-
Notifications
You must be signed in to change notification settings - Fork 0
/
train_matrix_compare_for_5folds.py
488 lines (394 loc) · 18.7 KB
/
train_matrix_compare_for_5folds.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
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
import pandas as pd
import numpy as np
from tensorflow.keras.utils import to_categorical
from utils import *
from mhcflurry.allele_encoding import AlleleEncoding
from mhcflurry.amino_acid import BLOSUM62_MATRIX
from numpy.testing import assert_equal
from mhcflurry.encodable_sequences import EncodableSequences
from mhcflurry import amino_acid
from model import model1_MHCflurry_CNN, model2_NetMHCpan4_CNN, model3_MSF_CNN, model4_ANN40_CNN, model5_DeepHLA_CNN
import tensorflow as tf
def readSeqInfo(filename):
'''
读取伪序列的序列信息
'''
fr = open(filename)
raw_datas = fr.readlines()
seq_dict = {}
for i in range(len(raw_datas)):
if raw_datas[i].startswith('>'):
label = raw_datas[i].split(' ')[0].replace('>HLA_', '') # 第一行以空格分隔 第0列是HLA名字将>HLA_去掉，len(raw_datas)是文件总行数
seq = raw_datas[i + 1].replace('\n', '') # 第二行伪序列
seq_dict[label] = seq
return seq_dict
def readSeqInfo2(filename):
'''
读取伪序列的序列信息 MHCflurry, NetMHCpan4.0
'''
fr = open(filename)
raw_datas = fr.readlines()
seq_dict = {}
for i in range(len(raw_datas)):
temp = raw_datas[i].split(' ')
seq_dict[temp[0]] = temp[1].strip()
return seq_dict
'''
encoding = AlleleEncoding(
["A*02:01"],
{
"A*02:01": "ACDD",
}
)
'''
def Concat_reshape(data1, data2):
data3 = np.concatenate((data1, data2), axis=1)
data3 = np.expand_dims(data3, axis=3)
# data3 = np.concatenate((data3, data3, data3, data3), axis=3)
return data3
def MHCflurry_matrix(filename, file_data): # (36+15)*21*1
seq_dict = readSeqInfo2(filename)
df = pd.read_csv(file_data)
encoding = AlleleEncoding(
# alleles=df['allele'].to_list(),
alleles=df['HLA'].str.replace(':', '').to_list(),
allele_to_sequence=seq_dict,
borrow_from=None
)
encoder_alle = encoding.fixed_length_vector_encoded_sequences("BLOSUM62")
encoded_peptides = EncodableSequences.create(df['peptide'].to_list())
encoder_peptides = encoded_peptides.variable_length_to_fixed_length_vector_encoding("BLOSUM62")
X = Concat_reshape(encoder_peptides, encoder_alle)
Y = np.array(df['label'].to_list())
return X, Y
def MSF_matrix(i): # 9*6*5
tt = np.load("./data/npy/k_folds/ms_fold_{}.npz".format(i))
return tt['x'], tt['y']
def NetMHCpan4_matrix(filename, file_data):
df = pd.read_csv(file_data)
seq_dict = readSeqInfo2(filename)
index_encoded_matrix = amino_acid.index_encoding(df['9mer'].to_list(), amino_acid.AMINO_ACID_INDEX)
encoder_peptide = amino_acid.fixed_vectors_encoding(index_encoded_matrix,
