Transfer learning using EfficientNetB7.ipynb

{"nbformat":4,"nbformat_minor":0,"metadata":{"colab":{"name":"Transfer learning using EfficientNetB7.ipynb","version":"0.3.2","provenance":[{"file_id":"1-s0kfBhD6Ks47U6WBJLH_zKMW4v2I_vZ","timestamp":1567104630333},{"file_id":"1n0jt6RTVQqGugvDJ3uPxI0NPR3QL90CZ","timestamp":1567102653052},{"file_id":"1kJPlztxWuc_fd76jnlmfMCFYCOsNSn7w","timestamp":1565943125418}],"collapsed_sections":["a1gfATjCbqD0"]},"kernelspec":{"name":"python3","display_name":"Python 3"},"accelerator":"GPU"},"cells":[{"cell_type":"code","metadata":{"id":"8zCg1FVHF_xK","colab_type":"code","outputId":"2d5d3f57-bd55-430b-bb1b-d7c03308f968","executionInfo":{"status":"ok","timestamp":1567780673444,"user_tz":-60,"elapsed":26380,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":615}},"source":["# keras.backend.clear_session()\n","!pip install -U git+https://github.com/qubvel/efficientnet\n","import efficientnet.keras as efn \n","\n","from __future__ import absolute_import, division, print_function, unicode_literals\n","import datetime\n","\n","import keras\n","from keras import backend as K\n","from keras import optimizers\n","from keras.models import Sequential\n","from keras.layers import GlobalAveragePooling2D,GlobalMaxPooling2D,Conv2D,Input\n","from keras.layers.core import Activation, Dense, Flatten,Dropout\n","from keras.preprocessing.image import ImageDataGenerator\n","from keras.models import Model\n","from keras.callbacks import EarlyStopping, ModelCheckpoint, TensorBoard\n","from keras.models import load_model\n","\n","import numpy as np\n","import matplotlib.pyplot as plt\n","import matplotlib.image as mpimg\n","import pandas as pd\n","import os\n","import os.path\n","import shutil \n","from google.colab import drive\n","import tensorboardcolab as tbc\n","%load_ext tensorboard\n","drive.mount('/content/gdrive',force_remount=True)"],"execution_count":1,"outputs":[{"output_type":"stream","text":["Collecting git+https://github.com/qubvel/efficientnet\n","  Cloning https://github.com/qubvel/efficientnet to /tmp/pip-req-build-if_npziq\n","  Running command git clone -q https://github.com/qubvel/efficientnet /tmp/pip-req-build-if_npziq\n","Requirement already satisfied, skipping upgrade: keras_applications<=1.0.8,>=1.0.7 in /usr/local/lib/python3.6/dist-packages (from efficientnet==1.0.0b3) (1.0.8)\n","Requirement already satisfied, skipping upgrade: scikit-image in /usr/local/lib/python3.6/dist-packages (from efficientnet==1.0.0b3) (0.15.0)\n","Requirement already satisfied, skipping upgrade: h5py in /usr/local/lib/python3.6/dist-packages (from keras_applications<=1.0.8,>=1.0.7->efficientnet==1.0.0b3) (2.8.0)\n","Requirement already satisfied, skipping upgrade: numpy>=1.9.1 in /usr/local/lib/python3.6/dist-packages (from keras_applications<=1.0.8,>=1.0.7->efficientnet==1.0.0b3) (1.16.4)\n","Requirement already satisfied, skipping upgrade: matplotlib!=3.0.0,>=2.0.0 in /usr/local/lib/python3.6/dist-packages (from scikit-image->efficientnet==1.0.0b3) (3.0.3)\n","Requirement already satisfied, skipping upgrade: scipy>=0.17.0 in /usr/local/lib/python3.6/dist-packages (from scikit-image->efficientnet==1.0.0b3) (1.3.1)\n","Requirement already satisfied, skipping upgrade: pillow>=4.3.0 in /usr/local/lib/python3.6/dist-packages (from scikit-image->efficientnet==1.0.0b3) (4.3.0)\n","Requirement already satisfied, skipping upgrade: networkx>=2.0 in /usr/local/lib/python3.6/dist-packages (from scikit-image->efficientnet==1.0.0b3) (2.3)\n","Requirement already satisfied, skipping upgrade: imageio>=2.0.1 in /usr/local/lib/python3.6/dist-packages (from scikit-image->efficientnet==1.0.0b3) (2.4.1)\n","Requirement already satisfied, skipping upgrade: PyWavelets>=0.4.0 in /usr/local/lib/python3.6/dist-packages (from scikit-image->efficientnet==1.0.0b3) (1.0.3)\n","Requirement already satisfied, skipping upgrade: six in /usr/local/lib/python3.6/dist-packages (from h5py->keras_applications<=1.0.8,>=1.0.7->efficientnet==1.0.0b3) (1.12.0)\n","Requirement already satisfied, skipping upgrade: pyparsing!=2.0.4,!=2.1.2,!=2.1.6,>=2.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib!=3.0.0,>=2.0.0->scikit-image->efficientnet==1.0.0b3) (2.4.2)\n","Requirement already satisfied, skipping upgrade: python-dateutil>=2.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib!=3.0.0,>=2.0.0->scikit-image->efficientnet==1.0.0b3) (2.5.3)\n","Requirement already satisfied, skipping upgrade: cycler>=0.10 in /usr/local/lib/python3.6/dist-packages (from matplotlib!=3.0.0,>=2.0.0->scikit-image->efficientnet==1.0.0b3) (0.10.0)\n","Requirement already satisfied, skipping upgrade: kiwisolver>=1.0.1 in /usr/local/lib/python3.6/dist-packages (from matplotlib!=3.0.0,>=2.0.0->scikit-image->efficientnet==1.0.0b3) (1.1.0)\n","Requirement already satisfied, skipping upgrade: olefile in /usr/local/lib/python3.6/dist-packages (from pillow>=4.3.0->scikit-image->efficientnet==1.0.0b3) (0.46)\n","Requirement already satisfied, skipping upgrade: decorator>=4.3.0 in /usr/local/lib/python3.6/dist-packages (from networkx>=2.0->scikit-image->efficientnet==1.0.0b3) (4.4.0)\n","Requirement already satisfied, skipping upgrade: setuptools in /usr/local/lib/python3.6/dist-packages (from kiwisolver>=1.0.1->matplotlib!=3.0.0,>=2.0.0->scikit-image->efficientnet==1.0.0b3) (41.2.0)\n","Building wheels for collected packages: efficientnet\n","  Building wheel for efficientnet (setup.py) ... \u001b[?25l\u001b[?25hdone\n","  Created wheel for efficientnet: filename=efficientnet-1.0.0b3-cp36-none-any.whl size=17630 sha256=1a5d56fc7abb6906a6f761f3a37c478da75e7dcd96ccce0df7f3e95a2cc69705\n","  Stored in directory: /tmp/pip-ephem-wheel-cache-xk1zptlw/wheels/64/60/2e/30ebaa76ed1626e86bfb0cc0579b737fdb7d9ff8cb9522663a\n","Successfully built efficientnet\n","Installing collected packages: efficientnet\n","Successfully installed efficientnet-1.0.0b3\n"],"name":"stdout"},{"output_type":"stream","text":["Using TensorFlow backend.\n"],"name":"stderr"},{"output_type":"stream","text":["Go to this URL in a browser: https://accounts.google.com/o/oauth2/auth?client_id=947318989803-6bn6qk8qdgf4n4g3pfee6491hc0brc4i.apps.googleusercontent.com&redirect_uri=urn%3Aietf%3Awg%3Aoauth%3A2.0%3Aoob&scope=email%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdocs.test%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fdrive.photos.readonly%20https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fpeopleapi.readonly&response_type=code\n","\n","Enter your authorization code:\n","··········\n","Mounted at /content/gdrive\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"5kSrmJnUGryx","colab_type":"code","outputId":"294835cc-c39f-48c7-ecb3-f76283b7027a","executionInfo":{"status":"ok","timestamp":1567780746295,"user_tz":-60,"elapsed":8133,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":122}},"source":["train_dir ='/content/gdrive/My Drive/MScthesis/train/'\n","validation_dir = '/content/gdrive/My Drive/MScthesis/validation/'\n","\n","# Directory with our training residential pictures\n","train_residential_dir = os.path.join(train_dir, 'residential')\n","print ('Total training residential images:', len(os.listdir(train_residential_dir)))\n","\n","# Directory with our training non-residential pictures\n","train_nonresidential_dir = os.path.join(train_dir, 'nonresidential')\n","print ('Total training non-residential images:', len(os.listdir(train_nonresidential_dir)))\n","\n","# Directory with our validation residential pictures\n","validation_residential_dir = os.path.join(validation_dir, 'residential')\n","print ('Total validation residential images:', len(os.listdir(validation_residential_dir)))\n","\n","# Directory with our validation non-residential pictures\n","validation_nonresidential_dir = os.path.join(validation_dir, 'nonresidential')\n","print ('Total validation non-residential images:', len(os.listdir(validation_nonresidential_dir)))\n","\n","'''\n","Originally, there are around \n","5900 train residential images\n","770 train non-residential images  \n","545 validation residential images\n","120 validation non-residential images\n","So random guess on validation set accuracy should be 81.9% \n","Total test non-residential images: 180\n","Total test residential images: 504\n","So random guess on test set accuracy should be 73.6%\n","'''"],"execution_count":3,"outputs":[{"output_type":"stream","text":["Total training residential images: 7357\n","Total training non-residential images: 3803\n","Total validation residential images: 681\n","Total validation non-residential images: 551\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["'\\nOriginally, there are around \\n5900 train residential images\\n770 train non-residential images  \\n545 validation residential images\\n120 validation non-residential images\\nSo random guess on validation set accuracy should be 81.9% \\nTotal test non-residential images: 180\\nTotal test residential images: 504\\nSo random guess on test set accuracy should be 73.6%\\n'"]},"metadata":{"tags":[]},"execution_count":3}]},{"cell_type":"code","metadata":{"id":"9EBCSCmtGsaj","colab_type":"code","outputId":"a0b0d336-a0a1-4127-8db1-7a065be47f50","executionInfo":{"status":"ok","timestamp":1567780760965,"user_tz":-60,"elapsed":1280,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":51}},"source":["batch_size = 64\n","image_size = 160\n","# Rescale all images by 1./255 and apply image augmentation\n","train_datagen = ImageDataGenerator(rescale=1./255,\n","                  # rotation_range=3,\n","                  # shear_range=0.2,\n","                  # zoom_range=0.2,\n","                  # fill_mode='nearest',\n","                  # width_shift_range=0.1,\n","                  # horizontal_flip=True,\n","                  # height_shift_range=0.1,\n","                                  )\n","\n","validation_datagen = ImageDataGenerator(rescale=1./255,\n","                    # rotation_range=3,\n","                    # shear_range=0.2,\n","                    # zoom_range=0.2,\n","                    # fill_mode='nearest',\n","                    # width_shift_range=0.1,\n","                    # horizontal_flip=True,\n","                    # height_shift_range=0.1,\n","                                       )\n","\n","# Flow training images in batches of 20 using train_datagen generator\n","train_generator = train_datagen.flow_from_directory(train_dir,  # Source directory for the training images\n","                                                    target_size=(image_size, image_size),\n","                                                    batch_size=batch_size,\n","                                                    class_mode='binary')  # Since we use binary_crossentropy loss, we need binary labels\n","\n","# Flow validation images in batches of 20 using test_datagen generator\n","validation_generator = validation_datagen.flow_from_directory(validation_dir, # Source directory for the validation images\n","                                                              target_size=(image_size, image_size),\n","                                                              batch_size=batch_size,\n","                                                              shuffle=False,\n","                                                              class_mode='binary')"],"execution_count":4,"outputs":[{"output_type":"stream","text":["Found 11160 images belonging to 2 classes.\n","Found 1232 images belonging to 2 classes.\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"8llm0GXiGuR7","colab_type":"code","outputId":"76c1ce55-af39-4f5e-8f38-a5a549d65b5b","executionInfo":{"status":"ok","timestamp":1567780804372,"user_tz":-60,"elapsed":41084,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":428}},"source":["IMG_SHAPE = (image_size, image_size, 3)\n","# Create the base model from the pre-trained model EfficientNet-B7\n","base_model = efn.EfficientNetB7(input_shape=IMG_SHAPE,\n","                                include_top=False,\n","                                weights='imagenet')"],"execution_count":5,"outputs":[{"output_type":"stream","text":["WARNING: Logging before flag parsing goes to stderr.\n","W0906 14:39:23.511636 140004315789184 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:66: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n","\n","W0906 14:39:23.559433 140004315789184 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:541: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n","\n","W0906 14:39:23.568977 140004315789184 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:4479: The name tf.truncated_normal is deprecated. Please use tf.random.truncated_normal instead.\n","\n","W0906 14:39:23.603376 140004315789184 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n","\n","W0906 14:39:23.605497 140004315789184 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:197: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n","\n","W0906 14:39:27.081121 140004315789184 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:2041: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.\n","\n","W0906 14:39:27.515170 140004315789184 deprecation.py:506] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3733: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n","Instructions for updating:\n","Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n","W0906 14:39:39.083956 140004315789184 nn_ops.py:4224] Large dropout rate: 0.5125 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n","W0906 14:39:39.694496 140004315789184 nn_ops.py:4224] Large dropout rate: 0.525 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n","W0906 14:39:39.978321 140004315789184 nn_ops.py:4224] Large dropout rate: 0.5375 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n","W0906 14:39:40.261285 140004315789184 nn_ops.py:4224] Large dropout rate: 0.55 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n","W0906 14:39:40.551572 140004315789184 nn_ops.py:4224] Large dropout rate: 0.5625 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n"],"name":"stderr"},{"output_type":"stream","text":["Downloading data from https://github.com/Callidior/keras-applications/releases/download/efficientnet/efficientnet-b7_weights_tf_dim_ordering_tf_kernels_autoaugment_notop.h5\n","258441216/258434480 [==============================] - 8s 0us/step\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"C-G1SJ7LGxtq","colab_type":"code","colab":{}},"source":["# base_model.trainable = False\n","# base_model.summary()"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"eVdr-ZUDVcaO","colab_type":"code","colab":{}},"source":["import tensorflow as tf\n","#recall\n","# def recall_m(y_true, y_pred):\n","#         true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))\n","#         possible_positives = K.sum(K.round(K.clip(y_true, 0, 1)))\n","#         recall = true_positives / (possible_positives + K.epsilon())\n","#         return recall\n","\n","# #precision\n","# def precision_m(y_true, y_pred):\n","#         true_positives = K.sum(K.round(K.clip(y_true * y_pred, 0, 1)))\n","#         predicted_positives = K.sum(K.round(K.clip(y_pred, 0, 1)))\n","#         precision = true_positives / (predicted_positives + K.epsilon())\n","#         return precision\n","\n","# #f1_score\n","# def f1_m(y_true, y_pred):\n","#     precision = precision_m(y_true, y_pred)\n","#     recall = recall_m(y_true, y_pred)\n","#     return 2*((precision*recall)/(precision+recall+K.epsilon()))\n","\n","#focal loss\n","def focal_loss(gamma=2., alpha=.25):\n","\tdef focal_loss_fixed(y_true, y_pred):\n","\t\tpt_1 = tf.where(tf.equal(y_true, 1), y_pred, tf.ones_like(y_pred))\n","\t\tpt_0 = tf.where(tf.equal(y_true, 0), y_pred, tf.zeros_like(y_pred))\n","\t\treturn -K.mean(alpha * K.pow(1. - pt_1, gamma) * K.log(pt_1)) - K.mean((1 - alpha) * K.pow(pt_0, gamma) * K.log(1. - pt_0))\n","\treturn focal_loss_fixed\n","\n"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"isJDbW0tUVAU","colab_type":"code","colab":{}},"source":["# EARLY_STOP_PATIENCE must be < NUM_EPOCHS\n","EARLY_STOP_PATIENCE = 5\n","# These steps value should be proper FACTOR of no.-of-images in train & valid folders respectively\n","# Early stopping & checkpointing the best model in ../working dir & restoring that as our model for prediction\n","\n","cb_early_stopper = EarlyStopping(monitor = 'val_loss', patience = EARLY_STOP_PATIENCE)"],"execution_count":0,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"a1gfATjCbqD0","colab_type":"text"},"source":["## LR finder"]},{"cell_type":"code","metadata":{"id":"l8_kVXybtptV","colab_type":"code","colab":{}},"source":["from keras.callbacks import Callback\n","class LR_Finder(Callback):\n","    \n","    def __init__(self, start_lr=1e-5, end_lr=10, step_size=None, beta=.98):\n","        super().__init__()\n","        \n","        self.start_lr = start_lr\n","        self.end_lr = end_lr\n","        self.step_size = step_size\n","        self.beta = beta\n","        self.lr_mult = (end_lr/start_lr)**(1/step_size)\n","        \n","    def on_train_begin(self, logs=None):\n","        self.best_loss = 1e9\n","        self.avg_loss = 0\n","        self.losses, self.smoothed_losses, self.lrs, self.iterations = [], [], [], []\n","        self.iteration = 0\n","        logs = logs or {}\n","        K.set_value(self.model.optimizer.lr, self.start_lr)\n","        \n","    def on_batch_end(self, epoch, logs=None):\n","        logs = logs or {}\n","        loss = logs.get('loss')\n","        self.iteration += 1\n","        \n","        self.avg_loss = self.beta * self.avg_loss + (1 - self.beta) * loss\n","        smoothed_loss = self.avg_loss / (1 - self.beta**self.iteration)\n","        \n","        # Check if the loss is not exploding\n","        if self.iteration>1 and smoothed_loss > self.best_loss * 4:\n","            self.model.stop_training = True\n","            return\n","\n","        if smoothed_loss < self.best_loss or self.iteration==1:\n","            self.best_loss = smoothed_loss\n","        \n","        lr = self.start_lr * (self.lr_mult**self.iteration)\n","        \n","        self.losses.append(loss)\n","        self.smoothed_losses.append(smoothed_loss)\n","        self.lrs.append(lr)\n","        self.iterations.append(self.iteration)\n","        \n","        \n","        K.set_value(self.model.optimizer.lr, lr)  \n","        \n","    def plot_lr(self):\n","        plt.xlabel('Iterations')\n","        plt.ylabel('Learning rate')\n","        plt.plot(self.iterations, self.lrs)\n","        \n","    def plot(self, n_skip=10):\n","        plt.ylabel('Loss')\n","        plt.xlabel('Learning rate (log scale)')\n","        plt.plot(self.lrs[n_skip:-5], self.losses[n_skip:-5])\n","        plt.xscale('log')\n","        \n","    def plot_smoothed_loss(self, n_skip=10):\n","        plt.ylabel('Smoothed Losses')\n","        plt.xlabel('Learning rate (log scale)')\n","        plt.plot(self.lrs[n_skip:-5], self.smoothed_losses[n_skip:-5])\n","        plt.xscale('log')\n","        \n","    def plot_loss(self):\n","        plt.ylabel('Losses')\n","        plt.xlabel('Iterations')\n","        plt.plot(self.iterations[10:], self.losses[10:])"],"execution_count":0,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"tLuTfuGyb0Bj","colab_type":"text"},"source":["## add classification head"]},{"cell_type":"code","metadata":{"id":"B29tVoEdGzbq","colab_type":"code","colab":{}},"source":["dropout_rate = 0.5\n","model = Sequential([\n","  base_model,\n","  GlobalAveragePooling2D(),\n","  Dense(256,activation='relu',name='fc1'),\n","  Dropout(rate = dropout_rate, name=\"dropout_out\"),\n","  Dense(1, activation='sigmoid',name='predictions')\n","])"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"KZUvj0CV_8GY","colab_type":"code","colab":{"base_uri":"https://localhost:8080/","height":644},"outputId":"4e586f78-60c6-4e6d-a1be-39f2e270b62b","executionInfo":{"status":"ok","timestamp":1567780822891,"user_tz":-60,"elapsed":1624,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}}},"source":["from keras.utils import plot_model\n","plot_model(model, show_shapes=True,to_file='/content/models.png')"],"execution_count":7,"outputs":[{"output_type":"execute_result","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAskAAAJzCAYAAAD9ZI7+AAAABmJLR0QA/wD/AP+gvaeTAAAgAElE\nQVR4nOzdaVhUV7Y//u+BKqiBYlIUGoRAgYqzRtOC2sbQIdfYoIgEoqaD6fY6xCBibJwwqKhROsBV\nIWkzkEEliPKoiWLbdtoYb9RfEjQYokYxyKSgKIIUQwHr/yKX+lsUIAUFxbA+z8OL7LOHdQbN8rDP\n3gIRERhjjDHGGGMaJsYOgDHGGGOMse6Gk2TGGGOMMcaa4CSZMcYYY4yxJjhJZowxxhhjrAmRsQPo\nK86dO4e4uDhjh8EYY4yxHszLywsRERHGDqNP4DfJXSQ/Px8HDx40dhiMsV7k4MGDKCgoMHYYPcr5\n8+dx/vx5Y4fBWLucP38e586dM3YYfQa/Se5iaWlpxg6BMdZLCIKAFStW4KWXXjJ2KD1GUFAQAP67\nmPVMjc8v6xr8JpkxxhhjjLEmOElmjDHGGGOsCU6SGWOMMcYYa4KTZMYYY4wxxprgJJkxxhhjjLEm\nOElmjLE+7vjx47CyssIXX3xh7FC6vVOnTmHNmjU4dOgQ3NzcIAgCBEHAK6+8olPX19cXCoUCpqam\nGD58ODIzM40Qsf4aGhoQHx8Pb2/vFuuo1Wps3boV7u7uMDMzg7W1NUaMGIHc3FytemfPnsWkSZMg\nk8ng4OCAyMhI1NTU9Mr4tm/fjqFDh0IqlUIul2Po0KGIiopCeXm5ps7Ro0exfft21NfXt2sM1rU4\nSWaMsT6OiIwdQo/w1ltvYefOnVi7di0CAwNx8+ZNKJVK9OvXD3v37sWxY8e06p88eRJpaWnw8/ND\ndnY2xo0bZ6TI2+769ev4wx/+gIiICKhUqhbrBQcH49NPP8W+ffugUqlw5coVKJVKPHr0SFMnOzsb\nvr6+8PHxwd27d5Geno6PPvoIS5Ys6ZXxffPNN1i4cCHy8vJQXFyMzZs3Y/v27ZgzZ46mjr+/PyQS\nCXx8fFBWVtaucVgXItYlUlNTiS83Y8yQAFBqaqqxwzAolUpFXl5endb/nDlzaM6cOXq327ZtGw0e\nPJiqqqq0ypVKJe3bt49MTEzI0dGRysrKtI5nZGTQzJkzOxRzV7l06RLNnj2b9u7dS2PGjKHRo0c3\nWy8lJYUEQaCsrKxW+wsODiZXV1dqaGjQlMXGxpIgCHTlypVeF19AQIDO8xEUFEQAqKioSKs8LCyM\nvLy8SK1W6zVGe59f1j78Jpkxxli38eGHH6KkpMTYYWi5ceMGoqKisHHjRkgkEp3j3t7eCA8PR2Fh\nId58800jRGgYo0ePxqFDhzBv3jyYm5u3WO/dd9/FuHHjMHLkyBbr1NXV4dixY5g6dSoEQdCUT58+\nHUSEI0eO9Lr40tPTdZ4PR0dHANB6gw0A0dHRuHTpEhISEvQeh3UdTpIZY6wPO3v2LJydnSEIAnbv\n3g0ASEpKglwuh0wmw5EjRzB9+nRYWlrCyckJKSkpmrY7d+6ERCLBgAEDsHjxYjg4OEAikcDb2xsX\nLlzQ1AsLC4OZmRns7e01Za+//jrkcjkEQcC9e/cAAOHh4Vi5ciVycnIgCALc3d0BACdOnIClpSW2\nbNnSFZdEx86dO0FE8Pf3b7FOTEwMBg8ejA8++ACnTp1qtT8iQlxcHDw9PWFubg4bGxvMmjULV69e\n1dRp6z0AgPr6emzYsAHOzs6QSqUYNWoUUlNTO3bSLaitrcX58+cxZsyYVuvdvHkTjx49grOzs1a5\nUqkEAGRlZfWJ+K5fvw5ra2u4uLholdvY2GDq1KlISEjg6U7dGCfJjDHWh02ePBnffvutVtnSpUux\nYsUKVFVVQaFQIDU1FTk5OXBzc8PChQuhVqsB/Jb8hoaGQqVSYfny5cjNzUVmZibq6urw/PPPIz8/\nH8BvSWbTrbMTExOxceNGrbKEhAT4+flBqVSCiHDjxg0A0Hzk1NDQ0CnX4EmOHTuGIUOGQCaTtVhH\nKpXi448/homJCRYuXIjKysoW60ZHR2PNmjVYt24dSkpKcObMGeTn52PKlCkoLi4G0PZ7AACrV6/G\njh07EB8fj9u3b8PPzw9z587F999/b7iL8H+KiopQW1uLH374AdOmTdP8w8jT0xOJiYmahO/OnTsA\nAIVCodVeIpFAKpVqzrM3xqdWq1FYWIjdu3fj1KlT2LVrF8zMzHTqjR07FoWFhfjxxx/bPRbrXJwk\nM8YYa5G3tzcsLS1hZ2eHkJAQVFZWIi8vT6uOSCTSvBUdNmwYkpKSUFFRgeTkZIPEMGPGDJSXlyMq\nKsog/emjsrISv/76q+YNY2u8vLywYsUK5ObmYvXq1c3WqaqqQlxcHGbPno358+fDysoKI0eOxHvv\nvYd79+5hz549Om1auwfV1dVISkpCQEAAAgMDYW1tjfXr10MsFhvs+j+ucdqAnZ0dtmzZguzsbBQX\nF2PWrFlYtmwZ9u/fDwCaFSJMTU11+hCLxaiqqjJ4bN0lvkGDBsHJyQnR0dHYsWMHgoODm63n4eEB\nALh8+XK7x2Kdi5NkxhhjbdL4Nuzxt5jNGT9+PGQymdb0gZ6qpKQERNTqW+THxcTEYMiQIUhMTMTZ\ns2d1jmdnZ+PRo0cYP368VvmECRNgZmamNU2lOU3vwbVr16BSqTBixAhNHalUCnt7+065/o1zgYcP\nHw5vb2/Y2trCysoKGzduhJWVlSbJb5ybW1dXp9NHbW0tpFKpwWPrLvHl5+ejpKQE+/fvxyeffIKx\nY8c2O8++8ZnqrLfqrOM4SWaMMWZw5ubmuHv3rrHD6LDq6moAaPVDscdJJBIkJydDEAS89tprOm8k\nG5f9srCw0GlrbW2NiooKveJrnNaxfv16zZrNgiDg1q1brS6R1l4ODg4AoJlH3sjMzAwuLi7IyckB\nAM3888fXCAYAlUqF6upqTT+9MT6xWAw7Ozv4+vri888/R3Z2NrZu3apTrzERb3zGWPfDSTJjjDGD\nUqvVKCsrg5OTk7FD6bDGREafzR+8vLwQERGB69evY/PmzVrHrK2tAaDZZLg918zOzg4AEB8fDyLS\n+jl37pxefbWFhYUFPDw88PPPP+scq6urg5WVFQDA1dUVCoUCt27d0qrTOM981KhRBo+tO8bn7u4O\nU1NTZGdn6xyrra0FgE57q846jpNkxhhjBnX69GkQESZOnKgpE4lET5ym0R0NGDAAgiDg4cOHerXb\nvHkzhg4diosXL2qVjxgxAhYWFjof1V24cAG1tbV4+umn9Rpn0KBBkEgkuHTpkl7tOiI4OBgXL17E\nzZs3NWUqlQq3bt3SLLsmEonw4osv4syZM1ofXGZkZEAQhFZXCumJ8ZWWlmLu3Lk65devX0d9fT0G\nDRqkc6zxmRo4cKBeY7Guw0kyY4yxDmloaMCDBw9QV1eHrKwshIeHw9nZGaGhoZo67u7uuH//Pg4f\nPgy1Wo27d+/qvMUDAFtbWxQVFSE3NxcVFRVQq9XIyMgw2hJwMpkMbm5uKCgo0Ktd47SLph+GSSQS\nrFy5Eunp6di7dy/Ky8tx+fJlLFmyBA4ODli0aJHe4yxYsAApKSlISkpCeXk56uvrUVBQgNu3bwMA\nQkJCMHDgQINtix0REQEXFxeEhoYiLy8PpaWliIyMRFVVldYHi1FRUSguLsZbb72FyspKnDt3DrGx\nsQgNDcWQIUM09XpDfHK5HCdPnsRXX32F8vJyqNVqXLx4Ea+++irkcjkiIiJ02jQ+U62t58yMzBg7\nmPRFvOMeY8zQYIAd93bt2kX29vYEgGQyGfn7+1NiYiLJZDICQB4eHpSTk0N79uwhS0tLAkAuLi70\nyy+/EBHRokWLSCwWk6OjI4lEIrK0tKRZs2ZRTk6O1jilpaU0bdo0kkgk5OrqSm+88QatWrWKAJC7\nuzvl5eUREVFmZia5uLiQVCqlyZMn0507d+j48eOkUCgoJiamQ+dK1L4dy8LCwkgsFpNKpdKUpaen\nk1KpJADUv39/WrZsWbNtV61apbPjXkNDA8XGxpKHhweJxWKysbGhgIAAunbtmqaOPvegpqaGIiMj\nydnZmUQiEdnZ2VFgYCBlZ2cT0W87wQGgDRs2tHqe586do0mTJpGDgwMBIABkb29P3t7e9PXXX2vV\nzc/Pp5dffplsbGzI3NycnnnmGcrIyNDp8+uvv6ZnnnmGzM3NycHBgVatWkXV1dVadXpLfP7+/uTq\n6koWFhZkbm5OSqWSQkJC6PLly83WnzFjBjk6Omrt+PckvONe1xKIeBXrrnDgwAEEBwfzouGMMYMR\nBAGpqak6axB3pcWLFyMtLQ2lpaVGi0EfQUFBAIC0tLQ2t7lx4wY8PT2RnJyM+fPnd1ZonaahoQHP\nPvssQkND8dprrxk7HB19Mb7S0lI4OTkhJiYGK1eubHO79jy/rP14ugVjjLEO0eejtp7I3d0dmzZt\nwqZNm3S2F+7u6uvrcfjwYVRUVCAkJMTY4ejoq/FFR0djzJgxCAsLM1ifzPA4SWaMMcaeYM2aNQgK\nCkJISIjeH/EZ0+nTp3Ho0CFkZGS0ea3nrtQX44uLi8OlS5dw/PhxiMVig/TJOgcnyT1cTU0Nli9f\nDnt7e8hkMpw4caLF8uPHj8PKygpffPFFm/tvT5ueZMKECTA1NcWYMWOeWPfZZ5/VWof08Z/m1jx9\nkt5wbc+fPw9PT0+YmJhAEAQMHDgQMTExxg5Ly6FDh+Dm5qa5V/b29j3yV+bd0dq1a5GcnIyHDx/C\n1dUVBw8eNHZInWrLli0ICwvDtm3bjB1Km/n4+GDfvn2adYG7m74W35EjR1BTU4PTp0/DxsbGIH2y\nziMydgCsY9555x2cOHECV69exYEDBzS/CmyuvD1rMfb2OdTfffcd/vjHP+osPK+vyZMn692mN1zb\niRMn4sqVK/iv//ov/POf/8S1a9c068B2F4GBgQgMDIS7uzvu3buHO3fuGDukXmPr1q3NbpLQm/n6\n+sLX19fYYbAeaubMmZg5c6axw2BtxElyD3f48GGMHz8e1tbW+O///u8nluv7a8IZM2Z02q8Wq6qq\n4OPjg2+//dbofQuC8MQ6EokE5eXlUCgUWuWLFy9u14dTnXlt9dWZ96Kr9aZzYYwxZjw83aKHKygo\naHZOU0vl3cmHH37Y7H72xui7LdfqxIkTOglyfn4+fvrpJzz33HN6x9iddOa96Gq96VwYY4wZDyfJ\n3Vh9fT02bNgAZ2dnSKVSjBo1CqmpqQCAf/3rX3B3d8ft27fxySefaObFtlR+9uxZODs7QxAE7N69\nW2uczz77DOPHj4dEIoFcLsdTTz2FzZs3t9imtbiSkpIgl8shk8lw5MgRTJ8+HZaWlnByckJKSoqm\nj/DwcKxcuRI5OTkQBAHu7u5tbvukGJrr+0lu3LiBoUOHQi6XQyqVYsqUKTh79uwT27399ttYvnz5\nE+s11dy1bev579y5ExKJBAMGDMDixYvh4OAAiUQCb29vXLhwQVMvLCwMZmZmWnPpXn/9dcjlcgiC\noJli0tL1OnHiRLs3cOhu56Kvb775BsOGDYOVlRUkEglGjhyJf/7znwCAv/71r5r5zUqlUrOj2oIF\nCyCTyWBlZYWjR48CaP053bFjB2QyGRQKBUpKSrBy5Uo4Ojri2rVr7YqZMcaYgRl1leY+pD2bibz5\n5ptkbm5OBw8epAcPHtDatWvJxMSEvvvuO02dgQMH0quvvqrTtrny/Px8AkC7du3SlMXHxxMA2rZt\nG5WWltL9+/fpH//4B82bN6/FNk+Ka926dQSA/v3vf9PDhw+ppKSEpkyZQnK5nGprazX9BAYGklKp\n1IqxrW2fFENzfbfEx8eH3Nzc6NdffyW1Wk0//fQT/f73vyeJRKJZrL85BQUFNGzYMKqvr2/TOE01\nd23bev6LFi0iuVxOP//8M1VXV1N2djZNmDCBFAqFZlMGIqJ58+bRwIEDtcaNjY0lAHT37l1NWXPX\n68svvySFQkGbNm164rm88MILBIAePHjQLc+FiEipVJKVldUTz4WIKC0tjaKjo+n+/ftUWlpKEydO\npH79+mmNYWpqSoWFhVrt5s6dS0ePHtX8d1v/rCxfvpx27dpFs2fPpitXrrQpRiLDbCbS1/BmDKwn\n4+e3a/Gb5G6quroaSUlJCAgIQGBgIKytrbF+/XqIxWIkJycbZAy1Wo2NGzdi2rRpWL16NWxtbWFj\nY4O//OUvmDBhQofj8vb2hqWlJezs7BASEoLKykrk5eW1KbbW2nbGtVEoFHjqqacgEokwfPhwvP/+\n+6iursaePXtabPP222/jjTfegImJ4f8YteXaiUQieHp6wtzcHMOGDUNSUhIqKioM9nzMmDED5eXl\niIqK6lA/3eFc9DVnzhy89dZbsLGxga2tLfz9/VFaWoq7d+8CAJYsWYL6+nqt+MrLy/Hdd9/hxRdf\nBKDfc/r2229j2bJlOHToEIYOHdp1J8oYY6xF/OFeN3Xt2jWoVCqMGDFCUyaVSmFvb4+rV68aZIys\nrCyUlZXhhRde0Co3NTVtcQpBe+MyMzMD8Ftirq+mbbvi2owcORJWVlbIyspq9nhRURGOHj2K2NhY\ng4zXmrZeu/Hjx0MmkxnsGnSGnnoujXPWGzfNeO655zB48GB89NFHWLt2LQRBwOeff46QkBCYmpoC\n6JrnFACCg4MRHBxssP76irZ8rMtYdzRnzhxjh9BncJLcTVVWVgIA1q9fj/Xr12sdc3BwMMgY5eXl\nAKDXkl1dEVdnxPDll1/Cz89Pq2zevHnYu3dvi+OIxeIWk7nt27dj4cKFkEgk+oTe6czNzTVvO3s6\nY57LsWPHEBsbi+zsbJSXl+s8B4IgYPHixYiIiMC///1v/PGPf8Snn36Kffv2aep01Z+V8PBweHl5\nGay/3i4+Ph4AsGLFCiNHwpj+Gp9f1jU4Se6m7OzsAPz2ByI8PLxTxvjd734HAHqtEdwVcXVGDH/6\n05/0Wpe4rq4O9+/fh7Ozs86xO3fuYP/+/d3uAyu1Wo2ysjI4OTkZO5QO6+pzOXPmDH744QesWLEC\neXl5CAgIwOzZs/HRRx/hd7/7HXbt2oW//e1vWm1CQ0Oxdu1afPDBBxg0aBAsLS3h4uKiOd5Vf1a8\nvLzatQRhX5WWlgYAfM1Yj9T4/LKuwXOSu6lBgwZBIpHg0qVLnTbGU089BVtbW5w8ebJbxdUdYvjP\nf/6DhoYGjBs3TufY9u3bMX/+fNja2nba+O1x+vRpEBEmTpyoKROJRO2a4mJsXX0uP/zwA+RyOQDg\n8uXLUKvVWLp0Kdzc3CCRSJr91byNjQ2Cg4Nx+PBh/P3vf8fChQu1jneHPyuMMcbaj5PkbkoikWDB\nggVISUlBUlISysvLUV9fj4KCAty+fdsgY5ibm2Pt2rU4c+YMwsLCUFhYiIaGBlRUVODnn3/u9Lhs\nbW1RVFSE3NxcVFRUtDkBaksM+vZdW1uLhw8foq6uDpmZmQgLC4OLiwtCQ0O16hUXF+Ojjz7qFr+q\nbWhowIMHD1BXV4esrCyEh4fD2dlZK2Z3d3fcv38fhw8fhlqtxt27d3Hr1i2dvpq7XhkZGe1eAq67\nnUtL1Go1iouLcfr0aU2S3Pjbg1OnTqG6uhrXr1/XWo7ucUuWLEFNTU2z03m64s8wY4yxTmTk1TX6\njPYsAVdTU0ORkZHk7OxMIpGI7OzsKDAwkLKzsyk3N5fGjh1LAEgkEtG4cePo4MGDLZbv2rWL7O3t\nCQDJZDLy9/fXjLN7924aOXIkSSQSkkgkNHbsWEpMTGyxTWtxJSYmkkwmIwDk4eFBOTk5tGfPHrK0\ntCQA5OLiollWLTMzk1xcXEgqldLkyZNpw4YNbW7bWgzN9X3nzp0Wr3NycjJNmzaNBgwYQCKRiPr1\n60cvv/wy3bp1S6duREQEzZ8/X6/72Jzmrq0+127RokUkFovJ0dGRRCIRWVpa0qxZsygnJ0drnNLS\nUpo2bRpJJBJydXWlN954g1atWkUAyN3dXbPEWnPX6/jx46RQKCgmJqbF8zh//jwNHz6cTExMCADZ\n29vTli1butW5vPvuu6RUKglAqz/p6emasSIjI8nW1pasra0pKCiIdu/eTQBIqVRqLUtHRDR27Fha\ns2ZNs9ented0+/btJJVKCQANGjSIPvvss7Y8OlrAS8DpjZfQYj0ZP79dSyDSY6Ima7cDBw4gODhY\nr3mxjLVk8eLFSEtLQ2lpqbFD6bCefi4zZszA7t274erq2uVjC4KA1NRUnl+rh6CgIAA8t5P1TPz8\ndi2ebsFYD9W4HFlv0JPO5fHpG1lZWZBIJEZJkBljjHUuTpIZM5CrV69qtitu7SckJMTYobIOiIyM\nxPXr1/HLL79gwYIF2Lx5s7FDYl3o1KlTWLNmDQ4dOgQ3NzfNn+tXXnlFp66vry8UCgVMTU0xfPhw\nZGZmGiFi/TU0NCA+Ph7e3t4t1lGr1di6dSvc3d1hZmYGa2trjBgxArm5uVr1zp49i0mTJkEmk8HB\nwQGRkZGoqanplfFt374dQ4cOhVQqhVwux9ChQxEVFaVZbhUAjh49iu3bt/eoFwN9mpGne/QZ7ZmT\nzFhz1qxZQ2ZmZgSAnnrqKUpLSzN2SO3WE89l3bp1ZGJiQoMGDdLagtoYwHOS9daROZ0bNmwgPz8/\nKi8v15QplUrq168fAaAvv/xSp01GRgbNnDmz3fF2tV9++YUmTZpEAGj06NEt1gsICKAhQ4bQ+fPn\nSa1WU1FREfn7+9Ply5c1dX766SeSSqUUFRVFjx49om+//Zb69+9PCxYs6JXxzZgxg/7+979TSUkJ\nVVRU0IEDB0gsFtPzzz+vVS8hIYGmTp1KDx480HsMnpPctThr6yKcJDPGDM3YSbJKpSIvL68eNUZ7\nk4xt27bR4MGDqaqqSqtcqVTSvn37yMTEhBwdHamsrEzreE9Kki9dukSzZ8+mvXv30pgxY1pMQlNS\nUkgQBMrKymq1v+DgYHJ1daWGhgZNWWxsLAmCQFeuXOl18QUEBOg8H0FBQQSAioqKtMrDwsLIy8uL\n1Gq1XmNwkty1eLoFY4yxdvnwww9RUlLS48d4khs3biAqKgobN25sdpdNb29vhIeHo7CwEG+++aYR\nIjSM0aNH49ChQ5g3bx7Mzc1brPfuu+9i3LhxGDlyZIt16urqcOzYMUydOlVrnfHp06eDiHDkyJFe\nF196errO8+Ho6AgAePTokVZ5dHQ0Ll26hISEBL3HYV2Hk2TGGOsjiAhxcXHw9PSEubk5bGxsMGvW\nLFy9elVTJywsDGZmZrC3t9eUvf7665DL5RAEQbNDZ3h4OFauXImcnBwIggB3d3fs3LkTEokEAwYM\nwOLFi+Hg4ACJRAJvb2+ttaY7MgYAnDhxosvW8AaAnTt3gojg7+/fYp2YmBgMHjwYH3zwAU6dOtVq\nf225D0lJSZDL5ZDJZDhy5AimT58OS0tLODk5ISUlRau/+vp6bNiwAc7OzpBKpRg1ahRSU1M7dtIt\nqK2txfnz5zFmzJhW6928eROPHj3S2bVUqVQC+O2j174Q3/Xr12Ftba21Gyfw22ZEU6dORUJCAq96\n1Y1xkswYY31EdHQ01qxZg3Xr1qGkpARnzpxBfn4+pkyZguLiYgC/JYRNl5RLTEzExo0btcoSEhLg\n5+cHpVIJIsKNGzcQFhaG0NBQqFQqLF++HLm5ucjMzERdXR2ef/555Ofnd3gM4P9fDaWhocFwF6cV\nx44dw5AhQyCTyVqsI5VK8fHHH8PExAQLFy5EZWVli3Xbch+WLl2KFStWoKqqCgqFAqmpqcjJyYGb\nmxsWLlyotcrK6tWrsWPHDsTHx+P27dvw8/PD3Llz8f333xvuIvyfoqIi1NbW4ocffsC0adM0/xDy\n9PREYmKiJuG7c+cOAEChUGi1l0gkkEqlmvPsjfGp1WoUFhZi9+7dOHXqFHbt2gUzMzOdemPHjkVh\nYSF+/PHHdo/FOhcnyYwx1gdUVVUhLi4Os2fPxvz582FlZYWRI0fivffew71797Bnzx6DjSUSiTRv\nSYcNG4akpCRUVFQgOTnZIP3PmDED5eXliIqKMkh/ramsrMSvv/6qecPYGi8vL6xYsQK5ublYvXp1\ns3Xacx+8vb1haWkJOzs7hISEoLKyEnl5eQCA6upqJCUlISAgAIGBgbC2tsb69eshFosNdr0f1zht\nwM7ODlu2bEF2djaKi4sxa9YsLFu2DPv37wcAzQoRpqamOn2IxWJUVVUZPLbuEt+gQYPg5OSE6Oho\n7NixA8HBwc3W8/DwAABcvny53WOxzsVJMmOM9QHZ2dl49OgRxo8fr1U+YcIEmJmZtbj1tiGMHz8e\nMplMazpBT1FSUgIiavUt8uNiYmIwZMgQJCYm4uzZszrHO3ofGt9INr5JvnbtGlQqFUaMGKGpI5VK\nYW9v3ynXu3Eu8PDhw+Ht7Q1bW1tYWVlh48aNsLKy0iT5jXNz6+rqdPqora2FVCo1eGzdJb78/HyU\nlJRg//79+OSTTzB27Nhm59U3PlOd9VaddRwnyYwx1geUlZUBACwsLHSOWVtbo6KiolPHNzc3x927\ndzt1jM5QXV0NAK1+KPY4iUSC5ORkCIKA1157TeeNpKHvQ+O0jvXr12utx37r1i2oVCq9+moLBwcH\nANDMG29kZmYGFxcX5OTkAIBmvvnjawQDgEqlQnV1taaf3hifWCyGnZ0dfH198fnnnyM7Oxtbt27V\nqdeYiDc+Y6z74SSZMcb6AGtrawBoNgkrKyuDk5NTp42tVqs7fYzO0pjI6LP5g5eXFyIiInD9+nWd\nzWYMfR/s7OwAAPHx8aDflnXV/Jw7d06vvtrCwsICHh4e+Pnnn3WO1dXVwcrKCgDg6uoKhUKBW7du\nadVpnFc+atQog8fWHeNzd3eHqakpsrOzdY7V1tYCQKe9VWcdx0kyY4z1ASNGjICFhYXOx1wXLlxA\nbW0tnn76aU2ZSCTS+jCso06fPg0iwsSJEzttjM4yYMAACIKAhw8f6tVu8+bNGDp0KC5evKhVrs99\naItBgwZBIpHg0qVLerXriODgYFy8eBE3b97UlKlUKty6dUuz7JpIJMKLL2HvLh8AACAASURBVL6I\nM2fOaH1gmZGRAUEQWl0ppCfGV1pairlz5+qUX79+HfX19Rg0aJDOscZnauDAgXqNxboOJ8mMMdYH\nSCQSrFy5Eunp6di7dy/Ky8tx+fJlLFmyBA4ODli0aJGmrru7O+7fv4/Dhw9DrVbj7t27Om/cAMDW\n1hZFRUXIzc1FRUWFJultaGjAgwcPUFdXh6ysLISHh8PZ2RmhoaEGGSMjI6PLloCTyWRwc3NDQUGB\nXu0ap100/TBMn/vQ1nEWLFiAlJQUJCUloby8HPX19SgoKMDt27cBACEhIRg4cKDBtsWOiIiAi4sL\nQkNDkZeXh9LSUkRGRqKqqkrrg8WoqCgUFxfjrbfeQmVlJc6dO4fY2FiEhoZiyJAhmnq9IT65XI6T\nJ0/iq6++Qnl5OdRqNS5evIhXX30VcrkcEREROm0an6nW1nNmRtb1+5f0TbzjHmPM0KDnjnsNDQ0U\nGxtLHh4eJBaLycbGhgICAujatWta9UpLS2natGkkkUjI1dWV3njjDVq1ahUBIHd3d8rLyyMioszM\nTHJxcSGpVEqTJ0+mO3fu0KJFi0gsFpOjoyOJRCKytLSkWbNmUU5OjsHGOH78OCkUCoqJidH7mrVn\nx7KwsDASi8WkUqk0Zenp6aRUKgkA9e/fn5YtW9Zs21WrVunsuNeW+5CYmEgymYwAkIeHB+Xk5NCe\nPXvI0tKSAJCLiwv98ssvRERUU1NDkZGR5OzsTCKRiOzs7CgwMJCys7OJ6Led4ADQhg0bWj3Pc+fO\n0aRJk8jBwYEAEACyt7cnb29v+vrrr7Xq5ufn08svv0w2NjZkbm5OzzzzDGVkZOj0+fXXX9MzzzxD\n5ubm5ODgQKtWraLq6mqtOr0lPn9/f3J1dSULCwsyNzcnpVJJISEhWlthP27GjBnk6OiotePfk/CO\ne11LIOJVrLvCgQMHEBwczIuGM8YMRhAEpKam6qw5bEyLFy9GWloaSktLjR1Ks4KCggAAaWlpbW5z\n48YNeHp6Ijk5GfPnz++s0DpNQ0MDnn32WYSGhuK1114zdjg6+mJ8paWlcHJyQkxMDFauXNnmdu15\nfln78XQLxhhjBqXPR249gbu7OzZt2oRNmzbpbC/c3dXX1+Pw4cOoqKhASEiIscPR0Vfji46Oxpgx\nYxAWFmawPpnhcZLMGGOMPcGaNWsQFBSEkJAQvT/iM6bTp0/j0KFDyMjIaPNaz12pL8YXFxeHS5cu\n4fjx4xCLxQbpk3UOTpIZY4wZxNq1a5GcnIyHDx/C1dUVBw8eNHZIBrVlyxaEhYVh27Ztxg6lzXx8\nfLBv3z7NusDdTV+L78iRI6ipqcHp06dhY2NjkD5Z5xEZOwDGGGO9w9atW5vdNKE38fX1ha+vr7HD\nYD3UzJkzMXPmTGOHwdqI3yQzxhhjjDHWBCfJjDHGGGOMNcFJMmOMMcYYY01wkswYY4wxxlgT/OFe\nFztw4ICxQ2CM9SLnzp0zdgg9SuNWwPx3MeuJCgoK4OTkZOww+gzeca+LNO64xxhjjDHWXnPmzOEd\n97oIJ8mMMdZNdMdtphljrK/iOcmMMcYYY4w1wUkyY4wxxhhjTXCSzBhjjDHGWBOcJDPGGGOMMdYE\nJ8mMMcYYY4w1wUkyY4wxxhhjTXCSzBhjjDHGWBOcJDPGGGOMMdYEJ8mMMcYYY4w1wUkyY4wxxhhj\nTXCSzBhjjDHGWBOcJDPGGGOMMdYEJ8mMMcYYY4w1wUkyY4wxxhhjTXCSzBhjjDHGWBOcJDPGGGOM\nMdYEJ8mMMcYYY4w1wUkyY4wxxhhjTXCSzBhjjDHGWBOcJDPGGGOMMdYEJ8mMMcYYY4w1wUkyY4wx\nxhhjTXCSzBhjjDHGWBOcJDPGGGOMMdYEJ8mMMcYYY4w1wUkyY4wxxhhjTXCSzBhjjDHGWBOcJDPG\nGGOMMdYEJ8mMMcYYY4w1wUkyY4wxxhhjTXCSzBhjjDHGWBOcJDPGGGOMMdYEJ8mMMcYYY4w1IRAR\nGTsIxhjraxYtWoRr165plWVmZsLV1RU2NjaaMlNTU3zyySdwcnLq6hAZY6xPExk7AMYY64sGDhyI\nPXv26JRnZWVp/bebmxsnyIwxZgQ83YIxxoxg7ty5T6xjZmaG0NDQzg+GMcaYDp5uwRhjRjJixAj8\n/PPPaO2v4WvXrmHw4MFdGBVjjDGA3yQzxpjR/PnPf4apqWmzxwRBwOjRozlBZowxI+EkmTHGjOTl\nl19GfX19s8dMTU3x6quvdnFEjDHGGvF0C8YYMyJvb29cuHABDQ0NWuWCICA/Px+Ojo5Giowxxvo2\nfpPMGGNG9Morr0AQBK0yExMTTJ48mRNkxhgzIk6SGWPMiIKCgnTKBEHAn//8ZyNEwxhjrBEnyYwx\nZkT9+/eHj4+P1gd8giAgICDAiFExxhjjJJkxxoxs/vz5mmXgTE1N8cILL6Bfv35Gjooxxvo2TpIZ\nY8zIZs+eDTMzMwAAEWH+/PlGjogxxhgnyYwxZmRyuRx/+tOfAPy2y56fn5+RI2KMMcZJMmOMdQPz\n5s0DAAQEBEAulxs5GsYYY7xOMtMRFBSEgwcPGjsMxhhjrMtwOsSaEhk7ANY9TZw4EStWrDB2GIx1\nufj4eAAwyvO/d+9ehISEQCTqWX81nzt3DgkJCUhNTTV2KIzprfH5ZaypnvU3MesyTk5OeOmll4wd\nBmNdLi0tDQCM8vz7+/tDIpF0+biGkJCQwH9nsB6Lk2TWHJ6TzBhj3URPTZAZY6w34iSZMcYYY4yx\nJjhJZowxxhhjrAlOkhljjDHGGGuCk2TGGGOMMcaa4CSZMcY6wfHjx2FlZYUvvvjC2KF0e6dOncKa\nNWtw6NAhuLm5QRAECIKAV155Raeur68vFAoFTE1NMXz4cGRmZhohYv01NDQgPj4e3t7eLdZRq9XY\nunUr3N3dYWZmBmtra4wYMQK5ubla9c6ePYtJkyZBJpPBwcEBkZGRqKmp6ZXxbd++HUOHDoVUKoVc\nLsfQoUMRFRWF8vJyTZ2jR49i+/btqK+vb9cYjLWEk2TGGOsEvDFB27z11lvYuXMn1q5di8DAQNy8\neRNKpRL9+vXD3r17cezYMa36J0+eRFpaGvz8/JCdnY1x48YZKfK2u379Ov7whz8gIiICKpWqxXrB\nwcH49NNPsW/fPqhUKly5cgVKpRKPHj3S1MnOzoavry98fHxw9+5dpKen46OPPsKSJUt6ZXzffPMN\nFi5ciLy8PBQXF2Pz5s3Yvn075syZo6nTuHSij48PysrK2jUOY80ixpqYM2cOzZkzx9hhMGYUvfH5\nV6lU5OXl1Wn9p6amUnv+d7Jt2zYaPHgwVVVVaZUrlUrat28fmZiYkKOjI5WVlWkdz8jIoJkzZ3Yo\n5q5y6dIlmj17Nu3du5fGjBlDo0ePbrZeSkoKCYJAWVlZrfYXHBxMrq6u1NDQoCmLjY0lQRDoypUr\nvS6+gIAAnecjKCiIAFBRUZFWeVhYGHl5eZFardZrjPY+v6z34zfJjDHWy3344YcoKSkxdhhabty4\ngaioKGzcuLHZ9aG9vb0RHh6OwsJCvPnmm0aI0DBGjx6NQ4cOYd68eTA3N2+x3rvvvotx48Zh5MiR\nLdapq6vDsWPHMHXqVAiCoCmfPn06iAhHjhzpdfGlp6frPB+Ojo4AoPUGGwCio6Nx6dIl3hiEGQwn\nyYwxZmBnz56Fs7MzBEHA7t27AQBJSUmQy+WQyWQ4cuQIpk+fDktLSzg5OSElJUXTdufOnZBIJBgw\nYAAWL14MBwcHSCQSeHt748KFC5p6YWFhMDMzg729vabs9ddfh1wuhyAIuHfvHgAgPDwcK1euRE5O\nDgRBgLu7OwDgxIkTsLS0xJYtW7rikujYuXMniAj+/v4t1omJicHgwYPxwQcf4NSpU632R0SIi4uD\np6cnzM3NYWNjg1mzZuHq1auaOm29BwBQX1+PDRs2wNnZGVKpFKNGjeq0bbdra2tx/vx5jBkzptV6\nN2/exKNHj+Ds7KxVrlQqAQBZWVl9Ir7r16/D2toaLi4uWuU2NjaYOnUqEhISeLoTMwhOkhljzMAm\nT56Mb7/9Vqts6dKlWLFiBaqqqqBQKJCamoqcnBy4ublh4cKFUKvVAH5LfkNDQ6FSqbB8+XLk5uYi\nMzMTdXV1eP7555Gfnw/gtySz6TbQiYmJ2Lhxo1ZZQkIC/Pz8oFQqQUS4ceMGAGg+cmpoaOiUa/Ak\nx44dw5AhQyCTyVqsI5VK8fHHH8PExAQLFy5EZWVli3Wjo6OxZs0arFu3DiUlJThz5gzy8/MxZcoU\nFBcXA2j7PQCA1atXY8eOHYiPj8ft27fh5+eHuXPn4vvvvzfcRfg/RUVFqK2txQ8//IBp06Zp/mHk\n6emJxMRETcJ3584dAIBCodBqL5FIIJVKNefZG+NTq9UoLCzE7t27cerUKezatQtmZmY69caOHYvC\nwkL8+OOP7R6LsUacJDPGWBfz9vaGpaUl7OzsEBISgsrKSuTl5WnVEYlEmreiw4YNQ1JSEioqKpCc\nnGyQGGbMmIHy8nJERUUZpD99VFZW4tdff9W8YWyNl5cXVqxYgdzcXKxevbrZOlVVVYiLi8Ps2bMx\nf/58WFlZYeTIkXjvvfdw79497NmzR6dNa/eguroaSUlJCAgIQGBgIKytrbF+/XqIxWKDXf/HNU4b\nsLOzw5YtW5CdnY3i4mLMmjULy5Ytw/79+wFAs0KEqampTh9isRhVVVUGj627xDdo0CA4OTkhOjoa\nO3bsQHBwcLP1PDw8AACXL19u91iMNeIkmTHGjKjxbdjjbzGbM378eMhkMq3pAz1VSUkJiKjVt8iP\ni4mJwZAhQ5CYmIizZ8/qHM/OzsajR48wfvx4rfIJEybAzMxMa5pKc5reg2vXrkGlUmHEiBGaOlKp\nFPb29p1y/RvnAg8fPhze3t6wtbWFlZUVNm7cCCsrK02S3zg3t66uTqeP2tpaSKVSg8fWXeLLz89H\nSUkJ9u/fj08++QRjx45tdp594zPVWW/VWd/CSTJjjPUQ5ubmuHv3rrHD6LDq6moAaPVDscdJJBIk\nJydDEAS89tprOm8kG5f9srCw0GlrbW2NiooKveJrnNaxfv16zZrNgiDg1q1brS6R1l4ODg4AoJlH\n3sjMzAwuLi7IyckBAM3888fXCAYAlUqF6upqTT+9MT6xWAw7Ozv4+vri888/R3Z2NrZu3apTrzER\nb3zGGOsITpIZY6wHUKvVKCsrg5OTk7FD6bDGREafzR+8vLwQERGB69evY/PmzVrHrK2tAaDZZLg9\n18zOzg4AEB8fDyLS+jl37pxefbWFhYUFPDw88PPPP+scq6urg5WVFQDA1dUVCoUCt27d0qrTOM98\n1KhRBo+tO8bn7u4OU1NTZGdn6xyrra0FgE57q876Fk6SGWOsBzh9+jSICBMnTtSUiUSiJ07T6I4G\nDBgAQRDw8OFDvdpt3rwZQ4cOxcWLF7XKR4wYAQsLC52P6i5cuIDa2lo8/fTTeo0zaNAgSCQSXLp0\nSa92HREcHIyLFy/i5s2bmjKVSoVbt25pll0TiUR48cUXcebMGa0PLjMyMiAIQqsrhfTE+EpLSzF3\n7lyd8uvXr6O+vh6DBg3SOdb4TA0cOFCvsRhrDifJjDHWDTU0NODBgweoq6tDVlYWwsPD4ezsjNDQ\nUE0dd3d33L9/H4cPH4Zarcbdu3d13uIBgK2tLYqKipCbm4uKigqo1WpkZGQYbQk4mUwGNzc3FBQU\n6NWucdpF0w/DJBIJVq5cifT0dOzduxfl5eW4fPkylixZAgcHByxatEjvcRYsWICUlBQkJSWhvLwc\n9fX1KCgowO3btwEAISEhGDhwoMG2xY6IiICLiwtCQ0ORl5eH0tJSREZGoqqqSuuDxaioKBQXF+Ot\nt95CZWUlzp07h9jYWISGhmLIkCGaer0hPrlcjpMnT+Krr75CeXk51Go1Ll68iFdffRVyuRwRERE6\nbRqfqdbWc2aszYyxgwnr3nrjjmOMtZUhnv9du3aRvb09ASCZTEb+/v6UmJhIMpmMAJCHhwfl5OTQ\nnj17yNLSkgCQi4sL/fLLL0REtGjRIhKLxeTo6EgikYgsLS1p1qxZlJOTozVOaWkpTZs2jSQSCbm6\nutIbb7xBq1atIgDk7u5OeXl5RESUmZlJLi4uJJVKafLkyXTnzh06fvw4KRQKiomJ6dC5ErVvx7Kw\nsDASi8WkUqk0Zenp6aRUKgkA9e/fn5YtW9Zs21WrVunsuNfQ0ECxsbHk4eFBYrGYbGxsKCAggK5d\nu6apo889qKmpocjISHJ2diaRSER2dnYUGBhI2dnZRPTbTnAAaMOGDa2e57lz52jSpEnk4OBAAAgA\n2dvbk7e3N3399ddadfPz8+nll18mGxsbMjc3p2eeeYYyMjJ0+vz666/pmWeeIXNzc3JwcKBVq1ZR\ndXW1Vp3eEp+/vz+5urqShYUFmZubk1KppJCQELp8+XKz9WfMmEGOjo5aO/49Ce+4x1oiEPGK20xb\nUFAQACAtLc3IkTDW9brD87948WKkpaWhtLTUaDHo48CBAwgODtZrA4cbN27A09MTycnJmD9/fidG\n1zkaGhrw7LPPIjQ0FK+99pqxw9HRF+MrLS2Fk5MTYmJisHLlyja3a8/zy/oGnm7BGGPdkD4ftfVE\n7u7u2LRpEzZt2qSzvXB3V19fj8OHD6OiogIhISHGDkdHX40vOjoaY8aMQVhYmMH6ZH0bJ8msy9TU\n1GD58uWwt7eHTCbDiRMnWiw/fvw4rKys8MUXX7S5//a06UkmTJgAU1PTJ24NCwDPPvus1tJVj/80\nt0xWaw4dOgQ3NzdN+ydtPhEXFwdBEGBiYoKhQ4fizJkzeo3X1F//+lcoFAoIgtDmD6n+/ve/az4O\ne++99zo0Pus8a9asQVBQEEJCQvT+iM+YTp8+jUOHDiEjI6PNaz13pb4YX1xcHC5duoTjx49DLBYb\npE/GOElmXeadd97BiRMncPXqVSQkJGjeHjVX3p5fe/X2X5V99913mDZtWof7mTx5sl71AwMDcfPm\nTc3uaB988EGLKyrU19dj586dAIDnnnsOV69exR/+8IcOxfvBBx/g/fff16vNm2++qbMtdE+xdu1a\nJCcn4+HDh3B1dcXBgweNHVKn2rJlC8LCwrBt2zZjh9JmPj4+2Ldvn2Zd4O6mr8V35MgR1NTU4PTp\n07CxsTFIn4wBgMjYAbC+4/Dhwxg/fjysra3x3//9308s1/fN0owZMzrtbVRVVRV8fHw6JfHSt29B\nEJ5YRyKRoLy8HAqFQqt88eLFeOmll9oVJwA8/fTT+OGHH3D48GHN3N3HHTp0CI6Ojs2usMDaZuvW\nrc1uktCb+fr6wtfX19hhsB5q5syZmDlzprHDYL0Qv0lmXaagoKDZX4O1VN6dfPjhh81ugWqMvtty\nrU6cOKGTIOfn5+Onn37Cc889p3eMjZYuXQoAePfdd5s9HhcXp9cHM23Vln8YMMYYY4bESTIziPr6\nemzYsAHOzs6QSqUYNWoUUlNTAQD/+te/4O7ujtu3b+OTTz7RzIttqfzs2bNwdnaGIAjYvXu31jif\nffYZxo8fD4lEArlcjqeeegqbN29usU1rcSUlJUEul0Mmk+HIkSOYPn06LC0t4eTkhJSUFE0f4eHh\nWLlyJXJyciAIAtzd3dvc9kkxNNf3k9y4cQNDhw6FXC6HVCrFlClTcPbs2Se2e/vtt7F8+XKtshMn\nTui1Vu5zzz0HT09P/Oc//8G1a9e0jv3v//4vVCpVi28EiQhxcXHw9PSEubk5bGxsMGvWLFy9elWn\nXmxsLIYMGQJzc3NYWVlh1apVOv21dl0ZY4yxDjPe6nOsu2rPOrFvvvkmmZub08GDB+nBgwe0du1a\nMjExoe+++05TZ+DAgfTqq6/qtG2uPD8/nwDQrl27NGXx8fEEgLZt20alpaV0//59+sc//kHz5s1r\nsc2T4lq3bh0BoH//+9/08OFDKikpoSlTppBcLqfa2lpNP4GBgaRUKrVibGvbJ8XQXN8t8fHxITc3\nN/r1119JrVbTTz/9RL///e9JIpFo1ndtTkFBAQ0bNozq6+u1yr/88ktSKBS0adOmJ46tVCrp119/\npf/5n/8hABQeHq51PCAggJKTk6miooIAkI+Pj9bxDRs2kJmZGX322WdUVlZGWVlZNG7cOOrfvz/d\nuXNHU2/dunUkCAK988479ODBA1KpVJSYmEgA6OLFi5p6T7qu169fJwD07rvvPvHcHsfrhOuP15ll\nPRk/v6wl/CaZdVh1dTWSkpIQEBCAwMBAWFtbY/369RCLxUhOTjbIGGq1Ghs3bsS0adOwevVq2Nra\nwsbGBn/5y18wYcKEDsfl7e0NS0tL2NnZISQkBJWVlcjLy2tTbK217Yxro1Ao8NRTT0EkEmH48OF4\n//33UV1djT179rTY5u2338Ybb7wBExPtP/IzZsxAeXn5E1eseFzjbleffPIJqqqqAAA3b97Ed999\n1+wWssBv867j4uIwe/ZszJ8/H1ZWVhg5ciTee+893Lt3TxN7VVUV4uPj8cc//hERERGwtraGVCqF\nra2tVn9d8cwxxhjr2/jDPdZh165dg0qlwogRIzRlUqkU9vb2Or9Kb6+srCyUlZXhhRde0Co3NTXV\nmULQ0bjMzMwAoMUVHFrTtG1XXJuRI0fCysoKWVlZzR4vKirC0aNHERsba5DxrKysMHfuXLz//vv4\n/PPPsWDBAsTHx2Pp0qUwMzNDbW2tTpvs7Gw8evQI48eP1yqfMGECzMzMcOHCBQC/TSVRqVTw8fFp\nNYbOvq4FBQU4cOBAh/vpK86dOwcAfM1Yj9T4/DLWFCfJrMMqKysBAOvXr8f69eu1jjk4OBhkjPLy\ncgCAtbV1t4qrM2L48ssv4efnp1U2b9487N27t8VxxGJxi0n99u3bsXDhQkgkEn1Cb9XSpUvx/vvv\n47333kNAQADS0tJw5cqVFuuXlZUBQLNrNFtbW6OiogLAb8kpANjZ2bU6fmff2/PnzyM4OLjD/fQ1\nfM0YY70JT7dgHdaY0MTHx4OItH4M9S/03/3udwCAe/fudau4OiOGP/3pTzp1W0uQ6+rqcP/+fTg7\nO+scu3PnDvbv369ZlcJQxowZg4kTJ+L//b//h0WLFiEoKKjV9Ukb/3HTmAw/rqysDE5OTgCgSeRr\nampaHb+z7+2cOXN0+uWfln8aP5g0dhz8wz/t+eEPfllLOElmHTZo0CBIJJI274bWHk899RRsbW1x\n8uTJbhVXd4jhP//5DxoaGjBu3DidY9u3b8f8+fN15vQaQmPiffDgQaxYsaLVuiNGjICFhQW+//57\nrfILFy6gtrYWTz/9tKaeiYkJvv7661b76w73ljHGWO/GSTLrMIlEggULFiAlJQVJSUkoLy9HfX09\nCgoKcPv2bYOMYW5ujrVr1+LMmTMICwtDYWEhGhoaUFFRgZ9//rnT47K1tUVRURFyc3NRUVHR5vnK\nbYlB375ra2vx8OFD1NXVITMzE2FhYXBxcUFoaKhWveLiYnz00UetJrAZGRl6LQH3uJdeegn9+/dH\nQEAA3NzcWq0rkUiwcuVKpKenY+/evSgvL8fly5exZMkSODg4YNGiRQB+e0McGBiIgwcP4sMPP0R5\neTmysrJ0PkrsimeOMcZYH0eMNdGeJbBqamooMjKSnJ2dSSQSkZ2dHQUGBlJ2djbl5ubS2LFjCQCJ\nRCIaN24cHTx4sMXyXbt2kb29PQEgmUxG/v7+mnF2795NI0eOJIlEQhKJhMaOHUuJiYkttmktrsTE\nRJLJZASAPDw8KCcnh/bs2UOWlpYEgFxcXDTLqmVmZpKLiwtJpVKaPHkybdiwoc1tW4uhub4fXw6t\nqeTkZJo2bRoNGDCARCIR9evXj15++WW6deuWTt2IiAiaP39+q/ft+PHjpFAoKCYmpsU66enppFQq\nCQD179+fli1bpjn2t7/9jb799lvNf69fv15zH0xMTGjYsGH0zTffEBFRQ0MDxcbGkoeHB4nFYrKx\nsaGAgAC6du2a1ngVFRX017/+lfr160cWFhaa6w2AnJyc6Mcff3zidX3nnXdo4MCBBIDkcjnNnj27\n1evwOF4CTn+8hBbryfj5ZS0RiIiMkZyz7qtxu+G0tDQjR8JY1+PnX38HDhxAcHAw+H8nrCfi55e1\nhKdbMMYYY4wx1gQnyYwxxhhjjDXBSTJjjLFu59SpU1izZg0OHToENzc3CIIAQRDwyiuv6NT19fWF\nQqGAqakphg8fjszMTCNE3HYxMTGa83n85/HNcYzZ36ZNmzBs2DBYWlrC3Nwc7u7u+Nvf/oZHjx61\ne1y1Wo2tW7fC3d0dZmZmsLa2xogRI5Cbm6tV7+zZs5g0aRJkMhkcHBwQGRmptSTk0aNHsX37dtTX\n17fr3BjTByfJjDHGupW33noLO3fuxNq1axEYGIibN29CqVSiX79+2Lt3L44dO6ZV/+TJk0hLS4Of\nnx+ys7ObXQ6Rtd1XX32FZcuWITc3F/fu3cPWrVuRkJCgma/fHsHBwfj000+xb98+qFQqXLlyBUql\nUivxzs7Ohq+vL3x8fHD37l2kp6fjo48+wpIlSzR1/P39IZFI4OPjo9mkiLHOwkkyY4x1M1VVVfD2\n9u7xY7TH22+/jc8//xwHDhyAQqHQOrZz506YmJhg0aJFePjwoZEiNIzPPvtMZ1OLn376qVv0Z2Fh\ngUWLFsHW1hYKhQIvvfQSAgICcOLECeTn5+s97ueff47Dhw8jLS0Nv//97yESieDg4IAjR45ovXXe\nvHkz7O3tsXHjRsjlcnh5eSEyMhIff/yx1nbzy5cvx+jRo/Hiiy+irq6uXefIWFtwkswYY93Mhx9+\niJKSkh4/hr5u3LiBqKgobNy4sdlt1L29vREeHo7CwkK8+eabRoiwb/jyyy9hamqqVda/f38AgEql\n0ru/d999F+PGjcPIkSNbrFNXV4djx45h6tSpEARBUz59+nQQEY4czqrRAgAAIABJREFUOaJVPzo6\nGpcuXUJCQoLe8TDWVpwkM8ZYBxER4uLi4OnpCXNzc9jY2GDWrFlab7/CwsJgZmYGe3t7Tdnrr78O\nuVwOQRA0W66Hh4dj5cqVyMnJgSAIcHd3x86dOyGRSDBgwAAsXrwYDg4OkEgk8Pb2xoULFwwyBgCc\nOHGi3ZvLGMLOnTtBRPD392+xTkxMDAYPHowPPvgAp06darW/ttyXpKQkyOVyyGQyHDlyBNOnT4el\npSWcnJyQkpKi1V99fT02bNgAZ2dnSKVSjBo1qs9saVxYWAipVApXV1e92tXW1uL8+fMYM2ZMq/Vu\n3ryJR48ewdnZWatcqVQCALKysrTKbWxsMHXqVCQkJPDSbazTcJLMGGMdFB0djTVr1mDdunUoKSnB\nmTNnkJ+fjylTpqC4uBjAbwngSy+9pNUuMTERGzdu1CpLSEiAn58flEoliAg3btxAWFgYQkNDoVKp\nsHz5cuTm5iIzMxN1dXV4/vnnNb8C78gYADQfQzU0NBju4ujh2LFjGDJkCGQyWYt1pFIpPv74Y5iY\nmGDhwoWorKxssW5b7svSpUuxYsUKVFVVQaFQIDU1FTk5OXBzc8PChQu1dsBcvXo1duzYgfj4eNy+\nfRt+fn6YO3euznbrbbFmzRrY2NjAzMwMrq6umDVrFr777ju9++ms/h6nUqnw1VdfYeHChTAzM9Nr\n3KKiItTW1uKHH37AtGnTNP/A8/T0RGJioibBvXPnDgDoTLGRSCSQSqWa+/W4sWPHorCwED/++KNB\nzpOxpjhJZoyxDqiqqkJcXBxmz56N+fPnw8rKCiNHjsR7772He/fu6Wyp3REikUjzVnTYsGFISkpC\nRUUFkpOTDdL/jBkzUF5ejqioKIP0p4/Kykr8+uuvmjeHrfHy8sKKFSuQm5uL1atXN1unPffF29sb\nlpaWsLOzQ0hICCorK5GXlwcAqK6uRlJSEgICAhAYGAhra2usX78eYrFY7+v/6quv4ujRo8jPz8ej\nR4+QkpKCvLw8TJ06FdnZ2Xr11Rn9NbV161Y4ODggJiZG73EbP8yzs7PDli1bkJ2djeLiYsyaNQvL\nli3D/v37AUCzgkXTaR4AIBaLUVVVpVPu4eEBAP8fe/ceFmW19g/8OzJnzirqJIII5Nm01ATrp253\nlrI94BHbtiO3vmjbCCVDyiMimbbVV4O6OtGbJgLpRkvRLreR2zd0m4ckLA8YKHhAEeQwIDPM+v3R\ny2xnBnCG06B+P9fFH65Zz1r3c8J7FutZD7Kyspq8j0R1YZJMRNQE2dnZKC8vx+DBg03KhwwZArlc\nbjIdorkNHjwYarXaZPrAg6qwsBBCiAZHke8VGxuLnj17Ij4+HkeOHLH4vKnnpXbEtHYk+dy5c9Bq\ntSYPmqlUKnTp0sXm49+tWzcMGjQITk5OkMvlGDZsGBITE1FZWYn4+Hib2mqJ9u61a9cupKSk4MCB\nAxajvNb0q1AoAAB9+/ZFYGAg2rdvD1dXV6xatQqurq7GLyu1c9DrehCvuroaKpXKorz2WqlrlJmo\nOTBJJiJqgtplqJycnCw+c3NzQ1lZWYv2r1AocPPmzRbtozVUVVUB+E9SdT9KpRKJiYmQSCSYPXu2\nxUhjc5+X2mkdS5cuNVkTOC8vr1EPs5nr378/HBwccP78+Sa31Vzt7dixA2vXrkVGRga6d+/eqH41\nGg0AGOfD15LL5fD29kZOTg4AGOfRl5aWmtTTarWoqqoytnOv2sS59toham5MkomImsDNzQ0A6ky6\nSkpK4Onp2WJ963S6Fu+jtdQmPLa8JCIgIACLFi3ChQsXsHr1apPPmvu8eHh4AAA2btxoseRZZmam\nTW3VxWAwwGAwWP0loaXb27JlC7Zt24ZDhw7hsccea3S/Tk5O8Pf3x9mzZy3q6vV6uLq6AgB8fHzg\n7OyMvLw8kzq18+UHDBhgsX11dTUA1DnKTNQcmCQTETVBv3794OTkZPHw1rFjx1BdXY2nnnrKWCaV\nSk0eBGuqjIwMCCEwbNiwFuujtXTq1AkSicTm9Y9Xr16NXr164dSpUybltpwXa3Tr1g1KpRKnT5+2\nabu6PP/88xZlx48fhxACAQEBdm1PCIGoqChkZWUhLS2tzpF4W/udMWMGTp06hUuXLhnLtFot8vLy\njMvCSaVSjBs3DocPHzZ5cDQ9PR0SiaTOFU9qr5XOnTvbtI9E1mKSTETUBEqlEpGRkdi1axe2bduG\n0tJSZGVlYf78+dBoNAgLCzPW9fPzw+3bt5GWlgadToebN29ajJwBQPv27XH16lXk5uairKzMmPQa\nDAYUFxdDr9fjzJkziIiIgJeXF0JDQ5ulj/T0dLstAadWq9GjRw/k5+fbtF3ttAvzB75sOS/W9vPK\nK68gKSkJCQkJKC0tRU1NDfLz83Ht2jUAQEhICDp37nzf12IXFBRgx44dKCkpgU6nQ2ZmJubMmQMv\nLy+Tt8vZo72zZ89i3bp1+PjjjyGTySxeOf3ee+/Z3O+iRYvg7e2N0NBQXL58GUVFRYiKikJlZaXJ\ng5fLli3DjRs3sGLFClRUVCAzMxPr169HaGgoevbsaRFr7bXS0PrLRE0iiMxMnTpVTJ061d5hENlF\nY65/g8Eg1q9fL/z9/YVMJhPu7u4iODhYnDt3zqReUVGRGDVqlFAqlcLHx0e89tprYvHixQKA8PPz\nE5cvXxZCCHHy5Enh7e0tVCqVeOaZZ8T169dFWFiYkMlkomvXrkIqlQoXFxcxadIkkZOT02x97Nu3\nTzg7O4vY2Fib9j85OVk0x38n4eHhQiaTCa1WayzbtWuX8PX1FQBEx44dxYIFC+rcdvHixWLixIkm\nZdacl/j4eKFWqwUA4e/vL3JycsRHH30kXFxcBADh7e0tzp8/L4QQ4u7duyIqKkp4eXkJqVQqPDw8\nxJQpU0R2drYQQojg4GABQCxfvrzB/YyMjBS+vr7C0dFRSKVS4enpKebOnSuuXr1qUs8e7WVlZQkA\n9f6sX7/e5n6FEOLKlSti5syZwt3dXSgUCjF06FCRnp5uUe/7778XQ4cOFQqFQmg0GrF48WJRVVVV\nZ6xBQUGia9euwmAwNHh87qe5rl96+EiE4CrcZGratGkAgNTUVDtHQtT62ur1P2/ePKSmpqKoqMje\noVhISUnBjBkzmvxSh4sXL6J3795ITEzErFmzmim61mMwGDBy5EiEhoZi9uzZD3179lRUVARPT0/E\nxsYiMjKySW011/VLDx9OtyAiekDY8lDbg8jPzw8xMTGIiYkxrq/7oKipqUFaWhrKysoQEhLy0Ldn\nbytXrsTAgQMRHh5u71DoIcYkmYiI2ozo6GhMmzYNISEhNj/EZ08ZGRnYuXMn0tPTrV7r+UFuz542\nbNiA06dPY9++fZDJZPYOhx5iTJKJiNq4t956C4mJibhz5w58fHzw1Vdf2TukFrVmzRqEh4fjnXfe\nsXcoVhs9ejS+/PJL43q/D3t79rJ7927cvXsXGRkZcHd3t3c49JCT2jsAIiJqWFxcHOLi4uwdRqsa\nM2YMxowZY+8wqI2ZOHEiJk6caO8w6BHBkWQiIiIiIjNMkomIiIiIzDBJJiIiIiIywySZiIiIiMgM\nH9yjOh09etT4UgWiR8nRo0cBgNe/DWpfD8xjRg8iW1+FTo8OvnGPLGzYsAGZmZn2DoPokZOeno5B\ngwY98Mt0ET2I2tpbNsn+mCQTEbUREokEycnJmD59ur1DISJ65HFOMhERERGRGSbJRERERERmmCQT\nEREREZlhkkxEREREZIZJMhERERGRGSbJRERERERmmCQTEREREZlhkkxEREREZIZJMhERERGRGSbJ\nRERERERmmCQTEREREZlhkkxEREREZIZJMhERERGRGSbJRERERERmmCQTEREREZlhkkxEREREZIZJ\nMhERERGRGSbJRERERERmmCQTEREREZlhkkxEREREZIZJMhERERGRGSbJRERERERmmCQTEREREZlh\nkkxEREREZIZJMhERERGRGSbJRERERERmmCQTEREREZlhkkxEREREZIZJMhERERGRGSbJRERERERm\nmCQTEREREZlhkkxEREREZEZq7wCIiB5FJSUlEEJYlFdUVKC4uNikzMnJCTKZrLVCIyIiABJR129p\nIiJqUX/4wx/w3Xff3beeg4MDCgoK0Llz51aIioiIanG6BRGRHcycORMSiaTBOu3atcP/+3//jwky\nEZEdMEkmIrKDqVOnQipteMabRCLBX/7yl1aKiIiI7sUkmYjIDtzd3TFmzBg4ODjUW6ddu3YIDg5u\nxaiIiKgWk2QiIjuZNWsWDAZDnZ9JpVIEBQXB1dW1laMiIiKASTIRkd1MmDABCoWizs9qamowa9as\nVo6IiIhqMUkmIrITtVqN4ODgOpd3U6lUGDdunB2iIiIigEkyEZFdvfjii9DpdCZlMpkMU6dOhUql\nslNURETEJJmIyI6ef/55i3nHOp0OL774op0iIiIigEkyEZFdyWQyhISEQC6XG8vc3NwwevRoO0ZF\nRERMkomI7GzmzJmorq4G8HvSPGvWrPuuoUxERC2Lr6UmIrIzg8GAxx57DDdu3AAAHDlyBMOHD7dz\nVEREjzaOJBMR2Vm7du3w0ksvAQA0Gg0CAwPtHBEREVn8PS8/Px8//PCDPWIhInpkdezYEQDw9NNP\nIzU11c7REBE9Wrp164aAgACTMovpFikpKZgxY0arBkZEREREZC9Tp061GKCo98kQTlUmImpdX331\nFaZOnWrvMMgK06