Implementing a Neural Network

(Adapted from Stanford University's CS231n Open Courseware)

Complete and hand in this completed worksheet (including its outputs and any supporting code outside of the worksheet) with your assignment submission. For more details see the HW page on the course website.

In this exercise we will develop a neural network with fully-connected layers to perform classification, and test it out on the CIFAR-10 dataset.


In [1]:
# A bit of setup

import numpy as np
import matplotlib.pyplot as plt

from cs231n.classifiers.neural_net import TwoLayerNet

%matplotlib inline
plt.rcParams['figure.figsize'] = (10.0, 8.0) # set default size of plots
plt.rcParams['image.interpolation'] = 'nearest'
plt.rcParams['image.cmap'] = 'gray'

# for auto-reloading external modules
# see http://stackoverflow.com/questions/1907993/autoreload-of-modules-in-ipython
%load_ext autoreload
%autoreload 2

def rel_error(x, y):
  """ returns relative error """
  return np.max(np.abs(x - y) / (np.maximum(1e-8, np.abs(x) + np.abs(y))))

We will use the class TwoLayerNet in the file cs231n/classifiers/neural_net.py to represent instances of our network. The network parameters are stored in the instance variable self.params where keys are string parameter names and values are numpy arrays. Below, we initialize toy data and a toy model that we will use to develop your implementation.


In [2]:
# Create a small net and some toy data to check your implementations.
# Note that we set the random seed for repeatable experiments.

input_size = 4
hidden_size = 10
num_classes = 3
num_inputs = 5

def init_toy_model():
  np.random.seed(0)
  return TwoLayerNet(input_size, hidden_size, num_classes, std=1e-1)

def init_toy_data():
  np.random.seed(1)
  X = 10 * np.random.randn(num_inputs, input_size)
  y = np.array([0, 1, 2, 2, 1])
  return X, y

net = init_toy_model()
X, y = init_toy_data()

Forward pass: compute scores

Open the file cs231n/classifiers/neural_net.py and look at the method TwoLayerNet.loss. This function is very similar to the loss functions you have written for the SVM and Softmax exercises: It takes the data and weights and computes the class scores, the loss, and the gradients on the parameters.

Implement the first part of the forward pass which uses the weights and biases to compute the scores for all inputs.


In [3]:
scores = net.loss(X)
print 'Your scores:'
print scores
print
print 'correct scores:'
correct_scores = np.asarray([
  [-0.81233741, -1.27654624, -0.70335995],
  [-0.17129677, -1.18803311, -0.47310444],
  [-0.51590475, -1.01354314, -0.8504215 ],
  [-0.15419291, -0.48629638, -0.52901952],
  [-0.00618733, -0.12435261, -0.15226949]])
print correct_scores
print

# The difference should be very small. We get < 1e-7
print 'Difference between your scores and correct scores:'
print np.sum(np.abs(scores - correct_scores))


Your scores:
[[-0.81233741 -1.27654624 -0.70335995]
 [-0.17129677 -1.18803311 -0.47310444]
 [-0.51590475 -1.01354314 -0.8504215 ]
 [-0.15419291 -0.48629638 -0.52901952]
 [-0.00618733 -0.12435261 -0.15226949]]

correct scores:
[[-0.81233741 -1.27654624 -0.70335995]
 [-0.17129677 -1.18803311 -0.47310444]
 [-0.51590475 -1.01354314 -0.8504215 ]
 [-0.15419291 -0.48629638 -0.52901952]
 [-0.00618733 -0.12435261 -0.15226949]]

Difference between your scores and correct scores:
3.68027209324e-08

Forward pass: compute loss

In the same function, implement the second part that computes the data and regularizaion loss.


In [4]:
loss, _ = net.loss(X, y, reg=0.1)
correct_loss = 1.30378789133

# should be very small, we get < 1e-12
print 'Difference between your loss and correct loss:'
print np.sum(np.abs(loss - correct_loss))


Difference between your loss and correct loss:
1.79412040779e-13

Backward pass

Implement the rest of the function. This will compute the gradient of the loss with respect to the variables W1, b1, W2, and b2. Now that you (hopefully!) have a correctly implemented forward pass, you can debug your backward pass using a numeric gradient check:


In [5]:
from cs231n.gradient_check import eval_numerical_gradient

# Use numeric gradient checking to check your implementation of the backward pass.
# If your implementation is correct, the difference between the numeric and
# analytic gradients should be less than 1e-8 for each of W1, W2, b1, and b2.

loss, grads = net.loss(X, y, reg=0.1)

# these should all be less than 1e-8 or so
for param_name in grads:
  f = lambda W: net.loss(X, y, reg=0.1)[0]
  param_grad_num = eval_numerical_gradient(f, net.params[param_name])
  print '%s max relative error: %e' % (param_name, rel_error(param_grad_num, grads[param_name]))


(0, 0) -0.0968491846498
(0, 1) 0.00400207207374
(0, 2) -0.205514233587
(0, 3) 0.187987696121
(0, 4) 0.160532817883
(0, 5) -0.00977227880838
(0, 6) 0.00950138416833
(0, 7) 0.00268935298475
(0, 8) -0.030101163806
(0, 9) -0.00567656861516
(1, 0) 0.0445600319088
(1, 1) 0.0145432350607
(1, 2) 0.69552028592
(1, 3) -0.28861463619
(1, 4) -0.266985547204
(1, 5) 0.00333724328172
(1, 6) 0.014941290738
(1, 7) 0.0193443064234
(1, 8) -0.0654659517263
(1, 9) -0.53292646649
(2, 0) 0.01169830528
(2, 1) 0.00653668594808
(2, 2) -0.231617332491
(2, 3) -0.0626370572965
(2, 4) -0.00141470395398
(2, 5) -0.0145431567455
(2, 6) 0.000458085169797
(2, 7) -0.00290017021509
(2, 8) 0.53567254179
(2, 9) 0.369734381578
(3, 0) 0.0771772360686
(3, 1) 0.00378212519081
(3, 2) -0.313771155547
(3, 3) 0.226870743791
(3, 4) 0.206680562354
(3, 5) 0.00156398969331
(3, 6) 0.0123034068045
(3, 7) -0.00217994859941
(3, 8) -0.678938963583
(3, 9) 0.0985600677739
W1 max relative error: 5.954442e-04
(0, 0) -0.513760559562
(0, 1) 0.167236198534
(0, 2) 0.304787632488
(1, 0) 0.0195082539456
(1, 1) -0.00509602182674
(1, 2) -0.00438024301275
(2, 0) 0.29223366631
(2, 1) 0.118900468582
(2, 2) -0.432014014229
(3, 0) -0.23310773134
(3, 1) 0.186298418869
(3, 2) 0.0396330094299
(4, 0) 0.778188126538
(4, 1) -0.354223607624
(4, 2) -0.441125933515
(5, 0) 0.00428381869977
(5, 1) 0.00066567222845
(5, 2) 0.00302521898643
(6, 0) -0.00634272094846
(6, 1) -0.00362691165989
(6, 2) -0.00672410447322
(7, 0) 0.0617288478555
(7, 1) -0.129899812507
(7, 2) 0.0391831819213
(8, 0) 0.63205875811
(8, 1) -0.0843986286858
(8, 2) -0.56619285318
(9, 0) 0.889345768362
(9, 1) -0.604702071416
(9, 2) -0.288541635407
W2 max relative error: 3.757048e-04
(0,) 0.209970298459
(1,) -0.14318955901
(2,) -0.0667774968655
b2 max relative error: 8.250815e-06
(0,) -0.00704840343868
(1,) 0.0
(2,) 0.00310497081379
(3,) -0.00723988951012
(4,) -0.00573376650603
(5,) 0.0
(6,) 0.0
(7,) -0.00243719404658
(8,) 0.0412160637397
(9,) 0.0223617785178
b1 max relative error: 5.723668e-06

Train the network

To train the network we will use stochastic gradient descent (SGD), similar to the SVM and Softmax classifiers. Look at the function TwoLayerNet.train and fill in the missing sections to implement the training procedure. This should be very similar to the training procedure you used for the SVM and Softmax classifiers. You will also have to implement TwoLayerNet.predict, as the training process periodically performs prediction to keep track of accuracy over time while the network trains.

Once you have implemented the method, run the code below to train a two-layer network on toy data. You should achieve a training loss less than 0.2.


In [6]:
net = init_toy_model()
stats = net.train(X, y, X, y,
            learning_rate=1e-1, reg=1e-5,
            num_iters=100, verbose=False)

print 'Final training loss: ', stats['loss_history'][-1]

# plot the loss history
plt.plot(stats['loss_history'])
plt.xlabel('iteration')
plt.ylabel('training loss')
plt.title('Training Loss history')
plt.show()


Final training loss:  0.0171496079387

Load the data

Now that you have implemented a two-layer network that passes gradient checks and works on toy data, it's time to load up our favorite CIFAR-10 data so we can use it to train a classifier on a real dataset.


