|
a |
|
b/Transfer Learning Resnet 50.ipynb |
|
|
1 |
{ |
|
|
2 |
"cells": [ |
|
|
3 |
{ |
|
|
4 |
"cell_type": "markdown", |
|
|
5 |
"metadata": {}, |
|
|
6 |
"source": [] |
|
|
7 |
}, |
|
|
8 |
{ |
|
|
9 |
"cell_type": "markdown", |
|
|
10 |
"metadata": {}, |
|
|
11 |
"source": [] |
|
|
12 |
}, |
|
|
13 |
{ |
|
|
14 |
"cell_type": "code", |
|
|
15 |
"execution_count": 1, |
|
|
16 |
"metadata": {}, |
|
|
17 |
"outputs": [], |
|
|
18 |
"source": [ |
|
|
19 |
"# import the libraries as shown below\n", |
|
|
20 |
"\n", |
|
|
21 |
"from tensorflow.keras.layers import Input, Lambda, Dense, Flatten\n", |
|
|
22 |
"from tensorflow.keras.models import Model\n", |
|
|
23 |
"from tensorflow.keras.applications.resnet50 import ResNet50\n", |
|
|
24 |
"from tensorflow.keras.applications.resnet50 import preprocess_input\n", |
|
|
25 |
"from tensorflow.keras.preprocessing import image\n", |
|
|
26 |
"from tensorflow.keras.preprocessing.image import ImageDataGenerator,load_img\n", |
|
|
27 |
"from tensorflow.keras.models import Sequential\n", |
|
|
28 |
"import numpy as np\n", |
|
|
29 |
"from glob import glob\n", |
|
|
30 |
"import matplotlib.pyplot as plt" |
|
|
31 |
] |
|
|
32 |
}, |
|
|
33 |
{ |
|
|
34 |
"cell_type": "code", |
|
|
35 |
"execution_count": 2, |
|
|
36 |
"metadata": {}, |
|
|
37 |
"outputs": [], |
|
|
38 |
"source": [ |
|
|
39 |
"# re-size all the images to this\n", |
|
|
40 |
"IMAGE_SIZE = [224, 224]\n", |
|
|
41 |
"\n", |
|
|
42 |
"train_path = 'Datasets/Datasets/train'\n", |
|
|
43 |
"valid_path = 'Datasets/Datasets/test'\n" |
|
|
44 |
] |
|
|
45 |
}, |
|
|
46 |
{ |
|
|
47 |
"cell_type": "code", |
|
|
48 |
"execution_count": 3, |
|
|
49 |
"metadata": {}, |
|
|
50 |
"outputs": [], |
|
|
51 |
"source": [ |
|
|
52 |
"\n", |
|
|
53 |
"resnet = ResNet50(input_shape=IMAGE_SIZE + [3], weights='imagenet', include_top=False)\n", |
|
|
54 |
"\n", |
|
|
55 |
"\n" |
|
|
56 |
] |
|
|
57 |
}, |
|
|
58 |
{ |
|
|
59 |
"cell_type": "code", |
|
|
60 |
"execution_count": 5, |
|
|
61 |
"metadata": {}, |
|
|
62 |
"outputs": [], |
|
|
63 |
"source": [ |
|
|
64 |
"# don't train existing weights\n", |
|
|
65 |
"count = 0\n", |
|
|
66 |
"for layer in resnet.layers:\n", |
|
|
67 |
" layer.trainable = False\n", |
|
|
68 |
" count += 1" |
|
|
69 |
] |
|
|
70 |
}, |
|
|
71 |
{ |
|
|
72 |
"cell_type": "code", |
|
|
73 |
"execution_count": 6, |
|
|
74 |
"metadata": {}, |
|
|
75 |
"outputs": [ |
|
|
76 |
{ |
|
|
77 |
"data": { |
|
|
78 |
"text/plain": [ |
|
|
79 |
"175" |
|
|
80 |
] |
|
|
81 |
}, |
|
|
82 |
"execution_count": 6, |
|
|
83 |
"metadata": {}, |
|
|
84 |
"output_type": "execute_result" |
|
|
85 |
} |
|
|
86 |
], |
|
|
87 |
"source": [ |
|
|
88 |
"count" |
|
|
89 |
] |
|
|
90 |
}, |
|
|
91 |
{ |
|
|
92 |
"cell_type": "code", |
|
|
93 |
"execution_count": 5, |
|
|
94 |
"metadata": {}, |
|
|
95 |
"outputs": [], |
|
|
96 |
"source": [ |
|
|
97 |
" # useful for getting number of output classes\n", |
|
|
98 |
"folders = glob('Datasets/Datasets/train/*')" |
|
|
99 |
] |
|
|
100 |
}, |
|
|
101 |
{ |
|
|
102 |
"cell_type": "code", |
|
|
103 |
"execution_count": 6, |
|
|
104 |
"metadata": {}, |
|
|
105 |
"outputs": [ |
|
|
106 |
{ |
|
|
107 |
"data": { |
|
|
108 |
"text/plain": [ |
|
|
109 |
"['Datasets/Datasets/train\\\\Lymphocytes',\n", |
|
|
110 |
" 'Datasets/Datasets/train\\\\Promyelocytes',\n", |
|
|
111 |
" 'Datasets/Datasets/train\\\\Segmented Neutrophils']" |
|
|
112 |
] |
|
|
113 |
}, |
|
|
114 |
"execution_count": 6, |
|
|
115 |
"metadata": {}, |
|
|
116 |
"output_type": "execute_result" |
|
|
117 |
} |
|
|
118 |
], |
|
|
119 |
"source": [ |
|
|
120 |
"folders" |
|
|
121 |
] |
|
|
122 |
}, |
|
|
123 |
{ |
|
|
124 |
"cell_type": "code", |
|
|
125 |
"execution_count": 7, |
|
|
126 |
"metadata": {}, |
|
|
127 |
"outputs": [], |
|
|
128 |
"source": [ |
|
|
129 |
"# our layers - you can add more if you want\n", |
|
|
130 |
"x = Flatten()(resnet.output)" |
|
|
131 |
] |
|
|
132 |
}, |
|
|
133 |
{ |
|
|
134 |
"cell_type": "code", |
|
|
135 |
"execution_count": 8, |
|
|
136 |
"metadata": {}, |
|
|
137 |
"outputs": [], |
|
|
138 |
"source": [ |
|
|
139 |
"prediction = Dense(len(folders), activation='softmax')(x)\n", |
|
|
140 |
"\n", |
|
|
141 |
"# create a model object\n", |
|
|
142 |
"model = Model(inputs=resnet.input, outputs=prediction)" |
|
|
143 |
] |
|
|
144 |
}, |
|
|
145 |
{ |
|
|
146 |
"cell_type": "code", |
|
|
147 |
"execution_count": 9, |
|
|
148 |
"metadata": {}, |
|
|
149 |
"outputs": [ |
|
|
150 |
{ |
|
|
151 |
"name": "stdout", |
|
|
152 |
"output_type": "stream", |
|
|
153 |
"text": [ |
|
|
154 |
"Model: \"model\"\n", |
|
|
155 |
"__________________________________________________________________________________________________\n", |
|
|
156 |
" Layer (type) Output Shape Param # Connected to \n", |
|
|
157 |
"==================================================================================================\n", |
|
|
158 |
" input_1 (InputLayer) [(None, 224, 224, 3 0 [] \n", |
|
|
159 |
" )] \n", |
|
|
160 |
" \n", |
|
|
161 |
" conv1_pad (ZeroPadding2D) (None, 230, 230, 3) 0 ['input_1[0][0]'] \n", |
|
|
162 |
" \n", |
|
|
163 |
" conv1_conv (Conv2D) (None, 112, 112, 64 9472 ['conv1_pad[0][0]'] \n", |
|
|
164 |
" ) \n", |
|
|
165 |
" \n", |
|
|
166 |
" conv1_bn (BatchNormalization) (None, 112, 112, 64 256 ['conv1_conv[0][0]'] \n", |
|
|
167 |
" ) \n", |
|
|
168 |
" \n", |
|
|
169 |
" conv1_relu (Activation) (None, 112, 112, 64 0 ['conv1_bn[0][0]'] \n", |
|
|
170 |
" ) \n", |
|
|
171 |
" \n", |
|
|
172 |
" pool1_pad (ZeroPadding2D) (None, 114, 114, 64 0 ['conv1_relu[0][0]'] \n", |
|
|
173 |
" ) \n", |
|
|
174 |
" \n", |
|
|
175 |
" pool1_pool (MaxPooling2D) (None, 56, 56, 64) 0 ['pool1_pad[0][0]'] \n", |
|
|
176 |
" \n", |
|
|
177 |
" conv2_block1_1_conv (Conv2D) (None, 56, 56, 64) 4160 ['pool1_pool[0][0]'] \n", |
|
|
178 |
" \n", |
|
|
179 |
" conv2_block1_1_bn (BatchNormal (None, 56, 56, 64) 256 ['conv2_block1_1_conv[0][0]'] \n", |
|
|
180 |
" ization) \n", |
|
|
181 |
" \n", |
|
|
182 |
" conv2_block1_1_relu (Activatio (None, 56, 56, 64) 0 ['conv2_block1_1_bn[0][0]'] \n", |
|
|
183 |
" n) \n", |
|
|
184 |
" \n", |
|
|
185 |
" conv2_block1_2_conv (Conv2D) (None, 56, 56, 64) 36928 ['conv2_block1_1_relu[0][0]'] \n", |
|
|
186 |
" \n", |
|
|
187 |
" conv2_block1_2_bn (BatchNormal (None, 56, 56, 64) 256 ['conv2_block1_2_conv[0][0]'] \n", |
|
|
188 |
" ization) \n", |
|
|
189 |
" \n", |
|
|
190 |
" conv2_block1_2_relu (Activatio (None, 56, 56, 64) 0 ['conv2_block1_2_bn[0][0]'] \n", |
|
|
191 |
" n) \n", |
|
|
192 |
" \n", |
|
|
193 |
" conv2_block1_0_conv (Conv2D) (None, 56, 56, 256) 16640 ['pool1_pool[0][0]'] \n", |
|
|
194 |
" \n", |
|
|
195 |
" conv2_block1_3_conv (Conv2D) (None, 56, 56, 256) 16640 ['conv2_block1_2_relu[0][0]'] \n", |
|
|
196 |
" \n", |
|
|
197 |
" conv2_block1_0_bn (BatchNormal (None, 56, 56, 256) 1024 ['conv2_block1_0_conv[0][0]'] \n", |
|
|
198 |
" ization) \n", |
|
|
199 |
" \n", |
|
|
200 |
" conv2_block1_3_bn (BatchNormal (None, 56, 56, 256) 1024 ['conv2_block1_3_conv[0][0]'] \n", |
|
|
201 |
" ization) \n", |
|
|
202 |
" \n", |
|
|
203 |
" conv2_block1_add (Add) (None, 56, 56, 256) 0 ['conv2_block1_0_bn[0][0]', \n", |
|
|
204 |
" 'conv2_block1_3_bn[0][0]'] \n", |
|
|
205 |
" \n", |
|
|
206 |
" conv2_block1_out (Activation) (None, 56, 56, 256) 0 ['conv2_block1_add[0][0]'] \n", |
|
|
207 |
" \n", |
|
|
208 |
" conv2_block2_1_conv (Conv2D) (None, 56, 56, 64) 16448 ['conv2_block1_out[0][0]'] \n", |
|
|
209 |
" \n", |
|
|
210 |
" conv2_block2_1_bn (BatchNormal (None, 56, 56, 64) 256 ['conv2_block2_1_conv[0][0]'] \n", |
|
|
211 |
" ization) \n", |
|
|
212 |
" \n", |
|
|
213 |
" conv2_block2_1_relu (Activatio (None, 56, 56, 64) 0 ['conv2_block2_1_bn[0][0]'] \n", |
|
|
214 |
" n) \n", |
|
|
215 |
" \n", |
|
|
216 |
" conv2_block2_2_conv (Conv2D) (None, 56, 56, 64) 36928 ['conv2_block2_1_relu[0][0]'] \n", |
|
|
217 |
" \n", |
|
|
218 |
" conv2_block2_2_bn (BatchNormal (None, 56, 56, 64) 256 ['conv2_block2_2_conv[0][0]'] \n", |
|
|
219 |
" ization) \n", |
|
|
220 |
" \n", |
|
|
221 |
" conv2_block2_2_relu (Activatio (None, 56, 56, 64) 0 ['conv2_block2_2_bn[0][0]'] \n", |
|
|
222 |
" n) \n", |
|
|
223 |
" \n", |
|
|
224 |
" conv2_block2_3_conv (Conv2D) (None, 56, 56, 256) 16640 ['conv2_block2_2_relu[0][0]'] \n", |
|
|
225 |
" \n", |
|
|
226 |
" conv2_block2_3_bn (BatchNormal (None, 56, 56, 256) 1024 ['conv2_block2_3_conv[0][0]'] \n", |
|
|
227 |
" ization) \n", |
|
|
228 |
" \n", |
|
|
229 |
" conv2_block2_add (Add) (None, 56, 56, 256) 0 ['conv2_block1_out[0][0]', \n", |
|
|
230 |
" 'conv2_block2_3_bn[0][0]'] \n", |
|
|
231 |
" \n", |
|
|
232 |
" conv2_block2_out (Activation) (None, 56, 56, 256) 0 ['conv2_block2_add[0][0]'] \n", |
|
|
233 |
" \n", |
|
|
234 |
" conv2_block3_1_conv (Conv2D) (None, 56, 56, 64) 16448 ['conv2_block2_out[0][0]'] \n", |
|
|
235 |
" \n", |
|
|
236 |
" conv2_block3_1_bn (BatchNormal (None, 56, 56, 64) 256 ['conv2_block3_1_conv[0][0]'] \n", |
|
|
237 |
" ization) \n", |
|
|
238 |
" \n", |
|
|
239 |
" conv2_block3_1_relu (Activatio (None, 56, 56, 64) 0 ['conv2_block3_1_bn[0][0]'] \n", |
|
|
240 |
" n) \n", |
|
|
241 |
" \n", |
|
|
242 |
" conv2_block3_2_conv (Conv2D) (None, 56, 56, 64) 36928 ['conv2_block3_1_relu[0][0]'] \n", |
|
|
243 |
" \n", |
|
|
244 |
" conv2_block3_2_bn (BatchNormal (None, 56, 56, 64) 256 ['conv2_block3_2_conv[0][0]'] \n", |
|
|
245 |
" ization) \n", |
|
|
246 |
" \n", |
|
|
247 |
" conv2_block3_2_relu (Activatio (None, 56, 56, 64) 0 ['conv2_block3_2_bn[0][0]'] \n", |
|
|
248 |
" n) \n", |
|
|
249 |
" \n", |
|
|
250 |
" conv2_block3_3_conv (Conv2D) (None, 56, 56, 256) 16640 ['conv2_block3_2_relu[0][0]'] \n", |
|
|
251 |
" \n", |
|
|
252 |
" conv2_block3_3_bn (BatchNormal (None, 56, 56, 256) 1024 ['conv2_block3_3_conv[0][0]'] \n", |
|
|
253 |
" ization) \n", |
|
|
254 |
" \n", |
|
|
255 |
" conv2_block3_add (Add) (None, 56, 56, 256) 0 ['conv2_block2_out[0][0]', \n", |
|
|
256 |
" 'conv2_block3_3_bn[0][0]'] \n", |
|
|
257 |
" \n", |
|
|
258 |
" conv2_block3_out (Activation) (None, 56, 56, 256) 0 ['conv2_block3_add[0][0]'] \n", |
|
|
259 |
" \n", |
|
|
260 |
" conv3_block1_1_conv (Conv2D) (None, 28, 28, 128) 32896 ['conv2_block3_out[0][0]'] \n", |
|
|
261 |
" \n", |
|
|
262 |
" conv3_block1_1_bn (BatchNormal (None, 28, 28, 128) 512 ['conv3_block1_1_conv[0][0]'] \n", |
|
|
263 |
" ization) \n", |
|
|
264 |
" \n", |
|
|
265 |
" conv3_block1_1_relu (Activatio (None, 28, 28, 128) 0 ['conv3_block1_1_bn[0][0]'] \n", |
|
|
266 |
" n) \n", |
|
|
267 |
" \n", |
|
|
268 |
" conv3_block1_2_conv (Conv2D) (None, 28, 28, 128) 147584 ['conv3_block1_1_relu[0][0]'] \n", |
|
|
269 |
" \n", |
|
|
270 |
" conv3_block1_2_bn (BatchNormal (None, 28, 28, 128) 512 ['conv3_block1_2_conv[0][0]'] \n", |
|
|
271 |
" ization) \n", |
|
|
272 |
" \n", |
|
|
273 |
" conv3_block1_2_relu (Activatio (None, 28, 28, 128) 0 ['conv3_block1_2_bn[0][0]'] \n", |
|
|
274 |
" n) \n", |
|
|
275 |
" \n", |
|
|
276 |
" conv3_block1_0_conv (Conv2D) (None, 28, 28, 512) 131584 ['conv2_block3_out[0][0]'] \n", |
|
|
277 |
" \n", |
|
|
278 |
" conv3_block1_3_conv (Conv2D) (None, 28, 28, 512) 66048 ['conv3_block1_2_relu[0][0]'] \n", |
|
|
279 |
" \n", |
|
|
280 |
" conv3_block1_0_bn (BatchNormal (None, 28, 28, 512) 2048 ['conv3_block1_0_conv[0][0]'] \n", |
|
|
281 |
" ization) \n", |
|
|
282 |
" \n", |
|
|
283 |
" conv3_block1_3_bn (BatchNormal (None, 28, 28, 512) 2048 ['conv3_block1_3_conv[0][0]'] \n", |
|
|
284 |
" ization) \n", |
|
|
285 |
" \n", |
|
|
286 |
" conv3_block1_add (Add) (None, 28, 28, 512) 0 ['conv3_block1_0_bn[0][0]', \n", |
|
|
287 |
" 'conv3_block1_3_bn[0][0]'] \n", |
|
|
288 |
" \n", |
|
|
289 |
" conv3_block1_out (Activation) (None, 28, 28, 512) 0 ['conv3_block1_add[0][0]'] \n", |
|
|
290 |
" \n", |
|
|
291 |
" conv3_block2_1_conv (Conv2D) (None, 28, 28, 128) 65664 ['conv3_block1_out[0][0]'] \n", |
|
|
292 |
" \n", |
|
|
293 |
" conv3_block2_1_bn (BatchNormal (None, 28, 28, 128) 512 ['conv3_block2_1_conv[0][0]'] \n", |
|
|
294 |
" ization) \n", |
|
|
295 |
" \n", |
|
|
296 |
" conv3_block2_1_relu (Activatio (None, 28, 28, 128) 0 ['conv3_block2_1_bn[0][0]'] \n", |
|
|
297 |
" n) \n", |
|
|
298 |
" \n", |
|
|
299 |
" conv3_block2_2_conv (Conv2D) (None, 28, 28, 128) 147584 ['conv3_block2_1_relu[0][0]'] \n", |
|
|
300 |
" \n", |
|
|
301 |
" conv3_block2_2_bn (BatchNormal (None, 28, 28, 128) 512 ['conv3_block2_2_conv[0][0]'] \n", |
|
|
302 |
" ization) \n", |
|
|
303 |
" \n", |
|
|
304 |
" conv3_block2_2_relu (Activatio (None, 28, 28, 128) 0 ['conv3_block2_2_bn[0][0]'] \n", |
|
|
305 |
" n) \n", |
|
|
306 |
" \n", |
|
|
307 |
" conv3_block2_3_conv (Conv2D) (None, 28, 28, 512) 66048 ['conv3_block2_2_relu[0][0]'] \n", |
|
|
308 |
" \n", |
|
|
309 |
" conv3_block2_3_bn (BatchNormal (None, 28, 28, 512) 2048 ['conv3_block2_3_conv[0][0]'] \n", |
|
|
310 |
" ization) \n", |
|
|
311 |
" \n", |
|
|
312 |
" conv3_block2_add (Add) (None, 28, 28, 512) 0 ['conv3_block1_out[0][0]', \n", |
|
|
313 |
" 'conv3_block2_3_bn[0][0]'] \n", |
|
|
314 |
" \n", |
|
|
315 |
" conv3_block2_out (Activation) (None, 28, 28, 512) 0 ['conv3_block2_add[0][0]'] \n", |
|
|
316 |
" \n", |
|
|
317 |
" conv3_block3_1_conv (Conv2D) (None, 28, 28, 128) 65664 ['conv3_block2_out[0][0]'] \n", |
|
|
318 |
" \n", |
|
|
319 |
" conv3_block3_1_bn (BatchNormal (None, 28, 28, 128) 512 ['conv3_block3_1_conv[0][0]'] \n", |
|
|
320 |
" ization) \n", |
|
|
321 |
" \n", |
|
|
322 |
" conv3_block3_1_relu (Activatio (None, 28, 28, 128) 0 ['conv3_block3_1_bn[0][0]'] \n", |
|
|
323 |
" n) \n", |
|
|
324 |
" \n", |
|
|
325 |
" conv3_block3_2_conv (Conv2D) (None, 28, 28, 128) 147584 ['conv3_block3_1_relu[0][0]'] \n", |
|
|
326 |
" \n", |
|
|
327 |
" conv3_block3_2_bn (BatchNormal (None, 28, 28, 128) 512 ['conv3_block3_2_conv[0][0]'] \n", |
|
|
328 |
" ization) \n", |
|
|
329 |
" \n", |
|
|
330 |
" conv3_block3_2_relu (Activatio (None, 28, 28, 128) 0 ['conv3_block3_2_bn[0][0]'] \n", |
|
|
331 |
" n) \n", |
|
|
332 |
" \n", |
|
|
333 |
" conv3_block3_3_conv (Conv2D) (None, 28, 28, 512) 66048 ['conv3_block3_2_relu[0][0]'] \n", |
|
|
334 |
" \n", |
|
|
335 |
" conv3_block3_3_bn (BatchNormal (None, 28, 28, 512) 2048 ['conv3_block3_3_conv[0][0]'] \n", |
|
|
336 |
" ization) \n", |
|
|
337 |
" \n", |
|
|
338 |
" conv3_block3_add (Add) (None, 28, 28, 512) 0 ['conv3_block2_out[0][0]', \n", |
|
|
339 |
" 'conv3_block3_3_bn[0][0]'] \n", |
|
|
340 |
" \n", |
|
|
341 |
" conv3_block3_out (Activation) (None, 28, 28, 512) 0 ['conv3_block3_add[0][0]'] \n", |
|
|
342 |
" \n", |
|
|
343 |
" conv3_block4_1_conv (Conv2D) (None, 28, 28, 128) 65664 ['conv3_block3_out[0][0]'] \n", |
|
|
344 |
" \n", |
|
|
345 |
" conv3_block4_1_bn (BatchNormal (None, 28, 28, 128) 512 ['conv3_block4_1_conv[0][0]'] \n", |
|
|
346 |
" ization) \n", |
|
|
347 |
" \n", |
|
|
348 |
" conv3_block4_1_relu (Activatio (None, 28, 28, 128) 0 ['conv3_block4_1_bn[0][0]'] \n", |
|
|
349 |
" n) \n", |
|
|
350 |
" \n", |
|
|
351 |
" conv3_block4_2_conv (Conv2D) (None, 28, 28, 128) 147584 ['conv3_block4_1_relu[0][0]'] \n", |
|
|
352 |
" \n", |
|
|
353 |
" conv3_block4_2_bn (BatchNormal (None, 28, 28, 128) 512 ['conv3_block4_2_conv[0][0]'] \n", |
|
|
354 |
" ization) \n", |
|
|
355 |
" \n", |
|
|
356 |
" conv3_block4_2_relu (Activatio (None, 28, 28, 128) 0 ['conv3_block4_2_bn[0][0]'] \n", |
|
|
357 |
" n) \n", |
|
|
358 |
" \n", |
|
|
359 |
" conv3_block4_3_conv (Conv2D) (None, 28, 28, 512) 66048 ['conv3_block4_2_relu[0][0]'] \n", |
|
|
360 |
" \n", |
|
|
361 |
" conv3_block4_3_bn (BatchNormal (None, 28, 28, 512) 2048 ['conv3_block4_3_conv[0][0]'] \n", |
|
|
362 |
" ization) \n", |
|
|
363 |
" \n", |
|
|
364 |
" conv3_block4_add (Add) (None, 28, 28, 512) 0 ['conv3_block3_out[0][0]', \n", |
|
|
365 |
" 'conv3_block4_3_bn[0][0]'] \n", |
|
|
366 |
" \n", |
|
|
367 |
" conv3_block4_out (Activation) (None, 28, 28, 512) 0 ['conv3_block4_add[0][0]'] \n", |
|
|
368 |
" \n", |
|
|
369 |
" conv4_block1_1_conv (Conv2D) (None, 14, 14, 256) 131328 ['conv3_block4_out[0][0]'] \n", |
|
|
370 |
" \n", |
|
|
371 |
" conv4_block1_1_bn (BatchNormal (None, 14, 14, 256) 1024 ['conv4_block1_1_conv[0][0]'] \n", |
|
|
372 |
" ization) \n", |
|
|
373 |
" \n", |
|
|
374 |
" conv4_block1_1_relu (Activatio (None, 14, 14, 256) 0 ['conv4_block1_1_bn[0][0]'] \n", |
|
|
375 |
" n) \n", |
|
|
376 |
" \n", |
|
|
377 |
" conv4_block1_2_conv (Conv2D) (None, 14, 14, 256) 590080 ['conv4_block1_1_relu[0][0]'] \n", |
|
|
378 |
" \n", |
|
|
379 |
" conv4_block1_2_bn (BatchNormal (None, 14, 14, 256) 1024 ['conv4_block1_2_conv[0][0]'] \n", |
|
|
380 |
" ization) \n", |
|
|
381 |
" \n", |
|
|
382 |
" conv4_block1_2_relu (Activatio (None, 14, 14, 256) 0 ['conv4_block1_2_bn[0][0]'] \n", |
|
|
383 |
" n) \n", |
|
|
384 |
" \n", |
|
|
385 |
" conv4_block1_0_conv (Conv2D) (None, 14, 14, 1024 525312 ['conv3_block4_out[0][0]'] \n", |
|
|
386 |
" ) \n", |
|
|
387 |
" \n", |
|
|
388 |
" conv4_block1_3_conv (Conv2D) (None, 14, 14, 1024 263168 ['conv4_block1_2_relu[0][0]'] \n", |
|
|
389 |
" ) \n", |
|
|
390 |
" \n", |
|
|
391 |
" conv4_block1_0_bn (BatchNormal (None, 14, 14, 1024 4096 ['conv4_block1_0_conv[0][0]'] \n", |
|
|
392 |
" ization) ) \n", |
|
|
393 |
" \n", |
|
|
394 |
" conv4_block1_3_bn (BatchNormal (None, 14, 14, 1024 4096 ['conv4_block1_3_conv[0][0]'] \n", |
|
|
395 |
" ization) ) \n", |
|
|
396 |
" \n", |
|
|
397 |
" conv4_block1_add (Add) (None, 14, 14, 1024 0 ['conv4_block1_0_bn[0][0]', \n", |
|
|
398 |
" ) 'conv4_block1_3_bn[0][0]'] \n", |
|
|
399 |
" \n", |
|
|
400 |
" conv4_block1_out (Activation) (None, 14, 14, 1024 0 ['conv4_block1_add[0][0]'] \n", |
|
|
401 |
" ) \n", |
|
|
402 |
" \n", |
|
|
403 |
" conv4_block2_1_conv (Conv2D) (None, 14, 14, 256) 262400 ['conv4_block1_out[0][0]'] \n", |
|
|
404 |
" \n", |
|
|
405 |
" conv4_block2_1_bn (BatchNormal (None, 14, 14, 256) 1024 ['conv4_block2_1_conv[0][0]'] \n", |
|
|
406 |
" ization) \n", |
|
|
407 |
" \n", |
|
|
408 |
" conv4_block2_1_relu (Activatio (None, 14, 14, 256) 0 ['conv4_block2_1_bn[0][0]'] \n", |
|
|
409 |
" n) \n", |
|
|
410 |
" \n", |
|
|
411 |
" conv4_block2_2_conv (Conv2D) (None, 14, 14, 256) 590080 ['conv4_block2_1_relu[0][0]'] \n", |
|
|
412 |
" \n", |
|
|
413 |
" conv4_block2_2_bn (BatchNormal (None, 14, 14, 256) 1024 ['conv4_block2_2_conv[0][0]'] \n", |
|
|
414 |
" ization) \n", |
|
|
415 |
" \n", |
|
|
416 |
" conv4_block2_2_relu (Activatio (None, 14, 14, 256) 0 ['conv4_block2_2_bn[0][0]'] \n", |
|
|
417 |
" n) \n", |
|
|
418 |
" \n", |
|
|
419 |
" conv4_block2_3_conv (Conv2D) (None, 14, 14, 1024 263168 ['conv4_block2_2_relu[0][0]'] \n", |
|
|
420 |
" ) \n", |
|
|
421 |
" \n", |
|
|
422 |
" conv4_block2_3_bn (BatchNormal (None, 14, 14, 1024 4096 ['conv4_block2_3_conv[0][0]'] \n", |
|
|
423 |
" ization) ) \n", |
|
|
424 |
" \n", |
|
|
425 |
" conv4_block2_add (Add) (None, 14, 14, 1024 0 ['conv4_block1_out[0][0]', \n", |
|
|
426 |
" ) 'conv4_block2_3_bn[0][0]'] \n", |
|
|
427 |
" \n", |
|
|
428 |
" conv4_block2_out (Activation) (None, 14, 14, 1024 0 ['conv4_block2_add[0][0]'] \n", |
|
|
429 |
" ) \n", |
|
|
430 |
" \n", |
|
|
431 |
" conv4_block3_1_conv (Conv2D) (None, 14, 14, 256) 262400 ['conv4_block2_out[0][0]'] \n", |
|
|
432 |
" \n", |
|
|
433 |
" conv4_block3_1_bn (BatchNormal (None, 14, 14, 256) 1024 ['conv4_block3_1_conv[0][0]'] \n", |
|
|
434 |
" ization) \n", |
|
|
435 |
" \n", |
|
|
436 |
" conv4_block3_1_relu (Activatio (None, 14, 14, 256) 0 ['conv4_block3_1_bn[0][0]'] \n", |
|
|
437 |
" n) \n", |
|
|
438 |
" \n", |
|
|
439 |
" conv4_block3_2_conv (Conv2D) (None, 14, 14, 256) 590080 ['conv4_block3_1_relu[0][0]'] \n", |
|
|
440 |
" \n", |
|
|
441 |
" conv4_block3_2_bn (BatchNormal (None, 14, 14, 256) 1024 ['conv4_block3_2_conv[0][0]'] \n", |
|
|
442 |
" ization) \n", |
|
|
443 |
" \n", |
|
|
444 |
" conv4_block3_2_relu (Activatio (None, 14, 14, 256) 0 ['conv4_block3_2_bn[0][0]'] \n", |
|
|
445 |
" n) \n", |
|
|
446 |
" \n", |
|
|
447 |
" conv4_block3_3_conv (Conv2D) (None, 14, 14, 1024 263168 ['conv4_block3_2_relu[0][0]'] \n", |
|
|
448 |
" ) \n", |
|
|
449 |
" \n", |
|
|
450 |
" conv4_block3_3_bn (BatchNormal (None, 14, 14, 1024 4096 ['conv4_block3_3_conv[0][0]'] \n", |
|
|
451 |
" ization) ) \n", |
|
|
452 |
" \n", |
|
|
453 |
" conv4_block3_add (Add) (None, 14, 14, 1024 0 ['conv4_block2_out[0][0]', \n", |
|
|
454 |
" ) 'conv4_block3_3_bn[0][0]'] \n", |
|
|
455 |
" \n", |
|
|
456 |
" conv4_block3_out (Activation) (None, 14, 14, 1024 0 ['conv4_block3_add[0][0]'] \n", |
|
|
457 |
" ) \n", |
|
|
458 |
" \n", |
|
|
459 |
" conv4_block4_1_conv (Conv2D) (None, 14, 14, 256) 262400 ['conv4_block3_out[0][0]'] \n", |
|
|
460 |
" \n", |
|
|
461 |
" conv4_block4_1_bn (BatchNormal (None, 14, 14, 256) 1024 ['conv4_block4_1_conv[0][0]'] \n", |
|
|
462 |
" ization) \n", |
|
|
463 |
" \n", |
|
|
464 |
" conv4_block4_1_relu (Activatio (None, 14, 14, 256) 0 ['conv4_block4_1_bn[0][0]'] \n", |
|
|
465 |
" n) \n", |
|
|
466 |
" \n", |
|
|
467 |
" conv4_block4_2_conv (Conv2D) (None, 14, 14, 256) 590080 ['conv4_block4_1_relu[0][0]'] \n", |
|
|
468 |
" \n", |
|
|
469 |
" conv4_block4_2_bn (BatchNormal (None, 14, 14, 256) 1024 ['conv4_block4_2_conv[0][0]'] \n", |
|
|
470 |
" ization) \n", |
|
|
471 |
" \n", |
|
|
472 |
" conv4_block4_2_relu (Activatio (None, 14, 14, 256) 0 ['conv4_block4_2_bn[0][0]'] \n", |
|
|
473 |
" n) \n", |
|
|
474 |
" \n", |
|
|
475 |
" conv4_block4_3_conv (Conv2D) (None, 14, 14, 1024 263168 ['conv4_block4_2_relu[0][0]'] \n", |
|
|
476 |
" ) \n", |
|
|
477 |
" \n", |
|
|
478 |
" conv4_block4_3_bn (BatchNormal (None, 14, 14, 1024 4096 ['conv4_block4_3_conv[0][0]'] \n", |
|
|
479 |
" ization) ) \n", |
|
|
480 |
" \n", |
|
|
481 |
" conv4_block4_add (Add) (None, 14, 14, 1024 0 ['conv4_block3_out[0][0]', \n", |
|
|
482 |
" ) 'conv4_block4_3_bn[0][0]'] \n", |
|
|
483 |
" \n", |
|
|
484 |
" conv4_block4_out (Activation) (None, 14, 14, 1024 0 ['conv4_block4_add[0][0]'] \n", |
|
|
485 |
" ) \n", |
|
|
486 |
" \n", |
|
|
487 |
" conv4_block5_1_conv (Conv2D) (None, 14, 14, 256) 262400 ['conv4_block4_out[0][0]'] \n", |
|
|
488 |
" \n", |
|
|
489 |
" conv4_block5_1_bn (BatchNormal (None, 14, 14, 256) 1024 ['conv4_block5_1_conv[0][0]'] \n", |
|
|
490 |
" ization) \n", |
|
|
491 |
" \n", |
|
|
492 |
" conv4_block5_1_relu (Activatio (None, 14, 14, 256) 0 ['conv4_block5_1_bn[0][0]'] \n", |
|
|
493 |
" n) \n", |
|
|
494 |
" \n", |
|
|
495 |
" conv4_block5_2_conv (Conv2D) (None, 14, 14, 256) 590080 ['conv4_block5_1_relu[0][0]'] \n", |
|
|
496 |
" \n", |
|
|
497 |
" conv4_block5_2_bn (BatchNormal (None, 14, 14, 256) 1024 ['conv4_block5_2_conv[0][0]'] \n", |
|
|
498 |
" ization) \n", |
|
|
499 |
" \n", |
|
|
500 |
" conv4_block5_2_relu (Activatio (None, 14, 14, 256) 0 ['conv4_block5_2_bn[0][0]'] \n", |
|
|
501 |
" n) \n", |
|
|
502 |
" \n", |
|
|
503 |
" conv4_block5_3_conv (Conv2D) (None, 14, 14, 1024 263168 ['conv4_block5_2_relu[0][0]'] \n", |
|
|
504 |
" ) \n", |
|
|
505 |
" \n", |
|
|
506 |
" conv4_block5_3_bn (BatchNormal (None, 14, 14, 1024 4096 ['conv4_block5_3_conv[0][0]'] \n", |
|
|
507 |
" ization) ) \n", |
|
|
508 |
" \n", |
|
|
509 |
" conv4_block5_add (Add) (None, 14, 14, 1024 0 ['conv4_block4_out[0][0]', \n", |
|
|
510 |
" ) 'conv4_block5_3_bn[0][0]'] \n", |
|
|
511 |
" \n", |
|
|
512 |
" conv4_block5_out (Activation) (None, 14, 14, 1024 0 ['conv4_block5_add[0][0]'] \n", |
|
|
513 |
" ) \n", |
|
|
514 |
" \n", |
|
|
515 |
" conv4_block6_1_conv (Conv2D) (None, 14, 14, 256) 262400 ['conv4_block5_out[0][0]'] \n", |
|
|
516 |
" \n", |
|
|
517 |
" conv4_block6_1_bn (BatchNormal (None, 14, 14, 256) 1024 ['conv4_block6_1_conv[0][0]'] \n", |
|
|
518 |
" ization) \n", |
|
|
519 |
" \n", |
|
|
520 |
" conv4_block6_1_relu (Activatio (None, 14, 14, 256) 0 ['conv4_block6_1_bn[0][0]'] \n", |
|
|
521 |
" n) \n", |
|
|
522 |
" \n", |
|
|
523 |
" conv4_block6_2_conv (Conv2D) (None, 14, 14, 256) 590080 ['conv4_block6_1_relu[0][0]'] \n", |
|
|
524 |
" \n", |
|
|
525 |
" conv4_block6_2_bn (BatchNormal (None, 14, 14, 256) 1024 ['conv4_block6_2_conv[0][0]'] \n", |
|
|
526 |
" ization) \n", |
|
|
527 |
" \n", |
|
|
528 |
" conv4_block6_2_relu (Activatio (None, 14, 14, 256) 0 ['conv4_block6_2_bn[0][0]'] \n", |
|
|
529 |
" n) \n", |
|
|
530 |
" \n", |
|
|
531 |
" conv4_block6_3_conv (Conv2D) (None, 14, 14, 1024 263168 ['conv4_block6_2_relu[0][0]'] \n", |
|
|
532 |
" ) \n", |
|
|
533 |
" \n", |
|
|
534 |
" conv4_block6_3_bn (BatchNormal (None, 14, 14, 1024 4096 ['conv4_block6_3_conv[0][0]'] \n", |
|
|
535 |
" ization) ) \n", |
|
|
536 |
" \n", |
|
|
537 |
" conv4_block6_add (Add) (None, 14, 14, 1024 0 ['conv4_block5_out[0][0]', \n", |
|
|
538 |
" ) 'conv4_block6_3_bn[0][0]'] \n", |
|
|
539 |
" \n", |
|
|
540 |
" conv4_block6_out (Activation) (None, 14, 14, 1024 0 ['conv4_block6_add[0][0]'] \n", |
|
|
541 |
" ) \n", |
|
|
542 |
" \n", |
|
|
543 |
" conv5_block1_1_conv (Conv2D) (None, 7, 7, 512) 524800 ['conv4_block6_out[0][0]'] \n", |
|
|
544 |
" \n", |
|
|
545 |
" conv5_block1_1_bn (BatchNormal (None, 7, 7, 512) 2048 ['conv5_block1_1_conv[0][0]'] \n", |
|
|
546 |
" ization) \n", |
|
|
547 |
" \n", |
|
|
548 |
" conv5_block1_1_relu (Activatio (None, 7, 7, 512) 0 ['conv5_block1_1_bn[0][0]'] \n", |
|
|
549 |
" n) \n", |
|
|
550 |
" \n", |
|
|
551 |
" conv5_block1_2_conv (Conv2D) (None, 7, 7, 512) 2359808 ['conv5_block1_1_relu[0][0]'] \n", |
|
|
552 |
" \n", |
|
|
553 |
" conv5_block1_2_bn (BatchNormal (None, 7, 7, 512) 2048 ['conv5_block1_2_conv[0][0]'] \n", |
|
|
554 |
" ization) \n", |
|
|
555 |
" \n", |
|
|
556 |
" conv5_block1_2_relu (Activatio (None, 7, 7, 512) 0 ['conv5_block1_2_bn[0][0]'] \n", |
|
|
557 |
" n) \n", |
|
|
558 |
" \n", |
|
|
559 |
" conv5_block1_0_conv (Conv2D) (None, 7, 7, 2048) 2099200 ['conv4_block6_out[0][0]'] \n", |
|
|
560 |
" \n", |
|
|
561 |
" conv5_block1_3_conv (Conv2D) (None, 7, 7, 2048) 1050624 ['conv5_block1_2_relu[0][0]'] \n", |
|
|
562 |
" \n", |
|
|
563 |
" conv5_block1_0_bn (BatchNormal (None, 7, 7, 2048) 8192 ['conv5_block1_0_conv[0][0]'] \n", |
|
|
564 |
" ization) \n", |
|
|
565 |
" \n", |
|
|
566 |
" conv5_block1_3_bn (BatchNormal (None, 7, 7, 2048) 8192 ['conv5_block1_3_conv[0][0]'] \n", |
|
|
567 |
" ization) \n", |
|
|
568 |
" \n", |
|
|
569 |
" conv5_block1_add (Add) (None, 7, 7, 2048) 0 ['conv5_block1_0_bn[0][0]', \n", |
|
|
570 |
" 'conv5_block1_3_bn[0][0]'] \n", |
|
|
571 |
" \n", |
|
|
572 |
" conv5_block1_out (Activation) (None, 7, 7, 2048) 0 ['conv5_block1_add[0][0]'] \n", |
|
|
573 |
" \n", |
|
|
574 |
" conv5_block2_1_conv (Conv2D) (None, 7, 7, 512) 1049088 ['conv5_block1_out[0][0]'] \n", |
|
|
575 |
" \n", |
|
|
576 |
" conv5_block2_1_bn (BatchNormal (None, 7, 7, 512) 2048 ['conv5_block2_1_conv[0][0]'] \n", |
|
|
577 |
" ization) \n", |
|
|
578 |
" \n", |
|
|
579 |
" conv5_block2_1_relu (Activatio (None, 7, 7, 512) 0 ['conv5_block2_1_bn[0][0]'] \n", |
|
|
580 |
" n) \n", |
|
|
581 |
" \n", |
|
|
582 |
" conv5_block2_2_conv (Conv2D) (None, 7, 7, 512) 2359808 ['conv5_block2_1_relu[0][0]'] \n", |
|
|
583 |
" \n", |
|
|
584 |
" conv5_block2_2_bn (BatchNormal (None, 7, 7, 512) 2048 ['conv5_block2_2_conv[0][0]'] \n", |
|
|
585 |
" ization) \n", |
|
|
586 |
" \n", |
|
|
587 |
" conv5_block2_2_relu (Activatio (None, 7, 7, 512) 0 ['conv5_block2_2_bn[0][0]'] \n", |
|
|
588 |
" n) \n", |
|
|
589 |
" \n", |
|
|
590 |
" conv5_block2_3_conv (Conv2D) (None, 7, 7, 2048) 1050624 ['conv5_block2_2_relu[0][0]'] \n", |
|
|
591 |
" \n", |
|
|
592 |
" conv5_block2_3_bn (BatchNormal (None, 7, 7, 2048) 8192 ['conv5_block2_3_conv[0][0]'] \n", |
|
|
593 |
" ization) \n", |
|
|
594 |
" \n", |
|
|
595 |
" conv5_block2_add (Add) (None, 7, 7, 2048) 0 ['conv5_block1_out[0][0]', \n", |
|
|
596 |
" 'conv5_block2_3_bn[0][0]'] \n", |
|
|
597 |
" \n", |
|
|
598 |
" conv5_block2_out (Activation) (None, 7, 7, 2048) 0 ['conv5_block2_add[0][0]'] \n", |
|
|
599 |
" \n", |
|
|
600 |
" conv5_block3_1_conv (Conv2D) (None, 7, 7, 512) 1049088 ['conv5_block2_out[0][0]'] \n", |
|
|
601 |
" \n", |
|
|
602 |
" conv5_block3_1_bn (BatchNormal (None, 7, 7, 512) 2048 ['conv5_block3_1_conv[0][0]'] \n", |
|
|
603 |
" ization) \n", |
|
|
604 |
" \n", |
|
|
605 |
" conv5_block3_1_relu (Activatio (None, 7, 7, 512) 0 ['conv5_block3_1_bn[0][0]'] \n", |
|
|
606 |
" n) \n", |
|
|
607 |
" \n", |
|
|
608 |
" conv5_block3_2_conv (Conv2D) (None, 7, 7, 512) 2359808 ['conv5_block3_1_relu[0][0]'] \n", |
|
|
609 |
" \n", |
|
|
610 |
" conv5_block3_2_bn (BatchNormal (None, 7, 7, 512) 2048 ['conv5_block3_2_conv[0][0]'] \n", |
|
|
611 |
" ization) \n", |
|
|
612 |
" \n", |
|
|
613 |
" conv5_block3_2_relu (Activatio (None, 7, 7, 512) 0 ['conv5_block3_2_bn[0][0]'] \n", |
|
|
614 |
" n) \n", |
|
|
615 |
" \n", |
|
|
616 |
" conv5_block3_3_conv (Conv2D) (None, 7, 7, 2048) 1050624 ['conv5_block3_2_relu[0][0]'] \n", |
|
|
617 |
" \n", |
|
|
618 |
" conv5_block3_3_bn (BatchNormal (None, 7, 7, 2048) 8192 ['conv5_block3_3_conv[0][0]'] \n", |
|
|
619 |
" ization) \n", |
|
|
620 |
" \n", |
|
|
621 |
" conv5_block3_add (Add) (None, 7, 7, 2048) 0 ['conv5_block2_out[0][0]', \n", |
|
|
622 |
" 'conv5_block3_3_bn[0][0]'] \n", |
|
|
623 |
" \n", |
|
|
624 |
" conv5_block3_out (Activation) (None, 7, 7, 2048) 0 ['conv5_block3_add[0][0]'] \n", |
|
|
625 |
" \n", |
|
|
626 |
" flatten (Flatten) (None, 100352) 0 ['conv5_block3_out[0][0]'] \n", |
|
|
627 |
" \n", |
|
|
628 |
" dense (Dense) (None, 3) 301059 ['flatten[0][0]'] \n", |
|
|
629 |
" \n", |
|
|
630 |
"==================================================================================================\n", |
|
|
631 |
"Total params: 23,888,771\n", |
|
|
632 |
"Trainable params: 301,059\n", |
|
|
633 |
"Non-trainable params: 23,587,712\n", |
|
|
634 |
"__________________________________________________________________________________________________\n" |
|
|
635 |
] |
|
|
636 |
} |
|
|
637 |
], |
|
|
638 |
"source": [ |
|
|
639 |
"\n", |
|
|
640 |
"# view the structure of the model\n", |
|
|
641 |
"model.summary()\n" |
|
|
642 |
] |
|
|
643 |
}, |
|
|
644 |
{ |
|
|
645 |
"cell_type": "code", |
|
|
646 |
"execution_count": 10, |
|
|
647 |
"metadata": {}, |
|
|
648 |
"outputs": [], |
|
|
649 |
"source": [ |
|
|
650 |
"# tell the model what cost and optimization method to use\n", |
|
|
651 |
"model.compile(\n", |
|
|
652 |
" loss='categorical_crossentropy',\n", |
|
|
653 |
" optimizer='adam',\n", |
|
|
654 |
" metrics=['accuracy']\n", |
|
|
655 |
")\n" |
|
|
656 |
] |
|
|
657 |
}, |
|
|
658 |
{ |
|
|
659 |
"cell_type": "code", |
|
|
660 |
"execution_count": 11, |
|
|
661 |
"metadata": {}, |
|
|
662 |
"outputs": [], |
|
|
663 |
"source": [ |
|
|
664 |
"# Use the Image Data Generator to import the images from the dataset\n", |
|
|
665 |
"from tensorflow.keras.preprocessing.image import ImageDataGenerator\n", |
|
|
666 |
"\n", |
|
|
667 |
"train_datagen = ImageDataGenerator(rescale = 1./255,\n", |
|
|
668 |
" shear_range = 0.2,\n", |
|
|
669 |
" zoom_range = 0.2,\n", |
|
|
670 |
" horizontal_flip = True)\n", |
|
|
671 |
"\n", |
|
|
672 |
"test_datagen = ImageDataGenerator(rescale = 1./255)" |
|
|
673 |
] |
|
|
674 |
}, |
|
|
675 |
{ |
|
|
676 |
"cell_type": "code", |
|
|
677 |
"execution_count": 12, |
|
|
678 |
"metadata": {}, |
|
|
679 |
"outputs": [ |
|
|
680 |
{ |
|
|
681 |
"name": "stdout", |
|
|
682 |
"output_type": "stream", |
|
|
683 |
"text": [ |
|
|
684 |
"Found 511 images belonging to 3 classes.\n" |
|
|
685 |
] |
|
|
686 |
} |
|
|
687 |
], |
|
|
688 |
"source": [ |
|
|
689 |
"# Make sure you provide the same target size as initialied for the image size\n", |
|
|
690 |
"training_set = train_datagen.flow_from_directory('Datasets/Datasets/train',\n", |
|
|
691 |
" target_size = (224, 224),\n", |
|
|
692 |
" batch_size = 32,\n", |
|
|
693 |
" class_mode = 'categorical')" |
|
|
694 |
] |
|
|
695 |
}, |
|
|
696 |
{ |
|
|
697 |
"cell_type": "code", |
|
|
698 |
"execution_count": 13, |
|
|
699 |
"metadata": {}, |
|
|
700 |
"outputs": [ |
|
|
701 |
{ |
|
|
702 |
"name": "stdout", |
|
|
703 |
"output_type": "stream", |
|
|
704 |
"text": [ |
|
|
705 |
"Found 37 images belonging to 3 classes.\n" |
|
|
706 |
] |
|
|
707 |
} |
|
|
708 |
], |
|
|
709 |
"source": [ |
|
|
710 |
"test_set = test_datagen.flow_from_directory('Datasets/Datasets/test',\n", |
|
|
711 |
" target_size = (224, 224),\n", |
|
|
712 |
" batch_size = 32,\n", |
|
|
713 |
" class_mode = 'categorical')" |
|
|
714 |
] |
|
|
715 |
}, |
|
|
716 |
{ |
|
|
717 |
"cell_type": "code", |
|
|
718 |
"execution_count": 14, |
|
|
719 |
"metadata": {}, |
|
|
720 |
"outputs": [ |
|
|
721 |
{ |
|
|
722 |
"name": "stderr", |
|
|
723 |
"output_type": "stream", |
|
|
724 |
"text": [ |
|
|
725 |
"<ipython-input-14-69229fe26ea3>:3: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n", |
|
|
726 |
" r = model.fit_generator(\n" |
|
|
727 |
] |
|
|
728 |
}, |
|
|
729 |
{ |
|
|
730 |
"name": "stdout", |
|
|
731 |
"output_type": "stream", |
|
|
732 |
"text": [ |
|
|
733 |
"Epoch 1/50\n", |
|
|
734 |
"16/16 [==============================] - 67s 4s/step - loss: 7.0150 - accuracy: 0.3718 - val_loss: 2.0991 - val_accuracy: 0.3243\n", |
|
|
735 |
"Epoch 2/50\n", |
|
|
736 |
"16/16 [==============================] - 59s 4s/step - loss: 3.2412 - accuracy: 0.3777 - val_loss: 1.9412 - val_accuracy: 0.6216\n", |
|
|
737 |
"Epoch 3/50\n", |
|
|
738 |
"16/16 [==============================] - 59s 4s/step - loss: 1.6098 - accuracy: 0.4325 - val_loss: 1.0955 - val_accuracy: 0.4865\n", |
|
|
739 |
"Epoch 4/50\n", |
|
|
740 |
"16/16 [==============================] - 59s 4s/step - loss: 1.0142 - accuracy: 0.5108 - val_loss: 1.1076 - val_accuracy: 0.4595\n", |
|
|
741 |
"Epoch 5/50\n", |
|
|
742 |
"16/16 [==============================] - 59s 4s/step - loss: 0.8595 - accuracy: 0.6458 - val_loss: 0.5441 - val_accuracy: 0.8378\n", |
|
|
743 |
"Epoch 6/50\n", |
|
|
744 |
"16/16 [==============================] - 59s 4s/step - loss: 0.6768 - accuracy: 0.7221 - val_loss: 0.5361 - val_accuracy: 0.7838\n", |
|
|
745 |
"Epoch 7/50\n", |
|
|
746 |
"16/16 [==============================] - 57s 4s/step - loss: 0.8178 - accuracy: 0.6595 - val_loss: 0.4321 - val_accuracy: 0.9189\n", |
|
|
747 |
"Epoch 8/50\n", |
|
|
748 |
"16/16 [==============================] - 58s 4s/step - loss: 0.6431 - accuracy: 0.7358 - val_loss: 0.6896 - val_accuracy: 0.5405\n", |
|
|
749 |
"Epoch 9/50\n", |
|
|
750 |
"16/16 [==============================] - 58s 4s/step - loss: 0.5735 - accuracy: 0.8023 - val_loss: 0.4200 - val_accuracy: 0.9189\n", |
|
|
751 |
"Epoch 10/50\n", |
|
|
752 |
"16/16 [==============================] - 59s 4s/step - loss: 0.5606 - accuracy: 0.7632 - val_loss: 0.4029 - val_accuracy: 0.8378\n", |
|
|
753 |
"Epoch 11/50\n", |
|
|
754 |
"16/16 [==============================] - 59s 4s/step - loss: 0.5615 - accuracy: 0.7554 - val_loss: 0.4212 - val_accuracy: 0.8108\n", |
|
|
755 |
"Epoch 12/50\n", |
|
|
756 |
"16/16 [==============================] - 59s 4s/step - loss: 0.4745 - accuracy: 0.8317 - val_loss: 0.3112 - val_accuracy: 0.9730\n", |
|
|
757 |
"Epoch 13/50\n", |
|
|
758 |
"16/16 [==============================] - 59s 4s/step - loss: 0.4810 - accuracy: 0.8297 - val_loss: 0.3584 - val_accuracy: 0.9189\n", |
|
|
759 |
"Epoch 14/50\n", |
|
|
760 |
"16/16 [==============================] - 59s 4s/step - loss: 0.4991 - accuracy: 0.8102 - val_loss: 0.3882 - val_accuracy: 0.9189\n", |
|
|
761 |
"Epoch 15/50\n", |
|
|
762 |
"16/16 [==============================] - 59s 4s/step - loss: 0.4639 - accuracy: 0.8258 - val_loss: 0.2682 - val_accuracy: 0.9730\n", |
|
|
763 |
"Epoch 16/50\n", |
|
|
764 |
"16/16 [==============================] - 59s 4s/step - loss: 0.4838 - accuracy: 0.8180 - val_loss: 0.2745 - val_accuracy: 0.8919\n", |
|
|
765 |
"Epoch 17/50\n", |
|
|
766 |
"16/16 [==============================] - 59s 4s/step - loss: 0.6137 - accuracy: 0.7084 - val_loss: 0.6421 - val_accuracy: 0.8108\n", |
|
|
767 |
"Epoch 18/50\n", |
|
|
768 |
"16/16 [==============================] - 59s 4s/step - loss: 0.8219 - accuracy: 0.7065 - val_loss: 0.3786 - val_accuracy: 0.8649\n", |
|
|
769 |
"Epoch 19/50\n", |
|
|
770 |
"16/16 [==============================] - 59s 4s/step - loss: 0.6018 - accuracy: 0.7299 - val_loss: 0.2404 - val_accuracy: 0.9730\n", |
|
|
771 |
"Epoch 20/50\n", |
|
|
772 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3808 - accuracy: 0.8611 - val_loss: 0.2123 - val_accuracy: 0.9189\n", |
|
|
773 |
"Epoch 21/50\n", |
|
|
774 |
"16/16 [==============================] - 59s 4s/step - loss: 0.4001 - accuracy: 0.8532 - val_loss: 0.4761 - val_accuracy: 0.8108\n", |
|
|
775 |
"Epoch 22/50\n", |
|
|
776 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3991 - accuracy: 0.8630 - val_loss: 0.2407 - val_accuracy: 0.9189\n", |
|
|
777 |
"Epoch 23/50\n", |
|
|
778 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3520 - accuracy: 0.9022 - val_loss: 0.1834 - val_accuracy: 0.9730\n", |
|
|
779 |
"Epoch 24/50\n", |
|
|
780 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3321 - accuracy: 0.8787 - val_loss: 0.2143 - val_accuracy: 0.9459\n", |
|
|
781 |
"Epoch 25/50\n", |
|
|
782 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3811 - accuracy: 0.8767 - val_loss: 0.1881 - val_accuracy: 0.9730\n", |
|
|
783 |
"Epoch 26/50\n", |
|
|
784 |
"16/16 [==============================] - 59s 4s/step - loss: 0.5631 - accuracy: 0.7593 - val_loss: 0.3551 - val_accuracy: 0.8108\n", |
|
|
785 |
"Epoch 27/50\n", |
|
|
786 |
"16/16 [==============================] - 58s 4s/step - loss: 0.4562 - accuracy: 0.7965 - val_loss: 0.2522 - val_accuracy: 0.8919\n", |
|
|
787 |
"Epoch 28/50\n", |
|
|
788 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3618 - accuracy: 0.8552 - val_loss: 0.1567 - val_accuracy: 0.9730\n", |
|
|
789 |
"Epoch 29/50\n", |
|
|
790 |
"16/16 [==============================] - 59s 4s/step - loss: 0.2896 - accuracy: 0.8982 - val_loss: 0.1706 - val_accuracy: 0.9730\n", |
|
|
791 |
"Epoch 30/50\n", |
|
|
792 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3698 - accuracy: 0.8708 - val_loss: 0.1615 - val_accuracy: 1.0000\n", |
|
|
793 |
"Epoch 31/50\n", |
|
|
794 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3848 - accuracy: 0.8376 - val_loss: 0.3338 - val_accuracy: 0.8919\n", |
|
|
795 |
"Epoch 32/50\n", |
|
|
796 |
"16/16 [==============================] - 59s 4s/step - loss: 0.4109 - accuracy: 0.8023 - val_loss: 0.2486 - val_accuracy: 0.9189\n", |
|
|
797 |
"Epoch 33/50\n", |
|
|
798 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3151 - accuracy: 0.8787 - val_loss: 0.2090 - val_accuracy: 0.8919\n", |
|
|
799 |
"Epoch 34/50\n", |
|
|
800 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3107 - accuracy: 0.8787 - val_loss: 0.1483 - val_accuracy: 0.9730\n", |
|
|
801 |
"Epoch 35/50\n", |
|
|
802 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3467 - accuracy: 0.8748 - val_loss: 0.3154 - val_accuracy: 0.8108\n", |
|
|
803 |
"Epoch 36/50\n", |
|
|
804 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3902 - accuracy: 0.8317 - val_loss: 0.2794 - val_accuracy: 0.8108\n", |
|
|
805 |
"Epoch 37/50\n", |
|
|
806 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3609 - accuracy: 0.8532 - val_loss: 0.2058 - val_accuracy: 0.9189\n", |
|
|
807 |
"Epoch 38/50\n", |
|
|
808 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3740 - accuracy: 0.8493 - val_loss: 0.1270 - val_accuracy: 1.0000\n", |
|
|
809 |
"Epoch 39/50\n", |
|
|
810 |
"16/16 [==============================] - 59s 4s/step - loss: 0.2473 - accuracy: 0.9276 - val_loss: 0.1159 - val_accuracy: 0.9730\n", |
|
|
811 |
"Epoch 40/50\n", |
|
|
812 |
"16/16 [==============================] - 60s 4s/step - loss: 0.3530 - accuracy: 0.8630 - val_loss: 0.3245 - val_accuracy: 0.8378\n", |
|
|
813 |
"Epoch 41/50\n", |
|
|
814 |
"16/16 [==============================] - 59s 4s/step - loss: 0.3621 - accuracy: 0.8552 - val_loss: 0.1345 - val_accuracy: 0.9730\n", |
|
|
815 |
"Epoch 42/50\n", |
|
|
816 |
"16/16 [==============================] - 59s 4s/step - loss: 0.2466 - accuracy: 0.9237 - val_loss: 0.1629 - val_accuracy: 0.9189\n", |
|
|
817 |
"Epoch 43/50\n", |
|
|
818 |
"16/16 [==============================] - 59s 4s/step - loss: 0.2859 - accuracy: 0.8904 - val_loss: 0.1132 - val_accuracy: 0.9730\n", |
|
|
819 |
"Epoch 44/50\n", |
|
|
820 |
"16/16 [==============================] - 59s 4s/step - loss: 0.2872 - accuracy: 0.8982 - val_loss: 0.1673 - val_accuracy: 0.9189\n", |
|
|
821 |
"Epoch 45/50\n", |
|
|
822 |
"16/16 [==============================] - 59s 4s/step - loss: 0.2445 - accuracy: 0.9237 - val_loss: 0.1134 - val_accuracy: 0.9730\n", |
|
|
823 |
"Epoch 46/50\n", |
|
|
824 |
"16/16 [==============================] - 59s 4s/step - loss: 0.2351 - accuracy: 0.9237 - val_loss: 0.1080 - val_accuracy: 0.9730\n", |
|
|
825 |
"Epoch 47/50\n", |
|
|
826 |
"16/16 [==============================] - 59s 4s/step - loss: 0.2560 - accuracy: 0.9100 - val_loss: 0.1088 - val_accuracy: 0.9730\n", |
|
|
827 |
"Epoch 48/50\n", |
|
|
828 |
"16/16 [==============================] - 59s 4s/step - loss: 0.4450 - accuracy: 0.8023 - val_loss: 0.1547 - val_accuracy: 0.9459\n", |
|
|
829 |
"Epoch 49/50\n", |
|
|
830 |
"16/16 [==============================] - 59s 4s/step - loss: 0.2329 - accuracy: 0.9315 - val_loss: 0.1111 - val_accuracy: 1.0000\n", |
|
|
831 |
"Epoch 50/50\n", |
|
|
832 |
"16/16 [==============================] - 59s 4s/step - loss: 0.2553 - accuracy: 0.9002 - val_loss: 0.1550 - val_accuracy: 0.9730\n" |
|
|
833 |
] |
|
|
834 |
} |
|
|
835 |
], |
|
|
836 |
"source": [ |
|
|
837 |
"# fit the model\n", |
|
|
838 |
"# Run the cell. It will take some time to execute\n", |
|
|
839 |
"r = model.fit_generator(\n", |
|
|
840 |
" training_set,\n", |
|
|
841 |
" validation_data=test_set,\n", |
|
|
842 |
" epochs=50,\n", |
|
|
843 |
" steps_per_epoch=len(training_set),\n", |
|
|
844 |
" validation_steps=len(test_set)\n", |
|
|
845 |
")" |
|
|
846 |
] |
|
|
847 |
}, |
|
|
848 |
{ |
|
|
849 |
"cell_type": "code", |
|
|
850 |
"execution_count": 42, |
|
|
851 |
"metadata": {}, |
|
|
852 |
"outputs": [ |
|
|
853 |
{ |
|
|
854 |
"data": { |
|
|
855 |
"text/plain": [ |
|
|
856 |
"{'loss': [7.015007019042969,\n", |
|
|
857 |
" 3.2411556243896484,\n", |
|
|
858 |
" 1.609810471534729,\n", |
|
|
859 |
" 1.0142147541046143,\n", |
|
|
860 |
" 0.8594744801521301,\n", |
|
|
861 |
" 0.6768103241920471,\n", |
|
|
862 |
" 0.8177692890167236,\n", |
|
|
863 |
" 0.6431403160095215,\n", |
|
|
864 |
" 0.5735422372817993,\n", |
|
|
865 |
" 0.5605964064598083,\n", |
|
|
866 |
" 0.5614750385284424,\n", |
|
|
867 |
" 0.4744795262813568,\n", |
|
|
868 |
" 0.48096776008605957,\n", |
|
|
869 |
" 0.499077171087265,\n", |
|
|
870 |
" 0.46386879682540894,\n", |
|
|
871 |
" 0.4837878942489624,\n", |
|
|
872 |
" 0.6136612892150879,\n", |
|
|
873 |
" 0.8219000101089478,\n", |
|
|
874 |
" 0.6018136143684387,\n", |
|
|
875 |
" 0.3807554244995117,\n", |
|
|
876 |
" 0.4001126289367676,\n", |
|
|
877 |
" 0.39909622073173523,\n", |
|
|
878 |
" 0.3519579768180847,\n", |
|
|
879 |
" 0.3321286737918854,\n", |
|
|
880 |
" 0.381106436252594,\n", |
|
|
881 |
" 0.5630961060523987,\n", |
|
|
882 |
" 0.4562273323535919,\n", |
|
|
883 |
" 0.36175066232681274,\n", |
|
|
884 |
" 0.2896212041378021,\n", |
|
|
885 |
" 0.3697938919067383,\n", |
|
|
886 |
" 0.38481977581977844,\n", |
|
|
887 |
" 0.41089171171188354,\n", |
|
|
888 |
" 0.3151184022426605,\n", |
|
|
889 |
" 0.3107375502586365,\n", |
|
|
890 |
" 0.3467033803462982,\n", |
|
|
891 |
" 0.3902468681335449,\n", |
|
|
892 |
" 0.3609190285205841,\n", |
|
|
893 |
" 0.3740045130252838,\n", |
|
|
894 |
" 0.24729087948799133,\n", |
|
|
895 |
" 0.35296347737312317,\n", |
|
|
896 |
" 0.3620986342430115,\n", |
|
|
897 |
" 0.2466459721326828,\n", |
|
|
898 |
" 0.2858617901802063,\n", |
|
|
899 |
" 0.2872360646724701,\n", |
|
|
900 |
" 0.24449540674686432,\n", |
|
|
901 |
" 0.23513354361057281,\n", |
|
|
902 |
" 0.2560056149959564,\n", |
|
|
903 |
" 0.4450059235095978,\n", |
|
|
904 |
" 0.2329322248697281,\n", |
|
|
905 |
" 0.2553144097328186],\n", |
|
|
906 |
" 'accuracy': [0.37181997299194336,\n", |
|
|
907 |
" 0.37769079208374023,\n", |
|
|
908 |
" 0.43248531222343445,\n", |
|
|
909 |
" 0.5107632279396057,\n", |
|
|
910 |
" 0.645792543888092,\n", |
|
|
911 |
" 0.7221134901046753,\n", |
|
|
912 |
" 0.6594911813735962,\n", |
|
|
913 |
" 0.7358121275901794,\n", |
|
|
914 |
" 0.8023483157157898,\n", |
|
|
915 |
" 0.7632094025611877,\n", |
|
|
916 |
" 0.7553815841674805,\n", |
|
|
917 |
" 0.8317025303840637,\n", |
|
|
918 |
" 0.8297455906867981,\n", |
|
|
919 |
" 0.8101761341094971,\n", |
|
|
920 |
" 0.8258317112922668,\n", |
|
|
921 |
" 0.8180038928985596,\n", |
|
|
922 |
" 0.7084148526191711,\n", |
|
|
923 |
" 0.7064579129219055,\n", |
|
|
924 |
" 0.7299413084983826,\n", |
|
|
925 |
" 0.8610567450523376,\n", |
|
|
926 |
" 0.8532289862632751,\n", |
|
|
927 |
" 0.8630136847496033,\n", |
|
|
928 |
" 0.9021526575088501,\n", |
|
|
929 |
" 0.878669261932373,\n", |
|
|
930 |
" 0.8767123222351074,\n", |
|
|
931 |
" 0.7592955231666565,\n", |
|
|
932 |
" 0.7964774966239929,\n", |
|
|
933 |
" 0.8551859259605408,\n", |
|
|
934 |
" 0.8982387185096741,\n", |
|
|
935 |
" 0.8708415031433105,\n", |
|
|
936 |
" 0.8375734090805054,\n", |
|
|
937 |
" 0.8023483157157898,\n", |
|
|
938 |
" 0.878669261932373,\n", |
|
|
939 |
" 0.878669261932373,\n", |
|
|
940 |
" 0.8747553825378418,\n", |
|
|
941 |
" 0.8317025303840637,\n", |
|
|
942 |
" 0.8532289862632751,\n", |
|
|
943 |
" 0.8493150472640991,\n", |
|
|
944 |
" 0.927592933177948,\n", |
|
|
945 |
" 0.8630136847496033,\n", |
|
|
946 |
" 0.8551859259605408,\n", |
|
|
947 |
" 0.9236790537834167,\n", |
|
|
948 |
" 0.8904109597206116,\n", |
|
|
949 |
" 0.8982387185096741,\n", |
|
|
950 |
" 0.9236790537834167,\n", |
|
|
951 |
" 0.9236790537834167,\n", |
|
|
952 |
" 0.9099804162979126,\n", |
|
|
953 |
" 0.8023483157157898,\n", |
|
|
954 |
" 0.931506872177124,\n", |
|
|
955 |
" 0.9001957178115845],\n", |
|
|
956 |
" 'val_loss': [2.0990655422210693,\n", |
|
|
957 |
" 1.9411567449569702,\n", |
|
|
958 |
" 1.0954822301864624,\n", |
|
|
959 |
" 1.1076319217681885,\n", |
|
|
960 |
" 0.5441134572029114,\n", |
|
|
961 |
" 0.5360944271087646,\n", |
|
|
962 |
" 0.43207740783691406,\n", |
|
|
963 |
" 0.689609944820404,\n", |
|
|
964 |
" 0.4200369715690613,\n", |
|
|
965 |
" 0.402942955493927,\n", |
|
|
966 |
" 0.4212215542793274,\n", |
|
|
967 |
" 0.31122156977653503,\n", |
|
|
968 |
" 0.3583977222442627,\n", |
|
|
969 |
" 0.3882371783256531,\n", |
|
|
970 |
" 0.26823410391807556,\n", |
|
|
971 |
" 0.27452144026756287,\n", |
|
|
972 |
" 0.6421178579330444,\n", |
|
|
973 |
" 0.37860390543937683,\n", |
|
|
974 |
" 0.24037198722362518,\n", |
|
|
975 |
" 0.21231956779956818,\n", |
|
|
976 |
" 0.4761490821838379,\n", |
|
|
977 |
" 0.24066762626171112,\n", |
|
|
978 |
" 0.18336910009384155,\n", |
|
|
979 |
" 0.21427571773529053,\n", |
|
|
980 |
" 0.1881323754787445,\n", |
|
|
981 |
" 0.3551054298877716,\n", |
|
|
982 |
" 0.2521588206291199,\n", |
|
|
983 |
" 0.15665383636951447,\n", |
|
|
984 |
" 0.17057940363883972,\n", |
|
|
985 |
" 0.16153518855571747,\n", |
|
|
986 |
" 0.33384138345718384,\n", |
|
|
987 |
" 0.24861186742782593,\n", |
|
|
988 |
" 0.20901203155517578,\n", |
|
|
989 |
" 0.1482536792755127,\n", |
|
|
990 |
" 0.31539833545684814,\n", |
|
|
991 |
" 0.27941542863845825,\n", |
|
|
992 |
" 0.20582647621631622,\n", |
|
|
993 |
" 0.12700620293617249,\n", |
|
|
994 |
" 0.11588176339864731,\n", |
|
|
995 |
" 0.32454395294189453,\n", |
|
|
996 |
" 0.1344558298587799,\n", |
|
|
997 |
" 0.1628752201795578,\n", |
|
|
998 |
" 0.11320976167917252,\n", |
|
|
999 |
" 0.16728711128234863,\n", |
|
|
1000 |
" 0.11341923475265503,\n", |
|
|
1001 |
" 0.10802261531352997,\n", |
|
|
1002 |
" 0.10879490524530411,\n", |
|
|
1003 |
" 0.15465836226940155,\n", |
|
|
1004 |
" 0.11110226064920425,\n", |
|
|
1005 |
" 0.1549922525882721],\n", |
|
|
1006 |
" 'val_accuracy': [0.3243243098258972,\n", |
|
|
1007 |
" 0.6216216087341309,\n", |
|
|
1008 |
" 0.4864864945411682,\n", |
|
|
1009 |
" 0.45945945382118225,\n", |
|
|
1010 |
" 0.837837815284729,\n", |
|
|
1011 |
" 0.7837837934494019,\n", |
|
|
1012 |
" 0.9189189076423645,\n", |
|
|
1013 |
" 0.5405405163764954,\n", |
|
|
1014 |
" 0.9189189076423645,\n", |
|
|
1015 |
" 0.837837815284729,\n", |
|
|
1016 |
" 0.8108108043670654,\n", |
|
|
1017 |
" 0.9729729890823364,\n", |
|
|
1018 |
" 0.9189189076423645,\n", |
|
|
1019 |
" 0.9189189076423645,\n", |
|
|
1020 |
" 0.9729729890823364,\n", |
|
|
1021 |
" 0.8918918967247009,\n", |
|
|
1022 |
" 0.8108108043670654,\n", |
|
|
1023 |
" 0.8648648858070374,\n", |
|
|
1024 |
" 0.9729729890823364,\n", |
|
|
1025 |
" 0.9189189076423645,\n", |
|
|
1026 |
" 0.8108108043670654,\n", |
|
|
1027 |
" 0.9189189076423645,\n", |
|
|
1028 |
" 0.9729729890823364,\n", |
|
|
1029 |
" 0.9459459185600281,\n", |
|
|
1030 |
" 0.9729729890823364,\n", |
|
|
1031 |
" 0.8108108043670654,\n", |
|
|
1032 |
" 0.8918918967247009,\n", |
|
|
1033 |
" 0.9729729890823364,\n", |
|
|
1034 |
" 0.9729729890823364,\n", |
|
|
1035 |
" 1.0,\n", |
|
|
1036 |
" 0.8918918967247009,\n", |
|
|
1037 |
" 0.9189189076423645,\n", |
|
|
1038 |
" 0.8918918967247009,\n", |
|
|
1039 |
" 0.9729729890823364,\n", |
|
|
1040 |
" 0.8108108043670654,\n", |
|
|
1041 |
" 0.8108108043670654,\n", |
|
|
1042 |
" 0.9189189076423645,\n", |
|
|
1043 |
" 1.0,\n", |
|
|
1044 |
" 0.9729729890823364,\n", |
|
|
1045 |
" 0.837837815284729,\n", |
|
|
1046 |
" 0.9729729890823364,\n", |
|
|
1047 |
" 0.9189189076423645,\n", |
|
|
1048 |
" 0.9729729890823364,\n", |
|
|
1049 |
" 0.9189189076423645,\n", |
|
|
1050 |
" 0.9729729890823364,\n", |
|
|
1051 |
" 0.9729729890823364,\n", |
|
|
1052 |
" 0.9729729890823364,\n", |
|
|
1053 |
" 0.9459459185600281,\n", |
|
|
1054 |
" 1.0,\n", |
|
|
1055 |
" 0.9729729890823364]}" |
|
|
1056 |
] |
|
|
1057 |
}, |
|
|
1058 |
"execution_count": 42, |
|
|
1059 |
"metadata": {}, |
|
|
1060 |
"output_type": "execute_result" |
|
|
1061 |
} |
|
|
1062 |
], |
|
|
1063 |
"source": [ |
|
|
1064 |
"r.history" |
|
|
1065 |
] |
|
|
1066 |
}, |
|
|
1067 |
{ |
|
|
1068 |
"cell_type": "code", |
|
|
1069 |
"execution_count": 43, |
|
|
1070 |
"metadata": {}, |
|
|
1071 |
"outputs": [ |
|
|
1072 |
{ |
|
|
1073 |
"data": { |
|
|
1074 |
"image/png": "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\n", |
|
|
1075 |
"text/plain": [ |
|
|
1076 |
"<Figure size 432x288 with 1 Axes>" |
|
|
1077 |
] |
|
|
1078 |
}, |
|
|
1079 |
"metadata": { |
|
|
1080 |
"needs_background": "light" |
|
|
1081 |
}, |
|
|
1082 |
"output_type": "display_data" |
|
|
1083 |
}, |
|
|
1084 |
{ |
|
|
1085 |
"data": { |
|
|
1086 |
"image/png": "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\n", |
|
|
1087 |
"text/plain": [ |
|
|
1088 |
"<Figure size 432x288 with 1 Axes>" |
|
|
1089 |
] |
|
|
1090 |
}, |
|
|
1091 |
"metadata": { |
|
|
1092 |
"needs_background": "light" |
|
|
1093 |
}, |
|
|
1094 |
"output_type": "display_data" |
|
|
1095 |
}, |
|
|
1096 |
{ |
|
|
1097 |
"data": { |
|
|
1098 |
"text/plain": [ |
|
|
1099 |
"<Figure size 432x288 with 0 Axes>" |
|
|
1100 |
] |
|
|
1101 |
}, |
|
|
1102 |
"metadata": {}, |
|
|
1103 |
"output_type": "display_data" |
|
|
1104 |
} |
|
|
1105 |
], |
|
|
1106 |
"source": [ |
|
|
1107 |
"# plot the loss\n", |
|
|
1108 |
"plt.plot(r.history['loss'], label='train loss')\n", |
|
|
1109 |
"plt.plot(r.history['val_loss'], label='val loss')\n", |
|
|
1110 |
"plt.legend()\n", |
|
|
1111 |
"plt.show()\n", |
|
|
1112 |
"plt.savefig('LossVal_loss')\n", |
|
|
1113 |
"\n", |
|
|
1114 |
"# plot the accuracy\n", |
|
|
1115 |
"plt.plot(r.history['accuracy'], label='train acc')\n", |
|
|
1116 |
"plt.plot(r.history['val_accuracy'], label='val acc')\n", |
|
|
1117 |
"plt.legend()\n", |
|
|
1118 |
"plt.show()\n", |
|
|
1119 |
"plt.savefig('AccVal_acc')" |
|
|
1120 |
] |
|
|
1121 |
}, |
|
|
1122 |
{ |
|
|
1123 |
"cell_type": "code", |
|
|
1124 |
"execution_count": 44, |
|
|
1125 |
"metadata": {}, |
|
|
1126 |
"outputs": [], |
|
|
1127 |
"source": [ |
|
|
1128 |
"# save it as a h5 file\n", |
|
|
1129 |
"\n", |
|
|
1130 |
"\n", |
|
|
1131 |
"from tensorflow.keras.models import load_model\n", |
|
|
1132 |
"\n", |
|
|
1133 |
"model.save('model_resnet50.h5')" |
|
|
1134 |
] |
|
|
1135 |
}, |
|
|
1136 |
{ |
|
|
1137 |
"cell_type": "code", |
|
|
1138 |
"execution_count": null, |
|
|
1139 |
"metadata": {}, |
|
|
1140 |
"outputs": [], |
|
|
1141 |
"source": [] |
|
|
1142 |
}, |
|
|
1143 |
{ |
|
|
1144 |
"cell_type": "code", |
|
|
1145 |
"execution_count": 45, |
|
|
1146 |
"metadata": {}, |
|
|
1147 |
"outputs": [], |
|
|
1148 |
"source": [ |
|
|
1149 |
"\n", |
|
|
1150 |
"y_pred = model.predict(test_set)\n" |
|
|
1151 |
] |
|
|
1152 |
}, |
|
|
1153 |
{ |
|
|
1154 |
"cell_type": "code", |
|
|
1155 |
"execution_count": 46, |
|
|
1156 |
"metadata": {}, |
|
|
1157 |
"outputs": [ |
|
|
1158 |
{ |
|
|
1159 |
"data": { |
|
|
1160 |
"text/plain": [ |
|
|
1161 |
"array([[9.23960090e-01, 6.27789088e-03, 6.97619542e-02],\n", |
|
|
1162 |
" [8.90219398e-03, 9.59880054e-01, 3.12177557e-02],\n", |
|
|
1163 |
" [1.61567070e-02, 4.08069427e-05, 9.83802438e-01],\n", |
|
|
1164 |
" [9.58632350e-01, 1.72755762e-03, 3.96401100e-02],\n", |
|
|
1165 |
" [6.30823313e-04, 9.99005139e-01, 3.64071195e-04],\n", |
|
|
1166 |
" [3.00111622e-01, 5.42214394e-01, 1.57673940e-01],\n", |
|
|
1167 |
" [5.87674556e-03, 9.92273271e-01, 1.84986857e-03],\n", |
|
|
1168 |
" [1.87689245e-01, 7.37681389e-01, 7.46293962e-02],\n", |
|
|
1169 |
" [3.62052053e-01, 3.98187386e-03, 6.33966029e-01],\n", |
|
|
1170 |
" [5.56720328e-03, 9.94185507e-01, 2.47379445e-04],\n", |
|
|
1171 |
" [1.45750595e-02, 9.82338071e-01, 3.08688660e-03],\n", |
|
|
1172 |
" [8.55647251e-02, 1.00188379e-04, 9.14335132e-01],\n", |
|
|
1173 |
" [9.73771155e-01, 2.72319582e-03, 2.35056598e-02],\n", |
|
|
1174 |
" [2.23941237e-01, 7.66084254e-01, 9.97449458e-03],\n", |
|
|
1175 |
" [4.23269905e-02, 1.43688906e-03, 9.56236124e-01],\n", |
|
|
1176 |
" [2.27208659e-02, 9.60315347e-01, 1.69637445e-02],\n", |
|
|
1177 |
" [7.32151372e-03, 9.92544055e-01, 1.34420086e-04],\n", |
|
|
1178 |
" [1.86562967e-02, 3.09444949e-05, 9.81312752e-01],\n", |
|
|
1179 |
" [7.61464536e-02, 9.06794369e-01, 1.70592573e-02],\n", |
|
|
1180 |
" [9.98593390e-01, 1.00536190e-05, 1.39657059e-03],\n", |
|
|
1181 |
" [1.96318626e-02, 3.34709010e-04, 9.80033457e-01],\n", |
|
|
1182 |
" [9.96856213e-01, 1.50083515e-05, 3.12873721e-03],\n", |
|
|
1183 |
" [1.78742498e-01, 7.77231038e-01, 4.40265015e-02],\n", |
|
|
1184 |
" [6.96946159e-02, 3.67836491e-03, 9.26627040e-01],\n", |
|
|
1185 |
" [2.09010303e-01, 1.29479694e-03, 7.89694965e-01],\n", |
|
|
1186 |
" [9.77899432e-01, 1.07577248e-02, 1.13428310e-02],\n", |
|
|
1187 |
" [9.97433245e-01, 1.55827456e-05, 2.55123922e-03],\n", |
|
|
1188 |
" [1.64430931e-01, 8.12096775e-01, 2.34723687e-02],\n", |
|
|
1189 |
" [1.45679014e-02, 1.03409820e-05, 9.85421777e-01],\n", |
|
|
1190 |
" [6.93119541e-02, 9.23750520e-01, 6.93753688e-03],\n", |
|
|
1191 |
" [6.68204367e-01, 1.24834567e-01, 2.06961080e-01],\n", |
|
|
1192 |
" [8.40640366e-02, 1.05693369e-04, 9.15830314e-01],\n", |
|
|
1193 |
" [3.69445324e-01, 4.51182276e-01, 1.79372355e-01],\n", |
|
|
1194 |
" [9.51326191e-01, 1.81014072e-02, 3.05723399e-02],\n", |
|
|
1195 |
" [9.99626637e-01, 2.68646500e-06, 3.70592374e-04],\n", |
|
|
1196 |
" [2.74331737e-02, 1.94599816e-05, 9.72547352e-01],\n", |
|
|
1197 |
" [8.99137735e-01, 1.90306595e-03, 9.89592150e-02]], dtype=float32)" |
|
|
1198 |
] |
|
|
1199 |
}, |
|
|
1200 |
"execution_count": 46, |
|
|
1201 |
"metadata": {}, |
|
|
1202 |
"output_type": "execute_result" |
|
|
1203 |
} |
|
|
1204 |
], |
|
|
1205 |