amino_acid.ENCODING_DATA_FRAMES["BLOSUM62"])
alle = []
temp = df['HLA'].str.replace(':', '').to_list()
for i in range(len(temp)):
alle.append(seq_dict[temp[i]])
index_encoded_matrix = amino_acid.index_encoding(alle, amino_acid.AMINO_ACID_INDEX)
encoder_alle = amino_acid.fixed_vectors_encoding(index_encoded_matrix, amino_acid.ENCODING_DATA_FRAMES["BLOSUM62"])
# (90803, 34, 21, 1)
X = Concat_reshape(encoder_peptide, encoder_alle)
Y = np.array(df['label'].to_list())
return X, Y
def DeepHLA_transform(HLA, peptide):
aa_idx = {'A': 1, 'C': 2, 'D': 3, 'E': 4, 'F': 5, 'G': 6, 'H': 7, 'I': 8, 'K': 9, 'L': 10, 'M': 11, 'N': 12,
'P': 13, 'Q': 14, 'R': 15, 'S': 16, 'T': 17, 'V': 18, 'W': 19, 'Y': 20, 'X': 21}
data = HLA + peptide
seq = data + 'X' * (49 - len(data))
seq = [aa_idx[x] - 1 for x in seq] # one-hot need -1
return seq
def DeepHLA_matrix(filename, file_data): # (90901, 49, 21, 1)
df = pd.read_csv(file_data)
HLA_seq = pd.read_csv(filename, sep='\t')
seqs = {}
for i in range(len(HLA_seq)):
seqs[HLA_seq.HLA[i]] = HLA_seq.sequence[i]
df['cost_cents'] = df.apply(
lambda row: DeepHLA_transform(
HLA=seqs[row['HLA'].replace(':', '')],
peptide=row['peptide']),
axis=1)
X = np.vstack(df.cost_cents)
X_one_hot = to_categorical(X, num_classes=21)
X_one_hot = np.expand_dims(X_one_hot, axis=3)
Y = np.array(df['label'].to_list())
# X = np.concatenate((X_one_hot, X_one_hot, X_one_hot), axis=3) #无需扩展数据量
return X_one_hot, Y
def ANN40_matrix(filename, file_data): # (90901, 43, 21, 4)
X_list = []
aa_idx = {'A': 1, 'C': 2, 'D': 3, 'E': 4, 'F': 5, 'G': 6, 'H': 7, 'I': 8, 'K': 9, 'L': 10, 'M': 11, 'N': 12,
'P': 13, 'Q': 14, 'R': 15, 'S': 16, 'T': 17, 'V': 18, 'W': 19, 'Y': 20, 'X': 21}
df = pd.read_csv(file_data)
seq_dict = readSeqInfo2(filename)
# 将字符串通过aaidx转化为数字
for peptide in df['9mer'].to_list():
numbers = [aa_idx[char] for char in peptide]
# 将数字转化为one-hot编码
one_hot = np.zeros((len(numbers), len(aa_idx)))
for i, num in enumerate(numbers):
one_hot[i, num - 1] = 0.9 # 将1变为0.9
one_hot[i, np.arange(len(aa_idx)) != num - 1] = 0.05 # 将0变为0.05
X_list.append(one_hot)
alle = []
temp = df['HLA'].str.replace(':', '').to_list()
for i in range(len(temp)):
alle.append(seq_dict[temp[i]])
index_encoded_matrix = amino_acid.index_encoding(alle, amino_acid.AMINO_ACID_INDEX)
encoder_alle = amino_acid.fixed_vectors_encoding(index_encoded_matrix, amino_acid.ENCODING_DATA_FRAMES["BLOSUM62"])
# (90803, 34, 21)
X = Concat_reshape(X_list, encoder_alle)
Y = np.array(df['label'].to_list())
return X, Y
def train(index, times, load_weight=False):
FILE_PATH = "./data/csv/k_folds/ms_fold_{}.csv"
file_path_alle_NetMHCpan = "./data/pseudo/NetMHCpan_pseudo"
