質問編集履歴
1
ソースコードを挿入し直しました
test
CHANGED
|
File without changes
|
test
CHANGED
|
@@ -22,206 +22,210 @@
|
|
|
22
22
|
|
|
23
23
|
エラーメッセージ
|
|
24
24
|
|
|
25
|
+
Traceback (most recent call last):
|
|
26
|
+
|
|
27
|
+
File "lstm_tokutyou.py", line 55, in <module>
|
|
28
|
+
|
|
29
|
+
model.add(Dense(n_out, kernel_initializer=weight_variable))
|
|
30
|
+
|
|
31
|
+
File "/usr/local/lib/python3.7/site-packages/keras/engine/sequential.py", line 182, in add
|
|
32
|
+
|
|
33
|
+
output_tensor = layer(self.outputs[0])
|
|
34
|
+
|
|
35
|
+
File "/usr/local/lib/python3.7/site-packages/keras/engine/base_layer.py", line 463, in __call__
|
|
36
|
+
|
|
37
|
+
self.build(unpack_singleton(input_shapes))
|
|
38
|
+
|
|
39
|
+
File "/usr/local/lib/python3.7/site-packages/keras/layers/core.py", line 895, in build
|
|
40
|
+
|
|
41
|
+
constraint=self.kernel_constraint)
|
|
42
|
+
|
|
43
|
+
File "/usr/local/lib/python3.7/site-packages/keras/engine/base_layer.py", line 279, in add_weight
|
|
44
|
+
|
|
45
|
+
weight = K.variable(initializer(shape, dtype=dtype),
|
|
46
|
+
|
|
47
|
+
TypeError: weight_variable() got an unexpected keyword argument 'dtype'
|
|
48
|
+
|
|
25
49
|
```
|
|
26
50
|
|
|
27
|
-
|
|
51
|
+
|
|
28
|
-
|
|
29
|
-
File "lstm_tokutyou.py", line 55, in <module>
|
|
30
|
-
|
|
31
|
-
model.add(Dense(n_out, kernel_initializer=weight_variable))
|
|
32
|
-
|
|
33
|
-
File "/usr/local/lib/python3.7/site-packages/keras/engine/sequential.py", line 182, in add
|
|
34
|
-
|
|
35
|
-
output_tensor = layer(self.outputs[0])
|
|
36
|
-
|
|
37
|
-
File "/usr/local/lib/python3.7/site-packages/keras/engine/base_layer.py", line 463, in __call__
|
|
38
|
-
|
|
39
|
-
self.build(unpack_singleton(input_shapes))
|
|
40
|
-
|
|
41
|
-
File "/usr/local/lib/python3.7/site-packages/keras/layers/core.py", line 895, in build
|
|
42
|
-
|
|
43
|
-
constraint=self.kernel_constraint)
|
|
44
|
-
|
|
45
|
-
File "/usr/local/lib/python3.7/site-packages/keras/engine/base_layer.py", line 279, in add_weight
|
|
46
|
-
|
|
47
|
-
weight = K.variable(initializer(shape, dtype=dtype),
|
|
48
|
-
|
|
49
|
-
TypeError: weight_variable() got an unexpected keyword argument 'dtype'
|
|
50
52
|
|
|
51
53
|
### 該当のソースコード
|
|
52
54
|
|
|
53
55
|
|
|
54
56
|
|
|
55
|
-
```
|
|
57
|
+
```python
|
|
56
58
|
|
|
57
59
|
ソースコード
|
|
58
60
|
|
|
61
|
+
import csv
|
|
62
|
+
|
|
63
|
+
import re
|
|
64
|
+
|
|
65
|
+
import datetime
|
|
66
|
+
|
|
67
|
+
import time
|
|
68
|
+
|
|
