-
Notifications
You must be signed in to change notification settings - Fork 15
/
Copy pathn1_2cnv1fc.py
246 lines (170 loc) · 10.3 KB
/
n1_2cnv1fc.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
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
#import tensorflow as tf
from tensorflow.compat import v1 as tf
tf.compat.v1.disable_eager_execution()
import numpy as np
from collections import namedtuple
import datetime
import ocr_utils
from n0_network import base_network as b_network
class network(b_network):
''' definition of the network
'''
def __init__(self, truthed_features, dtype=np.float32):
self._sess = tf.InteractiveSession()
lst = []
extra_features_width = 0 # width of extra features
"""# ==============================================================================
Placeholders
Compute the size of various layers
Create a tensorflow Placeholder for each feature of data returned from the
dataset
"""# ==============================================================================
for i,nm in enumerate(truthed_features.feature_names):
# features[0], is always the target. For instance it may be m_label_one_hot
# the second features[1] is the 'image' that is passed to the convolution layers
# Any additional features bypass the convolution layers and go directly
# into the fully connected layer.
# The width of the extra features is calculated in order to allocate
# the correct widths of weights, # and inputs
# names are assigned to make the look pretty on the tensorboard graph.
if i == 0:
nm = 'y_'+nm
else:
nm = 'x_'+nm
if i>1:
extra_features_width += truthed_features.feature_width[i]
lst.append(tf.placeholder(dtype, shape=[None, truthed_features.feature_width[i]], name=nm))
# ph is a named tuple with key names like 'image', 'm_label', and values that
# are tensors. The display name on the Chrome graph are 'y_m_label', 'x_image,
# x_upper_case etc.
Place_Holders = namedtuple('Place_Holders', truthed_features.feature_names)
self._ph = Place_Holders(*lst) # unpack placeholders into named Tuple
self._keep_prob = tf.placeholder(dtype,name='keep_prob')
self._nRows = truthed_features.num_rows #image height
self._nCols = truthed_features.num_columns #image width
nFc = 1024 # size of fully connected layer
nConv1 = 32 # size of first convolution layer
nConv2 = 64 # size of second convolution layer
nTarget = truthed_features.feature_width[0] # the number of one_hot features in the target, 'm_label'
n_h_pool2_outputs = int(self._nRows/4) * int(self._nCols/4) * nConv2 # second pooling layer
n_h_pool2_outputsx = n_h_pool2_outputs + extra_features_width # fully connected
"""# ==============================================================================
Build a Multilayer Convolutional Network
Weight Initialization
"""# ==============================================================================
def weight_variable(shape, dtype):
initial = tf.truncated_normal(shape, stddev=0.1,dtype=dtype)
return tf.Variable(initial)
def bias_variable(shape, dtype):
initial = tf.constant(0.1, shape=shape, dtype=dtype)
return tf.Variable(initial)
"""# ==============================================================================
Convolution and Pooling
keep our code cleaner, let's also abstract those operations into functions.
"""# ==============================================================================
def conv2d(x, W):
return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME')
def max_pool_2x2(x):
return tf.nn.max_pool(x, ksize=[1, 2, 2, 1],
strides=[1, 2, 2, 1], padding='SAME')
"""# ==============================================================================
First Convolutional Layer
"""# ==============================================================================
with tf.name_scope("w_conv1") as scope:
W_conv1 = weight_variable([5, 5, 1, nConv1],dtype)
b_conv1 = bias_variable([nConv1],dtype)
with tf.name_scope("reshape_x_image") as scope:
self._x_image = tf.reshape(self._ph.image, [-1,self._nCols,self._nRows,1])
image_summ = tf.summary.image("x_image", self._x_image)
"""# ==============================================================================
We then convolve x_image with the weight tensor, add the bias, apply the ReLU function,
and finally max pool.
"""# ==============================================================================
with tf.name_scope("convolve_1") as scope:
h_conv1 = tf.nn.relu(conv2d(self._x_image, W_conv1) + b_conv1)
with tf.name_scope("pool_1") as scope:
h_pool1 = max_pool_2x2(h_conv1)
"""# ==============================================================================
Second Convolutional Layer
In order to build a deep network, we stack several layers of this type. The second
layer will have 64 features for each 5x5 patch.
"""# ==============================================================================
with tf.name_scope("convolve_2") as scope:
W_conv2 = weight_variable([5, 5, nConv1, nConv2],dtype)
b_conv2 = bias_variable([64],dtype)
h_conv2 = tf.nn.relu(conv2d(h_pool1, W_conv2) + b_conv2)
with tf.name_scope("pool_2") as scope:
h_pool2 = max_pool_2x2(h_conv2)
"""# ==============================================================================
Densely Connected Layer
Now that the image size has been reduced to 7x7, we add a fully-connected layer
with neurons to allow processing on the entire image. We reshape the tensor
from the pooling layer into a batch of vectors, multiply by a weight matrix, add
a bias, and apply a ReLU.
"""# ==============================================================================
with tf.name_scope("W_fc1_b") as scope:
W_fc1 = weight_variable([n_h_pool2_outputsx, nFc],dtype)
b_fc1 = bias_variable([nFc],dtype)
h_pool2_flat = tf.reshape(h_pool2, [-1, n_h_pool2_outputs])
# append the features, the 2nd on, that go directly to the fully connected layer
for i in range(2,truthed_features.num_features ):
print(i)
print(self._ph[i])
h_pool2_flat = tf.concat([h_pool2_flat, self._ph[i]],1)
h_fc1 = tf.nn.relu(tf.matmul(h_pool2_flat, W_fc1) + b_fc1)
"""# ==============================================================================
Dropout
"""# ==============================================================================
with tf.name_scope("drop") as scope:
h_fc1_drop = tf.nn.dropout(h_fc1, self._keep_prob)
"""# ==============================================================================
Readout Layer
"""# ==============================================================================
with tf.name_scope("softmax") as scope:
W_fc2 = weight_variable([nFc, nTarget],dtype)
b_fc2 = bias_variable([nTarget],dtype)
y_conv=tf.nn.softmax(tf.matmul(h_fc1_drop, W_fc2) + b_fc2)
with tf.name_scope("xent") as scope:
# 1e-8 added to eliminate the crash of training when taking log of 0
cross_entropy = -tf.reduce_sum(self._ph[0]*tf.log(y_conv+ 1e-8 ))
ce_summ = tf.summary.scalar("cross entropy", cross_entropy)
with tf.name_scope("train") as scope:
self._train_step = tf.train.AdamOptimizer(1e-4).minimize(cross_entropy)
with tf.name_scope("test") as scope:
self._correct_prediction = tf.equal(tf.argmax(y_conv,1), tf.argmax(self._ph[0],1))
self._prediction = tf.argmax(y_conv,1)
self._accuracy = tf.reduce_mean(tf.cast(self._correct_prediction, dtype))
accuracy_summary = tf.summary.scalar("accuracy", self._accuracy)
weight_summary = tf.summary.histogram("weights", W_fc2)
"""# ==============================================================================
Start TensorFlow Interactive Session
"""# ==============================================================================
self._sess.run(tf.initialize_all_variables())
self._merged = tf.summary.merge_all()
tm = ""
tp = datetime.datetime.now().timetuple()
for i in range(4):
tm += str(tp[i])+'-'
tm += str(tp[4])
self._writer = tf.summary.FileWriter("/tmp/ds_logs/"+ tm, self._sess.graph)
def computeSize(s,tens):
sumC = 1
tShape = tens.get_shape()
nDims = len(tShape)
for i in range(nDims):
sumC *= tShape[i]
print ('\t{}\t{}'.format(s,sumC),flush=True)
return sumC
print ('network size:',flush=True)
total = computeSize("W_fc1",W_fc1)+ \
computeSize ("b_fc1",b_fc1) + \
computeSize ("W_conv1",W_conv1) + \
computeSize ("b_conv1",b_conv1) + \
computeSize ("W_conv2",W_conv2) + \
computeSize ("b_conv2",b_conv2) + \
computeSize ("W_fc2",W_fc2) + \
computeSize ("b_fc2",b_fc2)
print('\ttotal\t{}'.format(total),flush=True )
def reset_graph(self):
tf.reset_default_graph() # only necessary when iterating through fonts
self._sess.close()