Created
May 31, 2021 07:58
-
-
Save michelkana/58147829b35d3d4254da5c97d336f6e3 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
from keras.datasets import cifar10 | |
from keras.utils import np_utils | |
from keras.models import Sequential, Model | |
from keras.layers import Dense, Dropout, Flatten | |
from keras.layers.convolutional import Conv2D, MaxPooling2D | |
from keras import regularizers | |
from keras.layers import BatchNormalization | |
from keras.optimizers import RMSprop | |
from keras.preprocessing.image import ImageDataGenerator | |
# load and prepate cifar images | |
(xtrain, ytrain), (xtest, ytest) = cifar10.load_data() | |
x_train = xtrain.astype('float32') | |
x_test = xtest.astype('float32') | |
x_train /= 255 | |
x_test /= 255 | |
nb_classes = 10 | |
y_train = np_utils.to_categorical(ytrain, nb_classes) | |
y_test = np_utils.to_categorical(ytest, nb_classes) | |
# augment images with a generator | |
cifar_gen = ImageDataGenerator( | |
rotation_range=15, | |
width_shift_range=0.1, | |
height_shift_range=0.1, | |
horizontal_flip=True, | |
) | |
cifar_gen.fit(x_train) | |
# create convolutional neural network (CNN) | |
def create_cnn(input_shape=x_train.shape[1:], nb_classes=nb_classes, | |
nb_blocks=3, nb_filters=32, filter_size=(3,3), | |
pool_size=(2,2), weight_decay=1e-4, padding='same', | |
dropout=.2, output_activation='softmax'): | |
model = Sequential() | |
for i in range(nb_blocks): | |
if i==0: | |
model.add(Conv2D(nb_filters, filter_size, activation='relu', | |
padding=padding, kernel_regularizer=regularizers.l2(weight_decay), | |
input_shape=input_shape)) | |
else: | |
model.add(Conv2D(nb_filters, filter_size, activation='relu', | |
padding=padding, kernel_regularizer=regularizers.l2(weight_decay))) | |
model.add(BatchNormalization()) | |
model.add(Conv2D(nb_filters, filter_size, activation='relu', | |
padding=padding, kernel_regularizer=regularizers.l2(weight_decay))) | |
model.add(BatchNormalization()) | |
model.add(MaxPooling2D(pool_size=pool_size)) | |
model.add(Dropout(dropout)) | |
model.add(Flatten()) | |
model.add(Dense(nb_classes, activation=output_activation)) | |
return model | |
cnn = create_cnn(nb_filters=32) | |
# compile CNN | |
opt_rms = RMSprop(lr=0.001, decay=1e-6) | |
cnn.compile(loss='categorical_crossentropy', optimizer=opt_rms, metrics=['accuracy']) | |
# train CNN | |
batch_size = 64 | |
epochs = 100 | |
steps = x_train.shape[0] | |
cnn_history = cnn.fit_generator(cifar_gen.flow(x_train, y_train, batch_size=batch_size), | |
steps_per_epoch=steps, | |
epochs=epochs, | |
verbose=1, | |
validation_data=(x_test,y_test)) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
When I try to run the code, it shows the below error:
ImportError: cannot import name 'RMSprop' from 'keras.optimizers' (/Users/jianbfan/opt/anaconda3/lib/python3.8/site-packages/keras/optimizers.py)