Last active
June 21, 2020 16:45
-
-
Save abearman/916673e9a0f12bcf5ac2723b8b5eb819 to your computer and use it in GitHub Desktop.
Showing that a simple ResNet50 Keras network run on CIFAR10 does not have reproducible validation loss with learning rate = 0
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
import keras | |
import numpy as np | |
from keras.datasets import cifar10 | |
from keras.applications.resnet50 import ResNet50 | |
from keras.layers import GlobalAveragePooling2D, Dense | |
from keras.preprocessing.image import ImageDataGenerator | |
from keras.models import Model | |
from skimage.transform import resize | |
from IPython import embed | |
NUM_CLASSES = 10 | |
BATCH_SIZE = 2 | |
NUM_EPOCHS = 15 | |
use_data_aug = True | |
# img_arr is of shape (n, h, w, c) | |
def resize_image_arr(img_arr): | |
x_resized_list = [] | |
for i in range(img_arr.shape[0]): | |
img = img_arr[0] | |
resized_img = resize(img, (224, 224)) | |
x_resized_list.append(resized_img) | |
return np.stack(x_resized_list) | |
(x_train, y_train), (x_test, y_test) = cifar10.load_data() | |
x_train = x_train[0:2] | |
y_train = y_train[0:2] | |
x_test = x_test[0:2] | |
y_test = y_test[0:2] | |
# Resize image arrays | |
x_train = resize_image_arr(x_train) | |
x_test = resize_image_arr(x_test) | |
# Convert class vectors to binary class matrices. | |
y_train = keras.utils.to_categorical(y_train, NUM_CLASSES) | |
y_test = keras.utils.to_categorical(y_test, NUM_CLASSES) | |
# Normalize the data | |
x_train = x_train.astype('float32') | |
x_test = x_test.astype('float32') | |
x_train /= 255 | |
x_test /= 255 | |
base_model = ResNet50(include_top=False, weights='imagenet') | |
# add a global spatial average pooling layer | |
x = base_model.output | |
x = GlobalAveragePooling2D()(x) | |
# let's add a fully-connected layer | |
x = Dense(512, activation='relu')(x) | |
# and a logistic layer -- 10 classes for CIFAR10 | |
predictions = Dense(NUM_CLASSES, activation='softmax')(x) | |
# this is the model we will train | |
model = Model(inputs=base_model.input, outputs=predictions) | |
# initiate RMSprop optimizer | |
opt = keras.optimizers.rmsprop(lr=0.0, decay=0.0) | |
# Let's train the model using RMSprop | |
model.compile(loss='binary_crossentropy', | |
optimizer=opt, | |
metrics=['accuracy']) | |
if not use_data_aug: | |
model.fit(x_train, y_train, | |
batch_size=BATCH_SIZE, | |
epochs=NUM_EPOCHS, | |
validation_data=(x_test, y_test), | |
shuffle=False) | |
else: | |
datagen = ImageDataGenerator( | |
featurewise_center=False, # set input mean to 0 over the dataset | |
samplewise_center=False, # set each sample mean to 0 | |
featurewise_std_normalization=False, # divide inputs by std of the dataset | |
samplewise_std_normalization=False, # divide each input by its std | |
zca_whitening=False, # apply ZCA whitening | |
rotation_range=0, # randomly rotate images in the range (degrees, 0 to 180) | |
width_shift_range=0, # randomly shift images horizontally (fraction of total width) | |
height_shift_range=0, # randomly shift images vertically (fraction of total height) | |
horizontal_flip=True, # randomly flip images | |
vertical_flip=False) # randomly flip images | |
# Compute quantities required for feature-wise normalization | |
# (std, mean, and principal components if ZCA whitening is applied). | |
datagen.fit(x_train) | |
# Fit the model on the batches generated by datagen.flow(). | |
model.fit_generator(datagen.flow(x_train, y_train, | |
batch_size=BATCH_SIZE), | |
epochs=NUM_EPOCHS, | |
validation_data=(x_test, y_test), | |
workers=1) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment