exercise-dropout-and-batch-normalization.ipynb

{"metadata":{"jupytext":{"cell_metadata_filter":"-all","formats":"ipynb"},"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"name":"python","version":"3.10.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"gpu","dataSources":[{"sourceId":1480608,"sourceType":"datasetVersion","datasetId":829369}],"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"**This notebook is an exercise in the [Intro to Deep Learning](https://www.kaggle.com/learn/intro-to-deep-learning) course.  You can reference the tutorial at [this link](https://www.kaggle.com/ryanholbrook/dropout-and-batch-normalization).**\n\n---\n","metadata":{}},{"cell_type":"markdown","source":"# Introduction #\n\nIn this exercise, you'll add dropout to the *Spotify* model from Exercise 4 and see how batch normalization can let you successfully train models on difficult datasets.\n\nRun the next cell to get started!","metadata":{}},{"cell_type":"code","source":"# Setup plotting\nimport matplotlib.pyplot as plt\nplt.style.use('seaborn-v0_8')\n# Set Matplotlib defaults\nplt.rc('figure', autolayout=True)\nplt.rc('axes', labelweight='bold', labelsize='large',\n       titleweight='bold', titlesize=18, titlepad=10)\nplt.rc('animation', html='html5')\n\n# Setup feedback system\nfrom learntools.core import binder\nbinder.bind(globals())\nfrom learntools.deep_learning_intro.ex5 import *","metadata":{"execution":{"iopub.status.busy":"2024-01-05T02:31:08.619844Z","iopub.execute_input":"2024-01-05T02:31:08.620193Z","iopub.status.idle":"2024-01-05T02:31:09.045986Z","shell.execute_reply.started":"2024-01-05T02:31:08.620165Z","shell.execute_reply":"2024-01-05T02:31:09.045166Z"},"trusted":true},"execution_count":1,"outputs":[]},{"cell_type":"markdown","source":"First load the *Spotify* dataset.","metadata":{}},{"cell_type":"code","source":"import pandas as pd\nfrom sklearn.preprocessing import StandardScaler, OneHotEncoder\nfrom sklearn.compose import make_column_transformer\nfrom sklearn.model_selection import GroupShuffleSplit\n\nfrom tensorflow import keras\nfrom tensorflow.keras import layers\nfrom tensorflow.keras import callbacks\n\nspotify = pd.read_csv('../input/dl-course-data/spotify.csv')\n\nX = spotify.copy().dropna()\ny = X.pop('track_popularity')\nartists = X['track_artist']\n\nfeatures_num = ['danceability', 'energy', 'key', 'loudness', 'mode',\n                'speechiness', 'acousticness', 'instrumentalness',\n                'liveness', 'valence', 'tempo', 'duration_ms']\nfeatures_cat = ['playlist_genre']\n\npreprocessor = make_column_transformer(\n    (StandardScaler(), features_num),\n    (OneHotEncoder(), features_cat),\n)\n\ndef group_split(X, y, group, train_size=0.75):\n    splitter = GroupShuffleSplit(train_size=train_size)\n    train, test = next(splitter.split(X, y, groups=group))\n    return (X.iloc[train], X.iloc[test], y.iloc[train], y.iloc[test])\n\nX_train, X_valid, y_train, y_valid = group_split(X, y, artists)\n\nX_train = preprocessor.fit_transform(X_train)\nX_valid = preprocessor.transform(X_valid)\ny_train = y_train / 100\ny_valid = y_valid / 100\n\ninput_shape = [X_train.shape[1]]\nprint(\"Input shape: {}\".format(input_shape))","metadata":{"execution":{"iopub.status.busy":"2024-01-05T02:31:09.047539Z","iopub.execute_input":"2024-01-05T02:31:09.047965Z","iopub.status.idle":"2024-01-05T02:31:22.997916Z","shell.execute_reply.started":"2024-01-05T02:31:09.047937Z","shell.execute_reply":"2024-01-05T02:31:22.996808Z"},"trusted":true},"execution_count":2,"outputs":[{"name":"stderr","text":"/opt/conda/lib/python3.10/site-packages/scipy/__init__.py:146: UserWarning: A NumPy version >=1.16.5 and <1.23.0 is required for this version of SciPy (detected version 1.24.3\n  warnings.warn(f\"A NumPy version >={np_minversion} and <{np_maxversion}\"\n","output_type":"stream"},{"name":"stdout","text":"Input shape: [18]\n","output_type":"stream"}]},{"cell_type":"markdown","source":"# 1) Add Dropout to Spotify Model\n\nHere is the last model from Exercise 4. Add two dropout layers, one after the `Dense` layer with 128 units, and one after the `Dense` layer with 64 units. Set the dropout rate on both to `0.3`.","metadata":{}},{"cell_type":"code","source":"# YOUR CODE HERE: Add two 30% dropout layers, one after 128 and one after 64\nmodel = keras.Sequential([\n    layers.Dense(128, activation='relu', input_shape=input_shape),\n    layers.Dropout(rate=0.3),\n    layers.Dense(64, activation='relu'),\n    layers.Dropout(rate=0.3),\n    layers.Dense(1)\n])\n\n# Check your answer\nq_1.check()","metadata":{"lines_to_next_cell":2,"execution":{"iopub.status.busy":"2024-01-05T02:32:14.681124Z","iopub.execute_input":"2024-01-05T02:32:14.682101Z","iopub.status.idle":"2024-01-05T02:32:14.742756Z","shell.execute_reply.started":"2024-01-05T02:32:14.682068Z","shell.execute_reply":"2024-01-05T02:32:14.741812Z"},"trusted":true},"execution_count":13,"outputs":[{"output_type":"display_data","data":{"text/plain":"<IPython.core.display.Javascript object>","application/javascript":"parent.postMessage({\"jupyterEvent\": \"custom.exercise_interaction\", \"data\": {\"outcomeType\": 1, \"valueTowardsCompletion\": 0.25, \"interactionType\": 1, \"questionType\": 2, \"questionId\": \"1_Q1\", \"learnToolsVersion\": \"0.3.4\", \"failureMessage\": \"\", \"exceptionClass\": \"\", \"trace\": \"\"}}, \"*\")"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"Correct","text/markdown":"<span style=\"color:#33cc33\">Correct</span>"},"metadata":{}}]},{"cell_type":"code","source":"# Lines below will give you a hint or solution code\n#q_1.hint()\n#q_1.solution()","metadata":{"execution":{"iopub.status.busy":"2024-01-05T02:31:23.993152Z","iopub.execute_input":"2024-01-05T02:31:23.993489Z","iopub.status.idle":"2024-01-05T02:31:23.997563Z","shell.execute_reply.started":"2024-01-05T02:31:23.993460Z","shell.execute_reply":"2024-01-05T02:31:23.996667Z"},"trusted":true},"execution_count":4,"outputs":[]},{"cell_type":"markdown","source":"Now run this next cell to train the model see the effect of adding dropout.","metadata":{}},{"cell_type":"code","source":"model.compile(\n    optimizer='adam',\n    loss='mae',\n)\nhistory = model.fit(\n    X_train, y_train,\n    validation_data=(X_valid, y_valid),\n    batch_size=512,\n    epochs=50,\n    verbose=0,\n)\nhistory_df = pd.DataFrame(history.history)\nhistory_df.loc[:, ['loss', 'val_loss']].plot()\nprint(\"Minimum Validation Loss: {:0.4f}\".format(history_df['val_loss'].min()))","metadata":{"execution":{"iopub.status.busy":"2024-01-05T02:32:21.501616Z","iopub.execute_input":"2024-01-05T02:32:21.502404Z","iopub.status.idle":"2024-01-05T02:32:25.188292Z","shell.execute_reply.started":"2024-01-05T02:32:21.502370Z","shell.execute_reply":"2024-01-05T02:32:25.187303Z"},"trusted":true},"execution_count":14,"outputs":[{"name":"stdout","text":"Minimum Validation Loss: 11.4386\n","output_type":"stream"},{"output_type":"display_data","data":{"text/plain":"<Figure size 800x550 with 1 Axes>","image/png":"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"},"metadata":{}}]},{"cell_type":"markdown","source":"# 2) Evaluate Dropout\n\nRecall from Exercise 4 that this model tended to overfit the data around epoch 5. Did adding dropout seem to help prevent overfitting this time?","metadata":{}},{"cell_type":"code","source":"# View the solution (Run this cell to receive credit!)\nq_2.check()","metadata":{"execution":{"iopub.status.busy":"2024-01-05T02:31:35.740919Z","iopub.execute_input":"2024-01-05T02:31:35.741313Z","iopub.status.idle":"2024-01-05T02:31:35.749859Z","shell.execute_reply.started":"2024-01-05T02:31:35.741269Z","shell.execute_reply":"2024-01-05T02:31:35.748908Z"},"trusted":true},"execution_count":6,"outputs":[{"output_type":"display_data","data":{"text/plain":"<IPython.core.display.Javascript object>","application/javascript":"parent.postMessage({\"jupyterEvent\": \"custom.exercise_interaction\", \"data\": {\"outcomeType\": 1, \"valueTowardsCompletion\": 0.25, \"interactionType\": 1, \"questionType\": 4, \"questionId\": \"2_Q2\", \"learnToolsVersion\": \"0.3.4\", \"failureMessage\": \"\", \"exceptionClass\": \"\", \"trace\": \"\"}}, \"*\")"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"Correct: \n\n\nFrom the learning curves, you can see that the validation loss remains near a constant minimum even though the training loss continues to decrease. So we can see that adding dropout did prevent overfitting this time. Moreover, by making it harder for the network to fit spurious patterns, dropout may have encouraged the network to seek out more of the true patterns, possibly improving the validation loss some as well).","text/markdown":"<span style=\"color:#33cc33\">Correct:</span> \n\n\nFrom the learning curves, you can see that the validation loss remains near a constant minimum even though the training loss continues to decrease. So we can see that adding dropout did prevent overfitting this time. Moreover, by making it harder for the network to fit spurious patterns, dropout may have encouraged the network to seek out more of the true patterns, possibly improving the validation loss some as well).\n"},"metadata":{}}]},{"cell_type":"markdown","source":"Now, we'll switch topics to explore how batch normalization can fix problems in training.\n\nLoad the *Concrete* dataset. We won't do any standardization this time. This will make the effect of batch normalization much more apparent.","metadata":{}},{"cell_type":"code","source":"import pandas as pd\n\nconcrete = pd.read_csv('../input/dl-course-data/concrete.csv')\ndf = concrete.copy()\n\ndf_train = df.sample(frac=0.7, random_state=0)\ndf_valid = df.drop(df_train.index)\n\nX_train = df_train.drop('CompressiveStrength', axis=1)\nX_valid = df_valid.drop('CompressiveStrength', axis=1)\ny_train = df_train['CompressiveStrength']\ny_valid = df_valid['CompressiveStrength']\n\ninput_shape = [X_train.shape[1]]","metadata":{"execution":{"iopub.status.busy":"2024-01-05T02:32:41.294092Z","iopub.execute_input":"2024-01-05T02:32:41.294881Z","iopub.status.idle":"2024-01-05T02:32:41.309768Z","shell.execute_reply.started":"2024-01-05T02:32:41.294842Z","shell.execute_reply":"2024-01-05T02:32:41.308702Z"},"trusted":true},"execution_count":15,"outputs":[]},{"cell_type":"markdown","source":"Run the following cell to train the network on the unstandardized *Concrete* data.","metadata":{}},{"cell_type":"code","source":"model = keras.Sequential([\n    layers.Dense(512, activation='relu', input_shape=input_shape),\n    layers.Dense(512, activation='relu'),    \n    layers.Dense(512, activation='relu'),\n    layers.Dense(1),\n])\nmodel.compile(\n    optimizer='sgd', # SGD is more sensitive to differences of scale\n    loss='mae',\n    metrics=['mae'],\n)\nhistory = model.fit(\n    X_train, y_train,\n    validation_data=(X_valid, y_valid),\n    batch_size=64,\n    epochs=100,\n    verbose=0,\n)\n\nhistory_df = pd.DataFrame(history.history)\nhistory_df.loc[0:, ['loss', 'val_loss']].plot()\nprint((\"Minimum Validation Loss: {:0.4f}\").format(history_df['val_loss'].min()))","metadata":{"execution":{"iopub.status.busy":"2024-01-05T02:33:05.490821Z","iopub.execute_input":"2024-01-05T02:33:05.491522Z","iopub.status.idle":"2024-01-05T02:33:13.780234Z","shell.execute_reply.started":"2024-01-05T02:33:05.491487Z","shell.execute_reply":"2024-01-05T02:33:13.779182Z"},"trusted":true},"execution_count":17,"outputs":[{"name":"stdout","text":"Minimum Validation Loss: 693.7224\n","output_type":"stream"},{"output_type":"display_data","data":{"text/plain":"<Figure size 800x550 with 1 Axes>","image/png":"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"},"metadata":{}}]},{"cell_type":"markdown","source":"Did you end up with a blank graph? Trying to train this network on this dataset will usually fail. Even when it does converge (due to a lucky weight initialization), it tends to converge to a very large number.\n\n# 3) Add Batch Normalization Layers\n\nBatch normalization can help correct problems like this.\n\nAdd four `BatchNormalization` layers, one before each of the dense layers. (Remember to move the `input_shape` argument to the new first layer.)","metadata":{}},{"cell_type":"code","source":"# YOUR CODE HERE: Add a BatchNormalization layer before each Dense layer\nmodel = keras.Sequential([\n    layers.BatchNormalization(input_shape=input_shape),\n    layers.Dense(512, activation='relu'),\n    layers.BatchNormalization(),\n    layers.Dense(512, activation='relu'),\n    layers.BatchNormalization(),\n    layers.Dense(512, activation='relu'),\n    layers.BatchNormalization(),\n    layers.Dense(1),\n])\n\n# Check your answer\nq_3.check()","metadata":{"lines_to_next_cell":0,"execution":{"iopub.status.busy":"2024-01-05T02:34:33.772887Z","iopub.execute_input":"2024-01-05T02:34:33.773705Z","iopub.status.idle":"2024-01-05T02:34:33.902958Z","shell.execute_reply.started":"2024-01-05T02:34:33.773670Z","shell.execute_reply":"2024-01-05T02:34:33.902024Z"},"trusted":true},"execution_count":18,"outputs":[{"output_type":"display_data","data":{"text/plain":"<IPython.core.display.Javascript object>","application/javascript":"parent.postMessage({\"jupyterEvent\": \"custom.exercise_interaction\", \"data\": {\"outcomeType\": 1, \"valueTowardsCompletion\": 0.25, \"interactionType\": 1, \"questionType\": 2, \"questionId\": \"3_Q3\", \"learnToolsVersion\": \"0.3.4\", \"failureMessage\": \"\", \"exceptionClass\": \"\", \"trace\": \"\"}}, \"*\")"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"Correct","text/markdown":"<span style=\"color:#33cc33\">Correct</span>"},"metadata":{}}]},{"cell_type":"code","source":"# Lines below will give you a hint or solution code\n#q_3.hint()\n#q_3.solution()","metadata":{"execution":{"iopub.status.busy":"2024-01-05T02:31:44.568550Z","iopub.execute_input":"2024-01-05T02:31:44.568853Z","iopub.status.idle":"2024-01-05T02:31:44.573182Z","shell.execute_reply.started":"2024-01-05T02:31:44.568827Z","shell.execute_reply":"2024-01-05T02:31:44.572197Z"},"trusted":true},"execution_count":10,"outputs":[]},{"cell_type":"markdown","source":"Run the next cell to see if batch normalization will let us train the model.","metadata":{}},{"cell_type":"code","source":"model.compile(\n    optimizer='sgd',\n    loss='mae',\n    metrics=['mae'],\n)\nEPOCHS = 100\nhistory = model.fit(\n    X_train, y_train,\n    validation_data=(X_valid, y_valid),\n    batch_size=64,\n    epochs=EPOCHS,\n    verbose=0,\n)\n\nhistory_df = pd.DataFrame(history.history)\nhistory_df.loc[0:, ['loss', 'val_loss']].plot()\nprint((\"Minimum Validation Loss: {:0.4f}\").format(history_df['val_loss'].min()))","metadata":{"execution":{"iopub.status.busy":"2024-01-05T02:34:39.005812Z","iopub.execute_input":"2024-01-05T02:34:39.006230Z","iopub.status.idle":"2024-01-05T02:34:52.136420Z","shell.execute_reply.started":"2024-01-05T02:34:39.006171Z","shell.execute_reply":"2024-01-05T02:34:52.135298Z"},"trusted":true},"execution_count":19,"outputs":[{"name":"stdout","text":"Minimum Validation Loss: 3.7919\n","output_type":"stream"},{"output_type":"display_data","data":{"text/plain":"<Figure size 800x550 with 1 Axes>","image/png":"iVBORw0KGgoAAAANSUhEUgAAAxUAAAIcCAYAAACAZYN2AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8WgzjOAAAACXBIWXMAAA9hAAAPYQGoP6dpAACadklEQVR4nOzdd3jb1d3+8beW997b2Xb2JpuwAmGEvSmUltJBgS5aaH9tKTwt0DJaHlYHtE+BQtmUVXYY2XsnznK897ZlW/P3h2wlJvGMbcn2/boursTSV9LH4XjcOudzjsHtdrsRERERERHpI6OvCxARERERkaFNoUJERERERE6KQoWIiIiIiJwUhQoRERERETkpChUiIiIiInJSFCpEREREROSkKFSIiIiIiMhJUagQEREREZGTolAhIiIiIiInxezrAk6koqLB1yVgNBqIiQmluroJl0uHjktHGh/SHY0R6YrGh3RHY0S6MtjjIz4+vPuaBryKIcpoNGAwGDAaDb4uRfyQxod0R2NEuqLxId3RGJGu+OP4UKgQEREREZGTolAhIiIiIiInRaFCREREREROikKFiIiIiIicFIUKERERERE5KQoVIiIiIiJyUhQqRERERETkpChUiIiIiIjISVGoEBERERGRk6JQISIiIiIiJ0WhQkRERERETopChYiIiIjICdx667d59NGHfV3GkKBQISIiIiIiJ0WhQkRERERETorZ1wWIiIiIiPi7+vp6Hn30IVav/hK73caMGbP54Q/vID09A4DS0hIeeeQP7NixDYfDTlJSCt///u0sWLCY+vp6/vjHP7Bx4zqs1mYSEhK4/vpvcP75F/r4s+o/ChUiIiIiMuisLQ5KqpsG9TWTY0IJCerbr7/33fcbCgsL+P3vHyEkJJSnnnqMn/70Bzz//CuYzWYeeeT32O12nnjibwQFBXHkSC7BwSEAPP30Uxw5cpiHHvpfIiOjKCwsoLW1tT8/NZ9TqBARERGRQWVtcfCzp9ZgbXUM6uuGBJr5w/cW9jpYFBTks2rVFzz11DNMnTodgLvv/h8uvfR8vvjiM8444yzKykpZuvQMxo4dB0Bqapr38WVlpYwfn0V29iQAkpNT+ukz8h8KFSIiIiIiXcjLy8VkMjFp0hTvbZGRUWRkZJKXlwvA5ZdfzUMP3c/GjeuYM2ceS5eewbhx4wG4+OLL+eUvf8b+/Tmccso8liw5zRtOhguFChEREREZVCFBnhmDobT8qTsrVlzMKafMZ+3aVWzYsJ7nnvsHt976Qy6//GoWLFjEq6++w7p1q9m4cT0/+MEtXHrpFdx66w8HpBZfUKjoxJfbi2lqdXLqtGQCzNokS0RERKQ/hQSZGZsS6esyeiQzczROp5M9e3Z5Zxjq6mrJz89j1KjR3usSE5O4+OLLufjiy/nznx/n7bff5PLLrwYgOjqac8+9gHPPvYBp02bw5JP/q1Ax3NkdLp55Zy8ut5uPN+TznYsmMzo5wtdliYiIiIgPpKdnsGTJUn7/+9/x05/+gpCQEP7858eJj09gyZLTAHj00YeZP38h6ekZNDQ0sGXLJjIzPYHj6af/TFZWNqNHj8Vms7FmzSoyM0f57hMaAAoVJ2AxG1k2N40PNhRQXtvMfc9t5orTxrJsbjoGg8HX5YmIiIjIIPv5z+/m0Ucf4s47f4jdbmf69Fk8+OCjmM2eX6ddLiePPPJ7KirKCQkJZd68Bdx++48BMJvN/OUvT1BSUkxgYBDTp8/gnnvu8+Wn0+8Mbrfb7esivqqiosHXJWA2G9l6qJonX91Oq90JwPSxsdx0wSTCgi0+rk58zWw2Eh0dSk1NEw6Hy9fliB/SGJGuaHxIdzRGpCuDPT7i48O7vaZXzQIvvPACK1asYNasWcyaNYurrrqKzz//3Hv/9ddfT1ZWVof/fv3rX/e+cj9xxpx07v3WKaTFhwGw/VAVd/99A/sLan1bmIiIiIiIH+nV8qekpCTuuOMOMjMzcbvdvPnmm3z/+9/njTfeYPx4z5ZZV155Jbfffrv3McHBwf1b8SBLjg3lV1+fzb8/PcjKLUXUNLTy+xe2cPGSMZy/IBOjlkOJiIiIyAjXq1BxxhlndPj4Rz/6ES+++CLbtm3zhoqgoCDi4+P7r0I/YDGbuP7sLCZmRPOP/+6judXBG18c5lBRHTevmERokJZDiYiIiMjI1ee9Up1OJ++++y5Wq5WZM2d6b3/77beZN28eF1xwAQ8//DDNzc39Uqg/mJOdwG++MZfRyZ51ZTsOVXHPPzaSV+r7HhAREREREV/p9e5POTk5XH311bS2thISEsITTzzBuHGe48gvuOACUlJSSEhIICcnh4ceeojc3Fwef/zxXr2G0WjAaPTtsiKTydjhz3bJcaH88utzef7DHFZuKaKyroX7nt/Mjedms2T68DtyXU6ss/Eh0k5jRLqi8SHd0RiRrvjj+Oj17k82m42SkhIaGhr44IMPeOWVV3j++ee9weJYa9eu5cYbb+Sjjz4iIyOjx6/hdruHxNatH2/I56nXtmNr67o/Z34m37lkKhazyceViYiIiIgMnpPeUvbGG28kIyODe++997j72pdGPf300yxZsqTHz1lV1egXMxUREcHU1zfjdHa+VVdeaQP/++oOKmo9y7zGpERw62XTiIsMGqxSxQd6Oj5k5NIYka5ofEh3NEakK4M9PqKjQ7u95qQPv3O5XNhsthPet3fvXoBeN267XG5cLv84PsPpdHW5/29qXCi/vnEOf3t7DzsOVXG4uJ67n1nPbZdNY1zq0Dh6Xvquu/EhojEiXdH4kO5ojEhX/Gl89Goh1sMPP8zGjRspLCwkJyeHhx9+mA0bNrBixQry8/N54okn2LVrF4WFhXzyySfceeedzJ07l+zs7IGq3y+EBlm4/fJpXLxkNAagwWrnDy9sZf2eMl+XJiIiIiIy4Ho1U1FVVcWdd95JeXk54eHhZGVl8cwzz7Bo0SJKSkpYu3Ytzz77LFarleTkZM4++2xuueWWgardrxgNBi5cNJrUuDD+9vZubA4Xf3lrN2U1VlYsHDUkekREREREpH9dfvkKrrzyGq688tpur128eA733fcQp5562sAX1s96FSruu+++Tu9LTk7m+eefP+mChrrZWfHERs7i0Vd3UNdo480vcymrtnLjuROxmP2nQ19EREREpL/ot9wBMCopgl/dMIeMhDAA1u4u46F/b6XBeuLeExERERGRoUyhYoDERARx19dmMWNcHAAHCuv43bObKalq8nFlIiIiItIT//nP61x00XJcro7N0Hfd9WPuu+8eiooKueuuH7NixdksW7aEb33rBjZuXN9vr3/o0EFuv/27nHHGIs4770x+//vfYbVavfdv3ryJm2++gbPOWszy5afxve99k9LSEgAOHNjPbbd9h2XLTuXss5fyzW9+jX379vRbbV910rs/SeeCAszceulUXl55kA83FlBe28wD/9rCT66aQUZiuK/LExEREfGZZkczpU0Vg/qaSaHxBJuDe3z96aefxZ/+9CBbtmxizpxTAKivr2P9+rU8+OCjWK1W5s9fxLe/fQsWSwDvv/8ud975Y1544TWSkpJOqtbm5mZ+/ONbmTJlKk8//U9qamp44IHf8sc//oG7774Xh8PBnXf+mBUrLuE3v7kPu93O3r27AU8f7733/pIJE7K4446fYzQaOXBgPybTwP3qr1AxwIxGA1efOZ7EmBCe/yDHuzPUj66aztgUbTkrIiIiI0+zo5lfrXmAZkfzoL5usDmY/1l4V4+DRUREBPPnL+Sjj973hoqVKz8hMjKKWbPmYDQaGT9+gvf6m2/+Hl98sZLVqz/nssuuOqlaP/rofWw2G7/85b0EB3vq/fGPf8qdd/6YW2/9AQZDBI2NjSxcuJjU1DQARo0a7X18WVkZ1157A5mZowBIT+/5QdR9oeVPg+T0mal8+8LJGA0GrK0OHvr3NnLya3xdloiIiIh0Ydmyc/n880+957J99NH7nHnm2RiNRqxWK48//ieuu+5yli8/jWXLlpCXd4SystKTft28vFzGjRvvDRQAU6fOwOVykZ9/hKioKM4/fwU/+clt/OxnP+Lll1+ksrLSe+1VV13LAw/8Dz/4wS0899z/UVRUeNI1dUUzFYNo3qREAsxGnvrPLlptTh55eTu3XjqVqWNifV2aiIiIyKBpnzHw9+VPAIsWLcHtdrN27SqysyexfftWbrvtxwA88cSf2LhxPd///g9JS0snMDCQX/7yTux2x0CUf5xf/eoeLrvsKtavX8unn37E3/72FH/84xNMmTKVm276DsuWLWft2lWsW7eGv//9L/zmN/exdOnpA1KLQsUgmzkhntsvn8bjr+3E5nDxv6/u4LsXTWF2Vu9OHRcREREZyoLNwYyOHNglOf0hMDCQpUvP4MMP/0thYQEZGZlkZXkOdt65czvnnbfC+4u61WqltLQYmH3Sr5uZOZr33nuH5uZm72zFzp3bMBqNZGSM8l43YUI2EyZkc/313+A73/kGH3/8PlOmTAUgIyOTjIxMrrrqOu6++xe8995bAxYqtPzJB6aMjuXHV80gKMCE0+XmqTd3sXb3yU+TiYiIiEj/87zjv5p3332LZcuWe29PS8vg888/5cCBHA4c2M899/w/XC53v7zm2WefS0BAAL/73d0cPnyQLVs28cc/Psg555xHbGwsBQUFPPnkY+zatYPS0hI2bFhHYWE+mZmjaW1t4ZFHfs+WLZsoLS1hx45t7Nu3h8zM0d2/cB9ppsJHJqRHccfVM/njy9toanHw9Nt7cLncLJqa7OvSREREROQYs2fPJTw8gvz8vA6h4rbbfsT999/Ld7/7TSIjo7juuq/T1NQ/xwcEBQXxyCOP8+ijD/Gtb32doKAgli49g9tu+xEAwcHB5OUd4d1336a+vo7Y2DguvfRKLrroUpxOJ3V1dfz2t3dTU1NNZGQUS5eezk03fadfajsRg9vt7p841Y8qKhp8XQJms5Ho6FBqappwOI7uTWy1W1lVvJ5WRyvpEWlkhqcRFRiJwWDo0+sUlDfy8L+3Um+1YzQYuO2yqUxvO9tC/Fdn40OkncaIdEXjQ7qjMSJdGezxER/f/VEImqnoIafLyRdFa/lv7sc0Oawd7osICCcjPI3MiDQyI9IZGzmaIHNgj543PSGMO66ZyQPPb8Ha6uCpN3dxxzUzGZeq7WZFREREZGhQqOiG2+1mR8Vu3jj0LuXWo9t0WYwW7C47APW2BnZV7WVX1V4AQi0hnJN5BqemLsBisnT7GmnxYdx++TQefmkbNoeLR1/Zzl1fm01qXOjAfFIiIiIiMqg+/PC/PPjgfSe8LzExmeeff3mQK+pfWv7UCbPZSB3VPLPxJXJqDnlvzwhP47LxKxgTmUlpUzl5DYXk1xeS11BAUUMxDrfTe21UYCTnjT6L+UlzMBlN3b7m1gMVPPH6LlxuN9Hhgfy/62cTExE0IJ+fnBxNS0t3NEakKxof0h2NkeHHam2iurr6hPeZzWaSknreV+uPy58UKjrxdu77fJC7Ejeef56owEguGnsucxJnYDSceNMsh8vBgdrDvHv4Q3Lr8723J4TEccHoc5iZMLXTx7b7cnsx//jvPgCSY0P4+ddmExbc/WyHDC59s5fuaIxIVzQ+pDsaI9IVfwwVWv50Ag6Xg/dzPwUg0BTAsozTOTNjCQGmgC4fZzaamRgzgezo8eyo3MPbh9+npKmMcmslf9/9L9LzUrgy6xLGRGZ2+hxLpqdQb7Xx2ueHKamy8ugr27nj6pkEBnQ/0yEiIiIi4guaqejEhrLNNNHAKXFzCDWF9ek5XG4XG0u38m7uh1S11ABgNBi5ZOx5nJ6+pNMdo9xuNy9+coCPN3mOU582NpZbL52K2aRjRfyF3kGS7miMSFc0PqQ7GiPSFX+cqdBvqZ1YmDqXK6esIDIwos/PYTQYmZc8m1/P/ylXTriYIFMgLreL1w6+wzO7nqfZ0XLCxxkMBq4+czzzJiUCsONQFa9+duiE14qIiIiI+JpCxSAwG80sTVvIz+beTkpoEgBbK3byh03/S1FjyQkfYzQYuOn8iUzMjAbgw40FbNpXPmg1i4iIiIj0lELFIEoMieeOObdyStIsAMqtlTy46XHWl2w+4fVmk5HvXDSZ6HDPmRfPvLeXkqr+OaVRRERERKS/KFQMskBTADdMvIprsi7FbDBhd9l5du9LvLjvNewux3HXR4QE8L2Lp2AyGmi1OXnyjV202pwneGYREREREd9QqPABg8HA4tT5/Hj2LcQEeZY3rSpez193/BO7037c9eNSI7nyjHEAFFU28c/39+GH/fUiIiIiMkIpVPhQZkQ6d839AZNisgDYU53DX3c9e8JgcdbsNOZmJwCwbk8ZK7cWDWqtIiIiIiKdUajwsVBLCN+ddiPT46cAsKcqh7/teu64pVAGg4Ebz80mOTYEgBc/PsDh4vpBr1dERERE5KsUKvyAyWjim5OvZXrcZAB2V+3j6Z3PHhcsggPN3HLJVAItJpwuN0++uZMGq80XJYuIiIiIeClU+Amz0cw3p1zHtLZgsatqH0/vPH7GIjUulK+f61kuVV3fyl/f3oNL/RUiIiIi4kMKFX7EbDRz05TrmBo3EYBdVXt5ZtdzOL4SLOZPSuLMWWkA7M6tZuUW9VeIiIiIiO8oVPgZT7C4nimxnmCxs3IvT+96/rhgcdWZ40hPCAPg1c8PUVnXPOi1ioiIiIiAQoVfshjNfGvq9UyJzQZgZ+UeXj/4TodrzCYj3zgvG4MBWm1Onv0gR9vMioiIiIhPKFT4KU+wuIHs6PEAfF64hg2lWzpcMyopgnNOyQBg1+Fq1u0uG/Q6RUREREQUKvyYxWjmG1OuJTowCoAX971GcWNph2suWjyahKhgz/2fHKC+SbtBiYiIiMjgUqjwc2GWUG6eej1mgwmby87fdj5Ls+No/0SgxcTXz/Usk2pstvPCx/t9VaqIiIiIjFAKFUNAZkQ6V064GIDy5kqe2/tKh/6JiZnRnDo9BYANe8vZdqDSF2WKiIiIyAilUDFELEw5hfnJcwDYXrGLj/M/73D/laePJSosAIDnPszB2uI47jlERERERAaCQsUQYTAYuGrCJaSHeWYk/nPov+RUH/TeHxJk4fqzPYfi1TS08urnh3xSp4iIiIiMPAoVQ0iAycK3pl5PsDkYN27+vvtf1LTUeu+fOSGeudkJAHy2tYic/BofVSoiIiIiI4lCxRATFxzLjZOuBqDR3sQzu57H6XJ677922QRCg8wA/N9/9+FwunxSp4iIiIiMHAoVQ9CUuImcO+pMAHLr8/ko/zPvfZGhAVx9pudsi7KaZr7YXuyLEkVERERkBFGoGKLOG72MsZGjAPhv7seUNB09+G7BlCQyEsMAeGv1EVptzhM9hYiIiIhIv1CoGKKMBiPXZV+O2WjG4Xbyr72v4HK72u4zcPnSsQDUN9n4aFOBL0sVERERkWFOoWIISwxN4PzRywDPMqjPCld775s8OobsjCgA/rs+j8Zmuy9KFBEREZERQKFiiDsz/VTSw1MBePvQ+1Q2VwGeLWgva5utaG518t66PJ/VKCIiIiLDm0LFEGcymvha9hUYDUZsLjsv7HvNe9r22NRIZo6PA+CTzYVU17f4slQRERERGaYUKoaBtPAUzs48HYCcmoOsKdngve/SU8dgMIDd4eKt1Ud8VKGIiIiIDGcKFcPE8lFnkhTiOfju9QPvUttaB0BqfBgLpyQBsGpHCSVVTT6rUURERESGJ4WKYcJiNHPdxCswYKDF2cK/c173LoO6aPFozCYDLrebN77M9XGlIiIiIjLcKFQMI2MiMzktfREAOyv3srl8OwBxkcGcPjMNgE37ysktqfdZjSIiIiIy/ChUDDMrxiwnLigGgFf2/4dmh6c5+/yFmQQGmAB4/fNDPqtPRERERIYfhYphJtAUwDXZlwHQaG/is4JVAESEBLD8lAwAdh+pYe+Rap/VKCIiIiLDi0LFMJQdM56s6HEAfFLwBVa7FYCz56YTFmwB4PUvD/usPhEREREZXhQqhqkLxpwDQLOjhU8KvgQgONDMefMzAThUVM+hojqf1SciIiIiw4dCxTA1JjKTybHZAKws+JJGm2cr2aUzUghq6634aFOBz+oTERERkeFDoWIYu2D02QC0Om18lP8Z4JmtWDwtGYBN+yp0yraIiIiInDSFimEsIyKN6XGTAfi8cA11rQ0AnDU7DQPgcrv5dEuRDysUERERkeFAoWKYO3/M2RgwYHfZ+TDvUwASokOYMT4OgM+3FdFqd/qyRBEREREZ4hQqhrnUsGRmJUwDYFXROmpaagFYNicdgKYWB2t3lfqqPBEREREZBhQqRoDzRi/DgAGH28n7bbMVWRlRpCeEAZ6Gbbfb7csSRURERGQIU6gYAZJCEzglaRYAa4o3UNlcjcFg8M5WlFRZ2a3D8ERERESkjxQqRohzR52F0WDE5Xbx3yMfAzBvUgLhIZ7D8D7aWOjL8kRERERkCFOoGCHiQ2KZnzQHgA2lWyizVmAxmzh9ZioAOw9XUVLV5MsSRURERGSI6lWoeOGFF1ixYgWzZs1i1qxZXHXVVXz++efe+1tbW7nnnnuYN28eM2fO5LbbbqOysrLfi5a+WT7qTEwGk2e2ItczW3H6zFRMRgMAH2/SbIWIiIiI9F6vQkVSUhJ33HEHr7/+Oq+99hrz58/n+9//PgcOHADgvvvuY+XKlfzpT3/iueeeo7y8nFtvvXVACpfeiw2OZlHKKQBsLt9ObWsdkWGBzJuUCMDqXSU0tdh9WaKIiIiIDEG9ChVnnHEGS5cuZdSoUYwePZof/ehHhISEsG3bNhoaGnjttde46667WLBgAVOmTOG+++5j69atbNu2bYDKl946PX0xAC63izXFG4Cj28va7C6+2F7ss9pEREREZGgy9/WBTqeT999/H6vVysyZM9m1axd2u52FCxd6rxk7diwpKSls27aNGTNm9Pi5jUYDxrYlOb5iMhk7/DlcpEQkMjF2Anur9rO6eD3njz2LsWmRZGVEkZNfyyebCzlvQSYm4/D6vPvbcB0f0n80RqQrGh/SHY0R6Yo/jo9eh4qcnByuvvpqWltbCQkJ4YknnmDcuHHs3bsXi8VCREREh+tjY2OpqKjo1WvExIRiMPg2VLSLiAj2dQn97vzs09m7ej+1rfUcaj7EvLSZXHbGeO77v41U17eyr7CexdNTfV3mkDAcx4f0L40R6YrGh3RHY0S64k/jo9ehYvTo0bz55ps0NDTwwQcfcOedd/L888/3a1HV1U1+MVMRERFMfX0zTqfLp7X0tzHBY4gOjKSmtY739q5kQugEJqREEBcZRGVdC299fojJGVG+LtOvDefxIf1DY0S6ovEh3dEYka4M9viIjg7t9ppeh4qAgAAyMzMBmDJlCjt37uTZZ5/l3HPPxW63U19f32G2oqqqivj4+F69hsvlxuXyjxOenU4XDsdw+2I2sChlHu/kfsje6gMU1ZWSGJrAkmnJvPFlLvvyaqioaSY6PNDXhfq94Tk+pD9pjEhXND6kOxoj0hV/Gh8nvRDL5XJhs9mYMmUKFouFtWvXeu87fPgwxcXFveqnkMGxMOUUjAbP//4vi9cBMG9yEgBuYP2eMl+VJiIiIiJDTK9CxcMPP8zGjRspLCwkJyeHhx9+mA0bNrBixQrCw8O57LLLeOCBB1i3bh27du3iF7/4BTNnzlSo8EORgRHMiJ8CwLqSzdicNhKighmb6pllWre71JfliYiIiMgQ0qvlT1VVVdx5552Ul5cTHh5OVlYWzzzzDIsWLQLgF7/4BUajkdtvvx2bzcbixYu5++67B6RwOXmnpi5gS/kOmh3NbCrbzsKUucyflMShonryyxspqmwiNa77NXQiIiIiMrIZ3G63fzQvHKOiosHXJWA2G4mODqWmpslv1qr1N7fbzW83PEJpUxnp4ancOed2Gprt/Pix1bjcbs5fkMllS8f6uky/NBLGh5wcjRHpisaHdEdjRLoy2OMjPj6822v8Z3NbGXQGg4ElqfMBKGgoIq+hgIiQAKaMiQE8fRV+mDlFRERExM8oVIxw85JmEWAKAOCLQk+T/fxJiQBU1rVwsKjOZ7WJiIiIyNCgUDHCBZuDmZs4E4At5dtptDcxc3w8gRYTAOt2axcoEREREemaQoWwJHUBAHaXg3UlmwgMMDFzQhwAG/eV49ChOyIiIiLSBYUKIT08hTGRngMNvyxah8vtYv4kz5kVjc12dh2u9mV5IiIiIuLnFCoEODpbUdlcxb7qA0weHU14iAWAdXt0ZoWIiIiIdE6hQgCYmTCNMIvnTIrPC1djMho5ZaKnYXvbgUqaWx2+LE9ERERE/JhChQBgMZpZlDIPgF1V+yhtKmf+ZE+osDlcbNlf4cvyRERERMSPKVSI16lpCzAZPLs+fVrwJWOSI0iICgZg3R7tAiUiIiIiJ6ZQIV5RgZHe7WU3lG6m0d7kna3Yc6SausZWX5YnIiIiIn5KoUI6OCNjCeDZXvaLwjXMn+zZBcrthg17y31ZmoiIiIj4KYUK6SA1LJmJMRMA+KJoLTGRFkYlhQPaBUpERERETkyhQo5zZvqpADTam9hQutk7W5Fb0kBptdWXpYmIiIiIH1KokONkx4wnJdQTJD4t+JK52XEYDJ77NqhhW0RERES+QqFCjmMwGDgjwzNbUWatoLA1lwlpUQDsOFzlw8pERERExB8pVMgJzUmcQWSAp5fik/wvmDo2FoDc4nrqrTZfliYiIiIifkahQk7IYjSzNG0RAAdqD5OQ7AkSbmB3brUPKxMRERERf6NQIZ1anDqfAKMFgJ2NG4kOD/T8/ZCWQImIiIjIUQoV0qlQSwgLUuYCsKV8B1lj20LF4SpcLrcvSxMRERERP6JQIV06PW0JBgy43C6cMYcBaGpxkFtS7+PKRERERMRfKFRIl+JDYpkePxmAg807MZkdAOzQEigRERERaaNQId06M2MpAC3OVpLGeZq0tbWsiIiIiLRTqJBujYnMJCM8DQBXVCEAeaUN1DW2+rIsEREREfETChXSI3MTZwBQ6yrHENgEwC5tLSsiIiIiKFRID81KnI4BAwDhKZWA+ipERERExEOhQnokKjCScVGjATDHlQBuduVW43S5fFuYiIiIiPicQoX02Jy2JVDNhloMwY00tzo4VKStZUVERERGOoUK6bEZCVMxGjxDxhJXAngOwhMRERGRkU2hQnoszBLKpJgJAAQmlAFu9VWIiIiIiEKF9M7stiVQDlMTxrBaCsobqWnQ1rIiIiIiI5lChfTKtLjJWIwWAEwxWgIlIiIiIgoV0ktB5kCmxk0EwBJXBrjYqSVQIiIiIiOaQoX02pzEmQC4za0YI6rZfaQah1Nby4qIiIiMVAoV0muTYrMINgcDYIotocXm5EBhnY+rEhERERFfUaiQXrMYzcyInwKAKboMDC71VYiIiIiMYAoV0iftB+EZzA6MkRXqqxAREREZwRQqpE8mRI8lPCAM8CyBKqpsorKu2cdViYiIiIgvKFRInxgNRmYlTAfAFFUORgd7jtT4uCoRERER8QWFCukz7xIokwtTdDkH1awtIiIiMiIpVEifjY7IIDYoGvAchHegSKFCREREZCRSqJA+MxgMzG6brTBGVlJWV0t9k823RYmIiIjIoFOokJPiXQJldGOKKeOgZitERERERhyFCjkpKaFJJATHA2CKKeVAYa1vCxIRERGRQadQISfFYDAwK3EaAMaIanKKy31ckYiIiIgMNoUKOWmzEjyhwmBwU9R6iFa708cViYiIiMhgUqiQk5YSmkSUJQYAQ3QpR0rqfVyRiIiIiAwmhQo5aQaDgTlJnoPwjBFV7C4o83FFIiIiIjKYFCqkX8xtCxUGo5udlXt8XI2IiIiIDCaFCukXqWHJBLkjACh3H8Lldvu4IhEREREZLAoV0i8MBgMTwicC4A6r5FBppY8rEhEREZHBolAh/ebUzFmAZwnUqrxtvi1GRERERAaNQoX0m+yEURhsoQDk1O/1cTUiIiIiMlgUKqTfGAwGEgxjAGgwFWG1N/u4IhEREREZDAoV0q+mxk72/MXgZm3Bdt8WIyIiIiKDQqFC+tUpmRNwtQQDsKFkm2+LEREREZFBoVAh/SolLhRTQwoARS1HaHZoCZSIiIjIcKdQIf3KYDCQETABALfBxc5KNWyLiIiIDHcKFdLvpiSNxtXqWQK1qVR9FSIiIiLDnbk3F//lL3/hww8/5PDhwwQFBTFz5kzuuOMOxowZ473m+uuvZ8OGDR0ed9VVV3Hvvff2T8Xi98anR+E8nIgx+Qj7avbT7Ggh2Bzk67JEREREZID0KlRs2LCB6667jqlTp+J0OnnkkUe46aabePfddwkJCfFed+WVV3L77bd7Pw4ODu6/isXvjUqKwFCbAslHcLqd7KzcwylJs3xdloiIiIgMkF6FimeeeabDxw888AALFixg9+7dzJ0713t7UFAQ8fHx/VOhDDkWs5GMyDQKW4MwBrawtXynQoWIiIjIMNarUPFVDQ0NAERGRna4/e233+att94iPj6e008/nVtuuaVXsxVGowGj0XAypZ00k8nY4U/pneyMaPIKkjAmH2FPdQ4ObAQNoyVQGh/SHY0R6YrGh3RHY0S64o/jo8+hwuVycd999zFr1iwmTJjgvf2CCy4gJSWFhIQEcnJyeOihh8jNzeXxxx/v8XPHxIRiMPg2VLSLiNDSrb6YNTGJ/+5IgOQjOFwOqlyVTInO8nVZ/U7jQ7qjMSJd0fiQ7miMSFf8aXz0OVTcc889HDhwgBdeeKHD7VdddZX371lZWcTHx3PjjTeSn59PRkZGj567urrJL2YqIiKCqa9vxul0+bSWoSg5KgiXNRK3GwwG2FV0gNSANF+X1W80PqQ7GiPSFY0P6Y7GiHRlsMdHdHRot9f0KVTce++9fPbZZzz//PMkJSV1ee306dMByMvL63GocLncuFzuvpTW75xOFw6Hvph7K8hiIiUmgqrmMAwhjeTVFQ7Lf0eND+mOxoh0ReNDuqMxIl3xp/HRq4VYbrebe++9l48++oh//vOfpKend/uYvXs9h5+pcXvkGZcaiavJ02+T11Do42pEREREZKD0aqbinnvu4Z133uHJJ58kNDSUiooKAMLDwwkKCiI/P5+3336bpUuXEhUVRU5ODvfffz9z584lOzt7QD4B8V/j0yJZvTEC4ouobK7CarcSYgnp/oEiIiIiMqT0KlS8+OKLgOeAu2Pdf//9XHrppVgsFtauXcuzzz6L1WolOTmZs88+m1tuuaX/KpYhY0xKhHemAiC/oYjsmPE+rEhEREREBkKvQkVOTk6X9ycnJ/P888+fVEEyfCRGh2CxR+J2GzAY3BQoVIiIiIgMS/6zua0MO0ajgfS4SNzNYYD6KkRERESGK4UKGVAZieG4miIAyK9XqBAREREZjhQqZEClJ4Z5+yqqWqppslt9XJGIiIiI9DeFChlQmcfMVADkawmUiIiIyLCjUCEDKjUuFENzBG6X54T0gvoiH1ckIiIiIv1NoUIGVIDFRHJMuJq1RURERIYxhQoZcBnH9FUUKFSIiIiIDDsKFTLgjt0BqqqlhkZ7k48rEhEREZH+pFAhAy4jMRyX9ejJ2uqrEBERERleFCpkwKUnhOG2hnubtdVXISIiIjK8KFTIgAsLthAbHoK7ORzQtrIiIiIiw41ChQwKT7O2TtYWERERGY4UKmRQeJq1PX0VNa21NNgafVyRiIiIiPQXhQoZFMfOVADkN6hZW0RERGS4UKiQQZGZGI67+WiztpZAiYiIiAwfChUyKKLDAwkNDMBt9TRr6xA8ERERkeFDoUIGhcFg6HBehbaVFRERERk+FCpk0GQec7J2bWsd9bYGH1ckIiIiIv1BoUIGTXpimHcHKFBfhYiIiMhwoVAhgyYjMRx3cxhul2fY6RA8ERERkeFBoUIGTXJMCAEms7dZW9vKioiIiAwPChUyaIxGA6nxOllbREREZLhRqJBBlXlMX0WdrZ661nofVyQiIiIiJ0uhQgZVxjE7QIH6KkRERESGA4UKGVQZieG4W45t1lZfhYiIiMhQp1Ahgyo1PhQDxqPN2uqrEBERERnyFCpkUAVaTCTHhnr7KrT8SURERGToU6iQQZeReHQHqHpbg5q1RURERIY4hQoZdBkJ4bisR5u1C9RXISIiIjKkKVTIoMtIDGs7WdsAQEFDsY8rEhEREZGToVAhgy4jMRzcRtzNnmbtgkbNVIiIiIgMZQoVMujCgi3ERAR6+yoKtfxJREREZEhTqBCfOLavoqqlhia71ccViYiIiEhfKVSITxy7AxSoWVtERERkKFOoEJ9IiQvF3RyO2+35WKFCREREZOhSqBCfSIoJAZcJd3MYAIWN2gFKREREZKhSqBCfSIwOAfD2VWimQkRERGToUqgQnwgMMBEdfnQHqHJrJS2OVh9XJSIiIiJ9oVAhPpMUE4K7babCjZuixhIfVyQiIiIifaFQIT6TFBOCyxru/VhLoERERESGJoUK8ZmkmBBwWnC1ePorFCpEREREhiaFCvGZpNi2Zu22voqCRoUKERERkaFIoUJ8JjHGEyra+ypKmsqwO+2+LElERERE+kChQnwmLiIIs8ngnalwuV0UN5X6uCoRERER6S2FCvEZo9FAQnSI96wKUF+FiIiIyFCkUCE+lRQTAo4AjI5gAAp0sraIiIjIkKNQIT6VFPOVZm3NVIiIiIgMOQoV4lOJMZ4ZCntjGABFjSU4XU5fliQiIiIivaRQIT6VHBMKgKspEgCHy0GptdyXJYmIiIhILylUiE+1z1S0L38CKGxQX4WIiIjIUKJQIT4VHhJAaJAZ7IFYCALUVyEiIiIy1ChUiM95TtY2EGCPBiBfoUJERERkSFGoEJ9LivbsAOVoDAegqLEYl9vly5JEREREpBcUKsTnPDMV0FTt+bPF2Uplc5UvSxIRERGRXlCoEJ9L/MpMBaivQkRERGQoUagQn2ufqXC3hhBgCACgQDtAiYiIiAwZChXic4nRwRgAMBBuiAM0UyEiIiIylChUiM9ZzCZiIz3byVradoAqaCzC7Xb7siwRERER6aFehYq//OUvXHbZZcycOZMFCxZwyy23cPjw4Q7XtLa2cs899zBv3jxmzpzJbbfdRmVlZb8WLcNPUoxnCZS9IQyAJruVmtZaH1YkIiIiIj3Vq1CxYcMGrrvuOl5++WX+8Y9/4HA4uOmmm7Bard5r7rvvPlauXMmf/vQnnnvuOcrLy7n11lv7vXAZXhLbQkV9VbD3NvVViIiIiAwNvQoVzzzzDJdeeinjx48nOzubBx54gOLiYnbv3g1AQ0MDr732GnfddRcLFixgypQp3HfffWzdupVt27YNRP0yTCR5Q0UAZqMZUF+FiIiIyFBhPpkHNzQ0ABAZGQnArl27sNvtLFy40HvN2LFjSUlJYdu2bcyYMaNHz2s0GjAaDSdT2kkzmYwd/pSBlRof2vY3IwmBiRQ3F1HYWITZ7J///hof0h2NEemKxod0R2NEuuKP46PPocLlcnHfffcxa9YsJkyYAEBlZSUWi4WIiIgO18bGxlJRUdHj546JCcVg8G2oaBcREdz9RXLSssYc/f8dZUmguLmIsuZyoqNDu3iU72l8SHc0RqQrGh/SHY0R6Yo/jY8+h4p77rmHAwcO8MILL/RnPQBUVzf5xUxFREQw9fXNOJ0un9YyEpjcbgLMRmwOF06r5wukoqma8qo6LMaTmlAbEBof0h2NEemKxod0R2NEujLY46Mnb/L26be1e++9l88++4znn3+epKQk7+1xcXHY7Xbq6+s7zFZUVVURHx/f4+d3udy4XP6xnajT6cLh0BfzYEiIDqGwopHWhiAIBjduShsqSA5N9HVpndL4kO5ojEhXND6kOxoj0hV/Gh+9Wojldru59957+eijj/jnP/9Jenp6h/unTJmCxWJh7dq13tsOHz5McXFxj/spZORqP1m7vibAe1u5tefL5kRERETEN3o1U3HPPffwzjvv8OSTTxIaGurtkwgPDycoKIjw8HAuu+wyHnjgASIjIwkLC+O3v/0tM2fOVKiQbrXvAFVZbsCcasTldlGmUCEiIiLi93oVKl588UUArr/++g6333///Vx66aUA/OIXv8BoNHL77bdjs9lYvHgxd999dz+VK8NZUoynl6LV5iYhMJrKlirKrTo4UURERMTf9SpU5OTkdHtNYGAgd999t4KE9FpSzNEmoDBjNJVUafmTiIiIyBDgP5vbyojXPlMBEOAMB9BMhYiIiMgQoFAhfiMkyEJEiAUAV4tn1qLB3ojV3uzLskRERESkGwoV4lfam7Wb64O8t5U3awmUiIiIiD9TqBC/ktgWKuqqjrb7aAmUiIiIiH9TqBC/0n5WRVUVBJo851WoWVtERETEvylUiF9JivaECpcbogNiAXRWhYiIiIifU6gQv9I+UwEQYogEtPxJRERExN8pVIhfiY8KxmgwAGCyhwGe5U9ut9uXZYmIiIhIFxQqxK+YTUbiojw7PzmtnlkLm8tOna3el2WJiIiISBcUKsTvtG8r21Qb6L2trEl9FSIiIiL+SqFC/E5iW7N2deUx28rqrAoRERERv6VQIX4nMSYYgPoGN+GW9r4KNWuLiIiI+CuFCvE77TMVAJGWGEBnVYiIiIj4M4UK8TuJ0cHevwe7PdvK6qwKEREREf+lUCF+JyYiCLPJs62ssW1b2aqWGhwuhy/LEhEREZFOKFSI3zEaDcRHeWYr7E2eP11uF1XN1b4sS0REREQ6oVAhfqm9r6LxmG1ly5vVrC0iIiLijxQqxC+17wBVVWnEgGcplPoqRERERPyTQoX4pYT2mYomJzFB0YB2gBIRERHxVwoV4peO3QEq0ty+rayWP4mIiIj4I4UK8UvHnlUR6IoANFMhIiIi4q8UKsQvRUcEYjG3DU9bKAB1tgaaHS0+rEpERERETkShQvyS0WAgoW1bWVvD0aVQFVoCJSIiIuJ3FCrEbyW09VU01gZ4b9MSKBERERH/o1AhfisxxtNXUVEJAUYLAGU6q0JERETE7yhUiN9q3wGqqdlBbFAsoJkKEREREX+kUCF+K+GYHaAiTO3byipUiIiIiPgbhQrxW8eeVRHgCgc8Z1W43W5flSQiIiIiJ6BQIX4rKjyQgLZtZd0tnm1lW5yt1NsafVmWiIiIiHyFQoX4LaPB4N0BqqUxyHt7ubXcVyWJiIiIyAkoVIhfaz9Zu77q2G1ltQOUiIiIiD9RqBC/lhDjmamorHYQZvEsgSprVrO2iIiIiD9RqBC/1j5T0dTiIC4oDtAOUCIiIiL+RqFC/NqxO0CFGqMALX8SERER8TcKFeLXjj2rwuLwbCtb0VyF0+X0VUkiIiIi8hUKFeLXosICCLSYgKPbyrrcLqpaanxZloiIiIgcQ6FC/JrhmG1lmxsCvberr0JERETEfyhUiN9r76uorTRjwAAoVIiIiIj4E4UK8XuJMZ6+ivIaGzFB0QCU6gA8EREREb+hUCF+z7v8qdVBUnASAHn1hb4sSURERESOoVAhfi/xmB2gok2eUFHUWEKr0+arkkRERETkGAoV4vfalz8BBNljAXDjJr++wFcliYiIiMgxFCrE70WEWAgM8Gwr62gMx2jwDNvcunxfliUiIiIibRQqxO8ZDAbvDlBVtQ5Sw5IByK1XqBARERHxBwoVMiS091WUVTczOiIDgNz6PNxuty/LEhEREREUKmSISIzxzFSU1VgZ1RYqGmyNVOtkbRERERGfU6iQIaF9pqLF5iQuINl7u5ZAiYiIiPieQoUMCcduK+toCibU4vk4ty7PVyWJiIiISBuFChkSEtqWPwGU1x7bV6GZChERERFfU6iQISE82EJwoBmA8ppmRkVkAlDYUIzdafdlaSIiIiIjnkKFDAnHbitbVm1ldKRnpsLpdlLQWOzL0kRERERGPIUKGTIS2kNFTTOZEekYMADqqxARERHxNYUKGTLam7XLa5oJMgWSHJoIqK9CRERExNcUKmTIaD+rotXupK7J5j2v4kidQoWIiIiILylUyJBx7Layx/ZV1LTWUtta56uyREREREY8hQoZMhJjjgkVNc3emQrQbIWIiIiILylUyJARFmwhNMizrWxZjZWk0ASCTEEAHK5Xs7aIiIiIryhUyJCS0LYEqqy6GaPByKiIdEAzFSIiIiK+1OtQsXHjRr773e+yePFisrKy+Pjjjzvcf9ddd5GVldXhv5tuuqnfCpaRLaltCVRptRXA21eR31CI0+X0WV0iIiIiI5m5tw+wWq1kZWVx2WWXceutt57wmiVLlnD//fd7Pw4ICOh7hSLHSIpt31bWisvl9vZV2F0OihpLyIhI82V5IiIiIiNSr0PF0qVLWbp0aZfXBAQEEB8f3+eiRDqT3DZT4XC6qaxrZlTk0Wbt3Pp8hQoRERERHxiQnooNGzawYMECzjnnHO6++25qamoG4mVkBEo6Zgeo0morYZZQEkLiAJ2sLSIiIuIrvZ6p6M6SJUtYtmwZaWlpFBQU8Mgjj3DzzTfz0ksvYTKZevQcRqMBo9HQ36X1islk7PCn+IeU+FAMgBsor23GbDYyJjKTcmslR+rzMZsH5/+Xxod0R2NEuqLxId3RGJGu+OP46PdQcf7553v/3t6ofdZZZ3lnL3oiJiYUg8G3oaJdRESwr0uQr4iPCaG82kp1o53o6FCmpIxnXclmKpqrMAW7iAgKH7RaND6kOxoj0hWND+mOxoh0xZ/GR7+Hiq9KT08nOjqavLy8HoeK6uomv5ipiIgIpr6+GafT5dNapKPEqGDKq63kFddRU9NEoiXZe9/W/H1Mi5804DVofEh3NEakKxof0h2NEenKYI+P6OjQbq8Z8FBRWlpKbW1trxq3XS43Lpd7AKvqOafThcOhL2Z/khgdzE6guKoJh8NFYlACAUYLNpedg9VHmBSdPWi1aHxIdzRGpCsaH9IdjRHpij+Nj16HiqamJvLzjx40VlhYyN69e4mMjCQyMpLHH3+cc845h7i4OAoKCnjwwQfJzMxkyZIl/Vq4jFzJbdvK1jXaaG51EBxoJjMinQO1h8mt1yF4IiIiIoOt16Fi165d3HDDDd6P28+juOSSS/jNb37D/v37efPNN2loaCAhIYFFixbxgx/8QGdVSL/56g5Qo5MjGBWRwYHaw+TV5+NyuzAa/KdxSURERGS463WomDdvHjk5OZ3e/8wzz5xUQSLdSYo9uq6vPVS0n6zd6rRR0lRGalhyZw8XERERkX6mt3NlyIkKCyAwwLM9cWmVFYBREZne+w/XHfFFWSIiIiIjlkKFDDkGg4GkaM8SqNJqT6iIDAz3HoK3r/qgz2oTERERGYkUKmRISortGCoAJsZMACCn5iBOl9MndYmIiIiMRAoVMiS1N2uXVVtxuT3bD7eHimZHM/kNhT6rTURERGSkUaiQIak9VNgcLqrrWwAYHzUWk8HTa7Gner/PahMREREZaRQqZEj66rayAEHmQMZEehq29ylUiIiIiAwahQoZkjqEiqqjfRXZbUugjtQXYLU3D3pdIiIiIiORQoUMSYEBJqLDA4GOzdqT2kKFy+1if412gRIREREZDAoVMmS1z1YcGyrSwlMItXhu36slUCIiIiKDQqFChqwTbStrNBjJjh4PeEKFu21nKBEREREZOAoVMmS1z1RU17fSajt6LsXE2CwAqlpqqGiu9EltIiIiIiOJQoUMWcnHNGuX1Rx7CN5479/3Vh8Y1JpERERERiKFChmyTrStLEBUYCTJoYkA7K3OGfS6REREREYahQoZsmIig7CYPUP42G1l4ejp2vtrDuFwOQa9NhEREZGRRKFChiyjwUBi9PHN2nA0VLQ6beTW5Q96bSIiIiIjiUKFDGntO0CVfCVUjIsajdloBrS1rIiIiMhAU6iQIe3YsyqO3T42wBTAuMjRgEKFiIiIyEBTqJAhrX0HqFabk9pGW4f7JsZ6lkAVNBTRaGsa9NpERERERgqFChnS2pc/Qed9FW7c7KvR1rIiIiIiA0WhQoa0zraVBUgJTSIiIBzQEigRERGRgaRQIUNacKCZyNAA4PhtZQ0Gg3e2Ym/V/g49FyIiIiLSfxQqZMg7tln7q7LbTteus9VT0lQ2qHWJiIiIjBQKFTLktfdVlFYf34zdPlMBsE9LoEREREQGhEKFDHntMxWVdS3YHa4O94UHhJEelgLAHoUKERERkQGhUCFDXnuocLuhvOZES6A8sxUHaw9jd9oHtTYRERGRkUChQoa8rraVBZjUdl6F3eXgw7yVg1aXiIiIyEihUCFDXlxkECajAThxqBgXNYbMiHQA3jvyMTsr9wxqfSIiIiLDnUKFDHkmo5GE6GDg+G1lAYwGIzdPuZ4wSygA/9zzb8qtFYNao4iIiMhwplAhw0J7X0XJCWYqAKKDorhpytcwGow0O1r4287naHG0DmaJIiIiIsOWQoUMC95tZausnR5yNyF6LJeMPQ+A4qZSXtj3qg7EExEREekHChUyLLTPVFhbHTRYO9/h6fT0JcxOmA7A5vLtfFrw5aDUJyIiIjKcKVTIsJAcG+r9+4matdsZDAaum3gFKaFJALx56D321xwc8PpEREREhjOFChkW2mcqoOtQARBoCuDmqTcQbA7C5XbxzK5/UdNSO8AVioiIiAxfChUyLIQFW4gIsQBQXNnU7fUJIXHcOOkaDBhotDfxzK5/qb9CREREpI8UKmTYSI0PA6CoorFH10+Jm8jyUWcAkFufR5m1fMBqExERERnOFCpk2EiJ8/RVFPZgpqLdopR53r8frsvv95pERERERgKFChk2UuM9oaKu0UZjc+c7QB0rOiiKqMBIAHLrjgxUaSIiIiLDmkKFDBtpcWHev/ekr6LdmMhMAA7X5fV7TSIiIiIjgUKFDBspcUd3gOppXwXA6LZQUWotx2rveucoERERETmeQoUMGyFBFqLDAwEo6sNMBUBuvfoqRERERHpLoUKGlfa+iqKKnoeKtLAULEYzALlaAiUiIiLSawoVMqyktu0AVVTZ1ONzJ8xGMxnhaYD6KkRERET6QqFChpXUtmbtxmY79dae7QAFMCZyFABH6vNxuV0DUZqIiIjIsKVQIcNK+/In6G2zdgYArU4bxY2l/V6XiIiIyHCmUCHDSkrssaGi530Vo49p1tYSKBEREZHeUaiQYSUwwER8VBDQux2gIgLCiQuOBSC3XqFCREREpDcUKmTYae+rKKrs+fIngNEROgRPREREpC8UKmTYae+rKO7FDlBw9LyKyuYqGmy9CyQiIiIiI5lChQw77dvKNrc6qWlo7fHj1FchIiIi0jcKFTLspMaHef9e2Itm7ZTQRAJNAYAOwRMRERHpDYUKGXaSYkIwGgxA7/oqTEYToyI8W8tqpkJERESk5xQqZNixmI0kxgQDUNyLmQo4ugQqv6EAh8vR77WJiIiIDEcKFTIstfdVFPZiW1k42qxtdzkoaizp97pEREREhiOFChmWUtpCRUllE65e7AA1um35E2gJlIiIiEhPKVTIsJTW1qxtc7iorG3u8eNCLCEkhSQAatYWERER6SmFChmW2s+qACjqZV9F+xIozVSIiIiI9IxChQxLCdHBmE2eHaB621fR3qxd01pLTUttf5cmIiIiMuwoVMiwZDIaSYrxzFYUVfTudOwxxxyCl1uf3691iYiIiAxHvQ4VGzdu5Lvf/S6LFy8mKyuLjz/+uMP9brebRx99lMWLFzNt2jRuvPFGjhw50l/1ivRYWtsSqOJezlQkhMQTbPZsSau+ChEREZHu9TpUWK1WsrKyuPvuu094/9/+9jeee+45fvOb3/Dyyy8THBzMTTfdRGtr60kXK9Ib3h2gqqw4nK4eP85oMDI6UofgiYiIiPRUr0PF0qVL+dGPfsSyZcuOu8/tdvPss8/yve99j7POOovs7Gz+8Ic/UF5eftyMhshAa2/WdrrclNX0fAcogDERowAoaCjC7rT3d2kiIiIiw0q/9lQUFhZSUVHBwoULvbeFh4czffp0tm7d2p8vJdKt1LZtZaH3S6DaZyqcbif5DUX9WpeIiIjIcGPuzyerqKgAIDY2tsPtsbGxVFZW9vh5jEYDRqOhP0vrNZPJ2OFPGXqSYkMIsBix2V2UVDVhNvf8/+W4mEwMGHDj5khDHllxYzrcr/Eh3dEYka5ofEh3NEakK/44Pvo1VPSXmJhQDAbfhop2ERHBvi5BTkJGUgQHC2opq20hOjq0+wd4hZIRlUpebSGF1qJOH6vxId3RGJGuaHxIdzRGpCv+ND76NVTEx8cDUFVVRUJCgvf2qqoqsrOze/w81dVNfjFTERERTH19M85eNPmKf0mODuZgQS1HiuuoqendEqhRYenk1RayqyyH8qo6LMajXy4aH9IdjRHpisaHdEdjRLoy2OOjJ2/M9muoSEtLIz4+nrVr1zJx4kQAGhsb2b59O9dcc02Pn8flcuNyufuztD5zOl04HPpiHqqSYz1fBKXVVppb7FjMph4/dmrsZD4vXIvV0czW0l3MSph23DUaH9IdjRHpisaHdEdjRLriT+Oj16GiqamJ/PyjB4IVFhayd+9eIiMjSUlJ4YYbbuCpp54iMzOTtLQ0Hn30URISEjjrrLP6tXCRnmjfAcrt9mwtm5EY3uPHZsWMIyowktrWOtaVbDphqBARERGRPoSKXbt2ccMNN3g/vv/++wG45JJLeOCBB7j55ptpbm7m17/+NfX19cyePZunn36awMDA/qtapIdS445O1xVVNvUqVBgNRuYlzeaDvE/ZU5VDbWsdUYGRA1GmiIiIyJDW61Axb948cnJyOr3fYDDwgx/8gB/84AcnVZhIf4gODyQ40Exzq4Oiit71VADMT/aECjduNpRu4ezM0wegyr5rslv5onAtsxOnkRAS7+tyREREZITyn32oRAaAwWDwzlYUVTT2+vEJIfGMjRwFwLqSTbjd/tHr0+7Ng+/xTu4HPLX9H7jc/rGmUkREREYehQoZ9tr7Kop6eQBeu/nJcwEos1aQW5/fzdWek+Wb7FYabU002a1Y7c00O1pocbTQ6rThdDn7VMeJXmd31V4Aypsr2V21r1+eV0RERKS3/PKcCpH+1D5TUVnXQovNQVBA74b9rISpvLL/TWwuO+tKNjImMrPTa91uN8/sep6tFTs7vcZsMJESlkR6eBoZ4amkh6eSEpbcYcvaniizllNna/B+vLJgFVPjJvXqOURERET6g0KFDHvHNmsXljcxLq13zdZB5iBmJkxjfelmNpdt5/LxF2I2B53w2nUlm7oMFAAOt5P8hiLyG4pY3Xab0WAkJTSJmQlTOSfzjB4d/riv+mCHj3NqDlLUWEJqWHKPPi8RERGR/qJQIcNeZlIEJqMBp8vNrtyqXocKgAXJc1hfupkWZyvbKnaxMG3OcdfUttbx2sF3AEgIiePcUWfhdrtx4277E9y4qGutp6ChmIKGImpaawFwuV0UNhZT2FhMVvR4RkdmdFtTTo0nVMQGRVNna8DhcvBZwWqum3h5rz8/ERERkZOhUCHDXkiQmayMKPYcqWHbgUouXjKm188xLmoMcUExVLZUs7Zk03Ghwu1281LOmzQ7mjFg4GvZVzI2alS3z9tga6SgoYiChiLey/0Ih9vJjsrd3YYKp8vJ/ppDAEyLm0yrs5U1JRvZULaFC8cuJzwgrNefo4iIiEhfqVFbRoQZ4+IAyC9vpKqupdePNxgMzE/2BIn9NQepaq7ucP+W8u3sqNwNwNK0hT0KFADhAWFMis3inFFnkBUzHoDtFbu7fVx+QyEtTs/nkRUzjtPSFwPgcDlYXby+R68tIiIi0l8UKmREaA8VANsOVvbpOeYlz8aAp9dhbfEm7+2NtiZe3v8fwLMUacWY5X16/ulxkwFPA3ZZU3mX17YvfTIajIyLGkNqWDIToscB8EXhWhwuR59qEBEREekLhQoZEeKigkmL9ywJ2nagok/PERMUTVbbL+5rizd5z4V45cB/aLR7tqu9Nvtygsx9Oz1+Stwkb2jZUbmny2v3VR8AYFREOsFtTeNntM1W1Nnq2VredbO4iIiISH9SqJARY8Z4z2zFvvxarC19eye/fQlUZXM1eysOsqNiD5vKtgGwMHku2W1LmPoiMjCcURGeXoqulkDZnDZy6/IAyIo++nqTY7OJC44FPNvL+ttBfSIiIjJ8KVTIiDGzLVS07wLVF9Pjp3hnBt7b/yn/2vMaAJEBEVwy7oKTrnF6vGcJ1JH6fOpaG054zaHaIzjcngP02mdOwLMU6rS0RQDkNRT06KA+ERERkf6gUCEjRmZSOJFhAUDf+yoCTBZmJ0wHYGPRdmpb6wC4OusSQizBJ13jtLbD69y42dXJEqh9NZ6lTwFGy3G7RC1InkOQyRN6VhZ8edL1iIiIiPSEQoWMGEaDwduwveNgFQ6nq0/PMz95boeP5yTOYFrbDMPJSgxNIDEkAYDtlSdeAtXepD0uagzmr5zCHWQOYkGKZ4nWtopd1LTU9ktdIiIiIl1RqJARpT1UWFsdHCys69NzjIpIJznU84t/mCWUy8df2G/1wdHZipzqA7Q4Om5/22hvorChGPBsJXsip6UtwoABl9vF54Vr+rU2ERERkRNRqJARZWJmNAEWz7Dv6xIog8HA1yZdwczkKXx3xtf7/aC59r4Kh9vJnur9He7bX3MIz9nckB194qbwuOBYbzBZXbyeVqetX+sTERER+SqFChlRAiwmJo+KAWDbgco+75A0Lno0Pz/1+4yP7v3p3N3JjEgnIiAcgB1f2QUqp20r2TBLKClhSZ0+R/theFZHMxtKN/d7jSIiIiLHUqiQEWfm+HgAymubKa6y+ria4xkNRqa2zTTsqtqH0+X03revrZ8iK3ocRkPnX77j2w7EA1hZsNp7poaIiIjIQFCokBFn2rjYtiPm+n4Q3kBrXwLV7GjmQO1hAKqaq6ls9myFe+xWsidiMBg4PX0J4Dmhu/2wPBEREZGBoFAhI05ESABj0yKBvvdVDLQJ0eMINHm2v93RtgtU+65PAFk9OGRvTsJ0wi2efo+VBasGoEoRERERD4UKGZFmtu0Cdbionrom/2tkthjNTIrNBmBHxR7cbrc3VMQGxRAXHNP9c5gsLEmdD8Ce6hxKm8oHrmAREREZ0RQqZESa0Xa6thvY4aezFdPjPEugalprKWgoIqfaEyqyO9lK9kQWpy7AbDAB8Fnh6v4vUkRERASFChmhkmJCSIz2nIDtr0ugJsdme5ux3z/yCQ32RgCyOtlK9kQiA8OZnTgDgPUlm2iy+19juoiIiAx9ChUyIhkMBu9sxe7camx2ZzePGHwhlmAmRI0FOp6uPSF6bK+e57T0RQDYXHbWFG/ovwJFRERE2ihUyIjVfrq2zeFiT16Nj6s5sWltu0C1SwtL6fVhexnhaYyLGg14lkAdu0WtiIiISH9QqJARa1xaJKFBZsBzEJ4/aj8Zu113W8l2pn172drWOrZV7Ory2ty6PPbXHOrT64iIiMjIpFAhI5bJaGTaWM9sxbaDlTic/ndAXHRQFBnhqd6Pe7KV7IlMi5tEbFA0AJ8Vnnh7WbfbzX9zP+GhzU/w6Na/sKlsW59eS0REREYehQoZ0WZneU7Xrm+y8dbqXB9Xc2LT46cCYDFavMuYestoMHJamqe34nBdHkfq8zvc73K7+Pf+N3gn9wPvbS/nvElda0MfqxYREZGRRKFCRrQZ4+OYmOl5B/+9tfnkltT7uKLjnZlxKstHncm3pnzNeyBeXyxImet9/LGH4dmcdp7e+RyritYBEBccC0CTw8qLOa/hdrtPonoREREZCRQqZEQzGgx847xsggJMuNxunn5nD3aHfzUyW4xmVow5hylxE0/qeYLNwSxIngvAlvId1LbW0WS38ti2v3l3lxoXNZo759zundXYWbmHDaVbTu4TEBERkWFPoUJGvLjIYK4+09OrUFJl5Y0v/HMZVH9YmrYIAwZcbhdvH/6AR7Y8xeG6IwDMjJ/KrdO/RYglmAvHnkt824zFKwf+Q21rnQ+rFhEREX+nUCECLJmWzJQxMQB8sCGfA4W1vi1ogCSExDElLhuAdSWbKG0qA2Bp2kK+OeU6LCYLAIGmAK6feBUGDDQ7WvjX3le1DEpEREQ6pVAhgucwvBuXZxMcaMYNPPPuXlpt/rUMqr+cnrakw8cXjTmXK8Zf5D29u93YqFGc0bYV7Z7qHNaU6OA8EREROTGFCpE2MRFBXHuWZxlUeU0zr34+PM9qmBA9lkkxWQSYArh+4pWcPep0DAbDCa+9YMw5JIYkAPD6gXeoah6cQwLdbjd7qnLYVLoVl9v/tvoVERGRjsy+LkDEnyycksTmnAq2Hazkk82FzJoQ790dargwGAzcMv2bON1OzMauvwUEmCxcP/FKHt78BC3OVv617xVunfGt42Y1+ovb7WZfzQHePvwBefUFAOTW53P5+As7DT4iIiLie5qpEDmGwWDg68uzvCdt/+O9vTS3OnxcVf8zGAzdBop2oyMzWJZ5GgA5NQdZVbR+QGo6VHuER7f+hce3Pe0NFACfFa5mZcGXA/KavuB2u8mpPqgzQEREZFhRqBD5isiwQK4/JwuAyroWXll50McV+d55o5eREpoEwKsH3mJlwap+a9wuaCjiye1/55EtT3Kg9jAAYZZQLhl3PokhnsMJXz/4LlvKd/TL6/naW4ff53+3/ZXHt/1Nze8iIjJsKFSInMApExOZk+3pJfhsWzGHi/3vULzBZDGauWHSVQSYAnC6nbx64C3+tvNZrHZrn5/T6XLyYs7rPLDxUXZX7QM8Z2msGLOcexbcxVkZS7ll+k2EW8Jw4+afe/7Nwdqhvd3vodojfJT3GQDFTaUcrsvzbUEiIiL9RKFCpBPXnTWe4EATAM9/mIPLNbLfVU4PT+XOObd5Zyy2V+7m/o2PkluX3+vnana08OT2v3tP8Q4wBbA88wzuXXAXy0edQZA5EIC44Bi+N/0bBBgtOFwO/rrjn5Q1lfffJzWIWhytPLv3JdwcHUcby7b6sCIREZH+o1Ah0onIsEAuWjwGgCOlDXy5o9jHFfleUmgiP51zKwuTTwGguqWGR7Y8ySf5X/R4KU9tax1/3PIU+2oOAJAdPZ57F9zFirHLCbEEH3d9ZkQ635xyHQYMNDmsPLH979Tbhl4/wpuH3qOyuQqA2CBP8/+W8u04XcNz62IRERlZFCpEunDm7FRS40MBeO3zwzQ2231cke8FmAK4buLl3DjpGgJNAbjcLl4/+A5/2fl/NNgau3xsUWMJD256nKLGEgDmJ8/hlunfJDwgrMvHTY2bxFVZFwNQ1VLNU9v/QavT1i+fz2DYW7WfL4vWAjA9fgpXZV0CQJPdyt7q/b4sTUREpF8oVIh0wWQ08rVlEwBobLbzxheHfVyR/5ibNJM75/6A1LBkAHZW7uUXq3/LE9ueYU3xRpq+0m+xr/oAj2x+itrWOgDOH72Mr2Vfgclo6tHrLUldwLKM0wDIbyjk6Z3PdRtiBorD5aC6pYbcunwqm6u7vLbJbuX5fa8Angb0a7IuJTt6PGEWT1jVEigRERkOdE6FSDeyMqKZNymR9XvK+GxrEadOT2FsWqSvy/ILiSHx3DH7Vl47+DaritbhcrvYU53DnuocXsx5jeyY8cyKn4bT7eSl/W/icrswGoxcl30585Pn9Pr1Lhy7nJrWWjaVbWNPdQ6/Wft7lo86k9PSF2Pp4Ra5nbG7HDTaGmm0N9Foa6LB3kijrZEGexMNtkbqWuups9VT11pPo72pw2PnJs7iorHLiQ6KOu55X9r3pjdIXZt9mXdWZnbidD4vXMOOit20OFq9fSQiIiJDkcHth3saVlT4fr202WwkOjqUmpomHA6d6DvS1TS08ou/raPV5mRsagS/unEusTFhGh/HKG4sZXP5draUbae8ufKE1wSZArl56g1kx4zv8+vYXQ7+nfM660o2eW+LC4rh4nHnMyN+Sq8PybM77bx64C3WlGw8qdO7LUYLZ2UsZVnmaQSaAjCbjexv2s/Dq/8KwLyk2dww6Srv9bl1eTy0+QkAvj7pak5JmtXn15ahRz9jpDsaI9KVwR4f8fHh3V6jUNEJfTHLV/13fR6vrDwEwM0rJnHhaeM1Pk7A7XZT1FjClvIdbCnfTkVbc3JUYCS3TP+md7nUyTpSn89rB97usC3ruKjRXDZ+BRnhaT16jpqWWv6267kOh+0dy4CBUEsIYQFhRAVEEBkYQURAOJGBnr+HW8LYVLaV1cUbvLs6RQZEcOHY5UxNyOZ/1j1CQ2sj0YFR/L95PyLYfLQR3e1285t1f6CyuYpJMVl8f8ZNJ/GvIUONfsZIdzRGpCsKFT2kUCH+yOF08etnNlBabSUiNIC//vwsbC02jY8uuN1uChtLOFKfz/T4yUQEdP9NqbfPv6V8B28eeo/qlhrAEwTmJM5k+ajTSQpN7PSxB2tzeXrX0b6MKbETmZc8mzBLKGGWUMIDwgi1hGA0dN96VtRYwusH3vHuaAUQaAqk1dkKwG0zbj7h7Mw7hz/gv0c+wWgwct+iX3bbsN7V61e31DA1blKfHi+DTz9jpDsaI9IVhYoeUqgQf7U7t5qHX9oGwIVLxnD50jEaH37A7rTzacGXfJD3qXdXKAMGZsRP4ZxRZ5IenuK91u1282XROl458B/vcqfzRp3FuaPP6lGA6Izb7WZ31T5eP/gOZdYK7+2npS/iivEXnfAxZU3l3Lv+IQCuGH8Rp6Uv6vXrNtmt/HrNA7Q4W7hh4lXMS57dt09ABpV+xkh3NEakK/4YKtSoLdILk0fHMDsrns05FbyzOpdJGVGMT4vs9Tp+6V8Wk4VzRp3B/OS5vH/kY9aUbMThcrC1YidbK3YyJTa7LVyk8nLOG6wp2Qh4ejxumHQV0+OnnHQNBoOBKXETmRgzgS+L1vFx/mckRyRw2YTzoZO3bhJDE8gITyW/oYiNZVv7FCo2lG6hxdkCwMf5n3NK0iyNx07k1xfycf7nlFkrSAyJJyUsiZTQJFLCkogJij6pUCkiMtJppqITeodAOlNV18L/+9s6bG3jIi4yiFkT4pmdFc/Y1EiM+oXO52pb6/gk/wtWFa3D5jp6tki4JYwGu2e5U0JIHN+Z+vUul0idjJ5+D/kk/wteP/gOAHfP/xkJIXE9fg232819G/5IcVOp97bOllqNZAdrc3n/yCddngkSaAogOTSJ0ZEZnJI4i/Tw1AENZ/oZI93RGJGu+ONMhUJFJ/TFLF1Zu7uUf7y3F4ez45dPRGgAs8bHMTs7gUmZ0XrH2McabU2sLFzFZwWrve/mg6d/4sbJV3donO5vPf0eUttaxy9X34cbN+ePXsZ5o5f1+DWO1Ofz4KbHO9w2OTabW6Z/s891Dxdut5u91ft5/8inHKrL9d4eYApgYvR4KluqKW0qx+k+8YnmSaGJzEuaxSlJs4gK7P8tpPUzRrqjMSJdUajoIYUK8XdmsxFLoIXPNuWzcW85uw5XeWcu2s0YF8c3z59IWLDFR1VKu2ZHM58XrmVL+XZmxk/jnFGnD/hSl958D/nfrX8lp+YgCSFx/HreT3scRl/Y9yqrizcQYLSwMOUUPitcDcCv5t1BUmjCSX8OQ1VefQH/znmd/IYi723B5mBOS1vEaemLvAcPOl1OyqwVlDSVUtxYSmFjCTk1B7C7HN7HGTCQFT2OecmzmRE/hQBTQL/UqJ8x0h2NEemKP4YK9VSI9FFYSACLpiYzb2IirTYnu3Kr2Ly/gu0HK2ludbLtYCV3/30D314xiayMaF+XO6IFm4NZPuoMlo86w9elnNDcxJnk1Byk3FpJfkMhmRHp3T6mxdHKprJtAMxKnM7ZmafzZdE6nG4nKwtXcU3WpQNctX9qcbTy1PZ/eJe5hQeEcWb6qSxOnU+wOajDtSajydNXEZbE7LZVcFZ7M1srdrC+ZDOH6o7gxs2+mgPsqznAJ2Ep3DH7+1hMeqNAROSr1JUm0g8CA0zMzkrg2ysm88dbF3P6zFTAc2jeH17cyptfHsbl8rtJQfETMxKmYG47EXxj2dYePWZL+Q7vTleLUk4hMjCCOYkzAFhfsvm4U79His8KV3kDxUVjz+XeBT9nWeZpxwWKzoRYglmUMo8fz76F38y/k/NGLyMuKAaAwsZiVhasGrDaRcS/OV1O3sv9iE/zv/B1KX5JoUKknwVYTFx/Tha3XDyFkEAzbje8tfoIf3hxK9X1LSd8jMvtpraxlVbbidd3y/AWbA5mauxEADaXbe/Ryd5rijcAnrX/oyMyATg9fQkAdped1UXrB6ha/9Vkt/Jx/ucATIgay7KM0wg4iVmF+JBYzh+9jN8suJOxkaMBeD/vE+pafb9EV0QG36ri9byb+xGvHXyH3GMOXhUPLX8SGSBzshMYlRzOX9/aw8GiOvYX1HL33zdw0eLR2BwuKutaqKxtpqKuhaq6FhxOFxGhAfz0mpmkxoX6unwZZHOTZrK1Yif1tgZyag4yMWZCp9cWN5aSW+/5gbYoea63ByM9PIXxUWM4UHuYzwtXc2bGqd4ZkJHgo7zPaHZ4gvuFY5f320YJBoOBy8ev4A+bHqPVaeOdw+9z3cQr+uW5RWRoaHa08F7uR96Pd1buZXRkpg8r8j+aqRAZQHGRwdx53UwuWJiJAWhqcfDCxwd49bNDfLa1iF251ZRVW3E4Pe9M1zfZeOSlbVTWNvu2cBl0k2KzvbtRfVawmq720FhT4pmlMBlMnJLU8bC7M9pmK+psDWwp3zFA1fqfutZ6b6P61LhJ/f7DPiMizXuw4NqSTRQc0wQuIsPfR3mfdVhWurtqnw+r8U8KFSIDzGQ0cumpY7nj6hlEhwe23WYgISqYiZnRnDo9mUtPHcOKhaMATx/GQy9to67J5sOqZbBZjGbmtwWEXVV7+e+Rj094nd3lYEPJFgCmx08mLKDjrNaUuInEB8cC8GnBl12Gk+Hk/SOfYHfZMWBgxZhzBuQ1LhyznABTAG7cvHbg7QH7t82rL2Bn5Z4eLYMTGWgFDcU02kZmj1a7mpZaPi3w9FG07wBX2FhMbWudL8vyOyNnXlzExyaOiuHB7y2krslGZGgARuPxSzOCA828vPIg5TXN/PGlbfzs2pmEBGmnmZHiwrHLOViXS0FDEe/mfkRSaCKzEqZ1uGZHxS6aHFYAFqacctxzGA1GTk9fwsv736SgoYiDtbmMjx4zKPX7SmVzFauKPT0kcxJnkBqWPCCvExkYwTmZZ/D24fc5UHuY7RW7mJEwtV9f43BdHo9sfhI3bmYnTOf6iVf6ZLcpl9vFu7kfUdRYwteyrzguvMrIsKZ4A//a9yoxQdH84pQfDujZPv7sncMfYnc5MGDgxknX8Ned/wRgd+U+FqXO83F1/kMzFSKDyGg0EB0eeMJAAbB8XgbnL/As28gvb+TRV3fQalfz9kgRYArgO1O/TmSAZz/wZ/e8RF59QYdr1hRvBCAmKJqs6HEnfJ55SbO9P/xXFnzZp1rsTnv3F/mJdw5/hMvtwmgwcv7oswf0tc5IX0JMkGeL6NcPvtvhTIuTZXPaeG7PS7jxzIBsLt/O/277Kw22xn57jZ6wuxw8s+t53j/yCTsr9/BR/meD+voyMPLrC9lRsbvHM2z1tgZeP/gOANUtNbx56L8DWZ7fKmwoZn3pZgAWJM9levxkEoLjAC2B+qp+DxWPPfYYWVlZHf5bvnx5f7+MyLB16aljOK1tS9oDhXU8+cYub89Fuxabg8055Tz9zh5+/pe1vLPmyIhZ5jLcRQdF8Z1pN2IxmrG77Pxlxz+9U+yVzdXsqzkAwMLkuZ0e4BdkDmRxiufdsx2Ve6iwVvWqhg+OfModX/yaf+x+AafLv0NtcWMpm9q24V2UMo/4kNgBfb0Ak4WLx54HQFVLNZ/14xaz/zn0X8qbKwG8sy2H6/J4aNPjlDaV99vrdKXF0cJT2//Otopd3tvWlWzqVXhyu909+n7kdDmpaanlSH0+2yt2sb5kc78tJ6m3NfBh3krWl2zul+cb6mpb63hky1P8Zec/+aSgZ9uhvnHwXe/GBwCritZxsDa3i0cMT28eeg83bgKMFi4Y43nTYnJcNgB7v3JY5kg3IMufxo8fzz/+8Q/vxyaTaSBeRmRYMhgMfG3ZBKwtdjbsLWfn4SqefmcPV50xnu2HKtl2oJI9R2o6BI3XvzhMZV0L158zAZNRE5BDXWZEOtdPvJK/736BOls9f9nxT34067usLfHMUhgwMD95TpfPsTRtIZ8UfIHL7eKzwlVcMeGiHr32h0dW8tbh9wHYVLaNEHMwV064uN92Uupvbx/+ADduLEbLoB1uOCthGp8VruZw3RHeP/IJ85JnExHQ/WmzXdlfc/CYRvOJfHvq13nz4Ht8UvAFlS3VPLT5Cb499XomdDI71R8abU08uf3v5DV4ZsdSw5Ipaiyh0d7EjopdzG47B6UrZdYKHt3yF+ptDQSaAgg0BRBgCiDQFEiAKYAAo4UmexO1tnoabU3eWZl2RoORqbETWZw6n+yY8Z0G587Uttbxcf7nrCpaj93lmW2zmCzHLSMcadYUb/D+e7x16H3GR43p8pDNAzWH2FDq6d2aET+V/TUHsTqaeWHfa/x87g9GzAGQe6v2s7d6PwBnZiwlMjACgCmxE1lZsAqb08bB2sNd7tY3kgxIqDCZTMTHxw/EU4uMCEajgW9dMAlrq4Ndh6vZsLecDXuPf6cyIsRCYICJitoWvtheTIPVxncvmozF3Pcgn1/WQFFFW1OeocMfYIBAi4mwYAthwRZCgy2EBpkHNcjUNLTywL82YzGb+PnXZhE6THtOZifOoLSpnPeOfEx+QyHP7n3Zuy/6pNgsooOiunx8dFAUsxKmsalsG2tKNjIrYTpjo0Z1+ZhP87/gP4c9SxwMGHDj5ouitcQFx3Jmxqn98Wn1q9y6PHZU7gbgtLRFRAVGDsrrHrvFbIuzlXcOf8i12Zf1+fmaHS08t/cVAELNIVyTdTlGg5FLx19AfEgsL+//D82OZh7b9jTXZl/Ogm4CZV/UtNTy2LanKbN6vs/MS5rNtdmXcc+6B6luqWFV0foehYp3Dn9Ana0egBZnKy3O1l7V4XK72F65m+2Vu4kLimFR6jwWJM8lPCCsy8dVt9TwUd5nrCnZiOMr7xz/O+d1xkWNPungN1Q5XU5Wt51rA+B0O/n7rn9x1yk/POGhkA6Xg3/nvAFAqCWEa7IvZWfFHp7f9wpl1nI+yPuUCwZoMwR/4nK7eOPQuwCEB4RxVsZS731jo0YTYArA5rSxu2qfQkWbAQkVeXl5LF68mMDAQGbMmMFPfvITUlJSBuKlRIYts8nI9y+ZysMvbeNg4dElAcmxIcwYF8fM8fGMSYmgxebksdd2kFNQy9YDlTz80nZuv2xqrxu8y2usvPrZITblVPS61pBAM+EhFk6dkcLyUzIG9F3tlz49QEWtZ0r+/fX5XLZ07IC9lq+dO/osSq3lbCnfwdZjtoc9UYP2iZyZfiqby7Zjc9r445anOHf0WSzPPAOT8fjQ+XnhGl5rWz8dGxTNd6bdyDO7/kWZtZw3Dr5LTFA0M/u5KflkvXX4AwCCTEEsyzxtUF87MyKdeUmzWV+6mTXFGzg1dQFp4X37Off6gXeobqkB4KqsS4gMPPrL75LUBcQGxfDMrudpcbby/N6X2V9zkPTwVOKCYogNjiE2KIYgc2CfP5eypnIe2/Y0Na21gKdv5JJx52M0GFmUcgpvH/6A/bWHKLNWkBjS+RuGZdYKtpbvBGBKbDajIjJoddpodba2/WnD5rIRYg4mKjCSyMAIIgMiiAyMICowAofLydqSjawr2USjvYnKlmr+c+i/vHP4QybHZhMVGEmQOZBgUxBB5kCCzEEEmgLYXbWPdSWbcbo9S/UMGJiZMJUJ0eP4d87rNNmt/Hvf69w89Qa/nXEbSDur9nqXlc1OmM7m8u1UtlTz75zXuXHSNcf9m3xa8CWlbeHy4rHnEWYJZX7yHDaUbWV/zUE+yFvJrITppIQlDfrnMpg2lG6hqLEEgPNHn93ha8xiNDMxejzbK3ezu3Ifl4+/0Fdl+hWDu58XYn/++edYrVZGjx5NRUUFTzzxBGVlZbz99tuEhXX9TkO7qqrGThtZB4vJZCQiIpj6+macTm3rJx0N5viwtjh4d+0RQoLMzJoQT3Ls8buw2BxO/vzmLjbt8wSC9IQw7rhmpncL2640tdh5a9URPtqYj8N58t8OzluQyVVnjBuQH967c6v5/b+2eD8OsBh5+PuLiAzr+y9UA6W/xojNaeOhjU+SV18IeN4x+/2pvzphMDiRzaXbeW7PK9610WOjRnHT1GuJDY7xXvNl4Tqe3/Mq4JnhuGPO94gLiaXSWsUDGx6jwdaIxWjmJ3O+x+go3x721GBrpKChmP3Vh/hv7icAXDhuOeePOWvQa6ltqeNXqx7A5rITHhDGaekLOTVtARGB3b8j3j4+Vh/awqOb/gbAnKQZ3Dztaye8vqihhMe3PkN1S+0J7w+zhBIXHMP0hMmclXmqd9vL7hysyeWpbf/n3X//4nHnsnz0Gd6v37rWeu764re43C6WZS7l8qwVnT7Xs7tfZnXRBkwGE79b8vNuZ9M6Y3c52Fq2ky8K13Kg5nCPH2fAwNzkmZw3+kySwxIBeHHv63xWsAaAb0y5hvkps7t6Cr/SX99D/rT5r+yt2k9UYCT3LfkF/9z9EuvbtqW+YfKVLEo9+iZFVXM1v1n9IDaXnbFRo7hj7i3eJWgV1kruWfMQdpeD0ZEZ/OyUW3u9PG2osDnt/HrVA9S01pEcmsCvFvzkuO+5qwrX89wezwzjvYvuJDF0cFfoDPbvqdHR3e8A1++h4qvq6+s5/fTTueuuu7jiip6dQOp2u0fkuwkiJ8PpcvOX13fw37VHAEiIDube7ywkNf7EYd7hdPH+2iO88EEODVbPmRgmo4HzF43m4qXjCAwwHdds6XZDc6uDBquN+iYbDVYbDU026q021uwopqDMs0vNxUvH8s0Vk/v169jucHHbQyspqmgkNNhCU7NnffCKJWP49sX+9Q56f6u21vLzjx+gprmOyyadx1VTO//F7kTKm6p4bO3fyany/IIWYgnm23OuY2HGbD7LXctTG57DjZvo4EjuOf3HJIUneB97sOoIv1n5CDannYjAMH531s9IDBucH54Op4Otpbs5VH2EIzWF5NYWUNPcsZE3MjCcx86/lyDL8cs4BsM7OR/z7LbXvB9bjGYWZ57C+RPOICMqtcvHNrY28ZP3/4ealjoigyJ4ZPmvCA/s/M232uY6/r71ZQ5U5lLdXHtcP0K72JBorp9+KQvSZ3f6NVjSUM6LO/7DukLPL5cGDNw85xrOGrvkuGsfWvUXNhRtIzwwjD+vuO+E6+krm6q57d1f4XS7OGPMIr4798ThqLcK60r46NCX7Ks8SLO9hWZ7C1ZHS4fdyYwGI6dmzuOSSctJPmbsArQ4WvnpB7+jrLGCEEswjyz/NTEhUZ2/Xn0JHx38ksSwOCbGjyczMhXjEO5TK2ko5wfv3Q3AlVMu4PLJ59Nsb+HOD++jtLGCQFMA9599F2kRno0B/rDqz2wq2o7RYOT3Z/+czKi0Ds/31r4PeX67Z2nUN2ZeybkTTh/cT2iQvLHnfV7c+R8Afrb4e8xJPb4np9pay3ff/jkAX59xOednnTmoNfqjAQ8VAJdddhkLFy7kJz/5SY+u10yF+Dt/HR9ut5s3v8zljS88vzyGBVsYnx6JxWTEYjZiPubP7QcrKamyeh87OyueK88Yd8KZkJ6ob7Lx+39toaDcEyyWz8vgmrPG91uweGfNEV7+9CAAN6+YxM5DVazbU4bZZOAP31tIXFTX+6cXVTTyny9zWTQtmenj4vqlpq709xipa61nf81hZiVM7fEsxbGcLifvHf6Ydw9/7P1ldHJsFnuq9uPGTURAOD+Z+z2SQhOOe+zWsp38ZfuzuHGTGBLPnfNuI9QSctKfU2ea7c18UbiOT/NXdbobkNFgJDk0kYvHn8u0+EkDVktP7Krcx8d5X7C3an+H2yfGjOf0jMWMjswgLCC0w7u6JpOR/9vzb1bleda63zLjG0xPmNzj17S7HFQ311DZXE1lcxWVzdUcqj3Codoj3mvGRY3myuyLyIw4+othfWsD7xz+iC8L13kP1wsyB3HDpCuYnTS908/vsS1PA/CtaV9jbtKM46759743WZm/CgMG7l18JwkhA/s15nA5aHG00uJoIdgS3OV4PFiTy0MbPWd/TI7N4rZZ3zru+5Lb7WZN8UZe3PuGt6EZINgcxLio0YyPHsP46DFkRqT16euvL/rje8grOW/xcd4XGA1G7l/y/4gK8vQd5dcX8sD6x3C6naSFJXPXvNvZU7WfJ7d5Ntk5K/NUrsg6fkmP0+Xk/vX/S0FDEYGmAH6z8KfEBEf3/ZP0Q3Wt9fx61e9pcbYyIXoMP57zvU5/jv127SMUNBQzMXYCP5z97UGt0x9nKgb88LumpiYKCgp61bjtcrlxufxje0yn04XD4T+/NIp/8cfxsWLhKMKCLTz/QQ6NzXa27q/s8vrMxHCuPnMcWRmeHwx9/XxCAs3ccfUMHnxxG4UVjby/Ph+n083VZ578Uqjq+hbe/NITlMalRTJvUiKjkyPYsLcch9PN618c5pvnTez08Y3Ndh56cRtV9S1syinnZ9fMYlza4DT19tcYCTWFMTNuGm4XOFx9eT4D545axviocfzf7hepaa1ld1UO4Fk6c9uMm4kLjDthrVNjJ3Pp+At47cDblFkreHLrP7gm6zIiAsIJNgf1W3CsaallZcEqVhev79DgG2AKIC0smbSwVNLCk0kLSyE5NImAtnfMff01mB01geyoCRQ3lrKyYBUbyrbgcDnYW32AvdWeLYDNBhORgZFEB0USFRhJoDmA1UWeQDE/eQ6TYyb26vMwYCQ2MJbYwFiyosYDnl+Md1Tu5vUD71DZUs3B2lzuX/co85PncHbm6Wws28on+Z/T6mybmTSYODVtAcszzyQsILTT158QOY6YoGiqW2r4In8tM+M6vmtbb2tgVeE6AGYnTicmIGYQ/p8YCTIGExTgeTOhq9cbFZ7J6emL+bTgS3ZX5fBFwToWpRw9sKzF0cpL+9/w7nZkNBi9gavZ0cLOyr3srNzrvS/QFIDJYMJsNHv/sxhMGI0mcHPMDNLRvwWbg0kJTSQlNInksCSSQxNP2CR9In39HmJz2llT5NkxbnrcZMLM4d7nSQlJ4eJx5/HagbcpbCzhhT2vs6/G86ZNVGAk52ae1clrGrg2+zIe3PQ4rU4bz+95je9N+8awWl3yyr63aXG2YsDAxWPPx+l0QyezgpNjsiloKOZA9SEaW5pPqrepr/zp95B+DxW///3vOf3000lJSaG8vJzHHnsMo9HIBRdc0N8vJSKdOH1mKnGRQazcUkSLzYHd6cLhcONwujx/d7oICTRzzikZLJiShLGffiCEhwTws2tn8uCLWykob+SjTQW4cXPNmSc3Y/HvTw5gs7swGOBryyZgNBhIiglh0dQkvtxRwuqdJZw7L+OEsywul5u/vLWbqnpPT4HD6ebx13fwq6/PJTbSN0tmfGlc1Gh+ccqPeDHnNbaU7yDEHMxtM27utuny9LTFVDZX83nhag7W5vI/6x8CPL8shweEExEQTkRgGLFBMSxMOaVXp1oXNZbwcf7nbCrb5v1lDmBs5CjOyljKlLiJQ2LtdkpYEtdNvJwLxy5nVdE6Pi9a4z24zuF2UtVSTVVLdYfHRAdFcfn43i1n64zBYGB6/BQmxWazsuBL3j/yCa1OG2tLNnq3I243J3EGK8YsJ+6Y3prOdNewvbJglXev/rMz/XM5zIoxy9ldlUOZtZzXDrxNdvR4YoNjKGos8W5IABAXHMtNU64jJjCaQ3W5HKzN5WDtYQoainHjxuV2dTi7oTf2t/3S3i46MIqUsCRmxk9lfvKcfv/FfHP5dqyOZgBOTVtw3P2npy0mp/ogu6r2suaY8XHZ+BUEdRF4MsLTOCN9CR/nf87uqn28mPMa6eFpxAXHEBcUS0xQ1KDN5vS3/TUH2eg99+aULrfdBc95Fe/nfYrD7SSn5gDT46cMRpl+q9+XP/3oRz9i48aN1NbWEhMTw+zZs/nRj35ERkZGj5+joqKhP0vqE7PZSHR0KDU1TX6TAMV/aHx0zTMzsJX8tqVQZ81O6/NSqF25VTzy0nbv81y77OjWfVV1Lfz8r2txON3MzU7gexcf/w399S8O8c4az1as2RlR7MuvBTzN7D//2iyCAgZmwtbfx4jb7aagsYiYwGjCAnq25M3ldvH33S902ImqM5Niszg74zTGRY054f93u9PO1oqdfFm0jsN1R7y3G/D8YnxWxqmMjvRtU/jJsrsc5NUXUN1SQ21LHTWttZ7/Wuqoba3DYDTw7anXMyZi9IC8fl1rPW8dfp91JZu8t2VHj+eiceeSEZ7WxSNP/Fy/XHMfLreLMzNO5dJxnjcKrfZmfrXmPlqcrUyNm8R3p93Yn59CvzpSn89Dm57AjZsJUWOZnTidVw+85Q1EsxOmc032ZSecQWh2tHC4Lo/ixhJsLjsOlwOHy4HT7Wz7u7PD7lO0/a1dg62B4qZS6m0n/v1mYfIpXJ11SYdfxk/2e8gfNj5GXkMBiSEJ/GreT074ddhoa+K+DX/0bgM8KSaLW6Z/s9vv1Tanjd+tf4TKr4Rk8Hz+0UFRxAZFExEQTnhAmOc/SxhhbX+PC47ply1+nS4n2yp2sbNyDzPipzDjJHaoc7gc3L/hT5RaywmzhPLr+T/tdpmny+3irlX30mS3sijlFK7NvvyE1+XVF/BO7oeEW8I8s67hKaSGpZzUMtLB/hkTH9/9/69B6anoLYUK8XcaH91rbLbz0L+3kt/WvJ0SF8q0sbFMGxPLuLRIzKbu33m2O1z8+u8bKKu2EhEawH03zztuq9wXPt7Px5s8OyPdfeNcMpOOfuPbeqCCx17zbHE5LjWSn107k/fX5/N6W8/J7AnxfO+SKX2eqWlstlNR20xIoJngIDMhgWbv5zVcx4jb7aakqYya1lrqWxtosDVSb2vw/nekPr/DCbOZEeksyziN6fGTMRqMlFsrWFW0nnUlm2hyHO3psRgtLEiew+npSwZ8Pb4/GMzxkVdfwKaybUyKzTqp/fT/uvNZtlfsIswSym8X/T8sRjP/zf2Ed3I9W/veMftWRkf2/A1EX3jr0Pt8kPdph9vMRjOXj7+QxSnzBnwZT6O9iZLGMkqaSiluKmNvVY73F/MJ0eO4ecrXCGn7RfNkxkh+fSG/3/S/AFw+/kJOT1/c6bX7aw7xxLanMRst3DX3Bz0+lb6osYRX979FcVOpd/ewnjJgYEnqfC4cu5xgc9f9cCdS19rA6uJ1rCpa7w1EZoOJX8z7cZfbHnflw7yV/OeQ55yer2VfwYKUuT163P/tfpGNZVuJCozktwt/cdwYKm4s5Y9bnvLOGh0rOjCKtPAUMsJTWZQyv8N20t1RqOghhQrxdxofPdPYbOfhf28jr6zj13RwoIlJo2KYNiaWKWNiO9369t21R3jtc08A+NYFE1k45fglNXVNNu7681pa7U6mjY3lh1d4mk3Lqq3c+8+NNLc6iQgN4O4b5xIdHojb7eavb+9h/Z4ywNODcsmpY3r1ebncbj7eVMjrnx/C9pX//4EBJkKDzIQGWRifEc287HjGpUYOqzXHXWmwNfJ54Wo+L1zT4YdoQnAcUYGR7K891OH6+OBYFqfOZ37SnB7PmAwHQ/F7yO6qHJ7c/gwA35h8LVPjJvGrNffRZLeSFT2O22cObqNqX9hdDh7c9Jj3/IGEkDhumvy1Pp8xcrKaHc08s+tf3lObE0Pi+d60bxIfEnvCMdJga2RV0Tpyag6yMOUUTkmadcLn/dfeV1hTspEAo4XfLfolIZauf3Gvaq7GZDT1+QDJFkcLVS013k0DqlqqqWquocHeSKOtkXpbY4cG+HYRAeFcPv5CZiVM6/Z7pNvt5nBdHl8UrWFr+U7vzNCxsqLHcduMm3v9/ba6pYb/WfcQNpedMZGj+NGs7/Z4yeXG0q38354XAbhr7g9JP2YsVTXX8PDmJ6iz1WM0GIkLjqHCWnXCndtCLSFcm305M3q4hMofQ8WAN2qLyMgVFmzhrutmsWZXCTsOVbE3rwabw0Vzq5PNORVsbjtoLzo8kFFJ4WQmhbf9GYHD4eLt1UcAmJAWyYLJJ17zHxkawLK5abyzJo8dh6rYX1BLZmI4j7+xk+ZWJ0aDge9dNNkbXAwGA984N5vyGiu5JQ28veYIqfGhnDIxsUefU1mNlb+/u5cDhSfelajV5qTV5qS6vpWC8kY+3VRAQnQwS6Yls2hqMlF+eKZGfwoPCOOCMedwVsZprC3ZyCf5X1DTWkt5cyXlzZ5NA4wGI9PjJrM4dT4ToscOiX4J8exm1d6wvbpoPfW2BprsntmmczLP8HF1PWMxmvnm5Gt5Yd/rpIYlcdHYc7vsHxhoweZgvjftG7xy4C2+LFpLmbWCBzc/xrenfp3suKMHexY2FLOycBWbyrZ5Tww/UHuYPVU5XJV1cYd3+612KxvLtgEwJ3Fmt4EC6HBuTV8EmYNIDUvuspeq1Wlrm9msZ2XBKraU76De1sDfd/+LdSWbuCrrkuN6fJwuJ7n1+eypymFn5R6Km0o73J8dPZ6laQvJqy/g/bxPyak5yKaybcxNmtmr+l898DY2lx2jwcjVWZf06nvSpNgsDBhw42Z31T5vqGiwNfL49r95Z1Kuy76c+clzaHXaKG4sobCxmMKGYgoai8mrL6DJbuVvO59lfvIcrhh/oU/HZV9ppqITQ/FdJBk8Gh99Y7M72V9Qy45DVew4VEV57fHTwe0CzEZsDhdGg4HffGMuaQmd799vbbHzs6fWYm11MCEtkuiIIO9MxNVnjOPsU45fklHT0Mr//HMjtY02LGYjd103i9HJEZ2+hsvt5pNNhbx2zOxERmIYl546Brfbc0hhU4sda6sDa4uDBqudHYcqaWo5uhTIaDAwbWwsS6YnMzYlkpAgc4+WgQ1lTpeTzeXb+bxwDS3OVuYkzGBhylwiAzv/tx4Jhur3kPePfMLbbSeZh1pCaLJbGR2RwU9mf3/EzMYNBLfbzWeFq3ntwNu4cWM2mPja5CtIiIriP7s/Iqem4wxfiDnYOxMYGxTDjZOvYUxbD9LKglW8euAtAO6ce3uv+2cGy67Kvby8/02q2k6TtxgtnDfqLOYkzSCn+iC7q3PYV32A5q8sGwoyBTI/eQ6npi4gsW0LbLvTzm83PEJlcxXhAWH8et5PexSm2ut4aodnK93T0xf36XTshzc/yeG6I4yJzOQns79Pi6OFR7f+lfwGz9LcS8adz1kZSzt9/L7qAzy392XvFtqxQTHcMOkqxkV13m/ljzMVChWdGKrf8GVwaHz0j7JqK/vya8grbeBIaQOFFY3Hnep99tx0rj5zfLfPdexSqXanTEzgOxd2fghfbkk9D/xrC3aHi8iwAC49dQxxEUHERgYRExHk/YW/vMbK39/bx/6CWsBzSOCKhaM4b0Fmp6HAbDYSHBrIx+uO8NmWInLaHvtVgRYTIUFmQoPMhARZiAixsHhaMtPGDv++gpFsqH4PObZhu913p93I1DjfnhUyXOys3MPfd7+ArW3b32NZjGZOSZrNaWmLiAmK4uX9/2F96WbAM/t33qizODvzdH634RHKrBWMisjgp3NuHexPoVdsThvv5X7MJwVfdBhTX2U2mBgXNYbp8VM4JWnmCd/F31u1n8e3e85TOTV1AVdlXdKD17fzu/UPU9lSTWRAOL+a/9Meb/V7rPePfMrbh9/HgIHfLvoFz+55iZy23b6WZZzGxePO6/Y5rHYr/855g83lno1JDBhYlnka549ehtl4/MIihYoeUqgQf6fxMTAcThdFFU0cKa0nr7QBs9nIZaeOJTCg++0JW21O7vzLWuqbPD+MU+NC+X83zO52d6cNe8v48392H3e7AYgICyAuIoiCikZs9rbZiYQwvnn+RDISu/4G+9UxUlpt5csdxazeWeqtsSuzs+K59qwJnfabyNA2lL+HtDdsA6SGJfPzuT/ULEU/Kmgo5s87/uF91zoqMIJTUxeyKGXecX1HG0u38u+cN2hxera5TQpJoLRte9zrJ17J/OQ5g1t8HxU1lvDvnNc5XJfnvS02KJpJsdlMjs1ifNTYHp0B8fdd/2Jz+XYMGPjpnFu73RL23cMf8t6RjwFPn9CcxBl9qr+woZj7N/4JgLigGG/z/fzkOXwt+4pefX1sLN3KS/vf8G5dnB6eyg9mfvu4hnaFih5SqBB/p/Hhn1ZuLeK5D3IIDjTxq6/PJSmmZ9v1fbypgDe+zKW51dHpNSajgQsWjuL8LmYnjtXZGHE4XRworKO2sRVriwNri52mFod3+dTh4nrq2kJHYICJS5eM4czZaRiNfd+hqsXmICYiqN/OI5GTN5S/h+ypyuGJ9obtSdcwp5fr16V7da31fFm8lvGJmWSHZeF2df61W9lczf/tfoHc+nzvbaHmEH676P95D4gcClxuFzsq91DfWs+E6HEkhsT3OqzWttbxP+seosXZSnp4Kj+bc1un/RHl1kp+t+ERHC5Hnxu827ndbn655j5vEASYGjeJm6dc36czO2paanl2z0vejS2+Punq45ryFSp6SKFC/J3Gh39yu93syq0mMTqYhOje7f/tdrtpanFQVddCZV0L1fUtVNW3UFXXgslk4Lz5md3OThyrr2OkudXB618c5tMthbR/d85MDOeG5Vld9nx8VUlVE++ty2Pd7jKcLjcBZiNJMSEkx4WSEhtCcmwoyXGhJMeE9DmwSN8N5e8h7ev/nW4nZ6QvUaP9AOnNGHG6nPz3yMe8f+RT3Lg5K2Mpl4w7f5Aq9S/H9pRcMf4iTktf1OH+9pPn/3Pofcqs5ZgMJv7fKT/y9mf01Qv7XmV18QYAxkaO5tYZ3zqpUOdyu1hbvJGSpjLOG32Wd6vhdgoVPaRQIf5O40O6c7JjJLeknmc/yCGv1PP90ACcNiuV+ZMSyUwMJ8By4ne/ckvqeW9tHlv2V5xg08LjxUUGsWLRKBZOScJk9L9fDstqrLS0OgkPsRAWbOn08z5Z9VYbH24ooLqhhdiIIOKjgomLDCLuK/01/UXfQwQ8G0AY4ITvkPdljBypz+dw7REWp84nwBTQz9UODU6Xkwc3P05BQxFBpkB+Nf8OogIjPW86Ve3l3dyPKGgo8l6/fNSZrBhzzkm/bn59IY9seZK0sFRumf7NHjeK95VCRQ8pVIi/0/iQ7vTHGHG53HyypZA3vjhMi+3onuwmo4G0hDDGpkQwNiWSMSkRVNe38O66PPYcqTlag8nA4qnJZGVEU1ptpaSqieJKK6XVVhzOjjUlRgdz4eLRzJuY2OeZC2uLg8q6ZuKjggkOPLkdy+0OF89/mMOXO0o63B5gNhLWFjDCQwIYlxrJtLGxZCaF92l5l9PlYuWWIt78MhdrJ8vfDAaICQ/EbDbhdLpwOF04nO62P12AgYVTErlu2QQs5p6Fnr6MD4fTxaGiOkKCLKR3sRuaeJTXNmMxGf2uL8nucLLjUDUb9pax/WAlyXGh/PCK6USGdgwB+jnTc+W1zUSGBhDY9qZDXn0BD256HDduZiVMY17SbN7N/ci7GxNAVGAky0edyaKUU/ptts3ucvD/27vzqCjvew3gz2zsMCzDNoAoICDL6OC+R5OYGGPrEnOvN6vH1pM0S28be3PbprEaE2172hN7m1RrEo/XLJakMY3bjRqXNEaDCwgqKIgoOzPDMsAMs73v/WNkdGQbMwoYns85HmTeF+Y3wxfmfea3ySWyfpljxFDhIYYKGuxYH9SX21kjTa0WbP+yFCcvNMCTv9i+PjLMGhOH+8cndHtBJQgidC1mlFcbsfv4FdTor++Gq1YFYsG0EchJi/ToIl0URZTXGnE4vxonihtcy+2GBvk4h1h1DrWKCMCw6GAE+fc9HKC5zYK3dhThUrWx7wd7TUigD7KTwqFJViFzeDgC/PoONReuNuGD/RdRpbv++ONUgWhqtfQYMHqTEq/E84uyERLQ9zvEntaH3SGg+EoTTpQ0IP+izrVEcUq8Eg+MT4B2ZGSfIdBosqJW346EqGCPnpe+WG0O7D52BTaHgPvHdV9jt5MgiBBE0ePeIocg4POvK7DrmwpIJBJMzorG/CnDb3lI5O3kEJw/x2/P1+P0RR3MFveN22IjArDy37Vuz+VgeJ0RBBH5pTqoVYGIjRh8m1MKoojtB0px4FQVAv3kmDFGjXtz4hEe4oe/X9iBr6qPdfmaUF8lHkichcnqCVB0s6rS3YKhwkMMFTTYsT6oL3eiRswWOypqjSivNeJStRHlNS0wmq7vUhvkr8B94+Jx79h4BPp5NpZXEER8W1yPf359GQ1N19eDT4gKwvj0KKhVgVCrAhEZ6uc2PMpsseP4uTocLqhBZUObR/clk0owY7Qa8yYnIjyk+2UbL9W04K1Pi9Dc5pysrkmOwAMThqHdbEOb2YZWsw1tJuf/9S1mXKo2QrjpZUwmlSBJHYKEqCDERgQiJiIAseEBCAv2hUQiQaOxA7mHypBX3OD6msToYDx2fypS4p07Cps6bNC3dEDX3AF9ixn6lg4IgvPCVi6TQCaTQiGTQC6T4sK1vVcAIDLUD/+5ZHSfF2C91YfdIaDkahNOFDfg9A1BojuRoX64f1wCpmliXSudiaKIyoY2nLlkQGGZHuU1RogAAv3kmD91BGbnxH3n4Vx1jSa8veMsqnTOn7mPXIp7x8XjoUmJHtdcX6w2By7XGnGxqgWlVc24VN0Ci1XANE0sFs5I6vKO/o0ajR342+fncPGmzSmlEgmmZMXg4anDERV6Z4eldGoz21BypQnnKhqRf1Hn9rsKOINwYnQwisqdtRMd5o9fLNW6fjcG+nVGFEVs23cRh/OrEeArx2s/mjioen3sDgHv7i527UnUSSaVYHx6FKbnRGJrxUa0Wp21qvQJxpzhszE1dgIUd9EE9p4wVHiIoYIGO9YH9aU/akQURRhaOnCpxghRFKEdGenR8rvdcQgCvjlbh51HK6Bv6ehyXC5zTvRWqwIgl0lx6oIOFtv1d1t9FTJMzIjGmBQVmlo7UGMwoc7QjtpGExqNlpu+lwTTR6sxb5J7uPi6sBb/+0WJa6+SeZMTsXB6Uq/vxJs6bDhX0YTCS3oUlTf2ulyvr0KG6HB/1DWaXEsEB/krsGhmEmZo1N952JcgiPj4cBm+yKsEAPj7yvHcwixkDO95l+Lu6sPUYceRgmrsO1mJljb3xxHgK0dOaiTGpkWi1mDCgVOVbs9rgK8c00fHwmITcKZMj6ZW9+f8RpGhflg8Mxnj06NuaZhGXnE9tuwtgcXq6HIs0E+OhyYl4t6x8d3Oe7HZBVTp2lB1bXnmzt6HGz+aLHaUVbegorYVDqH7SxM/HxkenjIc94+L7zLUrKBMj/d2F6PN7Lx4T4lTQq0KxNGiWtf3k0okmJIdg/lThiPyNocLm92B0qoWnK9owvmKRlypa+0yrynQT46xaZGYMCoa6cPCIJEAnx+twD+/vgzA+bP5xVItVEr/AX+d2XeiEtu/LHV9rkmOwE8f0QyK5YM7rHa8teMszl12Lt2arA5BSKAPCkr1bs95YpIVvnEVmJCQeW2eyd0fJjoxVHiIoYIGO9YH9eVurRG7Q8DXhbU4lF+NGn17jxd3nRKignDPGDUmZcb0OI/CbLGjRt+Og6ercfx8nWsIl1zm7Ll4cOIw7DtRiQMnneOdfRRSLJ+XgfHpt7YaiyCKuFLXiqJLBlysakZdN4Gmk0QC3KONw8LpSR4NyfLE4YJqvP/FRQiiCJlUgiceSMOM0epuz72xPnRNZuw/WYnD+dVuc2f8feXISVVhfHo0MoaHufUu2B0CTl3Q4Yu8q6io6/41UyIBRsYpMTpFhejwAOw+dgWXa68PKUtSh+DRWSlITQjt9XHZ7AJyD5bhy9POn49MKsGSWSkYlRiGHV+Vo6BM7zo3LNgXP5w2AglRQaioa8WVOiMq6lpRreu7lm4mk0owLDoYqQlKGNttOHauznVMpfTDklkpGJcWCYcg4pPDl7DvhDPUSQA8NDkRP5w2AnKZFPpmM3Ydu+IWLmRSCUYlhiExJhjDY4KRGB2MCKXfLV