Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Initialization)

时间:2023-03-08 16:55:12
Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Initialization)

声明:所有内容来自coursera,作为个人学习笔记记录在这里.

Initialization

Welcome to the first assignment of "Improving Deep Neural Networks".

Training your neural network requires specifying an initial value of the weights. A well chosen initialization method will help learning.

If you completed the previous course of this specialization, you probably followed our instructions for weight initialization, and it has worked out so far. But how do you choose the initialization for a new neural network? In this notebook, you will see how different initializations lead to different results.

A well chosen initialization can:

  • Speed up the convergence of gradient descent
  • Increase the odds of gradient descent converging to a lower training (and generalization) error

To get started, run the following cell to load the packages and the planar dataset you will try to classify.

In [3]:
Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Initialization)
import numpy as np
import matplotlib.pyplot as plt
import sklearn
import sklearn.datasets
from init_utils import sigmoid, relu, compute_loss, forward_propagation, backward_propagation
from init_utils import update_parameters, predict, load_dataset, plot_decision_boundary, predict_dec %matplotlib inline
plt.rcParams['figure.figsize'] = (7.0, 4.0) # set default size of plots
plt.rcParams['image.interpolation'] = 'nearest'
plt.rcParams['image.cmap'] = 'gray' # load image dataset: blue/red dots in circles
train_X, train_Y, test_X, test_Y = load_dataset()
Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Initialization)
aaarticlea/png;base64,iVBORw0KGgoAAAANSUhEUgAAAbcAAAD8CAYAAAD0f+rwAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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" alt="" />

You would like a classifier to separate the blue dots from the red dots.

1 - Neural Network model

You will use a 3-layer neural network (already implemented for you). Here are the initialization methods you will experiment with:

  • Zeros initialization -- setting initialization = "zeros" in the input argument.
  • Random initialization -- setting initialization = "random" in the input argument. This initializes the weights to large random values.
  • He initialization -- setting initialization = "he" in the input argument. This initializes the weights to random values scaled according to a paper by He et al., 2015.

Instructions: Please quickly read over the code below, and run it. In the next part you will implement the three initialization methods that this model() calls.

In [13]:
Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Initialization)
def model(X, Y, learning_rate = 0.01, num_iterations = 15000, print_cost = True, initialization = "he"):
"""
Implements a three-layer neural network: LINEAR->RELU->LINEAR->RELU->LINEAR->SIGMOID. Arguments:
X -- input data, of shape (2, number of examples)
Y -- true "label" vector (containing 0 for red dots; 1 for blue dots), of shape (1, number of examples)
learning_rate -- learning rate for gradient descent
num_iterations -- number of iterations to run gradient descent
print_cost -- if True, print the cost every 1000 iterations
initialization -- flag to choose which initialization to use ("zeros","random" or "he") Returns:
parameters -- parameters learnt by the model
""" grads = {}
costs = [] # to keep track of the loss
m = X.shape[1] # number of examples
layers_dims = [X.shape[0], 10, 5, 1] # Initialize parameters dictionary.
if initialization == "zeros":
parameters = initialize_parameters_zeros(layers_dims)
elif initialization == "random":
parameters = initialize_parameters_random(layers_dims)
elif initialization == "he":
parameters = initialize_parameters_he(layers_dims) # Loop (gradient descent) for i in range(0, num_iterations): # Forward propagation: LINEAR -> RELU -> LINEAR -> RELU -> LINEAR -> SIGMOID.
a3, cache = forward_propagation(X, parameters) # Loss
cost = compute_loss(a3, Y) # Backward propagation.
grads = backward_propagation(X, Y, cache) # Update parameters.
parameters = update_parameters(parameters, grads, learning_rate) # Print the loss every 1000 iterations
if print_cost and i % 1000 == 0:
print("Cost after iteration {}: {}".format(i, cost))
costs.append(cost) # plot the loss
plt.plot(costs)
plt.ylabel('cost')
plt.xlabel('iterations (per hundreds)')
plt.title("Learning rate =" + str(learning_rate))
plt.show() return parameters
Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Initialization)

2 - Zero initialization

There are two types of parameters to initialize in a neural network:

  • the weight matrices (W[1],W[2],W[3],...,W[L−1],W[L])(W[1],W[2],W[3],...,W[L−1],W[L])
  • the bias vectors (b[1],b[2],b[3],...,b[L−1],b[L])(b[1],b[2],b[3],...,b[L−1],b[L])

Exercise: Implement the following function to initialize all parameters to zeros. You'll see later that this does not work well since it fails to "break symmetry", but lets try it anyway and see what happens. Use np.zeros((..,..)) with the correct shapes.

In [14]:
Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Initialization)
# GRADED FUNCTION: initialize_parameters_zeros 

def initialize_parameters_zeros(layers_dims):
"""
Arguments:
layer_dims -- python array (list) containing the size of each layer. Returns:
parameters -- python dictionary containing your parameters "W1", "b1", ..., "WL", "bL":
W1 -- weight matrix of shape (layers_dims[1], layers_dims[0])
b1 -- bias vector of shape (layers_dims[1], 1)
...
WL -- weight matrix of shape (layers_dims[L], layers_dims[L-1])
bL -- bias vector of shape (layers_dims[L], 1)
""" parameters = {}
L = len(layers_dims) # number of layers in the network for l in range(1, L):
### START CODE HERE ### (≈ 2 lines of code)
parameters['W' + str(l)] = np.zeros((layers_dims[l], layers_dims[l-1]))
parameters['b' + str(l)] = np.zeros((layers_dims[l], 1))
### END CODE HERE ###
return parameters
Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Initialization)
In [15]:
parameters = initialize_parameters_zeros([3,2,1])
print("W1 = " + str(parameters["W1"]))
print("b1 = " + str(parameters["b1"]))
print("W2 = " + str(parameters["W2"]))
print("b2 = " + str(parameters["b2"]))
 
W1 = [[ 0.  0.  0.]
[ 0. 0. 0.]]
b1 = [[ 0.]
[ 0.]]
W2 = [[ 0. 0.]]
b2 = [[ 0.]]

Expected Output:

**W1** [[ 0. 0. 0.] [ 0. 0. 0.]]
**b1** [[ 0.] [ 0.]]
**W2** [[ 0. 0.]]
**b2** [[ 0.]]

Run the following code to train your model on 15,000 iterations using zeros initialization.

In [16]:
Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Initialization)
parameters = model(train_X, train_Y, initialization = "zeros")
print ("On the train set:")
predictions_train = predict(train_X, train_Y, parameters)
print ("On the test set:")
predictions_test = predict(test_X, test_Y, parameters)
Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Initialization)
Cost after iteration 0: 0.6931471805599453
Cost after iteration 1000: 0.6931471805599453
Cost after iteration 2000: 0.6931471805599453
Cost after iteration 3000: 0.6931471805599453
Cost after iteration 4000: 0.6931471805599453
Cost after iteration 5000: 0.6931471805599453
Cost after iteration 6000: 0.6931471805599453
Cost after iteration 7000: 0.6931471805599453
Cost after iteration 8000: 0.6931471805599453
Cost after iteration 9000: 0.6931471805599453
Cost after iteration 10000: 0.6931471805599455
Cost after iteration 11000: 0.6931471805599453
Cost after iteration 12000: 0.6931471805599453
Cost after iteration 13000: 0.6931471805599453
Cost after iteration 14000: 0.6931471805599453
aaarticlea/png;base64,iVBORw0KGgoAAAANSUhEUgAAAcMAAAEWCAYAAAAadfxCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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%0AERGNlzCMiIjG+/8qs5fH0fJiOQAAAABJRU5ErkJggg==" alt="" />
On the train set:
Accuracy: 0.5
On the test set:
Accuracy: 0.5

The performance is really bad, and the cost does not really decrease, and the algorithm performs no better than random guessing. Why? Lets look at the details of the predictions and the decision boundary:

In [17]:
print ("predictions_train = " + str(predictions_train))
print ("predictions_test = " + str(predictions_test))
predictions_train = [[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0]]
predictions_test = [[0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0]]
In [18]:
plt.title("Model with Zeros initialization")
axes = plt.gca()
axes.set_xlim([-1.5,1.5])
axes.set_ylim([-1.5,1.5])
plot_decision_boundary(lambda x: predict_dec(parameters, x.T), train_X, train_Y)
aaarticlea/png;base64,iVBORw0KGgoAAAANSUhEUgAAAcoAAAEWCAYAAADmYNeIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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%0AkTz+G43ycvD61JRrRDiEhUN+fmAAk*kd+9ojR7nw3r1fV1YyKar2ArsX5z/MvCvbsbYpm1XVJ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%0AC363uAqGAFWIAAAAAElFTkSuQmCC" alt="" />

The model is predicting 0 for every example.

