three_layer_net.py

from __future__ import print_function

import numpy as np
import matplotlib.pyplot as plt
from past.builtins import xrange
from layers import *
class ThreeLayerNet(object):
    """
    A three-layer fully-connected neural network with a residual block. 
    The net has an input dimension of N, a hidden layer dimension of H, 
    a hidden2 layer dimension of H2, and performs classification over C classes.
    We train the network with a svm loss function and L2 regularization on the
    weight matrices. The network uses a ReLU nonlinearity after the first and second
    fully connected layer.

    In other words, the network has the following architecture:
           |-  residual fully connected layer                     -|
    input - fully connected layer - ReLU - fully connected layer -  ->  relu - FC - SVM

    The outputs of the second fully-connected layer are the scores for each class.
    """

    def __init__(self, input_size, hidden_size, hidden2_size, output_size, std=1e-4, use_Res=False):
        """
        Initialize the model. Weights are initialized to small random values and
        biases are initialized to zero. Weights and biases are stored in the
        variable self.params, which is a dictionary with the following keys:

        W1: First layer weights; has shape (D, H)
        b1: First layer biases; has shape (H,)
        W2: Second layer weights; has shape (H, H2)
        b2: Second layer biases; has shape (H2,)
        W3: Third layer weights; has shape (H2, C)
        b3: Third layer biases; has shape(C,)
        Wr: Residual layer weights; has shape (D, H2)
        bb: Residual layer biases; has shape(H2,)

        Inputs:
        - input_size: The dimension D of the input data.
        - hidden_size: The number of neurons H in the first hidden layer.
        - hidden2_size: The number of neurons H2 in the second hidden layer.
        - output_size: The number of classes C.
        - use_Res: Apply residual block shortcut (or not).
        """
        self.params = {}
        self.params['W1'] = std * np.random.randn(input_size, hidden_size)
        self.params['b1'] = np.zeros(hidden_size)
        self.params['W2'] = std * np.random.randn(hidden_size, hidden2_size)
        self.params['b2'] = np.zeros(hidden2_size)
        self.params['W3'] = std * np.random.randn(hidden2_size, output_size)
        self.params['b3'] = np.zeros(output_size)
        self.use_Res = use_Res
        if use_Res == True:
            self.params['Wr'] = std * np.random.randn(input_size, hidden2_size)
            self.params['br'] = np.zeros(hidden2_size)

    def loss(self, X, y=None, reg=0.0):
        """
        Compute the loss and gradients for a two layer fully connected neural
        network.

        Inputs:
        - X: Input data of shape (N, D). Each X[i] is a training sample.
        - y: Vector of training labels. y[i] is the label for X[i], and each y[i] is
          an integer in the range 0 <= y[i] < C. This parameter is optional; if it
          is not passed then we only return scores, and if it is passed then we
          instead return the loss and gradients.
        - reg: Regularization strength.

        Returns:
        If y is None, return a matrix scores of shape (N, C) where scores[i, c] is
        the score for class c on input X[i].

        If y is not None, instead return a tuple of:
        - loss: Loss (data loss and regularization loss) for this batch of training
          samples.
        - grads: Dictionary mapping parameter names to gradients of those parameters
          with respect to the loss function; has the same keys as self.params.
        """
        # Unpack variables from the params dictionary
        W1, b1 = self.params['W1'], self.params['b1']
        W2, b2 = self.params['W2'], self.params['b2']
        W3, b3 = self.params['W3'], self.params['b3']
        if self.use_Res == True:
            Wr, br = self.params['Wr'], self.params['br']
        N, D = X.shape

        scores = None
        #############################################################################
        # TODO: Perform the forward pass, computing the class scores for the input. #
        # Store the result in the scores variable, which should be an array of      #
        # shape (N, C).                                                             #
        #############################################################################

        layer1_relu_out, cache1_relu = affine_relu_forward(X, W1, b1)
        layer2_out, cache2 = affine_forward(layer1_relu_out, W2, b2)

        
        if self.use_Res == True:
            layerr_out, cacher = affine_forward(X, Wr, br)
            layer2_out += layerr_out
        
        layer2_relu_out, cache2_relu = relu_forward(layer2_out)
        layer3_out, cache3 = affine_forward(layer2_relu_out, W3, b3)

        scores = layer3_out


        pass
        #############################################################################
        #                              END OF YOUR CODE                             #
        #############################################################################