8wOwQBRZca8a/CGpy93AhbN7/rQf4KZI0Ix4RR0chKCu+2l2jnNxXY8ZVz886IED/8139oEasKHLC/IQWlevzPPwohwvl72hn0l88bhanZsXfsfhuNHRBEESplz4Gv1WTFmx8XumpZkxyBZxdkwVchQ32TCQdOVOHrolq3Nz3CQ3wxXaPGdE1sj72kg9E3Z2tRVm3E4plJXXoCGSo8xFBBgx3rg/ryfagRhyCgocmMGr0JNYZ21OrbUaNvR4vJiqwR4bhHG4ek2JBbuhCrazRh59HLOH6+vtv5IREhfnhhcfYtLd/bmw6rHfWNZtQ2tqPOYEKNwQSFTIoHJiTctvu40bnLjXj7syLXmPkHJw7DdE0sfBUyKORS+NzwsdXiwPZ9JfimqM7tgjs1Xok5E4ZBkxzR5zAlURRRWtWCL/KuoqBUD39fObKSwjE6RYXspAi3wCSKIk6UNOCTw5fceqPGpKiQOSLcNf8lNMjH9TPVNZvx18/OuoJLWLAvnl2QhZQ4pevrL1Y245Mjl1B205Cjvkgkzp4DmVQCiVQChUyKYdFBSI0Pxch4JZLUSreet8u1Rnz0Zanb/YyMV8JmF1ztCwn0wY8fzkDmiK69RLpmM3Yfq8DRm57vToF+ciReCxgJ0UEYFhWMmB6WXK5vNOFfhbU4era2S6+SQi5FakIoMoaHIXN4OOKjgjyan7T32yv4+JBzX4KwYF/86omxSE+O7Pe/IVfqWrH+g9Ow2BwI8lfgv/5Di7c+LUJ9k/m2D4MSRRFX69tw+qIO+aV617C6+MhAjEuLwrhrQzA7GVo68Me/F6Cu0QQAmJwZjWUPjerye9LeYcNXBTU4cKrKrddOIgFGJ6swY4wamqSIQb2c9oGTlfjwgLOn6MfzMzA5M8btOEOFhxgqaLBjfVBfWCO9qzW0Y+c3Ffj2hnCRPiwUzy7IQrAHE50Hs2p9OzZ8fKbbYWSdFHKp27vaEgA5qZF4cOIwJN9wwX4rrDYHZDJJn0sD2+wCDp6uwq5vKrqdr+HrI0NMeABiwgNQdMngmrSenRSBHz08qtufjyiKOHPJgH8cuYTqGya+R4T4OXsCrvUIDIsORqC/HFKJ5DsNoxFFEScv6PDxobIuz2/m8DD86OEMKIN6v+BtaDbj0OkqXKox4mp9q2soXHcUciniIwOREBWEhKhgKORSfHO2Dhcrm93OCwn0weTMaGiSIpASr/R4FbCb7cu7iu0HywA4g8XqFZMR6i/vt78hTa0WrP3fk2hqtUAuk+AXS7UYGR+Ki5XN+N0HpyECGJ0cgRe9GAZldwi4UNmM/Is6FJTpe+xN7KRWBWJcWiSS1CHY+n8XXCFhzvgEPDo7pdfAZncIKCjV40hBNc7dsDIe4Hx+p2bHIlkdgjhVIMKVg2eD0GNn67B513kAzl65V58e36VHlaHCQwwVNNixPqgvrBHP1Bra8eWpKigDfTB3kme7ld8NjO1W/M+nhX2uYKWQSTElOwYPTBjm8Q7wt0ub2YY9x67g2+L6XudgSCTAohlJmDspsc+LLkEQUXK1CYIoIjE6+I4FRJvdgX0nKrH72BVYbQIWzhjhUfu6a29dowlX6ltxpe7av/pWt2FoPZFKJNAkR2C6JhbZHvQqeerLU1X4YP9F1+fZSRGYpY2DJvnOvrPeYbVj/QencbXe2Vtw87vjHx0oxf6TziFmP344A5OzYrr9Pt3pXP2qp8UHZFIJUhNCkZMaCYkEOHVBh5KrTT2udvfIPcmYO3HYLQWbhmYzviqowdeFNV0mzQPOOVdqVQDUqkDEqYIwMkGJZPWtBXxdsxlmix0qpR8CvuPCBQWlevzl0yIIoghloA9++XhOtyuXMVR4iKGCBjvWB/WFNUKCKKKithWmDhssNgFWuwNWmwNWmwC7ICI0xA/ZI8IQdJtWTfKG2WJHfZMJtQYT6gwm1DY6P/oqpFg8MxnpiWED3cRumS12WO1CrytC3SpBFKFv6UBlfRsqG1pR2dCGyoY2V89IVJg/pmtiMSUr9o6thnSkoBrv77voNkxLpfTDPdo4TNfEdglrgiCipd2KRmMHjCYrBMHZq+M+Gd45XykqzB9RoQFuywsLgoi/fFrkmh/zg6nDsWB6ktt9WGwOrHo3Dw3NZgT6OYdBhfbSKyQIIi5cbUJeSQNOXdC5JtB38vORITspAtqRKmiSI7pchBvbrThdqsOpkgYUX2l2bhQoAZ56ML3HuUqeuLH3ouRqc69zfR6YkIAl96T0GeYEQcTfD5a5QhfgnBPVuYGmSukPVagfRidH9Lq0ccmVJvwp9wzsDgEBvnL892M5iO9hTxqGCg8xVNBgx/qgvrBGqDesj7uPqcOOVrMVUaH+/bICktFkxfFiHfYeu+w2b0MukyInVQUAaGy1oMnYgeY26y1PhA/yVyA6zB9RYf7osDqQX+oMFBMzorFifka3j/HGYVBjUlR4YXF2l/OqdG3415lafHu+rkuPwI2LD4xKDINC7lnvTpvZhrPlBkSFBSBJHXJLj7M3NruA+kYTqvXtqL42Z6xa14aGZrOrl0Q7UoUV8zN7XFnPbLFj0+fnXMtK90YCYMxIFeaMT0BqQqjbc1dRZ8TvP8xHh9UBH4UUK/9d6zZ36WYMFR5iqKDBjvVBfWGNUG9YH9SXzhrR6VuRd74eB09Xd5nLcbulxCnxi6Vjep0T8uGBi66V2lbMz8CkzBiYLXbkFdfjX4W1KK9xH/Ln5yODdmQkxo+KQubwcI+DxEAytlvxl0+LUFbtXBQgMSYYLy7WdOmZMrR0YMMnZ1wbaKYmhOLhyYlobLVA39IBw7U9bvQtHWhutbgvdxsTjDnjEzA+PQq6ZjPWvX8abWYbZFIJfrpEg6wREb22kaHCQwwVNNixPqgvrBHqDeuD+tJdjVQ1tOFgfjXOX26Ev58c4cG+CA/xQ3iIL8KDnR+VgT6QSaWQSiWQSgCJVAKpxPnPbLGjodmMhiYTGprMaGgyo77JDF2zGdHh/vj5v43pc0d4i9WBVe9dHwalHRmJEyUNbku4yqQSaFMjMTnDuYzud524PpBsdgfe21Pi2lwvLNgXP31E41o1rrzGiD//o9C1N87U7Bg89WB6j3NrGo0d+PJ0FY7k17gWPwCA0CAfiABa2qyQSIBnf5iFcR4sp81Q4SGGChrsWB/UF9YI9Yb1QX0ZzDVy4WoTfvdhfpfb41SBmK6JxaSsmD7Dyd1AFEX88+vL+PxoBQDnymjP/jATHVYH3t1d7FrBbfHMJDw0KdGjYXEdVjuOFtVh/8lKNDSZ3Y49Pdfz+SKDMVR0v1MREREREVE30oaF4cEJw/B/eVfh6yPDxFFRmK5RI0l9a/vWDHYSiQQLpichKswfW/Y4d5Pf8Emha76Fj1yKHz2c4VHPQic/HznuHRuPWdo4nCnTY9+JSpTXGvHIzGSvJqAPBgwVRERERHRLlsxKxpSsGKhC/eDn8/2+nJySFYuIED/85dMi13K4ykAfvPiIBiNiv9vEcem1IWLa1EiIovi9CGPf7yogIiIiottOIpH0uNzp91HasDC88uQ4bNlTDIVChmVz0xEe4ndbvvf3IVAADBVERERERH2KDg/Afz8+dqCbMWgN/nW9iIiIiIhoUGOoICIiIiIirzBUEBERERGRVxgqiIiIiIjIKwwVRERERETkFYYKIiIiIiLyCkMFERERERF5haGCiIiIiIi8wlBBREREREReYaggIiIiIiKvMFQQEREREZFXGCqIiIiIiMgrDBVEREREROQVhgoiIiIiIvIKQwUREREREXmFoYKIiIiIiLzCUEFERERERF5hqCAiIiIiIq8wVBARERERkVcYKoiIiIiIyCsSURTFgW4EERERERHdvdhTQUREREREXmGoICIiIiIirzBUEBERERGRVxgqiIiIiIjIKwwVRERERETkFYYKIiIiIiLyCkMFERERERF5haGCiIiIiIi8wlBBREREREReYaggIiIiIiKvMFT04IMPPsDs2bORnZ2NJUuWoLCwcKCbRANg06ZNWLx4MbRaLSZPnoyf/OQnKC8vdzvHYrFg9erVmDhxIrRaLV544QXo9foBajENpL/97W9IS0vD66+/7rqN9TG01dfXY+XKlZg4cSI0Gg3mz5+PoqIi13FRFLFhwwZMmzYNGo0GTz/9NCoqKgauwdSvHA4H3nzzTcyePRsajQb33Xcf3nrrLYii6DqHNTJ0nDhxAs888wymTZuGtLQ0HDhwwO24J7XQ3NyMl156CTk5ORg3bhx+9atfob29vV/az1DRjT179mDdunV47rnnsGPHDqSnp2P58uUwGAwD3TTqZ3l5eXjssceQm5uLLVu2wG63Y/ny5TCZTK5z3njjDRw6dAhvvvkmtm3bhoaGBjz//PMD2GoaCIWFhdi+fTvS0tLcbmd9DF0tLS1YunQpFAoFNm/ejN27d+Pll1+GUql0nbN582Zs27YNv/3tb5Gbmwt/f38sX74cFotlAFtO/WXz5s346KOP8Oqrr2LPnj1YuXIl3nnnHWzbts3tHNbI0GAymZCWloZVq1Z1e9yTWli5ciXKysqwZcsWbNy4ESdPnsSrr77aPw9ApC4eeeQRcfXq1a7PHQ6HOG3aNHHTpk0D2CoaDAwGg5iamirm5eWJoiiKRqNRzMzMFPfu3es6p6ysTExNTRXz8/MHqJXU39ra2sQ5c+aIR48eFR9//HFx7dq1oiiyPoa6P/zhD+LSpUt7PC4Igjh16lTxnXfecd1mNBrFrKwscdeuXf3RRBpgK1asEH/5y1+63fb888+LL730kiiKrJGhLDU1Vdy/f7/rc09qofP1pbCw0HXOkSNHxLS0NLGuru6Ot5k9FTexWq04d+4cpkyZ4rpNKpViypQpyM/PH8CW0WDQ2toKAK53Gs+ePQubzeZWL8nJyVCr1SgoKBiIJtIAWLNmDWbOnOlWBwDrY6g7ePAgsrKy8OKLL2Ly5MlYsGABcnNzXcerqqqg0+nc6iM4OBijR4/m680QodVqcfz4cVy+fBkAUFJSglOnTmHGjBkAWCN0nSe1kJ+fj5CQEGRnZ7vOmTJlCqRSab8M45ff8Xu4yzQ1NcHhcCAiIsLt9oiIiC5j6WloEQQBb7zxBnJycpCamgoA0Ov1UCgUCAkJcTs3IiICOp1uIJpJ/Wz37t04f/48Pvnkky7HWB9DW2VlJT766CMsW7YMzzzzDIqKirB27VooFAosXLjQVQPdvd5w3s3QsGLFCrS1tWHu3LmQyWRwOBz42c9+hh/84AcAwBohF09qQa/XIzw83O24XC6HUqnsl9cchgoiD61evRqlpaX48MMPB7opNEjU1tbi9ddfx3vvvQdfX9+Bbg4NMqIoIisrCz//+c8BABkZGSgtLcX27duxcOHCAW4dDQZ79+7Fzp078cc//hEpKSkoLi7GunXrEBUVxRqhuw6HP90kLCwMMpmsy6Rsg8EAlUo1QK2igbZmzRocPnwYW7duRUxMjOt2lUoFm80Go9Hodr7BYEBkZGR/N5P62blz52AwGLBo0SJkZGQgIyMDeXl52LZtGzIyMlgfQ1xkZCSSk5PdbktKSkJNTY3rOAC+3gxhv//977FixQrMmzcPaWlpWLBgAZ566ils2rQJAGuErvOkFlQqFRobG92O2+12tLS09MtrDkPFTXx8fJCZmYljx465bhMEAceOHYNWqx3AltFAEEURa9aswf79+7F161YkJCS4Hc/KyoJCoXCrl/LyctTU1GDMmDH93Frqb5MmTcLOnTvx2Wefuf5lZWVh/vz5rv+zPoaunJwc11j5ThUVFYiLiwMAxMfHIzIy0q0+2tracObMGb7eDBEdHR2QSCRut8lkMteSsqwR6uRJLWi1WhiNRpw9e9Z1zvHjxyEIAjQazR1vI4c/dWPZsmV4+eWXkZWVBY1Gg61bt8JsNmPRokUD3TTqZ6tXr8auXbvw9ttvIzAw0DUmMTg4GH5+fggODsbixYuxfv16KJVKBAUFYe3atdBqtbxoHAKCgoJc82s6BQQEIDQ01HU762Poeuqpp7B06VJs3LgRc+fORWFhIXJzc7FmzRoAgEQiwZNPPom//vWvSExMRHx8PDZs2ICoqCjcd999A9x66g+zZs3Cxo0boVarXcOftmzZgsWLFwNgjQw17e3tuHr1quvzqqoqFBcXQ6lUQq1W91kLycnJmD59On7zm99g9erVsNlseO211zBv3jxER0ff8fZLRPGGHVbI5f3338e7774LnU6HUaNG4ZVXXsHo0aMHulnUz27ec6DTunXrXCHTYrFg/fr12L17N6xWK6ZNm4ZVq1ZxeMsQ9cQTTyA9PR2//vWvAbA+hrpDhw7hT3/6EyoqKhAfH49ly5bh0UcfdR0XRRF//vOfkZubC6PRiLFjx2LVqlUYMWLEALaa+ktbWxs2bNiAAwcOwGAwICoqCvPmzcNzzz0HHx8fAKyRoeTbb7/Fk08+2eX2hQsXYv369R7VQnNzM1577TUcPHgQUqkUc+bMwSuvvILAwMA73n6GCiIiIiIi8grnVBARERERkVcYKoiIiIiIyCsMFURERERE5BWGCiIiIiIi8gpDBREREREReYWhgoiIiIiIvMJQQUREREREXmGoICIiIiIirzBUEBERERGRVxgqiIiIiIjIKwwVRERERETkFYYKIiIiIiLyyv8DkYlM9XovGMcAAAAASUVORK5CYII="},"metadata":{}}]},{"cell_type":"markdown","source":"# 4) Evaluate Batch Normalization\n\nDid adding batch normalization help?","metadata":{}},{"cell_type":"code","source":"# View the solution (Run this cell to receive credit!)\nq_4.check()","metadata":{"lines_to_next_cell":0,"execution":{"iopub.status.busy":"2024-01-05T02:31:55.871857Z","iopub.execute_input":"2024-01-05T02:31:55.872164Z","iopub.status.idle":"2024-01-05T02:31:55.880211Z","shell.execute_reply.started":"2024-01-05T02:31:55.872138Z","shell.execute_reply":"2024-01-05T02:31:55.879082Z"},"trusted":true},"execution_count":12,"outputs":[{"output_type":"display_data","data":{"text/plain":"<IPython.core.display.Javascript object>","application/javascript":"parent.postMessage({\"jupyterEvent\": \"custom.exercise_interaction\", \"data\": {\"outcomeType\": 1, \"valueTowardsCompletion\": 0.25, \"interactionType\": 1, \"questionType\": 4, \"questionId\": \"4_Q4\", \"learnToolsVersion\": \"0.3.4\", \"failureMessage\": \"\", \"exceptionClass\": \"\", \"trace\": \"\"}}, \"*\")"},"metadata":{}},{"output_type":"display_data","data":{"text/plain":"Correct: \n\n\nYou can see that adding batch normalization was a big improvement on the first attempt! By adaptively scaling the data as it passes through the network, batch normalization can let you train models on difficult datasets.","text/markdown":"<span style=\"color:#33cc33\">Correct:</span> \n\n\nYou can see that adding batch normalization was a big improvement on the first attempt! By adaptively scaling the data as it passes through the network, batch normalization can let you train models on difficult datasets.\n"},"metadata":{}}]},{"cell_type":"markdown","source":"# Keep Going #\n\n[**Create neural networks**](https://www.kaggle.com/ryanholbrook/binary-classification) for binary classification.","metadata":{}},{"cell_type":"markdown","source":"---\n\n\n\n\n*Have questions or comments? Visit the [course discussion forum](https://www.kaggle.com/learn/intro-to-deep-learning/discussion) to chat with other learners.*","metadata":{}}]}