Neural Network on Keras

Create a new Jupyter notebook with python 2.7 kernel. Name it as TensorFlow Deep Learning. Let’s import all the required modules. Import libraries and modules.

Keras

%pylab inline

import numpy as np
np.random.seed(123)  # for reproducibility

from keras.models import Sequential
from keras.layers import Dense, Dropout, Activation, Flatten
from keras.layers import Convolution2D, MaxPooling2D
from keras.utils import np_utils

Populating the interactive namespace from numpy and matplotlib. Load image data from MNIST.

from keras.datasets import mnist
 
# Load pre-shuffled MNIST data into train and test sets
(X_train, y_train), (X_test, y_test) = mnist.load_data()

Shape of the dataset

print X_train.shape

60,000 samples in the training set, and the images are 28 pixels x 28 pixels each. Plotting the first sample in matplotlib

from matplotlib import pyplot as plt
plt.imshow(X_train[0])

Preprocess input data for Keras. When using the Theano backend, you must explicitly declare a dimension for the depth of the input image. For example, a full-color image with all 3 RGB channels will have a depth of 3. Our MNIST images only have a depth of 1, but we must explicitly declare that. In other words, we want to transform our dataset from having shape (n, width, height) to (n, depth, width, height).

X_train = X_train.reshape(X_train.shape[0], 1, 28, 28)
X_test = X_test.reshape(X_test.shape[0], 1, 28, 28)

Show X_train’s dimensions again

print X_train.shape

The final preprocessing step for the input data is to convert our data type to float32 and normalize our data values to the range [0, 1].

X_train = X_train.astype('float32')
X_test = X_test.astype('float32')
X_train /= 255
X_test /= 255

Preprocess class labels for Keras. Check shape of our class label data.

print y_train.shape

Convert 1-dimensional class arrays to 10-dimensional class matrices

Y_train = np_utils.to_categorical(y_train, 10)
Y_test = np_utils.to_categorical(y_test, 10)

Check shape of our class label data again

print Y_train.shape

Define model architecture. Declaring a sequential model format

model = Sequential()

Declare the input layer which is (1, 28, 28) that corresponds to the (depth, width, height) of each digit image

model.add(Convolution2D(32, (3, 3), activation='relu', input_shape=(1,28,28), data_format='channels_first'))

Shape of the current model output

print model.output_shape

Add more layers to our model. For Dense layers, the first parameter is the output size of the layer. Keras automatically handles the connections between layers.

Note that the final layer has an output size of 10, corresponding to the 10 classes of digits.

Also note that the weights from the Convolution layers must be flattened (made 1-dimensional) before passing them to the fully connected Dense layer.

model.add(Convolution2D(32, (3, 3), activation='relu'))
model.add(MaxPooling2D(pool_size=(2,2)))
model.add(Dropout(0.25))

model.add(Flatten())
model.add(Dense(128, activation='relu'))
model.add(Dropout(0.5))
model.add(Dense(10, activation='softmax'))

Compile model. Declare the loss function and the optimizer (SGD, Adam, etc.).

model.compile(loss='categorical_crossentropy',
              optimizer='adam',
              metrics=['accuracy'])

Fit model on training data. Declare the batch size and number of epochs to train for, then pass in training data

model.fit(X_train, Y_train, batch_size=32, epochs=10, verbose=1)

Evaluate model on test data.

score = model.evaluate(X_test, Y_test, verbose=0)

Predict the image

Y_test = model.predict_classes(X_test, verbose=2)