{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# Deep Pool-based Active Learning: Scikit-activeml with Skorch" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "> **_Google Colab Note:_** If the notebook fails to run after installing the needed packages, try to restart the runtime (Ctrl + M) under Runtime -> Restart session.\n", "\n", "[![Open in Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/scikit-activeml/scikit-activeml.github.io/blob/gh-pages/latest/generated/tutorials_colab//01_deep_pool_al_with_skorch.ipynb)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "**Notebook Dependencies**\n", "\n", "
\n", " Disclaimer: This notebook is currently not compatible with skactiveml=0.6.0. As of now, skorch=1.1.0 does not fully support torch_load_kwargs. Therefore, we resort to installing skorch from source.\n", "
\n", "\n", "\n", "Uncomment the following cell to install all dependencies for this tutorial." ] }, { "cell_type": "code", "execution_count": 1, "metadata": { "ExecuteTime": { "end_time": "2025-04-07T10:04:45.744998Z", "start_time": "2025-04-07T10:04:45.739776Z" } }, "outputs": [], "source": [ "# !pip install scikit-activeml skorch torch torchvision torchaudio tqdm pandas\n", "# !pip install --upgrade git+https://github.com/skorch-dev/skorch.git" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "
" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "In this brief tutorial, we show an example use-case of our package `skactiveml` with the Python package [skorch](https://skorch.readthedocs.io/en/stable/), which is a scikit-learn wrapper for Pytorch models. This way, we are able to implement and test deep learning models in combination with query strategies implemented in our framework." ] }, { "cell_type": "code", "execution_count": 2, "metadata": { "ExecuteTime": { "end_time": "2025-04-07T10:04:50.043805Z", "start_time": "2025-04-07T10:04:45.936107Z" } }, "outputs": [], "source": [ "import matplotlib as mlp\n", "import matplotlib.pyplot as plt\n", "import numpy as np\n", "import torch\n", "import warnings\n", "\n", "from copy import deepcopy\n", "\n", "from torch.ao.nn.qat import Conv2d\n", "\n", "from skactiveml.classifier import SklearnClassifier\n", "from skactiveml.pool import UncertaintySampling, QueryByCommittee, RandomSampling, BatchBALD, DiscriminativeAL\n", "from skactiveml.utils import call_func, unlabeled_indices\n", "from sklearn.datasets import fetch_openml\n", "from sklearn.ensemble import VotingClassifier\n", "from sklearn.model_selection import train_test_split\n", "from sklearn.preprocessing import StandardScaler\n", "from skorch import NeuralNetClassifier\n", "from torch import nn\n", "import torch.nn.functional as F\n", "from torchvision.transforms import transforms\n", "from tqdm import tqdm\n", "\n", "mlp.rcParams[\"figure.facecolor\"] = \"white\"\n", "\n", "MISSING_LABEL = -1\n", "RANDOM_STATE = 0\n", "FONTSIZE = 12\n", "\n", "np.random.seed(RANDOM_STATE)\n", "torch.manual_seed(RANDOM_STATE)\n", "torch.cuda.manual_seed(RANDOM_STATE)\n", "device = 'cuda' if torch.cuda.is_available() else 'cpu'\n", "\n", "warnings.filterwarnings(\"ignore\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Loading Digit Data Set\n", "For simplicity, we use the `sklearn` function `fetch_openml` to load the MNIST dataset. The dataset contains 28 x 28px images of handwritten digits from 0 to 9 and the task is to recognize the digits in the images." ] }, { "cell_type": "code", "execution_count": 3, "metadata": { "ExecuteTime": { "end_time": "2025-04-07T10:04:59.306288Z", "start_time": "2025-04-07T10:04:50.048955Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "# Load digit data set.\n", "X, y_true = fetch_openml(\"mnist_784\", version=1, return_X_y=True, as_frame=False)\n", "\n", "# Visualize first 10 images.\n", "fig, axes = plt.subplots(figsize=(15, 5), nrows=1, ncols=10)\n", "for i in range(10):\n", " axes[i].set_title(f'Digit: {y_true[i]}', fontsize=FONTSIZE)\n", " axes[i].imshow(X[i].reshape(28,28))\n", "plt.show()\n", "\n", "# Standardize data.\n", "X = StandardScaler().fit_transform(X)\n", "\n", "# Reshape samples to n_samples x n_channels x width x height to fit skorch\n", "# requirements.\n", "X = X.reshape((len(X), 1, 28, 28))\n", "\n", "# Set data types according to skorch requirements.\n", "X, y_true = X.astype(np.float32), y_true.astype(np.int64)\n", "\n", "# Identify list of possible classes.\n", "classes = np.unique(y_true)\n", "\n", "# Make a 66-34 train-test split.\n", "X_train, X_test, y_train, y_test = train_test_split(\n", " X, y_true, train_size=0.66, random_state=RANDOM_STATE\n", ")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Data Augmentation\n", "Data augmentation is used to artificially generate data from existing data points. In case of image data, this can be done by rotating or scaling the image. Other basic transformations are:\n", "* padding,\n", "* vertical and horizontal flipping,\n", "* translation,\n", "* cropping,\n", "* darkening, brightening, color modification,\n", "* grayscaling,\n", "* changing contrast,\n", "* adding noise,\n", "* random erasing.\n", "\n", "For more information on data augmentation take a look at the [Data Augmentation Tutorial](https://www.tensorflow.org/tutorials/images/data_augmentation) of [TensorFlow](https://www.tensorflow.org) or visit [Transforming And Augmenting Images](https://pytorch.org/vision/stable/transforms.html) on [PyTorch](pytorch.org)." ] }, { "cell_type": "code", "execution_count": 4, "metadata": { "ExecuteTime": { "end_time": "2025-04-07T10:05:01.902630Z", "start_time": "2025-04-07T10:04:59.355283Z" } }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAABLkAAACkCAYAAAB7PPybAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjEsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvc2/+5QAAAAlwSFlzAAAPYQAAD2EBqD+naQAAY51JREFUeJztvQmcXFWZ/n+qqqur9046SW9ZCLIFSEggEBJAEEVRUUEZt1FZZMAFVEZn/A0OKAhjXGYcRwXEDYRx1ME/KKDgYFgDCYFAdtKQfe3udKf3vavu/3Nu7Drv+96+J1WdTnfd28/386nk3jq37vacc+65p8/7nIjjOI4CAAAAAAAAAAAAACDARMf7BAAAAAAAAAAAAAAAOFLQyQUAAAAAAAAAAAAAAg86uQAAAAAAAAAAAABA4EEnFwAAAAAAAAAAAAAIPOjkAgAAAAAAAAAAAACBB51cAAAAAAAAAAAAACDwoJMLAAAAAAAAAAAAAAQedHIBAAAAAAAAAAAAgMCDTi4AAAAAAAAAAAAAEHjQyZUFt956q4pEIiP67X333ef+dseOHepoofetj6GPBcBY5j0w9kDXcAJdwwl0DS/QNpxA13ACXcMJdA0n9x2BrhOmk2vjxo3qk5/8pJo+fbpKJBKqtrZWfeITn3C/B+PLXXfd5Wbgs88+W01Uuru73U7UZ555RoUF6Apdwwp0DSdh1FUDbcOpLXSFrmEFuoYT6BpOunNU1wnRyfXQQw+pM844Qy1btkxdffXVboa85ppr1NNPP+1+//DDD2e0n5tvvln19PSM6Bw+9alPub895phjRvT7MPPrX/9azZ49W61atUpt2bJFTdQK4rbbbsu5CuJIgK7QNaxA13ASRl010Dac2kJX6BpWoGs4ga7hpDtHdQ19J9fWrVvdDqa3vOUtat26deqOO+5wO7huv/12d11/r9O3bdvmu4+uri73/7y8PFVQUDCi84jFYu5vRxruGFa2b9+uXnzxRfX9739fTZs2za0sQPCBruEEuoYT6BpeoG04ga7hBLqGE+gaTqBrbhP6Tq7vfe97bg/jT3/6UzcDUqZOnaruuecetxPru9/9LvPd2rRpk/r7v/97NXnyZHXeeeexNIoenfXFL37R3Vdpaan6wAc+oPbu3etup7e3xZTqnt/3ve99avny5WrRokVuJ5judLv//vvZMQ4ePKj+6Z/+Sc2bN0+VlJSosrIy9Z73vEetXbtWBR1dIeh7fMkll6i/+7u/81QQuldY3zfZO+znP/bggw+qU045xb2Xc+fOdUfpXXXVVe69lr/993//d3XnnXe697yoqEi9613vUrt371aO47idoDNmzFCFhYXq0ksvdTWQPP744+qtb32rKi4udrXX1yDDX/WxtWY6T1x22WXuss6HWs9kMpk+n6G8qXvC9bnJ/LN582b3/lRUVLjXduaZZ6pHHnnEc076+G9/+9vd89bnrzt1U6mUGmugK3SlQFfoCl3HXlcNtA2nttAVulKgK3SFrtAVugqckFNbW+vMnj3buo1OnzFjhrv8jW98w9G35ZRTTnEuvfRS56677nLuvPNOlkb5yEc+4n73qU99yt1Or8+fP9/9Tm8/xL333ut+t3379vR3xxxzjHPSSSc5VVVVzte+9jXnxz/+sXPGGWc4kUjE2bBhQ3q7l19+2TnuuOOcf/mXf3Huuece55vf/KYzffp0p7y83Nm7d296O71vfQx9rKAwZ84c55prrnGXn3vuOff8V61alU5/+umn3e/0/5ThrvWxxx5z791pp53mfP/733duueUWZ/Lkyc7cuXPdey1/u2DBAldnve3NN9/s5OfnO4sXL3a1OOecc5wf/vCHzhe/+EV3n1dffTU7/v333+9+/+53v9v50Y9+5HznO99x89GkSZOYxldeeaVTUFDgnHrqqc6nP/1p5+6773Yuv/xy9/g6b2k6Ozvd7/V3H/zgB50HHnjA/axdu9ZN13lBa63PVR9H55Pzzz/fPf5DDz2UPtb+/fudadOmudd86623Ot/73vecE044wb0fMu8dbaArdKVAV+gKXcdeVw20Dae20BW6UqArdIWu0BW6ckLdydXa2ureGN1ZZeMDH/iAu117e3u6I+vjH/+4ZzvZybV69Wp3/cYbb2TbXXXVVRl3cunvdMEYorGx0UkkEs5XvvKV9He9vb1OMplkx9D70dvpDq+gdnK98sor7vk++eST7noqlXI7G7/0pS+NqIKYN2+e+/uOjo70d88884y73XAVhC5MOo8McdNNN7nf607KgYGB9Pc6L+jKQ+ug0fvXFcG1117Lzqm+vt4tyPR7XUHofVKdNKeffrqzcOHC9PqBAwc8eWaId7zjHe61DR1/6F7pSkxXAEPofKj38dJLL7H8pM9pLCt+6ApdoSt0ha7jq6sG2oZTW+gKXaErdIWu0BW62gl1uGJHR4f7vx6GZ2Movb29Pf3dZz/72cPu/4knnnD///znP8++/8IXvpDxOephiXq44BB6yN9JJ53EPML0bJDR6CGp9PDA5uZmd8ig3u7VV19VQUUP66yqqlIXXnihu66HN370ox9Vv/3tb9PDIDNl3759av369eqKK65w780QF1xwgRvmORwf/vCHVXl5eXp9aGYMPQun9l+j3/f397vDNTVPPvmkam1tVR//+MdVU1NT+qN91/S2ekIDicxPWnObD9wQeojpU089pT7ykY+4+XnoWDoPXHzxxerNN99Mn9ef//xntXjxYjf0leYnPYvoWAJdoevhgK7QdQjoevSAtuHUFrpC18MBXaHrEND16ABdt+W8ruYuhJChzquhzq5sOsOOPfbYw+5/586dbueT3Pb444/P+BxnzZrl+U7H97a0tKTXdTzqf/3Xf7mzQmqTO1p4pkyZooKIvgZdEejKQV/TELqA/cd//Ic7E6aOL84UrYXfvdffDdcZKO/9UGUxc+bMYb8f0kQXSo2OGx4O7ZlG0fHH0g9OauyHnqlDj7i85ZZb3M9wNDY2qunTp7v3YLgpbHVn6FgBXaFrJkBX6DoEdD06QNtwagtdoWsmQFfoOgR0HX2g6+RA6BrqTi4tbE1NjTuLog2drm8wFVabno0Fuud0OHSmGOJb3/qWmzk+/elPu2Zy2rhNd67deOON42Zme6Tont39+/e7lYT+DNdDrisIv9kos+0lz+beH06ToXv+wAMPqOrqas92tAfdtr9MGDqWNvnTvd7DkU2n6tEGumYGdM0e6Dr6QNdw6qqBtuHUFrpmBnTNHug6+kDXzICu2QNdJ3gnl0bPXvizn/3MncFwaJZEyvPPP+/ODPCZz3wm630fc8wxroC6F/eEE05gPZejye9//3u3t/gXv/gF+14PN9SzOgYRXQFUVla6M0NIHnroIXdGiZ/85Cdub/HQtQ7X60218Lv3o63Hcccd5/6vz/+iiy4alX36VYR61gxNPB4/7LH0PRjqoafU1dWpsQK6cqDrIaCrF+h6ZEBXO9A2nNpCVw50PQR09QJdjwzo6g90DYauofbk0vzzP/+zOypLd2LpGFAZK6rjTPXUm3q7bBnqldRhhJQf/ehHajTRvah0ZNfQNKNDcaxBo6enx60EdAeknlJUfm644QY3hFRPL6ozvb7+5557ju1D3vPa2lp3utX7779fdXZ2pr9/9tln3Tjn0UTrrkf96RF2AwMDnvQDBw5kvU+dB4erCHUl9La3vU3dc8897l8NbMd673vfq1auXKlWrVrF0uWUtkcL6OoFuh4CunqPBV2PDOjqD7QNp7bQ1Qt0PQR09R4Luh4Z0HV4oGtwdA39SC49wupXv/qVa1ymzduuueYa10NLj97SI6O0AdpvfvObdM9mNixcuFBdfvnl6gc/+IHbgaYN03SGfOONN6w9m9miC9I3v/lNdfXVV6tzzjnHzfBa9KEe0qChC76uAD7wgQ8Mm67vo47/1deoTfy0uZ7uONT3U+v02GOPuTG8El1gL730UnXuuee690rHC//4xz92Kw5aaRwpunK4++671ac+9Sl1xhlnqI997GPu+e7atUv96U9/co+vj5sNuiNWT0Lwu9/9Tp144oluSKo+b/3RfynQoxB1/r322mtd3RsaGtSKFSvUnj171Nq1a919fPWrX3WHn7773e9WX/rSl1RxcbH66U9/6layhwvZHQ2gqxfoCl2hK3QdS1010Dac2kJXL9AVukJX6ApdM6dsIunqTBDWrVvnTqNZU1PjxONxp7q62l1fv349205Pf6lvi54OUzKURunq6nKuv/56p6KiwikpKXEuu+wyp66uzt3u29/+dno7PU2onAJTTwl6ySWXeI5zwQUXuJ8h9LSbX/nKV9xzLywsdM4991xnxYoVnu2Gm5I0F3n/+9/vFBQUuPfOj6uuusrVqampydXi8ssvd4qKipzJkyc7n/nMZ5wNGzYMe62//e1vnTlz5jiJRMKZO3eu88gjj7i/1d/J+/S9732P/XZoqtcHH3yQfT+k3csvv+zZ/uKLL3anN9XXc9xxx7nnraeVpdOvFhcXZ5SXXnzxRXdKVj3Vq5yKdevWrc4VV1zh5lt9X6ZPn+68733vc37/+9978rnOE/p89Da3336784tf/GJMptWFrtAVuh4CukLX8dJVA23DqS10ha7Q9RDQFbpCV+h6OCL6n5F1jwE/1qxZo04//XT13//93+MybTjgLFiwwO2l1tOmgvAAXcMJdA0n0DW8QNtwAl3DCXQNJ9A1nEDXkRN6T66xiM2V6PBFPfvh+eefPy7nNFHRscWDg4Psu2eeecYdCqljgkEwga7hBLqGE+gaXqBtOIGu4QS6hhPoGk6g6+iDkVxHyG233aZWr17tzn6op918/PHH3c91113nGq2BsUP7rOnZGz75yU+6Jn6bN292Z7coLy9XGzZsUFOmTBnvUwQjALqGE+gaTqBreIG24QS6hhPoGk6gaziBrqNP6I3njzbaCF4PIbz99ttdY7hZs2apW2+9Vf3rv/7reJ/ahENP1aonA/j5z3/uzsigTesuueQS9e1vfxuVQ4CBruEEuoYT6BpeoG04ga7hBLqGE+gaTqDr6IORXAAAAAAAAAAAAAAg8MCTCwAAAAAAAAAAAAAEnqPWyXXnnXeq2bNnq4KCAnX22WerVatWHa1DgTEEuoYT6BpOoGt4gbbhBLqGE+gaTqBreIG24QS6ThyOSrji7373O3XFFVe4hmk6A+nZBh988EFVV1enKisrrb9NpVJq3759qrS0VEUikdE+NZAFOmt0dHS4Bnh6tsgj0VUDbXMD6BpOoGt4GU1toWvuAF3DCericAJdJ4auGtTF4QDP2IlTZv02HHUWLVrkXH/99en1ZDLp1NbWOkuXLj3sb3fv3q073fDJoY/W5Eh1hba594Gu4fxA1/B+RkNb6Jp7H+gazg/q4nB+oGu4dUVdHL4PnrEq9GV2OEZ9dsX+/n61evVqddNNN6W/071selrMFStWeLbv6+tzP7R3TnPuwn9SebHEaJ8eyILBZJ96YfW/u73W2epq0/acRV9VeXnQdrwYHOxTL6767qjreu6Z/wxdx1nXF1753qjruvDir6m8eMEYXAHwY3CgV63+y7dGpK2frkvO+X9HXl4P7Sp78EfQdJld8eJ3ck9XMG662rRd9LabVF4e6uLxYnCwV616Zumo63r2Bf8CXcdZ15ee/barq2a06uKz3o7ymgvavvzUyMqs7zvsYrzD5sR77MpD77E2Rr2Tq6mpSSWTSVVVVcW+1+ubN2/2bL906VJ12223eU8slkDlkCPoYZnZ6mrVNg/a5gLQNZyMuq7xAnRyBVjbo1pe0ckVTl1Bjj5jUReHVleU2XFnKARt9OpilNdcAc/YcHK4sNFxn11R96i2tbWlP7t371a5RMRxxvwTFnJdWxAwXXXZ+Nsnkkz5fwaSvp9of+Yf62/7BtmHbSvOR6Uc83Esn3EG5TWcHLGutsHiGeJ5xo37IPcMPzlMTpbXMdAkknIy+gRZ55zUFhwxE0ZX2ubR5TFpPjLN+gkIE0bXCQZ0DT6jPpJr6tSpKhaLqYaGBva9Xq+urvZsn0gk3A/IbbLVVQNtcx/oGk6ga3jBMzacQNdwgro4nEDX8IK6OJxA14nHqI/kys/PVwsXLlTLli1jMxLo9SVLloz24cAYAV3DCXQNJ9A1vEDbcAJdwwl0DSfQNbxA23ACXSceoz6SS/PlL39ZXXnllerMM89UixYtcqfo7OrqUldffbXKFTIOCxRDZiO2n41SyBGLMBXxpo4t/DRqEp2jML1pEHQFAdfVVoYc/3I5WJKfXo7qEEFCrLM/