ZNAwCO+tugdhCO+QW1NbX3sznOSSYiaiOYIBMRtR1MkomIiIiIzDBJJiIiIiIy\nwySZiIiIiMgMk2QiIiIiIjNMkomIiIiIzDBJJiIispN9+/bB1dUVX3/9tb1DafMOHjyI6Oho7Ny5\nEz169IBEIoFEIjG+rfJeY8aMgbOzMxwcHNC3b1+cPHnSDhFbLzY21rg/9/7069evTbQXExODPn36\nwMXFBQqFAn5+fnjzzTdRXl7e6H51Oh3i4uLg5+cHuVwONzc39OvXD7m5uSb1jhw5guHDh0OtVkOj\n0SAqKgp37941fr5nzx68++67qKmpadS+NYRJMhERkZ1wzWDrrFixAps3b8Zbb72FKVOm4NKlS/D1\n9UWHDh2wbds27N2716T+t99+i9TUVIwfPx7Z2dl48skn7RT5w+HQoUNYsGABcnNzcevWLcTFxWHT\npk31ri9sjRkzZuCLL77Al19+Ca1Wi19++QW+vr4miXd2djbGjBmD0aNH4+bNm9i1axc+++wzzJ8/\n31hnwoQJUCqVGD16NEpKSpq0n+aYJBMREdlJUFAQ7ty5g/Hjx9s7FFRWViIwMNDeYVhYu3YtduzY\ngZSUFDg7O5t8tnnzZrRr1w5hYWG4c+eOnSJsHlu3boUQwuTn559/bhPtOTk5ISwsDO3bt4ezszOm\nT5+O4OBg7N+/H1euXLG53x07diAtLQ2pqal4+umnIZVKodFosHv3bpNR59WrV6NLly5YtWoVHB0d\nERAQgKioKHz++ef49ddfjfVef/11PPHEExg3bhz0en2j9rEuTJKJiIgIn376KQoLC+0dhomLFy9i\n2bJlWLVqFZRKpcXngYGBiIiIQEFBAd544w07RPho+Oabb+Dg4GBS1rFjRwCAVqu1ub0PPvgATz75\nJPr3719vHb1ej71792LEiBGQSCTG8rFjx0IIgd27d5vUX7lyJU6fPo1NmzbZHE99mCQTERHZwZEj\nR+Dl5QWJRIL3338fAJCQkABHR0eo1Wrs3r0bY8eOhYuLCzw9PZGUlGTcdvPmzVAqlejUqRPmzZsH\njUYDpVKJwMBAHDt2zFgvPDwccrkcXbp0MZb97W9/g6OjIyQSCW7dugUAiIiIQGRkJHJyciCRSODn\n5wcA2L9/P1xcXLBmzZrWOCQWNm/eDCEEJkyYUG+d2NhYPP744/jkk09w8ODBBtsTQmDDhg3o3bs3\nFAoF3N3dMWnSJJNRSWvPAQDU1NRg+fLl8PLygkqlwoABA5CcnNy0nX5AFBQUQKVSwcfHx6btqqur\ncfToUQwcOLDBepcuXUJ5eTm8vLxMyn19fQEAZ86cMSl3d3fHiBEjsGnTpmabxsQkmYiIyA6eeeYZ\n/PDDDyZlr776KhYuXIjKyko4OzsjOTkZOTk56NGjB+bOnQudTgfg9+Q3NDQUWq0Wr7/+OnJzc3Hy\n5Eno9Xo899xzxj+Bb968GdOnTzfpIz4+HqtWrTIp27RpE8aPHw9fX18IIXDx4kUAMD4MZTAYWuQY\n3M/evXvRs2dPqNXqeuuoVCp8/vnnaNeuHebOnYuKiop6665cuRLR0dF4++23UVhYiMOHD+PKlSt4\n9tlncePGDQDWnwMAWLJkCdatW4eNGzfi2rVrGD9+PF588UX8+OOPNu9rdHQ03N3dIZfL4ePjg0mT\nJuH48eM2t9NS7d1Lq9Xi0KFDmDt3LuRyuU39Xr16FdXV1Thx4gRGjRpl/ILXu3dvxMfHGxPc69ev\nA4DFFBulUgmVSmU8X/caNGgQCgoK8NNPPzXLfjJJJiIiaoMCAwPh4uICDw8PhISEoKKiApcvXzap\nI5VKjaOiffr0QUJCAsrKypCYmNgsMQQFBaG0tBTLli1rlvZsUVFRgd9++804ctiQgIAALFy4ELm5\nuViyZEmddSorK7FhwwZMnjwZs2bNgqurK/r3748PP/wQt27dwkcffWSxTUPnoKqqCgkJCQgODsaU\nKVPg5uaGpUuXQiaT2Xz8X375ZezZswdXrlxBeXk5kpKScPnyZYwYMQLZ2dk2tdUS7ZmLi4uDRqNB\nbGyszf3WPpjn4eGBNWvWIDs7Gzdu3MCkSZOwYMECbN++HQCMK1iYT/MAAJlMhsrKSotyf39/AEBW\nVlaT9xFgkkxERNTm1Y7W3TuKWZfBgwdDrVabTB94UBUWFkII0eAo8r1iY2PRs2dPxMfH48iRIxaf\nZ2dno7y8HIMHDzYpHzJkCORyuck0lbqYn4Nz585Bq9WaPGimUqnQpUsXm49/t27dMGjQIDg5OUEu\nl2PYsGFITExEZWUl4uPjbWqrJdq7165du5CSkoIDBw5YjPJa069CoQAA9O3bF4GBgWjfvj1cXV2x\natUquLq6Gr+s1M5Br+tBvOrqaqhUKovy2mulrlHmxmCSTERE9BBRKBS4efOmvcNosqqqKgD/Saru\nR6lUIjExERKJBLNnz7YYaaxdHszJycliWzc3N5SVldkUX+20jqVLl5qsCZyXl9eoh9nM9e/fHw4O\nDjh//nyT22qu9nbs2IG1a9ciIyMD3bt3b1S/Go0GAIzz4WvJ5XJ4e3sjJycHAIzz6EtLS03qabVa\nVFVVGdu5V23iXHvtNBWTZCIiooeETqdDSUkJPD097R1Kk9UmPLa8JCIgIACLFi3ChQsXsHr1apPP\n3NzcAKDOZLgxx8zDwwMAsHHjRoslzzIzM21qqy4GgwEGg8HqLwkt3d6WLVuwbds2HDp0CI899lij\n+3VycoK/vz/Onj1rUVev18PV1RUA4OPjA2dnZ+Tl5ZnUqZ0vP2DAAIvtq6urAaDOUebGYJJMRET0\nkMjIyIAQAsOGDTOWSaXS+07TaIs6deoEiURi8/rHq1evRq9evXDq1CmT8n79+sHJycniobpjx46h\nuroaTz31lE39dOvWDUqlEqdPn7Zpu7o8//zzFmXHjx+HEAIBAQF2bU8IgaioKGRlZSEtLa3OkXhb\n+50xYwZOnTqFS5cuGcu0Wi3y8vKMy8JJpVKMGzcOhw8fNnlwND09HRKJpM4VT2qvlc6dO9u0j/Vh\nkkxERPSAMhgMKC4uhl6vx5kzZxAREQEvLy+EhoYa6/j5+eH27dtIS0uDTqfDzZs3LUbnAKB9+/a4\nevUqcnNzUVZWBp1Oh/T0dLstAadWq9GjRw/k5+fbtF3ttAvzB76USiUiIyOxa9cubNu2DaWlpcjK\nysL8+fOh0WgQFhZmcz+vvPIKkpKSkJCQgNLSUtTU1CA/Px/Xrl0DAISEhKBz5873fS12QUEBduzY\ngZKSEuh0OmRmZmLOnDnw8vIyebucPdo7e/Ys1q1bh48//hgymczildPvvfeezf0uWrQI3t7eCA0N\nxeXLl1FUVISoqChUVlaaPHi5bNky3LhxAytWrEBFRQUyMzOxfv16hIaGomfPnhax1l4rDa2/bBNh\nJjk5WdRRTERERP9n6tSpYurUqU1qY8uWLaJLly4CgFCr1WLChAkiPj5eqNVqAUD4+/uLnJwc8dFH\nHwkXFxcBQHh7e4vz588LIYQICwsTMplMdO3aVUilUuHi4iImTZokcnJyTPopKioSo0aNEkqlUvj4\n+IjXXntNLF68WAAQfn5+4vLly0IIIU6ePCm8vb2FSqUSzzzzjLh+/brYt2+fcHZ2FrGxsU3aVyEa\nl1+Eh4cLmUwmtFqtsWzXrl3C19dXABAdO3YUCxYsqHPbxYsXi4kTJ5qUGQwGsX79euHv7y9kMplw\nd3cXwcHB4ty5c8Y6tpyDu3fviqioKOHl5SWkUqnw8PAQU6ZMEdnZ2UIIIYKDgwUAsXz58gb3MzIy\nUvj6+gpHR0chlUqFp6enmDt3rrh69apJPXu0l5WVJQDU+7N+/Xqb+xVCiCtXroiZM2cKd3d3oVAo\nxNChQ0V6erpFve+//14MHTpUKBQKodFoxOLFi0VVVVWdsQYFBYmuXbsKg8HQ4PExV9/9LBHCdMXl\nlJQUzJgxg++TJyIiqse0adMAAKmpqXaLYd68eUhNTUVRUZHdYrBFY/KLixcvonfv3khMTMSsWbNa\nMLqWYTAYMHLkSISGhmL27NkPfXv2VFRUBE9PT8TGxiIyMtKmbeu7nzndgoiI6AFly0NtDyI/Pz/E\nxMQgJibGuL7ug6KmpgZpaWkoKytDSEjIQ9+eva1cuRIDBw5EeHh4s7XZaknye++9Z5yE/+GHH9q0\n7ZAhQ+Dg4HDfVxi2ZAzUNtR3Dvft2wdXV1d8/fXXLdZ3TEwM+vTpAxcXFygUCvj5+eHNN9+87y/u\nOXPmwNnZGRKJpEkPeBgMBmzcuBGBgYGNbuNe58+fx2uvvYa+ffvCxcUFcrkcHh4e6NWrFyZPnox/\n/OMfxroP0v27fft2SCSSZjtObdXOnTvRo0cPk7mBMpkMXbt2xZ///Gf88ssvLdZ3W78P6zo2EokE\ncrkcnTp1wsiRI7F+/XoUFxe3WJzUfKKjozFt2jSEhITY/BCfPWVkZGDnzp1IT0+3eq3nB7k9e9qw\nYQNOnz6Nffv2QSaTNV/D5vMvWnJO8oULFwQA8cEHH9i87ejRo8UTTzxh1xiobajrHH7zzTfCxcVF\n7Nmzp8X6HTFihIiPjxdFRUWitLRUJCcnC5lMJl544YX7bpuUlCQAiFOnTjWq7/Pnz4vhw4cLAM1y\nHyQmJgq5XC6eeeYZsX//flFcXCyqqqpETk6O+Prrr0VQUJAICwsz2eZBuX+DgoKM8xUvXLjQ5D7b\nOl9fX+Hq6iqEEKK8vFzs2bNHeHl5CScnJ/Hrr7+2WL8Pwn1477ExGAyiuLhYfPfddyI0NFRIJBKh\n0WjE8ePHGxVHc8xJboro6Gghl8sFANG9e3eRmppqt1is1dT84sCBAyIqKqoZI6KHQVpamoiLixN6\nvb7RbdR3Pz9Q0y0kEom9Q6A2KigoCHfu3MH48eNbrA8nJyeEhYWhffv2cHZ2xvTp0xEcHIz9+/fj\nypUrLdbvTz/9hCVLlmD+/PlNHo0FgKNHj2LOnDkIDAzEd999h+effx5ubm5QKBTo0aMH/vSnP2Hz\n5s3NELmp1rh/i4qKcPbsWaxatQoA8MUXX7R4n22Jo6Mjxo8fj//+7/9GeXk5tmzZ0qr9t+X7UCKR\nwM3NDSNHjkRiYiJSUlJw48YNY8wPmri4ONy9exdCCPz222+YOnWqvUNqcWPGjMHatWvtHQa1MRMn\nTkR0dHSdr69uqhZLkoUQSE1NrfNd6I3VrEPoj4iWOA+PgrqO2zfffGNxE3bs2BEA7vt2paYkiE88\n8QR27tyJP//5z82yqPyaNWtQU1ODd955B1KptM46PXr0aPYpSa1x/6akpCAoKAgTJkyAUqnE1q1b\nH5iHkJvzXh06dCgA4Oeff25yW/bU3PfhvaZOnYrQ0FAUFhZy+h0R1alZkuSamhrExcWhZ8+eUKlU\n6NixI3x8fBAXF4fp06c3uK0QAhs2bEDv3r2hUCjg7u6OSZMm1fne84sXL6JXr15wdHSESqXCs88+\na/F+9n/961/o06cPXF1doVQq0b9/fxw4cKA5drPBtnv37g2JRIJ27drhqaeeMv6yfvPNN431P//8\ncwC/H6/ly5fDy8sLKpUKAwYMQHJyMgBg3bp1UKvVcHZ2RmFhISIjI9G1a1ecO3fuvvtm7XloqH9r\nbd68GUqlEp06dcK8efOg0WigVCoRGBiIY8eOmdS19hzbci3c68iRI/Dy8oJEIsH7778PAEhISICj\noyPUajV2796NsWPHwsXFBZ6enkhKSjLZvinXb0FBAVQqFXx8fEz2Y/369ejZsycUCgVcXV2xePHi\n+x7Tptq/f/991zOtrq7GwYMH0b59e5OXDTRWW7t/t2/fjsmTJ8PZ2RljxoxBbm4u/vWvf5nUeRTu\nVb1eD8D0db6P2n1ojdq1hNPT023ajogeEebzLxozZ2jNmjXCwcFB7N69W2i1WnHixAnRuXNnMXLk\nSJN6dc1hW758uZDL5WLr1q2ipKREnDlzRjz55JOiY8eO4vr168Z6o0ePFj169BC//fab0Ol04uef\nfxZPP/20UCqVxvUKhRAiNTVVrFy5Uty+fVsUFRWJYcOGiQ4dOjQYg7Uaaluv14vu3bsLLy8vi3kx\nCxcuFBs3bjT++4033hAKhUJ89dVXori4WLz11luiXbt2xrlxb7/9tgAgXn/9dbFlyxYxefJk8csv\nv9x336w9D/fr31phYWHC0dFRnD17VlRVVYns7GwxZMgQ4exsqSe+AAAgAElEQVTsbFx3Uwjrz7G1\n9eo6h1euXBEAxJYtW4xltcfxn//8p7hz544oLCwUzz77rHB0dBTV1dU2HzdzFRUVwtnZWYSHh5uU\nv/3220IikYi///3vori4WGi1WhEfH9+kOcm1nn766Xrn9n7zzTfC2dlZxMTE1Lv9+fPnBQAxbNgw\nm/tu6/dvXl6e8PDwMN5/W7duFQDEX//6V5N6D9u9eu+821q1+7548WKbz9XDch/Wd2zuVVpaKgCI\nbt26NdhHXew9J/lBxPcwUFtV3/3cLEnykCFDxNChQ03K/uu//ku0a9dO3L1711hm/ktVq9UKJycn\nERISYrLtv//9bwHA5D/7uh78OXPmjAAg3njjjXpji4uLEwBEYWFhnTE0hXnbGzduFABESkqKsU5F\nRYXw8vISd+7cEUIIUVlZKdRqtck+a7VaoVAoxKuvviqE+M9/KpWVlTb1b815sKZ/a4WFhVn8B3T8\n+HEBQKxatcrYtjXn2JZrwdb/nO89jrXJ6sWLF41l1l6/5t5++23x+OOPi9LSUmOZVqsVarVaPPfc\ncyZ1m/rgXq2GkmRr/PjjjwKA+OMf/2jztm39/n3nnXfEK6+8Yvz3nTt3hEKhEC4uLiYvIhDi4bpX\nzR/c++qrr0Tnzp1Fp06dRH5+vnG7R+k+rOvY1EcikQg3N7cG69SFSbLtmCRTW1Xf/Vz3hEQbVVVV\nQalUmpTV1NRAJpM1OJE6Ozsb5eXlGDx4sEn5kCFDIJfLLf5sb65///5wdXXFmTNn6q1TOw+yJdaS\nNG97zpw5WLlyJTZt2mRcmHrbtm2YNGkSXFxcAADnzp2DVqtFv379jO2oVCp06dLlvn/OvF//1pyH\n5uy/LoMHD4ZarTa2Ze05buq1YC25XA4A0Ol0xrLGXL+7du1CSkoKvv32Wzg7OxvLL168CK1Wi9Gj\nRzdLvM3NyckJAFBRUVHn5ykpKYiKikJubi4AoFevXvj+++/RqVMni7pt7f7dvn074uLijP92cXHB\nmDFj8PXXX2P37t0m64A+bPfqnTt3IJFI4ODggC5dumDcuHFYsWIFunbtCuDRuw+tVVFRASGE8Zzb\n6ujRo8brh+6v9pXBPGbU1hw9erTOKYjNMid53LhxOHHiBHbv3o3Kykr8+OOPSEtLw5/+9KcGk+SS\nkhIA//mP+15ubm4oKyu7b98ymczkF+3evXsxcuRIeHh4QKFQ4M0332zEHtXtfm07OTnhv/7rv/DD\nDz/g3//+NwDggw8+MFnYujY5Wbp0qcn6nXl5efd96OR+/VtzHprSv7UUCgVu3rwJwPpz3BzXQmPZ\nev3u2LEDa9euRUZGBrp3727yWe1/Ah4eHi0Wb1N4e3tDoVDg4sWLdX4+ffp0/Pbbb/D29kbnzp3x\nyy+/1JkgA23r/v3555+RlZWF8ePHm1zXtev1mq9y8bDdq66urhBCQK/XIz8/H5999hm8vb2Nnz9q\n96G1zp8/D+D3L4NEROaaZSR55cqVOHHiBEJDQ1FeXg6NRoPp06c3+AAR8PsvXQB1/uItKSmBp6dn\ng9vr9Xrcvn0bXl5eAIDLly8jODgYkydPxmeffYbHHnsMW7ZsaZZE2dq2w8PDsWnTJmzcuBHz589H\nt27d4Ovra/y8NnnauHEjIiIimrV/a85DY/u3lk6nMzl31p7jpl4LTWHL9btlyxYcOHAAhw4dqjOR\nqB0Ju3v3bovF2xRKpRJ//OMfsXfv3nq/OVurLd2/X375JWbOnInt27eblBcXF6Nr16749ttvcf36\ndXTp0sX42aN0rz5q96G19u/fDwAYO3Zso7YfNmyYXV9L/aCpfS01jxm1NfX9daNZkuTs7Gzk5OTg\n5s2b9S4pVZd+/frByckJP/74o0n5sWPHUF1djaeeeqrB7b/77jsYDAY8+eSTAICsrCzodDq8+uqr\n6NGjB4DmW5vV2rY9PT0xffp0JCcn4+rVq1ixYoXJ5926dYNSqbT5zWvW9G/NeWhs/9bKyMiAEMKY\nfFl7jpt6LTSFNcdNCIElS5aguLgYaWlp9dbr168f2rVrh++//x7z589vsZibYtWqVfj222+xePFi\nHDp0qNFLs7WV+1cIgR07dmDbtm0Wn7m7u2PatGn44osvsH37dixatMj42aN0rz5q96E1rl+/jo0b\nN8LT0xOzZ89udDtE9PBqlukWCxYsgJeXl83vVVcqlYiMjMSuXbuwbds2lJaWIisrC/Pnz4dGo0FY\nWJhJ/erqaty5cwd6vR4nT55EeHg4vL29jcv41I5IHTx4EFVVVbhw4UKzzaGzpe3IyEjo9XoUFxfj\nD3/4g8U+v/LKK0hKSkJCQgJKS0tRU1OD/Px8XLt2rUn9W3MeGtt/fQwGA4qLi6HX63HmzBlERETA\ny8vLeE6sPce2XgvNyZrjdvbsWaxbtw4ff/wxZDKZxetu33vvPQC/j/5NmTIFX331FT799FOUlpbi\nzJkzrbJOdXp6+n2XgAOAp556Clu3bsWJEycwcuRI7N+/H9euXYNer0deXh62bt2K27dv37e/tnL/\n/vDDD3BxccHw4cPr/Lz2y0pdLxZ5VO7VR+0+vJcQAuXl5TAYDBBC4ObNm0hOTsbw4cPh4OCAtLS0\nRs9JJqKHnPmTfI15+vTQoUOiQ4cOAoDxRyaTid69e4udO3cKIYT4+9//Ljp37iwACEdHRzF58mQh\nxO+vCl2/fr3w9/cXMplMuLu7i+DgYHHu3DmTPhITE8WoUaNEp06dhFQqFR06dBAzZ84UeXl5JvWi\noqJE+/bthZubm5g2bZp4//33BQDh6+srIiIi6ozBWg21fe+SZ0IIMWrUKPHJJ5/U2c7du3dFVFSU\n8PLyElKpVHh4eIgpU6aI7Oxs8e677wqVSmVclmjr1q1W92/Nebhf/7YICwsTMplMdO3aVUilUuHi\n4iImTZokcnJyTOpZe46tqVfXdbRlyxbRpUsXAUCo1WoxYcIEER8fL9RqtQAg/P39RU5Ojvjoo4+E\ni4uLACC8vb2NS49Zc9yysrJMPjf/Wb9+vTHGsrIyMWfOHNGhQwfh5OQknnnmGbF8+XIBQHh6eoqf\nfvrJpuOcmZkphg8fLjQajbG/Ll26iMDAQPH9998b6+3bt084OzuL2NhYq9r97bffREREhOjbt69w\ndHQUSqVS+Pj4iGeffVYsWbJEHD58uMHjbsu5ban7969//atwdHQUUqlUPPHEE+LkyZMm7a1evdrk\nuHXt2lXEx8eb1HlQ79X//d//FY8//rhxe41GI6ZNm1bv+X6U7sM9e/aIAQMGCLVaLeRyuWjXrp0A\nYFzJYujQoSImJkYUFRXVe7zuh6tb2I6rW1Bb1aJLwMXHx4uIiAiTsrt374qFCxcKhUJhsfwStYzW\nPg9hYWGiffv2zdqmPfD6pdbGa87Sg3ZMmCTbjkkytVUttgTc9evXER4ebjFvTi6Xw8vLCzqdDjqd\nDiqVqqldUQPsdR5aYmm91sTrl1obrzlLPCZE1BY1eU6ySqWCTCbDp59+ihs3bkCn0+Hq1av45JNP\nsHz5coSEhLTZ+V6//vqrxZy2un7uXV+1rWqO8/AwHQ9r2eP6fRSPM/3Hg/w7s6XwmJA9HDx4ENHR\n0di5cyd69Ohh/N370ksvWdQdM2YMnJ2d4eDggL59++LkyZN2iNh6MTEx6NOnD1xcXKBQKODn54c3\n33zTYs5/bGxsnf//3Ls+ey2dToe4uDj4+flBLpfDzc0N/fr1M66rX+vIkSMYPnw41Go1NBoNoqKi\nTFZ82rNnD959990HY5DNfGi5MX8OOXz4sPjjH/8oXFxchIODg3B1dRWBgYEiPj5e6HS6xo9/k01a\n8zxER0cLuVwuAIju3buL1NTUZm2/NfH6pdbGa87Sg3ZMON3Cdm1pusXy5cvF+PHjTd7U6Ovra5wX\n/80331hsk56eLiZOnNiaYTbaiBEjRHx8vCgqKhKlpaUiOTlZyGQy8cILL5jUW716dZ3z+/v27WvR\nZnBwsOjZs6c4evSo0Ol04urVq2LChAkiKyvLWOfnn38WKpVKLFu2TJSXl4sffvhBdOzY0eRtqEII\nsWnTJjFixAhRXFzcMgfARi06J5mIiOhR0haSZK1WKwICAh6YPtpKfvHOO++Ixx9/3OJ18r6+vuLL\nL78U7dq1E127dhUlJSUmnz9ISXJQUJDQ6/UmZdOnTxcATBYaWL16tclDx/VJSkoSEolEnDlzpsF6\nM2bMED4+PsJgMBjL1q9fLyQSifjll19M6oaHh4uAgIA28SW4vvu5WZaAIyIiotb16aeforCw8IHv\nozVdvHgRy5Ytw6pVqyxegw4AgYGBiIiIQEFBAd544w07RNg8vvnmG4s3VXbs2BEAGvV23Q8++ABP\nPvkk+vfvX28dvV6PvXv3YsSIESZrw48dOxZCCOzevduk/sqVK3H69Gls2rTJ5nhaC5NkIiKiViCE\nwIYNG9C7d28oFAq4u7tj0qRJ+PXXX411wsPDIZfLTd4O+be//Q2Ojo6QSCS4desWACAiIgKRkZHI\nycmBRCKBn58fNm/eDKVSiU6dOmHevHnQaDRQKpUIDAw0Wau7KX0Av7+p0Jo12duizZs3QwiBCRMm\n1FsnNjYWjz/+OD755BMcPHiwwfasOacJCQlwdHSEWq3G7t27MXbsWLi4uMDT0xNJSUkm7dXU1GD5\n8uXw8vKCSqXCgAEDkJyc3LSd/j8FBQVQqVTw8fGxabvq6mocPXoUAwcObLDepUuXUF5eblwrvlbt\nm0zPnDljUu7u7o4RI0Zg06ZNEELYFFNrYZJMRETUClauXIno6Gi8/fbbKCwsxOHDh3HlyhU8++yz\nuHHjBoDfk7jp06ebbBcfH49Vq1aZlG3atAnjx4+Hr68vhBC4ePEiwsPDERoaCq1Wi9dffx25ubk4\nefIk9Ho9nnvuOVy5cqXJfQD/WdXIYDA038FpJXv37kXPnj2hVqvrraNSqfD555+jXbt2mDt3Lioq\nKuqta805ffXVV7Fw4UJUVlbC2dkZycnJyMnJQY8ePTB37lzodDpje0uWLMG6deuwceNGXLt2DePH\nj8eLL75o8RZMW2m1Whw6dAhz586FXC43+Sw6Ohru7u6Qy+Xw8fHBpEmTcPz4cePnV69eRXV1NU6c\nOIFRo0YZv3z17t0b8fHxxgT3+vXrAABnZ2eT9pVKJVQqlfF43GvQoEEoKCjATz/91KT9aylMkomI\niFpYZWUlNmzYgMmTJ2PWrFlwdXVF//798eGHH+LWrVvN+lZOqVRqHNns06cPEhISUFZWhsTExGZp\nPygoCKWlpVi2bFmztNdaKioq8NtvvxlHNhsSEBCAhQsXIjc3F0uWLKmzTmPOaWBgIFxcXODh4YGQ\nkBBUVFTg8uXLAICqqiokJCQgODgYU6ZMgZubG5YuXQqZTNbkcxcXFweNRoPY2FiT8pdffhl79uzB\nlStXUF5ejqSkJFy+fBkjRoxAdnY2ABhXxPDw8MCaNWuQnZ2NGzduYNKkSViwYAG2b98OAMYVLMyn\neQCATCZDZWWlRbm/vz8AICsrq0n711KYJBMREbWw7OxslJeXY/DgwSblQ4YMgVwut+oV7I01ePBg\nqNVqkykAj6LCwkIIIRocRb5XbGwsevbsifj4eBw5csTi86ae09oR3dqR5HPnzkGr1Zosv6ZSqdCl\nS5cmnbtdu3YhJSUFBw4csBjl7datGwYNGgQnJyfI5XIMGzYMiYmJqKysRHx8PABAoVAAAPr27YvA\nwEC0b98erq6uWLVqFVxdXY1fBmrneOv1eosYqqur61znvPZc1DXK3BYwSSYiImphJSUlAAAnJyeL\nz9zc3FBWVtai/SsUCty8ebNF+2jrqqqqAPwn6bsfpVKJxMRESCQSzJ4922IktLnPae20jqVLl5qs\nWZyXl9eoh+0AYMeOHVi7di0yMjLQvXt3q7bp378/HBwccP78eQCARqMBAONc9VpyuRze3t7IyckB\nAOMc99LSUpN6Wq0WVVVVxnbuVZs4156btoZJMhERUQtzc3MDgDoTp5KSEnh6erZY3zqdrsX7eBDU\nJmS2vMQiICAAixYtwoULF7B69WqTz5r7nHp4eAAANm7cCPH7Er3Gn8zMTJvaAoAtW7Zg27ZtOHTo\nEB577DGrtzMYDDAYDMYvE05OTvD398fZs2ct6ur1eri6ugIAfHx84OzsjLy8PJM6tXPZBwwYYLF9\ndXU1ALTZt2kySSYiImph/fr1g5OTk8UDWMeOHUN1dTWeeuopY5lUKjV5mKupMjIyIITAsGHDWqyP\nB0GnTp0gkUhw584dm7ZbvXo1evXqhVOnTpmU23JOrdGtWzcolUqL17PbSgiBqKgoZGVlIS0trc6R\n7lrPP/+8Rdnx48chhEBAQICxbMaMGTh16hQuXbpkLNNqtcjLyzMuCyeVSjFu3DgcPnzY5KHO9PR0\nSCSSOlcUqT0XnTt3tn1HWwGTZCIiohamVCoRGRmJXbt2Ydu2bSgtLUVWVhbmz58PjUaDsLAwY10/\nPz/cvn0baWlp0Ol0uHnzpsXoHAC0b98eV69eRW5uLsrKyoxJr8FgQHFxMfR6Pc6cOYOIiAh4eXkh\nNDS0WfpIT09/IJeAU6vV6NGjB/Lz823arnbahfkDabacU2v7eeWVV5CUlISEhASUlpaipqYG+fn5\nuHbtGgAgJCQEnTt3bvC12GfPnsW6devw8ccfQyaTWbxy+r333jPWLSgowI4dO1BSUgKdTofMzEzM\nmTMHXl5emD9/vrHeokWL4O3tjdDQUFy+fBlFRUWIiopCZWWlyYONy5Ytw40bN7BixQpUVFQgMzMT\n69evR2hoKHr27GkRa+25aGj9Zbsyf7tIW3kjDhERUVvVmDfuGQwGsX79euHv7y9kMplwd3cXwcHB\n4ty5cyb1ioqKxKhRo4RSqRQ+Pj7itddeE4sXLxYAhJ+fn/GNaSdPnhTe3t5CpVKJZ555Rly/fl2E\nhYUJmUwmunbtKqRSqXBxcRGTJk0SOTk5zdbHvn37hLOzs4iNjbVp/9tCfhEeHi5kMpnQarXGsl27\ndglfX18BQHTs2FEsWLCgzm0XL15s8cY9a85pfHy8UKvVAoDw9/cXOTk54qOPPhIuLi4CgPD29hbn\nz58XQghx9+5dERUVJby8vIRUKhUeHh5iypQpIjs7Wwjx+6uhAYjly5fXu49ZWVl1vmq69mf9+vXG\nupGRkcLX11c4OjoKqVQqPD09xdy5c8XVq1ct2r1y5YqYOXOmcHd3FwqFQgwdOlSkp6db1Pv+++/F\n0KFDhUKhEBqNRixevFhUVVXVGWtQUJDo2rWryRv67KG++1kihOkKzikpKZgxY0abXdiZiIjI3qZN\nmwYASE1NtXMkpubNm4fU1FQUFRXZOxQLbSG/uHjxInr37o3ExETMmjXLbnE0lsFgwMiRIxEaGorZ\ns2fbO5wmKSoqgqenJ2JjYxEZGWnXWOq7nzndgoiI6CFiy4Npjxo/Pz/ExMQgJibGuP7vg6KmpgZp\naWkoKytDSEiIvcNpspUrV2LgwIEIDw+3dyj1YpJMREREj4zo6GhMmzYNISEhNj/EZ08ZGRnYuXMn\n0tPTrV7rua3asGEDTp8+jX379kEmk9k7nHoxSSYiInoIvPXWW0hMTMSdO3fg4+ODr776yt4htVlr\n1qxBeHg43nnnHXuHYrXRo0fjyy+/NK5H/KDavXs37t69i4yMDLi7u9s7nAZJ7R0AERERNV1cXBzi\n4uLsHcYDY8yYMRgzZoy9w3jkTJw4ERMnTrR3GFbhSDIRERERkRkmyUREREREZpgkExERERGZYZJM\nRERERGSGSTIRERERkZl6V7eQSCStGQcREdEDh/9X2o7HjNqiqVOnWpRZvJY6Pz8fP/zwQ6sFRURE\nv5sxYwYiIiIQEBBg71CIiB4p3bp1s/jda5EkExGRfUgkEiQnJ2P69On2DoWI6JHHOclERERERGaY\nJBMRERERmWGSTERERERkhkkyEREREZEZJslERERERGaYJBMRERERmWGSTERERERkhkkyEREREZEZ\nJslERERERGaYJBMRERERmWGSTERERERkhkkyEREREZEZJslERERERGaYJBMRERERmWGSTERERERk\nhkkyEREREZEZJslERERERGaYJBMRERERmWGSTERERERkhkkyEREREZEZJslERERERGaYJBMRERER\nmWGSTERERERkhkkyEREREZEZJslERERERGaYJBMRERERmWGSTERERERkhkkyEREREZEZJslERERE\nRGaYJBMRERERmWGSTERERERkhkkyEREREZEZqb0DICJ6FCUlJaGsrMyi/ODBgygpKTEpCw4OhoeH\nR2uFRkREACRCCGHvIIiIHjWhoaH4n//5H8hkMmNZ7a9jiUQCAKipqYGTkxMKCwuhUCjsEicR0aOK\n0y2IiOxg5syZAACdTmf80ev10Ov1xn87ODhg2rRpTJCJiOyAI8lERHag1+vRuXNn3L59u8F6//zn\nP/GHP/yhlaIiIqJaHEkmIrIDqVSKmTNnmky3MNexY0eMGDGiFaMiIqJaTJKJiOxk5syZ0Ol0dX4m\nk8nw0ksvwcHBoZWjIiIigNMtiIjsRggBLy8v5Ofn1/n5v//9bwwZMqSVoyIiIoAjyUREdiORSDBr\n1qw6p1x069YNgwcPtkNUREQEMEkmIrKruqZcyGQyhIaGGpeCIyKi1sfpFkREdtarVy+cO3fOpOzn\nn39G37597RQRERFxJJmIyM5eeuklkykXffr0YYJMRGRnTJKJiOxs1qxZ0Ov1AH6favHyyy/bOSIi\nIuJ0CyKiNmDw4ME4ceIEJBIJcnNz4eXlZe+QiIgeaRxJJiJqA/7yl78AAJ5++mkmyEREbYDU3gEQ\nPQgyMzOxYcMGe4dBD7GqqipIJBLcvXsX06ZNs3c49BALCAjAokWL7B0GUZvHkWQiK1y5cgVfffWV\nvcOgOhw9ehRHjx61dxhNplQq0blzZ3h6erZ4X/n5+byeH1FHjx5FZmamvcMgeiBwJJnIBqmpqfYO\ngczUjro+DOfm4sWL8PPza/F+UlJSMGPGjIfimJFt+FcKIutxJJmIqI1ojQSZiIiswySZiIiIiMgM\nk2QiIiIiIjNMkomIiIiIzDBJJiIiIiIywySZiAjAvn374Orqiq+//treobR5Bw8eRHR0NHbu3Ike\nPXpAIpFAIpHgpZdesqg7ZswYODs7w8HBAX379sXJkyftELH1YmJi0KdPH7i4uEChUMDPzw9vvvkm\nysvLTerFxsYa9/ven379+lm0qdPpEBcXBz8/P8jlcri5uaFfv37Izc01qXfkyBEMHz4carUaGo0G\nUVFRuHv3rvHzPXv24N1330VNTU2L7DsRmWKSTEQEQAhh7xAeCCtWrMDmzZvx1ltvYcqUKbh06RJ8\nfX3RoUMHbNu2DXv37jWp/+233yI1NRXjx49HdnY2nnzySTtFbp1Dhw5hwYIFyM3Nxa1btxAXF4dN\nmzY1aem0GTNm4IsvvsCXX34JrVaLX375Bb6+viaJd3Z2NsaMGYPRo0fj5s2b2LVrFz777DPMnz/f\nWGfChAlQKpUYPXo0SkpKmrSfRHR/TJKJiAAEBQXhzp07GD9+vL1DQWVlJQIDA+0dhoW1a9dix44d\nSElJgbOzs8lnmzdvRrt27RAWFoY7d+7YKcKmc3JyQlhYGNq3bw9nZ2dMnz4dwcHB2L9/P65cuWJS\nd+vWrRBCmPz8/PPPJnV27NiBtLQ0pKam4umnn4ZUKoVGo8Hu3btNRp1Xr16NLl26YNWqVXB0dERA\nQACioqLw+eef49dffzXWe/311/HEE09g3Lhx0Ov1LXswiB5xTJKJiNqYTz/9FIWFhfYOw8TFixex\nbNkyrFq1Ckql0uLzwMBAREREoKCgAG+88YYdImwe33zzDRwcHEzKOnbsCADQarU2t/fBBx/gySef\nRP/+/euto9frsXfvXowYMQISicRYPnbsWAghsHv3bpP6K1euxOnTp7Fp0yab4yEi6zFJJqJH3pEj\nR+Dl5QWJRIL3338fAJCQkABHR0eo1Wrs3r0bY8eOhYuLCzw9PZGUlGTcdvPmzVAqlejUqRPmzZsH\njUYDpVKJwMBAHDt2zFgvPDwccrkcXbp0MZb97W9/g6OjIyQSCW7dugUAiIiIQGRkJHJyciCRSIwv\nGNm/fz9cXFywZs2a1jgkFjZv3gwhBCZMmFBvndjYWDz++OP45JNPcPDgwQbbE0Jgw4YN6N27NxQK\nBdzd3TFp0iSTUVNrzwEA1NTUYPny5fDy8oJKpcKAAQOQnJzctJ3+PwUFBVCpVPDx8bFpu+rqahw9\nehQDBw5ssN6lS5dQXl4OLy8vk3JfX18AwJkzZ0zK3d3dMWLECGzatInThIhaEJNkInrkPfPMM/jh\nhx9Myl599VUsXLgQlZWVcHZ2RnJyMnJyctCjRw/MnTsXOp0OwO/Jb2hoKLRaLV5//XXk5ubi5MmT\n0Ov1eO6554x/ot+8eTOmT59u0kd8fDxWrVplUrZp0yaMHz8evr6+EELg4sWLAGB8WMtgMLTIMbif\nvXv3omfPnlCr1fXWUalU+Pzzz9GuXTvMnTsXFRUV9dZduXIloqOj8fbbb6OwsBCHDx/GlStX8Oyz\nz+LGjRsArD8HALBkyRKsW7cOGzduxLVr1zB+/Hi8+OKL+PHHH5u031qtFocOHcLcuXMhl8tNPouO\njoa7uzvkcjl8fHwwadIkHD9+3Pj51atXUV1djRMnTmDUqFHGL1C9e/dGfHy8McG9fv06AFhMYVEq\nlVCpVMbjca9BgwahoKAAP/30U5P2j4jqxySZiOg+AgMD4eLiAg8PD4SEhKCiogKXL182qSOVSo2j\non369EFCQgLKysqQmJjYLDEEBQWhtLQUy5Yta5b2bFFRUYHffvvNOLLZkICAACxcuBC5ublYsmRJ\nnXUqKyuxYcMGTJ48GbNmzYKrqyv69++PDz/8ELdu3Zl+nSMAACAASURBVMJHH31ksU1D56CqqgoJ\nCQkIDg7GlClT4ObmhqVLl0ImkzX5+MfFxUGj0SA2Ntak/OWXX8aePXtw5coVlJeXIykpCZcvX8aI\nESOQnZ0NAMYH8zw8PLBmzRpkZ2fjxo0bmDRpEhYsWIDt27cDgHEFC/NpHgAgk8lQWVlpUe7v7w8A\nyMrKatL+EVH9mCQTEdmgdjTx3lHMugwePBhqtdpk+sCDqrCwEEKIBkeR7xUbG4uePXsiPj4eR44c\nsfg8Ozsb5eXlGDx4sEn5kCFDIJfLTaap1MX8HJw7dw5ardbkQTiVSoUuXbo06fjv2rULKSkpOHDg\ngMUob7du3TBo0CA4OTlBLpdj2LBhSExMRGVlJeLj4wEACoUCANC3b18EBgaiffv2cHV1xapVq+Dq\n6mr8MlA7x7uuB/Gqq6uhUqksymvPRV2jzETUPJgkExG1EIVCgZs3b9o7jCarqqoC8J+k736USiUS\nExMhkUgwe/Zsi5HQ2uXLnJycLLZ1c3NDWVmZTfHVTutYunSpyZrFeXl5jXrYDvh9VYq1a9ciIyMD\n3bt3t2qb/v37w8HBAefPnwcAaDQaADDON68ll8vh7e2NnJwcADDOUy8tLTWpp9VqUVVVZWznXrWJ\nc+25IaLmxySZiKgF6HQ6lJSUwNPT096hNFltQmbLSywCAgKwaNEiXLhwAatXrzb5zM3NDQDqTIYb\nc8w8PDwAABs3brRYki0zM9OmtgBgy5Yt2LZtGw4dOoTHHnvM6u0MBgMMBoPxy4STkxP8/f1x9uxZ\ni7p6vR6urq4AAB8fHzg7OyMvL8+kTu189AEDBlhsX11dDQB1jjITUfNgkkxE1AIyMjIghMCwYcOM\nZVKp9L7TNNqiTp06QSKR2Lz+8erVq9GrVy+cOnXKpLxfv35wcnKyeKju2LFjqK6uxlNPPWVTP926\ndYNSqcTp06dt2s6cEAJRUVHIyspCWlpanSPdtZ5//nmLsuPHj0MIgYCAAGPZjBkzcOrUKVy6dMlY\nptVqkZeXZ1wWTiqVYty4cTh8+LDJg5np6emQSCR1rihSey46d+5s+44SkVWYJBMRNQODwYDi4mLo\n9XqcOXMGERER8PLyQmhoqLGOn58fbt++jbS0NOh0Oty8edNi9BAA2rdvj6tXryI3NxdlZWXQ6XRI\nT0+32xJwarUaPXr0QH5+vk3b1U67MH8gTalUIjIyErt27cK2bdtQWlqKrKwszJ8/HxqNBmFhYTb3\n88orryApKQkJCQkoLS1FTU0N8vPzce3aNQBASEgIOnfu3OBrsc+ePYt169bh448/hkwms3jl9Hvv\nvWesW1BQgB07dqCkpAQ6nQ6ZmZmYM2cOvLy8TN6St2jRInh7eyM0NBSXL19GUVERoqKiUFlZafJg\n47Jly3Djxg2sWLECFRUVyMzMxPr16xEaGoqePXtaxFp7Lhpaf5mImoZJMhE98t5//30MGTIEABAV\nFYWJEyciISEBGzduBPD7n7svXbqEjz/+GJGRkQCAF154ARcuXDC2UVVVhf79+0OlUuHZZ5/F448/\nju+++85kHu+rr76KUaNGYebMmejZsydWr15t/HN5QECAcbm4+fPno1OnTujTpw/GjRuH27dvt8px\naEhQUBCys7NN5hf/4x//gJ+fH3JycjBkyBC89tprFtsNGzYMixYtsihfsWIF4uLiEBMTg44dO2LE\niBHo3r07MjIy4OjoCAA2nYNNmzZh4cKFePfdd9GhQwdoNBpERESguLgYwO/TEwoLCy1ezHEvW9Yc\nfuGFF7B06VJ4enpCrVZj+vTpGD58OI4ePYoOHToY67m7u+Nf//oXPD09MXDgQHTt2hX//ve/sXfv\nXpP1k/v27YsDBw7g22+/RYcOHTBlyhTMnj0bH3zwQZ39Hz9+HF27dq1zKgYRNQ+J4ErkRPeVkpKC\nGTNmcOH+NmjatGkAgNTUVLvFMG/ePKSmpqKoqMhuMdiiMdfzxYsX0bt3byQmJmLWrFktGF3LMBgM\nGDlyJEJDQzF79mx7h9MkRUVF8PT0RGxsrPELg7Xawv1C9KDgSDIRUTOw5aG2B5Gfnx9iYmIQExNj\nXP/3QVFTU4O0tDSUlZUhJCTE3uE02cqVKzFw4ECEh4fbOxSihxqTZCIiskp0dPT/Z+/Oo6K48j2A\nf1ug6WZfBG1tUaDdBfcJoI76fHFjcBfQIQnx6UONQZBMFINREYmaDHBIIJ6YHDJRo4AYMHFJnscw\niWfUmChqMG4gqOiIoggCytL1/nDopLtRu2Uplu/nHP+YW7eqfl0VnW9X37oXc+fORWBgoNEv8Ykp\nOzsbGRkZOHjwoMFzPbdWcXFxyMnJwYEDB2BmZiZ2OUTtGkMyUTN6/Pgxli9fjq5du8LCwgKHDh0y\neF+1Wo34+Hj4+Pi88PkzMjLg5uam9wKSVCqFs7Mzxo0bh/fff18zbpOMt3r1aqSkpODBgwdwdXXF\nnj17xC6pWW3cuBGhoaF47733xC7FYBMmTMDOnTs18xG3VVlZWXj8+DGys7Nhb28vdjlE7R5DMlEz\n+vvf/45Dhw7hwoULSEhIMPhn6suXL+PPf/4zVqxY8cKLIQDA7NmzkZ+fD3d3d9ja2kIQBKjVahQX\nFyMtLQ2urq5YuXIlBg4cqDcdFxkmNjYWjx8/hiAIuHr1KubMmSN2Sc1u4sSJ2LRpk9hldDjTp09H\nZGRkg8tXE1HTMxW7AKL2LDMzEyNGjICdnR3+93//16B9zpw5g+joaCxZsgQVFRVN/rKgRCKBnZ0d\nxo0bh3HjxsHX1xcBAQHw9fXFpUuXNAscEBERdWR8kkzUjG7cuGH0uMHBgwcjIyMDf/3rXw1eBrgx\n5syZg+DgYBQXF2Pr1q3Nfj4iIqK2gCGZqBn83//9H1QqFW7duoV//OMfkEgkWqt3bd++HSNGjIBM\nJoOlpSV69eqlt3Tv8xw6dKjJFpeoX/Di4MGDmra6ujq8++67cHFxgVwuh6enJ1JTUwE8mb/W0tIS\nFhYWyMrKwpQpU2BjYwOlUoldu3ZpHfuf//wn/vSnP8HCwgI2Njbw8PBAWVnZc89BREQkJoZkombw\n8ssv48qVK+jSpQtee+01CIKgGY+ckJCAV199FXPmzMHNmzdx48YNrF69GhcvXjTqHPVTjv1xGdsX\nVb+owR+Xzl21ahW2bNmC+Ph43Lp1C35+fpg/fz5+/vlnLF26FOHh4aiqqoK1tTVSU1ORl5cHNzc3\nLFq0SLP0ckVFBaZNm4Y5c+bg3r17uHz5Mvr06YPq6urnnoOIiEhMDMlELaimpgbr16/H+PHjsWrV\nKjg4OMDe3h7/8z//o1nxzVC+vr4oKyvDmjVrGl2XtbU1JBIJysvLATxZPS45ORkzZ87E7NmzYWdn\nh6ioKJiZmSElJUVrXx8fH9jY2MDJyQmBgYGoqKjAtWvXAAAFBQUoKyvDwIEDIZPJ0KVLF2RkZKBz\n585GnYOIiKil8cU9ohZ09uxZlJaWYtKkSVrtJiYmWL58uUhVQfOCoI2NDQDg4sWLqKysxKBBgzR9\n5HI5unbtigsXLjz1OFKpFAA0T5Ld3Nzg7OyMoKAgLF++HMHBwejVq1ejztGQPXv2QCKRGLUPgdes\ng+oIM7AQNQWGZKIWVD8W187OTuRKtF26dAkA0K9fPwBPQjMAREVFISoqSquvQqEw+LhyuRxHjhzB\nqlWrsHHjRkRHR8Pf3x8pKSlNdg4A8PLyQnh4uFH7dGTHjh1DQkICx393QPHx8WKXQNRmMCQTtaBu\n3boBAO7evStyJdrqFzmZMmUKAMDJyQnAk/9DDQsLa9SxBw4ciK+//hp37txBXFwcNm3ahIEDB2qW\nB26KcyiVSvj7+zfqGB1NQkICr1kHlJ6eLnYJRG0GxyQTtaBevXrBwcEB3333ndilaPz73/9GfHw8\nlEolFixYAADo0aMHZDIZcnJyGnXsmzdv4vz58wCeBO/33nsPw4YNw/nz55vsHERERM2BIZmoBZmb\nm2P16tX44YcfEBoaiqKiIqjVapSXl2vCpKEOHjxo1BRw9TNsqNVqCIKAO3fuIDU1FaNGjYKJiQky\nMzM1Y5JlMhlef/117Nq1C8nJySgrK0NdXR1u3LiBW7duGVzjzZs3sXjxYly4cAHV1dU4ffo0CgsL\n4eXl1WTnICIiag4MyUTNoLCwEMOGDcPt27exc+dODB8+HBkZGQCAiIgIfPTRR8jOzoZKpYKlpSXG\njh2L7OxsAMDx48cxevRodOvWDSdOnMCZM2egUCgwatQo/PDDD0bV8fXXX2Pw4MG4desWHj16BFtb\nW5iYmMDExAR9+vRBXFwcgoODkZubi+HDh2vtm5CQgPDwcGzevBmOjo5QKBQICwvD/fv3kZycrBnb\n6Onpifz8fGzbtg0REREAgMmTJ+Py5ctwcnJCXV0dfHx8YGFhgb/85S9YvHgxli1b9txzEBERiUki\nNPWat0TtUFpaGgICApp8iWhqvLlz5wLgWEtj8L/njot/X4gMxyfJREREREQ6GJKJiKjRDh8+jMjI\nSGRkZMDNzQ0SiQQSiQSvvPKKXt+JEyfC2toaJiYmGDhwIE6dOiVCxYaLjo7GgAEDYGNjA3Nzc6hU\nKrz99tuaVTTrxcTEaD73H//8cS7wejU1NYiNjYVKpYJUKoWdnR0GDRqEgoICAMC+ffuwefNmzcqa\nRNTyGJKJiKhR1q5di8TERKxevRqzZ89Gfn4+3N3d4ejoiB07dmD//v1a/b/77jukp6fDz88Pubm5\nGDZsmEiVG+bIkSNYtmwZCgoKcPfuXcTGxiIhIUEzdOFFBAQE4IsvvsDOnTtRWVmJ3377De7u7prg\nPW3aNMhkMkyYMAGlpaVN9VGIyAgMyUREjVRVVQUfH582f44XsWnTJuzevRtpaWmwtrbW2paYmIhO\nnTohJCQEDx48EKnCxrOyskJISAgcHBxgbW0Nf39/zJw5E4cOHcL169e1+m7fvh2CIGj9+fXXX7X6\n7N69G5mZmUhPT8dLL70EU1NTKBQKZGVlaT11Xr58OQYPHoypU6eitra2RT4rEf2OIZmIqJE+++wz\nFBcXt/lzGOvKlStYs2YN1q9fD5lMprfdx8cHYWFhKCoqwltvvSVChU3jm2++gYmJiVZb586dAQCV\nlZVGH+/jjz/GsGHD4OHh8dy+69atQ05ODhISEow+DxE1DkMyEXU4giAgLi4O/fv3h7m5Oezt7TFj\nxgxcuHBB0yc0NBRSqRRdu3bVtL3xxhuwtLSERCLRrJoYFhaGiIgI5OXlQSKRQKVSITExETKZDM7O\nzli8eDEUCgVkMhl8fHxw4sSJJjkH8GSlRGPmym5qiYmJEAQB06ZNe2qfmJgY9OnTB59++ikOHz78\nzOMZcl+Sk5NhaWkJCwsLZGVlYcqUKbCxsYFSqcSuXbu0jldXV4d3330XLi4ukMvl8PT0bLKluIuK\niiCXy+Hq6mrUftXV1Th+/DiGDBliUH97e3uMHTsWCQkJnI2EqIUxJBNRh7Nu3TpERkbinXfeQXFx\nMX744Qdcv34dY8aMwe3btwE8CYC6yzYnJSVh/fr1Wm0JCQnw8/ODu7s7BEHAlStXEBoaiuDgYFRW\nVmL58uUoKCjAqVOnUFtbi5dfflnzE31jzgFA81KXWq1uuotjhP3796Nv376wsLB4ah+5XI7PP/8c\nnTp1wqJFi1BRUfHUvobcl6VLlyI8PBxVVVWwtrZGamoq8vLy4ObmhkWLFqGmpkZzvFWrVmHLli2I\nj4/HrVu34Ofnh/nz5+Pnn39u1OeurKzEkSNHsGjRIkilUq1tkZGRsLe3h1QqhaurK2bMmIGTJ09q\ntt+8eRPV1dX45ZdfMH78eM0XqP79+yMpKanBIDx06FAUFRXhzJkzjaqbiIzDkExEHUpVVRXi4uIw\na9YsBAUFwdbWFh4eHti6dSvu3r2LTz75pMnOZWpqqnkqOmDAACQnJ6O8vBwpKSlNcnxfX1+UlZVh\nzZo1TXI8Y1RUVODq1atwd3d/bl9vb2+Eh4ejoKAAq1atarDPi9wXHx8f2NjYwMnJCYGBgaioqMC1\na9cAAI8ePUJycjJmzpyJ2bNnw87ODlFRUTAzM2v09Y+NjYVCoUBMTIxW+2uvvYZ9+/bh+vXrePjw\nIXbt2oVr165h7NixyM3NBQDNi3lOTk7YuHEjcnNzcfv2bcyYMQPLli3Dl19+qXe+3r17AwDOnTvX\nqLqJyDgMyUTUoeTm5uLhw4cYMWKEVvvIkSMhlUq1hkM0tREjRsDCwkJr+EBbVVxcDEEQnvkU+Y9i\nYmLQt29fJCUl4ejRo3rbG3tf6p/o1j9JvnjxIiorK7VehJPL5ejatWujrv/evXuRlpaGb7/9Vu9F\nxR49emDo0KGwsrKCVCqFl5cXUlJSUFVVhaSkJABPlqYHgIEDB8LHxwcODg6wtbXF+vXrYWtr2+CX\ngfprXP80nYhaBkMyEXUo9dNpWVlZ6W2zs7NDeXl5s57f3Nwcd+7cadZztIRHjx4B+D30PY9MJkNK\nSgokEgkWLFiAqqoqre1NfV/qh3VERUVpzVlcWFj4Qi/bAU9mpdi0aROys7PRq1cvg/bx8PCAiYkJ\nLl26BABQKBQAoBlvXk8qlaJnz57Iy8vTO4ZcLgfw+zUnopbBkExEHYqdnR0ANBi6SktLoVQqm+3c\nNTU1zX6OllIf3IxZ7MLb2xsrVqzA5cuXsWHDBq1tTX1fnJycAADx8fF6U7IdO3bMqGMBwIcffogd\nO3bgyJEj6Natm8H7qdVqqNVqzZcJKysr9O7dG+fPn9frW1tbC1tbW7326upqAL9fcyJqGQzJRNSh\nDBo0CFZWVnovb504cQLV1dUYPny4ps3U1FTrRbDGys7OhiAI8PLyarZztBRnZ2dIJBKj5z/esGED\n+vXrh9OnT2u1G3NfDNGjRw/IZDLk5OQYtZ8uQRCwcuVKnDt3DpmZmQ0+6a43adIkvbaTJ09CEAR4\ne3tr2gICAnD69Gnk5+dr2iorK1FYWNjgtHD117hLly6N+ShEZCSGZCLqUGQyGSIiIrB3717s2LED\nZWVlOHfuHJYsWQKFQoGQkBBNX5VKhXv37iEzMxM1NTW4c+cOCgsL9Y7p4OCAmzdvoqCgAOXl5ZrQ\nq1arcf/+fdTW1uLs2bMICwuDi4sLgoODm+QcBw8eFG0KOAsLC7i5ueHGjRtG7Vc/7EJ33mFj7ouh\n53n99dexa9cuJCcno6ysDHV1dbhx4wZu3boFAAgMDESXLl2euSz2+fPnsWXLFmzbtg1mZmZ6S05/\n8MEHmr5FRUXYvXs3SktLUVNTg2PHjmHhwoVwcXHBkiVLNP1WrFiBnj17Ijg4GNeuXUNJSQlWrlyJ\nqqqqBl9srL/GhsyrTERNhyGZiDqctWvXIjY2FtHR0ejcuTPGjh2LXr16ITs7G5aWlpp+S5cuxfjx\n4zFv3jz07dsXGzZs0Pzk7e3trZnKbcmSJXB2dsaAAQMwdepU3Lt3D8CTMaQeHh6Qy+UYM2YM+vTp\ng++//15rHG9jzyEmX19f5Obmao0v/uqrr6BSqZCXl4eRI0fizTff1NvPy8sLK1as0Gs35L4kJycj\nPj4eAODp6Yn8/Hxs27YNERERAIDJkyfj8uXLAJ5MnRceHo7NmzfD0dERCoUCYWFhuH//PoAnwxiK\ni4uRlZX11M9ozNzEkydPRlRUFJRKJSwsLODv749Ro0bh+PHjcHR01PSzt7fHjz/+CKVSiSFDhqB7\n9+746aefsH///gbnTz558iS6d+8OT09Pg2shosaTCJydnOi50tLSEBAQwMn8W6G5c+cCANLT00Wu\nRNvixYuRnp6OkpISsUvR01T/PV+5cgX9+/dHSkoKgoKCmqi6lqNWqzFu3DgEBwdjwYIFYpfToJKS\nEiiVSsTExGi+CDRGa/37QtQa8UkyEVEzMealtrZIpVIhOjoa0dHRmvl/24q6ujpkZmaivLwcgYGB\nYpfzVOvWrcOQIUMQGhoqdilEHQ5DMhERvbDIyEjMnTsXgYGBRr/EJ6bs7GxkZGTg4MGDBs/13NLi\n4uKQk5ODAwcOwMzMTOxyiDochmQioia2evVqpKSk4MGDB3B1dcWePXvELqlZbdy4EaGhoXjvvffE\nLsVgEyZMwM6dO9G1a1exS2lQVlYWHj9+jOzsbNjb24tdDlGHZCp2AURE7U1sbCxiY2PFLqNFTZw4\nERMnThS7jHZj+vTpmD59uthlEHVofJJMRERERKSDIZmIiIiISAdDMhERERGRDoZkIiIiIiIdfHGP\nyAhpaWlil0A66pfs5b0x3LFjxwDwmnVEN27cgFKpFLsMojaBK+4RGaB+hTIiorZuzpw5XHGPyAAM\nyURErYREIkFqair8/f3FLoWIqMPjmGQiIiIiIh0MyUREREREOhiSiYiIiIh0MCQTEREREelgSCYi\nIiIi0sGQTERERESkgyGZiIiIiEgHQzIRERERkQ6GZCIiIiIiHQzJREREREQ6GJKJiIiIiHQwJBMR\nERER6WBIJiIiIiLSwZBMRERERKSDIZmIiIiISAdDMhERERGRDoZkIiIiIiIdDMlERERERDoYkomI\niIiIdDAkExERERHpYEgmIiIiItLBkExEREREpIMhmYiIiIhIB0MyEREREZEOhmQiIiIiIh0MyURE\nREREOhiSiYiIiIh0MCQTEREREelgSCYiIiIi0sGQTERERESkgyGZiIiIiEgHQzIRERERkQ6GZCIi\nIiIiHRJBEASxiyAi6mhCQkJw8eJFrbZTp07B1dUV9vb2mjYTExP84x//gFKpbOkSiYg6NFOxCyAi\n6oi6dOmCTz75RK/97NmzWv/bzc2NAZmISAQcbkFEJIL58+c/t49UKkVwcHDzF0NERHo43IKISCSD\nBg3C+fPn8ax/hi9evIg+ffq0YFVERATwSTIRkWheffVVmJiYNLhNIpFg8ODBDMhERCJhSCYiEsm8\nefNQV1fX4DYTExO89tprLVwRERHV43ALIiIR+fj44MSJE1Cr1VrtEokE169fR/fu3UWqjIioY+OT\nZCIiEb3yyiuQSCRabZ06dcLo0aMZkImIRMSQTEQkorlz5+q1SSQSvPrqqyJUQ0RE9RiSiYhE1Llz\nZ0yYMEHrBT6JRIKZM2eKWBURETEkExGJLCgoSDMNnImJCSZNmgRHR0eRqyIi6tgYkomIRDZr1ixI\npVIAgCAICAoKErkiIiJiSCYiEpmlpSX+8pe/AHiyyp6fn5/IFREREUMyEVEr8Ne//hUAMHPmTFha\nWopcDRERcZ5kahXS0tIQEBAgdhlERCSyOXPmID09XewyiGAqdgFEf5Samip2CUSiCQgIwJtvvonR\no0eLXUqbER8fDwAIDw8XuRJqCvX3k6g1YEimVsXf31/sEohEExAQgNGjR/PvgRHqnzjymrUPfIJM\nrQnHJBMRERER6WBIJiIiIiLSwZBMRERERKSDIZmIiIiISAdDMhERERGRDoZkIqJ25sCBA7C1tcXX\nX38tdimt3uHDhxEZGYmMjAy4ublBIpFAIpHglVde0es7ceJEWFtbw8TEBAMHDsSpU6dEqNhw0dHR\nGDBgAGxsbGBubg6VSoW3334bDx8+1OoXExOj+dx//DNo0CC9Y9bU1CA2NhYqlQpSqRR2dnYYNGgQ\nCgoKAAD79u3D5s2bUVdX1xIfkahZMSQTEbUzXCPKMGvXrkViYiJWr16N2bNnIz8/H+7u7nB0dMSO\nHTuwf/9+rf7fffcd0tPT4efnh9zcXAwbNkykyg1z5MgRLFu2DAUFBbh79y5iY2ORkJCAuXPnvvAx\nAwIC8MUXX2Dnzp2orKzEb7/9Bnd3d03wnjZtGmQyGSZMmIDS0tKm+ihEomBIJiJqZ3x9ffHgwQP4\n+fmJXQqqqqrg4+Mjdhl6Nm3ahN27dyMtLQ3W1tZa2xITE9GpUyeEhITgwYMHIlXYeFZWVggJCYGD\ngwOsra3h7++PmTNn4tChQ7h+/bpW3+3bt0MQBK0/v/76q1af3bt3IzMzE+np6XjppZdgamoKhUKB\nrKwsrafOy5cvx+DBgzF16lTU1ta2yGclag4MyURE1Gw+++wzFBcXi12GlitXrmDNmjVYv349ZDKZ\n3nYfHx+EhYWhqKgIb731lggVNo1vvvkGJiYmWm2dO3cGAFRWVhp9vI8//hjDhg2Dh4fHc/uuW7cO\nOTk5SEhIMPo8RK0FQzIRUTty9OhRuLi4QCKR4KOPPgIAJCcnw9LSEhYWFsjKysKUKVNgY2MDpVKJ\nXbt2afZNTEyETCaDs7MzFi9eDIVCAZlMBh8fH5w4cULTLzQ0FFKpFF27dtW0vfHGG7C0tIREIsHd\nu3cBAGFhYYiIiEBeXh4kEglUKhUA4NChQ7CxscHGjRtb4pLoSUxMhCAImDZt2lP7xMTEoE+fPvj0\n009x+PDhZx5PEATExcWhf//+MDc3h729PWbMmIELFy5o+hh6DwCgrq4O7777LlxcXCCXy+Hp6YnU\n1NTGfej/KCoqglwuh6urq1H7VVdX4/jx4xgyZIhB/e3t7TF27FgkJCRw+A+1WQzJRETtyOjRo/Gv\nf/1Lq23p0qUIDw9HVVUVrK2tkZqairy8PLi5uWHRokWoqakB8CT8BgcHo7KyEsuXL0dBQQFOnTqF\n2tpavPzyy5qf6BMTE/WWgU5KSsL69eu12hISEuDn5wd3d3cIgoArV64AgOalLrVa3SzX4Hn279+P\nvn37wsLC4ql95HI5Pv/8c3Tq1AmLFi1CRUXFU/uuW7cOkZGReOedd1BcXIwffvgB169fx5gxY3D7\n9m0Aht8DAFi1ahW2bNmC+Ph43Lp1C35+fpg/fz5+/vnnRn3uyspKHDlyBIsWLYJUKtXaFhkZCXt7\ne0ilUri6umLGjBk4efKkZvvNmzdRXV2NX375BePHj9d8gerfvz+SkpIaDMJDhw5FUVERzpw506i6\nicTCkExE1IH4+PjAxsYGTk5OCAwMREVFBa5dCs93awAAIABJREFUu6bVx9TUVPNUdMCAAUhOTkZ5\neTlSUlKapAZfX1+UlZVhzZo1TXI8Y1RUVODq1atwd3d/bl9vb2+Eh4ejoKAAq1atarBPVVUV4uLi\nMGvWLAQFBcHW1hYeHh7YunUr7t69i08++URvn2fdg0ePHiE5ORkzZ87E7NmzYWdnh6ioKJiZmTX6\n+sfGxkKhUCAmJkar/bXXXsO+fftw/fp1PHz4ELt27cK1a9cwduxY5ObmAoDmxTwnJyds3LgRubm5\nuH37NmbMmIFly5bhyy+/1Dtf7969AQDnzp1rVN1EYmFIJiLqoOqfJv7xKWZDRowYAQsLC63hA21V\ncXExBEF45lPkP4qJiUHfvn2RlJSEo0eP6m3Pzc3Fw4cPMWLECK32kSNHQiqVag1TaYjuPbh48SIq\nKyu1XoSTy+Xo2rVro67/3r17kZaWhm+//VbvRcUePXpg6NChsLKyglQqhZeXF1JSUlBVVYWkpCQA\ngLm5OQBg4MCB8PHxgYODA2xtbbF+/XrY2to2+GWg/hrXP00namsYkomI6LnMzc1x584dsctotEeP\nHgH4PfQ9j0wmQ0pKCiQSCRYsWICqqiqt7fXTnFlZWenta2dnh/LycqPqqx/WERUVpTVncWFh4Qu9\nbAc8mZVi06ZNyM7ORq9evQzax8PDAyYmJrh06RIAQKFQAIBmvHk9qVSKnj17Ii8vT+8YcrkcwO/X\nnKitYUgmIqJnqqmpQWlpKZRKpdilNFp9cDNmsQtvb2+sWLECly9fxoYNG7S22dnZAUCDYfhFrpmT\nkxMAID4+Xm9KtmPHjhl1LAD48MMPsWPHDhw5cgTdunUzeD+1Wg21Wq35MmFlZYXevXvj/Pnzen1r\na2tha2ur115dXQ3g92tO1NYwJBMR0TNlZ2dDEAR4eXlp2kxNTZ87TKM1cnZ2hkQiMXr+4w0bNqBf\nv344ffq0VvugQYNgZWWl91LdiRMnUF1djeHDhxt1nh49ekAmkyEnJ8eo/XQJgoCVK1fi3LlzyMzM\nbPBJd71JkybptZ08eRKCIMDb21vTFhAQgNOnTyM/P1/TVllZicLCwganhau/xl26dGnMRyESDUMy\nERFpUavVuH//Pmpra3H27FmEhYXBxcUFwcHBmj4qlQr37t1DZmYmampqcOfOHRQWFuody8HBATdv\n3kRBQQHKy8tRU1ODgwcPijYFnIWFBdzc3HDjxg2j9qsfdqE777BMJkNERAT27t2LHTt2oKysDOfO\nncOSJUugUCgQEhJi9Hlef/117Nq1C8nJySgrK0NdXR1u3LiBW7duAQACAwPRpUuXZy6Lff78eWzZ\nsgXbtm2DmZmZ3pLTH3zwgaZvUVERdu/ejdLSUtTU1ODYsWNYuHAhXFxcsGTJEk2/FStWoGfPnggO\nDsa1a9dQUlKClStXoqqqqsEXG+uvsSHzKhO1RgzJRETtyEcffYSRI0cCAFauXInp06cjOTkZ8fHx\nAABPT0/k5+dj27ZtiIiIAABMnjwZly9f1hzj0aNH8PDwgFwux5gxY9CnTx98//33WuN4ly5divHj\nx2PevHno27cvNmzYoPlZ3dvbWzNd3JIlS+Ds7IwBAwZg6tSpuHfvXotch2fx9fVFbm6u1vjir776\nCiqVCnl5eRg5ciTefPNNvf28vLywYsUKvfa1a9ciNjYW0dHR6Ny5M8aOHYtevXohOzsblpaWAGDU\nPUhISEB4eDg2b94MR0dHKBQKhIWF4f79+wCeDGMoLi5GVlbWUz+jMXMTT548GVFRUVAqlbCwsIC/\nvz9GjRqF48ePw9HRUdPP3t4eP/74I5RKJYYMGYLu3bvjp59+wv79+xucP/nkyZPo3r07PD09Da6F\nqDWRCJzlm1qBtLQ0BAQEcNJ56tAkEglSU1P15iBuSYsXL0Z6ejpKSkpEq8EYc+fOBQCkp6cbvM+V\nK1fQv39/pKSkICgoqLlKazZqtRrjxo1DcHAwFixYIHY5DSopKYFSqURMTIzmi4AhXuR+EjUXPkkm\nIiItxrzU1hapVCpER0cjOjpaM/9vW1FXV4fMzEyUl5cjMDBQ7HKeat26dRgyZAhCQ0PFLoXohTEk\nU7uxcOFCWFtbQyKRNPqlF2oaGRkZcHNz0xsPKZVK4ezsjHHjxuH999/X/IxM1FIiIyMxd+5cBAYG\nGv0Sn5iys7ORkZGBgwcPGjzXc0uLi4tDTk4ODhw4ADMzM7HLIXphDMnUbnz66afYtm2b2GXQH8ye\nPRv5+flwd3eHra0tBEGAWq1GcXEx0tLS4OrqipUrV2LgwIGNXnKXGm/16tVISUnBgwcP4Orqij17\n9ohdUrPauHEjQkND8d5774ldisEmTJiAnTt3omvXrmKX0qCsrCw8fvwY2dnZsLe3F7scokZhSCZq\npaqqquDj4yN2GQCathaJRAI7OzuMGzcOKSkpSEtLw+3bt+Hr69umnug9TWu6b8aKjY3F48ePIQgC\nrl69ijlz5ohdUrObOHEiNm3aJHYZ7cb06dMRGRmpNwsIUVvEkEztikQiEbuEJvPZZ5+huLhY7DIA\nNG8tc+bMQXBwMIqLi7F169ZmOUdLak33jYiIXhxDMrVZgiDg/fffR9++fWFubg5bW1v87W9/0+qz\nZcsWWFhYwNraGsXFxYiIiED37t1x8eJFCIKAuLg49O/fH+bm5rC3t8eMGTNw4cIFzf6JiYmQyWRw\ndnbG4sWLoVAoIJPJ4OPjgxMnTujV87zjhYaGQiqVav1U+sYbb8DS0hISiUSz5GtYWBgiIiKQl5cH\niUQClUpl9LVp7loOHTrUZHPd1s+/e/DgQQAd974REVErIhC1AqmpqYKx/zm+8847gkQiEf7+978L\n9+/fFyorK4WkpCQBgHD69GmtfgCE5cuXCx9++KEwa9Ys4bfffhPeffddQSqVCtu3bxdKS0uFs2fP\nCsOGDRM6d+4s/Pvf/9bsHxISIlhaWgrnz58XHj16JOTm5gojR44UrK2thWvXrmn6GXq8v/71r0KX\nLl20Psv7778vABDu3LmjaZs9e7bg7u5u1DVpyVq++eYbwdraWoiOjn5uPe7u7oKtre1Tt5eVlQkA\nhB49emjaOuJ9AyCkpqa+0L4d1Zw5c4Q5c+aIXQY1Ed5Pak34JJnapKqqKsTHx+O///u/sWLFCtjZ\n2UEul8PBweGp+2zatAnLli1DRkYGevbsibi4OMyaNQtBQUGwtbWFh4cHtm7dirt37+KTTz7R2tfU\n1FTzpHHAgAFITk5GeXk5UlJSNPUYc7zm1FK1+Pr6oqysDGvWrGn0sepnJSkvL9fb1lHuGxERtS6m\nYhdA9CKuXLmCyspKTJgw4YX2z83NxcOHDzFixAit9pEjR0Iqler9JK9rxIgRsLCw0Pwk39jjNaXW\nVIuhKioqIAgCbGxsntmvPd+3eseOHWvxc7Zl9Usfp6WliVwJNYUbN25AqVSKXQYRAIZkaqPq/4/R\nycnphfYvLS0FAFhZWelts7Oza/CJpi5zc3PcuXOnyY7XVFpTLYa6dOkSAKBfv37P7Nee71u9hIQE\nJCQktPh527qAgACxS6Am0hFmVaG2gSGZ2iSZTAYAePz48Qvtb2dnBwANhqDS0tLnPsmoqanR6tfY\n4zWl1lSLoQ4dOgQAmDJlyjP7tef7Vk/sZanbGi5j3L7U30+i1oBjkqlNGjRoEDp16oR//vOfL7y/\nlZWV3gIWJ06cQHV1NYYPH/7M/bOzsyEIAry8vIw+nqmpKWpqal6obkO0ploM8e9//xvx8fFQKpVY\nsGDBM/u25/tGREStC0MytUlOTk6YPXs29uzZg88++wxlZWU4e/aswS9ayWQyREREYO/evdixYwfK\nyspw7tw5LFmyBAqFAiEhIVr91Wo17t+/j9raWpw9exZhYWFwcXHRTF1mzPFUKhXu3buHzMxM1NTU\n4M6dOygsLNSr0cHBATdv3kRBQQHKy8sNDmgtVcvBgweNmgJOEAQ8fPgQarUagiDgzp07SE1NxahR\no2BiYoLMzMznjkluz/eNiIhaGVHn1iD6jxeZAq68vFxYuHCh4OjoKFhZWQmjR48W3n33XQGAoFQq\nhTNnzgibN28W5HK5Znqx7du3a/ZXq9XC+++/L/Tu3VswMzMT7O3thZkzZwoXL17UOk9ISIhgZmYm\ndO/eXTA1NRVsbGyEGTNmCHl5eVr9DD1eSUmJMH78eEEmkwmurq7Cm2++Kfztb38TAAgqlUozPdmp\nU6eEnj17CnK5XBg9erTWdGTP0xK1HDhwQLC2thZiYmKeWse+ffsET09PwcLCQpBKpUKnTp0EAIJE\nIhHs7OyEP/3pT0J0dLRQUlKitV9HvW/gFHBG45Rh7QvvJ7UmEkEQBNESOtF/pKWlISAgAK3xP8fF\nixcjPT0dJSUlYpdCRmiL900ikXBMspE4Jrl94f2k1oTDLYgMUFdXJ3YJ9AJ434iI6EUxJBO1ARcu\nXIBEInnun8DAQLFLJSIiahcYkomeYfXq1UhJScGDBw/g6uqKPXv2iFJHv379IAjCc//s3r1blPpa\nm9Zy36j1O3z4MCIjI5GRkQE3NzfNF85XXnlFr+/EiRNhbW0NExMTDBw4EKdOnRKhYsNFR0djwIAB\nsLGxgbm5OVQqFd5++208fPhQq19MTEyDX7oHDRqkd8yamhrExsZCpVJBKpXCzs4OgwYNQkFBAQBg\n37592Lx5M3/FoXaBIZnoGWJjY/H48WMIgoCrV69ykvs2gveNDLF27VokJiZi9erVmD17NvLz8+Hu\n7g5HR0fs2LED+/fv1+r/3XffIT09HX5+fsjNzcWwYcNEqtwwR44cwbJly1BQUIC7d+8iNjYWCQkJ\njZqLOCAgAF988QV27tyJyspK/Pbbb3B3d9cE72nTpkEmk2HChAmaxXqI2iqGZCIiAgBUVVXBx8en\nzZ/DEJs2bcLu3buRlpYGa2trrW2JiYno1KkTQkJC8ODBA5EqbDwrKyuEhITAwcEB1tbW8Pf3x8yZ\nM3Ho0CFcv35dq+/27dv1fpn69ddftfrs3r0bmZmZSE9Px0svvQRTU1MoFApkZWVpPXVevnw5Bg8e\njKlTp6K2trZFPitRc2BIJiIiAMBnn32G4uLiNn+O57ly5QrWrFmD9evXa1bv/CMfHx+EhYWhqKgI\nb731lggVNo1vvvkGJiYmWm2dO3cGAFRWVhp9vI8//hjDhg2Dh4fHc/uuW7cOOTk5XGKd2jSGZCKi\nNkoQBMTFxaF///4wNzeHvb09ZsyYgQsXLmj6hIaGQiqVomvXrpq2N954A5aWlpBIJLh79y4AICws\nDBEREcjLy4NEIoFKpUJiYiJkMhmcnZ2xePFiKBQKyGQy+Pj44MSJE01yDuDJsuTGLEzTWImJiRAE\nAdOmTXtqn5iYGPTp0weffvopDh8+/MzjGXIfkpOTYWlpCQsLC2RlZWHKlCmwsbGBUqnErl27tI5X\nV1eHd999Fy4uLpDL5fD09ERqamrjPvR/FBUVQS6Xw9XV1aj9qqurcfz4cQwZMsSg/vb29hg7diwS\nEhJa5dSeRIZgSCYiaqPWrVuHyMhIvPPOOyguLsYPP/yA69evY8yYMbh9+zaAJ4FQd97lpKQkrF+/\nXqstISEBfn5+cHd3hyAIuHLlCkJDQxEcHIzKykosX74cBQUFOHXqFGpra/Hyyy9rfrJvzDmA36fq\nU6vVTXdxnmH//v3o27cvLCwsntpHLpfj888/R6dOnbBo0SJUVFQ8ta8h92Hp0qUIDw9HVVUVrK2t\nkZqairy8PLi5uWHRokVaKzOuWrUKW7ZsQXx8PG7dugU/Pz/Mnz9fb/l0Y1VWVuLIkSNYtGgRpFKp\n1rbIyEjY29tDKpXC1dUVM2bMwMmTJzXbb968ierqavzyyy8YP3685gtT//79kZSU1GAQHjp0KIqK\ninDmzJlG1U0kFoZkIqI2qKqqCnFxcZg1axaCgoJga2sLDw8PbN26FXfv3jV4iXZDmJqaap6SDhgw\nAMnJySgvL0dKSkqTHN/X1xdlZWVYs2ZNkxzvWSoqKnD16lW4u7s/t6+3tzfCw8NRUFCAVatWNdjn\nRe6Dj48PbGxs4OTkhMDAQFRUVODatWsAgEePHiE5ORkzZ87E7NmzYWdnh6ioKJiZmTX6esfGxkKh\nUCAmJkar/bXXXsO+fftw/fp1PHz4ELt27cK1a9cwduxY5ObmAoDmxTwnJyds3LgRubm5uH37NmbM\nmIFly5bhyy+/1Dtf7969AQDnzp1rVN1EYmFIJiJqg3Jzc/Hw4UOMGDFCq33kyJGQSqVawyGa2ogR\nI2BhYaE1nKCtKC4uhiAIz3yK/EcxMTHo27cvkpKScPToUb3tjb0P9U90658kX7x4EZWVlVovwsnl\ncnTt2rVR13vv3r1IS0vDt99+q/eiYo8ePTB06FBYWVlBKpXCy8sLKSkpqKqqQlJSEgDA3NwcADBw\n4ED4+PjAwcEBtra2WL9+PWxtbRv8MlB/jeufphO1NQzJRERtUP30WlZWVnrb7OzsUF5e3qznNzc3\nx507d5r1HM3h0aNHAH4Pfc8jk8mQkpICiUSCBQsWoKqqSmt7U9+H+mEdUVFRWnMWFxYWvtDLdsCT\nWSk2bdqE7Oxs9OrVy6B9PDw8YGJigkuXLgEAFAoFAGjGl9eTSqXo2bMn8vLy9I4hl8sB/H7Nidoa\nhmQiojbIzs4OABoMYaWlpVAqlc127pqammY/R3OpD27GLHbh7e2NFStW4PLly9iwYYPWtqa+D05O\nTgCA+Ph4vSnZjh07ZtSxAODDDz/Ejh07cOTIEXTr1s3g/dRqNdRqtebLhJWVFXr37o3z58/r9a2t\nrYWtra1ee3V1NYDfrzlRW8OQTETUBg0aNAhWVlZ6L3OdOHEC1dXVGD58uKbN1NRU68WwxsrOzoYg\nCPDy8mq2czQXZ2dnSCQSo+c/3rBhA/r164fTp09rtRtzHwzRo0cPyGQy5OTkGLWfLkEQsHLlSpw7\ndw6ZmZkNPumuN2nSJL22kydPQhAEeHt7a9oCAgJw+vRp5Ofna9oqKytRWFjY4LRw9de4S5cujfko\nRKJhSCYiaoNkMhkiIiKwd+9e7NixA2VlZTh37hyWLFkChUKBkJAQTV+VSoV79+4hMzMTNTU1uHPn\nDgoLC/WO6eDggJs3b6KgoADl5eWa0KtWq3H//n3U1tbi7NmzCAsLg4uLC4KDg5vkHAcPHmyxKeAs\nLCzg5uaGGzduGLVf/bAL3XmHjbkPhp7n9ddfx65du5CcnIyysjLU1dXhxo0buHXrFgAgMDAQXbp0\neeay2OfPn8eWLVuwbds2mJmZ6S05/cEHH2j6FhUVYffu3SgtLUVNTQ2OHTuGhQsXwsXFBUuWLNH0\nW7FiBXr27Ing4GBcu3YNJSUlWLlyJaqqqhp8sbH+GhsyrzJRa8SQTETURq1duxaxsbGIjo5G586d\nMXbsWPTq1QvZ2dmwtLTU9Fu6dCnGjx+PefPmoW/fvtiwYYPmJ3Bvb2/NVG5LliyBs7MzBgwYgKlT\np+LevXsAnowp9fDwgFwux5gxY9CnTx98//33WuN6G3uOluTr64vc3Fyt8cVfffUVVCoV8vLyMHLk\nSLz55pt6+3l5eWHFihV67Ybch+TkZMTHxwMAPD09kZ+fj23btiEiIgIAMHnyZFy+fBnAk6nywsPD\nsXnzZjg6OkKhUCAsLAz3798H8GQYQ3FxMbKysp76GY2Zm3jy5MmIioqCUqmEhYUF/P39MWrUKBw/\nfhyOjo6afvb29vjxxx+hVCoxZMgQdO/eHT/99BP279/f4PzJJ0+eRPfu3eHp6WlwLUStiUTgLN/U\nCqSlpSEgIICTzlOHJpFIkJqaqjfnsJgWL16M9PR0lJSUiF1Kg+bOnQsASE9PN3ifK1euoH///khJ\nSUFQUFBzldZs1Go1xo0bh+DgYCxYsEDschpUUlICpVKJmJgYzRcBQ7zI/SRqLnySTEREz2TMS25t\ngUqlQnR0NKKjozXz/7YVdXV1yMzMRHl5OQIDA8Uu56nWrVuHIUOGIDQ0VOxSiF4YQzIREXU4kZGR\nmDt3LgIDA41+iU9M2dnZyMjIwMGDBw2e67mlxcXFIScnBwcOHICZmZnY5RC9MIZkIiJq0OrVq5GS\nkoIHDx7A1dUVe/bsEbukJrVx40aEhobivffeE7sUg02YMAE7d+5E165dxS6lQVlZWXj8+DGys7Nh\nb28vdjlEjWIqdgFERNQ6xcbGIjY2VuwymtXEiRMxceJEsctoN6ZPn47p06eLXQZRk+CTZCIiIiIi\nHQzJREREREQ6GJKJiIiIiHQwJBMRERER6eCLe9Sq1E8kT9RRxcfHcyEFIxw/fhwA/+1oL44fPw4v\nLy+xyyACwJBMrUSPHj0wZ84cscsgEpWlpSVMTfnPsjEYqNoXLy8veHt7i10GEQAuS01E1Gq0xmWp\niYg6Ko5JJiIiIiLSwZBMRERERKSDIZmIiIiISAdDMhERERGRDoZkIiIiIiIdDMlERERERDoYkomI\niIiIdDAkExERERHpYEgmIiIiItLBkExEREREpIMhmYiIiIhIB0MyEREREZEOhmQiIiIiIh0MyURE\nREREOhiSiYiIiIh0MCQTEREREelgSCYiIiIi0sGQTERERESkgyGZiIiIiEgHQzIRERERkQ6GZCIi\nIiIiHQzJREREREQ6GJKJiIiIiHQwJBMRERER6WBIJiIiIiLSwZBMRERERKSDIZmIiIiISAdDMhER\nERGRDoZkIiIiIiIdDMlERERERDoYkomIiIiIdDAkExERERHpMBW7ACKijqi0tBSCIOi1V1RU4P79\n+1ptVlZWMDMza6nSiIgIgERo6F9pIiJqVv/1X/+F77///rn9TExMUFRUhC5durRAVUREVI/DLYiI\nRDBv3jxIJJJn9unUqRP+/Oc/MyATEYmAIZmISARz5syBqemzR7xJJBK8+uqrLVQRERH9EUMyEZEI\n7O3tMXHiRJiYmDy1T6dOnTBz5swWrIqIiOoxJBMRiSQoKAhqtbrBbaampvD19YWtrW0LV0VERABD\nMhGRaKZNmwZzc/MGt9XV1SEoKKiFKyIionoMyUREIrGwsMDMmTMbnN5NLpdj6tSpIlRFREQAQzIR\nkajmz5+PmpoarTYzMzPMmTMHcrlcpKqIiIghmYhIRJMmTdIbd1xTU4P58+eLVBEREQEMyUREojIz\nM0NgYCCkUqmmzc7ODhMmTBCxKiIiYkgmIhLZvHnzUF1dDeBJaA4KCnruHMpERNS8uCw1EZHI1Go1\nunXrhtu3bwMAjh49ilGjRolcFRFRx8YnyUREIuvUqRNeeeUVAIBCoYCPj4/IFREREX/Po3YhLS1N\n7BKIGqVz584AgJdeegnp6ekiV0PUOD4+PlAqlWKXQdQoHG5B7YJEIhG7BCIi+o/U1FT4+/uLXQZR\no/BJMrUb/EeZxJaWloaAgAC86LOHPXv2YM6cOU1cVesnkUj497cd4UMLai84JpmIqJXoiAGZiKi1\nYkgmIiIiItLBkExEREREpIMhmYiIiIhIB0MyEREREZEOhmQiIiIiIh0MyURErcyBAwdga2uLr7/+\nWuxSWr3Dhw8jMjISGRkZcHNzg0QigUQi0axg+EcTJ06EtbU1TExMMHDgQJw6dUqEig0XHR2NAQMG\nwMbGBubm5lCpVHj77bfx8OFDrX4xMTGaz/3HP4MGDdI7Zk1NDWJjY6FSqSCVSmFnZ4dBgwahoKAA\nALBv3z5s3rwZdXV1LfERiVo1hmQiolaGazwZZu3atUhMTMTq1asxe/Zs5Ofnw93dHY6OjtixYwf2\n79+v1f+7775Deno6/Pz8kJubi2HDholUuWGOHDmCZcuWoaCgAHfv3kVsbCwSEhIwd+7cFz5mQEAA\nvvjiC+zcuROVlZX47bff4O7urgne06ZNg0wmw4QJE1BaWtpUH4WoTWJIJiJqZXx9ffHgwQP4+fmJ\nXQqqqqrg4+Mjdhl6Nm3ahN27dyMtLQ3W1tZa2xITE9GpUyeEhITgwYMHIlXYeFZWVggJCYGDgwOs\nra3h7++PmTNn4tChQ7h+/bpW3+3bt0MQBK0/v/76q1af3bt3IzMzE+np6XjppZdgamoKhUKBrKws\nrafOy5cvx+DBgzF16lTU1ta2yGclao0YkomI6Kk+++wzFBcXi12GlitXrmDNmjVYv349ZDKZ3nYf\nHx+EhYWhqKgIb731lggVNo1vvvkGJiYmWm2dO3cGAFRWVhp9vI8//hjDhg2Dh4fHc/uuW7cOOTk5\nSEhIMPo8RO0FQzIRUSty9OhRuLi4QCKR4KOPPgIAJCcnw9LSEhYWFsjKysKUKVNgY2MDpVKJXbt2\nafZNTEyETCaDs7MzFi9eDIVCAZlMBh8fH5w4cULTLzQ0FFKpFF27dtW0vfHGG7C0tIREIsHdu3cB\nAGFhYYiIiEBeXh4kEglUKhUA4NChQ7CxscHGjRtb4pLoSUxMhCAImDZt2lP7xMTEoE+fPvj0009x\n+PDhZx5PEATExcWhf//+MDc3h729PWbMmIELFy5o+hh6DwCgrq4O7777LlxcXCCXy+Hp6YnU1NTG\nfej/KCoqglwuh6urq1H7VVdX4/jx4xgyZIhB/e3t7TF27FgkJCRw+A91WAzJREStyOjRo/Gvf/1L\nq23p0qUIDw9HVVUVrK2tkZqairy8PLi5uWHRokWoqakB8CT8BgcHo7KyEsuXL0dBQQFOnTqF2tpa\nvPzyy5qf6BMTE+Hv7691jqSkJKxfv16rLSEhAX5+fnB3d4cgCLhy5QoAaF7qUqvVzXINnmf//v3o\n27cvLCwsntpHLpfj888/R6dOnbBo0SJUVFQ8te+6desQGRmJd955B8XFxfjhhx9w/fp1jBkzBrdv\n3wZg+D0AgFWrVmHLli2Ij4/HrVu34Ofnh/nz5+Pnn39u1OeurKzEkSNHsGjRIkilUq1tkZGRsLe3\nh1QqhaurK2bMmIGTJ09qtt+8eRPV1dXga2RTAAAYqklEQVT45ZdfMH78eM0XqP79+yMpKanBIDx0\n6FAUFRXhzJkzjaqbqK1iSCYiakN8fHxgY2MDJycnBAYGoqKiAteuXdPqY2pqqnkqOmDAACQnJ6O8\nvBwpKSlNUoOvry/KysqwZs2aJjmeMSoqKnD16lW4u7s/t6+3tzfCw8NRUFCAVatWNdinqqoKcXFx\nmDVrFoKCgmBrawsPDw9s3boVd+/exSeffKK3z7PuwaNHj5CcnIyZM2di9uzZsLOzQ1RUFMzMzBp9\n/WNjY6FQKBATE6PV/tprr2Hfvn24fv06Hj58iF27duHatWsYO3YscnNzAUDzYp6TkxM2btyI3Nxc\n3L59GzNmzMCyZcvw5Zdf6p2vd+/eAIBz5841qm6itoohmYiojap/mvjHp5gNGTFiBCwsLLSGD7RV\nxcXFEAThmU+R/ygmJgZ9+/ZFUlISjh49qrc9NzcXDx8+xIgRI7TaR44cCalUqjVMpSG69+DixYuo\nrKzUehFOLpeja9eujbr+e/fuRVpaGr799lu9FxV79OiBoUOHwsrKClKpFF5eXkhJSUFVVRWSkpIA\nAObm5gCAgQMHwsfHBw4ODrC1tcX69etha2vb4JeB+mtc/zSdqKNhSCYi6gDMzc1x584dsctotEeP\nHgH4PfQ9j0wmQ0pKCiQSCRYsWICqqiqt7fXTnFlZWenta2dnh/LycqPqqx/WERUVpTVncWFh4Qu9\nbAc8mZVi06ZNyM7ORq9evQzax8PDAyYmJrh06RIAQKFQAIBmvHk9qVSKnj17Ii8vT+8YcrkcwO/X\nnKijYUgmImrnampqUFpaCqVSKXYpjVYf3IxZ7MLb2xsrVqzA5cuXsWHDBq1tdnZ2ANBgGH6Ra+bk\n5AQAiI+P15uS7dixY0YdCwA+/PBD7NixA0eOHEG3bt0M3k+tVkOtVmu+TFhZWaF37944f/68Xt/a\n2lrY2trqtVdXVwP4/ZoTdTQMyURE7Vx2djYEQYCXl5emzdTU9LnDNFojZ2dnSCQSo+c/3rBhA/r1\n64fTp09rtQ8aNAhWVlZ6L9WdOHEC1dXVGD58uFHn6dGjB2QyGXJycozaT5cgCFi5ciXOnTuHzMzM\nBp9015s0aZJe28mTJyEIAry9vTVtAQEBOH36NPLz8zVtlZWVKCwsbHBauPpr3KVLl8Z8FKI2iyGZ\niKidUavVuH//Pmpra3H27FmEhYXBxcUFwcHBmj4qlQr37t1DZmYmampqcOfOHRQWFuody8HBATdv\n3kRBQQHKy8tRU1ODgwcPijYFnIWFBdzc3HDjxg2j9qsfdqE777BMJkNERAT27t2LHTt2oKysDOfO\nncOSJUugUCgQEhJi9Hlef/117Nq1C8nJySgrK0NdXR1u3LiBW7duAQACAwPRpUuXZy6Lff78eWzZ\nsgXbtm2DmZmZ3pLTH3zwgaZvUVERdu/ejdLSUtTU1ODYsWNYuHAhXFxcsGTJEk2/FStWoGfPnggO\nDsa1a9dQUlKClStXoqqqqsEXG+uvsSHzKhO1RwzJREStyEcffYSRI0cCAFauXInp06cjOTkZ8fHx\nAABPT0/k5+dj27ZtiIiIAABMnjwZly9f1hzj0aNH8PDwgFwux5gxY9CnTx98//33WuN4ly5divHj\nx2PevHno27cvNmzYoPlZ3dvbWzNd3JIlS+Ds7IwBAwZg6tSpuHfvXotch2fx9fVFbm6u1vjir776\nCiqVCnl5eRg5ciTefPNNvf28vLywYsUKvfa1a9ciNjYW0dHR6Ny5M8aOHYtevXohOzsblpaWAGDU\nPUhISEB4eDg2b94MR0dHKBQKhIWF4f79+wCeDGMoLi5GVlbWUz+jMXMTT548GVFRUVAqlbCwsIC/\nvz9GjRqF48ePw9HRUdPP3t4eP/74I5RKJYYMGYLu3bvjp59+wv79+xucP/nkyZPo3r07PD09Da6F\nqD2RCJwlnNoBiUSC1NRUvblfiVpSWloaAgICRF18YfHixUhPT0dJSYloNRjL2L+/V65cQf/+/ZGS\nkoKgoKBmrq7pqdVqjBs3DsHBwViwYIHY5TSopKQESqUSMTExmi8ChuK/x9Re8EkyEVE7Y8xLbW2R\nSqVCdHQ0oqOjNfP/thV1dXXIzMxEeXk5AgMDxS7nqdatW4chQ4YgNDRU7FKIRMOQTNQCFi5cCGtr\na0gkEq0Xeg4cOABbW1t8/fXXjTp+Ux2nOWVkZMDNzU1vbKVUKoWzszPGjRuH999/X/OTNNGzREZG\nYu7cuQgMDDT6JT4xZWdnIyMjAwcPHjR4rueWFhcXh5ycHBw4cABmZmZil0MkGoZkohbw6aefYtu2\nbXrtTfWzfFsYNTV79mzk5+fD3d0dtra2EAQBarUaxcXFSEtLg6urK1auXImBAwc2evnejmr16tVI\nSUnBgwcP4Orqij179ohdUrPauHEjQkND8d5774ldisEmTJiAnTt3omvXrmKX0qCsrCw8fvwY2dnZ\n+P/27jwmyuONA/h3ufbgEFQQFBEBxQPwAqOIPzVEWiWKikZabYKmBrAph9hUEVEBUTRlCQqxtoYm\n2tAVtGJVjLEVjWlLTVSwWFBQUEQLIvci5/P7w7Ltcggri6vL80n8Z955Z56ZyW4eXmfnNTMz03Q4\njGmUnqYDYGwo8/b2VvkpWFNTEzw9PfHrr78OqJ13gUAggKmpKRYuXIiFCxfC29sba9euhbe3N+7d\nu9fj2a2sd3FxcYiLi9N0GG+Vl5cXvLy8NB2G1vDx8YGPj4+mw2DsncBPkhl7SwQCgVraOXbsGCoq\nKtTS1rtm9erV8Pf3R0VFBY4cOaLpcBhjjA1hnCSzIScpKQkikQgWFhYIDAyElZUVRCIR3N3dkZOT\no6h34MABSCQSGBsbo6KiAuHh4RgzZgwKCwvR3t6OqKgo2NjYQCwWw8XFBTKZTHEvEeHgwYNwdHSE\nUCjEsGHD8MUXXyjFcf36ddjY2EAgEODw4cNK144fPw5XV1eIRCIYGhrC1tYWMTExCA0NRXh4OIqL\niyEQCODg4NBrO0SEhIQETJ48GUKhEGZmZlixYgUKCgoUdVJSUmBoaAiJRILMzEwsWbIEJiYmsLa2\nRlpamlJMV69exezZsyGRSGBiYgJnZ2fU1dUBAC5evKi2c3M7z/LNyspSlL1uvtU1hr7WlDHG2BBD\njGkBACSTyfpdPyAggAwNDenu3bv08uVLys/PJzc3NzI2NqZHjx4p6u3YsYMAUEhICB06dIhWrVpF\nf/31F23dupWEQiFlZGRQdXU1RUREkI6ODt24cUNxn0AgoK+++oqqq6tJLpdTcnIyAaBbt24p2n/8\n+DEBoEOHDinKpFIpAaB9+/ZRVVUVvXjxgr7++mtat24dERH5+vqSvb290nh6aicqKooMDAzo+PHj\nVFNTQ3l5eTRz5kwaOXIkPXv2rNsYf/75Z6qtraWKigqaP38+GRoaUktLCxERNTQ0kImJCcXHx1NT\nUxM9e/aMVq1aRZWVlUREdO7cOTI2Nqbo6Og+597e3p6GDRvW6/W6ujoCQGPHjlWU9We+BzqGvvro\nD5lMRvy1qjpVP7/s3cbrybQFf5szrfAmSXLXRO3GjRsEgPbs2aMo60y+mpqaFGVNTU0kkUjIz89P\nUSaXy0koFNLmzZtJLpeTRCKhxYsXK7WflpbWZ5Lc0tJCpqamtGjRIqV729raKDExkYj6lyTL5XIy\nMjJSipGI6I8//iAASslsT2PsTOiLioqIiOjPP/8kAHTu3Lke51MVfSXJREQCgYBMTU2JqO/5VscY\n+tNHf3CS/GY4qdIuvJ5MW/AP9xj7h6urKyQSidJ2hJ4UFhZCLpfDyclJUSYWi2FpaYmCggIUFRVB\nLpfD09NT5Rjy8vJQU1ODDz74QKlcV1cXISEh/W4nPz8fDQ0NcHV1VSp3c3ODgYGB0raSnhgYGAAA\nWltbAQB2dnawsLDA+vXrERISAn9/f9ja2vY7HlU0NjaCiGBiYgKg7/lWxxjetI/erFmzRuV7hjqp\nVIr09HRNh8EYYwq8J5mx/xAKhaisrHxtncbGRgBAZGSk0nm/paWlkMvlKCsrAwCYm5ur3H/n/lhT\nU1OV7/2vmpoaAICRkVG3a6ampqivr1epPbFYjF9++QUeHh7Yu3cv7Ozs4Ofnp/RaYHW5d+8eAGDS\npEkA+p5vdYxBXX0wxhjTHvwkmbF/tLa2oqamBtbW1q+t15n8SqVShIaGdrt+5coVAEBzc7PKMYwe\nPRoA8Pz5c5Xv/a/OJLunZLg/Y+zJ1KlT8dNPP6GyshIJCQnYv38/pk6dip07dw4o1q4uXrwIAFiy\nZAmAvudbFb2NofPNZ+roAwA/EVWRQCBAWFgYv8ZYS6jrJB/GNI2fJDP2j+zsbBAR5syZ89p6Y8eO\nhUgkUnpz3n85OTlBR0cHV69eVTkGW1tbDB8+HJcuXVL53q4xGBkZdXspR05ODlpaWjBr1iyV2isv\nL8fdu3cBvEpa9+3bh5kzZyrK1OXZs2eQSqWwtrbGxo0bAfQ93/31ujGoqw/GGGPag5NkNmR1dHSg\nuroabW1tyMvLQ2hoKGxsbBRHkPVGJBJhw4YNSEtLQ0pKCurq6tDe3o6ysjI8ffoU5ubm8PX1RUZG\nBo4dO4a6ujrk5eXh6NGjfcYkFAoRERGBa9euITg4GE+ePEFHRwfq6+sVCd7w4cNRXl6OkpIS1NfX\nK/bcdo0xPDwcp0+fxokTJ1BXV4c7d+4gKCgIVlZWCAgIUGmuysvLERgYiIKCArS0tODWrVsoLS1V\n/EGRlZWl0hFwRISGhgZ0dHSAiFBZWQmZTIZ58+ZBV1cXZ86cUexJ7mu+1TEGdfXBGGNMi2j2d4OM\nqQfe4HQLfX19GjNmDOnp6ZGJiQmtWLGCiouLFXXi4+NJLBYrjiM7fvy44lpzczN9+eWXZGNjQ3p6\nemRubk6+vr6Un59PRET19fX06aef0ogRI8jIyIg8PDwoKiqKAJC1tTXl5ubSoUOHyNLSkgCQRCKh\n5cuXK9o/fPgwOTs7k0gkIpFIRDNmzKDk5GQiIrp58yaNGzeOxGIxeXh4UGRkZI/tdHR00MGDB2nC\nhAmkr69PZmZmtHLlSiosLFT0k5ycTBKJhADQhAkTqLi4mI4ePUomJiYEgMaNG0f37t2jkpIScnd3\nJzMzM9LV1aXRo0fTjh07qK2tjYiILly4QMbGxhQbG9vrnJ89e5ZcXFxIIpGQgYEB6ejoEADFSRaz\nZ8+m6Ohoqqqq6nbv6+ZbXWPoa037g0+3eDOqfn7Zu43Xk2kLARGRRrJzxtRIIBBAJpP1e09jYGAg\n0tPTUVVVNciRsaHk5MmTWLt2LfhrVTWqfn7Zu43Xk2kL3m7Bhqz29nZNh8AYY4yxdxQnyYwxxrTe\n5cuXsX37dpw6dQp2dnaKY/4++eSTbnW9vLxgbGwMXV1dTJ06FTdv3tRAxKrr6OiAVCqFu7t7t2tn\nz55FfHw8PxxgTAWcJLMhJyIiAqmpqaitrcX48eORkZGh6ZAYY4No165dSEpKQkREBHx9ffHgwQPY\n29tjxIgROHHiBM6fP69U/9KlS0hPT8eyZcuQn5+PmTNnaijy/rt//z7+97//YcuWLT2e7b18+XKI\nRCJ4enoqzlFnjL0eJ8lsyImLi0NzczOICA8fPsTq1as1HRJjatPU1NTjk8T3rQ912b9/P3744Qec\nPHkSxsbGSteSkpKgo6ODgIAA1NbWaijCgcvNzcW2bdsQFBSE6dOn91ovJCQE06ZNw9KlS9HW1vYW\nI2Ts/cRJMmOMaZFjx46hoqLive9DHYqKirBz507s2bMHIpGo23V3d3eEhobiyZMn2Lp1qwYiVI9p\n06bh1KlTWLduHYRC4Wvr7t69G7dv30ZiYuJbio6x9xcnyYwxpkFEhISEBEyePBlCoRBmZmZYsWIF\nCgoKFHWCg4NhYGAAS0tLRdlnn30GQ0NDCAQCxRsaQ0NDER4ejuLiYggEAjg4OCApKQkikQgWFhYI\nDAyElZUVRCIR3N3dkZOTo5Y+gFdvSlTlrOy3ISkpCUSE5cuX91onNjYWEydOxLfffovLly+/tr3+\nrFVKSgoMDQ0hkUiQmZmJJUuWwMTEBNbW1khLS1Nqr729HVFRUbCxsYFYLIaLiwtkMtnABt0HMzMz\nLFiwAImJiXwKC2N94CSZMcY0aPfu3di+fTt27NiBiooKXLt2DY8fP8b8+fPx999/A3iV7HU9Tis5\nORl79uxRKktMTMSyZctgb28PIkJRURGCg4Ph7+8PuVyOkJAQlJSU4ObNm2hra8PixYvx+PHjAfcB\n/HtaTEdHh/omZ4DOnz8PR0dHSCSSXuuIxWJ899130NHRwaZNm9DY2Nhr3f6s1ebNmxEWFoampiYY\nGxtDJpOhuLgYdnZ22LRpk9LLf7Zt24YDBw5AKpXi6dOnWLZsGT7++ONub8pUtxkzZuDJkyfIzc0d\n1H4Ye99xkswYYxrS1NSEhIQErFq1CuvXr8ewYcPg7OyMI0eO4Pnz5/16S2N/6enpKZ6ATpkyBSkp\nKaivr0dqaqpa2vf29kZdXR127typlvYGqrGxEQ8fPoS9vX2fdefOnYuwsDCUlJRg27ZtPdZ5k7Vy\nd3eHiYkJzM3N4efnh8bGRjx69AgA8PLlS6SkpGDlypXw9fWFqakpIiMjoa+vr7Y16c2ECRMAAHfu\n3BnUfhh733GSzBhjGpKfn4+Ghga4uroqlbu5ucHAwEBpO4S6ubq6QiKRKG0V0CYVFRUgotc+Rf6v\n2NhYODo6Ijk5GdevX+92faBrZWBgAACKJ8mFhYWQy+VwcnJS1BGLxbC0tBz0Nemck86n34yxnnGS\nzBhjGtJ5FJeRkVG3a6ampqivrx/U/oVCISorKwe1D015+fIlAPT5Q7ZOIpEIqampEAgE2LhxI5qa\nmpSuq3utOrd1REZGKs5sFggEKC0t7fEIN3USi8UA/p0jxljPOElmjDENMTU1BYAeE6yamhpYW1sP\nWt+tra2D3ocmdSaCqrw8Y+7cudiyZQvu37+PmJgYpWvqXitzc3MAgFQqBREp/fvtt99UaktVLS0t\nAP6dI8ZYzzhJZowxDXFycoKRkVG3H2rl5OSgpaUFs2bNUpTp6ekp/ehroLKzs0FEmDNnzqD1oUkW\nFhYQCAQqn38cExODSZMm4datW0rlqqxVf4wdOxYikQi3b99W6T516JyTUaNGvfW+GXufcJLMGGMa\nIhKJEB4ejtOnT+PEiROoq6vDnTt3EBQUBCsrKwQEBCjqOjg44MWLFzhz5gxaW1tRWVmJ0tLSbm0O\nHz4c5eXlKCkpQX19vSLp7ejoQHV1Ndra2pCXl4fQ0FDY2NjA399fLX1kZWW9U0fASSQS2NnZoays\nTKX7Ordd6Orqdivv71r1t58NGzYgLS0NKSkpqKurQ3t7O8rKyvD06VMAgJ+fH0aNGqX212J3zomz\ns7Na22VM23CSzBhjGrRr1y7ExcUhOjoaI0eOxIIFC2Bra4vs7GwYGhoq6m3evBmLFi3CRx99BEdH\nR8TExCj+u3zu3LmKo9yCgoJgYWGBKVOmYOnSpXjx4gWAV/tPnZ2dIRaLMX/+fEycOBFXrlxR2rM7\n0D7eNd7e3sjPz1faX/zjjz/CwcEBxcXFcHNzw+eff97tvjlz5mDLli3dyvuzVikpKZBKpQAAFxcX\nPHjwAN988w3Cw8MBAB9++CHu378P4NVxemFhYYiPj8eIESNgZWWF0NBQVFdXA3i1LaKiogKZmZmv\nHefvv/8ODw8PjB49Gjk5OcjNzYWVlRXmzZuHa9eudat/48YNjBkzBi4uLv2ZRsaGLAHxaeJMCwgE\nAshksm7nvDL2Np08eRJr1659517SEBgYiPT0dFRVVWk6lB4N1ue3qKgIkydPRmpqKtavX6/Wtt+G\njo4OLFy4EP7+/ti4caNa2qyqqoK1tTViY2MVibu68fcx0xb8JJkxxoYAVX7Api0cHBwQHR2N6Oho\nNDQ0aDoclbS3t+PMmTOor6+Hn5+f2trdvXs3pk+fjuDgYLW1yZi24iSZMcaY1tq+fTvWrFkDPz8/\nlX/Ep0nZ2dk4deoUsrKy+n3Wc18SEhJw+/ZtXLhwAfr6+mppkzFtxkkyY4xpsYiICKSmpqK2thbj\nx49HRkaGpkN66/bu3Yvg4GDs27dP06H0m6enJ77//ntYWlqqpb3MzEw0NzcjOzsbZmZmammTMW2n\np+kAGGOMDZ64uDjExcVpOgyN8/LygpeXl6bD0BgfHx/4+PhoOgzG3iv8JJkxxhhjjLEuOElmjDHG\nGGOsC06SGWOMMcYY64KTZMYYY4wxxrrgJJkxxhhjjLEu+I17TCsIBAJNh8AYY+wf/MY9pg34CDim\nFWQymaZDYIwx9g93d3dNh8DYgPGTZMYYY4wxxrrgPcmMMcYYY4x1wUkyY4wxxhhjXXCSzBhjjDHG\nWBd6ANI1HQRjjDHGGGPvkv8DzPeFj9MT/JYAAAAASUVORK5CYII=\n","text/plain":["<IPython.core.display.Image