"source": [ |
|
|
1206 |
"y_pred" |
|
|
1207 |
] |
|
|
1208 |
}, |
|
|
1209 |
{ |
|
|
1210 |
"cell_type": "code", |
|
|
1211 |
"execution_count": 47, |
|
|
1212 |
"metadata": {}, |
|
|
1213 |
"outputs": [], |
|
|
1214 |
"source": [ |
|
|
1215 |
"import numpy as np\n", |
|
|
1216 |
"y_pred = np.argmax(y_pred, axis=1)" |
|
|
1217 |
] |
|
|
1218 |
}, |
|
|
1219 |
{ |
|
|
1220 |
"cell_type": "code", |
|
|
1221 |
"execution_count": 48, |
|
|
1222 |
"metadata": {}, |
|
|
1223 |
"outputs": [ |
|
|
1224 |
{ |
|
|
1225 |
"data": { |
|
|
1226 |
"text/plain": [ |
|
|
1227 |
"array([0, 1, 2, 0, 1, 1, 1, 1, 2, 1, 1, 2, 0, 1, 2, 1, 1, 2, 1, 0, 2, 0,\n", |
|
|
1228 |
" 1, 2, 2, 0, 0, 1, 2, 1, 0, 2, 1, 0, 0, 2, 0], dtype=int64)" |
|
|
1229 |
] |
|
|
1230 |
}, |
|
|
1231 |
"execution_count": 48, |
|
|
1232 |
"metadata": {}, |
|
|
1233 |
"output_type": "execute_result" |
|
|
1234 |
} |
|
|
1235 |
], |
|
|
1236 |
"source": [ |
|
|
1237 |
"y_pred" |
|
|
1238 |
] |
|
|
1239 |
}, |
|
|
1240 |
{ |
|
|
1241 |
"cell_type": "code", |
|
|
1242 |
"execution_count": null, |
|
|
1243 |
"metadata": {}, |
|
|
1244 |
"outputs": [], |
|
|
1245 |
"source": [] |
|
|
1246 |
}, |
|
|
1247 |
{ |
|
|
1248 |
"cell_type": "code", |
|
|
1249 |
"execution_count": 49, |
|
|
1250 |
"metadata": {}, |
|
|
1251 |
"outputs": [], |
|
|
1252 |
"source": [ |
|
|
1253 |
"from tensorflow.keras.models import load_model\n", |
|
|
1254 |
"from tensorflow.keras.preprocessing import image" |
|
|
1255 |
] |
|
|
1256 |
}, |
|
|
1257 |
{ |
|
|
1258 |
"cell_type": "code", |
|
|
1259 |
"execution_count": 50, |
|
|
1260 |
"metadata": {}, |
|
|
1261 |
"outputs": [], |
|
|
1262 |
"source": [ |
|
|
1263 |
"model=load_model('model_resnet50.h5')" |
|
|
1264 |
] |
|
|
1265 |
}, |
|
|
1266 |
{ |
|
|
1267 |
"cell_type": "code", |
|
|
1268 |
"execution_count": null, |
|
|
1269 |
"metadata": {}, |
|
|
1270 |
"outputs": [], |
|
|
1271 |
"source": [] |
|
|
1272 |
}, |
|
|
1273 |
{ |
|
|
1274 |
"cell_type": "code", |
|
|
1275 |
"execution_count": 73, |
|
|
1276 |
"metadata": {}, |
|
|
1277 |
"outputs": [], |
|
|
1278 |
"source": [ |
|
|
1279 |
"img=image.load_img('Datasets/Datasets/Test/Segmented Neutrophils/SNE_10863929.jpg',target_size=(224,224))\n", |
|
|
1280 |
"\n" |
|
|
1281 |
] |
|
|
1282 |
}, |
|
|
1283 |
{ |
|
|
1284 |
"cell_type": "code", |
|
|
1285 |
"execution_count": 74, |
|
|
1286 |
"metadata": {}, |
|
|
1287 |
"outputs": [ |
|
|
1288 |
{ |
|
|
1289 |
"data": { |
|
|
1290 |
"text/plain": [ |
|
|
1291 |
"<module 'tensorflow.keras.preprocessing.image' from 'C:\\\\Users\\\\junnajamshed\\\\anaconda3\\\\lib\\\\site-packages\\\\keras\\\\api\\\\_v2\\\\keras\\\\preprocessing\\\\image\\\\__init__.py'>" |
|
|
1292 |
] |
|
|
1293 |
}, |
|
|
1294 |
"execution_count": 74, |
|
|
1295 |
"metadata": {}, |
|
|
1296 |
"output_type": "execute_result" |
|
|
1297 |
} |
|
|
1298 |
], |
|
|
1299 |
"source": [ |
|
|
1300 |
"image" |
|
|
1301 |
] |
|
|
1302 |
}, |
|
|
1303 |
{ |
|
|
1304 |
"cell_type": "code", |
|
|
1305 |
"execution_count": 75, |
|
|
1306 |
"metadata": {}, |
|
|
1307 |
"outputs": [ |
|
|
1308 |
{ |
|
|
1309 |
"data": { |
|
|
1310 |
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAOAAAADgCAIAAACVT/22AADp7UlEQVR4nMT9WbMsSZIeiH2qaubuEXGWu+ZW1VVd3Y1GDdAYkoMRCoVPI+QD+cRfwB/KF4qMQAQyQ2AIkjMA0Y1Gd3VVZVVm3u1ssbibmarywdw9POKcezOzCiBMbp6M8HA3t0VN94Xe/sv/29vf3bz9zS7vLW390FrZ9UwE2EGzkJuZiAAopZAKFo0b1F9FJMaYDmgiB1Yil0DUSgihW8UY43rdXf683VysLl9cipBEhvmyK/eTr/WNIIK7uxMHNzMzqs3h7sRcfwXT8lmC1Q6JCPUfAHczA0B0dvNJs9NhEJ3MV1Xr4+Mo3JkBsmkABHe4q+rD7b7v07e/ev/wvtx/w4d+6+rcs5kppxTU67yI3D2agMxda69EEYADYIIwsRIRO0idmRsXJfShxEhdiJfds9Dayz8NX/zZZ7HbrBoC4EwgcsLZIgsE7tlNGSCKOq5GHYyG49QAkPk8TQCwk6XyqU0L4stnVfUETpg/sddwmTsE4HpcXncPr372fHW1Xl3d7W4Ob765DfucSjGDgxOhoVhf5u4MkhBO31QCC4Fgrrno6qBNA2FmliZehg1HipfcXsTLF6sXLy/atpUQAFNVoZNBf0/zk/mbmbvX582MWT71qFmF1LpM56vzYxoz18crrDMHdycw3EEEs3oeROTZy8s8JDHqX9pv6C6/C64oUVV1Z31XOvPxvMHH3ojZ3QCvwDn+5+4KIhAIDncqpGC+iN2q6zqJ4ZleXnU/+fmL9rIl8RmMvA7ptNkMVQajCl0jUAJgRx0P6smZ+6mn/fwsY36QiMy0gtSywx/aRuAGRrxTB+5EgHvwmDavQ7x4fthtVl/x/jdy87ub/V1S5w0Hw53DmJmIEKLqKfohMPO88Z/Za1FmEmJpOMYrXV20m8/bsObmMnRBQmDDuLVwezzVjzWfAPTsyg+BtuOzEypd/nqKF/DkNnysWzdU3ElOmJE66mIXjnj+xdqUTLT763W/Td++eVd6v2iuNSYYNFeUSYMqEQkxkcCNTTECKAHOzEQ8zYApgpkEpWvs6qr9/J+t23V78XKtpMSgTOOzdeJ+ghStYkMFOcI00wVIOXwxPYzrNuK2c9oyot7aeaV4C3h9YnNnHHnW6llw9/GuxUa7eyAiEJp1EyJxvH6gaK6uD2kw3SsHroeC69OnuJphBNTFI2YuTCEwOwuaRpprX102F89WoeOwCkjmnt1jndGPBdATcrNYi++F0Xl78AOg+WwYn3hicSfDHbDxSv1rzMwUGOBnn63LnR/u+f5NQLGSjALDzBWgyhcQEdO4xhVxjDjsSHDrLjC7K0Ni9HhJ8QVfvlxxCEpuTuzTIcR0ep0BTPBNTuoOAsOcQTM8zhh0Xkx3B08fcPJh+ci8CMuvxwVf3Lzs59GvlVWYYJQA2PwT+bt/jdSDKiRr6lC+S/1vD4eb9O53N71uDodDKcXdzazYCYC2kcyMmUWkaRq56LoNtS9odREvXnSXl5fMMKhDidAARFKcAZbYwPJH9nvxdeJBzcDMM9P5n5AH1VNGI/gZgJ7efMqDamFmgKZjK3ZE0iaAOfeAGcL96uA9Dn+H/Xfp679593A7qGrf9wCYeTskZm5aJiKHBTMseNA6dwGJk4iERjbr5qs/ubj8RXPx2bpp3Z3gkUMDoLG6vW7uToCG+opxwMHcnc1JnYmK6vhTpYE2rlsFJqVP8aAisuRB3RWnxwmnzRft7Ffm4w14zINmGDctewVRah5SXIn/1P0Vrr9Yr9+Ehw98uAnIZIPIJXLOlStl5siNBLQbkkhxxYfL1G3a5y/atpP1Jgg7EQkEEADg4ABXXOhmZ9TnnCicfBUJAEy1rqabgcjcxq+nh7uefYejMm8Tqq6syOl7wGrLNV2seF1KmTkEIkKUkXrWv1FL/RVERAEjHXRVxColBAAMPD+wqskzp0DtwOHvL4e+z9s9Ka3btm2a5PlgicU79hKNHGRk5mIcE4UQuiYGoRgDP7fVM1r/o+bixaqNDXSCPgOAoSgzhxCkLmmFIljF1aIjEXWCwhHY6mxcAQhzhZE6O8bEjwIA7BQETSva52l9Rj6ywl4hYweA+tdcxBkOEJSdOJaSHMaMEBia67OV6TBqRhio//T9v2aHm5ED7ii2xDoHy/vtvn84eLbhkCi3daeZOcaI6z7GKG3DzM70LK5PNv9UBDY7kgkiqlD+sQM38ehHQKqP1FcfOfdTur9si1/19PrJu8zKk+RpOtDzBlAdxdkrluBLKDRhA5xJgJ4BgBq4DIf87uv7YZff/PrD4S4/fLvPuRiBYsPMVOCuzBx4JMS8unC2uLaXX60vnq0//5NV6KS9ijATsBIvB68VbpirHDkvDk1tybufrgRoIvEjwJ8e5UfMlcy75u5AWf7qgakqENQAJJRgEJDCC0MsAAaqEiEE7cmO4GQ3yd/9a7hDrcp4JKGC0gimBYBBGDBoHjhjIhnMjJI4RMgKTqqA3RECUwMKgEANE3Pt7uZHSJqBbF7ZRxj0TAlFSwCd1+WHAKhZOe1ZTr+ea7uWAOp+vsFnN5/uaBmlSQCnFGBcTFK4ApZYLVv/AXnnN9/s7t8/5Kz7294KeeYQgrvDNRCeXV/Il9y0/OIn681rQVC3EChEtEQM5gplCw5vHOGo6Fis8xKexvtPz9usk6qzOFMVna0zTVqRqcPjze6ujBmDElFC4WJk7kwWmLIzgwgscFcr7RI7nO4PwszVocqwFTrnZc0F7qiUTZ3XjIVai2kDD9AAJ3GAAxBGgr4AzenvuQz+MfT5uNFp+17B6OzZT3wF0zjfpwXMUyHg7JaJHE7Uid0JIHfnM07XjNiZyQFiBGYPLM/YN4gdX//0Ivf54ZubPFgavIAACGkM9PrV5vrnnZNxo6FzFkKO8EAWHHAY47if8yI/5vMWN3xqoZb49WOrsaDp8yE83r8Uhtzdzd1d1f0IoGBzMJhBJMyNItdhEzGI6FRpGLzoNBkCE5jhDjOYac607ksxJyYSc2pLQyKAjCghmkPVS5VBOV8cB01WhZhTinlcu1lk+SEAWo/EiJ+IngSmj7Xv6d9suSsfg+ZxnPbpDZban1k9ryf9MAcEIuVRw2xGPGDl63VBiF0OL//klTg5cU+ByJkLmTKXGEQVbF20NQq5J/IAb6wpA5VVGXV80yyODHcl9I/HeUT8pzq1im7n9uQynl2nSfc5r2H9MAxJU845D/tDSgmpwVDqrWhDbOsEqVs1bRtDp0xVbGCA3Fcnr/C3/6OZQUcjTUEiG/W9DKqUor6VmQETiRAGuLgFODiAI4jdiOwAI4DcCBPC9Emadj1RXp6xGvOER5iYQGEkJXymSDvlB3ziw+oinjEx53jv/JU4AdBHOv9Jk+ALXdJ0xj6qjiUEkE+iw5ETwPIMzCe2QpUwCOomTj7pHJiZOGCaPHC+aizuDjfykdafiZ4n7M2MFCYaeGbg0Xp9BDg7WXNiwYg1j0qSuRVz4gIqh32+uduXt3q4O6Re84OJCRqrChARUdVAwcjAHlahXTUxanvRxFaay9BetkIsMRCRuYtIgEZGKSAisAj7UK0QMCIP0IdKrIkIRty0YK92OYfBBKQwOPFsuDij5jTJWxMh/RT5+CHY9L9IW1KuJdr4BPb9EW1WhxGYGOozhNH3kQuzUUKfMNzJr2c2xlLKjAg/xgMsr8/zehKhzuyWmWX00SNUhru0/X0u95r3VnKxZAQvbtN5M1XNXqr+PBTrD7kRdHsNq7DRGCC+NpKOmQFWUABuwSEQTMtwKPubIR90v01l8JyslbA8K/H6PnZxfdmRCEeKsYET2IhGuarO6GTDJgw0X58wwQnz/qN29/zmH0Hwn1Y4f+JmOmGjT6BzpolPdPJjhjQxwfWPm9aVYQCqTmSfGOTHQG3+dfl1VojWD4/l9BM2dGFJefINM6ybmTB9+P394a7c/2443KShqGomc2I1UfOj0kZV4aGenNwb1JLIPvQs/vCGt9eHcF1efPZqdbEK66juAVZQPO8sHcp+O9y9S6lPh+1QBstZV93lDKDMzH1ar90ThSZII80mk4hEhhAxFlA4UjXiybBqtsSg7u44ruwMpp/m4s+W5riyP+SZjzz7ve2MiPuJ7mbs8DEG/TFMMjDKtAQCw8tkmgfgZiSfwtDfe8BO33PUSVWO5fH9RxiFLjfobEYz9SillFK0t7u3h+3N0N8Uzl5UzRQAO7nDi+nIQ8IMIHVTOJPCFDlnFhC5WWQTTqULGVlWRNxxeLiT22+2b/72drgr6d69qeohJXcB38Z3mLgWZrYPYdeWh8s+RKJA8YJX6/bq2UXsYtPFsD7qHX74qp0t9A+E0T+8/aGE+AyDYlLlTD9NHlh/QKsKHWEnqBtRJCIee1P3qsd+etgsAocbVWHhjK08G1JFmbNcEcK5k9H5yfcjs3HWKgIupWy322EYPvyHfndbSnKUXGxQtgIHGCTk0ixENyIySl5ABjIJCKUzBrvzcIAVdNvwod9uL/bPS/fi80341b/8dvvQ7+6SFcDDMGyXKBN5noMDGsXSIaVDW2ce1tiuy+4+tZt48XzzEpfoipIxc6DOoVUd6HCDqpwsR9CFewGO6HOcA2ZDqJvb7Ob35C7RqNcavZYIo/vVxPbGeU1rd1hIVJgUeLNUsWQ53G1SJdpE8CeNB0A+CcKnjB0RkYwUbZKvJjXKqH/zxWwI3Izyh0NgCA64w+rUCHExnjOWAyX5tDu12dQzzdfPqNMZY3pcRrIlqi7MR7UuwJrrBEXEzNhtn/vDXj/8Q+4/lN1tLqVUkHUnQEjVUZiLCJesIuJMxcwJVASAc7UzF8ucRxEKJe2AeMjgW0pqOnC4eXNX1FW9wNSSTj6XAEopTKfCWimYEKq7S6YyqLuVXsVjv04hBmrqlpQzUYn9bDlO7De6NJQzsx9VHhMA/IHtsT7rh1P5szs/zS5Xt4TqM+AL9u54UhYE9Ie3Iyv/1Pn83t5oUswR0Y/ihgREfnzjcuREBOd+lw8Pw+F+t9+lYcgVMdvSoE/HZ+eROI6na5rCeH8FquIZysq03/V3NwjlIL3mXnuDGoFVZj8AVWVynk4SzTa0YkRuZp20rlaGfoiSbvFwv3398+dXX66ZLXsfqJtQWgXQ8+Vd7vdSCQczkZlh+fFE3yd7ozsmD48z29V4w/e1JX/JzHZuHjqZUZBQx2+lzLqtGdst5Y8fxTjW5RvVYROo/WGw/qOaEC9R7/xZVd29f0i3v9/tb/vhPlsyM1NVM5tNrJjQirtLADOcHTzRyhO96ejwUB8fUJg7LuGwNfIh9K5Z1RQQjiyK0ZF2eo3XY1EB1JzMvHppmJlABQy1UswMJnF/lbqrENsYA4NONuPMuFLPzfwiWmht6r5M2/MHIM9KwrwijJmg+ezfdKIb+mRPkwrilO4vEMnJa3lcYzOR453uTpPF/IdM5wmwm68s1SA/ADrHOf5BnP3oLDIvF46ct6rm5P2upF2xoZL1kzZR3KMcSXT0LCM+2et59+vX4iaqZkwDJfawowHkpuYJRUsl0BVhiAgvzr2ZQdhnqsG8z/uWWyH2gt4Tvw3f6fvDcH/xbPX55695cyKnn5GXpfMbEdnEGo68ry9cSZhP/T1+wNoCM4CKnDoj4rjo39+W4yeCfYpM26RJEZEav7HoZprXJF788MmM2OEULk8P86eexYI6PWGG+ERbzN3MKoASUc657/v37/e3b+7LvcLE3A6Hw3zzPM3qIEkEIiVyYgLD4EJcydsoZDovj27TNW7iipwQiILtjImQHe5BxEiIMOreyIh4ZkBRZQgmuE2OWKFkMzFmkLnR3fq+22bOF30YdptfaNPGVRA4I7OfOhSSLJQywPTNAXezSqaPyPXodF3PogIOdhi7CagAqAcek+AyH98l6ANLqQgA5rCHJRt03FowkdAkbxCUjjISRgQzPXziIrjAuGGKk6keFfXek6Vgnd8IoEzxPfWvFB/DPdzdPfM0HSYAQQHACU4wAvKJUxzz5LpWz5eV5a/kp9SAw2goIq2Xxl8JYBfXHqagw0M+vE+7XyXv2Uiz9WbmjbDDzNyqkrMwg5lEqsE7MDMc7Nw0zaHsR6WaRjMLjc5+VCBoCe4OKDEdigXzAieFju7X45L5hJJPRBkAIlJRSBWYMCJXJyJj1kK5VwNuPuziZ7GhiEjw6sR5cuJnzmZcrx+FJJ3HRfSnBNI/RqSa+M7/3Bzess16n7oUgY6xU25WWbQjpzRFDj05yRlZ4uPS+tzOcPxjEnd8KbNlc3ZVpD73+1Tdgis/OiFzIsdoUCeSMGomATDPjJbNBtWJ0DET2ajrJSLRUxNroCptBSIQgjA5kYsQi4cQ2AMWAOogMzNTEJo2aKl9AU7mNBC8WFtYihdsYyd4Bf6s9QBpGz4Fpplk4Ck98Pc1hgNuNd7kHEz/GPj8pBTyRwH+9795ZgPMfXIoOPXRBEZn7EpHxjsWrUqxVZ9ARI9V8cvmcxAVZsHFZ6zkp5THQG5k2Q83w/7DMPS5ZHMjOMNBWsDMlRVykIAZIs5CZkbCLKPtytmbEAHAyQxmgKm7ujOcVQ1UdbqVYCIQbBwlkzccTFkoBJFAIYyq3FnRoSXW+ddJBXIzmIqZuUMl52RsDjh63H29K9ts4aK5aCKj8xO/VJYjnlBV/nGbP3OZ1U/7FPR/DId33mbGC8DEHjz68T9xm3mMUQTGkTtnZi0nrDyPVJcqmJ753NStCcfg208thapKZfFHFH6cIBGxyAm/y6TZUp/2D6nstBRTdTOvsrubEbEQVc6LGO5qDnJmJiULITAzT0GWROROpm5WnRjNnQDmwqSjF86IQc8YImKfzEbEjCqNHkUZRYyhLmUppTrBmYKI3QAhdzVXNefCmkrqkVMRFTH7BP5x/wjF+nTzmVKcMQ8/TqT6L96WXK+7V93z4xvOGj11XI5s+g9+70IzcPrUJJnNbzezUsyKq3pFfvWfqhOdmLBGdyjzMZAMi+AwIh4lbXYzImOqURJEJHA3qNfgmQqg0rQObZiInJGZWhGPjYTATRNiqyFEU5iBOaw3nlIxhTtJEShKdnc3BTNTz05urAXuBZb6dBf5t4f11q9/GsJVYUeUcBSHpnUgkTMFilXHtiocmAnPUj/DMXp8mWM0Gj2RF2Dmbt11Qg8Mokq5juKX61ENwoTZVjQuZln29hRCncQIIiKuLhEiQlUinMn0Ai8Twb1gspm5O1HA7DNeZfwpytcBDuLu6nPU6Lgy5gbAIvmosWZmDjGeLMVT4b/zXw6Sq4kc7B7UepHqVEFmXtnHeUn3ZsPDkG+Hsh1SSqVk1QxAhJgFYwiBckBhF3jTBGaWgKZpRL0JkYSJmYKEY88CIBUTDW5WBiNyCVxKdSBEKXaSiAGLE1O3pGYNYSJ3Yg7CBqAUM3VVr8BmpkRwYzsP97FSSr/vqUG7azZXq7r6Y3Tt47cuYfTUSdt9lF2W4sJ881MnH+f3fLwdb1jyZD+gPUJUP5SPPhc9JzB6PNTvxYXsVX/PBOJHZOh8Kc4PWPVYHdmZ+eYZo8+8R6WZOWtKlbirnyIU5lmfPWLeeQBEFCNLrBSeSIhP0380Yl7cjNxAxb0aEZncyOwpAF2unZsQhIWISAI0W4gsIu7EQl6UxQHRYqUUcZltCWYmLqxUtlZYsQVqrKraaKkyzJOnSTU4T5iEZzFlpj7zelEIAMidnlJZnz5F8zLV9z25zXMPnxbWPg0u7s5TDgt6dACf6OoUIOaRL99F3+NujSo4E4grbTm7/fTbLNePBMELRlEVcMgYA+pnIxnDmwRDr/3BtBxBdilKTtRgEoYWGxdaxJaYyZhCYJGzZQylFCtE5FbI4Eji7mZkgnMAZebZw6GKg2YIgYhdRKxAhDgGN3JXRRFnAmlgKTak43K7e86ZuRm22bLGLl78rIsrqchQS2EKS/5mKTjP1BlHaLMpaNgAWM4TVa4Rv2cmSNDIwSwQLNGTwtMMUv4UrJ9DA52TiNNfj0fo7MFHV9zd+QgHRDQGp9uEnDByDez+KUHHphDwOrszTfwZT1KDTY/nnEDscCPiusKzfxaW6tLxL3K2oTdVMsUCBKv9s4aA14BwllBJr4TAMUaJJbQsIggSYzyLKHQjjsGLhyBQL24ci7trIYifA+iSw57x3KSxdxFxBzNXHpHIiVjGkAQ5AxSCEAm0lGJlKNXbpTK/ZibhPNvE0/hjsbXzylbqc4zwPEMbp8iysgfLF53f/H3G8Y+1R5j7Rzw4T83da0QOTqdMy534AV2ND54Oyc/x6VMDJlsK+4+XaPpqqmr69GH2Sd5aoLZqD4eZxSp0C0sIIQTHiRRrZABDQM5gc6UQzAxwD1UfUbFm3X0i4knlE0KQwACYAvPIKbkboMyIDdruopQSims2smQxwKo/AKsqGWUvsRERHu5S/w6I+mwFWAFJdp0SvziARggeCMFRHAOOiayq2k+q9Uo4gshs4SLpbtN8qwJDU7WmTKGnLKg609Ed4KjE8Wr9q/mVTgnWtPrHI+QL7fcksZ3ByoJYn/pusR91ame7y8yjo2AV+OjU92rBcU1PeeUlziC43ucmJ4zjeX6hkwGbCkFqPyKAC7NVsyQAx5goxd3MbRiYew6HAYMZvFgGzwzBSD5AQE3hQ1IjtlkQIhCUo8VI3JYQ3TT6xOaaWfDg5GjIGpBwoNyWYMr9Q59uUw3B9kW85HgO6wyLKgnUDU7MJDF4KU5g5ihNCEGKaNYQvB4FNSqZ3QnORQszB+cqLe1vt77xy2Ylps6NjEs2GlNhPhrZ3OaQgCM7K3JEFfieduIY9cig8gRWOtXy0KSABJHpEz7n3/f+J287e2sdErmPKfE+0c7Q/3jePpLJ5weOzU/hfn7RRG188fn4VM6j06eZAexuizM5UtQ6HVmEzKNyGk4Ai8QY46CjN1JlA9RDjORsmw1TQC+Ut0UP7rew4ZGQdLYoRrAxLY8DJPV8MFOFaB7dSgoXsLOASN3djOf9rlYNACVlK05OZkbMNOM0x1HqG2nNuSg9Y50fsh+jBPCj1PU/2JD1B4PFx9/sP7CzmZTjRy7ID28zXFYGf744L+lZNoeFFDH+NbPKNLlX9T+5+xi872SKUuYkkCPwkCB0waGxC4YSwPuk5VBSSjA6UdTT6MREzCTCIoJGWGo2VFc3gZCIhECj2crADgYYzh4b5EwlVyrJ1ftMVUUohMDZkV0owBUSxDImRZ0D4AAqqFoPa8wmMWghx5yxpB9rM2hObjXnwbXLr1pKjQ7g80P/NHw/iXs+dtvcHt0+utbPt3+iqzMMusR/fzDrXHs4j1EZgfLY67J/XqTTISK38d8oh9EkkxgRoZTRYF6Kl1K8cCkGqHLNeByWXckqr59tVDV08X6b8psy3OXUp7wvxXI4m+98XOZhVW4HOB7ZOSaQiJyPQjcLmJmPvg0+LwcRwUe+yc/DsStTNW/VEUPMIzmjv59uZ458j1jFJ3breNsp4//4bUcu84/xJjnnU35EP3/4S//QdgoRM9TS/K8uVOWP5+2beEV39+pVwmw1g0H1HwUQY4wxmgwSYEQOLkpl6xicysjjBQoiDCESArkTF3dhlhp6gqHEtqkp+qplPkhgcndjFGhkB4ILBcpA7+RGXgjG5GQCCBMbeFA8jyCxXPZtCEl3gjjPGETqBucxJZVYtZzZpLQfD8Ds18hHpAIclSvVC6ASlJFpd7gPGB1KKj9zmuKUtRqTRrGM5+UngM6Mpu46Y50ZAYxTwEnCUyKfOXozcy+LTR4d82ZiXfW50+yo+hfYFE5kehI1dZaawbWyjF7FMvMjEgGODuz1ER/P4EyX53k5EVkZMGlziYgl2phcm0yRLW+6Vj3tU58Z7lb/1ehNV674UwvBUYgBcUMUKowoXHbFo7UwNaMcRcjZwSWuAq/aEm0g7d/dP/zqPg82DEPJWoqZcVggTjATC0Lkqk2VwCzMgtm37/HE5sViZkVxAEwwAiMV5WoFcy8pe3NJTXCCu8u0WIt2Eu016ad93P16RqF4hBDxCKnMrvvj8CpcEgAF+JG58gTIzin7J0WyIzYdvy67PQrgj3HtTK0w4p+xtyexYz0S80+fxqB0Nndhd8cs7jwFl0vOYWZviGg2DtVug7iLWySRINkHtUlpStMj48SJ4GSqCkdOwqTKpsWlsLMHdakp9wXCrbpFCvdvd4d93r3J+aGkfDAzTOlVTgCUxpTeEBEJVbCvG2buBHI62dCK3utajDNcanrnRfeafYwJROZuNdHdmYvcuSrzRH7kR0lA/8BG5bEE9p+j/UBW9eyRT8/xD1gBnzR1T7nmjX3SQvZ6/Lr5JxbiEDgGuJJVK95oD5ng2ypvagSwVogthUIgFDA72EsG2DhUzxAGkaoe9tbfp9R72mWyUYeKsXMPVQk6q0JrTtEQZQxbYxchCRVHwkZX6xOUoKqlaM6lFM2mWc3Mi8HnJLxOkcWjcBsBdXchPksOL36CQSE8ErPj32No1Jmu6Rw/nRmHTI6O4o5HeqqZRT5aXI9d+adUVI/oycm3ecsfY9Aj41uPX5AZgT15FOfDj+8D+gXb4AB04pEwOgudd7vEoDOyH2UM1yUGBcrqat3f+r0O4kyo8fijJrBiUqphmyBHZhCcsxmThzGrGnF0MBFKDFGEwbTd78q+7e+VjMtDEip1F0y9JoUN4okZIKZALMLUgDm7FUfbxLYJIQpLNaN5Qwyy4gCTS2NpKEo2cM6adllVzVCKayFTYhRE8UZMeS2XG5eL7Lj0Al4pTUkafUI2AWREjpm4TxvhZuQNFuFvQFoutGEMsR+XuGq84WOEa5mF+kBERqd8pdZHRnYRdKJ3Iz4Jk7cJXmchtGolq/NECMcwyAwHnA3kqN42R5peecYFwM2RQ8uv83uX0IlqcqkSRWVkJYCoEMhADrfqYSRV4xNxBH3ULVycjTl8d251AQ0n0fejhBrD6tqGl/CYrRwZ8QkLBGIDKUhB5oULm4tzS4US52jEMCsowUJS4nUCp7SL6T7QLqVhqErWuptmI4qRmlp7HL0zABaIiAgkUIgnuYzH9ZpHBhAJo3r91xXHZKVl9/HgwtxdjYwDkTDViioLNDAimDN7+oxF6sLhUy6eC4aVJqboD23f45tx3mZsd4bYTrDm95HmM0x2PqLFCfkh45lh+mx4fwCHsByVgMAcAktk64mZlzmyzQsZ4DVVx5EPrvCtqlKT5xupqjTBHapuxSy7plH/7xMRW44/1IIbXt1LjTX3Fj04kVelfKBFs6pNUB8duRWatSQtySxbzijFx9yLDAmRmRuS0FCz8e46xi4we6AwypdL4jL7Ts7SEtVIcIDsLGLpbLPqca4RWwB8ErRHfHZmgD5TM008xaQjOXqdVaQ3r/V87eS4foTXPBNKfLIITp/PWZQljJ51VUkDz91aHfHsMVq1Oyc4b1an15wAR9ryY0C0MhvzqMQDtdpdN92zqKmwjkjOxioadXxcpZQ5RaOqVm6HFNWMIwqQpR1o7+lw6LcQ9Sm2CYCXqVJF7TyMQoNXAHcoXM1VnEHmOGVbtNJNM1TMb6TFSzFXda2M4lyvhpzMvaqwXFrIiji61DxjJIBi4vym7Z/xqIF4pPJefRY/ZRZiiLvz93mm1Xa2/3YOoNNttDg1i4v4wdiIFspjLKiQ/wBr7Xmbj8HH8egf0u33teVpEWOwhhbNKhziSSC8mS0IILmd2UYqFqD5r7ohuauXvWkPiVKX30wfL2yA1wwLMHUCgQQlaiI3ykTUZACVBxUh4wg1zwXmblay94eiyay4Fy9F6kmoczM3BjUsTeura4mX4AAyIYQxC4AfU2UAXnkodwfMJxwwntFTuJLQnEyiFKKJvyPSc0PLqRLqdAUyTxp6wAF2q48y08QzHIGsqjuWGHTJPi5ftATiJYJ8AgoWQ50h+GR3z4wURLMyHNP4Zgxa2cpZmpw7nMbwI+Lia7zw/LVD0zclULh4tco7ur8dYowiknM2M7VR924mzBziZCUSCUGMxiI29UPf76UnStAkxVjzMI2zVo8qVUKvLhwhjCl6K1w4nA/7xGahEWVtjC17faAQQD2BVV3VU8o2iBmVzKqqINU8q5nMLHRtF1cUTC5D9/KCvXElbmq4PRXVYAAxGMCADKi7uEdTHm4sa3K/ZyqsPWJIJqAr8RWccBU4sIRBAwlYEC9ME4uYRlWPoS4ra6m6wGmna9pmGo1jo2SjRxPSqDSphS5rSBoPNdrQXQAENZ9xwALWJ27s6GQ0ZlAkIdSCWeyzwocdmNU/ABEj4EgljMfEGyOrU/n7mS2bTh1Vu4NBMdUQAEH1LBHVeHIqX8EiPsd8HmPlxm7nz5hUe6ihOKpE5GARIU0vf3IJeLl5+f7mLTF7CCVnp6Y6/dRO2FsyBWAZBmt5w56YQEIFAXdNUjWzUgZ3Z5flmXRf+uhI4CmTeZ2SjsvGMPfiVgn6dCDNM0FUXVVzKtCa+O+YNWrGBEQkCGQuDTebprtoFqsGuE8OfAq4sjklYweiJy2q5Z3kg6aHUgbXXkN0CiYXDXXu7G0DrBpfN6khERJLo+WG5phrAEaMs2wR8+qPSIKeTjGykG+ecFvxj1P5j7Gk/2XbctmP2PSHiG4+JdeeYrkCMTWh6xoOxZncSo2OxznuP2HJ1DKNynLS4m5jJv+KEWre5yWyXz4bQghLL5VcPQoLinnOg8WG2ZlHMsdSq/nWqE43zQDVr3ZMaVFJJHdgiXb904urLy66L2JwYWbAMfprAVRIkE13++Q77e/LzT98ONzuH97d8sOrVEqxbIRiWdAxLEYRdnf/tjFqQC/j67/4fH2xuvx8iA2RDY100kQvo+kf8BpB+HjpP7YrsxpyvMdnnWtVXdGEBkcO9ZQrXehlnoptmno+xvqcPjuLicB/IlhfPu4+yyI+swFP3lxfrWpy0oOJMXNDm+ifX928Hrqey9CYFg++BCxmdldiAnz0pQxOAiLKvZdckMeRVDBQyyGE2WB2DqBLYc3dwe7O7rXmwZjyYfbVECd3aKnxpmM2RvfZq8Vxkmfa4iqsnnftRXQyZpm0d16z95KwQnPywwe9+493/b3tf1+8Z9ptHmxfvLi4MSyYaHHXnFMt5jTIlR8yfvuAna1Wbf5fba5ebJ69aty1qLKHmkqHiHxM7nDeHm/Pcm/mD24n6Q+8qqR5NL0tsS8vSmp84gA8+aLH4zgdxh+Lj+dd/1FAv4QVqmYaQ/WR4Ih4SaEhKijMhmMsGo2+RA7YaDxnloYoMNw1aR6IJv2+L/wZ/COeQKFt25rFeaTUUkwdikAcOVqyebRmNeviVN3CWatz1UguuBqfRKQmHrt+3Vx8tnrx5WW7joECUACq4jkz3Evf9+++2e3f6vu/H3bf7Ie+9IOCTDHwSrS6STuZuUDNbaBanMyHD9+t2m4d5cObXii9fddfffbwZ//85fqrQOKrmm6EiM692k6W/snlmAG3sgq5ENU0tVRZdHZ3Nw9BmLnWoMYj3POJLQcwJ8Vd3HyOQU+B4xNdfk+bMcsSND+WEueMyPJUWcXMyExa4d4BFFFu8fJPL/c3h4FMc3Jj1VyRIo2Vj02EWRCChMiFSsnm6rqXMjQU8gI6IYFVtRJ6ETlbw/Dis/X+YUj7RtVLsl6UsRc1c+4tIvTHvpgrvNZjbaYOjTFWeSIE6aTNCTEIUXn2vF39Yn3x/KJZBZdMTaHSgMTUocrm7353d/vr/d1f+2Ffvv3wbh8STf5KZu77NHluT+WRFi1GK6b3PVfseK1tf59236Y//+dfXL1ap58PF7yOCiMDE1sxwDlkB1Br2h6lZqPR4FQ1AGxQOAnn6rnoYqaA1+hvwxi3iWIGA8/o7UT95O7IGQBqioLJO3vSvnBVT9Z4G1V1JMKxNMyZRfMc7U3+viO0TZ1MN48HbELnY19TGE+VX0eJW0RmSb12icXiuJqq1dJZSsCgo3FPET2sX+DFX16++/omJA7Dtdp9ws4Fve45rMidA0kgFy8ovhNVqJq7IiSzE0OU5RMetOQ5bw/MLKyvO3XLeSBzd28oCHeAOrGCaskZn3KCzgquEWQRCEJzcZ1iIHAbQ+TuqllvmqaNzNVuW72e4ewwDMPh9s329u1ht8OQXflsWU/kmMcSyci0TWzfoCU49gd68+2Hh/3ui+dXelEgOYYVYuf7PRE5GYMcfEwB/hTHM2q+cES8j3HkH8AXnr1lyYPOH2xKy3rGk5y/7hGS/iEcxadHNa3Dp+a1fEvlb7ouXFyu+nU/aEEiSBCGII6efpPdF4AbmY65GB6vntsUncvnKRABhM9++ezypr37Zui3eft+VwZmglIu0MKHgODENqlhR7PmjIHQinCNR4lROmp5Td1r6S67559vrp91HAKzAIE8wPrixYn328Pf/i9/f/tvfXfX58QuPKyMis+yGi3s0fVFZ5EG5lqd+Ote3iNFtLF4+Vtr4zZv0+t/fPH8l60NFg4ZLRNBqheeTz6mj0Gz9lxNA+ZVPmIRgAGr6f6qPev44FNg9+TWPglCZwD6+Bw+3S2d/HQGNz8KVqcoyJmmf/pmm8VfIoqg159fXq5jQ7ubd7uH34oOHTN3GlkcMRERg9wN2edETu4+KZiXjbEIz/KFHpeZQ7vxtttcvtz0h3z7Ie7fDulODvfGRkwFhStIAmDmRDaDzlg1LEACQqTVKpag1591F6/i6mq1vliJGXmGgyhA1cShdvvb3Yevt9/9fzTfalbPXNR1P/SdSHXDQ2XH/Ihmzlz46qIuMWhwUahh4MFSIv/bh7dvbn9xePln/+xncrEa+p2AhEiMAQNX1xqbN/Wk5yhkTiAxmDkHIgIxu1dCP2f6rOMcj+sMYfSUudLd+bT+pE3VNjDS36k47wJulo8vv9JizERkp6Tmo/D1dDvJ9Ps9t04osf6NCC7aXITXf3ax/qx5g7ubt70XKsmZDqVlGMHcEyGbOZnBR7My+7nhWiY2YxzMzMsSUQjV872jtpFNbDn4QchyUfUGwUKc00ACiDHOxeJFJESEwE0bYsObTVeu+NnrdXfp7aqhUdFD5AbKrl6o5EN689t3N78e8L412RYq6m7u5MieaVHvUCgsAfR7lr64sVLDyZwcoYT4Ae//dn/9k7vVyyE6AYEnUW65yo8xVt1lmozKgLEcLTdnvnB2akCfPz/u9knUeIY4F/d8arLL8/CJzn9Ie3ScPtXPkk0EQBAnJea44Zb86nm32yYdBF4IlnkqGaIGBfjxu47tEf8zXhxptf/6f1j+rJYO2+H+w2EYtN+VdTYz5KSlWCnFSleT2oUoXdfkl0pEXdc1TbNery+eWxXhK1RxdDZGUQvY23C4