vRwZ5GnZxHrbylSU7NeJRfz3I/dJtnXknxyOsAznlK6ZIuWnt2C8oxdyyEcqcM9YUewijvnCDS1jifQ5Jn/nfzzrs9FWlkS5Y+XcGdv8kDO6jrCsueFIvvsc/QJsbY9F/dc9dflR1jlndB0tbFoehbZnrhJIXVNjUA5llU7yi8wdrCzKYmk7V/K+czQIpLa5RKbZbIyrC+g6sTgqnVwf/ehH1YEDB9TXv/51VV9frxYsWKCeeOIJj9kbCBbQNZxA13ACXcMLtA0n0DWcQNdwAl3DC7QNJ9B1YnFUOrk0N9xwg/sB4QK6hhPoGk6ga3iBtuEEuoYT6BpOoGt4gbbhBLpOHMZ9dkUAAAAAAAAAAAAAAHJ2JFeu4fHgssR6s/hy+TviOWH19ZJp2XgYME8S4e9DuyWjUd/4dbadPLUJ5Jlg5Sh4hBwRE0UXy31PFpoqqXO68dnSDBaY5Zix2XI5cP5AejnaxmdDOebxWHq5YE87S4v09PmfZowXokjM7EeJNIeUxeiA41venDjZh1uHEM3z5D5DlDds1V8q0+2k0cfIjpcV1J7JU0+TNNufi4ImnTM6vluUne8lhVcp9caVd6eXX+jlP7zij59PLx/3+16WFu1PZnxutGjJssy0E2ZeTOZs/LpCrPNI/Eyl/yHPK1kc3OK9Iz3AmJRSc7qfHPLrylksOns8mGjb0/I7j2/eCH2VrP57kqA/O0cDS3mL0jJk8dI6suP7J0Wi5hiOzA+Z+nUdZX+uwDMGrzpeXzYnozJoPTXIaidzi2E1UYc0TdDLBgAAAAAAAAAAAABhAp1cAAAAAAAAAAAAACDwhDpc0TY8noa+FH2/gaU9dPyT6eVvNZ3E0p5smJNebnl0OkurfYLsJyWG6ovjs2G5cignCX2KyLAoGtJEhvm6xPzjZ+jwbhnKGLrwxUzDKeQQbjq61rYP+TvLUGnrvfUk+Q/vtQ7PZyE0crx5DmjrCd31v3ctJ5gQxXd+ZgVL+07VmvTyE908JPGX9eell1/e9BaW1jTP7HNy/iSWVlzXZFYOtvLzjIrQwvy4WckTaaTMOnFerbIyLMueMvuJiPHetqm1QxOeKMph/ZU8PO3qk1emlzd21rC0V/48N708/elulpZKEH2cbEISeZITyyzEyfNDWyhj0LS0hv2TNFEvdteYMlp8SgtLu7lxXnr5f18/g++y0Ox033lFLK3mxZ70cryxw14P0zIpQ4HJugxlZGkylDGWYVhbgDU+bBgqSbOFqMi8Eu0Z8H0myLBDtkt5n2mZlHYOZL8ezWmouNCV6ux5ZNpCGcOGTVtaT5/Nn78OeeQVkEeqpnyb0T2/jfsMOHH/v7Nbw8OtaTb9RFqQ9czUcuUwbdtofyqjtprV8kXUvbZyIvfD3oVkHU7rYpEW8QtdHG4/YSErKwHlX986I8xH2ZwbswsQm7KmreVdJ+iWAEcj7NBWR9tIZhY+mtX7ZgCGTeXgKQEAAAAAAAAAAAAAkB3o5AIAAAAAAAAAAAAAgQedXAAAAAAAAAAAAAAg8ITKkyti83kQafvfWpZevnrKcpZ27roPpZf37Z/M0qItxpcnPoUfv/GtlenlaavbeOLeA/7nLae6jvt7/yjq9yPTmM+IPEbEP8w5iJ4T2Uxv7fc7GcY/QKanFx4h2UylTO+h9D/jUyL7x6JLnxi2qfydLVA/w2l8jyrSJixpAtAd4XtF+d+XFrH11jMK08vLtnCvvOguk1bY5X9fO2v58VLxaenlstcGedqBZn5CfeS3MeHJVWA8SuRddiJxf+8ZOn22+B3zl/EkRoLlw5X0925qOcHcn83n/Yyl7R/sTC//aNBorFk1z3gy1feVsrSql423V157nzgZ/2nKUzbvpjz+uxStU+XvLH5drG7OQRmzwuJ7mYybi2ttKmFpv1/z1vTytI3ckCJpLPRUj3jGNiwyeWDqel4GE8s3sfVoxeTh/fT0aSfMupPPm0FOKuavq2Px12OemDwtJ3R2Rua1tvdtotx95vvp5ZJoAUv7WsNp6eU3Ok17SLPpLyeml4sa+PGmvtqeXo41cG/EiCyvpA0k/Q9Zm8jWRiB+be6mtCxLjzYqnv/jKphYvHrks6qz1hTMTdffxdLW9Zv69us7L2VpW/90XHq5+iX+jM1r6/M9n1RClEuiCytrGqqf9IeilxGztJ1y8ZkqIXWs1XdLehDTmyDbICQt1sU90+i2kWTS91yUzTvY1SvmW2/S8p2ivnmHvjHbSe2oXlJzejOC7s+V4auHR3OqnfRBpb5O0pOLbnsYj1/alo8O8IOkiN9eKj9mqW/lhZDtouH1vbTdZ+opGx0U/t79/r6X1vrBgsdSltx3T5+AZWgUO50c8e7CSC4AAAAAAAAAAAAAEHjQyQUAAAAAAAAAAAAAAk/gwxX5VNf+w3DlEN2eSrP+izfOYWmd9Sa0ouZZ3g+Y12OGDvbzCBnVV2a2bVxUztKmbODhErFX68xp0vBEEVoRKSjwD2Hih+fD+sX4VIf0Z8rh/9nMQhqEkMT22eae9Zfxa83vMD/M6+E7KWwyQ7XjjSZMyjMMdIAPufdMbUzvrwwppb+TIW8sNMp/mmVbKKNneG+OaxfpJdPKK6UmbTH3pOV0fp1/rZuTXi59mYfQTF1nwiUkA6Wmmot3cu16ppkQjPYzalha8W5ehiPb96WXnR5xPKJJROjK80DY4l2GxzM8npYfGQYz22z8neYTWNpPXrgwvVy0S4SvFJn9dNfwA+4939QBtct5mUis28VPrtzU91ER1paiYW1kCLl7HSS0wjPiX4Q28h/S0BqZpnI7lM2qK0+MkuiWeAO/rzP/2mPS3tjLd0rKVqSUhzn2nFqbXm45nsQ1KqUqBk9m6/kvbjT7KS8T512Q0RTqEjZ0Xw7xd3KwHs4UoWvfFJLvTzNhwZrHukw9ef++JSytqbs4vdzawevowUqTIaKDvCwdONPoM+0Vfi7RN3fzLxKJYcPE3XMtMHkiIp69tN5xPGFRJESVhJD/7ZcqVFimoGchTqKePrDYZJJlPfz+/cOzN6SXo228no5VmP00zRPP7fX8+Plv7Df7Edqmigt99XNIaJRMo2FTnnqOJnnCpnJAd0/YoVnumMHrv7d+8aX08t9XrGRpj7UvSC8PiAv9nxfM+8/xv+aHTxYYLQu2ccsV2gaK5HHNIyKM2EnkW0KMjQjycegk/S0BWIikJ6otB7TLhhG+hNHn7/RvbGFp11U/k15+qOVMlvb442ell499uD1jm5+IDJ3r6fPVNULCjSMFsl2Vl5ENhCRwz1XLM5beW1p/abZ9wlzn9nffy9LeGOhKL9/TZGwfNP/328Xp5Zl/FnYrlncyR4Tv0zLp2ZbqYwv7lyHEdD2VeQjkkYKRXAAAAAAAAAAAAAAg8KCTCwAAAAAAAAAAAAAEHnRyAQAAAAAAAAAAAIDAEzhPLuuUmdJbg02hygNiB8vM+kAD9/2gPlyTn93B0px+49vk9PLpj6OVU9PLnXOrWFrLnCK2XpEy02lHVm8W15Tyn7KVrHunwaYGA5bYZct04XKK5VzBO12y+aK/nHsTnHHjmvTyXdO5N8GdrTPTy483zmVp2/9ybHp5xtN8uuTIgFmPNXHdIwPCP6unx9erQFH/NeL3cyiNbEumsT900JjvvfBMp51jMeyyzNKsKc8uRqYhjnBLDpWoM54c1S9yH4Hozgaz/17ul5VPNHAGxRTmx89KL/fMMH4ymsaFvF4orTZToZe8uodfU3ePv48B8SpyLHHxgcPm4yS9REgdEx0U+aHC1Kk/ed54cGmO/43xbMvf08B/19mdXk4dw+vbhrONx0/9Ip6Rqh2juSbx6hZfD8RomckDjiP8f0hdLC0NnAj1C5HeBGRl/Itndsj8S/31RFK8w9SZZdt4PZi/xXjvOP3cl88hU9U7La0sLfHcwfRyed48lnbgdK5PbYspr5HtorySZfkcpV4v0qsiEqUZWfgCqWBh815J5pvrHhzg1/m/DWcO68Gl6Vw5Lb08eSff5wBpAvVzu0MV7zTbHpzH/dOm9lSzdWcn8XBzhEcMrd+JD5C7KfXbk89MWigtbY3gFdjD4KmnHV//naIa41N63aPXsrTZj5v7Husz9bn0xOys4Xnp4Bxe307tMfkn+jpve0d7TV5ziD/XoXWzH9sj1uP/Q+tpWbWNl9QZegS2nMrX/6Pm1fTyyl7hS0by7TP13PdSlZr6d/c7eXmOkOI0ZRL3LC1bsdPsv5379tm8hNn7jSYZ82/3En816UU7UWDPVZG5O2aaOu2KyhdYGrElVfcO8PKSPNa0kfe8k1fGNctNuyp/az1LS3UaP6hDG/j7R6ukeR7Tx6YkJbohmKexx6SNLAatKpbvBaScd9bwZ9X02sb08o37uZ/akztOSi93N/F+hVi1udEN51WwtMqX2tLL0QO8XSV9uplvbMLfM9zjbUnXpa8X3ecYajcxaw0AAAAAAAAAAAAAECrQyQUAAAAAAAAAAAAAAk/gwhUPG8bG0kiimA45Um6GVBev4UM5Jy83w3CdARFKQUMrSFiFu7rPDO0sJMua1CVmCl9N8zwzLLiqsZZvW2+GKioxLTs7ppPnP8xZDg+V0yMHYTinZZguHUrffCofTpkfNWOs31v3Xpb2et2M9HK8hd+UKBlt23AWD1WreN2EKKaq+RDRiAi/Kly93Wzb0cm3JdNis2nssximzsKd3B35h0axae2HxvqORcjcCA8R7Uv6Xktho9lpbI+YzpqGqYjh8DTEWBLb35ReLmnlWjkxHgLXfowpb/ntPC2+iYdDjXrIaABmtachibaQHxnKmL/TlInK1bxOzd9l9HG6zDB6F1IXR+pMna2p3V+aXt7zd7NZ2v4lPKxtRq9Jj29v8J0iW+WJ+oIMzU6K8KdIzFIUclA7K5ZQA7/p5jXdlWY93i1CVGXIit/087Ru05CQlaItJnRR035MJVvfS8IwZjwhws/3k+m15bnQvCpDrX3PmoevBsLqwVImY/3mngy28VCKDVETthR5k4c3HbPMhLPEt/E2EGtLVRlrB03raSa0IiWaNQdP52EXU9o6fcOk2DXJMFRmB6AmDpZ2FauzRT7YeYkIceo2+fvYP/JnauKN+uFD93UdTpaLp05maQfP5mW28UxTb9c2TeGX0WTKu6ekRS1tIDatfcz3em22D+OFTR8nynV99+ZL0ssnlpF3CKXUo2vmp5crVvH28mRy2f3m9h86BrldbbNFCGTsmPRy+VNv8h/29fm2gah9gwsNI5YWLMLRY0JCyq+0eugj4j3ScgZLe7LdlNHlz3FrlkSL+V1/Gd/nnovM+01VubF30eR18+dofFWd72mzEDhit+JeRz/RWbaPSVJStLmC1ia2QctyMs5P/mCbef98pPE0ljZ5uWm/TmngetDy2zON77NhiWkPVa4SdeTr2/l6oXk3jTgiNJxu57HiMOJFROcMq6/GsFGMkVwAAAAAAAAAAAAAIPCgkwsAAAAAAAAAAAAABB50cgEAAAAAAAAAAACAwBN4Ty6GxWtI+lGkesylx6RlD/X3kfHjlDgPGI/SKW6Fl0hpHZ+ys/NC4zew/2I+NW/1U0SWRuIdMlpI75DR8Aw6Glg8Q+j6ILfIUn/dSaZY3c1NBmY8a36XOMh9A5IFJv67t4LHgvdONVr3lYk4ZBE27kTfkl4ueoHHrDu9ff7eEdSPQEyXzNKEFwPtqnYC7qMX7SE+S81c2Hwyzbzj8ZdJDe/p437hH//PvLyEl0jJRuMHpemdZHy4+suEDxz1lbBMx+tJo0hvIHreYurznIR56PlrvvXDXJ+iqvb0cuwFHv+vBrnnAIPc82ie0JyUl4rNvILfcyHXbs+Fxldo1oCZwt49n0ZebzNs3naRAE91PVLE/SAziKs4L1r250/EaBmRulJvj3buoTftNe4PtfcC42vRV809FgsbW8jxLNNnW56Vo+K1N5bI6sVSXgvrzRTzU17l9/XgeUaT2rV8p/GdxivR6ev390bcwT0MJ+0xnk4D849jafvP4XVC3weNh175du6ZWrx2HzmgvweZ9c+7UlZbnc0MSlQwsHl/EjkTc8yU85ruHWUmbfs+f69a8iz2HK+JlDut3xZeLvdcaNb3vYd71db8HxGto0scI+X/HKXH99TZEf/7Mk7lm3mmST814i08/Vl+nc1zTDn960HTBtbU/sXUm5Ne5P6V1EMtUszbXD1zqtPLXbXcm6/1BLPPWP/xLC3ewZ/bBduJt6ZoZ0Wo75LUx+adFw2OH6IHq5e08tWcLkv/wl3d3O9uw3rjmXbco6Y+1+RvN55tTinXvPks45fYMT3P9z1IM7Pd1MWRnfX+/tGeMknLq0XzHCmTIyZlee+h1yYua6DX3Pfyl7l3c9Vy8l5yoMXXCy8ylXtZOsS/svk9J7K0+Ox5bL3syddJInlHcvNgnq+3Kq2vPFl8nF5OMZILAAAAAAAAAAAAAASerDu5nnvuOfX+979f1dbWus76f/jDH1i6HmHx9a9/XdXU1KjCwkJ10UUXqTffFDNvgJyjpX2HWrP5v9XzL39XLXvxFnWgeRNLh67BpKVtu1q78QH1/KrvqKeev1kdaIKuYdN12fKbUV5DQlvTNrVpxb1q1eO3qxce/qpq3reBpUPX4NLasl2tW3u/emH5UvX0U19TBw6gzIaB1tbtat066Bo2Wg9uUxteuU+tfOoO9dzj/081NWxk6dA1uLquf/U+teKZf1PP/uVfoGuIaGvepja+fJ966a93qOf/9P9UUz20nehkHa7Y1dWl5s+frz796U+rD33oQ5707373u+qHP/yh+tWvfqWOPfZYdcstt6iLL75Ybdq0SRUU8GF340nhLjPssmQfH1rrCYUiRGgYmSWUwjO1phimPfkNE0bXcgIfBtxfa4aF5ze3ZDxVMhvKGbUN7RVp7sjSflVSVK1qKheq9XW/GTtd2XTtMs0ybJX8LhXnab2NZth2zQr+s7LXyFTXYvpxGkpYVMqH0dN7PSimqj94Cg/naD3e5K28Xj6MO3/DbnJNIhSLDv20hVpkQTI5oEpKqlVN1UK1fvP/jJ2uNI/ZTl0ORyfXmdfFMyqbWl7mB0tIorJMZczCQMU9dlp4uEZ+hwll66rk+ykuNiE1kR4eBmuFaB6RhdYShjpuumaKZVR50W5ebw60mvour8s/rs0TukbDxeVU00Tzwu28Dp0ylYcktpxslnun8ntT3BQ98jBUkeRY/rSUGuxXxeU1quqYs9Tml+73pP/gBz8Ye13p+VqiR/M6eL6vXGXC0/qnFFqG6osbRMuoKK+253SsbjdfP9sI2zOV551Cm3YWLVlYTJZ/Ikym+g+V2dqFasP6X+dWmZW3gExVHxGaO6mIfxoN/5aHoGHkws6BhufLfDTzT7xOaF5ownI6q4WuO027KtrKn+/2MppZGIwMH9P5wW075aquWUJtILr2cKuH4n3RzMLIRT3N6m1RnvPqeTh4+XYTOtV8Ktek5zhj8VG0tluceIZhTLJdzsqzaG8k+1VxWY2qnnGm2vTaA2PYJvZZFiSaeTnpIQ3o1EbzTNVMWrV7+NBSUWadg/xZmVhB9Fkyh/8uat5b2mdxzfsq+Pr0lNEu//U9mWtHqwmbxFlEsbltp9IaVTP9TLVxzX+PT3mN2DSX1izKN1yR7md/B9e8erlJzN9hwsjdQ/Sa8EWnzdhFaKYcOJheTh3LQ4Z3vJ+/F+16T3l6edbj/NRi9eLdlZ23f7mjmst3wsO9BSX/VmarZp6pXl89hmV2JJDrjpLnrcYhdkoyJD9ysM3/eUvqV6mrIvVw2Xb+TG09kbfPWKijCC8eMZGAdHK95z3vcT9+jU7dCL/55pvVpZde6n53//33q6qqKnfE18c+9rEjP2NwVJg6+UT3M1xsO3QNLlMrTnQ/wwFdgwt0DSeTq+e4Hz/uvvtu6BpQpkw5yf0MB8psCHQdphEPXYNLxbQ57mc4oGtwmTLtJPczHNA12FRUznE/wwFtJyaj6sm1fft2VV9f7w4BHKK8vFydffbZasUKMaTmb/T19an29nb2AbnFjh07stZVA21zG+gaTqBreGloaICuIWQkZRa65j6oi8MJdA0n0DW84Bk7MRnVTi6dgTS6Z5Si14fSJEuXLnUz2tBn5syZo3lKYBRobGzMWlcNtM1toGs4ga7hBrqGj5GUWeia+6AuDifQNZxA1/CCZ+zEJOtwxdHmpptuUl/+8pfT67qnNJuMRP2lpA+WLYJ39m9Npq5/B8/0ZbYplqnvlvSIsU41zfdZ8PJWs3I8MYWRSO8K4iHkmcI8x6bRPVJtKd7pV81irI9fd6zDaFS8h8cTO53Ey2HA3z8k1cGnp6dE13Ovj7JiPjy2aW4ivdx4Oo/zru00UzJHt+0ddc+QXNTVEwJLxJTZW5Hw85oXuT5tx/LpizOeWpgcxFNH0Clvk8IgTMS75/Wa9MFi7i3SerrxeZq03ngaHNrYfypl6sNl82oKWnmVUC+JGU/zstU8tzizfC7rW+LD5VCvROEtE+nm02dP3sj/GndwXrmvd1MR2a8TE34uYt0vzVNPB1lXoU+EThMuvUS6jWdMQRP3t8vYYZDuX5ZRMl22m9ZvPMA0pXvNb1uP4+W15MSa9HJ8L/cFspI71XD2uor6hcpFfV8ksX6hVg9pA9naSh6vFePhE5EeeoSIyCsyD0yuM/tpOo37YTrx2JF76ElYvo4Evi6mePQj6295iHvB7F9C7q3Nt5Z6JWqYJ5fIhOKZO3nlvvRyy0kz+KYJf29N5lVLl911i49exO5Vm3O6Ug9boV15gXnOtSYc//vsiDYI1SRfaBc197lgF68nC7eY/JGcZp6hmj3v4H5uvVPNfhO29ybrO5QaGTKPH4Xn8dF812E+gKLc5ZHXmc4e896hqaD+XdK7iewnWmx88FxI+Y01cs0rX+XeTQ0LTf7oreb7KW5ozaguHs4ix6SpcWU0y6uNeJfQtZXUb6lB/zaQrE9t/ROE/D38HSVRZdpDmpZF5j118rM7RlRex1u7IUb1taq6ujodUkHR60NpkkQiocrKytgH5BaVlZVZ66qBtrkNdA0n0DXcQNfwMZIyC11zH9TF4QS6hhPoGl7wjJ2YjGonl56tQGeWZcuWsZ7Pl156SS1ZsmQ0DwXGkNmzZ0PXEAJdwwl0DS96aD10DR8os+EEuoYT6BpOoGt4gbYTk6zDFTs7O9WWLVuY2fyaNWtURUWFmjVrlrrxxhvVHXfcoU444YT0FJ21tbXqsssuU+OKHDZI1mse28WTrEPu/YfZZxP6REMromJa2OaTTYhbzSZ5/Eynrc98rKAeLjqY7FM9vWYIY09fq4p31adDvHJWVx1S1MDvX3SQTs0qdKBDtYW2kWhmU5zLtIJtTWy9sMYM/eyu4tv2TTHaFu6IZTwM1LEM72XDQoXsrq49VNeW4Oi6pZmtJ5p5aAqD3hPLEGFPuUwl/afWFuFPhbtNmGpXpZnGXtM72Ry/9bQKHkawyQzbjvTyY7DQCksdJYfqD6aErr0tKp43zrraqhzHP3ygcsVB39BCvn//A0Rk6DEJEaXTZWuiLbxM5rdOSi+3Hc93U3hwqlne4x/CLLGVSRvJwT7V02nyfW/3QdXZuk9F/pZPPve5z425rrS+icpwW6qlJeTNA617U6KuJft05FD9XhOi6MhwRRHWVrLTxHK0H8PDZ5rmmbCLmmzCFY+AQa1tD9G256DqjBflRl1syaMFLeK+bjPhLPEOroGtrcRCFD0hZRnWg3rTTlMv95fxKe17a03ITHELtxXwhHb4EJFtBhIC6dh0jeSgrsPB2j0ixJjc6/hBbvVwzB+7zIqoB3hIoCVckIbu66MPiOchKdPRAR6u2HyqaZ+VrvE/hrd9lGFIjdjOrYu7eV3clUu6iufozkbT7ogkswibilvCiGlYaE+ff8jbLj46Jt7O69vOWrOfsqIC//cmW4ixhDblRZlN5RGbAbELj666vHaMsa4WOeT5poj1QUwUl6IGc909b/K6MN45mFndLMPaLPVk6ct72HpH7ez0cu8knneK6X7EPm1WD7bQzcMF/7nadhlt+8a7zFreS2heL9ov6tq/GE3yGyztTvEuOtRO9BxblrMe3iYureMWAQdPJ+83BTwMltlvZFHXjtguYKw7uV555RV14YUXpteH4lWvvPJKdd9996mvfvWrqqurS1133XWqtbVVnXfeeeqJJ55QBQWiYgM5RUfnPvXqxl+m19/c8ThLh67BpKNjr3p1g9F1yzboGkZd39wOXcNAZ8setWH5Pen1Hesfc/+fOmOB+79upCWTSegaQHSZXfPaz9PrW7b8maWjzAZY1zXQNWx0tO1R61b9NL2+bfOhungI6BpMOtr3qLUv/yy9vrXuTywduga7zK5fScrs6yizE52I42TojDZG6OGDehaDCxb9q8rL82Y8ZsKnoQbD4i8b9C9ykX7Rq01GbUR6+YgNp69/2JEesgecGTYKshnJ1fjJ03x77mv+PzNqzj0m6VV1pElkPG94A1a9Tv+y4elVH/46Bgd71bOr/k21tbWNSizykLbnn3OL0ZYaaXrM5clfiuRfV8l605n83HqnmPs3/RneCx7bstffjJGSxUiuSAk3XGw+138k15SNJt8Vrt7hr22B+euau57I99c2Hh1W50M7pX/hNLo+t+KOUdf1gsU3D6ur/PNLhJQNuuyu95nyFqHlUP+VZrIZyRXbXu8/Qk/WEbbRl1mM5IqcdGx6uWkhH8mVJH/sKGh1Mh/JRcus0I7qnBKaD2ee6pbXlaOv69nv+6bKiw/TCJC6poY3mtfEiGl/HhmJ4aZ19fuP5KKayBEC9K/OcmSBZSRXpISPCNz5cWMkOsh9VVX1qgHfkVypwrivPoNFRLt8oStdzfAPWoMDveqlx74+KtoO6frW879Oymtmv40O8Psc7TN1aKSX16e0/EbEKAA2gsNmXi6ev9mM5IqcOTe9vO98PrIgRn5a88R+fkxS16by+d8BU4X0GSt0ZX+pVBmhy+zzz33z6OlqMxknZVSmUZ17K/m+mk82+b7yVa5BYuNusyLbYyMdySWe06mppu7d824+anbK6yZfFW/ko0ucQtp24rrSdTmSgtbDTp7lL9eRo6Mr1faci24bvi62ISeEIKvRfl6eY6QMR3sG/NvTB8VEEqT+9RgeU93F8942kmvn509Vfsz+9W6+HzI6KCXaTqkCoq2oi5O2tpNPPfziX78x6rqe+45b02U2MpjZO020n9d32/+RTLCzjbdJj79rp2+9ySbmsY3kso0METruv5wPiXZIlqh9vN5/P3LUCNFOPmOZXnIwsGUkl1/b6YVlt466rkve5VNeLY88j+akno71cs07Z5p9N8/j11X7vCnLxWtEeSHtowh53rlYJgWRk3Xtu9SM5Eq08bI9ZfneYcunJlXk/4ylZZTq6K7T8ipljPiX2RX/NzplNv0Oex55h7WRymySD9mvkSy0jOQ60OJ7OFZ+5eQcDilnsj9iWoXvSK6py/fxbem7sMg7VEtP+4g+Rz0Rcipr3PfY5bcfVtdxns8LAAAAAAAAAAAAAIBxCFcMFLZpjmmv6ggHs9n8fWSPd0r+9YT81XnKejIPrB4lMt/8FabpPcextGnP7M1wut2cGqB3xMhee7o6dQ3v6W5/ixmp0XwqH7UxdcDMohHdIf6CTxF/jWSxzhIR31yxmsT7v3Oa73V4/2pmmQY7U6Ts4zWNK8ublimcpXeEhbx9ZNpbce8cVtwsPmzyL8lktJZnZIj4a0eszXiSxHuMj5Omv8zolRR/GMvGbyY0SMnp6D15ybZ7YKvj6GgtMeqOlslUN69fo6JuLtkzPb3cdDrfTetxZtRKIbej4Himps+R+ZNHm2zyL00T90P7Y6Q3k5mFllehFR1tKcunBzq6TxwirzvD887QOySQ2OpoghzBWNlb5DtK0QYb7SOfb1QrMSJPjhKJtrSnl8t28Xq45XhTXmM9