69
|
+
import numpy as np
|
|
70
|
+
|
|
71
|
+
import matplotlib.pyplot as plt
|
|
72
|
+
|
|
73
|
+
from keras.models import Sequential
|
|
74
|
+
|
|
75
|
+
from keras.layers.core import Dense, Activation
|
|
76
|
+
|
|
77
|
+
from keras.layers.recurrent import LSTM
|
|
78
|
+
|
|
79
|
+
from keras.layers.wrappers import Bidirectional
|
|
80
|
+
|
|
81
|
+
from keras.optimizers import Adam
|
|
82
|
+
|
|
83
|
+
from keras.callbacks import EarlyStopping
|
|
84
|
+
|
|
85
|
+
from sklearn import datasets
|
|
86
|
+
|
|
87
|
+
from sklearn.model_selection import train_test_split
|
|
88
|
+
|
|
89
|
+
from sklearn.utils import shuffle
|
|
90
|
+
|
|
91
|
+
|
|
92
|
+
|
|
93
|
+
data = np.loadtxt("lstm_tokutyou1002.csv",delimiter=",")
|
|
94
|
+
|
|
95
|
+
X = data[:,:1105]
|
|
96
|
+
|
|
97
|
+
y = data[:,-1]
|
|
98
|
+
|
|
99
|
+
Y = np.eye(2)[y.astype(int)]
|
|
100
|
+
|
|
101
|
+
del y
|
|
102
|
+
|
|
103
|
+
X_train, X_test, Y_train, Y_test = \
|
|
104
|
+
|
|
105
|
+
train_test_split(X, Y, test_size=0.3)
|
|
106
|
+
|
|
107
|
+
|
|
108
|
+
|
|
109
|
+
X_train, X_validation, Y_train, Y_validation = \
|
|
110
|
+
|
|
111
|
+
train_test_split(X_train, Y_train, test_size=0.2)
|
|
112
|
+
|
|
113
|
+
|
|
114
|
+
|
|
115
|
+
'''
|
|
116
|
+
|
|
117
|
+
モデル設定
|
|
118
|
+
|
|
119
|
+
'''
|
|
120
|
+
|
|
121
|
+
n_in = 151186
|
|
122
|
+
|
|
123
|
+
n_long = 1105
|
|
124
|
+
|
|
125
|
+
n_hidden = 128
|
|
126
|
+
|
|
127
|
+
n_out = 10
|
|
128
|
+
|
|
129
|
+
|
|
130
|
+
|
|
131
|
+
|
|
132
|
+
|
|
133
|
+
def weight_variable(shape, name=None):
|
|
134
|
+
|
|
135
|
+
return np.random.normal(scale=.01, size=shape)
|
|
136
|
+
|
|
137
|
+
|
|
138
|
+
|
|
139
|
+
|
|
140
|
+
|
|
141
|
+
early_stopping = EarlyStopping(monitor='val_loss', patience=10, verbose=1)
|
|
142
|
+
|
|
143
|
+
|
|
144
|
+
|
|
145
|
+
model = Sequential()
|
|
146
|
+
|
|
147
|
+
model.add(Bidirectional(LSTM(n_hidden),
|
|
148
|
+
|
|
149
|
+
input_shape=(n_long,n_in)))
|
|
150
|
+
|
|
151
|
+
model.add(Dense(n_out, kernel_initializer=weight_variable))
|
|
152
|
+
|
|
153
|
+
model.add(Activation('softmax'))
|
|
154
|
+
|
|
155
|
+
|
|
156
|
+
|
|
157
|
+
model.compile(loss='categorical_crossentropy',
|
|
158
|
+
|
|
159
|
+
optimizer=Adam(lr=0.001, beta_1=0.9, beta_2=0.999),
|
|
160
|
+
|
|
161
|
+
metrics=['accuracy'])
|
|
162
|
+
|
|
163
|
+
|
|
164
|
+
|
|
165
|
+
'''
|
|
166
|
+
|
|
167
|
+
モデル学習
|
|
168
|
+
|
|
169
|
+
'''
|
|
170
|
+
|
|
171
|
+
epochs = 300
