multi_label_toxity_classification.ipynb

{"cells":[{"cell_type":"code","execution_count":null,"metadata":{"_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","id":"bGxUiVeW_jV3"},"outputs":[],"source":["\n"]},{"cell_type":"code","execution_count":6,"metadata":{"id":"gkcDdB7N_jV6","executionInfo":{"status":"ok","timestamp":1672732881521,"user_tz":-540,"elapsed":1,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}}},"outputs":[],"source":["# import packages\n","import json\n","import numpy as np # linear algebra\n","import pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n","import os \n","import tensorflow as tf\n","import sklearn\n","import seaborn as sbs\n","import sklearn.naive_bayes \n","import sklearn.model_selection\n","import sklearn.metrics\n","import numpy as np # linear algebra\n","import pandas as pd"]},{"cell_type":"code","execution_count":7,"metadata":{"_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","id":"omVMhQCN_jV6","executionInfo":{"status":"ok","timestamp":1672732881522,"user_tz":-540,"elapsed":2,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}}},"outputs":[],"source":["json_1 = '/content/drive/MyDrive/datasets/CB_dataset/train.csv'\n","# json_2 = '../input/news-headlines-dataset-for-sarcasm-detection/Sarcasm_Headlines_Dataset_v2.json'\n"]},{"cell_type":"markdown","metadata":{"id":"L1r36JA9_jV7"},"source":["Load data into `Pandas.DataFrame`"]},{"cell_type":"code","execution_count":8,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":206},"id":"aGBTKP-m_jV8","executionInfo":{"status":"ok","timestamp":1672732886421,"user_tz":-540,"elapsed":4256,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"605643cf-e841-4e7d-b2a4-40d5958cd5d8"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["                                        comment_text  toxic  severe_toxic  \\\n","0  Explanation\\nWhy the edits made under my usern...      0             0   \n","1  D'aww! He matches this background colour I'm s...      0             0   \n","2  Hey man, I'm really not trying to edit war. It...      0             0   \n","3  \"\\nMore\\nI can't make any real suggestions on ...      0             0   \n","4  You, sir, are my hero. Any chance you remember...      0             0   \n","\n","   obscene  threat  insult  identity_hate  \n","0        0       0       0              0  \n","1        0       0       0              0  \n","2        0       0       0              0  \n","3        0       0       0              0  \n","4        0       0       0              0  "],"text/html":["\n","  <div id=\"df-af167ccd-d7df-4d50-b553-f5f49e1abcfd\">\n","    <div class=\"colab-df-container\">\n","      <div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>comment_text</th>\n","      <th>toxic</th>\n","      <th>severe_toxic</th>\n","      <th>obscene</th>\n","      <th>threat</th>\n","      <th>insult</th>\n","      <th>identity_hate</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>Explanation\\nWhy the edits made under my usern...</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>D'aww! He matches this background colour I'm s...</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>Hey man, I'm really not trying to edit war. It...</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>\"\\nMore\\nI can't make any real suggestions on ...</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>You, sir, are my hero. Any chance you remember...</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>\n","      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-af167ccd-d7df-4d50-b553-f5f49e1abcfd')\"\n","              title=\"Convert this dataframe to an interactive table.\"\n","              style=\"display:none;\">\n","        \n","  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n","       width=\"24px\">\n","    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n","    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n","  </svg>\n","      </button>\n","      \n","  <style>\n","    .colab-df-container {\n","      display:flex;\n","      flex-wrap:wrap;\n","      gap: 12px;\n","    }\n","\n","    .colab-df-convert {\n","      background-color: #E8F0FE;\n","      border: none;\n","      border-radius: 50%;\n","      cursor: pointer;\n","      display: none;\n","      fill: #1967D2;\n","      height: 32px;\n","      padding: 0 0 0 0;\n","      width: 32px;\n","    }\n","\n","    .colab-df-convert:hover {\n","      background-color: #E2EBFA;\n","      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n","      fill: #174EA6;\n","    }\n","\n","    [theme=dark] .colab-df-convert {\n","      background-color: #3B4455;\n","      fill: #D2E3FC;\n","    }\n","\n","    [theme=dark] .colab-df-convert:hover {\n","      background-color: #434B5C;\n","      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n","      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n","      fill: #FFFFFF;\n","    }\n","  </style>\n","\n","      <script>\n","        const buttonEl =\n","          document.querySelector('#df-af167ccd-d7df-4d50-b553-f5f49e1abcfd button.colab-df-convert');\n","        buttonEl.style.display =\n","          google.colab.kernel.accessAllowed ? 'block' : 'none';\n","\n","        async function convertToInteractive(key) {\n","          const element = document.querySelector('#df-af167ccd-d7df-4d50-b553-f5f49e1abcfd');\n","          const dataTable =\n","            await google.colab.kernel.invokeFunction('convertToInteractive',\n","                                                     [key], {});\n","          if (!dataTable) return;\n","\n","          const docLinkHtml = 'Like what you see? Visit the ' +\n","            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n","            + ' to learn more about interactive tables.';\n","          element.innerHTML = '';\n","          dataTable['output_type'] = 'display_data';\n","          await google.colab.output.renderOutput(dataTable, element);\n","          const docLink = document.createElement('div');\n","          docLink.innerHTML = docLinkHtml;\n","          element.appendChild(docLink);\n","        }\n","      </script>\n","    </div>\n","  </div>\n","  "]},"metadata":{},"execution_count":8}],"source":["toxic_data = pd.read_csv(json_1)\n","toxic_data.drop(['id'], axis=1, inplace=True)\n","toxic_data.head()"]},{"cell_type":"code","execution_count":9,"metadata":{"id":"gT6uwyAM_jV8","executionInfo":{"status":"ok","timestamp":1672732887476,"user_tz":-540,"elapsed":1065,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}}},"outputs":[],"source":["X = toxic_data[\"comment_text\"]\n","Y = toxic_data[[\"toxic\", \"severe_toxic\", \"obscene\", \"threat\", \"insult\", \"identity_hate\"]]"]},{"cell_type":"code","source":["\n","# encode class values as integers\n","from sklearn.preprocessing import LabelEncoder\n","from keras.utils.np_utils import to_categorical\n","# convert integers to dummy variables (i.e. one hot encoded)\n","dummy_y = to_categorical(Y)"],"metadata":{"id":"KVl6NbWAFNMs","executionInfo":{"status":"ok","timestamp":1672732893603,"user_tz":-540,"elapsed":629,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}}},"execution_count":10,"outputs":[]},{"cell_type":"code","source":["dummy_y[6]\n","\n","# it is basically like shru k 3, 1 hain to 1 index par 1 hua va sb k aagy, 4th 0 hai to 0th index 1 hua va, 1 mean on and 0 mean off. "],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"TpuJdDGgF2vc","executionInfo":{"status":"ok","timestamp":1672732901466,"user_tz":-540,"elapsed":498,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"2aa012d1-84b4-4421-8abd-752083a8c7c9"},"execution_count":11,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([[0., 1.],\n","       [0., 1.],\n","       [0., 1.],\n","       [1., 0.],\n","       [0., 1.],\n","       [1., 0.]], dtype=float32)"]},"metadata":{},"execution_count":11}]},{"cell_type":"markdown","metadata":{"id":"SmlXnRJx_jV9"},"source":["Now check the distribution (ratio) of our target `is_sarcastic`."]},{"cell_type":"code","execution_count":null,"metadata":{"id":"jqDOYc1F_jV9"},"outputs":[],"source":["# df.is_sarcastic.value_counts(normalize=True), df.is_sarcastic.value_counts()"]},{"cell_type":"code","execution_count":12,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":352},"id":"GPFYKLvN_jV-","executionInfo":{"status":"ok","timestamp":1672732905991,"user_tz":-540,"elapsed":1434,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"e178132e-a34f-421d-dbed-b5b5f805114d"},"outputs":[{"output_type":"stream","name":"stderr","text":["/usr/local/lib/python3.8/dist-packages/seaborn/distributions.py:2619: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n","  warnings.warn(msg, FutureWarning)\n"]},{"output_type":"execute_result","data":{"text/plain":["<matplotlib.axes._subplots.AxesSubplot at 0x7f874129adf0>"]},"metadata":{},"execution_count":12},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["# A birdview of headline length. Seems the majority has a length of 70, that's about 5 to 15 words, which makes sence.\n","sbs.distplot(toxic_data.comment_text.str.len())\n"]},{"cell_type":"markdown","metadata":{"id":"nKQvjOmd_jV-"},"source":["The data is well-balanced, which is rare in reality but great for our practicing purpose.\n","\n","Now we will try Naive Bayes as our base model.\n","\n","## Building Base Model: Naive Bayes\n","First step, we should tokenize texts into vectors."]},{"cell_type":"code","execution_count":13,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":341},"id":"cmjVf1UL_jV-","executionInfo":{"status":"ok","timestamp":1672732924113,"user_tz":-540,"elapsed":16995,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"4567133c-556a-47a0-8ef6-3201800bbbd0"},"outputs":[{"output_type":"stream","name":"stderr","text":["/usr/local/lib/python3.8/dist-packages/seaborn/distributions.py:2619: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).\n","  warnings.warn(msg, FutureWarning)\n"]},{"output_type":"execute_result","data":{"text/plain":["<matplotlib.axes._subplots.AxesSubplot at 0x7f873f1a68e0>"]},"metadata":{},"execution_count":13},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}],"source":["vocab_size = 10000 # max_features \n","tokenizer = tf.keras.preprocessing.text.Tokenizer(num_words=vocab_size)\n","tokenizer.fit_on_texts(X)\n","train_inputs = tokenizer.texts_to_sequences(X)\n","sbs.distplot([len(l) for l in train_inputs])"]},{"cell_type":"markdown","metadata":{"id":"CArQBl03_jV_"},"source":["See, the above graph confirms our guess: most headline has 5 - 15 words. \n"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"kqWhahaT_jV_","executionInfo":{"status":"ok","timestamp":1672235125261,"user_tz":-540,"elapsed":1071,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"2d769d70-b389-444c-c6af-633b9db218cd"},"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([ 688,   75,    1,  126,  130,  177,   29,  672, 4511, 1116,   86,\n","        331,   51, 2278,   50, 6864,   15,   60, 2756,  148,    7, 2937,\n","         34,  117, 1221, 2825,    4,   45,   59,  244,    1,  365,   31,\n","          1,   38,   27,  143,   73, 3462,   89, 3085, 4583, 2273,  985,\n","          0,    0,    0,    0,    0,    0], dtype=int32)"]},"metadata":{},"execution_count":16}],"source":["train_inputs = tf.keras.preprocessing.sequence.pad_sequences(train_inputs, padding='post', maxlen=50)\n","train_labels = dummy_y\n","\n","# Split data into train /validation \n","X_train, X_val, y_train, y_val = sklearn.model_selection.train_test_split(train_inputs, train_labels, test_size=0.2, random_state=0)\n","train_inputs[0]"]},{"cell_type":"code","source":["X_train.shape"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"0pPJQfZc-xpS","executionInfo":{"status":"ok","timestamp":1672235142525,"user_tz":-540,"elapsed":521,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"14976bde-dd0f-457a-ecdd-0961fc46b27a"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(127656, 50)"]},"metadata":{},"execution_count":19}]},{"cell_type":"code","source":[],"metadata":{"id":"IqMqpTQZNG7f"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"jtlOuSrr_jWA"},"source":["Validate how the model performs. \n","\n","Turs out that the `accuracy_score` is only 0.58, which is merely a little better than random guess.\n","But at least wew know what to expect now. We're going to try more powerful models, e.g., CNN, to see whether we can reach a `accuracy_score = 0.7` or higher."]},{"cell_type":"code","execution_count":null,"metadata":{"id":"s70tiYZI_jWA","outputId":"b1edaf3a-03e7-48cf-cabe-2f12dfc31efa"},"outputs":[{"name":"stdout","output_type":"stream","text":["Accuracy score 0.5718831898165481\n","Classification report\n","               precision    recall  f1-score   support\n","\n","           0       0.60      0.69      0.64      3007\n","           1       0.51      0.42      0.46      2335\n","\n","    accuracy                           0.57      5342\n","   macro avg       0.56      0.55      0.55      5342\n","weighted avg       0.56      0.57      0.56      5342\n","\n"]}],"source":["y_preds = nb.predict(X_val)\n","print(f\"Accuracy score\", sklearn.metrics.accuracy_score(y_val, y_preds))\n","print(f\"Classification report\\n\", sklearn.metrics.classification_report(y_val, y_preds))"]},{"cell_type":"markdown","metadata":{"id":"14eAPkj8_jWA"},"source":["# Another model: Keras LSTM"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"PkEn08H2_jWB","outputId":"82f3de98-ee3b-47b7-d244-1f3796951653"},"outputs":[{"name":"stdout","output_type":"stream","text":["Model: \"model\"\n","_________________________________________________________________\n","Layer (type)                 Output Shape              Param #   \n","=================================================================\n","input_1 (InputLayer)         [(None, 20)]              0         \n","_________________________________________________________________\n","embedding (Embedding)        (None, 20, 128)           1280000   \n","_________________________________________________________________\n","bidirectional (Bidirectional (None, 20, 64)            41216     \n","_________________________________________________________________\n","global_max_pooling1d (Global (None, 64)                0         \n","_________________________________________________________________\n","dense (Dense)                (None, 64)                4160      \n","_________________________________________________________________\n","dropout (Dropout)            (None, 64)                0         \n","_________________________________________________________________\n","dense_1 (Dense)              (None, 32)                2080      \n","_________________________________________________________________\n","dropout_1 (Dropout)          (None, 32)                0         \n","_________________________________________________________________\n","dense_2 (Dense)              (None, 1)                 33        \n","=================================================================\n","Total params: 1,327,489\n","Trainable params: 1,327,489\n","Non-trainable params: 0\n","_________________________________________________________________\n"]}],"source":["max_len = 50\n","text_input = tf.keras.Input(shape=(max_len, ))\n","embed_text = tf.keras.layers.Embedding(vocab_size, 128)(text_input)\n","\n","net = tf.keras.layers.Bidirectional(tf.keras.layers.LSTM(32, return_sequences=True))(embed_text)\n","net = tf.keras.layers.GlobalMaxPool1D()(net)\n","net = tf.keras.layers.Dense(64, activation='relu')(net)\n","net = tf.keras.layers.Dropout(0.4)(net)\n","net = tf.keras.layers.Dense(32, activation='relu')(net)\n","net = tf.keras.layers.Dropout(0.4)(net)\n","\n","output = tf.keras.layers.Dense(1, activation='sigmoid')(net)\n","model = tf.keras.models.Model(text_input, output)\n","model.summary()"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"_qo7lna8_jWB","outputId":"f466231c-d6d9-4096-db0d-fb1df77baa51"},"outputs":[{"name":"stdout","output_type":"stream","text":["Train on 19230 samples, validate on 2137 samples\n","Epoch 1/30\n","19200/19230 [============================>.] - ETA: 0s - loss: 0.5742 - accuracy: 0.6864\n","Epoch 00001: val_accuracy improved from -inf to 0.85400, saving model to model_best.hdf5\n","19230/19230 [==============================] - 8s 419us/sample - loss: 0.5738 - accuracy: 0.6866 - val_loss: 0.3461 - val_accuracy: 0.8540\n","Epoch 2/30\n","18688/19230 [============================>.] - ETA: 0s - loss: 0.2827 - accuracy: 0.8939\n","Epoch 00002: val_accuracy improved from 0.85400 to 0.86055, saving model to model_best.hdf5\n","19230/19230 [==============================] - 2s 85us/sample - loss: 0.2826 - accuracy: 0.8940 - val_loss: 0.3323 - val_accuracy: 0.8606\n","Epoch 3/30\n","18688/19230 [============================>.] - ETA: 0s - loss: 0.1828 - accuracy: 0.9386\n","Epoch 00003: val_accuracy improved from 0.86055 to 0.86196, saving model to model_best.hdf5\n","19230/19230 [==============================] - 2s 97us/sample - loss: 0.1835 - accuracy: 0.9384 - val_loss: 0.3691 - val_accuracy: 0.8620\n","Epoch 4/30\n","18688/19230 [============================>.] - ETA: 0s - loss: 0.1239 - accuracy: 0.9603\n","Epoch 00004: val_accuracy did not improve from 0.86196\n","19230/19230 [==============================] - 2s 81us/sample - loss: 0.1248 - accuracy: 0.9600 - val_loss: 0.4346 - val_accuracy: 0.8493\n","Epoch 5/30\n","18688/19230 [============================>.] - ETA: 0s - loss: 0.0884 - accuracy: 0.9729\n","Epoch 00005: val_accuracy did not improve from 0.86196\n","19230/19230 [==============================] - 2s 89us/sample - loss: 0.0883 - accuracy: 0.9729 - val_loss: 0.5302 - val_accuracy: 0.8489\n","Epoch 6/30\n","18688/19230 [============================>.] - ETA: 0s - loss: 0.0590 - accuracy: 0.9821\n","Epoch 00006: val_accuracy did not improve from 0.86196\n","19230/19230 [==============================] - 2s 86us/sample - loss: 0.0588 - accuracy: 0.9822 - val_loss: 0.6247 - val_accuracy: 0.8484\n","Epoch 7/30\n","18688/19230 [============================>.] - ETA: 0s - loss: 0.0450 - accuracy: 0.9871\n","Epoch 00007: val_accuracy did not improve from 0.86196\n","19230/19230 [==============================] - 2s 82us/sample - loss: 0.0448 - accuracy: 0.9872 - val_loss: 0.6932 - val_accuracy: 0.8498\n","Epoch 8/30\n","18688/19230 [============================>.] - ETA: 0s - loss: 0.0308 - accuracy: 0.9903\n","Epoch 00008: val_accuracy did not improve from 0.86196\n","19230/19230 [==============================] - 2s 82us/sample - loss: 0.0321 - accuracy: 0.9900 - val_loss: 0.8198 - val_accuracy: 0.8484\n"]}],"source":["model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\n","\n","mc = tf.keras.callbacks.ModelCheckpoint('model_best.hdf5', monitor='val_accuracy', \n","                                        verbose=1, save_best_only=True, mode='max')\n","es = tf.keras.callbacks.EarlyStopping(monitor=\"val_accuracy\", mode=\"max\", patience=5)\n","    \n","history = model.fit(X_train, y_train,\n","                    epochs=30, batch_size=256, callbacks=[mc, es], \n","                    validation_split=0.1, verbose=1)"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"qDDHNeg6_jWB","outputId":"5b11187c-a487-4b41-bbfe-25cfc9f5b7ac"},"outputs":[{"name":"stdout","output_type":"stream","text":["Test accracy score 0.8554848371396481\n"]}],"source":["model = tf.keras.models.load_model('model_best.hdf5')\n","y_preds = model.predict(X_val, batch_size=1024).round().astype(int)\n","print(\"Test accracy score\", sklearn.metrics.accuracy_score(y_val, y_preds))"]},{"cell_type":"markdown","metadata":{"id":"KnpzCm-v_jWB"},"source":["Now we can see, the result is better. The validation accuracy has improved from `0.58` to `0.8521`, which is not ideal, but good enough to detect sarcasm by reading a headline of an article.\n","\n","A close view will also finds out that this model is apparrenly overfitting our dataset. We will also do experiments on word embedding and BERT later to seen any improvement of our model. "]},{"cell_type":"markdown","metadata":{"id":"pDbtUUQ-_jWB"},"source":["# WordVector Embedding + LSTM\n","\n","We now use pretrained wordvector https://code.google.com/archive/p/word2vec/ of Google News (after all, we're dealing with article headlines). \n","\n","The vectors trained on a part of the Google News dataset (about 100 billion words). The model contains 300-dimensional vectors for 3 million words and phrases. The phrases were obtained using a simple data-driven approach described in 'Distributed Representations of Words and Phrases and their Compositionality' (https://code.google.com/archive/p/word2vec/)."]},{"cell_type":"code","execution_count":null,"metadata":{"id":"OmN3-8Az_jWC","outputId":"538dd03b-15dd-4e84-d663-64e904c69c53"},"outputs":[{"name":"stdout","output_type":"stream","text":["CPU times: user 73.7 ms, sys: 14 ms, total: 87.7 ms\n","Wall time: 215 ms\n"]}],"source":["%%time\n","import gensim.downloader as api\n","\n","def embed_word_vector(word_index, pretrained='word2vec-google-news-300'):\n","    embed_size = 300 # Google news vector is 300-dimensional\n","    vector = api.load(pretrained)\n","    zeros = [0] * embed_size\n","    matrix = np.zeros((vocab_size, embed_size)) \n","    \n","    for word, i in word_index.items():\n","        if i >= vocab_size or word not in vector: continue \n","        matrix[i] = vector[word]\n","    \n","    print(\"Embed word vector completed.\")\n","    return matrix"]},{"cell_type":"markdown","metadata":{"id":"_It_Cf5r_jWC"},"source":["Update embedding layer weights of our model"]},{"cell_type":"code","execution_count":null,"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":148},"id":"TGMeq1uM_jWC","executionInfo":{"status":"ok","timestamp":1672237367117,"user_tz":-540,"elapsed":4,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"0ff38719-98dd-4cb9-a675-4d82e5126a00"},"outputs":[{"output_type":"error","ename":"NameError","evalue":"ignored","traceback":["\u001b[0;31m---------------------------------------------------------------------------\u001b[0m","\u001b[0;31mNameError\u001b[0m                                 Traceback (most recent call last)","\u001b[0;32m<timed exec>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n","\u001b[0;31mNameError\u001b[0m: name 'embed_word_vector' is not defined"]}],"source":["%%time\n","pretrained = 'glove-wiki-gigaword-300'\n","matrix = embed_word_vector(tokenizer.word_index, pretrained=pretrained)"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"zfHOe3u8_jWC","outputId":"8a861568-f1d4-45e1-f457-a64c9d3b821d"},"outputs":[{"name":"stdout","output_type":"stream","text":["Model: \"model_1\"\n","_________________________________________________________________\n","Layer (type)                 Output Shape              Param #   \n","=================================================================\n","input_2 (InputLayer)         [(None, 20)]              0         \n","_________________________________________________________________\n","embedding_1 (Embedding)      (None, 20, 300)           3000000   \n","_________________________________________________________________\n","bidirectional_1 (Bidirection (None, 20, 256)           330240    \n","_________________________________________________________________\n","global_max_pooling1d_1 (Glob (None, 256)               0         \n","_________________________________________________________________\n","dense_3 (Dense)              (None, 64)                16448     \n","_________________________________________________________________\n","dropout_2 (Dropout)          (None, 64)                0         \n","_________________________________________________________________\n","dense_4 (Dense)              (None, 32)                2080      \n","_________________________________________________________________\n","dropout_3 (Dropout)          (None, 32)                0         \n","_________________________________________________________________\n","dense_5 (Dense)              (None, 1)                 33        \n","=================================================================\n","Total params: 3,348,801\n","Trainable params: 348,801\n","Non-trainable params: 3,000,000\n","_________________________________________________________________\n"]}],"source":["max_len = 20\n","text_input = tf.keras.Input(shape=(max_len, ))\n","embed_text = tf.keras.layers.Embedding(vocab_size, 300, weights=[matrix], trainable=False)(text_input)\n","\n","net = tf.keras.layers.Bidirectional(tf.keras.layers.GRU(128, return_sequences=True))(embed_text)\n","net = tf.keras.layers.GlobalMaxPool1D()(net)\n","net = tf.keras.layers.Dense(64, activation='relu')(net)\n","net = tf.keras.layers.Dropout(0.4)(net)\n","net = tf.keras.layers.Dense(32, activation='relu')(net)\n","net = tf.keras.layers.Dropout(0.4)(net)\n","\n","output = tf.keras.layers.Dense(1, activation='sigmoid')(net)\n","model = tf.keras.models.Model(text_input, output)\n","model.summary()"]},{"cell_type":"markdown","metadata":{"id":"b2oXVdx5_jWC"},"source":["Re run our model to see if any improvements we got."]},{"cell_type":"code","execution_count":null,"metadata":{"id":"d7yUlO7l_jWD","outputId":"fc38bed4-55f8-4d90-f3b4-52eae1608aae"},"outputs":[{"name":"stdout","output_type":"stream","text":["Train on 19230 samples, validate on 2137 samples\n","Epoch 1/30\n","18688/19230 [============================>.] - ETA: 0s - loss: 0.5886 - accuracy: 0.6775\n","Epoch 00001: val_accuracy improved from -inf to 0.79364, saving model to model_best_embed.hdf5\n","19230/19230 [==============================] - 4s 203us/sample - loss: 0.5869 - accuracy: 0.6798 - val_loss: 0.4357 - val_accuracy: 0.7936\n","Epoch 2/30\n","17920/19230 [==========================>...] - ETA: 0s - loss: 0.4074 - accuracy: 0.8244\n","Epoch 00002: val_accuracy improved from 0.79364 to 0.83060, saving model to model_best_embed.hdf5\n","19230/19230 [==============================] - 1s 43us/sample - loss: 0.4078 - accuracy: 0.8245 - val_loss: 0.3755 - val_accuracy: 0.8306\n","Epoch 3/30\n","18176/19230 [===========================>..] - ETA: 0s - loss: 0.3285 - accuracy: 0.8683\n","Epoch 00003: val_accuracy improved from 0.83060 to 0.84558, saving model to model_best_embed.hdf5\n","19230/19230 [==============================] - 1s 45us/sample - loss: 0.3284 - accuracy: 0.8680 - val_loss: 0.3431 - val_accuracy: 0.8456\n","Epoch 4/30\n","18944/19230 [============================>.] - ETA: 0s - loss: 0.2698 - accuracy: 0.8926\n","Epoch 00004: val_accuracy improved from 0.84558 to 0.86336, saving model to model_best_embed.hdf5\n","19230/19230 [==============================] - 1s 49us/sample - loss: 0.2693 - accuracy: 0.8929 - val_loss: 0.3164 - val_accuracy: 0.8634\n","Epoch 5/30\n","18944/19230 [============================>.] - ETA: 0s - loss: 0.2179 - accuracy: 0.9185\n","Epoch 00005: val_accuracy improved from 0.86336 to 0.86664, saving model to model_best_embed.hdf5\n","19230/19230 [==============================] - 1s 46us/sample - loss: 0.2178 - accuracy: 0.9185 - val_loss: 0.3164 - val_accuracy: 0.8666\n","Epoch 6/30\n","19200/19230 [============================>.] - ETA: 0s - loss: 0.1632 - accuracy: 0.9413\n","Epoch 00006: val_accuracy improved from 0.86664 to 0.86944, saving model to model_best_embed.hdf5\n","19230/19230 [==============================] - 1s 51us/sample - loss: 0.1633 - accuracy: 0.9413 - val_loss: 0.3531 - val_accuracy: 0.8694\n","Epoch 7/30\n","18688/19230 [============================>.] - ETA: 0s - loss: 0.1283 - accuracy: 0.9569\n","Epoch 00007: val_accuracy did not improve from 0.86944\n","19230/19230 [==============================] - 1s 44us/sample - loss: 0.1286 - accuracy: 0.9566 - val_loss: 0.4032 - val_accuracy: 0.8615\n","Epoch 8/30\n","19200/19230 [============================>.] - ETA: 0s - loss: 0.0890 - accuracy: 0.9705\n","Epoch 00008: val_accuracy did not improve from 0.86944\n","19230/19230 [==============================] - 1s 49us/sample - loss: 0.0890 - accuracy: 0.9705 - val_loss: 0.4082 - val_accuracy: 0.8685\n","Epoch 9/30\n","18176/19230 [===========================>..] - ETA: 0s - loss: 0.0649 - accuracy: 0.9773\n","Epoch 00009: val_accuracy did not improve from 0.86944\n","19230/19230 [==============================] - 1s 47us/sample - loss: 0.0657 - accuracy: 0.9773 - val_loss: 0.5035 - val_accuracy: 0.8676\n","Epoch 10/30\n","18176/19230 [===========================>..] - ETA: 0s - loss: 0.0575 - accuracy: 0.9799\n","Epoch 00010: val_accuracy did not improve from 0.86944\n","19230/19230 [==============================] - 1s 43us/sample - loss: 0.0573 - accuracy: 0.9799 - val_loss: 0.5802 - val_accuracy: 0.8601\n","Epoch 11/30\n","18688/19230 [============================>.] - ETA: 0s - loss: 0.0400 - accuracy: 0.9868\n","Epoch 00011: val_accuracy did not improve from 0.86944\n","19230/19230 [==============================] - 1s 46us/sample - loss: 0.0407 - accuracy: 0.9866 - val_loss: 0.5833 - val_accuracy: 0.8540\n"]}],"source":["model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\n","\n","mc = tf.keras.callbacks.ModelCheckpoint('model_best_embed.hdf5', monitor='val_accuracy', \n","                                        verbose=1, save_best_only=True, mode='max')\n","es = tf.keras.callbacks.EarlyStopping(monitor=\"val_accuracy\", mode=\"max\", patience=5)\n","    \n","history = model.fit(X_train, y_train,\n","                    epochs=30, batch_size=256, callbacks=[mc, es], \n","                    validation_split=0.1, verbose=1)"]},{"cell_type":"markdown","metadata":{"id":"4fYI1xwM_jWD"},"source":["From the training history, we can see performance of this model is pretty similiar to our previous one. Thus, we know pretrained NLP models are not going help us much. We have to find other ways to improve the score."]},{"cell_type":"code","execution_count":null,"metadata":{"id":"iJ93n2GV_jWD","outputId":"7e980315-dd86-4f26-a807-8ea67e7325de"},"outputs":[{"name":"stdout","output_type":"stream","text":["Test accracy score 0.8627854736053913\n"]}],"source":["model = tf.keras.models.load_model('model_best_embed.hdf5')\n","y_preds = model.predict(X_val, batch_size=1024).round().astype(int)\n","print(\"Test accracy score\", sklearn.metrics.accuracy_score(y_val, y_preds))"]},{"cell_type":"markdown","metadata":{"id":"cjgN44-d_jWD"},"source":["# BERT, NLP text classification killer"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"eQNlc5KY_jWD"},"outputs":[],"source":["!wget --quiet https://raw.githubusercontent.com/tensorflow/models/master/official/nlp/bert/tokenization.py"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"3otJJDgg_jWD","outputId":"0e3236ef-917d-4ec1-e82b-2e61f667f48d"},"outputs":[{"name":"stdout","output_type":"stream","text":["CPU times: user 8.92 s, sys: 1.36 s, total: 10.3 s\n","Wall time: 11.2 s\n"]}],"source":["%%time\n","import tensorflow_hub as hub \n","import tokenization\n","\n","module_url = 'https://tfhub.dev/tensorflow/bert_en_uncased_L-12_H-768_A-12/2'\n","bert_layer = hub.KerasLayer(module_url, trainable=True)"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"NfoS2aAy_jWE"},"outputs":[],"source":["vocab_file = bert_layer.resolved_object.vocab_file.asset_path.numpy()\n","do_lower_case = bert_layer.resolved_object.do_lower_case.numpy()\n","tokenizer = tokenization.FullTokenizer(vocab_file, do_lower_case)\n"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"cqAm2n2v_jWE"},"outputs":[],"source":["def bert_encode(texts, tokenizer, max_len=512):\n","    all_tokens = []\n","    all_masks = []\n","    all_segments = []\n","    \n","    for text in texts:\n","        text = tokenizer.tokenize(text)\n","            \n","        text = text[:max_len-2]\n","        input_sequence = [\"[CLS]\"] + text + [\"[SEP]\"]\n","        pad_len = max_len - len(input_sequence)\n","        \n","        tokens = tokenizer.convert_tokens_to_ids(input_sequence) + [0] * pad_len\n","        pad_masks = [1] * len(input_sequence) + [0] * pad_len\n","        segment_ids = [0] * max_len\n","        \n","        all_tokens.append(tokens)\n","        all_masks.append(pad_masks)\n","        all_segments.append(segment_ids)\n","    \n","    return np.array(all_tokens), np.array(all_masks), np.array(all_segments)\n","\n","def build_model(bert_layer, max_len=512):\n","    input_word_ids = tf.keras.Input(shape=(max_len,), dtype=tf.int32, name=\"input_word_ids\")\n","    input_mask = tf.keras.Input(shape=(max_len,), dtype=tf.int32, name=\"input_mask\")\n","    segment_ids = tf.keras.Input(shape=(max_len,), dtype=tf.int32, name=\"segment_ids\")\n","\n","    pooled_output, sequence_output = bert_layer([input_word_ids, input_mask, segment_ids])\n","    clf_output = sequence_output[:, 0, :]\n","    net = tf.keras.layers.Dense(64, activation='relu')(clf_output)\n","    net = tf.keras.layers.Dropout(0.2)(net)\n","    net = tf.keras.layers.Dense(32, activation='relu')(net)\n","    net = tf.keras.layers.Dropout(0.2)(net)\n","    out = tf.keras.layers.Dense(1, activation='sigmoid')(net)\n","    \n","    model = tf.keras.models.Model(inputs=[input_word_ids, input_mask, segment_ids], outputs=out)\n","    model.compile(tf.keras.optimizers.Adam(lr=1e-5), loss='binary_crossentropy', metrics=['accuracy'])\n","    \n","    return model"]},{"cell_type":"markdown","metadata":{"id":"HoXcS6VF_jWE"},"source":["Remember in previous sections, we have explored on the headline length. Most of them has a length between 50 to 150, and the majority has a length of 70. We set the length to 120."]},{"cell_type":"code","execution_count":null,"metadata":{"id":"OrRHnD10_jWE"},"outputs":[],"source":["max_len = 100\n","X_train, X_val, y_train, y_val = sklearn.model_selection.train_test_split(df.headline, df['is_sarcastic'], \n","                                                                          test_size=0.1, random_state=0)\n","X_train = bert_encode(X_train, tokenizer, max_len=max_len)\n","X_val = bert_encode(X_val, tokenizer, max_len=max_len)"]},{"cell_type":"markdown","metadata":{"id":"4kxEn8Pv_jWE"},"source":["Let's build the model and train on it."]},{"cell_type":"code","execution_count":null,"metadata":{"id":"cr45aADG_jWE","outputId":"906baa04-9f13-4e6c-9033-4b4df5a1583e"},"outputs":[{"name":"stdout","output_type":"stream","text":["Model: \"model_2\"\n","__________________________________________________________________________________________________\n","Layer (type)                    Output Shape         Param #     Connected to                     \n","==================================================================================================\n","input_word_ids (InputLayer)     [(None, 100)]        0                                            \n","__________________________________________________________________________________________________\n","input_mask (InputLayer)         [(None, 100)]        0                                            \n","__________________________________________________________________________________________________\n","segment_ids (InputLayer)        [(None, 100)]        0                                            \n","__________________________________________________________________________________________________\n","keras_layer (KerasLayer)        [(None, 768), (None, 109482241   input_word_ids[0][0]             \n","                                                                 input_mask[0][0]                 \n","                                                                 segment_ids[0][0]                \n","__________________________________________________________________________________________________\n","tf_op_layer_strided_slice (Tens [(None, 768)]        0           keras_layer[0][1]                \n","__________________________________________________________________________________________________\n","dense_6 (Dense)                 (None, 64)           49216       tf_op_layer_strided_slice[0][0]  \n","__________________________________________________________________________________________________\n","dropout_4 (Dropout)             (None, 64)           0           dense_6[0][0]                    \n","__________________________________________________________________________________________________\n","dense_7 (Dense)                 (None, 32)           2080        dropout_4[0][0]                  \n","__________________________________________________________________________________________________\n","dropout_5 (Dropout)             (None, 32)           0           dense_7[0][0]                    \n","__________________________________________________________________________________________________\n","dense_8 (Dense)                 (None, 1)            33          dropout_5[0][0]                  \n","==================================================================================================\n","Total params: 109,533,570\n","Trainable params: 109,533,569\n","Non-trainable params: 1\n","__________________________________________________________________________________________________\n"]}],"source":["model = build_model(bert_layer, max_len=max_len)\n","model.summary()"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"6zRZe2Xe_jWE","outputId":"f8fd0c5f-b2a4-4269-c6d7-6d6b5d467de7"},"outputs":[{"name":"stdout","output_type":"stream","text":["Train on 21634 samples, validate on 2404 samples\n","Epoch 1/3\n","21632/21634 [============================>.] - ETA: 0s - loss: 0.3036 - accuracy: 0.8695\n","Epoch 00001: val_accuracy improved from -inf to 0.91597, saving model to model.h5\n","21634/21634 [==============================] - 304s 14ms/sample - loss: 0.3036 - accuracy: 0.8696 - val_loss: 0.2100 - val_accuracy: 0.9160\n","Epoch 2/3\n","21632/21634 [============================>.] - ETA: 0s - loss: 0.1521 - accuracy: 0.9443\n","Epoch 00002: val_accuracy improved from 0.91597 to 0.92554, saving model to model.h5\n","21634/21634 [==============================] - 288s 13ms/sample - loss: 0.1520 - accuracy: 0.9443 - val_loss: 0.2174 - val_accuracy: 0.9255\n","Epoch 3/3\n","21632/21634 [============================>.] - ETA: 0s - loss: 0.0843 - accuracy: 0.9714\n","Epoch 00003: val_accuracy did not improve from 0.92554\n","21634/21634 [==============================] - 285s 13ms/sample - loss: 0.0843 - accuracy: 0.9714 - val_loss: 0.2542 - val_accuracy: 0.9193\n","CPU times: user 9min 26s, sys: 3min 14s, total: 12min 41s\n","Wall time: 14min 36s\n"]}],"source":["%%time\n","checkpoint = tf.keras.callbacks.ModelCheckpoint('model.h5', monitor='val_accuracy', save_best_only=True, verbose=1)\n","earlystopping = tf.keras.callbacks.EarlyStopping(monitor='val_accuracy', patience=3, verbose=1)\n","\n","train_history = model.fit(\n","    X_train, y_train, \n","    validation_split=0.1,\n","    epochs=3,\n","    callbacks=[checkpoint, earlystopping],\n","    batch_size=16,\n","    verbose=1\n",")"]},{"cell_type":"code","execution_count":null,"metadata":{"id":"oCZ1HduU_jWF","outputId":"98e3388a-c4b7-4614-97ef-d4473a2855c4"},"outputs":[{"name":"stdout","output_type":"stream","text":["Validation accuracy:  0.9284912017970798\n","CPU times: user 7.51 s, sys: 3.76 s, total: 11.3 s\n","Wall time: 11.3 s\n"]}],"source":["%%time\n","model.load_weights('model.h5')\n","y_preds = model.predict(X_val).round().astype(int)\n","print(\"Validation accuracy: \", sklearn.metrics.accuracy_score(y_val, y_preds))"]},{"cell_type":"markdown","metadata":{"id":"f-17JAxl_jWF"},"source":["* The validation accuracy is `92%`, impressive !\n","\n","The result can be further improved by: \n","1. increasing the length of `max_len`, i.e., the length of input headline length \n","2. fine-tuning parameters \n","3. trying other pretrained bert modules.\n","\n","To summarize, BERT is definitely the winner of text classification tasks."]},{"cell_type":"markdown","source":["# **My Own code from here**"],"metadata":{"id":"IXzrynyPDqIS"}},{"cell_type":"code","source":["import tensorflow as tf\n","\n","from numpy import array\n","from keras.preprocessing.text import one_hot\n","# from keras.preprocessing.sequence import pad_sequences\n","# from keras.models import Sequential\n","from keras.layers.core import Activation, Dropout, Dense\n","from keras.layers import Flatten, LSTM\n","from keras.layers import GlobalMaxPooling1D\n","from keras.models import Model\n","# from keras.layers.embeddings import Embedding\n","from sklearn.model_selection import train_test_split\n","# from keras.preprocessing.text import Tokenizer\n","# from keras.layers import Input\n","# from keras.layers.merge import Concatenate\n","\n","import pandas as pd\n","import numpy as np\n","import re\n","\n","import matplotlib.pyplot as plt\n","\n","# tokenizer = tf.keras.preprocessing.text.Tokenizer(num_words=vocab_size)"],"metadata":{"id":"YhvXZt15DogM"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["toxic_comments = pd.read_csv(\"/content/drive/MyDrive/datasets/CB_dataset/train.csv\")"],"metadata":{"id":"-n2Xteu9Doi8"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["print(toxic_comments.shape)\n","\n","toxic_comments.head()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":223},"id":"F7J5FG1kDonL","executionInfo":{"status":"ok","timestamp":1672425425492,"user_tz":-540,"elapsed":11,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"ebbb3633-0bf5-4553-cc9d-cce718572558"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["(159571, 8)\n"]},{"output_type":"execute_result","data":{"text/plain":["                 id                                       comment_text  toxic  \\\n","0  0000997932d777bf  Explanation\\nWhy the edits made under my usern...      0   \n","1  000103f0d9cfb60f  D'aww! He matches this background colour I'm s...      0   \n","2  000113f07ec002fd  Hey man, I'm really not trying to edit war. It...      0   \n","3  0001b41b1c6bb37e  \"\\nMore\\nI can't make any real suggestions on ...      0   \n","4  0001d958c54c6e35  You, sir, are my hero. Any chance you remember...      0   \n","\n","   severe_toxic  obscene  threat  insult  identity_hate  \n","0             0        0       0       0              0  \n","1             0        0       0       0              0  \n","2             0        0       0       0              0  \n","3             0        0       0       0              0  \n","4             0        0       0       0              0  "],"text/html":["\n","  <div id=\"df-129ef867-0efb-4ea5-a3a2-c652a8ce49f4\">\n","    <div class=\"colab-df-container\">\n","      <div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>id</th>\n","      <th>comment_text</th>\n","      <th>toxic</th>\n","      <th>severe_toxic</th>\n","      <th>obscene</th>\n","      <th>threat</th>\n","      <th>insult</th>\n","      <th>identity_hate</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>0000997932d777bf</td>\n","      <td>Explanation\\nWhy the edits made under my usern...</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>000103f0d9cfb60f</td>\n","      <td>D'aww! He matches this background colour I'm s...</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>000113f07ec002fd</td>\n","      <td>Hey man, I'm really not trying to edit war. It...</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>0001b41b1c6bb37e</td>\n","      <td>\"\\nMore\\nI can't make any real suggestions on ...</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>0001d958c54c6e35</td>\n","      <td>You, sir, are my hero. Any chance you remember...</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>\n","      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-129ef867-0efb-4ea5-a3a2-c652a8ce49f4')\"\n","              title=\"Convert this dataframe to an interactive table.\"\n","              style=\"display:none;\">\n","        \n","  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n","       width=\"24px\">\n","    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n","    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n","  </svg>\n","      </button>\n","      \n","  <style>\n","    .colab-df-container {\n","      display:flex;\n","      flex-wrap:wrap;\n","      gap: 12px;\n","    }\n","\n","    .colab-df-convert {\n","      background-color: #E8F0FE;\n","      border: none;\n","      border-radius: 50%;\n","      cursor: pointer;\n","      display: none;\n","      fill: #1967D2;\n","      height: 32px;\n","      padding: 0 0 0 0;\n","      width: 32px;\n","    }\n","\n","    .colab-df-convert:hover {\n","      background-color: #E2EBFA;\n","      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n","      fill: #174EA6;\n","    }\n","\n","    [theme=dark] .colab-df-convert {\n","      background-color: #3B4455;\n","      fill: #D2E3FC;\n","    }\n","\n","    [theme=dark] .colab-df-convert:hover {\n","      background-color: #434B5C;\n","      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n","      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n","      fill: #FFFFFF;\n","    }\n","  </style>\n","\n","      <script>\n","        const buttonEl =\n","          document.querySelector('#df-129ef867-0efb-4ea5-a3a2-c652a8ce49f4 button.colab-df-convert');\n","        buttonEl.style.display =\n","          google.colab.kernel.accessAllowed ? 'block' : 'none';\n","\n","        async function convertToInteractive(key) {\n","          const element = document.querySelector('#df-129ef867-0efb-4ea5-a3a2-c652a8ce49f4');\n","          const dataTable =\n","            await google.colab.kernel.invokeFunction('convertToInteractive',\n","                                                     [key], {});\n","          if (!dataTable) return;\n","\n","          const docLinkHtml = 'Like what you see? Visit the ' +\n","            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n","            + ' to learn more about interactive tables.';\n","          element.innerHTML = '';\n","          dataTable['output_type'] = 'display_data';\n","          await google.colab.output.renderOutput(dataTable, element);\n","          const docLink = document.createElement('div');\n","          docLink.innerHTML = docLinkHtml;\n","          element.appendChild(docLink);\n","        }\n","      </script>\n","    </div>\n","  </div>\n","  "]},"metadata":{},"execution_count":4}]},{"cell_type":"code","source":["print(\"Toxic:\" + str(toxic_comments[\"toxic\"][168]))\n","print(\"Severe_toxic:\" + str(toxic_comments[\"severe_toxic\"][168]))\n","print(\"Obscene:\" + str(toxic_comments[\"obscene\"][168]))\n","print(\"Threat:\" + str(toxic_comments[\"threat\"][168]))\n","print(\"Insult:\" + str(toxic_comments[\"insult\"][168]))\n","print(\"Identity_hate:\" + str(toxic_comments[\"identity_hate\"][168]))"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"iTaLJgyEDo1v","executionInfo":{"status":"ok","timestamp":1672425462695,"user_tz":-540,"elapsed":6,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"40f06587-b965-42c4-ac4c-911b88beee0e"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Toxic:1\n","Severe_toxic:0\n","Obscene:0\n","Threat:0\n","Insult:1\n","Identity_hate:0\n"]}]},{"cell_type":"code","source":["toxic_comments_labels = toxic_comments[[\"toxic\", \"severe_toxic\", \"obscene\", \"threat\", \"insult\", \"identity_hate\"]]\n","toxic_comments_labels.head()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":206},"id":"ilwxiHOPDo4B","executionInfo":{"status":"ok","timestamp":1672425462695,"user_tz":-540,"elapsed":4,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"a33038b5-7259-4de2-ce5a-9ad1b60a7cd7"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["   toxic  severe_toxic  obscene  threat  insult  identity_hate\n","0      0             0        0       0       0              0\n","1      0             0        0       0       0              0\n","2      0             0        0       0       0              0\n","3      0             0        0       0       0              0\n","4      0             0        0       0       0              0"],"text/html":["\n","  <div id=\"df-74bf3c5d-f6db-44c5-b444-881300783047\">\n","    <div class=\"colab-df-container\">\n","      <div>\n","<style scoped>\n","    .dataframe tbody tr th:only-of-type {\n","        vertical-align: middle;\n","    }\n","\n","    .dataframe tbody tr th {\n","        vertical-align: top;\n","    }\n","\n","    .dataframe thead th {\n","        text-align: right;\n","    }\n","</style>\n","<table border=\"1\" class=\"dataframe\">\n","  <thead>\n","    <tr style=\"text-align: right;\">\n","      <th></th>\n","      <th>toxic</th>\n","      <th>severe_toxic</th>\n","      <th>obscene</th>\n","      <th>threat</th>\n","      <th>insult</th>\n","      <th>identity_hate</th>\n","    </tr>\n","  </thead>\n","  <tbody>\n","    <tr>\n","      <th>0</th>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>1</th>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>2</th>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>3</th>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","    <tr>\n","      <th>4</th>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","      <td>0</td>\n","    </tr>\n","  </tbody>\n","</table>\n","</div>\n","      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-74bf3c5d-f6db-44c5-b444-881300783047')\"\n","              title=\"Convert this dataframe to an interactive table.\"\n","              style=\"display:none;\">\n","        \n","  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n","       width=\"24px\">\n","    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n","    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n","  </svg>\n","      </button>\n","      \n","  <style>\n","    .colab-df-container {\n","      display:flex;\n","      flex-wrap:wrap;\n","      gap: 12px;\n","    }\n","\n","    .colab-df-convert {\n","      background-color: #E8F0FE;\n","      border: none;\n","      border-radius: 50%;\n","      cursor: pointer;\n","      display: none;\n","      fill: #1967D2;\n","      height: 32px;\n","      padding: 0 0 0 0;\n","      width: 32px;\n","    }\n","\n","    .colab-df-convert:hover {\n","      background-color: #E2EBFA;\n","      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n","      fill: #174EA6;\n","    }\n","\n","    [theme=dark] .colab-df-convert {\n","      background-color: #3B4455;\n","      fill: #D2E3FC;\n","    }\n","\n","    [theme=dark] .colab-df-convert:hover {\n","      background-color: #434B5C;\n","      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n","      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n","      fill: #FFFFFF;\n","    }\n","  </style>\n","\n","      <script>\n","        const buttonEl =\n","          document.querySelector('#df-74bf3c5d-f6db-44c5-b444-881300783047 button.colab-df-convert');\n","        buttonEl.style.display =\n","          google.colab.kernel.accessAllowed ? 'block' : 'none';\n","\n","        async function convertToInteractive(key) {\n","          const element = document.querySelector('#df-74bf3c5d-f6db-44c5-b444-881300783047');\n","          const dataTable =\n","            await google.colab.kernel.invokeFunction('convertToInteractive',\n","                                                     [key], {});\n","          if (!dataTable) return;\n","\n","          const docLinkHtml = 'Like what you see? Visit the ' +\n","            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n","            + ' to learn more about interactive tables.';\n","          element.innerHTML = '';\n","          dataTable['output_type'] = 'display_data';\n","          await google.colab.output.renderOutput(dataTable, element);\n","          const docLink = document.createElement('div');\n","          docLink.innerHTML = docLinkHtml;\n","          element.appendChild(docLink);\n","        }\n","      </script>\n","    </div>\n","  </div>\n","  "]},"metadata":{},"execution_count":6}]},{"cell_type":"code","source":["from sklearn.preprocessing import MultiLabelBinarizer\n","\n","mlb = MultiLabelBinarizer()\n","y_one_hot = mlb.fit_transform(toxic_comments_labels.values)"],"metadata":{"id":"XS7BwlVt08Y8"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["y_one_hot.shape"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"B0wPr6xR3i4f","executionInfo":{"status":"ok","timestamp":1672425463080,"user_tz":-540,"elapsed":388,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"148f19b9-bdf3-4fd3-d784-e7ec97849978"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(159571, 2)"]},"metadata":{},"execution_count":8}]},{"cell_type":"code","source":["fig_size = plt.rcParams[\"figure.figsize\"]\n","fig_size[0] = 10\n","fig_size[1] = 8\n","plt.rcParams[\"figure.figsize\"] = fig_size\n","\n","toxic_comments_labels.sum(axis=0).plot.bar()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":558},"id":"0hHNj7RJDo6U","executionInfo":{"status":"ok","timestamp":1672425463080,"user_tz":-540,"elapsed":6,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"f67fea38-0231-4e8f-fa0f-655e370a6908"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["<matplotlib.axes._subplots.AxesSubplot at 0x7f6cca2f3e20>"]},"metadata":{},"execution_count":9},{"output_type":"display_data","data":{"text/plain":["<Figure size 720x576 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":["# You can see that the \"toxic\" comment has the highest frequency of occurrence followed by \"obscene\" and \"insult\", respectively."],"metadata":{"id":"5Xe9mpGCDo8t"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["**Creating Multi-label Text Classification Models**"],"metadata":{"id":"xiW1ZnXHFenz"}},{"cell_type":"code","source":["# Multi-lable Text Classification Model with Single Output Layer\n","\n","def preprocess_text(sen):\n","    # Remove punctuations and numbers\n","    sentence = re.sub('[^a-zA-Z]', ' ', sen)\n","\n","    # Single character removal\n","    sentence = re.sub(r\"\\s+[a-zA-Z]\\s+\", ' ', sentence)\n","\n","    # Removing multiple spaces\n","    sentence = re.sub(r'\\s+', ' ', sentence)\n","\n","    return sentence"],"metadata":{"id":"zNUVIYagDo_B"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["X = []\n","sentences = list(toxic_comments[\"comment_text\"])\n","for sen in sentences:\n","    X.append(preprocess_text(sen))\n","\n","y = toxic_comments_labels.values"],"metadata":{"id":"yMak4y8TDpBM"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# Here we do not need to perform any one-hot encoding because our output labels are already in the form of one-hot encoded vectors.\n","# In the next step, we will divide our data into training and test sets:"],"metadata":{"id":"c_m2BN7QDpDa"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=42)"],"metadata":{"id":"OPI2xbk5DpG4"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# We need to convert text inputs into embedded vectors."],"metadata":{"id":"oSV5ZvQnGHoV"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["vocab_size = 5000\n","tokenizer = tf.keras.preprocessing.text.Tokenizer(num_words=vocab_size)\n","tokenizer.fit_on_texts(X_train)\n","\n","X_train = tokenizer.texts_to_sequences(X_train)\n","X_test = tokenizer.texts_to_sequences(X_test)\n","\n","vocab_size = len(tokenizer.word_index) + 1\n","\n","maxlen = 200\n","\n","X_train = tf.keras.preprocessing.sequence.pad_sequences(X_train, padding='post', maxlen=maxlen)\n","X_test = tf.keras.preprocessing.sequence.pad_sequences(X_test, padding='post', maxlen=maxlen)"],"metadata":{"id":"HQmu9VibGIH1"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# We will be using GloVe word embeddings to convert text inputs to their numeric counterparts."],"metadata":{"id":"YhCJw0OnGIJ6"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# %%time\n","# import gensim.downloader as api\n","\n","# def embed_word_vector(word_index, pretrained='word2vec-google-news-300'):\n","#     embed_size = 300 # Google news vector is 300-dimensional\n","#     vector = api.load(pretrained)\n","#     zeros = [0] * embed_size\n","#     matrix = np.zeros((vocab_size, embed_size)) \n","    \n","#     for word, i in word_index.items():\n","#         if i >= vocab_size or word not in vector: continue \n","#         matrix[i] = vector[word]\n","    \n","#     print(\"Embed word vector completed.\")\n","#     return matrix"],"metadata":{"id":"7d_YrArYHbRe"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# %%time\n","# pretrained = 'glove-wiki-gigaword-300'\n","# matrix = embed_word_vector(tokenizer.word_index, pretrained=pretrained)"],"metadata":{"id":"Q-ZRRwiYHifc"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["from numpy import array\n","from numpy import asarray\n","from numpy import zeros\n","\n","embeddings_dictionary = dict()\n","\n","glove_file = open('/content/drive/MyDrive/datasets/CB_dataset/glove.6B.100d.txt', encoding=\"utf8\")\n","\n","for line in glove_file:\n","    records = line.split()\n","    word = records[0]\n","    vector_dimensions = asarray(records[1:], dtype='float32')\n","    embeddings_dictionary[word] = vector_dimensions\n","glove_file.close()\n","\n","embedding_matrix = zeros((vocab_size, 100))\n","for word, index in tokenizer.word_index.items():\n","    embedding_vector = embeddings_dictionary.get(word)\n","    if embedding_vector is not None:\n","        embedding_matrix[index] = embedding_vector"],"metadata":{"id":"jTcIImqwGIMA"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# The following script creates the model. Our model will have one input layer, one embedding layer, one LSTM layer with\n","#  128 neurons and one output layer with 6 neurons since we have 6 labels in the output."],"metadata":{"id":"6Hh-gS2fGIOI"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["deep_inputs = tf.keras.Input(shape=(maxlen, ))\n","embedding_layer = tf.keras.layers.Embedding(vocab_size, 100, weights=[embedding_matrix], trainable=False)(deep_inputs)\n","LSTM_Layer_1 = tf.keras.layers.Bidirectional(tf.keras.layers.GRU(128, return_sequences=True))(embedding_layer)\n","net = tf.keras.layers.GlobalMaxPool1D()(LSTM_Layer_1)\n","net = tf.keras.layers.Dense(124, activation='relu')(net)\n","net = tf.keras.layers.Dropout(0.4)(net)\n","output = tf.keras.layers.Dense(6, activation='sigmoid')(net)\n","model = tf.keras.models.Model(inputs=deep_inputs, outputs=output)\n","model.summary()\n","\n","# model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['acc'])\n","model.compile(loss='categorical_crossentropy', optimizer='adam', metrics=['acc'])"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"XVP1XTy8GIRv","executionInfo":{"status":"ok","timestamp":1672430911590,"user_tz":-540,"elapsed":1847,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"e1386af8-a0da-4720-f748-2f35256ad4fc"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Model: \"model_7\"\n","_________________________________________________________________\n"," Layer (type)                Output Shape              Param #   \n","=================================================================\n"," input_8 (InputLayer)        [(None, 200)]             0         \n","                                                                 \n"," embedding_7 (Embedding)     (None, 200, 100)          14824300  \n","                                                                 \n"," bidirectional_7 (Bidirectio  (None, 200, 256)         176640    \n"," nal)                                                            \n","                                                                 \n"," global_max_pooling1d_7 (Glo  (None, 256)              0         \n"," balMaxPooling1D)                                                \n","                                                                 \n"," dense_14 (Dense)            (None, 124)               31868     \n","                                                                 \n"," dropout_7 (Dropout)         (None, 124)               0         \n","                                                                 \n"," dense_15 (Dense)            (None, 6)                 750       \n","                                                                 \n","=================================================================\n","Total params: 15,033,558\n","Trainable params: 209,258\n","Non-trainable params: 14,824,300\n","_________________________________________________________________\n"]}]},{"cell_type":"code","source":["# The following script prints the architecture of our neural network:\n","\n","from keras.utils import plot_model\n","plot_model(model, to_file='model_plot4a.png', show_shapes=True, show_layer_names=True)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":754},"id":"lHdMW30PJhUX","executionInfo":{"status":"ok","timestamp":1672430913613,"user_tz":-540,"elapsed":1231,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"96d0fc5e-57c1-40f8-a894-f6e3cc64daae"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAdMAAALhCAYAAAAEpE6WAAAABmJLR0QA/wD/AP+gvaeTAAAgAElEQVR4nOzdeVQUd7YH8G9DN71gN7tAMCiLGy7xuWQEZYwh40SNC4KK0WQ0yTzixCAuSDSKBpFocJSnkclzCTNPHBWEB0YlyRgHHU6MJ4kyICYGcUWjgLKvDdz3h6/72LJ10910g/dzDn/4q1tVt+pXVddaW0BEBMYYY4x1VYqFqTNgjDHGejoupowxxpieuJgyxhhjeuJiyhhjjOlJaOoEeprz589jx44dpk6DMcaMJiUlxdQp9Dh8ZqqjO3fu4NixY6ZOg2nhu+++w3fffWfqNHqUoqIi3r6fYdz/Xcdnpl3E/3Mzf3PnzgXAfaWL5ORkzJ8/n9fZM0rV/0x3fGbKGGOM6YmLKWOMMaYnLqaMMcaYnriYMsYYY3riYsoYY4zpiYupGTp16hRsbGzwxRdfmDoVvf3973/HuHHjIJfL0b9/fyxZsgT37983dVo66U39YQzvvvsuBAKB+m/RokWtYk6fPo21a9ciNTUVnp6e6tg33nijVeyUKVMgl8thaWmJYcOG4eLFi92xGF0WHR0NHx8fKBQKiMVieHt7Y82aNaiurm4Vm52djQkTJkAmk8HV1RWRkZFoaGjQOe748ePYtm0bmpubNcZLT0/X6AtHR0fDLzBrExdTM9Rbfsjn6NGjWLhwIebOnYuioiJkZGTg3LlzmDp1KpqamkydntZ6S38Yk729PTIzM3H16lUcOHBAY9jGjRuxa9curFu3DkFBQbh+/Tq8vLzg4OCApKQknDx5UiP+66+/RkpKCmbMmIH8/HyMHj26OxdFZ2fOnMGyZctw8+ZNlJaWIjY2FvHx8epXs1Ty8/MxZcoUBAQEoKSkBGlpafj888+xdOlSneNmzpwJiUSCgIAAlJeXq9tnzZqFoqIinDt3DtOmTTPugjNNxHRy9OhRepZWW21tLfn6+nZp3MmTJ9Nzzz1HLS0t6rZPP/2UAFB2drahUmxXcHAwBQcHG30+3Umf/tBGV7bv0NBQcnNza3PYxx9/TIMGDaK6ujqNdi8vLzp06BBZWFiQm5sblZeXawzPzMykWbNm6Za8iUyfPp2ampo02ubNm0cA6Pbt2+q2+fPnk4eHh8b+EBcXRwKBgH766Sed44iIwsLCyNfXl5RKZau8li9fTg4ODjoty7N2fDOgZD4zZR06cOAAiouLuzTunTt34OrqCoFAoG57/vnnAQC3bt0ySH7PGn36o7tdu3YNGzZswEcffQSJRNJquJ+fH8LDw3H37l2sXr3aBBkaxokTJ2BpaanRprq8WltbCwBoamrCyZMnMWnSJI39YerUqSAiZGRk6BSnsmnTJuTk5CA+Pt4oy8a0x8XUzGRnZ8Pd3R0CgQCffvopACAhIQHW1taQyWTIyMjA1KlToVAo0K9fPxw+fFg97q5duyCRSNC3b1+8++67cHV1hUQigZ+fHy5cuKCOCwsLg5WVFVxcXNRt7733HqytrSEQCFBaWgoACA8Px6pVq1BYWAiBQABvb2+dlsXT07PVgV91v9TT01O3FWMiPaE/vvzySygUCmzZsqU7VonWdu3aBSLCzJkz242JiYnBoEGDsH//fpw+fbrD6RERduzYgaFDh0IsFsPOzg6zZ8/Gzz//rI7Rtm8AoLm5GVFRUXB3d4dUKsXIkSNx9OhR/Rb6/929exdSqRQeHh4AgOvXr6O6uhru7u4acV5eXgCA3NxcneJU7OzsMGnSJMTHx/PtCFMz6YlxD9Qdl0Hu3LlDAGj37t3qtg8//JAA0DfffEMVFRVUXFxM/v7+ZG1tTY2Njeq40NBQsra2pitXrlB9fT3l5+fTuHHjSC6Xa1xyWrhwITk7O2vMNy4ujgBQSUmJui0oKIi8vLy6tBxZWVkkEolo165dVFlZSZcvX6ahQ4fS73//+y5NT1eGusxr7v1x4sQJksvlFB0drfeyGvIyr6enJ/n4+LQ5jpeXF924cYOIiL799luysLCgAQMGUHV1NRG1fZk3KiqKrKys6ODBg1ReXk65ubk0evRocnR0pPv376vjtO2b1atXk1gspmPHjlFZWRmtW7eOLCws6Pvvv9dp+Z9WU1NDcrmcwsLC1G1nz54lABQXF9cqXiqVUkBAgE5xT1q7di0BoEuXLmm082XebsWXeXsaPz8/KBQKODk5ISQkBDU1Nbh9+7ZGjFAoVP/v3cfHBwkJCaiqqkJiYmK35jpp0iRERkYiLCwMCoUCw4cPR1VVFfbv39+teRiTOfTH9OnTUVlZiQ0bNhhkeoZQU1ODGzduqM+oOuLr64sVK1bg5s2b+OCDD9qMqaurw44dOzBnzhwsWrQINjY2GDFiBD777DOUlpZi7969rcbpqG/q6+uRkJCAwMBABAUFwdbWFuvXr4dIJNK7X2JjY+Hq6oqYmBh1m+pJ3KcvBwOASCRCXV2dTnFPGjhwIAAgLy9Pr7yZfriY9mBWVlYAAKVS2WHc2LFjIZPJNC6HdYcPP/wQe/fuxTfffIPq6mpcv34dfn5+8PX1xZ07d7o1l+5g7v3RnYqLi0FEkMlkWsXHxMRg8ODB2LNnD7Kzs1sNz8/PR3V1NcaOHavRPm7cOFhZWWlcNm/L031z9epV1NbWYvjw4eoYqVQKFxcXvfolLS0NycnJ+OqrryCXy9XtqnvGbT3F3tjYCKlUqlPck1Tr+MGDB13Om+mPi+kzQiwWo6SkpNvm9+uvv2Lbtm34z//8T7z88suwtraGh4cH9u3bh3v37iEuLq7bcjFH3d0f3a2+vh7A4+XUhkQiQWJiIgQCAd56661WZ2Cq1z/69OnTalxbW1tUVVXplF9NTQ0AYP369RrvZd66dUv90JCujhw5gq1btyIrKwsDBgzQGKa6H15ZWanRXltbi/r6eri6uuoU9yRVgVWtc2YaXEyfAUqlEuXl5ejXr1+3zbOgoADNzc147rnnNNoVCgXs7e2Rn5/fbbmYG1P0R3dTHeCf/qhAR3x9fbFy5UoUFBRg8+bNGsNsbW0BoM2i2ZV16eTkBADYuXMniEjj7/z58zpNCwB2796NpKQknDlzptU2DwAeHh6Qy+WtnmK/du0aAGDkyJE6xT2psbERANo8a2Xdh4vpMyArKwtEhPHjx6vbhEJhp5cj9aE6uP36668a7VVVVXj06JH6FZlnkSn6o7v17dsXAoEAFRUVOo23efNmDBkyBJcuXdJoHz58OPr06YMffvhBo/3ChQtobGzEmDFjdJrP888/D4lEgpycHJ3GexoRITIyEnl5eUhPT2/zzBl43L/Tpk3DuXPn0NLSom7PzMyEQCBQP/GsbdyTVOvY2dlZr2Vh+uFi2gu1tLSgrKwMTU1NyM3NRXh4ONzd3bF48WJ1jLe3Nx49eoT09HQolUqUlJS0+e6nvb097t27h5s3b6KqqkrrA76HhwcmT56Mffv24dy5c6irq8OdO3cQGhoKAHj77bcNsqw9gbH7IzMz0+xejZHJZPD09ERRUZFO46ku9z79AI5EIsGqVauQlpaGpKQkVFZWIi8vD0uXLoWrq6t6u9JlPkuWLMHhw4eRkJCAyspKNDc3o6ioSP0fwJCQEDg7O3f4OcMrV67gk08+wb59+yASiTQuGQsEAmzfvl0du2HDBjx48AAbN25ETU0Nzp8/j7i4OCxevBiDBw/WOU5FtY5HjBih0zpgBma6J4l7JmM/Or57925ycXEhACSTyWjmzJm0Z88ekslkBIAGDhxIhYWFtHfvXlIoFASA+vfvT7/88gsRPX5NQSQSkZubGwmFQlIoFDR79mwqLCzUmM/Dhw9p8uTJJJFIyMPDg95//32KiIggAOTt7a1+bePixYvUv39/kkqlNHHiRI1XEDpTWlpK4eHh5O3tTWKxmPr06UMTJkyg//3f/zXcCuuAIV6N6Qn9cerUKZLL5RQTE6PXshIZ9tWYsLAwEolEVFtbq25LS0sjLy8vAkCOjo60bNmyNqcZERHR6tWYlpYWiouLo4EDB5JIJCI7OzsKDAykq1evqmN06ZuGhgaKjIwkd3d3EgqF5OTkREFBQZSfn09ERIGBgQSAoqKi2l32vLw8AtDu39OvuJw9e5ZefPFFEovF5OrqShEREVRfX99qutrGET3+ApObm5vGF5OI+NWYbpbMa01H5r6xhYaGkr29vanTMAvm8DnBntYfhiymBQUFJBQK6eDBg4ZKr1s1NzeTv78/HThwwNSptKu0tJQkEglt37691TAupt2K3zPtjXR56IMZ37PQH3V1dfjqq69QUFCgfiDG29sb0dHRiI6ObvMXVMxZc3Mz0tPTUVVVhZCQEFOn065NmzZh1KhRCAsLA/D4Hu69e/eQnZ2tfmiJdQ8upkxrP//8c6t7Qm39mfPBhxnHo0eP8Oqrr2LQoEF466231O1r167F3LlzERISovPDSKaUlZWF1NRUZGZmav2ubHfbsWMHcnJycOrUKYhEIgBARkYG3Nzc4O/v3+rXeJhxcTHtRdatW4fExERUVFTAw8MDx44dM+j0hwwZ0uo1grb+jhw5YtD59lTG7g9z8dlnn2n0f1JSksbwLVu2ICwsDB9//LGJMtRdQEAADh06pPG9ZHOSkZGBhoYGZGVlwc7OTt0+e/Zsjb5QfdeZGZ+AiL+OrIvk5GTMnz+fPyrdA6h+TzIlJcXEmfQcvH0/27j/uyyFz0wZY4wxPXExZYwxxvTExZQxxhjTExdTxhhjTE9cTBljjDE9CU2dQE8lEAhMnQLTEveV7nidMaYbLqZddPToUVOnwDqxc+dOAMCKFStMnEnPcf78ecTHx/P2/YxS9T/THRfTLpo3b56pU2CdUL1fyn2lm/j4eF5nzzAupl3D90wZY4wxPXExZYwxxvTExZQxxhjTExdTxhhjTE9cTBljjDE9cTE1su+++w5Dhw6FhYUFBAIBnJ2dERMTY+q0NKSmpsLT01P9e6QuLi5YtGiRqdNiPcS7776r8Xu2bW07p0+fxtq1a1tta2+88Uar2ClTpkAul8PS0hLDhg3DxYsXu2Mxuiw6Oho+Pj5QKBQQi8Xw9vbGmjVr2vxB9OzsbEyYMAEymQyurq6IjIxEQ0ODznHHjx/Htm3bWv3wfHp6ukZfODo6Gn6BWduI6eTo0aPUldX2+9//ngBQWVmZEbIyDC8vL7KxsTF1GgYTHBxMwcHBpk6jR+nK9h0aGkr29vaUmZlJV69epfr6eo3hUVFRNGPGDKqsrFS3eXl5kYODAwGgEydOtJpmZmYmzZo1q2sL0c0mTZpEe/bsoYcPH1JlZSUdPXqURCIRvfrqqxpxly9fJqlUShs2bKDq6mr69ttvydHRkZYsWdKluPj4eJo0aZLGMaWlpYWKioro3LlzNG3aNHJwcNBpWbp6fGOUzGemz6C6ujr4+fmZOo1nQnesa3PoT6lUildffRWDBg2CWCxWt2/duhVHjhxBcnIy5HK5xji7du2ChYUFQkNDUVFR0d0pG0yfPn0QGhoKe3t7yOVyzJs3D4GBgfjyyy9x584dddzmzZvh4uKCjz76CNbW1vD19UVkZCT++te/4ueff9Y5bvny5XjhhRcwbdo0NDU1AXj85So3Nzf4+/tj4MCB3bcSGF/mfRYdOHAAxcXFpk7jmdAd69pc+/PatWvYsGEDPvroI0gkklbD/fz8EB4ejrt372L16tUmyNAwTpw4AUtLS4021eXV2tpaAEBTUxNOnjyJSZMmaXyqcerUqSAiZGRk6BSnsmnTJuTk5PCHFswAF1MTSUhIgLW1NWQyGTIyMjB16lQoFAr069cPhw8fVsft2rULEokEffv2xbvvvgtXV1dIJBL4+fnhwoUL6riwsDBYWVnBxcVF3fbee+/B2toaAoEApaWlAIDw8HCsWrUKhYWFEAgE8Pb27lL+//rXv+Dj4wMbGxtIJBKMGDECX331FQDgnXfeUd+z8fLywqVLlwAAS5YsgUwmg42NDY4fPw4AaG5uRlRUFNzd3SGVSjFy5Ej1p+w++eQTyGQyyOVyFBcXY9WqVXBzc8PVq1e7lLM2iAg7duzA0KFDIRaLYWdnh9mzZ2ucEeizrrurP7/88ksoFAps2bLFaOuqM7t27QIRYebMme3GxMTEYNCgQdi/fz9Onz7d4fS06Rtt9yug421PX3fv3oVUKoWHhwcA4Pr166iuroa7u7tGnJeXFwAgNzdXpzgVOzs7TJo0CfHx8SAig+TOusiUF5l7IkPeM/3www8JAH3zzTdUUVFBxcXF5O/vT9bW1tTY2KiOCw0NJWtra7py5QrV19dTfn4+jRs3juRyOd2+fVsdt3DhQnJ2dtaYb1xcHAGgkpISdVtQUBB5eXm1ylGXe6YpKSm0adMmevToET18+JDGjx+vcX8mKCiILC0t6e7duxrjvf7663T8+HH1v1evXk1isZiOHTtGZWVltG7dOrKwsKDvv/9eYx0tX76cdu/eTXPmzKGffvpJqxy7cs80KiqKrKys6ODBg1ReXk65ubk0evRocnR0pPv376vj9FnX3dGfJ06cILlcTtHR0Totf1fvmbq5ubVq9/T0JB8fnzbH8fLyohs3bhAR0bfffksWFhY0YMAAqq6uJqK275lq2zfa7ledbXtdVVNTQ3K5nMLCwtRtZ8+eJQAUFxfXKl4qlVJAQIBOcU9au3YtAaBLly5ptC9fvpzvmXYfvmdqDvz8/KBQKODk5ISQkBDU1NTg9u3bGjFCoVD9P3IfHx8kJCSgqqoKiYmJJsk5ODgYGzduhJ2dHezt7TFz5kw8fPgQJSUlAIClS5eiublZI7/Kykp8//33mDZtGgCgvr4eCQkJCAwMRFBQEGxtbbF+/XqIRKJWy7V161YsW7YMqampGDJkiFGWqa6uDjt27MCcOXOwaNEi2NjYYMSIEfjss89QWlqKvXv3Gmxexu7P6dOno7KyEhs2bDDI9HRVU1ODGzduqM+oOuLr64sVK1bg5s2b+OCDD9qM6UrfdLRf6bLt6So2Nhaurq4aT+2rnsR9+nIwAIhEItTV1ekU9yTVvdG8vDy98mb64WJqZqysrAAASqWyw7ixY8dCJpNpXOIyJZFIBADqR/VffvllDBo0CJ9//rn68tORI0cQEhKiPlBcvXoVtbW1GD58uHo6UqkULi4uJlmu/Px8VFdXY+zYsRrt48aNg5WVlcZlWEMzt/7UV3FxMYgIMplMq/iYmBgMHjwYe/bsQXZ2dqvh+vbN0/uVsba9tLQ0JCcn46uvvtJ44Ep1z1j1oNCTGhsbIZVKdYp7kmodP3jwoMt5M/1xMe3BxGKx+kywu508eRIvvfQSnJycIBaLsWbNGo3hAoEA7777Lq5fv45vvvkGAPA///M/ePvtt9UxNTU1AID169drvBt369Yt9YMb3am8vBzA46czn2Zra4uqqiqjzt+U/Wlo9fX1AKDxZG9HJBIJEhMTIRAI8NZbb7U6AzN03xhj2zty5Ai2bt2KrKwsDBgwQGOY6t53ZWWlRnttbS3q6+vh6uqqU9yTVAVWtc6ZaXAx7aGUSiXKy8vRr1+/bpnfuXPn1L8Pevv2bQQGBsLFxQUXLlxARUUFtm3b1mqcxYsXQyKRYP/+/bh69SoUCgX69++vHu7k5ATg8e+OEpHG3/nz57tluZ5ka2sLAG0emI29rru7P41NdYB/+qMCHfH19cXKlStRUFCAzZs3awwzdN8YetvbvXs3kpKScObMGTz33HOthnt4eEAul+PWrVsa7deuXQMAjBw5Uqe4JzU2NgJAm2etrPvw75n2UFlZWSAijB8/Xt0mFAo7vTzcVT/++COsra0BPL43o1Qq8ac//Qmenp4AoPEYv4qdnR3mz5+PI0eOQC6X449//KPG8Oeffx4SiQQ5OTlGyVlXw4cPR58+ffDDDz9otF+4cAGNjY0YM2aMus3Q67q7+9PY+vbtC4FAoPP7o5s3b8aJEydw6dIljSdadekbbRhq2yMifPDBBygrK0N6ejqEwrYPqUKhENOmTcO5c+fQ0tICC4vH5zGZmZkQCATqJ561jXuSah07OzvrtSxMP3xm2kO0tLSgrKwMTU1NyM3NRXh4ONzd3bF48WJ1jLe3Nx49eoT09HQolUqUlJS0+h8uANjb2+PevXu4efMmqqqqOjxgK5VKPHjwAFlZWepiqjrInT59GvX19SgoKGj3ntXSpUvR0NCAEydOYMaMGRrDJBIJlixZgsOHDyMhIQGVlZVobm5GUVERfv31V11Xkd4kEglWrVqFtLQ0JCUlobKyEnl5eVi6dClcXV0RGhqqjtV3XRu7PzMzM036aoxMJoOnpyeKiop0Gk91uffpB3B06Rtt59PZthcSEgJnZ+cOP2d45coVfPLJJ9i3bx9EIpHGJWOBQIDt27erYzds2IAHDx5g48aNqKmpwfnz5xEXF4fFixdj8ODBOsepqNbxiBEjdFoHzMBM8hBxD6bro+PfffcdDRs2jCwsLAgAubi40JYtW2jPnj0kk8kIAA0cOJAKCwtp7969pFAoCAD179+ffvnlFyJ6/OqBSCQiNzc3EgqFpFAoaPbs2VRYWKgxr4cPH9LkyZNJIpGQh4cHvf/++xQREUEAyNvbW/3axcWLF6l///4klUpp4sSJ9Je//IW8vLwIQId/aWlp6nlFRkaSvb092dra0ty5c+nTTz8lAOTl5aXxegcR0X/8x3/Q2rVr21w/DQ0NFBkZSe7u7iQUCsnJyYmCgoIoPz+ftm3bRlKplADQ888/TwcPHtR6vRN17dWYlpYWiouLo4EDB5JIJCI7OzsKDAykq1evasR1dV3fv3/f6P15//59OnXqFMnlcoqJidFp+Q35akxYWBiJRCKqra1Vt6Wlpam3NUdHR1q2bFmb04yIiGj1aow2faPLftXRtkdEFBgYSAAoKiqq3WXPy8vrcJ95+hWXs2fP0osvvkhisZhcXV0pIiKi1ecXdYkjIpo+fTq5ublRS0uLRju/GtOtknmt6cgUG5vq26c91bRp0+j69evdPl9z/TavOfenIYtpQUEBCYVCnf8TZC6am5vJ39+fDhw4YOpU2lVaWkoSiYS2b9/eahgX027F75n2FLo8yGFqT142zs3NhUQiUX8Jhj3Wk/pTG3V1dfjqq69QUFCgfiDG29sb0dHRiI6ObvMXVMxZc3Mz0tPTUVVVhZCQEFOn065NmzZh1KhRCAsLA/D4Hu69e/eQnZ2tfmiJdQ8upszgIiMjUVBQgF9++QVLlixp9WQm630ePXqk/tD9W2+9pW5fu3Yt5s6di5CQkB71MfusrCykpqYiMzNT63dlu9uOHTuQk5ODU6dOqd/zzsjIUH/o/uTJkybO8NnCxdTMrVu3DomJiaioqICHhweOHTtm6pQ6JZPJMGTIELzyyivYtGkTfHx8TJ2S2eiJ/dmZzz77TOPVkqSkJI3hW7ZsQVhYGD7++GMTZai7gIAAHDp0SOPbyOYkIyMDDQ0NyMrKgp2dnbp99uzZGn2h+oYzMz4BEX8dWRfJycmYP38+f1S6B5g7dy4AICUlxcSZ9By8fT/buP+7LIXPTBljjDE9cTFljDHG9MTFlDHGGNMTF1PGGGNMT/xt3i5KTk42dQqsE6rPrHFfaU/1kXdeZ88mU/zARG/BT/PqSPW0G2OM9VZcFnSWwsWUMTPGryow1iPwqzGMMcaYvriYMsYYY3riYsoYY4zpiYspY4wxpicupowxxpieuJgyxhhjeuJiyhhjjOmJiyljjDGmJy6mjDHGmJ64mDLGGGN64mLKGGOM6YmLKWOMMaYnLqaMMcaYnriYMsYYY3riYsoYY4zpiYspY4wxpicupowxxpieuJgyxhhjeuJiyhhjjOmJiyljjDGmJy6mjDHGmJ64mDLGGGN64mLKGGOM6YmLKWOMMaYnLqaMMcaYnriYMsYYY3riYsoYY4zpiYspY4wxpicupowxxpieuJgyxhhjeuJiyhhjjOmJiyljjDGmJy6mjDHGmJ6Epk6AMfZYUVER/vCHP6C5uVndVlZWBrlcjpdeekkjdvDgwfjv//7vbs6QMdYeLqaMmYl+/frh1q1bKCwsbDXs7NmzGv/+7W9/211pMca0wJd5GTMjb775JkQiUadxISEh3ZANY0xbXEwZMyMLFy5EU1NThzHDhg2Dj49PN2XEGNMGF1PGzIiXlxdGjhwJgUDQ5nCRSIQ//OEP3ZwVY6wzXEwZMzNvvvkmLC0t2xzW1NSEuXPndnNGjLHOcDFlzMwsWLAALS0trdotLCwwfvx4DBgwoPuTYox1iIspY2bG1dUVEyZMgIWF5u5pYWGBN99800RZMcY6wsWUMTP0xhtvtGojIsyZM8cE2TDGOsPFlDEzFBwcrHHf1NLSEq+88gr69u1rwqwYY+3hYsqYGbKzs8Pvfvc7dUElIixatMjEWTHG2sPFlDEztWjRIvWDSCKRCLNnzzZxRoyx9nAxZcxMzZw5E2KxGAAwY8YM9OnTx8QZMcbaw8WUMTNlbW2tPhvlS7yMmTcBEZGpkzCk9r4cwxhjzDwEBwcjJSXF1GkYUkqv/NWY8PBw+Pr6mjoNZgLz58/vVf3f3NyMo0eP4vXXXzfaPHbu3AkAWLFihdHmwZiKanvrbXplMfX19cW8efNMnQYzgfnz5/e6/g8MDIREIjHa9FVnCL1pnTHz1cvOSNX4niljZs6YhZQxZhhcTBljjDE9cTFljDHG9MTFlDHGGNMTF1PGGGNMT1xMjWTcuHGwtLTEqFGjDD7td955B3K5HAKBADk5OTrHnTp1CjY2Nvjiiy8MnpuuXnrpJQgEgjb/TPnFH3NaR8+a06dPY+3atUhNTYWnp6d6e2jrl3SmTJkCuVwOS0tLDBs2DBcvXjRBxtqLjo6Gj48PFAoFxGIxvL29sWbNGlRXV7eKzc7OxoQJEyCTyeDq6orIyEg0NDToHHf8+HFs27YNzc3NRl22Zx0XUyP5/vvvMXnyZKNMe//+/di3b1+X43rKdzomTpxosvSFs/wAACAASURBVHn3lHXU22zcuBG7du3CunXrEBQUhOvXr8PLywsODg5ISkrCyZMnNeK//vprpKSkYMaMGcjPz8fo0aNNlLl2zpw5g2XLluHmzZsoLS1FbGws4uPjMXfuXI24/Px8TJkyBQEBASgpKUFaWho+//xzLF26VOe4mTNnQiKRICAgAOXl5d2ynM8iLqZGZo5fZJo+fToqKiowY8YMU6cCiUSCyspKEJHGX2hoKNasWWOyvMxpHdXV1cHPz8/UaRjd1q1bceTIESQnJ0Mul2sM27VrFywsLBAaGoqKigoTZai/Pn36IDQ0FPb29pDL5Zg3bx4CAwPx5Zdf4s6dO+q4zZs3w8XFBR999BGsra3h6+uLyMhI/PWvf8XPP/+sc9zy5cvxwgsvYNq0aWhqaurWZX5WcDE1MpFIZJTpaluku6OYExFSUlKwd+9encf98ssvWx0479y5g8uXL+Pll182VIo92oEDB1BcXGzqNIzq2rVr2LBhAz766KM236v18/NDeHg47t69i9WrV5sgQ8M4ceKExu/UAoCjoyMAoLa2FgDQ1NSEkydPYtKkSRr779SpU0FEyMjI0ClOZdOmTcjJyUF8fLxRlu1Z98wX0+bmZkRFRcHd3R1SqRQjR47E0aNHAQDx8fGwtraGhYUFxowZA2dnZ4hEIlhbW2P06NHw9/fH888/D4lEAltb2zbPpK5du4YhQ4bA2toaUqkU/v7+yM7O1joH4HGxiouLw+DBgyEWi2FjY4OIiIhW89ImLjs7G+7u7hAIBPj0008BAAkJCbC2toZMJkNGRgamTp0KhUKBfv364fDhw61yjY2NxeDBgyGVSuHo6AgPDw/ExsYa7As6W7duxfLlyw0yra7QZx3t2rULEokEffv2xbvvvgtXV1dIJBL4+fnhwoUL6riwsDBYWVnBxcVF3fbee+/B2toaAoEApaWlAB5/GnPVqlUoLCyEQCCAt7c3gMf/CVEoFNiyZUt3rBKj27VrF4gIM2fObDcmJiYGgwYNwv79+3H69OkOp0dE2LFjB4YOHQqxWAw7OzvMnj1b42xN1+2+o31UH3fv3oVUKoWHhwcA4Pr166iuroa7u7tGnJeXFwAgNzdXpzgVOzs7TJo0CfHx8XwbwxiolwFAR48e1Tp+9erVJBaL6dixY1RWVkbr1q0jCwsL+v7774mIaOPGjQSALly4QDU1NVRaWkqvvvoqAaCTJ09SSUkJ1dTUUFhYGAGgnJwc9bQDAgLI09OTbty4QUqlki5fvky/+c1vSCKR0C+//KJ1Dh9++CEJBAL685//TGVlZVRbW0t79uwhAHTp0iX1dLSNu3PnDgGg3bt3a4wLgL755huqqKig4uJi8vf3J2tra2psbFTHbdmyhSwtLSkjI4Nqa2vpxx9/JGdnZ3rppZd06KX2FRUVkY+PDzU3N3dpfF37vz36rKPQ0FCytramK1euUH19PeXn59O4ceNILpfT7du31XELFy4kZ2dnjfnGxcURACopKVG3BQUFkZeXl0bciRMnSC6XU3R0tN7LGhwcTMHBwXpPRx+enp7k4+PT5jAvLy+6ceMGERF9++23ZGFhQQMGDKDq6moiIsrMzKRZs2ZpjBMVFUVWVlZ08OBBKi8vp9zcXBo9ejQ5OjrS/fv31XHa9mln+2hX1dTUkFwup7CwMHXb2bNnCQDFxcW1ipdKpRQQEKBT3JPWrl3b6njQ3cxhezOC5Gf6zLS+vh4JCQkIDAxEUFAQbG1tsX79eohEIiQmJmrE+vj4QCaTwcHBAQsWLAAAuLu7w9HRETKZTP0TWU/+zxcA5HI5BgwYAKFQiGHDhmHfvn2or69XXxLtLIe6ujrs3LkTr7zyClauXAlbW1tIpVLY29trzEfbuM74+flBoVDAyckJISEhqKmpwe3bt9XD09PTMWbMGMycORNSqRSjR4/GrFmzcO7cOTQ2Nuo0r7Zs3boV77//PiwszHfT7GwdAYBQKFSfFfn4+CAhIQFVVVWttquumj59OiorK7FhwwaDTM+UampqcOPGDfUZVUd8fX2xYsUK3Lx5Ex988EGbMXV1ddixYwfmzJmDRYsWwcbGBiNGjMBnn32G0tLSNm9HdNSnuhwndBUbGwtXV1fExMSo21RP4j59ORh4fNuorq5Op7gnDRw4EACQl5enV96sNfM9YnWDq1evora2FsOHD1e3SaVSuLi4tCqKT7KysgIAjRv5qnujSqWyw3mOGDECNjY26kswneVw7do11NbWIiAgoMPpahunC9VyPrlM9fX1rS4RNTc3QyQStblT6+LevXs4fvw4Fi9erNd0ulNb66gtY8eOhUwm63C7elYVFxeDiCCTybSKj4mJweDBg7Fnz55Wt0yAx0+4VldXY+zYsRrt48aNg5WVlcbl9rY83addPU50Ji0tDcnJyfjqq680nhtQ3TNu60GhxsZGSKVSneKepFrHDx486HLerG3PdDGtqakBAKxfv17j/cZbt26pHwYwBpFIpN5RO8uhqKgIAODk5NThNLWN09e0adPw448/IiMjA3V1dfjhhx+Qnp6O1157Te9ium3bNvzxj3/stR92F4vFKCkpMXUaZqe+vh7A4/WjDYlEgsTERAgEArz11lutzsBUr3+09Z6yra0tqqqqdMrPGMeJI0eOYOvWrcjKysKAAQM0hqnuo1dWVmq019bWor6+Hq6urjrFPUlVYFXrnBnOM11MVYVn586drV7NOH/+vFHm2dTUhEePHqkfGugsB1Vhaetl7SdpG6evTZs24eWXX8bixYuhUCgwZ84czJs3T6v3Xjty//59/P3vf8ef/vQnA2VqXpRKJcrLy9GvXz9Tp2J2VAd4XT4q4Ovri5UrV6KgoACbN2/WGGZrawsAbRbNrvSBoY8Tu3fvRlJSEs6cOYPnnnuu1XAPDw/I5XLcunVLo/3atWsAgJEjR+oU9yTVrZi2zlqZfp7pYqp6ErejrwgZ2j//+U+0tLSoXy7vLIfhw4fDwsICZ8+e7XC62sbpKz8/H4WFhSgpKYFSqcTt27eRkJAAOzs7vaa7bds2LFq0SOd7vD1FVlYWiAjjx49XtwmFwk4vDz8L+vbtC4FAoPP7o5s3b8aQIUNw6dIljfbhw4ejT58++OGHHzTaL1y4gMbGRowZM0an+RjqOEFEiIyMRF5eHtLT09v9wpdQKMS0adNw7tw5tLS0qNszMzMhEAjUTzxrG/ck1Tp2dnbWa1lYa890MZVIJFiyZAkOHz6MhIQEVFZWorm5GUVFRfj1118NMo/GxkZUVFSgqakJFy9eRFhYGPr376++L9hZDk5OTggKCsKxY8dw4MABVFZWIjc3t9VDFNrG6WvZsmVwd3dv8/NnXfXgwQN8/vnnWLFihcGmaWotLS0oKytDU1MTcnNzER4eDnd3d437wd7e3nj06BHS09OhVCpRUlLS6iwDAOzt7XHv3j3cvHkTVVVVUCqVyMzM7DWvxshkMnh6eqpvVWhLdbn36dsLEokEq1atQlpaGpKSklBZWYm8vDwsXboUrq6uCA0N1Xk+nR0nQkJC4Ozs3OHnDK9cuYJPPvkE+/btg0gkavX5zO3bt6tjN2zYgAcPHmDjxo2oqanB+fPnERcXh8WLF2Pw4ME6x6mo1vGIESN0WgdMC93/BLFxQcdXIxoaGigyMpLc3d1JKBSSk5MTBQUFUX5+PsXHx5NMJiMANGDAAPrXv/5FW7duJRsbGwJAzs7OdOjQITpy5Ag5OzsTALKzs6PDhw8TEVFiYiJNnjyZ+vbtS0KhkBwcHGjBggV069YtrXMgIqqqqqJ33nmHHBwcqE+fPjRx4kSKiooiANSvXz/697//rXXc7t27ycXFhQCQTCajmTNn0p49e9TLOXDgQCosLKS9e/eSQqEgANS/f3/1qzxnzpwhBwcHAqD+E4lENHToUEpNTe1Sn61cuZIWLVrUpXGfpmv/t0XfdRQaGkoikYjc3NxIKBSSQqGg2bNnU2FhocZ8Hj58SJMnTyaJREIeHh70/vvvU0REBAEgb29v9Ws0Fy9epP79+5NUKqWJEyfS/fv36dSpUySXyykmJkavZSUyj1cVwsLCSCQSUW1trbotLS2NvLy8CAA5OjrSsmXL2hw3IiKi1asxLS0tFBcXRwMHDiSRSER2dnYUGBhIV69eVcfo0qed7aOBgYEEgKKiotpdxry8PI395um/p19xOXv2LL344oskFovJ1dWVIiIiqL6+vtV0tY0jIpo+fTq5ublRS0tLu3kamzlsb0aQ/MwXU6abPXv2UHh4uEZbQ0MDrVixgsRiscbB0BTMof9DQ0PJ3t7epDnowhwObgUFBSQUCungwYMmzaOrmpubyd/fnw4cOGDqVNpVWlpKEomEtm/fbtI8zGF7M4Jn+z1Tppv79+8jLCwMb7/9tka7lZUV3N3doVQq+R7g/+Nf6NCNt7c3oqOjER0dbdBbCN2hubkZ6enpqKqqQkhIiKnTademTZswatQohIWFmTqVXomLKdOaVCqFSCTCgQMH8ODBAyiVSty7dw/79+9HVFQUQkJCcO/evXZ/Uu3JP3M+6DDTWLt2LebOnYuQkJAe9TH7rKwspKamIjMzU+t3Zbvbjh07kJOTg1OnThnte+HPOi6mTGs2Njb4+uuvcfnyZQwaNAhSqRQ+Pj5ITEzE1q1b8be//Q1Dhgxp9fpAW39Hjhwx9eIYxbp165CYmIiKigp4eHjg2LFjpk6pR9myZQvCwsLw8ccfmzoVrQUEBODQoUMa31k2JxkZGWhoaEBWVpbeT92z9glNnQDrWfz9/fGPf/zD1GmYrdjYWMTGxpo6jR5typQpmDJliqnT6DVmzZqFWbNmmTqNXo/PTBljjDE9cTFljDHG9MTFlDHGGNMTF1PGGGNMT73yASRjfaSe9Qzc/7pRfWIuOTnZxJmwZ0FRUVGv/MEHAdFTP07ZwwkEAlOnwBhjrAPBwcFISUkxdRqGlNIrz0yPHj2KefPmmToNZgICgYD7X0dz584FgN52cGNmSrW99TZ8z5QxxhjTExdTxhhjTE9cTBljjDE9cTFljDHG9MTFlDHGGNMTF1PGGGNMT1xMO5CamgpPT88Of5dzwIABBpnXuHHjYGlpiVGjRhlkek965513IJfLIRAIkJOTo3PcqVOnYGNjgy+++MLgubGe7/Tp01i7dm2r/eWNN95oFTtlyhTI5XJYWlpi2LBhuHjxogky1l50dDR8fHygUCggFovh7e2NNWvWtPkD5tnZ2ZgwYQJkMhlcXV0RGRmJhoaGLsfpoqWlBTt37oSfn1+7MYbK7/jx49i2bRuam5v1yrnXoV4GAB09etSg0/Ty8iIbGxv1v5uamqi2tpYePHhAQ4cONdh8AgIC6IUXXjDY9J50+PBhAkCXLl3SOe7EiROkUCjo+PHjRsnNkIzR/71dcHAwBQcHd2ncqKgomjFjBlVWVqrbvLy8yMHBgQDQiRMnWo2TmZlJs2bN6nK+3WnSpEm0Z88eevjwIVVWVtLRo0dJJBLRq6++qhF3+fJlkkqltGHDBqqurqZvv/2WHB0dacmSJV2K08Uvv/xCEyZMIADtHj8MnV98fDxNmjSJysrKdM5Xn+3NjCVzMdXC08X0SYY8KAQEBNCoUaMMNr0n6VNMexJTF9Pa2lry9fXtUfPo6sHt448/pkGDBlFdXZ1Gu5eXFx06dIgsLCzIzc2NysvLNYb3pGI6ffp0ampq0mibN28eAaDbt2+r2+bPn08eHh7U0tKibouLiyOBQEA//fSTznHaysnJoTlz5lBSUhKNGjWq3WJqjPzCwsLI19eXlEqlTjn31mLKl3n1lJ6ebtDpiUQig05PRdvPLHbH5xiJCCkpKdi7d6/R59XdDhw4gOLi4h4/j85cu3YNGzZswEcffQSJRNJquJ+fH8LDw3H37l2sXr3aBBkaxokTJ2BpaanR5ujoCACora0FADQ1NeHkyZOYNGmSxv4zdepUEBEyMjJ0itPFCy+8gNTUVCxcuBBisbjNGGPlt2nTJuTk5CA+Pl7nvHsjLqYGFB8fD2tra1hYWGDMmDFwdnaGSCSCtbU1Ro8eDX9/fzz//POQSCSwtbXFmjVrWk3j2rVrGDJkCKytrSGVSuHv74/s7GyNmObmZkRFRcHd3R1SqRQjR47E0aNH1cOJCHFxcRg8eDDEYjFsbGwQERHRal7axGVnZ8Pd3R0CgQCffvopACAhIQHW1taQyWTIyMjA1KlToVAo0K9fPxw+fLhVrrGxsRg8eDCkUikcHR3h4eGB2NhYs/jkHxFhx44dGDp0KMRiMezs7DB79mz8/PPP6piwsDBYWVnBxcVF3fbee+/B2toaAoEApaWlAIDw8HCsWrUKhYWFEAgE8Pb2xq5duyCRSNC3b1+8++67cHV1hUQigZ+fHy5cuGCQeQDAl19+CYVCgS1bthh1fans2rULRISZM2e2GxMTE4NBgwZh//79OH36dIfT06YfdN3uOtpH9HH37l1IpVJ4eHgAAK5fv47q6mq4u7trxHl5eQEAcnNzdYozNGPlZ2dnh0mTJiE+Ph7Uuz7x3jUmOiU2GnTTZd7ly5dTXl5eq9iNGzcSALpw4QLV1NRQaWkpvfrqqwSATp48SSUlJVRTU0NhYWEEgHJyctTjBgQEkKenJ924cYOUSiVdvnyZfvOb35BEIqFffvlFHbd69WoSi8V07NgxKisro3Xr1pGFhQV9//33RET04YcfkkAgoD//+c9UVlZGtbW1tGfPnlaXb7WNu3PnDgGg3bt3a4wLgL755huqqKig4uJi8vf3J2tra2psbFTHbdmyhSwtLSkjI4Nqa2vpxx9/JGdnZ3rppZf06JH26dr/UVFRZGVlRQcPHqTy8nLKzc2l0aNHk6OjI92/f18dt3DhQnJ2dtYYNy4ujgBQSUmJui0oKIi8vLw04kJDQ8na2pquXLlC9fX1lJ+fT+PGjSO5XK5xqVCfeZw4cYLkcjlFR0drvewqXbns5unpST4+Pm0O8/Lyohs3bhAR0bfffksWFhY0YMAAqq6uJqK2L/Nq2w/abned7SNdVVNTQ3K5nMLCwtRtZ8+eJQAUFxfXKl4qlVJAQIBOcV31m9/8ps3LvMbMb+3atTrfFuLLvM+4iooKjad4/+u//qvDeB8fH8hkMjg4OGDBggUAAHd3dzg6OkImk2HRokUAoPE/bwCQy+UYMGAAhEIhhg0bhn379qG+vl59SbS+vh4JCQkIDAxEUFAQbG1tsX79eohEIiQmJqKurg47d+7EK6+8gpUrV8LW1hZSqRT29vYa89E2rjN+fn5QKBRwcnJCSEgIampqcPv2bfXw9PR0jBkzBjNnzoRUKsXo0aMxa9YsnDt3Do2NjTrNy9Dq6uqwY8cOzJkzB4sWLYKNjQ1GjBiBzz77DKWlpQa9DC0UCtVnXT4+PkhISEBVVRUSExMNMv3p06ejsrISGzZsMMj0OlJTU4MbN26oz1g64uvrixUrVuDmzZv44IMP2ozpSj90tN11to/oIzY2Fq6uroiJiVG3qZ50ffpyMPD4tk1dXZ1OcYZmzPwGDhwIAMjLyzNYvj0VF1Mt2djYgIjUf8uXL9d6XCsrKwCP70moqO6NKpXKDscdMWIEbGxs1JdYrl69itraWgwfPlwdI5VK4eLigp9//hnXrl1DbW0tAgICOpyutnG6UC3nk8tUX1/f6hJQc3MzRCJRmzttd8rPz0d1dTXGjh2r0T5u3DhYWVlpXIY1tLFjx0Imk7X6z1RPUFxcDCKCTCbTKj4mJgaDBw/Gnj17Wt2yAPTvh6e3u872ka5KS0tDcnIyvvrqK8jlcnW76p7xk/u3SmNjI6RSqU5xhmbM/FTbwIMHDwyWb0/FxbSL4uPjNXZWYxKJROoDRU1NDQBg/fr1GmfKt27dQm1trfqHnp2cnDqcprZx+po2bRp+/PFHZGRkoK6uDj/88APS09Px2muvmbyYlpeXAwD69OnTapitrS2qqqqMOn+xWIySkhKjzsMY6uvrAaDdB16eJpFIkJiYCIFAgLfeeqvVGY6h+6GzfaQrjhw5gq1btyIrK6vVu+Wq+9yVlZUa7bW1taivr4erq6tOcYZmzPxUBVa1TTzLuJiauaamJjx69Ej9UICq+O3cuVPjTJmIcP78efX/Ljt7CVzbOH1t2rQJL7/8MhYvXgyFQoE5c+Zg3rx52Ldvn1Hnqw1bW1sAaPNgXV5ejn79+hlt3kql0ujzMBbVAVSXl/Z9fX2xcuVKFBQUYPPmzRrDDN0Pne0jutq9ezeSkpJw5swZPPfcc62Ge3h4QC6X49atWxrt165dAwCMHDlSpzhDM2Z+qls1xjqr7km4mOrp119/xZIlS4w2/X/+859oaWnB6NGjAUD9NHB7XzIaPnw4LCwscPbs2Q6nq22cvvLz81FYWIiSkhIolUrcvn0bCQkJsLOzM+p8tTF8+HD06dMHP/zwg0b7hQsX0NjYiDFjxqjbhEJhp5fkdZGVlQUiwvjx4402D2Pp27cvBAIBKioqdBpv8+bNGDJkCC5duqTRrks/aKOzfURbRITIyEjk5eUhPT29zTNn4HG/TZs2DefOnUNLS4u6PTMzEwKBQP3Es7ZxhmbM/FTbgLOzs1Fy70m4mHYREaGurg6pqalQKBQGm25jYyMqKirQ1NSEixcvIiwsDP3798fixYsBPD6jXLJkCQ4fPoyEhARUVlaiubkZRUVF+PXXX+Hk5ISgoCAcO3YMBw4cQGVlJXJzc1s9xKFtnL6WLVsGd3f3Nj+/ZmoSiQSrVq1CWloakpKSUFlZiby8PCxduhSurq4IDQ1Vx3p7e+PRo0dIT0+HUqlESUlJq//BA4C9vT3u3buHmzdvoqqqSl0cW1paUFZWhqamJuTm5iI8PBzu7u7qftV3HpmZmd32aoxMJoOnp6f6VoG2VJd7n768r0s/aDufjvYRAAgJCYGzs3OHnzO8cuUKPvnkE+zbtw8ikajVp0S3b9+ujt2wYQMePHiAjRs3oqamBufPn0dcXBwWL16MwYMH6xynTX66MHR+KqptYMSIEQbJs0fr/ieIjQsGfDUmLS2NvLy8CECHf+vXryeix5/YkslkBIAGDBhA//rXv2jr1q1kY2NDAMjZ2ZkOHTpER44cIWdnZwJAdnZ2dPjwYSIiSkxMpMmTJ1Pfvn1JKBSSg4MDLViwgG7duqWRV0NDA0VGRpK7uzsJhUJycnKioKAgys/PJyKiqqoqeuedd8jBwYH69OlDEydOpKioKAJA/fr1o3//+99ax+3evZtcXFwIAMlkMpo5cybt2bNHvZwDBw6kwsJC2rt3LykUCgJA/fv3V7/Kc+bMGfWn5VR/IpGIhg4dSqmpqQbppyfp2v8tLS0UFxdHAwcOJJFIRHZ2dhQYGEhXr17ViHv48CFNnjyZJBIJeXh40Pvvv08REREEgLy9vdWvuFy8eJH69+9PUqmUJk6cSPfv36fQ0FASiUTk5uZGQqGQFAoFzZ49mwoLCw02j1OnTpFcLqeYmBid11lXXlUICwsjkUhEtbW16rYn9xdHR0datmxZm+NGRES0ejVGm37QZbvrbB8JDAwkABQVFdXuMubl5XW43z/9CsnZs2fpxRdfJLFYTK6urhQREUH19fWtpqtNnDb5ERGdP3+eJkyYQK6uruq8XFxcyM/Pj86ePWu0/FSmT59Obm5uGl9M6kxvfTWGiykzqj179lB4eLhGW0NDA61YsYLEYrHGwdgQzLH/Q0NDyd7e3tRptKsrB7eCggISCoV08OBBI2VlXM3NzeTv708HDhwwdSptMvf8iIhKS0tJIpHQ9u3bdRqvtxZTvszLjOb+/fsICwvD22+/rdFuZWUFd3d3KJXKHnGP0BB62y9seHt7Izo6GtHR0WZ5Cb8jzc3NSE9PR1VVFUJCQkydTivmnp/Kpk2bMGrUKISFhZk6FbPAxZQZjVQqhUgkwoEDB/DgwQMolUrcu3cP+/fvR1RUFEJCQgx6v5l1r7Vr12Lu3LkICQnR+WEkU8rKykJqaioyMzO1fle2O5l7fgCwY8cO5OTk4NSpU0b7nnhPw8WUGY2NjQ2+/vprXL58GYMGDYJUKoWPjw8SExOxdetW/O1vfzN1ika3bt06JCYmoqKiAh4eHjh27JipUzKoLVu2ICwsDB9//LGpU9FaQEAADh06pPEdZHNi7vllZGSgoaEBWVlZZvFUvrkQmjoB1rv5+/vjH//4h6nTMJnY2FjExsaaOg2jmjJlCqZMmWLqNFg3mTVrFmbNmmXqNMwOn5kyxhhjeuJiyhhjjOmJiyljjDGmJy6mjDHGmJ4ERL3rJ9IFAgHGjx/fIz8gzvR37Ngx7n8dfffddwCg8Z1gxozlu+++w/jx45GSkmLqVAwppdcV07lz55o6BcYM5v79+7h06RKmTp1q6lQYMxjVrwj1Ir2vmDLWmyQnJ2P+/PmtfmCdMWZWUvieKWOMMaYnLqaMMcaYnriYMsYYY3riYsoYY4zpiYspY4wxpicupowxxpieuJgyxhhjeuJiyhhjjOmJiyljjDGmJy6mjDHGmJ64mDLGGGN64mLKGGOM6YmLKWOMMaYnLqaMMcaYnriYMsYYY3riYsoYY4zpiYspY4wxpicupowxxpieuJgyxhhjeuJiyhhjjOmJiyljjDGmJy6mjDHGmJ64mDLGGGN64mLKGGOM6YmLKWOMMaYnLqaMMcaYnriYMsYYY3riYsoYY4zpiYspY4wxpicupowxxpieuJgyxhhjeuJiyhhjjOlJaOoEGGOPKZVKVFdXa7TV1NQAAMrKyjTaBQIBbG1tuy03xljHuJgyZiYePXoENzc3NDc3txpmb2+v8e/JkyfjzJkz3ZUaY6wTfJmXMTPh7OyM3/72t7Cw6Hi3FAgEWLBgQTdlxRjTBhdTxszIG2+80WmMpaUl5syZ0w3ZMMa0xcWUMTMSFBQES6ZpFgAAIABJREFUobD9uy+WlpZ49dVX4eDg0I1ZMcY6w8WUMTOiUCgwderUdgsqEWHRokXdnBVjrDNcTBkzM4sWLWrzISQAsLKywmuvvdbNGTHGOsPFlDEz89prr0Emk7VqF4lECAwMhLW1tQmyYox1hIspY2ZGIpFgzpw5EIlEGu1KpRILFy40UVaMsY5wMWXMDL3++utQKpUabQqFAr/73e9MlBFjrCNcTBkzQ6+88orGhxpEIhEWLFgAKysrE2bFGGsPF1PGzJBQKMSCBQvUl3qVSiVef/11E2fFGGsPF1PGzNSCBQvUl3qdnZ0xceJEE2fEGGsPF1PGzJSfnx/c3NwAAG+++WannxlkjJmOwT50X1RUhG+//dZQk2OMARg3bhzu3r0LBwcHJCcnmzodxnqVefPmGWxaAiIiQ0woOTkZ8+fPN8SkGGOMMaMzUPkDgBSD/wSbAZNjrEeaO3cuACAlJcUg0zt27BiCg4MNMi1zpfrPOB8/WHcwxskf34RhzMz19kLKWG/AxZQxxhjTExdTxhhjTE9cTBljjDE9cTFljDHG9MTFlDHGGNOTyYrpuHHjYGlpiVGjRnUae+rUKdjY2OCLL75oN+add96BXC6HQCBATk6OTuMak6nnv337dvTt2xcCgQCfffaZzuO/9NJLEAgEbf716dOnzXFOnz6NtWvX6pu6UWizPMePH8e2bdva/YHu7mLqbacnUW1zqamp8PT0VPfpG2+80Sp2ypQpkMvlsLS0xLBhw3Dx4kUTZKy96Oho+Pj4QKFQQCwWw9vbG2vWrEF1dXWr2OzsbEyYMAEymQyurq6IjIxEQ0NDl+N00dLSgp07d8LPz6/dGEPlZy776JNMVky///57TJ48WatYbd49279/P/bt29elcY3J1PNfvXq10b5M1da3Yjdu3Ihdu3Zh3bp1RpmnMamWZ+bMmZBIJAgICEB5ebnJ8jH1ttNTPLnNBQUF4fr16/Dy8oKDgwOSkpJw8uRJjfivv/4aKSkpmDFjBvLz8zF69GgTZa6dM2fOYNmyZbh58yZKS0sRGxuL+Ph49fvMKvn5+ZgyZQoCAgJQUlKCtLQ0fP7551i6dGmX4nRRUFCA3/72t1i5ciVqa2vbjDFkfuayj2ogAzl69CjpOrmAgAAaNWqUoVKgw4cPEwC6dOmSwaapi9raWvL19TXJvDtSUFBAAOgvf/mLzuP+/ve/p8rKylbtoaGh9M0332i0ffzxxzRo0CCqq6vrcq7GpsvyhIWFka+vLymVSp3mERwcTMHBwXrlaW6MvW135fhB1P425+XlRYcOHSILCwtyc3Oj8vJyjeGZmZk0a9YsvXLuLtOnT6empiaNtnnz5hEAun37trpt/vz55OHhQS0tLeq2uLg4EggE9NNPP+kcp62cnByaM2cOJSUl0ahRo+iFF15oM84Y+XV1H+3q9taBZJPfM1X9xJQhCAQCg02rKw4cOIDi4mKT5mBoX375JeRyuUbbnTt3cPnyZbz88svqtmvXrmHDhg346KOPIJFIOpwmESElJQV79+41Ss4d0XZ5AGDTpk3IyclBfHx8d6Zolsxx2+5sm/Pz80N4eDju3r2L1atXmyBDwzhx4gQsLS012hwdHQFAfRbY1NSEkydPYtKkSRrHwalTp4KIkJGRoVOcLl544QWkpqZi4cKFEIvFbcYYKz9z2kdNXkyvXbuGIUOGwNraGlKpFP7+/sjOzlYPz87Ohru7OwQCAT799FN1OxEhLi4OgwcPhlgsho2NDSIiIjSm3da4n3zyCWQyGeRyOYqLi/+PvTuPi6rq/wD+GZh9mBlA1kR2FVHMfSH9ieFjT5oLrpj2pD0WZoa4LyluiBqmPipaptEvLRXUwFS0l5WaieYaiCuE4pKCyD4sA3x/f/SbeRhmBmaYAcY679eLPzj33HO/995z7pl77oY5c+agdevWuH37NqqrqxEZGQl3d3eIRCJ07twZ+/fv1yhz9+7d6NGjB4RCISQSCTw9PbFq1SpERERgzpw5yMzMBIfDga+vb72xb9iwAR06dIBAIICdnR1GjhyJW7duqfNs27YNEokEYrEYSUlJeP311yGTyeDm5oa9e/dqxPTzzz/D398fcrkcQqEQAQEBOHHihOk7R4+1a9di5syZGmmbN28GEWH48OEa6dXV1YiOjkb79u0hEong4OAALy8vREdHq18yrW+fvPbaa+Dz+XBxcVGX98EHH0AikYDD4eDZs2dNtj4AYGdnhwEDBmDTpk3NPuSqq+4YWic2b94MoVAIJycnTJs2Da6urhAKhQgMDMSFCxfU+cLDww3avrrqNvDnDxOZTIbVq1c3xybRoq/O1RYVFYV27dph586dOHnyZL3lmbtdGnI8aaxHjx5BJBLBy8sLAPD777+jpKQE7u7uGvl8fHwAAKmpqUblM7emiq8l26gWc53jNnaY19vbm7KyskipVNL169epd+/eJBQK6c6dO+p8Dx48IAC0ZcsWddpHH31EHA6HPvnkE8rPzyeFQkGxsbFaw7z65gVAM2fOpC1bttCoUaPo5s2bNHfuXBIIBHTgwAHKz8+nxYsXk5WVFV28eJGIiDZu3EgAaM2aNZSXl0fPnz+nzz77jCZOnEhERKNHjyYfHx+NddS1/MjISOLz+bR7924qKCig1NRU6tatGzk4ONCTJ0+04vzhhx+osLCQcnJyqH///iSRSKiyslKdLyEhgZYvX07Pnz+nvLw86tOnD7Vq1Uo93ZRh3roePnxI/v7+VF1drZHu7e1N/v7+WvlXr15N1tbWlJSURAqFgi5fvkzOzs4UFBSkkU/fPpk4cSI5Oztr5I2JiSEAlJub22Tro7Jo0SKjLx2Ya5i3vrrbUJ0ICwsjiURCN27coPLyckpPT6eePXuSVCrVGBo0dPvqqttHjhwhqVRKK1euNHldG3P80FfniP4c5s3KyiIionPnzpGVlRV5enpSSUkJEeke5jV3u2zoeNJYpaWlJJVKKTw8XJ12+vRpAkAxMTFa+UUiEQUHBxuVr7F69+6tc5i3KeNrTBv9Sw7zSqVSeHp6gsvlomPHjvj8889RXl5e7xBgWVkZNm7ciEGDBmH27NmwtbWFSCSCvb29Ucteu3YtZsyYgYMHD8LT0xPbtm1DSEgIRo8eDVtbWyxZsgQ8Hg9xcXFQKpVYsWIFBg4ciIULF8Le3h52dnb497//jZ49exq8zLKyMmzYsAGjRo3CpEmTIJfLERAQgE8//RTPnj3Tud6BgYGQyWRwdHREaGgoSktLkZ2drZ4+ZswYLFu2DHZ2drC3t8fw4cORl5eH3Nxco7aHIdauXYsPP/xQ49uapaWlyMrKUv96rC0xMRHdu3fH8OHDIRKJ0K1bN4wYMQJnzpxBZWWlzvJV+8TPz8/s8etaXt31qa1t27YAgLS0tCaPxRgN1QkA4HK56rMsf39/bNu2DcXFxYiLizNLDEOHDkVRURGWLl1qlvKMUV+dq6tv376YNWsW7t27h4ULF+rMY+52WV5eXu/xxBTR0dFwdXVFVFSUOk11p2vd4WDgz0tpZWVlRuUzt6aMz1LaaIt3pnUFBARALpfXO9yQkZEBhUKB4OBgsy339u3bUCgU6NSpkzpNJBLBxcUFt27dQmpqKgoKCvDaa69pzGdtba1ziFCf9PR0lJSUoEePHhrpPXv2BJ/P1xiG04XP5wMAlEql3jyq69Dmvm388ePHOHz4MCZPnqyRnpOTAyKCWCzWmqe8vFxr+KW6uho8Hk9ng2lO+tanNtU6PX36tJmiMp4hdQIAevToAbFYrDFs+aKqr87pEhUVhfbt2yM2NlbjMpKKudtlQ8eTxjp06BDi4+Nx4sQJjWv/qmvGVVVVWvNUVlZCJBIZlc/cmjI+S2mjFteZAn92BvUdGB4+fAgAcHR0NNsyS0tLAQBLlizRePbw/v37UCgUKCoqAgDY2tqatBzVbdy6ntG0tbVFcXGx0WUePXoUQUFBcHR0hEAgwPz5802KUZ9169bh3Xff1brZo7y8HAB03nwwZMgQXL58GUlJSSgrK8OlS5eQmJiIN954o8U7U33rU5uq8arW8UUnEAiaZMSiudVX53QRCoWIi4sDh8PBO++8o3WGY+522dDxpDH27duHtWvX4tSpU/D09NSYprrurTpOqSgUCpSXl8PV1dWofObWlPFZShu1uM60qqoKz58/17oAXZvq4GfqQ8a1qTrmjRs3gog0/lJSUvDSSy8BgMk3vag6Y12Ns6CgAG5ubkaVl52djZCQELi4uODChQsoLCzEunXrTIpRlydPnuCbb77B9OnTtaapKrOuM+Hly5fj1VdfxeTJkyGTyTBq1CiMGzdO5zPBzam+9alNNRTdVL/Ym5NSqWxUHbNE9dU5ffr27YvZs2fj7t27WLVqlcY0c7fLho4nxtqyZQv27NmDH3/8UX0sqs3LywtSqRT379/XSM/IyAAAdO7c2ah85taU8VlKG7W4zvSnn35CTU1NvQ9Sd+rUCVZWVjh9+rTZltumTRsIhUKNtyfV5unpCXt7e3z//fcmLadTp06wsbHBpUuXNNIvXLiAyspKdO/e3ajy0tLSoFQqMX36dHh7e0MoFDbJI0Lr1q3DpEmTdF6XVr1hqbCwUGtaeno6MjMzkZubC6VSiezsbGzbtg12dnYGLZfL5TY4fNkY9a1Pbap1cnZ2NnsMze3UqVMgIvTp00ed1lTbt6nVV+fqs2rVKvj5+eHq1asa6eZulw0dTwxFRFiwYAHS0tKQmJio961jXC4XQ4YMwZkzZ1BTU6NOT05OBofDUd/xbGg+c2vK+CyljbZ4Z1pZWYnCwkJUVVXhypUrCA8Ph4eHR73XsRwdHTF69GgcOHAAu3btQlFREVJTU016blEoFGLKlCnYu3cvtm3bhqKiIlRXV+Phw4f4448/IBAIsHjxYpw5cwbh4eF49OgRampqUFxcjBs3bgAA7O3t8fjxY9y7dw/FxcU6D1JCoRBz5szBoUOHsGfPHhQVFSEtLQ3vv/8+XF1dERYWZlTcqjP4kydPory8HHfv3m3w+o6xnj59ii+++AKzZs3SOV0sFsPb21s9/F7bjBkz4O7urvPVZ4bw9fXF8+fPkZiYCKVSidzcXK1frcZqaH1qU61TQECASctsCTU1NcjPz0dVVRVSU1MREREBd3d3jbZl6PbVVbeTk5Nb7NGY+upcfVTDvXUvMZi7XTZ0PAGA0NBQODs71/s6wxs3buDjjz/G559/Dh6Pp/UKzPXr16vzLl26FE+fPsWyZctQWlqKlJQUxMTEYPLkyWjfvr3R+QyJzxjmjk/FYtqoue4LbsytxnFxcTRw4EBycnIiLpdLrVq1ogkTJtD9+/fVebZs2UIuLi4EgMRiMQ0fPpyIiIqLi2nq1KnUqlUrsrGxoX79+lFkZCQBIDc3N/rtt990zrtu3ToSiUQEgNq0aUO7d+9WL6uiooIWLFhA7u7uxOVyydHRkUaPHk3p6enqPFu3bqWAgAASCoUkFAqpa9euFBsbS0REV65cIQ8PDxKJRNSvXz9asmSJzthramooJiaG2rZtSzwej+zs7CgkJIRu376tXk5sbCyJxWICQG3btqXMzEzasWMHyWQyAkAeHh7qx4cWLFhA9vb2ZGtrS2PHjqWtW7cSAPLx8aGIiAhydnYmACSRSGjUqFFG7lmi2bNn06RJk+rNEx4eTjwejxQKhUb6jz/+SK1atSIA6j8ej0cdOnSggwcPEhHVu0/y8vJo4MCBJBQKycvLiz788EOaN28eASBfX1+NxzzMuT4qQ4cOpdatW2u8jaUh5ng0RlfdNaZOhIWFEY/Ho9atWxOXyyWZTEYjR46kzMxMjeUYun3r1u0nT57QsWPHSCqVUlRUlEnrStS444euOnfo0CHy8fEhAOTg4EAzZszQOe+8efO0Ho0xd7ts6HgSEhJCACgyMlLvOqalpWm0nbp/dR8hOX36NPXq1YsEAgG5urrSvHnzqLy8XKtcQ/IZEh8RUUpKCr3yyivk6uqqjsvFxYUCAwPp9OnTTRafSmPaaFM8GtOinSnz13H37l3icrkaHSHRnwefiIgIjbSKigqaNWsWCQQCrc7Xkjx79oyEQiGtX7/eqPks4XWCYWFhZG9v36IxGKMxxw99de5FUV1dTf3796ddu3a1dCg6WXp8RI1vo3/J50yZvwZfX1+sXLkSK1euVA/pPnnyBOHh4fj3v/+tkZfP58Pd3R1KpdKir9ctX74cXbp0QXh4eEuH0iiW9EWNpqCrzr0oqqurkZiYiOLiYoSGhrZ0OFosPT4VS2qjrDP9m7l165beT5DV/mtMA1q0aBHGjh2L0NBQFBYWQiQSgcfjYdeuXXj69CmUSiUeP36MnTt3IjIyEqGhoZDJZBa5Phs2bMC1a9dw7Ngxs74/mjGvunXuRXHq1CkcPHgQycnJBj8r25wsPT7AAtuouc5x2TAvo3LixAlasGABERGdOXOGBg0aRDKZjKytrUkul1NgYCDFxsYa/aWH5pKYmEjR0dFaX+owVEsP8y5atIj4fD4BIE9PT0pISGixWAxl6vGjdp1j/vpMbaNNMczLITLP24Hj4+Mxfvz4ln/ZMMO0MNV3JhMSElo4khcHO34wzakJ6lsCG+ZlGIZhGBOxzpRhGIZhTMQ6U4ZhGIYxEetMGYZhGMZEXHMXqLr5gmH+rs6fPw+AtQVjqF4Jx7YZ0xyMfQ2lIdiZKcMwDMOYyOxnpuxxAObvjj0aYzzVowpsmzHNQVXfzImdmTIMwzCMiVhnyjAMwzAmYp0pwzAMw5iIdaYMwzAMYyLWmTIMwzCMiV6IzvTgwYPw9vbW+qwWl8uFg4MDBg0ahEOHDmnNd+zYMcjlcnz33Xd6y546dSqkUik4HA6uXbtm1LxNqaWXv379ejg5OYHD4eDTTz/VmefkyZNYtGiRRtqdO3fw4YcfomPHjpDJZODz+XB0dISfnx9GjRqFb7/9Vp1X334VCoXw8vLCO++8g6ysLI3y//Of/+Cll14Ch8OBlZUV2rVrh5MnT2rkeeONNyCTyWBlZQU/Pz/88ssvOHz4MNatW/eX/8bn35WqLtatU2+99ZZW3sGDB0MqlcLa2hodO3bElStXWiBiw61cuRL+/v6QyWQQCATw9fXF/PnzdX7D9ezZs3jllVcgFovh6uqKBQsWoKKiotH5zBlfVFSUzs8jdurUSatMpVKJ6Oho+Pr6gs/nw9bWFp06dcK9e/cAwDLbs7m+P9Mcn2Dz8fEhuVyu/v/58+d08uRJ8vPzIwC0b98+jfxHjhwhmUxGhw8frrfcvXv3EgC6evWq0fM2lZZePhHR3bt3CQBt375da1pkZCQNGzaMioqK1GlxcXHE5/OpX79+dPz4ccrPz6fy8nLKzMyk7777joYOHUphYWFaZdXer9XV1fT06VP66quvSCwWk5OTEz179kxrHgDUu3dvvbH/9NNPFBwcrJG2adMmGjBgAOXn5xu8DRqjpT/B9iIy5fihqy76+PhQq1atCAAdOXJEa57k5GQaMWJEo+NtTgMGDKDY2FjKy8ujoqIi2r9/P/F4PPrnP/+pke/69eskEolo6dKlVFJSQufOnSMHBweaMmVKo/KZO75Vq1YRAK2/jh07apUZEhJC7du3p/Pnz5NSqaTHjx/T8OHDKS0tTZ3HlPbcFJ9ge6E7U5UTJ04QABo1alSjytXVmTYnhUJBffv2bZFl10dfZ7pmzRpq164dlZWVqdNSUlLI2tqagoKC9H6nNDMzs8HOtLb58+fr/JFE1LjOlIgoPDyc+vbt26TfUrWEzrQ56pQ5l9HY44euukj0Z536+uuvycrKilq3bk0FBQUa01+kznTo0KFa3+0cN24cAaDs7Gx12vjx48nLy4tqamrUaTExMcThcOjmzZtG5zN3fKtWraLdu3c3WN7evXuJw+FQampqg3kb256bojN9IYZ5G+Lp6QkAKCgoaNT8HA7HjNEYb9euXcjJyWnRGAyVkZGBpUuXYsWKFRAKher01atXo7q6GmvWrAGXq/tdIN7e3nqHjHXx9fUFADx58sS0oGtZvnw5rl27hk2bNpmtTEvUHHWqpeutvrqoEhgYiIiICDx69Ahz585tgQjN48iRI7C2ttZIc3BwAAAoFAoAQFVVFY4ePYoBAwZoHM9ef/11EBGSkpKMymfu+Iyxfft2dOvWDQEBAQ3mtaT2/JfoTFNTUwEAAwYMUKedPXsW7u7u4HA42Lp1qzqdiBATE4P27dtDIBBALpdj3rx5GuXpmvfjjz+GWCyGVCpFTk4O5syZg9atW+P27duorq5GZGQk3N3dIRKJ0LlzZ+zfv1+jzN27d6NHjx4QCoWQSCTw9PTEqlWrEBERgTlz5iAzMxMcDge+vr71xr5hwwZ06NABAoEAdnZ2GDlyJG7duqXOs23bNkgkEojFYiQlJeH111+HTCaDm5sb9u7dqxHTzz//DH9/f8jlcgiFQgQEBODEiRP1buvNmzeDiDB8+HB1WmVlJU6ePAl7e3v06dOn3vmNcffuXQDAyy+/bLYy7ezsMGDAAGzatMmiPkRtyL4NDw8Hn8+Hi4uLOu2DDz6ARCIBh8PBs2fPAEBnndq8eTOEQiGcnJwwbdo0uLq6QigUIjAwEBcuXDDLMgDg+PHjkMlkWL16dZNuL0B3XawrKioK7dq1w86dO7Wurddl7vZlyHGhsR49egSRSAQvLy8AwO+//46SkhK4u7tr5PPx8QHw32OkofnMHZ+hKisrcf78eXTp0sWg/BbVns11jtsSw7wKhYKSk5PJw8ODBg8eTCUlJRr5Hzx4QABoy5Yt6rSPPvqIOBwOffLJJ5Sfn08KhYJiY2O1hnn1zQuAZs6cSVu2bKFRo0bRzZs3ae7cuSQQCOjAgQOUn59PixcvJisrK7p48SIREW3cuJEA0Jo1aygvL4+eP39On332GU2cOJGIiEaPHk0+Pj4Nxh4ZGUl8Pp92795NBQUFlJqaSt26dSMHBwd68uSJVpw//PADFRYWUk5ODvXv358kEglVVlaq8yUkJNDy5cvp+fPnlJeXR3369KFWrVqpp+sa5vX29iZ/f3+NWO/cuUMAqE+fPvXtPr3q7tf8/Hz68ssvSSwW09ChQ3XOg0YO8xIRLVq0qEmH9RszzGvovp04cSI5OztrzBsTE0MAKDc3V52mq06FhYWRRCKhGzduUHl5OaWnp1PPnj1JKpVqDMeZsowjR46QVCqllStXGrX+jTl+6KqLKj4+PpSVlUVEROfOnSMrKyvy9PRUHyN0DfOau301dFxorNLSUpJKpRQeHq5OO336NAGgmJgYrfwikUjdFgzNZ+74iP4c5nVzcyNbW1vi8Xjk6elJI0aMoF9//VWdJysriwBQly5dKCgoiFxcXEggEJCfnx9t3bpVY2hapTHtmQ3zAigsLFTfBSYWi9W/DCdOnAgej1fvvGVlZdi4cSMGDRqE2bNnw9bWFiKRCPb29kbFsHbtWsyYMQMHDx6Ep6cntm3bhpCQEIwePRq2trZYsmQJeDwe4uLioFQqsWLFCgwcOBALFy6Evb097Ozs8O9//xs9e/Y0eJllZWXYsGEDRo0ahUmTJkEulyMgIACffvopnj17hh07dmjNExgYCJlMBkdHR4SGhqK0tBTZ2dnq6WPGjMGyZctgZ2cHe3t7DB8+HHl5ecjNzdUZQ2lpKbKystS/YlWKiooAADY2NgavT12196udnR2mTJmCxYsXa9z9ay5t27YFAKSlpZm97MZozL5tLC6Xqz7z8vf3x7Zt21BcXIy4uDizlD906FAUFRVh6dKlZilPH311UZe+ffti1qxZuHfvHhYuXKgzj7nbV3l5eb3HBVNER0fD1dUVUVFR6jTVnbh1h1sBgMfjoayszKh85o4PAN5++20cPnwYDx48QElJCfbu3Yvs7GwMGDAA6enpAKC+A9jR0RGrV69Geno6nj59ipEjR2LGjBn45ptvtJZnKe35hetM5XI5iAhEBKVSiYcPH2LWrFkIDw9H586d1cNQumRkZEChUCA4ONhs8dy+fRsKhULj9m6RSAQXFxfcunULqampKCgowGuvvaYxn7W1NWbOnGnwctLT01FSUoIePXpopPfs2RN8Pl9jqE4XPp8P4M9bzvVR/RjRd7t5Tk4OiAhisVgjXdWJlpaW6pwvPj4eXl5e6s6yQ4cOWtfaau/XefPmgYggl8sb/IHUGKr4nz59avayG8PUfWuKHj16QCwWawxlvgj01UV9oqKi0L59e8TGxuLs2bNa083dvho6LjTWoUOHEB8fjxMnTkAqlarTVdeMq6qqtOaprKyESCQyKp+54wOANm3aoGvXrrCxsQGfz0efPn0QFxeHsrIyxMbGAgAEAgEAoGPHjggMDIS9vT3kcjlWrFgBuVyu80eNpbTnF64zrY3L5aJ169aYMmUK1q9fj9u3b2PNmjV686u+Yefo6Gi2GFQdyJIlSzSenbp//z4UCoX6rM3W1tak5ahurtJ19mdra4vi4mKjyzx69CiCgoLg6OgIgUCA+fPn15u/vLwcwH8rvIqHhwcEAgEyMjJ0zjdu3DhkZWXBw8MDzs7OuHnzJpycnPQuZ+nSpXBxccHixYvx4MEDvflqamr0TquurtbbEasOGKr1aWlNsW+NIRAI9I5GWCp9dVEfoVCIuLg4cDgcvPPOO1pnYObeBw0dFxpj3759WLt2LU6dOqW+6VJFdY1bdbxRUSgUKC8vh6urq1H5zB2fPgEBAbC2tsadO3cAQL38uidFfD4fHh4eyMzM1CrDUtrzC92Z1qa68+vGjRt686h+lTX24WRdVB3zxo0b1WdWqr+UlBS89NJLALQrh7FUnbGuRl1QUAA3NzejysvOzkZISAhcXFxw4cIFFBYWYt26dfXOo6q0dc9chUIhBg0ahNzcXPWHsU0hlUqxdu1aFBcXY/r06Trz2Nvb4/EqlzvoAAAgAElEQVTjx3rLyMrKQps2bXROq6ysBACTf4Wbi7n3rTGUSmWTL6Mp6KuL9enbty9mz56Nu3fvYtWqVRrTzL0PGjouGGvLli3Ys2cPfvzxR/UxpTYvLy9IpVLcv39fI131A7dz585G5TN3fPrU1NSgpqZG/aPIxsYGbdu21Xkcr6qqglwu10q3lPb8l+lML1++DABo37693jydOnWClZUVTp8+bbbltmnTBkKhUOPtSbV5enrC3t4e33//vUnL6dSpE2xsbHDp0iWN9AsXLqCyshLdu3c3qry0tDQolUpMnz4d3t7eEAqFDT4ipHojUmFhoda0FStWgMfjYd68efUOJRvqX//6F3r37o0jR44gPj5ea/qrr76KR48e4dy5c1rTiAhffvklevfurbNsVfzOzs4mx2kOxuxbLpdrlu2rcurUKRCRxl3Y5l5GU6ivLtZn1apV8PPzw9WrVzXSzd2+GjouGIqIsGDBAqSlpSExMVHvfQlcLhdDhgzBmTNnNEZskpOTweFw1Hc8G5rP3PEB0LrUBQAXL14EEaFv377qtPHjx+Pq1av4/fff1WkKhQL379/X+biMpbTnF7IzLSsrQ01NDYgIjx8/RlxcHJYsWQIHBwfMmjVL73yOjo4YPXo0Dhw4gF27dqGoqAipqakm3eAhFAoxZcoU7N27F9u2bUNRURGqq6vx8OFD/PHHHxAIBFi8eDHOnDmD8PBwPHr0CDU1NSguLlb/+lKdZd27dw/FxcU6D2RCoRBz5szBoUOHsGfPHhQVFSEtLQ3vv/8+XF1dERYWZlTcqlvjT548ifLycty9e7fB60JisRje3t7q4fLaunfvjt27d+Py5csICgrC8ePH8ccff6Cqqgr379/H7t278fz5c4Pj43A42Lx5MzgcDsLDw5Gfn68xPSoqCra2thg7diy+/fZblJaWoqKiAr/99hvefPNNVFVV6XyVHPDf4X5DnmNrDsbsW19fXzx//hyJiYlQKpXIzc3VOssA9Nepmpoa5Ofno6qqCqmpqYiIiIC7uzsmT55slmUkJyc3y6Mx9dXF+qiGe+vegGPu9tXQcQEAQkND4ezsXO/rDG/cuIGPP/4Yn3/+OXg8ntar+NavX6/Ou3TpUjx9+hTLli1DaWkpUlJSEBMTg8mTJ2ucZBiaz9zxPXr0CPv27UNBQQGUSiVSUlIwdepUuLu74/3331fnmz17Njw8PDB58mRkZ2cjLy8PCxYsQFlZmc4byCymPZvrvuCmfDTm0KFD5OPjo/NVVAKBgNq2bUvTp0/XuL1/y5Yt5OLiQgBILBbT8OHDiYiouLiYpk6dSq1atSIbGxvq168fRUZGEgByc3Oj3377Tee869atI5FIRACoTZs2Gm/yqKiooAULFpC7uztxuVxydHSk0aNHU3p6ujrP1q1bKSAggIRCIQmFQuratSvFxsYSEdGVK1fIw8ODRCIR9evXj5YsWaIz9pqaGoqJiaG2bdsSj8cjOzs7CgkJodu3b6uXExsbS2KxmABQ27ZtKTMzk3bs2EEymYwAkIeHB925c4eIiBYsWED29vZka2tLY8eOpa1btxIA8vHxoYiICHJ2diYAJJFI1G+XCg8PJx6PRwqFQue+ysrKooiICOrYsSNJJBISCoXk5eVF/fv3p4ULF9KZM2fUeX/55Rdq166del++9NJLNG3aNI3yJk+eTADI1taW1qxZo7Wsd999l7y8vIjP55NIJCJ/f3+KjIzUekyqtqFDh1Lr1q113mZvDo15NMaQfUtElJeXRwMHDlRv1w8//JDmzZtHAMjX11fdBurWqSdPnlBYWBjxeDxq3bo1cblckslkNHLkSMrMzDTbMo4dO0ZSqZSioqKMWv/GHD901cXaxwoHBweaMWOGznnnzZun9WiMudtXQ8eFkJAQAkCRkZF61zEtLU3ncU/1V/cRl9OnT1OvXr1IIBCQq6srzZs3j8rLy7XKNSSfueObM2cO+fj4kEQiIS6XS25ubvTuu+/S48ePtcp98OABTZgwgezs7EggEFCvXr0oOTlZZwyNac9/+9cJMi3v7t27xOVyDXotmCV69uwZCYVCWr9+fZMtwxJeJ6hLWFgY2dvbt3QYOjXm+PGi18Xq6mrq378/7dq1q6VD0cnS4yNqfHtmz5kyLc7X1xcrV67EypUrdX61wtItX74cXbp0QXh4eEuH0iIs6isbJnqR62J1dTUSExNRXFyM0NDQlg5Hi6XHp2JJ7Zl1pozRFi1ahLFjxyI0NNToG0Ba0oYNG3Dt2jUcO3asSZ5fZZrfi1oXT506hYMHDyI5OdngZ2Wbk6XHB1hee2adKdMoq1evRnh4eL3P9VqSpKQkVFRU4NSpU7Czs2vpcJrd4sWLERcXh8LCQnh5eeHAgQMtHZLZvGh1EQCCg4Px9ddfa7wD2ZJYenyW2J45ROZ5O3B8fDzGjx/f8i8bZpgWNnbsWABAQkJCC0fy4mDHD6Y5NUF9S2BnpgzDMAxjItaZMgzDMIyJWGfKMAzDMCZinSnDMAzDmIh1pgzDMAxjIq65C2zoZekM83fB2oLx2DZjXlRm60wDAwOxf/9+cxXHMAyAlJQUbNq0ibUthrFwZnvOlGEY82PPXzLMC4E9Z8owDMMwpmKdKcMwDMOYiHWmDMMwDGMi1pkyDMMwjIlYZ8owDMMwJmKdKcMwDMOYiHWmDMMwDGMi1pkyDMMwjIlYZ8owDMMwJmKdKcMwDMOYiHWmDMMwDGMi1pkyDMMwjIlYZ8owDMMwJmKdKcMwDMOYiHWmDMMwDGMi1pkyDMMwjIlYZ8owDMMwJmKdKcMwDMOYiHWmDMMwDGMi1pkyDMMwjIlYZ8owDMMwJmKdKcMwDMOYiHWmDMMwDGMi1pkyDMMwjIlYZ8owDMMwJmKdKcMwDMOYiHWmDMMwDGMi1pkyDMMwjIlYZ8owDMMwJmKdKcMwDMOYiHWmDMMwDGMi1pkyDMMwjIm4LR0AwzB/ys3NxbfffquRdunSJQDAjh07NNKlUikmTJjQbLExDFM/DhFRSwfBMAxQUVEBJycnlJSUwNraGgCgap4cDkedT6lU4u2338aXX37ZEmEyDKMtgQ3zMoyFEAgEGDNmDLhcLpRKJZRKJaqqqlBVVaX+X6lUAgDefPPNFo6WYZjaWGfKMBbkzTffRGVlZb15bG1t8eqrrzZTRAzDGIJ1pgxjQQYOHAhHR0e903k8HiZNmgQul93uwDCWhHWmDGNBrKysMHHiRPB4PJ3TlUolu/GIYSwQ60wZxsJMmDBBfW20rpdeegl9+/Zt5ogYhmkI60wZxsL06tULHh4eWul8Ph9vv/22xp29DMNYBtaZMowFeuutt7SGeisrK9kQL8NYKNaZMowFmjhxotZQr6+vLwICAlooIoZh6sM6U4axQH5+fvD391cP6fJ4PEyZMqWFo2IYRh/WmTKMhfrXv/6lfhNSVVUVG+JlGAvGOlOGsVATJkxAdXU1AKBbt27w8vJq4YgYhtGHdaYMY6Hc3d3Ru3dvAMDbb7/dwtEwDFMfrdeopKSkYMOGDS0RC8MwdVRUVIDD4eD777/HmTNnWjochmEAJCQkaKVpnZk+ePAABw4caJaAGIb5r/Pnz+P8+fMaaW5ubnB2doZQKGyhqCzbw4cP2fGKaTb11Te9L/jU1fMyDNN0xo4dC0C77WVkZMDX17clQrJ48fHxGD9+PDteMc1CVd90YddMGcbCsY6UYSwf60wZhmEYxkSsM2UYhmEYE7HOlGEYhmFMxDpThmEYhjFRs3Wm69evh5OTEzgcDj799FOj5u3Zsyesra3RpUuXFouBaZi+7Xvs2DHI5XJ89913TR5DTU0NNm7ciMDAQIPyT506FVKpFBwOB9euXTN6eUFBQeBwODr/bGxsjC7PHJpze7/oTp48iUWLFuHgwYPw9vZW77u33npLK+/gwYMhlUphbW2Njh074sqVKy0QseFWrlwJf39/yGQyCAQC+Pr6Yv78+SgpKdHKe/bsWbzyyisQi8VwdXXFggULUFFR0eh85owvKipKZ/vq1KmTVplKpRLR0dHw9fUFn8+Hra0tOnXqhHv37gEADh8+jHXr1qnfLGZOzdaZzp07F+fOnWvUvBcvXsTAgQNbNAamYfq2LxE1y/Lv3r2L//mf/8Hs2bOhUCgMmmfnzp34/PPPmySefv36NUm5DWmu7f2iW7ZsGTZv3ozFixdj9OjR+P333+Hj44NWrVphz549OHr0qEb+77//HgkJCRg2bBjS09PRrVu3ForcMD/++CNmzJiBe/fu4dmzZ4iOjsamTZvUj2CppKenY/DgwQgODkZubi4OHTqEL774Au+//36j8pk7PmOMHz8eX331Fb7++msoFArcvHkTPj4+6g56+PDhEAqFCA4ORkFBQaOXoxPVsX//ftKRbBZ3794lALR9+3aj5w0ODqYuXbq0aAxMw1pq+167do1GjRpFe/bsoS5dutDLL79s8Lx79+4lAHT16lWjl/vaa69RUVGRVnpYWBj98MMPRpU1ZswYGjNmjNExWDKFQkF9+/ZtsvIbe7xas2YNtWvXjsrKyjTSfXx86OuvvyYrKytq3bo1FRQUaExPTk6mESNGmBRzcxk6dChVVVVppI0bN44AUHZ2tjpt/Pjx5OXlRTU1Neq0mJgY4nA4dPPmTaPzmTu+VatW0e7duxssb+/evcThcCg1NbXBvOHh4dS3b19SKpVGxVxPfYtvsjNTIkJCQgJ27NhhtjLrfiyZ+XvSVbdefvllHDx4EBMnToRAIDCqPNVnzhrj+PHjkEqlGmkPHjzA9evX8eqrrza63L+KXbt2IScnp6XD0JCRkYGlS5dixYoVOt8sFRgYiIiICDx69Ahz585tgQjN48iRI+qvDqk4ODgAgHrkpqqqCkePHsWAAQM02sHrr78OIkJSUpJR+cwdnzG2b9+Obt26GfTN3+XLl+PatWvYtGmT0cvRxyydaXV1NaKjo9G+fXuIRCI4ODjAy8sL0dHRGDduXL3zEhE2bNiADh06QCAQwM7ODiNHjsStW7e08mZkZMDPzw8SiQQikQj9+/fH2bNnNfL8/PPP8Pf3h1wuh1AoREBAAE6cOGHyOm7atAkSiQRWVlbo3r07nJ2dwePxIJFI0K1bN/Tv3x9t2rSBUCiEra0t5s+fb3BcX375JWxsbMDhcGBnZ4fExERcunQJHh4esLa2xptvvmlUrJs3b4ZQKISTkxOmTZsGV1dXCIVCBAYG4sKFCxp5Dd3+xuyn2s6ePQt3d3dwOBxs3boVALBt2zZIJBKIxWIkJSXh9ddfh0wmg5ubG/bu3asxvyl1SxciQkxMDNq3bw+BQAC5XI558+YZXU591q5di5kzZ5q1TEOZsr0NrTfh4eHg8/lwcXFRp33wwQeQSCTgcDh49uwZACAiIgJz5sxBZmYmOByO+uUTx48fh0wmw+rVq5tjk2jZvHkziAjDhw/XmycqKgrt2rXDzp07cfLkyXrLM6RtGFvnIyMj4e7uDpFIhM6dO2P//v2mrfT/e/ToEUQikfoLRL///jtKSkrg7u6ukc/HxwcAkJqaalQ+c8dnqMrKSpw/f97g+2rs7OwwYMAAbNq0yXyXRYw4jdVr9erVZG1tTUlJSaRQKOjy5cvk7OxMQUFBGvl0DQFGRkYSn8+n3bt3U0FBAaWmplK3bt3IwcGBnjx5os4XHBxM3t7elJWVRUqlkq5fv069e/cmoVBId+7cUedLSEig5cuX0/PnzykvL4/69OlDrVq1qjcGQy1btowA0IULF6i0tJSePXtG//znPwkAHT16lHJzc6m0tJTCw8MJAF27ds3guG7cuEFisZjefvttddqiRYto586dRsdJ9Ocwo0QioRs3blB5eTmlp6dTz549SSqVagyhGLr9Dc2na/s+ePCAANCWLVvUaR999BEBoB9++IEKCwspJyeH+vfvTxKJhCorK9X5DK1btfXu3VvvMO9HH31EHA6HPvnkE8rPzyeFQkGxsbGNHuat6+HDh+Tv70/V1dVGz2uuYV5Ttreh9WbixInk7OyssdyYmBgCQLm5ueq00aNHk4+Pj0a+I0eOkFQqpZUrV5q8ro05Xnl7e5O/v7/OaT4+PpSVlUVEROfOnSMrKyvy9PSkkpISItI9zGto2zB0H8ydO5cEAgEdOHCA8vPzafHixWRlZUUXL140aj3rKi0tJalUSuHh4eq006dPEwCKiYnRyi8SiSg4ONiofOaOj+jPYV43NzeytbUlHo9Hnp6eNGLECPr111/VebKysggAdenShYKCgsjFxYUEAgH5+fnR1q1bNYamVRYtWmR0u69vmNcsnWnPnj2pV69eGmnvvfceWVlZUUVFhTqt7oFWoVCQjY0NhYaGasz766+/EgCNxhYcHKx1gExNTSUANHfuXL2xRUdHEwDKycnRGYMxVJ1pcXGxOu1///d/CQClpaVpxb9v3z6D4yIi+uyzzwgA7dmzh7755huaPXu20TGqhIWFkVwu10i7ePEiAaAVK1YQkeHb35j9ZGxnWvualapTy8jIUKcZWrdq09eZKhQKEovF9I9//EMj3ZRrpnXNmDGj0deLm6MzbWh7G1JviEzrTM3J2ONVSUkJcTgcGjZsmM7ptTtTIqI5c+YQAJoxYwYRaXemxrQNQ/ZBWVkZicVijfIUCgUJBAKaPn26weupy0cffUTt2rXTuMb//fffEwDasGGDVn6ZTEaBgYFG5TN3fERE2dnZdOXKFSouLqaKigpKSUmhrl27kkgkouvXrxMRUVpaGgGgf/zjH/TLL79QXl4eFRQU0MKFC9XH1Lq++OILAkBfffWVwTE2+TXT8vJyrVPl6upq8Hg8rTHx2tLT01FSUoIePXpopPfs2RN8Pl9rSLKugIAAyOXyeocYVNdZm+JWaADg8/kA/rymUHeZSqXSqLjee+89jBkzBtOmTUN8fDw+/vhjs8bao0cPiMVi9fCTodvf1P1kKNW2rL3dGlu3dMnIyIBCoUBwcLDpwerw+PFjHD58GJMnT26S8s1N1/bWpW69eZHl5OSAiCAWiw3KHxUVhfbt2yM2NlbrkhJgetuouw9u374NhUKh8diHSCSCi4uLSdv/0KFDiI+Px4kTJzSu8auuGdc+fqlUVlZCJBIZlc/c8QFAmzZt0LVrV9jY2IDP56NPnz6Ii4tDWVkZYmNjAUB9n0THjh0RGBgIe3t7yOVyrFixAnK5XOe9O6o68PTpU5NiVzFLZzpkyBBcvnwZSUlJKCsrw6VLl5CYmIg33nij3gOe6tZkXc/j2draori4uMFl83g8jYPB0aNHERQUBEdHRwgEAq1rly3F0LhWr16NkpKSJrtpQyAQIDc3F4Dh298c+6mxGlu3dHn48CEAwNHRsSlCxbp16/Duu+/+JT+XVrvevMjKy8sBwOCb1IRCIeLi4sDhcPDOO++grKxMY7q520ZpaSkAYMmSJRrPVN6/f79RN+UAwL59+7B27VqcOnUKnp6eGtNU172Lioo00hUKBcrLy+Hq6mpUPnPHp09AQACsra1x584dAFAvX3W9XoXP58PDwwOZmZlaZah+AKjqhKnM0pkuX74cr776KiZPngyZTIZRo0Zh3LhxDT6/Z2trCwA6K1xBQQHc3Nzqnb+qqgrPnz9XXxTPzs5GSEgIXFxccOHCBRQWFmLdunWNXCvzMTQupVKJmTNnYsOGDUhJSUFUVJRZ41AqlRrb1dDtb+p+MkVj65Yuqk6usQ+Z1+fJkyf45ptvMH36dLOX3dLq1psXmeoAasxIVd++fTF79mzcvXsXq1at0phm7rah+qG3ceNGEJHGX0pKilFlAcCWLVuwZ88e/Pjjj3jppZe0pnt5eUEqleL+/fsa6RkZGQCAzp07G5XP3PHpU1NTg5qaGvWPIhsbG7Rt2xY3btzQyltVVQW5XK6VXllZCQAmn1Wr6P2eqTHS09ORmZmJ3NxccLmGF9mpUyfY2Njg0qVLGukXLlxAZWUlunfvXu/8P/30E2pqatQPT6elpUGpVGL69Onw9vYGYNpjD+ZiaFwffvgh3n33XYwaNQqPHj3CqlWrMHjwYPTt29cscZw6dQpEhD59+gAwfPubup9M0di6pUunTp1gZWWF06dPN/pBc33WrVuHSZMmwd7e3qzlWoK69QYAuFxug8PDlkj1hq7CwkKj5lu1ahWOHDmCq1evatzRau62oXoioDFv46qNiLBw4ULk5+cjMTFRb9vhcrkYMmQIzpw5g5qaGlhZ/Xl+lZycDA6Ho77j2dB85o4PAF577TWtJzIuXrwIItI4No4fPx6rV6/G77//rj7OKhQK3L9/H2+88YZWuao64OzsbFTs+pjlzHTGjBlwd3fX+Zqq+giFQsyZMweHDh3Cnj17UFRUhLS0NLz//vtwdXVFWFiYRv7KykoUFhaiqqoKV65cQXh4ODw8PNTXqFSV/OTJkygvL8fdu3fNdj3PFIbEFRsbi9atW2PUqFEAgOjoaPj7+2PixIlaQyuGqqmpQX5+PqqqqpCamoqIiAi4u7urt5eh29/Y/WROja1bujg6OmL06NE4cOAAdu3ahaKiIqSmppr8LPTTp0/xxRdfYNasWSbHaAkaqjfAn99Yff78ORITE6FUKpGbm6t11gIA9vb2ePz4Me7du4fi4mIolUokJye32KMxYrEY3t7e6iF/Q6mGe+teWjB32xAKhZgyZQr27t2Lbdu2oaioCNXV1Xj48CH++OMPAEBoaCicnZ3rfZ3hjRs38PHHH+Pzzz8Hj8fTehXf+vXr1XmXLl2Kp0+fYtmyZSgtLUVKSgpiYmIwefJktG/f3uh85o7v0aNH2LdvHwoKCqBUKpGSkoKpU6fC3d1d40fx7Nmz1f1BdnY28vLysGDBApSVlWHhwoVaMajqgCHPpRrEiLuV9Prxxx+pVatWBED9x+PxqEOHDnTw4EEiIvrkk0/I2dmZAJBEIqFRo0YREVFNTQ3FxMRQ27ZticfjkZ2dHYWEhNDt27c1lhEXF0cDBw4kJycn4nK51KpVK5owYQLdv39fI9+CBQvI3t6ebG1taezYsbR161YCQD4+PhQREaEzBkNs2rSJxGIxASBPT0/6+eefae3atSSXywkAOTs709dff0379u1TL8POzo727t3bYFxdunQhDodD9vb2dO7cOSIimjVrFllZWREAksvldOnSJaP2SVhYGPF4PGrdujVxuVySyWQ0cuRIyszM1Mhn6PY3JJ+ufbxlyxZycXEhACQWi2n48OEUGxur3pZt27alzMxM2rFjB8lkMgJAHh4e6sedDKlbREQpKSn0yiuvkKurqzqfi4sLBQYG0unTp9X5iouLaerUqdSqVSuysbGhfv36UWRkJAEgNzc3+u2334zazkREs2fPpkmTJhk9X13muJvX1O1taL3Jy8ujgQMHklAoJC8vL/rwww9p3rx5BIB8fX3Vj9FcuXKFPDw8SCQSUb9+/ejJkyd07NgxkkqlFBUVZdK6EjXueBUeHk48Ho8UCoU67dChQ+Tj40MAyMHBQX33bl3z5s3TejTGkLZhzD6oqKigBQsWkLu7O3G5XHJ0dKTRo0dTeno6ERGFhIQQAIqMjNS7jqq7W/X91X3E5fTp09SrVy8SCATk6upK8+bNo/Lycq1yDcln7vjmzJlDPj4+JJFIiMvlkpubG7377rv0+PFjrXIfPHhAEyZMIDs7OxIIBNSrVy9KTk7WGcPQoUOpdevWOh+b0afJH42JjY2liIgIjbSKigqaNWsWCQQCjUrLNI+wsDCyt7dv6TBM9neqW5bwOsEXrd405nh19+5d4nK5Br2izhJVV1dT//79adeuXS0dik6WHh8R0bNnz0goFNL69euNmq++ztTka6ZPnjxBeHi41hg/n8+Hu7s7lEollEql2S7yMoZrqseBmgurWy3jRa83DfH19cXKlSuxcuVKjBw5ssW+7tMY1dXVSEpKQnFxMUJDQ1s6HC2WHp/K8uXL0aVLF4SHh5utTJOvmYpEIvB4POzatQtPnz6FUqnE48ePsXPnTkRGRiI0NBQymcwcsZrdrVu39H4+q/afJVSKFylWc2mJuvV33M5/R4sWLcLYsWMRGhpq9M1ILenUqVM4ePAgkpOTDX5WtjlZenwAsGHDBly7dg3Hjh0z7/vejTiN1evMmTM0aNAgkslkZG1tTXK5nAIDAyk2Ntbot/Izplu0aBHx+Xz19d2EhISWDqnR/k51q6WHeV/EemPqV65OnDhBCxYsMGNEjCVLTEyk6Ohora/VGKq+YV4OkebrZeLj4zF+/Hj2TUSGaWaq7zgmJCS0cCQvDna8YppTPfUtodk+Ds4wDMMwf1WsM2UYhmEYE7HOlGEYhmFMxDpThmEYhjER60wZhmEYxkR6X9pgCS+IZ5i/I9b2jMe2GdPS9Ham+/fvb844GOZvb+PGjQDwl3lhfnNISUnBpk2b2PGKaRaq+qaL3s503LhxTRYQwzDaVM+XsrZnnE2bNrFtxjQbfZ0pu2bKMAzDMCZinSnDMAzDmIh1pgzDMAxjItaZMgzDMIyJWGfKMAzDMCayiM70zp07+PDDD9GxY0fIZDLw+Xw4OjrCz88Po0aNwrfffqvOu379ejg5OYHD4eDTTz81ajk9e/aEtbU1unTpYlK8umI4ePAgvL291d+7XLp0ab1lbNiwARwOB1ZWVvDz88OZM2dMiqm2urFwOBzweDy0bt0aEydOxM2bN822rLr07Z9jx45BLpfju+++a7Jlq9TU1GDjxo0IDAzUmqZr23A4HPD5fDg5OSEoKAgxMTHIz89v8jiZv56TJ09i0aJFWvXsrbfe0so7ePBgSKVSWFtbo2PHjrhy5UoLRGy4lStXwt/fHzKZDAKBAL6+vpg/fz5KSko08kVFRen8BnCnTp20ylQqlYiOjoavry/4fD5sbW3RqVMn3Lt3DwBw+PBhrFu37sX4YEkpxY8AACAASURBVL0R32trEnFxccTn86lfv350/Phxys/Pp/LycsrMzKTvvvuOhg4dSmFhYRrz3L17lwDQ9u3bjV5ecHAwvfzyyybHrS8GHx8fAkAuLi5UWVmpc96qqiry8PAgABQcHGxyLPr4+PiQXC4nIqKSkhI6fPgwubu7k42NDd26davJlqtr2xw5coRkMhkdPny4yZZLRHTnzh165ZVXCEC9+7n2tqmpqaH8/Hz66aefaPLkycThcMjV1ZUuXrzYpLHW1dLfM30RNffxqj6RkZE0bNgwKioqUqf5+PhQq1atCAAdOXJEa57k5GQaMWJEc4bZaAMGDKDY2FjKy8ujoqIi2r9/P/F4PPrnP/+pkW/VqlUEQOuvY8eOWmWGhIRQ+/bt6fz586RUKunx48c0fPhwSktLU+fZtGkTDRgwgPLz85t8HRtS3/dMW/TM9Pz585g6dSoCAwPx008/4bXXXoOtrS0EAgG8vb3xxhtvYPPmzWZfblO/LaV79+548uQJEhMTdU4/ePAgWrdu3aQx1CWRSDBs2DD85z//QUlJCbZs2dKsyx86dCgKCwsxbNiwJlvGb7/9hoULF+L99983avSBw+HA1tYWQUFBiIuLQ3x8PJ4+faqO+e+krKxM5xn9i7aM5rZ27Vrs27cP8fHxkEqlGtM2b94MKysrhIWFvdD1ycbGBmFhYbC3t4dUKsW4ceMQEhKC48eP48GDBxp5d+/eDSLS+Lt+/bpGnn379iExMREJCQno3bs3uFwuXF1dkZSUpHEWO3PmTLz88ssYMmQIqqqqmmVdG6NFO9PVq1ejuroaa9asAZer+/0R3t7eRg/nNoTH45m1vLqmT58OANi+fbvO6Rs2bMCcOXOaNAZ9evXqBQBaFftFQ0RISEjAjh071Gkvv/wyDh48iIkTJ0IgEDS67DFjxmDy5MnIyckxe92zdLt27UJOTs4Lv4zmlJGRgaVLl2LFihUQCoVa0wMDAxEREYFHjx5h7ty5LRCheRw5cgTW1tYaaQ4ODgAAhUJhdHnbt29Ht27dEBAQ0GDe5cuX49q1a3pfmGAJWqwzraysxMmTJ2Fvb48+ffqYXB4RYcOGDejQoQMEAgHs7OwwcuRI3Lp1SytvRkYG/Pz8IJFIIBKJ0L9/f5w9e1Yjz88//wx/f3/I5XIIhUIEBATgxIkTBsXy6quvokOHDvjpp59w+/ZtjWm//PILFAoFBg8erHPe+pb75ZdfwsbGBhwOB3Z2dkhMTMSlS5fg4eEBa2trvPnmmw3GpvplV7uzMXTbGbONazt79izc3d3B4XCwdetWAMC2bdsgkUggFouRlJSE119/HTKZDG5ubti7d6/G/NXV1YiOjkb79u0hEong4OAALy8vREdHN9mbbyZPngwASE5ObpLyzcWQfRIeHg4+nw8XFxd12gcffACJRAIOh4Nnz54BACIiIjBnzhxkZmaCw+HA19cXmzdvhlAohJOTE6ZNmwZXV1cIhUIEBgbiwoULZlkGABw/fhwymQyrV69u0u3VFDZv3gwiwvDhw/XmiYqKQrt27bBz506cPHmy3vIM2afGtp/IyEi4u7tDJBKhc+fOZnv94qNHjyASieDl5WXUfJWVlTh//rzBI0h2dnYYMGAANm3aBCJqTKhNz4gxYbO6c+cOAaA+ffoYPa+ua3KRkZHE5/Np9+7dVFBQQKmpqdStWzdycHCgJ0+eqPMFBweTt7c3ZWVlkVKppOvXr1Pv3r1JKBTSnTt31PkSEhJo+fLl9Pz5c8rLy6M+ffpQq1at6o2B6M9rJFlZWfSf//yHAFBERITG9JCQEIqLi6Pi4mKd10wbWu6NGzdILBbT22+/rU5btGgR7dy5U2s71b4uqLJ7924CQPPmzTN62xmaT9e2efDgAQGgLVu2qNM++ugjAkA//PADFRYWUk5ODvXv358kEonG9ebVq1eTtbU1JSUlkUKhoMuXL5OzszMFBQVprbNK7969Db5mqktRUREBoDZt2ujNY26NuWZq6D6ZOHEiOTs7a8wbExNDACg3N1edNnr0aPLx8dHIFxYWRhKJhG7cuEHl5eWUnp5OPXv2JKlUStnZ2WZZxpEjR0gqldLKlSuNWn9LuGbq7e1N/v7+OqepjgdEROfOnSMrKyvy9PSkkpISItJ9zdTQfWpo+5k7dy4JBAI6cOAA5efn0+LFi8nKysrkewJKS0tJKpVSeHi4RvqqVavIzc2NbG1ticfjkaenJ40YMYJ+/fVXdZ6srCwCQF26dKGgoCBycXEhgUBAfn5+tHXrVqqpqdFa3qJFiwgAXb161aS4TVHfNdMW60wvXbpEAGjQoEFGz1v3YK1QKMjGxoZCQ0M18v36668EQKOB6roBKTU1lQDQ3Llz9S4zOjqaAFBOTo7OGFRUjaegoIAkEgnZ2dmRQqEgIqLMzExyc3OjiooKvZ1pQ8slIvrss88IAO3Zs4e++eYbmj17ts55696AdODAAXJ2diYnJyd6+PChUdvOmG1sbGdaVlamTouNjSUAlJGRoU7r2bMn9erVS2O57733HllZWVFFRYXOdTe1MyUi4nA4ZGtrW28eczK2MzVmn5jamdbdVhcvXiQAtGLFCrMso7FaujMtKSkhDodDw4YN0zm9dmdKRDRnzhwCQDNmzCAi7c7UmH1qSPspKysjsVisUZ5CoSCBQEDTp09v/Ir///LbtWunccMVEVF2djZduXKFiouLqaKiglJSUqhr164kEono+vXrRESUlpZGAOgf//gH/fLLL5SXl0cFBQW0cOFC9bGtri+++IIA0FdffWVS3KawyBuQbGxsAAClpaU6p8fHx8PLy0t9W3WHDh30XmdJT09HSUkJevTooZHes2dP8Pl8jeEoXQICAiCXy5Gamqo3j+o6q6G3aMvlcrz55pvIz8/Hvn37APz5VZDp06eDz+cbVIa+5b733nsYM2YMpk2bhvj4eHz88cd65y8sLASHw4FcLsfMmTMxZMgQ/Prrr+oboAzddqZuY0Opto1SqVSnlZeXaw3tVFdXg8fjaV3DMZfS0lIQEWQyWZOUbw7NtU906dGjB8RicYND/H91OTk5ICKIxWKD8kdFRaF9+/aIjY3VurQEmL5P67af27dvQ6FQaNzQIxKJ4OLiYtK+O3ToEOLj43HixAmtG67atGmDrl27wsbGBnw+H3369EFcXBzKysoQGxsL4L+XmTp27IjAwEDY29tDLpdjxYoVkMvlGvdCqKi28dOnTxsdd1Nqsc7Uw8MDAoEAGRkZOqePGzcOWVlZ8PDwgLOzM27evAknJyedeQsKCgD8t4OuzdbWFsXFxQ3Gw+PxNA7gR48eRVBQEBwdHSEQCDB//nxDVkuD6kakTz/9FAUFBUhISMC0adPqncfQ5a5evRolJSUN3sghl8tBRKiqqsLDhw/xxRdfwMPDQz3d0G1njm3cWEOGDMHly5eRlJSEsrIyXLp0CYmJiXjjjTearDO9c+cOAMDPz69JyjeHltwnwJ8HxNzc3CZdhqUrLy8HAINveBMKhYiLiwOHw8E777yDsrIyjenm3qeqk5UlS5ZoPPN5//79Rt00BPx5F+7atWtx6tQpeHp6GjRPQEAArK2t1e3K1dUVANTX0lX4fD48PDyQmZmpVYZIJALw321uaVqsMxUKhRg0aBByc3Nx/vx5k8qytbUFAJ0VraCgAG5ubvXOX1VVhefPn8Pd3R0AkJ2djZCQELi4uODChQsoLCzEunXrjI6rS5cu/8fevYdFcaX5A/82dEPT0M1FERAEBUwMXteoUdREx4yzxomKeMFoEpM1Q1wN4oWgYIxBNBoc9dHIZjUuu6sZFcVFY8Rk1cVdV8dJfmokOBpDxGsUMCpX5fb+/sjSk7a5NBTQDX4/z8MfVJ06561Tp/vtqq5TjcGDB+Mvf/kLIiMjMXnyZLi7u9dZ3tJ2KyoqMG/ePKxbtw6nTp1CYmJio2OrYWnfKe1jJZYvX47f/OY3mDlzJgwGAyZOnIgpU6Zg69atLdbm4cOHAQBjxoxpsTaUsuYxqaioaPE22oKaN/jGPFRgyJAhWLBgAS5fvowVK1aYrGvuY+rp6Qngl6ti8thUlVOnTjWqLgDYtGkTduzYgWPHjqFz584Wb1ddXY3q6mrjhw4XFxd0794dFy5cMCtbWVkJV1dXs+Xl5eUA/tbntsaqU2M++OADaDQaxMTEmJwVNlavXr3g4uKCb775xmT56dOnUV5ejmeffbbe7f/rv/4L1dXV6N+/PwAgKysLFRUV+Md//EcEBgZCq9U2eW5qzdnp3r17G/zRZ0vbfeedd/DWW29h/vz5WLBgAVasWNGkFwZged8p7WMlsrOzkZOTg/z8fFRUVODatWtITk6u94OJErdv38b69evh5+eHN998s0XaaA6NOSZqtVrRa+xxmZmZEBGTO/Gbu422oOZpX42dP7pixQr06NEDZ8+eNVne3K+zLl26QKvV4ty5c43a7nEigtjYWGRlZSE9Pb3WM+cav/vd78yWff311xARDBkyxLhs6tSpOHv2LH788UfjstLSUly9erXW6TI1fezl5aVkV1qMVZPps88+i+3bt+P//b//hxEjRuDw4cP46aefUFlZiatXr2L79u34+eefG6xHq9Vi4cKF2LdvH3bs2IHCwkJkZWVh9uzZ8PHxQWRkpEn58vJyPHjwAJWVlThz5gyioqIQEBBgnA5Rc4Z65MgRPHz4EJcvX27y909TpkxBx44dERYWhsDAwHrLWtLu5s2b4evri4kTJwIAVq1ahZCQEEyfPh2FhYWNjs/SvmtsHzenuXPnwt/f3+yxZUqJCIqLi1FdXQ0RQX5+Pnbv3o2hQ4fC3t4e6enpNv2daWOOSXBwMH7++Wekp6ejoqIC+fn5uHr1qlmdHh4euHXrFnJzc1FUVGRMjtXV1bh37x4qKytx/vx5REdHw9/f3/iaUdpGRkZGm5wao9PpEBgYiBs3bjRqu5rLvY9/TdHcrzOtVos33ngDO3fuRHJyMgoLC1FVVYUbN27gp59+AgBERETAy8ur3scZXrhwAR999BG2bt0KjUZj9qjAtWvXGsvevHkTu3btwv3791FRUYFTp05h1qxZ8Pf3x+zZs43lFixYYHzfvXbtGu7evYvY2FiUlZVh8eLFZjHU9LEl81KtohF3K7WYK1euSHR0tPTs2VOcnZ1Fq9VKt27dZPjw4bJ48WL57//+b2PZP/7xj+Ll5SUAxNnZWSZOnCgivzwSLikpSbp37y4ajUbc3d0lLCxMLl26ZNJWSkqKjBw5Ujp16iRqtVo6dOgg06ZNk6tXr5qUi42NFQ8PD3Fzc5PJkyfLxx9/LAAkKChIoqOjzWLYt2+f8VGCHTt2NN6tJyLy7rvvysmTJ43/L126VLy9vQWA2NnZSUhIiPzP//xPg+3269dPVCqVeHh4GOubP3++2NnZCQBxdXWVb775Rv73f/9XnnrqKeNjvHx8fGTy5Ml19r+lfWdJudqOz6ZNm4z7q9PpZNy4cbJ582bR6XQCQLp37y45OTmyZcsWMRgMAkACAgKMU5WOHTtmfCRbzZ9Go5FnnnlG0tLSjG2fOnVKhg4dKj4+PsZy3t7eEhoaKsePHxcRkQMHDkifPn1Ep9OJg4ODse9q7twdNGiQJCQkyN27d+vsr5bSlKkxlh67u3fvysiRI42vrXfeeUdiYmIEgAQHBxunuJw5c0YCAgLEyclJhg0bJrdv35bIyEjRaDTi6+srarVaDAaDTJgwQXJycpqtjUOHDoler5fExMRG7b+17+YVEYmKihKNRmO8a19E6n0/+LWYmBizqTGWHNPGvH4ePXoksbGx4u/vL2q1Wjw9PSU8PFyys7NF5JfpegBk2bJlde5jzd23df0lJSUZyy5cuFCCgoLE2dlZ1Gq1+Pn5yVtvvSW3bt0yq/f69esybdo0cXd3F0dHRxk0aJBkZGTUGsPYsWPF19e31mkzrcUmp8YQWWrz5s1m83UfPXok8+fPF0dHR5M3sbbMVp/NGxkZKR4eHtYOo1a28H51+fJlUavVsn37dqvG0VRVVVUyfPhw2bZtm7VDqVNBQYFotVpZu3atVeOwyakxRJa4ffs2oqKi8A//8A8myx0cHODv74+Kioon7ns6a2gTv9phJcHBwUhISEBCQkKzfxXR0qqqqpCeno6ioiJERERYO5w6LV++HP369UNUVJS1Q6kTkynZNCcnJ2g0Gmzbtg137txBRUUFbt26hU8//RTLli1DRESETX+vSU+GJUuWYPLkyYiIiGhTD7PPzMxEWloaMjIyLJ4r29rWrVuHc+fO4dChQy3+XHUlmEzJprm6uuKrr77Cd999h6eeegpOTk4ICQlBSkoKVq9ejX/7t3+zdojtWlxcHFJSUvDgwQN069YNe/futXZINmvlypWIiorChx9+aO1QLDZq1Ch89tlnJs9UtiX79+/Ho0ePkJmZ2WJ37zeX2n+qhciGDB8+HP/5n/9p7TCeSKtWrcKqVausHUabMXr06Dp/xIIab/z48Rg/fry1w7AIz0yJiIgUYjIlIiJSiMmUiIhIISZTIiIiheq8ASk1NbU14yB64tU8Lo2vPcvVPJOafUatob5noKtETH8oMjU1FVOnTm3xoIiIiNqix9ImAOwxS6ZEZDtqPtzyZUpk0/bwO1MiIiKFmEyJiIgUYjIlIiJSiMmUiIhIISZTIiIihZhMiYiIFGIyJSIiUojJlIiISCEmUyIiIoWYTImIiBRiMiUiIlKIyZSIiEghJlMiIiKFmEyJiIgUYjIlIiJSiMmUiIhIISZTIiIihZhMiYiIFGIyJSIiUojJlIiISCEmUyIiIoWYTImIiBRiMiUiIlKIyZSIiEghJlMiIiKFmEyJiIgUYjIlIiJSiMmUiIhIISZTIiIihZhMiYiIFGIyJSIiUojJlIiISCEmUyIiIoXU1g6AiH5x48YNvP7666iqqjIuu3fvHvR6PUaMGGFS9umnn8Y///M/t3KERFQXJlMiG+Hn54erV68iJyfHbN3x48dN/n/++edbKywisgAv8xLZkNdeew0ajabBchEREa0QDRFZismUyIZMnz4dlZWV9Zbp2bMnQkJCWikiIrIEkymRDQkKCkKfPn2gUqlqXa/RaPD666+3clRE1BAmUyIb89prr8He3r7WdZWVlZg8eXIrR0REDWEyJbIx06ZNQ3V1tdlyOzs7DB48GF27dm39oIioXkymRDbGx8cHQ4cOhZ2d6cvTzs4Or732mpWiIqL6MJkS2aBXX33VbJmIYOLEiVaIhogawmRKZIMmTZpk8r2pvb09XnzxRXTq1MmKURFRXZhMiWyQu7s7fvvb3xoTqohgxowZVo6KiOrCZEpko2bMmGG8EUmj0WDChAlWjoiI6sJkSmSjxo0bB0dHRwDAyy+/DBcXFytHRER1YTIlslHOzs7Gs1Fe4iWybSoREWsH0Z6lpqZi6tSp1g6DiJ5gfJtvcXv4qzGtZPfu3dYO4Ymzfv16AMD8+fOtHEnTVVVVYffu3XjllVdapb1Tp05hw4YNHK/tRM3xpJbHZNpKpkyZYu0Qnjh79uwB0Pb7PiwsDFqtttXa27BhQ5vvM/obJtPWwe9MiWxcayZSImoaJlMiIiKFmEyJiIgUYjIlIiJSiMmUiIhIISbTNmDWrFnQ6/VQqVQ4d+6ctcNRpLq6GuvXr0doaKhF5R8+fIgePXpg6dKlLRxZ3Q4dOgRXV1d8/vnnVouhrThy5AiWLFmCtLQ0BAYGQqVSQaVS1forOKNHj4Zer4e9vT169uyJM2fOWCFiyyUkJCAkJAQGgwGOjo4IDg7Gu+++i+LiYpNyiYmJxv3+9V+vXr3M6qyoqMCqVasQHBwMBwcHuLm5oVevXsjNzQUAHDhwAGvWrEFVVVVr7CIpwGTaBnz66afYunWrtcNQ7PLly3j++eexYMEClJaWWrRNfHw8Ll261MKR1Y8T3i3z/vvvY+PGjYiLi0N4eDh+/PFHBAUFoUOHDtixYwe++OILk/JfffUV9uzZg5dffhnZ2dno37+/lSK3zLFjxzB37lzk5uaioKAAq1atwoYNGzB58uQm1zl16lT8+7//Oz777DOUlpbir3/9K4KCgowJety4cdBqtRg1ahTu37/fXLtCLYDJlFrFt99+i8WLF2P27Nno16+fRducPHkS3333XQtH1rCxY8fiwYMHePnll60dCsrKyiw+q29Nq1evxq5du5Camgq9Xm+ybuPGjbCzs0NkZCQePHhgpQiVc3FxQWRkJDw8PKDX6zFlyhSEhYXh8OHDuH79uknZ7du3Q0RM/h4fy7t27UJ6ejr27NmD5557Dmq1Gj4+Pti/f7/JWey8efPQt29fvPTSS6isrGyVfaXGYzJtI1QqlbVDUKRv375IS0vD9OnTjQ9vr09ZWRliYmI44fwx27ZtQ15enrXDMPHDDz/gvffewwcffFDrnNjQ0FBER0fj5s2bWLRokRUibB4HDx40+Y1ZAOjYsSMAWHyl5df+6Z/+Cf3790fv3r0bLLt8+XKcO3eOrwcbxmRqg0QESUlJePrpp+Ho6AhXV1fExMSYlauqqsKyZcvg7+8PJycn9OnTx/gYuOTkZDg7O0On02H//v0YM2YMDAYD/Pz8sHPnTpN6jh8/jkGDBkGn08FgMKB3794oLCxssI2WFB8fjzlz5sDT07PF26rPiRMn4O/vD5VKhY8//hiA5X27ceNGaLVadOrUCW+//TZ8fHyg1WoRGhqK06dPG8tFRUXBwcEB3t7exmVz5syBs7MzVCoVCgoKAADR0dFYuHAhcnJyoFKpEBwcDAA4fPgwDAYDVq5c2RpdYmbjxo0QEYwbN67OMomJiXjqqafw6aef4siRI/XWJyJYt24dnnnmGTg6OsLd3R0TJkzAxYsXjWUaM75bcgzfvHkTTk5O6NatW6O2Ky8vx5///GeLr9K4u7vjhRdewIYNG/i1g60SalG7d++WxnZzfHy8qFQq+eMf/yj37t2T0tJS2bx5swCQs2fPGsstWrRIHB0dZe/evXLv3j2Ji4sTOzs7+frrr431AJCjR4/KgwcPJC8vT4YPHy7Ozs5SXl4uIiLFxcViMBhkzZo1UlZWJrdv35aJEydKfn6+RW00xXPPPSd9+/atc/2JEydk3LhxIiKSn58vACQ+Pr7R7UyaNEkmTZrU5DhrXL9+XQDIpk2bjMss6VsRkcjISHF2dpYLFy7Iw4cPJTs7WwYOHCh6vV6uXbtmLDd9+nTx8vIyaTcpKUkAGI+FiEh4eLgEBQWZlDt48KDo9XpJSEhQvK9NGa+BgYESEhJS67qgoCC5cuWKiIicPHlS7OzspGvXrlJcXCwiIhkZGTJ+/HiTbZYtWyYODg6yfft2uX//vpw/f1769+8vHTt2lNu3bxvLWXoMWmIMi4iUlJSIXq+XqKgok+UrVqwQPz8/cXNzE41GI127dpXx48fLX/7yF2OZK1euCADp16+fjBgxQry9vcXR0VF69OghH3/8sVRXV5u1t2TJErP3gIY05XhSk6TyzNTGlJWVYf369XjxxRexYMECuLm5wcnJCR4eHiblHj58iOTkZISFhSE8PBxubm5YunQpNBoNUlJSTMqGhobCYDDA09MTERERKCkpwbVr1wAAubm5KCwsRM+ePaHVauHl5YW0tDR07NixUW005/5HR0cjOTm5RepvbvX1bQ21Wm08ywoJCUFycjKKioqarQ/Hjh2LwsJCvPfee81SX2OUlJTgypUrCAoKarDskCFDMH/+fOTm5mLx4sW1likrK8O6deswceJEzJgxA66urujduzc++eQTFBQUYMuWLWbb1HcMWnIMr1q1Cj4+PkhMTDRZ/vrrr+PAgQO4fv06iouLsXPnTly7dg0vvPACsrOzAcB4g5GnpydWrlyJ7Oxs3LlzBxMmTMDcuXPxpz/9yay97t27AwCysrIUxU0tg8nUxvzwww8oLS3FqFGj6i136dIllJaWmtyo4OTkBG9vb5PLYY9zcHAA8Mst+QAQGBiITp06YcaMGVi+fLnxlnwlbSgRFxeHP/zhD/D19W2R+lvS431blwEDBkCn07VYH7amvLw8iAh0Op1F5RMTE/H0009j8+bNOHHihNn67OxsFBcXY8CAASbLBw4cCAcHB5PL47V5/Bi01Bjet28fUlNT8eWXX5rdcNWlSxf83d/9HVxcXODg4IDBgwcjJSUFZWVl2Lx5MwAY7xvo2bMnQkND4eHhAVdXV3zwwQdwdXWt9UNDTR/fuXOnyXFTy2EytTE3btwAgAa/KywpKQEALF261GQu29WrVxt1M4STkxOOHTuGYcOGYeXKlQgMDERERATKysqarQ1LnThxAllZWZg1a1az121rHB0dkZ+fb+0wFHv48CEAWHRTGfDLQ/tTUlKgUqnw5ptvoqyszGR9zfQPFxcXs23d3NxQVFTUqPhaYgzv2rULq1evRmZmJrp27WrRNr1794a9vT2+//57AICPjw8AGL8Pr+Hg4ICAgADk5OSY1eHk5ATgb31OtoXJ1MbU3A356NGjesvVJNv169eb3YJ/6tSpRrXZs2dPfP7557h16xZiY2Oxe/durF27tlnbsMS2bdtw9OhR2NnZGd/0amJYuXIlVCoVvvnmm2Zvt7VVVFTg/v378PPzs3YoitW8wTfmoQJDhgzBggULcPnyZaxYscJknZubGwDUmjSb0mfNPYY3bdqEHTt24NixY+jcubPF21VXV6O6utr4ocPFxQXdu3fHhQsXzMpWVlbC1dXVbHl5eTmAv/U52RYmUxvTq1cv2NnZ4fjx4/WW69KlC7RareInIt26dcv4gvb09MSHH36I/v3748KFC83WhqVSUlLM3vBqzt7i4+MhImaX/9qizMxMiAgGDx5sXKZWqxu8PGyLOnXqBJVK1ej5oytWrECPHj1w9uxZk+W9evWCi4uL2Yem06dPOaaaagAAIABJREFUo7y8HM8++2yj2mmuMSwiiI2NRVZWFtLT02s9c67xu9/9zmzZ119/DRHBkCFDjMumTp2Ks2fP4scffzQuKy0txdWrV2udLlPTx15eXkp2hVoIk6mN8fT0RHh4OPbu3Ytt27ahsLAQ58+fN/sORavV4o033sDOnTuRnJyMwsJCVFVV4caNG/jpp58sbu/WrVt4++23cfHiRZSXl+Ps2bO4evUqBg8e3GxtPOmqq6tx7949VFZW4vz584iOjoa/vz9mzpxpLBMcHIyff/4Z6enpqKioQH5+Pq5evWpWl4eHB27duoXc3FwUFRWhoqICGRkZVpsao9PpEBgYaPx6wlI1l3sfn7ep1WqxcOFC7Nu3Dzt27EBhYSGysrIwe/Zs+Pj4IDIystHtNDSGIyIi4OXlVe/jDC9cuICPPvoIW7duhUajMXtU4Nq1a41lb968iV27duH+/fuoqKjAqVOnMGvWLPj7+2P27NnGcgsWLEBAQABmzpyJa9eu4e7du4iNjUVZWVmtN2jV9LEl81LJClr15uEnUFNuTS8qKpJZs2ZJhw4dxMXFRYYNGybLli0TAOLn5yfffvutiIg8evRIYmNjxd/fX9RqtXh6ekp4eLhkZ2fL5s2bRafTCQDp3r275OTkyJYtW8RgMAgACQgIkO+//15yc3MlNDRU3N3dxd7eXjp37izx8fFSWVnZYBuNcerUKRk6dKj4+PgIAAEg3t7eEhoaKsePH69zO2tPjdm0aZN4e3sLANHpdDJu3DiL+1bkl6kxGo1GfH19Ra1Wi8FgkAkTJkhOTo5JO3fv3pWRI0eKVquVbt26yTvvvCMxMTECQIKDg43TaM6cOSMBAQHi5OQkw4YNk9u3b8uhQ4dEr9dLYmKion0Vadp4jYqKEo1GI6WlpcZl+/btk6CgIAEgHTt2lLlz59a6bUxMjNnUmOrqaklKSpLu3buLRqMRd3d3CQsLk0uXLhnLNOYYNDSGw8LCBIAsW7aszn3Mysoyjtva/pKSkoxlFy5cKEFBQeLs7CxqtVr8/Pzkrbfeklu3bpnVe/36dZk2bZq4u7uLo6OjDBo0SDIyMmqNYezYseLr61vrtJm6cGpMq0llL7cwDmbraa55pkpERkaKh4eHVWNojKaM18uXL4tarZbt27e3UFQtq6qqSoYPHy7btm2zdih1KigoEK1WK2vXrm3Udnz/aTWcZ0rU0tr7L34EBwcjISEBCQkJZr+gYuuqqqqQnp6OoqIiREREWDucOi1fvhz9+vVDVFSUtUOhOjCZUpNcvHix1p+ZevzPlt+gqPksWbIEkydPRkRERJt6mH1mZibS0tKQkZFh8VzZ1rZu3TqcO3cOhw4dgkajsXY4VAcmU2qSHj16mN15W9vfrl27rB2q1cTFxSElJQUPHjxAt27dsHfvXmuH1KJWrlyJqKgofPjhh9YOxWKjRo3CZ599ZvJcZFuyf/9+PHr0CJmZmXB3d7d2OFQPtbUDIGqvVq1ahVWrVlk7jFY1evRojB492tphtBvjx4/H+PHjrR0GWYBnpkRERAoxmRIRESnEZEpERKQQkykREZFCvAGplaSmplo7hCdOzePX2PeWq3n4O/usfWiJH6Sg2qlERKwdRHuWmpqKqVOnWjsMInqC8W2+xe3hmWkr4WBufZMnTwYA7Nmzx8qRtB01H/44XtsHfphvPfzOlIiISCEmUyIiIoWYTImIiBRiMiUiIlKIyZSIiEghJlMiIiKFmEzbgbS0NAQGBpr9lqiDgwM6deqEESNGICkpCffu3bN2qNTOHTlyBEuWLDEbk6+++qpZ2dGjR0Ov18Pe3h49e/bEmTNnrBBx41VXV2P9+vUIDQ2tdX1CQgJCQkJgMBjg6OiI4OBgvPvuuw3+cPrDhw/Ro0cPLF261LjswIEDWLNmTbv/gfn2gMm0HQgPD8ePP/6IoKAguLq6QkRQXV2NvLw8pKamolu3boiNjUXPnj3xzTffWDtcaqfef/99bNy4EXFxcSZjskOHDtixYwe++OILk/JfffUV9uzZg5dffhnZ2dno37+/lSK33OXLl/H8889jwYIFKC0trbXMsWPHMHfuXOTm5qKgoACrVq3Chg0bjPOe6xIfH49Lly6ZLBs3bhy0Wi1GjRqF+/fvN9t+UPNjMm2nVCoV3NzcMGLECKSkpCA1NRV37tzB2LFj8eDBA2uH98QoKyur8wymLbXRkNWrV2PXrl1ITU2FXq83Wbdx40bY2dkhMjKyTY+9b7/9FosXL8bs2bPRr1+/Osu5uLggMjISHh4e0Ov1mDJlCsLCwnD48GFcv3691m1OnjyJ7777rtZ18+bNQ9++ffHSSy+hsrKyWfaFmh+T6RNi0qRJmDlzJvLy8vDJJ59YO5wnxrZt25CXl9fm26jPDz/8gPfeew8ffPABtFqt2frQ0FBER0fj5s2bWLRokRUibB59+/ZFWloapk+fDkdHxzrLHTx4EPb29ibLOnbsCAC1ns2WlZUhJiYGGzZsqLPO5cuX49y5c/WWIetiMn2CzJw5EwCQkZFhXFZVVYVly5bB398fTk5O6NOnD3bv3g0ASE5OhrOzM3Q6Hfbv348xY8bAYDDAz88PO3fuNKn7+PHjGDRoEHQ6HQwGA3r37o3CwsIG27A1IoJ169bhmWeegaOjI9zd3TFhwgRcvHjRWCYqKgoODg7w9vY2LpszZw6cnZ2hUqlQUFAAAIiOjsbChQuRk5MDlUqF4OBgbNy4EVqtFp06dcLbb78NHx8faLVahIaG4vTp083SBgAcPnwYBoMBK1eubNH+An458xQRjBs3rs4yiYmJeOqpp/Dpp5/iyJEj9dZnyTFozNi0hfF38+ZNODk5oVu3bmbr4uPjMWfOHHh6eta5vbu7O1544QVs2LCBj3q0VUItavfu3dJa3RwUFCSurq51ri8sLBQA0qVLF+OyRYsWiaOjo+zdu1fu3bsncXFxYmdnJ19//bWIiMTHxwsAOXr0qDx48EDy8vJk+PDh4uzsLOXl5SIiUlxcLAaDQdasWSNlZWVy+/ZtmThxouTn51vURkuZNGmSTJo0qVHbLFu2TBwcHGT79u1y//59OX/+vPTv3186duwot2/fNpabPn26eHl5mWyblJQkAIz7LSISHh4uQUFBJuUiIyPF2dlZLly4IA8fPpTs7GwZOHCg6PV6uXbtWrO0cfDgQdHr9ZKQkNCo/W/KeA0MDJSQkJBa1wUFBcmVK1dEROTkyZNiZ2cnXbt2leLiYhERycjIkPHjx5tsY+kxsGRsirTM+Hvuueekb9++FpUtKSkRvV4vUVFRZutOnDgh48aNExGR/Px8ASDx8fG11rNkyRIBIGfPnrU4ztZ8/3nCpfLM9Ami1+uhUqlQVFQE4Je7B5OTkxEWFobw8HC4ublh6dKl0Gg0SElJMdk2NDQUBoMBnp6eiIiIQElJCa5duwYAyM3NRWFhIXr27AmtVgsvLy+kpaWhY8eOjWrD2srKyrBu3TpMnDgRM2bMgKurK3r37o1PPvkEBQUF2LJlS7O1pVarjWdeISEhSE5ORlFRUbP1ydixY1FYWIj33nuvWeqrS0lJCa5cuYKgoKAGyw4ZMgTz589Hbm4uFi9eXGuZphyD+samLYy/VatWwcfHB4mJiWb7Gh0djeTkZIvq6d69OwAgKyur2WMk5ZhMnyAlJSUQERgMBgDApUuXUFpail69ehnLODk5wdvb2+SS2uMcHBwAABUVFQCAwMBAdOrUCTNmzMDy5cuRm5trLNvUNqwhOzsbxcXFGDBggMnygQMHwsHBweQybHMbMGAAdDqdzfVJQ/Ly8iAi0Ol0FpVPTEzE008/jc2bN+PEiRNm65Ueg8fHprXH3759+5Camoovv/zS7MasuLg4/OEPf4Cvr69FddX08Z07d5o9TlKOyfQJ8v333wMAevToAeCX5AoAS5cuNZmfevXq1Tpv+6+Nk5MTjh07hmHDhmHlypUIDAxEREQEysrKmq2N1lAz9cDFxcVsnZubm/GMvqU4OjoiPz+/Rdtobg8fPgSAem/I+TWtVouUlBSoVCq8+eabKCsrM1nf3MfAmuNv165dWL16NTIzM9G1a1eTdSdOnEBWVhZmzZplcX1OTk4A/tbnZFuYTJ8ghw8fBgCMGTMGAIw3PKxfvx4iYvJ36tSpRtXds2dPfP7557h16xZiY2Oxe/durF27tlnbaGlubm4AUOsb9v379+Hn59dibVdUVLR4Gy2h5g2+MQ8VGDJkCBYsWIDLly9jxYoVJuua+xhYa/xt2rQJO3bswLFjx9C5c2ez9du2bcPRo0dhZ2dnTPA1sa5cuRIqlcpsTnh5eTmAv/U52RYm0yfE7du3sX79evj5+eHNN98EAHTp0gVarRbnzp1TVPetW7dw4cIFAL+8eX344Yfo378/Lly40GxttIZevXrBxcXF7E3s9OnTKC8vx7PPPmtcplarjZcSm0NmZiZEBIMHD26xNlpCp06doFKpGj1/dMWKFejRowfOnj1rsrwxx8ASrT3+RASxsbHIyspCenp6rWfYAJCSkmKW3GuuSsTHx0NEzC511/Sxl5dXy+4ENQmTaTsjIiguLkZ1dbXxBbp7924MHToU9vb2SE9PN35nqtVq8cYbb2Dnzp1ITk5GYWEhqqqqcOPGDfz0008Wt3nr1i28/fbbuHjxIsrLy3H27FlcvXoVgwcPbrY2WoNWq8XChQuxb98+7NixA4WFhcjKysLs2bPh4+ODyMhIY9ng4GD8/PPPSE9PR0VFBfLz83H16lWzOj08PHDr1i3k5uaiqKjImByrq6tx7949VFZW4vz584iOjoa/v79x+pLSNjIyMlplaoxOp0NgYCBu3LjRqO1qLvc+Ph+zMcfA0nYaGn8RERHw8vJqlscZXrhwAR999BG2bt0KjUZj9ojPtWvXNrnumj7u3bu34jipBbTircNPpNa4Nf3AgQPSp08f0el04uDgIHZ2dgJAVCqVuLm5yaBBgyQhIUHu3r1rtu2jR48kNjZW/P39Ra1Wi6enp4SHh0t2drZs3rxZdDqdAJDu3btLTk6ObNmyRQwGgwCQgIAA+f777yU3N1dCQ0PF3d1d7O3tpXPnzhIfHy+VlZUNttGSmjI1prq6WpKSkqR79+6i0WjE3d1dwsLC5NKlSybl7t69KyNHjhStVivdunWTd955R2JiYgSABAcHG6e4nDlzRgICAsTJyUmGDRsmt2/flsjISNFoNOLr6ytqtVoMBoNMmDBBcnJymq2NQ4cOiV6vl8TExEbtf1PGa1RUlGg0GiktLTUu27dvnwQFBQkA6dixo8ydO7fWbWNiYsymxlhyDCwdmyINj7+wsDABIMuWLat3P0+dOiVDhw4VHx8fASAAxNvbW0JDQ+X48eMiIpKVlWVcV9tfUlJSnfU3NDVm7Nix4uvrK9XV1fXG+WucGtNqUtnLLYyD2XqakkxbQ2RkpHh4eFg7jFo1ZbxevnxZ1Gq1bN++vYWiallVVVUyfPhw2bZtm7VDqVNBQYFotVpZu3Zto7bj+0+r4TxTImtoT78CEhwcjISEBCQkJDT4yyi2pqqqCunp6SgqKkJERIS1w6nT8uXL0a9fP0RFRVk7FKoDkykRKbZkyRJMnjwZERERbeph9pmZmUhLS0NGRobFc2Vb27p163Du3DkcOnQIGo3G2uFQHZhMiVpRXFwcUlJS8ODBA3Tr1g179+61dkjNZuXKlYiKisKHH35o7VAsNmrUKHz22Wcmz0C2Jfv378ejR4+QmZkJd3d3a4dD9VBbOwCiJ8mqVauwatUqa4fRYkaPHo3Ro0dbO4x2Y/z48Rg/fry1wyAL8MyUiIhIISZTIiIihZhMiYiIFGIyJSIiUog3ILWSyZMnWzuEJ86f//xnAOz7xqh5ZB37rH1o7GMeqelUIiLWDqI9O3XqFNatW2ftMKiNun37Ns6ePWv8pR+iptizZ4+1Q2jv9jCZEtmw1NRUTJ06FXyZEtm0PfzOlIiISCEmUyIiIoWYTImIiBRiMiUiIlKIyZSIiEghJlMiIiKFmEyJiIgUYjIlIiJSiMmUiIhIISZTIiIihZhMiYiIFGIyJSIiUojJlIiISCEmUyIiIoWYTImIiBRiMiUiIlKIyZSIiEghJlMiIiKFmEyJiIgUYjIlIiJSiMmUiIhIISZTIiIihZhMiYiIFGIyJSIiUojJlIiISCEmUyIiIoWYTImIiBRiMiUiIlKIyZSIiEghJlMiIiKFmEyJiIgUYjIlIiJSSG3tAIjoFxUVFSguLjZZVlJSAgC4d++eyXKVSgU3N7dWi42I6sdkSmQjfv75Z/j6+qKqqspsnYeHh8n/I0eOxLFjx1orNCJqAC/zEtkILy8vPP/887Czq/9lqVKpMG3atFaKiogswWRKZENeffXVBsvY29tj4sSJrRANEVmKyZTIhoSHh0OtrvvbF3t7e/z93/89OnTo0IpREVFDmEyJbIjBYMCYMWPqTKgighkzZrRyVETUECZTIhszY8aMWm9CAgAHBwf8/ve/b+WIiKghTKZENub3v/89dDqd2XKNRoOwsDA4OztbISoiqg+TKZGN0Wq1mDhxIjQajcnyiooKTJ8+3UpREVF9mEyJbNArr7yCiooKk2UGgwG//e1vrRQREdWHyZTIBr344osmD2rQaDSYNm0aHBwcrBgVEdWFyZTIBqnVakybNs14qbeiogKvvPKKlaMiorowmRLZqGnTphkv9Xp5eWHYsGFWjoiI6sJkSmSjQkND4evrCwB47bXXGnzMIBFZDx9038xu3LiBkydPWjsMaicGDhyImzdvokOHDkhNTbV2ONROTJkyxdohtDsqERFrB9GepKamYurUqdYOg4ioTnzbb3Z7eGbaQjhYbc/kyZMBAHv27LFyJI2zd+9eTJo0ySpt13w45HhuH/hhv+XwSxgiG2etREpElmMyJSIiUojJlIiISCEmUyIiIoWYTImIiBRiMiUiIlKIydQGzZo1C3q9HiqVCufOnbN2OFY1YsQIqFSqWv9cXFysEtOhQ4fg6uqKzz//3CrttyVHjhzBkiVLkJaWhsDAQOOxe/XVV83Kjh49Gnq9Hvb29ujZsyfOnDljhYgbr7q6GuvXr0doaGit6xMSEhASEgKDwQBHR0cEBwfj3XffRXFxcb31Pnz4ED169MDSpUuNyw4cOIA1a9bU+ePxZD1Mpjbo008/xdatW60dhs2z1rNqOefSMu+//z42btyIuLg4hIeH48cff0RQUBA6dOiAHTt24IsvvjAp/9VXX2HPnj14+eWXkZ2djf79+1spcstdvnwZzz//PBYsWIDS0tJayxw7dgxz585Fbm4uCgoKsGrVKmzYsME477ku8fHxuHTpksmycePGQavVYtSoUbh//36z7Qcpx2RKLa6srKzOT+0N0Wq1KCwshIiY/EVGRuLdd99t5kgtM3bsWDx48AAvv/yyVdr/NSV925JWr16NXbt2ITU1FXq93mTdxo0bYWdnh8jISDx48MBKESr37bffYvHixZg9ezb69etXZzkXFxdERkbCw8MDer0eU6ZMQVhYGA4fPozr16/Xus3Jkyfx3Xff1bpu3rx56Nu3L1566SVUVlY2y76QckymNkqlUlk7hGazbds25OXlNWnbw4cPm70ZX79+Hd999x1+85vfNEd4bZqSvm0pP/zwA9577z188MEH0Gq1ZutDQ0MRHR2NmzdvYtGiRVaIsHn07dsXaWlpmD59OhwdHessd/DgQdjb25ss69ixIwDUejZbVlaGmJgYbNiwoc46ly9fjnPnztVbhloXk6kNEBEkJSXh6aefhqOjI1xdXRETE2NS5qOPPoJOp4Ner0deXh4WLlwIX19fXLp0CSKCdevW4ZlnnoGjoyPc3d0xYcIEXLx40bj9xo0bodVq0alTJ7z99tvw8fGBVqtFaGgoTp8+bRZPQ/VFRUXBwcEB3t7exmVz5syBs7MzVCoVCgoKAADR0dFYuHAhcnJyoFKpEBwcrLi/Vq9ejXnz5imupylOnDgBf39/qFQqfPzxxwCA5ORkODs7Q6fTYf/+/RgzZgwMBgP8/Pywc+dO47aWHgOlfXv48GEYDAasXLmyNbrEzMaNGyEiGDduXJ1lEhMT8dRTT+HTTz/FkSNH6q3PkvFo6TEAgKqqKixbtgz+/v5wcnJCnz59sHv3bmU73Ug3b96Ek5MTunXrZrYuPj4ec+bMgaenZ53bu7u744UXXsCGDRv4tYOtEGpWu3fvlsZ2a3x8vKhUKvnjH/8o9+7dk9LSUtm8ebMAkLNnz5qUAyDz5s2TTZs2ycSJE+Wvf/2rLFu2TBwcHGT79u1y//59OX/+vPTv3186duwot2/fNm4fGRkpzs7OcuHCBXn48KFkZ2fLwIEDRa/Xy7Vr14zlLK1v+vTp4uXlZbIvSUlJAkDy8/ONy8LDwyUoKKhRfVKXGzduSEhIiFRVVTV620mTJsmkSZMUx3D9+nUBIJs2bTIuqzk2R48elQcPHkheXp4MHz5cnJ2dpby83FjO0mOgpG8PHjwoer1eEhISFO9rU8ZzYGCghISE1LouKChIrly5IiIiJ0+eFDs7O+natasUFxeLiEhGRoaMHz/eZBtLx6Olx2DRokXi6Ogoe/fulXv37klcXJzY2dnJ119/3aj9/LXnnntO+vbta1HZkpIS0ev1EhUVZbbuxIkTMm7cOBERyc/PFwASHx9faz1Lliwxe49oSFOOJ1kklWemVlZWVob169fjxRdfxIIFC+Dm5gYnJyd4eHjUuc3q1asxd+5cpKWlISAgAOvWrcPEiRMxY8YMuLq6onfv3vjkk09QUFCALVu2mGyrVquNn/BDQkKQnJyMoqIipKSkGONpTH2tbfXq1XjnnXds9rc9Q0NDYTAY4OnpiYiICJSUlODatWsmZRo6BkqNHTsWhYWFeO+995qlvsYoKSnBlStXEBQU1GDZIUOGYP78+cjNzcXixYtrLdOU8VjfMXj48CGSk5MRFhaG8PBwuLm5YenSpdBoNM3W/w1ZtWoVfHx8kJiYaLav0dHRSE5Otqie7t27AwCysrKaPUZqPNt8R3qC/PDDDygtLcWoUaOatH12djaKi4sxYMAAk+UDBw6Eg4OD2SXcxw0YMAA6nc54yUxpfS3p1q1bOHDgAGbOnGm1GBrDwcEBAFBRUVFvucePQVuWl5cHEYFOp7OofGJiIp5++mls3rwZJ06cMFuvdDw+fgwuXbqE0tJS9OrVy1jGyckJ3t7erdL/+/btQ2pqKr788kuzewHi4uLwhz/8wfiD8A2p6eM7d+40e5zUeEymVnbjxg0AqPf7kfrU3B5f25xLNzc3FBUVNViHo6Mj8vPzm62+lrJmzRq89dZbtd7U0tb9+hi0ZQ8fPgSAem/I+TWtVouUlBSoVCq8+eabKCsrM1nf3OOxpKQEALB06VKTOctXr16tc2pLc9m1axdWr16NzMxMdO3a1WTdiRMnkJWVhVmzZllcn5OTE4C/9TlZF5OpldUkhkePHjVpezc3NwCo9U3l/v378PPzq3f7iooKk3JK62spt2/fxp/+9Cf84z/+o1Xab0mPH4O2rOYNvjEPFRgyZAgWLFiAy5cvY8WKFSbrmns81nxoXb9+vdl0q1OnTjWqrsbYtGkTduzYgWPHjqFz585m67dt24ajR4/Czs7OmOBrYl25ciVUKhW++eYbk23Ky8sB/K3PybqYTK2sV69esLOzw/Hjx5u8vYuLi9kL7fTp0ygvL8ezzz5b7/aZmZkQEQwePLjR9anV6gYvYTaXNWvWYMaMGfV+l9xWPX4MgNbt2+bUqVMnqFSqRs8fXbFiBXr06IGzZ8+aLFc6vh/XpUsXaLXaVnuymIggNjYWWVlZSE9Pr/OpXSkpKWbJveZKRXx8PETE7FJ3TR97eXm17E6QRZhMrczT0xPh4eHYu3cvtm3bhsLCQpw/f97iG320Wi0WLlyIffv2YceOHSgsLERWVhZmz54NHx8fREZGmpSvrq7GvXv3UFlZifPnzyM6Ohr+/v7G7yEbU19wcDB+/vlnpKeno6KiAvn5+bh69apZjB4eHrh16xZyc3NRVFTU6CRx584d/Mu//Avmz5/fqO1sVUPHAFDWtxkZGVabGqPT6RAYGGj8+sJSNZd7H5+P2djxbUk7b7zxBnbu3Ink5GQUFhaiqqoKN27cwE8//QQAiIiIgJeXV7M8zvDChQv46KOPsHXrVmg0GrNHYq5du7bJddf0ce/evRXHSc3AKjcRt2NNufW8qKhIZs2aJR06dBAXFxcZNmyYLFu2TACIn5+ffPvtt7JmzRpxcnISANKlSxfZvn27cfvq6mpJSkqS7t27i0ajEXd3dwkLC5NLly6ZtBMZGSkajUZ8fX1FrVaLwWCQCRMmSE5Ojkk5S+u7e/eujBw5UrRarXTr1k3eeecdiYmJEQASHBxsnOpx5swZCQgIECcnJxk2bJjJdAZLLFiwQGbMmNGobWrTHFNjNm3aJN7e3gJAdDqdjBs3TjZv3iw6nU4ASPfu3SUnJ0e2bNkiBoNBAEhAQIB8//33ImL5MVDSt4cOHRK9Xi+JiYmK9lWkaeM5KipKNBqNlJaWGpft27dPgoKCBIB07NhR5s6dW+u2MTExZlNjLBmPjTkGjx49ktjYWPH39xe1Wi2enp4SHh4u2dnZIiISFhYmAGTZsmX17uepU6dk6NCh4uPjIwAEgHh7e0toaKgcP35cRESysrKM62r7S0pKqrP+hqbGjB07Vnx9faW6urreOH+NU2NaTCp7tZnZ8mCNjIwUDw8Pa4dhNc01z1SJtnYMmjKeL1++LGq12uQDX1tSVVUlw4cPl23btlk7lDoVFBSIVquVtWvXNmo7W35/auM4z/RJw1+bsL72fgyCg4ORkJCAhISEBn8ZxdZUVVUhPT0dRUVFiIiIsHY4dVq+fDn69es5LyDpAAAgAElEQVSHqKgoa4dC/4fJlFrVxYsX6/xJtV//2fIbGTVsyZIlmDx5MiIiItrUw+wzMzORlpaGjIwMi+fKtrZ169bh3LlzOHToEDQajbXDof/DZPqEiIuLQ0pKCh48eIBu3bph7969VomjR48eZnct1va3a9cuq8TXkmzlGLSWlStXIioqCh9++KG1Q7HYqFGj8Nlnn5k8F9mW7N+/H48ePUJmZibc3d2tHQ79ikqET0luTqmpqZg6dSofPm2Dan4/cs+ePVaOpO3geG5feDxbzB6emRIRESnEZEpERKQQkykREZFCTKZEREQKqa0dQHtVc7ML2Y4///nPAHhsGqPmkXXss/ahsY95JMvxzJSIiEghnpm2EE6/sD2cGtN4NVMp2GftQ83xpObHM1MiIiKFmEyJiIgUYjIlIiJSiMmUiIhIISZTIiIihZhMbVxaWhoCAwPNfqLMwcEBnTp1wogRI5CUlIR79+5ZO1QiM0eOHMGSJUvMxvGrr75qVnb06NHQ6/Wwt7dHz549cebMGStE3HjV1dVYv349QkNDa12fkJCAkJAQGAwGODo6Ijg4GO+++26Dv/X68OFD9OjRA0uXLjUuO3DgANasWdPufxO3LWIytXHh4eH48ccfERQUBFdXV4gIqqurkZeXh9TUVHTr1g2xsbHo2bMnvvnmG2uHS2T0/vvvY+PGjYiLizMZxx06dMCOHTvwxRdfmJT/6quvsGfPHrz88svIzs5G//79rRS55S5fvoznn38eCxYsQGlpaa1ljh07hrlz5yI3NxcFBQVYtWoVNmzY0OCDMOLj43Hp0iWTZePGjYNWq8WoUaNw//79ZtsPUo7JtA1SqVRwc3PDiBEjkJKSgtTUVNy5cwdjx45tUz/EXJeysrI6P+W3da2xb7bQf6tXr8auXbuQmpoKvV5vsm7jxo2ws7NDZGRkmx6v3377LRYvXozZs2ejX79+dZZzcXFBZGQkPDw8oNfrMWXKFISFheHw4cO4fv16rducPHkS3333Xa3r5s2bh759++Kll15CZWVls+wLKcdk2g5MmjQJM2fORF5eHj755BNrh6PYtm3bkJeXZ+0wWkRr7Ju1+++HH37Ae++9hw8++ABardZsfWhoKKKjo3Hz5k0sWrTIChE2j759+yItLQ3Tp0+Ho6NjneUOHjwIe3t7k2UdO3YEgFrPZsvKyhATE4MNGzbUWefy5ctx7ty5estQ62IybSdmzpwJAMjIyAAAfPTRR9DpdNDr9cjLy8PChQvh6+uLS5cuQUSwbt06PPPMM3B0dIS7uzsmTJiAixcvGuvbuHEjtFotOnXqhLfffhs+Pj7QarUIDQ3F6dOnTdq2pL6oqCg4ODjA29vbuGzOnDlwdnaGSqVCQUEBACA6OhoLFy5ETk4OVCoVgoODW6rLLNLS+2ZpPyvtv8OHD8NgMGDlypUt2l/AL2NHRDBu3Lg6yyQmJuKpp57Cp59+iiNHjtRbnyXHIDk5Gc7OztDpdNi/fz/GjBkDg8EAPz8/7Ny506S+qqoqLFu2DP7+/nByckKfPn2we/duZTvdSDdv3oSTkxO6detmti4+Ph5z5syBp6dnndu7u7vjhRdewIYNG/hD37ZCqFnt3r1bWqJbg4KCxNXVtc71hYWFAkC6dOliXBYfHy8AZN68ebJp0yaZOHGi/PWvf5Vly5aJg4ODbN++Xe7fvy/nz5+X/v37S8eOHeX27dvG7SMjI8XZ2VkuXLggDx8+lOzsbBk4cKDo9Xq5du2asZyl9U2fPl28vLxM4k5KShIAkp+fb1wWHh4uQUFBivqrNpMmTZJJkyY1apvW2DdL+1lJGwcPHhS9Xi8JCQmN2v+mjOfAwEAJCQmpdV1QUJBcuXJFREROnjwpdnZ20rVrVykuLhYRkYyMDBk/frzJNpYeg5rxfvToUXnw4IHk5eXJ8OHDxdnZWcrLy43lFi1aJI6OjrJ37165d++exMXFiZ2dnXz99deN2s9fe+6556Rv374WlS0pKRG9Xi9RUVFm606cOCHjxo0TEZH8/HwBIPHx8bXWs2TJEgEgZ8+etTjOlnp/IknlmWk7odfroVKpUFRUZLZu9erVmDt3LtLS0hAQEIB169Zh4sSJmDFjBlxdXdG7d2988sknKCgowJYtW0y2VavVxjOCkJAQJCcno6ioCCkpKQB+uSTVmPraktbct4b6WamxY8eisLAQ7733XrPUV5eSkhJcuXIFQUFBDZYdMmQI5s+fj9zcXCxevLjWMk05BqGhoTAYDPD09ERERARKSkpw7do1AL/cIZucnIywsDCEh4fDzc0NS5cuhUajaba+bsiqVavg4+ODxMREs32Njo5GcnKyRfV0794dAJCVldXsMVLjMZm2EyUlJRARGAyGestlZ2ejuLgYAwYMMFk+cOBAODg4mF3CfdyAAQOg0+mMl9iU1mfLrLlvj/dzW5GXlwcRgU6ns6h8YmIinn76aWzevBknTpwwW6/0GDg4OAAAKioqAACXLl1CaWkpevXqZSzj5OQEb2/vVunrffv2ITU1FV9++aXZjVlxcXH4wx/+AF9fX4vqqunjO3fuNHuc1HhMpu3E999/DwDo0aNHveVqbqd3cXExW+fm5lbrme3jHB0dkZ+f32z12Spr79uv+7mtePjwIQDUe0POr2m1WqSkpEClUuHNN99EWVmZyfrmPgYlJSUAgKVLl5rM27569WqdU1uay65du7B69WpkZmaia9euJutOnDiBrKwszJo1y+L6nJycAPytz8m6mEzbicOHDwMAxowZU285Nzc3AKj1Tej+/fvw8/Ord/uKigqTckrrs2XW3LfH+7mtqHmDb8xDBYYMGYIFCxbg8uXLWLFihcm65j4GNTf1rF+/HiJi8nfq1KlG1dUYmzZtwo4dO3Ds2DF07tzZbP22bdtw9OhR2NnZGRN8TawrV66ESqUym0deXl4O4G99TtbFZNoO3L59G+vXr4efnx/efPPNesv26tULLi4uZi/M06dPo7y8HM8++2y922dmZkJEMHjw4EbXp1arjZfb2gJr7tvj/dwSbbSETp06QaVSNXr+6IoVK9CjRw+cPXvWZLnS8fq4Ll26QKvV4ty5c43arqlEBLGxscjKykJ6enqtZ9gAkJKSYpbca65KxMfHQ0TMLnXX9LGXl1fL7gRZhMm0DRERFBcXo7q62vhi2717N4YOHQp7e3ukp6c3+J2pVqvFwoULsW/fPuzYsQOFhYXIysrC7Nmz4ePjg8jISJPy1dXVuHfvHiorK3H+/HlER0fD39/fOBWnMfUFBwfj559/Rnp6OioqKpCfn4+rV6+axejh4YFbt24hNzcXRUVFVksgrblvDfWz0jYyMjJaZWqMTqdDYGAgbty40ajtai73Pj4fs7Hj1ZJ23njjDezcuRPJyckoLCxEVVUVbty4gZ9++gkAEBERAS8vr2Z5nOGFCxfw0UcfYevWrdBoNGaPBV27dm2T667p4969eyuOk5pB699B3L41963nBw4ckD59+ohOpxMHBwexs7MTAKJSqcTNzU0GDRokCQkJcvfuXZPt1qxZI05OTsbpMtu3bzeuq66ulqSkJOnevbtoNBpxd3eXsLAwuXTpkkkdkZGRotFoxNfXV9RqtRgMBpkwYYLk5OSYlLO0vrt378rIkSNFq9VKt27d5J133pGYmBgBIMHBwcZpIGfOnJGAgABxcnKSYcOGmUx/UKIpU2NaY98s7WclbRw6dEj0er0kJiY2av+bMp6joqJEo9FIaWmpcdm+ffskKChIAEjHjh1l7ty5tW4bExNjNjXGkmOwefNm0el0AkC6d+8uOTk5smXLFjEYDAJAAgIC5PvvvxcRkUePHklsbKz4+/uLWq0WT09PCQ8Pl+zsbBERCQsLEwCybNmyevfz1KlTMnToUPHx8REAAkC8vb0lNDRUjh8/LiIiWVlZxnW1/SUlJdVZf0NTY8aOHSu+vr5SXV1db5y/xqkxLSaVvdrM2tNgjYyMFA8PD2uH0Wyakkxbgy33c1PG8+XLl0WtVpt8gGtLqqqqZPjw4bJt2zZrh1KngoIC0Wq1snbt2kZt157en2wM55lS/fjrFK2jPfVzcHAwEhISkJCQ0OAvo9iaqqoqpKeno6ioCBEREdYOp07Lly9Hv379EBUVZe1Q6P8wmRJRs1uyZAkmT56MiIiINvUw+8zMTKSlpSEjI8PiubKtbd26dTh37hwOHToEjUZj7XDo/zCZUq3i4uKQkpKCBw8eoFu3bti7d6+1Q2qX2nM/r1y5ElFRUfjwww+tHYrFRo0ahc8++8zkGci2ZP/+/Xj06BEyMzPh7u5u7XDoV1QifEpyc0pNTcXUqVP58GkbVPP7kXv27LFyJG0Hx3P7wuPZYvbwzJSIiEghJlMiIiKFmEyJiIgUYjIlIiJSiMmUiIhIIbW1A2ivVCqVtUOgOvDYNB77jKh+TKbNLDQ0FLt377Z2GNROnDp1Chs2bOCYIrJxnGdKZMM4L5CoTeA8UyIiIqWYTImIiBRiMiUiIlKIyZSIiEghJlMiIiKFmEyJiIgUYjIlIiJSiMmUiIhIISZTIiIihZhMiYiIFGIyJSIiUojJlIiISCEmUyIiIoWYTImIiBRiMiUiIlKIyZSIiEghJlMiIiKFmEyJiIgUYjIlIiJSiMmUiIhIISZTIiIihZhMiYiIFGIyJSIiUojJlIiISCEmUyIiIoWYTImIiBRiMiUiIlKIyZSIiEghJlMiIiKFmEyJiIgUYjIlIiJSiMmUiIhIIbW1AyCiX+Tn5+M//uM/TJZ98803AIAtW7aYLNfr9Zg2bVqrxUZE9VOJiFg7CCICHj16hE6dOqG4uBj29vYAgJqXp0qlMparqKjA66+/jn/913+1RphEZG4PL/MS2QhHR0dMmjQJarUaFRUVqKioQGVlJSorK43/V1RUAABeeeUVK0dLRL/GZEpkQ1555RWUl5fXW8bNzQ2/+c1vWikiIrIEkymRDRk5ciQ8PT3rXK/RaDBjxgyo1bzdgciWMJkS2RA7OztMnz4dGo2m1vUVFRW88YjIBjGZEtmYadOmGb8bfVznzp0xZMiQVo6IiBrCZEpkYwYNGoSAgACz5Q4ODnj99ddN7uwlItvAZEpkg1599VWzS73l5eW8xEtko5hMiWzQ9OnTzS71BgcHo3fv3laKiIjqw2RKZIN69OiBkJAQ4yVdjUaDN954w8pREVFdmEyJbNRrr71mfBJS5f9v796DmjrTP4B/AyGEQMJFroqgBKoVUZd6TfWnHXfYsY4gopWOdmodd9HRUrwtCkoVAXVx0cHKOlaGmVVXQHDwUnEddXDH8bJ2xMrCekMBlSJg5SZBbs/vjw7pxqAkJuSAPJ8Z/jnnPe/z5D3v5OEk5z1pb+ePeBnrw7iYMtZHff755+jo6AAABAYGYvjw4QJnxBh7Ey6mjPVRXl5emDRpEgDgyy+/FDgbxtjb8GNUetnVq1eRkpIidBqsn3r16hVEIhHOnTuHf/3rX0Knw/qpY8eOCZ3Ce4+vTHvZ48ePkZOTI3QaA9K1a9dw7do1odMwiqenJ9zc3CCVSs0S78mTJzxf3yN8Ps2Hr0zNhP8zNL8FCxYA6P9j/+DBA/j6+polVnZ2NhYuXNjvx4z9qut8st7HV6aM9XHmKqSMsXfHxZQxxhgzEhdTxhhjzEhcTBljjDEjcTFljDHGjMTFtB9YtmwZ5HI5RCIRbt26JXQ6Runs7MTu3buhUqm63Z+QkACRSKTzN3r0aDNn+pszZ87A3t4ep06dEiyH/uL8+fPYuHEjcnNz4ePjozl/X3zxhU7boKAgyOVyWFpawt/fHzdv3hQgY8P1NIfj4+MxatQoKBQKWFtbw9fXF3/+85/R1NT01n5bWlowcuRIbNq0SbPt5MmT2Llzp+ZJWKzv4mLaDxw8eBDff/+90GkY7f79+/i///s/rFmzBs3NzUKnozciEjqFfuHbb79FamoqYmJiEBYWhocPH0KpVGLQoEE4fPgwfvjhB632586dw7FjxzBnzhwUFxcjMDBQoMz1p88cvnjxIlatWoWysjLU1tYiKSkJe/bs0SzVepPY2FjcvXtXa1twcDCkUilmzpyJuro6k70OZnpcTJlZ/PTTT9iwYQNWrFiBcePGvbXtoUOHQERaf//5z3/MlKmu2bNno76+HnPmzBEshy5qtfqNV0RC2rFjBzIzM5GdnQ25XK61LzU1FRYWFoiIiEB9fb1AGRpP3zlsZ2eHiIgIODk5QS6X47PPPkNoaCjOnj2Lx48fd3vMlStX3jjHv/nmG4wdOxaffvop2tvbTfJamOlxMe0nun6Kq78aO3YscnNzsWjRIlhbWwudTr+Vnp6O6upqodPQ8uDBA2zevBlbt27t9klNKpUKUVFRePr0KdatWydAhqah7xw+ffq05td+ujg7OwNAt1ezarUa69evx549e97Y55YtW3Dr1q23tmHC4mLaBxERkpOTMWLECFhbW8Pe3h7r16/XadfR0YG4uDh4eXnBxsYGY8aMQVZWFgAgLS0Ntra2kMlkOHHiBGbNmgWFQgFPT08cPXpUq59Lly5h4sSJkMlkUCgUCAgIQENDQ48xBoLLly/Dy8sLIpEI3333HQD9xzY1NRVSqRSurq5Yvnw5PDw8IJVKoVKpcP36dU27yMhISCQSuLu7a7atXLkStra2EIlEqK2tBQBERUVh7dq1KC0thUgk0jzM4ezZs1AoFEhMTDTHkOhITU0FESE4OPiNbRISEvDBBx/g4MGDOH/+/Fv7IyKkpKTgww8/hLW1NRwdHTF37lzcuXNH08aQ+d0X5vDTp09hY2PT7S//xMbGYuXKlXBxcXnj8Y6Ojpg+fTr27NnDXzv0VcR6VVZWFhk6zLGxsSQSieivf/0rvXjxgpqbm2nfvn0EgAoLCzXt1q1bR9bW1pSTk0MvXrygmJgYsrCwoBs3bmj6AUAXLlyg+vp6qq6upmnTppGtrS21trYSEVFTUxMpFArauXMnqdVqqqqqonnz5lFNTY1eMd7FpEmTaOzYsd3u27ZtG3l6epKDgwNZWVnRsGHDKCQkhP79738bHGf+/Pk0f/78d86zy+PHjwkA7d27V7NNn7ElIoqIiCBbW1sqKSmhlpYWKi4upgkTJpBcLqeKigpNu0WLFpGbm5tW3OTkZAKgORdERGFhYaRUKrXanT59muRyOcXHxxv9Wt9lvvr4+NCoUaO63adUKunRo0dERHTlyhWysLCgYcOGUVNTExER5efnU0hIiNYxcXFxJJFI6NChQ1RXV0e3b9+mwMBAcnZ2pqqqKk07fc+Buefw616+fElyuZwiIyN19l2+fJmCg4OJiKimpoYAUGxsbLf9bNy4Uec9oCfvcj7ZO8nmK9M+Rq1WY/fu3fj973+PNWvWwMHBATY2NnByctJq19LSgrS0NISGhiIsLAwODg7YtGkTrKyskJGRodVWpVJBoVDAxcUF4eHhePnyJSoqKgAAZWVlaGhogL+/P6RSKdzc3JCbmwtnZ2eDYpjKl19+iZMnT+Lx48doamrC0aNHUVFRgenTp6O4uLhXYhrjbWPbRSwWa66yRo0ahbS0NDQ2NppsDGfPno2GhgZs3rzZJP0Z4uXLl3j06BGUSmWPbadMmYLVq1ejrKwMGzZs6LaNWq1GSkoK5s2bh8WLF8Pe3h4BAQHYv38/amtrceDAAZ1j3nYOhJjDr0tKSoKHhwcSEhJ0XmtUVBTS0tL06sfPzw8AUFRUZPIcmfG4mPYxDx48QHNzM2bOnPnWdnfv3kVzc7PWkhEbGxu4u7trfRz2OolEAgBoa2sDAPj4+MDV1RWLFy/Gli1bUFZWZnQMYwwdOhS/+93vYGdnB4lEgsmTJyMjIwNqtRr79u3rlZim8vrYvsn48eMhk8l6bQzNqbq6GkQEmUymV/uEhASMGDEC+/btw+XLl3X2FxcXo6mpCePHj9faPmHCBEgkEq2Px7vz+jkQYg7/r+PHjyM7Oxv//Oc/dW7MiomJwZ/+9CcMGTJEr766xvjZs2cmz5MZj4tpH/PkyRMAeOv3J8CvVwQAsGnTJq31mOXl5QYtO7GxscHFixcxdepUJCYmwsfHB+Hh4VCr1SaLYayAgABYWlri3r17ZovZ26ytrVFTUyN0GkZraWkBAL1vKpNKpcjIyIBIJMLSpUuhVqu19nct/7Czs9M51sHBAY2NjQblJ+QczszMxI4dO1BQUIBhw4Zp7bt8+TKKioqwbNkyvfuzsbEB8NuYs76Fi2kf03U35KtXr97arqvY7t69W2cZydWrVw2K6e/vj1OnTqGyshLR0dHIysrCrl27TBrDGJ2dnejs7Hxv7gJua2tDXV0dPD09hU7FaF1v8IY8VGDKlClYs2YN7t+/j23btmntc3BwAIBui+a7jJlQc3jv3r04fPgwLl68iMGDB+vsT09Px4ULF2BhYaEp8F25JiYmQiQS4ccff9Q6prW1FcBvY876Fi6mfczo0aNhYWGBS5cuvbXd0KFDIZVKjX4iUmVlJUpKSgD8+sazfft2BAYGoqSkxGQxDPGHP/xBZ9uNGzdARJgyZYrZ8uhNBQUFICJMnjxZs00sFvf48XBf5OrqCpFIZPD60W3btmHkyJEoLCzU2j569GjY2dnpFJLr16+jtbUVH330kUFxzD2HiQjR0dEoKipCXl5et1fYAJCRkaFT3Ls+qYiNjQUR6XzU3TXGbm5uvfsi2DvhYtrHuLi4ICwsDDk5OUhPT0dDQwNu376tc+OFVCrFV199haNHjyItLQ0NDQ3o6OjAkydP8PPPP+sdr7KyEsuXL8edO3fQ2tqKwsJClJeXY/LkySaLYYinT58iMzMTdXV1aGtrw9WrV7Fs2TJ4eXlhxYoVvRKzt3V2duLFixdob2/H7du3ERUVBS8vLyxZskTTxtfXF7/88gvy8vLQ1taGmpoalJeX6/Tl5OSEyspKlJWVobGxEW1tbcjPzxdsaYxMJoOPj4/m6wl9dX3c+/p6TKlUirVr1+L48eM4fPgwGhoaUFRUhBUrVsDDwwMREREGx+lpDoeHh8PNzc0kjzMsKSnBX/7yF3z//fewsrLSeSzmrl273rnvrjEOCAgwOk/WC8x68/AA9C63pjc2NtKyZcto0KBBZGdnR1OnTqW4uDgCQJ6envTTTz8REdGrV68oOjqavLy8SCwWk4uLC4WFhVFxcTHt27ePZDIZASA/Pz8qLS2lAwcOkEKhIADk7e1N9+7do7KyMlKpVOTo6EiWlpY0ePBgio2Npfb29h5jGOLq1av08ccfk4eHBwEgAOTu7k4qlYouXbqkabd27VpSKpVka2tLYrGYPD096Y9//CNVVlYaFI/INEtj9u7dS+7u7gSAZDIZBQcH6z22RL8ujbGysqIhQ4aQWCwmhUJBc+fOpdLSUq04z58/p08++YSkUikNHz6cvv76a1q/fj0BIF9fX80ymps3b5K3tzfZ2NjQ1KlTqaqqis6cOUNyuZwSEhKMeq1E7zZfIyMjycrKipqbmzXbjh8/TkqlkgCQs7MzrVq1qttj169fr7M0prOzk5KTk8nPz4+srKzI0dGRQkND6e7du5o2hpyDnuZwaGgoAaC4uLi3vk595nBRUZFmX3d/ycnJb+y/p6Uxs2fPpiFDhlBnZ+db8/xfvDTGbLJ5lHsZT2bhmGqdqTEiIiLIyclJ0BwM8S7z9f79+yQWi+nQoUO9lFXv6ujooGnTplF6errQqbxRbW0tSaVS2rVrl0HH8fuP2fA6U8Z62/v+ix++vr6Ij49HfHx8j7+M0td0dHQgLy8PjY2NCA8PFzqdN9qyZQvGjRuHyMhIoVNhb8DFlL2TO3fudPtTaa//9eU3KGY6GzduxIIFCxAeHt6vHmZfUFCA3Nxc5Ofn671W1txSUlJw69YtnDlzBlZWVkKnw96Aiyl7JyNHjtS5G7G7v8zMTKFTFUxMTAwyMjJQX1+P4cOHIycnR+iUelViYiIiIyOxfft2oVPR28yZM3HkyBGt5yL3JSdOnMCrV69QUFAAR0dHodNhbyEWOgHG3ldJSUlISkoSOg2zCgoKQlBQkNBpvDdCQkIQEhIidBpMD3xlyhhjjBmJiyljjDFmJC6mjDHGmJG4mDLGGGNG4mLKGGOMGYnv5jUTkUgkdAoDFo+94XjMGDMMF1MzycrKEjqFAWf37t0AgNWrVwucSf9x9epV7Nmzh+fre6LrfLLex8XUTD777DOhUxhwjh07BoDH3lB79uzhMXuPcDE1D/7OlDHGGDMSF1PGGGPMSFxMGWOMMSNxMWWMMcaMxMWUMcYYMxIX0/dAbm4ufHx8dH5LVCKRwNXVFTNmzEBycjJevHghdKqMAQDOnz+PjRs36szdL774QqdtUFAQ5HI5LC0t4e/vj5s3bwqQsWHa2tqQlJQEX19fSCQSODg4YPTo0SgrKwMAnDx5Ejt37nzvfzh+IOFi+h4ICwvDw4cPoVQqYW9vDyJCZ2cnqqurkZ2djeHDhyM6Ohr+/v748ccfhU6XDXDffvstUlNTERMTozV3Bw0ahMOHD+OHH37Qan/u3DkcO3YMc+bMQXFxMQIDAwXKXH8LFy7E3//+dxw5cgTNzc3473//C6VSiaamJgBAcHAwpFIpZs6cibq6OoGzZabAxfQ9JRKJ4ODggBkzZiAjIwPZ2dl49uwZZs+ejfr6eqHTGzDUajVUKlW/j2EqO3bsQGZmJrKzsyGXy7X2paamwsLCAhEREf16jmZmZiIvLw/Hjh3DpEmTIBaL4eHhgRMnTmD06NGadt988w3Gjh2LTz/9FO3t7QJmzEyBi+kAMX/+fCxZsgTV1dXYv3+/0OkMGOnp6aiuru73MUzhwYMH2Lx5M7Zu3QqpVKqzX6VSISoqCk+fPsW6desEyNA0/va3vyEwMBABAQE9tt2yZQtu3brFD1Z4D3AxHUCWLFkCAI59vzcAAAm8SURBVMjPz9ds6+joQFxcHLy8vGBjY4MxY8ZoHiWXlpYGW1tbyGQynDhxArNmzYJCoYCnpyeOHj2q1felS5cwceJEyGQyKBQKBAQEoKGhoccYfQ0RISUlBR9++CGsra3h6OiIuXPn4s6dO5o2kZGRkEgkcHd312xbuXIlbG1tIRKJUFtbCwCIiorC2rVrUVpaCpFIBF9fX6SmpkIqlcLV1RXLly+Hh4cHpFIpVCoVrl+/bpIYAHD27FkoFAokJib26ngZIjU1FUSE4ODgN7ZJSEjABx98gIMHD+L8+fNv7U+fc2XIHDbFPG1tbcW1a9cwbtw4vdo7Ojpi+vTp2LNnD4jIoFisjyHWq7Kysshcw6xUKsne3v6N+xsaGggADR06VLNt3bp1ZG1tTTk5OfTixQuKiYkhCwsLunHjBhERxcbGEgC6cOEC1dfXU3V1NU2bNo1sbW2ptbWViIiamppIoVDQzp07Sa1WU1VVFc2bN49qamr0itFb5s+fT/PnzzfomLi4OJJIJHTo0CGqq6uj27dvU2BgIDk7O1NVVZWm3aJFi8jNzU3r2OTkZAKged1ERGFhYaRUKrXaRUREkK2tLZWUlFBLSwsVFxfThAkTSC6XU0VFhUlinD59muRyOcXHxxv0+ntzvvr4+NCoUaO63adUKunRo0dERHTlyhWysLCgYcOGUVNTExER5efnU0hIiNYx+p4rfeYwkWnm6aNHjwgAjRs3jmbMmEHu7u5kbW1NI0eOpO+++446Ozt1jtm4cSMBoMLCQr3j6Muc7z8DXDZfmQ4gcrkcIpEIjY2NAICWlhakpaUhNDQUYWFhcHBwwKZNm2BlZYWMjAytY1UqFRQKBVxcXBAeHo6XL1+ioqICAFBWVoaGhgb4+/tDKpXCzc0Nubm5cHZ2NiiG0NRqNVJSUjBv3jwsXrwY9vb2CAgIwP79+1FbW4sDBw6YLJZYLNZcUY0aNQppaWlobGw02ZjMnj0bDQ0N2Lx5s0n6M9bLly/x6NEjKJXKHttOmTIFq1evRllZGTZs2NBtm3c5V2+bw6aap103GLm4uCAxMRHFxcV49uwZ5s6di1WrVuEf//iHzjF+fn4AgKKiIr3jsL6Hi+kA8vLlSxARFAoFAODu3btobm7WuinCxsYG7u7uWh+VvU4ikQD49fZ/APDx8YGrqysWL16MLVu2aG7/NyaGEIqLi9HU1ITx48drbZ8wYQIkEonWx7CmNn78eMhksj43JqZSXV0NIoJMJtOrfUJCAkaMGIF9+/bh8uXLOvuNPVevz2FTzVNra2sAgL+/P1QqFZycnGBvb4+tW7fC3t6+2yLfNSbPnj3TOw7re7iYDiD37t0DAIwcORLAr8UVADZt2qS1PrW8vBzNzc1692tjY4OLFy9i6tSpSExMhI+PD8LDw6FWq00Wwxy6lijY2dnp7HNwcNBc0fcWa2tr1NTU9GoMobS0tAD4rdj0RCqVIiMjAyKRCEuXLoVardbab+pzZap56uHhAQCa77S7SCQSeHt7o7S0VOcYGxsbAL+NEeufuJgOIGfPngUAzJo1C8CvH0UBv/7uJxFp/V29etWgvv39/XHq1ClUVlYiOjoaWVlZ2LVrl0lj9DYHBwcA6PaNuK6uDp6enr0Wu62trddjCKmrYBjykIIpU6ZgzZo1uH//PrZt26a1z9TnylTz1M7ODn5+figpKdHZ197eDnt7e53tra2tAH4bI9Y/cTEdIKqqqrB79254enpi6dKlAIChQ4dCKpXi1q1bRvVdWVmpefNwcXHB9u3bERgYiJKSEpPFMIfRo0fDzs5O58EW169fR2trKz766CPNNrFYrPmI0BQKCgpARJg8eXKvxRCSq6srRCKRwetHt23bhpEjR6KwsFBruyHnSh+mnKcLFy5EYWEhHj58qNnW3NyM8vLybpfLdI2Jm5ub0bGZcLiYvmeICE1NTejs7AQRoaamBllZWfj4449haWmJvLw8zXemUqkUX331FY4ePYq0tDQ0NDSgo6MDT548wc8//6x3zMrKSixfvhx37txBa2srCgsLUV5ejsmTJ5sshjlIpVKsXbsWx48fx+HDh9HQ0ICioiKsWLECHh4eiIiI0LT19fXFL7/8gry8PLS1taGmpgbl5eU6fTo5OaGyshJlZWVobGzUFMfOzk68ePEC7e3tuH37NqKiouDl5aVZvmRsjPz8/D61NEYmk8HHxwdPnjwx6Liuj3stLS11tut7rvSN09M8DQ8Ph5ubW4+PM1yzZg28vb2xZMkSVFRU4Pnz54iOjoZare72hqquMdFnXSrrw8x+A/EAY45b00+ePEljxowhmUxGEomELCwsCACJRCJycHCgiRMnUnx8PD1//lzn2FevXlF0dDR5eXmRWCwmFxcXCgsLo+LiYtq3bx/JZDICQH5+flRaWkoHDhwghUJBAMjb25vu3btHZWVlpFKpyNHRkSwtLWnw4MEUGxtL7e3tPcboTe+yNKazs5OSk5PJz8+PrKysyNHRkUJDQ+nu3bta7Z4/f06ffPIJSaVSGj58OH399de0fv16AkC+vr6aJS43b94kb29vsrGxoalTp1JVVRVFRESQlZUVDRkyhMRiMSkUCpo7dy6VlpaaLMaZM2dILpdTQkKCQa+/N+drZGQkWVlZUXNzs2bb8ePHSalUEgBydnamVatWdXvs+vXrdZbG6HOu9J3DRD3P09DQUAJAcXFxPb7Wx48f0+eff06Ojo5kbW1NEydOpPz8/G7bzp49m4YMGdLtshlj8dIYs8nmUe5lPJmF8y7F1BwiIiLIyclJ6DS61Zvz9f79+yQWi+nQoUO90n9v6+jooGnTplF6errJ+qytrSWpVEq7du0yWZ//i99/zIbXmTImhIH4ayG+vr6Ij49HfHy8Zj1mf9HR0YG8vDw0NjYiPDzcZP1u2bIF48aNQ2RkpMn6ZMLgYsoYM5uNGzdiwYIFCA8P71cPsy8oKEBubi7y8/P1Xivbk5SUFNy6dQtnzpyBlZWVSfpkwuFiypgZxcTEICMjA/X19Rg+fDhycnKETsnsEhMTERkZie3btwudit5mzpyJI0eOaD0r2RgnTpzAq1evUFBQAEdHR5P0yYQlFjoBxgaSpKQkJCUlCZ2G4IKCghAUFCR0GoIJCQlBSEiI0GkwE+IrU8YYY8xIXEwZY4wxI3ExZYwxxozExZQxxhgzEt+AZCbZ2dlCpzDgdD2mjcdef10Pdecxez/0tR+TeJ+JiIiETuJ9lp2djYULFwqdBmNsAOO3+V53jIspY4wxZpxj/J0pY4wxZiQupowxxpiRuJgyxhhjRuJiyhhjjBnp/wF0hlEI3j3fvwAAAABJRU5ErkJggg==\n","text/plain":["<IPython.core.display.Image object>"]},"metadata":{},"execution_count":56}]},{"cell_type":"code","source":["# From the figure above, you can see that the output layer only contains 1 dense layer with 6 neurons. Let's now train our model:"],"metadata":{"id":"lBk4tbN1JhWn"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["history = model.fit(X_train, y_train, batch_size=128, epochs=5, verbose=1, validation_split=0.2)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"8SpVLMYOJhYl","executionInfo":{"status":"ok","timestamp":1672431062883,"user_tz":-540,"elapsed":148783,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"037c48a6-83ab-4e57-ffc6-123d5caa4632"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 1/5\n","798/798 [==============================] - 49s 54ms/step - loss: 0.3911 - acc: 0.9929 - val_loss: 0.3954 - val_acc: 0.9943\n","Epoch 2/5\n","798/798 [==============================] - 24s 31ms/step - loss: 0.3894 - acc: 0.9941 - val_loss: 0.3938 - val_acc: 0.9943\n","Epoch 3/5\n","798/798 [==============================] - 24s 30ms/step - loss: 0.3881 - acc: 0.9941 - val_loss: 0.3927 - val_acc: 0.9943\n","Epoch 4/5\n","798/798 [==============================] - 26s 33ms/step - loss: 0.3870 - acc: 0.9941 - val_loss: 0.3917 - val_acc: 0.9943\n","Epoch 5/5\n","798/798 [==============================] - 25s 32ms/step - loss: 0.3860 - acc: 0.9941 - val_loss: 0.3909 - val_acc: 0.9943\n"]}]},{"cell_type":"code","source":["score = model.evaluate(X_test, y_test, verbose=1)\n","\n","print(\"Test Score:\", score[0])\n","print(\"Test Accuracy:\", score[1])"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"JjHBZOjDJhat","executionInfo":{"status":"ok","timestamp":1672431072695,"user_tz":-540,"elapsed":9823,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"a5cb61ee-49f5-4710-90c3-7d192b6463e5"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["998/998 [==============================] - 10s 10ms/step - loss: 0.3907 - acc: 0.9941\n","Test Score: 0.390667587518692\n","Test Accuracy: 0.9941093325614929\n"]}]},{"cell_type":"code","source":["y_pred = model.predict(X_test)\n"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"MgJqdqP954qX","executionInfo":{"status":"ok","timestamp":1672431083746,"user_tz":-540,"elapsed":11062,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"8d8fb8c1-fb13-472e-fd75-8775da0c8185"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["998/998 [==============================] - 11s 10ms/step\n"]}]},{"cell_type":"code","source":["from sklearn.metrics import label_ranking_average_precision_score\n","\n","lrap = label_ranking_average_precision_score(y_test[0], y_pred[0])"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":339},"id":"46GGerBnU9iM","executionInfo":{"status":"error","timestamp":1672426118934,"user_tz":-540,"elapsed":856,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"9d421767-5476-42d0-a962-9d254901a89b"},"execution_count":null,"outputs":[{"output_type":"error","ename":"ValueError","evalue":"ignored","traceback":["\u001b[0;31m---------------------------------------------------------------------------\u001b[0m","\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)","\u001b[0;32m<ipython-input-33-53428f13c057>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0msklearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmetrics\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mlabel_ranking_average_precision_score\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 3\u001b[0;31m \u001b[0mlrap\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlabel_ranking_average_precision_score\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_pred\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0m","\u001b[0;32m/usr/local/lib/python3.8/dist-packages/sklearn/metrics/_ranking.py\u001b[0m in \u001b[0;36mlabel_ranking_average_precision_score\u001b[0;34m(y_true, y_score, sample_weight)\u001b[0m\n\u001b[1;32m   1064\u001b[0m         \u001b[0my_type\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;34m\"binary\"\u001b[0m \u001b[0;32mand\u001b[0m \u001b[0my_true\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mndim\u001b[0m \u001b[0;34m==\u001b[0m \u001b[0;36m2\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1065\u001b[0m     ):\n\u001b[0;32m-> 1066\u001b[0;31m         \u001b[0;32mraise\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"{0} format is not supported\"\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_type\u001b[0m\u001b[0;34m)\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   1067\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1068\u001b[0m     \u001b[0my_true\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mcsr_matrix\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_true\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;31mValueError\u001b[0m: binary format is not supported"]}]},{"cell_type":"code","source":[],"metadata":{"id":"oQqO4pDBW_Br"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["X = tokenizer.texts_to_sequences(['You are a fucking stupid head and I a tired of your shit'])\n","X = tf.keras.preprocessing.sequence.pad_sequences(X, padding='post', maxlen=maxlen)\n","#you must reoder the labls in alphabetical order (keras)\n","labels = [\"toxic\", \"severe_toxic\", \"obscene\", \"threat\", \"insult\", \"identity_hate\"]\n","pred = model.predict(X)\n","for i, p in enumerate(pred[0]):\n","  print(labels[i], p)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"KTLPX7ZnSqPA","executionInfo":{"status":"ok","timestamp":1672427622690,"user_tz":-540,"elapsed":351,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"8ce0d8d9-9817-429c-abc9-bc93dbabe736"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["1/1 [==============================] - 0s 64ms/step\n","toxic 0.020367365\n","severe_toxic 0.018566864\n","obscene 0.019043945\n","threat 0.018508988\n","insult 0.01898131\n","identity_hate 0.018555567\n"]}]},{"cell_type":"code","source":["pred\n","from sklearn.metrics import label_ranking_average_precision_score\n","\n","lrap = label_ranking_average_precision_score(dd, pred)"],"metadata":{"id":"0ZeD277MU7Q-"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["lrap"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"sarcc_RZdtml","executionInfo":{"status":"ok","timestamp":1672427791277,"user_tz":-540,"elapsed":341,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"b2c67f99-0faf-4602-bcf7-17756b5e758b"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["1.0"]},"metadata":{},"execution_count":51}]},{"cell_type":"code","source":["toxic_comments_labels.values[6]\n","dd = np.expand_dims(y_test[6], 0)\n","dd.shape"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"Nk5aEeW52aCz","executionInfo":{"status":"ok","timestamp":1672427770950,"user_tz":-540,"elapsed":2,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"f5504f8b-b2b5-47c6-8f80-1f574b21f1fb"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(1, 6)"]},"metadata":{},"execution_count":49}]},{"cell_type":"code","source":["# from sklearn.metrics import label_ranking_average_precision_score\n","# from sklearn.metrics import label_ranking_loss\n","# ranking_loss = label_ranking_loss(y_test, y_pred[0])\n","\n","# ranking_scores = label_ranking_average_precision_score(y_one_hot[0], pred)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":356},"id":"_5vumo0hWyr8","executionInfo":{"status":"error","timestamp":1672251727969,"user_tz":-540,"elapsed":4,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"cf0f60ab-dd59-434a-bf43-d7d20c26c82e"},"execution_count":null,"outputs":[{"output_type":"error","ename":"ValueError","evalue":"ignored","traceback":["\u001b[0;31m---------------------------------------------------------------------------\u001b[0m","\u001b[0;31mValueError\u001b[0m                                Traceback (most recent call last)","\u001b[0;32m<ipython-input-57-eb68781fcd8e>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;31m# from sklearn.metrics import label_ranking_average_precision_score\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      2\u001b[0m \u001b[0;32mfrom\u001b[0m \u001b[0msklearn\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmetrics\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mlabel_ranking_loss\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 3\u001b[0;31m \u001b[0mranking_loss\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mlabel_ranking_loss\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0my_pred\u001b[0m\u001b[0;34m[\u001b[0m\u001b[0;36m0\u001b[0m\u001b[0;34m]\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      4\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m      5\u001b[0m \u001b[0;31m# ranking_scores = label_ranking_average_precision_score(y_one_hot[0], pred)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;32m/usr/local/lib/python3.8/dist-packages/sklearn/metrics/_ranking.py\u001b[0m in \u001b[0;36mlabel_ranking_loss\u001b[0;34m(y_true, y_score, sample_weight)\u001b[0m\n\u001b[1;32m   1201\u001b[0m     \u001b[0my_type\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mtype_of_target\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_true\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1202\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0my_type\u001b[0m \u001b[0;32mnot\u001b[0m \u001b[0;32min\u001b[0m \u001b[0;34m(\u001b[0m\u001b[0;34m\"multilabel-indicator\"\u001b[0m\u001b[0;34m,\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m-> 1203\u001b[0;31m         \u001b[0;32mraise\u001b[0m \u001b[0mValueError\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m\"{0} format is not supported\"\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mformat\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_type\u001b[0m\u001b[0;34m)\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   1204\u001b[0m \u001b[0;34m\u001b[0m\u001b[0m\n\u001b[1;32m   1205\u001b[0m     \u001b[0;32mif\u001b[0m \u001b[0my_true\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m \u001b[0;34m!=\u001b[0m \u001b[0my_score\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mshape\u001b[0m\u001b[0;34m:\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;31mValueError\u001b[0m: binary format is not supported"]}]},{"cell_type":"code","source":["y_rank = []\n","for label_set in y_test:\n","    ranking = [i for i, label in enumerate(label_set) if label == 1]\n","    y_rank.append(ranking)"],"metadata":{"id":"Iyi88XOd99Bw"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["import tensorflow.keras as keras\n","lrap = keras.metrics.LabelRankingAveragePrecision()(y_test, pred)\n","lrap"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":200},"id":"136Hjg0BBcoC","executionInfo":{"status":"error","timestamp":1672253435018,"user_tz":-540,"elapsed":308,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"9b18450e-bb0b-407b-f329-bc30dd79a1ce"},"execution_count":null,"outputs":[{"output_type":"error","ename":"AttributeError","evalue":"ignored","traceback":["\u001b[0;31m---------------------------------------------------------------------------\u001b[0m","\u001b[0;31mAttributeError\u001b[0m                            Traceback (most recent call last)","\u001b[0;32m<ipython-input-64-7f57bc12e2c8>\u001b[0m in \u001b[0;36m<module>\u001b[0;34m\u001b[0m\n\u001b[1;32m      1\u001b[0m \u001b[0;32mimport\u001b[0m \u001b[0mtensorflow\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mkeras\u001b[0m \u001b[0;32mas\u001b[0m \u001b[0mkeras\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n\u001b[0;32m----> 2\u001b[0;31m \u001b[0mlrap\u001b[0m \u001b[0;34m=\u001b[0m \u001b[0mkeras\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mmetrics\u001b[0m\u001b[0;34m.\u001b[0m\u001b[0mLabelRankingAveragePrecision\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0;34m)\u001b[0m\u001b[0;34m(\u001b[0m\u001b[0my_test\u001b[0m\u001b[0;34m,\u001b[0m \u001b[0mpred\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      3\u001b[0m \u001b[0mlrap\u001b[0m\u001b[0;34m\u001b[0m\u001b[0;34m\u001b[0m\u001b[0m\n","\u001b[0;31mAttributeError\u001b[0m: module 'keras.api._v2.keras.metrics' has no attribute 'LabelRankingAveragePrecision'"]}]},{"cell_type":"code","source":["ranking_loss"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"2wXQKS7w6JGy","executionInfo":{"status":"ok","timestamp":1672250694459,"user_tz":-540,"elapsed":4,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"74f03db4-30d5-430e-f6aa-ccf6ee06623b"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["0.0054460633279370535"]},"metadata":{},"execution_count":50}]},{"cell_type":"code","source":["from sklearn.preprocessing import label_binarize\n","y_binarized = label_binarize(y_test, classes=[0, 1, 2, 3,4,5])"],"metadata":{"id":"jApUhSlu6aly"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":[],"metadata":{"id":"q7p322kR6arf"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["y_one_hot[1]"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"b1ZhvUbE3ssx","executionInfo":{"status":"ok","timestamp":1672250060018,"user_tz":-540,"elapsed":261,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"16431edb-d1a7-44a2-f965-7b6ecf7594eb"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"text/plain":["array([1, 0])"]},"metadata":{},"execution_count":46}]},{"cell_type":"code","source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['acc'])\n","plt.plot(history.history['val_acc'])\n","\n","plt.title('model accuracy')\n","plt.ylabel('accuracy')\n","plt.xlabel('epoch')\n","plt.legend(['train','test'], loc='upper left')\n","plt.show()\n","\n","plt.plot(history.history['loss'])\n","plt.plot(history.history['val_loss'])\n","\n","plt.title('model loss')\n","plt.ylabel('loss')\n","plt.xlabel('epoch')\n","plt.legend(['train','test'], loc='upper left')\n","plt.show()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":1000},"id":"SeZzeYkRJhct","executionInfo":{"status":"ok","timestamp":1672238617211,"user_tz":-540,"elapsed":740,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"ad2fe33f-969d-42b5-f3d2-5c85078e7f20"},"execution_count":null,"outputs":[{"output_type":"display_data","data":{"text/plain":["<Figure size 720x576 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}},{"output_type":"display_data","data":{"text/plain":["<Figure size 720x576 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":[],"metadata":{"id":"T7PVtJOFJheg"},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":["# **Multi-lable Text Classification Model with Multiple Output Layers**"],"metadata":{"id":"v0k5JYXtMoBp"}},{"cell_type":"code","source":["def preprocess_text(sen):\n","    # Remove punctuations and numbers\n","    sentence = re.sub('[^a-zA-Z]', ' ', sen)\n","\n","    # Single character removal\n","    sentence = re.sub(r\"\\s+[a-zA-Z]\\s+\", ' ', sentence)\n","\n","    # Removing multiple spaces\n","    sentence = re.sub(r'\\s+', ' ', sentence)\n","\n","    return sentence"],"metadata":{"id":"VRUhDg4dMaI_"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["X = []\n","sentences = list(toxic_comments[\"comment_text\"])\n","for sen in sentences:\n","    X.append(preprocess_text(sen))\n","\n","y = toxic_comments[[\"toxic\", \"severe_toxic\", \"obscene\", \"threat\", \"insult\", \"identity_hate\"]]"],"metadata":{"id":"aN3xmlB-MaLH"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.20, random_state=42)"],"metadata":{"id":"DPPHSER-MaOp"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["# First output\n","y1_train = y_train[[\"toxic\"]].values\n","y1_test =  y_test[[\"toxic\"]].values\n","\n","# Second output\n","y2_train = y_train[[\"severe_toxic\"]].values\n","y2_test =  y_test[[\"severe_toxic\"]].values\n","\n","# Third output\n","y3_train = y_train[[\"obscene\"]].values\n","y3_test =  y_test[[\"obscene\"]].values\n","\n","# Fourth output\n","y4_train = y_train[[\"threat\"]].values\n","y4_test =  y_test[[\"threat\"]].values\n","\n","# Fifth output\n","y5_train = y_train[[\"insult\"]].values\n","y5_test =  y_test[[\"insult\"]].values\n","\n","# Sixth output\n","y6_train = y_train[[\"identity_hate\"]].values\n","y6_test =  y_test[[\"identity_hate\"]].values"],"metadata":{"id":"omQk26WXMaQ2"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["vocab_size = 5000\n","tokenizer = tf.keras.preprocessing.text.Tokenizer(num_words=vocab_size)\n","tokenizer.fit_on_texts(X_train)\n","\n","X_train = tokenizer.texts_to_sequences(X_train)\n","X_test = tokenizer.texts_to_sequences(X_test)\n","\n","vocab_size = len(tokenizer.word_index) + 1\n","\n","maxlen = 200\n","\n","X_train = tf.keras.preprocessing.sequence.pad_sequences(X_train, padding='post', maxlen=maxlen)\n","X_test = tf.keras.preprocessing.sequence.pad_sequences(X_test, padding='post', maxlen=maxlen)"],"metadata":{"id":"Q6nivb5hMaSv"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["from numpy import array\n","from numpy import asarray\n","from numpy import zeros\n","\n","embeddings_dictionary = dict()\n","\n","glove_file = open('/content/drive/MyDrive/datasets/CB_dataset/glove.6B.100d.txt', encoding=\"utf8\")\n","\n","for line in glove_file:\n","    records = line.split()\n","    word = records[0]\n","    vector_dimensions = asarray(records[1:], dtype='float32')\n","    embeddings_dictionary[word] = vector_dimensions\n","glove_file.close()\n","\n","embedding_matrix = zeros((vocab_size, 100))\n","for word, index in tokenizer.word_index.items():\n","    embedding_vector = embeddings_dictionary.get(word)\n","    if embedding_vector is not None:\n","        embedding_matrix[index] = embedding_vector"],"metadata":{"id":"lHgMmomAMaU9"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["input_1 = tf.keras.Input(shape=(maxlen, ))\n","embedding_layer = tf.keras.layers.Embedding(vocab_size, 100, weights=[embedding_matrix], trainable=False)(input_1)\n","net = tf.keras.layers.LSTM(128)(embedding_layer)\n","# LSTM_Layer_1 = tf.keras.layers.Bidirectional(tf.keras.layers.GRU(128, return_sequences=True))(embedding_layer)\n","# net = tf.keras.layers.GlobalMaxPool1D()(LSTM_Layer_1)\n","# net = tf.keras.layers.Dense(124, activation='relu')(net)\n","# net = tf.keras.layers.Dropout(0.4)(net)\n","# output = tf.keras.layers.Dense(6, activation='sigmoid')(net)\n","\n","output1 = Dense(1, activation='sigmoid')(net)\n","output2 = Dense(1, activation='sigmoid')(net)\n","output3 = Dense(1, activation='sigmoid')(net)\n","output4 = Dense(1, activation='sigmoid')(net)\n","output5 = Dense(1, activation='sigmoid')(net)\n","output6 = Dense(1, activation='sigmoid')(net)\n","\n","model = Model(inputs=input_1, outputs=[output1, output2, output3, output4, output5, output6])\n","model.compile(loss='binary_crossentropy', optimizer='adam', metrics=['acc'])"],"metadata":{"id":"VR55GYPyMaW8"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":["print(model.summary())"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"FkxMhZb1MaZJ","executionInfo":{"status":"ok","timestamp":1672239675984,"user_tz":-540,"elapsed":329,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"7dc2819a-af3d-44c3-d5bb-8f615e6e0f74"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Model: \"model\"\n","__________________________________________________________________________________________________\n"," Layer (type)                   Output Shape         Param #     Connected to                     \n","==================================================================================================\n"," input_1 (InputLayer)           [(None, 200)]        0           []                               \n","                                                                                                  \n"," embedding (Embedding)          (None, 200, 100)     14824300    ['input_1[0][0]']                \n","                                                                                                  \n"," lstm (LSTM)                    (None, 128)          117248      ['embedding[0][0]']              \n","                                                                                                  \n"," dense (Dense)                  (None, 1)            129         ['lstm[0][0]']                   \n","                                                                                                  \n"," dense_1 (Dense)                (None, 1)            129         ['lstm[0][0]']                   \n","                                                                                                  \n"," dense_2 (Dense)                (None, 1)            129         ['lstm[0][0]']                   \n","                                                                                                  \n"," dense_3 (Dense)                (None, 1)            129         ['lstm[0][0]']                   \n","                                                                                                  \n"," dense_4 (Dense)                (None, 1)            129         ['lstm[0][0]']                   \n","                                                                                                  \n"," dense_5 (Dense)                (None, 1)            129         ['lstm[0][0]']                   \n","                                                                                                  \n","==================================================================================================\n","Total params: 14,942,322\n","Trainable params: 118,022\n","Non-trainable params: 14,824,300\n","__________________________________________________________________________________________________\n","None\n"]}]},{"cell_type":"code","source":["from keras.utils import plot_model\n","plot_model(model, to_file='model_plot4b.png', show_shapes=True, show_layer_names=True)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":277},"id":"orH9CzqGMabe","executionInfo":{"status":"ok","timestamp":1672239676602,"user_tz":-540,"elapsed":6,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"14daec17-b49a-4195-ad14-2ae19407cfe2"},"execution_count":null,"outputs":[{"output_type":"execute_result","data":{"image/png":"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\n","text/plain":["<IPython.core.display.Image object>"]},"metadata":{},"execution_count":16}]},{"cell_type":"code","source":["history = model.fit(x=X_train, y=[y1_train, y2_train, y3_train, y4_train, y5_train, y6_train], batch_size=8192, epochs=5, verbose=1, validation_split=0.2)"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"eEarsJwBMadU","executionInfo":{"status":"ok","timestamp":1672239732378,"user_tz":-540,"elapsed":35126,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"c43bf28f-236f-41f2-ffee-e2e23ad1dc48"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["Epoch 1/5\n","13/13 [==============================] - 12s 495ms/step - loss: 3.1738 - dense_loss: 0.5683 - dense_1_loss: 0.5170 - dense_2_loss: 0.5517 - dense_3_loss: 0.4805 - dense_4_loss: 0.5480 - dense_5_loss: 0.5082 - dense_acc: 0.9043 - dense_1_acc: 0.9900 - dense_2_acc: 0.9474 - dense_3_acc: 0.9969 - dense_4_acc: 0.9510 - dense_5_acc: 0.9916 - val_loss: 0.9301 - val_dense_loss: 0.3281 - val_dense_1_loss: 0.0885 - val_dense_2_loss: 0.2118 - val_dense_3_loss: 0.0271 - val_dense_4_loss: 0.1964 - val_dense_5_loss: 0.0781 - val_dense_acc: 0.9036 - val_dense_1_acc: 0.9901 - val_dense_2_acc: 0.9468 - val_dense_3_acc: 0.9967 - val_dense_4_acc: 0.9509 - val_dense_5_acc: 0.9902\n","Epoch 2/5\n","13/13 [==============================] - 6s 436ms/step - loss: 0.8780 - dense_loss: 0.3235 - dense_1_loss: 0.0635 - dense_2_loss: 0.2102 - dense_3_loss: 0.0218 - dense_4_loss: 0.2047 - dense_5_loss: 0.0542 - dense_acc: 0.9043 - dense_1_acc: 0.9900 - dense_2_acc: 0.9474 - dense_3_acc: 0.9969 - dense_4_acc: 0.9510 - dense_5_acc: 0.9916 - val_loss: 0.8646 - val_dense_loss: 0.3198 - val_dense_1_loss: 0.0558 - val_dense_2_loss: 0.2117 - val_dense_3_loss: 0.0225 - val_dense_4_loss: 0.1989 - val_dense_5_loss: 0.0558 - val_dense_acc: 0.9036 - val_dense_1_acc: 0.9901 - val_dense_2_acc: 0.9468 - val_dense_3_acc: 0.9967 - val_dense_4_acc: 0.9509 - val_dense_5_acc: 0.9902\n","Epoch 3/5\n","13/13 [==============================] - 6s 438ms/step - loss: 0.8497 - dense_loss: 0.3172 - dense_1_loss: 0.0574 - dense_2_loss: 0.2075 - dense_3_loss: 0.0213 - dense_4_loss: 0.1970 - dense_5_loss: 0.0492 - dense_acc: 0.9043 - dense_1_acc: 0.9900 - dense_2_acc: 0.9474 - dense_3_acc: 0.9969 - dense_4_acc: 0.9510 - dense_5_acc: 0.9916 - val_loss: 0.8623 - val_dense_loss: 0.3183 - val_dense_1_loss: 0.0576 - val_dense_2_loss: 0.2091 - val_dense_3_loss: 0.0227 - val_dense_4_loss: 0.1976 - val_dense_5_loss: 0.0571 - val_dense_acc: 0.9036 - val_dense_1_acc: 0.9901 - val_dense_2_acc: 0.9468 - val_dense_3_acc: 0.9967 - val_dense_4_acc: 0.9509 - val_dense_5_acc: 0.9902\n","Epoch 4/5\n","13/13 [==============================] - 6s 453ms/step - loss: 0.8477 - dense_loss: 0.3162 - dense_1_loss: 0.0576 - dense_2_loss: 0.2068 - dense_3_loss: 0.0214 - dense_4_loss: 0.1962 - dense_5_loss: 0.0495 - dense_acc: 0.9043 - dense_1_acc: 0.9900 - dense_2_acc: 0.9474 - dense_3_acc: 0.9969 - dense_4_acc: 0.9510 - dense_5_acc: 0.9916 - val_loss: 0.8575 - val_dense_loss: 0.3175 - val_dense_1_loss: 0.0563 - val_dense_2_loss: 0.2084 - val_dense_3_loss: 0.0226 - val_dense_4_loss: 0.1962 - val_dense_5_loss: 0.0565 - val_dense_acc: 0.9036 - val_dense_1_acc: 0.9901 - val_dense_2_acc: 0.9468 - val_dense_3_acc: 0.9967 - val_dense_4_acc: 0.9509 - val_dense_5_acc: 0.9902\n","Epoch 5/5\n","13/13 [==============================] - 6s 458ms/step - loss: 0.8443 - dense_loss: 0.3158 - dense_1_loss: 0.0563 - dense_2_loss: 0.2063 - dense_3_loss: 0.0212 - dense_4_loss: 0.1958 - dense_5_loss: 0.0489 - dense_acc: 0.9043 - dense_1_acc: 0.9900 - dense_2_acc: 0.9474 - dense_3_acc: 0.9969 - dense_4_acc: 0.9510 - dense_5_acc: 0.9916 - val_loss: 0.8545 - val_dense_loss: 0.3171 - val_dense_1_loss: 0.0556 - val_dense_2_loss: 0.2080 - val_dense_3_loss: 0.0224 - val_dense_4_loss: 0.1960 - val_dense_5_loss: 0.0554 - val_dense_acc: 0.9036 - val_dense_1_acc: 0.9901 - val_dense_2_acc: 0.9468 - val_dense_3_acc: 0.9967 - val_dense_4_acc: 0.9509 - val_dense_5_acc: 0.9902\n"]}]},{"cell_type":"code","source":["score = model.evaluate(x=X_test, y=[y1_test, y2_test, y3_test, y4_test, y5_test, y6_test], verbose=1)\n","\n","print(\"Test Score:\", score[0])\n","print(\"Test Accuracy:\", score[1])"],"metadata":{"colab":{"base_uri":"https://localhost:8080/"},"id":"ozA2svDyMaf4","executionInfo":{"status":"ok","timestamp":1672239746797,"user_tz":-540,"elapsed":10216,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"b2a92926-f810-466b-9548-df6917c1a0e8"},"execution_count":null,"outputs":[{"output_type":"stream","name":"stdout","text":["998/998 [==============================] - 10s 10ms/step - loss: 0.8505 - dense_loss: 0.3157 - dense_1_loss: 0.0563 - dense_2_loss: 0.2094 - dense_3_loss: 0.0164 - dense_4_loss: 0.2002 - dense_5_loss: 0.0524 - dense_acc: 0.9042 - dense_1_acc: 0.9899 - dense_2_acc: 0.9463 - dense_3_acc: 0.9977 - dense_4_acc: 0.9494 - dense_5_acc: 0.9908\n","Test Score: 0.8504884243011475\n","Test Accuracy: 0.3156925141811371\n"]}]},{"cell_type":"code","source":["import matplotlib.pyplot as plt\n","\n","plt.plot(history.history['dense_1_acc'])\n","plt.plot(history.history['val_dense_1_acc'])\n","\n","plt.title('model accuracy')\n","plt.ylabel('accuracy')\n","plt.xlabel('epoch')\n","plt.legend(['train','test'], loc='upper left')\n","plt.show()\n","\n","plt.plot(history.history['dense_1_loss'])\n","plt.plot(history.history['val_dense_1_loss'])\n","\n","plt.title('model loss')\n","plt.ylabel('loss')\n","plt.xlabel('epoch')\n","plt.legend(['train','test'], loc='upper left')\n","plt.show()"],"metadata":{"colab":{"base_uri":"https://localhost:8080/","height":573},"id":"5LemNcKRMah6","executionInfo":{"status":"ok","timestamp":1672239748536,"user_tz":-540,"elapsed":602,"user":{"displayName":"Furqan Ali","userId":"08783436873128206665"}},"outputId":"3da04a92-3d40-46ff-905a-ded5d0cab9c6"},"execution_count":null,"outputs":[{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"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\n"},"metadata":{"needs_background":"light"}},{"output_type":"display_data","data":{"text/plain":["<Figure size 432x288 with 1 Axes>"],"image/png":"iVBORw0KGgoAAAANSUhEUgAAAYIAAAEWCAYAAABrDZDcAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deXgc9Z3n8fe3Wy3JsmX5ki8d2Nw2Nrax5EBgEgIJMRBMCIdEQibMzgyZ3eEh2UnYwEzCPsnuzmQn+2SymWGSkIRnc/rgSpxgAoFAjiGAZWOwjQ0YAkg2toV8Hzq6+7t/dElqC8mWrG5Vq/vzep5+urrqV1XfLrv7ozr6V+buiIhI4YqEXYCIiIRLQSAiUuAUBCIiBU5BICJS4BQEIiIFTkEgIlLgFAQig2Rm/8/M/ucg275hZh8c7nJERoKCQESkwCkIREQKnIJA8kpwSOZ2M3vRzA6b2ffNbJqZPWJmB83scTObmNZ+mZltNrN9ZvaUmc1Jm7bIzNYH860ESvus6yNmtiGY92kzO/cka/5rM9tmZnvMbLWZzQzGm5n9i5ntNrMDZrbRzOYF064ws5eC2rab2edPaoOJoCCQ/HQt8CHgTOAq4BHg74FKUv/nbwMwszOB5cBng2lrgF+YWbGZFQM/A34ETALuC5ZLMO8i4F7g08Bk4DvAajMrGUqhZnYJ8E/ADcAM4E1gRTD5MuB9wfuoCNq0BdO+D3za3cuBecBvhrJekXQKAslH/+ruu9x9O/B74Fl3f97d24GHgEVBuwbgYXf/tbt3Af8HGAO8FzgfiAHfcPcud78fWJu2jluA77j7s+6ecPcfAB3BfEPxCeBed1/v7h3AncAFZjYL6ALKgbMBc/ct7v52MF8XMNfMxrv7XndfP8T1ivRQEEg+2pU2fLSf1+OC4Zmk/gIHwN2TQDNQFUzb7sf2yvhm2vApwOeCw0L7zGwfUBPMNxR9azhE6q/+Knf/DfBvwN3AbjO7x8zGB02vBa4A3jSz35rZBUNcr0gPBYEUsh2kvtCB1DF5Ul/m24G3gapgXLfatOFm4H+5+4S0R5m7Lx9mDWNJHWraDuDu33T3xcBcUoeIbg/Gr3X3q4GppA5hrRriekV6KAikkK0CrjSzS80sBnyO1OGdp4E/AnHgNjOLmdnHgCVp834X+Bsze09wUnesmV1pZuVDrGE58BdmtjA4v/CPpA5lvWFm9cHyY8BhoB1IBucwPmFmFcEhrQNAchjbQQqcgkAKlru/DNwE/CvwDqkTy1e5e6e7dwIfA24G9pA6n/Bg2rxNwF+TOnSzF9gWtB1qDY8DXwIeILUXchrQGEweTypw9pI6fNQGfC2Y9kngDTM7APwNqXMNIifFdGMaEZHCpj0CEZECpyAQESlwCgIRkQKnIBARKXBFYRcwVFOmTPFZs2aFXYaIyKiybt26d9y9sr9poy4IZs2aRVNTU9hliIiMKmb25kDTdGhIRKTAKQhERAqcgkBEpMCNunME/enq6qKlpYX29vawS8mq0tJSqquricViYZciInkkL4KgpaWF8vJyZs2axbGdReYPd6etrY2WlhZmz54ddjkikkfy4tBQe3s7kydPztsQADAzJk+enPd7PSIy8vIiCIC8DoFuhfAeRWTk5U0QnMiRzjhv7z8adhkiIjmnYILgaGeC1oMdHOmMZ3zZ+/bt49///d+HPN8VV1zBvn37Ml6PiMhQFEwQVJTFiJix93Bnxpc9UBDE48cPnTVr1jBhwoSM1yMiMhR5cdXQYBRFIlSMibHvSBfTK5xoJHPH2++44w5ee+01Fi5cSCwWo7S0lIkTJ7J161ZeeeUVPvrRj9Lc3Ex7ezuf+cxnuOWWW4De7jIOHTrE5ZdfzkUXXcTTTz9NVVUVP//5zxkzZkzGahQRGUjeBcGXf7GZl3Yc6Hdawp32zgQlsShFQwiCuTPH89+vOmfA6V/96lfZtGkTGzZs4KmnnuLKK69k06ZNPZd53nvvvUyaNImjR49SX1/Ptddey+TJk49Zxquvvsry5cv57ne/yw033MADDzzATTfdNOgaRUROVlYPDZnZUjN72cy2mdkd/Uy/2cxazWxD8PirbNYTNSNiRjyR3ft8L1my5Jhr/b/5zW+yYMECzj//fJqbm3n11VffNc/s2bNZuHAhAIsXL+aNN97Iao0iIt2ytkdgZlHgbuBDQAuw1sxWu/tLfZqudPdbM7Xe4/3lDtB6sJ2397dz5rRySmPRTK32GGPHju0Zfuqpp3j88cf54x//SFlZGRdffHG/vwUoKSnpGY5Goxw9qiucRGRkZHOPYAmwzd1fd/dOYAVwdRbXNygTyooxMnvSuLy8nIMHD/Y7bf/+/UycOJGysjK2bt3KM888k7H1iohkQjbPEVQBzWmvW4D39NPuWjN7H/AK8F/dvbmfNhkTi0YYP6aIvUe6mFZRSiQDP9KaPHkyF154IfPmzWPMmDFMmzatZ9rSpUv59re/zZw5czjrrLM4//zzh70+EZFMCvtk8S+A5e7eYWafBn4AXNK3kZndAtwCUFtbO+yVThpbzP6jhzlwtIsJZcXDXh7AT3/6037Hl5SU8Mgjj/Q7rfs8wJQpU9i0aVPP+M9//vMZqUlEZDCyeWhoO1CT9ro6GNfD3dvcvSN4+T1gcX8Lcvd73L3O3esqK/u909qQjCspIhaNsCcLvykQERltshkEa4EzzGy2mRUDjcDq9AZmNiPt5TJgSxbrSV8vk8YWc6gjTmc8MRKrFBHJWVkLAnePA7cCj5L6gl/l7pvN7CtmtixodpuZbTazF4DbgJuzVU9fE4NDQnsOd43UKkVEclJWzxG4+xpgTZ9xd6UN3wncmc0aBlJcFKG8NMbeI51MG1+inj1FpGAVTF9D/ZlUFqMrkeRge+Y7ohMRGS0KOgjKx8QoikTYe0QnjUWkcBV0EETMmDg2xoGjcbqG0e3EyXZDDfCNb3yDI0eOnPS6RUSGq6CDAFInjR0f1l6BgkBERrOwf1AWutJYlLHFRew93EnluJM7aZzeDfWHPvQhpk6dyqpVq+jo6OCaa67hy1/+MocPH+aGG26gpaWFRCLBl770JXbt2sWOHTv4wAc+wJQpU3jyySez8A5FRI4v/4LgkTtg58YhzVKbTNLRlSRZHCXaXxBMnw+Xf3XA+dO7oX7ssce4//77ee6553B3li1bxu9+9ztaW1uZOXMmDz/8MJDqg6iiooKvf/3rPPnkk0yZMmVINYuIZErBHxoCKIoYZgzrPEG3xx57jMcee4xFixZx3nnnsXXrVl599VXmz5/Pr3/9a77whS/w+9//noqKigxULiIyfPm3R3Ccv9wHYsC+vUfZc6STOdPLKYqefD66O3feeSef/vSn3zVt/fr1rFmzhi9+8Ytceuml3HXXXf0sQURkZGmPIDBpbAx3Z9/Rof/SOL0b6g9/+MPce++9HDp0CIDt27eze/duduzYQVlZGTfddBO3334769evf9e8IiJhyL89gpM0priIMbEoew53Mnls8ZBOGqd3Q3355Zfz8Y9/nAsuuACAcePG8eMf/5ht27Zx++23E4lEiMVifOtb3wLglltuYenSpcycOVMni0UkFObuYdcwJHV1dd7U1HTMuC1btjBnzpxhL7vtUAfb9x3l9KnjKCvOzYzM1HsVkcJiZuvcva6/aTo0lGZCWYyImbqnFpGCoiBIE41EqBgTY9+RLhLJ0bWnJCJysvImCDJ1iGvS2GKS7uw/iZPG2TbaDuOJyOiQF0FQWlpKW1tbRr4oy4qjlBRFc+7wkLvT1tZGaWlp2KWISJ7JzTOiQ1RdXU1LSwutra0ZWd7B9jj7j3ZxZHcJsWH8piDTSktLqa6uDrsMEckzeREEsViM2bNnZ2x5bYc6OP+fnuDPL5jFlz6iK3REJL/lzp+7OWTyuBIumzudB9e30KF7GotInlMQDKChvoa9R7p4bPOusEsREckqBcEALjp9ClUTxrBybXPYpYiIZJWCYACRiNFQX8Mftr3DW226cYyI5C8FwXFct7iaiMGqJu0ViEj+UhAcx8wJY3j/mZXct66ZeAbuVSAikosUBCfQUF/LrgMd/PaVzPxGQUQk1ygITuDSOVOZMq6E5c/p8JCI5CcFwQnEohGuW1zNky/vZteB9rDLERHJOAXBIDTU15BIOvevawm7FBGRjFMQDMLsKWN5z+xJrGpqJqnuqUUkzygIBunGJbW82XaEZ15vC7sUEZGMUhAM0tJ50xlfWsQK/dJYRPKMgmCQSmNRrllUxa827WRvjt2rQERkOBQEQ9C4pJbORJKHnt8edikiIhmjIBiCOTPGs6C6gpVrm3XbSBHJGwqCIWqor+XlXQfZ0Lwv7FJERDIiq0FgZkvN7GUz22Zmdxyn3bVm5mZWl816MuGqBTMYE4uqe2oRyRtZCwIziwJ3A5cDc4EbzWxuP+3Kgc8Az2arlkwqL41x1YIZrH5hB4c64mGXIyIybNncI1gCbHP31929E1gBXN1Pu/8B/G9g1PTf0FBfy5HOBL98YUfYpYiIDFs2g6AKSD9+0hKM62Fm5wE17v7w8RZkZreYWZOZNbW2ht8L6Hm1Ezhj6jj9pkBE8kJoJ4vNLAJ8Hfjcidq6+z3uXufudZWVldkv7gTMjMYltWxo3sfWnQfCLkdEZFiyGQTbgZq019XBuG7lwDzgKTN7AzgfWD0aThgDXLOoiuJohBXqnlpERrlsBsFa4Awzm21mxUAjsLp7orvvd/cp7j7L3WcBzwDL3L0pizVlzKSxxVx2zjQeen477V2JsMsRETlpWQsCd48DtwKPAluAVe6+2cy+YmbLsrXekdRYX8v+o108unln2KWIiJy0omwu3N3XAGv6jLtrgLYXZ7OWbHjvaZOpmTSGlWubuXph1YlnEBHJQfpl8TBEIkZDXQ1Pv9bGm22Hwy5HROSkKAiG6brFNUQM/dJYREYtBcEwTa8o5QNnTeW+dS3EE8mwyxERGTIFQQY0Lqml9WAHv9m6O+xSRESGTEGQAR84q5Kp5SU6PCQio5KCIAOKohGuW1zNky/vZuf+UdNlkogIoCDImIb6GpIO96/TXoGIjC4Kggw5ZfJY3nvaZFY2NZNM6u5lIjJ6KAgyqKG+huY9R3n6tbawSxERGTQFQQZ9+JzpTCiLsWLtW2GXIiIyaAqCDCqNRblmURWPbd7FnsOdYZcjIjIoCoIMa6ivoTOR5MH1LWGXIiIyKAqCDDt7+ngW1kxg5dpm3HXSWERyn4IgCxrra3h19yHWv7Uv7FJERE5IQZAFVy2YydjiKCue00ljEcl9CoIsGFtSxFULZvLLF9/mYHtX2OWIiByXgiBLGuprONqV4BcvvB12KSIix6UgyJKFNRM4e3o5K/WbAhHJcQqCLDEzGupreKFlPy/tOBB2OSIiA1IQZNE1i6ooLopor0BEcpqCIIsmlBWz9JzpPPT8dtq7EmGXIyLSLwVBljUuqeFAe5xfbdoZdikiIv1SEGTZ+bMnc8rkMpbrNwUikqMUBFkWiRg31NXw7J/28HrrobDLERF5FwXBCLh+cTXRiLGqSR3RiUjuURCMgKnjS7nk7Kncv66FrkQy7HJERI6hIBghjfU1vHOogye27A67FBGRYygIRsj7z6xk2vgS/aZARHKOgmCEFEUj3FBXw29faWXHvqNhlyMi0kNBMIJuqKsh6XCfThqLSA5REIygmkllXHT6FFY1NZNI6u5lIpIbFAQjrKG+hu37jvIf294JuxQREUBBMOIuO2caE8tirFzbHHYpIiKAgmDElRRF+dh51Tz20k7aDnWEXY6ISHaDwMyWmtnLZrbNzO7oZ/rfmNlGM9tgZn8ws7nZrCdXNNTX0JVwHly/PexSRESyFwRmFgXuBi4H5gI39vNF/1N3n+/uC4F/Br6erXpyyZnTyjmvdgIr1r6Fu04ai0i4srlHsATY5u6vu3snsAK4Or2Bu6ffumssUDDfio1Lanmt9TBNb+4NuxQRKXCDCgIz+4yZjbeU75vZejO77ASzVQHpZ0RbgnF9l/23ZvYaqT2C2wZY/y1m1mRmTa2trYMpOeddOX8G40qKWPGcThqLSLgGu0fwn4K/3i8DJgKfBL6aiQLc/W53Pw34AvDFAdrc4+517l5XWVmZidWGbmxJEVctmMnDG3dwoL0r7HJEpIANNggseL4C+JG7b04bN5DtQE3a6+pg3EBWAB8dZD154cYlNbR3JVm9YUfYpYhIARtsEKwzs8dIBcGjZlYOnKg/5bXAGWY228yKgUZgdXoDMzsj7eWVwKuDrCcvzK+qYM6M8axQR3QiEqLBBsFfAncA9e5+BIgBf3G8Gdw9DtwKPApsAVa5+2Yz+4qZLQua3Wpmm81sA/B3wKdO5k2MVmZGY30Nm7YfYNP2/WGXIyIFqmiQ7S4ANrj7YTO7CTgP+L8nmsnd1wBr+oy7K234M0OoNS99dGEV/7hmCyvXNjOvqiLsckSkAA12j+BbwBEzWwB8DngN+GHWqiogFWUxrpg/g59t2M7RzkTY5YhIARpsEMQ99cunq4F/c/e7gfLslVVYGuprONgeZ83Gt8MuRUQK0GCD4KCZ3UnqstGHzSxC6jyBZMB7Zk9i9pSx6ohOREIx2CBoADpI/Z5gJ6lLQb+WtaoKjJnRUF/Dc2/s4bXWQ2GXIyIFZlBBEHz5/wSoMLOPAO3urnMEGfSx86ooihirtFcgIiNssF1M3AA8B1wP3AA8a2bXZbOwQjO1vJRL50zl/nUtdMZP9BMNEZHMGeyhoX8g9RuCT7n7n5PqUO5L2SurMDXW19J2uJMntuwKuxQRKSCDDYKIu+9Oe902hHllkN53ZiUzKkpZocNDIjKCBvtl/isze9TMbjazm4GH6fNDMRm+aMS4vq6G373aSsveI2GXIyIFYrAni28H7gHODR73uPsXsllYobp+cTUA9zW1hFyJiBSKwXYxgbs/ADyQxVoEqJlUxkWnT+G+pmZuu/QMopETdfIqIjI8x90jMLODZnagn8dBMztwvHnl5N24pJYd+9v5/av5cRMeEcltxw0Cdy939/H9PMrdffxIFVloPjhnGpPGFuvuZSIyInTlTw4qLopw7XlVPL5lF60HO8IuR0TynIIgRzXU1xJPOg+u10ljEckuBUGOOn3qOOpnTWTl2mZSHb+KiGSHgiCHNdTX8vo7h3nuT3vCLkVE8piCIIddMX865SVF6p5aRLJKQZDDyoqLuHrRTB7e+Db7j3aFXY6I5CkFQY5rrK+lI57k5xu2h12KiOQpBUGOm1dVwTkzx7P8OZ00FpHsUBCMAo31NWx5+wCbtuvH3CKSeQqCUWDZwipKYxFWrH0r7FJEJA8pCEaBijExrpg/g59v2MGRznjY5YhInlEQjBKN9bUc6ojz8Itvh12KiOQZBcEoUT9rIqdWjtVvCkQk4xQEo4SZ0VhfQ9Obe9m2+2DY5YhIHlEQjCIfO6+aooipe2oRySgFwSgyZVwJH5o7jQef305HPBF2OSKSJxQEo0zjklr2HO7k8Zd2h12KiOQJBcEoc9HpU6iaMEa/KRCRjFEQjDLRiHF9XTW/f/UdmvccCbscEckDCoJR6Pq6GszgviadNBaR4VMQjEJVE8bw/jMrWdXUQiKpjuhEZHiyGgRmttTMXjazbWZ2Rz/T/87MXjKzF83sCTM7JZv15JPG+hp2Hmjnt6/opLGIDE/WgsDMosDdwOXAXOBGM5vbp9nzQJ27nwvcD/xzturJN5ecPY0p44r1mwIRGbZs7hEsAba5++vu3gmsAK5Ob+DuT7p79xnPZ4DqLNaTV4qLIly7uJontu5m98H2sMsRkVEsm0FQBaT/udoSjBvIXwKP9DfBzG4xsyYza2ptbc1giaNbQ10NiaTzwDrdvUxETl5OnCw2s5uAOuBr/U1393vcvc7d6yorK0e2uBx2auU4lsyexMq1b+nuZSJy0rIZBNuBmrTX1cG4Y5jZB4F/AJa5e0cW68lLjfU1vNF2hGde3xN2KSIySmUzCNYCZ5jZbDMrBhqB1ekNzGwR8B1SIaDLX07CFfNnUF5axEr90lhETlLWgsDd48CtwKPAFmCVu282s6+Y2bKg2deAccB9ZrbBzFYPsDgZQGksyjWLqlizaSf7j3SFXY6IjEJF2Vy4u68B1vQZd1fa8Aezuf5C0VBfww//+CYPPd/CzRfODrscERllcuJksQzPOTMrmF9VwYq1zTppLCJDpiDIE41Lati68yAvtuwPuxQRGWUUBHli2YKZjIlF1T21iAyZgiBPlJfGuPLcGazesIPDHfGwyxGRUURBkEduXFLD4c4ED7/4dtiliMgooiDII+fVTuT0qeN0eEhEhkRBkEfMjMb6Gta/tY9Xdh0MuxwRGSUUBHnmmkVVxKKm7qlFZNAUBHlm8rgSLjtnOg8+30JHPBF2OSIyCigI8lBjfQ37jnTx2OZdYZciIqOAgiAPXXjaFKonjtFJYxEZFAVBHopEjIa6Gv5jWxtvtR058QwiUtAUBHnqurpqIgarmnTSWESOT0GQp2ZUjOHis6Zy37pm4olk2OWISA5TEOSxhvoadh3o4KmXdZ9nERmYgiCPXXL2VCrLS1ixVoeHRGRgCoI8FotGuG5xNU++vJtdB9rDLkdEcpSCIM/dUFdDIuncv64l7FJEJEcpCPLc7CljOf/USaxc20wyqbuXici7KQgKwI1LanlrzxGeeb0t7FJEJAcpCArAh8+ZTsWYmE4ai0i/FAQFoDQW5ZpFVfxq0072Hu4MuxwRyTEKggLRUF9DZyLJQ89vD7sUEckxCoICMWfGeBbUTGDl2mbcddJYRHopCApIY30NL+86yPPN+8IuRURyiIKggFy1YCZlxVFW6u5lIpJGQVBAxpUUcdW5M/nFizs41BEPuxwRyREKggLTsKSGI50JfvnCjrBLEZEcoSAoMItqJnDmtHEs128KRCSgICgwZkZDfS0vNO9jy9sHwi5HRHKAgqAAfWxRFcXRCCu1VyAiKAgK0sSxxXx43nQeen477V2JsMsRkZApCApUY30N+4928ejmnWGXIiIhUxAUqAtOnUztpDJW6DcFIgUvq0FgZkvN7GUz22Zmd/Qz/X1mtt7M4mZ2XTZrkWNFIkZDfQ1/fL2NN945HHY5IhKirAWBmUWBu4HLgbnAjWY2t0+zt4CbgZ9mqw4Z2HWLq4kYrGrSXoFIIcvmHsESYJu7v+7uncAK4Or0Bu7+hru/CCSzWIcMYNr4Ui45eyr3rWuhK6F/ApFClc0gqALS/9RsCcYNmZndYmZNZtbU2tqakeIkpbG+ltaDHTy5dXfYpYhISEbFyWJ3v8fd69y9rrKyMuxy8srFZ1UytbxEvykQKWDZDILtQE3a6+pgnOSQomiE6+uqefLl3by9/2jY5YhICLIZBGuBM8xstpkVA43A6iyu7/iSCdANWfp1Q10NSYf7m1rCLkVEQpC1IHD3OHAr8CiwBVjl7pvN7CtmtgzAzOrNrAW4HviOmW3OVj003QtfOw1+eDU8+g/wwgrYuQniuofvKZPHcuHpk1nZ1EwyqbAUKTRF2Vy4u68B1vQZd1fa8FpSh4yyb8qZcNYVsGsTrP0exNtT4yMxqDwbps+D6fNhWvBcNmlEysoVDfW13Lb8eZ5+rY2LzpgSdjkiMoKyGgQ55dT3px4AiTi0bUuFws6NqefXnoQXlve2L5+ZCofuYJg+HyadCpFoOPVn2WVzpzGhLMbytW8pCEQKTOEEQbpoEUw9O/WYn/aD5kOtsGtj6pDRrk2p59d+A8ngbl6xMpg659hwmHYOlJSH8z4yqDQW5WOLqvnRM2+w53Ank8YWh12SiIyQwgyCgYyrhHGXwGmX9I6Ld0Dry717Djs3wpbVsP4HvW0mzuoNh+7nCbVgNuJvYTga6mu49z/+xIPrW/irPzs17HJEZIQoCE6kqARmnJt6dHOHAzuCYHixdw9i68NAcLK1pCK1t9BzeGkeTJ0LsTGhvI3BOGt6OYtqJ7BybTN/edFsbJQFmYicHAXByTCDiqrU48wP947vPAy7t6T2Grr3IDb8FDoPBfNFYPIZx557mDYPyqfnzN5DY30NX3hgI+vf2sviUwrrhLlIoVIQZFLxWKiuSz26JZOw74208w4boXktbHqgt03Z5GMPK02bB5VnQTQ24m/hI+fO5Cu/eIkVzzUrCEQKhIIg2yKR1NVGk06Fuct6xx/dB7s294bDzo3w3Hch0RHM131Z6/xj9yCyfFnr2JIili2cyc+e38FdV82lvHTkw0hERpaCICxjJsCsC1OPbv1e1voEvJDWS3f3Za3pexAZvqy1ob6W5c81s/qFHXziPadkbLkikpsUBLkkRy5rXVBdwdnTy/nJM28xb2YFpbEopbFI6rkoSkksQklRRCeTRfKE+Sjrf6eurs6bmprCLiN88Q5o3XrsuYddm+Do3t42w7is9Yd/fIO7fj5wjx9mUFLUGw6lsQgl3c+xaDA+cmyIBOO6p5ekTy+K9mnbvbzecbHoqOgsVyQnmdk6d6/rb5r2CEarohKYsSD16OYOB7YH4bBxWJe1fnxJLWdOK+dwR5z2riTtXQna44me4Y6uBO3x7uFkMK13+v6jXezuShuXNv1kRSOWFj6p55JjgiQtcIIgKRkgZLrHlfTZ0zk2wKJEI9rrkfynIMgnZlBRnXqctbR3fOdh2PXSseFwgstai6bN4/zZJ7isNZkET4Ang95dg9fdPb0eMy017MkEnV1ddHYl6OzqDJ67gkeczq4uuuJxOuNxurrixHueu4gnEnTFU+PiieA5niAZjxPviJOIJ0gk4iSSCRLxBMlknHgiycFkgkMkiZIkimPdw5bEcKIkieBEgvHdwxGSxMwpjjhFEYhFgmFzYuYUBcPdj6g5mAVzG24RnNRroOd1kggetEtND4bNUvMdM52edkmL9M5jkdQm7llm7/roXkfauJ51ps3z7jqPna+nFqynjmTfaWa4W591dQ9D0vvU4d3vLfXADLNI8Nz9iICBEekdF42ARVLvznrHd88fiRz72syImKXm6W4TLMcxIhYhYsGyIql2ETMikdR4zIgYqXEGhmHdryN9Xne3MXrW2+9ruscHy45Y8NzSrCkAAAeASURBVH5612UDPHfPN2VcCRVjMn8Bh4KgEBSPhZr61KNbz2WtaeHQ97LW4vJUSHjfL/xg+CQYUBI8RkQ0eAyg+0uK4Es2aRHcoj1fij0Pi5B0I+EREh4MEyHhRgIj7hG6O26N4MR6v1K7vz57vlYjx37lBtN7v777Trd+xkV1d9esSnp3+HX/C9ATst3Dvf+qFuxv9z/PsdPfPT7Z/YfCAOtMrTc1/OKiz/Jn13w64+9XQVCojrmsNe1W0t2Xte7cCHv/lBpn0dSfN5FoMBwJhiOp15FI73DPtJOdJxJMy9Q8QR39zhNN/bWYtllGVZeC7sEj+e4H6eP7thlgHvc+8/UzfVDrS1/HCZaHD+65Zx3p0xjUvN7zSPY80/Oantekt+tn3tQ3c+9yumvytHYE7ZzuNr3L6X0vjgWvzVMh78kk0F1Hav6e7U1vjaefkn6vr8xREMix+rusVXJTcChklNxxNjQWPGRg+h8kIlLgFAQiIgVOQSAiUuAUBCIiBU5BICJS4BQEIiIFTkEgIlLgFAQiIgVu1PU+amatwJsnOfsU4J0MlpMpqmtoVNfQ5WptqmtohlPXKe5e2d+EURcEw2FmTQN1wxom1TU0qmvocrU21TU02apLh4ZERAqcgkBEpMAVWhDcE3YBA1BdQ6O6hi5Xa1NdQ5OVugrqHIGIiLxboe0RiIhIHwoCEZECl5dBYGZLzexlM9tmZnf0M73EzFYG0581s1k5UtfNZtZqZhuCx1+NUF33mtluM9s0wHQzs28Gdb9oZuflSF0Xm9n+tO111wjUVGNmT5rZS2a22cw+00+bEd9eg6wrjO1VambPmdkLQV1f7qfNiH8eB1lXKJ/HYN1RM3vezH7Zz7TMby93z6sHqbsNvgacChQDLwBz+7T5L8C3g+FGYGWO1HUz8G8hbLP3AecBmwaYfgXwCKkbPZ0PPJsjdV0M/HKEt9UM4LxguBx4pZ9/xxHfXoOsK4ztZcC4YDgGPAuc36dNGJ/HwdQVyucxWPffAT/t798rG9srH/cIlgDb3P11d+8EVgBX92lzNfCDYPh+4FIzy/bd7AZTVyjc/XfAnuM0uRr4oac8A0wwsxk5UNeIc/e33X19MHwQ2AJU9Wk24ttrkHWNuGAbHApexoJH3ytURvzzOMi6QmFm1cCVwPcGaJLx7ZWPQVAFNKe9buHdH4ieNu4eB/YDk3OgLoBrg8MJ95tZdu5UPXSDrT0MFwS794+Y2TkjueJgl3wRqb8m04W6vY5TF4SwvYLDHBuA3cCv3X3A7TWCn8fB1AXhfB6/Afw3IDnA9Ixvr3wMgtHsF8Asdz8X+DW9qS/9W0+q/5QFwL8CPxupFZvZOOAB4LPufmCk1nsiJ6grlO3l7gl3XwhUA0vMbN5IrPdEBlHXiH8ezewjwG53X5ftdaXLxyDYDqQnd3Uwrt82ZlYEVABtYdfl7m3u3hG8/B6wOMs1DdZgtumIc/cD3bv37r4GiJnZlGyv18xipL5sf+LuD/bTJJTtdaK6wtpeaevfBzwJLO0zKYzP4wnrCunzeCGwzMzeIHX4+BIz+3GfNhnfXvkYBGuBM8xstpkVkzqZsrpPm9XAp4Lh64DfeHDmJcy6+hxHXkbqOG8uWA38eXA1zPnAfnd/O+yizGx697FRM1tC6v9zVr9AgvV9H9ji7l8foNmIb6/B1BXS9qo0swnB8BjgQ8DWPs1G/PM4mLrC+Dy6+53uXu3us0h9R/zG3W/q0yzj26toODPnInePm9mtwKOkrtS51903m9lXgCZ3X03qA/MjM9tG6mRkY47UdZuZLQPiQV03Z7suADNbTuqKkilm1gL8d1Inz3D3bwNrSF0Jsw04AvxFjtR1HfCfzSwOHAUaRyDQLwQ+CWwMji8D/D1Qm1ZXGNtrMHWFsb1mAD8wsyip4Fnl7r8M+/M4yLpC+Tz2J9vbS11MiIgUuHw8NCQiIkOgIBARKXAKAhGRAqcgEBEpcAoCEZECpyAQGUGW6gH0XT1KioRJQSAiUuAUBCL9MLObgv7qN5jZd4IOyg6Z2b8E/dc/YWaVQduFZvZM0DnZQ2Y2MRh/upk9HnTytt7MTgsWPy7oxGyrmf1kBHq+FTkuBYFIH2Y2B2gALgw6JUsAnwDGkvp15znAb0n90hngh8AXgs7JNqaN/wlwd9DJ23uB7m4mFgGfBeaSuj/FhVl/UyLHkXddTIhkwKWkOhhbG/yxPoZUV8VJYGXQ5sfAg2ZWAUxw998G438A3Gdm5UCVuz8E4O7tAMHynnP3luD1BmAW8Ifsvy2R/ikIRN7NgB+4+53HjDT7Up92J9s/S0facAJ9DiVkOjQk8m5PANeZ2VQAM5tkZqeQ+rxcF7T5OPAHd98P7DWzPwvGfxL4bXCXsBYz+2iwjBIzKxvRdyEySPpLRKQPd3/JzL4IPGZmEaAL+FvgMKkbmHyR1KGihmCWTwHfDr7oX6e3t9FPAt8Jeo7sAq4fwbchMmjqfVRkkMzskLuPC7sOkUzToSERkQKnPQIRkQKnPQIRkQKnIBARKXAKAhGRAqcgEBEpcAoCEZEC9/8Bzw1h96oeS7EAAAAASUVORK5CYII=\n"},"metadata":{"needs_background":"light"}}]},{"cell_type":"code","source":[],"metadata":{"id":"ZAglUMUJMakS"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":[],"metadata":{"id":"jlp5KWOSMamL"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":[],"metadata":{"id":"9-K4F0VOMao7"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":[],"metadata":{"id":"fr41TKb3Maqs"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":[],"metadata":{"id":"wihjhai8MauK"},"execution_count":null,"outputs":[]},{"cell_type":"code","source":[],"metadata":{"id":"Il0I0FezJhh5"},"execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"display_name":"Python 3","language":"python","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":"3.7.6"},"colab":{"provenance":[]},"accelerator":"GPU","gpuClass":"standard"},"nbformat":4,"nbformat_minor":0}