tensorflow_keras_cifar10

tensorflow_keras_cifar10.py

import os, json, shutil

import keras
from keras.callbacks import TensorBoard
from keras.datasets import cifar10
from keras.preprocessing.image import ImageDataGenerator
from keras.models import Sequential
from keras.layers import Dense, Dropout, Activation, Flatten
from keras.layers import Conv2D, MaxPooling2D

# https://github.com/tensorflow/tensorflow/issues/24828
try:
    from tensorflow.compat.v1 import ConfigProto
    from tensorflow.compat.v1 import InteractiveSession
    config = ConfigProto()
    config.gpu_options.allow_growth = True
    session = InteractiveSession(config=config)
except:
    pass

NUM_CLASSES = 10
DATA_AUGMENTATION = False

NUM_EPOCHS    = 3
LEARNING_RATE = 0.0001
BATCH_SIZE    = 32
WEIGHT_DECAY  = 1e-6
OPTIMIZER     = 'Adam' # Adam, RMSPROP, SGD
TENSORBOARD_ENABLED = "False"

######################## AUTOML SETTINGS ##########################
# SEARCH_SPACE:
# {"OPTIMIZER": choice(["Adam", "SGD", "RMSprop"]),
#  "LEARNING_RATE": choice([0.0001, 0.00025]), 
#  "BATCH_SIZE": choice([32, 64]), 
#  "WEIGHT_DECAY": choice([0.0005, 0.005])}
"""
NUM_EPOCHS    = int(variables.get("NUM_EPOCHS"))
LEARNING_RATE = float(variables.get("LEARNING_RATE"))
BATCH_SIZE    = int(variables.get("BATCH_SIZE"))
WEIGHT_DECAY  = float(variables.get("WEIGHT_DECAY"))
OPTIMIZER     = variables.get("OPTIMIZER")
TENSORBOARD_ENABLED = variables.get("TENSORBOARD_ENABLED")

input_variables = variables.get("INPUT_VARIABLES")
if input_variables is not None and input_variables != '':
    input_variables = json.loads(input_variables)
    LEARNING_RATE = input_variables["LEARNING_RATE"]
    BATCH_SIZE = input_variables["BATCH_SIZE"]
    WEIGHT_DECAY = input_variables["WEIGHT_DECAY"]
    OPTIMIZER = input_variables["OPTIMIZER"]
"""
###################################################################

OPTIMIZER = OPTIMIZER.upper()

# Create a TensorBoard instance with the path to the logs directory
if TENSORBOARD_ENABLED.lower() == "true":
    DOCKER_LOG_PATH = variables.get("DOCKER_LOG_PATH")
    PA_JOB_ID = "job_ID_" + variables.get("PA_JOB_ID")
    PATH = os.path.join(DOCKER_LOG_PATH, PA_JOB_ID)
    print('PATH: ', PATH)
    if os.path.isdir(PATH):
        try:
            print('Removing existing path')
            shutil.rmtree(PATH)
        except:
            print('Error while deleting directory')
    os.mkdir(PATH)
    tensorboard = TensorBoard(log_dir=PATH)

print('-' * 30)
print('NUM_CLASSES:       ', NUM_CLASSES)
print('DATA_AUGMENTATION: ', DATA_AUGMENTATION)
print('NUM_EPOCHS:        ', NUM_EPOCHS)
print('LEARNING_RATE:     ', LEARNING_RATE)
print('BATCH_SIZE:        ', BATCH_SIZE)
print('WEIGHT_DECAY:      ', WEIGHT_DECAY)
print('OPTIMIZER:         ', OPTIMIZER)
print('-' * 30)

# The data, split between train and test sets:
(x_train, y_train), (x_test, y_test) = cifar10.load_data()
print('x_train shape:', x_train.shape)
print(x_train.shape[0], 'train samples')
print(x_test.shape[0], 'test samples')

# 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)
print(x_train.shape[1:])

# Implemnting above stack into code.
model = Sequential()
# Adding layers to the sequential model 
# .add() will push the layer into the stack(sequential model)
# filter -> 32
# strides ->(3,3)
model.add(Conv2D(32, (3, 3), padding='same',input_shape=x_train.shape[1:]))
# Reulu is an activation function faster than sigmoid and reduce the liklehod of vanishing gradient 
model.add(Activation('relu'))

# We dont have to define parameters again.
model.add(Conv2D(32, (3, 3)))
model.add(Activation('relu'))
# Maxpooling is done for downsampling the input and reducing its dimensionality
# and also help over-fitting by providing an abstracted form of the representation
model.add(MaxPooling2D(pool_size=(2, 2)))
# Drop put is a regularization technique where randomly selected neurons are ignored during training
# Here 25% connection will drop
model.add(Dropout(0.25))

# Depth is changed into 64
model.add(Conv2D(64, (3, 3), padding='same'))
model.add(Activation('relu'))
model.add(Conv2D(64, (3, 3)))
model.add(Activation('relu'))
model.add(MaxPooling2D(pool_size=(2, 2)))
model.add(Dropout(0.25))

# Conver matrix into a vector by streching it.
model.add(Flatten())

# Dense layer is use to collect all the info 
model.add(Dense(512))

model.add(Activation('relu'))

model.add(Dropout(0.5))
# Number of target classes
model.add(Dense(NUM_CLASSES))
# Softmax convert probabilty into binary number.
model.add(Activation('softmax'))

# Initiate RMSprop optimizer
if OPTIMIZER == 'RMSPROP':
    opt = keras.optimizers.rmsprop(lr=LEARNING_RATE, decay=WEIGHT_DECAY)
elif OPTIMIZER == 'SGD':
    opt = keras.optimizers.SGD(lr=LEARNING_RATE, decay=WEIGHT_DECAY)
else:
    opt = keras.optimizers.Adam(lr=LEARNING_RATE, decay=WEIGHT_DECAY)

# Let's train the model using RMSprop
model.compile(loss='categorical_crossentropy',
              optimizer=opt,
              metrics=['accuracy'])

x_train = x_train.astype('float32')
x_test = x_test.astype('float32')
x_train /= 255
x_test /= 255

if not DATA_AUGMENTATION:
    print('Not using data augmentation.')
    if TENSORBOARD_ENABLED.lower() == "true":
    	model.fit(x_train, y_train,
             batch_size =BATCH_SIZE,
              epochs=NUM_EPOCHS,
              validation_data=(x_test, y_test),
              shuffle=True, callbacks=[tensorboard])
    else:
        model.fit(x_train, y_train,
             batch_size =BATCH_SIZE,
              epochs=NUM_EPOCHS,
              validation_data=(x_test, y_test),
              shuffle=True)
else:
    print('Using real-time data augmentation.')
    # This will do preprocessing and realtime data augmentation:
    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=0.1,  # randomly shift images horizontally (fraction of total width)
        height_shift_range=0.1,  # 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), 
      steps_per_epoch=x_train.shape[0], 
      epochs=NUM_EPOCHS, 
      validation_data=(x_test, y_test)
    )

# Score trained model.
scores = model.evaluate(x_test, y_test, verbose=1)
loss = scores[0]
acc = scores[1]
print('Test loss:', loss)
print('Test accuracy:', acc)

######################## AUTOML SETTINGS ##########################
"""
token = variables.get("TOKEN")
# Convert from JSON to dict
token = json.loads(token)

# Return the loss value
result_map = {'token': token, 'loss':  loss}
print('result_map: ', result_map)

resultMap.put("RESULT_JSON", json.dumps(result_map))

# To appear in Job Analytics
resultMap.put("LOSS", str(loss))
resultMap.put("ACCURACY", str(acc))
"""
###################################################################