{"cells":[{"cell_type":"markdown","metadata":{"id":"n1-hDaq3vSHZ"},"source":["## Get Data"]},{"cell_type":"code","execution_count":1,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":1076,"status":"ok","timestamp":1640148499708,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"WqJULMydYxd7","outputId":"f5d9daf0-d5ff-44c7-d19f-8ca947e38c62"},"outputs":[{"name":"stdout","output_type":"stream","text":["Drive already mounted at /content/drive; to attempt to forcibly remount, call drive.mount(\"/content/drive\", force_remount=True).\n"]}],"source":["#@title Mount Google Drive (You don't need to run this if you are running notebooks on your laptop)\n","\n","from google.colab import drive\n","\n","# The following command will prompt a URL for you to click and obtain the\n","# authorization code\n","\n","drive.mount(\"/content/drive\")"]},{"cell_type":"code","execution_count":2,"metadata":{"executionInfo":{"elapsed":8,"status":"ok","timestamp":1640148499710,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"7vZsd4-2I-rF"},"outputs":[],"source":["data_path=\"/content/drive/MyDrive/Project/Dataset/Updated Dataset/\""]},{"cell_type":"code","execution_count":3,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":72278,"status":"ok","timestamp":1640148571984,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"xyz2OnkUzLRJ","outputId":"7a8d6e87-030f-498a-b893-9181fd7e5658"},"outputs":[{"name":"stdout","output_type":"stream","text":["1000\n","2000\n","3000\n","4000\n","5000\n","6000\n","7000\n","8000\n"]}],"source":["import json\n","import os\n","# import pandas as pd\n","import numpy as np\n","# df=pd.DataFrame()\n","count=0\n","dataset=[]\n","labels=[]\n","for folder in os.listdir(data_path)[:8]:\n","  for filename in os.listdir(data_path+folder)[::]:\n","    with open(data_path+folder+\"/\"+filename) as f:\n","      data = json.load(f)\n","      if(int(data['label'])\u003e0):\n","        # df=df.append(pd.json_normalize(data))\n","        li=[]\n","        for i in list(data['data'].keys()):\n","          npdata=np.array(data['data'][i])\n","          li.append(np.pad(npdata,(0,10000-len(npdata)),'constant'))  #some windows smaller than 10000\n","        ar=np.array(li)\n","        dataset.append(ar)\n","        labels.append(int(data['label']))\n","      count+=1\n","      if(count%1000==0):\n","        print(count)\n","dataset1=np.array(dataset)"]},{"cell_type":"code","execution_count":4,"metadata":{"executionInfo":{"elapsed":1300,"status":"ok","timestamp":1640148573280,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"TJzxbW9ldfFD"},"outputs":[],"source":["np.random.shuffle(dataset1)"]},{"cell_type":"code","execution_count":5,"metadata":{"executionInfo":{"elapsed":4955,"status":"ok","timestamp":1640148578231,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"XHGK4KfZ00Sh"},"outputs":[],"source":["#Required Imports\n","%matplotlib inline\n","# !pip install einops\n","# !pip install seaborn\n","# !pip install prettytable\n","# !pip install sklearn\n","# !pip install levenshtein\n","# !pip install tensorflow\n","import gc\n","import os\n","\n","import numpy as np\n","import pandas as pd\n","import matplotlib.pyplot as plt\n","import seaborn as sns\n","\n","from collections import Counter\n","import itertools\n","import random\n","from prettytable import PrettyTable\n","from IPython.display import Image\n","\n","from sklearn.preprocessing import LabelEncoder\n","from sklearn.metrics.pairwise import euclidean_distances\n","# from Levenshtein import distance as levenshtein_distance\n","\n","from tensorflow.keras.models import Model\n","from tensorflow.keras.regularizers import l2\n","from tensorflow.keras.constraints import max_norm\n","from tensorflow.keras.utils import to_categorical\n","from tensorflow.keras.preprocessing.text import Tokenizer\n","from tensorflow.keras.preprocessing.sequence import pad_sequences\n","from tensorflow.keras.callbacks import EarlyStopping\n","from tensorflow.keras.layers import Input, Dense, Dropout, Flatten, Activation\n","from tensorflow.keras.layers import Conv1D, Add, MaxPooling1D, BatchNormalization\n","from tensorflow.keras.layers import Embedding, Bidirectional, LSTM, GlobalMaxPooling1D\n"]},{"cell_type":"code","execution_count":6,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":9,"status":"ok","timestamp":1640148578233,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"ukuJ4-nOA5zF","outputId":"75e1cad9-e188-4afc-a761-728f8a15cbfe"},"outputs":[{"name":"stdout","output_type":"stream","text":["(969, 23, 10000)\n","(323, 23, 10000)\n","(324, 23, 10000)\n","969\n","323\n","324\n","\n"]}],"source":["x_train=dataset1[:int(len(dataset1)*.6)]\n","x_val=dataset1[int(len(dataset1)*.6):int(len(dataset1)*.8)]\n","x_test=dataset1[int(len(dataset1)*.8):]\n","#shape of data\n","print(x_train.shape)\n","print(x_val.shape)\n","print(x_test.shape)\n","y_train=labels[:int(len(labels)*.6)]\n","y_val=labels[int(len(labels)*.6):int(len(labels)*.8)]\n","y_test=labels[int(len(labels)*.8):]\n","#shape of labels\n","print(len(y_train))\n","print(len(y_val))\n","print(len(y_test))\n","print(\"\")\n","# One hot encoding of outputs\n"]},{"cell_type":"code","execution_count":7,"metadata":{"executionInfo":{"elapsed":1589,"status":"ok","timestamp":1640148579816,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"_LhrjhYrQSF4"},"outputs":[],"source":["from numpy.core.numeric import False_\n","is1=[]\n","for i in y_train:\n","  if i==1:\n","    is1.append(False)\n","  else:\n","    is1.append(True)\n","  \n","is1[::5]==True\n","x_train=np.asarray(x_train)[is1]\n","y_train=np.asarray(y_train)[is1]\n","\n","\n","is1=[]\n","for i in y_val:\n","  if i==1:\n","    is1.append(False)\n","  else:\n","    is1.append(True)\n","  \n","is1[::5]==True\n","x_val=np.asarray(x_val)[is1]\n","y_val=np.asarray(y_val)[is1]\n","\n","is1=[]\n","for i in y_test:\n","  if i==1:\n","    is1.append(False)\n","  else:\n","    is1.append(True)\n","  \n","is1[::5]==True\n","x_test=np.asarray(x_test)[is1]\n","y_test=np.asarray(y_test)[is1]\n","del is1"]},{"cell_type":"code","execution_count":8,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":11,"status":"ok","timestamp":1640148579817,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"-zRE_17aYezn","outputId":"e3fed282-65eb-4641-8772-33773f080560"},"outputs":[{"name":"stdout","output_type":"stream","text":["(607, 10)\n"]}],"source":["y_train_orig=y_train\n","y_val_orig=y_val\n","y_test_orig=y_test\n","\n","from tensorflow.keras.utils import to_categorical\n","y_train = to_categorical(y_train)\n","y_val = to_categorical(y_val)\n","y_test = to_categorical(y_test)\n","print(y_train.shape)"]},{"cell_type":"code","execution_count":9,"metadata":{"executionInfo":{"elapsed":7,"status":"ok","timestamp":1640148579817,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"gx3t9wgcE4W9"},"outputs":[],"source":["#Utility functions: plot model's accuracy and loss\n","\n","# https://realpython.com/python-keras-text-classification/\n","import matplotlib.pyplot as plt\n","plt.style.use('ggplot')\n","\n","def plot_history(history):\n","  acc = history.history['accuracy']\n","  val_acc = history.history['val_accuracy']\n","  loss = history.history['loss']\n","  val_loss = history.history['val_loss']\n","  x = range(1, len(acc) + 1)\n","\n","  plt.figure(figsize=(12, 5))\n","  plt.subplot(1, 2, 1)\n","  plt.plot(x, acc, 'b', label='Training acc')\n","  plt.plot(x, val_acc, 'r', label='Validation acc')\n","  plt.title('Training and validation accuracy')\n","  plt.legend()\n","\n","  plt.subplot(1, 2, 2)\n","  plt.plot(x, loss, 'b', label='Training loss')\n","  plt.plot(x, val_loss, 'r', label='Validation loss')\n","  plt.title('Training and validation loss')\n","  plt.legend()"]},{"cell_type":"markdown","metadata":{"id":"ca4BTauuEE6H"},"source":["#ResNet\n"]},{"cell_type":"code","execution_count":10,"metadata":{"executionInfo":{"elapsed":7,"status":"ok","timestamp":1640148579818,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"9X1MCAy-kT35"},"outputs":[],"source":["#Add your code to define a residual block below, as shown in the diagram above\n","def residual_block(data, filters, d_rate):\n","  \"\"\"\n","  _data: input\n","  _filters: convolution filters\n","  _d_rate: dilation rate\n","  \"\"\"\n","  \n","  skipdata = data\n","  bnorm1 = BatchNormalization()(data)\n","  activation1 = Activation('relu')(bnorm1)\n","  conv1 = Conv1D(filters, 1, dilation_rate=d_rate, padding='same', kernel_regularizer=l2(0.001))(activation1)\n","  bnorm2 = BatchNormalization()(conv1)\n","  activation2 = Activation('relu')(bnorm2)\n","  conv2 = Conv1D(filters, 3, padding='same', kernel_regularizer=l2(0.001))(activation2)\n","  x = Add()([conv2, skipdata])\n","\n","  return x"]},{"cell_type":"code","execution_count":11,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"executionInfo":{"elapsed":530,"status":"ok","timestamp":1640148580343,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"TPrgszi_kT36","outputId":"4550fbbb-d656-450d-f367-448e36add522"},"outputs":[{"name":"stdout","output_type":"stream","text":["Model: \"model\"\n","__________________________________________________________________________________________________\n"," Layer (type)                   Output Shape         Param #     Connected to                     \n","==================================================================================================\n"," input_1 (InputLayer)           [(None, 23, 10000)]  0           []                               \n","                                                                                                  \n"," conv1d (Conv1D)                (None, 23, 128)      1280128     ['input_1[0][0]']                \n","                                                                                                  \n"," batch_normalization (BatchNorm  (None, 23, 128)     512         ['conv1d[0][0]']                 \n"," alization)                                                                                       \n","                                                                                                  \n"," activation (Activation)        (None, 23, 128)      0           ['batch_normalization[0][0]']    \n","                                                                                                  \n"," conv1d_1 (Conv1D)              (None, 23, 128)      16512       ['activation[0][0]']             \n","                                                                                                  \n"," batch_normalization_1 (BatchNo  (None, 23, 128)     512         ['conv1d_1[0][0]']               \n"," rmalization)                                                                                     \n","                                                                                                  \n"," activation_1 (Activation)      (None, 23, 128)      0           ['batch_normalization_1[0][0]']  \n","                                                                                                  \n"," conv1d_2 (Conv1D)              (None, 23, 128)      49280       ['activation_1[0][0]']           \n","                                                                                                  \n"," add (Add)                      (None, 23, 128)      0           ['conv1d_2[0][0]',               \n","                                                                  'conv1d[0][0]']                 \n","                                                                                                  \n"," batch_normalization_2 (BatchNo  (None, 23, 128)     512         ['add[0][0]']                    \n"," rmalization)                                                                                     \n","                                                                                                  \n"," activation_2 (Activation)      (None, 23, 128)      0           ['batch_normalization_2[0][0]']  \n","                                                                                                  \n"," conv1d_3 (Conv1D)              (None, 23, 128)      16512       ['activation_2[0][0]']           \n","                                                                                                  \n"," batch_normalization_3 (BatchNo  (None, 23, 128)     512         ['conv1d_3[0][0]']               \n"," rmalization)                                                                                     \n","                                                                                                  \n"," activation_3 (Activation)      (None, 23, 128)      0           ['batch_normalization_3[0][0]']  \n","                                                                                                  \n"," conv1d_4 (Conv1D)              (None, 23, 128)      49280       ['activation_3[0][0]']           \n","                                                                                                  \n"," add_1 (Add)                    (None, 23, 128)      0           ['conv1d_4[0][0]',               \n","                                                                  'add[0][0]']                    \n","                                                                                                  \n"," max_pooling1d (MaxPooling1D)   (None, 7, 128)       0           ['add_1[0][0]']                  \n","                                                                                                  \n"," dropout (Dropout)              (None, 7, 128)       0           ['max_pooling1d[0][0]']          \n","                                                                                                  \n"," flatten (Flatten)              (None, 896)          0           ['dropout[0][0]']                \n","                                                                                                  \n"," dense (Dense)                  (None, 10)           8970        ['flatten[0][0]']                \n","                                                                                                  \n","==================================================================================================\n","Total params: 1,422,730\n","Trainable params: 1,421,706\n","Non-trainable params: 1,024\n","__________________________________________________________________________________________________\n"]}],"source":["#input\n","x_input = Input(shape=(23,10000))\n","\n","#initial conv\n","conv = Conv1D(128, 1, padding='same')(x_input) \n","\n","# per-residue representation\n","res1 = residual_block(conv, 128, 2)\n","res2 = residual_block(res1, 128, 3)\n","\n","#Max Pooling\n","x = MaxPooling1D(3)(res2)\n","x = Dropout(0.3)(x)\n","\n","#softmax classifier\n","x = Flatten()(x)\n","x_output = Dense(10, activation='softmax', kernel_regularizer=l2(0.0001))(x)\n","\n","#Compile your model\n","model2 = Model(inputs=x_input, outputs=x_output)\n","# model2.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n","model2.summary()"]},{"cell_type":"code","execution_count":12,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":1000},"executionInfo":{"elapsed":144975,"status":"ok","timestamp":1640148725314,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"hsYrxG7WIIcg","outputId":"fed2edd1-b23d-48e0-9d22-7eac745a41d8"},"outputs":[{"name":"stdout","output_type":"stream","text":["Epoch 1/50\n","19/19 [==============================] - 5s 166ms/step - loss: 80.0354 - accuracy: 0.1351 - val_loss: 112.7291 - val_accuracy: 0.1488\n","Epoch 2/50\n","19/19 [==============================] - 2s 126ms/step - loss: 81.5618 - accuracy: 0.2488 - val_loss: 99.6529 - val_accuracy: 0.1369\n","Epoch 3/50\n","19/19 [==============================] - 2s 126ms/step - loss: 67.4522 - accuracy: 0.2949 - val_loss: 82.4899 - val_accuracy: 0.1429\n","Epoch 4/50\n","19/19 [==============================] - 2s 126ms/step - loss: 47.3411 - accuracy: 0.3493 - val_loss: 83.6552 - val_accuracy: 0.1607\n","Epoch 5/50\n","19/19 [==============================] - 2s 129ms/step - loss: 54.1027 - accuracy: 0.3493 - val_loss: 94.0403 - val_accuracy: 0.1369\n","Epoch 6/50\n","19/19 [==============================] - 2s 130ms/step - loss: 48.7582 - accuracy: 0.3773 - val_loss: 107.9749 - val_accuracy: 0.1369\n","Epoch 7/50\n","19/19 [==============================] - 2s 127ms/step - loss: 37.6530 - accuracy: 0.3839 - val_loss: 75.3816 - val_accuracy: 0.1845\n","Epoch 8/50\n","19/19 [==============================] - 2s 126ms/step - loss: 48.1563 - accuracy: 0.3839 - val_loss: 127.7328 - val_accuracy: 0.1607\n","Epoch 9/50\n","19/19 [==============================] - 4s 216ms/step - loss: 44.2747 - accuracy: 0.4168 - val_loss: 111.4504 - val_accuracy: 0.1250\n","Epoch 10/50\n","19/19 [==============================] - 4s 182ms/step - loss: 30.0869 - accuracy: 0.4563 - val_loss: 98.4091 - val_accuracy: 0.1607\n","Epoch 11/50\n","19/19 [==============================] - 2s 126ms/step - loss: 26.1734 - accuracy: 0.4959 - val_loss: 91.6387 - val_accuracy: 0.1131\n","Epoch 12/50\n","19/19 [==============================] - 2s 128ms/step - loss: 30.4648 - accuracy: 0.5107 - val_loss: 100.8223 - val_accuracy: 0.1548\n","Epoch 13/50\n","19/19 [==============================] - 2s 128ms/step - loss: 33.7870 - accuracy: 0.4794 - val_loss: 109.4796 - val_accuracy: 0.1488\n","Epoch 14/50\n","19/19 [==============================] - 2s 127ms/step - loss: 37.8746 - accuracy: 0.4712 - val_loss: 106.5345 - val_accuracy: 0.1667\n","Epoch 15/50\n","19/19 [==============================] - 2s 128ms/step - loss: 33.2583 - accuracy: 0.5305 - val_loss: 91.6072 - val_accuracy: 0.1905\n","Epoch 16/50\n","19/19 [==============================] - 2s 128ms/step - loss: 22.8103 - accuracy: 0.5272 - val_loss: 90.5906 - val_accuracy: 0.1548\n","Epoch 17/50\n","19/19 [==============================] - 2s 126ms/step - loss: 29.7808 - accuracy: 0.5420 - val_loss: 85.4209 - val_accuracy: 0.1488\n","Epoch 18/50\n","19/19 [==============================] - 2s 127ms/step - loss: 21.2672 - accuracy: 0.5618 - val_loss: 91.4589 - val_accuracy: 0.1190\n","Epoch 19/50\n","19/19 [==============================] - 2s 127ms/step - loss: 19.9186 - accuracy: 0.5601 - val_loss: 79.5150 - val_accuracy: 0.1429\n","Epoch 20/50\n","19/19 [==============================] - 2s 126ms/step - loss: 17.5948 - accuracy: 0.6013 - val_loss: 85.2976 - val_accuracy: 0.1667\n","Epoch 21/50\n","19/19 [==============================] - 2s 127ms/step - loss: 22.5501 - accuracy: 0.6178 - val_loss: 86.7816 - val_accuracy: 0.1488\n","Epoch 22/50\n","19/19 [==============================] - 2s 125ms/step - loss: 16.5704 - accuracy: 0.6178 - val_loss: 89.5121 - val_accuracy: 0.1369\n","Epoch 23/50\n","19/19 [==============================] - 2s 125ms/step - loss: 15.7924 - accuracy: 0.6277 - val_loss: 88.5604 - val_accuracy: 0.1429\n","Epoch 24/50\n","19/19 [==============================] - 2s 125ms/step - loss: 16.1570 - accuracy: 0.6573 - val_loss: 91.9523 - val_accuracy: 0.1190\n","Epoch 25/50\n","19/19 [==============================] - 2s 128ms/step - loss: 18.5162 - accuracy: 0.6310 - val_loss: 97.7506 - val_accuracy: 0.1310\n","Epoch 26/50\n","19/19 [==============================] - 2s 126ms/step - loss: 11.1697 - accuracy: 0.6969 - val_loss: 84.7251 - val_accuracy: 0.1548\n","Epoch 27/50\n","19/19 [==============================] - 2s 127ms/step - loss: 10.6742 - accuracy: 0.6771 - val_loss: 83.1427 - val_accuracy: 0.1310\n","Epoch 28/50\n","19/19 [==============================] - 2s 126ms/step - loss: 10.9837 - accuracy: 0.6672 - val_loss: 79.1992 - val_accuracy: 0.1786\n","Epoch 29/50\n","19/19 [==============================] - 2s 127ms/step - loss: 12.7917 - accuracy: 0.6755 - val_loss: 86.4403 - val_accuracy: 0.1488\n","Epoch 30/50\n","19/19 [==============================] - 2s 126ms/step - loss: 11.5348 - accuracy: 0.6557 - val_loss: 89.0582 - val_accuracy: 0.1250\n","Epoch 31/50\n","19/19 [==============================] - 2s 126ms/step - loss: 13.6367 - accuracy: 0.6870 - val_loss: 95.6918 - val_accuracy: 0.1429\n","Epoch 32/50\n","19/19 [==============================] - 2s 126ms/step - loss: 10.6153 - accuracy: 0.7100 - val_loss: 101.5226 - val_accuracy: 0.1250\n","Epoch 33/50\n","19/19 [==============================] - 2s 128ms/step - loss: 13.1379 - accuracy: 0.6969 - val_loss: 92.1399 - val_accuracy: 0.1310\n","Epoch 34/50\n","19/19 [==============================] - 2s 125ms/step - loss: 9.2758 - accuracy: 0.7381 - val_loss: 89.5472 - val_accuracy: 0.1310\n","Epoch 35/50\n","19/19 [==============================] - 2s 126ms/step - loss: 12.0026 - accuracy: 0.7035 - val_loss: 87.6059 - val_accuracy: 0.1548\n","Epoch 36/50\n","19/19 [==============================] - 2s 126ms/step - loss: 8.8568 - accuracy: 0.7265 - val_loss: 82.4135 - val_accuracy: 0.1488\n","Epoch 37/50\n","19/19 [==============================] - 2s 127ms/step - loss: 9.1014 - accuracy: 0.7331 - val_loss: 81.9083 - val_accuracy: 0.1726\n","Epoch 38/50\n","19/19 [==============================] - 2s 126ms/step - loss: 8.8432 - accuracy: 0.7545 - val_loss: 85.2458 - val_accuracy: 0.1488\n","Epoch 39/50\n","19/19 [==============================] - 2s 126ms/step - loss: 9.8360 - accuracy: 0.7298 - val_loss: 87.4490 - val_accuracy: 0.1488\n","Epoch 40/50\n","19/19 [==============================] - 2s 126ms/step - loss: 9.0065 - accuracy: 0.7512 - val_loss: 85.6212 - val_accuracy: 0.1667\n","Epoch 41/50\n","19/19 [==============================] - 2s 126ms/step - loss: 9.7850 - accuracy: 0.7331 - val_loss: 85.4930 - val_accuracy: 0.2202\n","Epoch 42/50\n","19/19 [==============================] - 2s 127ms/step - loss: 12.4122 - accuracy: 0.7232 - val_loss: 95.7436 - val_accuracy: 0.1607\n","Epoch 43/50\n","19/19 [==============================] - 2s 126ms/step - loss: 10.7785 - accuracy: 0.7562 - val_loss: 102.5942 - val_accuracy: 0.1131\n","Epoch 44/50\n","19/19 [==============================] - 2s 125ms/step - loss: 10.3326 - accuracy: 0.7100 - val_loss: 96.4126 - val_accuracy: 0.1548\n","Epoch 45/50\n","19/19 [==============================] - 2s 126ms/step - loss: 13.5701 - accuracy: 0.7100 - val_loss: 95.0694 - val_accuracy: 0.1310\n","Epoch 46/50\n","19/19 [==============================] - 2s 128ms/step - loss: 8.8110 - accuracy: 0.7694 - val_loss: 90.9561 - val_accuracy: 0.0952\n","Epoch 47/50\n","19/19 [==============================] - 2s 126ms/step - loss: 9.1100 - accuracy: 0.7628 - val_loss: 80.8264 - val_accuracy: 0.1548\n","Epoch 48/50\n","19/19 [==============================] - 2s 126ms/step - loss: 8.2203 - accuracy: 0.7809 - val_loss: 85.5354 - val_accuracy: 0.1488\n","Epoch 49/50\n","19/19 [==============================] - 2s 126ms/step - loss: 11.3471 - accuracy: 0.7858 - val_loss: 75.4379 - val_accuracy: 0.1845\n","Epoch 50/50\n","19/19 [==============================] - 2s 128ms/step - loss: 9.2335 - accuracy: 0.7875 - val_loss: 115.8502 - val_accuracy: 0.1667\n"]},{"data":{"image/png":"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\n","text/plain":["\u003cFigure size 864x360 with 2 Axes\u003e"]},"metadata":{},"output_type":"display_data"}],"source":["from tensorflow.keras.optimizers import Adam\n","model2.compile(optimizer=Adam(learning_rate=.001), loss='categorical_crossentropy', metrics=['accuracy'])\n","history=model2.fit( #train model and save training steps in history variable\n","    x=x_train,\n","    y=y_train,  \n","    batch_size=32,  \n","    epochs=50,  \n","    validation_data=(x_val, y_val)\n",")\n","plot_history(history)"]},{"cell_type":"code","execution_count":13,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":651},"executionInfo":{"elapsed":1095,"status":"ok","timestamp":1640148726405,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"M1Qn2_XpGnTw","outputId":"1490b3a6-2a5f-4d2a-d478-82a1eca681d5"},"outputs":[{"name":"stderr","output_type":"stream","text":["/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:12: RuntimeWarning: invalid value encountered in true_divide\n","  if sys.path[0] == '':\n"]},{"name":"stdout","output_type":"stream","text":["Accuracy: 20.833333333333336%\n"]},{"data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAigAAAJDCAYAAAAhPu8cAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdd3QU1fvH8ffsJiGVkAAJndB77wLSgggCogioWBGQooKiwE+KggUsFAuoWBBRUFSKYEEDCIgKSJESeu8QCCE92c3+/ojfaEwMAZKtn9c5OYfMzJ15btwxT557545hs9lsiIiIiDgRk6MDEBEREfk3JSgiIiLidJSgiIiIiNNRgiIiIiJORwmKiIiIOB0lKCIiIuJ0lKCIiIhIgZk9ezYDBw5k1KhROfYtX76cvn37cuXKlaueRwmKiIiIFJj27dvz7LPP5tgeExPDjh07KFGiRL7OowRFRERECkzt2rUJDAzMsX3evHn0798fwzDydR4lKCIiIlKoNm/eTGhoKBEREflu41V44WTaFbuisC/htI7Emx0dgkP1qFDF0SE41K7Y/Y4OwaHqhlR3dAgOlWQ55+gQHOa2H/W375pure16Pb8K99jtWss/eoSoqKis7yMjI4mMjPzP41NTU1myZAnjx4+/pusUeoIiIiIi7uNqCcm/nTt3jvPnz/PMM88AcPHiRcaMGcOUKVMoVqzYf7ZTgiIiIiKFpkKFCnzwwQdZ3w8fPpwpU6ZQtGjRPNspQREREXFxhuE8w2ozZ84kOjqa+Ph4hgwZQt++fenYseM1n0cJioiIiBSYkSNH5rl/1qxZ+TqPEhQREREXZ7jhQ7nu1yMRERFxeaqgiIiIuDhnmoNSUNyvRyIiIuLyVEERERFxcaqgiIiIiNiBKigiIiIuLr8v4HMlqqCIiIiI01EFRURExOW5X73B/XokIiIiLk8JioiIiDgdDfGIiIi4OD1mLCIiImIHqqCIiIi4OFVQREREROxAFRQREREXZ7hhvcH9eiQiIiIuTxUUERERF6c5KCIiIiJ2oAqKiIiIi1MFRURERMQOVEERERFxcaqgiIiIiNiBS1RQZr34OX9s2ENwSCAzFzyTtf27Rev5/usNmEwmmtxUiwce75Gt3alj55k+fn7W9+dOXeTuwbfS/e6bmf/2Crb+tpdK1cvwxHP3ArD2+y3ExyXS/e6b7dOxfLp8PpaFr35GfGw8hmHQslsr2t7ZjuVzlhH9+268vMwUL1OCfk/fg1+gf77aAqx4/xv2bd5DmSpluWfMfQBsifqDxCsJ3Hxne3t3U/6Dp3/+5W/dOo8hIMAXk8mE2cvEgkUTchzzx6a9vDb1CywWK8VCAvlw3mguXYpn1BOziI9PYvgTd9ChUyMARj72Ns9OvI+wsGL27kq+eJsM3mhZDx+TCbNhsPZsDB8fOEHj4sE8WjMCkwHJlgym7jjA6aSUbG29DIOn6lWhRnAgNhu8FX2YPy9dwdtk8GKTWpT09WHZsbMsO34WgFF1q/DN8bMcuJLoiK7eMAPD0SEUOJdIUNrf1oyud7XhzckLs7bt3HKQTet2M33+03j7eBF3KT5Hu7IVw5g2fxQAVmsGg3tMpnm7uiQmJHN430lmfPY0s1/6gmMHz1CqXAnWfLuJ8TMH261f+WUym+jx6O2Uq1aelKQUZg6bRrUmNajeuAbdHumO2WxmxfvfsGphFN0H9cxX2+ASwZw6eJJRc8awaNrnnDlymhJlSrB55UYGTRnioJ5Kbjz98y/ZzZn7NCEhQbnui7+SxMsvfMas90ZSukxxLl28AsAP323krn7t6BjZmMeHvkGHTo1Yu2Y7NWuVd9rkBCA9w8ZTG3eRYs3AbBi81aoeGy/EMrJuFcb/sYfjicncXqEU91ctxys7DmZr271COACPrN9OMR9vXmlWmyEb/qRZiRB2XrrCZ4dO8lareiw7fpYqQf6YDMNlkxN35RJDPHUaVSGwaPbKwMrFv3LHAx3x9snMsYJDc79h/2fnHwcIL1ucsNKhmAwDqyUDm81Gamo6Zi8T3yz4ma592uDlZS6sbly3osWDKVetPAC+/r6EVwjnSkwcNZrWxGzOjLdirQjiYuLy3dYwDKwWKzabjfTUNMxmMz9/uYY2vdpidsKfgSfz9M+/5N/3326kU2RjSpcpDkBo8aIAeHmZSUlOIz3NgtlkwmKxsmB+FA8OuNWR4eZLijUDyKyImA0DbGCzQcBfn9UAbzMXU9JytKsY6M+2v/6feDktnYR0CzWCA7HYMvA1m/AyGVlVh4erV+Sj/cft1KPCYRgmu33Zi0skKLk5c/wCe/48zNgBbzBh6CwORuf94drw0zba3JJZ1vQL8KXxTTV5+oHphBQvSkCgHwd2H6NFu3r2CP2GXDp7kVMHT1KhZsVs2zet3EjNZrXy3dbX35dazWszY8hrBIUWxTfAl+N7j1G3df3CDF8KiKd+/j2dYRgMGzSDe/tM5utFa3PsP3b0HFeuJDHwoVe5t89kli/7FYCut7Xg5zXbGTJoOgMG38aiz9dwW49W+PkVsXcXrpkJeL9NA5ZENmdLzGX2xCXw+s6DTGlWm0UdmtK5TBgLDp/K0e7QlURuCg/FZEApvyJUDw4kzK8If8RcppS/L7Na1WfxsdPcFBbKgSsJXEzNmeSIY7nEEE9urNYMEuKSmPLhExyMPsG0cfOZvfhZDCPnOFx6uoXN63fTf+htWdt63d+RXvd3BGD2S19w96BbiVr2O9s37SeiSmnuGtDZbn3Jr9TkVOZNnsvtQ+/AN8A3a3vUZz9iNpto3KnJNbXt0K8THfp1AmDRtM/p8mBXNn73G/u27KNM5TJE9r+lcDsk180TP/8Cc+ePISw8hEsXrzBk4HQiKpemSdPqWfutVit7oo/x3oejSElN48F7p1C/QWUqRpTirXdGAHAlLpG5H3zH9DeGM3niPK5cSeL+h26hQcMqjupWnjKAQb/8SYCXmRea1CIi0J+7KpXh/zZHsycugX6VyjKsViVe35l9iOe7k+eoEOjPe60bcC45lV2xV7DabGTY4MXt+wEwGwavNqvN+C17GVYrgjDfIvx46gK/nr/kgJ7Kv7lsBaV4WDAtOtTHMAyq1amAYTK4cjn38cNtv+2lco1yFCueswx+eN9JbECZiiX5dfWfPP3SA5w9dZHTxy8Ucg+ujdViZd6kj2jcsQn12jbI2r555Ub2bNzNvWPvz/WXU15t/+fUwZOAjZLlwvhz3XYemPAQMadjuHDSuX4G8jdP+/xLprDwECBz6KZjZCN27zySY3+r1nXw8y9CSEgQjZtWZ/++k9mOmfPuCgYOvo0fvttEw8ZVeeHlAbw36xu79eF6JVqsbL8YR4uSIVQJ8mdPXAIAa85coE6xnJ/tDBvM3nOEQb/8yfgtewn09uJkYnK2Y3pVLMWPpy5QOySQhHQrk7fto2+lMnbpT0HTEI8TaX5zXXZtycyYTx+/gCXdQtFiAbke+8uPf5e3/+3zOT9wz+BbsVoyyLDaADBMBmmp6YUT+HWw2WwsmraQ8ArhtLurQ9b2vZv3sGbRah6ePAgfX59ravtPP3z8HV0e7EaG1YotI/NnYDIZpKvk6bQ86fMvmZKTUklMTMn692+/RlOlatlsx7Tv2JDtWw9gsVhJTk5l147DVKpcOmv/sWPnOH8ulqbNa5KSkobJZAIDUpz0Xg/28cqaa+JjMtGkRDDHEpMI9Pai3F+V4KYlinE8MSlH2yImE77mzF9xTUoEY7XZOJbwd4IS6GWmZVgoP546j6/JjA0bNsDH7LK/Ft2OSwzxTJ8wn91bDxF/OZFBPSbTb1AXOvZozuwXv2Dkva/h5WXm8Yn3YBgGly7EMfvlRYyfMQiAlORU/ty0n0fH3pXjvBvX7qRKzfKElgwGoFL1MjzZ/zUqVilNRDXnyaKP7j7Clqg/KF2pNNMffRWArgO6s3T2YizpFuaMmQ1AhVoR3DWyL3ExcXw5/XMGvvzof7at1aI2ALs27KBc9fIEl8j8GZSpUpbXB71C6cplKFOlbC7RiL15+udfMl28eIWnnpgFZA7xdb2tOa3b1uXLL34GoE+/9lSuUoab2tSl7x3PYzIZ3NG7LVWr/X0fz3pjCcNH3AHArd2a8+QTs5j7wfcMfex2u/cnP4oX8WFs/WqYDAOTAT+fucjv52N5fedBJjWuic0G8ekWXt1xAICbwkKpERzI3APHKVbEm1eb1cGGjZiUNKZsP5Dt3A9Uq8CnB09gAzbFxHJ7xVJ81LYR3/z12LGrcceF2gybzWYrzAvsil1RmKd3akfiPfuJiB4VnHNM2152xe53dAgOVTek+tUPcmNJlnOODsFhbvvR/X5ZXqs13Vrb9Xqlao+x27XORr9il+tctYJy6tQpNm/ezKVLmZOGQkNDadq0KeXKlSv04ERERCQ/3C8pzLNHS5cuZebMmQBUrVqVqlWrAvDGG2+wdOnSwo9OREREPFKeFZQ1a9Ywbdo0vLyyH9a9e3eeeuopevXqlWu7qKgooqKiALhvTJsCClVERERy445zUPJMUAzDIDY2lpIlS2bbHhsb+5+PtAJERkYSGRkJFN4clLTUdCYMnUV6mgWrNYNWHetz96DMVRF/+WkbZ09epGWHesx68QsO7zvJvUO6cnv/v59iWb5wLVHfbMQwDCpUKcVj4+/Gp4h3ocTqKBnWDGYOn0ZwiWAeeVFLmHuKvO4NcX/Pj5/LurU7CA0N4qtlkx0djl2VD/BjYqO/5z6V9vNl7oHjfH30jAOjkuuVZ4Ly0EMPMXnyZEqXLk3x4plLJ8fExHD27FkeeeQRuwT4X7x9vHj+7aH4+RfBYrEyfvDbNG5Vi+p1K7Ltt71069uGoKL+PPJULzau3ZWt7cXzcXy36BdmLhxNEV9vXh/3Cb/8tI2O3Zs7qDeFY/2StYRXCCflXy/REveW170h7q9Hr9b0u7cjE/7vQ0eHYncnEpMZ9MufQOb8hS87NeOXs56x6JrHVVAaNmzIG2+8wcGDB7NNkq1atWrm8/MOZBgGfv6ZyzRbLVYsFiuQue7Hkf2nqFyjHIZhEBwaxJYNe3K0t1qtpKWm4+VlIi0lLetRS3dx+cJl9myMptO9nVn39c+ODkfs6L/uDfEMTZpW5/SpGEeH4XCNSxTjdGIK51JSHR2KXKerPsVjMpmoXt05Hxe0WjMY/dAMzp6M4dberaletyKH954kolqZPIegiocF07N/e4b0egGfIt40aF6dhi1q2DHywrfsnSV0H9STlGRVTzxRbveGiCfpWLoEq854zorIhqc9xePszGYT0+aPYs43EzkQfZzjh86w7fe9NGpVM892CVeS2LxuN7MXj+P9Fc+RkpLG2u+32Cnqwhf9+24CiwVSrnp5R4ciDpLbvSHiKbwMg5vCQ1l75qKjQ5Eb4NIJyv8EBPlRt0lVtv2+lz837r9qNWTH5gOElQklOCQQLy8zLdvXZ9/Oo/YJ1g6O7j5M9G+7eOm+SXz20icc3H6ABVPnOzoscYB/3hsinqJFyRD2xyUQm+Y5r2xwx3fxuMRS97mJi03Ay8tMQJAfqSnp7Ni0n57922O17iIoOPd3kvxPifBi7N91jNSUNHyKeLPzjwNUqek+C891e6QH3R7pAcDBPw+w9ss13Dv2fgdHJfaS273xvzcXi3iCjmVKsPqM5uG4OpdNUGJjrvD2CwuxWm3YbDZu6tQAq8VK/WZ/z5eJvXiF0Q/NJDkxBcNksOLz9bzx+Wiq161Iq471efrB6ZjNZipVL0vnXq0c2BuRgpPbvdG0TW1HhyV2MvbpOWzZvI/LlxPo0vEZhgzvyR292zo6LLvxNZtoUqIY03cdcnQodpXXvEtX5Vbv4pn90hdE3t7SaSYE6l08ehePJ9O7ePQuHk9m73fxlK9vvzVvTuyYaJfruGwFJTfDxvVzdAgiIiJSANwqQREREfFE7rhQm/v1SERERFyeKigiIiIuTgu1iYiIiNiBKigiIiIuTnNQREREROxAFRQREREXpwqKiIiIiB2ogiIiIuLi9BSPiIiIiB2ogiIiIuLqNAdFREREpPCpgiIiIuLi9BSPiIiIiB2ogiIiIuLiDMNwdAgFThUUERERcTpKUERERMTpaIhHRETExWmhNhERERE7UAVFRETExekxYxERERE7UAVFRETE1ekxYxEREZHCpwqKiIiIq3PDcoMSFBERESkws2fPZuvWrQQHBzNt2jQA5s+fz5YtW/Dy8iI8PJxhw4YREBCQ53ncMOcSERHxMIZhv6+raN++Pc8++2y2bfXr12fatGm8/vrrlC5dmiVLllz1PEpQREREpMDUrl2bwMDAbNsaNGiA2WwGoHr16ly6dOmq59EQj0ghKemb4egQHGpX7H5HhyAO8u0twY4OwfO40FM8q1ev5qabbrrqcUpQREREJN+ioqKIiorK+j4yMpLIyMh8tV28eDFms5m2bdte9VglKCIiIq7OjhM2riUh+aeff/6ZLVu2MHHiRIx8VHw0B0VEREQK1fbt21m2bBljxoyhSJEi+WqjCoqIiIiLsznRHJSZM2cSHR1NfHw8Q4YMoW/fvixZsgSLxcILL7wAQLVq1Rg8eHCe51GCIiIiIgVm5MiRObZ17Njxms+jIR4RERFxOqqgiIiIuDrnGeEpMKqgiIiIiNNRBUVERMTVmdyvhKIKioiIiDgdVVBERERcnRM9ZlxQVEERERERp6MKioiIiKtzvwKKKigiIiLifFRBERERcXV6ikdERESk8KmCIiIi4ur0FI+IiIhI4VMFRURExNW5XwFFFRQRERFxPqqgiIiIuDo9xSMiIiJS+FRBERERcXXuV0BRBUVEREScjxIUERERcToa4hEREXFxNi3UJiIiIlL4VEERERFxdW74mLFLJCizXvycPzbsITgkkJkLngHgyP5TvPfKV6SnWTCbTQx6pjfV6lTI0faTt5az5dc92DJsNGhenQFP9cKSbmXq6I+4eD6OW++8iVvvag3AO1O+pMsdrahcs5xd+3c1l8/HsvDVz4iPjccwDFp2a0XbO9ux8pPv2fjd7wQGBwDQdUB3arWonaP93s17WDZ7MRkZNlp0bUnHuyMB+GzKfM4eOU2tFnXo9kh3AKI++5FSEaWo27q+/TooeZr63CJ+XRdNSGgg875+GoDZ01fw67povLzNlC1XnLGT+hFU1C9H275dX8YvoAhmk4HZy8z7C0YA8M7Mb9m4YS/VapRh3Iv3APDjt1u4HJtE3/va2q9z+eDJ978n9z038VeSmDRxHocOnsIw4LkXHqZBwypZ+79b8Tsff/g9Nhv4B/jy7IT7qFGzPJcuxTPqiVnExycx/Ik76NCpEQAjH3ubZyfeR1hYMUd1SfLgEkM87W9rxoQZg7Jtm//2Cvo+cgvT5o+i3+Bbmf/2ihzt9u44wt4dR5n+6dPMWPAMB/ecYPfWQ2z/fS+16ldi+qejWPvDFgCOHjhNRkaGU96gJrOJHo/ezugP/4/H3xzJhm9+4eyxswDc3LsdT703mqfeG51rcpJhzWDJW18x8OVHeeaDsWxbs5Wzx85y+vBpvH28GTVnDCf2Hyc5MZkrF+M4tveYkhMnc2vPprw2e2C2bU1bVuPjr0bx8ZejKFexJJ9+tPo/27/x/hA+WvRUVnKSEJ/MgT2n+PjLUXh5e3HowBlSU9L5btkf3NnvpkLty/Xw5Pvfk/uem1enLOSmNnVYsuJFvvj6eSpXLp1tf5myJfjg49F8uXQSg4Z058XnPwHgh+82cle/dsz/fByfzf8JgLVrtlOzVnn3SU4MO37ZiUskKHUaVSGwqH/2jQYkJ6YAkJSQTEjJojnaGYZBepoFS7oVS7oFq8VKsdAgzF5mUlPTsVoysNlsACx873vuGXxrofflehQtHky5auUB8PX3JbxCOFdi4vLV9vi+YxQvU4LipUvg5e1Fw/aN2P3rTsxmE+lp6WRkZGC1WDGZTKyc9z1dHnDOn4Ena9ikMkX/9flvflMNvLzMANSpX4EL5/L3eQAwmQwsFis2m43U5DS8vMx8/snP9L67NV7e5gKNvSB48v3vyX3/t/j4JLZuOcAdvTMrfN4+XgT962fTsFFViv5VUa5fvzLnzsUC4OVlJiU5LbPqZDJhsVhZMD+KBwc4f789mUsM8eRmwMhevDByDvPeWo7NZuOlOY/nOKZGvQjqNqnCwO7Pgw1uvas15SqFU7p8Cdb+sIX/G/gGt/fvwOZ1u6hcoxyhJYPt35FrdOnsRU4dPEmFmhU5svswG5atZ8tPmylXvTw9Hu2Ff1D2GzYuJo5iJUOyvi9WohjH9h4jvGIpAoMDmDn0dRpHNiPm1AVsNltWIiSu47ulm+nYpUHuOw0YNfR9DAN69m5Jz7ta4h/gS8s2NXmk3wwat6hGQKAv0TtP8ODgzvYN/AZ46v0Pntv30ydjCAkJ5Llxc9m/7wS16lRk9Nh78PMvkuvxSxf/Quu2dQHoelsLnh39Pl9/tY4RT93Fos/XcFuPVvj55d7WJbnhUzwum6CsXPwrD424nVYd67MhajuzX1rE828PyXbMmRMxnDx6njnfTARg8hPvEb39MLUbVubJyfcBYLFYeWHEHMa++jBzZy4j5txl2ndtQrOb69q9T1eTmpzKvMlzuX3oHfgG+HJTjzZ07t8FDFj58fcsf28p/Z6+N9/nu33YnVn//nDC+9w1oi9Rn/3I6cOnqd6kBi27tSqMbkgB+uT9VZjNJjp3a5zr/llzh1MyPJjYSwk8NWQOFSqF0bBJZe59uAP3PtwBgFcmfcmAYbewYvFGNv+2n8rVS/PgoEh7duOaeeL9/z+e2neLNYO9e44zZty91KtfmVenLOSjD75n+BO9chy7eeNeli5ez0fzxwIQFOTPW+9kDnFeiUtk7gffMf2N4UyeOI8rV5K4/6Fbss1lEefgEkM8ufn5uz9o2aEeADd1asDB6OM5jtm4difV61bEz78Ifv5FaNSqJvt3Hs12zA9fb6Bd1ybs33WMgEBfnnrxfr5ZuNYeXbgmVouVeZM+onHHJtRrm/nXclBIECazCZPJRItuLTm+L+fPILhEMJcvxGZ9fznmMsElsv+1tOvXnZSrVo60lFQunonhgQkPsWPddtJS0gq3U3JDvl+2md/WRzPh5Xsx/uOvp5Lhmf+tQ0IDaduhLnt2Zf+M7N97CpvNRoWIMNb8tINJr93P6RMXOXHsQqHHfyM87f7/J0/te3h4CGHhIdSrXxmAyFuasHfPsRzH7d93gsnPzWPGW49RrFhgjv1z3l3BwMG38cN3m2jYuCovvDyA92Z9U+jxFzqTYb8ve3XJblcqYCElirJ76yEAdv5xgNLlS+Y4pmR4MXZvPYTVYsVisRK97RBlI8Kz9idcSWLLL9G079aU1NR0DJMJw4C01HS79SM/bDYbi6YtJLxCOO3u6pC1/crFv+cd7Nqwk9IRpXO0LV+jAjGnYrh45iKWdAvbf95GnVZ//4VktVhZv3gtHfp2Ij01PesXnS3DhtViKcReyY3YuGEvC+b9zJSZD+Pr55PrMcnJaST9NVchOTktszpStVS2Yz6ctZKBw7pgSbeSkZEBgGEySE1xrnvg3zzp/v83T+17iZLBlCoVytEjmQ8IbPp9D5WrlMl2zJnTF3l6xGxemPIIFSNK5TjHsWPnOH8ulqbNa5KSkobJZAIDUlL1x5gzMmz/mylVSHbF5pxhfq2mT5jP7q2HiL+cSHBoEP0GdaFshZJ8NGMZVqsVHx9vBo2+kyo1y3Nwzwl+XPwrw8b1w2rN4P3XviZ622EMw6Bhyxo8PPL2rPPOnbmMZm3rULdJVdJS05n6zEdcvBBHlzta0a3vjT9qeSS+YCYcHtl1mFlPvknpSqWzEoiuA7qzbc1WTh/KfNwuJDyUu0b2pWjxYOJi4vhy+ucMfPlRAPZsjGbZO0uwZWTQrEsLIvvfknXudYt/xi/Aj2ZdWmCz2fjs5U84e/QsNZvXovugnjcUd48Knl0yPZe8t0DOM2nsZ2z74xBxlxMJDQ3i4aG38NlHq0lLsxAcnDnnqHb9ijw9vjcx5+N4ZdJXvDbrEU6fvMi4p+YBYLVkENm1EQ8M6pR13vWrd3Fg32kGDM38PMyavpxNv+6nSrXSTJyS/6HC/3IhpWD+/nHV+78guGrfKwcVzpyWfXuOM+m5eVjSLZQtV5JJLz7Myh82A9CnX3smTfyYVT9tpXTp4gCYvUwsWDQhq/3op95l+Ig7qFgxnEsXr/DkE7NIiE9m6GO3E3lLkwKN1d/Lvp+hqnd8YrdrHVzygF2u4xIJiqsqqATFVSlBKZgExVUVVIIirqewEhRXogTlxrnsJFkRERH5ixs+xXPdf+KsWbOmIOMQERERyXLdFZRFixbRoUOHXPdFRUURFRUFwH1j2lzvJW5IWmo6E4bOIj3NgtWaQauO9bl7kHstypNhzWDm8GkElwjmkRcHA7BtzVYuno6hftsGfPH6Ak4ePEnXh2+jfZ+OWe2SE5JYNP0Lzh49gwH0ffoeImpXclAvpKCdO3uZl8d/zqVL8RgY9Ojdgj79nWNOhT14wr2fF0/v//Pj57Ju7Q5CQ4P4atlkR4djP25YQckzQXn66adz3W6z2YiL+++VKyMjI4mMzFxHwVFzULx9vHj+7aH4+RfBYrEyfvDbNG5Vi+p1KzoknsKwfslawiuEk5KUkrVt7+Y9tO11M35B/tw+vDe7N+zM0W7p7CXUbFqTByc+jCXdQrpmsLsVs9nEsFHdqVGrHEmJKQy85w2ataxORJXwqzd2A55w7+fF0/vfo1dr+t3bkQn/96GjQ5EblGeCEhcXx7hx4wgICMi23WazMWHChP9o5RwMw8haYfB/j9q5k8sXLrNnYzSd7u3Muq9/BjL/u5w+dIqy1cphGAZBIUHs2bg7W7vkxGQO7zzE3c9kPqXh5e2Fl7emIrmTEiWLUuKv5c/9A3ypWDmMC+fjPCZBcfd7/2o8vf9Nmlbn9KkYR4chBSDP30yNGzcmJSWFiIiIHPtq1875YjpnY7VmMPqhGZw9GcOtvVu71V8Qy95ZQvdBPUlJ/rt6curgScpULvOfi3YBXDpzkcDgQL54bQGnD5+mXLXy3B8PGVoAACAASURBVD7sDoq405LPkuXMqUsc2Hua2vVyvu3WnbnzvZ8fnt5/j+SGD83l2aWhQ4dSs2bNXPeNGDGiUAIqSGaziWnzRzHnm4kciD7O8UNnHB1SgYj+fTeBxQIpVz37e3P2bd5LzWa18mybYc3g1IGTtOrRmqfefQYfXx/WfLGqMMMVB0lKSmXC05/w+DM9CQj0dXQ4duWu935+eXr/xT14RG0/IMiPuk2qsu33vVSoknO1VVdzdPdhon/bxd5N0VjSLKQkpbBg6nwuX7jMg88NyLNtcMliBJcMpmKtCADq39yA1Z8rQXE3lnQrE0Z9QudujWjXqZ6jw3EYd7v3r5Wn99+juOEkWTcsCmWKi00gMT4ZgNSUdHZs2k/Ziu4xBt/tkR5MWDiJcZ8+R/9xD1C1YTXuePwuMqwZBBQNyLNt0dCiFCsZwvkT5wA4sG0/4W7yc5FMNpuNVyYtomKlMPrd387R4didO9/7+eHp/Rf34bYVlNiYK7z9wkKsVhs2m42bOjWgaRvnnzdzvfZv2Ue1xtWzvr9y6QpvDJ9GSlIKhmGwfvFanvng//AN8KXX8DtZMOVTrBYLoaWLX9MbkMX57dx+lJUrtlK5WikG9J0OwKDHu9Kqbd7Df+7C0+79f/P0/o99eg5bNu/j8uUEunR8hiHDe3JHbw94zN79Ciha6r4w2XOp+0XTPqdF15ZUrB1ht2tejZa611L34pm01L0Dlrrv95ndrnXwi/52uY7bVlA8Td9Rdzs6BBERcRCbyf1KKPoTR0RERJyOKigiIiKuTk/xiIiIiBQ+VVBERERcnfsVUFRBEREREeejCoqIiIir01M8IiIiIoVPFRQRERFXp6d4RERERAqfEhQRERFxOhriERERcXXuN8KjCoqIiIg4H1VQREREXJ0eMxYREREpfKqgiIiIuDpVUEREREQKnyooIiIiLs7mfgUUVVBERETE+aiCIiIi4uo0B0VERESk8KmCIiIi4ur0skARERGRwqcKioiIiKvTHBQRERGRwqcKioiIiKtzw3KDEhQREREpMLNnz2br1q0EBwczbdo0ABISEpgxYwYXLlygZMmSPPnkkwQGBuZ5HjfMuURERMRR2rdvz7PPPptt29KlS6lXrx5vvvkm9erVY+nSpVc9jxIUERERV2cY9vu6itq1a+eojmzevJl27doB0K5dOzZv3nzV8xT6EE/dkOqFfQmnVTfE0RGII4X71XR0CA4V7ufoCETEWcTFxRESkvlLsVixYsTFxV21jeagiIiIuDo7PmYcFRVFVFRU1veRkZFERkbmu71hGBj5qMQoQREREZF8u9aEBCA4OJjY2FhCQkKIjY2laNGiV22jOSgiIiIuzmYYdvu6Hk2bNmXt2rUArF27lmbNml21jSooIiIiUmBmzpxJdHQ08fHxDBkyhL59+9KrVy9mzJjB6tWrsx4zvhrDZrPZCjfU/YV7ehEREadj3wdEKj25zG7XOjLjdrtcR0M8IiIi4nQ0xCMiIuLq9LJAERERkcKnCoqIiIiru86na5yZKigiIiLidFRBERERcXWagyIiIiJS+FRBERERcXXuV0BRBUVEREScjxIUERERcToa4hEREXFxNk2SFRERESl8qqCIiIi4OlVQRERERAqfKigiIiKuTkvdi4iIiBQ+VVBERERcnRuWG9ywSyIiIuLqVEERERFxdZqDIiIiIlL4VEERERFxdVoHRURERKTwqYIiIiLi6lRBEXFthw+f5M47R9Kjx+Ns27YXAIvFykMPjSc5OcXB0YmIyP8oQRGP8sUXP/Dss4N4//3n+OijxQAsXPgdPXu2x8/P18HRiYhcH5th2O3LXjTEIy7h5MlzDBr0PE2a1Gbbtj2Ehxdn9uzxfPPNz3zxxQ+kp1uoWLE0r776FH5+vowdO4PAQH927TrIhQuxPPPMw9x6a2u8vLxISUklOTkVLy8vrlxJYM2aTXzwwSRHd1FERP5BFRRxGceOnaZ//9v49tvZBAUFsHLlr3Tu3Iqvv57BN9+8ReXK5fnqq5+yjj9/PpYFC17hvfcmMm3aPAD69+/Ge+99ydixM3j00T7Mnv0Fjz7aF5NJt4KIiDNRBUVcRrly4dSqVRmAOnWqcurUeQ4cOMbMmZ8SH59IYmIybdo0zjo+MrIlJpOJqlUrEBNzGYAyZcKYP38KkJnwnD0bQ5Uq5XjmmWmkp1sYMeI+KlUqa//OiYjcCDf8G8sNuyTuysfHO+vfZrMJq9XK2LEzmThxCMuXv81jj91DWlparseDLcf5ZsyYz8iR9zN//nL69LmFZ555mLffXliYXRARkXxSgiIuLTExmZIlQ0hPt7B8+dp8t9u0aSdhYaFERJQhOTkVk8mEyWSQkpJaiNGKiBQSw7Dfl51oiEdc2ogR99GnzyhCQ4Np0KAGiYlJV21js9l4551FzJgxGoB+/W7l6adfx2LJ4PnnhxZ2yCIikg+GzWbLWfsuUPsL9/QiIiJOp7pdr1bxlVV2u9axMZ3scp2rDvGcOnWKnTt3kpKSfRGr7du3F1pQIiIi4tnyHOL57rvvWLlyJWXLluXdd9/loYceolmzZgAsXLiQhg0b5touKiqKqKgoAKZOHVDAIYuIiEg2brjUfZ4JyqpVq3jllVfw9fXl/PnzTJ8+nQsXLtCtWzfyGhmKjIwkMjLyr+80xCOFa926Lbz00vtkZGTQp09nBg/uA8C3367j+PEzDBnSl5demsPatVvw9S3C1KkjqFOnqoOjFhGRvOQ5xGOz2fD1zVz+OywsjOeff55t27Yxb968PBMUEXuxWq1MnvwuH3zwPN9+O4sVK9Zx8OBxIDNxadu2CevWbeHo0dP8+ON7vPDCcJ5//h0HRy0iUsAMO37ZSZ4JSnBwMEePHs363tfXl7FjxxIfH8/x48cLOzaRq9qx4wAVK5amfPlS+Ph4c9ttN7Nq1UZsNht79hymTp0qrFr1O716dcQwDBo2rMmVK4mcP3/J0aGLiEge8hzieeyxxzCbzdm2mc1mHnvssX8M4Yg4zrlzFylVqkTW9+HhxdmxYz/R0YeoWbMShmHkOKZUqeKcO3eRsLBQR4QsIlLgbJ42B6V48eL/ua9mzZoFHoxIQVm/fis339zE0WGIiMh10kqy4tLCw4tz9mxM1vfnzl0kPLw4GzZso3XrRrkec/Zs5jEiIm7DDVeSVYIiLq1evWocPXqaEyfOkpaWzrffrqNdu6ZYLFZCQooC0LFjC5YuXY3NZmP79r0EBflreEdExMlpqXtxaV5eZiZOHMLAgc9htWbQu3ckR4+e5qab/l6jp127pqxd+wedOw/Gz68IL788woERi4gUAjecg6Kl7sXtjBv3Jn363ELDhponJSKOYt+l7iu8kf+Xpd6o4yPa2eU6qqCI23nppSccHYKIiNwgJSgiIiKuzv1GeDRJVkRERJyPKigiIiIuzuSG5QY37JKIiIi4OlVQREREXJwd10+zG1VQRERExOmogiIiIuLiVEERERERsQNVUERERFyc4YYlFFVQRERExOmogiIiIuLi3LCAogqKiIiIOB9VUERERFycKigiIiIidqAKioiIiIsz3LDc4IZdEhEREVenBEVEREScjoZ4REREXJwmyYqIiIjYgSooIiIiLs6kCoqIiIhI4VMFRURExMVpDoqIiIiIHaiCIiIi4uLcsYKiBEVEREQKzIoVK1i9ejWGYVC+fHmGDRuGj4/PNZ9HQzwiIiIuzjAMu33l5dKlS3z//fdMnTqVadOmkZGRwa+//npdfVKCIiIiIgUmIyODtLQ0rFYraWlphISEXNd5NMQjIiLi4pzlZYGhoaH06NGDoUOH4uPjQ4MGDWjQoMF1navQE5Tlxw8V9iWc1vLjfo4OwaHmtCnj6BAcqnSdeY4OwaG2/9HC0SE41LwDnn3/e7rR9as7OoRCExUVRVRUVNb3kZGRREZGApCQkMDmzZuZNWsW/v7+TJ8+nXXr1nHzzTdf83VUQREREXFx9nyK558Jyb/t3LmTsLAwihYtCkCLFi3Yv3//dSUoTlIUEhEREVdXokQJDhw4QGpqKjabjZ07d1K2bNnrOpcqKCIiIi7OWdZBqVatGi1btmTMmDGYzWYiIiL+s9pyNUpQREREpMD07duXvn373vB5NMQjIiIiTkcVFBERERfnLEM8BUkVFBEREXE6qqCIiIi4OJMqKCIiIiKFTxUUERERF6c5KCIiIiJ2oAqKiIiIi1MFRURERMQOVEERERFxcYYbPsajCoqIiIg4HVVQREREXJzmoIiIiIjYgSooIiIiLk4VFBERERE7UAVFRETExamCIiIiImIHSlBERETE6WiIR0RExMW54TptqqCIiIiI81EFRURExMVpkqyIiIiIHaiCIiIi4uIMNyw3uESC8sXrC4jeGE1gsUCeeX8sAH+u3c6P83/g/PFzPPHWk5SvUeE/22dYM5g5fBrBJYJ55MXBAHw2ZT5nj5ymVos6dHukOwBRn/1IqYhS1G1dv/A7dQ32fTSPizt24h0URLMXngMg+t05JJ09B4AlKRkvfz+aPj8hX20BDn/5NZd27SawfHlqDnwYgHO//U56QgLlOkfaoVeSH0V8vFjyySB8fMx4mU2s+HE3r89axdJPBhEQUASAEqEBbN95koef+CxH+wXvPUjj+uXZtPUYDwyfn7V91it9qFmtFFFr9zLljZ8AGPloe/YeOMcPq/fYp3PXoW/Xl/ELKILZZGD2MvP+ghHZ9i/8+Gd++m4rAFZrBseOnOebNc+TYc1g3FPzSIhPZuDwW2nbsS4A/zdyLqOevZMSYcF270t+rJv9KSe27MI3OIje08dl27dz+So2fbKE/h9OxbdoYLZ98Rcuseq1OdgybGRYrdTu2o5at7TFmp7OT6/MIfHSZWp1aUvtLjcD8Mu7C6h5S1tKVC5vt77lh6f339O5RILS9JYWtL69LQtf/ft/wKUiSvHgcw/z1cxFV22/fslawiuEk5KUAsDpw6fx9vFm1JwxvDdmNsmJyaSnpHFs7zEi+99SaP24XuGtW1GmUwf2fjA3a1vtIYOz/n3oiy8x+/nlu60lKZn44ydoOmki+z7+hISTp/ALK8nZX36l3pMjcj2POEZqmoW7BnxIUlIaXl4mls0fzOr1++n1wPtZx3ww8x5W/kdSMfuj9fj5+XB/n2ZZ22pVDyc5xUKnO9/i8/cfJiiwCH6+PjSqX56Z7/1c2F26YW+8P4RiIQG57rvnofbc81B7ADasjWbRp+soGuzPVwt+4fY+Lbm5Yz1GP/YhbTvWZcPaaKrVKOu0yQlAtfYtqX1rO9a+/Um27QkxsZz6cw8BJUJybedfrCg9XhqF2dub9ORUFo96iQpN6xFz6DjhtarQ8I5bWD5+OrW73MzFoyexZdic8pezp/f/WmgOioNUqV8F/yD/bNvCK5YirHz4VdtevnCZPRujad61ZdY2s9lEelo6GRkZWC1WTCYTK+d9T5cHbi3w2AtCsRrV8Q7wz3WfzWbjwuYthLVoluv+XNsaBjarFZvNRkZaGiazmZMrf6Jsp46YvMwFHb7coKSkNAC8vcx4e5mx2WxZ+wIDitC6eRW+X5V7gvLLxsMkJKZm22axZODn64VhGHh7mbBm2Bj9eCdef3tV4XXCAVZ9v43IWxsB4OVlIiU5nfR0CyazCYvFypefrefev5IZZ1W6dlWKBOa89zd+/DXN7uuF8R+/lczeXpi9vQGwWtKxZWR+ZkxeZqypaWRYrVnHbv18BY3vvq0Qor9xnt5/T+cSFZQbseydJXQf1JOU5JSsbeEVSxEYHMDMoa/TOLIZMacuYLPZKFfN9TLouP0H8C4ahH/41ZO1//Hy8yW0Xl22THqRkFo1Mfv5ceXwESr20E3qjEwmg5VfDqdShVDmLtzItp0ns/Z17VSLXzYeypGE5OXA4QtcjE3kx6+G89U326hUoTiGYbBzz+nCCL9gGTBq6PsYBvTs3ZKed7XM9bCU5DQ2/rqPkf93BwCRXRsx+f8WsHzxRoaM6MbSRb/R5bbG+Pr52DP6AnFs8w78Q4tRPKJcnsclxMTy45R3uHL2As3v70VAaDH8goM4uHYT3zw7jfo9O3Fs8w6KVy5PQGgxO0V/4zy9///lv5I1V+bWCUr077sJLBZIuerlOfjngWz7bh92Z9a/P5zwPneN6EvUZz9y+vBpqjepQcturewd7nU5v2kzYS2aX3O7Cl27UKFrFwD2ffwJEb16cmbdL8TujiagXFklK04kI8NG595vUzTIl4/e7E+NqmHsO3gegF7dGrDg6z+u+ZwTp36X9e95s+5n9PNLGTG4PbVrlGLdbwf57KtrP6c9zJo7nJLhwcReSuCpIXOoUCmMhk0q5zhuw7po6jWMoGhw5l/fgUF+vPr2IwDEX0nis4/W8OKMB3l10pfExyfT7/6bqdsgwp5duS6W1DT+XLySW8c/dtVjA0uEcOe0Z0m8dJmoV9+nUstG+BUrSoeRmXPOMixWfnhxFpFjBvP7x1+TGBNL1XbNqdjMuebg/ZOn99/TuMQQz/U6uvsw0b/t4qX7JvHZS59wcPsBFkydn+2YXb/upFy1cqSlpHLxTAwPTHiIHeu2k5aS5qCo889mtRKzdRthzZpe9znijx0HG/iXCufCH1uoPXQwyRcukHTuXAFGKgXhSnwKGzYdpkOb6gCEFvOnYb1yRK3dd93n7NKhFjt2nyLA34eK5UN5dNTndL+lLn6+3gUVdoEqGZ45XyQkNJC2HeqyZ9fxXI9b/cN2Ov01vPNv8+ZEcf/ATqz6fjv1GlXi2RfuZu67PxVazAXpytkLxJ+/yJJnpvDFsIkkXrzM0tGvkBR75T/bBIQWI6RCac7uOZRte/TKdVRt15wL+4/gE+BHhycHsGv56sLuwg3x9P7nxTDs92Uvbp2gdHukBxMWTmLcp8/Rf9wDVG1YjXvH3p+132qxsn7xWjr07UR6anpWicyWYcNqsTgq7HyLjd6Df6lSFAnNfaJYfhxd+g0RvXpmzknJyADAMExkpDl/guYJiof4UzTIFwDfIl60a1WVg0cuAND9lrpErd1Latr1fVa9vEwMeuAmZn+0Hl9f76y5LSaTgbe3881FSk5OIykxJevfm3/bT+WqpXIclxCfzPYth2nToU6OfSeOXeDCuTgaNatCSkoaJpOBAaSmphd2+AUitGJZ+n84lX6zJ9Nv9mQCihej16tj8A8pmu24xIuxWFIz7+HUhCTO7T1EcJmwrP2pCUmc2LKLau2aY0nL/H+fYYAlzbl/Dp7ef0/jEkM8n740j0M7DpEYl8AL9zzHLQ90xT/In6WzviYhLoEPx8+hTJWyDJ46lLiYOL6c/jkDX370qufd8M16mnZuho+vD6UrlyEtJY3XB71Czea18MtlYpajRL/3AXH79pGekMBvT48h4vYelG7bhvOb/sgxOTY19jL7582n3sjH82wLELN1O0ERFSkSkjn+Gli+PH9MnERAuXIElne9+TjuKKxkEG+8fBdmkwmTyeCblTuzKia3d63H2x+uy3Z8gzplub9vc55+bgkASz8ZRNVKJfH392HLqtGMmriYnzccBODhe1qyaNlWklPSid53Fj8/b1YveZxV6/dzJT4FZxN7MZ5xT80DwGrJILJrI1q0rsmyL38D4PY+mcOy61fvolmr6vjlMr/kg7d/YOBjmZPhI7s24tmRH/PZR2sYMMz5nt4DWDNzLmd2HyAlPoGFj46ncd9u1Oh0U67HXjh0jL0//kLbof25fPIsGz9ZgmEY2Gw26vXoRGjFslnHbvvqexr27oJhMlG2QS2if1jH4lEvU/OWNvbqWr54ev+vhRtOQcGw/fORgEKw/Pj3hXl6p7b8eO6P/nqKOW3KODoEhypdZ56jQ3Co7X+0cHQIDjXvgGff/55udP3Odr1e+2832O1aP9/W2i7XuWoF5eDBzL+2qlatysmTJ9m+fTtlypShcePGhR6ciIiIXJ07VlDyTFC+/PJLtm/fjtVqpX79+hw4cIA6deqwbNkyjh49yp133plXcxEREZHrkmeC8vvvv/Paa6+Rnp7O4MGDeeedd/D396dnz548++yz/5mgREVFERUVBUDrYe0KPmoRERHJYvK0CorZbMZkMlGkSBHCw8Px98+cOOrj45PnojCRkZFERma+z8WRc1ByewePp7AkJbHv4/kknjqFYRhUf+gBgqtWcXRYcp2mv3AnndvVIOZSIh16vQnA6Mcj6dKhFhk2GxcvJjBi3NecuxAPZD6h893CofTo/16u7/IBmPlSb1o1jeBKQuYibyPHfc3uvWcc08ECdO7sZV4e/zmXLsVjYNCjdwv69G/r6LDs6othE/H2LYJhMmEym7j9lTGODsluPLnv7ibPBMXLy4vU1FSKFCnC1KlTs7YnJSVhMjn/E8r/fgePJzm48AtC69ahzrBHybBY9Niwi1u0dCtzF/zOm1Puyto2+6P1vPpWZqXykf6teGpoR8ZMXgZA88YV2bTt2H++y2frjhMATJ72A9/+uNv+HSpEZrOJYaO6U6NWOZISUxh4zxs0a1mdiCr5X23ZHXR7fkSOl+h5Ck/uuzvJM8uYNGkSRYpkvjH1nwmJxWJh+PDhhRvZDcrtHTyewpKUTNz+A5RqmznT2uTlhZe/8zw2Ldfu9y1HiY1Lyrbtn8vb+/t5Z3tHT4c21Vm9fj+Q97t83FGJkkWpUStzGXT/AF8qVg7jwvk4B0clUrhMhv2+7CXPCoq3d+6rSRYtWpSiRYvmus9Z5PYOHk+REhODd1AQ+z6aR+KJkwRGVKDqPf0w/5VsivsY+0Rn7urZkPiEVO56+IOs7a2bV2b67MxVMfN6l8/YJzrz1JCO/LLxEC9NX0laujXHNVzZmVOXOLD3NLXrVXB0KHZm8MOLbwMGNTu3pmZn113f49p5ct/di/OP01yHf76DxxPZMqzEHztOmQ7taPL8eMw+RTj+3Q+ODksKwdQ3f6Jp5GssXrGdh+/NXKisVFhRLsclkZySuSrm/97l07jjqzSqV44aVTNX1Hx5xo+07T6Trv1mUyzYj+EDb3ZYPwpDUlIqE57+hMef6UlAoK+jw7Gr7i88Sa9Xx9Jl3DD2rFzPmeiDjg7Jbjy17ybDZrcvu/XJbleyo/y8g8edFQkJoUhICEUrVwKgRNPGJBzL/Z0l4h4Wf/snt3XOXNq9Q5tq/LzhQI5j/v0un/MxmRNq09KtfL5kK43q5v12WFdiSbcyYdQndO7WiHad6jk6HLsLKJ65OrRfcBAVm9cn5uBRxwZkR57cd3fjlgnK1d7B4+58goMpEhpC0tmzAFzesxf/MqUdHJUUtEoVimf9u0uHWlnv6OnQplrW/JO83uUTViIoq33XTrXYe9A9XhBps9l4ZdIiKlYKo9/9nrfMQXpKKml/DW2np6Ry6s+9hJT3jFWdPbnvHjcHRVxXtXvvZs+cD7FZrfiWKEGNAQ86OiS5AbNf68tNzSoTWsyfLatG8/qsVXS6uTpVIkqSkWHj5JnLjJm0DJPJIKJCcQ4eiQHyfpfPrFf7UDwkAMMw2L33DKP/egLI1e3cfpSVK7ZSuVopBvSdDsCgx7vSqm0tB0dmH8lx8ax67X0AMqxWqrRpSrlGtR0clX14ct/dkd7FU4j0Lh7P+MvlvzjiXTzNG1ekd/eGWY8bO5LexePZ97+ns/e7eG778Re7XetbO71UURUUETeyaesxNm095ugwRERumBIUERERF2fPp2vsxS0nyYqIiIhrUwVFRETExbnjywJVQRERERGnowqKiIiIi3PHaoM79klERERcnCooIiIiLk5zUERERETsQAmKiIiIOB0N8YiIiLg4Qwu1iYiIiBQ+VVBERERcnCbJioiIiNiBKigiIiIuzh2rDe7YJxEREXFxqqCIiIi4OJOe4hEREREpfKqgiIiIuDg9xSMiIiJiB6qgiIiIuDh3rDa4Y59ERETExamCIiIi4uI0B0VERETEDlRBERERcXHuuA6KEhQREREpMImJibz77rucOHECwzAYOnQo1atXv+bzKEERERGRAjN37lwaNmzIqFGjsFgspKamXtd5NAdFRETExZkM+33lJSkpiT179tCxY0cAvLy8CAgIuK4+FXoFZc9lzy3SHLji7egQxIF6vNfZ0SE41IWUBEeH4FDvfpDo6BAcZsjA6/uFJK7v/PnzFC1alNmzZ3Ps2DEqV67MQw89hK+v7zWfy3OzBxERETdhz+GQqKgooqKisr6PjIwkMjISAKvVypEjRxgwYADVqlVj7ty5LF26lLvvvvuar6MERURERPLtnwnJvxUvXpzixYtTrVo1AFq2bMnSpUuv6zpKUERERFycszxmXKxYMYoXL87p06cpU6YMO3fupFy5ctd1LiUoIiIiUmAGDBjAm2++icViISwsjGHDhl3XeZSgiIiIuDhnWuo+IiKCqVOn3vB59JixiIiIOB1VUERERFycM1VQCooqKCIiIuJ0VEERERFxce5YbXDHPomIiIiLUwVFRETExTnLOigFSRUUERERcTqqoIiIiLg4PcUjIiIiYgdKUERERMTpaIhHRETExbljtcEd+yQiIiIuThUUERERF6dJsiIiIiJ2oAqKiIiIizO0UJuIiIhI4VMFRURExMVpDoqIiIiIHaiCIiIi4uLcsdrgjn0SERERF6cKioiIiIsz6SkeERERkcKnCoqIiIiLc8eneFwiQVk3+1NObNmFb3AQvaePA2DL5ys4tnkHhmHgGxzEzcPvIyC0WI62mz5dyomtuwFo1PtWKrduAsDPb3zMpeOnqdCkLk3v7QnAtq9/IKR8aSKaN7BTz/JndL2qtAwL4XJaOgPWbwegSpA/T9atgp+XmbNJqbz0536SLNZc25uAd1s3ICY1jWf/2APAuAbVqRTkz+/nL/HB/uMA3FelHEcSkthw7pJd+iX5s++jeVzcuZc/0wAAIABJREFUsRPvoCCavfBc1vZTq1ZzavXPGCYTofXrUaVP72ztks6eJfrd97O+T7kQQ0SvHpTrHMnhL7/m0q7dBJYvT82BDwNw7rffSU9IoFznSPt07DosX7iWqG82YhgGFaqU4rHxd+NTxDtr/8rFv/LD1xswmUz4+vkw5P/6UL5SKfb+eYT3Xv0ab28zIyffR5kKJUmMT2bauE8YP3MQJpNzFpNLF/Pj9fsbUyKoCDYb/8/efYdHUX0NHP/O7qb3SkJCEgKE3kMHaUGKdF5RUOyigvKzodhQxIIFOwoKAhYQCyggIgRBDCK9pgEJEBIgvbfNlvePYDAmVJNtOR+ePM/uzNzZe0h29u65Zfjmz1Ms+z2ZD+6KJNzfDQB3JzsKSisY+ebWGuXvGdCMib1CMRrh2LkCZn69H63OwLt3dKVloDu/xZ7n7fWV14TpN0Zw7Fwhm4+cM2mMl9PQr/0NnVU0UFoM6EmbYf35/aMvqra1Hz2YrreOBCB2wzYOfv8LfaZOqlYuZd9RspPPMO6tWegrdGx46X2CO7ehKDMHtb0d4+c/yy8vf4i2uBSdVkvm8VN0njDMpLFdjY2pGaw5fY5nOrao2vZk++YsTDjFoZwChgf7c0vTIJYeT6m1/ISmjUkpLsVZowYg3M2ZcoOB+2IO8lb3trho1DioVbT2dOOrpFSTxCSuXqM+vWg8eCAJi5dWbctNSCTrwCEiX3oBlZ0d2oKCGuWcAwKIfOkFAIwGAzufeBrfzp3RlZRSmHKGyDmzSVz2BUWpaTj5+3E+5k/aP/Y/k8V1rbIz8tnwbQzvrXwKB0c73n7uC2I2H2DQyO5Vx/Qb2oWh43sDsGf7UZa9v5YX3pvK2hW/8/y795FxNodNa3Zy1/9G8/3SzYy/c7DFNk4AdAYDr605SmxqPi4OGtbOHEBMYiYzlu2tOubZse0oLKuoUbaRhyN39g/nxte2UF5h4MO7uzGqSzCxqXmUVegZ8cZWvpjWGzdHDY72ajqFebNg0zFThndFDf3afy1sMYNiue/Mfwhs0xwHV+dq2+ydnaoe68rLgZq/nbzU8wS0aY5KrcbO0QHvkCBSD8ajUqvRayswGgwY9HoUlYp9q36my8Sb6juU63I4t4CCCl21bcEuThzKqfxQ2puVxw0BPrWW9XW0p6efFz+fSa/apjMYcVCpUACNoqA3Grm7RQjLLtHAEebl2TICO5fqf//ntv5OyIhhqOwqswf27u6XPUduXAJO/n44+vqAomDU6zEajRi0WlRqNam/biZo8CBUFxqxlkqv16Mtr0Cv06Mt0+Lt51Ftv7OLY9XjsjItyoXrglqjorysgvLyCtQaFedTs8hKz6Nd1+Ymrf+1yiwoJzY1H4Dich0n0gsJ8HCsdsyIzo1Zt6/2LxZqlYKjnRq1SsHJTk16QSkVegOOdmoUBTRqFXqDkcdGtOa9DfH1Hs+1aujX/obOKjIol7J3xVpObN+NnbMTI16cUWO/d1gQB777hfajBqMr13Iu9hieTQII790FR3dXfnzqDZr3707B+UwwGPENb2KGKK7PqaIS+jTyZkd6DgMCffF3cqj1uIdbN2VRwimc/vHBk1JcSp62gk/7dmRTWiZBzk6oFIXjBcWmqr74j0rS08k/dpyTq39EZWdH+MT/w71p2CWPz9y9B//u3QDQODni3b4d++a8glfrVqidnChIPknoKMu+SPv4ezD6tgE8OHYu9g52dOweQaceLWsc98v3MaxbuR1dhY6XPnoIgPF3DuaDOSuwd7Djfy9NZvkH65j0wHBTh/CfBHk70zbIg4Onc6u2dWvmQ3ZhOacya7530/PLWPzbCWLmDKWsQk9MQgYxCZkA5BRpWTdzIGv2nCHUzxWVolQ1hKxBQ772NyRW3UCJnDyayMmjObTmV+I3bqfLLdUvsMEdW5N14jTrnpuPo7sr/hFNUS6kc3ve/X9Vx22at5A+U2/l4A8byTmdRuMOrWgV1ceksVyrNw+f4JE2TbmjeRN2pOdQYTDUOObvcSvHCorp6F39G/aC+JNVj1/t2pp3jiZxW7Ngmru7sDcrr1rGRVgeo95ARXExnZ+bReHJU8Qv/JTu815FUWp+mzTodGQdOkTTCeOqtoUMH0rI8KEAJC77grCxozm3PYbc2DhcgoMssrFSVFDCnu2xfLz6OVzcnHj72eX8/ss++g/vWu244f/Xl+H/15c/ft3PD8uieWT2JJpGBDFvSWX3VeyBJLx8K98P85/7Ao1GzZ0zRuPp42bymK6Ws72aj+/tztzVRygqu5hNHd01mLX70mot4+5kR1T7QPrP2URBSQUf3dOdMZHB/LQ3lbmrj1Qd99nUnjz3zUGm3RhB6yAPYhIyWLXzdL3H9F805Gv/pVh27vP6WEUXz5U069uNk7sO1rqv04RhjHv7GYbPfgSj0YhHoH+1/af3HMY3vAm6snIK0rMY9Pi9nPrrALpyrSmqft3OFJfy1J44HthxiN/OZXK2pKzGMe283Ont783KAV2Z3bklnX08ePYf41gA+vh7c6ygCCeNiiBnR+YcSKR/gA8OFtwvL8DB2xPfrl1QFAX38KagKFQUFdV6bM6Ro7iFhGDvUbMbqPB0ChjBOaARmXv30eahqZRmZlKSbnkN1MN7juPf2BsPL1c0GjU9B3Qg8cipSx7fZ0gndv9+tNo2o9HID0uj+b+7h/Dt4l+Z8vBIosb05Odv/6jn2l8/jUrh43u7s3bvGX49fHEAq1qlMLRDID8fqL17p09LP1KzS8gp0qIzGPn10Fm6NvWudkxU+wCOnsnDxUFNqK8Ljyzdw/BOjXG0s46Pu4Z47W9IrPZTKP9cRtXj03sP49m4UY1jDHoDZYWVF+2c02nkpJwlqGOri/t1emJ/3kqHMUPQaSuq+quNBiN6na7G+SyJp33l2AMFmNKsCetSztc4ZnHiaSZu3cukbft4+UAiB7Lzee3Q8ar9akVhQtPGfJOUhoNKxd/L/KgUBY0tjriyIb6dO5GXkAhAyfl0jDo9dq6utR6bsWsP/j261brv1I9rCRs7unJMyoUsnKKoMGgt7yLt28iTY0dPU16mxWg0cmTvcYLDqn/onE3JrHq8b0c8gU18q+3ftmEvXXq3xs3DmfLyClQqBUWloC2vOcjUUsyb3Jmk9CKWbE2qtr1PSz+SMoo4n1fzywnA2dxSOoV5VTU2ekf4cSL9YiNWo1K4e0AzFkUfx8FOjfHCBUCtUrDTWO5HQ0O/9l+KSjGa7MdUrKKLZ+t7SzkXe5yywiJWPvA8XSaOIPVALHlnM1AUBVc/b/rcfysAmUmnSdgUQ7+HbsOg1/PzC+8BYOfsyIBH7kSlvvjNIO7X7TTv3wONgz3eoUHotFpWP/4qwV3a4vCvQYnm9HynCDp5e+Bhr+HbgZEsO56Ck0bNmNBAAP44n80vqZVvWh8He55s34xn9l55wNvY0AB+Tc2g3GAgqbAEB7WKJf06sSsjl+JLTFkWphe3aDH5iYlUFBWx88mnCRszioC+fUhcupw9L8xBpVHT8t67UBSF8tw8ji3/kvaPPgKAvryc3Lh4Iu64vcZ5s/YfxC0sFAevyimark2asHf2HFyCg3FtYnl98hHtQuk1qANP3vkOarWaphFBDBnbi5WfbqR5q2C63dCOX77fweE9x9Bo1Li4OfHw7IuzO8rLtGz9eQ+zP3gAgFGT+vPq44vRaDQ8+vJt5grrsiLDvRnfPYSEtHzWPzUQgLfXx7EtLp2RXYJrDI71d3dk3qRO3LPoLw6dzmXjwbOse2oAOr2RuLR8vvnzVNWxU24IZ/WuM5RV6Ek4W4CTvZpfZg1kW1w6haWW0WBr6Nf+hk4xGo312hx68/Dm+jy9RfsltWH/oW8d4WfuKpjV1Jiz5q6CWc1oW3uXU0Mx+kXr/CZeFx68z8XcVTC7pzoMMenrzTtkus/aWR1NE9s15/E++uij+qiHEEIIIUSVy3bxvPHGG9WeG41GYmNjKS6unNL29NNP11/NhBBCCHFVbHHY4GUbKDk5OQQFBTF48GAURcFoNJKcnMyoUaMue9Lo6Giio6MB8J48uO5qe41WTZuNnaMDikqFSq1izBsNo0HVxMWJ2Z0jqp4HOjmy9HgKP5yynCWsRf3SlZSQuOxLitPSUBSFiLvuwKN5M3NXq05oyyt44aEFVGh16PUGeg3qwK33V64CGrP5AOdTswlpFsDKRRtRqRTUahV3PzqG1p3CAcg8n8snr31LVnoeiqLw3Dv34d/Y+3IvafHemNyZgW0DyC4sZ/i83wDwcLbjw7u6EeztTGpOCQ8v3UOBhYwtqU8N9bpviy7bQHn99dfZsGEDq1evZsqUKYSFhWFvb0+bNm0ue9KoqCiioirv52HuMSgjXvofju61z26wVWeKS7k/5hBQ2Yf33eBuxJyX++s0JCdWrsK7XVvaTnsAg05nkbNyrpedvYaXPnoIJ2cHdDo9z0/9iC69WhPRLpQDOxMYMbEvjUP86davLYqicOr4WeY//wUfrpoFwIdzVjLhrsF07NGS0pJyVDbw1fP7XSl8sT2Zt2+/uCbMg1ER/Hksk4XRx3kwqgUPDWnBG2vjzFhL02mI13219f8Z13DZMSgqlYqRI0cybdo0Vq9ezZIlS9DrZXaHNeni68nZ4jLSy8rNXRVhIrqSUvKPHSegX+WCUyqNBo2z7QzYVhQFJ+fKlZP1Oj26CzPOjEYjJ4+lEd4yGCdnh6pF68r/seT9mZPn0ev1dLywAq2TswMOjvZmiKJu7UnKJq+kenZkSPsAfthdefuKH3anMKR9oDmqJsR1u6ppxj4+Pjz++OPs378fJyenKxewGAobX/kIUGg1pA+thvQ1d4VMblCgL1vOZV75QGEzyrKysHNzI/Hz5RSfScU1LITmk25B7VD77RCskV5v4Km73uV8ahbDJvQhol0oyQmphLVoXNUw2bXtCF998jMFuUU8O/8+oHKdFBc3J958ehkZ57Jp3y2C26fdhFptuet+XC9fN0cyCyq/mGQWlOPr5niFEraiYV73bSARWMM1rYPSpUsXunTpUl91qXMj5z6Gi48npfmFbJz7ER5BAQS2seybg9UljaLQu5E3nyVa9rLVom4ZDXoKT6fQfPKtuIc35cSKVaRs2EjTcWPMXbU6o1armP/lExQXlvLG00tJSTrHgb8S6Nzr4mJcPQa0p8eA9sQeSGLloo289NGDGPQG4g+e5K0vHsevkSfzn/+SrT/vIWp0DzNGYxpGTLfAljk19Ou+LbG9rw3/4OJTuQCVk4cbod07kHXilHkrZGI9/Lw4ll9Ertb2B8aJixy8vHDw8qpcAh/wjexC0WnbvFO1i5sT7bo258BfCRzadazWmwe27dyM9LPZFOQV4ePvSVhEYwKCfFBr1HTv346TibUvFW/tsgrL8HOvzJr5uTuQXdgwunkb6nXfFleStdkGSkVZOdrSsqrHaYcS8GrS2My1Mq1BjX357VyWuashTMzewwMHby9Kzlfe/iAvPgHnxrYz/iA/t4jiwlIAyssqOLz7GIFN/NDr9bh5VC4Qdu5MFn+vQZmckIquQoebhwvNWjehuLCU/NzKReSO7j1BcNOaS6Xbguij55nQPQSACd1D2Hyk5u0wbI1c922LVSx1fz1K8wvZ8tZnABj0epr1jSS48+VnH9kSR7WKrr6evHM06coHC5vTYvKtxH+6BKNej6OvLy3vudPcVaozuVkFfDR3JXq9EaPRSO/BHdHr9HTodnFq/V9bD7Ptl71oNGrsHex4fO4UFEVBrVa485FRvPTwQsBIeMtgosb0NF8wdeT9OyPp0dwXL1d7drw8lPc3JLBw8zE+urs7E3uGkpZbOc3Y1jXk674tjkGRpe7rkSx1L0vdN2SmXOr+41dXETWmJxHtQk32mlciS903bKZe6v7DuE0me61H2txoktex2QyKEKLhmPbcLeaughCijkkDRQghhLBy6isfYnVsdpCsEEIIIayXZFCEEEIIK2eLg2QlgyKEEEIIiyMZFCGEEMLKmXIBNVORDIoQQgghLI5kUIQQQggrp5YxKEIIIYQQ9U8yKEIIIYSVk1k8QgghhBAmIBkUIYQQwspJBkUIIYQQwgQkgyKEEEJYOcmgCCGEEEKYgGRQhBBCCCunlpVkhRBCCCHqnzRQhBBCCGFxpItHCCGEsHK2mG2wxZiEEEIIYeUkgyKEEEJYOZlmLIQQQghhApJBEUIIIaycZFCEEEIIIUxAMihCCCGElZOF2oQQQgghrsBgMPDUU08xb9686z6HZFCEEEIIK2dpY1A2bNhAUFAQpaWl130OyaAIIYQQos5kZ2ezf/9+Bg8e/J/OIxkUIYQQwspZUgZl2bJl3H777f8pewImaKCcKLCr75ewWMmx5eauglmV3Jhu7iqY1dyuBeauglndutXH3FUwK1VqsrmrYDatPR3MXQVRj6Kjo4mOjq56HhUVRVRUFAD79u3Dw8OD8PBwYmNj/9PrSAZFCCGEsHKmzKD8s0Hyb4mJiezdu5cDBw6g1WopLS3lgw8+YMaMGdf8OtJAEUIIIUSdmDx5MpMnTwYgNjaWdevWXVfjBKSBIoQQQlg9tQWNQakr0kARQgghRJ1r27Ytbdu2ve7yMs1YCCGEEBZHMihCCCGElVPJUvdCCCGEEPVPMihCCCGElbPFbIMtxiSEEEIIKycZFCGEEMLKWdJS93VFMihCCCGEsDiSQRFCCCGsnC0u1CYZFCGEEEJYHMmgCCGEEFZO1kERQgghhDAByaAIIYQQVk5m8QghhBBCmIBkUIQQQggrJxkUIYQQQggTkAyKEEIIYeVsMdtgizEJIYQQwspJA0UIIYQQFke6eIQQQggrp8ggWSGEEEKI+icZFCGEEMLK2WACRTIoQgghhLA8kkERQgghrJyMQRFCCCGEMAGryKAkfr6c7MNHsHNzo9vcFwE49dM6zm2Pwc7NFYCm48fi06F9jbI5R45yYuW3GI0GAvv1JWTEMADiP11CcVoa3h3aEz5hHACn1/2MS1AQvl06mSiyqxPo5sC7I9rg62yPESMrDp1l6f5Unu3fjMHNfKkwGDmdV8rMX+IpKNfVKB8ztRfFWj16oxG9wcioL/cCMOuGZgwI9yEuo5DHN8QDMK5NI7yc7Ph8X6pJY7xa58/l8MIzS8jOLkBRFCbcfAOTp0RVO2bv7gQee2QBjYN8ARgU1YUHpo0iJ6eQJ2YsoLCwhOkzxjFwcGcAHn34I56dfTv+/p4mj+d6TBz+Gk4uDqhVCmqNms9W/K/afqPRyAdv/sRfMQk4ONrxzMu30LJ1MCmnMnj5mRXodHqefH4C7TqGodPpmTl9Ma+/dzeOTvZmiujSnmrfnJ7+XuRpK7jnj4MAzO7UkiaujgC4ajQU6XTcH3OoRtmVA7pSotdjMBrRG+HBHZXHTG0ZSnc/L5IKinn98HEAohr74WGv4YdT50wU2dUJ9HHmrRl98PV0xGiEbzYfZ/nPCbz/RD+aNnYHwN3FnoJiLaOf+LlG+W0Lx1FcWoHeYESvNzLuqQ0AzJzSmf6dg4g/lcPMD/4EYMwNTfFyd2DZ+gTTBXgFq95eQdyuOFw9XZn52SwASgqK+fLV5eSez8ErwJspz9+Fs5tzjbJ7Nu1my9ebABh82410u7E7Oq2OpS8uJi8rj96j+tJndF8Avnt3Fb1G9ia4RRPTBVfHbDHbYBUNlEZ9etF48EASFi+ttj14yGCaDLvxkuWMBgPHv15JhycexcHLi/1zX8enUweMegMqezsi58zm0Pz30JWUotdqKTh5ktBRN9V3ONdMbzDyytbjHM0owsVOzfo7uhFzOoc/TufyxvZk9EYjs25oxrQeoczbnlTrOW5ddYDc0oqq5272ato1cmPYst28MbQVLX1dOJVXys3tArnj+5oXe0uh1qh4/KmJtG4TSnFxGZNvnkuPXm1o1rxxteM6d23BBx/PqLZt44Zd/N8t/RkU1YVHHnqfgYM78/vWg7Rq3cRqGid/e/+zB/H0cql1318xCaSmZLFi7dPEHUnhnVdXs+irGaz9/i9mPDWGgMZefPDmT7wyP4yfvtvJjSO6WGTjBGBjagZrTp/jmY4tqra9fDCx6vFDrcIo1ukvWf6xv45SUHGx0e6iUdPCw5X7Yg7yZPvmNHVzJq24jOHB/jy1J65+gvgPdAYjry/fR2xyDi6OGn58+yZ2HDrH/+b/UXXMM3d1pbBYe8lz3D57M7mF5VXPXZ3taBvuw8jH1/PatJ5EhHhy+nwhEwY14565W+o1nmsVeWMP+ozpx8o3v67a9tuqLbToHMGgW6P47ZtofvsmmpH3j65WrqSgmM1f/sqjCx4HReG9afNp26sdJ48k07RdOIMmRbHg0ffpM7ovZ5PSMBoMVt04sVVW0ejybBmBnUvNFvKVFCSfxMnfHyc/P1QaDf7dI8k+cAhFrcagrcBoMGDU61FUCqd+XEvYmNFXPqkZZBRrOZpRBEBxhZ4T2cU0cnXgj1M56I1GAA6cyyfQzeGqz2kwgubC3aWcNCp0BiMPdAth2f5UdAZj3QdRR/z8PGndJhQAFxdHmoYHkpmRe1VlNRo1ZaVaKrQ61CoVOp2eFV9Gc+c9w+qzyiYXsy2WoSO7oigKbTuEUlRYRlZmAeoL8ZeXVaDRqCksKGXH73EMHdXV3FW+pMO5BdUaGP82INCXLWczr/p8BiNoLnTWO6or/+5vCW/M6tPnqt5LliQzt5TY5BwAist0JKXm08in+rVwRO9Q1sWcuupzGg1G7NQX/g/sNej0Bu4b04YvNiSi01vW/0GzDs1qZEdi/zxC5JBuAEQO6Ubsn0dqlEvcm0BE1wic3V1wdnMmomsEiXviUWlUaMu0GHQG/v51b1y2gaF3jqj3WOqbohhN9mMqVtFAuZS037ax98WXSfx8ORXFxTX2a/PycPD2qnru4OVFeV4eLo0DsXNzZd/Lr+LTsQOlGZlgNOIWGmLK6l+XYHdH2jZy4+C5gmrbJ7ZrzLaT2bUXMsJXN3di/ZRIJnWozDQUV+jZejKbDXd2I6NYS2G5jk6B7mw6kVXfIdSZs2lZJMan0K5DeI19hw8mMXHcS0x/4D2STqQBMPymHmzbepAH73+He6bexLffbOWmUb1wcrr6hp1FUOCJhz7jvknvsfb7v2rszsoowD/gYkbIr5EHWRn5jL+lN18t+Y3XXviGKfcOYvln0Uy5dzAqlXVeBjp4uZOrrSCtpKzW/Ubgre5tWdSnIyObNAKgVK9nV2Yun/XtSHaZlmKdjtaebuxIzzFhza9PkJ8LbZp6c+jYxfdotzb+ZOWVcfpcYa1ljEZY9uJgfnxrBLcMqcxCFZfp2LY/jbXzbyIzr5TCkgo6tvAlevcZk8TxXxXmFuLu4wGAm7c7hbk1Y8/PzsfT7+K138PXk/zsfCK6tiQ3PYcPZrxL33E3EPvnUYJaBOPh62Gy+ourZxVdPLVpPKB/VXfMqR/Xkrzqe1rec+dVl28+6Zaqx0c++IiIKbdzev0Gis+k4tWmNYH9+9V5nf8rZzs1C8e04+XfjlOkvZjWfrhnKDqjkTVx6bWWm7ByH+lFWnyc7fjq5k4k5ZSwOzWPRbtTWLQ7BYA3hrbinZhkbm0fSL8wbxIyi/nwr1OmCOu6lBSX8eSjH/PkrFtwdXWqtq9Vm1A2bH4DZxdH/th+mMceWcDaX17Dzc2ZDz+pHK9RkF/M0sUbeOf96bw8ezkFBSVMuetGOnZqZo5wrsmCpdPxa+RBbk4Rjz/4KSFN/enUtWYj7d8aBXrxwZKHAEhNySIzPZ/QcH9eeW4lFRV67ps+lCahfvVd/TozqPHlsyczdh4hq1yLp70db3dvS0pRKYdzC/gmOY1vkisbrU+2b87SYymMCG5ENz9PkgqK+SrJ8sZfOTtqWPBUf175fA9F/+iqHdk3jPUxJy9Z7tbnNpKeU4q3hyPLXxxMclo+e+Iy+OzHOD77sbJL67VpPXn/m0NMjGpO346BJJzO4+Pva2YlLJGiKCjXMH1FrVZz27N3AKDX6fn0mYXcPede1i5cQ25GHpFR3Wjbu119Vbde2eAkHuvNoNh7uKOoVCgqFYE39KXg5Kmax3h6Up5zMf1fnpuLg2f1sQZZBw7iFhqKvrycsoxM2jw0lcx9+9GXX7pP1xw0KoWFY9rxY3w6G49fvCj/X9sABjfz5X/rYy9ZNr2oMpbskgp+PZ5Fp0C3avvb+ruiAEm5JYxo6c/0dbGEeDoR5ulUy9nMr6JCx5OPfsLwm3oyeEjN7glXVyecXSoHUfa7oQM6nZ7cf33L+nTheu6behMbN+ymU5fmzH3tHhYtWGuS+v9Xfo0qv+15ebvSb2A74o+mVNvv6+9Oxvm8queZ6fn4+lf/hvjZRxu5b/pQflgRw8hx3Xno0ZtYunBz/Ve+jqgU6Bfgw9Zzl874ZV14D+dpK/gjPZtWnq7V9jd3d0EBzhSXMiDQhzkHEmns4kiQs2N9Vv2aadQKC2b2Z+32k2zadTHLoVYpDO0Zws87Tl+ybHpOKQA5+WVs3nWGDi18q+1v09QLBYXktAKG9wplxvw/CAlwJfRf1whL4ublRkF2PgAF2fm4/uv3CuDh40Fe5sVrf35WHh4+1d8Df66NITIqkpT40zi6ODHluTv5/fut9Vt5cU2stoFSnpdf9Thr/0FcghrXOMa9aRil6RmUZmZh0OnI2L0Xn04dq/YbdHpSN2+hybChGCq0VRPJK8emXLrf2xzeHNaKE9klLN578QLVP8ybB7uHcu/qw5TpDLWWc7JT4WKnrnp8Q5g3iZnVu8Oe6BvO2zHJ2KlUqC+MSzEajThdKGdJjEYjc2Yvp2l4IFPuqn2AdFZmPsYLHcxHDydjNBjx/MdF7PTpdDLSc4ns3oqyMm1lF4cCZRbWKK1NaamWkuKyqsd7dh4jvHlAtWP69m/Lr+v3YTQaiT02mlImAAAgAElEQVR8GhdXR3z93Kv2H9ybhK+fO01C/Sgrq0BRKSgqhfIyy4//b119PDlTVErWJersqFbhpFZXPY709eRkYUm1Y+6JCOHzYymoFQXV3+99Y+XxluT16b04kZbP5+viq23v0zGQ5LQCzmeX1FrOyUGDi6Om6nHfjoEcT8mrdsyjkzrx7sqDaNQKqr/f+4bK4y1Vm17t2Lt5DwB7N++hbe+aszdbRrYicV8iJYUllBSWkLgvkZaRrar2lxSWELcrlq5DuqEt11ZmYRSo0FbUOJe1UBTT/ZiK5f4V/kPcosXkJyZSUVTEziefJmzMKPISjlF85gwoCo4+PrS443YAynPzOLb8S9o/+giKWk3z227lyLvvYzQYCOjbp1pD5uzWrQT07oXawR6X4GD0Wi17Z8/Bu307NM7XPii3vkQGeTChbSDxmUVsuLNycNhb25N5aXAL7NUqvppYOS36wNkCntuciL+LPW8Oa8VdPxzG19meT8dWvoE1KoWf4tP5/dTF/vYbm/ty+HwhGRdmAcRlFPHrXd1JyCwiPrPIxJFe2cH9J/h57U5aRARxy/g5ADz86DjOn6uM6eZbBhC9aR/frdqGWq3C0dGe19+eWi0NvOD9NUz/X+XU8mEjuvPYjAUsXfwLDz08xvQBXaPc7EKee3w5AHqdgajhnenRpxU/fbcTgDE396Jnv1bsjIln0qh5ODja88yciVXljUYjX3y2hZferHy/jJrQg7nPrkSv1/PEs+NNH9AVPN8pgk7eHnjYa/h2YCTLjqewITWjsnvnX9kTHwd7nmzfjGf2xuNlb8fcrq0BUCsK0Wcz2ZN18cO5TyNvEvOLyL7QKD1RUMySfp1ILighqbD2D3xz6NrKj3EDmpFwKpe18yu7tOd/fYDf95/lpj5hrPujeveOv5cTr03rxX2v/oavpyMfP90fAI1Kxdo/TrL9wNmqY6O6N+FoUjYZuZVZlvhTOfz87kgSTueRcOrqBp7Xt69eXU7S4SSK84uYO+lFbrxjOINujeLLucvY/ctfeDXyZsrzlV37ZxJT2Ln+TyY+cSvO7i4Mue1G3n/4HQCG3DYUZ/eLs942f/UrgycPQaVS0TKyFX/+FMP8qW/Sa2Rvs8QpaqcYjfU7dH1qzLb6PL1F+3Vn7VmNhiL+MTtzV8GsCiss4yJvLrdu9TF3FczqzOJkc1fBbN59z/fKB9m4USHDTfp6h3PWm+y1OniPNMnrXFMGJSEhgRMnTtCkSRM6dux45QJCCCGEENfhsp2tzzzzTNXj6OholixZQmlpKd9//z0//vjjJctFR0cza9YsZs2aVXc1FUIIIUSDcdkMil5/cSrrli1beOGFF3B3d2fUqFE899xzjB07ttZyUVFRREVVLj9uzi4eXUkJicu+pDgtDUVRiLjrDjyaW/400rpwd5dgJnVojKLAysNnLXbp+vry0vNL2f77Yby93fj+p5fNXR2TSj+fx2vPf0NOTiEKCqMm9ODm2yxv2nx9aeLixOzOEVXPA50cWXo8xeKWsb8Wr0/vxaDIYLLzyxjx6DoAhvcKYcYtHWkW7MH4pzdwNKn6Wi4/vjWCic9sRHthAP2iZwbQpJFbVXkPV3vef+IGgv1cSM0sZsbb2ym4zIq01sSgN/De9Pl4+Hpw7ytTzV0dk1DZ4Dzjy2ZQjEYjRUVFFBYWYjQacXevnAng6OiIWm15Mzz+7cTKVXi3a0v3V1+m60sv4NI40NxVMokIXxcmdWjM6K/2MmzZHgY38yXUQqcM15dRY/uwYNGj5q6GWajVKqY9MZIvV89k4ZcPs2bVn5xKqn2NHFt0priU+2MOcX/MIR6IOUS5wUDMectfiO1yVm9NqrEM/bGUPKa9+Tt7aln/KNjflfSckqrGyY09mlBcWn1m4gPj2rHz8DmiHv6JnYfP8cD4tvUXgIn9seZ3GoU0Mnc1xH902QZKSUlJVVdNUVERubmVg/7Kysqo57G1/5mupJT8Y8cJ6NcHAJVGY1Ezc+pTc29nDp4roExnQG80sutMHsMirGcBrrrQNTICD4/a71Vj63z93GnZOhgAZxdHQsP9yczIv0Ip29TF15OzxWWkl5Vf+WALticug7zC6jEkpRVw8mxBrcff0Llx1YwdZ0cN94xuU2PxtajuwazeVjmQd/W2ZIZ0t4170eRl5hG/K47uw3uauyompZjwx1Qu28WzYMGCWrcrisLMmTPrpUJ1pSwrCzs3NxI/X07xmVRcw0JoPukW1A5Wtqz5dTiWVczMfs3wdNRQpjMwMNyHw+drv5AJ23YuLYfjCWdp097yb+NQHwYF+rLl3NXfq8dW3NC5Ma8urbxr+WOTOrFkbRyl/7rTua+nE5kXphhn5pbiayNZ1p8+WcPI+0dTVlr7LRCE9biuFYkcHBzw9/ev67rUKaNBT+HpFBoP7E/Xl55Hbe9AyoaN5q6WSZzIKWHh7tN8dXMnvvi/TsRmFKJv2DOeG6SSknJeePILHpk5GhdXy1od1RQ0ikLvRt78fu4S96iyUXYaFQE+zpxJL6J1mBchAa5s3nXl++xYelb8asT9FYurpyvBEbaRDboWslCbFXHw8sLBywv38KYA+EZ24UwDaaAArDpyjlVHKgcFzuwXzvlC605xi2ujq9DzwhNfMGREZ/oPrrnSZkPQw8+LY/lF5Frx6qDXI7K1P3vjMwDo3NKPds182LZwHBq1gre7I1+/PITbZm8mK68UP6/KLIqflxPZ+dafcTgVm0zczqMk7I5Dp9VRVlLGinlfMnnWFHNXTVwHm22g2Ht44ODtRcn58zgHBJAXn4BzAxkkC+DjbEd2SQWN3RwY1sKPcV/vM3eVhIkYjUbemPMtoU39uWVKf3NXx2wGNfblt8vcq8dW/XP8yYpfj7Hi12NA5d2QP3tuELfNrrzn0pY9qYwfEM6iNbGMHxBO9G7rn+k34t5RjLh3FAAnDh3n9++2NpjGiQ1O4rHdBgpAi8m3Ev/pEox6PY6+vtd0t2Nrt3BMe7wc7agwGJgdfYyCcsu6t1B9m/Xkp+zbk0heXhFDB83kwemjGTehYUy1PXLwFL+u3094iwDumVi51Pf9jwynV7/WZq6Z6TiqVXT19eSdo0nmrkqdePexvvRo1wgvN0diPhvP+98cJq+onBfv64a3uyOLnxtE/Mlc7p67hR7tGvHeN4eueM5Fq4/ywZM3cPPg5qRlFjNj/nYTRCLE1ZOl7uuRLHUvS903ZLLUvemXug/wcebVh3py7yu/mfy1/0mWujf9UvcJeaZb6r6Vp2mWures23YKIYS4buezS8zeOBGirth0F48QQgjREDS4lWSFEEIIIcxBMihCCCGElbPBBIpkUIQQQghheSSDIoQQQlg5RbH+lYD/TTIoQgghhLA4kkERQgghrJyMQRFCCCGEMAFpoAghhBDC4kgXjxBCCGHlFBvs45EMihBCCCEsjmRQhBBCCCtni9kGW4xJCCGEEFZOMihCCCGElZMxKEIIIYQQJiAZFCGEEMLK2WACRTIoQgghhLA8kkERQgghrJyMQRFCCCGEMAHJoAghhBBWzgYTKJJBEUIIIYTlkQyKEEIIYeVUNphCkQyKEEIIISyOZFCEEEIIK2eDCRRpoAghhBCibmRlZbFgwQLy8vJQFIWoqChGjBhxXeeSBooQQggh6oRarWbKlCmEh4dTWlrKrFmz6NChA8HBwdd8LmmgCCGEEFZOUYzmrgIAXl5eeHl5AeDk5ERQUBA5OTnX1UCRQbJCCCGEqHMZGRmcPHmS5s2bX1d5xWg01muzyylkUn2e3qIFRV5fv5utOLG6h7mrYFbtv0g3dxXM6rUBReaugjCTdSlO5q6C2X3ad4BJXy+9dK3JXuvIDmeio6OrnkdFRREVFVXtmLKyMl588UXGjx9Pjx7X91kgXTxCCCGEuGq1NUj+SafTMX/+fPr163fdjROQBooQQghh9SzlZoFGo5GFCxcSFBTEyJEj/9O5pIEihBBCiDqRmJjI9u3bCQkJYebMmQBMmjSJLl26XPO5pIEihBBCWDkLSaDQqlUrvv322zo5l8ziEUIIIYTFkQyKEEIIYeVsMdtgizEJIYQQwspJBkUIIYSwcpYyi6cuSQZFCCGEEBZHMihCCCGE1bO9FIpkUIQQQghhcSSDIoQQQlg5RTIoQgghhBD1TxooQgghhLA40sUjhBBCWDlFsb18g+1FJIQQQgirJxkUIYQQwurJIFkhhBBCiHonGRQhhBDCysk0YyGEEEIIE5AMihBCCGH1JIMihBBCCFHvJIMihBBCWDlZB0UIIYQQwgQkgyKEEEJYPRmDIoQQQghR7ySDIoQQQlg5W1wHxSoaKMGB3ix+dxr+fh4YjfD5ii0s+Hwj7VuH8OFr9+Li4sjp1EzunrGAwqLSGuUfuXc4d00ahNFoJDbhDFOfXEh5eQVL359O21Yh/LJlPy++uQqApx8ZR1ziGdZt2mvqMC8p0MeZt2b0wdfTEaMRvtl8nOU/J/DopI5EdWuCwWgkJ7+Mpz78k4zcmvEDuDrZsfGDUWzedYY5i/dgr1Gx8JkBBPi48PXGRL7eeAyAVx7sycpNx4hNzjFliOIyXu7dghuCvMkpq2D8uv0AvHVDK8LcnQBws9dQqNVx8/oDNcr2aezF093CUSsKq0+cZ8nRVADm9W1JCy8Xfk/N5oMDpwGY2r4JJ/JK+O1Mtokiu3bbf9jGrl/+AgUCwwK5ZeZk7Oztqvbv+XUX6z9bi4ePBwB9xvSjx4heZJxJ5+vXv8Sg0zPh0YmEtWmKXq9n8TOLuPvl+7B3tDdXSNekocWf+Plysg8fwc7NjW5zXwQg6dvvyT50GJVGg6OfH63uuRONs3ONsqmbojn3Rwyg4BIcRKt77kRlZ0f8p0soTkvDu0N7wieMA+D0up9xCQrCt0snU4YnrsAqGig6vYFZr3zFwaOncHVx5M+fX2PLH0f45M2pzHrla2J2xXPHxAE89sBIXp7/XbWyjRt5Me3uYXQe/CRl5RV89fH/uHlULw4ePUVpmZbuQ59m/dfP4u7mhLOTA906N+OND9eYKdLa6QxGXl++j9jkHFwcNfz49k3sOHSOxT/G8d7KQwDcMaIVD0/swOxFu2o9x6OTOrI7NqPqeb/Ojdkbn8knP/zGt68N4+uNx2gV5oVapUjjxML8dCKdlQlnebVPy6ptM7cnVD1+smtTiir0NcqpFHiuRzOmbj7K+ZJyvhnRia1nctAoUKY3MGHdfj6NaoernRpHjYr2vm58euSMSWK6HvlZefzx43aeWjwLOwd7vpi7jINb99NtaI9qx3Xs35nxj/xftW1//fwnY6eNx6uRNz99vJqwF5uyc90OugzuarEfzv/WEONv1KcXjQcPJGHx0qptXm3aED5hHIpaTfJ3P5Dy8y+E3zyhWrny3FzStvxG5NyXUNvbE/fJp2Ts2oNraAgqezsi58zm0Pz30JWUotdqKTh5ktBRN5k6vDplixkUqxiDcj4jj4NHTwFQVFxGwok0Ggd407xpIDG74gH47Y/DjB3RvdbyGo0aJ0d71GoVTk72nEvPpUKnw8nRHkVRsNOo0esNvPDEzbzyzvemCuuqZeaWVjUaist0JKXm08jHmaLSiqpjnB01GI3GWsu3DffG19OJmENnq7ZV6Aw4OaixU6tQLvxdPzapI++uPFh/gYjrsi+jgPxy3SX3Dw3zY8PJjBrb2/u4kVJYRmpRGTqDkV9OZTKwiTcVRiOOahUKoFEp6I1GpncK5eNDp+sxirph0BuoKK9Ar9dTUa7F/UKm4EpUajXaMi0V5VrUGjWlRSXE7Yyl65Bu9VzjutXQ4vdsGYGdS/XsiHe7NihqNQDuzcIpz82rtaxRb8CgrcCo16PXarH39ERRqyu3GQwY9XoUlcKpH9cSNmZ0vccirp1VZFD+KSTYl05tw9hz4ATxx1IZdWMk6zbtZfxNPQkO9Klx/Nn0XN77dD3H/vqI0jItW7YfZssfRwDIyilk54bXWLk6hmZhAagUpaohZKmC/Fxo09SbQ8eyAHh8cifGDQinsKSC22dvqnG8osCzd3Xlifd30LtDQNX2HYfOMbZ/ON/PG85nP8UyuFswsck5l+wiEpapq7872aVaUgrLauzzd3bgfHF51fP0Ei0dfN04mV9KTlkF347szLrkDELcnFChEJ9TbMqqXzMPX08G/N9AXrltDnYOdkR0bUXLyFY1jjsSc5iTR5LwDfZnzINj8fT3os+Yfqx84yv0FTomPHoLm7/axKDJUahUVvEdDZD4a3MuZgf+3SJrbHfw8iJ46BD+euoZ1HZ2eLVtg3e7NgDYubmy7+VXadSrJ6UZmWA04hYaYuqqi6tgVQ0UF2cHVi56jJlzvqCwqJQHZi5i/pw7mfW/cfy8eT/aiprfMj09XBg5JJLWfWaQV1DCik/+x63j+vLNmhhmzvmi6rjvP3+SR55ZwlMPj6VDm1C2/HGEpSt/M2V4V+TsqGHBU/155fM9VdmTd1Yc5J0VB3lwfDumDG/J+6sOVytz+7CWbNufxvnskmrb9QYjj78XA4BGrbB0dhQPvr6VZ+/qSmM/F9ZsS2bLnlTTBCau2/Cm/mw4lXnN5d7cm1z1+MOBbXj5rxPc374JLb1c2Hkujx+On6/LataJksISju48yrNfzsbJ1Ykv5i5lX/ReukZd/IBq06sdnQd2RWOvYef6Hax8awUPvTUdL38vps1/BICstEzys/JoFBLAinlfodfpGHbXCPyC/c0V2lVp6PH/2+n1G1BUavx79qixr6K4mOyDh+jxxqtonJyJ+2QR6Tv/olGvnjSfdEvVcUc++IiIKbdzev0Gis+k4tWmNYH9+5kyjDpk3Y3N2lhNRBqNmpWLHmPVmh38tHEPAMeSzjLq9tfpc9NzfPvTDk6eTq9RblDfdpw6k0FWTiE6nZ4fN+6hZ9eIaseMHNKVA0dO4uLsQHhoI26f9j7jRnTHyYL6ZjVqhQUz+7N2+0k27ao5TuCn7ckM7RVaY3unln5MGd6KbQvHMevOrowbEM7M2ztXO+a2YS1Zsy2JTi39KCypYMb8P7h3dJt6i0XUDbUCUSE+/HqJBkpGSTkBLg5Vzxs525NeUl7tmIFNvInLKcJZo6aJmyNPbk9gSKgvjmrLuzQc338MnwBvXD1dUWvUtO/bgVNxJ6sd4+Lugsa+8ntXj+G9SDtW873yy9KfGXbXTcSs2U6P4T256f7RbPryV5PE8F809Pj/6XzMn2QfOkzr++9FUWqOvciLS8DR1xd7NzdUGjW+XTtTcCK52jFZBw7iFhqKvrycsoxM2jw0lcx9+9GXa00VhrgCy7sKXcLCt6aSeOIsHyzeULXNz8cdAEVRmDVjHJ99taVGuTNpWXTv0qKqsTGwTzsST6RV7ddo1Dx873De+WQdTo72VeM41GoV9vaWk2B6fXovTqTl8/m6+KptoYFuVY+jujchOS2/Rrkn3ovhhgdWM+DBNcxbvo8125J566uLsz3cXewZGBnMmm3JONmrMRiNGI1GHOzV9RuQ+M96BnpxMr+U9JLaL6hHswsJdXMkyNUBjUpheJgf285cHACtURRubx3E0qOpOGhU/D2ESa2AncryBtx5+ntyOv402jItRqOR4weO0yikUbVjCrIvvgdidx7F/1/7kw6dwN3HA79gP7TlWhSVgkpRqLCCD6WGHv/fco4c5czGTbSbMR21Q+1fIh18vClITkZfXvl/lRefgHPji13cBp2e1M1baDJsKIYKLX8PxKscm3Lp8V6WTFEUk/2YiuV8Al9G724tuW3CDRyJT+GvX14H4MU3V9G8aQAP3HEjAD9t3M0X324DILCRFx+/cT/j7nqTPQeTWLNhFzs3vIZOb+BQ7CmWrLjYkHnwzhv56vvtlJZpORKfgrOTA3s2vcGvWw+SX1BSoy7m0LWVH+MGNCPhVC5r51eONJ//9QFuHtyc8CAPDAYjZzOLeWHRXwC0a+bN5KERPPvxX1c89yMTO/DJ90cwGmH7wbPcPrwlG94dxYpNx+o1JnH13ujXkm6NPPF01BA9oTsLDp1mzYl0hof5seFU9cGxfk72zOnVgmm/xaI3wmu7k1gY1Q61orDmRDpJ+Rf/pm9tFcjapHTK9AaO5RbjqFGxelQX/kjLobCWWUHmFto6jA79OvLutLdRqVUENQum54jebFy2gSYRIbTt3Y6YH7cTuzMWlVqFs5szt86cXFXeaDQSvWITU567E4CeI3qxYt5XGPR6xs+42VxhXbWGGH/cosXkJyZSUVTEziefJmzMKFI2bMRYoePw/PcAcA8PJ+KO2yjPzePY8i9p/+gjuIc3xa9rF/a9/AqKSo1rSBMCb7jYdXN261YCevdC7WCPS3Aweq2WvbPn4N2+Xa1TloV5KMZLTf2oI04hk+rz9BYtKHKEuatgVidW1+wbbkjaf1Gzy7EheW1AkbmrIMxkXYqTuatgdp/2HWDS1yvWbTfZa7lobjDJ61w2g3L8+HGCgoJwdnZGq9Xy448/kpycTHBwMOPHj8dZWppCCCGEqAeXHYPyySef4OBQOchu6dKllJSUMHbsWBwcHPj4449NUkEhhBBCXJ5iwn+mctkMitFoRH1hQZzk5GTeeOMNAFq1asXMmTMvWS46Opro6Og6rKYQQgghGpLLNlCaNGnC1q1bGThwIKGhoSQlJdGsWTPOnj2LRnPpolFRUURFRQHw/grTjUFZ+NYDDB/cmczsAiKHPFVt3//uv4l5L9xOcMepZOcWmqxOpvT69F4MigwmO7+MEY+uA2B4rxBm3NKRZsEejH96A0eTZBn7hmBK68aMbxGA0QjH84p5YccxtIZ6HW5mUQx6A+9Nn4+Hrwf3vjLV3NUxOVuOX1dSQuKyLylOS0NRFCLuugOP5s0oSErm3B87aHnXFADKsnPY88JLhI0eSZNhN1Jy/jxxCz+rOk9ZZhZhY0cRPCTKXKHUMauZlHvVLttAefDBB1m6dCmrV6/Gzc2N559/Hh8fH3x8fHjggQdMVcer9uV3v7Nw+a8sfndate3Bgd4MvqE9KanXvqCVNVm9NYmvfknkrRl9qrYdS8lj2pu/88qDDXvAakPi72TP5FZBjF27j3K9gbdvaMXwpn78lFRzOXxb9cea32kU0oiykpor7DYEthz/iZWr8G7XlrbTHsCg02HQVk6RzjlyFO92bauOS1r1XbXnzgEBRL70AlA5nXjnE0/j27n6mlDCsly2geLs7Mz06dMpKSkhIyMDg8GAt7c3np6epqrfNdmxO4GQYN8a29988Q6ee20F3y1+0gy1Mp09cRkE+blU25aUVmCm2ghz0qgUHNQqdAYDjhoVGZdYK8UW5WXmEb8rjsGTh7D9h23mro7J2XL8upJS8o8dp+U9dwGg0mhQXcjm58YnEHzjEACy9h/E0dcHtYNDrefJjUvAyd8PR9+at0exVrZ4s8CrWgfF2dmZsLCweq5K/Rg5pCtnz+dwJD7F3FURwiQySrUsi01l84TulOkN7Dyby85ztd9QzRb99MkaRt4/mrJS28seXA1bjr8sKws7NzcSP19O8ZlUXMNCaD7pFgzaChS1Go2zE/qyMlJ+2UjHJx7lzK+baz1P5u49+He37BslClvstPoHJ0d7nnp4LC/P/87cVRHCZNztNQxs4sOw1XsY/N0unDQqRjb1M3e1TCLur1hcPV0Jjmhi7qqYha3HbzToKTydQuOB/en60vOo7R1I2bCRnNg4vNtW3p7j1E/rCb4xCrWjY63nMOh0ZB06hF9kV1NWvd7JSrJWJjy0EaFN/Ni9sXL2UVCgNzs3vEa/0c+TnllzWXghbEHPQE/SisrILa+8oWR0SjYd/d1Zf9K2x2ABnIpNJm7nURJ2x6HT6igrKWPFvC+ZPGuKuatmErYev4OXFw5eXriHNwXAN7ILZzZspCwrm+AbKwe7Fpw8Sea+/SR/txpdSQmKoqCysyNo8ECgcqyKW0gI9h7uZotDXB2bbqDEJp4htMuDVc8TdnxAn5HP2ewsHiEAzhWX08HPDUe1ijK9gR6BnsRlN4xVXUfcO4oR944C4MSh4/z+3Vab+XC+GrYev72HBw7eXpScP49zQMCFe+wEkhsbj2tIZdao86yLS2Cc+mkdageHqsYJQMauPfj3sMXunQY6BsVaLP/wEfr1ao2vlxsndn3E3He+Z/mqbeaulsm8+1hferRrhJebIzGfjef9bw6TV1TOi/d1w9vdkcXPDSL+ZC53z615U0VhO45kFbL5dBbfjuyMzmAkIaeI746dM3e1hKgTLSbfSvynSzDq9Tj6+hIycgQVhUVX1fWgLy8nNy6eiDtuN0FNxX8l9+KpR3IvnoY9tVnuxdMwsjaiJlPei+f0up9x8ve3uKyIqe/FU6a/8s1h64qjuqdJXsemMihCCCEaltBRN5m7ChZBscE5L7YXkRBCCCGsnmRQhBBCCKtne4NkJYMihBBCCIsjGRQhhBDCyplyATVTkQyKEEIIISyOZFCEEEIIqycZFCGEEEKIeicZFCGEEMLKyTooQgghhBAmIBkUIYQQwurJGBQhhBBCiHonGRQhhBDCyimSQRFCCCGEqH+SQRFCCCGsnKwkK4QQQghhAtJAEUIIIYTFkS4eIYQQwurZXr7B9iISQgghhNWTDIoQQghh5WSasRBCCCGECUgGRQghhLB6kkERQgghhKh3kkERQgghrJws1CaEEEIIYQKSQRFCCCGsnuXkGw4ePMjSpUsxGAwMHjyYsWPHXtd5LCciIYQQQlg1g8HAkiVLePbZZ3n33XfZsWMHqamp13UuaaAIIYQQVk4x4b/LOXHiBAEBATRq1AiNRkPv3r3Zs2fPdcUkDRQhhBBC1ImcnBx8fHyqnvv4+JCTk3Nd56r3MSilKSvr+yUuKzo6mqioKLPWwZwkfvPFf+SOCLO87t/kdy/xmyv+USFmedlqGt7v3/G9NTcAAASrSURBVHTXm+joaKKjo6ueR0VF1cv/tc1nUP75n9gQSfwNN/6GHDtI/BJ/w46/PkVFRTFv3ryqn382Try9vcnOzq56np2djbe393W9js03UIQQQghhGs2aNePcuXNkZGSg0+n4888/iYyMvK5zyTRjIYQQQtQJtVrNPffcw6uvvorBYGDgwIE0adLkus5l8w2UhtUHWZPE33Djb8ixg8Qv8Tfs+M2pS5cudOnS5T+fRzEajcY6qI8QQgghRJ2RMShCCCGEsDg23cVTV8vtWqOPP/6Y/fv34+Hhwfz5881dHZPKyspiwYIF5OXloSgKUVFRjBgxwtzVMhmtVsuLL76ITqdDr9fTs2dPJk6caO5qmZTBYGDWrFl4e3sza9Ysc1fH5KZPn46joyMqlQq1Ws28efPMXSWTKS4uZuHChZw5cwZFUXjooYeIiDDvlH9xfWy2gfL3crvPP/88Pj4+PPPMM0RGRhIcHGzuqpnEgAEDGDZsGAsWLDB3VUxOrVYzZcoUwsPDKS0tZdasWXTo0KHB/O7t7P6/vft7ZbeP4zj+JN+MaaOVk6t2sDjBwUpCcUI4cazIwXIkSkpSDuxADvxYrdW0M/4ER8qpY6TRtNDiQFpJCFPXtvvg2+3ovk+Wrs82r8df8Dy6evf5XNf7+kM4HMblcmHbNmtrawSDwV/1kD48PMSyLD4/P02nGBMOh/F4PKYzHLe3t0cwGGRpaQnbtvn6+jKdJCWq2iuen1y3W4k6OjpoamoynWFES0sLgUAAgIaGBizLKnmTYSWqqanB5XIBkM/nyefzVfkr9v/z9PTE2dkZw8PDplPEYR8fH1xdXTE0NARAXV0dbrfbcJWUqmpPUP5r3e719bXBIjEhm82SyWRoa2szneKoQqHAysoKj4+PjI2N0d7ebjrJMfv7+0xPT//q0xOAjY0NAEZGRn7NFy3ZbBaPx8Pu7i53d3cEAgFCodD3wC6VpWpPUERyuRyRSIRQKERjY6PpHEfV1tayvb1NIpHg9vaW+/t700mOOD09xev1fp+g/Vbr6+tsbm6yurrK0dERqVTKdJIj8vk8mUyG0dFRtra2qK+v5+DgwHSWlKhqB5SfXLcrlce2bSKRCIODg/T29prOMcbtdtPZ2cn5+bnpFEek02lOTk6Yn58nGo1yeXlJLBYzneW4f591Xq+Xnp4ebm5uDBc5w+fz4fP5vk8M+/r6yGQyhqukVFU7oPzkul2pLMVikUQigWVZjI+Pm85x3OvrK+/v78DfL3qSySSWZRmucsbU1BSJRIJ4PM7i4iJdXV0sLCyYznJULpf7vt7K5XIkk0n8/jL4e58Dmpub8fl8PDw8AHBxcfFrXo6vRlX7DspPrtutRNFolFQqxdvbG7Ozs0xMTHy/OFbt0uk0x8fH+P1+lpeXAZicnPyRzYaV4Pn5mXg8TqFQoFgs0t/fT3d3t+ksccjLyws7OzvA3yuPgYEBgsGg4SrnzMzMEIvFsG2b1tZW5ubmTCdJibRJVkRERMpO1V7xiIiISOXSgCIiIiJlRwOKiIiIlB0NKCIiIlJ2NKCIiIhI2dGAIiIiImVHA4qIiIiUHQ0oIiIiUnb+ASmKofsFMhiQAAAAAElFTkSuQmCC\n","text/plain":["\u003cFigure size 720x720 with 2 Axes\u003e"]},"metadata":{},"output_type":"display_data"}],"source":["from sklearn.metrics import confusion_matrix\n","def cm_testdata(model,data_test,label_test,num_labels,letter_labels):\n","    # have model generate predictions on test data\n","    y_pred=model2.predict(data_test)\n","    \n","    # find classes with the largest predicted probabilities\n","    y_pred_max = np.argmax(y_pred,axis=1)\n","    label_max=np.argmax(label_test, axis=1)\n","    # plot confusion matrix\n","    cm = confusion_matrix(y_true=label_max, y_pred=y_pred_max)\n","    cm_sum = np.sum(cm, axis=1, keepdims=True)\n","    cm_perc = cm / cm_sum.astype(float) * 100\n","    annot = np.empty_like(cm).astype(str)\n","    nrows, ncols = cm.shape\n","    for i in range(nrows):\n","        for j in range(ncols):\n","            c = cm[i, j]\n","            p = cm_perc[i, j]\n","            if i == j:\n","                s = cm_sum[i]\n","                annot[i, j] = '%.1f%%\\n%d/%d' % (p, c, s)\n","            elif c == 0:\n","                annot[i, j] = ''\n","            else:\n","                annot[i, j] = '%.1f%%\\n%d' % (p, c)\n","    fig, ax = plt.subplots(figsize=(10,10))\n","    sns.heatmap(cm, cmap= \"YlGnBu\", annot=annot, fmt='', ax=ax) \n","\n","    #accuracy\n","\n","    correct = (y_pred_max == (label_max))\n","    accuracy = correct.sum() / correct.size *100\n","    print(\"Accuracy: \"+str(accuracy)+\"%\")\n","\n","num_labels=np.unique(y_test)\n","letter_labels = range(1,11)\n","cm_testdata(model2,x_test,y_test,num_labels,letter_labels)"]},{"cell_type":"markdown","metadata":{"id":"hflelrALVtaO"},"source":["#LSTM"]},{"cell_type":"markdown","metadata":{"id":"Ixz4ceqdfTtV"},"source":["![lstmHW.jpg](data:image/jpeg;base64,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)"]},{"cell_type":"code","execution_count":14,"metadata":{"executionInfo":{"elapsed":9,"status":"ok","timestamp":1640148726407,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"FFh8XsQ9FDAH"},"outputs":[],"source":["# Hint: use an embedding layer to project your sequence data to a higher dimension,\n","# and use the diagram above for some ideas on layers to include in your model.\n","# Remember to compile your model after designing it.\n","# from tensorflow.keras.layers import Input, Dense, Dropout, Flatten, Activation\n","# x_input = Input(shape=(23,))\n","# embed = Embedding(24,40)(x_input)\n","# lstm = LSTM(50, kernel_regularizer=l2(0.01), recurrent_regularizer=l2(0.01), bias_regularizer=l2(0.01))(embed)\n","# x = Dropout(0.2)(lstm)\n","\n","# # softmax classifier\n","# x_output = Dense(10, activation='softmax')(x)\n","\n","# model1 = Model(inputs=x_input, outputs=x_output)\n","# model1.compile(optimizer='adam', loss='categorical_crossentropy', metrics=['accuracy'])\n","\n","# model1.summary()\n","\n"]},{"cell_type":"code","execution_count":15,"metadata":{"executionInfo":{"elapsed":10,"status":"ok","timestamp":1640148726408,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"h0L2502g6iS_"},"outputs":[],"source":["x_train_lstm = x_train.reshape(x_train.shape[0],10000,23)\n","x_val_lstm = x_val.reshape(x_val.shape[0],10000,23)\n","x_test_lstm = x_test.reshape(x_test.shape[0],10000,23)"]},{"cell_type":"code","execution_count":16,"metadata":{"executionInfo":{"elapsed":10,"status":"ok","timestamp":1640148726409,"user":{"displayName":"Neil Kennedy","photoUrl":"https://lh3.googleusercontent.com/a/default-user=s64","userId":"10401538321009628157"},"user_tz":300},"id":"NGe5QmEpV-3-"},"outputs":[],"source":["try:\n","  del dataset1\n","  del x_train\n","  del x_val\n","  del x_test\n","except:\n","  pass"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"background_save":true,"base_uri":"https://localhost:8080/"},"id":"aPAaDxAA7etY"},"outputs":[{"name":"stdout","output_type":"stream","text":["Model: \"sequential\"\n","_________________________________________________________________\n"," Layer (type)                Output Shape              Param #   \n","=================================================================\n"," lstm (LSTM)                 (None, 40)                10240     \n","                                                                 \n"," dense_1 (Dense)             (None, 10)                410       \n","                                                                 \n","=================================================================\n","Total params: 10,650\n","Trainable params: 10,650\n","Non-trainable params: 0\n","_________________________________________________________________\n","Epoch 1/10\n","19/19 - 78s - loss: nan - val_loss: nan - 78s/epoch - 4s/step\n","Epoch 2/10\n","19/19 - 74s - loss: nan - val_loss: nan - 74s/epoch - 4s/step\n","Epoch 3/10\n","19/19 - 73s - loss: nan - val_loss: nan - 73s/epoch - 4s/step\n","Epoch 4/10\n","19/19 - 73s - loss: nan - val_loss: nan - 73s/epoch - 4s/step\n","Epoch 5/10\n","19/19 - 73s - loss: nan - val_loss: nan - 73s/epoch - 4s/step\n","Epoch 6/10\n","19/19 - 73s - loss: nan - val_loss: nan - 73s/epoch - 4s/step\n","Epoch 7/10\n","19/19 - 73s - loss: nan - val_loss: nan - 73s/epoch - 4s/step\n","Epoch 8/10\n","19/19 - 73s - loss: nan - val_loss: nan - 73s/epoch - 4s/step\n","Epoch 9/10\n","19/19 - 73s - loss: nan - val_loss: nan - 73s/epoch - 4s/step\n","Epoch 10/10\n","19/19 - 73s - loss: nan - val_loss: nan - 73s/epoch - 4s/step\n"]}],"source":["from keras import Sequential\n","model3 = Sequential()\n","model3.add(LSTM(40, input_shape=(10000, 23)))\n","model3.add(Dense(10))\n","model3.summary()\n","model3.compile(loss='categorical_crossentropy', optimizer='adam')\n","history = model3.fit(x_train_lstm, np.asarray(y_train), epochs=10, batch_size=32, validation_data=(x_val_lstm, np.asarray(y_val)), verbose=2, shuffle=False)"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"background_save":true},"id":"ugPLLi72WP9p"},"outputs":[],"source":["#Save your model in case you need to run again to save time with training\n","# model1.save_weights('/gdrive/My Drive/LSTM_Model1.h5')\n","model3.save_weights('LSTM_Model1.h5')"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"background_save":true},"id":"DeL6E2XwIesQ"},"outputs":[],"source":["model3.load_weights('LSTM_Model1.h5')"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"background_save":true},"id":"colfaSU4IesQ"},"outputs":[{"ename":"NameError","evalue":"ignored","output_type":"error","traceback":["\u001b[0;31m---------------------------------------------------------------------------\u001b[0m","\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)","\u001b[0;32m\u003cipython-input-20-5c4baee970a5\u003e\u001b[0m in \u001b[0;36m\u003cmodule\u003e\u001b[0;34m()\u001b[0m\n\u001b[0;32m----\u003e 1\u001b[0;31m \u001b[0mextractor\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mModel\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmodel1\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0minputs\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0moutputs\u001b[0m\u001b[0;34m=\u001b[0m\u001b[0mmodel3\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mlayers\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m1\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0moutput\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m\u001b[1;32m      2\u001b[0m \u001b[0mextractor\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0mx_val\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;31mNameError\u001b[0m: name 'model1' is not defined"]}],"source":["extractor=Model(inputs=model1.inputs, outputs=model3.layers[1].output)\n","extractor(x_val).shape"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"background_save":true},"id":"VThFcWCiSs4-"},"outputs":[],"source":["from sklearn.metrics import confusion_matrix\n","def cm_testdata(model,data_test,label_test,num_labels,letter_labels):\n","    # have model generate predictions on test data\n","    y_pred=model3.predict(data_test)\n","    \n","    # find classes with the largest predicted probabilities\n","    y_pred_max = np.argmax(y_pred,axis=1)\n","    label_max=np.argmax(label_test, axis=1)\n","    # plot confusion matrix\n","    cm = confusion_matrix(y_true=label_max, y_pred=y_pred_max)\n","    cm_sum = np.sum(cm, axis=1, keepdims=True)\n","    cm_perc = cm / cm_sum.astype(float) * 100\n","    annot = np.empty_like(cm).astype(str)\n","    nrows, ncols = cm.shape\n","    for i in range(nrows):\n","        for j in range(ncols):\n","            c = cm[i, j]\n","            p = cm_perc[i, j]\n","            if i == j:\n","                s = cm_sum[i]\n","                annot[i, j] = '%.1f%%\\n%d/%d' % (p, c, s)\n","            elif c == 0:\n","                annot[i, j] = ''\n","            else:\n","                annot[i, j] = '%.1f%%\\n%d' % (p, c)\n","    fig, ax = plt.subplots(figsize=(10,10))\n","    sns.heatmap(cm, cmap= \"YlGnBu\", annot=annot, fmt='', ax=ax) \n","\n","    #accuracy\n","\n","    correct = (y_pred_max == (label_max))\n","    accuracy = correct.sum() / correct.size *100\n","    print(\"Accuracy: \"+str(accuracy)+\"%\")\n","\n","num_labels=6\n","letter_labels = range(1,7)\n","cm_testdata(model3,x_test_lstm,y_test,num_labels,letter_labels)"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"background_save":true},"id":"E6rV8fIZTxm4"},"outputs":[],"source":["y_pred=model3.predict(x_test_lstm)"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"background_save":true},"id":"LbPstuStUAbP"},"outputs":[],"source":["np.argmax(y_test,axis=1)"]}],"metadata":{"colab":{"collapsed_sections":[],"name":"Network.ipynb","version":""},"interpreter":{"hash":"2c64d0c5992f3f181c8bce812f63a0c7071bcd02b2731ba7dc467e66152e260a"},"kernelspec":{"display_name":"Python 3.9.7 64-bit (windows store)","name":"python3"},"language_info":{"codemirror_mode":{"name":"ipython","version":3},"file_extension":".py","mimetype":"text/x-python","name":"python","nbconvert_exporter":"python","pygments_lexer":"ipython3","version":"3.9.7"}},"nbformat":4,"nbformat_minor":0}