NormalNN/tf1.py at master · A4Vision/NormalNN · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
import numpy as np
# Python
import tensorflow as tf
hello = tf.constant('Hello, TensorFlow!')
sess = tf.Session()
print(sess.run(hello))
node1 = tf.constant(3.0, tf.float32)
node2 = tf.constant(4.0) # also tf.float32 implicitly
print(node1, node2)
sess = tf.Session()
print sess.run([node1, node2])
node3 = tf.add(node1, node2)
print("node3: ", node3)
print("sess.run(node3): ",sess.run(node3))

a = tf.placeholder(tf.float32)
b = tf.placeholder(tf.float32)
adder_node = a + b  # + provides a shortcut for tf.add(a, b)
sess.close()
sess = tf.Session()
print "run(adder_node)", sess.run(adder_node, feed_dict={a: [1, 123], b: [213, 123]})
triple_3 = adder_node * 3
print "run(triple_3)", sess.run(triple_3, feed_dict={a: 3, b: 1})

W = tf.Variable([.3], tf.float32)
b = tf.Variable([-.3], tf.float32)
x = tf.placeholder(tf.float32)
linear_model = W * x + b
init = tf.global_variables_initializer()
sess.run(init)
print 'sess.run(linear_model, {x: [1, 2, 3, 4]})', sess.run(linear_model, {x: [1, 2, 3, 4]})

y = tf.placeholder(tf.float32)
squared_deltas = tf.square(linear_model - y)
loss = tf.reduce_sum(squared_deltas)
fixW = tf.assign(W, [-1.])
fixb = tf.assign(b, [1.])
sess.run([fixW, fixb])
print(sess.run(loss, {x: [1,2,3,4], y: [0,-1,-2,-3]}))


fixW = tf.assign(W, [4.2])
fixb = tf.assign(b, [0.])
sess.run([fixW, fixb])
optimizer = tf.train.GradientDescentOptimizer(0.01)
train = optimizer.minimize(loss)
for i in range(6):
    print sess.run([W, b, loss], {x:[1,2,3,4], y:[0,-1,-2,-3]})
    sess.run(train, {x:[1,2,3,4], y:[0,-1,-2,-3]})


W = tf.Variable([1, 3.], tf.float32)
x = tf.placeholder(tf.float32, shape=2)
z = W * x
s = tf.reduce_sum(z)


droped_x = tf.nn.dropout(x, 0.5)
matrix = tf.Variable(np.array([[1, 2, 3], [4, 5, 6.], [7, 8, 9.]]), dtype= tf.float32)
droped_matrix = tf.nn.dropout(matrix, 0.5)
# print 'shape', matrix.shape, droped_x.shape
init = tf.global_variables_initializer()
pow_x = tf.pow(x, 2)
sess.run(init)

mult_matrix = tf.constant(10.2) * matrix

# print 'matrix mult', sess.run(mult_matrix, {x: [1, 3]})
# print 'pow x', sess.run(pow_x, {x: [1, 3]})
pow_matrix = matrix * matrix
# print 'pow matrix', sess.run(pow_matrix)
# print 'run(s)', sess.run([s, droped_x, droped_matrix], {x: [1, 2.]})

x = tf.placeholder(tf.float32, shape=(5,))
y = tf.placeholder(tf.float32, shape=(5,))

def f(x, y):
    return x ** 2 + 3 * y * x - y ** 2 - 3
#
c = [[-3, 0, -1], [0, 3, 0], [1, 0, 0]]


def f2(x, y, c):
    res = 0
    for i in xrange(len(c)):
        for j in xrange(len(c[0])):
            res += c[i][j] * x ** i * y ** j
    return res


def f3(x, y, c):
    res = 0
    x_cur = 1

    for i in xrange(len(c)):
        y_cur = 1
        for j in xrange(len(c[0])):
            res += c[i][j] * x_cur * y_cur
            y_cur *= y
        x_cur *= x
    return res


def polynomial2d_tf(x, y, c):
    assert x.shape == y.shape
    shape = x.shape
    res = tf.constant(0, tf.float32, shape)
    x_cur = tf.constant(1, tf.float32, shape)
    x_degree = len(c)
    y_degree = len(c[0])
    for i in xrange(x_degree):
        y_cur = tf.constant(1, tf.float32, shape)
        for j in xrange(y_degree):
            res += c[i][j] * x_cur * y_cur
            y_cur *= y
        x_cur *= x
    return res


p = polynomial2d_tf(x, y, c)
q = (p / 2.) ** 2
z = p[1] + p[2]
w = tf.reduce_sum(p) - p[1]
concat1 = tf.stack([w, w, w], 0)
concat2 = tf.stack([w, w, w])
concat_val1, concat_val2, w_value, z_value, divided, poly_value = sess.run([concat1, concat2, w, z, q, p], feed_dict={x: [21.] * 5, y: [-2.] * 5})
print concat_val1, concat_val2
print f(21, -2), f2(21, -2, c), f3(21, -2, c), poly_value, divided, z_value, w_value

assert f(21, -2) == f2(21, -2, c) == f3(21, -2, c) == poly_value[0] == poly_value[1]