/Nmb2N/u77tFTYzkiBGWxwOYH2RdxOpi/9EFdKSfu630ZGKHsNwXWmaFltIbhj6Dq1bx4+29wDxY973PjL7V1Dzd5H9udARD6vCjesadTE9JnHt8vWnrxqWNDxtxaPFM85yLJY2WU1Fmy3dynZvmmXLZvKSKpVWsISNTUlm8w5Ey7BmFF+QiK+85bT6K0TFlW02bNL+DD0NPJSweprT9bKkfbBFImuoH1qeX9181j6X1nHBoNkNNwb4O/xFO0qsqG2/PkLzDRgZ4+clvMx5dkXbihSif1Jay6xklmyTHEFEr0U7eRh4oMSO5Ws8ykUma8jUHzDEC0M7FSC4AAAAAAAAAAAAAEHjQyQUAAAAAAAAAAAAAAk/gwhXl8Dg28tZmYDvSMCGP2WKGoU/SxHpAhNOwRLHKzMKzuCYb9FQtvoI5RYYjIaNdfFj7pPV9vsOYo+3dGQ21tA6/l8NrZdjMXmN4G+ubypI6p5siV7xBFL8RDv3khoYiT9pM13MBGTVC760M56QTLcjiJKYmZ9B9ylAYop00rvZAynvhAV6+B6V5Jz0G0dUTHpJhGfYM6WbD8cdpyLDlsFER6sJ0zqbeskzBzIbSi7CXFFn3mGxaQ+z4al5XZufqMdK07DMnsUSusbxniyyV95VqJ++PLc1WDOm2om6LWDSQj9G8Xn9z54zlkmEVsRwMGx8NhK6JXS2+k/awiU5sf0KV9TBtS4m2Ei3LGjZvg9SOPFI9oZT0GSFbvvRkbfWDTfNcxWI8733+ZrhPGW5qCV2jYTNOFw+LSYlJQaxkFl1rxTtxgMptPOba/ic860+mXDadMcl30gUlw5gyzTvyOUp1JROCaAr38hCrvOPM+XRO59dQvtUYeMfaev2N9mlIsWjb59j8aYdHymhrktAmkNA/lR/zveexVhmz6oOYtIe1l4SFR6qrh29Lynr5dmFK/1aT52Y8zS8wr80/9Nk2aZXNAiEnYf0TWYhOr02+B1F95POPWSaJCfVIn0RKlNeYaDuV7De2Pn1lIny/kLzryHZ+DoKRXAAAAAAAAAAAAAAg8KCTCwAAAAAAAAAAAAAEHnRyAQAAAAAAAAAAAIDAEzhPrqyg8bs2vy7PlK1kPSWnpvWf6pquH87fJ5KX5+/nQg5ZdCA5sinbs4BNsTzWcc6WqeSZP4RtqlThu8Xi+Dt4DLnqs3ijRSwx/rZpdUXcukOmvZ/0Jj9e81wTp95yzgyWNunVRv9zo6cpY+hphsnFKbGzgGoZGRDx5uTaPFdpK7PUeETcn0iMxJvT5WFItbSmlxO7jY+EprvSTFc/WCDOTk5vHVZs1Y+tbqJ1QEyUZeanZplb2lOHRoata4dbT8Xo9OrCO6LPyegagjCV8oi1I74LHt+tTKf7lveHeNg5ZUV8U3I8eV+jpP529tZbD9lTZfabJJY0mngXmX5e5B2H+g0FXddRaTuJNHJ/nETc/9kkvSrp7j3PUKKr8Avx+JkmiJjyMU3sQtpn8XKeaC42l9Au2gVZ+PSFCotHpMcHktap0pOL6O6I9oki7SHmvTYcxGdPtuvoQ7+oKZXxeafIeXs8fWn1noNF3XO+Gfr4THuBtyUj1I9OaMf0kppT/y7pz0R9L4kfsSbWw9epnVaeKHrdNaY8l/Ryv7BoL8k7Ms9F/e8F95EU9bvKbWQ+ZE0SW5p8VSUethFSBj3vqrKeJjpLP0RPXWx5PiYT5uQ6p3PP2slNxJuPtr/C3F7OxuuRIvJ9xPLOrIgnF30P9ZRR6t3lrotyT47RN4nr0zO9JL1ctLOd74f2JajcIKS5CQAAAAAAAAAAAABMJNDJBQAAAAAAAAAAAAACDzq5AAAAAAAAAAAAAEDgCb4nF40lzcYvhMYSy7h36jkhd0L9fqSHD1mPxOW58G0jxHvGyYvYNj0KHlxiNypHYbqIsyRx3I6K+Xpy2fdv8VsbFMcj8c0p6cEl1inxdp4WIeYwPVP48SflxXxjpCM0hl5eLzOW4Mdn3mK54i9Dr03G42dqjCG9m5KZlVlnVg1LShYbPWLrtmTsV+IUco+BnmnmGAUH+fEjPdxngieOUBOLT8OYIW5PJJWhB6LcDbkHzCNN61Ne5O/XRfxdBku56VL+2u2+nhORAr5t+3H+51bQkvTVSp7PhEDWSxl6rXl8YAqNBvsvqOA/S/rvP0qKUlHTVJaWaOblrGeqad6kREsn0dxveY7nSD05nv4+FE8dHfP3esm0PvO01agv0OF8m8wxEi38+GU7TVrvJJHnQqprtu09KrsnzZZf6Dptq4jyzPyfNIXGvzLa2cV/J9ZpPdE7WRyD2mf2pzL3R8yVdk+G0CaQ13uUbiefR2Sd+DEdzpOL+anJckmenU7/gL/Hj8fDlueB4gaz3lnDde2eas6naB+vqKOk7eTxiFOj8AwLmodtNhdNry1uedWXHnq2thvxzHNXST7rmCnKa57ZT36naBPT/DEo2nVZtB2DDM3PtnaUp91L/EwjwmPaoc9OWc5j5LktDyLyh0NtSUV1ntedHFkfBNt27MrdBGylAwAAAAAAAAAAAICwgU4uAAAAAAAAAAAAABB4AheuaA2PyHQ6c4kcskqG9aVmTBPHSPmGxfVPMaE1+S+/4fs7SdPcQt+0wnoy1aokm2HZAe/O9AxBJ9PMRi1TlXsgQzgjMTH8mQ7LTIrpV+nwWjEUO5tQWDoMtOiAyBMk/EoO+XdIeICcvjsVtOH4ZPht5HChKRneV1pmnaoKe94hNJ5lppXPm3saP5yQuagx6T90PjZ8SJW73mFCMhzbsHHLeUaSMlyXHHuc4hUjKUsYjEizhcWwukkMzW5YXJ5e7qkSoXIpc90pUZYLTzvZLItylmjlwvZPMumxXn78wt0d5DxF+BN9bgSrCFqReduqXYbhvbZh7d01MjTbLEcH+Y3NJzNWD3RzPXoqTFiUpr+M/FYcvnOm2ba8a8D3+jONng4k9PaRsuSuxk0FExV1T4Q0tFgZECFt8jnl0DZQHq8HI4UR/+erCCel09oX1O1naV01s9LL8W4R6trNQztGHXo4JyDl21ZmJdTGI59r27R4mm9+SZEwugJR9xbt4eGK0Q6jbetJMqzO7Ldoj2gXs1BKSz0dBOj5invpWK6Th32J9iPZJws7HWad/c5ymtJagNHcylZLdph2Vve0Ur4fGmHVf5i2dabkcPkbVWh5le0xoit9f3ApMCFvDqnr3XWia6y+maWl2kl7SHO8qW+7q+X7qDm34h3id/R4Mv+xdqR/2yBw5XqkyHfIYtNfMDBrqr+FRxF/xhas3jZ8qLH7/OV9EF3VJA/0cg3iLT1mhRwvVwl41wcAAAAAAAAAAAAAAOjkAgAAAAAAAAAAAAAhAJ1cAAAAAAAAAAAAACDwBM6TS2LzC7F6D2TobbX7Iv/4cenZQ714onPmsbSiA8KLoDEzf4jY3ib+BZk+1ErAfJoOO5WyjL8esMQCU61tvmXiXjIPETE1eqTIxCw7HZ2+HiHutsRXaDAhYuGpTYzoYu6Ya+KrSzcI3TMkEB4yo3GOHk8uczPrz53EkgZKzHJU2Lcl8/1j/Cs2c68eOqV572QhHslyvRV8P/2zja7xfW326wgrmdbFIq2XWA70zez39SCJ9HM9iuqNVt3VPM3j9UJ84ZIJfvzWU01eKtsqp7snu5C2EmGVVZQR6qdh83aRuka6TZ1Zsotv2jfZ7GfSVl7Pl9W1+frVtCzg5T6Vb/aT4BYxKjpIfLdkHT1ATcGy8L0MMB5vrUyncbfUwyoa991nqoK3q7pnmvWS9dxny+nt48cgHpxOeQl/hs4yxy/fzvNOtMN4iTjEk8ZzHZ42Az0Zi0cM8Z7KKaSWI/Urol51+VzbpkXkXgsfveiA+V3iIH/l6JhhPBc1sf4yX59FP8842XZLSa8qupqjEvnheY6QPObx6bQNX6B5WvheMkNR6R1cZR7ATnGC73LTNv/zlj655PiyfVSyx1xHtJ34/WiIl5THuylMnkysjrFs5vHQs+2T7FT4I3bPNs/Kg6fEfbND4qApj5riev4e2zGD6CPLK1ntnsXr++JtbRkNtfF6cJP8Lw+Xi9mB9hfIS8n0GSuI9JjnYecM7j/ceiLxqhbN5dh841NbvC/l71+qq/Bi/xubSpg6PNYrDIgpOdJ2wkguAAAAAAAAAAAAABB40MkFAAAAAAAAAAAAAAJP4MMV7UMbLWQ4jXJUhCTS4emecEWyPnUjH3bbM5UPjz84p8ASGmCG9XUsMlO0akrXNZADyqmSoxmF6eXksE6NZyrZDPWU29FrlWEYZKi2I8IVe2dPTi/3lcswQzINdjMfopnYz6fHjXQZ7btrC3yHAtPQCk3Nih7/YeN0+m7PsG2V28jztQ3THanmZJ+yPNE8QMMTD21Lt+NpBS9sZuvRs05ML/dO4UP36dDg/A5+/IaFJtS1SoQY5LXyUFc/vOFwuVqI/4Y8PUs4EAu7EFMSl5NwtYEyPqy+cL8RrPplXt/m7zRTX7eeVcPSGmq50HQYefEunlZ/gQm7KN0lptqmdYtnaLYKDSzvSe1I+IwM12Lh35J+U4dWvsLrz71vMyESed08P0S6THnpPolPn91Xxo+XOGiEnbyZ54/6c4rSy6m4WdaUv9nlW87Yvcj1evcwsFBTEfdC21JZtavIPgcreSjaQFn+sOGimoMnm6bogQUzWVpMVJFFDea3cZE/KLI+t4ck+rePIoPkGCJMLmfbUjYyPGdvaBTJE6T8agr2mfuSLOS/K91hlidt4WIOlPBXkKZ5eb7xPYlWc+L1S2g8jVK1zw5mFFItCaR+I4HWxTGeh1Ol5F6KUM9koXnm7juf3/OC004zuxfRiYXNSd/Q8WQB17Vnmknrm8FDzhMNnSOrby3hfxlGBuYm8lpG2F4eKDE3s+M4Yc1Cwo0LmvhNL2zgz9HW40t9wxWL95rf7r2An86MpHnGFxwQFXzU//3NCj18mMq1513HPI8Srbyc5bebex7r5b+reXxvern3LdNYWnd1wveZK5/VTacbi4Cpr/FTi9hshMaJgDfTAAAAAAAAAAAAAADIspNr6dKl6qyzzlKlpaWqsrJSXXbZZaquro5t09vbq66//no1ZcoUVVJSoi6//HLV0EBGH4GcY/ve59SqdT9RT790h3pu1bfV2s2/Vl09B9g20DWY7NjzrFq19ifqmRW3q+dXLlXrNv1adfeYUS4aaBtgXVferp57aalaC11DwZ66p9Tap3+oVjx6s1r1p9vU6yt/pbo7Gtk20DV47NzxjHrl5TvVc8/eqpY//29q/boHVHcXnrFBZ8euZ9XLr96lnn3hNrV8+b+p9esfUN3d0DUM7Nr6tHr1xR+pF/7vFrVi2TfVxtW/Uj1dfEIgaBs8dm5/Rq1e9WP1/NPfUC8+fbva8Nr9qhu6Bp7dW55Wr73wI/XiX25RK5/8ptr0yq9Udyfq4olOVp1czz77rJtBVq5cqZ588kk1MDCg3vWud6muLjO8/x//8R/Vo48+qh588EF3+3379qkPfehDR+PcwSjR2rZDzag+W5017zp1+qlXKieVUms2/kolkyYGC7oGkxZX20XqzPnXqQVzr1JOKqnWb/o12wbaBo+W9h1qRs0ideZpusxepRwnqda9Dl2DTlvTNlX9lnPU/AtuUKeed61bXje98HOVHERdHGRaW7er6TMWq4ULP6cWLPi0SjkptWbNvXjGBpzWtu1qRu1itXDBZw/pqttO0DUUtB3cpmpnLVELllyv5p31D8pxUmrja/ezbaBt8Ght3aZqZyxRZ5z1eXXawmsOPWPXPMC2ga4BLa/HLFHzz7lezT37H9y6eMMqtJ0mOll5cj3xxBNs/b777nNHdK1evVqdf/75qq2tTf3iF79Q//M//6Pe/va3u9vce++96uSTT3Y7xhYvXqxyEuEDQ6eonrqWz8PZcqLxlZi6gccSx18yHj5OksfKFi2cw9Z7phT5xh1PfsP4HTQs4j40eV3Gh6RgX4dlSvvMg