InceptionV3的Keras输入错误，预期input_1具有形状None

Question

ValueError: Error when checking input: expected input_1 to have shape (None, 299, 299, 3) but got array with shape (32, 229, 229, 3)

以下是代码的核心部分：

train_datagen = ImageDataGenerator()
train_batches = train_datagen.flow_from_directory('../train/',
                                 target_size=(229, 229))

test_datagen = ImageDataGenerator()
test_batches = test_datagen.flow_from_directory('../test/',
                                 target_size=(229, 229))
#
model = keras.applications.InceptionV3(weights="imagenet")
#
clf_model.compile(optimizer='rmsprop',
              loss='categorical_crossentropy',
              metrics=['accuracy'])
train_history = clf_model.fit_generator(train_batches,
                    steps_per_epoch=train_batches.samples/train_batches.batch_size, 
                                        epochs=1, 
                                        validation_data = test_batches,
                    validation_steps=test_batches.samples/test_batches.batch_size )

看起来很奇怪。我看到其他人的错误就像是

Error when checking input: expected input_1 to have shape (None, 10, 3) but got array with shape (1, 3, 10) 

或

Error when checking input: expected input_1 to have shape (None, 10, 3) but got array with shape (1, 4, 16) 

然而，在我的错误中，形状看起来很好。 Keras抱怨形状不是没有...... 请帮忙。谢谢

Answer 1

您的输入形状不是数学：

train_datagen = ImageDataGenerator()
train_batches = train_datagen.flow_from_directory('../train/',
                             target_size=(229, 229))

test_datagen = ImageDataGenerator()
test_batches = test_datagen.flow_from_directory('../test/',
                             target_size=(229, 229))

将train_satches和test_batches的target_size =（229,229）更改为target_size =（299,299）。