        # If the targets are not given then jump out, we're done
        if y is None:
            return scores

        # Compute the loss
        loss = None
        #############################################################################
        # TODO: Finish the forward pass, and compute the loss. This should include  #
        # both the data loss and L2 regularization for W1 and W2. Store the result  #
        # in the variable loss, which should be a scalar. Use the Softmax           #
        # classifier loss.                                                          #
        #############################################################################
        data_loss, dout = svm_loss(scores, y)
        reg_loss = 0.5 * reg * (np.sum(W3*W3) + np.sum(W2*W2) + np.sum(W1*W1))


        if self.use_Res == True:
           reg_loss += 0.5 * reg * (np.sum(Wr*Wr))

        loss = data_loss + reg_loss
        pass
        #############################################################################
        #                              END OF YOUR CODE                             #
        #############################################################################

        # Backward pass: compute gradients
        grads = {}

        #############################################################################
        # TODO: Compute the backward pass, computing the derivatives of the weights #
        # and biases. Store the results in the grads dictionary. For example,       #
        # grads['W1'] should store the gradient on W1, and be a matrix of same size #
        #############################################################################
        dlayer2_relu_out, dW3, db3 = affine_backward(dout, cache3)
        dlayer2_out = relu_backward(dlayer2_relu_out, cache2_relu)
        dlayer1_relu_out, dW2, db2 = affine_backward(dlayer2_out, cache2)
        dx, dW1, db1 = affine_relu_backward(dlayer1_relu_out, cache1_relu)

        grads['b1'] = db1
        grads['W1'] = dW1 + 1 * reg * W1
        grads['b2'] = db2
        grads['W2'] = dW2 + 1 * reg * W2
        grads['b3'] = db3
        grads['W3'] = dW3 + 1 * reg * W3

        if self.use_Res == True:
            dx, dWr, dbr = affine_backward(dlayer2_out, cacher)
            grads['Wr'] = dWr
            grads['br'] = dbr

        pass
        #############################################################################
        #                              END OF YOUR CODE                             #
        #############################################################################

        return loss, grads

    def train(self, X, y, X_val, y_val,
              learning_rate=1e-3, learning_rate_decay=0.95,
              reg=5e-6, num_iters=100,
              batch_size=200, verbose=False):
        """
        Train this neural network using stochastic gradient descent.

        Inputs:
        - X: A numpy array of shape (N, D) giving training data.
        - y: A numpy array f shape (N,) giving training labels; y[i] = c means that
          X[i] has label c, where 0 <= c < C.
        - X_val: A numpy array of shape (N_val, D) giving validation data.
        - y_val: A numpy array of shape (N_val,) giving validation labels.
        - learning_rate: Scalar giving learning rate for optimization.
        - learning_rate_decay: Scalar giving factor used to decay the learning rate
          after each epoch.
        - reg: Scalar giving regularization strength.
        - num_iters: Number of steps to take when optimizing.
        - batch_size: Number of training examples to use per step.
        - verbose: boolean; if true print progress during optimization.
        """
        num_train = X.shape[0]
        iterations_per_epoch = max(num_train / batch_size, 1)

        # Use SGD to optimize the parameters in self.model
        loss_history = []
        train_acc_history = []
        val_acc_history = []

        for it in xrange(num_iters):
            X_batch = None
            y_batch = None

