Deep-Learning-in-Python/Midterm-Single-layer+.py at master · DolanGWU/Deep-Learning-in-Python · 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
131
132
133
134
135
136

# coding: utf-8

# In[14]:

#pre-code
import tensorflow as tf
import numpy as np
import matplotlib.pyplot as plt
get_ipython().magic('matplotlib inline')
import os;
print (os.getcwd())


# In[8]:

#Batch Making for data file
#create a while loop and read the text file line by line
import numpy as np
import pickle
with open('bbs-train.txt') as f:
    count = 0
    batch = []
    counter = 0
    for line in f:
        if count<150:
            batch.append(np.array(line.strip().split(" "),dtype=np.float64))
            count = count + 1
        else:
            filen = "C:\\Users\\12562\\batches\\datafile"+str(counter)
            pickle.dump(batch,open(filen,"wb"))
            #with open('batches\\datafile' + str(counter),"w") as h:
            #    h.write("\n".join(str(batch)))
            #print(str(batch[0]))
            counter = counter + 1
            #write batch into a file
            #label_{counter}
            batch.clear()
            count = 0
#print(batch)


# In[9]:

#Batch making for labels
#create a while loop and read the text file line by line
import numpy as np
with open('label-train.txt') as d:
    count = 0
    batch = []
    counter = 0
    for line in d:
        if count<150:
            batch.append(int(line.strip().split("   ")[1]))
            count = count + 1
        else:
            filen = "C:\\Users\\12562\\batches\\belfile"+str(counter)
            pickle.dump(batch,open(filen,"wb"))
            counter = counter + 1
            #write batch into a file
            #label_{counter}
            batch.clear()
            count = 0
#print(batch)


# In[10]:

#Batch making test data and label
import pickle
import numpy as np
listdata=[]
listlabel=[]
for counter in range(24,30):
    batch_xtest = np.array(pickle.load(open("C:\\Users\\12562\\batches\\datafile"+str(counter),"rb")))
    batch_ytest = np.array(pickle.load(open("C:\\Users\\12562\\batches\\belfile"+str(counter),"rb"))).reshape(-1,1)
    listdata.append(batch_xtest)
    listlabel.append(batch_ytest)
X_test = np.vstack(listdata)
filen = "C:\\Users\\12562\\batches\\testdatafile"
pickle.dump(X_test,open(filen,"wb"))
Y_test = np.vstack(listlabel)
filen = "C:\\Users\\12562\\batches\\testlabelfile"
pickle.dump(Y_test,open(filen,"wb"))


# In[19]:

#Model making with batched data
import tensorflow as tf
from tensorflow.python.framework import ops
ops.reset_default_graph()
import numpy as np
import pickle
x = tf.placeholder("float", [None, 800])
y_ = tf.placeholder("float", [None,1])

# inference
W = tf.Variable(tf.random_normal([800, 250]))
b = tf.Variable(tf.random_normal([250]))

y= tf.nn.sigmoid(tf.matmul(x,W) + b)

cross_entropy = -tf.reduce_sum(y_*tf.log(y+0.00001))
train_step = tf.train.GradientDescentOptimizer(0.001).minimize(cross_entropy)

correct_prediction = tf.equal(tf.round(y), y_)
accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))

sess = tf.Session()
sess.run(tf.global_variables_initializer())
for i in range(1000):
   # print("epoch"+str(i))
    #read the batch one by one from the folder
    #counter=0
    for counter in range(23):
        #print(counter)
        batch_x = np.array(pickle.load(open("C:\\Users\\12562\\batches\\datafile"+str(counter),"rb")))
        batch_y = np.array(pickle.load(open("C:\\Users\\12562\\batches\\belfile"+str(counter),"rb"))).reshape(-1,1)
        #print(batch_x.shape)
        #print(batch_y.shape)
        sess.run([train_step,cross_entropy], feed_dict={x: batch_x, y_: batch_y})
    if i%2 == 0:
        batch_x = np.array(pickle.load(open("C:\\Users\\12562\\batches\\testdatafile","rb")))
        batch_y = np.array(pickle.load(open("C:\\Users\\12562\\batches\\testlabelfile","rb"))).reshape(-1,1)
        loss,acc = sess.run([cross_entropy,accuracy], feed_dict={x: batch_x, y_: batch_y})
        print(str(i)+":")
        print(loss)
        print(acc)
        print()
    #print(sess.run([cross_entropy]))
        #counter = counter + 1
print("done")


# In[ ]: