回答編集履歴
3
class_modeの設定を見逃していました．
answer
CHANGED
|
@@ -4,26 +4,17 @@
|
|
|
4
4
|
|
|
5
5
|
```Python
|
|
6
6
|
seed = 1 # 共通の乱数seedを使うことでshuffleしても同じ順に取り出されるようにする．
|
|
7
|
-
|
|
7
|
+
train_datagen = ImageDataGenerator(rescale=1.0 / 255, seed = seed)
|
|
8
|
-
|
|
8
|
+
train_data = train_datagen.flow_from_directory(
|
|
9
9
|
directory = train_data_dir, target_size = (img_height, img_width), color_mode='grayscale',
|
|
10
10
|
class_mode='input', batch_size=batch_size, shuffle=True, seed = seed
|
|
11
11
|
)
|
|
12
|
-
train_datagen_output = ImageDataGenerator(rescale=1.0 / 255, seed = seed)
|
|
13
|
-
train_data_output = train_datagen_output.flow_from_directory(
|
|
14
|
-
directory = train_data_dir, target_size = (img_height, img_width), color_mode='grayscale',
|
|
15
|
-
class_mode='input', batch_size=batch_size, shuffle=True, seed = seed
|
|
16
|
-
)
|
|
17
12
|
|
|
18
|
-
train_generator = zip(train_datagen_input, train_datagen_output) # 1つにまとめる
|
|
19
|
-
|
|
20
13
|
autoencoder.fit(
|
|
21
|
-
|
|
14
|
+
train_data,
|
|
22
15
|
epochs=50,
|
|
23
16
|
batch_size=128,
|
|
24
17
|
shuffle=True,
|
|
25
18
|
# validation_data=(test_data, test_data), # ここも同様に変更してください．
|
|
26
19
|
)
|
|
27
20
|
```
|
|
28
|
-
|
|
29
|
-
ちなみに`train_datagen_input`と`train_datagen_output`をまとめたくなる気持ちはわかりますが`zip()`の実装的に画像が正しく取り出されないので，冗長に見えますが上のように記述するようにしてください．
|
2
fix typo
answer
CHANGED
|
@@ -18,7 +18,7 @@
|
|
|
18
18
|
train_generator = zip(train_datagen_input, train_datagen_output) # 1つにまとめる
|
|
19
19
|
|
|
20
20
|
autoencoder.fit(
|
|
21
|
-
|
|
21
|
+
train_generator,
|
|
22
22
|
epochs=50,
|
|
23
23
|
batch_size=128,
|
|
24
24
|
shuffle=True,
|
1
fixed correct answer
answer
CHANGED
|
@@ -18,8 +18,7 @@
|
|
|
18
18
|
train_generator = zip(train_datagen_input, train_datagen_output) # 1つにまとめる
|
|
19
19
|
|
|
20
20
|
autoencoder.fit(
|
|
21
|
-
|
|
21
|
+
train_datagenerator,
|
|
22
|
-
train_data,
|
|
23
22
|
epochs=50,
|
|
24
23
|
batch_size=128,
|
|
25
24
|
shuffle=True,
|