In [7]:
from cs231n.data_utils import load_CIFAR10

def get_CIFAR10_data(num_training=49000, num_validation=1000, num_test=1000):
    """
    Load the CIFAR-10 dataset from disk and perform preprocessing to prepare
    it for the two-layer neural net classifier. These are the same steps as
    we used for the SVM, but condensed to a single function.  
    """
    # Load the raw CIFAR-10 data
    cifar10_dir = 'cs231n/datasets/cifar-10-batches-py'
    X_train, y_train, X_test, y_test = load_CIFAR10(cifar10_dir)
        
    # Subsample the data
    mask = range(num_training, num_training + num_validation)
    X_val = X_train[mask]
    y_val = y_train[mask]
    mask = range(num_training)
    X_train = X_train[mask]
    y_train = y_train[mask]
    mask = range(num_test)
    X_test = X_test[mask]
    y_test = y_test[mask]

    # Normalize the data: subtract the mean image
    mean_image = np.mean(X_train, axis=0)
    X_train -= mean_image
    X_val -= mean_image
    X_test -= mean_image

    # Reshape data to rows
    X_train = X_train.reshape(num_training, -1)
    X_val = X_val.reshape(num_validation, -1)
    X_test = X_test.reshape(num_test, -1)

    return X_train, y_train, X_val, y_val, X_test, y_test


# Invoke the above function to get our data.
X_train, y_train, X_val, y_val, X_test, y_test = get_CIFAR10_data()
print 'Train data shape: ', X_train.shape
print 'Train labels shape: ', y_train.shape
print 'Validation data shape: ', X_val.shape
print 'Validation labels shape: ', y_val.shape
print 'Test data shape: ', X_test.shape
print 'Test labels shape: ', y_test.shape


Train data shape:  (49000, 3072)
Train labels shape:  (49000,)
Validation data shape:  (1000, 3072)
Validation labels shape:  (1000,)
Test data shape:  (1000, 3072)
Test labels shape:  (1000,)

Train a network

To train our network we will use SGD with momentum. In addition, we will adjust the learning rate with an exponential learning rate schedule as optimization proceeds; after each epoch, we will reduce the learning rate by multiplying it by a decay rate.


In [8]:
input_size = 32 * 32 * 3
hidden_size = 50
num_classes = 10
net = TwoLayerNet(input_size, hidden_size, num_classes)

# Train the network
stats = net.train(X_train, y_train, X_val, y_val,
            num_iters=1000, batch_size=200,
            learning_rate=1e-4, learning_rate_decay=0.95,
            reg=0.5, verbose=True)

# Predict on the validation set
val_acc = (net.predict(X_val) == y_val).mean()
print 'Validation accuracy: ', val_acc


iteration 0 / 1000: loss 2.302954
iteration 100 / 1000: loss 2.302550
iteration 200 / 1000: loss 2.297648
iteration 300 / 1000: loss 2.259602
iteration 400 / 1000: loss 2.204170
iteration 500 / 1000: loss 2.118565
iteration 600 / 1000: loss 2.051535
iteration 700 / 1000: loss 1.988466
iteration 800 / 1000: loss 2.006591
iteration 900 / 1000: loss 1.951473
Validation accuracy:  0.287

Debug the training

With the default parameters we provided above, you should get a validation accuracy of about 0.29 on the validation set. This isn't very good.

One strategy for getting insight into what's wrong is to plot the loss function and the accuracies on the training and validation sets during optimization.

Another strategy is to visualize the weights that were learned in the first layer of the network. In most neural networks trained on visual data, the first layer weights typically show some visible structure when visualized.


In [9]:
# Plot the loss function and train / validation accuracies
plt.subplot(2, 1, 1)
plt.plot(stats['loss_history'])
plt.title('Loss history')
plt.xlabel('Iteration')
plt.ylabel('Loss')

plt.subplot(2, 1, 2)
plt.plot(stats['train_acc_history'], label='train')
plt.plot(stats['val_acc_history'], label='val')
plt.title('Classification accuracy history')
plt.xlabel('Epoch')
plt.ylabel('Clasification accuracy')
plt.show()



In [10]:
from cs231n.vis_utils import visualize_grid

# Visualize the weights of the network

def show_net_weights(net):
  W1 = net.params['W1']
  W1 = W1.reshape(32, 32, 3, -1).transpose(3, 0, 1, 2)
  plt.imshow(visualize_grid(W1, padding=3).astype('uint8'))
  plt.gca().axis('off')
  plt.show()

show_net_weights(net)


Tune your hyperparameters

What's wrong?. Looking at the visualizations above, we see that the loss is decreasing more or less linearly, which seems to suggest that the learning rate may be too low. Moreover, there is no gap between the training and validation accuracy, suggesting that the model we used has low capacity, and that we should increase its size. On the other hand, with a very large model we would expect to see more overfitting, which would manifest itself as a very large gap between the training and validation accuracy.

