-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathDataDistributor.py
129 lines (103 loc) · 3.73 KB
/
DataDistributor.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
"""
Name: DataDistributor.py
Aim: Convert MNIST data into binary classification data
Author: Siddarth C
Date: September, 2021
"""
from mnist import MNIST
import numpy as np
import random
import os
import shutil
random.seed(100)
if os.path.exists('ClientData'):
shutil.rmtree('ClientData')
print('Deleted exisitng *ClientData* folder')
os.mkdir('ClientData')
print('Created *ClientData* folder')
mndata = MNIST('samples')
print('Loading Training Data...')
images, labels = mndata.load_training()
print('Processing Training Data...')
images = np.array(images).reshape((60000, 28, 28))
labels = np.array(labels)
inds = labels.argsort()
images = images[inds]
labels = labels[inds]
sorted = [images[i*6000:(i+1)*6000] for i in range(10)]
trainx = []
trainy = []
mnist_200 = [i for i in range(10) for j in range(20)]
mnist_10_20_300 = []
for i in range(10):
dummy = [j for j in range(6000)]
random.shuffle(dummy)
mnist_10_20_300.append([dummy[j*300:(j+1)*300] for j in range(20)])
trainx = []
trainy = []
client_id = 0
safe_value = 0
while len(mnist_200) > 0:
n1 = random.randint(0, len(mnist_200) - 1)
n2 = random.randint(0, len(mnist_200) - 1)
no1 = mnist_200[n1]
no2 = mnist_200[n2]
safe_value += 1
if safe_value > 10e6:
print('Error. Please run the program again!')
exit()
if no1 != no2:
client_id += 1
indiv_client_x = []
indiv_client_y = []
no1_index = mnist_10_20_300[no1][0]
no2_index = mnist_10_20_300[no2][0]
mnist_10_20_300[no1] = mnist_10_20_300[no1][1:]
mnist_10_20_300[no2] = mnist_10_20_300[no2][1:]
client_prob = random.randint(100,200)/1000
for no, no_index in zip([no1, no2], [no1_index, no2_index]):
z = 0
for no_i in no_index:
img = sorted[no][no_i]
data_prob = random.random()
red_version = np.stack((img, np.zeros_like(img), np.zeros_like(img)), axis = 2)
green_version = np.stack((np.zeros_like(img), img, np.zeros_like(img)), axis = 2)
if (data_prob > client_prob and no < 5) or (data_prob < client_prob and no > 5):
indiv_client_x.append(red_version)
else:
indiv_client_x.append(green_version)
indiv_client_y.append(1*(no>4))
os.mkdir('ClientData/Client' + str(client_id))
np.save('ClientData/Client' + str(client_id) + '/y.npy', np.stack(indiv_client_y))
np.save('ClientData/Client' + str(client_id) + '/x.npy', np.stack(indiv_client_x))
del mnist_200[n1]
if n2 > n1:
del mnist_200[n2 - 1]
else:
del mnist_200[n2]
print('Distributed training data among clients!')
print()
if os.path.exists('TestData'):
shutil.rmtree('TestData')
print('Deleted exisitng *TestData* folder')
os.mkdir('TestData')
print('Created *TestData* folder')
print('Loading Test Data...')
images, labels = mndata.load_testing()
print('Processing Test Data...')
images = np.array(images).reshape((10000, 28, 28))
labels = np.array(labels)
testx = []
testy = []
for i, l in zip(images, labels):
red_version = np.stack((i, np.zeros_like(i), np.zeros_like(i)), axis = 2)
green_version = np.stack((np.zeros_like(i), i, np.zeros_like(i)), axis = 2)
data_prob = random.randint(100,200)/1000
if (data_prob > 0.9 and l > 5) or (data_prob < 0.9 and l < 5):
testx.append(green_version)
else:
testx.append(red_version)
testy.append(1*(l>4))
np.save('TestData/x.npy', testx)
np.save('TestData/y.npy', testy)
print('Test data saved!')