forked from opencog/python-destin
-
Notifications
You must be signed in to change notification settings - Fork 0
/
test_destin.py
129 lines (120 loc) · 4.75 KB
/
test_destin.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
__author__ = 'teddy'
import cPickle as pickle
from network import *
from load_data import *
"""
Here I don't move the image, I rather let a typical node move around the image
This is to make use of the methods from the previous UniformDeSTIN version
"""
# *****Define Parameters for the Network and nodes
# Network Params
num_layers = 4
patch_mode = 'Adjacent'
image_type = 'Color'
network_mode = True
# For a Node: specify Your Algorithm Choice and Corresponding parameters
'''
# ******************************************************************************************
#
# Incremental Clustering
#
num_nodes_per_layer = [[8, 8], [4, 4], [2, 2], [1, 1]]
num_cents_per_layer = [25, 25, 25, 25]
print "Uniform DeSTIN with Clustering"
algorithm_choice = 'Clustering'
alg_params = {'mr': 0.01, 'vr': 0.01, 'sr': 0.001, 'DIMS': [],
'CENTS': [], 'node_id': [],
'num_cents_per_layer': num_cents_per_layer}
# ******************************************************************************************
'''
# '''
# ******************************************************************************************
# Hierarchy Of AutoEncoders
print "Uniform DeSTIN with AutoEncoders"
num_nodes_per_layer = [[8, 8], [4, 4], [2, 2], [1, 1]]
num_cents_per_layer = [25, 25, 25, 25]
algorithm_choice = 'AutoEncoder'
inp_size = 48
hid_size = 100
alg_params = [[inp_size, hid_size], [4 * hid_size, hid_size],
[4 * hid_size, hid_size], [4 * hid_size, hid_size]]
# ******************************************************************************************
# '''
# Declare a Network Object
DESTIN = Network(
num_layers, algorithm_choice, alg_params, num_nodes_per_layer, patch_mode, image_type)
DESTIN.setmode(network_mode)
DESTIN.set_lowest_layer(0)
# Load Data
[data, labels] = loadCifar(10)
del labels
# data = np.random.rand(5,32*32*3)
# Initialize Network; there is is also a layer-wise initialization option
DESTIN.init_network()
for I in range(data.shape[0]): # For Every image in the data set
if I % 200 == 0:
print("Training Iteration Number %d" % I)
for L in range(DESTIN.number_of_layers):
if L == 0:
img = data[I][:].reshape(32, 32, 3)
DESTIN.layers[0][L].train_typical_node(img, [4, 4], algorithm_choice)
# This is equivalent to sharing centroids or kernels
DESTIN.layers[0][L].share_learned_parameters()
DESTIN.layers[0][L].load_input(img, [4, 4])
DESTIN.layers[0][L].do_layer_learning()
else:
DESTIN.layers[0][L].train_typical_node(
DESTIN.layers[0][L - 1].nodes, [2, 2], algorithm_choice)
DESTIN.layers[0][L].share_learned_parameters()
DESTIN.layers[0][L].load_input(
DESTIN.layers[0][L - 1].nodes, [2, 2])
DESTIN.layers[0][L].do_layer_learning()
print("Testing Started")
network_mode = False
DESTIN.setmode(network_mode)
# On the training set
[data, labels] = loadCifar(10)
del labels
for I in range(data.shape[0]): # For Every image in the data set
if I % 1000 == 0:
print("Testing Iteration Number %d" % I)
for L in range(DESTIN.number_of_layers):
if L == 0:
img = data[I][:].reshape(32, 32, 3)
DESTIN.layers[0][L].load_input(img, [4, 4])
DESTIN.layers[0][L].do_layer_learning()
else:
DESTIN.layers[0][L].load_input(
DESTIN.layers[0][L - 1].nodes, [2, 2])
DESTIN.layers[0][L].do_layer_learning()
DESTIN.update_belief_exporter()
if I in range(199, 50999, 200):
Name = 'train/' + str(I + 1) + '.txt'
file_id = open(Name, 'w')
pickle.dump(np.array(DESTIN.network_belief['belief']), file_id)
file_id.close()
# Get rid-off accumulated training beliefs
DESTIN.clean_belief_exporter()
[data, labels] = loadCifar(6)
del labels
# On the test set
for I in range(data.shape[0]): # For Every image in the data set
if I % 1000 == 0:
print("Testing Iteration Number %d" % (I+50000))
for L in range(DESTIN.number_of_layers):
if L == 0:
img = data[I][:].reshape(32, 32, 3)
DESTIN.layers[0][L].load_input(img, [4, 4])
DESTIN.layers[0][L].do_layer_learning() # Calculates belief for
else:
DESTIN.layers[0][L].load_input(
DESTIN.layers[0][L - 1].nodes, [2, 2])
DESTIN.layers[0][L].do_layer_learning()
DESTIN.update_belief_exporter()
if I in range(199, 10199, 200):
Name = 'test/' + str(I + 1) + '.txt'
file_id = open(Name, 'w')
pickle.dump(np.array(DESTIN.network_belief['belief']), file_id)
file_id.close()
# Get rid-off accumulated training beliefs
DESTIN.clean_belief_exporter()