file_path_alle_flurry = "./data/pseudo/MHCflurry_pseudosequences"
file_path_alle_DeepHLA = "./data/pseudo/DeepHLA_MHC_pseudo.dat"
# create models
MHCflurry_CNN = model1_MHCflurry_CNN(2)
NetMHCpan4_CNN = model2_NetMHCpan4_CNN(2)
MSF_CNN = model3_MSF_CNN(2)
ANN40_CNN = model4_ANN40_CNN(2)
DeepHLA_CNN = model5_DeepHLA_CNN(2)
if load_weight: # 如果进行多次训练，则load_weight = True,表明获取上次的参数进行再次训练
epoch = 7
MHCflurry_CNN.load_weights(
"./model/input_matrix_compare_5folds/MHCflurry_CNN_{}_{}.h5".format(epoch, times))
NetMHCpan4_CNN.load_weights(
"./model/input_matrix_compare_5folds/NetMHCpan4_CNN_{}_{}.h5".format(epoch, times))
MSF_CNN.load_weights(
"./model/input_matrix_compare_5folds/MSF_CNN_{}_{}.h5".format(epoch, times))
ANN40_CNN.load_weights(
"./model/input_matrix_compare_5folds/ANN40_CNN_{}_{}.h5".format(epoch, times))
DeepHLA_CNN.load_weights(
"./model/input_matrix_compare_5folds/DeepHLA_CNN_{}_{}.h5".format(epoch, times))
epochs = 13
for epoch in range(8, epochs): # 训练轮次
for i in index: # 从index获取训练数据
print("EPOCH:{}, MS_index:{}-----------".format(epoch, i))
file_data = FILE_PATH.format(i) # loads ms_fold_i.csv
X, Y = MSF_matrix(i)
Y = tf.one_hot(Y, 2)
MSF_CNN.fit(X, Y)
X, Y = MHCflurry_matrix(file_path_alle_flurry, file_data)
Y = tf.one_hot(Y, 2)
MHCflurry_CNN.fit(X, Y)
X, Y = NetMHCpan4_matrix(file_path_alle_NetMHCpan, file_data)
Y = tf.one_hot(Y, 2)
NetMHCpan4_CNN.fit(X, Y)
X, Y = ANN40_matrix(file_path_alle_NetMHCpan, file_data)
Y = tf.one_hot(Y, 2)
ANN40_CNN.fit(X, Y)
X, Y = DeepHLA_matrix(file_path_alle_DeepHLA, file_data)
Y = tf.one_hot(Y, 2)
DeepHLA_CNN.fit(X, Y)
MHCflurry_CNN.save_weights(
"./model/input_matrix_compare_5folds/MHCflurry_CNN_{}_{}.h5".format(epoch, times))
NetMHCpan4_CNN.save_weights(
"./model/input_matrix_compare_5folds/NetMHCpan4_CNN_{}_{}.h5".format(epoch, times))
MSF_CNN.save_weights(
"./model/input_matrix_compare_5folds/MSF_CNN_{}_{}.h5".format(epoch, times))
ANN40_CNN.save_weights(
"./model/input_matrix_compare_5folds/ANN40_CNN_{}_{}.h5".format(epoch, times))
DeepHLA_CNN.save_weights(
"./model/input_matrix_compare_5folds/DeepHLA_CNN_{}_{}.h5".format(epoch, times))
def valid2(epoch, times, index):
f = open("./logs/log_input_matrix_compare_5folds_{}.csv".format(index[0]), "w")
f.write("times,precision, recall, fscore, mcc, val_acc,\n")
f.close()
FILE_PATH = "./data/csv/k_folds/ms_fold_{}.csv"
f = open("./logs/log_input_matrix_compare_5folds_{}.csv".format(index[0]), "a")
file_path_alle_DeepHLA = "./data/pseudo/DeepHLA_MHC_pseudo.dat"
file_path_alle_NetMHCpan = "./data/pseudo/NetMHCpan_pseudo"
file_path_alle_flurry = "./data/pseudo/MHCflurry_pseudosequences"
# create models
MHCflurry_CNN = model1_MHCflurry_CNN(2)
