%H%M', time.localtime())  # 借助时间给log 命名
    set_paths(args)
    os.environ['CUDA_VISIBLE_DEVICES'] = args.gpu

    model_trainer = custom_model_trainer(args, curmodel)

    # ----- create model ----- #
    model = curmodel
    model = model.cuda()

    num_params = 0
    for param in model.parameters():
        num_params += param.numel()
    args.num_params = num_params

    cudnn.benchmark = True

    # log 相关
    log_path = os.path.join(args.save_path, 'log.txt')
    logging.basicConfig(filename=log_path, level=logging.INFO,
                        format='[%(asctime)s.%(msecs)03d] %(message)s', datefmt='%H:%M:%S')
    logging.getLogger().addHandler(logging.StreamHandler(sys.stdout))
    logging.info(str(args))
    # ----- define optimizer ----- #
    optimizer = torch.optim.Adam(model.parameters(), 0.0001, betas=(0.9, 0.99),
                                 weight_decay=1e-4)
    criterion = torch.nn.NLLLoss(ignore_index=2).cuda()
    # a=args.transform['train']
    # ----- load data ----- #
    data_transforms = {'train': get_transforms(args.transform['train']),
                       'test': get_transforms(args.transform['test'])}

    dir_list = ['images', 'labels_cluster', 'labels_voronoi']
    post_fix = ['_label_vor.png', '_label_cluster.png']     # 后四个输入的后缀名
    num_channels = [3, 3, 3]      # 分别对应dir_list文件夹中的图像通道数
    datasets = []
    for client in args.source:
        train_set = DataFolder(dir_list, post_fix, num_channels, client, data_transforms['train'])  # 读入数据集
        train_loader = DataLoader(train_set, batch_size=args.batch, shuffle=True, num_workers=1)
        datasets.append(train_loader)

    valSets = []
    for curClient in args.source:
        valSet = getVal(client = curClient, type='val')
        valSets.append(valSet)

    federated_manager = custom_federated_api(args, model_trainer)
    federated_manager.train(datasets, model, optimizer, criterion, args, valSets)

    rename_and_cleanup_round_dirs(args.save_path)

    # ----------------------- test -----------------------
    pthPath = args.save_path + '/' + 'round_best'
    pthNums = count_tar_files(pthPath)
    ndarray_list = []

    for pid in range(pthNums):
        pthName = 'client_' + str(pid+1)
        curResult = test.main(curTime, curmodel, path=pthName, testSet=args.source[pid])
        ndarray_list.append(curResult)

    avg_list = [float() for _ in range(len(ndarray_list[0]))]

    for curList in ndarray_list:
        for idx in range(len(curList)):
            avg_list[idx] = avg_list[idx] + curList[idx]

    for idx in range(len(avg_list)):
        avg_list[idx] = avg_list[idx] / len(ndarray_list)

    logging.info('Average Acc: {r[0]:.4f}\nF1: {r[1]:.4f}\nIoU: {r[2]:.4f}\nDice: {r[3]:.4f}\nAJI: {r[4]:.4f}\n'.format(r=avg_list))


if __name__ == "__main__":
    curTime = time.strftime('%y%m%d-%H%M', time.localtime())
    notes = 'model = ResUNet34_DA_eA0v5(); '
    name2 = notes.split('model = ')[-1].split('()')[0]
    curTime = curTime + '_' + name2
    main(curTime, ResUNet34_DA_eA0v5(), notes)