object>"]},"metadata":{"tags":[]},"execution_count":7}]},{"cell_type":"code","metadata":{"id":"CTYJtYu2G1Dr","colab_type":"code","outputId":"d9d33bea-006f-4566-f680-b3af1bbaa184","executionInfo":{"status":"ok","timestamp":1567067820638,"user_tz":-60,"elapsed":26193,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":411}},"source":["model.compile(optimizers.SGD(lr=1e-3, decay=1e-5, momentum=0.9, nesterov=True),\n","#     optimizers.adam(lr=5e-4),\n","#     optimizer=optimizers.SGD(lr=1e-3, decay=1e-5, momentum=0.9, nesterov=True),\n","#               loss=[focal_loss(alpha=.25, gamma=2)],\n","            loss='binary_crossentropy',\n","            metrics=['acc'])\n","model.summary()"],"execution_count":0,"outputs":[{"output_type":"stream","text":["W0829 08:36:59.870680 139698953435008 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/optimizers.py:790: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n","\n","W0829 08:36:59.885616 139698953435008 deprecation.py:323] From /usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/nn_impl.py:180: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n","Instructions for updating:\n","Use tf.where in 2.0, which has the same broadcast rule as np.where\n"],"name":"stderr"},{"output_type":"stream","text":["_________________________________________________________________\n","Layer (type)                 Output Shape              Param #   \n","=================================================================\n","efficientnet-b7 (Model)      (None, 5, 5, 2560)        64097680  \n","_________________________________________________________________\n","global_average_pooling2d_1 ( (None, 2560)              0         \n","_________________________________________________________________\n","fc1 (Dense)                  (None, 256)               655616    \n","_________________________________________________________________\n","dropout_out (Dropout)        (None, 256)               0         \n","_________________________________________________________________\n","predictions (Dense)          (None, 1)                 257       \n","=================================================================\n","Total params: 64,753,553\n","Trainable params: 59,738,905\n","Non-trainable params: 5,014,648\n","_________________________________________________________________\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"IBSmonXLG79E","colab_type":"code","colab":{}},"source":["epochs = 5\n","steps_per_epoch = train_generator.n // batch_size\n","validation_steps = validation_generator.n // batch_size\n","history = model.fit_generator(train_generator,\n","                              steps_per_epoch = steps_per_epoch,\n","                              epochs=epochs,\n","                              workers=4,\n","                              validation_data=validation_generator,\n","                              validation_steps=validation_steps,\n","                              callbacks=[cb_early_stopper])"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"TTiNj3MAxrZR","colab_type":"code","colab":{}},"source":["acc = history.history['acc']\n","val_acc = history.history['val_acc']\n","\n","loss = history.history['loss']\n","val_loss = history.history['val_loss']\n"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"bjNZrOaLxsAp","colab_type":"code","colab":{}},"source":["acc += history.history['acc']\n","val_acc += history.history['val_acc']\n","\n","loss += history.history['loss']\n","val_loss += history.history['val_loss']"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"2f4MebwEG-B8","colab_type":"code","colab":{}},"source":["plt.figure(figsize=(8, 8))\n","plt.subplot(2, 1, 1)\n","plt.plot(acc, label='Training Accuracy')\n","plt.plot(val_acc, label='Validation Accuracy')\n","plt.legend(loc='lower right')\n","plt.ylabel('Accuracy')\n","plt.ylim([min(plt.ylim()),1])\n","plt.title('Training and Validation Accuracy')\n","\n","plt.subplot(2, 1, 2)\n","plt.plot(loss, label='Training Loss')\n","plt.plot(val_loss, label='Validation Loss')\n","plt.legend(loc='upper right')\n","plt.ylabel('Cross Entropy')\n","plt.ylim([0,max(plt.ylim())])\n","plt.title('Training and Validation Loss')\n","plt.show()"],"execution_count":0,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"9nTU1qLgHCxT","colab_type":"text"},"source":["Fine-tuning"]},{"cell_type":"code","metadata":{"id":"0CHDXfBjv5y4","colab_type":"code","colab":{}},"source":["base_model.trainable = True\n","set_trainable = False\n","# for layer in base_model.layers:\n","#     if layer.name == 'multiply_16':\n","#         set_trainable = True\n","#     if set_trainable:\n","#         layer.trainable = True\n","#     else:\n","#         layer.trainable = False"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"2tbxS2zPHHg3","colab_type":"code","outputId":"72208dbd-9238-42f2-9a0c-eba96bb9ba61","executionInfo":{"status":"ok","timestamp":1567067818270,"user_tz":-60,"elapsed":88190,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":34}},"source":["# Check how many layers are in the base model\n","print(\"Number of layers in the base model: \", len(base_model.layers))"],"execution_count":0,"outputs":[{"output_type":"stream","text":["Number of layers in the base model:  806\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"PEQkjBr9oZ73","colab_type":"code","colab":{}},"source":["# Fine tune from this layer onwards\n","fine_tune_at = 403\n","\n","# Freeze all the layers before the `fine_tune_at` layer\n","for layer in base_model.layers[:fine_tune_at]:\n","  layer.trainable =  False\n","!mkdir models"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"R7KDKgatCGhB","colab_type":"code","colab":{}},"source":["EARLY_STOP_PATIENCE = 5\n","import tensorboardcolab as tbc\n","from keras.callbacks import EarlyStopping, ModelCheckpoint, TensorBoard\n","# These steps value should be proper FACTOR of no.-of-images in train & valid folders respectively\n","# Training images processed in each step would be no.-of-train-images / STEPS_PER_EPOCH_TRAINING\n","# Early stopping & checkpointing the best model in ../working dir & restoring that as our model for prediction\n","cb_early_stopper = EarlyStopping(monitor = 'val_loss', patience = EARLY_STOP_PATIENCE, restore_best_weights=True)\n","cb_checkpointer  = ModelCheckpoint(filepath = '/content/models/EfficientNetB7-{val_loss:.4f}-{val_acc:.4f}.h5', monitor = 'val_acc', save_best_only = True, mode = 'auto')"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"3SQojQzxHILL","colab_type":"code","colab":{}},"source":["model.compile(optimizer = optimizers.adam(lr=2e-5),\n","              # optimizers.SGD(lr=5e-4, decay=1e-6, momentum=0.9, nesterov=True)\n","              loss='binary_crossentropy',\n","              metrics=['acc'])\n","# logdir = os.path.join(\"logs\", datetime.datetime.now().strftime(\"%Y%m%d-%H%M%S\"))\n","# tensorboard= keras.callbacks.TensorBoard(logdir, histogram_freq=1)\n","model.summary()"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"kmbE3udCHJ8z","colab_type":"code","outputId":"4a919963-5c09-47b3-a05b-322e4181bb9d","executionInfo":{"status":"ok","timestamp":1566382656855,"user_tz":-60,"elapsed":456,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":35}},"source":["len(model.trainable_weights)"],"execution_count":0,"outputs":[{"output_type":"execute_result","data":{"text/plain":["359"]},"metadata":{"tags":[]},"execution_count":18}]},{"cell_type":"code","metadata":{"id":"uYttZFz-4w54","colab_type":"code","colab":{}},"source":["# tbc.TensorBoardColab(graph_path=logdir);"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"qHe0ugNLIx4N","colab_type":"code","colab":{}},"source":["# %tensorboard --logdir logs"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"qNZwZ6QuHPQd","colab_type":"code","outputId":"997fd006-c1ed-439b-db33-5a8b8b177e19","executionInfo":{"status":"ok","timestamp":1567077837878,"user_tz":-60,"elapsed":10005839,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":867}},"source":["steps_per_epoch = train_generator.n // batch_size\n","validation_steps = validation_generator.n // batch_size\n","history_fine = model.fit_generator(train_generator,\n","                  steps_per_epoch = steps_per_epoch,\n","                  epochs=30,\n","                  workers=4,\n","                  validation_data=validation_generator,\n","                  validation_steps=validation_steps,\n","                  callbacks=[cb_checkpointer,cb_early_stopper])"],"execution_count":0,"outputs":[{"output_type":"stream","text":["Epoch 1/30\n","174/174 [==============================] - 1395s 8s/step - loss: 0.6613 - acc: 0.6161 - val_loss: 0.6728 - val_acc: 0.5477\n","Epoch 2/30\n","174/174 [==============================] - 360s 2s/step - loss: 0.6180 - acc: 0.6715 - val_loss: 0.6557 - val_acc: 0.5479\n","Epoch 3/30\n","174/174 [==============================] - 357s 2s/step - loss: 0.5914 - acc: 0.6912 - val_loss: 0.6159 - val_acc: 0.5899\n","Epoch 4/30\n","174/174 [==============================] - 357s 2s/step - loss: 0.5630 - acc: 0.7144 - val_loss: 0.5791 - val_acc: 0.6678\n","Epoch 5/30\n","174/174 [==============================] - 357s 2s/step - loss: 0.5324 - acc: 0.7404 - val_loss: 0.5520 - val_acc: 0.6969\n","Epoch 6/30\n","174/174 [==============================] - 357s 2s/step - loss: 0.4999 - acc: 0.7629 - val_loss: 0.5108 - val_acc: 0.7440\n","Epoch 7/30\n","174/174 [==============================] - 357s 2s/step - loss: 0.4722 - acc: 0.7892 - val_loss: 0.4858 - val_acc: 0.7748\n","Epoch 8/30\n","174/174 [==============================] - 357s 2s/step - loss: 0.4371 - acc: 0.8064 - val_loss: 0.4691 - val_acc: 0.7928\n","Epoch 9/30\n","174/174 [==============================] - 357s 2s/step - loss: 0.4130 - acc: 0.8169 - val_loss: 0.4560 - val_acc: 0.8099\n","Epoch 10/30\n","174/174 [==============================] - 356s 2s/step - loss: 0.4003 - acc: 0.8251 - val_loss: 0.4441 - val_acc: 0.8134\n","Epoch 11/30\n","174/174 [==============================] - 356s 2s/step - loss: 0.3771 - acc: 0.8341 - val_loss: 0.4351 - val_acc: 0.8236\n","Epoch 12/30\n","174/174 [==============================] - 357s 2s/step - loss: 0.3557 - acc: 0.8477 - val_loss: 0.4240 - val_acc: 0.8288\n","Epoch 13/30\n","174/174 [==============================] - 356s 2s/step - loss: 0.3484 - acc: 0.8500 - val_loss: 0.4140 - val_acc: 0.8348\n","Epoch 14/30\n","174/174 [==============================] - 356s 2s/step - loss: 0.3292 - acc: 0.8607 - val_loss: 0.4089 - val_acc: 0.8382\n","Epoch 15/30\n","174/174 [==============================] - 356s 2s/step - loss: 0.3232 - acc: 0.8617 - val_loss: 0.4082 - val_acc: 0.8442\n","Epoch 16/30\n","174/174 [==============================] - 356s 2s/step - loss: 0.3011 - acc: 0.8709 - val_loss: 0.4387 - val_acc: 0.8390\n","Epoch 17/30\n","174/174 [==============================] - 356s 2s/step - loss: 0.2943 - acc: 0.8763 - val_loss: 0.4549 - val_acc: 0.8339\n","Epoch 18/30\n","174/174 [==============================] - 356s 2s/step - loss: 0.2870 - acc: 0.8785 - val_loss: 0.4515 - val_acc: 0.8356\n","Epoch 19/30\n","174/174 [==============================] - 356s 2s/step - loss: 0.2736 - acc: 0.8847 - val_loss: 0.4145 - val_acc: 0.8493\n","Epoch 20/30\n","174/174 [==============================] - 356s 2s/step - loss: 0.2558 - acc: 0.8917 - val_loss: 0.3684 - val_acc: 0.8639\n","Epoch 21/30\n","174/174 [==============================] - 356s 2s/step - loss: 0.2486 - acc: 0.8969 - val_loss: 0.4442 - val_acc: 0.8438\n","Epoch 22/30\n","174/174 [==============================] - 355s 2s/step - loss: 0.2345 - acc: 0.9021 - val_loss: 0.4312 - val_acc: 0.8536\n","Epoch 23/30\n","174/174 [==============================] - 355s 2s/step - loss: 0.2264 - acc: 0.9066 - val_loss: 0.4214 - val_acc: 0.8562\n","Epoch 24/30\n","174/174 [==============================] - 356s 2s/step - loss: 0.2170 - acc: 0.9119 - val_loss: 0.4814 - val_acc: 0.8399\n","Epoch 25/30\n","174/174 [==============================] - 356s 2s/step - loss: 0.2045 - acc: 0.9169 - val_loss: 0.4922 - val_acc: 0.8373\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"Ahzthli0hmNw","colab_type":"code","colab":{}},"source":["acc = history_fine.history['acc']\n","val_acc = history_fine.history['val_acc']\n","\n","loss = history_fine.history['loss']\n","val_loss = history_fine.history['val_loss']"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"vLVU9191HSVt","colab_type":"code","colab":{}},"source":["acc += history_fine.history['acc']\n","val_acc += history_fine.history['val_acc']\n","\n","loss += history_fine.history['loss']\n","val_loss += history_fine.history['val_loss']"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"N7FglUXAHS87","colab_type":"code","outputId":"e5253216-a4e1-49c6-9cde-2890daa1e67f","executionInfo":{"status":"ok","timestamp":1567089214348,"user_tz":-60,"elapsed":1498,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":499}},"source":["plt.figure(figsize=(8, 8))\n","plt.subplot(2, 1, 1)\n","plt.plot(acc, label='Training Accuracy')\n","plt.plot(val_acc, label='Validation Accuracy')\n","plt.ylim([min(plt.ylim()),1])\n","# plt.plot([epochs-1,epochs-1], plt.ylim(), label='Start Fine Tuning')\n","plt.legend(loc='lower right')\n","plt.title('Training and Validation Accuracy')\n","\n","plt.subplot(2, 1, 2)\n","plt.plot(loss, label='Training Loss')\n","plt.plot(val_loss, label='Validation Loss')\n","plt.ylim([0,max(plt.ylim())])\n","# plt.plot([epochs-1,epochs-1], plt.ylim(), label='Start Fine Tuning')\n","plt.legend(loc='upper right')\n","plt.title('Training and Validation Loss')\n","plt.show()"],"execution_count":0,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAecAAAHiCAYAAADSwATnAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n","text/plain":["<Figure size 576x576 with 2 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"markdown","metadata":{"id":"CLHzoJsZQn6R","colab_type":"text"},"source":["## plot confusion matrix"]},{"cell_type":"code","metadata":{"id":"bR1knhX9IxOz","colab_type":"code","colab":{}},"source":["model.trainable = False\n","set_trainable = False"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"hkhrU4B03n7v","colab_type":"code","outputId":"5413504a-e95c-4fe6-92dc-e3d7bb1117f0","executionInfo":{"status":"ok","timestamp":1567162250633,"user_tz":-60,"elapsed":19508,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":34}},"source":["from sklearn.metrics import classification_report, confusion_matrix\n","import seaborn as sns\n","#Confution Matrix and Classification Report\n","# threshold = 0.60\n","batch_size=16\n","Y_pred = model.predict_generator(test_gen, test_gen.n // batch_size+1,workers=4,verbose=1).ravel()\n","# y_pred = np.where(Y_pred > threshold, 1,0)\n","target_names = ['nonresidential','residential']\n","# print(classification_report(validation_generator.classes, Y_pred, target_names=target_names))"],"execution_count":0,"outputs":[{"output_type":"stream","text":["40/40 [==============================] - 17s 437ms/step\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"U6vUJMmksQad","colab_type":"code","outputId":"1dc23d32-7128-4ec0-f084-b6d84c6147c2","executionInfo":{"status":"ok","timestamp":1567105076387,"user_tz":-60,"elapsed":12131,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":204}},"source":["from sklearn.metrics import classification_report, confusion_matrix\n","from numpy import *\n","random.seed(5)\n","import seaborn as sns\n","#Confution Matrix and Classification Report\n","threshold = 0.50\n","batch_size= 16\n","Y_pred = model.predict_generator(test_gen, test_gen.n // batch_size+1,workers = 4,verbose=1)\n","y_pred = np.where(Y_pred > threshold, 1,0)\n","print('Classification Report')\n","target_names = ['nonresidential','residential']\n","print(classification_report(test_gen.classes, y_pred, target_names=target_names))"],"execution_count":0,"outputs":[{"output_type":"stream","text":["40/40 [==============================] - 12s 295ms/step\n","Classification Report\n","                precision    recall  f1-score   support\n","\n","nonresidential       0.67      0.54      0.60       156\n","   residential       0.86      0.91      0.88       471\n","\n","      accuracy                           0.82       627\n","     macro avg       0.76      0.72      0.74       627\n","  weighted avg       0.81      0.82      0.81       627\n","\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"AToEvxbRveKb","colab_type":"code","outputId":"5249742e-2832-4d7b-ac74-4184ea892cf1","executionInfo":{"status":"ok","timestamp":1567105084637,"user_tz":-60,"elapsed":636,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":513}},"source":["confusion_mtx = confusion_matrix(test_gen.classes, y_pred) \n","# plot the confusion matrix\n","f,ax = plt.subplots(figsize=(8, 8))\n","sns.heatmap(confusion_mtx, annot=True, linewidths=0.01,cmap=\"Greens\",linecolor=\"gray\", fmt= '.1f',ax=ax)\n","plt.xlabel(\"Predicted Label\")\n","plt.ylabel(\"True Label\")\n","plt.title(\"Confusion Matrix\")\n","plt.show()"],"execution_count":0,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAdgAAAHwCAYAAAAFJn7CAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAwusJEk1+P97caZ6i+v2VwAAAABJRU5ErkJggg==\n","text/plain":["<Figure size 576x576 with 2 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"code","metadata":{"id":"xpTc0O1AP_eY","colab_type":"code","colab":{}},"source":["from sklearn.metrics import roc_curve,auc,roc_auc_score\n","from sklearn.metrics import precision_recall_curve,f1_score,average_precision_score\n","from matplotlib import pyplot as plt\n","fpr_keras, tpr_keras, thresholds_keras = roc_curve(test_gen.classes, Y_pred)\n","auc_keras = auc(fpr_keras, tpr_keras)"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"81CfM8NP_w0M","colab_type":"code","outputId":"062b2895-306b-44f9-bab6-2499607e1cf8","executionInfo":{"status":"ok","timestamp":1567092578397,"user_tz":-60,"elapsed":1186,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":295}},"source":["plt.figure(1)\n","plt.plot([0, 1], [0, 1], linestyle='--',label='Dummy Classifier (area = 0.500)')\n","plt.plot(fpr_keras, tpr_keras, label='EfficientNetB7 (area = {:.3f})'.format(auc_keras),marker='.')\n","# plt.plot(fpr_rf, tpr_rf, label='RF (area = {:.3f})'.format(auc_rf))\n","plt.xlabel('False positive rate')\n","plt.ylabel('True positive rate')\n","plt.title('ROC curve')\n","plt.legend(loc='best')\n","plt.show()"],"execution_count":0,"outputs":[{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAYUAAAEWCAYAAACJ0YulAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nDOEUgiAIggz/A2NqMOx9c/31AAAAAElFTkSuQmCC\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"code","metadata":{"id":"emdtnMYTvNG-","colab_type":"code","outputId":"908d98f2-fca9-4cab-adbe-bdf0a594811b","executionInfo":{"status":"ok","timestamp":1567105125157,"user_tz":-60,"elapsed":433,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":34}},"source":["# calculate precision-recall curve\n","precision, recall, thresholds = precision_recall_curve(test_gen.classes, Y_pred)\n","# calculate F1 score\n","f1 = f1_score(test_gen.classes, y_pred)\n","# calculate precision-recall AUC\n","rp_auc = auc(recall, precision)\n","# calculate average precision score\n","ap = average_precision_score(test_gen.classes, Y_pred)\n","print('f1=%.3f auc=%.3f ap=%.3f' % (f1, rp_auc, ap))"],"execution_count":0,"outputs":[{"output_type":"stream","text":["f1=0.883 auc=0.932 ap=0.932\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"3GmvUg3PIsJv","colab_type":"code","outputId":"8813755d-55e9-470e-8cf1-1a3bab94bf5d","executionInfo":{"status":"ok","timestamp":1567162255678,"user_tz":-60,"elapsed":1489,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":312}},"source":["from sklearn.metrics import roc_curve,auc,roc_auc_score\n","from sklearn.metrics import precision_recall_curve,f1_score,average_precision_score\n","from matplotlib import pyplot as plt\n","precision, recall, _ = precision_recall_curve(test_gen.classes, Y_pred)\n","average_precision = average_precision_score(test_gen.classes, Y_pred)\n","plt.step(recall, precision, color='b', alpha=0.2,\n","         where='post')\n","plt.fill_between(recall, precision, step='post', alpha=0.2,\n","                 color='b',label='EfficientNetB7 (AP = {:.3f})'.format(average_precision))\n","plt.plot([0, 1], [0.75,0.75], linestyle='--',label='Dummy Classifier (AP = 0.75)')\n","plt.xlabel('Recall')\n","plt.ylabel('Precision')\n","plt.legend(loc='best')\n","plt.ylim([0.0, 1.05])\n","plt.xlim([0.0, 1.0])\n","plt.title('2-class Precision-Recall curve: AP={0:0.2f}'.format(average_precision))"],"execution_count":0,"outputs":[{"output_type":"execute_result","data":{"text/plain":["Text(0.5, 1.0, '2-class Precision-Recall curve: AP=0.93')"]},"metadata":{"tags":[]},"execution_count":12},{"output_type":"display_data","data":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAYoAAAEWCAYAAAB42tAoAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n","text/plain":["<Figure size 432x288 with 1 Axes>"]},"metadata":{"tags":[]}}]},{"cell_type":"markdown","metadata":{"id":"lPols3pca-Hm","colab_type":"text"},"source":["Save the model"]},{"cell_type":"code","metadata":{"id":"BXJJ0QO1ZTEe","colab_type":"code","colab":{}},"source":["!cp -r /content/models gdrive/My\\ Drive/MScthesis"],"execution_count":0,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"D6FUjl3pVyLV","colab_type":"text"},"source":["load the model, rerun the model with weight"]},{"cell_type":"code","metadata":{"id":"9X9qTuL2N2Qz","colab_type":"code","colab":{}},"source":["model.load_weights('/content/gdrive/My Drive/MScthesis/MobileNetV2.0.3666-0.8656.h5')\n","# model.compile(optimizer = keras.optimizers.Adam(lr=1e-5),\n","#               loss='binary_crossentropy',\n","#               metrics=['acc',f1_m,precision_m, recall_m])\n","# model.summary()"],"execution_count":0,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"Z1YwBlKeR_-4","colab_type":"text"},"source":["# Save the model"]},{"cell_type":"code","metadata":{"id":"1eyavNZl0vi8","colab_type":"code","colab":{}},"source":["!cp -r /content/models gdrive/My\\ Drive/MScthesis"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"ZIr-Pm4LPXFF","colab_type":"code","colab":{}},"source":["model.save('models/keras/model.h5')"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"ZX9oVDo9O1kR","colab_type":"code","colab":{}},"source":["import numpy as np\n","import pandas as pd\n","import seaborn as sns\n","import matplotlib.pyplot as plt\n","from scipy import stats\n","sns.set(color_codes=True)\n"],"execution_count":0,"outputs":[]},{"cell_type":"markdown","metadata":{"id":"PE56q1uyyOyp","colab_type":"text"},"source":["## Load the model"]},{"cell_type":"code","metadata":{"id":"DYueny_ifAtP","colab_type":"code","colab":{}},"source":["from keras.models import load_model"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"V7py_aH1eUct","colab_type":"code","outputId":"ddd8197f-63e1-4fd6-d629-be028eb6bfc7","executionInfo":{"status":"ok","timestamp":1567162148646,"user_tz":-60,"elapsed":145281,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":474}},"source":["model=load_model('/content/gdrive/My Drive/MScthesis/models/EfficientNetB7-0.3684-0.8639.h5')"],"execution_count":0,"outputs":[{"output_type":"stream","text":["WARNING: Logging before flag parsing goes to stderr.\n","W0830 10:47:43.924714 139851012175744 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:541: The name tf.placeholder is deprecated. Please use tf.compat.v1.placeholder instead.\n","\n","W0830 10:47:44.067563 139851012175744 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:4479: The name tf.truncated_normal is deprecated. Please use tf.random.truncated_normal instead.\n","\n","W0830 10:47:44.107858 139851012175744 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:66: The name tf.get_default_graph is deprecated. Please use tf.compat.v1.get_default_graph instead.\n","\n","W0830 10:47:44.109041 139851012175744 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:190: The name tf.get_default_session is deprecated. Please use tf.compat.v1.get_default_session instead.\n","\n","W0830 10:47:44.111710 139851012175744 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:197: The name tf.ConfigProto is deprecated. Please use tf.compat.v1.ConfigProto instead.\n","\n","W0830 10:47:47.591369 139851012175744 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:2041: The name tf.nn.fused_batch_norm is deprecated. Please use tf.compat.v1.nn.fused_batch_norm instead.\n","\n","W0830 10:47:48.232280 139851012175744 deprecation.py:506] From /usr/local/lib/python3.6/dist-packages/keras/backend/tensorflow_backend.py:3733: calling dropout (from tensorflow.python.ops.nn_ops) with keep_prob is deprecated and will be removed in a future version.\n","Instructions for updating:\n","Please use `rate` instead of `keep_prob`. Rate should be set to `rate = 1 - keep_prob`.\n","W0830 10:48:02.763881 139851012175744 nn_ops.py:4224] Large dropout rate: 0.5125 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n","W0830 10:48:03.134909 139851012175744 nn_ops.py:4224] Large dropout rate: 0.525 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n","W0830 10:48:03.505711 139851012175744 nn_ops.py:4224] Large dropout rate: 0.5375 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n","W0830 10:48:03.868708 139851012175744 nn_ops.py:4224] Large dropout rate: 0.55 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n","W0830 10:48:04.243109 139851012175744 nn_ops.py:4224] Large dropout rate: 0.5625 (>0.5). In TensorFlow 2.x, dropout() uses dropout rate instead of keep_prob. Please ensure that this is intended.\n","W0830 10:48:43.766532 139851012175744 deprecation_wrapper.py:119] From /usr/local/lib/python3.6/dist-packages/keras/optimizers.py:793: The name tf.train.Optimizer is deprecated. Please use tf.compat.v1.train.Optimizer instead.\n","\n","W0830 10:48:43.779799 139851012175744 deprecation.py:323] From /usr/local/lib/python3.6/dist-packages/tensorflow/python/ops/nn_impl.py:180: add_dispatch_support.<locals>.wrapper (from tensorflow.python.ops.array_ops) is deprecated and will be removed in a future version.\n","Instructions for updating:\n","Use tf.where in 2.0, which has the same broadcast rule as np.where\n"],"name":"stderr"}]},{"cell_type":"code","metadata":{"id":"3UJEQvO8m0Z2","colab_type":"code","outputId":"f31d8ed9-ca8a-477d-8a50-8af6f53d1006","executionInfo":{"status":"ok","timestamp":1567162152835,"user_tz":-60,"elapsed":148014,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":50}},"source":["test_dir='/content/gdrive/My Drive/MScthesis/test/'\n","# Directory with our test non-residential pictures\n","test_nonresidential_dir = os.path.join(test_dir, 'nonresidential')\n","print ('Total test non-residential images:', len(os.listdir(test_nonresidential_dir)))\n","\n","# Directory with our test residential pictures\n","test_residential_dir = os.path.join(test_dir, 'residential')\n","print ('Total test residential images:', len(os.listdir(test_residential_dir)))"],"execution_count":0,"outputs":[{"output_type":"stream","text":["Total test non-residential images: 156\n","Total test residential images: 471\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"gWcgscgIi8XL","colab_type":"code","outputId":"e2533202-5058-449e-ef2e-27d873b71dd2","executionInfo":{"status":"ok","timestamp":1567162152836,"user_tz":-60,"elapsed":146730,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":50}},"source":["image_size = 160\n","batch_size = 16\n","test_idg = ImageDataGenerator(rescale=1./255)\n","test_gen = test_idg.flow_from_directory(\n","    '/content/gdrive/My Drive/MScthesis/test/',\n","    target_size=(image_size, image_size),\n","    batch_size = batch_size,\n","    shuffle = False,\n","    class_mode='binary',\n",")\n","len(test_gen.filenames)"],"execution_count":0,"outputs":[{"output_type":"stream","text":["Found 627 images belonging to 2 classes.\n"],"name":"stdout"},{"output_type":"execute_result","data":{"text/plain":["627"]},"metadata":{"tags":[]},"execution_count":7}]},{"cell_type":"code","metadata":{"id":"AYdzwbDnjoyb","colab_type":"code","outputId":"ed81ebc9-74fa-4c70-c702-48975d39ee20","executionInfo":{"status":"ok","timestamp":1567162230099,"user_tz":-60,"elapsed":220469,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":34}},"source":["test_steps = test_gen.n // batch_size\n","loss, acc = model.evaluate_generator(test_gen, steps = test_steps, workers = 4, verbose=1)"],"execution_count":0,"outputs":[{"output_type":"stream","text":["39/39 [==============================] - 78s 2s/step\n"],"name":"stdout"}]},{"cell_type":"code","metadata":{"id":"5z2EA-K3IpP6","colab_type":"code","outputId":"a01af717-1d83-4e02-ad11-5e1764b94f66","executionInfo":{"status":"ok","timestamp":1567162231116,"user_tz":-60,"elapsed":1011,"user":{"displayName":"yuan li","photoUrl":"https://lh3.googleusercontent.com/a-/AAuE7mCpnhepyXFu9kSJ63OIuZ-f8GLO3Z90r-0iMADZ=s64","userId":"01872871457478507435"}},"colab":{"base_uri":"https://localhost:8080/","height":34}},"source":["acc,loss"],"execution_count":0,"outputs":[{"output_type":"execute_result","data":{"text/plain":["(0.8173076923076923, 0.4805700835079337)"]},"metadata":{"tags":[]},"execution_count":9}]},{"cell_type":"markdown","metadata":{"id":"6uywmIniSQ5z","colab_type":"text"},"source":["Make predictions on sample test images"]},{"cell_type":"code","metadata":{"id":"bTKQDeAjSRwF","colab_type":"code","colab":{}},"source":["validation_img_paths = [\"validation/alien/11.jpg\",\n","                        \"validation/alien/22.jpg\",\n","                        \"validation/predator/33.jpg\"]\n","img_list = [Image.open(input_path + img_path) for img_path in validation_img_paths]"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"PmgS1LW6SXaY","colab_type":"code","colab":{}},"source":["validation_batch = np.stack([preprocess_input(np.array(img.resize((224,224))))\n","                             for img in img_list])"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"ffr_SJvRSX1w","colab_type":"code","colab":{}},"source":["pred_probs = model.predict(validation_batch)\n","pred_probs"],"execution_count":0,"outputs":[]},{"cell_type":"code","metadata":{"id":"ph8ASM4gSZ4A","colab_type":"code","colab":{}},"source":["fig, axs = plt.subplots(1, len(img_list), figsize=(20, 5))\n","for i, img in enumerate(img_list):\n","    ax = axs[i]\n","    ax.axis('off')\n","    ax.set_title(\"{:.0f}% Alien, {:.0f}% Predator\".format(100*pred_probs[i,0],\n","                                                            100*pred_probs[i,1]))\n","    ax.imshow(img)"],"execution_count":0,"outputs":[]}]}