            #########################################################################
            # TODO: Create a random minibatch of training data and labels, storing  #
            # them in X_batch and y_batch respectively.                             #
            #########################################################################
            indices = np.random.choice(num_train, batch_size, replace=False)
            X_batch = X[indices]
            y_batch = y[indices]
            pass
            #########################################################################
            #                             END OF YOUR CODE                          #
            #########################################################################

            # Compute loss and gradients using the current minibatch
            loss, grads = self.loss(X_batch, y=y_batch, reg=reg)
            loss_history.append(loss)

            #########################################################################
            # TODO: Use the gradients in the grads dictionary to update the         #
            # parameters of the network (stored in the dictionary self.params)      #
            # using stochastic gradient descent. You'll need to use the gradients   #
            # stored in the grads dictionary defined above.                         #
            #########################################################################
            self.params['W1'] += -grads['W1'] * learning_rate
            self.params['W2'] += -grads['W2'] * learning_rate
            self.params['W3'] += -grads['W3'] * learning_rate
            self.params['b1'] += -grads['b1'] * learning_rate
            self.params['b2'] += -grads['b2'] * learning_rate
            self.params['b3'] += -grads['b3'] * learning_rate
            if self.use_Res == True:
                self.params['Wr'] += -grads['Wr'] * learning_rate
                self.params['br'] += -grads['br'] * learning_rate
            pass
            #########################################################################
            #                             END OF YOUR CODE                          #
            #########################################################################

            if verbose and it % 100 == 0:
                print('iteration %d / %d: loss %f val_acc %f train_acc %f' % (it, 
                    num_iters, loss, (self.predict(X_val) == y_val).mean(), 
                    (self.predict(X) == y).mean()))

            # Every epoch, check train and val accuracy and decay learning rate.
            if it % iterations_per_epoch == 0:
                # Check accuracy
                train_acc = (self.predict(X_batch) == y_batch).mean()
                val_acc = (self.predict(X_val) == y_val).mean()
                train_acc_history.append(train_acc)
                val_acc_history.append(val_acc)

                # Decay learning rate
                learning_rate *= learning_rate_decay

        return {
            'loss_history': loss_history,
            'train_acc_history': train_acc_history,
            'val_acc_history': val_acc_history,
        }

    def predict(self, X):
        """
        Use the trained weights of this two-layer network to predict labels for
        data points. For each data point we predict scores for each of the C
        classes, and assign each data point to the class with the highest score.

        Inputs:
        - X: A numpy array of shape (N, D) giving N D-dimensional data points to
          classify.

        Returns:
        - y_pred: A numpy array of shape (N,) giving predicted labels for each of
          the elements of X. For all i, y_pred[i] = c means that X[i] is predicted
          to have class c, where 0 <= c < C.
        """
        y_pred = None

        ###########################################################################
        # TODO: Implement this function; it should be VERY simple!                #
        ###########################################################################
        W1, b1 = self.params['W1'], self.params['b1']
        W2, b2 = self.params['W2'], self.params['b2']
        W3, b3 = self.params['W3'], self.params['b3']
        if self.use_Res == True:
            Wr, br = self.params['Wr'], self.params['br']
        scores = None
        #############################################################################
        # TODO: Perform the forward pass, computing the class scores for the input. #
        # Store the result in the scores variable, which should be an array of      #
        # shape (N, C).                                                             #
        #############################################################################
        layer1_relu_out, _ = affine_relu_forward(X, W1, b1)
        layer2_out, _ = affine_forward(layer1_relu_out, W2, b2)

        
        if self.use_Res == True:
            layerr_out, _ = affine_forward(X, Wr, br)
            layer2_out += layerr_out
        
        layer2_relu_out, _ = relu_forward(layer2_out)
        layer3_out, _ = affine_forward(layer2_relu_out, W3, b3)

        scores = layer3_out

        y_pred = np.argmax(scores, axis=1)
        pass
        ###########################################################################
        #                              END OF YOUR CODE                           #
        ###########################################################################

        return y_pred