|
|
172
|
+
|
|
173
|
+
batch_size = 250
|
|
174
|
+
|
|
175
|
+
|
|
176
|
+
|
|
177
|
+
hist = model.fit(X_train, Y_train,
|
|
178
|
+
|
|
179
|
+
batch_size=batch_size,
|
|
180
|
+
|
|
181
|
+
epochs=epochs,
|
|
182
|
+
|
|
183
|
+
validation_data=(X_validation, Y_validation),
|
|
184
|
+
|
|
185
|
+
callbacks=[early_stopping])
|
|
186
|
+
|
|
187
|
+
|
|
188
|
+
|
|
189
|
+
'''
|
|
190
|
+
|
|
191
|
+
学習の進み具合
|
|
192
|
+
|
|
193
|
+
'''
|
|
194
|
+
|
|
195
|
+
acc = hist.history['val_acc']
|
|
196
|
+
|
|
197
|
+
loss = hist.history['val_loss']
|
|
198
|
+
|
|
199
|
+
|
|
200
|
+
|
|
201
|
+
plt.rc('font', family='serif')
|
|
202
|
+
|
|
203
|
+
fig = plt.figure()
|
|
204
|
+
|
|
205
|
+
plt.plot(range(len(loss)), loss,
|
|
206
|
+
|
|
207
|
+
label='loss', color='black')
|
|
208
|
+
|
|
209
|
+
plt.xlabel('epochs')
|
|
210
|
+
|
|
211
|
+
plt.show()
|
|
212
|
+
|
|
213
|
+
|
|
214
|
+
|
|
215
|
+
'''
|
|
216
|
+
|
|
217
|
+
予測精度の評価
|
|
218
|
+
|
|
219
|
+
'''
|
|
220
|
+
|
|
221
|
+
loss_and_metrics = model.evaluate(X_test, Y_test)
|
|
222
|
+
|
|
223
|
+
print(loss_and_metrics)
|
|
224
|
+
|
|
225
|
+
|
|
226
|
+
|
|
59
227
|
```
|
|
60
228
|
|
|
61
|
-
import csv
|
|
62
|
-
|
|
63
|
-
import re
|
|
64
|
-
|
|
65
|
-
import datetime
|
|
66
|
-
|
|
67
|
-
import time
|
|
68
|
-
|
|
69
|
-
import numpy as np
|
|
70
|
-
|
|
71
|
-
import matplotlib.pyplot as plt
|
|
72
|
-
|
|
73
|
-
from keras.models import Sequential
|
|
74
|
-
|
|
75
|
-
from keras.layers.core import Dense, Activation
|
|
76
|
-
|
|
77
|
-
from keras.layers.recurrent import LSTM
|
|
78
|
-
|
|
79
|
-
from keras.layers.wrappers import Bidirectional
|
|
80
|
-
|
|
81
|
-
from keras.optimizers import Adam
|
|
82
|
-
|
|
83
|
-
from keras.callbacks import EarlyStopping
|
|
84
|
-
|
|
85
|
-
from sklearn import datasets
|
|
86
|
-
|
|
87
|
-
from sklearn.model_selection import train_test_split
|
|
88
|
-
|
|
89
|
-
from sklearn.utils import shuffle
|
|
90
|
-
|
|
91
|
-
|
|
92
|
-
|
|
93
|
-
data = np.loadtxt("lstm_tokutyou1002.csv",delimiter=",")
|
|
94
|
-
|
|
95
|
-
X = data[:,:1105]
|
|
96
|
-
|
|
97
|
-
y = data[:,-1]
|
|
98
|
-
|
|
99
|
-
Y = np.eye(2)[y.astype(int)]
|
|
100
|
-
|
|
101
|
-
del y
|
|
102
|
-
|
|
103
|
-
X_train, X_test, Y_train, Y_test = \
|
|
104
|
-
|
|
105
|
-
train_test_split(X, Y, test_size=0.3)
|
|
106
|
-
|
|
107
|
-
|
|
108
|
-
|
|
109
|
-
X_train, X_validation, Y_train, Y_validation = \
|
|
110
|
-
|
|
111
|
-
