Q&A
実現したいこと
2枚の画像を入力して1つの出力を得る（分類される）CNNの設計を行っています。
具体的には，VGG16の特徴抽出層（重みは初期化しないでImagenetで学習済みのものを使用）を用いて2枚の画像から特徴を抽出し，それを全結合層で推論するネットワークです。
しかし，以下のエラーが発生しました。
エラーメッセージ
--------------------------------------------------------------------------- ValueError Traceback (most recent call last) <ipython-input-62-d0881f532e33> in <module>() ----> 1 NN3_conv = Flatten()(merged) 2 NN3_conv = Dense(8192)(NN3_conv) 3 NN3_conv = BatchNormalization()(NN3_conv) 4 NN3_conv = Dense(8192)(NN3_conv) 5 NN3_conv = BatchNormalization()(NN3_conv) ~\Anaconda3\lib\site-packages\keras\engine\base_layer.py in __call__(self, inputs, **kwargs) 504 if all([s is not None 505 for s in to_list(input_shape)]): --> 506 output_shape = self.compute_output_shape(input_shape) 507 else: 508 if isinstance(input_shape, list): ~\Anaconda3\lib\site-packages\keras\layers\core.py in compute_output_shape(self, input_shape) 499 raise ValueError('The shape of the input to "Flatten" ' 500 'is not fully defined ' --> 501 '(got ' + str(input_shape[1:]) + '). ' 502 'Make sure to pass a complete "input_shape" ' 503 'or "batch_input_shape" argument to the first ' ValueError: The shape of the input to "Flatten" is not fully defined (got (None, None, 1024)). Make sure to pass a complete "input_shape" or "batch_input_shape" argument to the first layer in your model.
ソースコード
from keras.models import Model, Sequential from keras.layers import Input, Dense, Dropout, Activation, Flatten from keras.layers import add, concatenate from keras.utils import plot_model from keras.applications.vgg16 import VGG16 import keras NN1=keras.applications.vgg16.VGG16(include_top=False, weights='imagenet', input_tensor=None, input_shape=None, pooling=None) NN2=keras.applications.vgg16.VGG16(include_top=False, weights='imagenet', input_tensor=None, input_shape=None, pooling=None) merged = concatenate([NN1.output, NN2.output]) NN3_conv = Flatten()(merged) NN3_conv = Dense(8192)(NN3_conv) NN3_conv = BatchNormalization()(NN3_conv) NN3_conv = Dense(8192)(NN3_conv) NN3_conv = BatchNormalization()(NN3_conv) NN3_conv = Dense(CLASS_NUM, activation="softmax")(NN3_conv) model = Model([NN1.input, NN2.input], NN3_conv)
何が原因なのかよく分かりません。
どなたかご教授よろしくお願いします。
環境・バージョン
OS
windows10
バージョン
python3.6.4
keras2.3.1
回答1件
0
ベストアンサー
Imagenetの基本サイズを指定したところ正常にFlattenできました。 注意点は、同じVGG16のモデルを２回利用するため、普通に並べてしまうとkerasの内部でlayerの名前がバッティングします。ということで、名前だけ途中で書き換えています。
python
1from keras.models import Model, Sequential 2from keras.layers import Input, Dense, Dropout, Activation, Flatten 3from keras.layers import add, concatenate, BatchNormalization 4from keras.utils import plot_model 5from keras.applications.vgg16 import VGG16 6import keras 7 8NN1=keras.applications.vgg16.VGG16(include_top=False, weights='imagenet', input_tensor=None, input_shape=(224,224,3), pooling=None) 9NN2=keras.applications.vgg16.VGG16(include_top=False, weights='imagenet', input_tensor=None, input_shape=(224,224,3), pooling=None) 10for layer in NN1.layers: 11 layer.name = layer.name + str("NN1") 12for layer in NN2.layers: 13 layer.name = layer.name + str("NN2") 14 15merged = concatenate([NN1.output, NN2.output]) 16 17NN3_conv1 = Flatten()(merged) 18NN3_conv2 = Dense(8192)(NN3_conv1) 19NN3_conv3 = BatchNormalization()(NN3_conv2) 20NN3_conv4 = Dense(8192)(NN3_conv3) 21NN3_conv5 = BatchNormalization()(NN3_conv4) 22NN3_conv6 = Dense(1000, activation="softmax")(NN3_conv5) 23 24model = Model([NN1.input, NN2.input], NN3_conv6) 25model.summary()
console
1__________________________________________________________________________________________________ 2Layer (type) Output Shape Param # Connected to 3================================================================================================== 4input_1NN1 (InputLayer) (None, 224, 224, 3) 0 5__________________________________________________________________________________________________ 6input_2NN2 (InputLayer) (None, 224, 224, 3) 0 7__________________________________________________________________________________________________ 8block1_conv1NN1 (Conv2D) (None, 224, 224, 64) 1792 input_1NN1[0][0] 9__________________________________________________________________________________________________ 10block1_conv1NN2 (Conv2D) (None, 224, 224, 64) 1792 input_2NN2[0][0] 11__________________________________________________________________________________________________ 12 13・・中略・・ 14 15block1_conv2NN1 (Conv2D) (None, 224, 224, 64) 36928 block1_conv1NN1[0][0] 16__________________________________________________________________________________________________ 17block1_conv2NN2 (Conv2D) (None, 224, 224, 64) 36928 block1_conv1NN2[0][0] 18__________________________________________________________________________________________________ 19block1_poolNN1 (MaxPooling2D) (None, 112, 112, 64) 0 block1_conv2NN1[0][0] 20__________________________________________________________________________________________________ 21block1_poolNN2 (MaxPooling2D) (None, 112, 112, 64) 0 block1_conv2NN2[0][0] 22__________________________________________________________________________________________________ 23block2_conv1NN1 (Conv2D) (None, 112, 112, 128 73856 block1_poolNN1[0][0] 24__________________________________________________________________________________________________ 25block2_conv1NN2 (Conv2D) (None, 112, 112, 128 73856 block1_poolNN2[0][0] 26__________________________________________________________________________________________________ 27block2_conv2NN1 (Conv2D) (None, 112, 112, 128 147584 block2_conv1NN1[0][0] 28 29flatten_1 (Flatten) (None, 50176) 0 concatenate_1[0][0] 30__________________________________________________________________________________________________ 31dense_1 (Dense) (None, 8192) 411049984 flatten_1[0][0] 32__________________________________________________________________________________________________ 33batch_normalization_1 (BatchNor (None, 8192) 32768 dense_1[0][0] 34__________________________________________________________________________________________________ 35dense_2 (Dense) (None, 8192) 67117056 batch_normalization_1[0][0] 36__________________________________________________________________________________________________ 37batch_normalization_2 (BatchNor (None, 8192) 32768 dense_2[0][0] 38__________________________________________________________________________________________________ 39dense_3 (Dense) (None, 1000) 8193000 batch_normalization_2[0][0] 40================================================================================================== 41Total params: 515,854,952 42Trainable params: 515,822,184 43Non-trainable params: 32,768 44__________________________________________________________________________________________________ 45
投稿2020/11/22 05:44
総合スコア351
あなたの回答
tips
プレビュー
バッドをするには、ログインかつ
こちらの条件を満たす必要があります。
2020/11/26 07:03