CoCalc Logo Icon
StoreFeaturesDocsShareSupportNewsAboutSign UpSign In
amanchadha

Real-time collaboration for Jupyter Notebooks, Linux Terminals, LaTeX, VS Code, R IDE, and more,
all in one place. Commercial Alternative to JupyterHub.

GitHub Repository: amanchadha/coursera-deep-learning-specialization
 Path: blob/master/C2 - Improving Deep Neural Networks Hyperparameter tuning, Regularization and Optimization/Week 2/testCases.py
Views: 4802
1
import numpy as np

2


3
def update_parameters_with_gd_test_case():

4
    np.random.seed(1)

5
    learning_rate = 0.01

6
    W1 = np.random.randn(2,3)

7
    b1 = np.random.randn(2,1)

8
    W2 = np.random.randn(3,3)

9
    b2 = np.random.randn(3,1)

10


11
    dW1 = np.random.randn(2,3)

12
    db1 = np.random.randn(2,1)

13
    dW2 = np.random.randn(3,3)

14
    db2 = np.random.randn(3,1)

15
    

16
    parameters = {"W1": W1, "b1": b1, "W2": W2, "b2": b2}

17
    grads = {"dW1": dW1, "db1": db1, "dW2": dW2, "db2": db2}

18
    

19
    return parameters, grads, learning_rate

20


21
"""

22
def update_parameters_with_sgd_checker(function, inputs, outputs):

23
    if function(inputs) == outputs:

24
        print("Correct")

25
    else:

26
        print("Incorrect")

27
"""

28


29
def random_mini_batches_test_case():

30
    np.random.seed(1)

31
    mini_batch_size = 64

32
    X = np.random.randn(12288, 148)

33
    Y = np.random.randn(1, 148) < 0.5

34
    return X, Y, mini_batch_size

35


36
def initialize_velocity_test_case():

37
    np.random.seed(1)

38
    W1 = np.random.randn(2,3)

39
    b1 = np.random.randn(2,1)

40
    W2 = np.random.randn(3,3)

41
    b2 = np.random.randn(3,1)

42
    parameters = {"W1": W1, "b1": b1, "W2": W2, "b2": b2}

43
    return parameters

44


45
def update_parameters_with_momentum_test_case():

46
    np.random.seed(1)

47
    W1 = np.random.randn(2,3)

48
    b1 = np.random.randn(2,1)

49
    W2 = np.random.randn(3,3)

50
    b2 = np.random.randn(3,1)

51


52
    dW1 = np.random.randn(2,3)

53
    db1 = np.random.randn(2,1)

54
    dW2 = np.random.randn(3,3)

55
    db2 = np.random.randn(3,1)

56
    parameters = {"W1": W1, "b1": b1, "W2": W2, "b2": b2}

57
    grads = {"dW1": dW1, "db1": db1, "dW2": dW2, "db2": db2}

58
    v = {'dW1': np.array([[ 0.,  0.,  0.],

59
        [ 0.,  0.,  0.]]), 'dW2': np.array([[ 0.,  0.,  0.],

60
        [ 0.,  0.,  0.],

61
        [ 0.,  0.,  0.]]), 'db1': np.array([[ 0.],

62
        [ 0.]]), 'db2': np.array([[ 0.],

63
        [ 0.],

64
        [ 0.]])}

65
    return parameters, grads, v

66
    

67
def initialize_adam_test_case():

68
    np.random.seed(1)

69
    W1 = np.random.randn(2,3)

70
    b1 = np.random.randn(2,1)

71
    W2 = np.random.randn(3,3)

72
    b2 = np.random.randn(3,1)

73
    parameters = {"W1": W1, "b1": b1, "W2": W2, "b2": b2}

74
    return parameters

75


76
def update_parameters_with_adam_test_case():

77
    np.random.seed(1)

78
    v, s = ({'dW1': np.array([[ 0.,  0.,  0.],

79
         [ 0.,  0.,  0.]]), 'dW2': np.array([[ 0.,  0.,  0.],

80
         [ 0.,  0.,  0.],

81
         [ 0.,  0.,  0.]]), 'db1': np.array([[ 0.],

82
         [ 0.]]), 'db2': np.array([[ 0.],

83
         [ 0.],

84
         [ 0.]])}, {'dW1': np.array([[ 0.,  0.,  0.],

85
         [ 0.,  0.,  0.]]), 'dW2': np.array([[ 0.,  0.,  0.],

86
         [ 0.,  0.,  0.],

87
         [ 0.,  0.,  0.]]), 'db1': np.array([[ 0.],

88
         [ 0.]]), 'db2': np.array([[ 0.],

89
         [ 0.],

90
         [ 0.]])})

91
    W1 = np.random.randn(2,3)

92
    b1 = np.random.randn(2,1)

93
    W2 = np.random.randn(3,3)

94
    b2 = np.random.randn(3,1)

95


96
    dW1 = np.random.randn(2,3)

97
    db1 = np.random.randn(2,1)

98
    dW2 = np.random.randn(3,3)

99
    db2 = np.random.randn(3,1)

100
    

101
    parameters = {"W1": W1, "b1": b1, "W2": W2, "b2": b2}

102
    grads = {"dW1": dW1, "db1": db1, "dW2": dW2, "db2": db2}

103
    

104
    return parameters, grads, v, s

105
    

106