├── Animals_CNN.ipynb
├── Animals_keras.ipynb
├── Facial-Expression-Recognition-Challenge.ipynb
├── Facial-Expression-Recognition.ipynb
├── MNIST_keras.ipynb
├── MNIST_keras_CNN.ipynb
├── MNIST_keras_updated.ipynb
├── Pima_Neural.ipynb
├── README.md
├── VggPretrained.ipynb
├── dcgan_color_images.ipynb
├── dcgan_fashion_mnist.ipynb
├── expression_recognition.ipynb
├── imagecaptioning.ipynb
└── visualize_cnns.ipynb


/MNIST_keras.ipynb:
--------------------------------------------------------------------------------
  1 | {
  2 |   "nbformat": 4,
  3 |   "nbformat_minor": 0,
  4 |   "metadata": {
  5 |     "colab": {
  6 |       "name": "MNIST_keras.ipynb",
  7 |       "provenance": [],
  8 |       "include_colab_link": true
  9 |     },
 10 |     "kernelspec": {
 11 |       "name": "python3",
 12 |       "display_name": "Python 3"
 13 |     },
 14 |     "accelerator": "GPU"
 15 |   },
 16 |   "cells": [
 17 |     {
 18 |       "cell_type": "markdown",
 19 |       "metadata": {
 20 |         "id": "view-in-github",
 21 |         "colab_type": "text"
 22 |       },
 23 |       "source": [
 24 |         "<a href=\"https://colab.research.google.com/github/mohanrajmit/DL-Training/blob/master/MNIST_keras.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
 25 |       ]
 26 |     },
 27 |     {
 28 |       "cell_type": "code",
 29 |       "metadata": {
 30 |         "id": "mE7zM-zj_3GW"
 31 |       },
 32 |       "source": [
 33 |         "import numpy as np\n",
 34 |         "import pandas as pd\n",
 35 |         "import matplotlib.pyplot as plt\n",
 36 |         "%matplotlib inline\n",
 37 |         "from keras.datasets import mnist\n",
 38 |         "from keras.models import Sequential\n",
 39 |         "from keras.layers import Dense, Dropout, Activation\n",
 40 |         "from keras.utils import to_categorical\n",
 41 |         "\n",
 42 |         "np.random.seed(35)"
 43 |       ],
 44 |       "execution_count": 1,
 45 |       "outputs": []
 46 |     },
 47 |     {
 48 |       "cell_type": "code",
 49 |       "metadata": {
 50 |         "id": "4R_C1zvA_3Ge",
 51 |         "colab": {
 52 |           "base_uri": "https://localhost:8080/"
 53 |         },
 54 |         "outputId": "abb820e3-0bda-4bed-fdbd-e012848331df"
 55 |       },
 56 |       "source": [
 57 |         "(X_train, y_train), (X_test, y_test) = mnist.load_data()\n",
 58 |         "print(\"X_train original shape\", X_train.shape)\n",
 59 |         "print(\"y_train original shape\", y_train.shape)\n",
 60 |         "print(\"X_test original shape\", X_test.shape)\n",
 61 |         "print(\"y_test original shape\", y_test.shape)"
 62 |       ],
 63 |       "execution_count": 2,
 64 |       "outputs": [
 65 |         {
 66 |           "output_type": "stream",
 67 |           "name": "stdout",
 68 |           "text": [
 69 |             "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
 70 |             "\u001b[1m11490434/11490434\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 0us/step\n",
 71 |             "X_train original shape (60000, 28, 28)\n",
 72 |             "y_train original shape (60000,)\n",
 73 |             "X_test original shape (10000, 28, 28)\n",
 74 |             "y_test original shape (10000,)\n"
 75 |           ]
 76 |         }
 77 |       ]
 78 |     },
 79 |     {
 80 |       "cell_type": "code",
 81 |       "metadata": {
 82 |         "id": "BgPvv_Mo_3Gl",
 83 |         "colab": {
 84 |           "base_uri": "https://localhost:8080/",
 85 |           "height": 470
 86 |         },
 87 |         "outputId": "2167cf31-6a8b-46a2-fcb5-8146546e1656"
 88 |       },
 89 |       "source": [
 90 |         "plt.imshow(X_test[1001], cmap='gray')\n",
 91 |         "plt.title(y_test[1001])"
 92 |       ],
 93 |       "execution_count": 3,
 94 |       "outputs": [
 95 |         {
 96 |           "output_type": "execute_result",
 97 |           "data": {
 98 |             "text/plain": [
 99 |               "Text(0.5, 1.0, '0')"
100 |             ]
101 |           },
102 |           "metadata": {},
103 |           "execution_count": 3
104 |         },
105 |         {
106 |           "output_type": "display_data",
107 |           "data": {
108 |             "text/plain": [
109 |               "<Figure size 640x480 with 1 Axes>"
110 |             ],
111 |             "image/png": "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\n"
112 |           },
113 |           "metadata": {}
114 |         }
115 |       ]
116 |     },
117 |     {
118 |       "cell_type": "code",
119 |       "metadata": {
120 |         "id": "yRXjnJNy_3Gq"
121 |       },
122 |       "source": [
123 |         "X_train = X_train.reshape(60000,784)\n",
124 |         "X_test = X_test.reshape(10000,784)\n",
125 |         "\n",
126 |         "X_train = X_train.astype('float32')\n",
127 |         "X_test = X_test.astype('float32')\n",
128 |         "\n",
129 |         "X_train/=255\n",
130 |         "X_test/=255"
131 |       ],
132 |       "execution_count": 4,
133 |       "outputs": []
134 |     },
135 |     {
136 |       "cell_type": "code",
137 |       "metadata": {
138 |         "id": "6jOAvg1P_3Gt",
139 |         "colab": {
140 |           "base_uri": "https://localhost:8080/"
141 |         },
142 |         "outputId": "921d8c62-0ed0-4add-80ba-71cf6ab76a6a"
143 |       },
144 |       "source": [
145 |         "number_of_classes = 10\n",
146 |         "\n",
147 |         "Y_train = to_categorical(y_train, number_of_classes)\n",
148 |         "Y_test = to_categorical(y_test, number_of_classes)\n",
149 |         "\n",
150 |         "y_train[10], Y_train[10]"
151 |       ],
152 |       "execution_count": 5,
153 |       "outputs": [
154 |         {
155 |           "output_type": "execute_result",
156 |           "data": {
157 |             "text/plain": [
158 |               "(3, array([0., 0., 0., 1., 0., 0., 0., 0., 0., 0.]))"
159 |             ]
160 |           },
161 |           "metadata": {},
162 |           "execution_count": 5
163 |         }
164 |       ]
165 |     },
166 |     {
167 |       "cell_type": "code",
168 |       "metadata": {
169 |         "id": "BsjgvmUN_3Gv",
170 |         "outputId": "80dcfd26-3777-42ed-d9ab-60ff00ad0b1f",
171 |         "colab": {
172 |           "base_uri": "https://localhost:8080/"
173 |         }
174 |       },
175 |       "source": [
176 |         "model = Sequential()\n",
177 |         "\n",
178 |         "model.add(Dense(512, input_dim=784,activation='relu'))\n",
179 |         "# An \"activation\" is just a non-linear function applied to the output\n",
180 |         "# of the layer above. Here, with a \"rectified linear unit\",\n",
181 |         "# we clamp all values below 0 to 0.\n",
182 |         "#model.add(Activation('relu'))\n",
183 |         "# Dropout helps protect the model from memorizing or \"overfitting\" the training data\n",
184 |         "#model.add(Dropout(0.2))\n",
185 |         "\n",
186 |         "model.add(Dense(256,activation='relu'))\n",
187 |         "#model.add(Activation('relu'))\n",
188 |         "#model.add(Dropout(0.2))\n",
189 |         "\n",
190 |         "model.add(Dense(128,activation='relu'))\n",
191 |         "#model.add(Activation('relu'))\n",
192 |         "#model.add(Dropout(0.2))\n",
193 |         "\n",
194 |         "model.add(Dense(10,activation='softmax'))\n",
195 |         "# This special \"softmax\" activation among other things,\n",
196 |         "# ensures the output is a valid probaility distribution, that is\n",
197 |         "# that its values are all non-negative and sum to 1.\n",
198 |         "#model.add(Activation('softmax'))"
199 |       ],
200 |       "execution_count": 6,
201 |       "outputs": [
202 |         {
203 |           "output_type": "stream",
204 |           "name": "stderr",
205 |           "text": [
206 |             "/usr/local/lib/python3.10/dist-packages/keras/src/layers/core/dense.py:87: UserWarning: Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.\n",
207 |             "  super().__init__(activity_regularizer=activity_regularizer, **kwargs)\n"
208 |           ]
209 |         }
210 |       ]
211 |     },
212 |     {
213 |       "cell_type": "code",
214 |       "source": [
215 |         "model.summary()"
216 |       ],
217 |       "metadata": {
218 |         "id": "zkUSV8DUKYM3",
219 |         "outputId": "b48188e5-12a2-4cb9-b75e-ba555c91f9b3",
220 |         "colab": {
221 |           "base_uri": "https://localhost:8080/",
222 |           "height": 276
223 |         }
224 |       },
225 |       "execution_count": 7,
226 |       "outputs": [
227 |         {
228 |           "output_type": "display_data",
229 |           "data": {
230 |             "text/plain": [
231 |               "\u001b[1mModel: \"sequential\"\u001b[0m\n"
232 |             ],
233 |             "text/html": [
234 |               "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"sequential\"</span>\n",
235 |               "</pre>\n"
236 |             ]
237 |           },
238 |           "metadata": {}
239 |         },
240 |         {
241 |           "output_type": "display_data",
242 |           "data": {
243 |             "text/plain": [
244 |               "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓\n",
245 |               "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                        \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape               \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m        Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
246 |               "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩\n",
247 |               "│ dense (\u001b[38;5;33mDense\u001b[0m)                        │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)                 │         \u001b[38;5;34m401,920\u001b[0m │\n",
248 |               "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
249 |               "│ dense_1 (\u001b[38;5;33mDense\u001b[0m)                      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)                 │         \u001b[38;5;34m131,328\u001b[0m │\n",
250 |               "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
251 |               "│ dense_2 (\u001b[38;5;33mDense\u001b[0m)                      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m128\u001b[0m)                 │          \u001b[38;5;34m32,896\u001b[0m │\n",
252 |               "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
253 |               "│ dense_3 (\u001b[38;5;33mDense\u001b[0m)                      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m10\u001b[0m)                  │           \u001b[38;5;34m1,290\u001b[0m │\n",
254 |               "└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘\n"
255 |             ],
256 |             "text/html": [
257 |               "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓\n",
258 |               "┃<span style=\"font-weight: bold\"> Layer (type)                         </span>┃<span style=\"font-weight: bold\"> Output Shape                </span>┃<span style=\"font-weight: bold\">         Param # </span>┃\n",
259 |               "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩\n",
260 |               "│ dense (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                        │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>)                 │         <span style=\"color: #00af00; text-decoration-color: #00af00\">401,920</span> │\n",
261 |               "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
262 |               "│ dense_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)                 │         <span style=\"color: #00af00; text-decoration-color: #00af00\">131,328</span> │\n",
263 |               "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
264 |               "│ dense_2 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">128</span>)                 │          <span style=\"color: #00af00; text-decoration-color: #00af00\">32,896</span> │\n",
265 |               "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
266 |               "│ dense_3 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">10</span>)                  │           <span style=\"color: #00af00; text-decoration-color: #00af00\">1,290</span> │\n",
267 |               "└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘\n",
268 |               "</pre>\n"
269 |             ]
270 |           },
271 |           "metadata": {}
272 |         },
273 |         {
274 |           "output_type": "display_data",
275 |           "data": {
276 |             "text/plain": [
277 |               "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m567,434\u001b[0m (2.16 MB)\n"
278 |             ],
279 |             "text/html": [
280 |               "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">567,434</span> (2.16 MB)\n",
281 |               "</pre>\n"
282 |             ]
283 |           },
284 |           "metadata": {}
285 |         },
286 |         {
287 |           "output_type": "display_data",
288 |           "data": {
289 |             "text/plain": [
290 |               "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m567,434\u001b[0m (2.16 MB)\n"
291 |             ],
292 |             "text/html": [
293 |               "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">567,434</span> (2.16 MB)\n",
294 |               "</pre>\n"
295 |             ]
296 |           },
297 |           "metadata": {}
298 |         },
299 |         {
300 |           "output_type": "display_data",
301 |           "data": {
302 |             "text/plain": [
303 |               "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
304 |             ],
305 |             "text/html": [
306 |               "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
307 |               "</pre>\n"
308 |             ]
309 |           },
310 |           "metadata": {}
311 |         }
312 |       ]
313 |     },
314 |     {
315 |       "cell_type": "code",
316 |       "source": [
317 |         "from keras.callbacks import EarlyStopping\n",
318 |         "es = EarlyStopping(monitor='val_accuracy', mode='max',patience=10,min_delta=0.01)"
319 |       ],
320 |       "metadata": {
321 |         "id": "gScFBC25PY_C"
322 |       },
323 |       "execution_count": 8,
324 |       "outputs": []
325 |     },
326 |     {
327 |       "cell_type": "code",
328 |       "source": [
329 |         "!mkdir /content/weight"
330 |       ],
331 |       "metadata": {
332 |         "id": "FL9gq5CXyDYf",
333 |         "outputId": "b954f92a-f0c8-41ba-c621-f3e454c2647b",
334 |         "colab": {
335 |           "base_uri": "https://localhost:8080/"
336 |         }
337 |       },
338 |       "execution_count": 32,
339 |       "outputs": [
340 |         {
341 |           "output_type": "stream",
342 |           "name": "stdout",
343 |           "text": [
344 |             "mkdir: cannot create directory ‘/content/weight’: File exists\n"
345 |           ]
346 |         }
347 |       ]
348 |     },
349 |     {
350 |       "cell_type": "code",
351 |       "source": [
352 |         "from tensorflow.keras.callbacks import ModelCheckpoint\n",
353 |         "import os\n",
354 |         "\n",
355 |         "#fname = os.path.sep.join([\"/content/weight\",\"weights-{epoch:03d}-{val_loss:.4f}.h5\"])"
356 |       ],
357 |       "metadata": {
358 |         "id": "rUiS8PS6Q1WU"
359 |       },
360 |       "execution_count": 38,
361 |       "outputs": []
362 |     },
363 |     {
364 |       "cell_type": "code",
365 |       "source": [
366 |         "# prompt: modelcheck points for val loss\n",
367 |         "\n",
368 |         "# Change the file extension to '.keras'\n",
369 |         "fname = os.path.sep.join([\"/content/weight\",\"weights-{epoch:03d}-{val_loss:.4f}.keras\"])\n",
370 |         "\n",
371 |         "mc = ModelCheckpoint(filepath=fname,monitor=\"val_loss\",mode='min',save_best_only=True,verbose=1)"
372 |       ],
373 |       "metadata": {
374 |         "id": "AZROKw7rUMjx"
375 |       },
376 |       "execution_count": 39,
377 |       "outputs": []
378 |     },
379 |     {
380 |       "cell_type": "code",
381 |       "metadata": {
382 |         "id": "UfM4obWE_3Gy"
383 |       },
384 |       "source": [
385 |         "model.compile(loss='categorical_crossentropy', optimizer='adam',metrics=['accuracy'])"
386 |       ],
387 |       "execution_count": 40,
388 |       "outputs": []
389 |     },
390 |     {
391 |       "cell_type": "code",
392 |       "metadata": {
393 |         "id": "4bgNGsohMEr2",
394 |         "colab": {
395 |           "base_uri": "https://localhost:8080/",
396 |           "height": 276
397 |         },
398 |         "outputId": "0d129662-95f5-428f-c57d-1cad6a31e229"
399 |       },
400 |       "source": [
401 |         "model.summary()"
402 |       ],
403 |       "execution_count": 41,
404 |       "outputs": [
405 |         {
406 |           "output_type": "display_data",
407 |           "data": {
408 |             "text/plain": [
409 |               "\u001b[1mModel: \"sequential\"\u001b[0m\n"
410 |             ],
411 |             "text/html": [
412 |               "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"sequential\"</span>\n",
413 |               "</pre>\n"
414 |             ]
415 |           },
416 |           "metadata": {}
417 |         },
418 |         {
419 |           "output_type": "display_data",
420 |           "data": {
421 |             "text/plain": [
422 |               "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓\n",
423 |               "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                        \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape               \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m        Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
424 |               "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩\n",
425 |               "│ dense (\u001b[38;5;33mDense\u001b[0m)                        │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)                 │         \u001b[38;5;34m401,920\u001b[0m │\n",
426 |               "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
427 |               "│ dense_1 (\u001b[38;5;33mDense\u001b[0m)                      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)                 │         \u001b[38;5;34m131,328\u001b[0m │\n",
428 |               "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
429 |               "│ dense_2 (\u001b[38;5;33mDense\u001b[0m)                      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m128\u001b[0m)                 │          \u001b[38;5;34m32,896\u001b[0m │\n",
430 |               "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
431 |               "│ dense_3 (\u001b[38;5;33mDense\u001b[0m)                      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m10\u001b[0m)                  │           \u001b[38;5;34m1,290\u001b[0m │\n",
432 |               "└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘\n"
433 |             ],
434 |             "text/html": [
435 |               "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓\n",
436 |               "┃<span style=\"font-weight: bold\"> Layer (type)                         </span>┃<span style=\"font-weight: bold\"> Output Shape                </span>┃<span style=\"font-weight: bold\">         Param # </span>┃\n",
437 |               "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩\n",
438 |               "│ dense (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                        │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>)                 │         <span style=\"color: #00af00; text-decoration-color: #00af00\">401,920</span> │\n",
439 |               "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
440 |               "│ dense_1 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)                 │         <span style=\"color: #00af00; text-decoration-color: #00af00\">131,328</span> │\n",
441 |               "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
442 |               "│ dense_2 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">128</span>)                 │          <span style=\"color: #00af00; text-decoration-color: #00af00\">32,896</span> │\n",
443 |               "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
444 |               "│ dense_3 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">10</span>)                  │           <span style=\"color: #00af00; text-decoration-color: #00af00\">1,290</span> │\n",
445 |               "└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘\n",
446 |               "</pre>\n"
447 |             ]
448 |           },
449 |           "metadata": {}
450 |         },
451 |         {
452 |           "output_type": "display_data",
453 |           "data": {
454 |             "text/plain": [
455 |               "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m567,434\u001b[0m (2.16 MB)\n"
456 |             ],
457 |             "text/html": [
458 |               "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">567,434</span> (2.16 MB)\n",
459 |               "</pre>\n"
460 |             ]
461 |           },
462 |           "metadata": {}
463 |         },
464 |         {
465 |           "output_type": "display_data",
466 |           "data": {
467 |             "text/plain": [
468 |               "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m567,434\u001b[0m (2.16 MB)\n"
469 |             ],
470 |             "text/html": [
471 |               "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">567,434</span> (2.16 MB)\n",
472 |               "</pre>\n"
473 |             ]
474 |           },
475 |           "metadata": {}
476 |         },
477 |         {
478 |           "output_type": "display_data",
479 |           "data": {
480 |             "text/plain": [
481 |               "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
482 |             ],
483 |             "text/html": [
484 |               "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
485 |               "</pre>\n"
486 |             ]
487 |           },
488 |           "metadata": {}
489 |         }
490 |       ]
491 |     },
492 |     {
493 |       "cell_type": "code",
494 |       "metadata": {
495 |         "id": "3P6VqYcf_3G2"
496 |       },
497 |       "source": [
498 |         "size = int(len(X_train) * 0.8)\n",
499 |         "\n",
500 |         "train_x, val_x = X_train[:size], X_train[size:]\n",
501 |         "train_y, val_y = Y_train[:size], Y_train[size:]"
502 |       ],
503 |       "execution_count": 42,
504 |       "outputs": []
505 |     },
506 |     {
507 |       "cell_type": "code",
508 |       "source": [
509 |         "train_x.shape"
510 |       ],
511 |       "metadata": {
512 |         "id": "Cx3EJrbRL1hk",
513 |         "outputId": "39759e45-33f6-481f-b69c-2255dd029fda",
514 |         "colab": {
515 |           "base_uri": "https://localhost:8080/"
516 |         }
517 |       },
518 |       "execution_count": 43,
519 |       "outputs": [
520 |         {
521 |           "output_type": "execute_result",
522 |           "data": {
523 |             "text/plain": [
524 |               "(48000, 784)"
525 |             ]
526 |           },
527 |           "metadata": {},
528 |           "execution_count": 43
529 |         }
530 |       ]
531 |     },
532 |     {
533 |       "cell_type": "code",
534 |       "source": [
535 |         "val_x.shape"
536 |       ],
537 |       "metadata": {
538 |         "id": "A2Osopl6L4Tp",
539 |         "outputId": "98e17ae8-045d-438f-938b-587d34843e36",
540 |         "colab": {
541 |           "base_uri": "https://localhost:8080/"
542 |         }
543 |       },
544 |       "execution_count": 44,
545 |       "outputs": [
546 |         {
547 |           "output_type": "execute_result",
548 |           "data": {
549 |             "text/plain": [
550 |               "(12000, 784)"
551 |             ]
552 |           },
553 |           "metadata": {},
554 |           "execution_count": 44
555 |         }
556 |       ]
557 |     },
558 |     {
559 |       "cell_type": "code",
560 |       "source": [
561 |         "X_test.shape"
562 |       ],
563 |       "metadata": {
564 |         "id": "tmbnxhfLL6i3",
565 |         "outputId": "b9793ce9-71e4-48b3-dcf4-67d2db3d0a91",
566 |         "colab": {
567 |           "base_uri": "https://localhost:8080/"
568 |         }
569 |       },
570 |       "execution_count": 45,
571 |       "outputs": [
572 |         {
573 |           "output_type": "execute_result",
574 |           "data": {
575 |             "text/plain": [
576 |               "(10000, 784)"
577 |             ]
578 |           },
579 |           "metadata": {},
580 |           "execution_count": 45
581 |         }
582 |       ]
583 |     },
584 |     {
585 |       "cell_type": "code",
586 |       "metadata": {
587 |         "id": "1IQyPm1g_3G6",
588 |         "colab": {
589 |           "base_uri": "https://localhost:8080/"
590 |         },
591 |         "outputId": "37ac720f-2559-4b54-c17f-7e7b8c1e47de"
592 |       },
593 |       "source": [
594 |         "hist = model.fit(train_x, train_y, batch_size=128, epochs=100,callbacks=[es,mc], validation_data=(val_x, val_y))"
595 |       ],
596 |       "execution_count": 46,
597 |       "outputs": [
598 |         {
599 |           "output_type": "stream",
600 |           "name": "stdout",
601 |           "text": [
602 |             "Epoch 1/100\n",
603 |             "\u001b[1m370/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - accuracy: 0.9961 - loss: 0.0117\n",
604 |             "Epoch 1: val_loss improved from inf to 0.12152, saving model to /content/weight/weights-001-0.1215.keras\n",
605 |             "\u001b[1m375/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 4ms/step - accuracy: 0.9961 - loss: 0.0117 - val_accuracy: 0.9772 - val_loss: 0.1215\n",
606 |             "Epoch 2/100\n",
607 |             "\u001b[1m351/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - accuracy: 0.9966 - loss: 0.0105\n",
608 |             "Epoch 2: val_loss improved from 0.12152 to 0.10669, saving model to /content/weight/weights-002-0.1067.keras\n",
609 |             "\u001b[1m375/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 3ms/step - accuracy: 0.9967 - loss: 0.0104 - val_accuracy: 0.9785 - val_loss: 0.1067\n",
610 |             "Epoch 3/100\n",
611 |             "\u001b[1m365/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - accuracy: 0.9983 - loss: 0.0057\n",
612 |             "Epoch 3: val_loss did not improve from 0.10669\n",
613 |             "\u001b[1m375/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 3ms/step - accuracy: 0.9983 - loss: 0.0058 - val_accuracy: 0.9784 - val_loss: 0.1165\n",
614 |             "Epoch 4/100\n",
615 |             "\u001b[1m366/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - accuracy: 0.9980 - loss: 0.0068\n",
616 |             "Epoch 4: val_loss did not improve from 0.10669\n",
617 |             "\u001b[1m375/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 3ms/step - accuracy: 0.9979 - loss: 0.0069 - val_accuracy: 0.9782 - val_loss: 0.1101\n",
618 |             "Epoch 5/100\n",
619 |             "\u001b[1m361/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - accuracy: 0.9979 - loss: 0.0083\n",
620 |             "Epoch 5: val_loss improved from 0.10669 to 0.09601, saving model to /content/weight/weights-005-0.0960.keras\n",
621 |             "\u001b[1m375/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 3ms/step - accuracy: 0.9979 - loss: 0.0083 - val_accuracy: 0.9804 - val_loss: 0.0960\n",
622 |             "Epoch 6/100\n",
623 |             "\u001b[1m364/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - accuracy: 0.9976 - loss: 0.0073\n",
624 |             "Epoch 6: val_loss did not improve from 0.09601\n",
625 |             "\u001b[1m375/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 3ms/step - accuracy: 0.9976 - loss: 0.0074 - val_accuracy: 0.9745 - val_loss: 0.1271\n",
626 |             "Epoch 7/100\n",
627 |             "\u001b[1m371/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - accuracy: 0.9975 - loss: 0.0085\n",
628 |             "Epoch 7: val_loss did not improve from 0.09601\n",
629 |             "\u001b[1m375/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 3ms/step - accuracy: 0.9975 - loss: 0.0085 - val_accuracy: 0.9781 - val_loss: 0.1210\n",
630 |             "Epoch 8/100\n",
631 |             "\u001b[1m356/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - accuracy: 0.9971 - loss: 0.0097\n",
632 |             "Epoch 8: val_loss did not improve from 0.09601\n",
633 |             "\u001b[1m375/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 3ms/step - accuracy: 0.9972 - loss: 0.0096 - val_accuracy: 0.9795 - val_loss: 0.1096\n",
634 |             "Epoch 9/100\n",
635 |             "\u001b[1m371/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━\u001b[0m \u001b[1m0s\u001b[0m 3ms/step - accuracy: 0.9978 - loss: 0.0064\n",
636 |             "Epoch 9: val_loss did not improve from 0.09601\n",
637 |             "\u001b[1m375/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 4ms/step - accuracy: 0.9978 - loss: 0.0065 - val_accuracy: 0.9790 - val_loss: 0.1161\n",
638 |             "Epoch 10/100\n",
639 |             "\u001b[1m352/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m━━\u001b[0m \u001b[1m0s\u001b[0m 2ms/step - accuracy: 0.9978 - loss: 0.0061\n",
640 |             "Epoch 10: val_loss did not improve from 0.09601\n",
641 |             "\u001b[1m375/375\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 3ms/step - accuracy: 0.9978 - loss: 0.0062 - val_accuracy: 0.9783 - val_loss: 0.1356\n"
642 |           ]
643 |         }
644 |       ]
645 |     },
646 |     {
647 |       "cell_type": "code",
648 |       "metadata": {
649 |         "id": "iVMEuC0-_3HB",
650 |         "colab": {
651 |           "base_uri": "https://localhost:8080/"
652 |         },
653 |         "outputId": "e8b97c0c-7e2a-425f-c624-d686e6c2b44d"
654 |       },
655 |       "source": [
656 |         "score = model.evaluate(X_test, Y_test)\n",
657 |         "print()\n",
658 |         "print('Test accuracy: ', score[1])"
659 |       ],
660 |       "execution_count": 17,
661 |       "outputs": [
662 |         {
663 |           "output_type": "stream",
664 |           "name": "stdout",
665 |           "text": [
666 |             "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m2s\u001b[0m 4ms/step - accuracy: 0.9754 - loss: 0.1180\n",
667 |             "\n",
668 |             "Test accuracy:  0.979200005531311\n"
669 |           ]
670 |         }
671 |       ]
672 |     },
673 |     {
674 |       "cell_type": "code",
675 |       "metadata": {
676 |         "id": "elo2tQ9C_3HE",
677 |         "colab": {
678 |           "base_uri": "https://localhost:8080/"
679 |         },
680 |         "outputId": "78eb9b90-8a8e-4fb1-f9eb-1d4b0714c966"
681 |       },
682 |       "source": [
683 |         "predictions = model.predict(X_test)\n",
684 |         "\n",
685 |         "predictions = list(predictions)\n",
686 |         "actuals = list(y_test)\n",
687 |         "\n",
688 |         "sub = pd.DataFrame({'Actual': actuals, 'Predictions': predictions})\n",
689 |         "sub.to_csv('output.csv', index=False)"
690 |       ],
691 |       "execution_count": 18,
692 |       "outputs": [
693 |         {
694 |           "output_type": "stream",
695 |           "name": "stdout",
696 |           "text": [
697 |             "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 2ms/step\n"
698 |           ]
699 |         }
700 |       ]
701 |     },
702 |     {
703 |       "cell_type": "code",
704 |       "source": [
705 |         "np.argmax(predictions[1])"
706 |       ],
707 |       "metadata": {
708 |         "id": "V7WVw-ylOE1d",
709 |         "outputId": "e30223b5-b097-44cc-e2db-02d55e593f94",
710 |         "colab": {
711 |           "base_uri": "https://localhost:8080/"
712 |         }
713 |       },
714 |       "execution_count": 19,
715 |       "outputs": [
716 |         {
717 |           "output_type": "execute_result",
718 |           "data": {
719 |             "text/plain": [
720 |               "2"
721 |             ]
722 |           },
723 |           "metadata": {},
724 |           "execution_count": 19
725 |         }
726 |       ]
727 |     },
728 |     {
729 |       "cell_type": "code",
730 |       "metadata": {
731 |         "id": "jd888zf-mDjy"
732 |       },
733 |       "source": [],
734 |       "execution_count": 19,
735 |       "outputs": []
736 |     },
737 |     {
738 |       "cell_type": "code",
739 |       "metadata": {
740 |         "id": "CiSKGn2jVhkK",
741 |         "colab": {
742 |           "base_uri": "https://localhost:8080/",
743 |           "height": 207
744 |         },
745 |         "outputId": "51999b98-8f3f-4fe3-cc40-4f361f1734e4"
746 |       },
747 |       "source": [
748 |         "plt.figure(figsize=(14,3))\n",
749 |         "plt.subplot(1, 2, 1)\n",
750 |         "plt.suptitle('Optimizer : Adam', fontsize=10)\n",
751 |         "plt.ylabel('Loss', fontsize=16)\n",
752 |         "plt.plot(hist.history['loss'], color='b', label='Training Loss')\n",
753 |         "plt.plot(hist.history['val_loss'], color='r', label='Validation Loss')\n",
754 |         "plt.legend(loc='upper right')\n",
755 |         "\n",
756 |         "plt.subplot(1, 2, 2)\n",
757 |         "plt.ylabel('Accuracy', fontsize=16)\n",
758 |         "plt.plot(hist.history['accuracy'], color='b', label='Training Accuracy')\n",
759 |         "plt.plot(hist.history['val_accuracy'], color='r', label='Validation Accuracy')\n",
760 |         "plt.legend(loc='lower right')\n",
761 |         "plt.show()"
762 |       ],
763 |       "execution_count": 20,
764 |       "outputs": [
765 |         {
766 |           "output_type": "display_data",
767 |           "data": {
768 |             "text/plain": [
769 |               "<Figure size 1400x300 with 2 Axes>"
770 |             ],
771 |             "image/png": "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\n"
772 |           },
773 |           "metadata": {}
774 |         }
775 |       ]
776 |     },
777 |     {
778 |       "cell_type": "code",
779 |       "metadata": {
780 |         "id": "yXNDej7bnSZz"
781 |       },
782 |       "source": [],
783 |       "execution_count": null,
784 |       "outputs": []
785 |     }
786 |   ]
787 | }


--------------------------------------------------------------------------------
/MNIST_keras_CNN.ipynb:
--------------------------------------------------------------------------------
  1 | {
  2 |   "nbformat": 4,
  3 |   "nbformat_minor": 0,
  4 |   "metadata": {
  5 |     "colab": {
  6 |       "name": "MNIST_keras_CNN-99.55%.ipynb",
  7 |       "provenance": [],
  8 |       "include_colab_link": true
  9 |     },
 10 |     "kernelspec": {
 11 |       "name": "python3",
 12 |       "display_name": "Python 3"
 13 |     },
 14 |     "accelerator": "GPU"
 15 |   },
 16 |   "cells": [
 17 |     {
 18 |       "cell_type": "markdown",
 19 |       "metadata": {
 20 |         "id": "view-in-github",
 21 |         "colab_type": "text"
 22 |       },
 23 |       "source": [
 24 |         "<a href=\"https://colab.research.google.com/github/mohanrajmit/DL-Training/blob/master/MNIST_keras_CNN.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
 25 |       ]
 26 |     },
 27 |     {
 28 |       "metadata": {
 29 |         "id": "VBPBbqSYId80"
 30 |       },
 31 |       "cell_type": "code",
 32 |       "source": [
 33 |         "import keras\n",
 34 |         "from keras.datasets import mnist\n",
 35 |         "from keras.models import Sequential\n",
 36 |         "from keras.layers import Dense, Dropout, Flatten\n",
 37 |         "from keras.layers import Conv2D, MaxPooling2D\n",
 38 |         "from keras import backend as K"
 39 |       ],
 40 |       "execution_count": 1,
 41 |       "outputs": []
 42 |     },
 43 |     {
 44 |       "metadata": {
 45 |         "id": "8Ot3_ZvXId9D"
 46 |       },
 47 |       "cell_type": "code",
 48 |       "source": [
 49 |         "batch_size = 128\n",
 50 |         "num_classes = 10\n",
 51 |         "epochs = 10\n",
 52 |         "\n",
 53 |         "# input image dimensions\n",
 54 |         "img_rows, img_cols = 28, 28\n"
 55 |       ],
 56 |       "execution_count": 2,
 57 |       "outputs": []
 58 |     },
 59 |     {
 60 |       "metadata": {
 61 |         "id": "mqD4qoB_Im51",
 62 |         "colab": {
 63 |           "base_uri": "https://localhost:8080/"
 64 |         },
 65 |         "outputId": "f3f60cfe-0f55-4ce3-d609-9efacff5f3b4"
 66 |       },
 67 |       "cell_type": "code",
 68 |       "source": [
 69 |         "# the data, split between train and test sets\n",
 70 |         "(x_train, y_train), (x_test, y_test) = mnist.load_data()\n",
 71 |         "\n",
 72 |         "if K.image_data_format() == 'channels_first':\n",
 73 |         "    x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols)\n",
 74 |         "    x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols)\n",
 75 |         "    input_shape = (1, img_rows, img_cols)\n",
 76 |         "else:\n",
 77 |         "    x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1)\n",
 78 |         "    x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1)\n",
 79 |         "    input_shape = (img_rows, img_cols, 1)\n"
 80 |       ],
 81 |       "execution_count": 3,
 82 |       "outputs": [
 83 |         {
 84 |           "output_type": "stream",
 85 |           "name": "stdout",
 86 |           "text": [
 87 |             "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
 88 |             "11490434/11490434 [==============================] - 0s 0us/step\n"
 89 |           ]
 90 |         }
 91 |       ]
 92 |     },
 93 |     {
 94 |       "cell_type": "code",
 95 |       "source": [
 96 |         "x_train.shape"
 97 |       ],
 98 |       "metadata": {
 99 |         "id": "Kxu9LpKZieH9",
100 |         "outputId": "57077d71-a863-4520-e3e8-e03dc0ebdc3b",
101 |         "colab": {
102 |           "base_uri": "https://localhost:8080/"
103 |         }
104 |       },
105 |       "execution_count": 4,
106 |       "outputs": [
107 |         {
108 |           "output_type": "execute_result",
109 |           "data": {
110 |             "text/plain": [
111 |               "(60000, 28, 28, 1)"
112 |             ]
113 |           },
114 |           "metadata": {},
115 |           "execution_count": 4
116 |         }
117 |       ]
118 |     },
119 |     {
120 |       "cell_type": "code",
121 |       "source": [
122 |         "x_test.shape"
123 |       ],
124 |       "metadata": {
125 |         "id": "67DPoINvig3K",
126 |         "outputId": "0daa24f2-f51d-4d51-f7a1-d647e4b0e896",
127 |         "colab": {
128 |           "base_uri": "https://localhost:8080/"
129 |         }
130 |       },
131 |       "execution_count": 5,
132 |       "outputs": [
133 |         {
134 |           "output_type": "execute_result",
135 |           "data": {
136 |             "text/plain": [
137 |               "(10000, 28, 28, 1)"
138 |             ]
139 |           },
140 |           "metadata": {},
141 |           "execution_count": 5
142 |         }
143 |       ]
144 |     },
145 |     {
146 |       "metadata": {
147 |         "id": "Nrpi7I6lId9V",
148 |         "colab": {
149 |           "base_uri": "https://localhost:8080/"
150 |         },
151 |         "outputId": "801306e4-908f-4bb1-fd22-85613c12ad11"
152 |       },
153 |       "cell_type": "code",
154 |       "source": [
155 |         "x_train = x_train.astype('float32')\n",
156 |         "x_test = x_test.astype('float32')\n",
157 |         "x_train /= 255\n",
158 |         "x_test /= 255\n",
159 |         "print('x_train shape:', x_train.shape)\n",
160 |         "print(x_train.shape[0], 'train samples')\n",
161 |         "print(x_test.shape[0], 'test samples')"
162 |       ],
163 |       "execution_count": 6,
164 |       "outputs": [
165 |         {
166 |           "output_type": "stream",
167 |           "name": "stdout",
168 |           "text": [
169 |             "x_train shape: (60000, 28, 28, 1)\n",
170 |             "60000 train samples\n",
171 |             "10000 test samples\n"
172 |           ]
173 |         }
174 |       ]
175 |     },
176 |     {
177 |       "cell_type": "code",
178 |       "source": [
179 |         "from keras.utils import to_categorical"
180 |       ],
181 |       "metadata": {
182 |         "id": "PtXbgBtsc0YP"
183 |       },
184 |       "execution_count": 8,
185 |       "outputs": []
186 |     },
187 |     {
188 |       "metadata": {
189 |         "scrolled": false,
190 |         "id": "sb95fYcVId9g"
191 |       },
192 |       "cell_type": "code",
193 |       "source": [
194 |         "# convert class vectors to binary class matrices\n",
195 |         "y_train = to_categorical(y_train, num_classes)\n",
196 |         "y_test = to_categorical(y_test, num_classes)"
197 |       ],
198 |       "execution_count": 9,
199 |       "outputs": []
200 |     },
201 |     {
202 |       "cell_type": "code",
203 |       "source": [
204 |         "y_train[100]"
205 |       ],
206 |       "metadata": {
207 |         "id": "O6n8VwiRc_YM",
208 |         "outputId": "0442645d-cdaf-4328-998a-24da1d2921a8",
209 |         "colab": {
210 |           "base_uri": "https://localhost:8080/"
211 |         }
212 |       },
213 |       "execution_count": 10,
214 |       "outputs": [
215 |         {
216 |           "output_type": "execute_result",
217 |           "data": {
218 |             "text/plain": [
219 |               "array([0., 0., 0., 0., 0., 1., 0., 0., 0., 0.], dtype=float32)"
220 |             ]
221 |           },
222 |           "metadata": {},
223 |           "execution_count": 10
224 |         }
225 |       ]
226 |     },
227 |     {
228 |       "metadata": {
229 |         "id": "LyHYFqR0Id9q"
230 |       },
231 |       "cell_type": "code",
232 |       "source": [
233 |         "model = Sequential()\n",
234 |         "model.add(Conv2D(32, kernel_size=(5, 5),\n",
235 |         "                 activation='relu',\n",
236 |         "                 input_shape=input_shape))\n",
237 |         "model.add(Conv2D(64, (3, 3),padding='same' ,activation='relu'))\n",
238 |         "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
239 |         "model.add(Conv2D(128, (3, 3),padding='same' ,activation='relu'))\n",
240 |         "model.add(MaxPooling2D(pool_size=(2, 2)))\n",
241 |         "#model.add(Dropout(0.25))\n",
242 |         "model.add(Flatten())\n",
243 |         "model.add(Dense(128, activation='relu'))\n",
244 |         "#model.add(Dropout(0.5))\n",
245 |         "model.add(Dense(num_classes, activation='softmax'))\n",
246 |         "\n"
247 |       ],
248 |       "execution_count": 31,
249 |       "outputs": []
250 |     },
251 |     {
252 |       "metadata": {
253 |         "id": "DfJk083kId95",
254 |         "colab": {
255 |           "base_uri": "https://localhost:8080/"
256 |         },
257 |         "outputId": "f79d59d1-3f6a-4789-fd97-a53c25a8b3b3"
258 |       },
259 |       "cell_type": "code",
260 |       "source": [
261 |         "model.summary()"
262 |       ],
263 |       "execution_count": 32,
264 |       "outputs": [
265 |         {
266 |           "output_type": "stream",
267 |           "name": "stdout",
268 |           "text": [
269 |             "Model: \"sequential_3\"\n",
270 |             "_________________________________________________________________\n",
271 |             " Layer (type)                Output Shape              Param #   \n",
272 |             "=================================================================\n",
273 |             " conv2d_7 (Conv2D)           (None, 24, 24, 32)        832       \n",
274 |             "                                                                 \n",
275 |             " conv2d_8 (Conv2D)           (None, 24, 24, 64)        18496     \n",
276 |             "                                                                 \n",
277 |             " max_pooling2d_4 (MaxPoolin  (None, 12, 12, 64)        0         \n",
278 |             " g2D)                                                            \n",
279 |             "                                                                 \n",
280 |             " conv2d_9 (Conv2D)           (None, 12, 12, 128)       73856     \n",
281 |             "                                                                 \n",
282 |             " max_pooling2d_5 (MaxPoolin  (None, 6, 6, 128)         0         \n",
283 |             " g2D)                                                            \n",
284 |             "                                                                 \n",
285 |             " flatten_3 (Flatten)         (None, 4608)              0         \n",
286 |             "                                                                 \n",
287 |             " dense_6 (Dense)             (None, 128)               589952    \n",
288 |             "                                                                 \n",
289 |             " dense_7 (Dense)             (None, 10)                1290      \n",
290 |             "                                                                 \n",
291 |             "=================================================================\n",
292 |             "Total params: 684426 (2.61 MB)\n",
293 |             "Trainable params: 684426 (2.61 MB)\n",
294 |             "Non-trainable params: 0 (0.00 Byte)\n",
295 |             "_________________________________________________________________\n"
296 |           ]
297 |         }
298 |       ]
299 |     },
300 |     {
301 |       "metadata": {
302 |         "id": "iS6_SY_SId-B"
303 |       },
304 |       "cell_type": "code",
305 |       "source": [
306 |         "model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['accuracy'])"
307 |       ],
308 |       "execution_count": 33,
309 |       "outputs": []
310 |     },
311 |     {
312 |       "metadata": {
313 |         "id": "I0CqfWbMId-J",
314 |         "colab": {
315 |           "base_uri": "https://localhost:8080/"
316 |         },
317 |         "outputId": "b3b15099-e73c-4f0f-c061-eb6f9d9febff"
318 |       },
319 |       "cell_type": "code",
320 |       "source": [
321 |         "hist = model.fit(x_train, y_train,\n",
322 |         "          batch_size=batch_size,\n",
323 |         "          epochs=epochs,\n",
324 |         "          verbose=1,\n",
325 |         "          validation_data=(x_test, y_test))"
326 |       ],
327 |       "execution_count": 34,
328 |       "outputs": [
329 |         {
330 |           "output_type": "stream",
331 |           "name": "stdout",
332 |           "text": [
333 |             "Epoch 1/10\n",
334 |             "469/469 [==============================] - 9s 13ms/step - loss: 0.1599 - accuracy: 0.9500 - val_loss: 0.0404 - val_accuracy: 0.9871\n",
335 |             "Epoch 2/10\n",
336 |             "469/469 [==============================] - 5s 10ms/step - loss: 0.0420 - accuracy: 0.9872 - val_loss: 0.0348 - val_accuracy: 0.9896\n",
337 |             "Epoch 3/10\n",
338 |             "469/469 [==============================] - 6s 12ms/step - loss: 0.0286 - accuracy: 0.9911 - val_loss: 0.0252 - val_accuracy: 0.9916\n",
339 |             "Epoch 4/10\n",
340 |             "469/469 [==============================] - 5s 11ms/step - loss: 0.0216 - accuracy: 0.9930 - val_loss: 0.0291 - val_accuracy: 0.9898\n",
341 |             "Epoch 5/10\n",
342 |             "469/469 [==============================] - 5s 11ms/step - loss: 0.0159 - accuracy: 0.9946 - val_loss: 0.0298 - val_accuracy: 0.9916\n",
343 |             "Epoch 6/10\n",
344 |             "469/469 [==============================] - 5s 11ms/step - loss: 0.0120 - accuracy: 0.9960 - val_loss: 0.0265 - val_accuracy: 0.9918\n",
345 |             "Epoch 7/10\n",
346 |             "469/469 [==============================] - 5s 10ms/step - loss: 0.0104 - accuracy: 0.9968 - val_loss: 0.0253 - val_accuracy: 0.9918\n",
347 |             "Epoch 8/10\n",
348 |             "469/469 [==============================] - 5s 11ms/step - loss: 0.0093 - accuracy: 0.9970 - val_loss: 0.0243 - val_accuracy: 0.9927\n",
349 |             "Epoch 9/10\n",
350 |             "469/469 [==============================] - 5s 11ms/step - loss: 0.0071 - accuracy: 0.9976 - val_loss: 0.0298 - val_accuracy: 0.9914\n",
351 |             "Epoch 10/10\n",
352 |             "469/469 [==============================] - 5s 11ms/step - loss: 0.0085 - accuracy: 0.9972 - val_loss: 0.0315 - val_accuracy: 0.9920\n"
353 |           ]
354 |         }
355 |       ]
356 |     },
357 |     {
358 |       "metadata": {
359 |         "id": "1Reuwv4kId-P",
360 |         "colab": {
361 |           "base_uri": "https://localhost:8080/"
362 |         },
363 |         "outputId": "0088b28f-7085-4145-93df-3077e1f5ac8c"
364 |       },
365 |       "cell_type": "code",
366 |       "source": [
367 |         "score = model.evaluate(x_test, y_test, verbose=0)\n",
368 |         "print('Test loss:', score[0])\n",
369 |         "print('Test accuracy:', score[1])"
370 |       ],
371 |       "execution_count": 28,
372 |       "outputs": [
373 |         {
374 |           "output_type": "stream",
375 |           "name": "stdout",
376 |           "text": [
377 |             "Test loss: 0.030728774145245552\n",
378 |             "Test accuracy: 0.9923999905586243\n"
379 |           ]
380 |         }
381 |       ]
382 |     },
383 |     {
384 |       "cell_type": "code",
385 |       "source": [
386 |         "import matplotlib.pyplot as plt\n"
387 |       ],
388 |       "metadata": {
389 |         "id": "6kq1Lv-3lPz2"
390 |       },
391 |       "execution_count": 29,
392 |       "outputs": []
393 |     },
394 |     {
395 |       "cell_type": "code",
396 |       "source": [
397 |         "plt.figure(figsize=(14,3))\n",
398 |         "plt.subplot(1, 2, 1)\n",
399 |         "plt.suptitle('Optimizer : Adam', fontsize=10)\n",
400 |         "plt.ylabel('Loss', fontsize=16)\n",
401 |         "plt.plot(hist.history['loss'], color='b', label='Training Loss')\n",
402 |         "plt.plot(hist.history['val_loss'], color='r', label='Validation Loss')\n",
403 |         "plt.legend(loc='upper right')\n",
404 |         "\n",
405 |         "plt.subplot(1, 2, 2)\n",
406 |         "plt.ylabel('Accuracy', fontsize=16)\n",
407 |         "plt.plot(hist.history['accuracy'], color='b', label='Training Accuracy')\n",
408 |         "plt.plot(hist.history['val_accuracy'], color='r', label='Validation Accuracy')\n",
409 |         "plt.legend(loc='lower right')\n",
410 |         "plt.show()"
411 |       ],
412 |       "metadata": {
413 |         "id": "PUn1HMfUlMj8",
414 |         "outputId": "824cb838-9435-45fe-bd52-c172c046d4b0",
415 |         "colab": {
416 |           "base_uri": "https://localhost:8080/",
417 |           "height": 303
418 |         }
419 |       },
420 |       "execution_count": 30,
421 |       "outputs": [
422 |         {
423 |           "output_type": "display_data",
424 |           "data": {
425 |             "text/plain": [
426 |               "<Figure size 1400x300 with 2 Axes>"
427 |             ],
428 |             "image/png": "iVBORw0KGgoAAAANSUhEUgAABJsAAAEwCAYAAAAQFHCXAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAACizElEQVR4nOzdeViUZfvG8e+AAiKLOyiiKGporrmllZpZmCZulZnlkln5qrm8lVpmar80c8m1LF+XNpcsd80lct/TrDQ1M3fBLQVX1vn9cccgAgqyzADn5zieY4aZh2euGdB5OOe+r9titVqtiIiIiIiIiIiIZAInexcgIiIiIiIiIiK5h8ImERERERERERHJNAqbREREREREREQk0yhsEhERERERERGRTKOwSUREREREREREMo3CJhERERERERERyTQKm0REREREREREJNMobBIRERERERERkUyjsElERERERERERDKNwiYRERHJU4YNG0bNmjUzdIxjx45hsVjYu3dvptTkKCwWC4sXL7Z3GSIiIpLDKWwSERERh3Ly5EleeuklSpUqhYuLC2XLlqVv375cvHgx3cdKKTx54403CA0NzVCN/v7+hIWFUbVq1QwdJ6sFBwfj7OzMrl277F2KiIiI5CEKm0RERMRh/P3339SpU4fDhw8zd+5c/vrrL6ZNm0ZoaCgNGjTgn3/+yfBjeHh4ULRo0Qwdw9nZGV9fX/Lly5fhelISFxdHfHx8ho5x4sQJtm7dSu/evZk5c2YmVSYiIiJydwqbRERExGH06tULFxcX1qxZQ+PGjSlTpgxPPvkkP/74I6dPn+add96x7RsQEMD7779Px44dKViwIH5+fkydOjXJ/QBt27bFYrHYvr59Gl3Xrl1p06YNI0eOxMfHh0KFCjFixAhiY2N58803KVKkCKVLl2bWrFm277l9Gl3Xrl2xWCzJtvXr1wMQFRXFG2+8gZ+fHwULFqR+/fq2+wBmz55NoUKFWLp0KVWqVMHV1ZUTJ05k6LWcNWsWTz31FD179mTu3LncuHEjyf2HDx+mUaNGuLm5UaVKFdauXZvsGAMHDqRSpUq4u7tTvnx53n33XWJiYmz3J7yWM2fOpEyZMnh4ePCf//yHuLg4PvroI3x9fSlRogQffPBBhp6LiIiI5CwKm0RERMQh/PPPP6xevZr//Oc/FChQIMl9vr6+dOrUifnz52O1Wm23jxkzhho1avDLL78waNAg+vbtawtNEqaOzZo1i7CwsDtOJfvpp584c+YMGzduZPz48bz33ns89dRTFC5cmB07dvDaa6/x6quvcurUqRS/f+LEiYSFhdm2vn37UqJECYKCggDo3bs327ZtY968efz2228888wzNG/enMOHD9uOcf36dUaPHs3//vc/9u/fT4kSJVJ8rICAAIYNG3bH19JqtTJr1ixeeOEFgoKCqFChAt99953t/vj4eNq1a4eLiws7duxg2rRpDBw4MNlxPD09mT17Nn/88QcTJ05k+vTpfPzxx0n2OXLkCD/88AOrVq1i7ty5zJgxg5YtW3Lq1Ck2bNjA6NGjGTJkCDt27LhjzSIiIpJ7ZM3YbxEREZF0Onz4MFarlcqVK6d4f+XKlbl06RLnz5+3BTEPPfQQgwYNAqBSpUps2bKFjz/+mMcff5zixYsDUKhQIXx9fe/42EWKFGHSpEk4OTlx33338dFHH3H9+nXefvttAAYPHsyHH37I5s2bee6555J9v7e3N97e3gAsXLiQzz77jB9//BFfX19OnDjBrFmzOHHiBKVKlQJM36hVq1Yxa9YsRo4cCUBMTAyffPIJNWrUuGOtgYGBFCtW7I77/Pjjj1y/fp3g4GAAXnjhBWbMmMGLL75ou//gwYOsXr3aVtPIkSN58sknkxxnyJAhtusBAQG88cYbzJs3j7feest2e3x8PDNnzsTT05MqVarw6KOPcujQIVauXGl7PUePHs26deuoX7/+HesWERGR3EFhk4iIiDiUW0cu3U2DBg2SfT1hwoR0P+b999+Pk1PigG8fH58kzb+dnZ0pWrQo586du+NxfvnlF1588UWmTJnCQw89BMDvv/9OXFwclSpVSrJvVFRUkt5RLi4uVK9e/a61pqW5+cyZM+nQoYOtp1THjh158803OXLkCIGBgRw4cAB/f39b0ATJX0uA+fPnM2nSJI4cOcLVq1eJjY3Fy8sryT4BAQF4enravvbx8cHZ2TnZ63m3105ERERyD4VNIiIi4hAqVKiAxWLhwIEDtG3bNtn9Bw4coHDhwrYRS5kpf/78Sb62WCwp3nanpt3h4eGEhITw8ssv0717d9vtV69exdnZmd27d+Ps7Jzkezw8PGzXCxQogMViycjTAMx0xEWLFhETE8Onn35quz0uLo6ZM2emuX/Stm3b6NSpE8OHDyc4OBhvb2/mzZvHuHHjkuyXGa+diIiI5C4Km0RERMQhFC1alMcff5xPPvmE/v37J+nbFB4ezjfffEPnzp2TBDLbt29Pcozt27cnmYaXP39+4uLisrz2mzdv0rp1a4KCghg/fnyS+2rVqkVcXBznzp3jkUceyfJavvnmG0qXLs3ixYuT3L5mzRrGjRvHiBEjqFy5MidPniQsLIySJUsCyV/LrVu3UrZs2SRN2Y8fP57l9YuIiEjOpwbhIiIi4jCmTJlCVFQUwcHBbNy4kZMnT7Jq1Soef/xx/Pz8ko3K2bJlCx999BF//vknU6dOZcGCBfTt29d2f0BAAKGhoYSHh3Pp0qUsq/vVV1/l5MmTTJo0ifPnzxMeHk54eDjR0dFUqlSJTp060blzZxYuXMjRo0fZuXMno0aNYsWKFel+rMcee4wpU6akev+MGTN4+umnqVq1apKte/fuXLhwgVWrVtGsWTMqVapEly5d+PXXX9m0aVOSUAmgYsWKnDhxgnnz5nHkyBEmTZrEokWL0l2viIiI5D0Km0RERMRhVKxYkZ9//pny5cvz7LPPEhgYyCuvvMKjjz7Ktm3bKFKkSJL9//vf//Lzzz9Tq1Yt/u///o/x48fbmmIDjBs3jrVr1+Lv70+tWrWyrO4NGzYQFhZGlSpVKFmypG3bunUrYFbE69y5M//973+57777aNOmDbt27aJMmTLpfqwjR45w4cKFFO/bvXs3v/76K+3bt092n7e3N4899hgzZszAycmJRYsWcePGDerVq8fLL7+cLMgLCQmhf//+9O7dm5o1a7J161befffddNcrIiIieY/Fmp4unCIiIiIOIiAggH79+tGvXz97lyIiIiIit9DIJhERERERERERyTQKm0REREREREREJNNoGp2IiIiIiIiIiGQajWwSEREREREREZFMo7BJREREREREREQyjcImERERERERERHJNAqbREREREREREQk0yhsEhERERERERGRTKOwSUREREREREREMo3CJhERERERERERyTQKm0REREREREREJNMobBIRERERERERkUyjsElERERERERERDKNwiYREREREREREck0CptERERERERERCTTKGwSEREREREREZFMk8/eBeQ28fHxnDlzBk9PTywWi73LERERkVRYrVauXLlCqVKlcHLS52/2pPMnERGRnCGt508KmzLZmTNn8Pf3t3cZIiIikkYnT56kdOnS9i4jT9P5k4iISM5yt/MnhU2ZzNPTEzAvvJeXl52rERERkdRERkbi7+9ve+8W+9H5k4iISM6Q1vMnhU2ZLGHot5eXl06WREREcgBN27I/nT+JiIjkLHc7f1KDAhERERERERERyTQKm0REREREREREJNMobBIRERERERERkUyjnk0iImJ38fHxREdH27sMyWXy58+Ps7OzvcsQERERyXMUNomIiF1FR0dz9OhR4uPj7V2K5EKFChXC19dXTcBFREREspHCJhERsRur1UpYWBjOzs74+/vj5KTZ3ZI5rFYr169f59y5cwCULFnSzhWJiIiI5B0Km3KI06dh9Wro1g304ayI5BaxsbFcv36dUqVK4e7ubu9yJJcpUKAAAOfOnaNEiRKaUiciIiI2cXFw7Vry7eZNcHGBAgXA3d1c3no9n1KUNNHLlANER0NQEFy9CjVqQO3a9q5IRCRzxMXFAeDi4mLnSiS3SggxY2JiFDaJiIjkIFYrREWlHAglbNev3/n+O+0bFXVvdeXLl3oQldLX97Jfwtdubjl3sInCphzAxQWeeAIWLoSlSxU2iUjuo346klX0uyUiIpK1YmIgIgIiI80AicwMhbKjpafFAgULJm5ubmbAx40bZrt+3Yx2ShAbC1eumC07uLnde2D1+ONQs2b21Hk7hU05REhIYtg0fLi9qxEREREREZGcLD7ehEORkYlhUURE0uupXd56/dYgJqu4uCQNhBI2d/eUb0/PvmkZPRQfb0ZCJYRPtwZRCdfvdF9a90v4OjY28bFv3rz31/jTTxU2JTN16lTGjBlDeHg4NWrUYPLkydSrVy/Ffffv38/QoUPZvXs3x48f5+OPP6Zfv35J9hk2bBjDb0tp7rvvPg4ePGj7+ubNm/z3v/9l3rx5REVFERwczCeffIKPj0+mP7/0atECnJxg7144cQLKlLF3RSIiIiIiImIPN2+mLRS6U4B05YqZqpZZ3N3BwyNrAiF790lyckocLVSkSNY/XmzsvYVUt18PCsr6WlPjkGHT/PnzGTBgANOmTaN+/fpMmDCB4OBgDh06RIkSJZLtf/36dcqXL88zzzxD//79Uz3u/fffz48//mj7Ot9tv7H9+/dnxYoVLFiwAG9vb3r37k27du3YsmVL5j25e1S8ODRsCJs3w7Jl0KuXvSsSEZHMFBAQQL9+/ZJ9WJKa9evX8+ijj3Lp0iUKFSqUpbWJiIhI5rp2Dc6fhwsXzOXly+kbXRQdnXm15MsH3t5m8/JK32XCdS8v+wdCuUm+fODpabacyiF/HcaPH0+PHj3o1q0bANOmTWPFihXMnDmTQYMGJdu/bt261K1bFyDF+xPky5cPX1/fFO+LiIhgxowZzJkzh6ZNmwIwa9YsKleuzPbt23nwwQcz+rQyrHVrEzYtXaqwSUTEXu7WA+i9995j2LBh6T7url27KFiwYJr3b9iwIWFhYXh7e6f7sdJDoZaIiMidxcfDP/8kDY8SLlO67cIFM+okM3h63jkESstlTm5CLY7L4cKm6Ohodu/ezeDBg223OTk50axZM7Zt25ahYx8+fJhSpUrh5uZGgwYNGDVqFGX+nY+2e/duYmJiaNasmW3/oKAgypQpw7Zt21INm6Kiooi6pY19ZGRkhmq8k5AQePNNWLfOpNleXln2UCIikoqwsDDb9fnz5zN06FAOHTpku83Dw8N23Wq1EhcXl2wkbUqKFy+erjpcXFxS/QBFRERE7t3Nm3cPjG69/s8/99bI2tXVzGApVsxMzbrb6KHbLz09zfQuEUfkcGHThQsXiIuLS9YnycfHJ0l/pfSqX78+s2fP5r777iMsLIzhw4fzyCOPsG/fPjw9PQkPD8fFxSXZp7Y+Pj6Eh4enetxRo0Yl6wWVVSpVgvvug0OHYNUqePbZbHlYEZFsY7WaOeb24O6etk/1bg14vL29sVgsttsSRgGtXLmSIUOG8Pvvv7NmzRr8/f0ZMGAA27dv59q1a1SuXJlRo0Yl+YDj9ml0FouF6dOns2LFClavXo2fnx/jxo0jJCQkyWMljDiaPXs2/fr1Y/78+fTr14+TJ0/y8MMPM2vWLEqWLAlAbGwsAwYM4Msvv8TZ2ZmXX36Z8PBwIiIiWLx48T29bpcuXaJv374sW7aMqKgoGjduzKRJk6hYsSIAx48fp3fv3mzevJno6GgCAgIYM2YMLVq04NKlS/Tu3Zs1a9Zw9epVSpcuzdtvv20b2SwiIpJRVquZopbWUUfnz5spbveiUKHE8OjWy5RuK1bM9DfSiCLJrRwubMoqTz75pO169erVqV+/PmXLluXbb7+le/fu93zcwYMHM2DAANvXkZGR+Pv7Z6jWOwkJgTFjzFQ6hU0ikttcv25OvOzh6lXTgDIzDBo0iLFjx1K+fHkKFy7MyZMnadGiBR988AGurq58+eWXtGrVikOHDtlG2KZk+PDhfPTRR4wZM4bJkyfTqVMnjh8/TpFUOlNev36dsWPH8tVXX+Hk5MQLL7zAG2+8wTfffAPA6NGj+eabb2zTxCdOnMjixYt59NFH7/m5du3alcOHD7N06VK8vLwYOHAgLVq04I8//iB//vz06tWL6OhoNm7cSMGCBfnjjz9so7/effdd/vjjD3744QeKFSvGX3/9xY3MmlcgIiK5TlSU6Vl0+XJij6OLF+8eIsXFpf+x8uW7e2B06/WiRSF//sx+xiI5l8OFTcWKFcPZ2ZmzZ88muf3s2bOZOl2gUKFCVKpUib/++gswn1RHR0dz+fLlJKOb7va4rq6uuLq6Zlpdd5MQNq1YATEx+g9NRMQRjRgxgscff9z2dZEiRahRo4bt6/fff59FixaxdOlSevfunepxunbtSseOHQEYOXIkkyZNYufOnTRv3jzF/WNiYpg2bRqBgYEA9O7dmxEjRtjunzx5MoMHD6Zt27YATJkyhZUrV97z80wImbZs2ULDhg0B+Oabb/D392fx4sU888wznDhxgvbt21OtWjUAypcvb/v+EydOUKtWLerUqQOY0V0iIpI7Wa2mT1FCUJQQFqXn63td/h3MlLO0jjoqXtxMU9OoI5F753Bhk4uLC7Vr1yY0NJQ2bdoAEB8fT2ho6B1PyNPr6tWrHDlyhBdffBGA2rVrkz9/fkJDQ2nfvj0Ahw4d4sSJEzRo0CDTHjejGjQwqfnFi7BlCzRpYu+KREQyj7u7GWFkr8fOLAnhSYKrV68ybNgwVqxYQVhYGLGxsdy4cYMTJ07c8TjVq1e3XS9YsCBeXl6cO3cu1f3d3d1tQRNAyZIlbftHRERw9uxZ6tWrZ7vf2dmZ2rVrE38vjSaAAwcOkC9fPurXr2+7rWjRotx3330cOHAAgNdff52ePXuyZs0amjVrRvv27W3Pq2fPnrRv3549e/bwxBNP0KZNG1toJSIijiU+3rxHZyQsio3NnFq8vMyUtUKFoHDhtI06cnPLnMcWkbRxuLAJYMCAAXTp0oU6depQr149JkyYwLVr12w9HDp37oyfnx+jRo0CTFPxP/74w3b99OnT7N27Fw8PDypUqADAG2+8QatWrShbtixnzpzhvffew9nZ2faJsbe3N927d2fAgAEUKVIELy8v+vTpQ4MGDRxiJboEzs7w1FPwxRdmKp3CJhHJTSyWzJvKZk+3ryr3xhtvsHbtWsaOHUuFChUoUKAATz/9NNF3Wbc4/23DVy0Wyx2DoZT2t1qt6aw+c7388ssEBwezYsUK1qxZw6hRoxg3bhx9+vThySef5Pjx46xcuZK1a9fy2GOP0atXL8aOHWvXmkVEciur1QRAZ87A2bNJg6C7hUUREeb7M8rZ2YRE3t6JgdHtX9/pPk9PcwwRcWwOGTZ16NCB8+fPM3ToUMLDw6lZsyarVq2yNQ0/ceIETre03T9z5gy1atWyfT127FjGjh1L48aNWb9+PQCnTp2iY8eOXLx4keLFi/Pwww+zffv2JKv/fPzxxzg5OdG+fXuioqIIDg7mk08+yZ4nnQ6tWyeGTePGaXiniIij27JlC127drVNX7t69SrHjh3L1hq8vb3x8fFh165dNGrUCIC4uDj27NlDzZo17+mYlStXJjY2lh07dthGJF28eJFDhw5RpUoV237+/v689tprvPbaawwePJjp06fTp08fwKzC16VLF7p06cIjjzzCm2++qbBJROQeREVBWBicPm3CpNOnU76e0YU4XFzMaKI7BUJ3+rpgQf39IpIXOGTYBKbPRGrT5hICpAQBAQF3/eR23rx5d31MNzc3pk6dytSpU9Ncpz08/rhZJvPIEThwAG45nxcREQdUsWJFFi5cSKtWrbBYLLz77rv3PHUtI/r06cOoUaOoUKECQUFBTJ48mUuXLmFJw1n/77//jqenp+1ri8VCjRo1aN26NT169OCzzz7D09OTQYMG4efnR+vWrQHo168fTz75JJUqVeLSpUusW7eOypUrAzB06FBq167N/fffT1RUFMuXL7fdJyIiRny8aXKdEBqlFiRduJD2YxYuDL6+5jK9YZGmo4lIWjhs2CSp8/CAxx6DlSthyRKFTSIijm78+PG89NJLNGzYkGLFijFw4EAiIyOzvY6BAwcSHh5O586dcXZ25pVXXiE4OBjnNMxHSBgNlcDZ2ZnY2FhmzZpF3759eeqpp4iOjqZRo0asXLnSNqUvLi6OXr16cerUKby8vGjevDkff/wxYPo0Dh48mGPHjlGgQAEeeeSRNH04JCKSW1y7dveRSGfOmIWB0sLVFUqVAj8/s6V0vVQpKFAga5+XiIjFau9mDrlMZGQk3t7eRERE4OXllWWP89ln8Npr8OCDsG1blj2MiEiWunnzJkePHqVcuXK46aPSbBcfH0/lypV59tlnef/99+1dTpa40+9Ydr1ny93pZyG5TWys6Yl0tyApIiJtx7NYoESJuwdJRYpoipqIZK20vmdrZFMO9dRT5nLHDggPN8NgRURE7uT48eOsWbOGxo0bExUVxZQpUzh69CjPP/+8vUsTEckxLl9ODIxSC5POnjXT39LCwyP18CjhesmScNsaECIiDk1hUw7l5wd16sDPP8OKFdC9u70rEhERR+fk5MTs2bN54403sFqtVK1alR9//FF9kkRE7uDYMdi4MXE7fDht3+fsbEKiW4OjlIIkDeYTkdxIYVMOFhJiwqalSxU2iYjI3fn7+7NlyxZ7lyEi4rCsVvjzz6Th0okTyfcrXDjlXki33laihAmcRETyIoVNOVjr1jB0KKxda5YwdXe3d0UiIiIiIjlHfDzs25c0XDp7Nuk+zs5mRkGjRtC4MTRsaMImERFJncKmHKxaNShbFo4fh9BQaNXK3hWJiIiIiDiu2Fj45RcTKm3YAJs3w6VLSfdxdYX69U2w1KiRWZDHw8M+9YqI5FQKm3Iwi8VMpZs8GZYsUdgkIiIiInKrqCjYtcsESxs3wtatcPVq0n0KFoSHHjLBUqNGULcuaIFUEZGMUdiUwyWETcuWmWHATk72rkhERERExD6uXYPt2xPDpe3bTeB0q0KF4JFHEsOlWrW00puISGZT2JTDNWpkVrA4dw527jTDfEVERERE8oKICDMVLqHf0s8/m6lytypRIjFYatwYqlbVB7QiIllNYVMO5+ICTz4J8+ebVekUNomI5AxNmjShZs2aTJgwAYCAgAD69etHv379Uv0ei8XCokWLaNOmTYYeO7OOIyKS3c6fh02bEsOlvXvNCnK3Kl3ahEoJPZcqVTLtJ+RfVqt58bZtM8vmlS0LAQFmGb18+vNQ0slqhTNnTDO0ffvMPzYvL/D2TvnSy0u/Z9nBajWriOXLZxrR2YF+yrlASEhi2DRypL2rERHJ3Vq1akVMTAyrVq1Kdt+mTZto1KgRv/76K9WrV0/XcXft2kXBggUzq0wAhg0bxuLFi9m7d2+S28PCwiicxUspzZ49m379+nH58uUsfRwRyd1On066UtwffyTfp0KFxGCpUSOTnShcSoHVCqtWwQcfwJYtye93djZJXUL4dPulv7/5pFvyrrg4+PNPk/L+8kvi5YUL6TuOu3vKYVRqAVVKlwUL5p5/6PHxZg7w1atJtytXkt+W1n2uXTP/5mfOhG7d7PK0FDblAk8+aQLL/fvhyBEIDLR3RSIiuVf37t1p3749p06donTp0knumzVrFnXq1El30ARQvHjxzCrxrnx9fbPtsURE0spqhaNHE4OlDRvg77+T71e1amKw9MgjZnCO3EF8PCxebEKmPXvMba6u0LKlmYd4/LjZYmISr2/cmPw4FkvSkVC3B1JlykCBAtn3vCRrXb8Ov/9uAqWEUOm33+DGjeT7OjtD5cpQvboJJCMiIDIy+WXC916/brawsHuvz8kpcaTUvQZWXl7pH/UTG5tyMJSRgOj69Xt/He7m9hURspHCplygcGHzZvvTT6ZR+B1mYIiISAY99dRTFC9enNmzZzNkyBDb7VevXmXBggWMGTOGixcv0rt3bzZu3MilS5cIDAzk7bffpmPHjqke9/ZpdIcPH6Z79+7s3LmT8uXLM3HixGTfM3DgQBYtWsSpU6fw9fWlU6dODB06lPz58zN79myGDx8OmGlzYMKwrl27JptG9/vvv9O3b1+2bduGu7s77du3Z/z48Xj8u9Z3165duXz5Mg8//DDjxo0jOjqa5557jgkTJpD/Hrvqnjhxgj59+hAaGoqTkxPNmzdn8uTJ+Pj4APDrr7/Sr18/fv75ZywWCxUrVuSzzz6jTp06HD9+nN69e7N582aio6MJCAhgzJgxtGjR4p5qERH7sFrh4MHEYGnjRjOS6VZOTqaBd0K49PDDUKyYferNcWJjzfSHkSMTh4S5u0PPnvDf/0LJkon7xsdDeDgcO2bCppQub940P6DTp82yfinx8Ul5VFTZsmb7931FHMyFC8lHKx06ZH4vbufuDjVqmH+YNWuay/vvT1vQGB1tgpbUwqhbL+90X1ycqe3yZbNlhKtr8hDKw8P8vqcUEt28mbHHuxMnJ/D0NI9/+5ba7Xfbx44BsMKmXCIkxIRNS5YobBKRHCxhfrk9uLunaTh2vnz56Ny5M7Nnz+add96xBTkLFiwgLi6Ojh07cvXqVWrXrs3AgQPx8vJixYoVvPjiiwQGBlKvXr27PkZ8fDzt2rXDx8eHHTt2EBERkWIvJ09PT2bPnk2pUqX4/fff6dGjB56enrz11lt06NCBffv2sWrVKn788UcAvL29kx3j2rVrBAcH06BBA3bt2sW5c+d4+eWX6d27N7Nnz7btt27dOkqWLMm6dev466+/6NChAzVr1qRHjx53fT4pPb/WrVvj4eHBhg0biI2NpVevXnTo0IH169cD0KlTJ2rVqsWnn36Ks7Mze/futQVbvXr1Ijo6mo0bN1KwYEH++OMPWzAmIo4rLs4MlEgIljZtMj2YbpU/P9StmxguPfSQ+ftP0iE6Gr76CkaNMtMewLyIffqYPxRSSuucnMyopVKloGHD5PdbreaHdacw6upVOHvWbDt2pFxb0aKph1EBAeYPfck6Vqv5Wd0aKu3dC6dOpbx/iRJJQ6WaNc28VWfne3t8FxfzO1C06L19P5jncOPGvQdVCZcJI36ioszv9u3/Gd2Ns7MJd+41BEppH1fX3DM1EIVNuUarVua9Y9Mm+OcfKFLE3hWJiNyD69ft96nn1atm/n8avPTSS4wZM4YNGzbQpEkTwIwaat++Pd7e3nh7e/PGG2/Y9u/Tpw+rV6/m22+/TVPY9OOPP3Lw4EFWr15NqX/nh4wcOZInn3wyyX63jqwKCAjgjTfeYN68ebz11lsUKFAADw8P8uXLd8dpc3PmzOHmzZt8+eWXtp5RU6ZMoVWrVowePdo20qhw4cJMmTIFZ2dngoKCaNmyJaGhofcUNoWGhvL7779z9OhR/P39Afjyyy+5//772bVrF3Xr1uXEiRO8+eabBAUFAVCxYkXb9584cYL27dtTrVo1AMqXL5/uGkQk+/z5J3z6KXz5pTlPvZWbGzRokLhSXP36JvuXe3DjBsyYAR99BCdPmtuKFoX+/aFXLyhU6N6PbbGY4KFECUjpfcxqhUuX7hxGXb4MFy+abffulB/H2/vOYVSRIrnqj/EsFR0NBw4kDZZ+/dUELSmpUCF5sHTr6DdHYbGY/yTc3TNWX1ycGamUUhh15YoZEXSnkMjFRb+Ld6GwKZcoX97MX9+3D374ATp1sndFIiK5V1BQEA0bNmTmzJk0adKEv/76i02bNjFixAgA4uLiGDlyJN9++y2nT58mOjqaqKgo3NP4F9SBAwfw9/e3BU0ADRo0SLbf/PnzmTRpEkeOHOHq1avExsbilc4hAAcOHKBGjRpJmpM/9NBDxMfHc+jQIVvYdP/99+N8yyeZJUuW5Pfff0/XY936mP7+/ragCaBKlSoUKlSIAwcOULduXQYMGMDLL7/MV199RbNmzXjmmWcI/Lcp4euvv07Pnj1Zs2YNzZo1o3379vfUJ0tEsk5cHKxYAVOnwpo1ibd7epqpcAkjl+rUUc/pDLtyBaZNg3HjzKgiMH+Ev/EGvPpqmj9IyRCLxQRBRYrAAw+kvE9Cf6jUAqkLF8w+v/5qtpQULJg8hPLzM31Fbt/y0i9WZKR5zW4drbR/vwmcbufiYv5wvDVUqlHD/OPMS5ydTQCbkRBW7khhUy4SEmLCpqVLFTaJSA7l7m6/Robp/Ci9e/fu9OnTh6lTpzJr1iwCAwNp3LgxAGPGjGHixIlMmDCBatWqUbBgQfr160d0Sid992jbtm106tSJ4cOHExwcjLe3N/PmzWPcuHGZ9hi3ur03k8ViIT6lXg6ZZNiwYTz//POsWLGCH374gffee4958+bRtm1bXn75ZYKDg1mxYgVr1qxh1KhRjBs3jj59+mRZPSKSNufPm8E10z61EnPiDLXZzXB207zEbu53PkiBBjVx6vQ8tGhhhjXJvbt0CSZPhokTE4eMlSkDgwaZ1acc7fX19jYNpFP7cODatTuHUeHhZp/9+812N+7uKYdQqW2FCiVed7TXLoHVappq3z4NLmG65O28vRMDpYRwKSgobwVxYjcKm3KRkBDT/++HH0yIrf9DRCTHsViy5xPYTPDss8/St29f5syZw5dffknPnj1t/Zu2bNlC69ateeGFFwDTo+jPP/+kSpUqaTp25cqVOXnyJGFhYZT8d4j49u3bk+yzdetWypYtyzvvvGO77fjx40n2cXFxIS4u7q6PNXv2bK5du2Yb3bRlyxacnJy477770lRveiU8v5MnT9pGN/3xxx9cvnw5yWtUqVIlKlWqRP/+/enYsSOzZs2ibdu2APj7+/Paa6/x2muvMXjwYKZPn66wScROrPFW9i4/xYbxu7m+aTc143eznT34cjZxp3P/Xi48Agu/Nz2E2rWD55+HRx81SytL2pw/Dx9/DFOmmFFNABUrwttvm0+c73HhBrsrWBCqVDFbSm7ehBMnkoZQx46ZEOryZRO+XbpkRkcl9IC8fj151/m0cHNLX1B165ZZDZnj4uDw4eSNu1PrLeTvn3S0Uq1aZuSXpnqJneh/9Vykbl3w9TX/327YAI8/bu+KRERyLw8PDzp06MDgwYOJjIyka9eutvsqVqzId999x9atWylcuDDjx4/n7NmzaQ6bmjVrRqVKlejSpQtjxowhMjIySaiU8BgnTpxg3rx51K1blxUrVrBo0aIk+wQEBHD06FH27t1L6dKl8fT0xPW2JX47derEe++9R5cuXRg2bBjnz5+nT58+vPjii7YpdPcqLi6OvXv3JrnN1dWVZs2aUa1aNTp16sSECROIjY3lP//5D40bN6ZOnTrcuHGDN998k6effppy5cpx6tQpdu3aRfv27QHo168fTz75JJUqVeLSpUusW7eOypUrZ6hWEUkjq9X8wb9nDzHbd3N25W7cD+6mVux5at2+q5MTlipVoHZts1WsaFa0mTvXNCSePdtsJUpAhw4meKpfX38cp+b0aRg7Fj77LHEJ+WrVTMj0zDP33rQ5p3Bzg0qVzHYncXFmWllC+JSe7fJl8zt+86YZQRQWlv46XV3vPGoqte32EUu//ZbyoilOTlC5ctLRSjVqaJlGcTgKm3IRJyfTKHz6dDOVTmGTiEjW6t69OzNmzKBFixZJ+isNGTKEv//+m+DgYNzd3XnllVdo06YNEak15byNk5MTixYtonv37tSrV4+AgAAmTZpE8+bNbfuEhITQv39/evfuTVRUFC1btuTdd99l2LBhtn3at2/PwoULefTRR7l8+TKzZs1KEooBuLu7s3r1avr27UvdunVxd3enffv2jB8/PkOvDcDVq1epVSvpn5+BgYH89ddfLFmyhD59+tCoUSOcnJxo3rw5kydPBsDZ2ZmLFy/SuXNnzp49S7FixWjXrh3Dhw8HTIjVq1cvTp06hZeXF82bN+fjjz/OcL15xdSpUxkzZgzh4eHUqFGDyZMnp9q4PiYmhlGjRvHFF19w+vRp7rvvPkaPHp3kd/HKlSu8++67LFq0iHPnzlGrVi0mTpxI3bp1s+spSVaxWs0Ikt27E7c9e0xvHSA/UPrfXWNx5kzh+ynQ8AGKNzfhkqVGjeRTlJs3hw8/hM2bTej07bdw7pyZDjZ5MpQrBx07muDp/vuz9ek6rKNHYfRomDUrsQdP3bowZAg89ZT5I0ASOTsnBjjpFR+fNKi6dcRUWoKquDizull4uNkyyt3dTDu8dbRS1ap2Xc5eJK0sVqvVau8icpPIyEi8vb2JiIhId5PWzLB8uQmc/P3NuYE+GBIRR3bz5k2OHj1KuXLlcHPU/giSo93pd8ze79n2MH/+fDp37sy0adOoX78+EyZMYMGCBRw6dIgSJUok23/gwIF8/fXXTJ8+naCgIFavXs2AAQPYunWrLUjs0KED+/bt49NPP6VUqVJ8/fXXfPzxx/zxxx/4+fmlqa68+LNwOFarCTUSAqWEcOn25eOAGPKxn/vZTW2OFqlNpedq02JQdYr538MfwNHRsHatCZ4WLzY9eRJUq2ZCp+eeM82g85qDB2HUKPjmGxNigOmq/s475lNlneg7FqvVTGu8WyCV2n2FCydfDa5ixdw/Yk1ynLS+ZytsymT2Plm6ccOscHrjhhl9WbNmtpcgIpJmCpskqylsSqp+/frUrVuXKVOmAKafmL+/P3369GHQoEHJ9i9VqhTvvPMOvXr1st3Wvn17ChQowNdff82NGzfw9PRkyZIltGzZ0rZP7dq1efLJJ/m///u/FOuIiooiKirK9nVkZCT+/v556mdhV1Yr/P138hFLly4l3zV/fs6VqMr6yNqsu1Kb3dTmd6rRJNiNXr1Mn+9M+1v42jVYtswETz/8ADExifc1bGiCp2eeMdPucrO9e00j1u++Mz8rgCeeMCFTo0Z2LU2yiNWq8FByjLSeP2kaXS5ToIB5L1qyxEylU9gkIiIiANHR0ezevZvBgwfbbnNycqJZs2Zs27Ytxe+JiopKFtIVKFCAzZs3AxAbG0tcXNwd90nJqFGjbNMiJYvFx5uVqm4PllKa1ps/v5my88ADHC9Wmy/21WbcmmpEnja93goVMouczelpBlxkuoIFzSim554zI6q+/94ET+vXw9atZuvbF5o1M8FTmzam0XhusX07fPCBmaqQoHVrEzJpWmrupqBJciGHneA7depUAgICcHNzo379+uzcuTPVfffv30/79u0JCAjAYrEwYcKEZPuMGjWKunXr4unpSYkSJWjTpg2HDh1Ksk+TJk2wWCxJttdeey2zn1qWCwkxl0uX2rcOERERcRwXLlwgLi4uWeN3Hx8fwlPpLRIcHMz48eM5fPgw8fHxrF27loULFxL2b9NcT09PGjRowPvvv8+ZM2eIi4vj66+/Ztu2bbZ9UjJ48GAiIiJs28mTJzPvieZl8fFw6JAJaN54w6zyVriwaajcsaNpLr1unQmaXFygTh149VX4/HPYvZubF67yVd+fefC3zwkY9SrvLatDZJQrNWuanqCnT8P48VkUNN2uSBHo0cM0FD950jxwnTpmOtnq1dClC/j4mJFOixaZhs45kdVqwrRmzaBBAxM0OTmZwO2338zUQgVNIpIDOeTIpvnz5zNgwIAk/QSCg4NT7Sdw/fp1ypcvzzPPPEP//v1TPOaGDRvo1asXdevWJTY2lrfffpsnnniCP/74w7bUM0CPHj0YMWKE7Wv325sa5gAtW5pwfPdus9BH6dJ3/x4RERGR202cOJEePXoQFBSExWIhMDCQbt26MXPmTNs+X331FS+99BJ+fn44OzvzwAMP0LFjR3bv3p3qcV1dXZOtjCjpFB8Pf/6ZdMTSL7+YnjG3c3U1q1UlrApXu7ZZXt7FBTB9PqdNg//9z9b7m/z5TY7Tq5fJQOw68MLPD/r3N9uff8K8eTBnjgnWvvvObN7e0K6dGfH06KOO3+fGaoVVq+D//s+M2ALIlw9efBEGDbr7imsiIg7OIcOm8ePH06NHD7p16wbAtGnTWLFiBTNnzkyxn0DdunVtK56kdD/AqlWrknw9e/ZsSpQowe7du2l0y9xnd3d3fH19M+up2IWPjzkp2LrVfDiSAwdniYiISCYrVqwYzs7OnD17NsntZ8+eTfXcp3jx4ixevJibN29y8eJFSpUqxaBBgyhfvrxtn8DAQDZs2MC1a9eIjIykZMmSdOjQIck+kkFxcSkHS1evJt/Xzc30UXjggaTBUv78SXaLj4cf18DUqeZ8MT7e3F66tDl3fPllc07pcCpVgqFD4d13zWswd67ZTp82q7XNmgW+vvDssyZ4qlfPsaYoxceb0UoffGCmM4IJA7t3h7fegrJl7VqeiEhmcbiw6V76CdyLhOWnixQpkuT2b775hq+//hpfX19atWrFu+++e8fRTSk1uHQEISEmbFq6VGGTiDg+rVUhWSU+4S9owcXFhdq1axMaGkqbNm0A8/qEhobSu3fvO36vm5sbfn5+xMTE8P333/Pss88m26dgwYIULFiQS5cusXr1aj766KOseBp5x40bsGYNLFxoTuguX06+T4ECJli6dcRS5cpmhEwqLl+G2bPhk0/g8OHE2x97zIxiatXqjt/uOCwWE6g98ACMHg2bNpnQacECs+T8pElmK1/eTCF8/nkTutlLbCzMn28af//xh7nN3d2cqP/3v1CqlP1qExHJAg73VnKnfgIHDx7MlMeIj4+nX79+PPTQQ1StWtV2+/PPP0/ZsmUpVaoUv/32GwMHDuTQoUMsXLgw1WM5aoPLkBAzAjc01Iym9vS0d0UiIsnlz58fi8XC+fPnKV68OBZH+vRZcjSr1Up0dDTnz5/HyckJl3+nC+V1AwYMoEuXLtSpU4d69eoxYcIErl27ZhtN3rlzZ/z8/Bg1ahQAO3bs4PTp09SsWZPTp08zbNgw4uPjeeutt2zHXL16NVarlfvuu4+//vqLN998k6CgINsxJR0iI2HlShMwrVxpVmdL4O5ulkO/NVi67740J0O//mpGMX3zDVy/bm7z9ISuXeE//4GgoMx/OtnGyQkaNzbbpEkmpJs714wg+vtvM4rogw/MVMKOHU0/pOwaQRQdDV9+CR9+aBq1g2lq3qcP9OsHxYplTx0iItnM4cKm7NCrVy/27duXbJWUV155xXa9WrVqlCxZkscee4wjR44QGBiY4rEGDx7MgAEDbF8nLN1rb0FBUKEC/PWXeb9t397eFYmIJOfs7Ezp0qU5deoUx44ds3c5kgu5u7tTpkwZnJwcdk2UbNWhQwfOnz/P0KFDCQ8Pp2bNmqxatcr2Id+JEyeSvFY3b95kyJAh/P3333h4eNCiRQu++uorChUqZNsnIiKCwYMHc+rUKYoUKUL79u354IMPyH/btC1JxcWLZuTSwoXmpC06OvG+MmXMSVy7dqZHQjr7EEVHmwXdpk6FLVsSb69a1YxieuEF8PDIpOfhKFxc4KmnzHbtmnlt586FH34widuvv5pPZB9+2ARPzzwDxYtnfh03bsCMGfDRR6bBOUDRoqbvVK9eZmk/EZFczGJ1sLkL0dHRuLu7891339mGeAN06dKFy5cvs2TJkjt+f0BAAP369aNfv34p3t+7d2+WLFnCxo0bKVeu3B2Pde3aNTw8PFi1ahXBwcFpqj8yMhJvb28iIiLwsvNSrP/9r1m4o3Nn+OILu5YiInJHcXFxxMTE2LsMyWWcnZ3Jly9fqiPmHOk9O6/Lcz+LsDCzgtrChWYlsri4xPsqVTIBU/v2ZorYPYz4PHUKPvvMrCCX0KIrXz6TWfXqBY884lhtjLLFxYsmeZs7FzZsMA26wQR4Tzxhgqc2bTI+HeDKFdNtfdy4xBe/ZEmzOuArr+TCdE9E8pq0vmc73MimjPQTuBOr1UqfPn1YtGgR69evv2vQBLB3714ASpYsec+Pa08hISZsWrHCTBPPEfPvRSRPcnZ2xtnRVw4SEcmIo0dNuLRwIWzblhh2gOm71K6d2apUuackyGqFdevMKKYlSxLzq5IlTcbxyit5vC1Q0aKJL8SpU6Z/0ty5ptn6Dz+Yzc3NnEB37AhPPmkad6fVpUsweTJMnAj//GNuK1PGjKLq1s0cW0QkD3HI+CG9/QSio6P5499Ge9HR0Zw+fZq9e/fi4eFBhQoVADN1bs6cOSxZsgRPT0/Cw8MB8Pb2pkCBAhw5coQ5c+bQokULihYtym+//Ub//v1p1KgR1atXt8OrkHEPPQRFipgPcrZtM59iiYiIiEg2OXDAhEvff29WTrvVgw+a0Utt20Iq7RrSIjLStAT65BPzcAkaNza9mNq2TbYQnZQubaYA/Pe/ZpW/uXNhzhxz/dtvzebtbX4+zz8PTZqkPoXx3Dn4+GOT8l25Ym6rWBHefhs6ddKLLyJ5lsNNo0swZcoUxowZY+snMGnSJOrXrw9AkyZNCAgIYPbs2QAcO3YsxZFKjRs3Zv369QCpDqGfNWsWXbt25eTJk7zwwgvs27ePa9eu4e/vT9u2bRkyZEi6hnM72jDwzp3hq6/MyN0xY+xdjYiIiONwtPfsvCzX/CysVhMqJQRMty5uk9DEun17M13Lzy9DD7V/v8k3vvoKrl41txUsCC++aKbK3bIGjqSF1Qp79pjgae5cOHMm8T5fX9NUvGNHqFvXjDw7fdqcXH/+uenPBOZFf+cd0wdKo3VFJJdK63u2w4ZNOZWjnSx99515v6tUCQ4dsnc1IiIijsPR3rPzshz9s4iPN0PIE6bI3brYQf788PjjZnpc69YZXnksJsYssDZ1qmk7lCAoyIxi6tzZDMiRDIqLg02bTOi0YIGZIpcgMNCsBLh4cWIz97p1TcjUqpUJFUVEcjGFTXbiaCdLV66Y85roaDO0OkcvaysiIpKJHO09Oy/LcT+LmBjYuNGMXlq0CP5tzwCAu7vp99OuHbRsmSnpT1iYGUDz+eeJA26cnEx+1asXNG2aBxt+Z5foaFi92gRPS5bA9euJ9zVqZEKmxx/XD0BE8owc2yBcMpenJzz6qHmPXLpUYZOIiIjIPbl5E3780QRMS5cmNoEGEyi1amUCpuBgEzhlktOnTf/wCxfM1yVKQI8e8Oqr4O+faQ8jqXFxMT/bVq3MfMWlS81UyVatTNgkIiIpUtiUB4SEJIZNb71l72pEREREcoirV80qZd9/b5b3TWiOBGboeJs2pgdT06YmlMgC8+eboKlcOfi//zMPl55F0iQTeXiYhuHPP2/vSkREHJ7CpjygVSszxHrrVjh/HooXt3dFIiIiIg7q0iVYtsz0X1q92oxoSuDnZ0YvtW9vlv3Nl/Wn0kuWmMv+/ZVxiIhIzqGwKQ/w94datcyI3xUroGtXe1ckIiIi4kDOnjUNnxcuhJ9+gtjYxPsCA0241K6daQSdjQ2gz5+HzZvN9ZCQbHtYERGRDFPYlEe0bm3CpqVLFTaJiIiIcOJE4gpymzfDrWvmVK2aGDBVq2a35s/Ll5vF7mrVgrJl7VKCiIjIPVHYlEeEhMCwYYmjwd3c7F2RiIiISDb780/Tf2nhQvj556T31a1rwqV27aBSJfvUd5vFi81lmzb2rEJERCT9FDblETVrQunScOoUhIaalXhFREREcjWrFX77zYRL338P+/cn3mexwCOPmHCpbVsoU8Z+dabg2jVYs8ZcV9gkIiI5jcKmPMJiMaObPvnETKVT2CQiIuJ44uPjccrGnkC52pEjEBxsLhPkywePPWYCptatwcfHfvXdxdq1ZjR6QICZySciIpKT6GwmD0loLLlsmZn/LyIiIo6lbNmyfPDBB5w7d87epeR8ZcrAP/+Y3gGtW8OXX8K5c7BqFbzyikMHTZB0Cp2dWkaJiIjcM4VNeUiTJuDhAWFhsHu3vasRERGR250+fZqhQ4dSpkwZXnzxRbZv327vknKu/PlNs8rz501y8+KLULiwvatKk9hY8+EgaAqdiIjkTAqb8hBXV2je3FxfutS+tYiIiEhyO3bs4IUXXsBisfDNN9/w0EMPUbduXb744guioqLsXV7OU7eu+aQth9m82QzKKloUHnrI3tWIiIikn8KmPKZ1a3OpsElERMTxJARLp06dYuTIkfj7+7N7925eeuklSpcuzeDBgzlx4oS9y5QstmSJuWzVyrSZEhERyWkUNuUxLVqAs7NZmOXYMXtXIyIiIikpWrQogwYN4ujRoyxevJhmzZrxzz//MHr0aAIDA2nbti2hoaH2LlOygNWa2K8p4UNCERGRnEZhUx5TpAg8/LC5rtFNIiIijs1isRASEsLq1as5ePAgr7zyCnFxcSxdupQnnniC+++/nxkzZhCvlT9yjYQPBAsUgCeesHc1IiIi90ZhUx6UsCqdwiYREZGc4fjx4/zvf//j+++/B8BqteLj48OBAwd45ZVXqF27NqdOnbJzlZIZEkY1PfEEuLvbtRQREZF7prApD2rVylxu2ACXL9u1FBEREbmDNWvWEBISQoUKFRgzZgzXrl3jpZdeYu/evZw5c4Y1a9bw4IMP8uuvv9K/f397lyuZICFs0ip0IiKSk6nlYB5UsSJUrgwHDsCqVfDcc/auSERERBJERkYya9YsPv30Uw4fPozVasXPz4+ePXvy6quvUrRoUdu+zZo1o2nTptSsWZOffvrJjlVLZjh+HPbuBScneOope1cjIiJy7xQ25VEhISZsWrpUYZOIiIij6NmzJ9988w3Xrl3DarXSoEEDXn/9dZ5++mmcnZ1T/B4nJyfq1KnD/v37s7layWwJq9A9/DAUK2bfWkRERDJCYVMe1bo1jB4NK1dCTAzkz2/vikREROSzzz7DxcWF559/nr59+1KnTp00fV+jRo2wWq1ZXJ1kNU2hExGR3MJi1ZlJpoqMjMTb25uIiAi8vLzsXU6q4uKgVCk4dw5CQ6FpU3tXJCIikr0c8T172LBh9OzZEx8fH3uXkq0c8WeR3S5eBB8fc4525AiUL2/vikRERJJL63u2GoTnUc7Oib0AEoZsi4iIiH0NGzYszwVNYqxYYYKm6tUVNImISM6nsCkPCwkxl0uXgsa3iYiI2N+lS5fYuHEjp0+fTnWf06dPs3HjRi5rSdlcJeHDv9at7VuHiIhIZlDYlIc1awZubnDsGOzbZ+9qREREZOLEiTz66KOEhYWluk9YWBiPPvooU6dOzcbKJCvduGFWCAb1axIRkdxBYVMeVrCgCZzAjG4SERER+1q5ciXly5e/Y2PwOnXqUK5cOZYvX56NlUlW+vFHuH4d/P2hVi17VyMiIpJxDhs2TZ06lYCAANzc3Khfvz47d+5Mdd/9+/fTvn17AgICsFgsTJgw4Z6OefPmTXr16kXRokXx8PCgffv2nD17NjOflsO5dSqdiIiI2NexY8e477777rpfUFAQR48ezYaKJDvcugqdxWLPSkRERDKHQ4ZN8+fPZ8CAAbz33nvs2bOHGjVqEBwczLlz51Lc//r165QvX54PP/wQX1/fez5m//79WbZsGQsWLGDDhg2cOXOGdu3aZclzdBStWpnLnTvhDiP2RUREJBskrPByN15eXurZlEvExcGyZea6+jWJiEhu4ZBh0/jx4+nRowfdunWjSpUqTJs2DXd3d2bOnJni/nXr1mXMmDE899xzuLq63tMxIyIimDFjBuPHj6dp06bUrl2bWbNmsXXrVrZv355lz9XefH2hfn1zXaPxRURE7Kt48eIcPHjwrvsdOnSIIkWKZENFktW2bYPz56FQIWjUyN7ViIiIZA6HC5uio6PZvXs3zRKaCQFOTk40a9aMbdu2Zdkxd+/eTUxMTJJ9goKCKFOmzB0fNyoqisjIyCRbTpMwlS5hFRQRERGxjwcffJC9e/eycePGVPfZtGkTv/zyCw8++GA2ViZZJWEK3VNPQf78di1FREQk0zhc2HThwgXi4uLw8fFJcruPjw/h4eFZdszw8HBcXFwoVKhQuh531KhReHt72zZ/f/97qtGeEsKmH3+Ea9fsW4uIiEhe1rNnT6xWK08//TRLUvgUaMmSJTz99NNYLBZee+01O1QomclqTdqvSUREJLdwuLAppxk8eDARERG27eTJk/YuKd3uvx/KlYOoKFi71t7ViIiI5F1Nmzald+/eXLhwgXbt2uHj40OjRo1o1KgRvr6+tGvXjvPnz/Paa6/xxBNP2LtcyaD9++HIEXB1heBge1cjIiKSeRwubCpWrBjOzs7JVoE7e/Zsqs2/M+OYvr6+REdHJ2u2ebfHdXV1xcvLK8mW01gsWpVORETEUUyaNImJEydStGhRzp8/z+bNm9m8eTPnzp2jaNGifPzxx0ydOtXeZUomSBi81qwZeHjYtxYREZHM5HBhk4uLC7Vr1yY0NNR2W3x8PKGhoTRo0CDLjlm7dm3y58+fZJ9Dhw5x4sSJe37cnCRh9ZPly82qKCIiImI/ffr0ISwsjO3btzN//nzmz5/Ptm3bCAsLo2/fvhk69tSpUwkICMDNzY369euzc+fOVPeNiYlhxIgRBAYG4ubmRo0aNVi1alWSfeLi4nj33XcpV64cBQoUIDAwkPfffx+r1ZqhOvMCTaETEZHcKp+9C0jJgAED6NKlC3Xq1KFevXpMmDCBa9eu0a1bNwA6d+6Mn58fo0aNAkwD8D/++MN2/fTp0+zduxcPDw8qVKiQpmN6e3vTvXt3BgwYQJEiRfDy8qJPnz40aNAgTzTgfPhhswrK+fOwYwc0bGjvikRERPI2Z2dn6tWrR7169TLtmPPnz2fAgAFMmzaN+vXrM2HCBIKDgzl06BAlSpRItv+QIUP4+uuvmT59OkFBQaxevZq2bduydetWatWqBcDo0aP59NNP+eKLL7j//vv5+eef6datG97e3rz++uuZVntuc/Ik/PyzGWHeqpW9qxEREclcDhk2dejQgfPnzzN06FDCw8OpWbMmq1atsjX4PnHiBE5OiYOyzpw5YzvhARg7dixjx46lcePGrF+/Pk3HBPj4449xcnKiffv2REVFERwczCeffJI9T9rO8ueHFi1gzhwzlU5hk4iISO4zfvx4evToYfuwbdq0aaxYsYKZM2cyaNCgZPt/9dVXvPPOO7Ro0QIwDcx//PFHxo0bx9dffw3A1q1bad26NS1btgQgICCAuXPn3nHEVFRUFFFRUbavc+JqvhmV0LqgYUO4bQ0bERGRHM8hwyaA3r1707t37xTvSwiQEgQEBKRpqPadjgng5ubG1KlT82wfhJAQEzYtWQIffmjvakRERPK2gwcPcujQISIjI1M9z+ncuXOajxcdHc3u3bsZPHiw7TYnJyeaNWvGtm3bUvyeqKgo3NzcktxWoEABNm/ebPu6YcOGfP755/z5559UqlSJX3/9lc2bNzN+/PhUaxk1ahTDhw9Pc+25UUK/poRWBiIiIrmJw4ZNkv2aN4d8+eDgQfjzT6hUyd4ViYiI5D3bt2/nlVdeYf/+/anuY7VasVgs6QqbLly4QFxcXJJR3QA+Pj4cPHgwxe8JDg5m/PjxNGrUiMDAQEJDQ1m4cCFxtzR4HDRoEJGRkQQFBeHs7ExcXBwffPABnTp1SrWWwYMHM2DAANvXkZGR+Pv7p/m55HSXL8O6dea6+jWJiEhu5HANwsV+vL2hSRNzfdkyu5YiIiKSJ/355588/vjj7Nu3jwcffJBy5coB8Nxzz1G7dm2cnZ0BaNu2bbqCpns1ceJEKlasSFBQEC4uLvTu3Ztu3bolaWfw7bff8s033zBnzhz27NnDF198wdixY/niiy9SPW5uWM03I1auhNhYqFIFKla0dzUiIiKZL0NhU1xcHJGRkcTGxia5/caNGwwfPpy2bdvSv39/zpw5k6EiJfuEhJjLhD4CIiIikn1Gjx7NtWvX+OSTT9iyZQuPPPIIAN988w07d+7kl19+oWbNmhw+fJgpU6ak69jFihXD2dmZs2fPJrn97Nmz+Pr6pvg9xYsXZ/HixVy7do3jx49z8OBBPDw8KF++vG2fN998k0GDBvHcc89RrVo1XnzxRfr3729byEWS0yp0IiKS22UobBoxYgSFCxdOMs/farXSpEkTRowYwZIlS5g0aRINGjTg0qVLGS5Wsl5C2LR5M1y8aN9aRERE8pp169YRGBjIa6+9luL9999/P8uXL+fIkSN88MEH6Tq2i4sLtWvXJjQ01HZbfHw8oaGhNGjQ4I7f6+bmhp+fH7GxsXz//fe0vqXR0PXr15OMdAKzkl58fHy66ssroqLghx/MdYVNIiKSW2UobAoNDcXX19f2qRvAsmXL2LVrFxUrVmTChAk88cQTnDp1iunTp2e4WMl6ZctCjRoQH2+GeIuIiEj2CQsLo2rVqravE6bNRUdH224rWbIkjRs3ZuHChek+/oABA5g+fTpffPEFBw4coGfPnly7ds22Ol3nzp2TNBDfsWMHCxcu5O+//2bTpk00b96c+Ph43nrrLds+rVq14oMPPmDFihUcO3aMRYsWMX78eNq2bZvu+vKCn36Cq1ehVCmoXdve1YiIiGSNDIVNR48eJSgoKMltS5YswWKx8M033/D666+zbNkyihcvznfffZehQiX7aCqdiIiIfRQoUIB8+RLXb/H09ARINvXNy8uLkydPpvv4HTp0YOzYsQwdOpSaNWuyd+9eVq1aZWsafuLECcLCwmz737x5kyFDhlClShXatm2Ln58fmzdvplChQrZ9Jk+ezNNPP81//vMfKleuzBtvvMGrr77K+++/n+768oKEKXStW4OTuqeKiEguZbGmtpZuGri5udG+fXu++eYb221BQUFcv36dEydO2G5r06YN27dvJzw8PGPV5gCRkZF4e3sTERGRY5td/vwz1K0LHh5w4QK4utq7IhERkczniO/Z1apVw93dnR07dgAwdepUXn/9debOncuzzz4LmJYF9913H9evX+fUqVP2LDfTOOLPIivEx5sRTWfPwurV8MQT9q5IREQkfdL6np2hz1Py5cvHtWvXbF9funSJw4cP89BDDyXZz9PTk4iIiIw8lGSjBx4wJ0JXryYuyysiIiJZr379+vzxxx/cuHEDgObNmwPQv39/VqxYwe+//07Pnj05cuQIdevWtWepcg927DBBk5dX4grAIiIiuVGGwqby5cuzfft2WwPI5cuXY7Vaefjhh5Psd+7cOYoXL56Rh5Js5OQErVqZ65pKJyIikn1atGjBzZs3Wb58OQCBgYG88sorhIWFERISQs2aNfn8889xcXHh//7v/+xcraTXkiXmskULcHGxby0iIiJZKUNhU0hICOfOnaN169ZMnDiRgQMH4uzsTKuEpAIz1PuXX36hXLlyGS5Wss+tfZvufaKliIiIpEe7du2IiYnhmWeesd02depUxo4dS7169ahQoQIhISFs2LCB+++/346Vyr1I6NekVehERCS3y3f3XVL31ltvsWTJElasWMGKFSsAGDRoEGXKlLHts3nzZi5cuJBstJM4tqZNoWBBOH0afvnFTK0TERGR7Ofk5MSAAQMYMGCAvUuRDDh4EA4dgvz54ckn7V2NiIhI1spQ2OTl5cXOnTv57rvvOHv2LHXr1qVx48ZJ9rl48SJ9+/alQ4cOGSpUspebGwQHw8KFZnSTwiYREZGs99JLL1GsWDE++ugje5cimSxhVNNjj5meTSIiIrlZhsImMEv0vvjii6ne36ZNG9porHCOFBKSGDYNG2bvakRERHK/r7/+mtatW9u7DMkCCWGTfrwiIpIXZKhn091ERERgVcOfHKtFC9Ms/Jdf4MQJe1cjIiKS+/n6+mKxWOxdhmSysDCzEh0k9sUUERHJzTIUNu3bt49Jkybx559/Jrl93bp1lCtXjiJFilCiRAlmz56dkYcROyleHBo2NNeXLbNvLSIiInnB448/zpYtW4iJibF3KZKJElb3rV8fSpWyby0iIiLZIUNh06RJkxgwYAAFChSw3Xbx4kXatGnD8ePHsVqtXLx4kZdffplffvklw8VK9rt1VToRERHJWsOGDSMqKooePXpw5coVe5cjmUSr0ImISF6ToZ5NW7Zs4f7778ff399221dffcWVK1d49dVXGT16NEuXLqVz585MnjyZmTNnZrhgyV4hIfDWW7BuHURGqqGliIhIVpo1axbNmzfnyy+/ZMWKFTRr1oyAgIAkH+wlsFgsvPvuu3aoUtIjMhJCQ811hU0iIpJXWKwZaKpUrFgxGjRowLJb5li1bNmSNWvWEB4eTtGiRQGoXbs2169f58CBAxmv2MFFRkbi7e1NREQEXrkkmQkKMkv1fvstPPOMvasRERHJHI74nu3k5ITFYrljz8uE+y0WC3FxcdlYXdZxxJ9FZpk/H557DipVMudTIiIiOVla37MzNLIp4UFutWPHDmrWrGkLmgAqVqzIypUrM/JQYkchITBmjJlKp7BJREQk6wwdOlQNwnOZJUvMpUY1iYhIXpKhsMnLy4vTp0/bvj5w4AD//PMPnTp1SravTpxyroSwacUKiI2FfBn6rREREZHUDBs2zN4lSCaKjjbnT6CwSURE8pYMNQivWbMmW7du5a+//gJgxowZWCwWGjdunGS/o0ePUrJkyYw8lNhRgwZQtChcugSbN9u7GhEREZGcYf1607PJx8esRCciIpJXZChsevXVV4mJiaF27drUqlWLjz/+mBIlStCyZUvbPleuXGHv3r1UrVo1w8WKfTg7w1NPmetalU5EREQkbRJWoQsJAacMnXWLiIjkLBmaEPXMM89w4MABRo8eza+//kpAQABffvklrq6utn2+/fZbYmJiko12kpwlJAS++MKETePGgWZFioiIZL4RI0akeV+tRufY4uMTP6TTFDoREclrMrQaXYLo6GgiIyMpVqxYsvtOnDjBpUuXCAwMxMPDI6MP5fBy62oqV69CsWIQFQX790OVKvauSEREJGMc8T37TqvR3dr/UqvROb5du6BePfDwgPPnwc3N3hWJiIhkXLasRpfAxcUlxaAJoEyZMpQpUyYzHkbsyMMDHnsMVq40n9IpbBIREcl87733Xoq3x8fHc/z4cdatW8fJkyfp3r07pUuXzubqJD0SptA9+aSCJhERyXsybV2x6Ohodu/ebVudzs/Pj9q1a+Pi4nLPx5w6dSpjxowhPDycGjVqMHnyZOrVq5fq/gsWLODdd9/l2LFjVKxYkdGjR9OiRQvb/amtiPfRRx/x5ptvAhAQEMDx48eT3D9q1CgGDRp0z88jtwgJSQyb9HKIiIhkvtTCpgQ3btygR48erF69mj179mRTVXIvEsImTaETEZG8KMOtCmNjY3n33XcpUaIEDz/8MB06dKBDhw48/PDDlChRgqFDhxIbG5vu486fP58BAwbw3nvvsWfPHmrUqEFwcDDnzp1Lcf+tW7fSsWNHunfvzi+//EKbNm1o06YN+/bts+0TFhaWZJs5cyYWi4X27dsnOdaIESOS7NenT590158bJTQJ374dzp61by0iIiJ5UYECBfj888+Jiopi6NCh9i5HUnH4MPzxB+TLB7d87ikiIpJnZChsio+PJyQkhJEjRxIZGUmhQoWoVasWtWrVolChQkRGRvLBBx/QunVr4uPj03Xs8ePH06NHD7p160aVKlWYNm0a7u7uzJw5M8X9J06cSPPmzXnzzTepXLky77//Pg888ABTpkyx7ePr65tkW7JkCY8++ijly5dPcixPT88k+xUsWDD9L04u5OcHdeqA1QrLl9u7GhERkbzJ3d2dOnXqsFxvxg5ryRJz2aQJFCpkz0pERETsI0Nh0//+9z9WrVpF2bJl+e6777h48SI///wzP//8MxcvXuT777+nbNmyrFq1ihkzZqT5uAlT8po1a5ZYqJMTzZo1Y9u2bSl+z7Zt25LsDxAcHJzq/mfPnmXFihV079492X0ffvghRYsWpVatWowZM+aOI7OioqKIjIxMsuVmISHmMmF1FREREcl+Tk5OqY72FvvTFDoREcnrMhQ2ffnllxQoUICffvqJdu3aJbu/bdu2hIaG4urqyhdffJHm4164cIG4uDh8fHyS3O7j40N4eHiK3xMeHp6u/b/44gs8PT2T1f36668zb9481q1bx6uvvsrIkSN56623Uq111KhReHt72zZ/f/+0PMUcKyFsWrsWrl+3by0iIiJ50ZkzZ9i8eXOy8x5xDGfPwtat5nrCeZOIiEhek6EG4fv27aNJkyYEBASkuk+5cuVo2rQpmzdvzshDZbqZM2fSqVMn3G5bHmTAgAG269WrV8fFxYVXX32VUaNG4erqmuw4gwcPTvI9kZGRuTpwql4dypaF48chNBRatbJ3RSIiIrnHxo0bU73vypUrHDhwgKlTpxIZGUnnzp2zsTJJq2XLTMuB2rUhF58SioiI3FGGwqaoqCi8vb3vup+npydRUVFpPm6xYsVwdnbm7G1dqM+ePYuvr2+K3+Pr65vm/Tdt2sShQ4eYP3/+XWupX78+sbGxHDt2jPvuuy/Z/a6urimGULmVxWI+pZs82UylU9gkIiKSeZo0aZLq6rkJrFYrderU4f3338+mqiQ9Evo1aQqdiIjkZRkKm/z9/dm2bRtxcXE4OzunuE9cXBzbt2+ndOnSaT6ui4sLtWvXJjQ0lDb/vlPHx8cTGhpK7969U/yeBg0aEBoaSr9+/Wy3rV27lgYNGiTbd8aMGdSuXZsaNWrctZa9e/fi5OREiRIl0lx/bpcQNi1bBvHx4JThNQ1FREQEoFGjRqmGTS4uLvj5+dGsWTOeffZZ8uXL0GmcZIGrV02rAVDYJCIieVuGzlKCg4P55JNP6Nu3Lx9//DH58+dPcn90dDT9+/fnxIkT9OrVK13HHjBgAF26dKFOnTrUq1ePCRMmcO3aNbp16wZA586d8fPzY9SoUQD07duXxo0bM27cOFq2bMm8efP4+eef+fzzz5McNzIykgULFjBu3Lhkj7lt2zZ27NjBo48+iqenJ9u2baN///688MILFC5cOF3152aNGoGXl+lJsGsX1K9v74pERERyh/Xr19u7BMmA1ashKgoCA+H+++1djYiIiP1kKGwaNGgQc+bM4dNPP2XJkiU899xzlCtXDoC///6b+fPnc+bMGYoUKcLAgQPTdewOHTpw/vx5hg4dSnh4ODVr1mTVqlW2ZpgnTpzA6ZYhNQ0bNmTOnDkMGTKEt99+m4oVK7J48WKqVq2a5Ljz5s3DarXSsWPHZI/p6urKvHnzGDZsGFFRUZQrV47+/fsn6ckk4OICTz4J8+eboeIKm0RERESSrkJ3l9mQIiIiuZrFarVaM3KAXbt28cwzz3DixIlkw76tVitlypTh+++/p3bt2hkqNKeIjIzE29ubiIgIvLy87F1OlpkzBzp1Mp/a7dtn72pERETSL6+8Z+cEueFnERMDJUrA5cuwcSM88oi9KxIREcl8aX3PznC3nbp16/Lnn3/y1Vdf8dJLLxEcHExwcDAvvfQSX331FX/++SfOzs53XF1Fcp4nnwRnZ9i/H44csXc1IiIiucOUKVNwdnZm2bJlqe6zbNkynJ2d+eyzz7KxMrmbTZtM0FSsGDRsaO9qRERE7CtTOku6uLjQqVMnOnXqlOL9PXv2ZNeuXcTGxmbGw4kDKFzY9G5at840Cr+lL7uIiIjcoyVLllC8eHFatmyZ6j4tWrSgWLFiLFq0iFdffTUbq5M7SZhCFxJiPpATERHJy7JtHbEMztYTB9S6tblcutS+dYiIiOQWBw8epGrVqkn6Ut7O2dmZatWqceDAgWysTO7Eak3ar0lERCSv06L1cs9atTKXGzfCpUv2rUVERCQ3OH/+PL6+vnfdz9fXl3PnzmVDRZIWv/wCJ0+Cuzs0a2bvakREROxPYZPcs/LloWpViIuDH36wdzUiIiI5n6enJ2fOnLnrfmfOnMHd3T0bKpK0SBjVFBwMBQrYtRQRERGHoLBJMiQkxFwuWWLfOkRERHKDGjVqsHXrVk6ePJnqPidPnmTr1q1Uq1YtGyuTO0k4D9IUOhEREUNhk2RIQtj0ww8QHW3fWkRERHK6559/nujoaNq1a0d4eHiy+8PDw2nfvj0xMTE8//zzdqhQbvf33/Dbb6Yp+B36uouIiOQpCpskQ+rWBR8fuHIFNmywdzUiIiI5W5cuXXjooYfYvXs3gYGBPPvsswwdOpShQ4fy7LPPUqFCBX7++WcefPBBXnrppXt6jKlTpxIQEICbmxv169dn586dqe4bExPDiBEjCAwMxM3NjRo1arBq1aok+wQEBGCxWJJtvXr1uqf6cpqEUU2NGkHRovatRURExFHkS8/OX3755T09yPnz5+/p+8TxOTmZRuH/+59Zle7xx+1dkYiISM7l7OzMihUr6NatG4sWLeK7777DYrEAiSv7tm7dmlmzZpEvX7pO4wCYP38+AwYMYNq0adSvX58JEyYQHBzMoUOHKFGiRLL9hwwZwtdff8306dMJCgpi9erVtG3blq1bt1KrVi0Adu3aRVxcnO179u3bx+OPP84zzzxzLy9BjpPQrylhlV4REREBizXhzCUNnJycbCc86WG1WrFYLElORHKryMhIvL29iYiIwMvLy97lZIvly03gVKYMHDsG9/ArIiIiku0c/T37t99+Y9WqVRw/fhyAMmXK0Lx5c2rUqHHPx6xfvz5169ZlypQpAMTHx+Pv70+fPn0YNGhQsv1LlSrFO++8k2SUUvv27SlQoABff/11io/Rr18/li9fzuHDh9N83ujoP4vUXLhgRnjHx8PRoxAQYO+KREREslZa37PT9ZFYmTJl7ilsktztscfMyisnTpieBRk4BxYREZF/Va9enerVq2fa8aKjo9m9ezeDBw+23ebk5ESzZs3Ytm1bit8TFRWFm5tbktsKFCjA5s2bU32Mr7/+mgEDBtzxnDEqKoqoqCjb15GRkel5Kg5j+XITNNWsqaBJRETkVukKm44dO5ZFZUhOVqAAPPGE6VmwdKnCJhEREUd04cIF4uLi8PHxSXK7j48PBw8eTPF7goODGT9+PI0aNSIwMJDQ0FAWLlyY6mj1xYsXc/nyZbp27XrHWkaNGsXw4cPv6Xk4koQpdFqFTkREJCk1CJdMkbAqXUKTTBEREUm/VatW0bRpU3766adU9wkNDaVp06asXbs2y+uZOHEiFStWJCgoCBcXF3r37k23bt1wckr5FHLGjBk8+eSTlCpV6o7HHTx4MBEREbbt5MmTWVF+lrp+HdasMdfVr0lERCQphU2SKVq2NL2adu+GU6fsXY2IiEjONGvWLHbu3EndunVT3adevXrs2LGD2bNnp+vYxYoVw9nZmbNnzya5/ezZs/j6+qb4PcWLF2fx4sVcu3aN48ePc/DgQTw8PChfvnyyfY8fP86PP/7Iyy+/fNdaXF1d8fLySrLlNGvWwI0bULasRnWLiIjcTmGTZAofH3jwQXN9+XL71iIiIpJT/fzzz9SsWRNPT89U9/H09KRWrVrs3LkzXcd2cXGhdu3ahIaG2m6Lj48nNDSUBg0a3PF73dzc8PPzIzY2lu+//57WKQzlmTVrFiVKlKBly5bpqiunShjN3aaNFkcRERG5ncImyTQJU+mWLrVvHSIiIjlVWFgYZcqUuet+/v7+hIWFpfv4AwYMYPr06XzxxRccOHCAnj17cu3aNbp16wZA586dkzQQ37FjBwsXLuTvv/9m06ZNNG/enPj4eN56660kx42Pj2fWrFl06dKFfPnS1RI0R4qNhWXLzHX1axIREUku958NSLZp3RoGD4bQULh6FTw87F2RiIhIzuLi4sKVK1fuut/Vq1dT7Zt0Jx06dOD8+fMMHTqU8PBwatasyapVq2xNw0+cOJHkuDdv3mTIkCH8/fffeHh40KJFC7766isKFSqU5Lg//vgjJ06c4KWXXkp3TTnRli1w8SIUKQIPP2zvakRERByPxWq1Wu1dRG4SGRmJt7c3ERERObL/QEZYrVCpEvz1F3z/PbRrZ++KREREUueI79l169blyJEjnDp1Cnd39xT3uX79OqVLl6ZMmTLs3bs3ewvMIo74s7iT/v1hwgTo3Bm++MLe1YiIiGSftL5naxqdZBqLRVPpREREMqJVq1ZcvnyZ3r17k9LngVarlT59+hAREZFi3yTJelZr0n5NIiIikpxGNmWynPbJXGbbsAGaNIGiRSE8HPJA2wYREcmhHPE9+/Lly1StWpWwsDBq1arFSy+9RFBQEAAHDx5k5syZ/PLLL/j6+vL7779TpEgRO1ecORzxZ5Ga334zq8+5ucGFC1CwoL0rEhERyT5pfc9WFCCZ6qGHoHBh08dg2zZ45BF7VyQiIpJzFCpUiBUrVtCqVSv27NnDL7/8kuR+q9VK6dKlWbp0aa4JmnKaxYvN5RNPKGgSERFJjcImyVT58kHLlvD112YqncImERGR9KlRowYHDx5k+vTprF69muPHjwNQpkwZmjdvzssvv0xBpRx2kxA2aRajiIhI6jSNLpPlpGHgWWXBAnj2WdMs/NAhe1cjIiKSspz6nn3x4kW+/PJLZs6cye+//27vcjJFTvlZHD8OAQHg5GTaBRQvbu+KREREspem0YndNG8OLi7w558mbLrvPntXJCIikrNZrVZWrVrFjBkzWL58OTExMfYuKU9KWADloYcUNImIiNyJQ69GN3XqVAICAnBzc6N+/frs3LnzjvsvWLCAoKAg3NzcqFatGitXrkxyf9euXbFYLEm25s2bJ9nnn3/+oVOnTnh5eVGoUCG6d+/O1atXM/255WaenvDoo+a6VqUTERG5d0ePHuXdd9+lTJkyPPXUUyxcuJDo6Ghq1arFxx9/bO/y8pyEKXRahU5EROTOHDZsmj9/PgMGDOC9995jz5491KhRg+DgYM6dO5fi/lu3bqVjx450796dX375hTZt2tCmTRv27duXZL/mzZsTFhZm2+bOnZvk/k6dOrF//37Wrl3L8uXL2bhxI6+88kqWPc/cKiTEXCpsEhERSZ+oqCi++eYbmjZtSsWKFRk5ciSnT58G4K233mLfvn38/PPPvP7663auNG/55x+z6i6oX5OIiMjdOGzPpvr161O3bl2mTJkCQHx8PP7+/vTp04dBgwYl279Dhw5cu3aN5cuX22578MEHqVmzJtOmTQPMyKbLly+zOOFjqdscOHCAKlWqsGvXLurUqQPAqlWraNGiBadOnaJUqVJ3rTun9BzIaidPQpky6mkgIiKOy9Hes3fv3s2MGTOYN28eERERWK1W8uXLR4sWLfjtt984fvw4cXFx9i4zSzjazyIlX30FnTtD1aqQS1pliYiIpFta37MdcmRTdHQ0u3fvplmzZrbbnJycaNasGdu2bUvxe7Zt25Zkf4Dg4OBk+69fv54SJUpw33330bNnTy5evJjkGIUKFbIFTQDNmjXDycmJHTt2pPi4UVFRREZGJtkE/P2hVi2Ij4cVK+xdjYiIiGO6dOkSkydPpmbNmtSrV4/PPvuMy5cvc9999/HRRx9x6tQpFi9ejJ+fn71LzfOWLDGXmkInIiJydw7ZIPzChQvExcXh4+OT5HYfHx8OHjyY4veEh4enuH94eLjt6+bNm9OuXTvKlSvHkSNHePvtt3nyySfZtm0bzs7OhIeHU6JEiSTHyJcvH0WKFElynFuNGjWK4cOH38vTzPVCQuCXX8xUuq5d7V2NiIiI4ylZsiQxMTFYrVY8PDzo0KEDL730Eg0aNLB3aXKLGzdg1SpzXWGTiIjI3Tlk2JRVnnvuOdv1atWqUb16dQIDA1m/fj2PPfbYPR1z8ODBDBgwwPZ1ZGQk/v7+Ga41NwgJgeHDYfVquHkT3NzsXZGIiIhjiY6OxmKxULp0ab766isaN25s75IkBaGhcO0alC4NDzxg72pEREQcn0NOoytWrBjOzs6cPXs2ye1nz57F19c3xe/x9fVN1/4A5cuXp1ixYvz111+2Y9zegDw2NpZ//vkn1eO4urri5eWVZBOjVi1zUnb9Ovz0k72rERERcTzVqlXDarVy6tQpmjZtSs2aNZk0aVKSaf5ifwntPlu3BovFrqWIiIjkCA4ZNrm4uFC7dm1CQ0Ntt8XHxxMaGprqsPIGDRok2R9g7dq1dxyGfurUKS5evEjJkiVtx7h8+TK7d++27fPTTz8RHx9P/fr1M/KU8iSLRavSiYiI3Mmvv/7Kzp07eeWVV/D09OS3336jf//++Pn50aFDB1avXo2DruWSZ8TFJZ7HaAqdiIhI2jjsanTz58+nS5cufPbZZ9SrV48JEybw7bffcvDgQXx8fOjcuTN+fn6MGjUKgK1bt9K4cWM+/PBDWrZsybx58xg5ciR79uyhatWqXL16leHDh9O+fXt8fX05cuQIb731FleuXOH333/H1dUVgCeffJKzZ88ybdo0YmJi6NatG3Xq1GHOnDlpqjsnrKaSnVavhubNoVQps0Kdk0PGmyIikhc52nv2jRs3+Pbbb5kxYwabN28GwGKx4Ofnx40bN/jnn3+0Gp0dbNkCDz8M3t5w/jzkz2/vikREROwnR69GB9ChQwfGjh3L0KFDqVmzJnv37mXVqlW2JuAnTpwgLCzMtn/Dhg2ZM2cOn3/+OTVq1OC7775j8eLFVK1aFQBnZ2d+++03QkJCqFSpEt27d6d27dps2rTJFjQBfPPNNwQFBfHYY4/RokULHn74YT7//PPsffIpuXDBLO2WwzRpAh4ecOYM3DJgTERERG5ToEABunTpwsaNGzl06BBvvfUWPj4+tpHYAA899BCff/45ERERdq4270iYQvfUUwqaRERE0sphRzblVFn2yVzLlrB3L7RvD88+Cw0b5phhQs88A999B0OGwPvv27saERERw5FH0ySIi4tjxYoV/O9//+OHH34gLi4Oi8WCq6srISEhzJs3z94lZgpH/VlYrVCxIhw5AgsWwNNP27siERER+8rxI5vkFlFRsGuXGR40eTI88gj4+8Prr8OmTQ4/4kl9m0RERO6Ns7MzISEhLF26lJMnT/LBBx8QGBjIzZs3WbBggb3Ly/X++MMETS4uEBxs72pERERyDoVNOYGrq2l4tGwZdO5smgYkBE+NGpkl3/r0cdjgqWVLcHaG336DY8fsXY2IiEjO5Ovry+DBg/nzzz9Zt24dL7zwgr1LyvUSptA1awaennYtRUREJEdR2JRTuLqaZgFffAFnz8Ly5YnBU1gYTJmSNHjauNEsn+IAihQxjTXB5GUiIiKSMY0bN+aLL76wdxm53pIl5lKr0ImIiKSPwqacyNXVDBe6NXjq0iVp8NS4sQmeeveGDRvsHjxpKp2IiIjkJKdOmS4GFgu0amXvakRERHIWhU05XULwNHs2nDsHK1aY4KlQIQgPh6lTzZJwpUtDr152C54STtLWrwctoOMAIiPhk0/MaLmhQ81aziIiImKT8AHZgw+Cr699axEREclpFDblJi4u0KKFCZ7OnjXBU9euicHTJ5+Y4MnPzwRP69dnW/BUsSJUrgyxsfDDD9nykJKSvXvh1VehVCnzO7BihVkisGxZMwru6FF7VygiIuIQEvo1aQqdiIhI+ilsyq0SgqdZs0zwtHIldOsGhQubrz/5BB591ARP//kPrFuX5cGTptLZyY0b8OWX0KAB1KoFn38O166Z9G/4cKhb1+wzdapJBTt1Mt3cRURE8qjLl82pEShsEhERuRcKm/ICFxd48kmYOdOMcPrhh6TB06efQtOmZrRLFgZPCWHTwoXQvTv8+KPdW0nlbocPw3//a6ZQdukC27dD/vzw3HNmVNv+/WYK3Y4dEBoKTzxhfiBz5kCNGias3LABrFZ7PxMREZFs9cMPZjR25cpQqZK9qxEREcl5FDblNS4u0Ly5CZ7OnjVnUy+9ZIKnc+eSBk89e8JPP5mzrUxQvz7UqQNRUebhH3/cDKx6/XXYtk2ZRqaIiYHvvzcvbqVKMH48/POPmSY3ciScPAlz55oG8haL+R6LxfzMV6+GPXugQwdwcjK/G02aQMOGZi5BfLw9n5mIiEi2SZhC17q1XcsQERHJsSxWq/7Ez0yRkZF4e3sTERGBl5eXvctJu5gYEywtWACLFpmAIkHx4tCuHTz7LDRqBPny3fPDxMXBpk0m7/juu6QPU66cGXTTsSNUq5aB55IXnToF06ebLSzM3GaxmObxr71mAkZn57Qf78gRGDfOpIJRUea2oCB480144QUTWoqI5HA59j07F3Kkn0VUFBQrBlevmkHB9evbtRwRERGHktb3bIVNmcyRTpbuWVqCp2eeMaNjMhA8RUfD2rVm1taSJaaNUIL77zehU8eOUL58Bp5LbhYfb17AadNg2bLEOYklSsDLL8Mrr5gRTRlx9ixMmmT6OSUsI+jnB/37m+N7embs+CIidpQr3rNzCUf6Wfzwg5lJXrKk+SzHSfMAREREbBQ22YkjnSxlipgY08MpIXi6eDHxvuLFoW1bEzw1aZKh4OnaNVi+3Ix4+uEHE0QlqFfPhE4dOpgTvzzvwgXT+P2zz8wIpASNG5upj23bZv7Io8hI01j844/hzBlzW6FCZkW71183AZeISA6T696zczBH+lm89pp5i33tNdNdQEQkJXFxccTExNi7DJFMlz9/fpzvMCtGYZOdONLJUqaLiTGNpb/9NnnwVKyYCTmefTbDwdPly6aJ+Ny5ZoBVQqsgi8UsoNexI7Rvb9pM5RlWK2zdas56FyxITOO8vU3z79deM11Ms1pUFHz9NYwZA4cOmdvc3EzD+Tfe0DA0EclRcvV7dg7jKD+L+HgzgDc8HFatguBgu5UiIg7KarUSHh7O5cuX7V2KSJYpVKgQvr6+WBL6/N5CYZOdOMrJUpZLCJ4WLDDJUErB0zPPmHQoA8FTeLjJtubONX0TEuTPb9oQdexoVrkrWPDen4pDu3LFhDuffgq//554e+3aZhTTc8/Z58nHx5u5jx9+CDt3mtucnEzYOHAg1KyZ/TWJiKRTnnnPzgEc5WexfTs0aGBmiZ8/D66uditFRBxUWFgYly9fpkSJEri7u6f4x7hITmW1Wrl+/Trnzp2jUKFClExhapHCJjtxlJOlbBUbmzR4unAh8b6iRRNHPGUweDp6FObNM8HTrbmLu7sJnDp2NAFUruhd/euvphfT11+bDqUABQqYJ9mzp1nWzxFYrbBhA4webT4CThAcbEKnJk0SV70TEXEwefI920E5ys9i0CDzltahgznnEBG5VVxcHH/++SclSpSgaNGi9i5HJMtcvHiRc+fOUalSpWRT6hQ22YmjnCzZTWysCR++/Tb14KlFCzPyJSDgnoOI/ftN6DR3Lvz9d+LthQubKXYdO5oWRulZgM3ubt40gd2nn8K2bYm3BwWZaXKdOzv23MG9e+Gjj2D+/MS5j3XrmjP31q1z2A9DRPKCPP+e7UAc5WdRuTIcPGjOL557zm5liIiDunnzJkePHiUgIIACBQrYuxyRLHPjxg2OHTtGuXLlcHNzS3KfwiY7cZSTJYeQEDwljHg6fz7p/Z6eUK0a1KgB1aubrWpVSMfrZrXCrl3mpHDePDPtLkHJkmZAVceOpsm4ww6w+esv04l01qzE6Yj58plV/3r2NKmZwxafgr//hnHjYOZME6ABVKoEb74JL76oOQki4jD0nu04HOFncfCgCZvy5zenLN7edilDRBxYQtiU0h/gIrnJnX7XFTbZiSOcLDmk2FjYuBG++840uv7jD9P3KSXlyiWGT9WrmzCqfPm7joyJizPZ1ty58P33cOlS4n3ly5tPKJ9/Hu6/PxOf172KjYVly8woprVrE28vUwZeeQW6dwdfX/vVlxnOnYPJk2HKFNP1HUwC2L8/vPpqukJFEZGsoPdsx+EIP4vRo81g3CeegNWr7VKCiDg4hU2SVyhsckCOcLKUI8TEmNXMfvst6Xb6dMr7u7ubUU+3hlDVqkGRIinuHh1tThTnzjV9rK9fT7yvWjUz2um550yula1On4b//Q+mT098rhYLPPmkmSrXokXum2525Yp5vuPHJz5nb2/4z3+gb1/w8bFvfSKSZ+k923E4ws+iQQPTIPyTT8zAYhGR2ylskrwiM8Imp6wuUiRF+fOb8Oj5582KZitXwqlTpsfTunUwcSK89JJphO3mZtKinTtNUPP666bxdNGi4O8PLVvC22+beXR//AGxsbi4QKtWMGeOGWAzd675On9+01z87bfNaKcGDWDSpKTT7zJdfLwZvdSuHZQtC8OGmdCleHHzEeqRI7BihSkwtwVNYKZLDhhgptfNnGl6UEVEwKhR5vV47TUzlVBEJDbWjIQ8edL8f75zJ4SGwuLFpi+cZIqpU6cSEBCAm5sb9evXZ2fCqqIpiImJYcSIEQQGBuLm5kaNGjVYdeuCEP86ffo0L7zwAkWLFqVAgQJUq1aNn3/+OSufRqYKC0tc9TYkxL61iIjkFAEBAUyYMCHN+69fvx6LxcLlhFkPkqtpZFMmc4RP5nKduDgTRtw+CurYsZT3d3WFKlWSjoKqXh1KlOCff0z7qLlzTaaV8Nvv5GQWy+vY0WRCmdKH++JF04fps8+ShimNGpmPTNu2zZv9i+LjYelSM18h4czeyQmeftqsYPfAA/atT9ImLs40NTlzxoSnZ84kXo+LM0FiQEDiVrp0hlajFAcVE2NGL165YlbOTOl6eu5L6POWkj59zKcDmSgvvmfPnz+fzp07M23aNOrXr8+ECRNYsGABhw4dokSJEsn2HzhwIF9//TXTp08nKCiI1atXM2DAALZu3UqtWrUAuHTpErVq1eLRRx+lZ8+eFC9enMOHDxMYGEhgYGCa6rL3z+Lzz80M73r1YMeObH94EckhcurIJstd+r++9957DBs2LN3HPX/+PAULFsTd3T1N+0dHR/PPP//g4+Nz15oyS1BQEEePHuX48eP45vQ2JdlI0+gckL1PlvKUiAjYty8xfPr1VzNs6erVlPf38UnSjPx8yerM/zWIrxe4JjmxdHExs9o6djSDjdL4f6dhtZoA5dNPzYp8UVHmdi8vs5rca685SNMoB2C1wqZNZmTbDz8k3v744yZ0ato0ZzVGzy2sVvNv6/YA6fbrYWEmVEorZ2cTON0aQCmMyn5RUZkXDF29mvh/XGbLn9+MivTwMJeenmap0TfeyNSHyYvv2fXr16du3bpMmTIFgPj4ePz9/enTpw+DBg1Ktn+pUqV455136NWrl+229u3bU6BAAb7++msABg0axJYtW9i0adM912Xvn0WLFuataORIGDw42x9eRHKInBo2hd8yjWP+/PkMHTqUQ4cO2W7z8PDAw8MDAKvVSlxcHPlywXnZ5s2b6dSpEw8//DDVq1dn4MCBdq0nJiaG/Pnz27WGtMqMsCnn/wZJ3uXtDQ89ZLYE8fFmxNPto6D++gvOnoU1a8wGFAd658tH76AgroRUZ1dUdb49WIOlx6uzZElJliyxULAgtG5tgqcnnjBBVIquXDFz9j791IReCWrVMqOYOnY0fzRJIovFjPJq1Mj8jD76yEyFXLvWbLVrm2mGbdvmzumF9nDjxp0DpITrN26k7XhOTibE9fODUqUSLy0WOH7c/Fs8dsxcj442l8ePm07+t1MYlX43b5p5wufOmf/fzp5NvH7rZUREYkiU2sIMGeXikhgK3RoQ3e3r1O7Li6M+s0F0dDS7d+9m8C1pipOTE82aNWPbtm0pfk9UVFSyk8wCBQqwefNm29dLly4lODiYZ555hg0bNuDn58d//vMfevTokWotUVFRRN0SVkZGRt7r08qwyEgzWxPMe76ISHpYrUn7w2YXd/e0fy5764geb29vLBaL7bb169fz6KOPsnLlSoYMGcLvv//OmjVr8Pf3Z8CAAWzfvp1r165RuXJlRo0aRbNmzWzHCggIoF+/fvTr1w8wI6imT5/OihUrWL16NX5+fowbN46Qf+cnJzzWpUuXKFSoELNnz6Zfv37Mnz+ffv36cfLkSR5++GFmzZpFyZIlAYiNjWXAgAF8+eWXODs78/LLLxMeHk5ERASLFy++4/OeMWMGzz//PI0bN6Zv377JwqZTp07x5ptvsnr1aqKioqhcuTJTp06lfv36ACxbtowRI0bw+++/4+HhwSOPPMKiRYtsz3XRokW0adPGdrxChQoxYcIEunbtyrFjxyhXrhzz5s3jk08+YceOHUybNo1WrVrRu3dvNm7cyKVLlwgMDOTtt9+mY8eOtuPEx8czduxYPv/8c06ePImPjw+vvvoq77zzDk2bNqVKlSq2D43AjDDz8/Pjhx9+4LHHHkvLr0S20Jm75C5OTqYZU/nycMs/fK5dg/37k46C+u030xtk3z489+2jKXNoCkwDrhUoyq/W6uy8Vp3f5lRn2JzqvFr4flo+XYCOHU0+4uyMGUn16afw9dfmDzkwPaaee86ETHXranROWlSvbl7D//s/GDcOZsyA3bvhmWegYkUzmqFzZ/PaSnKxsSZYuNtopFuXaLybIkVMcHRriJRwmXC9RIm0BUDx8aYxWkL4dPumMMqwWs1fvamFRrcHShn549zNLXNCoYTrqSbx4kguXLhAXFwcPrctzODj48PBgwdT/J7g4GDGjx9Po0aNCAwMJDQ0lIULFxJ3y8jGv//+m08//ZQBAwbw9ttvs2vXLl5//XVcXFzo0qVLiscdNWoUw4cPz7wnlwGrVpn/gipWhMqV7V2NiOQ016/b5zPlq1ehYMHMO96gQYMYO3Ys5cuXp3Dhwpw8eZIWLVrwwQcf4OrqypdffkmrVq04dOgQZcqUSfU4w4cP56OPPmLMmDFMnjyZTp06cfz4cYqksrDT9evXGTt2LF999RVOTk688MILvPHGG3zzzTcAjB49mm+++YZZs2ZRuXJlJk6cyOLFi3n00Ufv+HyuXLnCggUL2LFjB0FBQURERLBp0yYeeeQRAK5evUrjxo3x8/Nj6dKl+Pr6smfPHuLj4wFYsWIFbdu25Z133uHLL78kOjqalStX3tPrOm7cOGrVqoWbmxs3b96kdu3aDBw4EC8vL1asWMGLL75IYGAg9erVA2Dw4MFMnz6djz/+mIcffpiwsDDb+/TLL79M7969GTduHK7/fjj39ddf4+fnR9OmTdNdX1bK4WfmImlUsKBpxPDvP2DA/GF36lTyUVCHDlHwxkUaso6GrLPtHnfJicPTK/Lb9OpM8giiVcGfqHB2S+LxKlUy0+S6dEl1lTy5i4AAmDwZhg6FKVPM9cOHTSON996Dfv3Ma+ztbe9Ks4fVCv/8kxgWpRYmnT1rAp20KFDABEW3B0e3XpYsafbLLE5OiY/VsGHy+1MKo24dGXXs2N3DKCenu4dR9hi2HBdnFj642+ijhMv0TkvLn9+EfiVKmFFmKV0WKpQ0GPLwsM9rITnSxIkT6dGjB0FBQVgsFgIDA+nWrRszZ8607RMfH0+dOnUYOXIkALVq1WLfvn1MmzYt1bBp8ODBDBgwwPZ1ZGQk/v7+WftkUrFkibls00afD4lI3jVixAgef/xx29dFihShRo0atq/ff/99Fi1axNKlS+ndu3eqx+natattlM7IkSOZNGkSO3fupHnz5inuHxMTw7Rp02w9/nr37s2IESNs90+ePJnBgwfTtm1bAKZMmZKm0GfevHlUrFiR+/9tYfLcc88xY8YMW9g0Z84czp8/z65du2xBWIUKFWzf/8EHH/Dcc88l+WDk1tcjrfr160e7du2S3PbGLW0B+vTpw+rVq/n222+pV68eV65cYeLEiUyZMsX2HhoYGMjDDz8MQLt27ejduzdLlizh2WefBWD27Nl07do12/pgpZVDh01Tp05lzJgxhIeHU6NGDSZPnmxL+1KyYMEC3n33XY4dO0bFihUZPXo0LVq0AMwv8ZAhQ1i5ciV///033t7eNGvWjA8//JBSpUrZjhEQEMDx48eTHHfUqFEp9jGQHM5iMavZJaxol+DmTThwIEkAZf31V5zPnyeIQwRxCK4CVyGGfCymDUtK9sT1oUdpXNRCo0gIUNaUMcWLw/Dh8OabZgXCceNMMDhokGmo8dprJnj6d3itw7FaTWhw/Xr6tsuXk4dJ0dFpe0xnZ/N6pDYKKeG6t7fj/TWVljDq7Nk7j4yKioITJ8y2cWPKj5FZYdTNm3cfdZRweeFC4koEaeXhkXpw5OOT9HqhQo738xSHVaxYMZydnTl79myS28+ePZtq09TixYuzePFibt68ycWLFylVqhSDBg2ifPnytn1KlixJlSpVknxf5cqV+f7771OtxdXV1faJrD1FR5sFYSHpgGgRkbRyd0+9ZWxWP25mqlOnTpKvr169yrBhw1ixYgVhYWHExsZy48YNTpw4ccfjVK9e3Xa9YMGCeHl5ce7cuVT3d3d3T7KYRMmSJW37R0REcPbs2SQZgLOzM7Vr17aNQErNzJkzeeGFF2xfv/DCCzRu3JjJkyfj6enJ3r17qVWrVqojrvbu3XvH6eBpdfvrGhcXx8iRI/n22285ffo00dHRREVF2ZqsHzhwgKioqFSnw7m5ufHiiy8yc+ZMnn32Wfbs2cO+fftYunRphmvNbA4bNs2fP58BAwYkWS0lODg41dVStm7dSseOHRk1ahRPPfUUc+bMoU2bNuzZs4eqVaty/fp19uzZw7vvvkuNGjW4dOkSffv2JSQkJNnSvCNGjEjyi+Xp6Znlz1cciJub6bX07yo7ABar1fzh+NtvxO75jTOhf7DveiATr7zEj/tLEh8GzIKZs8z+ZcpA48aJW2Cg/h68Jx4eJlT6z3/MEoIffWSWQ//oI5gwwYwie/NNM/chrWJjTU+iuwU/166lPyy6dcvMtReKF085OLo1WCpe3AQquZGTkwnSSpaEBg2S358VYVTZsuYfbUqBUsKU2bSyWKBo0TsHSLdeZvbZo8i/XFxcqF27NqGhobYeE/Hx8YSGht7xU2owJ7d+fn7ExMTw/fff2z5NBXjooYeSNJoF+PPPPylbtmymP4fMtmGDaWtWogT826JDRCRdLJbMnc5mLwVvexJvvPEGa9euZezYsVSoUIECBQrw9NNPE32XD0Jvb4BtsVjuGAyltH9G1zD7448/2L59Ozt37kzSpykuLo558+bRo0cPCtxlFP/d7k+pzpgUemPe/rqOGTOGiRMnMmHCBKpVq0bBggXp16+f7XW92+OCmUpXs2ZNTp06xaxZs2jatKlDvuc6bNg0fvx4evToQbdu3QCYNm0aK1asYObMmSmOMpo4cSLNmzfnzTffBMwwv7Vr1zJlyhSmTZuGt7c3a9euTfI9U6ZMoV69epw4cSLJvFNPT08tiyhJWSzg6wu+vuR74gnKDIIyQAvMSeqWLeaEdcMG+Pln8/fsV1+ZDUwe0KhRYvgUFKTwKV1cXEyw9OKLsHw5jB4NW7fC9Olm5NNTT5k/5tMSAqV1pFBmyZ/fhAcFC5rLO22enolBUkKI5OurZsl3kx1h1O3y5099tNHtl8WK5fx+UpJrDBgwgC5dulCnTh3q1avHhAkTuHbtmu18q3Pnzvj5+TFq1CgAduzYwenTp6lZsyanT59m2LBhxMfH89Zbb9mO2b9/fxo2bMjIkSN59tln2blzJ59//jmff/65XZ5jeiT0lg0J0VoUIiK32rJlC127drVNX7t69SrHjh3L1hq8vb3x8fFh165dNGrUCDCB0Z49e6hZs2aq3zdjxgwaNWrE1KlTk9w+a9YsZsyYQY8ePahevTr/+9//+Oeff1Ic3VS9enVCQ0Nt74+3K168OGFhYbavDx8+zPU0dIrfsmULrVu3to26io+P588//7SNEK5YsSIFChQgNDSUl19+OcVjVKtWjTp16jB9+nTmzJmTpFm4I3HIs997WS1l27ZtSeb+g2lqeacO9REREVgsFgoVKpTk9g8//JD333+fMmXK8Pzzz9O/f/9Ul350pNVUxD68vc2Syf/O2OTqVZODbNxowqedO81sqHnzzAZmEMqt4VPVqrl3UEqmcnIyfxGEhMDmzfDhh2b+w7Jl93a8OwU/aQmH7rYVKKDeOI7gXsMoiyX1IMkRpyOKpEGHDh04f/48Q4cOJTw8nJo1a7Jq1Spb0/ATJ07gdMsb0s2bNxkyZAh///03Hh4etGjRgq+++irJuVPdunVZtGgRgwcPZsSIEZQrV44JEybQqVOn7H566WK1Ju3XJCIiiSpWrMjChQtp1aoVFouFd999965T17JCnz59GDVqFBUqVCAoKIjJkydz6dKlVPsTxcTE8NVXXzFixAiqVq2a5L6XX36Z8ePHs3//fjp27MjIkSNp06YNo0aNomTJkvzyyy+UKlWKBg0a8N577/HYY48RGBjIc889R2xsLCtXrrSNlGratClTpkyhQYMGxMXFMXDgwGSjtFJSsWJFvvvuO7Zu3UrhwoUZP348Z8+etYVNbm5uDBw4kLfeegsXFxceeughzp8/z/79++nevXuS59K7d28KFixoCwQdjUOGTfeyWkp4eHiK+4eHh6e4/82bNxk4cCAdO3bEy8vLdvvrr7/OAw88QJEiRdi6dSuDBw8mLCyM8ePHp3gcR1pNRRyDhwc88YTZwMzY2r49MXzatg3On4fvvzcbmH7ijzySGEDVrKlPWO/q4YfNKKfffzdhk7Nz+oIgNzeFBWLcLYwSyWV69+6d6rS59evXJ/m6cePG/PHHH3c95lNPPcVTTz2VGeVlm927TXu8ggXBgVaKFhFxCOPHj+ell16iYcOGFCtWjIEDB9plYMXAgQMJDw+nc+fOODs788orrxAcHIxzKn8sLV26lIsXL6YYwFSuXJnKlSszY8YMxo8fz5o1a/jvf/9LixYtiI2NpUqVKrbRUE2aNGHBggW8//77fPjhh3h5edlGVwGMGzeObt268cgjj1CqVCkmTpzI7t277/p8Ej7ACQ4Oxt3dnVdeeYU2bdoQERFh2+fdd98lX758DB06lDNnzlCyZElee+21JMfp2LEj/fr1o2PHjrg56IrdFmtGJ0RmgTNnzuDn58fWrVtpcMuJ/1tvvcWGDRvYsWNHsu9xcXHhiy++sHW+B/jkk08YPnx4skaYMTExtG/fnlOnTrF+/fokYdPtZs6cyauvvsrVq1dTbGSZ0sgmf39/IiIi7nhcybuiomDXrsTwacsW0x7oVl5eJktJCJ9q19YAGRGRzBYZGYm3t7fesx2APX4WQ4bABx/A00/DggXZ8pAiksPdvHmTo0ePUq5cOYf9Az+3i4+Pp3Llyjz77LO8//779i7Hbo4dO0ZgYCC7du3igQceyPTj3+l3Pa3v2Q45suleVkvx9fVN0/4xMTE8++yzHD9+nJ9++umuJzT169cnNjaWY8eOcd999yW731FWU5Gcw9XVBEkPPwxvvw0xMbBnT2LPp82bITISVq40G5hPXRs2NMFTo0ZQr57a+IiIiGREQqeF1q3tWoaIiNzB8ePHWbNmDY0bNyYqKoopU6Zw9OhRnn/+eXuXZhcxMTFcvHiRIUOG8OCDD2ZJ0JRZHLJLzK2rpSRIWC2lQSpTHBo0aJBkf4C1a9cm2T8haDp8+DA//vgjRYsWvWste/fuxcnJKcUV8EQyQ/78ZgWct94y7Yf++ccM7R8/3pwAFyliRj6tXWs+hW3UyKx2/uijMGwYrFtnpuqJiIhI2hw+DPv3mxnYLVvauxoREUmNk5MTs2fPpm7dujz00EP8/vvv/Pjjj1SuXNnepdnFli1bKFmyJLt27WLatGn2LueOHHJkE6R/tZS+ffvSuHFjxo0bR8uWLZk3bx4///yzbSWUmJgYnn76afbs2cPy5cuJi4uz9XMqUqQILi4ubNu2jR07dvDoo4/i6enJtm3b6N+/Py+88AKFCxe2zwsheY6zMzzwgNn69ze9i/fvTxz5tHGjWYV9/XqzgQms6tVLbDjesKHpHSUiIiLJJTQGb9IEdIonIuK4/P392bJli73LcBhNmjTBATshpchhw6b0rpbSsGFD5syZw5AhQ3j77bepWLEiixcvtnWgP336NEuXLgVItkziunXraNKkCa6ursybN49hw4YRFRVFuXLl6N+/f7JV7kSyk5MTVKtmtt69zeo5hw4lhk8bNpjV7rZsMdvIkSawql07MXx6+GGzeJaIiIhoFToREZGs5pANwnMyNRuV7Ga1wpEjiQ3HN2yA48eT7uPkBDVqJIZPjzwCaZhFKiKSq+k923Fk58/i3Dnw9TXvn8ePQ5kyWfpwIpKLqEG45BW5tkG4iKSdxQIVKpjtpZfMbcePJ06527AB/voLfvnFbBMmmH2qVTP9n6pUgbJlEzdPT7s9FRERkSy3bJkJmh54QEGTiIhIVlHYJJILlS0LnTubDeD0aRM8JYRPBw7A77+b7XaFC5uT71sDqFu34sVNwCUiIpITJaxCpyl0IiIiWUdhk0ge4OcHHTuaDcwUgo0bYetW+PtvMxLq+HG4dClx+/XXlI9VoMCdw6hSpSCf/mcREREHdPWqWd0VFDaJiIhkJf1JKJIHlSgBTz9ttltduZIYPKW0hYXBjRumQfmhQykf29kZSpdOGkDdGk6VKWMCKxERkey2Zg1ERUH58vDvGjIiIpJGTZo0oWbNmkz4ty9HQEAA/fr1o1+/fql+j8ViYdGiRbTJYMKfWceR7KOwSURsPD3NyXdqJ+BRUXDypAmeTpxIHkadPAkxMYlfp6ZEidRHRpUtC4UKZcnTExGRPC5hCl3r1poSLiJ5R6tWrYiJiWHVqlXJ7tu0aRONGjXi119/pXr16uk67q5duyhYsGBmlQnAsGHDWLx4MXv37k1ye1hYGIULF87Ux0rNjRs38PPzw8nJidOnT+Pq6potj5vbKGwSkTRzdU1sRp6SuDgID7/z6Khr18w0vnPnYNeulI/j5XXnMKpECbPCnoiISFrFxMDy5ea6PhgXkbyke/futG/fnlOnTlG6dOkk982aNYs6deqkO2gCKF68eGaVeFe+vr7Z9ljff/89999/P1arlcWLF9OhQ4dse+zbWa1W4uLiyJcD+5TozzURyTTOzqY/VMOGpj/UoEHw6aewciXs32+m6V24ALt3w6JFZmW8/v2hXTuoXRuKFTPHiYw0zcuXL4epU+Gtt6BDB3jwQShZEtzdoWJFaNYMuneH4cPNfnPnwurVsHOnWYHvn39MACYiIrJpk+lJWKyYeZ8SEckrnnrqKYoXL87s2bOT3H716lUWLFhA9+7duXjxIh07dsTPzw93d3eqVavG3Llz73jcgIAA25Q6gMOHD9OoUSPc3NyoUqUKaxOa5N1i4MCBVKpUCXd3d8qXL8+7775LTEwMALNnz2b48OH8+uuvWCwWLBaLrWaLxcLihOGpwO+//07Tpk0pUKAARYsW5ZVXXuHq1au2+7t27UqbNm0YO3YsJUuWpGjRovTq1cv2WHcyY8YMXnjhBV544QVmzJiR7P79+/fz1FNP4eXlhaenJ4888ghHjhyx3T9z5kzuv/9+XF1dKVmyJL179wbg2LFjWCyWJKO2Ll++jMViYf369QCsX78ei8XCDz/8QO3atXF1dWXz5s0cOXKE1q1b4+Pjg4eHB3Xr1uXHH39MUldUVBQDBw7E398fV1dXKlSowIwZM7BarVSoUIGxY8cm2X/v3r1YLBb++uuvu74m9yLnxWMikmNZLFC0qNkeeCDlfa5dS3mKXsJ25oyZzvfXX2ZLy2N6e0ORIv/f3r0HRXXdcQD/7i4sLAgrakGoIMTgm4CRR9DGx4AhSmywRMFBBSE6bYWAhBgwoiZRfMVHCFZrRzFOQtG0Yoy2OBQzahCVaMho6yNGqhTlYZCn4bW7/WPD6vJGV+7V/X5m7uzu2bN7f3cPsD9/nnvug83GpmePzc0Ne/xERCScL7/U3s6cyQtZEJEBaTTA/ft9v18Lix6fD2xiYoIFCxZg7969eO+99yD55XVffPEFVCoV5s6di7q6OowfPx7vvvsurK2tcfToUcyfPx/Dhg2Dt7d3t/tQq9X43e9+Bzs7O5w9exbV1dUdruVkZWWFvXv3wsHBARcvXsSiRYtgZWWFZcuWISQkBJcuXUJ2draukKJUKtu9R319PQICAuDr64uCggKUl5fjzTffRHR0tF5B7euvv4a9vT2+/vprXL9+HSEhIfDw8MCiRYs6PY4ff/wR+fn5OHjwIDQaDZYuXYqbN29i6NChAICSkhJMmjQJU6ZMwfHjx2FtbY28vDy0tLQAAHbs2IH4+HisX78e06dPR3V1NfLy8rr9/NpKTEzERx99hOeeew42NjYoLi7GjBkzsHbtWpiZmWHfvn2YOXMmrl69CicnJwDAggULkJ+fj9TUVLi7u6OoqAh3796FRCJBZGQk0tPTkZCQoNtHeno6Jk2ahOc7O23lMfGrlohExdISGDVKu3WkuRkoKWm/VlRlpXa7d+/B/bo67fd/VZV2u3Gjd7EoFL0rTrVu1tZcC4SISEw0Gv31moiIDOb+faBfv77fb12dNnHuocjISGzatAknTpzAlClTAGiLDcHBwVAqlVAqlXqFiJiYGBw7dgwHDhzoUbHpX//6F65cuYJjx47BwcEBAJCSkoLp06fr9VuxYoXuvrOzMxISEpCZmYlly5ZBoVCgX79+MDEx6fK0uYyMDDQ0NGDfvn26NaPS0tIwc+ZMbNiwAXZ2dgAAGxsbpKWlQSaTYeTIkQgMDERubm6XxaY9e/Zg+vTpuvWhAgICkJ6ejtWrVwMAtm/fDqVSiczMTJiamgIAhg8frnv9mjVr8PbbbyM2NlbX5uXl1e3n19YHH3yAadOm6R4PGDAA7u7uuscffvghsrKycPjwYURHR+PatWs4cOAAcnJy4O/vDwB47rnndP0jIiKwcuVKnDt3Dt7e3mhubkZGRka72U6GxGITET1VTE0BZ2ft1p2mJm2RqbX41LYY1dVjtVp75b2SEu3WG1Jp7wtUNjbaTS5/hA+FiIi6VFionTWrUAAP5e5EREZj5MiRmDBhAvbs2YMpU6bg+vXrOHXqFD744AMAgEqlQkpKCg4cOICSkhI0NTWhsbERFhYWPXr/y5cvw9HRUVdoAgBfX992/fbv34/U1FT8+OOPqKurQ0tLC6ytrXt1LJcvX4a7u7ve4uQTJ06EWq3G1atXdcWmMWPGQCaT6frY29vj4sWLnb6vSqXCp59+io8//ljXNm/ePCQkJGDlypWQSqUoLCzEyy+/rCs0Pay8vBy3b9+Gn59fr46nI56ennqP6+rqsHr1ahw9ehR37txBS0sLfv75Z9y6dQuA9pQ4mUyGyZMnd/h+Dg4OCAwMxJ49e+Dt7Y2vvvoKjY2NmD179mPH2hkWm4jomSWXaxcTt7Xt3evUau36Uh0Vo7orWP38s/b1P/2k3XqrXz/taYYDBnR+27bNxoanhBARdaV1VlNAgPbMEyIig7Gw0M4yEmK/vRQVFYWYmBhs374d6enpGDZsmK44sWnTJnz88cfYtm0b3NzcYGlpibi4ODQ1NRks5Pz8fISFheH9999HQECAbobQ5s2bDbaPh7UtCEkkEqjV6k77Hzt2DCUlJe0WBFepVMjNzcW0adOgUCg6fX1XzwGA9JcrHGk0Gl1bZ2tItb3KX0JCAnJycvDRRx/h+eefh0KhwBtvvKEbn+72DQBvvvkm5s+fj61btyI9PR0hISE9LiY+Cv7zhIioDalUu86TUgm4uPTutQ0NPZ899fDjqirtaR51ddrt5s3e7ffhdal6WqTq31+7qDsR0bOudb0mXoWOiAxOIunV6WxCmjNnDmJjY5GRkYF9+/bhD3/4g279pry8PLz++uuYN28eAO0aTNeuXcPo0aN79N6jRo1CcXEx7ty5A3t7ewDAmTNn9PqcPn0aQ4cOxXvvvadru9km6ZXL5VB1c4WfUaNGYe/evaivr9cVZfLy8iCVSjFixIgexduR3bt3IzQ0VC8+AFi7di12796NadOm4YUXXsCnn36K5ubmdsUsKysrODs7Izc3F1OnTm33/q1X77tz5w7GjRsHAHqLhXclLy8PERERmDVrFgDtTKf//ve/uufd3NygVqtx4sQJ3Wl0bc2YMQOWlpbYsWMHsrOzcfLkyR7t+1Gx2EREZEDm5tor5v3yHdtjKpX+KX8//dT+fke31dXa11dXa7eiop7vUyLRFpx6W6RSKrUFOSKip0FREfD999q/W6+9JnQ0RETC6devH0JCQpCUlISamhpERETonnN1dcXf/vY3nD59GjY2NtiyZQvKysp6XGzy9/fH8OHDER4ejk2bNqGmpqZd0cbV1RW3bt1CZmYmvLy8cPToUWRlZen1cXZ2RlFREQoLCzFkyBBYWVnBzMxMr09YWBhWrVqF8PBwrF69GhUVFYiJicH8+fN1p9D1VkVFBb766iscPnwYY8eO1XtuwYIFmDVrFiorKxEdHY1PPvkEoaGhSEpKglKpxJkzZ+Dt7Y0RI0Zg9erV+P3vfw9bW1tMnz4dtbW1yMvLQ0xMDBQKBV566SWsX78eLi4uKC8v11vDqiuurq44ePAgZs6cCYlEguTkZL1ZWs7OzggPD0dkZKRugfCbN2+ivLwcc+bMAQDIZDJEREQgKSkJrq6uHZ7maEgsNhERiYBM9uBKfb3R0qItUnVVkOqoaFVbq51Jde+ednvoaq3denhNqo4KUv37A1ZW2lMCW7e2jy0sWLAior7ROqvp5Zd7/zeWiOhZExUVhd27d2PGjBl66yutWLECN27cQEBAACwsLLB48WIEBQWhuvV/NrshlUqRlZWFqKgoeHt7w9nZGampqXj11Vd1fX77299i6dKliI6ORmNjIwIDA5GcnKxbfBsAgoODcfDgQUydOhVVVVVIT0/XK4oBgIWFBY4dO4bY2Fh4eXnBwsICwcHB2LJlyyN/Lq2LjXe03pKfnx8UCgU+++wzvPXWWzh+/DjeeecdTJ48GTKZDB4eHpg4cSIAIDw8HA0NDdi6dSsSEhIwaNAgvPHGG7r32rNnD6KiojB+/HiMGDECGzduxCuvvNJtfFu2bEFkZCQmTJiAQYMG4d1330VNTY1enx07dmD58uX44x//iJ9++glOTk5Yvny5Xp+oqCikpKRg4cKFj/Ix9YpE8/AJg/TYampqoFQqUV1d3euFzoiI+kpzc+9nUVVWAvX1htl/64zzrgpSvW3rxRWAiQDwO1tMnuRYTJkCnDgBbN0KdHAVbiKiHmtoaEBRURFcXFxgbm4udDhEvXbq1Cn4+fmhuLi4y1lgXf2s9/Q7mzObiIiMkKkpYGen3XqjsbHzwlTr/aqqB2tPPbzV1j5YP/Ph9akM5eEC1uMWrlpnXykUXHyd6Gl29y5w6pT2/uuvCxsLERGRUBobG1FRUYHVq1dj9uzZj3y6YW8whSYioh4zM3u0NalaqdXaK/a1LUB1VJTqTRugX8AqLTXcMZuaaotOCsWDAtSTeGxhoV3zyxhmZ6lU2tl1XW1NTfqPHRyA4cOFjpyeNkePav/uuLv3/oIPREREz4q//vWviIqKgoeHB/bt29cn+2SxiYiI+oxUqp19ZGnZ+1lVnXm4gGWo4tXDM65aix1tTot/YszNDVvIUqu7L+z0ZGtb/Hmc7VFO4F+yBEhLM/znTc+2Q4e0t5zVRERExiwiIqLd2ldPGotNRET0VHtSBayGBm0R6+efgfv3H9zv7eOe9G1pebDvhgbtdu+eYY7laWFq2vVmayt0hPS0uX8fOHZMez8oSNBQiIiIjA6LTURERG1IpdqZQRYWfbO/lpZHL1R195xU2n0hp3WTy3ve15DvJ5MZx+mD1Lfu3gWmTgWuXwc8PISOhoiIyLiw2ERERCQwExPtwuRWVkJHQvTscHLSrtnU0sJiJhEZllqtFjoEoifKED/jLDYRERER0TOLV5QkIkORy+WQSqW4ffs2fvWrX0Eul0PCajY9QzQaDZqamlBRUQGpVAq5XP7I78WvXyIiIiIiIqJuSKVSuLi44M6dO7h9+7bQ4RA9MRYWFnBycoJUKn3k92CxiYiIiIiIiKgH5HI5nJyc0NLSApVKJXQ4RAYnk8lgYmLy2LP2WGwiIiIiIiIi6iGJRAJTU1OYmpoKHQqRaD36nKg+sH37djg7O8Pc3Bw+Pj44d+5cl/2/+OILjBw5Eubm5nBzc8M//vEPvec1Gg1WrlwJe3t7KBQK+Pv744cfftDrU1lZibCwMFhbW6N///6IiopCXV2dwY+NiIiIiIiIiOhZJNpi0/79+xEfH49Vq1bhwoULcHd3R0BAAMrLyzvsf/r0acydOxdRUVH47rvvEBQUhKCgIFy6dEnXZ+PGjUhNTcXOnTtx9uxZWFpaIiAgAA0NDbo+YWFh+Pe//42cnBwcOXIEJ0+exOLFi5/48RIRERERERERPQskGo1GI3QQHfHx8YGXlxfS0tIAaC+95+joiJiYGCQmJrbrHxISgvr6ehw5ckTX9tJLL8HDwwM7d+6ERqOBg4MD3n77bSQkJAAAqqurYWdnh7179yI0NBSXL1/G6NGjUVBQAE9PTwBAdnY2ZsyYgf/9739wcHDoNu6amhoolUpUV1fD2traEB8FERERPQH8zhYPjgUREdHToaff2aJcs6mpqQnnz59HUlKSrk0qlcLf3x/5+fkdviY/Px/x8fF6bQEBATh06BAAoKioCKWlpfD399c9r1Qq4ePjg/z8fISGhiI/Px/9+/fXFZoAwN/fH1KpFGfPnsWsWbPa7bexsRGNjY26x9XV1QC0A0BERETi1fpdLdL/dzMqrWPA/ImIiEjcepo/ibLYdPfuXahUKtjZ2em129nZ4cqVKx2+prS0tMP+paWluudb27rqY2trq/e8iYkJBgwYoOvT1rp16/D++++3a3d0dOzs8IiIiEhEamtroVQqhQ7DqNXW1gJg/kRERPS06C5/EmWx6WmSlJSkN6NKrVajsrISAwcOfOxLBT6spqYGjo6OKC4u5vRyEeB4iAfHQlw4HuLC8eiaRqNBbW1tj06TpyfLwcEBxcXFsLKyYv70DON4iAfHQlw4HuLC8ehaT/MnURabBg0aBJlMhrKyMr32srIyDB48uMPXDB48uMv+rbdlZWWwt7fX6+Ph4aHr03YB8paWFlRWVna6XzMzM5iZmem19e/fv+sDfAzW1tb8gRcRjod4cCzEheMhLhyPznFGkzhIpVIMGTLkib0/fwfEheMhHhwLceF4iAvHo3M9yZ9EeTU6uVyO8ePHIzc3V9emVquRm5sLX1/fDl/j6+ur1x8AcnJydP1dXFwwePBgvT41NTU4e/asro+vry+qqqpw/vx5XZ/jx49DrVbDx8fHYMdHRERERERERPSsEuXMJgCIj49HeHg4PD094e3tjW3btqG+vh4LFy4EACxYsAC//vWvsW7dOgBAbGwsJk+ejM2bNyMwMBCZmZn49ttvsWvXLgCARCJBXFwc1qxZA1dXV7i4uCA5ORkODg4ICgoCAIwaNQqvvvoqFi1ahJ07d6K5uRnR0dEIDQ3lFHsiIiIiIiIioh4QbbEpJCQEFRUVWLlyJUpLS+Hh4YHs7GzdAt+3bt2CVPpgYtaECROQkZGBFStWYPny5XB1dcWhQ4cwduxYXZ9ly5ahvr4eixcvRlVVFX7zm98gOzsb5ubmuj6ff/45oqOj4efnB6lUiuDgYKSmpvbdgXfCzMwMq1atanfKHgmD4yEeHAtx4XiIC8eDjB1/B8SF4yEeHAtx4XiIC8fDMCQaXu+XiIiIiIiIiIgMRJRrNhERERERERER0dOJxSYiIiIiIiIiIjIYFpuIiIiIiIiIiMhgWGwiIiIiIiIiIiKDYbGJiIiIiIiIiIgMhsWmp8T27dvh7OwMc3Nz+Pj44Ny5c0KHZHTWrVsHLy8vWFlZwdbWFkFBQbh69arQYdEv1q9fD4lEgri4OKFDMVolJSWYN28eBg4cCIVCATc3N3z77bdCh2V0VCoVkpOT4eLiAoVCgWHDhuHDDz8ELz5Lxoj5k/CYP4kb8ydhMXcSD+ZPhsdi01Ng//79iI+Px6pVq3DhwgW4u7sjICAA5eXlQodmVE6cOIElS5bgzJkzyMnJQXNzM1555RXU19cLHZrRKygowJ///Ge88MILQoditO7du4eJEyfC1NQU//znP/Gf//wHmzdvho2NjdChGZ0NGzZgx44dSEtLw+XLl7FhwwZs3LgRn3zyidChEfUp5k/iwPxJvJg/CYu5k7gwfzI8iYalOtHz8fGBl5cX0tLSAABqtRqOjo6IiYlBYmKiwNEZr4qKCtja2uLEiROYNGmS0OEYrbq6Orz44ov405/+hDVr1sDDwwPbtm0TOiyjk5iYiLy8PJw6dUroUIzea6+9Bjs7O+zevVvXFhwcDIVCgc8++0zAyIj6FvMncWL+JA7Mn4TH3ElcmD8ZHmc2iVxTUxPOnz8Pf39/XZtUKoW/vz/y8/MFjIyqq6sBAAMGDBA4EuO2ZMkSBAYG6v2OUN87fPgwPD09MXv2bNja2mLcuHH4y1/+InRYRmnChAnIzc3FtWvXAADff/89vvnmG0yfPl3gyIj6DvMn8WL+JA7Mn4TH3ElcmD8ZnonQAVDX7t69C5VKBTs7O712Ozs7XLlyRaCoSK1WIy4uDhMnTsTYsWOFDsdoZWZm4sKFCygoKBA6FKN348YN7NixA/Hx8Vi+fDkKCgrw1ltvQS6XIzw8XOjwjEpiYiJqamowcuRIyGQyqFQqrF27FmFhYUKHRtRnmD+JE/MncWD+JA7MncSF+ZPhsdhE9AiWLFmCS5cu4ZtvvhE6FKNVXFyM2NhY5OTkwNzcXOhwjJ5arYanpydSUlIAAOPGjcOlS5ewc+dOJkx97MCBA/j888+RkZGBMWPGoLCwEHFxcXBwcOBYEJGgmD8Jj/mTeDB3EhfmT4bHYpPIDRo0CDKZDGVlZXrtZWVlGDx4sEBRGbfo6GgcOXIEJ0+exJAhQ4QOx2idP38e5eXlePHFF3VtKpUKJ0+eRFpaGhobGyGTyQSM0LjY29tj9OjRem2jRo3C3//+d4EiMl7vvPMOEhMTERoaCgBwc3PDzZs3sW7dOiZLZDSYP4kP8ydxYP4kHsydxIX5k+FxzSaRk8vlGD9+PHJzc3VtarUaubm58PX1FTAy46PRaBAdHY2srCwcP34cLi4uQodk1Pz8/HDx4kUUFhbqNk9PT4SFhaGwsJCJUh+bOHFiu0tZX7t2DUOHDhUoIuN1//59SKX6X+8ymQxqtVqgiIj6HvMn8WD+JC7Mn8SDuZO4MH8yPM5segrEx8cjPDwcnp6e8Pb2xrZt21BfX4+FCxcKHZpRWbJkCTIyMvDll1/CysoKpaWlAAClUgmFQiFwdMbHysqq3XoPlpaWGDhwINeBEMDSpUsxYcIEpKSkYM6cOTh37hx27dqFXbt2CR2a0Zk5cybWrl0LJycnjBkzBt999x22bNmCyMhIoUMj6lPMn8SB+ZO4MH8SD+ZO4sL8yfAkGo1GI3QQ1L20tDRs2rQJpaWl8PDwQGpqKnx8fIQOy6hIJJIO29PT0xEREdG3wVCHpkyZwkv3CujIkSNISkrCDz/8ABcXF8THx2PRokVCh2V0amtrkZycjKysLJSXl8PBwQFz587FypUrIZfLhQ6PqE8xfxIe8yfxY/4kHOZO4sH8yfBYbCIiIiIiIiIiIoPhmk1ERERERERERGQwLDYREREREREREZHBsNhEREREREREREQGw2ITEREREREREREZDItNRERERERERERkMCw2ERERERERERGRwbDYREREREREREREBsNiExERERERERERGQyLTUREREREREREZDAsNhERERERERERkcGw2ERERERERERERAbzf7M0hIxCpzPKAAAAAElFTkSuQmCC\n"
429 |           },
430 |           "metadata": {}
431 |         }
432 |       ]
433 |     },
434 |     {
435 |       "cell_type": "code",
436 |       "source": [],
437 |       "metadata": {
438 |         "id": "bc6ej33Ngdyb"
439 |       },
440 |       "execution_count": null,
441 |       "outputs": []
442 |     }
443 |   ]
444 | }


--------------------------------------------------------------------------------
/README.md:
--------------------------------------------------------------------------------
1 | # DL-Training
2 | jupyter notebook for mnist digit recognition using CNN, CNN feature map visualization and facical expression recognition
3 | 
4 | updated readme file
5 | 


--------------------------------------------------------------------------------