In general, initializing all the weights to zero results in the network failing to break symmetry. This means that every neuron in each layer will learn the same thing, and you might as well be training a neural network with n[l]=1n[l]=1 for every layer, and the network is no more powerful than a linear classifier such as logistic regression.

What you should remember:

  • The weights W[l]W[l] should be initialized randomly to break symmetry.
  • It is however okay to initialize the biases b[l]b[l] to zeros. Symmetry is still broken so long as W[l]W[l] is initialized randomly.

3 - Random initialization

To break symmetry, lets intialize the weights randomly. Following random initialization, each neuron can then proceed to learn a different function of its inputs. In this exercise, you will see what happens if the weights are intialized randomly, but to very large values.

Exercise: Implement the following function to initialize your weights to large random values (scaled by *10) and your biases to zeros. Use np.random.randn(..,..) * 10 for weights and np.zeros((.., ..)) for biases. We are using a fixed np.random.seed(..) to make sure your "random" weights match ours, so don't worry if running several times your code gives you always the same initial values for the parameters.

In [19]:
Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Initialization)
# GRADED FUNCTION: initialize_parameters_random

def initialize_parameters_random(layers_dims):
"""
Arguments:
layer_dims -- python array (list) containing the size of each layer. Returns:
parameters -- python dictionary containing your parameters "W1", "b1", ..., "WL", "bL":
W1 -- weight matrix of shape (layers_dims[1], layers_dims[0])
b1 -- bias vector of shape (layers_dims[1], 1)
...
WL -- weight matrix of shape (layers_dims[L], layers_dims[L-1])
bL -- bias vector of shape (layers_dims[L], 1)
""" np.random.seed(3) # This seed makes sure your "random" numbers will be the as ours
parameters = {}
L = len(layers_dims) # integer representing the number of layers for l in range(1, L):
### START CODE HERE ### (≈ 2 lines of code)
parameters['W' + str(l)] = np.random.randn(layers_dims[l], layers_dims[l-1]) * 10
parameters['b' + str(l)] = np.zeros((layers_dims[l], 1))
### END CODE HERE ### return parameters
Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Initialization)
In [20]:
parameters = initialize_parameters_random([3, 2, 1])
print("W1 = " + str(parameters["W1"]))
print("b1 = " + str(parameters["b1"]))
print("W2 = " + str(parameters["W2"]))
print("b2 = " + str(parameters["b2"]))
W1 = [[ 17.88628473   4.36509851   0.96497468]
[-18.63492703 -2.77388203 -3.54758979]]
b1 = [[ 0.]
[ 0.]]
W2 = [[-0.82741481 -6.27000677]]
b2 = [[ 0.]]

Expected Output:

**W1** [[ 17.88628473 4.36509851 0.96497468] [-18.63492703 -2.77388203 -3.54758979]]
**b1** [[ 0.] [ 0.]]
**W2** [[-0.82741481 -6.27000677]]
**b2** [[ 0.]]

Run the following code to train your model on 15,000 iterations using random initialization.

In [21]:
parameters = model(train_X, train_Y, initialization = "random")
print ("On the train set:")
predictions_train = predict(train_X, train_Y, parameters)
print ("On the test set:")
predictions_test = predict(test_X, test_Y, parameters)
Cost after iteration 0: inf
Cost after iteration 1000: 0.6237287551108738
Cost after iteration 2000: 0.5981106708339466
Cost after iteration 3000: 0.5638353726276827
Cost after iteration 4000: 0.550152614449184
Cost after iteration 5000: 0.5444235275228304
Cost after iteration 6000: 0.5374184054630083
Cost after iteration 7000: 0.47357131493578297
Cost after iteration 8000: 0.39775634899580387
Cost after iteration 9000: 0.3934632865981078
Cost after iteration 10000: 0.39202525076484457
Cost after iteration 11000: 0.38921493051297673
Cost after iteration 12000: 0.38614221789840486
Cost after iteration 13000: 0.38497849983013926
Cost after iteration 14000: 0.38278397192120406
aaarticlea/png;base64,iVBORw0KGgoAAAANSUhEUgAAAcMAAAEWCAYAAAAadfxCAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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%0ARU5ErkJggg==" alt="" />
On the train set:
Accuracy: 0.83
On the test set:
Accuracy: 0.86

If you see "inf" as the cost after the iteration 0, this is because of numerical roundoff; a more numerically sophisticated implementation would fix this. But this isn't worth worrying about for our purposes.

Anyway, it looks like you have broken symmetry, and this gives better results. than before. The model is no longer outputting all 0s.

In [22]:
print (predictions_train)
print (predictions_test)
[[1 0 1 1 0 0 1 1 1 1 1 0 1 0 0 1 0 1 1 0 0 0 1 0 1 1 1 1 1 1 0 1 1 0 0 1 1
1 1 1 1 1 1 0 1 1 1 1 0 1 0 1 1 1 1 0 0 1 1 1 1 0 1 1 0 1 0 1 1 1 1 0 0 0
0 0 1 0 1 0 1 1 1 0 0 1 1 1 1 1 1 0 0 1 1 1 0 1 1 0 1 0 1 1 0 1 1 0 1 0 1
1 0 0 1 0 0 1 1 0 1 1 1 0 1 0 0 1 0 1 1 1 1 1 1 1 0 1 1 0 0 1 1 0 0 0 1 0
1 0 1 0 1 1 1 0 0 1 1 1 1 0 1 1 0 1 0 1 1 0 1 0 1 1 1 1 0 1 1 1 1 0 1 0 1
0 1 1 1 1 0 1 1 0 1 1 0 1 1 0 1 0 1 1 1 0 1 1 1 0 1 0 1 0 0 1 0 1 1 0 1 1
0 1 1 0 1 1 1 0 1 1 1 1 0 1 0 0 1 1 0 1 1 1 0 0 0 1 1 0 1 1 1 1 0 1 1 0 1
1 1 0 0 1 0 0 0 1 0 0 0 1 1 1 1 0 0 0 0 1 1 1 1 0 0 1 1 1 1 1 1 1 0 0 0 1
1 1 1 0]]
[[1 1 1 1 0 1 0 1 1 0 1 1 1 0 0 0 0 1 0 1 0 0 1 0 1 0 1 1 1 1 1 0 0 0 0 1 0
1 1 0 0 1 1 1 1 1 0 1 1 1 0 1 0 1 1 0 1 0 1 0 1 1 1 1 1 1 1 1 1 0 1 0 1 1
1 1 1 0 1 0 0 1 0 0 0 1 1 0 1 1 0 0 0 1 1 0 1 1 0 0]]
In [23]:
plt.title("Model with large random initialization")
axes = plt.gca()
axes.set_xlim([-1.5,1.5])
axes.set_ylim([-1.5,1.5])
plot_decision_boundary(lambda x: predict_dec(parameters, x.T), train_X, train_Y)
aaarticlea/png;base64,iVBORw0KGgoAAAANSUhEUgAAAcoAAAEWCAYAAADmYNeIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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" alt="" />

Observations:

  • The cost starts very high. This is because with large random-valued weights, the last activation (sigmoid) outputs results that are very close to 0 or 1 for some examples, and when it gets that example wrong it incurs a very high loss for that example. Indeed, when log(a[3])=log(0)log⁡(a[3])=log⁡(0), the loss goes to infinity.
  • Poor initialization can lead to vanishing/exploding gradients, which also slows down the optimization algorithm.
  • If you train this network longer you will see better results, but initializing with overly large random numbers slows down the optimization.

In summary:

  • Initializing weights to very large random values does not work well.
  • Hopefully intializing with small random values does better. The important question is: how small should be these random values be? Lets find out in the next part!

4 - He initialization

Finally, try "He Initialization"; this is named for the first author of He et al., 2015. (If you have heard of "Xavier initialization", this is similar except Xavier initialization uses a scaling factor for the weights W[l]W[l] of sqrt(1./layers_dims[l-1]) where He initialization would use sqrt(2./layers_dims[l-1]).)

Exercise: Implement the following function to initialize your parameters with He initialization.

Hint: This function is similar to the previous initialize_parameters_random(...). The only difference is that instead of multiplying np.random.randn(..,..) by 10, you will multiply it by 2dimension of the previous layer−−−−−−−−−−−−−−−−−−√2dimension of the previous layer, which is what He initialization recommends for layers with a ReLU activation.

In [26]:
Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Initialization)
# GRADED FUNCTION: initialize_parameters_he

def initialize_parameters_he(layers_dims):
"""
Arguments:
layer_dims -- python array (list) containing the size of each layer. Returns:
parameters -- python dictionary containing your parameters "W1", "b1", ..., "WL", "bL":
W1 -- weight matrix of shape (layers_dims[1], layers_dims[0])
b1 -- bias vector of shape (layers_dims[1], 1)
...
WL -- weight matrix of shape (layers_dims[L], layers_dims[L-1])
bL -- bias vector of shape (layers_dims[L], 1)
""" np.random.seed(3)
parameters = {}
L = len(layers_dims) - 1 # integer representing the number of layers for l in range(1, L + 1):
### START CODE HERE ### (≈ 2 lines of code)
parameters['W' + str(l)] = np.random.randn(layers_dims[l], layers_dims[l-1]) * np.sqrt((2/layers_dims[l-1]))
parameters['b' + str(l)] = np.zeros((layers_dims[l], 1))
### END CODE HERE ### return parameters
Coursera Deep Learning 2 Improving Deep Neural Networks: Hyperparameter tuning, Regularization and Optimization - week1, Assignment(Initialization)
In [27]:
parameters = initialize_parameters_he([2, 4, 1])
print("W1 = " + str(parameters["W1"]))
print("b1 = " + str(parameters["b1"]))
print("W2 = " + str(parameters["W2"]))
print("b2 = " + str(parameters["b2"]))
W1 = [[ 1.78862847  0.43650985]
[ 0.09649747 -1.8634927 ]
[-0.2773882 -0.35475898]
[-0.08274148 -0.62700068]]
b1 = [[ 0.]
[ 0.]
[ 0.]
[ 0.]]
W2 = [[-0.03098412 -0.33744411 -0.92904268 0.62552248]]
b2 = [[ 0.]]

Expected Output:

**W1** [[ 1.78862847 0.43650985] [ 0.09649747 -1.8634927 ] [-0.2773882 -0.35475898] [-0.08274148 -0.62700068]]
**b1** [[ 0.] [ 0.] [ 0.] [ 0.]]
**W2** [[-0.03098412 -0.33744411 -0.92904268 0.62552248]]
**b2** [[ 0.]]

Run the following code to train your model on 15,000 iterations using He initialization.

In [28]:
parameters = model(train_X, train_Y, initialization = "he")
print ("On the train set:")
predictions_train = predict(train_X, train_Y, parameters)
print ("On the test set:")
predictions_test = predict(test_X, test_Y, parameters)
 
Cost after iteration 0: 0.8830537463419761
Cost after iteration 1000: 0.6879825919728063
Cost after iteration 2000: 0.6751286264523371
Cost after iteration 3000: 0.6526117768893807
Cost after iteration 4000: 0.6082958970572938
Cost after iteration 5000: 0.5304944491717495
Cost after iteration 6000: 0.4138645817071794
Cost after iteration 7000: 0.3117803464844441
Cost after iteration 8000: 0.23696215330322562
Cost after iteration 9000: 0.18597287209206836
Cost after iteration 10000: 0.1501555628037182
Cost after iteration 11000: 0.12325079292273548
Cost after iteration 12000: 0.09917746546525937
Cost after iteration 13000: 0.0845705595402428
Cost after iteration 14000: 0.07357895962677366
aaarticlea/png;base64,iVBORw0KGgoAAAANSUhEUgAAAb0AAAEWCAYAAADy9UlpAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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%0ACUOhJyIiCUOhJyIiCeP/A/RfZM1G9hqrAAAAAElFTkSuQmCC" alt="" />
On the train set:
Accuracy: 0.993333333333
On the test set:
Accuracy: 0.96
In [29]:
plt.title("Model with He initialization")
axes = plt.gca()
axes.set_xlim([-1.5,1.5])
axes.set_ylim([-1.5,1.5])
plot_decision_boundary(lambda x: predict_dec(parameters, x.T), train_X, train_Y)
aaarticlea/png;base64,iVBORw0KGgoAAAANSUhEUgAAAcoAAAEWCAYAAADmYNeIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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%0AR4cbjjbmrB*oprr7cHqp24vvw5T8POH+fw/uSH+ov/9MP/woAf//xBP/kD/+IZLOJEwSErHLm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" alt="" />

Observations:

  • The model with He initialization separates the blue and the red dots very well in a small number of iterations.

5 - Conclusions

You have seen three different types of initializations. For the same number of iterations and same hyperparameters the comparison is:

**Model** **Train accuracy** **Problem/Comment**
3-layer NN with zeros initialization 50% fails to break symmetry
3-layer NN with large random initialization 83% too large weights
3-layer NN with He initialization 99% recommended method

What you should remember from this notebook:

  • Different initializations lead to different results
  • Random initialization is used to break symmetry and make sure different hidden units can learn different things
  • Don't intialize to values that are too large
  • He initialization works well for networks with ReLU activations.