Tuning. Tuning the hyperparameters and developing intuition for how they affect the final performance is a large part of using Neural Networks, so we want you to get a lot of practice. Below, you should experiment with different values of the various hyperparameters, including hidden layer size, learning rate, numer of training epochs, and regularization strength. You might also consider tuning the learning rate decay, but you should be able to get good performance using the default value.

Approximate results. You should be aim to achieve a classification accuracy of greater than 48% on the validation set. Our best network gets over 52% on the validation set.

Experiment: You goal in this exercise is to get as good of a result on CIFAR-10 as you can, with a fully-connected Neural Network. For every 1% above 52% on the Test set we will award you with one extra bonus point. Feel free implement your own techniques (e.g. PCA to reduce dimensionality, or adding dropout, or adding features to the solver, etc.).


In [15]:
best_net = None # store the best model into this 

#################################################################################
# TODO: Tune hyperparameters using the validation set. Store your best trained  #
# model in best_net.                                                            #
#                                                                               #
# To help debug your network, it may help to use visualizations similar to the  #
# ones we used above; these visualizations will have significant qualitative    #
# differences from the ones we saw above for the poorly tuned network.          #
#                                                                               #
# Tweaking hyperparameters by hand can be fun, but you might find it useful to  #
# write code to sweep through possible combinations of hyperparameters          #
# automatically like we did on the previous exercises.                          #
#################################################################################
lr_arr = [5e-5, 1e-4, 5e-4]
lr = 1e-4
hd_arr = [200, 500]
best_val = -1
    
for idx, hd_sz in enumerate(hd_arr):
    print 'Learning rate is: %f. Hidden layer size is: %d' % (lr, hd_sz)
    input_size = 32 * 32 * 3
    net = TwoLayerNet(input_size, hd_sz, 10)

    # Train the network
    stats = net.train(X_train, y_train, X_val, y_val,
            num_iters=1000, batch_size=200,
            learning_rate=lr, learning_rate_decay=0.95,
            reg=0.5, verbose=True)

    # Predict on the validation set
    val_acc = (net.predict(X_val) == y_val).mean()
    print 'Validation accuracy: ', val_acc
    
    # Plot the loss function and train / validation accuracies
    plt.subplot(2, 3, 1 + 2*idx)
    plt.plot(stats['loss_history'])
    plt.title('Loss history')
    plt.xlabel('Iteration')
    plt.ylabel('Loss')

    plt.subplot(2, 3, 2 + 2*idx)
    plt.plot(stats['train_acc_history'])
    plt.plot(val_acc)
    plt.legend(['Training accuracy', 'Validation accuracy'], loc='lower right')
    plt.xlabel('Epoch')
    plt.ylabel('Clasification accuracy')
    
    if val_acc > best_val:
        best_val = val_acc
        best_net = net
#################################################################################
#                               END OF YOUR CODE                                #
#################################################################################


Learning rate is: 0.000100. Hidden layer size is: 200
iteration 0 / 1000: loss 2.304095
iteration 100 / 1000: loss 2.302272
iteration 200 / 1000: loss 2.288699
iteration 300 / 1000: loss 2.232743
iteration 400 / 1000: loss 2.106805
iteration 500 / 1000: loss 2.132397
iteration 600 / 1000: loss 2.066734
iteration 700 / 1000: loss 2.083656
iteration 800 / 1000: loss 1.903089
iteration 900 / 1000: loss 1.987995
Validation accuracy:  0.309
Learning rate is: 0.000100. Hidden layer size is: 500
iteration 0 / 1000: loss 2.306414
iteration 100 / 1000: loss 2.302076
iteration 200 / 1000: loss 2.273258
iteration 300 / 1000: loss 2.228679
iteration 400 / 1000: loss 2.117304
iteration 500 / 1000: loss 1.998819
iteration 600 / 1000: loss 1.980998
iteration 700 / 1000: loss 2.100760
iteration 800 / 1000: loss 1.876847
iteration 900 / 1000: loss 1.893668
Validation accuracy:  0.318

In [16]:
# visualize the weights of the best network
show_net_weights(best_net)


Run on the test set

When you are done experimenting, you should evaluate your final trained network on the test set; you should get above 48%.

We will give you extra bonus point for every 1% of accuracy above 52%.


In [17]:
test_acc = (best_net.predict(X_test) == y_test).mean()
print 'Test accuracy: ', test_acc


Test accuracy:  0.324

In [ ]: