/
TensorFlow-NN.py
157 lines (129 loc) · 5.21 KB
/
TensorFlow-NN.py
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
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
import tensorflow as tf
import scipy.io as io
import numpy as np
import os
# Parameters
learning_rate = 0.01
training_epochs = 1000
batch_size = 512
display_step = 10
# Network Parameters
n_hidden = 25 # 1st layer number of neurons
n_input = 400 # MNIST data input_features (img shape: 28*28)
n_classes = 10 # MNIST total classes (0-9 digits)
X = tf.placeholder(dtype=tf.float64, shape=[None, 400], name='X')
Y = tf.placeholder(dtype=tf.float64, shape=[None, 10], name='y')
parameters = {
'W1': tf.get_variable(dtype=tf.float64, name='W1', shape=(25, 400), initializer=tf.random_normal_initializer()),
'b1': tf.get_variable(dtype=tf.float64, name='b1', shape=(1, 25), initializer=tf.zeros_initializer()),
'W2': tf.get_variable(dtype=tf.float64, name='W2', shape=(10, 25), initializer=tf.random_normal_initializer()),
'b2': tf.get_variable(dtype=tf.float64, name='b2', shape=(1, 10), initializer=tf.zeros_initializer())
}
def one_hot(y):
y_new = []
for i in y:
if i == [10] or i == [0]:
y_new.append([0, 0, 0, 0, 0, 0, 0, 0, 0, 1])
elif i == [9]:
y_new.append([0, 0, 0, 0, 0, 0, 0, 0, 1, 0])
elif i == [8]:
y_new.append([0, 0, 0, 0, 0, 0, 0, 1, 0, 0])
elif i == [7]:
y_new.append([0, 0, 0, 0, 0, 0, 1, 0, 0, 0])
elif i == [6]:
y_new.append([0, 0, 0, 0, 0, 1, 0, 0, 0, 0])
elif i == [5]:
y_new.append([0, 0, 0, 0, 1, 0, 0, 0, 0, 0])
elif i == [4]:
y_new.append([0, 0, 0, 1, 0, 0, 0, 0, 0, 0])
elif i == [3]:
y_new.append([0, 0, 1, 0, 0, 0, 0, 0, 0, 0])
elif i == [2]:
y_new.append([0, 1, 0, 0, 0, 0, 0, 0, 0, 0])
elif i == [1]:
y_new.append([1, 0, 0, 0, 0, 0, 0, 0, 0, 0])
return y_new
def next_batch(num, data, labels):
'''
Return a total of `num` random samples and labels.
'''
idx = np.arange(0, len(data))
np.random.shuffle(idx)
idx = idx[:num]
data_shuffle = [data[i] for i in idx]
labels_shuffle = [labels[i] for i in idx]
return np.asarray(data_shuffle), np.asarray(labels_shuffle)
# Create model
def multilayer_perceptron(x):
# Hidden fully connected layer with 256 neurons
layer_1 = tf.matmul(x, tf.transpose(parameters['W1'])) + parameters['b1']
rel1 = tf.nn.relu(layer_1)
# Hidden fully connected layer with 256 neurons
out_layer = tf.matmul(rel1, tf.transpose(parameters['W2'])) + parameters['b2']
return out_layer
def predict(x):
# Hidden fully connected layer with 256 neurons
layer_1 = tf.matmul(x, tf.transpose(parameters['W1'])) + parameters['b1']
rel1 = tf.nn.relu(layer_1)
# Hidden fully connected layer with 256 neurons
layer2 = tf.matmul(rel1, tf.transpose(parameters['W2'])) + parameters['b2']
out_layer = tf.sigmoid(layer2)
return out_layer
# Construct model
logits = multilayer_perceptron(X)
# Define loss and optimizer
with tf.name_scope('Loss'):
loss_op = tf.reduce_mean(tf.nn.softmax_cross_entropy_with_logits(logits=logits, labels=Y, name='loss'))
optimizer = tf.train.AdamOptimizer(learning_rate)
train_op = optimizer.minimize(loss_op)
init = tf.global_variables_initializer()
tf.summary.scalar('Loss', loss_op)
saver = tf.train.Saver(max_to_keep=5)
with tf.Session() as sess:
data = io.loadmat('dataset/data.mat')
org_X = data['X']
org_y = data['y']
print("org_X shape", org_X.shape)
print("org_y shape", org_y.shape)
if os.listdir('./model').__len__() != 0:
print("[NOTE] Loading model...")
saver.restore(sess, './model/TensorFlow/model.ckpt')
print("[NOTE] Model restored successfully!")
else:
# Initializing the session
sess.run(init)
# Training cycle
file_writer = tf.summary.FileWriter('./logs', sess.graph)
file_writer.add_graph(tf.get_default_graph())
total_batch = int(org_y.shape[0] / batch_size)
for epoch in range(training_epochs):
avg_cost = 0.
# Loop over all batches
for i in range(total_batch):
batch_x, batch_y = next_batch(num=batch_size, data=org_X, labels=one_hot(org_y))
# Run optimization op (backprop) and cost op (to get loss value)
_, c = sess.run([train_op, loss_op], {X: batch_x, Y: batch_y})
# Compute average loss
avg_cost += c / batch_size
# Display logs per epoch s1tep
if epoch % display_step == 0:
print("Epoch:", '%04d' % (epoch + 1), "cost={:.9f}".format(avg_cost))
saver.save(sess, './model/TensorFlow/model.ckpt')
print("Optimization Finished!")
print("Predicting sample from train set!")
random_index = np.random.randint(0, 5000)
sample_x = org_X[random_index]
logits = np.argmax(predict(np.resize(sample_x, (1, 400))).eval()[0]) + 1
print("Predicted X:", logits)
print("True X:", org_y[random_index][0])
print("Prediction finished!")
p = []
logits = predict(org_X)
for i in logits.eval():
p.append(np.argmax(i) + 1)
p = np.asarray(p)
c = 0
for i in range(org_y.shape[0]):
if p[i] == org_y[i][0]:
c += 1
print("Accuracy: %.6f" % (c / org_y.shape[0]))