zzd/t//6f7rfPgyZdDscaiwEgJqe+XQ9TotM+omSrJpej1q9uGFHJI+jW1uMHFZd++UXVy9ePHv5v2+aqyBCXXqBxNbtMGsrAaYwiyw0WoBO3ny2Sec/n6TikBlS9ZTxn2uN4lQhVRFnzcK1ODlHGZ+IzpP1sSwRNp3nHWEsOKIzgX1mz55mb47utic6qZHL4qP5lIjYxsWvtRPe3tze/UM6vC+W8t3wtrPLw2GofhqlmPrJbo7uRDPKpzFu0cxEBEZUa8uwAo9S30iMq40wtZatXGtJ2Qw5q6mXUoiiWWGGBG6aZvVSQghN04gIMWc9zKtMRCwMNwg4kBd/+O5w926bs5rZQYdz5unxgi3aY6K5ZIXVFpktAZAp7DD0d7fs7t3tM2kCOgVlxJN8vE92/se0H4hBZ8D9GCv8n2MM/8nbcq8vLtZ5Azt4hrbWeO/Ldj7C0XlyJE00DXv8Wx00oJWZCX6aOlYzOxC7gtY79xwbGj3YyRTmZcyMzw5AnGIUsMNSSimGFU0eN8IMKw71SMX1zXf33/6rh9u3D4e9Zc8H76Mc89kSEehTHp8TJjtySydQpe5wYy9EBApSnJCL3D0Mh0H1/zm8/vmLL375LG22FChYN6MBd3c74d4+piP8RJuJ1yyGzx0uAfG4PVOOsdlEvLyZnhB0ToX6aQ3ok6zAPMflxT8GatkXcZwYKTERCcDMl4375xqiPrwtWeNuW02go2BEpzBGYxJ+YJFlG/NCuTnAVLVgHM4UwUbEHEjAbkQISKAAkurUBr2sbC5gcM8Y9X1EFMKapIyCLRGYgUIs/W3a3eTv/t+3999pypSoJNem6awc65oR4VHGsxOpoozZbcbw1tlnp+4xW3BA2Z1AhEDBweqWUXRI736r+4e0ftZd/UmgCLYTr4DzjTzHpt+DXOmUBz37O9/DC2PJUzT0eOdTLzkBUJmgvd76lK/B09D5R7YZQGs71vVjJvch6eZ55LhKvd9+27jvFyM5X8Uxmmiar9lxwCJCKMxMBAnE7OHM0ObuZmASHd1umIlHt14HRIHJfZg8dqs6ivFfmgB07MtU9f797uG71P/etsNuyAltYBbTc8Gcz3OAn3yVqcqHu2NRTXk0ydQ6GO7uMEIxZveqYC/IWpr+UB4+bDevr5pwZOAWvCPNqI4+OYwnrFInMLoATT/BDViA4OyLPl/BZN6fenv8kh/a/v9A2WsbXaIwHulADaJSzGpmeuQb6xr4I/HrTGCiSU0kIjznLCQHHmW3i8FnAaK4B+nsRIUxVj2pgEJIPhm4SARBUJNHkhsN+/4hv9G3/+b+w3f7N989IIJbVlFjJE0hVV99JnYiZ2UiokBgOIFOg4MDACpulBO5C4kxyFJuOEgQZb5P/QEmTqQ2GAXn6NKURhCTp9Cv3v377WrTPPvqmq8HoQYK4QaaEU6SO2PKmAU/ZqgCcExsO/L1VBeDZgWJwynVjzXVsfuYL5ZYCNV440RsVmgscVZ7mjDiUUZ2OromH3OO1vt0Xhl3uFfxQt3A5Ew8+eOOtyABBA/wGrZ1TMKKRxh9MQDgcU4rDahug2QApAlWlKiqoj3LnoeL8uD7dx/E9rmAOEj0nLOT1qz7k6WGqITq2snizNxQqDFaxN40vMgVTw4K5/G+Z+4881RnJDFt2QiulYd1QGsZHmcpcNe+v/m7/c1v7t799pAOYEj2rDUdQc2XU6NJheq2BSZmpgDUlA+nJ6cUIwpG5kFMIVFpTG1AztSu46WGvB8OB3ejgQaAA9sh7mJDl7YWLfv+cPv2wZletJs2FKIAL3ZWfgELHLmEmvnHWeE1AjFVr97xogSfvCW8nEpFzFg4ET/Oq7Pkw/Bj2sjIwgEyx3m6AZ84pbrNn+z70yxBGR0cxqlbzrFpAXjKXpQDf/f1d9t3w8PtftiNNlGf1ZkYYaSeojBqhY0FzCQk1QUWVELgMpZtHs9keFLAXK7XTMLca+LgkSRNX4+PwuCiYLdS8nY4vMnbdzkNbiARychVRqs+7NU+yuy1UpZUNM1eMaic0laiWH1gCOJC4AMADzVRuyZTh7vFuw99f9CSE4BuE5rQrpquwIvboNbvSns76JedSyBRcBxNAh/bHncsbNWP2FNg5AdGt6bqcFiVzRPTebKk7mMEwVk3fwyzOMnJjqNLwRJHzkTYZn34D+95+dWmgzkOtv6oanAEduXdh93utuQeWUHhREslc1QRkxOxMmCYfJ2oeuhB3akKQcd5uQc+JaZjhNKpguDIsfnRP8gdtb7HOAFzBiWoDujv9Lv/sP/2r7f7XRkcSjrQXikANEbgWk0rPvrqsyBWW31gMJwknAJotby4ew0kIQ7uLpFTydn0MBxKj5s38a//bbm9HTYeTcNqLfZhZZ9frH92i86Dlv2HFFLT/zQ3UUJkZXcw2TKU4hEGneSweR3GD3V3pnqHY27eqd44O4JXYlqFAGOZ9vuTxsy65n8kvGLB4bkxqDIMTt+jJ3mCZi6/MoBaA2katZcy5BSaRt3e/1o//HZve8LgCEfdRU0y7DYQjcjNyUPEmLlBwAwrVp1ZVM3cqxrYpuq0TxdRmNtReviIDeCENhmMLCXfb/32zTDstSglUoNao9A4aRVGzqf+r6YcCwJmUGBnc6CRk4GpKkZ+j9xAFM3MGBnGChYR8fvbdNi2XtYPsYc3JfN3t73F+OVPclRRhyezwcpQNBfrpBaglI8nGqk7dGZoXVKYPN2WUjIzmWoDLBLl1Zu9+d7MKE9h1h/YaJKzqjyyHHyFCICAk4DgJ9unR0hnuq0pM6gTDsNw827oHzQUUVVrVE7SVBHXyi5TyKeQVCVABVDXmjLX6RQ5VhgNTnG+Oo1jnDLRmL+LJsyeRY/6sMqOLvKAejRK2n+7v/tuf/N2NwRSApKRI2JtRZmZgoCduW2yc2COIAGxty2Ias7HKCLEJ3hlrlNfSlFVczdDyRBqAaSypwvrvqDy7/pVeLnlG5OCZv32Tu4f0stfdt4EO+RBPKf9+neBQ0ubuHaPROAxtyo4ApVOK8jUk7ml4kYde0h3pQxlkB4UGEJ786J56IGanlfNbK1hKJrUmtVFaJK0US4iIrXRaFUDodxhRN6EHmAgEIIg8phRmmuAxVzooJbfDFNM0nhZTmWXynpZDTuHlqookJpWPOsBADuLCDnPvNkZTTgTj2hyDFgeTpYCVM2OExHlmOLQe6JDfP/rYfe3H1ilGhwpUbGqXEOMJMHAIYRAAmcY1PssIjzFMFksTgIn4liBMudc+SUrEuYKpxjZyvPAvxN1yWlbSn9E1OPQ3w9vfnN7/2awPdRqDN0US9rCiQQuDnbvLjoJJB1JIBJErgAqPFYP0iWANk0AUN1TqgJY1dPgOZm7b1Yxk7941m5Wd4eH3aFhwNPwsFrLarNygytxYI4swsOh9NvDRQreijsTB7gDBT4AyAAhwHkYgqqWvd883JcM7B3ZM0w9mVm625eUcejQCgIQYWStHQKFkq18eMsr5iDSBcAb83gR3V2tMOPq6uLy81V1AWMxl35yFkFVmM7FYkZ1dqUzp2uOj3yt/BvxiGnmXHyn0acnWra5n+VFf6QGmq/U3NM5WjL3Ih9+9e7+6/vtYZ9SKqrFjYRhxGwS6pYKCUJgEoLUgPllnnJjDqZupqpTGpUa3m4wQ5gimn36ewKgZ6OkUZKfSYXhKDa5u6eD7m7y8JCrX+18BInIIxjE5IE4EFyyM0sIHCDCMQpPyUjP1mVemhACK4lxKUUEbsWUzSw2BKf1Km4uOfX9mtZMEXnfao6WGI27gwXsHFxT0b5UbY2xs2o1rNWDmpw8kxXdPZScc9nZvs+WiRLEOQ+mxO407E2z4VCIjQnUMdi1RHJxVkDVgqlYIlbUIO2aHJKZE2O/cSIO0SUYtVJRUp0uFhswEa5PlQ/FErkQTTZjWyplZ5T5mFc52+vlDZ+m+CWYJyDT9u12uE05aVEvDnNiI69FZbwWLnQRlsDE7OyA81h1w1Vr+m6uiQdH5fVYLsDhbGaBpXI/o2rzbFBLEvAYQEWOhEBVde/3X+f+HUqSIgdLCoyZaZmpawSwIByZYozxAiLSrpgCc6BIYcKdbmZM8USTwDUwf4x4lhJ8TKqWzHhNzaDl6qr8V38VP7zp337taX8wyj/9qnvx+qILA4OYQkZiCdo3eVsouwllL51HNzbjlODuu4TDXZ973d+lMqi7m4hl9X0hdy3QhsBEUZgQeLDGKbAwgzm2VyUNzN60EcaVwpkYgP1hEJEYYy7l3fvbDa6JqGkRI3cr7y4iEcVILCZhVGbRZKeBHmHrE9BZP4xVT6ZkU2dM5xnFO4PgUTJYMADz/ZX7AjCiM/dMyXo8vDnc/e6gD1bMirkB6p5Nm1mdXEk9g2oGU1aChyCq6m41yZSDzVTVtThAZqZjXgioegBPZ7TqOD+SwuNJKk+RYYbKzlo5vO3vvt7nPQrjDncdX9TkIHV6Xa0X0opEkS5ebJgDxVXg4Mwgb+bEzaWUWuR+XCaz2Ma6gkZGRIEcznCuVU7ARNkMw5//cvX5l/Yn/0SsD0pSru/iets9rAQMcNI9LKzTKu+T9jnnITvI8nDgobf721yK7R407wsriZIbMhfnUkpplANYxQtnDiSNRwcNF64Ghe8CkWjYIbgLNFIcBgrRGjcQGmFlYnKBO1yx3x2I6HAYcwlK9yBCF5dt28b1Raw+BVTrsrlEjMVRP7YRSwRZeYEZF57xmme7ubx4RtPPOQeay1W6u7l76Xe2a26/vU0HmIasrg53JDV371ZVb8DualX1YlyZbIO58zLIbhhyzlmLmzlT8DFzDQA3RRiVeT6q+nS0uo519UbHk0lICrNxpaoupJCrFHeX222JD7LfqbbJNW/6z5USXGtJexFYaEQoNhYCrTtpViQiTWzqupibFyVmIZIQ1UlEHNndQwgHdSaOxKsQELArB81GwbsVm0HNOo9UQgraPmv5kEvXa3Hxjg+8CiTBOeRVWAUPYaNo6SGlS19LDG9vcXiTMTh2SqqeixXrU2niqlk1OVHJBQgeJTMXTa7MxMTsBAvK5DBiWGAkQqCWnHyg7K0YBaulrSxzJubgArUMqqTaerfK9dyTst/vBunSULrN82hwI4sk6+waDICTORRw4o1PGZGY2YZMxMKjPnuK/h7ZWZFKSZUmE+tSoliGLdSE2ktpiRCqSMdcUTIcTkxOrGa6cv0bWu2uBntPEkDRylA9kwmW+i5Eh8MbMjNBcBcyBGJitsomk5hrGpJ5dZ2HFS2qmkt1XCwZZhagR/OXuz+2rcyDfoJr0QaWsx8ggbw/HIacc85ZbSgFFEeLak3c3HaIDXcrCRGrKkPIkVmuXpIzlWfiiZuAuzchsgPu6jXRYWDU0GhV1SYSoaaUciZ/1q6LlMwVDYtZMXOzMZPWVrPm9mKLd3+/U/UyoL/3NJjlwkQI4k4EUdWUUhMRhMzMLJk6D00uSUWqxK+k67Zz95KyFpcok6s8cjIiwoEAj10MTUcFJakVd2F0hYAQxUy1qDlbKfo2M/Nwk3YvgkReP994i9B6U6XYSjmIyA3uQlWlaFUqmlKtLVHp0Tow0248ao8397jLZ7fXzt0drqoBqzff3N59u8tZA/F+35sVABJIJMCtMo1mbgpTp1qtC0pGWbXkyhZazjpRQq81i2usvDtVL/kAVMIz20uemMA8wnMQVXErCg/kpqlWuK9iOzFq2tjRM0UoRMQGTcshUIgglrNFOfk78fhVoufJJOPCcGcC2CGohYwDu7MUMiECcWRiCShuDriRyGhSE2FmJS1qaa9D8mHQ4JSHkLOBiEJw82r9djdVr3SEK+vjbupWCGaMkUe0iJEbBkRkLnROJG6kCmYQCRRwL6moOinFdS0KwKZcNLupmWmGwoKTcolRhJOvQuEYeDSRVLrFPOvb6wEe40XdfQlTn5aHPrK5n/Ii9amg+mhcTWF3P6RDrmA7h+CSETODzAACqRHMxQwFPtbBNNOajQY1XydQE42MfVgtwjRSeVRXupoqGkTQRz7kJ6qHM/jVRBFWAJb+oc9prIJA5CFA4SCIkAQPQUKkpgltG1mMebTcPObNxw8Edw9BiEjV2eFMxIwgAFxzzblbAbSNYIcFNTEGe3GQBqbqpFdqslIfMWiwThO9+7bPQzClpjGiAApBIMJ90hqsUJnxksZqOMxMIINxcGYew+pK2G17Iu8u2iZKCFzLXNU41uJWEoXA7uqxIXOPwkyIMgx9deN29zxYydUIzESSHb6TwkjbrQTOn11evGYRWa1rDQuDjmUgR135JLbX5Tzq0U9E+4+iz+XmPnK/P7nZzJzANBax3n6ww50it4BkKzMmrrHFsSMDw5GVDQwoszJP5JlYVd1pQpla08qWAjODzWl1iRCCj0LidOYe5T87k+xOR32ABHAghOFBNXlFIWaFyTBWnwAzS6AYuhiaEGq4SqlZD6YQVZ5zocwsKYDKBwC1Wh6UUEm8nGpnqwP1ZNBzhToRxAns7jKZ1CrT1pSVFn23uwvtVdut93RAOQSOuSSz4rhw5MqWOJRLyDljChmzzkIQCexQM/N7IyGIO1khHmVRd4DXm5CzM4cQJTb9g23JEdgAB7On4JUwgwkSQc6EFh44B817jcZN8qEchgd9e1/aNn7x1fXVs0hk7MeK8Esp20c/FVlu1lL9+dhD5fHmfgKJVuznkxvu4d7yzil5okR0VOkTETMpisMdDIPB4DrlmHeHEYuqu5E75ZwJoVJ2MzOF0BQ8Vt3C5pHV/ea5DObkCbE0J4y1H2lKe95ET7KKQ+lz7i/7D/eRB2ui6cYJklw4ijgF41a1G6wxkwaAQMhYWEouAOKqKU0Pja5BnISdAoi8aK+GGKVm4yB3NhCxZThRNgWRhJC1IEgIgBUrJRAXL0KoT2QyCElH1ljhsis9FZbcadBD3sUcuOcC9RV5CMCQVYSDIJt5k/eSGoXrhVNTWZqghUXawJQv78e4OSYAuc/O4kROcnBDgFwMYFeidQlEZJXXKtZQVMuGPkTmhlJLqsrMTRR3bqgXkT6pKuIQVt+0JraHCeXmgqW4CBPApETgyT5IC4jETIhcKstItQBBjfCZANW5ws1IYVXPfFjHoItqZK5sjBvBBTA6PLCohVC8YwvMW3dy1PMZkARV6xkcDI1jgCGquzuxFzczFCc05lRKMXOAokgpVZKrkpkG0Em+90lpPAr1yzNKzH4a/UTOxd2GcrjPw+2hWnmMvJKDRqSaSkVkvW5WF6tGAgGqaqpWlFkrfjLzHLRSBQNZQAjVlsBEohQqnq2JltxhQm5kTCbkjlraxomarpUYDvtSyYPXYJVQPTTdzdUcDhRPRshgd2ILEABWHOaeM2tgMaWkmmEAMpib0EoTecylo+YKR9u2k+xsZsa10IMD7NFjddghZ1drYjPT3KZpzB3GoOAiEIo6WnHrakgDBAkgah1mu5TLUPa/216l1WazevYVibuLEBWCizdYcJPLD49x4bnyaJEtgkZLAX3sWUxkJJuaWYhr1Q9uxqQ1VYl7ZTzdLIUQAGPU4nSUF3UDAYAD1MyAAsBtUSimlNGuOXNW4UxaO0mOu8D8FaTPeVAwYEVRBtV+5HKdRx0xyJg5BA6BR1UFWeU2UPUoPgb4aRkVywZMMaFERKZCREbCrEumymsyciafVNI133lxc6s5LGATZ0001hStHFDJroXUgAJzD2EsMAyFF5SheFIT8aAGE47VUA8ABpKl1a3WD6nLWtxravPqv3PMve1k7uriPiWXAwB2IlQFGQchYXKGUa3SyzFykCghqHrhQ1ZjV6X04KH49rKPMXJcsRBQVVZPZ+ygR87552A3PbuMKvkEw3oKZGQEh1Z1yqytctiU0JTMqtra1EcX7Nq5NBXjj71pObKwsz7BJ0+ogEdy+wyXFc7mx8j9zNJERAYumXf32e9gbiNhGbGvikjbNiyqVkDFHWZVNyJFC+B1aczgPRxOUlTIRS21XCUpghu5VQO+1zQpIlqrMXgN4W0bZ9JSUkqFphJok26XocyBaBSxD70nVUUrxQBIp7YmIrKDuiIl0/skInQRXMAkdOkInrSEA7RhEaEpBS5GjOW1KumQLTCLsERKD4WIQmQihnumPG9A3VwH1EzNYBQaIubYNFT1kapa2NXY2V21FDdnafIe94f+oeR23cJjdxUQXHjMAllt9rwIKnwMZ2dA7FPQAI/Ji06Cq87R7RQVWAcZN2W1ido7HSgiFLMqSbqTu5pWrDwCjk1Fv+uokg5k5EakIBKfDtIUuzVqZ0MIbdsG+Gk6bdJZpXYE0+mzhFODhGpRSqX0u2z7gjFCiZjZCcLk7rkMMRAzD6nPRnmAFpAJjNy9qn4ApWJEjugWXNgYa2ZmUREiLmeyW3AaqysY2GmfkpnN4akcxKorlhOImpacyMz6VFS15HU2MmFXR1ENoQZCWCHPasRduxGRA+Wk2bWsnsfQsCCSs2l19qt7CUee151ZYhSnAvZsZcy2NjqPE0mt0isjYvDkRkglZyNz7ypouZnlnDpbEZGmxMxs6CxkU3Ivnoh5NVy40u23hytbdRexdCNPxiONONEuPWm4XgKoTcV86y7MkP0YH2Oh2A8hbJ6X65fr/rakWxWPOetozKzKHni1sZs5kZep4vV4bITq7lfPpJqawN1FOIQw1jwiWq1WXdcFs4SlUEwVXielmutJrsAx85YTAWTOiJS4b6xvqE3qZgQvWiG8sAihCUTg3BOKmMGG4gqyrGMNMWOxpgmlCtpwNrA1UcitQCpPTSZeCbGRu1tN0FlpOoCGkJVyUlOyIspJiYEGchDRIo17UXW3Btb4zkjdPeYuyeqQW25kAzeTA8hcw46TEFbUSramiYFjCAzhoc8rqFGjyVoPDdMtM7tyg7zhwrbeh+hkCaFaQEAQpsasU9N1tiKNk5BRDl7IyAtTKQ3MczawWptM0aiXbO4s7qSKrF0WgTbOhRnRAhsZDnm4DdRT/kIj06UQeXRuaJaZqsg7pUWZju6J7wjMppQ8aq4yVaIfb6Y4JSYfna/DVBSKDJohL+CDPV9dpZK9l6FsJbJSNKwJd/CxAjIA8kZLcXcWFyGuGAaOABCMi0LZQcxuBb5b0/NA7SrQ9ZePguYeR6CftQmSR+FA1VNKacRho/7KUQMQDKhlFVQVgxrUrBg511PGgqrQMQW3ABsJOzMCiicRCaFGBLiqnfk6+WkhrIWm17b3ctindhU2l41hYGeRhoDsUFXrSJhyNlnJarOJWGtvVrIBHELL5iwMgZuzhVV0VzNU8aiUzlaEq7BXfdAiKMiAS+xjINrTr0Q2sd3AY7PNzshtdhZK/iF8UGikWKtUPMsX5kaXIV6vd/u+NSspe8FGOt2jzz0RdaumhufGVeBA3oolp8IOhUOL931S1c0hWiBfi7mZpiDxyV2b28e0SD+kLbVRl88v+y+05J1+vbUBkLVZAXkmQ+ilzJEXk8A9sUM+edMB5AAzFSuNBKFRT9r4KzDLil/++fWX//jqewH03Mmgvma0zjuboT+U/pBNYU5ubOZO5CMIWRpGv5Y0lBru6WRE7F6YGVSIqFFfQUQEQSFwhaGECJHIzCAn5plEAvCSZ8WtL9roVzUgHyxEAouTulMafOjHBJK2YhbYNlvyELG73QcvV1fr9vKiL7kpJVsBa9PFuI7cUmY18wh2xcD3xmAKZlLULF883Pcledr2qtrdfOnxobscnn2G7SpTg4ZZnMNe+Du2YpnV3LP5g7wNgcO6mIBC06w6Zqfm0OMQmCgHAtQARq3poAwmKq7uXmMNiDklFFX7urTr2HxO7Vpq2BYekfIzCFvybE/C3/cCKDNbKpsvOrS0e98//L6///beEciFuC/Wz6WsRx4XM3cL96qrnJldpszMkeAEFSGh+Nkvrq6/WH32F6v2mh8BKJ0CqD/lHjrDKIiIbVRQH8lHxaCTTm6SEM2r3yUcRs4Od6sRM6WYFwE5uVAlDdGO4yGf66f4bJLFyHEvl7UeniZCVxIinAzsMNaiJWOy81c1gpuhDOi3uSPzNWX1VDL27ELcEAlCJKNS3RWrD1gWrLH2xNuvddjZ8Lfv77Z9Ll526op3N0N7OVw9E/sz6X7SoqVc/RweLO+uqo9CKnro08Uhry+bz396GYIX57d96lbh+ZcdRSuSucaOaQkU6ojZURMYkrNVWZjZiFSddoXM04tG3CKeFsB/IAieS0UfuafCQKNcorWXvn7FeUfdexoysRE5C/NcJdqr/mgMpB4x6FHTSkREbbuKgc0KGZomtB1e/KS7+rLrrkITw2MM6tPfT56nUXoNpFZ9NInEawVcI2dyQC27h0mFDKrFFxzq5FOcXI2MrIVI3CqvzzSm3oN5ISem0dFm9tGKNNrWhyFNouVRD3J1XVYbURaX5AQoC4cgUBvMXBXk2lBjxpyw6daXwYPLvmQl/vDWXny+jgFGeyUtkFpGwbKb6j49/9W/fXP72/27/5fbPmzX8Q5QItYA9wu6DffN736jv/73h+bwJmx4/VVkZn+LQ3+rqsSNOpnzytbEw/rqTWxQssuu09Bf/ePy7E9Xn/3Vi/U6AVCzBr5atWwGhhGYmUIolgkSiNxgZjFBmfu+SAcEHvHrRzDoDDR/GKE/6nOIIrVt3PIaP/+vX2+f5XYfvv7mph/UjbqLTcZw+qITG2z1varCNDNDmZ2F5fnLq5/99Iv2p/bZX1xKV/b7fVtehVmzMM5KjwAIHP2sxsKsY4Zpgrk7kh1K04d76/bhNvTqBVy4qgnMiJo5iwYzF2czuFbtOYXocDKFCCIV9wgQQdxRCphm534mBwvX6IWUkrsnEyqa+gxl9lCyeVErTo4g0ptaNLW9eAgIPXmhoXB1JGAQcoyFt+2+gwp/lvuDa4wxNjnni8+GuIEzBQgKBelsXzLtDutQ+OLdv+j/47/Kw1CKSL/SwvLM+yY6X0DZmoHsAOOw9zJ0JFS4d3b0xrrOLO0uFVe9lNw3dyDzgBBZI6V+Y4W2f2c33+RV2tl/Gxrhtm3BriFVUSOCqa34v3UjzSM/07O788N77brW1yQdm5nDmAOBregMVXhk7Xwk1J/YZRbZsir/XWMEx5wiJoeOrzv1XbO1n5RXv7h6eLCQdRWGzO9o39i6692GZOu4uuBhhwKmwCK7wfKhW6+kifshcROf5y405it0P8nhzw8vfnLtZJabdbtGLe47672ePHnzoB8fu4abbL3U6rPs0COg05SeYF6U0Y9/vGWK0J2KgY+vgtZXVbCejiBVZzyfEgFoWbBTi2pUtYkInKyWaHPSg7o5CkBMRCF3AnfNNBSCI5Cs4KYl5QBZP7+OkUAOFBDtKaH1hjZv/mX/+//49utv0v6GmhbSwT2TDM9eX1xdNZcvAgS74cP7X+/6rWuv7UbWm/byWXfY9a76LBawEIrB12tCvnR2brOx56LbfGB4PjQP23T3f3/zl++bn/2Tnxz+Wa/oX/IlvDP3nKZ8jpqISEIUsBtrzuqGjPvbvWobQqFjdh2Fn3NBn2hL99BR0ngEzT47PflYKn6z2XRdN/yf7n/y8yZ9q9t/177/VZMb5P1wQekZZzo87ITadUfMwhYv2zc3Dfey0vAZ8XPZ7L7Yf/mL1xcvuud/0l28arU4c6heCu4elkDpMz0+hc7FEE8HbW3q1YklBhMioyWA+mm1kXml6j2ljH6pNbqvFoVgHp3MOVCF+xBCCGwTfa/PKkbnLndXPb5l7JzBzmQ8DMVUhz0BtB9S1oGZD13PRSiEvfQw7242uATYBCYSB1d1wF2CM1Nx2d1g983263+VHn4X84ZlY921Xz+POcXQGkfSkLbDLaNsXbI2Clxdt3J1a9SnIPFKVsGxlypUkgOrcPHifbcJm6825vjwTe74YbstFJuLDeX33Tf/j/TuV7/95dWr5z9fpcH6YV+hp5Ye7BoiEhY2haEo26AlDHL7LueemqaJjXRdcJwYpZdA9oc1Po2cFhGfc22H8Jm/evmPG/9H+f6X/ubrYt/o7//u28MDy/CZZr7UIWows5SSx/DLC7aVykUIr7n7bPOLX7y+uFo1baS2FEtBLqr8DQdqjuv5xY+R6PIKPTKdoVgZzKuXH1Q+6Xc4HUH4wnNxPq8SaMKmwkJjzomFUunxwDApm5bQOdkkyI1KhqqX4sSiJBYCifTDvhysbVaRO7WM3ODKqTF1JyYt5lVFwQQmHcLut+X295qUsYmMnjptLnHxLJbMesgpo4DbEFmYe9jAPrgHi2FVSukPpQkBZO3LKNSEjsF08bq9el7ai3b9WZcG33/I7cbJRDvrrjRvtdB692H/8Gu7WDe5VdK6VuxGBpTiRDqVqgIRqRtZ9ILENgwZ4Latcgkt9w4z33a6ho9BcFrMJzbxrM3YJ+Z17My7Qb/oX64UV228fJW2tn8fbm6GdQxsDnNDw028intchOb5avOnl5svry6DSahJugTVE7gW+HNGTR72WBw+G8RyqsuWD7rfH7SEWiT0zFNvQSxGLyyr1b7c3S0EIaoIkpomiLgIiYgEDqGqHecyP8pNiyk3sU/tMYAeV606oGQtWbcDiP2+7w9mJFy+Dt99c58PabW+zFZWzbuX/7VsXktzEYTjBa1LMZAndxF5+9vdr//7AYnu4nDbfXjdSHjNF698dTmUgw0P6f09gSRcd03gV00Elz0PFA55e+kei5iRvnjefvnPQ/CgH+CqzcUOq3VYycC66/Xwgcq3W9Cz659ex6ublNr3tx59/Q//4q7/Vn/xf1m9en41z7SYWzI4E7lIYGYSaFIHTMOw94f7fVlb20ZwYUYTVlikTq5O7x9rNepjXt6zhJgzia9fl/osIsLqfbYGKV5QuHg27L7orv+RkGUJxdoVy+Xw4R5qsWtLK4Sf2bCP5EGHgFsPr9yMjMkbKCkXZpsOCIca7TnmhWK2cuJBfBzB+CFg9KmGqnkUyTGvk2zK9YfV4PsYeEimIOcApBmMVBW1Nj3U3ZghTsIaG4RGmlZWl0FEJLgIhY6kkQXWZLY8qggK3F2HwYuVROXgWixZsSylMJESq3NIPAzGb35L778etm+LNHz7ELdFKQ777QZ53bSHN+/LkJpN1/zqVyohXVz3Ifo//e9k/QUFp3UXHob9t/8ie3+5fx0ppz+TK1zr5csggbdb5GLbfVqv120Xms3++bPOr5Up4V2w3ORMQVLpE1n7XUnlnW8M3U698XW4GNB7zze/8f2t5oP36WXYCCEFfvbi55RvPuzu0u3tZvs/3R7u9/+b/2tZr7vVOrqJqZTqw85GAQZzY4qSNRO7s+iHmA5eVg/ds8bAGQWAMNXKEDOQ1X1REZgzQFqrA9ly01V1MkTXHadZtQfAxqq1GP9aE0awgXu7SQ7ZgN1VfefMHMNzBMCxSmxQUEtEFNcAyAU8ZXgRr6WrMHpi0xGD/hAGZT6LIz4jb7voK+jKXXqBFJ8SjYxRH080HrXuFCLHNsSGY8fSCAtYmJhpSnQ/42DVXHNf1Xw97pN/a/VVNTIv6lQzmpqrquZebt70777d9Q8NxbIdpDeijKQhBAmr3LSFNQ97wCVnTjcO+OX/vPuLddtdh8PgaRf7w27zssRntntAgMhziRexFNu9zzl52pUYiwQ0G8dawzq616hohMYJrr2Soxzw4e9oUGs0Y8VtgITo5JY1RFs987zTsAavnTuSEDavtXjZ77INcvt7z0m1dRGxU6I8iZwjemMiZjan4pYKxeKYYmp8cmY7123aKLfXiJ9HrmrHF+ERPzD/NAHPWc8AqheLEFVZeBSqQDTGeXxPjuixhYUZCvgIlT++d4prq6IiNXz18rJtU4zx/mafbkKfU/X3UXM9JSjV6b16KMUYQgAHip2v1nL9ouVAc7exaWjyqp0AtI6QqlGgFNOsJZmqm7LVpI6czdnA+77PzsPDZv8up9vOm43GvjCVg0E1vtitQnf1Qp9tAMLb35b9dijFc+7g4ff/Tj77qjR/RcM27n8XN5urzZcW48Ow1cKtx36P3j0Md9bfW6TNZSeg3gOha+/vDtvbMmzpOnJ8HmCy3Zl4QxZW9yi7vbW5ietvf/OhvIvU+PWfNOvLqNEunq/aCwlXbNCbN9ub983Dg4YYYgiiA9G1u+ecRCREplq6tTqG0FgqmCfsmATm1PfclSBHseZpBBSpFgmH1XLep5t/Vt3VTmutTGmnZk+/00JWeWYAiIgsT9pJwH2utDtfOXmvWwCqL5qC8qgyXXgwfQqozybJjMCSI8JKwkZ8yzwwmTGz2uNcajUVrhOBGRJZAmLD0kAaENnoXweqiezO3jsrlDA5CKvWOEBzd2InoBRyh2kwk3SAWNw0V/lSm+ftOsa7d7cll/ZZi1xyVk4MgqrmslMlIiEKZtjdlfUQUHy4Te4+GJdEMFK4HlKzCaG1zUZioe2hDP1u0zXDHvv+4NB84JygIYM0xIKoKKnk1K6abFkaJ7H7m14/kLS2fi0c4PCrZxrXgcSHRGXPwx2XQyQSgzoST9XtWYwpMrOP6aSOnLeP/o2ubjDPJZRsIuQ0h8JNyR+Xq+rARDzV/bEIsWT3sVA4YnRHmTztH6FmXuZ+IyLHDGMLgJwA9LwuVL1uIEPND1rTAgFVc/g9ADrzMWbG5IpUkKlF3Mjqat3npDYU95zzafGqI0PNzCGEGHl10Vw8C6El0IAgTnADgRQOnQ0GNc+3mVkppmruNZK1kntyBxlThASooqinXg6JvvnNh5v3veRu/XIXXxZp8eqr9Sau/uFXfnNXbr6l0F6aNyX3ZCHAiDpC2PL7b7+Nci/P0Dx8+4GGTu+vRJv97d1Q/B9/df3qy3YYDv5+r8z37iGEdJ/3DxfUrN+9eVMOK1hon/XXP71+ftXST0237e9/927zUujqUkK7zRTyJYcMKbl3sxx5tfl5Eik393l/T+n9ehMfqPFslgtifA0AzswG1ND4UDmoai2pBT9pSpDRl8yOcJD1IMQ+pTz/yJ6qOcHgBdX17Ym0ZAu4WVBaoJadmvHieTs1+rv76PVTfdHlI8zf1DMI1RceZAHOrjNw84x2R0sSV//4EaHzQm3BIoohpogPe9sPX15ePqxTtxc6rJKW0gym0X1Mf4dqwwEAY3ImjdEFJXDTEAsHGbkWARjJNB4rm3vxyKLmapSyqTod1EvRbOYEUInqJKVwSVtVTdbs3tuHv+lBUaP/7oPbTbgivvuM+7W/6O+edemOumQfGk9r/OTPfrrqh+1N8kPB7sPzDx+2z/9jeP5PmP6Kwv9CLxrV0uzsesXNtt+Xb+zOym5P8UAtxZshhyaun7HZ9sVr3Lw55P7y/p6v94eh7V//lRO5/Fpsz2B66G+pL+svo+V1HoqaNDFe/ULz89gnefiaH970w91N8FfsD21silnsdrrqrPHSOwdGFLcizpaLZgsh4BkFJ1HKWVNRQiHEfOCh0RjgPvJjNQJUcyIiqjHPRGWqcBlAcKjlCf6qkHTO783SM4Apmf9oYjRlrpVzieCuZ87ETJX9BNXw6YIFB5r52DkAgwZiAGyN6VP5QZefv4eJdTZXCsRNZKNN7JrNbrjLMQTbOfH5EaTR/Fr1lKjJFMw4QNwVY6piJzJonc3IuVtRU9dMJamqUjEtbqjMuYuQQl0JYC02vNe73+ftjtW0XVER2u36xjbmuKd03XG7DivXtKMVQoxyebVpUraU2nwYPnxmOAAtY91SN6gPWowpx4N6xq4cvrO+FH3v0nNSX7/k1WXsrjgXdQuV5/Eh5L+3/JLlT5pwDb7e+52Zc3no8n5oydcbOQCa96ZhFS4irUpK27vd7r4MDzToLgh1B/FBN6+5a9pANSSSXBUDDApVdme3QFEAy4UGA7EHNzejUkrICR2Ce4WhUa4FJi+gR/LGY7bq05v/+PH5kaVc5cc0HLPe9OR1NW35EhXPshceF/KiRQb1CkozusYjRptZ0PD62YWV4opC6Sd/+epv7r7OBzTUKemMgAGUUmoqkSqM5xTNMjesxYzQSKBjWspiUzGy+iLL0KIloQxFVS2TwZ2sRuFDXE1LseEQtg/563+ze/eBlZ9v/Y0Hk27XAKGncjBR+2bXrrrw8DC0qWlE0s9/93W+YmJar4aHB/Cwadv00H/397f9wGh0j56YN6/hZH1p9+/1gCYIy4rj4CG4u8G1W8nunRGk6bIX/823D+Geb3jVddHu+fbdu5Ji2V8XC8+/KFevrJTG0+rwwX/9bzVcPRz2+f7bwYeI1Ng6sDqGu5fPm//qv/vsKnalFBJmkKWSd0rk3Mr6ouUYhAPUcjEvKlxT5QLIOfswyMYiAOfRfjHKPUZmo4vFk0D2A1U6Tz44Q+EZjOKEPZgT/FbwxRJAbcL3lWl5BKC1fHGV1kUmv09/ctBOIKZm07rHUkoTXCS8+tnL97974NLtyr25+Wngi6o6aiQxuyEnA6DoNXYiJKGyKYUWaXXd3YqaWcmkuYaOSo2641qEOLM6SpbdPb/7Dt982+xsODRvFbRL0n5Ye4i7hL1uB/FGJO3dm1ebZ3eNbB2b/V0cDts///nF69dXf3v7cP9NHO7RfrsLTfvZV77pLiyX1oOQf3Nr/S0s+NUz20S9uTnYIbSlyVsXkXIvQUjae4ve0aXCf/cPyT1fhwvhF7Bydc1Onll3/sE4uL4Y7u39b3dUGiFcrtom0rrjq77l+PD8v6HXP79ofyKWi6k2XbSc2RHaxtypcazZWJ0OYOdWIxPElZgd8FIK0iBz0OakwqyoZyzfc7adM5H9sdCJI2o8UZ/PrapUcRqhPwExnYj0oxWw3oBjXPxymMchftzsOV4xAoyYahS9tIJd2Vyut6uh3J6nCajuxrMpsmb1LsUAUWig4F61HaZa5lJdI4SqqZpZzYpWGR83txpX4MaOatuknC1lKuJG/Xr9KoZV2GUzHnzIlDJbwIObcPsMq4PTLe5faNFViJchW35o2hjDVcNI+9uhH+wnaNvoJOyFydcSyIu6tkhBtVAJHkQCZdbkSIjrEFYcLjm86RLrPeXCpWz6prTkaXOpHO0BhqYnjwXDYBiyr/SCuLgdvGih4SW/CCv96s/l+Z9cNx2gNtMuImKJ5OqsBlc3ITCBBAQQ8yRbjjXsPgFS9AgM/xgAfbL/Jcj6iT3yrMryo2Q2k+6WaHKJnTXwrqNLEgBXhVSmcEwQrxpBCkqAg4pRI0SqBvMQWs0WX+DaDs1afnXzm3TP6jDmVDSEOHgCEIgJogMhJHfmIZAaigyemBEiw9k9qiQ3KrUkiVMZ4E6qo2LV1UzVQfBgREMAMg258EbaV7xa9XrTprga+GbVDZc/v3ouh1XAb//GPDd3hys++NBk216xPwuuD3e05fL1EK4/e37VNpqddsNhCLcZL3fy5Z8f0t51i65Du+LsD7GE2zt+m61t5fUvaHMJH6Q/6FUK3WUb26+S7/JfbkMvz75tndYxlPt95uCXz0vbhtW+7b++HvYH7vOFs6kSDo3QhsXh8TK++unDxev167/qmitlQSiCQr4nbgI6bzhrYnuI5CG2SsYl204ttAzL4p25azIUp633m7S+WNEzca4R+hUunNmpmjUWrRzh0gGInUDtnBhn4iZthkIATLH6g9dMCDUHxGzcj7WsucNqUoZK8cf/UATkkAq2DlafcivC5kJenz40C3KwBHb2SbVWZTd2Jrf1Jspz717L9j66JZEmhJxI2WcruRlozGJWHDCA06ASjomjxopUXnO5Wcl05hhlcKvZTaq/MwFWQgibdbN+ofs9aTbqi3LpNty0xPrw7Mvm5kO/3kqB3w596Vpq201Hpd/mnDQHpNAWimweCltBMfTMKUSi5JoGE6d21Vrh/aAJ9ux5fPllbFeWbqmJTdlbbAPHYH2j5S5wu1mxupqX9VWTkg1bQ8L9u9K+WUvCxXNcPmufBT1sU8MdDm1K+voivvjzdXfZhJWzECm7sqs51IKJjWa2Qj4MgzA8wbQGyTjYmeAOK665JKWcQik1/MIcesbRzeGQI5D9IKe8c8B48voZMl56Qj3G6+Rgr4YnPHKYw7Ese+30/HmjmiBjVK9WJ1ZnkMFFTUMVy2u0J9hTkuCrV+HL//WztNs/vPdSQEIqAx9QgU4JEJBzzW3ChTMnicICEXcXNwvRq0huZqpGFuZ1JIKTGdwIbE6EdFAYp4NL0MuLdvXVfv0Q7h/gAEnePmS6K+u2ba43r6+u34eDDf1XmYdD8V53qs/WBUn922F74wNEyVab0nXW9XR40/76r6ntoIegA5vc5xTEQ+RW2vb5tcSQ1HoO3nK73fY3b9+zieb1sOJVS3iO1OdelfdBh9W7N/BiHW2cv738SftP/4+fv/iTtnndlhsMD/T7/++QU/vlz339Cw1hMDdLQrQa9ijFOfZIdhCsm8gSuQk5FwwQFSaKIkyoxmJ3JzGUWo1AS1YzZnE3ozOQomNycZ/TkH2knaGwJbtZdTIfu3/mDJe0fnnniKoxetidA+jcU9X1nz1slbybUzVFcFm4QrHNiZwqQqbM3JSUiXD96tnLX3r8nT38ugwDHOJgJlc3AwxESsRUMoicGUVtLKLsMHMJShCAzbwUh5WlJEjRjSrL5UQoyS0bWyxWmPnyIh5Wnd0d2mf07MvNqvO7r+PbW+bPPzTX+dn11WXTvE5+vz3sRA65ff66E5Hbh3Dba7+zq8951Zbtu52lw4dDM3zrL19xSHF/w6lc5jwE6DMvLfHh1/pdf7l+9jJe7tT3nvC8W8Hk3YfE9MyZyvZOB7Pc6MOBuWlCAB2+/Jn+k//Dy/WzZvVV5pAbanZ3Oa79p/80MAe51stNICKkSMrFLO3ZHU0rYDEDlVDULebYNqxETWESD0GREWCUAXCDoBSYQ1QWNRNTFjS6MH/MpHb+auXE+HcmQz0G0GVvZ7A9o88KjjXPCiazyxkos6mN8AMnyCmsP6FmOvnuk6+7VV2agqxCJx419cKx4+LuzhS7Z41tYaseZu6SoV4rFZysAdejZeo13xncTclMmZ051CAnm5LGHBnt0WY22fGMqJaUcY0xVJdybhE3smmtb8J97xtRtZ3dmF/JUFJPJa9Wbu1ASUK7Fd+S62D6sEUxIm6aJvW9csuRg4qQwVaRQ8e4IGnAN3cf3u3l8rW9+PlAkspBClwcbQj3dwkw2w9k6HwTWlcHx9yt8Rf/9Pmrr2Jcc8GBVIKtm1JMyyFvndHKRvMQSMiESapLDBFCCCRkzoHZvBTT6j7k7MZOAjdzQiFnkEh1SAeLEStgQC3XqX5KlP+TyEPf25Z792PfGIBxmO6OR6oHWCZmmgL+a3o0n2p6dNJCxywOACK3SLlWD2w4vvpqky9XkXj1jh6+k1st7EVdMdaWLF5FL4cRIo3iOZETFa/mhckt/BjvwUzVGO0eiEkJ5kWNhdDAEnwAScdd2SS+pmdDb3f9s75vpPvb/dcNdbQvbbrRb9Iq1Kj7tpeO+WAX3x3afdrpxe6b1buOrp77VdzmrmnMwval7d9dr/p/9OrL5/9N1zzj6+crmL/7n6/+5n+8HX6rb782bppD490LadetbPDuu61kvMrPg6HbtF/8jC6umxd/Ie0zu/hJw5HVCgYKwml4uBsOUMSuWa1aF6XdRZ8zM7oVACNkDjCAm8Yc91wicZdEDwdqiLpoRe1GAVggig6CERUmOO/3KGTPr0W9WE0ltQARLZXLQq1Rcbb3GbZMgR0m9Qsm1DhzljaV0qv7QkSoyq6pRAFjkXDrEYz6mH2FghMcLifJHMNRD/9UW/Y1i2xn7YSrmPp19yBNvAyvftY8rB62+fdh8FKoQzSmnFRDJiI3r3keZ93YOAE/twtjIhA0+R44yRjukVgJTu4tuSOuQ3Np1KRoazlYH3qLu0Gb1Muz57xZdfs7/vB7zX0S5Pg6hCbGyPZ5kQPpdni4b/okq6Fza2OTyDSlErLFRl7+kj77y7W1pYseOO6+erj6aU437eHmpyhxX/7BukAc1d17dSUqGpm/uFi9/Mv7q+fN659HarPFA61by4NajkHyUGJTctZak1hzTHeqptIRBZGGqGMW8oACNarapAB4Ts5BGmKCZxvcXZiri06tKAJA9VimisgnM/bS+cPnXTtrAuLJC2T+3RcKeSyIeCl53iMz4zM/jEew9MNbOHOjerrfyTh1difL6TgW/nVExAgQbL5q4qtL/vyL7V8Pt+93t2/2ltkzYhsNXnNMBOKik9vimM/2xFWvqkXnJQjEEJYYDjkpQC7ExBFB3Ajxkp7/aXz+PpY7Hz4YPrdB73qO3F7/5GeX+nKPd/HdAQ0R3Fyb3Z2vXz68+MtVKd2r3dC9893Wtu8ecp+un8ftvabD4ctN+/rz6/VfiV4WIjchj/zz/931F38Rdm/yX//3v797k1/fXA9vBrlzbukvd58759UXuX2RXv9vt69/edG2Mftei69wuX94MOX9bbvNAal593XMQxKkzdXD5ReRd5AW4dJ5bU1DcrHhQCF6Ls4IJOLFhyETJCdIX8jQti1YTawkrbitgk7uKQay4pDRKo9TH45RW4Sjm+8RMlgqSkSVYHBEPbToalSqGESYx8rhT9dyPkpUP4bNCI+UqE9B5/KnpTVsMVWMioIjQFePBLfSNe0XX33eP79/ebf57m+3+9uye5+Gw6GYFVPnqjDj5RzOZLnZDhFjZObQqZNb0BAd5gohwI1tMHdEDJtn9PznzX/4H/bBVv1OET+7/sn729/sfvsfwme/WNFBNd8MFkrxVYzvfrP/zNPL16vc24ff84svt19+3n7n7f2Hpr+7s7Iq0uRW/WoTLvKKoxAh5+AW11f6Kl9cy3/7xTrvy8M77H4v5aBKvo7ZA7/6s6/a54Gu99EgQlpWYP7wXve/u97e5X//b94e7rUcYmyTgNomtp1efUHPV9o8kxcv2lLcKTRtEmG4IxvBPQwi0naig7LAssJIDCZqYl3X1RXLOYuIDl4aeCaKRmIhNHW7Z5YJi3pfZ02I1Y6sv+G47zN9n6/E2LrX5KiYs2w8BsE/lAf9UW2yzuPUZuXu4TSKpbhFDuQ1nRjRyjqK159fNI1GLrsPkkrZ94dKpWeGps7hTNarX2ucJxEFUSWHOJEzvIi6Ccy9wI3IjAgcKRcVRM9uumIaAEn7+Oavb0m4yXR1+bk7P+yT54ASkN0GPzysX3w1rFdhs1pr41q2hoaI96l/2O+GIXeXFwGk0KYRHVJGA1Bz0VBMVy/vrl+trefQNHS5HXKRkLkRHkKzNnc/5N6U97393b/9cP/h8Nu/2Uc8u169ala/DxQot33vu0PZXzxs9m34aeic4zpI6IlYyDU7E7tpEyQ04ZAG5hiFHTA1Z2KBeM3qrzVlc5BGRKKEGGsesYXGZvHhe4HGfawa8DGWj8BmOtm0H4PoGQSdWyg/cXNYDvGMBIyQcXQwJUw4f/xV3MzdQCTMQjyVZgCBOULhCiZqgrvHckmRXv285K/y4XC4//bq8NAPN40efH+bclUkMVn1Erc4T4aZjazamZqGiCiuLkOgwZLuVzDWnAs8rrqge0CbrpNwkV5t/35zyPcEuY3+9va2CS/k/u62TVHWCJ91vb4X5qYPwyYM+/bdf6R9wv79h/3dC7+O+Mnd1TXu/h0Lcx5+d9G/3v67SP/n1bB3hcXYmIpAL8EEyfngWqKs8MwBCA0I6/8fd3/WLUmSpAdisqmqmW93iYiMXCqrqrt6QwOYBgaY4RzizAN5OK/8uXwgH4YHJIcDYA5mgJneu6q6Kisz9rv4ZmaqKiJ8MHO/fj0iIyqrm3ygnjw37w231VVUVOQTkU+gmnZUByWuuwzA8u4trl+Vt//JXvzCh4GeXl+JEMjL5kfOWKXmblt9D0Zy/7r2/5d9iH79e7tn/7ppZhk5c7kmIuoJA7upExpjT+JuHLHuB+4AnxgRMbKP5IEBMLpDMUMljBNzwZEH/nEHosMXPoqBuo38GeNHNIV+JkzxtPgHAHyK/iCiMcOBw/a4H9phW3UAH/ujHsdD+e4IdJqB+VTBYibvGwSnAnrcyk+l8+GVHqVYTynv01ObwYlCHcUfAEIIIYTZbLZYDbnPw30eurK93XQbqUOtHZIj1qb3R7gSJyWiEDilgIh8CRxosw2bTcYetr/OlDB9aU3rSLDLZnj75Cflq38Sf/7n7zy3jLPNXQ+0bZEiprrJcg+XsyUThc83GtvhZti8NYzpiy8Xt79ev7yt2Ma8NovhZtNfxKt57f7Z18/bYG0KDAAjCZ/ASEzupQBqgKmE0qpzBwxYaADEojYAlVf12/8737/QtdaCPqBKM5MZG2+qqInE66a5ZO+1bZMr+F72u/Luf97d3ZTZkp7//qL5ItJ8WEZ0LE4klwGZyVGrkzFxcPPS15H31cFOwcnR7vJDXu9R+5xJ5+k4FVmAc/z81Ftyd8BDUhIBoCEcKjUmiXp0rj2GRY8+++ntjleWsw/OhPVcRj+04AARYEQwT+3Rww+AUzl+yOULHpEYY2gxNIv5ykuneVfB2ArEk8WNiEwVEWOkOAroE1FwzP2YI9E20cQZvIJCVdCooM3MP/uy3a7Ld3/Ta2VpxbWLgs8WF6q1s1t6YkAlPrMaa1ajJjar9JTr7pV3OxlkNWwB0h1bi1U4+Wc/XqlsmA7eMI0EfeiOwpHJkblWg+Klq8jJ2KURRfQ97V71t3+/f/G3GXLjVyP7LdfMWSgsUpMVjKlhUEB3lR1LYJZitHuB+T9KOze/K9d/ult8VeliiQ51qOgUiIABHSYBcazFwUFExy/bnR+MTkDAg695UFlnsviDxlFG3f1A2uHHYOWDo/JDLnh64vjn1ML1wQ1CpEO+Ex5Ynz/oEvkhQxTxKJ1++kx4SEWarmxORHSgKyMgSAih+kLxaq6qtdaxTbgBWnmwNBARigMAC4owolubiltBsrfZgtNMhlqcCHYNmNOgDiuI9cdfw7MnzV/qm/1bve0lcvv8arn632uw5sUvQ9FSNG8rYxhMonBsLjHPcP7ddb/eroeBJXoPP4lRkH/vv3s6/Oz1XBovuao6CzE7srubAnNk5jrssRJWhmIFyYMbYR7g1V/nu/+c7952uwHivKRLXbT90INZNrBZskVsCvpgvbOuZrR64iCKaNiCfUvVcV/81/9b9+Z1//U/XfCznFhS7zgM3GD4gtUA3ZGk1kqVzbGShzjR1R5m9aBBTyb6+8o4z8TloEH9/U9PfjGcSocRXI9aehKtx/JPzMct3syOrJrwGGEYf8qDdB5/Hl/jVIN+6On9pN70+PvpOnhPpx4ONUMO4AaGONagYAUxYQSGCi6Z3OHgFcLYy4IIRk7XQgIEqY1tMi1efateCZsoiZzu9u8cUy8A3sUYvv7q+s7z/mUWomVK+Yt3tk374Z67GbnkjOLuHZmD7nW7Ym1Mc854EyiFSnPIIcTPfnadL765pK/gJFfBHbT62OAJgLRU9oQIwrEHA7Occ7/z21d7exNDnT/9LHnr8+sa+qEbbOhhGAaAXSEYXDMDM4ZZM58lEOjVIRYIkOV1g+z7xc2txRf5SZEAIhjALCgxs4sRCrA7P5JIGulZjsS/+JAIfDzogwjo7zjwUND2nkS+d+SjHICHONHjbXz6F3/xPz06mQwqmQGiI9WxLvkoZxPwedRtj+97rLEc7+TT0x5N6ZNAAuKxsfN0fH1IVnqwQnC0RA05wQmA5eDVyvp+eP2Lftjp7u3Qq+8dat5arf1rqaaYgiKgsMTBML75Vb/+ywx7eK1dHvBuV9ayNba6Xj69ali8Fmvbtr2q/q6UTUVLQdIf/Gm4/CNvLuDJFysiEnDNAaoLGjNaIhNCRDEAd9ixFdF54ZA3233Z89//P+vbX+2HO1j+tFAkWBhGX8zA86rr883dLmDq3u6++HKGi+ht9F1fBwrXFRHzgMO22q1Sl7mhnbXdK1gJ/ujfpM9/fJG+2K8ug2efLWaGwBJCBu1KuQ1E1j4hlSoxzFbczsKT540kgJPek+PgQwrOgSKej1NwIkt4sBc/Bl7+Nlv5yXZKcKLjRqUDk+KyMVf9OATO0598bLAJR07H4yN+6glO3TpEPKsoePQOfl6peihLwEm6iaZqVvcDwPrglsHEN8bMTOQSkQenTq2gGsRLCCQhhl3uAVScjPXZ78cw07zL+X9YvLrb9juG9pmhRmnvNxm0xiIQpX4TnHqJ5Ue/F5crfP5f16dfLUopw3ZgjDJH0KoKjo06hViZHKkGUXLQ2gzSa1P6fV3/Bf7q727+/n8W0Niuum5mYODvyDvohnDXv0qzsHw+W1y0fdT0UwyXVLR/+7f3P/+PkXsgQWsAqV6t+Nn1FXBtjX0edtv1L/+vw+vL8tW/WsT/Q8JVXoIKIIE5gyXgWQ9MFATYhbVpU2pERBDNgeh7asxOJ+4jk/up+X8Y38fzePjlwco8M1UnQ/ZkvFc0p4TogBWcwAU+noP1A8cjr+ts0Mm+MFpHU5wYAfAxLQCY2SPpHytTzNGQnBRHAnwUJgCw7MUKtXzxvKk1dq+H1Rta/9Jnw4Ip7uoaE6bUkhnkITV2+Tw2K37+I2xnNThT9kaTDiIWsc1jdNoBkXi0lNF8bC6BiBgcEetGXvzN/f2LIjYzoow91kYL0I5sZ91Ww1Vo581sKU0LpTdAcUerTk5YiT27C9SoKLmnAQaCElI7W3IHUl/buzcd/g19/l8u56sI4z4OrG5Ego0jIrEYjLSsOPaABEAHOnenD5Pyg4TvH2WceUL+4MycP6GcCbs7s4wAmIORaX54gd9ikT1aK98f0XpkCY3jREDxAf4AAHBwPim+c3dAGIvrmZnZgRxduY693qRk0FKtz+DAzMOY29JXYGDGn/yb5ss8e/J3u5d/8arsy+9dLtqvY7NkLrV0+9nF7PJHl6EBtt5VBRMagqsWdXfKkJbm5A4F0WBwU3Igs0AO2bwWGLb+4t8NL/+c1BO1t9xa/DKm0HZle3FN4XnotQ8XvLoQbvpAJQ6Qf0NryF1R2c++eFpe35MDDjog1TiDeN3MWlxEu8chzfi7Oa1fl+4X9ep/XP8hXG5/YlwpZdO+OiInBUCq5C6EBFinAnNgBD5L2zzOwYNN9dHJfSwnP/jgE/E406CGH9egR+H10c5WPmgsffTo7qe1bABTFszJcznh5NmYKY10d2PPVBq/JtARMEWIxuMpiDAS0YMxuCCpW4dViqqKOxMQWDeoVUSvNRNRJSLgms2bHVLFbTVWa9lyrjUzz9HBM5JHAEhJ1VjBREytCF43s/r7f0J/+M+eSogDog8ujqjklrJ54B0MCBoImxRQN0TM1MAwDPllIkJekCzQCcwlMgOA5lLcKgQtef9t+eYvN+Czoen0i+1q2T77seTcylCBtqmNK2rqaiuhQA7362FfS1OpH3Z5HfO2uTVYNIxO0dnEn3ydnvyehMYd99aR3S4u5ht5asOafvX/elfr5r/64ik41h3aQAhsHUjkcAGdZq/oCTSQOsSRFYMeyY2pj7JytAyn5JFRHT2WQNUTRB0AiQ8datTM8KMhybO44Njl3N3HbZCQzap7ZUEMrIeMlvGxfmio8xNNauAUAPsh1oFVdDcmHztSVS0oQMTmnvd1uKu1FkQnBhHKmAmC9WYVsZLjZJWSCRkoIQepXkGLIyKPextSRavMARgFOJqCAzvbYc9mAISqYADmZs7gSgYASGNcgxyqGVEVLQ4IOFVl+9hVuuhAwOpQuTZXFWaAF808ttYT1AoVYhPAebOt3jchJHKonTFwWOJq0XacqnLcG7eZDdnI1Pbb3vOMhGQ2i41ZQi7BvR9omDnqxosZAMXAOEaGEJwB0QMJjekbx94R8L2T9sG99RPjxIv44RbCmRR9LDAqR7f6ANp+/KkeU9i/7zidPOwHX3nS1u/dhiiCO5AB1FpKT4oV6z33u+H+7Xb3YoxROQvGGEvIAAAuYABANQBGgi36TrBISSWlyGKmGdBdhB24kPUBexbPYRYDLnabkqvb0gIDI7mKG8i+ShxjIVFQ3LObgkFIEhFDi0RihXArTihcbNQ4DESIyd99o29elvbL2eX1XpMDXMCmuft2WG++Rbb0xeXQh19/s+9/gZz6+ZWGmT3/8YIud7MZhxlu+pzWGT/fJEorv8pdLIqwRigcKDL3FgvlhOLdorvcJv8OB3dHSNElAiIoCwIaewBpY5MSjuk1Y9P1s28d8VHCx/uT8lFJeI/R7sNS8MHxSEDHPKFJDn0ivjlK75SwPP38R0LF8BB2/b5BRHAOHjjgxBelFWuB0unm1X7Yls27fb4TREd08ypScOYAQGRMAdEgAoCOaIkpaKmFkGVioUJzBCckMDJl3Q9EhCmYqhXAjJgM0GtFr2bZhQAJKCAQiIiCjUUpAB6iaAVXgFoRUVNFACBEJnD3Qq9+fnvzctdczGSOzjCsq2203yphIAJENqMyxJSEWF2LVy19FyqaglMNc2wWwEtOJHOnEGMVZjfLuL8p+03pe9W9AAGjm1fiSETgUEpBoikFG9FHsnGCkeDzaPu9PyNnzsoPG35ghPr4ZH/6Mg95VXCo3TsCNgeO+iM+/3ENemjneHjjDy2UQ5gV9AN12eO59t47GXRuWgbNg6/vyv0vu26X12/3ml0zVu5EBMABLSXH4kjuXCFkRORqDCCNyUpUod1T7gYXqca11nl0ikDiWbyyaQGrJSxwtiQoXiEJZgfoBysZE5FJcXKjgUgiNQ7ZwWJCorDva+0QAdrAzAjBrJoiGKMD/PLfbb/5t9VrwD+93c0vspfNy6w3QwaTJ0+y7V7e3g5DvXuX3AZCmNUYB4G3gtsq1+iXFq7l6fWiDZel6yFvmhXpssmFYe/dvW220FcIfc9Oi/v55XNf/lNviNiAC2klYIJIRAhLoDnwwo4MtEjHUvRHs/FYg/6A8UiDvi82Zxr0TPrxXIMetvBjOxnAowY9BX38qGAfbnR2p8nvOXhNRnwUfXU9xKVGf+txOrOPgTB3QCN34AAYoGZgc99nAPBw/6Lev9q/+eXN0EmtdRjKiPybQ7EhBGHCoShxRUMyEhdEtAyOwIDtbHD3wSI7Fc15ECvkWTx6uW4YMfbDPjY8eGMxXOyLdXU/SxRUd6UQy5Vcr4UbAkO3UoZBxWcWgliJoADFkTIJ4jyioCP61slpcOvy8OI/3+eaKYinFWwXIZu9flty5llKpRPg2xsC46+u6y1eNpRTczcAdbezn3f7y7v5ky/Ams6eSOhSX8sgQhx4Kz2532vY6qyVuE217hzgKuAf/uzi4p+RFBNnwmioRl7zIIFUaTlrZ7OUaDSfp0lBO5MjdX9I1CDD0Wl1BEPASqdeEaDiSUgc+cAt5pOqOp368yDq6DsfIzj1KMUIB1d7DMgg0eEZp3uJu+NB3D6A/nxyjCXIP+yUYyWBmhmhqZkD1n7Y3+/7TV97yzkfyJQdD/kHqjoGxuIIAJwiDEe1PSFchohEwIJevNY6dLsWY2qo0tigsYSWmVooZWzk1Aqk4CDiVWu1FEMUQWJnJ2JQMxy5quSY7KilmrkDZa25DsU0NMmZcileb9ABpZCXCmYcwREIzbAo1NhDoObiihB2XYc15qK37zIGbbB1EbQkajZgt84haN16D4YilcpSiAIuruLVj2KzckdQdzcb+9CNsQtmOm7uP3Q+Hwb6ROv4/5NxtEHPHJuJc/57z/uEaTKWd47yfV7neWbW+IPFM5oYxd3VKiCXgTe73f2r/cu/vM9bqx3stRyjpkSkboB24K6AcR6Yudbq7hIYHMfuIO5OPFYoYUxoBrkrdbfFEOI1pghWYBhq1zlJixLy2/vNaxHStOgsFUmrfpfLUKRNKbY+y2SBlc3MtUpCCY0TGLgaJIpb2RmguSvS4ooTtuq1trsgKcZw/eWcOZTqr7edZk9d0oxD1oWC93pfky+lfVrTMrvWlpZQqPyVvfm9PZLHGIe+9Hd+4VgybJ5h29a4U+2aWrT5L2z2LzTMJCxmWEHVkBEjJxQWjAlSkpimiPz3TaKPXfQOG/0Bo5yU12j0A/yuFurpsPFeH3atzExERu2rqvC4numHZ9Q/Goca+WOiwEfH+HzHn+42Pq5W2L3rd2+GYaM1myH72Ft0agXmaorkiMxMOJVl0TGfH8CQyCdYzpFG85eRlBCLFcu6CEsRAagC4B6K0jBkL5j7kLOYRMJqTm1xIpYxRj0STRR0c8fqqEECCgGCmZoZqA5WwQmtFRPiOwpVVS9WgpRCZIyKXLBH35mZo4IruHHqRIfc7TMWlwvBtCYjwATG4NpRJnaIZmawkPvCRdAYZJfjDkuNpVatXBoKkojoiFiOroa7IxKLEx/5Fj8mXg+xnA/N4HFT+ofIqB/GUQbev8WJefBoCDE/JBoTnQUVzx/r3CQdnS8/VKn7R8491aDTs4Kq1WHIr1+t3/3FdrhT37Mh3NM9eaOmI484IgIYHaJHiBjCkWqUEJFYgUDBxup6FsQKqgrkgAYEzCHvcC3asiBjjALVi2ZzT5ex/czDAnMz6wvVN7qYh/kyMilLLtltcGTmhTuTkxoWdSzVzAxbAkp1Tze/rNub4cXfGGjIjlXw8qdNRR2kZyYDtyJl3/veuDKxO+Pskj5/5kOq+9vShiSheXGzd5fLz+f6wjwU/twCUQG9pS5xK+vwaii0jqCvw7696J9C29UIbWGoQHrozReDGzBj00pMeCS3wffn7tMa1B5r0N99sz+oz2NI85FgTKSQOpX72WOxmXo7P9z8jLfbDABH9gQAAnpIdCci5ImNajpYBFRddZT1s648dDAlwRAAkcGYdnvbvSl3f5U3NzXnrIHUzbZkUMdAw7SCBYnQsXpQYo4xuHuIJoIAXrKDukhAxOrWY0ZxGRyHABDirGoTul0Hu+QMzZdUAnQRmKq47mfZt029lV01EJ7tt5p4kcJlRPDKnCSpCFEMQD40Q9qJ9RYkVNS3ttZfXd5/k3/1H2/32yEDvYKNmXU/p9v9PUZcfBlmDYXcf57jvdLbhRmRFl97xxexeepJoeR2XYam6mrndTd02QI/cdinqHEZNlKvbiE0iNJ1b6D04cntcvETXP3vNheLC/bYeDIx90yVCbhl4sabi+RtNHmUZwljuyHEMR3zaEEdj1FDGBEqcHYAnsiafOIR9rHLwGTuA00SN8kDAR4/en8YgB/gKB833gcsCAO4AiiijQJ4+mC/yxb/vab3qex/akcwM3WtWetQyzC5RGqubiesltOaEyQkJDYiGuvmiEiEx2ZTLuJjGyBgBkQkREMEBUczDoKO0WowYkJFcMpt03Y3vNvm+xtcv7zTChZCSDFdxvnnKUZwNpqYq9HJ1aupNX3bec5BM1jWanfz25f95l3pimdHc2ibkNV0n2GbKukG9qUNV5J2dV9okIaMtQnuNZHC+sYcYD/0wSLHmUPt6l5BpNmpDkNm7w1UPQBFYBKGzqxf/ahZfMmzRQIjr+hsAC4yqhiSwBLGVPspaeEDE/cgVQ+ZRIhHzvkPT9yDC3+c2ZOLfHyuTwXmBwXx4QMC+il7BQ/j/Tvh6Wt//JEBSsml1v2bYfNqv7/vSim11lwnAT0Gb8dlxoSEyAhNFBFJkUMIMU1s4l2HpZTca8kTdzijWITRuXUCEG8XIbqh29oRofBW/+rf9nff5dtXiC7MWGyHtB5+8nR1WYW8W6EjtMBEQoQDlKI6rGF7WXP17a+prsN3/364eb02w0zRQ9NSJewJhjhPrI4G2xvfQrnroc5vJcVm0ZhWKsU21GO4HQKwzp8p3SHNIJP7DJuVLD+77XcwrK/Klpt2KE8CEcteP7vC+Rfy9b+Zzy/D/CL271RA4qonojZFEjTzkFxmnBohAHZ7bx4ehbmPAjGB0z5t6NO2fjZhB2fLzyg2Ju/77D6PwcoxefJEbZ9J1aMTD+gMflCDnp3/vpzR2PD1RL0/HHyCBqjqx7OfqmnZ6/51373rPU+XUlV1E5Expjq+OxEiOLgygjDGQKnhGIkIWZCZqcXSI1MmdXMbBjQ3d3QGZLBur4ZO0Y3AHGODwL/5efftXxXYLYb5xrggapQaA928zn/+77qnX6cf/cuWWzXz6sAAA3g1r43t/ia8/UX3q//xnWfe+jBPTEz90BELJb24RkPJu5z9Pszbp59dDWq3v9n2N7PUNvsNlgJSMXBeLFOc85CHVaAnP2mcC1z68vrq4lm6eRk277qNUcDgoSlvrOtL3/V/8C8Wn//+vPkTJLd805OTtAJz9orFFZ0AIC4KRgjRGZXg/YpIgJMuhkf05gGeQzjI6HvjvRTPoz7+wESfZTPpg2fzPnkYHLJGTnHW4/gHevHnN/rtgTcz06x5l7VX9MmWnxI9cYIaxgce1SSjM1KgY3E8mFcGRiQO7g6hJ2UvZoiMBmOGHAAQupuBkiuaoVJFC2VIZkqEMWghMiuzhuezZsi2ecdI+uRnw0zI3MENEQaoRXX/2jf/q+7eKEbhJS/RZgGt6rDWEAPPZbYAYNxooY1wAeszKrYcONYg2CuYI1NDwsiEso/QNTSnVpRJFhovKS5MfAnuDr0zduq0Qe8xyWz+ZN484xiFKpmSBBAGr6gGXjWKBBZOGiKJ8ESc+SGD8DhNpxbqceJONOjjUJ89Soo7DYy72ccr4c9WwifGyaJydxlrQKdbjnu32egSIZJRHYEbYgQwUzQzGGuiieDw0KNs6eMGo/y4PmQCStHc1UxBSXvf3G6g0ODDiMAgGbkDGFNwt7E5nYhzdDQiImyZG1Z2QyNF1qDA0lSZ85C17rhmI7DqigY+tbJJYMYVRuPBc/BY0sKeXc7zW24HVor9LISL0rSe7u+3Pn/7rln+nXz904X+/m0ZlPYy/I3cf9d/8x/THmsO3fN/1bZXYcBhMZ95r5cviI1KartbANi2V89Af1WLlO+aAn0FXXzN5NruLWe9N4Mt12qLoBJjU2aDdVo5bxBgsJur3eub+zcFwgLnfbnbC/G8xp99vrj6Lyg+m+EQ3VXC0MTAQlDVKxLPPPQ4y8vLBTOxYQgNgOlZhxd3ALAptdICHLwoc3dDPs7mKI4MJ1pNJ0zQxqyvsSvbcXKzKxOPmIq7Tx2aH/bYYyf6UzE9wD5j5Psgy6eyiiMF+IOu/sA4wR4+ZRE/7BQfWiuHEx8ZBkdb52gYnS3uUX0iwojqHQzgaaFPGAoSIIqIiBsbTES5j1zU4y9WFQJIIE6FUsnYG2AF6jtl9RJiqWyD7V5sN2yL5ynve1vXzbu8ue82WppF4kQNo6hSSOIIxBoRjWYJ1N2KLaKuuUVvvVSkvGxKgJV5zVywgTZb4cLUohCCV3HKYo7QQYZyW27K2jx74pIqzIZ5yUOMMvuqiUuXQF4BASQSRhjVFpIDAwcJ4QF68wn1+wHY0On3/z2Te4jqfRJ1OtWv/zCQX5D5aFC62bmJ4HG84UFA68P93rvxqDKPy+TsgINvbgAPxrKIlEGJyKw88ijBAY2ZmDFGQXEClDB6RS5j9xevY7WX90CCxH4gr/KjwWBm7grgqgWYkCAiG9Trp5H/bOhv869+Ld07kBLhZagV+nletI24yH3Jqm/U8x6Gbd3eau2Zf+yrOSJQ1L3e9f061eUQWoYohrh67q6838hqtl9+cRH48t2321o8IfkdZwZ/IrNI17f5jmuMQ0jRkcoV6YuqlVicZo4+eLN89jUvLzIOZL+UYZcv/iCt/jtpn1rM5MEwEs/T2G0wVDZzitYsuJ1z00QiGrc7AOPRhPPDZI02z1gu6+5H32aCs/Hw2weklJjB9SEP41yMCB1A7aCfDo1mjhD77xp0lWNi0/dcwk7eYFLTp27g6XgQTfiAgJ7BGbXWkXh3FLLj3ccrk4xBTmBBYuAAjCxCRIBkpwebGVlEtNRAL0QoI6vo2HbWzJAYEGqthCIi7sWqexyufor+o9g0s7e/3mz7vtxhYvpJ3NFnpUrCBrqg9S7dfKd9p0aMzCrrnVYWX7ZXQ5b1htpBcyKIzCLDT3jxx0Sble3L/OKVuDEmzfPuJtzRFhM8uUhIfPvOsg55W/tbQlG56JrnDe7JdwbOqW3WEaTNdUXlNvS2efqH6eqfiT7p8a2AFRYEIprJuJ3EGMy9WUhsIaUwEVswIikigj3sje4POT5ojgAujwTUFYhG6cSTqT+MjypCchybeR+chokE6cAx9rtrU/FT4/f9uC3WUbgO7tXHbGE/48l5PE7EFwFwVG9wUHin0NUDmHWAikWEUUTGSk56HyiBw7o4XuE4xm/OvNJYnaM+2tkghGLxi/Kk5bZ43WJr0soCrqqiKQFg3d+0gw1VScmYiBSpCSi+M15n2FQHA8mO7LGR/C7z0z1YzF1oZstSBBjjDLQU6zE2GAHyoJue9wNRplYjS4Hi6UukLVVAJOu08C4NJa8r5xuqW1/+WTP7UgJR2iUkczVg0OBExIDjFti0EhsKacyaBQBHOuRIwMPXcjZZx9AfnuBP8D2qysymi09f9PnsHpcBHiZ0+uRQ7fNJaPyDQyAwOIBNoSiY7DxzV/fqY8c3IgAEohEveOzxTb8ws54E1hCPBUtjEweng6PnjiOdHRiixsi0rzu1WtXUfLzUiGWNz0JEgQyoUCMhirCUXNGdkUDIGElB1UEAQu9xpx0aoFaGCujoVg2IKEhhKtCzjXUlxuwS5p9z+3l7ScgSERF8M5pxowL+5Z93LVwV37MVXWf8cW32gkrfrW93kNcv+/jl80pO4pv+dvb36d3PuYEiF8PNPXmtZT0IR2mEDBdt3d2U228vfRuug1v0Nb4j8Oc3PxrmdS6y6W5zkZtdenJ798Js38yW++4nz/2L/9bmc5Se6qoQoA+BDGMH7lYCznoMTZBIccEiJJwczF1dAYlhzKS2CQkqbgBADnTsVHTYisAdZaxVfNAXR1kGAGSbxHKKm/IhuO0AAFQACBzdyBSJyxRzcnfzg49/dHYe+heDO5zxwh8afY8HyOGg71F4P2SgHUjEP5S0cgZChRCamSmVbHVMsD9+Ke6ec2HmiIKMzFJLiW2IsTGznHvCgABIY19vIvCRdCTGmFLdQTnEmMHdsysgYHXNqu64SG4uQuakCsSoWsApNCIihIucMzgwMVNYVim+6XUzaITMd79c3jVbQ1dNbvOQK1sRx+TCyi/XpRtqLP4MLodlVbTOB4bSpLBK10GGvrsZyp0hoUoM9PTimXtRKcUrGGRA5rTphm3eGc/pfpBUvv5v5k1sGYWBCFBEcCYMKHUMKganAgISEzPx2K6ARm/S38+OGG108kk87aPZ6edK9/GE+gkBGExA/eiSf9hIPb/yR2mTH7x2RPHTaIM7mB1VvR8w2Enz/3Ywp5/AbKcq9vSpEVECSRKOAGJnHXrcfbybmZmSqjLRoWLWEHH8FhARGR1BmFCOIS4GKI+gAEJCQkcAQ3Or7mrMbtUVXaauOM4O4HaSngLunjeVS54FB+YM3Hrgz4OhbV8NMNisbRYLYgGqWKtp1NSGQEpGAVNgo8ZYNLQIvbpx08xWF2RakyJGDKto7m30smg8++16H4GYGm2quEiu8xUtfxyYAwCZAQOioXN1YFeEgghAM6CIFIEYxk4sI1h44Lx4MECP6/+4S398Ls8ElB5PvR2NgUmYJj9rzCz7bSTk5OejTx/J2NgvftTreLCjT5/smA4zverZxc7cID22YXDEaYFOkNAhg+mAEGGIpEuaPU1uWO53eOSfdzezEAIC1mqqxcxo1lRV4/0iNCKkWsiBmBEdGZuWjNXATd0MxLGas4GauzuZOaoiGgEA6C6rKjmZQzE1b9IsQICyz4hoEVNKY6ZprXW/70ovabn0Z06l6qt765EDfvGMZokabq9/xrXW7jva7awllKhgw5t3v5n3LTMnxiiyaOO+7rveun3hAM+/mPXauSMDltwHprhcYk+RljhE6zxUnkF+9jN49k9Xlz9OosGKImA1teqGGohB0Q0Dw+xJGxKnloWRxz4ucOKgn1jk4xR9ryi8H0R8T3DO/jrVc25j1t/oDz3OiHtvvAdHnqnqRxDVB9rQ/M6IwPGiv5WiBSDCNAtlZqeRhoNvM17JEN2Mc65AhhlzJnce9bE7jE7baHUB8glQ+uDmi4GSG6MJoQNmBbXpfLVcHFhFCceW0XHM2/cJAQhcNLpavLamEej7rheptHxGl1dzkx0uMxTFWcIGYRdqyVIohtZbNMBOtZgHSiFVd27bliADFotJi1nHw65S7/BsE2sDxWs2G/yC07K1r/9AZj8NBCTEBQ0Bzc1UsSAJESEEDLOQ2iCRUQ5RFiRAnxJt8YF+9vhtwKlB9w+Y6FN/2s0O3jON3ITvZ68/OnecoAcb9wNHHBWlUDXkyZqGMYBOBE7g6GpAFY59uadHObnO4bWnNzc5uIPgDiSj9h7bGR697umRuEgFuPoqln6zaK7zdtt754GtoNSm1m68LyJVQHNgFt8qVJXg7VVC9M76VQTPXkNLcexJoYjqCCiGPFHSVGF3JTQXQMRU2xLDgOZuGCWWHCxZ9aEHRHTCGCqSEyiCXbTNdsD75u6f//OLEG3xxWeb/+Vu813YrTEsyuI61M58aDavsm5Y44BaPMDqItWfrC03dPuk7uvm7XbVctMKR7NKCXDzpr27v93sVUtatbP0l7UOvd5eNLYMw7eX/6Y8/Wr12b9YxJk3KdbiTUjo5urMLMQAQBIWVzRb0uXTiIhTQ3YC1wInzS3OMI0jN9OEHb4nGacL+3yjHoNDB8Ey1ccnloc/Du0GjhI8Jl4+2BgIAA+R1POMlsf0JzIBQ+cmxZSL6Y8t1vMXerR9PAA9fsj8wA91FpueNXJUUo3PPn/y9q/fcQ9YACtgneg14ST+q26qWNVUlZmUVEQkeOkcETvMrMCRcVSi5GbVvBITophXQwRwBgSHkgZ3JGXt1MxSE2MUJajgAC4QxllEBgC+/kP82//HXYAwy/LkWlK85ycQCm131L9FYyUizTZohYRUqXSiiD3ML389z3tfv/S+L3AN9pNoAXEOzSqkGUrYdGtf2lVY3LwcuvjX86fP4/z3bpFvv/hq9cf/9fPYkqwKsoUQmGsgQufSEQI7AAs2i7C44HYpk99z1CB64nwcuhqcoXjfN84qfz44cb8bVATvWZYfueTZ9nvw4k+Ry0NyycevC3C+fRAewTV3n+JSj8zzw97rAMBECCGwXLSyQN56UBQKxTVrhRN81N0J2QxsREkIINQYATHkrMwMBZwdeNqXD3uGEwEQQYEjVR8iFq4MgkZWAVQ9KRzZgWFiASOAsUXJ4rPWw8b2/OKv1/2adqJlT0huPuR9hU3rTnmoVQ0DtmJVMziBOBTESgTOjCSYu+LFaWahSc1VI11nAFRFMBezz9uvnz+Rr/91mF83cUFplYkBUAkZzSWAEICyEoIzohJRbCi2EtOU+XtUbN9nVv42Avrbmna/hYx+8FJ4AMI//hhnQ+BEGz/gVBNzLH08qPrg7I8njY3hDkT7I4vNI4caJjWLgECAjqFhavDpHy3K7WAF3YI7dN4HmnqTTZtTJUQwAy1ABEVriF6z545CCGmJVJ1LKXvo+2xWAY0YDtQUaIcGAOAggcaeQWxuBoDFnByIhYnIC/nIVBgIiTzwZ/9kfvtX+X/77zd51q2aL5+seHE5rL7AIbPu691byQq0MEz1csnPf/aEoxXo8uCQcXEZWjVLZREWEMyaTET9nuVtmCl0ddivdQYJFt9e/+yLp39KHvs2zSQAmDc08wKDQSPixuA+AsNhBrGh1ZPQLoLIo+l7JBmH5X349dMycVYmfh5Jejzpn8AAzjyq039HPKOIOs0VeX8IjOlLx5okHEuY4SFmcGKxft843dPHZLhT8+Dgv+Opy2Zex6yuCsOTH6/wN/zqN7d3Nz0BNrPkVk/z6kspiGwGzEAEoKX0pQ4UIjZN09XMjaU5QOZSyrhemNEVzRQRGdBHhI4wugAyBuCAAMyYqlt1nfyJqkaj3kYGKPabf/LfNL/RrL98/l3dzPr9EJxm/dVXz0WXw4tvy9aqkbfqWu43s5J2s0v6+g8vgs5yV/YvStmXfjDFfZsipdjv7Zu/vZe/wQZ813Sb++UXcfan/+f85R+Izjymq9ZaTxUrQCfDflCkmsgBXCuBSaDllaTEywvh2LiRWY8HpnZ4rLrcfez4dtrJ/COT+N4W/+GJ/q0E9IPjaEDy4zaCjy915ueIGY2ZGUAOqOAGSMg8dmQ/oF+HVyM9PibAA0/i5D5Xcnet055ORDCWtyAYKE/9aA0QAZw0AhrgSkBX83z/+SasKb1rAnvFXfWI4ObVxzYlNLgXQAdkB1RjRPfiBpRrH60N2VEJHAFbCZ0DlmxABQEBMwAyCEMAA3NDVhQyUHfvIrEHMRQLZLxhA4/BWVzIjdrL/qqf/VlIv1j/6C6+WuXLlBbNwnDPs82w9F62QxcFGVh3r/ehXJInLiVdvVHlzduU1wFI0KuRxiXvh7J9sw0pKAbswkVjS63tF4tM/rxZIWmlblYMXBRBEgU3TtmUNcc2cBCcXUQRQUwE6FinqjLHwzQ/NJ+AQ/LakfkAgSdmhpF6FXwUGh97m+vBuf6QMjqh5UIAQHmE/zhOqB75ASc6XsodDglJI30DwuFPPOHaPsorP/r0AwnL7hPcetab+P1xWng1opvwUBsFB8MP4IGJ/kT547gCABwQ/fM/uri4uPir3bfrt7umLnbQI0IMbO4KDgbHTCUzg7E7OEzhp973WhBspLgmqMUUiEGcAAgxjU3nSxkAiOPUtI6ZAIAHBiQAciQFB6/o6GBqxRxildXFYtmE9b9896u/evF083l+t3+7X/sCeQZtCCtmxTTv3U19me5vczdUuZjHp03tfFP3+6Hubj1trVlyeVPK4LC5vO/3wzYvYem2/exfyOXFIiYSCiIEMGYAgRE7KROCBwTiBLMLWi45xkgESGA2rd4PTs2p9fWJ4X5AqT+mFM+t2zMuhcfWgR8CitPvx/KKcZf+5POchBjeK/mwh4V4iDt88oLHNDk/9mmcYMVxsb5PnPbgKwGgE0Gcu3wWrr5oVWt3q4mrA1R0MCu1HG3ZiWtklFTFsZ2EowEQorIggMqkMCZy1BGWMTtdOeMXhYjIjoYjaxkAACMAOvkIwah46PpO0D/7k4XOLn/1f9vGkBiDDtosYrqw+ZPYoTYNxUY6tmFfSsV+26RnkaPOnxoFB8Whmg6274eavdVQMs6bJdz17dw+/5N5Ows4UWMxwNiwjdTNwQlBMAIDJ5qteL4MInRiONkH+6LDiWh+REZ/S8T60ZT9duNo1B1Nggle+n4N/X3P9n5NEiI6oE3stR4+dPrZkz8oRTxGLdEAKyDClNQNU+rXQ5RtRHQVEEmQiPWyfv1fry6/iH/7P7yoe89aUYh45F956OQJo7g5gpMpIqIjFHVXio0iOlGc3H8oY28+QB8LRca7AriZMQoRcTuBA+oATmopFwCHmJiI2amhYF6f/B4vv7q4Yvzm32836+G7/3R/8Xz1/E8augg5d7vGYGUtc5OaWo2hot+FQHwNsbHLJJtungff3wJ1VvVuie3QvfuT//bzH//x8+s/K7ExN9TBqwGSUkV3g8HdjBtHkhB4dR0WFyIJA4+OkU4Jdd8zu4eB53P0/qEj7PJJfrqzONM58daHK5Yefh7Byk+OI6zujh+sSXJ3REdyAP24gD7wAowPdBb2RTzBgu1hG5hcJZv8RHIA0o4Ld+1zjCnS/ovXf5/Xu+1t3le3wQo5jel5OPlfOAbrAciNvA7uomrGSuSDlbGYLgQBgFrcDBTcJqh5wgem2rEFUkUwEwV02O9St+/HvBMMkPtuNluFyBr61XJ++X+iRb189be78Gp2+6u7m9epD6R+2S2Gbr+btylBKENfob/5C0Nv9vvWu7TyXBu9fbML3ZPo3Ej56bP0/I//4On/0fgyp1mgDoDIKtZcWTzZSJilCNAgNa2nGT/7bMUtqhnZaOcoITJP/be/Z5YP2uv75XicDvdPczCda+KPbvF4oiwR4MEG9bED0UdbeR5uN2lQs3yE093Ggk0EZzAExDFHGmjcvP3gI8EhDcPxBCU+IvPj8zlmAEAKAETqABGsAhhggZrBKyBDYEcHzp0MtWrTtv6k9H/2HViiN3n1LpVsSDZQNmU8ZGYZG4Aj2JibDO7ESuKm6EYc2dH1sAB8VMLTa6OpIaIARiACAmBuEA0tRwCam9oOa2e0tBhDSEJcAC04QvbA7fKf7Ompw5/3i7sav61v77lJWjgPVuO7+b7ZevSyMHrLtfZ1N6DLVtsCNocmDVkw//SPnzb/ZYHV/Wwxa1IaslEFRGzSKEm8z/tIIXGwapj4+qsowijFK/JIuwDIksY3okCj8zFaLzpx3QBPrs/RGZi8pNPpQ2QHgKkFHZg/ivc8EjU4uCtH9VZH7WOIjkRQx338wD7rUzbPJKZHaR4vPrYAH7kg3jd9J49vClNi/fX/+xStHJlwHh5OAQ4djwyc4JHp/eDhH0Nkx7o5olFNuQAQKljWNSIyizv2fY9DU7L1W1MtbhX2HSAPdSD2xUUM7vu7/PLvb7t93veDbWfVbaha3AHI8nC6oN0dEVhGlAwpEBExIxIAGEJ0dzMwBXdUNQAQ4RgFydurNjUBAMpAALS5rbsNQOD2WcLo8yAihAij6bKjWvtCzg1KrfbtL4df/vtd98r4LoizhowOZk6AQVoA6MuuaAmBr7/ypgnPfj+1n9Hsn/NSnyGVEHtE90oQ4vFdzCwMDgApYprR9U+W8xWLSIxTJ23Cx0bn+M0fpmwSUDVUd/czxPGcK/58w/1A7eXJ5NLxNABwm8iwED9gyD5o7g8q7/NqfZte5ENmwFQ0dxZmeBBBf7COjxvBOfZ+HAQ4etmjzQkDEACCuRetkLmq7/pSi+/3Q7nflt6HjVtVctC+Q8TiVQLqNTQXVqoVN2OU2DgCABVyd/NDK6XHfgC6oSO4myu6KxGDAyKNpEVm4yTS2AJqtGGIKBDGgABASO7QzMV8LL93CAZECo4wLnpIHt2K4TBQX6le/GT++3bVv9H8qpLTJpuurezKcFPvh7VEjgsKwZ98Pv/Zv1pS4MVngEvv51sYeoexFBoDSz4VoykBXXmG6Sq2FxwkMDMQ4dmuejJbcNgTHzySD8nHx10i948e8Ni5QfxgT5vfZTzy1d5zoeSDOMXRSmWg9yXxaPmeYw2mI8w0sdr4Fgy1w9rp5n63/oV3XdluBlMwIygbVfAqiBgI99qr43w5iwm/e3c7u1oA+a4ngYYrFtn7eNooovyAPwMAAo+pkG4IwK4AThUnehKiqZMJEU7NFtBD5JQCEbXsTTQU5iiOLLO+aaIZFXJwxILu5qCjBpVutpQFCVUf+pLnTqs/svzH1VsFsdg/0w2UbR7eFW0wBE5zYvHZMm7aLXkljonScnhS/BYRo7SISA6n6bBEHOYRgq5+lOYXIc5FlI/QDOJ7HB4n6SBwxEWOSuSjkoHn4nuodPug6J3LLuJJB+EfgAa8f2E60c3v3VqON8bHT3bqSY129Jg/8j2PCwCgmpkZwB00l/ziNe7v+vUvurzRcqNdXNdazQCBzYwQHbhCNahqee5cVPN24AwSNK7VoxHmJsDQZ4+ClUV97I2oh/LRKfjhUxPTEX/2MeHMjhQaY6INjcX+KQViHDt7E1EQDQKUYHY5E5Evfnp5/2va3uTbm505eDBzBRh7hHqZb9UAkcClAKtJSyEIaqlasI0ZLk0XDs9xoIporTSC7FWf3l+SGCx68+0Qdmm3dMTqAOQQa9JoZiPXqQJskxMC5Zp3LLDkFnGkfRt9xPfdmRO392xezkTt7M9zqfqko/29GP4PgaveH8e64pHg7rELJWd5U3YIV06L0sXcXR2RBBEm9ubRaDBFF3ZwA4y6q9z6zraBl91Wb341bP9uv93u87qQC7k6VBEehlpKRmRgAKgMQAZkkkdj3SspoMh9MTYMselzhRZi7iuAlTFKxmxUtdQyAAAiqjcOhmA8dowxQCSEKf1ZBQlAnGQkARAjAgpGoTJzN2sSJWFq54Xbom3kP+rpTc/rzHvOQfKmAMP88+isBImRa9FaQUJC7h0Ajcgd3ayPRMDoVXMDCdHZCBAd1Zcj72iL3hB4DmMbOGBBdOqouqAFqQPmt+KVgasplQWmz9yTR+bWAD2jONTmiJwAAAqAO+FkYR9intNeT0zu/uC8jsJNOLm39VEIFJlA9USIR2jv4F/hQU4mUSx4qGwAeGBuGgUW2A4wy7RrTaANGoB75aPexZHf9DTYfrbFn0nzGYIFfkqcd75KRFqArlpBhhyt1fb+29v713f72+H+u163tR8GMwMWc9NiiGBmgGY2UQHCMTgE4KA0wfDVc2VB0MDizBSkFQJCG/pqpn2nZkY0dYBANUJHQK+qh7jr8RUYCR2IcMwREEQiaDjEEJDximpsGook86s0b5INskocLu53u/262j5znKNDX00LBlQwRcMWhYhUQh4yAKSUiBiJSxmQoGkah5HU1wAxcERQIkf2sY8ZQTCzUgatEAL1hrnqfgOlDze/7Lebkhb0Y7lcrEi7SiHXEFPbADLYSZrwKFhjQp3hYWM/MHsdjoH3E+fcR9hznGk8xHh+sIl5isbrgcEex3ziszq4A9gKPgZQxrMOpXMfu9OHuJmO6todaSoCOQE6H4bmzJHcQu3g5qarr7avf323e5dzp6Ccs1WHAuaWKxoVOmCf7mAI4fhdHLLyaGTXqcBEqIrZjRiY3RoTkTRnTmBm0pAp5Fy1ortHqONXXSv4SfRvyhAwRAQJwAmIsKFAQtIKRAaGrpnXpwTzCp9thzavq2uHxjDQbQ/7fV5U762UssvuvooX0REBWjFhpkgThYSZmRGjiNAYDhiDh+hjDwMEVBpL9T1X02Gc2ugVu7v+flv2Xbm9ddZkG26vS1o0wDmr73a0uZ3VFpoGAjJVYrSjxQkAYBO7rDuOU3/49EGb0CFl4ggv+TEODpNzBDDVwX3vZn0294+ledLcDyH4xwLFDEdiXHDws5bvP0RAzyEG18ca9LGJQwZGtcPc6e5lXn+z3bzp6x6KGYhXVkcY27DBSOY78cyjmZ7fbiIRZ0R0RzACdKuIjg5eiwFUZmFxmnj8xkppc3cm99HQcACH0T7BAwUzAQIBBZLIREjEKOgBi4xAXRwKQMfDC2aGNwMN74rt7O6XtdvW9V1ZfdGE2ID0BLA1a4jIDB3EPCiPOqGqAgADMQsiqlab+Fyn+R67m5tiqVArZS8ABMq1+t093d1aHrDbYyBoxKWpcVaNihnVgtqlKmyABhXM6DFy5O4ANErn8ft0N3cgevC+joKLB+kcfYqHqPeRK/T7LNH3zNlT1/tE2n6LcNFp/sanDv4edrvjmVN5oE/W+Vk0gW2z3r/9+WZ/V178Yk1K2gUHy9b3uHdiQgIlUBRnJXVzHDsTk7ueP5m7g5MDuXs1RfLoNHZMKFatmlNPUgAgzpI7SAq1oJnZHlW1Vhi75DhPNsPYZC0gI2OYUZgxMbgxMlgDRRTAb1589/I/8e6+/eY/4/629u+2m71Xl56COebhNny9mT/Dn/7s6vpytfiyW4AwetaK5jOIpxmGc5dSCioA2DbviSAEJiJAQwiIPBQr2fvsg2xVcbcO/Z7evVLqGTiFJEBqcNsuF+1yTpSGwTY3dX33ankdsZ0/fcJNEFB+JGoP2djjvnG+tx4Nuw84Uu4PNuipVp4E4OOSA6cCPaUgAUxRg8c2hdV6aoOih+le9ukSUDkrJ3UlAEM6qCPEU6U6MDQoUCswqQ46L/e/HPpXsn3X1b5TJGNTVUBPQwMAZlbVANEAfMxcrKMhGD3hqEvJDbUqtw/fIIDWyY409ZwNnR0dFCQERFBPSE6ks4Uz84aLKISBtUN3t9oiGUenYIhGTWxDRPGaFJsgLjnhPdXhLcLg3/719Tf/adtv7e0t7gdPqlveFXXWWcBY25e7G843Ev58c4dD+Jf58z+6ai5h9ePeub8xJsOo8mRYicpmtTcfXB08oCY0r2DEhqhtnjmG+/52oP0eUrct+3ezt38LMW7ihe8XV+BZspI2l7PLplWS4mJMlLPC24oFNm9WF/MkwXykV3BHczA/uLxTugvmMbbHSAgACGO1y0guBLXiSGUBapPtgyO57AQRP0gnIrBPpYU2GQZHmQYApMeq7XHIFSFMmBE6oOG0fg4wFlQ81PoCGChO9vAUEJjgSyRysx/GD5rQzbqRPHFbtP6Fvv7Lm5tvu5wzoE6V7MeoxqHR0Zh+JGFKZJ42vvF/7ATsIwk6PJD1jw4NgE01isaAh7xiAPBKDESOTIh09XQGAJqx9FVVydGdgJBEgKCZIZHxjIZEFes6525tL34Vfv1t3A1+881b7NlBN36z023F1q1B8kBrtcHsclEcAF5Cb9C1/+Hqf/23r5B0+UQkwNfXlz/+r67aK9pfFCJb3u1ZDBFNwXjMoeJx/+1nt7kj3UqE5Yu/2W/+o+WsdznH6/zkT9s4ZEIb7vfdbt1eYv8HDSqJBS8oIay/027r8+v95TOEFc0Pfs8hUPhofBzuGYXg4ED+A4AheK988h90sfcuflILcE7C+4mhFcWqwdDVF7+68b/k+zfdULNqRSrqcCqgE7QxBX5GHGGkpxp/GfukICI4Cp0grzBG0txVxx3D0Q0c7BiLr8aCxI5gxqg8hBAWlw2FGaIzsBOyxJAaIsKQFWGv9PZl9+52881/uN6/3m3u6h1BZwatvfh2vb/LoaYkT4Rv5hBB83JeFrPU956GoaKtW9173Wz3wkbAlme10l/+ov/1i9+Y5q7bufvXvmjbyMxEAXMGH2m40QxvfvqNaRMv580lNlcz/ZmnVFc91j7VmxQ/I4nkEWHui6/m1GRjhMLssa979VYHffdtfvdVMpN0VYmIRY6a73SMWvChAdo50IlIOHZ0O37Vp58+MiI/KnPvnftRsfm4jL336aSk3QHeZ4n++I0MwdEGLdu6fdnTWko1HDMU7aG0ACYHC6egIiIijBmZB3mFMPJ9CgGPvbAfwqfu7obugGWE3EkmI4rGYikfaVtsZAXRkAIRg6EgOFGzEBGJsza1DRFpkKHC7Qu7faNvvwtvf237rWNqYqvq9e5uuH27TtQuWDyX5oIuo0L1JrWpjSFYuGwN3LX3febFT7pup8X2GwUgX/V3t+sAcjX/IoXmm92a1UDVDYleA8C4r6kr/LUBUHoa5DI/Ibr+qmlnWLa5v5WZhXShMTEnxOS8hDCLAUQHqLX0ecg8oMNui/dvRCQ9uUgA5xbjwyA6Yorvf4iIB5TnQQ5OP34sNP9oWvETSvD7H8PfJ274+EBqu+5+83bXvS27n2+wxr31jOhoUzu2k2/O1AAdUcf8FJGIh7QpRGwSIY8srOTkDcqpBq0V3FAjuLsp1dy7u1Z3s5HrDwDMFBBVCW4Q2TS7X0YRKYsSEs6vqV0Yke9ju70tP//V7d/8T/jmN0zLt/sl2dPm/ubt5u5m+Nuvrhsm3V43LxcJX84lXOxiCK9ex/W7/vr58u/vb7XaZ7S0oVn2v7m+wBilPkWm9G7b3HbvkMtt87ICXPC16tgn3Ld0DQBICuDORdpOC9om1bW//eXu/vLFl3/Q/vhfPe2TaL1bMgYimIXCNE/77YDzOIOdbe/3oVlcfhW5EtzV9cuBFevXTETCPLK1nAviQQt+UEYftvh/sOy9d/F/pMD8qZIaNShSPPWwTEdioxFUI1Mj5loHEXbQrW9h62/+4n5/V3MfersnAOvc3asZ8KNEfwlyEkgAREJ0FmNBZpAmJqIkaOSQYEwfHgN6tVYYe2tXNqM81BhYVbVyLaCqOvVrIEU3dCVgQN8DIUqA4u6dL+bczaKjv3lp61+2v/7vb7abEi6hBJ7R/k7nw5+X1VsucYeI7WIFwWgV/R283fdmtavGobn57n5VAIAyD8a4bprBYigwX1TgmrY310/IhPalOnAHGyRn9Ejl2UoJQ9cFU1Rl7FBCCcvbatD3srrHu7/snn3leNG3V8s97ZLQZ6lNJLtBYUO9Z+gxyGL5FNJyIODMteSaB33deRs8RQdEI6bH/WE9ZyO0QGMfiYeNnNAQ2A+oNiAcshTgqLPODIZi/njff9gbEf0QZzl8+vHqoPN05qMwvn8oAjMxuLs5AZ9r0KOGO/9HRDdnr3e3en9by75yyDAcn/x7F9ARCzwmEBEhM4SAEpAFmAkDCh16nTtxBYKo6rWAVjAzZFHlOmalImoZwTzW6ojuqGaIZKUQAPjWu4r7uzybJYyyXZfbF93NL9b56rr3Mr92DLa+2+Xbbm4Juao6YOBGlN0txf3goLrIMKvDa2AnETEic2xbRgRT63oXIYuaS/QKoanFhtQyoQtSDHixyBJkmMmQsctS131KSUIccs1aW8yQgV4pdmTLQRJJIotmDGhxJJdMxCmmlCw1BgYeGR3VYNgYJqrJJRqgnVWVHa3OD0OMZ1GoT42jKQCH9pPjv37yxH/E8TGgfhx+CBL0fV/uhre/3HX3rlZBehQ2s2mhIeE5a9TYcvjgspMTYYgcAodI80UMDaYGiEioUaoTrmQmxl4Cq4cGzUxat8y12tBXQ4BqWNHdtU7+GLIruzsSagnelNaz377sakxhFr/7db/9BnWX+pW/6jc/iriYtVfflnUtV9erxZfh5TfvYoJ0GSn6M9yG72YlN4OXXH2x4GHbT96ew76/RQhJovVNV91Du90GdYuroag9WbijARiJ7V9EDrW93iyvI8+KrT/vuu7m3ZYcVo2s08p3ev+Xkb3MX2zpz5b8tMnP6gBDpCuqFJuYZqKlDJ2ajpsFEMmg8PYXu3YmC6H4LPChPvNktgAfVAK8v/OeapOzuT5VkHAekjmcdehANHEmwFFeP4UefPTjj3z4YRv0TCOOf+ach1vot72Qm2OupG7Hjpqjajx9+ZGQd+zUMW4LiMSMzCgizEyBuAECjpDK1PV0CtSaTu6/OwFyBQRCNa/mQCLFzNB8StR1RzdU9VrVHYI4Gudc+y5X8GGf90XS1SLN66wUM9p1CFt59mxx9XnAxmcNSyIikyDtFXa3jQMKF4cBsUgAQgRhJEBMUNxMazd4dWsZSYWQPSLQ0GdHd3J24H4u5kkLk7UNV8cMDqSCEhg2joq498p9aV/Cbdj7V/Tj5zMXzbsuSUpRjLyY4q7WHokhiAM5IlhvSlYLgAuAngvoRyf7487KWXBo/Kfzj8wOZu57tu9HxtmnH3HO4HxNfbiR1yHE4GPY0Nzcbb/fD6+W/f0G0AFFy4LiXt0mv5Bo7NIOD1+EHXf4kfx8zNiQwCEwcSI2jCgexBKESX2O91WgqdmwqwJEDFQNkIFYVQmqKhJBKeqOOYO6qVV3ZeYKnpqoRep+yyZ9l98Onj+fffmj8FVcbSr+8ufr3X/e/fj3BnnW7fr2ybPZvF3scDdbtd3Gf2Ww93wVN/PQrTUtlklVd8OumqfL5QqxX++XKc1Xq1f+enm5ZubtJpAkLxd9HdSrdvV6RWkWOIVhcL6b9/EGW3ryxYVVL932D3Pazio+ueu2+eWdhP9w3f385snPfPFFCKE0lNS7oajjvA6hahMTxEUxt0A8l8AMCmguDIJUzmZw+san+NC5qBxDR7/VLn8SYTo2HT50cjscMInyJ670Ow9BOBAhTej92HD0AAMr5JBLKZTD7d/1+VWpOQy1OCjYPRaIAEpk4GqVEQ6RCXB3djZEBSRHrpBaYCEKgGzSUGyKRGIVR+xIQwAAFJWJVJ7roRBLYmzyvqSEMRbiqgrsXCuSg+jUYdyd1MWNKyAUF7S6sdAYF4qElxJebe/wyTzP+nf/qe5+vst9s9nE4ZfVre7VO72VK73dD+hmUMFwc0cckm1gv9RcCCDOW4CbLlDjkYYwBByWlzhYkwuAgemwWFY0GrLC4BWHilW5VLf17S420bUkBkKneajpDQ+f5+11bG8h5Q28Dl168e9p/s9t9bUxDYGCZAKoaZ6Y9kQUk8QYkCDHsULEFLOCtTalxk6ZQejoTjqBoPrYxzA2AEB76Al4lFcAIHnoqPGg5I6gFJw0HXYHGkkZDOxA7IUnTA1A7oBIgAREoI9W0bki90OZ0MFZA3cY26p/kLjh7FquhsY1234zaJHj+lOYgmAOADjFNg7+0IeSBpARkYARDzR0wIeXP9R8OruZ63txDiZAJBBJkWzkqyse0ap7dTR3s4m7xsF03O5rzhVoQBAA2u2HxW5mxG+/e7vb5ASYWkLptnuNetFtiw2a2nZ1dbdoDSuCtdpLtZ4cCYW5RHEdtLop+FAdSm4uLArbkFlRXKQiFCeA2aK7nFcAa5oZedR7H8xTSAZQ3DSSyaJk6HLXkCVMPYI7br65oWWaf7HyEQMRQURhEg5EKEISmAgiJfJqxbwACz3aad3PNuUz6mp72MAfGWPwSTPxQ5JxsK8+3Jf10QX/AX7VJwQUyRiwdN69G4a3NgydgtfJTj7YAwBIRExeT0kiD2l/Bwt7qKgOTu6mhbVEAvbIyMQUZPQExjxRt4dvefw5ti5zNA5ITjhDFgAoZkgq6lUrmro7ukJQ1GJWsHRZVT3PRGJRvHkx6G1+9/OsN/Uy5aefX0AD3beZ7xMr3dythwapPl/SXUg9zUKptu9Tp+rubcRZA33vFaAyV1CpuNmknDNpyzspBXHQZpmWLaU25G7WdV0Y2kUTL9r1y73HwFrNHE1oNyz6rPvhbSlYcptJQKB/0V1Ik/51qzSAWts0RIRQkRxw5A4wJNQibpZ7tdoGCcfskGO54iNQ5UxV0ahMRpl56IZ2or2+f3wo3nOcoPOU0yOGcFCojz98hDp93Dr4FFBvCuTb2/39i10ozYB7QFQ3dyfhAHzsckRjA6lRKA/j9GVyMVV0d1CqVOtMKIwHGYCxwGiqurOrqQY/JjK7Tyz2CsLi7mPXBDOrlaBarFTJNLMaApAWdcMQSYuZGUORkNpFzLu6W3ehW0XLqcG7+y2rQ0Hvt2EGX/x0PiA2DZmtCBu4qPNWwn0vnWhpInVCpXILRAXyIvBVSi9176jUBHMpfXn2VTN/sgVX2FxtwEFmNzfDNtbPv2xWy1K7HWdOGLs7z8Ul1usnCDXVu3httsd9CU1/T/nFOv144VXHjmdIQDza7s6CiBbRWWDWhCCGnkci07Ov+sHKPBPQ6e+DOB699d9Cw53ty0eo4MPjdAv1M1aH92/3MRH95Bav7l6HXIeCEBDR6RCoIKIxU/Vw8JR+CQCHPlqPL+UGB92rcCLKaq7gjs7oU77I6CyNAmpmQA9+2+i6KeuU7kk22jmHC4KZEaAZsLs7oRvixBdsZgKJBWazgNgTkhAQVEBNc0FETjflLhWrNecUrerOcQ5gpRTygcOlAZEMDdCcJbrGJMhUBTwAtB3P7r1UpMvZcq9FNpuclU3a1VJ2rlSRFOouZ7fEFhM7sjIl8N60oHc5l11hvrAH+N2YI9HUCZIIhEgChjDm0OnHO1f9f3Gc4Kwfk9Qps+LRAR9tPHs+xE7MYQDwqVkTwCjYHnpcD3el3mihzmqXqzoCEUFvJQxEJI5EFIUUUIKP1A9aAymSmlZ2x75UdneG0jiQMhoNxcUZiJADy1gZruQTqZMTO5AauqtrteoOCDy2Qc4+sGBsuBRxctJqxKQKpgzmFkjQCOpQY4zdTItsA+6+K2HxFpthjwzt5xXFd7dl/wYXiZ1XrYQ27TJf3sMQ4nz4jcbV8k3thn0Ux9gjhhlc1lmVdogUap7tV6HFO0rIixXaikJabreNeU1cZbcopbZtkEC4z+vclO0TyZqavrmOCLFZhS2uc9010eOwkm59MUv7be++As6IrIYIHDghOzEyiVdzhspZTbsyy0oSTop7Piglj43QAIfUegSY2lLC6IsAjCYqAvjRxMIDQPiBmKrBo84fY8O3g4IGqw9HOuB5zic/Yvy0enik8Zkn/HZEjx406OEJHr2SutVqmtVsyhs/PR7H+itCImJmYgsBidENKgFkdgeFKcVp5Ed1Ba1eq9VaVRHMCRxxavN1NA8Q0RHIp6r2o+7EhzEy1PF4ZVM8WjbHr3I85dCuk7p7tbelKEhACYRSguLqacuqgLBdFxCXhbKbq6p6N2SJruDsRpDIKJHPHR15H2sOtmhFjdB8h12u3WceQlMVFNnr1gAgRJAAQAOsyYuWRKEJc9d57CsMfedBI/hAHoSCqqOxbccnVzM0wFH3H6qsEBxLBmKqxcca7H+8jI7zMYk+fCoX6XcdH/ak4SEoME7fMRqLh+Xy6EgF1V0dbov3njUDhHGjBjCkOpbVs4CIpEY81XYmIqTqfaf19uDlIRJRBSNAq+6OtXNcEjoLUwzcRHGnYxsaVWUDADA0IlewY5svnCBXQgZkF3F3F1FTIHpkVLi7mZtZ5KQNzJbR31j/Grtagvt6Y6uL9nJ50f5sbyXnjes2aNbtb0rTogdrrsNQhqtm3qsH4uXyOqV0X1+0O1T1oRENtn0dMqAi7gsMGXBTnv2TS4q8ve1RLDCIE7OnNjXmmbp+rjE2l+Tg5e4u864NmCR2d3dbZRgsCzXrX++elIY5xCgI4K5IAOhqBmMTdAtkXgu6IQJ9vOfLD/bNH5/7yHn6REbSp0gVHx3sR01kZuc59SNOCcCIMNGfPkjnwXY+5iUR1Gx1p2ikXsnDFOWCkchuREAgBI6JcMGzhcQYSqnAkLeoOnaHdzNTNAJkZAbEyjW7ZrVSlV0DUljYIaxKIkyoqg4KhkkiQD2apDBVgYGBHvoZHzYuRABUtZFhHKfkKTLsU1sWKQFxbpyJX75602/l+Wezxc8Uam9Fy2vyHocOvvrZUiWnTO9u7sPb5ey6YuP02aaE7WUn377c7beFbzQ45kK3eVfRIwoUvp37hXk0q7s8WG1SaDH0fV6v6epq2cAW95u6pncziXmGZqhr4AJhXmdOM+m2/QobHDzGIBLYydXGKi4iR8BxLxnLN9Syee/An7BBPx7C+egYg++/tYCfSdn3kKBMT3GIF47Fhu8XnY7jKKDHPz6ciWXu7mhjFfQRtoSjMTA6J0TE4iIsQgBjfsWYenIQI3JDGOnw0EYMF48g3Pi4fkgXB56YXqa9/EBbcDS5HB52gdGvOrqZB3l9MMtUNZdu6LLtNz5rnXjoykb7eVNmnQOwVrSavKrWTiJCcGTVkMlD0xDM65D6LD7TpkTMgisMSamyI5VAuExzj8G4VujJ3KynEJsmzYVMqeu5q4KKde1DD32l0GsMlGthN44LDAYBB+uzZwAKkQnZi5opHpK8cCocQncDV3dyUHMFbD8hOb/rOMVhPgAknY8fIKBnd3n/n07/EvDg4GCOOPZJmhzncatVUdg0incZe/JYMTOCjz3buAEwFOJIoZE0l2aR5k8iJfEKtfaAazMVETdzFrQKbigA7JlyqamqawUSLpWYCiKy8MG+BCGBCpoVBjz16M2sWrHqtWitWdVUUVULVEUCCEaFCK1Wj2JWueujpQrR3m3upUlWjUpMS3Le3b3jm2an2K1x293PIn/9bLFudiSo3+DFzYXzehcBkfxloAJ9CvIdzUtzs88URcXb7iI/hTAnfjdAFNtmbXh1fVFtF2rJgfFLX2xgf5NLp3ndmul8+a6dX3MQx2tET3MU3ARMTy+e5m3Pu0Xeb5qL1pzmuAIazAiRwFDVRHguLhLnFGggYYD0aHIP3Y6/R/G950sZwlSuABAPUZvxCOYxu/mQRnqWjHeIgk4XfAxonQkdPvbbD5R7Y2obueuRIBH8nKj0EzCTqtZaj7rtIwMRSZmAmVEBVB7w0VGu4AR4BzgvDsGTIP5YL4uHEqbqp3QXgIimoNVV3QxU1Y1UrRQ1h8m5cjjwA+CYKxCIxlJlcfHKDoGouNeSbegt77lmrsDadsNaNUN5rb4BWGbbRSphuAt5r8vPpS9EGEB81+0XbULkKEBRLapg0GIUYL7CFpPubIAa5iCkm7VqAmtQK1qSNpljkTowI8nQzEmQhBpU6Aeve7AFK7qhCdPYCEFV3c2dKQgyU2QUcj5sMr/rGCnqCA8a72AqwTGzCaZ0u9/5Fv/A8WkBHYbBvqc37anjAgDsJEgghuDOdUw0KaWYOTProcfU8VKn7vbpNQHAWQEYA3odc9imw0axVgVVKNlrca009DVny4OZAyILOzk5oxu5YXWrpqU4VGtDFKyGToEvriWI93vZvbWaoxev4neZ7TezunN4o1icFpeNq+4L1j4RbrZ7mhOzRdo/eT7XG1hz4WzDvnZ5Oxsu7Q3GlljQesU+VCZz86HIdi9sIaAi4YbebbYhRSSptTSL8OzLRR1g98Id8X6d13+vSZCfgZMaMbiYuWolBiTtnRhsHoFmDcSAdlyLH5q/j+YNjZpspNMgfHwwnvvLH5eT9wXjhx18WAn+Hp71CQEdjb8HI+88xwof/ZyVSuYDWCbYH/E5APCjLXxA1ymEMIabx0uNaNEIRdVaUdzdFRwFAobaw2hyjOo8D16GOnS572qtOgymFRDEFNwdzJAdiVjd0XrUYqTqOmTQ2D7RxWdw+YVYL/tNvr+17qYESfMVprntbuVitQ8z3s84DxCvs3Dl6mBiKlbLu/U+hERDMAcWz8hhpzNvzZu+uYOuASXorK/Ve8i7UO4ybwekhtDu8mCVnsYLwHszQ8xE3s5EwXOt+z7nakCw/bv9ZxdXs89IWC2jMjg4EcWARCTOZBpJEzMBIB27CTggwmNI/P3wzwfkwycZfVCfpymeJxb9x0XlTDIe/flR8T698g8W0LOT31f1pwcsrudxhnlsuF58tKyZoaKragh8qnPppPHr8Ss4rgetFQy1mldg51Mb4KBBVavXqqpqOq4QHnl1zMwfmjSPaMBhpalyAoo9JQe/kiK+DYjMCCxKwXVfY9vPFxEXwgNKuvfsZI1ZkzOGMLNSWObzJlkemksLnUkF8WjEs2dK9xDI26S8bAb37j6bF9/1Q8uxIV4uSGcawrNlLWrqNURo5xRCUAER01ArVu/cso2+ux2M70PBAQRmYpCxE7mfkKwjwof4sz42Tg6nH6Yi/1HHR0VfiPVhxbiP2VMjaQMBBohezY2zGYTGazYb6zPHtDqv1UoBd726uto+t8smDt/aflfv11aLAaBqJcYYZUTLCZnIOWQJbQgBIQC4sIO55kLGAKAZoJIZkAcHy1UNsoJntZLBlL0fLFvNqtVVUQlN0V3HHEHMUhSwEXNw4Iirmx3e3hYyFpLXb+++umjh5WyTrQ86S93yOamtYR4GMIrzV3dhVjVCXzuknkVi6dv1fshlv6T5lz8yCXfIyczrHmZzNKwdm3mCNXKbCtqLm2rzPWmse8GacEZPVxiboATd5tZvZBBtEnFITskY7nKHMfgKBXiRU/O6f/3Xbxf/+seq2yQrLJWIMJAzYiCXXhpO87lWYiQIB9auSVmeQTZjl+xDO4DTpDgAOBQf+4GaAacU0lF3TiD9eKyZH1nZ3B0ZH0nWQ/8sBwAcbcKTQNMDbTfRgR3xsLGffEQhnDeTBXycyfJ4JRWoFJgZxWAMCz16fXdV04rusFnv6wuCBK9/se7utd+yDVnVzao7Eo/8QcBCIUIIIAkoAgqSEEucCrj8QOrpjgZu5mquVosCEKjXoWr1XEs106k/PBxDHuPXV6qZOxSUxGDglYa97nvv2QSKguTcvnpVICSfM5TSbVoIC1pWagz7rWiLW71/3Zc9Pf+qqTO4W+8d6eJq+frtuya7iFQkJ1o20poju5OCW/4G65wNaL3fpVVL6hJzuKD5c18FYoE8IDpnwtA2MVC+g763WmJZFa/Vii0WjQTd/DLKq0yd4SKgDl4SMBEiOzIgUwMOg2oLYqRkeDCkxo3od9eE7uojyfKHqJYfRPN7Dd7zEx7kyk427vcNxQMzj5uh+1kfOnnkuL2fUkXOjMSAGd6v3GNBMzdzVdttB/v7jEmHl33tQCoNPkV73R1AERkJecrNGckQ3ckMyBCECBHHzZqIHBQB3NzdQc0MyM0n510NwREVxhyTEYR6eHK1sXAeDcgRatV+sKGMDRUcAEyx3zs1g7M3iqW4oGlBIFLVEAMCgQ3k5O6UBJNK8SCWQggAbWh37sXBHGtVdGNBcNyBlaFUNzOMvakqN+pMap75golhVmO09GyYy6WA59udZspDZMWhr2VbUxuauWBsoPRkhMKulQ8BNlf1bAoCpnVw9+gHFYgPIeJPS87vPI4W7Q+zR/0HGh6Ph5x66O5+FjpAhtgERgBXqOVsB3FXVSMiVdus+/U630YfhoqIaBvHNJKAMgMxmAIiiFAIFCPFGYfESOSIDlRrZeY6Ti8ymMNIDzWmiCrUqqWvVs3UkIgEHEDd3LBWG2uSVB2m3QarAQMiyd2g9/uSBxKkhhiJQsWMdvG88UWur+zzr6qkdLuGfg1IstMhhhKubRlm8anWhUZBv1OhKgbcG4Gi5BAw15AL+uCoYKp8FVCJBlumxawxc5gvErOUG9pvQQjayDGFtEKbUx0Kt76iiCv2qlJpKI1IUwwEPK+3uh3siSQMwoSAgijuDCidA3hNA141Qug0GvFjdPC8kfAPkotDOwSfzNnHicgna+DTVzUzPCYEuwPQkSTMzc49qNPdn+jsoeUhNjMaEGd3R0QeV7C+/+nk4PvIY+8VglasAExQvCYL00OSM7Opjgn8iIAELMIPHRwmPBinohHSyWRy9OPPA6OLE6LTIevP/eEnTKscDz3DEACLe1E3c3EUAhJGMGacXwS/8O7eudk0kRtp0aT6PsaGk4P3EIrMqYqSAJCBec2aDLyW2KLM0t19D94SRSvmGptQ7te7ftNdzJcea7Wi2oLicA9xVgAFJEIIFsI2OGcLGCXOQLx0e1JMlARC32+lUNG+5gJGiOI2VWoyU2CSyk7GTmQA5kB0SAQaRfO3DeG8Px6sRjwmsz369EH6ET9uSxyV+mFGJm7ICUh6r2IUvl8rT00UjvceS4iQCN0Ioe1Ct0LEEGG1kxtWdIexE4qW6mzMrFYcvFZFEtNKI3cSSIUy3dvR3BoMyJjmmNoQ58kYjTgyExIWwyAAiGQIDljB3MyrmqpmU/N9VVeFvq8lqxK4QelRM5sCVay1OgEyggO3JkiBpGhxs9tX8/3b1Sp9qx5Xoosv54PV3bykp5QV3t3a/Wb15AnNLmzJ8OaVwW4rA6WrWWyDWYD7wn3WhNkAGt4PEkLYhZ36/bXH7f2GiOMKc7wdePbVTxegTwyKhF2AZjckUE/XQ93Nd1q3bU6thRfz9PM6NNh/bheLrB7uCkdjh7opW95Vs+2cZ/lFWPwYm0tth5l5Ru65SchI1HFDzWehNgYIgniKJtFJazUAcK+HQmSGo4jYIy34YFMeAYHpcqPcHDFKPfxubobnhR50eilEPoo7AoyMtXCMY6PCSb7SWB1xOP69SBIeCvymn49vi04YICyk7jPXiezrGBYaQZ+j9vXH4aIjPjrmxYWAHLidh9BIM+cQmHlSzGZASu5gVccL1lrHQNF4hVyhFu8GHbLWYnXIJauqQ3GrFTCMb45MAICiREQJKlFXbLvpXbXheG8+u5DmGUSODXLbunUD954L70UvflTaud7d0bt3cw4yW66jD+ltQZaZtNm1VG2SW3Hfl6YIsHRBh6iIEBRq9b1ikrxcptklzKv0nXb7XVXEiBGUIy6etqEpd2W3qVXebZcwf/VM373ZLLbohM0sSSSzipC6nd2+6PkmzVcz9T0ACHOITOSBaLZoVqsWoMhJvsGpNvotkUtVnXg6xkjjmY9xhuo/Pvc9hPXsVITvj14eU58mROGTGfWnMnr2MSKioCRCIdQxO+lB8vxxSgF86M8j6kkCFEACSAAUoCn5GMaValWBHMzRp1YxRDjCH4gIIOrZHN3QHU1JtZoBoCPB+FQPdyQ0q27inAxdQmxnSCXd2r5YKRAYAyPWXsu+CgVAEMKQWGbYzhGgDoPqekhaYrekGIpBl80dKACQGZAYmUFeAF40mpUQkNNFDSkPkktCgl3j2ywbAIO4EI5bCmFxGUPL/do2cYnmpuw7mO3BelPR0lagmNqoEXxwHUx7KKVwnfCjGo3FORAJOjn5uMWfz+ipjP7/wZBR7Z9o0EcvRhxjw81FGu401OhW4L2Q+njkqWaFk/yXo4CGFkLiZhFiE8JsTPYzREJyFoTsCI42dnQ0VVX1WkxVc65dX4a+bve17w6pASzFsqNj4NHdBqIx8Y4sKKACd1m2Q0Zu5vO6fQsSGqW73bo1xdLni4s2dwkjPJ1ru6jR5941V19ta9dv39X+nsMw7ymD5gKwHzKJzIP7CkcmvWEY5iV0OpjZBc0CkjYWEw997fYCOusKMJGgC9Xl80UFRCfolXaBXm5LiO9yjbddmqW6rC44ND1ElSEoDoFiwhCM3BWhUfWhoAjEhOFCIJpaTcxkdKYmj8QCxwn9yNzzseHLuP++pyQf/flRhXo23B1HiocPHW56Uv2L+PGUVjk+x3gO4wFLw5ESF1klPt3Du/3F7XLj+4GGQsUrJmqLdyeBJsBDEv/otDAxADCjCAhbTJJa4YYwAiWCwcEcCdFIs1cujIhI6LTdbR1Q1UtnOZfdZq+d14z9Ten27obACgDoY7quV3RiYoex/58lUKYa4/Yu3N0qPNsGhuF/6XPe3axdQAUMdLjzwbG2PJs/BV3Ivsupr3uPyyW0Vl5p3qf6/E6kpf6S2jqk7fZeLtK8xtbe3pT7Aa/298ttUgB/1ukMRS/tqjSE4Y3vW3dESuCg2Obr8KS4dreaB6K3XIeGY6LLjnyIl4PdtMigO4udN0Z3Oc0pmQ/hal6gx0bYQLOqFjPmwE2TApmSGXt0gaNBdZj1o1weyS4PnS9P4qLvD5zYXQ4G4mOpcp2A+qm862BWjvdiBNXJB6IHn/0gG49cN5MKjhPn8gj5PzwDjj738TE+VdVJxIYxxtTGPVVGFicDVjd7j4r8dJ89ruNTe8DGvFJzsumzccWbGZKrApgxIhkOVb26FvNsVqxWVx0L8cwNfWp7OgFso+ZmYhJGhwHN0Pqh5CFohSCRyYApMLJ4EwkBdCBEB0RmTilCS0jFsmZFH7z21g3eF9UVXnzOYSVlnYjYXkC3R5vzzBOK9FVDiE4WGqRW+l1uyQ1gaIkULGtVbdqwaGfbfudiOA/chLCji/ngCQarUgiyVCsOAF4rApAHEUKKDRMRGrg7AQJOXECnWWBwsl/9MGzyKA2Hq/zgc8/GqZr7lI+P04qZ/nvQ98dLnQjSJ/lBFYGXFwv/itavX+KOIGOAAFaqZXxPQE/BhQPCSojIzMrsxKQoijJ4YXYHs0lhezUA8gqC2O+13/RarPZesuZ93ZUKTqpWazFziQBAx9UVGcdWs6P2qJFA2u0d3N8UyzTDiKxAjqgBHXxXByxDvbyecUz9tu7WNcU5OJd11iVhEIrU9EKZt3N7cj00PKzXlu+99LW/S8NG50tv0Tq6HEIG9rH0MDZ1VTkkuXnC8k3gDpLAgkkQ0ucxLkP4Ihrg64v14ptha3sfTHi2oHn9XLPWYa/osO9ytBZiTc8DGHoO1TIjCwVBR/OSc4pCKIwPabJn6uBEFB4J0HuezcFFJzoe8xGpwtMw6cH/mPCjY7HN5G997FI8RjnHNl84mnUnvt15wvJHh5mBOzO1q0QrsB6wGrEoYiU7W7PH/A//UAI2OoGT++i4u9ZHpkVRJwBQKF73u6HsTavVwbSYZqjVjkbVuIFMfPiKAMAMNl6VEQGUAkkEBDQjEreq7kbsClqMBq2dl6pN06S5QLstZcC+RGkFXZJaU7VqsysM3tgl7gbQwW647JBWtrwIEpGoN1XZIrgyWlMR0KivtatxIYBDt0Yv2swDMdRaK1IQCDNypNnFfPs3pl6QqjDF2FOcWfHCABXyUKSkIgPOvJQSjawqEDCOTeSBAenwEwDtH6j8frdcz5MShlFGT22MT+tjP/n5qfGpunhQBHaC+ZP55e8tTXnYDVYc0DXoGS5xpkHH5iwwhewVOgDDIXhWL6iOyU9GLW7Fhi4TQLcf8oC1qvZuBnUYu3DXUqacvVIGADJlUxzzUYCJUgJhINy4lD3s9haCzGJjYQdss9VFpardVmTmAxAWs1w9tyuzjLUOuXPOhArWoItJAa8QUrl7Lbgu+03XMbfuV0/WLJ3uVmYLunjHCUOIiG5m933sAjc75qp3tztifRICBONZaOOyFenerPNg21+wviWUKE8W1fRu08mw6/LAIKCsfbmIy27Rx8/YFKK1RAMBkoMwR8EmpigBx/4Z+ABHnkQ8TgYeJuZDGvShc9enesHAKWg1adAH02KSzoMadjPkj17QDhLqDuAkfAqLPtwLP8TNNMrccTUYGVNwNdXy9dfP7u3/09yXNbmRJOn5FZEHjipUkWwe09pu7czazszu2pqt9KYHPelny0wPkt5kpreRdqenNTvdZLPIOnBlZkS4ux4SQAFVRU7PJVOQBjMUEsjLMzz888/9u16tuftAEQPasEZ3NQJARCFOR/p5iEw82i+7k5l1g2XVFhsR8Q4z9+PuvLiWollyzjkXN1SllEopJQ3qjgjBc1bVbO7AAJADCGkIxQu7YZEJgBkhAPZ96bd16ZjXTk2CCUSukezlL41v5e4t11yFF/zbq2+e376M66j/pqrf5OcDpO+bDxudrYZ1DlDN5MW15GVpvPHGalqchQsnA82GRBdUk+J2Npt5s3XMdxsYctHS6jppgz6ve9Mq5owDV4iLFjLfXWlCdBxSdtSpS4d87e6rob6gs2ndb9O6bAXLBOXDX/3suS4AcQjSEKowhggSoZpwOwfE5BgdyQEelrVROXm7U3E5WNfDeHl3o/eA8+Gv44sdtY2BALDjRY06DTv9z13MZQX2aohI9CDAepQ62q8WwN0VnfZcH0DcaWYDjl/6fSofhxgHEImomTftWdPfdFaUIAS24gX2gQ48PI7d67jWNHMzKnmsc8dshk6goGqaS05aipZsZpazmoEWNwN3c0vuZjAuVnatv8ERaIR8ycxIECQYkap2nXlGJJaqqirmioVQ5rG7SKVPMaoFrKhJ20qN6guPWDUN8ULayvWjiRkRS11DFZSrGJgaKRWUUjRz9twPW/aGJbibuBhiMDcPqqpWaqz7oe+tsIG7gxEk7K0zYA7kIGzm0DEZQ0NgMXCnay7Ji7qquFMMHDnG6K6AFiKLcIgowSXcT5P4pD/9Uxz+Z3Gl/QYHfADvD+AvSVGR45AQTppFAIwEeAAABCQzqy7Dq59fDN277XVphxYwDYRZs405tRPAAEbqhruNS0bEUaOj4JARXQXR0IqXbCVlBzYDd1TVvktje/VdgUdxpV32TNXAXBKRIFRkzm5YoitIn9liyIybtTBSOw3N1OomhDhAtMlU7KcR27j5zRa6aiHzYVu6rrv4l8nsDO1V4Dd52nVdCOHKMGNqJhAZt1tqS9UGFsRizXVMELfdcLvZOOCUlbtQMt59hDTA+SVcXszW2w1s7fzlrBIXNsxoa0jhjsM00KT0Vu660FzX9VS7qW1z26jNGRWwIxiM0131t4v4oqqCoGJV06SNwjFWEqrUTnaqdsy8Y6mdmuTDJjN/GOvoYQh1vGbwMYc+2qg7HjS0nyLB/UnjdC3xI1Q+xskWgYiqug6v+NXP57ffb37452Uo4u4FDNENAZ+6WLsw3R0ckHRfmeQ6ljEXArOSMe/KjdwMUlJCGXU43cdwgFGBFYOJu7MwRrKaOgU1GELVb/n9x1JNETHEJoAATrFAHobSQOVQEmZfcLT69luCZIuG6XLoypB+W939Og0ZFq+5wXprOJBl1c16jYKVAWTxlIlMOKauLG8SV9WzxYUTCnd5ydobDykWbKbaLGCCZ9dvuz4N7JZ6BYdmMQ0RiEBCIKmnbzisEGPcrFIZumxcVuaDyZ14R7NzPP9lVb/WnDe1RZaq5B5hrB9HYjueup7K/P3xrZoeA1XHcKa7jaDqiWd8QCL5s4xjEMBdHh/W8Xi074iUZs9actlca75xSJBd07hueLijk0PPWfdTqQOAGuxodepj8xBVUx25UWQG5mPLGiAGLA7uqIhG7CSTgBFLTaiOBqVwMbaEsAnI5m12RqmIlBEQilCgZNmjhEkzuUz5rviACtGjSUva+fbGEH2O1N0OSSpltAJBvRSAXhuOgKKWrSdIxERYzLgMULJ5VpcAwgARjAsxu3tFyEBV9NBwmKAUUvdeBwwKMyurioVDveGsXS/aY+mBeqbC7Rm1zznMScgChqIIDoTou6zSvf35AeU5Gg/94Y+Ifj5z7x/PwId59JhxMmoW/vE7ejyO2MlyCLT9wFwetwEAAHZwU0BDciByQ5RJfVZR7F9PSveNfv/rq1k/6Uq/LltU4EoLuHlwbwCHUZgcDRmJ0d2hZBhlzsrYLER3jsJUXNHVELEJdVKCDmIJxJ6GjofIhSvguKDpvNm+QEUvtvbbL1a3Qk1uAlZ0U7bbqqrwCxeWKjak2TT3G6Pe3M5TvqFpV/07br7xm/8y2Xj2Cz97bZVW2vfvf0vfQT1b5voVJUvWedvO+7RZvvOhcXmeurKpGhQF1dRfQVacnU/Ylx6hfpMlZBto3V0C1VmHYU31PM2fTfJmUv63+eu6brj019SBbMPba51nbA1TmRqiSmriJH7wjF38xbz5imLkabNgwJI7dQDC2tiLYtkRZ0d3g4hymix5gDrdoz9jmLyXQN/fYn2aqjeGPqcmRwWBCBHG1oN0jIEfpwnGaXUPQu2i+6BgDoYABMioBQB8X8uOti86Gk1RFZlh10FFT5qHfXauvpenYZGqqiaTiVzmyYd6kw0Vo9RICJTVM9po1rvne/T0h2aV48OgY2g15iwdzcf+OEoA2Wzro0ADCiAKIYFUXk9p8lIm58Eq6AeQ1LKXStBz0Yw0AGsUqEbyqg49BWf2vkskDqAOQMzni9Zf0lXTpbuEqnABOANmiluCgYYh6R0rqGmXmNTY3XOCvIROvatD7FTUIGc2c9RYBQTMuZihWZXu1KHPyUPDLN4PKSexrYYNCgSKwo5WYCwrcEIkt2Taa8lDFWU2r8+eTYUjU+DxggO7Z3dTy3uj5AO9AeAPi4oeNOr8pOvcmcGTQdKff5z4gePYa0SHEE/odo++TQAOeyBAVV2NkUhkPp/DWRdmL97+r9X6OqxvLZkX7VmRARSL7qF7dzdVIPJdrznf7wtxlDQBQjcRkrBral2SkmBZ90AwP5/VX9L8snn+y/bip7Gahl/9j9X33yiX2WKRX7xAR1ktE1O9uWVNGa7UjTUizJUEDKNByTk5OyG1c5H5bPFxG/9r1O/isDX+p9xe8vxd7BL1GHVZh6qtzzzWObaTUGMadNsVzcG31WIe25Y8Lp189py2N9ZtANJ51cpd6ON1AvN2wSiWCm07BM1pSN3vSp6G+fMQArvidAr1xENAY44rPStt0XT21/LFT88v/6ZxBQUcPCGiMEowJB3Jh6WU8Zbd480Pcc/Ph0UPO8mevjsQisdb/7QZ/TFp1Z2/HvOZ9wka3BE19xS50/l+3EjGWfGw10eTKIEjYAFAcORAauk+wYV+/lfzMGm62/TdP1+tb9LqzsgFAbQMFsL+TN1pbOINCDRSAF3HyVPGXBujgpv7zke9rGex4fBsWjXh4s388udNe1HRswKxmOmLF61tS974/M08tHHld3kIX3Xx7mPuumH79nx11X98v6nbuqqjtA7AqyUXqCLCYKEjXPzjeVymd79abpbYfWtpQxGQGGQBkbYxBqprR+e6j5PM2cJatCf74bZ9dlFfVPk1eQPTUhFCGfBum7t1vvhJoYg5DVB52jJ0EmUSY5Hzupo3yCpNYlFAZA/cDnEqwBy/tQkiL+LLf5osvgphruQYiOqamLGtxZnBSQhCEJFdic7BxT+o0fm8gfIpzPIpdbh9YP6gvGc3/ph17T0ODyfP1GijI4B6AGXxiIZyAOo/7d8JQPfVq7tfGU0M3MFrAI1nFNsmp0lbCxP2S7NiKRfYK++O/kiLH5JhMK6WnPacKWBxAAhAzCzMoR5iJe25VGelfj2cfXkGEypgNEQq3jb2/MK2cRXPtNBQJ5FgiBDM8pRueZi03MTaJ05tclJwQqUypAykUbJTzS4/IemD/8qHDyDZ+Dm4WIjQcEc0dKUuym7uMTNgZDGBdZ22XfE1xS1VFAoCAISITSsIqXbqgmTz4CrqKXMeVLQLwo5imEMoHAkI2nBGDfNMbaAAat43bT17GcM5BkJCFMEQPRCKFGfxsW5rl0m+B4AQ8UG6mT/rhB/d4idioMPrUzDoCZHyDxp+hKHud/6IfvUgWeoOiDIC+WBHjPyjH0UkM2VmQABzKEbAo4w70IjMSwgCAV794lX31dB+f3vz7SZvNK0q3GDJmAsZQrHB2y0iqubR0VQQEXZVBoJUx4lECgurW1l8cRa+KnVTzSaCZIA28ADG4nW/1VL8Trt+mgrr9qorKftdO3YcH2PYWQPz13b+0tddr6oFQ+59In16R5t36IK9DP6185dtfdbU74fy7QAb2DQ9L0poq1KJgOB78LVu/G4qkzhpr3Bzi9tpdf7dDza94/mvsQ0+/flQltitBnVwAM2TqqvAYzdPMlVOpVzfpuiTllY3qTqrlFQYWmyD8s2mD/EsbfD6bvnmuT3/ul58hYMU8RkyESNLoIDGOVgiDMyNcK4ajFIBAJgzEgDgCO8hKIIhiN8TR+Ao9t2lfA4E8NEOhEB1F48TgQMyw0jXeARhjYjK3t06yglAeZDu9EPLz9HSwMENC99HI24wFprZffrej9KnQKPi7e6jT+Ogo13v+4SiPfVsnrYtDZW+eDmbVU2/TNfv79ZvMwxmHbk7llRnLqUICiLmnClGRBQBIpLA9STV0+r85aSZVLPLOjyviAA8u42cq8qyrZfD8mabs25uNacECVNvpYAO/Wigu/NU40AUZFJFd+/Vg2FufTsBCjDkAoaWSAIIe/0Crr/bkATuVZcao3BNCGhYnNWG2ZDBILtAKJxzhqGow4q6QgSbCjssGZkZGJbbDWxGnU3EhFYYQ6QKsOYmk0SOUrljurP324+5yZDX+YrY6Pwnk+mLSWyCunsaF4JQyngjLSAB3nvYT96vT9xD+AQgdTCI+wXlUST+eMK7B5VGw/h/xdh/Qmlu9zztsrSHP4KZ0cN5/+StMHujbcPNq7r9CqPFbq2bj8UG1SENV/3QFy0ELmakTZbIzbNYNaGayewZVLXEqQCBYy65SwnLQF5EEy1/96Hb6uomr2+LZitdAQDm4EVVtQT3PfEMEWllwMBRuXIimoYKAlYvuG4wFb95m2TTDHeOnAPoF/9BLl9fXP/6I177eTn/eLUyiBkGJQ0zvnze2Mfcr4utYmV1WGjbOLGvWuznVRgI+kFLmjTCDEvwnM0LsiNFqCI/+3IWniF47tebtAV/P92uhw/fbMLMuWXYFLili/PJ7N+3/AVkNFQhEkB1w2HIpWBdETcRAJBURPBRhYfv/gHtqj9OrWqvHbGfVh/e7JO0zRHpzt0fZPlHcPChF77/9L5E7nEI5Q83PrWhR+/vD+lTUoh4kGTkyGCIDvhk7HZ6vXQBYMIGaI7JcR0rahdnBAiWuAIA2qx7U5hMZmQKWFwUSCGQWePuDoyOmku5SX2fl9dpu+qv3i9t7ZrNEltGM9hSAoCiG4Xi5PXQHk51xzZA92yW3NHQlCJSTfNJVIc2xO2Kbu62mw60IFZ49g/t7Ov47r+9++Ht9wSTlRUHIMTQVOf/dhUum/UVpl9pGtyGQatUNbEVj2Sig1LhxvEMjSH/ztHrELmSHOpVG+rFa/eZ3v5rX4bohstVKYNJDDVO1h+WkyFMMb74e5p9PSXJugX0aMEDBwADRzeyEkyRCKpKmiaEQA+i67G1JxrgWLJxGgbByZoS4VFkfh+UwF6HEwCOWCBHG5/e+ce0qb3h/n4D/bzO59H07E8Slo8XBDiWb4x1o78PXzBXd2VCcGd145qYJMZd43lTIpCGAABjBqnBA0KBUSzA1N2hoLmnPl2/2wzboVt5Gdw76AfXAq7FTEHBMxpoAXdGQMqaYJw7iXY1gojMZMqICIoKKmwMhQihMXYbUhr6JpeYbszitprFyZezBEV/EMsBCHNfKNnytjmriRqsznqMBlwNFcmE6qRN8WSimYxxk4OZY5+1FArAtbqJFt5uspr317FbZmZWLe6ONeRtYRMu0M744utYVRUaUEGCYOHgXhGczDxnZSZiYMYnjOzIQp5U5tzf9BHTefy1fXQy9hw4eM6nrOLEsz/4rQO/7s/i/feAFwDIKFdwOKyS9ThU1LIdPx7PcZ+o2F2+/dkd0IGyv37EsWHMAABubg4QiR0Rq6ben+5oPQwOgJqgd+DN2rvrTfowvP+2G4bcd2mkjwwl+745KACokruj4SimiGwA4GalKACAMABYf9/8nAolxbHPVqyzVOn5C8ZrzEF/eOeeoX0ezv6uDj+dfvzPg/4OAlbXvi02yG9m/c9qmNaTl1JvPa+ggpj7wm0FUWy9cUgBA3yMUKA3FXdB3m55gG7SWnnfoyJ2PN1elnDTvFkT1uUd9oND7xevqp/8/eL8F9hkcxdiQfEACmpEFIIggnk2DUSAqBKAGREEduUzIzNyP4UQjk/jqIA+3l3zDEcQDfKeiLkzMgE7SiICuo5Iy3F6aW8zDyJ3ESjlfkHodIRv0okMDQAcdSU5BOz3k10IcAwR6K7h2SjbIoeZedxiJxZztCTHo9/65CL9aLl9HyT6SREWwH5iAABwdHI0k+TuZtrf6vJmvXzbp9thuOrXqaSUSh6jQs9+35gcANyP2rD47todH9vxwe9Oe8wlIQ5dkogSebFQNVfUH67ow3fd8xAuZ8/lH4clD9v3Cd9T6IPLmjFXE57/Dbj6zTdlebNlI+vjZjOUNsdJrcVz6kBkCgz9BqwLNZPOfK1rzIzkJZeXfcESoQlpOmx8IR5fxzf/cb7469icCSZjGuNFByAzPORNQgjuSiR1UyGBudLn5VVHT7+fTvGeL3kfRv95hur9Rf5xifhPud8H7RRGhv7+/77seLfpKOFxgnidoKpP7+DUj9w7CNxdp/2s+2ABBIA2VvBpts3Hsrzabt6nvC7DyhJbMVBwGBOmfmJ/Y/+Mo0N6+rE5bDNaOe8gEjbViFBXCu5nZ971k/U6d3c5CsHc4yX220JvETPhotYNJMrDRAEsTkJTxBKWDaFD2qahK6rebbOZCU6oKJNi5eykxYqN2W9y7oW9LhE2UiucXUJ8QdOvhb4wIJeamUAEwFQLh12Cw8ysFJNRuXd3RXcyxnBshifD9h89NWU8LNL5E8beOv1+9rkfj8OgYyN5sPGRaR5+8NhATxvF7xkjBwT1yA6Op+X7nd2bzv2sNh7TaOuw4yziWCS179sLZI6mhfIqL/81/eZ/Xm2XW09QzNQxD3nPgEZTzOUk6UF0uGH3DIFDGf7YuPRwJOo26oGDOiKKBEhYyKROiH52Ua1W3eZue/1b1HUTfkLtz8QmYf0+Ld+tNm8rTYEnogtR8Rnj/KwuqfTEACxLGfoSQpjPtJTUe0MgDiWTNXHQIYtX4iIIkjUqywr75WbyDH72n17IuVRfsgWggXEeQsBKPA+aO0ypIyIRYiZmRsjE4J7dRYTRTwz0kYWeGOjoVA8Y00Oi+59grn4E1/8Y8Osz0QvtQ5572zs6wV3q7LCO9HLPOEREP02HPTRQIji1zuNP7aBpNx7HgzM0U9e01f5jvv7NMi3dE7p7AcsBLanZmG1TLY6nakCIO9D4sGA6dmS0T/rj0eoFbQdAA5kbALsHQgJ0mE60m8L6jvs7jF+wTG36opq9giGtqBj2bupb9cwuTcVoSOjRTQ3nmWsEUqQkkFu1yAIU1t6BDmgq2NBYZg2iKVPROAmXfxvkTQkNigXtwAsUc3YMu6QdmyXc9wsiIkYSkaqqEMuPMYWTi/wXQyuPXSUeXo72/GjrfZj1RJyH4H6kFn/6mX//38eIZ1dID/uaeXy664Mfsgiqh7XgfiCiMyOSI4IrAoCC4yh1ONa9kLubg/Zqae03/yetbrYfv/uwXW9KKWbghqXoAdTcTdK7CpidaCyAwA78ekLdAY9YKXD/5BAiIzBwRkRmDiEgItfSgXYJfviIQ0YJNTVs6pvrLvUlvy+lVwByJTAvSoiKBCEgESHsniJTAACze2+Fo7w4MjFWtYlEiTx5RtU0XL6Znn2B6jhqMztwNW3QQciQjFDBAxHFQIxaRZ60VtWyuIwSgJlBxwYXh9lrAEDwCGCAxSi4O5rz2B8ZeBST3q0N9r5lt0Yy3CWimcHdjZBx35EZQPm4O7Of1pgj65gkwp3c8gPh4COgaresNEQHcgAHFTi4OESgDDiGWTjqBh//1I9g1P9lxma1WX0cfni3Sqs8DMPnn/XRmA4GOrZjRtI9PeeU1rA3yn3Iv6uCQiQEerCeNzMOHCNPJlEymgMFhkDWgFC2GXbQq3pS3c/KhgCjxPNnShfHwgzhwEJ17c1FEyuavaiqqbSXkSi7j+EMOdAwJAJwduKxLvXhuYcwNiE096cSQj9+4D1j6JN41f9n4/8C3u9HtCjWCjUAAAAASUVORK5CYII=\n", |
|
|
1311 |
"text/plain": [ |
|
|
1312 |
"<PIL.Image.Image image mode=RGB size=224x224 at 0x1B5111BEBB0>" |
|
|
1313 |
] |
|
|
1314 |
}, |
|
|
1315 |
"execution_count": 75, |
|
|
1316 |
"metadata": {}, |
|
|
1317 |
"output_type": "execute_result" |
|
|
1318 |
} |
|
|
1319 |
], |
|
|
1320 |
"source": [ |
|
|
1321 |
"img" |
|
|
1322 |
] |
|
|
1323 |
}, |
|
|
1324 |
{ |
|
|
1325 |
"cell_type": "code", |
|
|
1326 |
"execution_count": 76, |
|
|
1327 |
"metadata": {}, |
|
|
1328 |
"outputs": [ |
|
|
1329 |
{ |
|
|
1330 |
"data": { |
|
|
1331 |
"text/plain": [ |
|
|
1332 |
"array([[[235., 197., 184.],\n", |
|
|
1333 |
" [214., 170., 167.],\n", |
|
|
1334 |
" [193., 140., 156.],\n", |
|
|
1335 |
" ...,\n", |
|
|
1336 |
" [250., 221., 203.],\n", |
|
|
1337 |
" [249., 220., 202.],\n", |
|
|
1338 |
" [248., 219., 201.]],\n", |
|
|
1339 |
"\n", |
|
|
1340 |
" [[255., 227., 201.],\n", |
|
|
1341 |
" [244., 211., 194.],\n", |
|
|
1342 |
" [223., 184., 179.],\n", |
|
|
1343 |
" ...,\n", |
|
|
1344 |
" [250., 222., 201.],\n", |
|
|
1345 |
" [250., 222., 201.],\n", |
|
|
1346 |
" [249., 221., 200.]],\n", |
|
|
1347 |
"\n", |
|
|
1348 |
" [[254., 232., 195.],\n", |
|
|
1349 |
" [255., 232., 198.],\n", |
|
|
1350 |
" [249., 223., 196.],\n", |
|
|
1351 |
" ...,\n", |
|
|
1352 |
" [251., 223., 201.],\n", |
|
|
1353 |
" [251., 223., 201.],\n", |
|
|
1354 |
" [249., 221., 199.]],\n", |
|
|
1355 |
"\n", |
|
|
1356 |
" ...,\n", |
|
|
1357 |
"\n", |
|
|
1358 |
" [[253., 227., 200.],\n", |
|
|
1359 |
" [254., 228., 201.],\n", |
|
|
1360 |
" [255., 229., 202.],\n", |
|
|
1361 |
" ...,\n", |
|
|
1362 |
" [254., 230., 204.],\n", |
|
|
1363 |
" [254., 230., 202.],\n", |
|
|
1364 |
" [254., 230., 202.]],\n", |
|
|
1365 |
"\n", |
|
|
1366 |
" [[255., 230., 199.],\n", |
|
|
1367 |
" [255., 230., 200.],\n", |
|
|
1368 |
" [255., 229., 202.],\n", |
|
|
1369 |
" ...,\n", |
|
|
1370 |
" [255., 229., 204.],\n", |
|
|
1371 |
" [255., 229., 204.],\n", |
|
|
1372 |
" [255., 229., 204.]],\n", |
|
|
1373 |
"\n", |
|
|
1374 |
" [[255., 230., 199.],\n", |
|
|
1375 |
" [255., 230., 200.],\n", |
|
|
1376 |
" [255., 229., 202.],\n", |
|
|
1377 |
" ...,\n", |
|
|
1378 |
" [255., 229., 204.],\n", |
|
|
1379 |
" [255., 229., 204.],\n", |
|
|
1380 |
" [255., 229., 204.]]], dtype=float32)" |
|
|
1381 |
] |
|
|
1382 |
}, |
|
|
1383 |
"execution_count": 76, |
|
|
1384 |
"metadata": {}, |
|
|
1385 |
"output_type": "execute_result" |
|
|
1386 |
} |
|
|
1387 |
], |
|
|
1388 |
"source": [ |
|
|
1389 |
"x=image.img_to_array(img)\n", |
|
|
1390 |
"x" |
|
|
1391 |
] |
|
|
1392 |
}, |
|
|
1393 |
{ |
|
|
1394 |
"cell_type": "code", |
|
|
1395 |
"execution_count": null, |
|
|
1396 |
"metadata": {}, |
|
|
1397 |
"outputs": [], |
|
|
1398 |
"source": [] |
|
|
1399 |
}, |
|
|
1400 |
{ |
|
|
1401 |
"cell_type": "code", |
|
|
1402 |
"execution_count": null, |
|
|
1403 |
"metadata": {}, |
|
|
1404 |
"outputs": [], |
|
|
1405 |
"source": [] |
|
|
1406 |
}, |
|
|
1407 |
{ |
|
|
1408 |
"cell_type": "code", |
|
|
1409 |
"execution_count": 77, |
|
|
1410 |
"metadata": {}, |
|
|
1411 |
"outputs": [ |
|
|
1412 |
{ |
|
|
1413 |
"data": { |
|
|
1414 |
"text/plain": [ |
|
|
1415 |
"(224, 224, 3)" |
|
|
1416 |
] |
|
|
1417 |
}, |
|
|
1418 |
"execution_count": 77, |
|
|
1419 |
"metadata": {}, |
|
|
1420 |
"output_type": "execute_result" |
|
|
1421 |
} |
|
|
1422 |
], |
|
|
1423 |
"source": [ |
|
|
1424 |
"x.shape" |
|
|
1425 |
] |
|
|
1426 |
}, |
|
|
1427 |
{ |
|
|
1428 |
"cell_type": "code", |
|
|
1429 |
"execution_count": 78, |
|
|
1430 |
"metadata": {}, |
|
|
1431 |
"outputs": [], |
|
|
1432 |
"source": [ |
|
|
1433 |
"x=x/255" |
|
|
1434 |
] |
|
|
1435 |
}, |
|
|
1436 |
{ |
|
|
1437 |
"cell_type": "code", |
|
|
1438 |
"execution_count": 79, |
|
|
1439 |
"metadata": {}, |
|
|
1440 |
"outputs": [ |
|
|
1441 |
{ |
|
|
1442 |
"data": { |
|
|
1443 |
"text/plain": [ |
|
|
1444 |
"array([[[0.92156863, 0.77254903, 0.72156864],\n", |
|
|
1445 |
" [0.8392157 , 0.6666667 , 0.654902 ],\n", |
|
|
1446 |
" [0.75686276, 0.54901963, 0.6117647 ],\n", |
|
|
1447 |
" ...,\n", |
|
|
1448 |
" [0.98039216, 0.8666667 , 0.79607844],\n", |
|
|
1449 |
" [0.9764706 , 0.8627451 , 0.7921569 ],\n", |
|
|
1450 |
" [0.972549 , 0.85882354, 0.7882353 ]],\n", |
|
|
1451 |
"\n", |
|
|
1452 |
" [[1. , 0.8901961 , 0.7882353 ],\n", |
|
|
1453 |
" [0.95686275, 0.827451 , 0.7607843 ],\n", |
|
|
1454 |
" [0.8745098 , 0.72156864, 0.7019608 ],\n", |
|
|
1455 |
" ...,\n", |
|
|
1456 |
" [0.98039216, 0.87058824, 0.7882353 ],\n", |
|
|
1457 |
" [0.98039216, 0.87058824, 0.7882353 ],\n", |
|
|
1458 |
" [0.9764706 , 0.8666667 , 0.78431374]],\n", |
|
|
1459 |
"\n", |
|
|
1460 |
" [[0.99607843, 0.9098039 , 0.7647059 ],\n", |
|
|
1461 |
" [1. , 0.9098039 , 0.7764706 ],\n", |
|
|
1462 |
" [0.9764706 , 0.8745098 , 0.76862746],\n", |
|
|
1463 |
" ...,\n", |
|
|
1464 |
" [0.9843137 , 0.8745098 , 0.7882353 ],\n", |
|
|
1465 |
" [0.9843137 , 0.8745098 , 0.7882353 ],\n", |
|
|
1466 |
" [0.9764706 , 0.8666667 , 0.78039217]],\n", |
|
|
1467 |
"\n", |
|
|
1468 |
" ...,\n", |
|
|
1469 |
"\n", |
|
|
1470 |
" [[0.99215686, 0.8901961 , 0.78431374],\n", |
|
|
1471 |
" [0.99607843, 0.89411765, 0.7882353 ],\n", |
|
|
1472 |
" [1. , 0.8980392 , 0.7921569 ],\n", |
|
|
1473 |
" ...,\n", |
|
|
1474 |
" [0.99607843, 0.9019608 , 0.8 ],\n", |
|
|
1475 |
" [0.99607843, 0.9019608 , 0.7921569 ],\n", |
|
|
1476 |
" [0.99607843, 0.9019608 , 0.7921569 ]],\n", |
|
|
1477 |
"\n", |
|
|
1478 |
" [[1. , 0.9019608 , 0.78039217],\n", |
|
|
1479 |
" [1. , 0.9019608 , 0.78431374],\n", |
|
|
1480 |
" [1. , 0.8980392 , 0.7921569 ],\n", |
|
|
1481 |
" ...,\n", |
|
|
1482 |
" [1. , 0.8980392 , 0.8 ],\n", |
|
|
1483 |
" [1. , 0.8980392 , 0.8 ],\n", |
|
|
1484 |
" [1. , 0.8980392 , 0.8 ]],\n", |
|
|
1485 |
"\n", |
|
|
1486 |
" [[1. , 0.9019608 , 0.78039217],\n", |
|
|
1487 |
" [1. , 0.9019608 , 0.78431374],\n", |
|
|
1488 |
" [1. , 0.8980392 , 0.7921569 ],\n", |
|
|
1489 |
" ...,\n", |
|
|
1490 |
" [1. , 0.8980392 , 0.8 ],\n", |
|
|
1491 |
" [1. , 0.8980392 , 0.8 ],\n", |
|
|
1492 |
" [1. , 0.8980392 , 0.8 ]]], dtype=float32)" |
|
|
1493 |
] |
|
|
1494 |
}, |
|
|
1495 |
"execution_count": 79, |
|
|
1496 |
"metadata": {}, |
|
|
1497 |
"output_type": "execute_result" |
|
|
1498 |
} |
|
|
1499 |
], |
|
|
1500 |
"source": [ |
|
|
1501 |
"x" |
|
|
1502 |
] |
|
|
1503 |
}, |
|
|
1504 |
{ |
|
|
1505 |
"cell_type": "code", |
|
|
1506 |
"execution_count": 80, |
|
|
1507 |
"metadata": {}, |
|
|
1508 |
"outputs": [ |
|
|
1509 |
{ |
|
|
1510 |
"data": { |
|
|
1511 |
"text/plain": [ |
|
|
1512 |
"(1, 224, 224, 3)" |
|
|
1513 |
] |
|
|
1514 |
}, |
|
|
1515 |
"execution_count": 80, |
|
|
1516 |
"metadata": {}, |
|
|
1517 |
"output_type": "execute_result" |
|
|
1518 |
} |
|
|
1519 |
], |
|
|
1520 |
"source": [ |
|
|
1521 |
"x=np.expand_dims(x,axis=0)\n", |
|
|
1522 |
"img_data=preprocess_input(x)\n", |
|
|
1523 |
"img_data.shape" |
|
|
1524 |
] |
|
|
1525 |
}, |
|
|
1526 |
{ |
|
|
1527 |
"cell_type": "code", |
|
|
1528 |
"execution_count": 81, |
|
|
1529 |
"metadata": {}, |
|
|
1530 |
"outputs": [ |
|
|
1531 |
{ |
|
|
1532 |
"data": { |
|
|
1533 |
"text/plain": [ |
|
|
1534 |
"array([[1.2839345e-05, 9.9995458e-01, 3.2532727e-05]], dtype=float32)" |
|
|
1535 |
] |
|
|
1536 |
}, |
|
|
1537 |
"execution_count": 81, |
|
|
1538 |
"metadata": {}, |
|
|
1539 |
"output_type": "execute_result" |
|
|
1540 |
} |
|
|
1541 |
], |
|
|
1542 |
"source": [ |
|
|
1543 |
"model.predict(img_data)" |
|
|
1544 |
] |
|
|
1545 |
}, |
|
|
1546 |
{ |
|
|
1547 |
"cell_type": "code", |
|
|
1548 |
"execution_count": 82, |
|
|
1549 |
"metadata": {}, |
|
|
1550 |
"outputs": [], |
|
|
1551 |
"source": [ |
|
|
1552 |
"a=np.argmax(model.predict(img_data), axis=1)" |
|
|
1553 |
] |
|
|
1554 |
}, |
|
|
1555 |
{ |
|
|
1556 |
"cell_type": "code", |
|
|
1557 |
"execution_count": 83, |
|
|
1558 |
"metadata": {}, |
|
|
1559 |
"outputs": [ |
|
|
1560 |
{ |
|
|
1561 |
"data": { |
|
|
1562 |
"text/plain": [ |
|
|
1563 |
"array([1], dtype=int64)" |
|
|
1564 |
] |
|
|
1565 |
}, |
|
|
1566 |
"execution_count": 83, |
|
|
1567 |
"metadata": {}, |
|
|
1568 |
"output_type": "execute_result" |
|
|
1569 |
} |
|
|
1570 |
], |
|
|
1571 |
"source": [ |
|
|
1572 |
"a" |
|
|
1573 |
] |
|
|
1574 |
}, |
|
|
1575 |
{ |
|
|
1576 |
"cell_type": "code", |
|
|
1577 |
"execution_count": null, |
|
|
1578 |
"metadata": {}, |
|
|
1579 |
"outputs": [], |
|
|
1580 |
"source": [] |
|
|
1581 |
}, |
|
|
1582 |
{ |
|
|
1583 |
"cell_type": "code", |
|
|
1584 |
"execution_count": null, |
|
|
1585 |
"metadata": {}, |
|
|
1586 |
"outputs": [], |
|
|
1587 |
"source": [] |
|
|
1588 |
}, |
|
|
1589 |
{ |
|
|
1590 |
"cell_type": "code", |
|
|
1591 |
"execution_count": null, |
|
|
1592 |
"metadata": {}, |
|
|
1593 |
"outputs": [], |
|
|
1594 |
"source": [] |
|
|
1595 |
} |
|
|
1596 |
], |
|
|
1597 |
"metadata": { |
|
|
1598 |
"kernelspec": { |
|
|
1599 |
"display_name": "Python 3", |
|
|
1600 |
"language": "python", |
|
|
1601 |
"name": "python3" |
|
|
1602 |
}, |
|
|
1603 |
"language_info": { |
|
|
1604 |
"codemirror_mode": { |
|
|
1605 |
"name": "ipython", |
|
|
1606 |
"version": 3 |
|
|
1607 |
}, |
|
|
1608 |
"file_extension": ".py", |
|
|
1609 |
"mimetype": "text/x-python", |
|
|
1610 |
"name": "python", |
|
|
1611 |
"nbconvert_exporter": "python", |
|
|
1612 |
"pygments_lexer": "ipython3", |
|
|
1613 |
"version": "3.8.3" |
|
|
1614 |
} |
|
|
1615 |
}, |
|
|
1616 |
"nbformat": 4, |
|
|
1617 |
"nbformat_minor": 2 |
|
|
1618 |
} |