train_test_split(X_train, Y_train, test_size=0.2)
|
|
112
|
-
|
|
113
|
-
|
|
114
|
-
|
|
115
|
-
'''
|
|
116
|
-
|
|
117
|
-
モデル設定
|
|
118
|
-
|
|
119
|
-
'''
|
|
120
|
-
|
|
121
|
-
n_in = 151186
|
|
122
|
-
|
|
123
|
-
n_long = 1105
|
|
124
|
-
|
|
125
|
-
n_hidden = 128
|
|
126
|
-
|
|
127
|
-
n_out = 10
|
|
128
|
-
|
|
129
|
-
|
|
130
|
-
|
|
131
|
-
|
|
132
|
-
|
|
133
|
-
def weight_variable(shape, name=None):
|
|
134
|
-
|
|
135
|
-
return np.random.normal(scale=.01, size=shape)
|
|
136
|
-
|
|
137
|
-
|
|
138
|
-
|
|
139
|
-
|
|
140
|
-
|
|
141
|
-
early_stopping = EarlyStopping(monitor='val_loss', patience=10, verbose=1)
|
|
142
|
-
|
|
143
|
-
|
|
144
|
-
|
|
145
|
-
model = Sequential()
|
|
146
|
-
|
|
147
|
-
model.add(Bidirectional(LSTM(n_hidden),
|
|
148
|
-
|
|
149
|
-
input_shape=(n_long,n_in)))
|
|
150
|
-
|
|
151
|
-
model.add(Dense(n_out, kernel_initializer=weight_variable))
|
|
152
|
-
|
|
153
|
-
model.add(Activation('softmax'))
|
|
154
|
-
|
|
155
|
-
|
|
156
|
-
|
|
157
|
-
model.compile(loss='categorical_crossentropy',
|
|
158
|
-
|
|
159
|
-
optimizer=Adam(lr=0.001, beta_1=0.9, beta_2=0.999),
|
|
160
|
-
|
|
161
|
-
metrics=['accuracy'])
|
|
162
|
-
|
|
163
|
-
|
|
164
|
-
|
|
165
|
-
'''
|
|
166
|
-
|
|
167
|
-
モデル学習
|
|
168
|
-
|
|
169
|
-
'''
|
|
170
|
-
|
|
171
|
-
epochs = 300
|
|
172
|
-
|
|
173
|
-
batch_size = 250
|
|
174
|
-
|
|
175
|
-
|
|
176
|
-
|
|
177
|
-
hist = model.fit(X_train, Y_train,
|
|
178
|
-
|
|
179
|
-
batch_size=batch_size,
|
|
180
|
-
|
|
181
|
-
epochs=epochs,
|
|
182
|
-
|
|
183
|
-
validation_data=(X_validation, Y_validation),
|
|
184
|
-
|
|
185
|
-
callbacks=[early_stopping])
|
|
186
|
-
|
|
187
|
-
|
|
188
|
-
|
|
189
|
-
'''
|
|
190
|
-
|
|
191
|
-
学習の進み具合
|
|
192
|
-
|
|
193
|
-
'''
|
|
194
|
-
|
|
195
|
-
acc = hist.history['val_acc']
|
|
196
|
-
|
|
197
|
-
loss = hist.history['val_loss']
|
|
198
|
-
|
|
199
|
-
|
|
200
|
-
|
|
201
|
-
plt.rc('font', family='serif')
|
|
202
|
-
|
|
203
|
-
fig = plt.figure()
|
|
204
|
-
|
|
205
|
-
plt.plot(range(len(loss)), loss,
|
|
206
|
-
|
|
207
|
-
label='loss', color='black')
|
|
208
|
-
|
|
209
|
-
plt.xlabel('epochs')
|
|
210
|
-
|
|
211
|
-
plt.show()
|
|
212
|
-
|
|
213
|
-
|
|
214
|
-
|
|
215
|
-
'''
|
|
216
|
-
|
|
217
|
-
予測精度の評価
|
|
218
|
-
|
|
219
|
-
'''
|
|
220
|
-
|
|
221
|
-
loss_and_metrics = model.evaluate(X_test, Y_test)
|
|
222
|
-
|
|
223
|
-
print(loss_and_metrics)
|
|
224
|
-
|
|
225
229
|
|
|
226
230
|
|
|
227
231
|
### 試したこと
|