Get started with keras for deep learning.

• This post is for ultimate begginers in deep learning. It helps to get started with keras a python based deep learning package. Keras runs on backends like Theano, Tensorflow and CNTK. 
• In this post, I will explain the setting up of Keras in your computer using backend as Tensorflow and running a simple multilayer perceptron using keras.
• Step 0: Before starting make sure that you have installed,

 python 2.7 (prefered) or python 3,

 scipy, numpy, Matplotlib  

•  Step 1: Now install the Tensorflow, if you have compute GPU in your computer then you can install GPU version of Tensorflow otherwise install CPU version of Tensorflow.
  

  Instructions to install Tensorflow : LINK

  you can check whether the packages are installed or not by simply importing them.
  

 $ python

>>import tensorflow  
>>print tensorflow.__version__  
>>import numpy  
>>print numpy.__version__  
>>import scipy  
>>print scipy.__version__   

• Step 2: Installing Keras in Ubuntu

  $ sudo pip install keras

  $ python

  >>import keras

  >>using Tensorflow backend

  Example: Digit classification using multilayer perceptron (MLP) on MNIST dataset.

  #########Digit classification using MLP on MNIST dataset###

  import keras
  from keras.datasets import mnist
  from keras.models import Sequential
  from keras.layers import Dense, Dropout
  from keras.optimizers import RMSprop
  import matplotlib.pyplot as plt
  
  batch_size = 128
  num_classes = 10
  epochs = 20
  
  # the data, shuffled and split between train and test sets
  (x_train, y_train), (x_test, y_test) = mnist.load_data()
  
  x_train = x_train.reshape(60000, 784)
  x_test = x_test.reshape(10000, 784)
  x_train = x_train.astype('float32')
  x_test = x_test.astype('float32')
  x_train /= 255
  x_test /= 255
  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)
  
  model = Sequential()
  model.add(Dense(512, activation='relu', input_shape=(784,)))
  model.add(Dropout(0.2))
  model.add(Dense(512, activation='relu'))
  model.add(Dropout(0.2))
  model.add(Dense(num_classes, activation='softmax'))
  
  model.summary()
  
  model.compile(loss='categorical_crossentropy',
                optimizer=RMSprop(),
                metrics=['accuracy'])
  
  history = model.fit(x_train, y_train,
                      batch_size=batch_size,
                      epochs=epochs,
                      verbose=1,
                      validation_data=(x_test, y_test))
  score = model.evaluate(x_test, y_test, verbose=0)
  print('Test loss:', score[0])
  print('Test accuracy:', score[1])
  
  print(history.history.keys())
  # summarize history for accuracy
  plt.plot(history.history['acc'])
  plt.plot(history.history['val_acc'])
  plt.title('model accuracy')
  plt.ylabel('accuracy')
  plt.xlabel('epoch')
  plt.legend(['train', 'test'], loc='upper left')
  plt.show()
  # summarize history for loss
  plt.plot(history.history['loss'])
  plt.plot(history.history['val_loss'])
  plt.title('model loss')
  plt.ylabel('loss')
  plt.xlabel('epoch')
  plt.legend(['train', 'test'], loc='upper left')
  plt.show()

  OutPut: The model.summary() gives the details about
  the network architecture and number of trainable and non-
  trainable parameters. 

  ____________________________________________________________
  Layer (type)              Output Shape            Param #   
  ==========================================================
  dense_4 (Dense)           (None, 512)             401920    
  __________________________________________________________
  dropout_3 (Dropout)       (None, 512)             0         
  ___________________________________________________________
  dense_5 (Dense)           (None, 512)             262656    
  ___________________________________________________________
  dropout_4 (Dropout)       (None, 512)             0         
  ___________________________________________________________
  dense_6 (Dense)           (None, 10)              5130      
  ===========================================================
  Total params: 669,706
  Trainable params: 669,706
  Non-trainable params: 0 

  Training and testing accuracy vs epoch are showin in the following figure.
  

  

  Training and testing loss vs epoch are showin in the following figure.