Keras memory error issue

mem_err.py

from keras.models import Sequential
from keras.layers.core import Flatten, Dense, Dropout, Lambda
from keras.layers.normalization import BatchNormalization
from keras.layers.convolutional import Convolution2D, MaxPooling2D, ZeroPadding2D
from keras.optimizers import SGD, RMSprop, Adam
from theano.sandbox import cuda
from keras.preprocessing import image
cuda.use('gpu')


def get_batches(path, gen=image.ImageDataGenerator(), shuffle=True, batch_size=8, class_mode='categorical'):
    return gen.flow_from_directory(path, target_size=(224,224),
            class_mode=class_mode, shuffle=shuffle, batch_size=batch_size)
  
  
batch_size = 8
path = "./data/sample/";
train_batches = get_batches(path + "train", batch_size=batch_size)
valid_batches = get_batches(path + "valid", batch_size=batch_size, shuffle=False)
test_batches  = get_batches(path + "test", batch_size=batch_size, shuffle=False)


def get_model(p):
    return Sequential([
         BatchNormalization(input_shape=(3, 244, 244)),
        ZeroPadding2D((1, 1), input_shape=(3, 244, 244)),
        Convolution2D(64, 3, 3, activation='relu'),
        ZeroPadding2D((1, 1)),
        Convolution2D(64, 3, 3, activation='relu'),
        MaxPooling2D((2, 2), strides=(2, 2)),
        ZeroPadding2D((1, 1)),
        Convolution2D(128, 3, 3, activation='relu'),
        ZeroPadding2D((1, 1)),
        Convolution2D(128, 3, 3, activation='relu'),
        MaxPooling2D((2, 2), strides=(2, 2)),
        Flatten(),
        Dense(4096, activation='relu'),
        Dropout(p),
        Dense(train_batches.nb_class, activation='softmax')
    ])

  
model = get_model(0.5)
model.compile(optimizer=RMSprop(lr=1e-10),
        loss='categorical_crossentropy', metrics=['accuracy'])

model.fit_generator(train_batches, samples_per_epoch=train_batches.nb_sample, nb_epoch=1,
                validation_data=valid_batches, nb_val_samples=valid_batches.nb_sample, verbose=True)


# Error
# MemoryError: ('Error allocating 7803502592 bytes of device memory (out of memory).', "you might consider using 'theano.shared(..., borrow=True)'")