-
Notifications
You must be signed in to change notification settings - Fork 4
/
Generation.py
186 lines (163 loc) · 5.65 KB
/
Generation.py
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
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
import torch
import dataloader_seed
import argparse
from models import main_models
import numpy as np
from sklearn import metrics
temple_dir = "./seed_6_210/"
parser=argparse.ArgumentParser()
parser.add_argument('target',type=str, default = '2')
parser.add_argument('number',type=str, default = '1')
parser.add_argument('nargs',type=list, nargs='+', default=['s01', 's29', 's20', 's24', 's13' ])
parser.add_argument('batch',type=int, default=30)
opt = parser.parse_args()
path = temple_dir + opt.target +"/save_model/"
path1 = temple_dir + opt.target + "/"
# path = "./s04/save_model/"
# path1 = "./s04/"
use_cuda = True if torch.cuda.is_available() else False
device = torch.device('cuda:0') if use_cuda else torch.device('cpu')
####读取准确率#####找出结果中最好的三个
import numpy as np
final = []
name = []
or_all = []
auc_all = []
target = opt.target
print("target", target )
number = opt.number
all = ['1','10','13','2','24','26','3','31','4','42','44','6','7','8','9']
all.remove(str(number))
# print(all)
# def get_name(num):
# if num <10:
# name = "s0" + str(num)
# else:
# name = "s" + str(num)
#
# return name
for i in set(all):
# if i <10:
# dir = './' + target + "/s0" + str(i) + '.npz'
# else:
# dir = './' + target + "/s" + str(i) + '.npz'
dir = temple_dir + target + '/'+ str(i) + '.npz'
a = np.load(dir)
# print(a.keys())
data = a['accuracy_all2']
or_data = a['accuracy_all']
auc = a['auc_all']
# print(data.shape)
temp = np.max(data)
auc_temp = np.max(auc)
or_all.append(np.mean(or_data))
final.append(temp) ## 每个人对应source迁移最好的结果
auc_all.append(auc_temp) ## 每个人对应source中auc 最好的结果
name.append(i)
sort_list = np.argsort((-1) * np.array(or_all)) ## 返回的是排序后的坐标
print(or_all)
print(sort_list)
k = 14
a = sort_list ## 代表从大到小的作为排序
# print(a)
index = a[0:k]
# print(index)
max_value = np.max(final) ## 是
print(np.max(final)) ## 十四个人中最好的结果
temp = final[a[0]] ## 对应最好的
print("corres", temp) ## 对应最好的结果
aa = []
for i in range(len(index)):
aa.append(name[index[i]])
test_list = []
final_k = 5
for i in range(final_k):
test_list.append(aa[i])
print(test_list)
# save_dir = path1 = "./" + opt.target + '.xlsx'
# import xlsxwriter
# workbook = xlsxwriter.Workbook(save_dir)
# worksheet = workbook.add_worksheet()
# aa = final.__len__()
# for i in range(int(aa)):
# worksheet.write(i, 0, name[i]) # 第i行0列
# worksheet.write(i, 2, final[i]) # 第i行1列
# workbook.close()
def get_model(file):
encode = torch.load(path+"encode_%s.pth"%(file))
classifier = torch.load(path + "classifier_%s.pth" % (file))
return encode,classifier
# def get_model(file):
# encode = torch.load(path+"encode_%s.pth"%(file))
# classifier = torch.load(path + "classifier_%s.pth" % (file))
# model = classifier(encode)
# return model
from sklearn.preprocessing import label_binarize
test_dataloader = dataloader_seed.source_loader(opt.target)
opt.nargs = test_list
def Caculate(weight):
Encode=[]
Classifier=[]
for file in opt.nargs:
Ez, Cz=get_model(file)
# Ez = Ez.to(device)
# Cz = Cz.to(device)
# model_temp.to(device)
Encode.append(Ez)
Classifier.append(Cz)
acc = 0
auc = 0
for data, labels in test_dataloader:
data = data.to(device)
labels = (labels.long()).to(device)
Y = torch.zeros(opt.batch, 3).to(device)
for i in range(len(Encode)):
Encode[i].eval()
Classifier[i].eval()
Y+=Classifier[i](Encode[i](data))*weight[i]
acc = acc + (torch.max(Y, 1)[1] == labels).float().mean().item()
y_one_hot = labels.cpu().detach().numpy()
y_one_hot = label_binarize(y_one_hot, np.arange(3))
a = Y.cpu().detach().numpy()
auc += metrics.roc_auc_score(y_one_hot, a, average='micro')
# auc += metrics.roc_auc_score(labels.cpu(), torch.max(Y, 1)[1].cpu())
accuracy = round(acc / float(len(test_dataloader)), 3)
auc_temp = round(auc / float(len(test_dataloader)), 3)
print("accuracy: %.3f " % (accuracy))
return accuracy, auc_temp
print("len(opt.nargs)", len(opt.nargs))
weight=[1/len(opt.nargs)]*len(opt.nargs)
save_accuracy, save_auc = Caculate(weight)
save_dir = path1 = temple_dir + opt.target + '.xlsx'
import xlsxwriter
workbook = xlsxwriter.Workbook(save_dir)
worksheet = workbook.add_worksheet()
aa = final.__len__()
for i in range(int(aa)):
worksheet.write(i, 0, name[i]) # 第i行0列
worksheet.write(i, 1, final[i]) # 第i行1列
worksheet.write(i, 8, auc_all[i]) # 第i行1列
bb = or_all.__len__()
for j in range(bb):
worksheet.write(j, 5, or_all[j] ) # 第i行0列
# worksheet.write(i, 1, final[i]) # 第i行1列
worksheet.write(0, 2, "max")
worksheet.write(1, 2, max_value)
worksheet.write(0, 3, "mean")
worksheet.write(1, 3, save_accuracy)
worksheet.write(0, 4, "k")
worksheet.write(1, 4, final_k)
worksheet.write(0, 6, test_list[0])
# worksheet.write(1, 6, test_list[1])
# worksheet.write(2, 6, test_list[2])
# worksheet.write(3, 6, test_list[3])
# worksheet.write(4, 6, test_list[4])
worksheet.write(0, 7, "or_data")
worksheet.write(1, 7, temp)
worksheet.write(0, 9, "mean_auc")
worksheet.write(1, 9, save_auc)
workbook.close()
print("over")
# if __name__ =='__main__':
# weight=[1/len(opt.nargs)]*len(opt.nargs)
# Caculate(weight)