NetMHCpan4_CNN = model2_NetMHCpan4_CNN(2)
MSF_CNN = model3_MSF_CNN(2)
ANN40_CNN = model4_ANN40_CNN(2)
DeepHLA_CNN = model5_DeepHLA_CNN(2)
MHCflurry_CNN.load_weights(
"./model/input_matrix_compare_5folds/MHCflurry_CNN_{}_{}.h5".format(epoch, times))
NetMHCpan4_CNN.load_weights(
"./model/input_matrix_compare_5folds/NetMHCpan4_CNN_{}_{}.h5".format(epoch, times))
MSF_CNN.load_weights(
"./model/input_matrix_compare_5folds/MSF_CNN_{}_{}.h5".format(epoch, times))
ANN40_CNN.load_weights(
"./model/input_matrix_compare_5folds/ANN40_CNN_{}_{}.h5".format(epoch, times))
DeepHLA_CNN.load_weights(
"./model/input_matrix_compare_5folds/DeepHLA_CNN_{}_{}.h5".format(epoch, times))
for i in index:
print("---------{}--------".format(i))
file_data = FILE_PATH.format(i)
X, Y = MSF_matrix(i)
temp_Y = MSF_CNN.predict(X)
temp_Y = tf.argmax(temp_Y, axis=1)
precision, recall, fscore, mcc, val_acc = evaluate(np.array(temp_Y), Y)
f.write("MSF_CNN_{}, {}, {}, {}, {}, {}\n".format(times, precision, recall, fscore, mcc, val_acc))
X, Y = MHCflurry_matrix(file_path_alle_flurry, file_data)
temp_Y = MHCflurry_CNN.predict(X)
temp_Y = tf.argmax(temp_Y, axis=1)
precision, recall, fscore, mcc, val_acc = evaluate(np.array(temp_Y), Y)
f.write("MHCflurry_CNN_{}, {}, {}, {}, {}, {}\n".format(times, precision, recall, fscore, mcc, val_acc))
X, Y = ANN40_matrix(file_path_alle_NetMHCpan, file_data)
temp_Y = ANN40_CNN.predict(X)
temp_Y = tf.argmax(temp_Y, axis=1)
precision, recall, fscore, mcc, val_acc = evaluate(np.array(temp_Y), Y)
f.write("ANN40_CNN_{}, {}, {}, {}, {}, {}\n".format(times, precision, recall, fscore, mcc, val_acc))
X, Y = DeepHLA_matrix(file_path_alle_DeepHLA, file_data)
temp_Y = DeepHLA_CNN.predict(X)
temp_Y = tf.argmax(temp_Y, axis=1)
precision, recall, fscore, mcc, val_acc = evaluate(np.array(temp_Y), Y)
f.write("DeepHLA_CNN_{}, {}, {}, {}, {}, {}\n".format(times, precision, recall, fscore, mcc, val_acc))
X, Y = NetMHCpan4_matrix(file_path_alle_NetMHCpan, file_data)
temp_Y = NetMHCpan4_CNN.predict(X)
temp_Y = tf.argmax(temp_Y, axis=1)
precision, recall, fscore, mcc, val_acc = evaluate(np.array(temp_Y), Y)
f.write("NetMHCpan4_CNN_{}, {}, {}, {}, {}, {}\n".format(times, precision, recall, fscore, mcc, val_acc))
f.write("\n")
f.close()
def valid(epoch, times, index): # 将每个folds中的4个小folds进行合并统一测试
f = open("./logs/log_input_matrix_compare_5folds_{}.csv".format(index[0]//4), "w")
f.write("times,precision, recall, fscore, mcc, val_acc,\n")
f.close()
FILE_PATH = "./data/csv/k_folds/ms_fold_{}.csv"
f = open("./logs/log_input_matrix_compare_5folds_{}.csv".format(index[0]//4), "a")
file_path_alle_DeepHLA = "./data/pseudo/DeepHLA_MHC_pseudo.dat"
