Keras imageGenerator Exception: output of generato

I'm currently trying to follow the example here using a dataset I generated by myself. The back end is run using Theano. The directory structure is exactly the same:

image_sets/
    dogs/
        dog001.jpg
        dog002.jpg
        ...
    cats/
        cat001.jpg
        cat002.jpg
        ...
validation/
    dogs/
        dog001.jpg
        dog002.jpg
        ...
    cats/
        cat001.jpg

Here is my code for the keras convolutional neural network.

  img_width, img_height = 150, 150

img_width, img_height = 150, 150
train_data_dir = './image_sets'
validation_data_dir = './validation'
nb_train_samples = 267
print nb_train_samples
#number of validation images I have
nb_validation_samples =  2002
print nb_validation_samples
nb_epoch = 50
# from keras import backend as K
# K.set_image_dim_ordering('th')

model = Sequential()
model.add(Convolution2D(32, 3, 3, input_shape=(3,img_width, img_height)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))

model.add(Convolution2D(32, 3, 3))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))

model.add(Convolution2D(64, 3, 3))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))

model.add(Flatten())
model.add(Dense(64))
model.add(Activation('relu'))
model.add(Dropout(0.5))
model.add(Dense(1))
model.add(Activation('sigmoid'))

model.compile(loss='binary_crossentropy',
              optimizer='rmsprop',
              metrics=['accuracy'])

train_datagen = ImageDataGenerator(
        rescale=1./255,
        shear_range=0.2,
        zoom_range=0.2,
        horizontal_flip=True)

test_datagen = ImageDataGenerator(rescale=1./255)
train_generator = train_datagen.flow_from_directory(
        train_data_dir,
        target_size=(img_width, img_height),
        batch_size=32,
        class_mode='binary')
validation_generator = test_datagen.flow_from_directory(
        validation_data_dir,
        target_size=(img_width, img_height),
        batch_size=32,
        class_mode='binary')

model.fit_generator(
        train_generator,
        samples_per_epoch=nb_train_samples,
        nb_epoch=nb_epoch,
        validation_data=validation_generator,
        nb_val_samples=nb_validation_samples)
model.save_weights('first_try.h5')

回答1:

I ran into the same problem while running the code but i was using tensorflow as backend. My problem was that i was running it on an older version of keras.

Upgrade to keras 2.0 by

pip install --upgrade keras

Then update your fit_generator function as follows-

model.fit_generator(generator=train_generator,
                    steps_per_epoch=2048 // 16,
                    epochs=20,
                    validation_data=validation_generator,
                    validation_steps=832//16)

Here, 16 is your batch_size.

You can find the complete updated code by fchollet : Here.

回答2:

Your generator should be a python generator. You can read more of that here.

Briefly explained, a generator allows you to yield a series of value from the function called without its variables to be cleaned (as it is the case with a return statement for example).