file_path_alle_NetMHCpan = "./data/pseudo/NetMHCpan_pseudo"
file_path_alle_flurry = "./data/pseudo/MHCflurry_pseudosequences"
# create models
MHCflurry_CNN = model1_MHCflurry_CNN(2)
NetMHCpan4_CNN = model2_NetMHCpan4_CNN(2)
MSF_CNN = model3_MSF_CNN(2)
ANN40_CNN = model4_ANN40_CNN(2)
DeepHLA_CNN = model5_DeepHLA_CNN(2)
MHCflurry_CNN.load_weights(
"./model/input_matrix_compare_5folds/MHCflurry_CNN_{}_{}.h5".format(epoch, times))
NetMHCpan4_CNN.load_weights(
"./model/input_matrix_compare_5folds/NetMHCpan4_CNN_{}_{}.h5".format(epoch, times))
MSF_CNN.load_weights(
"./model/input_matrix_compare_5folds/MSF_CNN_{}_{}.h5".format(epoch, times))
ANN40_CNN.load_weights(
"./model/input_matrix_compare_5folds/ANN40_CNN_{}_{}.h5".format(epoch, times))
DeepHLA_CNN.load_weights(
"./model/input_matrix_compare_5folds/DeepHLA_CNN_{}_{}.h5".format(epoch, times))
all_X_MSF = []
all_Y_MSF = []
all_X_MHCflurry = []
all_Y_MHCflurry = []
all_X_ANN = []
all_Y_ANN = []
all_X_NetMHC = []
all_Y_NetMHC = []
all_X_DeepHLA = []
all_Y_DeepHLA = []
for i in index:
file_data = FILE_PATH.format(i)
X_MSF, Y_MSF = MSF_matrix(i)
all_X_MSF.append(X_MSF)
all_Y_MSF.append(Y_MSF)
X_MHCflurry, Y_MHCflurry = MHCflurry_matrix(file_path_alle_flurry, file_data)
all_X_MHCflurry.append(X_MHCflurry)
all_Y_MHCflurry.append(Y_MHCflurry)
X_ANN, Y_ANN = ANN40_matrix(file_path_alle_NetMHCpan, file_data)
all_X_ANN.append(X_ANN)
all_Y_ANN.append(Y_ANN)
X_DeepHLA, Y_DeepHLA = DeepHLA_matrix(file_path_alle_DeepHLA, file_data)
all_X_DeepHLA.append(X_DeepHLA)
all_Y_DeepHLA.append(Y_DeepHLA)
X_NetMHC, Y_NetMHC = NetMHCpan4_matrix(file_path_alle_NetMHCpan, file_data)
all_X_NetMHC.append(X_NetMHC)
all_Y_NetMHC.append(Y_NetMHC)
X = np.concatenate(all_X_MSF, axis=0)
Y = np.concatenate(all_Y_MSF, axis=0)
temp_Y = MSF_CNN.predict(X)
temp_Y = tf.argmax(temp_Y, axis=1)
precision, recall, fscore, mcc, val_acc = evaluate(np.array(temp_Y), Y)
f.write("MSF_CNN_{}, {}, {}, {}, {}, {}\n".format(times, precision, recall, fscore, mcc, val_acc))
X = np.concatenate(all_X_MHCflurry, axis=0)
Y = np.concatenate(all_Y_MHCflurry, axis=0)
temp_Y = MHCflurry_CNN.predict(X)
temp_Y = tf.argmax(temp_Y, axis=1)
precision, recall, fscore, mcc, val_acc = evaluate(np.array(temp_Y), Y)
f.write("MHCflurry_CNN_{}, {}, {}, {}, {}, {}\n".format(times, precision, recall, fscore, mcc, val_acc))
X = np.concatenate(all_X_ANN, axis=0)
Y = np.concatenate(all_Y_ANN, axis=0)
temp_Y = ANN40_CNN.predict(X)
temp_Y = tf.argmax(temp_Y, axis=1)
precision, recall, fscore, mcc, val_acc = evaluate(np.array(temp_Y), Y)
f.write("ANN40_CNN_{}, {}, {}, {}, {}, {}\n".format(times, precision, recall, fscore, mcc, val_acc))
X = np.concatenate(all_X_DeepHLA, axis=0)
Y = np.concatenate(all_Y_DeepHLA, axis=0)
temp_Y = DeepHLA_CNN.predict(X)
temp_Y = tf.argmax(temp_Y, axis=1)
precision, recall, fscore, mcc, val_acc = evaluate(np.array(temp_Y), Y)
f.write("DeepHLA_CNN_{}, {}, {}, {}, {}, {}\n".format(times, precision, recall, fscore, mcc, val_acc))
X = np.concatenate(all_X_NetMHC, axis=0)
Y = np.concatenate(all_Y_NetMHC, axis=0)
temp_Y = NetMHCpan4_CNN.predict(X)
temp_Y = tf.argmax(temp_Y, axis=1)
precision, recall, fscore, mcc, val_acc = evaluate(np.array(temp_Y), Y)
f.write("NetMHCpan4_CNN_{}, {}, {}, {}, {}, {}\n".format(times, precision, recall, fscore, mcc, val_acc))
f.write("\n")
f.close()
def main_5folds():
mapping = {}
for i in range(20): # 将20个csv文件进行合并，合并为5个folds，[0,4] [5,8] ...
if (i // 4 in mapping):
mapping[i // 4].append(i)
else:
mapping[i // 4] = [i]
print(mapping)
for i in range(5): # 5folds, times = 0, valid:[0,4], train:others
train_index = []
valid_index = []
for j in range(5):
if i != j:
train_index += mapping[j]
else:
valid_index = mapping[j]
train(train_index, times=i, load_weight=True)
# valid(valid_index, times=i)
def train_cnn_ms():
model_conext = model3_DE_CNN(2)
for j in range(3):
for i in range(20):
temp = np.load("./data/npy/ms_train/train_{}.npz".format(i))
Y = tf.one_hot(temp['y'], 2)
model_conext.fit(temp['x'], Y)
del temp
model_conext.save_weights("./save_model_weight/ms_cnn/cnn_{}.h5".format(j))
def train_cnn_af():
de_cnn_af = DE_CNN_AF(2)
for j in range(10):
for i in range(5):
Y = []
temp = np.load("./data/npy/af_train/af_{}.npz".format(i))
for k in temp['y']:
Y.append([1 - k, k])
de_cnn_af.fit(temp['x'], np.array(Y), batch_size=32)
del temp
de_cnn_af.save_weights("./save_model_weight/DE_CNN_AF/cnn_{}.h5".format(j))
def valid_af_de_CNN():
model_cnn = DE_CNN_AF(2)
X = np.load("./data/npy/af_valid/af_valid.npz")['x']
df = pd.read_csv("./data/csv/af_valid_data.csv")
Y_ture = df['log50k'].to_list()
for i in range(10):
ans = []
model_cnn.load_weights("./save_model_weight/DE_CNN_AF/cnn_{}.h5".format(i))
Y_pre = model_cnn.predict(X)
f = open("Y_pre.txt", 'w')
for j in range(len(Y_ture)):
ans.append(math.pow(Y_pre[j][1] - Y_ture[j], 2))
f.write("{}\n".format(Y_pre[j][1]))
f.close()
print("{} mean:{} std:{}".format(i, np.mean(ans), np.std(ans)))
def main_valid():
mapping = {}
for i in range(20): # 将20个csv文件进行合并，合并为5个folds，[0,4] [5,8] ...
if (i // 4 in mapping):
mapping[i // 4].append(i)
else:
mapping[i // 4] = [i]
for i in range(5): # 5folds, times = 0, valid:[0,4], train:others
train_index = []
valid_index = []
for j in range(5):
if i != j:
train_index += mapping[j]
else:
valid_index = mapping[j]
valid(epoch=12, times=i, index=valid_index)
main_valid()
# main_5folds()