5JPP+UK9rtTTviQev7lb6uOrnr3u5zQVU7dbQtFt8Wpk/uSKudeLAfnGG3JTNaHfkbkTEzhxk75NXwa13g38a4QHkzUu8njD0XimT2+UtTvxtY1LfKS7rRM7yMaUSefeLla/tLS9Hc5oS3Flm9tPnokTr3i9T5f34/22bzKK9tpynrxyu0sresCM+WupnM68bIQh696yUyJHDtIfCR0OZ1TmV7efhn3aDvhAX6uGet6ypUs7ZQTL1fPj7eu5BSlP5OTR3wWBmSaydARMSNxxWozOq2rdppvHRDr4j+kPlwtJwkfnRiv76c/QzwOmrkeb//E+vTyxt/MY2kp6kdl8+TKwh/i1HP/ga2fsPAjatWfv6m62vaNo67+dQ+7buH1ouijS3gMUi+R2P6DLK1qtakYeyv48+/g4iqTNpmfTErUp0UHzDHjB3iZrFztX5Z5XTs65h7zF1zN1k8++XL1wvJvqc6O/blRD4/Uz9TibdU5m3v4NM+l5UX8jDzv4u18n9PW8DLZXW3yRNuxMd86oXsaP0jR7Cnp5fxG84dZqbnn2St8iigL5l9ltotEDun6wrdUZ2eO6DqcRxXzoYz4e39Kjz26LDy5ql82DZuGhbwglm8z+sVbuKdP5/RyX8/S8i381Ir3m2PsvkJ4Pq3Iy6guzsaDa95Z17D1E+d9WK186vb0ek5omykWn6O2ucYHq/0Yf7/K/snC18nYManJm3laZy0vl/3lpHyJuobW291VvL5XEeP/E2/n72I2vyZWhkXxnX/6p80+YhF10rwPqxVP35G7ulo95aivokx0fJ+/RfXmXhbuNZ6xUo+SfbycJxPSl9QsT+IWtmrSlu70cu80Xie0fca8u+Z/m/tBpeLRYdtYLhbN5y4i5TWi1InzP6xe+uvtqqt9HNtOmZZJ6XVq+x3J20VbePTGYNE0Xz/TlsW16eWuKlHOhadezfOtJq2bP38bLjTvMw2LS/lz4IW2zPxMg+LJpTONpqLi0Eu+7uzSo7suuuii9DZz5sxRs2bNUitWrBh2H319faq9vZ19wPgyOHio8y6ed+iFfM2aNVnrqoG2ucdgknfMosyGq8wOAV3DweDAIV3z4oUjrouha+4xOHio4QhdQ6pr3sh11UDb3COJZ2woSf7tGTsEdA1XecUzdmIz4k4uPRTwxhtvVOeee66aO3eu+119fb3Kz89XkybxGTKqqqrcND+fr/Ly8vRn5kw+sw4YW/SQ7Dd2/FmVl85SxUWHemwbGxuz1lUDbXNP2ze3/VmVlc5If4cyG5IyC11Dqev2dY+o0orZqqisasR1MXTNPV23vPmYKi8/RhUXj/wZC11zUNctR66rBtrmnrZbX39UlZab2c7xjA1Jma17TJWWGx2gazh03bbpUVU2ebYqKkXbaSKTVbgiRXtzbdiwQS1fvvyITuCmm25SX/7yl9Pruqd0pBnJE6JHh3JmEb7nkGGehau2srS2Y03YYd8kPrQ2UWlCCTvnVbO0rioZMmOWq5e3sLRos+ktfksdP0bdHaaAnvCtqP80wbbpsy1TLNdtfVR1dTeqhXN52Mx4a3s0kMMn+0jdl29GXXpCSIt2id78AT7Ws7/WDMFvPY4PxU2S1fx2Pmz7wAIT3jFtLT+5vFYyzN+Szz2QtDe2PKq6uhrU/FOvUC+9+kOV87pmc510Wt3NZqpcTT4JR+4rP46l9ZOplIsreOiEHDOcJNNgx7u4drEmkyecOK9WaX45/jf+Va4MD7Hrahbrtj6murob1IJTrlArXxs7XWXIES1QMi1FQ34Sef6hxzIsuccMq5/xf2YItXt+J5qY4r1v5/HF/WVmP4NFfNh2vJPf1+ItRp9UCR9W/8q3FpLTHvTVS16vNVw8w0fRtrV/UN0dDWre+Z9TR0LWusqh81HHP6rNEsrFnkFREUuYF/OtPxM7zBD8WBfXtW+aCfftFkPuK17noRUFjT2+ZTLeav5678RlCAYNIxDXRO7NSDV+441H3Hr49DM+o46EI62HrdPPi5A9X+Q+yPqkl3g9nMyfnl5uO16ESxDpZv1+D9/lQV7unSUnppe7q3i+ipEoxFifmPp8vnn4dtXy300i8ydNWdeZeX4YTtfTj0zX0W8X83WabWUoZoroHnVEPU2sF2gbWZOoN/cs1s/tGw6eQttAvD3UN1mcLKmqK9aKdhY51dIXZJhjt3+ofPTIw4+3bPyj6upsUHMXXq1WL/++GjddLc9KK7Q8i3qrZKepJztm8RDjnipioSFsBmb/2dzz+HZu3t1x9iy+n0qTl6L9fD+0rd1yMk9LtBnRiePBISyXT+tmW9jUm6//UXV11Kt5C69Wr7zwnyrX28ReGwjLOx9BhmDn7zEWAVM28XfVpvnmedh8Cn//dKJx35DViKxoaFjdTh4e3vd7U/BTBX0Zl9dM6+ItG7SuDWr+ks+qI+GIdaUWQjK/kmuJWHSNJEX7hFShkR5+70q3mjDQxkW8juwvM8dItIj6u523kaPtpD5NcM0rV5rncbKE1+dOLDoGMYRj0Ml1ww03qMcee0w999xzasYMM3qgurpa9ff3q9bWVtZbqmcv0GnDkUgk3A8Yf+q2PaaaWurcDq6CRHk6BEr7rmWrqwba5g6687Lp4GZ1xvx/UPE840WGMhtsdAeX1nXhadA1TGxd+wd1sP51Ne+tn1OJwknpsMWR1MXQNXd4o+4R1dxUp04/41pVUHBkz1jomjvUvfGIam6uU6effuS6aqBt7rBl4x9U84HX1fyzP6vicTxjw8Kbm/+oDh7YrOaf9RkVz4euYSqvBxtfdzu40HYCWfWn6b/e6A6uhx9+WD311FPq2GOPZekLFy5U8XhcLVu2LP1dXV2d2rVrl1qyZMnonTUYVbSuuoPrwMFN6oxTP60KC/if1xYsWABdg6zt1kfVgeZN6vTTtLb8r6wos0HW9W9ldh50DZOubgfXvg1q7nnXqYJirivq4uDqqju4DhzYpBacfo0qLISuoamH33hEHWjapBYsgK5h01a/MDc1bFTzF12nCovwjA2LrrqDq6lxkzrtzGuha8jKa3P9RnXa4utUgdAVdfHEJC/bEEU9M8Ef//hHVVpamo5j1bGqhYWF7v/XXHONO7xPm9GXlZWpL3zhC24GyqmZRgBDd3A1NK1Tp835exWL5au+fj5rI3QNLrqDq+HAOnXaKZ9QsVjC1Tb5N3NcDbQNcJnVup6syyx0DQs6RPHAntfUyYuvVLG8AtXfi7o4DOhQtsaGtWruvE8eKq990DUMvPHmI6qhYZ2aB11Dx5ZNf1CN+9aoU8/QdXFC9fd1pEeFaKBtMNmy+Y+qoX6tmjv/UypvSFcyqQB0DSZbNx4qr6csvNKti4faTkOBedB1YpJVJ9fdd9/t/v+2t72Nfa+n4bzqqkNTKf/nf/6nikaj6vLLL3dnJrj44ovVXXfdpY4WNC43Isel0TjXFE90iCdIRE5vPug/3W5Rk9m2ZyrfZ89F0/39fOIiDr3DGdaD61Bi3Dce9oRvkZlAhJ8Cn/pcpFnil/c2rHL/f3XjL9n3Jx37vvTyWOuaFVn4rSniz5TXxBukk7YYP4I+Er+sSTSb+x5p454dqak89rl3itEsmRD+B/vM8Ut28em0eyvNsNg9bzPDpzUznib5pUfM92rxH9hb/zdt1/9C+TGe2to8bzyq0jwd43Hqit7nQT69uCINmKID/N61H2O0qr/QTL877AmQ9WkvHGBJDimzSvj/sF0MpiyeCv7l2VfXDbzMjquu1OtFyJMklbPU3ImZexcl00e76/l5vveudJsph8V7+T3f/gEzLXblK/x4Zdu4P4RDpsVOiSmyE60mv6SEKQj35LJMW5/FeOn67Ydm+dnw/D3s+7fM/+DR1ZWevrRZomVS6MrypfQ+If4M+nxZGvHBigxy7ajfT7Rb+GyR+nNypNjXg8vdb6/RzinI8y13Uju2Lp8tI/SS2Lf3Jff/Na/9nH1/4okfyIl6mPniZUMWvl7lZEr5vvISllbYTMp2n/Hh07S/82S23jGL5h3/qc+pr4h7qqRe3vkhMw26Jplv8RWx5PG9+w7Vw6+tyU1dM/J+JKTySQG3edwKHzuaDypX89nq2t5i2jVtJ/CflW7n65O2Gu0jA/w5Tv0Sp67rzrw8RzJrb0j271rp/r9uFa+LKbmmbUaIa47XmzIzdS1/32g8I9/XxyfWbv6o1nUG9ylqmy3aQOSQNS9yH6H8BvM8bj+Fm4Lv+pjJAyfcJetisu7xM/VZ1nXxnkN18drVP1NB0JXlX3GdKdJeihRw7Vj5oR6YwsupbA03Xc/rNu3gpnncu7C3kueB0m1muWIz1zXWM+Dr/Ve2vce3zcXbvSpjhsrr+pd4eT3u1MtyQ1fZPiK3JCXrU7JO28AeXcW7RqTT3NcpG7l2PeT9sngPrz+jbaI+LUz4e5ZSvzBRR7Nnp6dNPDI/xDHt5JJmk8NRUFCg7rzzTvcDgsE7zrl92O/1Xzfqtj/mLkPXYPKO8+7wPDC0rs+tMN9D2+DxjnNv9zw8tK7ProSuQebcD3532O/1CIJtax92l6Fr8Ljw7d8a9ntdZvUoLw10DR5vf9u/Ddu5Al2Dz/nv+c6w9fCLf/1Geh3aBo8L3rnUrPytzOry+sKyW9NfQ9fg8dZLvOV1qMzqUV4a6DrxGEOPewAAAAAAAAAAAAAAcmh2xfHEO705TZQhiZap6cmly/FpbKSrOF7ZM1vSy4n5s1laM5kqWYbr1CznIYmxA2beXKdATK9OhiA6MiyLDsuWQ3RJSIhnmCebenV8hg0eEZ5QGBqeJvNE1BJuQu5DFx+iOfnlxvRy/UVVLK3jWBMakzyJh1oM8KgZlconw/pF5FzxHjOUP6+R54mSZqP1YGIq/yG9DluYahCkpecoh7TSocriOllElRh+rUjIcaRvwPd3BWt3sbS8jpr0cvNpPES0v5SfW+WrZjh2ZGDQf3ivzHOZllk5jTALk1E5MfSXnwNf5dWvf/hASvxtxYk5/sPxyVB2OQ02nW5eMuMpkweiA/x3Up9kacI/JJGFwShLWJtMIys5INXRe/4S7cSFRnzCKty0VMQ3PCzT0DkZnijrC4eEN3nCSel2tmH1UvNcKHfjDb0HNExb359+U+4iImw8tt2ExRTN5JMW9ZWZG52czp99stylSNVf0MozS7SNhCJLrUg+m/V4Kz8GCcmQeZVPaR/A/CDbTrTFawtdjPjfhwgNa3TbOaYyjokQ40lbzfFK9/Dfxdt5aGqkLzlseKK7TjRyZD3N2rfiQixlP0yw6xYhTpRIn4jxJe9GRZsbWFJZhbFg6a3g+SFVZMp++0x+vP4yfog4KZaJXS0i0fy2bBMvl4WNvE3GTpvkAW+I6shC3nICeb6O5VlFnr+pfFE5lZh2jWgtc4sTEfJWsNfYQNS2xP3DmfXh+5PD1gFyW8+zOW5r9w5/np60wA8p8n8np21dec8jtB4WoYSRQaNXrJOHj5Z09Prf16JE5tZHVpuO3Ct3GMkFAAAAAAAAAAAAAAIPOrkAAAAAAAAAAAAAQOBBJxcAAAAAAAAAAAAACDyB8+SyIWNAqTeTw0OLeZypiDll060Kzwkad1ywrYmlTSfTqdqmU3fXiwp8j899YMS50VX5O5vXURCh91B4qrFp7WW8NzFEi8jfkXUZQ66In0jpHu4nMlAU9fVqShIpNcX1ZJr7tS2+x/D4mZD46rItHZl7N1mmvB82vj2HfEQ8nib0WrKYulZFSPx/kYhTT5D7LHxi4g3GG6/6SbM8vP8B0YDu0/2CeBNJPwxbDLvNKy93ZMoMWlxFtRUhRc3jS0ZWk1F/b4JMvZrktqm48PnK514v7FwtBo1e/wGffQQdiyeIhNXDxB/k0M/853F3ov66ZiGzOBfxxUjruegIPZeCVl6tbSfqmSafKaZ+i/YLfx/qIRrjdS31Ny04yH/XM9WUyf3ncUMfWbaqVhtvkfzdLf7trIQo59TTSTxbUrQNYdE8EB5ch4P6psjSRp9VHrsVUtalxy3x2PNA6v5Yj5hyXnopEZ8nq++Wx4OJrNjSwqCfj89RyscryeNTK/1MB0haH/dIK95v1rureUN3/1tLzS7MokvFRn780i2d/howfyh+brHeQf93qJG2nQL8buR5xpHy6rEoJX6mTpRrF0k6GT3vPeWc/k7kQekdRXWVfmHM18ny3Lb6OgVXRi/S65E1LsW29DkmfGrZu7D0sLURHZkfsF0f+a6jxoUwNc0BAAAAAAAAAAAAwAQFnVwAAAAAAAAAAAAAIPAEPlyRh0uINLIckdPW0xgJOayODvmTIVOWaeuz6k5kwwFlaKN/mnWooG2a9IAP07aFU7BQA8+tjfgP2xZT59JQtpL19fz4ZIh3/gl8ivPOWr6f/HYamyWG9xaSae1paMdhwtrYUO1sho/mou7snBzf6yRRp97wUqnrYCqzYf0ilJCGrx4WS5llyHqIhf5Y/q4QBO0yRZy6LZyP1nc0rNFNs41JH2lcm22XNnnkudn+RBRg6Tz4F1f7MyaW2U5HKuO4EwlRGKpsg9A6S9SZURIqLkMZI6RQONLqgSwXbN7P0qb0VaaXD5xRxNIGivlp7zvHhN7Mftg/5N8zbT17tkgbCFtIuQpPHpBY7Q34pvx5LHXP8HieEBp/HTyhUZmGJKoQPUfldYq3Gga1ORH3ma7LejpKLTzE4RPbD6SXq/orWFrDWaZgJvP58eKd/GFJ7SWcooR/yJt836L1i7Q5YAdQ4bVusTx/PXUVIUXLku0eiH1Si4hszk3mKxpubDtPa9m1HC/UsGpR1gEEESZuVc4ZpZZWFu8+uUAOnhIAAAAAAAAAAAAAANmBTi4AAAAAAAAAAAAAEHjQyQUAAAAAAAAAAAAAAk/gPbkoMiaYTn/qmbae2fuIuFYauyoDho+GJ5dkgvpuWZEakehjGW/uUHM2mRY3aZFUnq+vk5MSngLEH6JoRytLK9qhMp4SmXmGeLw/LFMijzSmneblXMwftqmfpZ8Z9VGJSiMB8kOhHd9ulPr1rdPMi3XhDZPpfkJFhpcpfdhyiVw+tzFjgmTXiQB7xnjyNn1O+k8bbysUsl2VIh5dkaTx6NHEGzvSy7VPmGWfHZPz5nVrKpHn/+yn6xYPMvzpN/tnU8a+Oln4I3naR7nelhkLaBvR48BD252iXJBNo84gTys2HlkR4b+nSJs4r7WHJU1/ssdfD9F2SxXl+/puZeqH5/GpHeF7UqCx+Cp6y4tZTFn8zKx+ekfg48Tr2xHvBtjI6lkVURMRPM4BAAAAAAAAAAAAQOBBJxcAAAAAAAAAAAAACDyhClfMKnyPDt/0DMkc4bC+bKZezXA4bahDEI8E21BlOmx7tEKMshi2m+l02llNncsOMEHyhOc6SfixLKMxS0iibTj88Ls/MsI6nTkAIDxYpqZnYUKeCi3q2z6JRkmafGbSZ7aTNzphMbawcRHKyJ63HjsAuh3q6JwEunAs4ZwpkfcjCRqO6x+qG8mP+b7TUPuXw777iGOkaIizCJ1j5U1ekjXN/3QmDLbmbKYWEdbtcJNBsMFILgAAAAAAAAAAAAAQeNDJBQAAAAAAAAAAAAACT86FKw4NcR9M9o3ZMa3DcCdwuOKQBp6wgyPVdnDstB01QhSuOHT/A6crm/XUUoYtQ+fDHK541HQd6B2V/YGRM6TBaGgb6Ho4ZIxmmT0iXa31qUhL+de10STZ2BKu6Ekb6eXLEKYUtTGIHf1wxchYP2NRF48nQ/c/F3WlM4B7yjMpp1EyY6InXFG2nUYrXJEXsKMfrpjl7Iq5rCs4MkZTW7SdcodMn7ERZ7RK9SixZ88eNXPmzPE+DUDYvXu3mjFjxhHvB9rmFtA1nEDX8DIa2kLX3AO6hhPUxeEEuoYT6Bpe8IydmLrmXCdXKpVS+/btc3vnZs2a5V5AWVnZeJ9WztDe3u4WsrG4L1qDjo4OVVtbywxlj0Tburo6dcopp0DXcdT2aOiKMusPdA0nQa2Loasd6BpeglwXo+0UTl1RZv2BruEEz9hw0p6DuuZcuKI+Wd0rp2+WRt8oZCIvY3VfysvLR1Xb6dOnu8vQ1Z+xuDejrSvK7OGBruEkaHUxdM0M6BpeglgXo+0UTl1RZg8PdA0neMaGk7Ic0hXG8wAAAAAAAAAAAAAg8KCTCwAAAAAAAAAAAAAEnpzt5EokEuob3/iG+z8Iz30J+vkfTYJ+b4J+/keLoN+XoJ//0SLo9yXo53+0CPp9Cfr5H02CfG+CfO5Hm6Dfm6Cf/9Ei6Pcl6Od/tAj6fQn6+U+k+5JzxvMAAAAAAAAAAAAAAIRmJBcAAAAAAAAAAAAAAJmCTi4AAAAAAAAAAAAAEHjQyQUAAAAAAAAAAAAAAg86uQAAAAAAAAAAAABA4EEnFwAAAAAAAAAAAAAIPDnbyXXnnXeq2bNnq4KCAnX22WerVatWqYnE0qVL1VlnnaVKS0tVZWWluuyyy1RdXR3bpre3V11//fVqypQpqqSkRF1++eWqoaFB5TLQFbqGEegaTqBreIG24QS6hhPoGk6gazgJq64aaLs0ONo6Ochvf/tbJz8/3/nlL3/pbNy40bn22mudSZMmOQ0NDc5E4eKLL3buvfdeZ8OGDc6aNWuc9773vc6sWbOczs7O9Daf/exnnZkzZzrLli1zXnnlFWfx4sXOOeec4+Qq0BW6hhXoGk6ga3iBtuEEuoYT6BpOoGs4CaOuGmjrBErbnOzkWrRokXP99den15PJpFNbW+ssXbrUmag0NjY6uk/y2WefdddbW1udeDzuPPjgg+ltXn/9dXebFStWOLkIdPUCXcMJdA0n0DW8QNtwAl3DCXQNJ9A1nIRBVw20DZa2OReu2N/fr1avXq0uuuii9HfRaNRdX7FihZqotLW1uf9XVFS4/+t7NDAwwO7TnDlz1KxZs3LyPkHX4YGu4QS6hhPoGl6gbTiBruEEuoYT6BpOgq6rBtoGT9uc6+RqampSyWRSVVVVse/1en19vZqIpFIpdeONN6pzzz1XzZ071/1O34v8/Hw1adKkQNwn6OoFuoYT6BpOoGt4gbbhBLqGE+gaTqBrOAmDrhpoGzxt88b0aGBEaPO2DRs2qOXLl4/3qYBRBLqGE+gaTqBreIG24QS6hhPoGk6gaziBruHl+hzXNudGck2dOlXFYjGPC79er66uVhONG264QT322GPq6aefVjNmzEh/r++FHjrZ2toaiPsEXTnQNZxA13ACXcMLtA0n0DWcQNdwAl3DSVh01UDb4Gmbc51ceojbwoUL1bJly9hwOL2+ZMkSNVHQkwLoDPTwww+rp556Sh177LEsXd+jeDzO7pOewnPXrl05eZ+g6yGgaziBruEEuoYXaBtOoGs4ga7hBLqGk7DpqoG2AdTWydEpOhOJhHPfffc5mzZtcq677jp3is76+npnovC5z33OKS8vd5555hln//796U93dzebolNP2/nUU0+5U3QuWbLE/eQq0BW6hhXoGk6ga3iBtuEEuoYT6BpOoGs4CaOuGmjrBErbnOzk0vzoRz9yb1B+fr47ZefKlSudiYTufxzuc++996a36enpcT7/+c87kydPdoqKipwPfvCDbkbLZaArdA0j0DWcQNfwAm3DCXQNJ9A1nEDXcBJWXTXQVgVG28jfThgAAAAAAAAAAAAAgMCSc55cAAAAAAAAAAAAAABkCzq5AAAAAAAAAAAAAEDgQScXAAAAAAAAAAAAAAg86OQCAAAAAAAAAAAAAIEHnVwAAAAAAAAAAAAAIPCgkwsAAAAAAAAAAAAABB50cgEAAAAAAAAAAACAwINOLgAAAAAAAAAAAAAQeNDJBQAAAAAAAAAAAAACDzq5AAAAAAAAAAAAAEDgQScXAAAAAAAAAAAAAFBB5/8Hl9Vg0PYZKZQAAAAASUVORK5CYII=", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAABLkAAACkCAYAAAB7PPybAAAAOnRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjEwLjEsIGh0dHBzOi8vbWF0cGxvdGxpYi5vcmcvc2/+5QAAAAlwSFlzAAAPYQAAD2EBqD+naQAAZN9JREFUeJztvQmYHVWZ/3/qLn1v752900k6O4SQsEVCSFiFn4iKoDgu4wYy4AIqozP+fo67+B9UZhxHBxBHBHEWHBxQREEx7BIIRAIJIQGykKXTnXR3el/uVv/nVNv3vO9bXZXbnV5uVX8/z3Ohqk/dqnPrW2epk/N+j2Xbtq0AAAAAAAAAAAAAAAgwkYnOAAAAAAAAAAAAAAAAxwoGuQAAAAAAAAAAAABA4MEgFwAAAAAAAAAAAAAIPBjkAgAAAAAAAAAAAACBB4NcAAAAAAAAAAAAACDwYJALAAAAAAAAAAAAAAQeDHIBAAAAAAAAAAAAgMCDQS4AAAAAAAAAAAAAEHgwyAUAAAAAAAAAAAAAAg8GuYbB17/+dWVZ1oi+e+eddzrf3bNnjxor9Ln1NfS1ABjPZw+MP9A1nEDXcAJdwwu0DSfQNZxA13ACXcPJnceg66QZ5Hr55ZfVhz70ITVnzhyVSCRUXV2d+uAHP+j8HUwst9xyi/MAn3HGGWqy0tPT4wyiPvbYYyosQFfoGlagazgJo64aaBtObaErdA0r0DWcQNdw0lOkuk6KQa57771XnXbaaWr9+vXqyiuvdB7Iq666Sj366KPO3++7776CzvPlL39Z9fb2jigPH/7wh53vzp8/f0TfDzP/+Z//qRYsWKA2btyoXn/9dTVZK4hvfOMbRVdBHAvQFbqGFegaTsKoqwbahlNb6Apdwwp0DSfQNZz0FKmuoR/k2rlzpzPAtGjRIvXSSy+pb33rW84A1w033ODs67/r9F27dnmeo7u72/l/LBZTyWRyRPmIRqPOd0ca7hhWdu/erZ5++mn1ve99T82YMcOpLEDwga7hBLqGE+gaXqBtOIGu4QS6hhPoGk6ga3ET+kGum266yRlh/PGPf+w8gJTp06er2267zRnE+u53v8t8t7Zt26b++q//Wk2ZMkWdddZZLI2iZ2d95jOfcc5VWVmp3vnOd6oDBw44x+nj/WJK9cjvO97xDvXUU0+p1atXO4NgetDtrrvuYtdobW1Vf/d3f6dWrlypKioqVFVVlbr44ovViy++qIKOrhD0PX7729+u3vOe97gqCD0qrO+bHB328h+755571PLly517uWLFCmeW3hVXXOHca/ndf/qnf1I333yzc8/LysrUW97yFrVv3z5l27YzCDp37lxVWlqqLr30UkcDyYMPPqjOPvtsVV5e7mivf4MMf9XX1prpZ+Kyyy5ztvVzqPXMZrP5/Aw+m3okXOdNPj/bt2937s/UqVOd3/amN71J3X///a486eu/+c1vdvKt868HdXO5nBpvoCt0pUBX6Apdx19XDbQNp7bQFbpSoCt0ha7QFboK7JBTV1dnL1iwwPcYnT537lxn+2tf+5qtb8vy5cvtSy+91L7lllvsm2++maVR3vve9zp/+/CHP+wcp/dPPvlk52/6+EHuuOMO52+7d+/O/23+/Pn28ccfb8+aNcv+h3/4B/vf/u3f7NNOO822LMveunVr/rjnnnvOXrx4sf3//t//s2+77Tb7m9/8pj1nzhy7urraPnDgQP44fW59DX2toLBs2TL7qquucrafeOIJJ/8bN27Mpz/66KPO3/T/KUP91gceeMC5dyeddJL9ve99z/7KV75iT5kyxV6xYoVzr+V3TznlFEdnfeyXv/xlu6SkxF6zZo2jxdq1a+0f/OAH9mc+8xnnnFdeeSW7/l133eX8/a1vfav9wx/+0P7Od77jPEc1NTVM449+9KN2Mpm0TzzxRPtjH/uYfeutt9qXX365c339bGm6urqcv+u/vetd77J//vOfO58XX3zRSdfPgtZa51VfRz8n55xzjnP9e++9N3+tgwcP2jNmzHB+89e//nX7pptuspcuXercD/nsjTXQFbpSoCt0ha7jr6sG2oZTW+gKXSnQFbpCV+gKXTmhHuRqa2tzbowerPLjne98p3NcR0dHfiDrAx/4gOs4Oci1adMmZ//6669nx11xxRUFD3Lpv+mCMcihQ4fsRCJhf/7zn8//ra+vz85ms+wa+jz6OD3gFdRBrueff97J78MPP+zs53I5Z7Dxs5/97IgqiJUrVzrf7+zszP/tsccec44bqoLQhUk/I4N88YtfdP6uBynT6XT+7/pZ0JWH1kGjz68rgquvvprlqbGx0SnI9O+6gtDnpDppTj31VHvVqlX5/cOHD7uemUEuuOAC57cNXn/wXulKTFcAg+jnUJ/j2WefZc+TztN4VvzQFbpCV+gKXSdWVw20Dae20BW6QlfoCl2hK3T1J9Thip2dnc7/9TQ8PwbTOzo68n/7xCc+cdTzP/TQQ87/P/WpT7G/f/rTny44j3paop4uOIie8nf88cczjzC9GmQkMiCVnh7Y0tLiTBnUx/35z39WQUVP65w1a5Y6//zznX09vfF973ufuvvuu/PTIAuloaFBbdmyRX3kIx9x7s0g5557rhPmORR/9Vd/paqrq/P7gytj6FU4tf8a/XsqlXKma2oefvhh1dbWpj7wgQ+o5ubm/Ef7rulj9YIGEvk8ac39fOAG0VNMH3nkEfXe977XeZ4Hr6WfgYsuuki99tpr+Xz97ne/U2vWrHFCX+nzpFcRHU+gK3Q9GtAVug4CXccOaBtObaErdD0a0BW6DgJdxwbouqvodTV3IYQMDl4NDnYNZzBs4cKFRz3/G2+84Qw+yWOXLFlScB7r6+tdf9PxvUeOHMnv63jUf/3Xf3VWhdQmd7TwTJs2TQUR/Rt0RaArB/2bBtEF7J//+Z+dlTB1fHGhaC287r3+21CDgfLeD1YW8+bNG/Lvg5roQqnRccNDoT3TKDr+WPrBSY290Ct16BmXX/nKV5zPUBw6dEjNmTPHuQdDLWGrB0PHC+gKXQsBukLXQaDr2ABtw6ktdIWuhQBdoesg0HX0ga5TAqFrqAe5tLCzZ892VlH0Q6frG0yF1aZn44EeOR0K/VAM8o//+I/Ow/Gxj33MMZPTxm16cO3666+fMDPbY0WP7B48eNCpJPRnqBFyXUF4rUY53FHy4dz7o2kyeM9//vOfq9raWtdxdATd73yFMHgtbfKnR72HYjiDqmMNdC0M6Dp8oOvoA13DqasG2oZTW+haGNB1+EDX0Qe6FgZ0HT7QdZIPcmn06oX//u//7qxgOLhKIuXJJ590Vgb4+Mc/Puxzz58/3xFQj+IuXbqUjVyOJr/85S+d0eLbb7+d/V1PN9SrOgYRXQHMnDnTWRlCcu+99zorSvzoRz9yRosHf+tQo95UC697P9p6LF682Pm/zv+FF144Kuf0qgj1qhmaeDx+1GvpezA4Qk/ZsWOHGi+gKwe6DgBd3UDXYwO6+gNtw6ktdOVA1wGgqxvoemxAV2+gazB0DbUnl+bv//7vnVlZehBLx4DKWFEdZ6qX3tTHDZfBUUkdRkj54Q9/qEYTPYpKZ3YNLjM6GMcaNHp7e51KQA9A6iVF5ee6665zQkj18qL6ode//4knnmDnkPe8rq7OWW71rrvuUl1dXfm/P/74406c82iiddez/vQMu3Q67Uo/fPjwsM+pn8GhKkJdCZ133nnqtttuc/7VwO9ab3vb29QzzzyjNm7cyNLlkrZjBXR1A10HgK7ua0HXYwO6egNtw6ktdHUDXQeAru5rQddjA7oODXQNjq6hn8mlZ1j97Gc/c4zLtHnbVVdd5Xho6dlbemaUNkD77//+7/zI5nBYtWqVuvzyy9X3v/99ZwBNG6bpB/LVV1/1HdkcLrogffOb31RXXnmlWrt2rfPAa9EHR0iDhi74ugJ45zvfOWS6vo86/lf/Rm3ip8319MChvp9apwceeMCJ4ZXoAnvppZeqdevWOfdKxwv/27/9m1Nx0ErjWNGVw6233qo+/OEPq9NOO029//3vd/K7d+9e9dvf/ta5vr7ucNADsXoRgl/84hfquOOOc0JSdb71R/9LgZ6FqJ/fq6++2tG9qalJbdiwQe3fv1+9+OKLzjm+8IUvONNP3/rWt6rPfvazqry8XP34xz92KtmjheyOBtDVDXSFrtAVuo6nrhpoG05toasb6ApdoSt0ha6FUzWZdLUnCS+99JKzjObs2bPteDxu19bWOvtbtmxhx+nlL/Vt0cthSgbTKN3d3fa1115rT5061a6oqLAvu+wye8eOHc5x3/72t/PH6WVC5RKYeknQt7/97a7rnHvuuc5nEL3s5uc//3kn76Wlpfa6devsDRs2uI4baknSYuSSSy6xk8mkc++8uOKKKxydmpubHS0uv/xyu6yszJ4yZYr98Y9/3N66deuQv/Xuu++2ly1bZicSCXvFihX2/fff73xX/03ep5tuuol9d3Cp13vuuYf9fVC75557znX8RRdd5Cxvqn/P4sWLnXzrZWXp8qvl5eUFPUtPP/20sySrXupVLsW6c+dO+yMf+Yjz3Or7MmfOHPsd73iH/ctf/tL1nOtnQudHH3PDDTfYt99++7gsqwtdoSt0HQC6QteJ0lUDbcOpLXSFrtB1AOgKXaErdD0alv7PyIbHgBebN29Wp556qvqP//iPCVk2HHBOOeUUZ5RaL5sKwgN0DSfQNZxA1/ACbcMJdA0n0DWcQNdwAl1HTug9ucYjNleiwxf16ofnnHPOhORpsqJjizOZDPvbY4895kyF1DHBIJhA13ACXcMJdA0v0DacQNdwAl3DCXQNJ9B19MFMrmPkG9/4htq0aZOz+qFedvPBBx90Ptdcc41jtAbGD+2zpldv+NCHPuSY+G3fvt1Z3aK6ulpt3bpVTZs2baKzCEYAdA0n0DWcQNfwAm3DCXQNJ9A1nEDXcAJdR5/QG8+PNdoIXk8hvOGGGxxjuPr6evX1r39dfelLX5rorE069FKtejGAn/zkJ86KDNq07u1vf7v69re/jcohwEDXcAJdwwl0DS/QNpxA13ACXcMJdA0n0HX0wUwuAAAAAAAAAAAAABB44MkFAAAAAAAAAAAAAALPmA1y3XzzzWrBggUqmUyqM844Q23cuHGsLgXGEegaTqBrOIGu4QXahhPoGk6gaziBruEF2oYT6Dp5GJNwxV/84hfqIx/5iGOYph8gvdrgPffco3bs2KFmzpzp+91cLqcaGhpUZWWlsixrtLMGhoF+NDo7Ox0DPL1a5LHoqoG2xQF0DSfQNbyMprbQtXiAruEEdXE4ga6TQ1cN6uJwgDZ28pRZrwNHndWrV9vXXnttfj+bzdp1dXX2jTfeeNTv7tu3Tw+64VNEH63JseoKbYvvA13D+YGu4f2MhrbQtfg+0DWcH9TF4fxA13Driro4fB+0sSr0ZXYoRn11xVQqpTZt2qS++MUv5v+mR9n0spgbNmxwHd/f3+986OicZtVb/kHF4snRzh4YBpl0n9r0h390Rq2Hq6uvthd/SUWh7YSR1bo++P+Nuq7LP/gVFS2BrhNFNtWntv3nDaOu66mXfBnltQjK7Au/+daItC32etjKDeRnPLHJv8JG0vz6dpTsWN7fc2GNb13speubLgxg38nvEbCC2Xd6/o9j0Hd6a3GU2eFgDSNYxLd8FQFOeX1o9PtOp7wLbexE67r5voH2VTNadfHxV34VfeIi6BfvuOObo9rGnozyWhRl9kVSZr0Y9UGu5uZmlc1m1axZs9jf9f727dtdx994443qG9/4hjtj8WTwOmohRU/LHK6uftrqygHahlDXkiQa9CIA5TW8jETbYtd1wge51MQNco2FroHsO4VskGsQ1MXhGuQaS13x0jzxDIagjVobq/vECegaxr4T3nWKg6OFjU746op6RLW9vT3/2bdv30RnSVm54v0EiWLUFhw70DWcQNdwMm66jnDCuZXlHz2zKv9JiU+/9yfaxz++x6Zy+U8kY3t+ZN70gNzgx/f3jwOBK69+z8FIvzcWH7/rq3BrqwekCvrkCv+4yjD90LKeFuVSpNFySY+TZdiVB5LviWbCyqxf/UvfMbIj/BztvWWkZS8gFGNdbNkFfqRefrraY/8pJopRVzA8Rn0m1/Tp01U0GlVNTU3s73q/trbWdXwikXA+oLgZrq4aaFv8QNdwAl3DC9rYcAJdwwnq4nACXcML6uJwAl0nH6M+k6ukpEStWrVKrV+/nq1IoPfPPPPM0b4cGCegaziBruEEuoYXaBtOoGs4ga7hBLqGF2gbTqDr5GPUZ3JpPve5z6mPfvSj6k1vepNavXq1s0Rnd3e3uvLKK9VE4Rvq5zNF0jXFme6KJDbV8limRpMYU1uEm7LwUz+/EJkWCaeu4NgJm66+U56HUyytsfeUkeV7UupK68rheLKMtI4NiO9LILQdTtvp8T1p/E5DAa2sT9pRvLxoe2hHLM/2MJrmnYNczCTm4hFPv66c7D2RS9iy0FtFqOsYhIYUHBI2WnW0b2YKT2N9p+HkrRh1PQqFlktXn9kurF8svxftz3pfTtbFVIYoT6N1gUzLUR89kUbLrOUj7kj8wMZd15E2eX71NNHLVd/6vu+M0GvN593E1R/yeRcacb8qYGV2VPqzUjvWjnpfT743ut4jR6Gulm2z5fc8jALFomsxY41x2OhYvveMyyDX+973PnX48GH11a9+VTU2NqpTTjlFPfTQQy6zNxAsoGs4ga7hBLqGF2gbTqBrOIGu4QS6hhdoG06g6+RiTAa5NNddd53zAeECuoYT6BpOoGt4gbbhBLqGE+gaTqBreIG24QS6Th4mfHVFAAAAAAAAAAAAAACKdiZXMcBijf3iyWX8KfMXEDHqJGZdxq2yY48Sv+4X/09jlF0x0cRvwBXLTDLkSsuNrj9XUTNCfyRQ5HHho+BNEE37fM/l0yIz5J1WsI+B3/fC/GwW7NVjj8rz4HtfxYlG5b6HwOerIEboX6mXI/dqK2O9GZ5GPLIiKf5FK232rYyf0aZoK+PUtIfv03M6+8n4kPnUZEtMYZbNaC7m7TtTFIySz0bBfSdXmj3q3j/D8XRi/SqX1ynN+FGuwRKHvvZEM1JvNFpOXX1f1p/27hdHMjwt3tZndnJHKbMl5pUkV8LLbITUC9kET7PIsbm40J2Izcqo80VV3Nij77vl0o76I4pyGe3z9kCUZdi3TSVlLyfKJa1IXd6JUdvze8zLS77v+PXr2IGq6BiOB5ffOyclIm3xct5ptPy69Bjh/XK9c5Lz5OI84zmS6Kp6J8jftmh9r8bIOysqfFLZJZkvuM/z4fM+5SfVaOsY9uEOAAAAAAAAAAAAADAJwCAXAAAAAAAAAAAAAAg8oQpX9F3y2GfqtSuNTNeMZHOeaa4wB3KsayqvnKbNljcXY42+S5j7pJGp2GxJZeeH0AO9l1gODPYIp+1TGSZgeqtv6EPAptuO+TRdv1AK17Rtn++RZyDa7xOCIe6/7/T4qHfeZEgEPdZVLn2m8AZt+jVjGMuLFxqS6F7qmoY/ibScz5T70fqnHXLabHw437NCGaLo0oCG9ov2MEKmw0e7Ujytj4QvZmQsheV9H1M8FjnS22924ryrY5cmhvg17vPaFv9ehFw/62MJ4A6JDZbmvuFutO/ktzS9CFujfSmZ5tt38gtXlGWb9J1oX2kgzTtkytcGImIHVschoeFpLh2UZzgwPZaGDrrCiGVaKuMdKkfLqKbf1AXRRInIN7l+eSlLypWb8pxN8jKbIyHGKhkp/hDjgvu2qvB3GpqW9dHcFa5oEmOd/d6a+4Quusol2Zah4zkRhpol7zi5hNCOhKXKfhW3iPAJj7SCFaLoakdp8+jzPMh+b4zsy7Rov/litJ+3v+kKXrayQhOWbZ/QcapXRopQQvtu8pze5ZXuBkLXEVhvDMsSYDh2LwIWwpwSfZmo93uQXwgxsxLw6YOP9nsQZnIBAAAAAAAAAAAAgMCDQS4AAAAAAAAAAAAAEHgwyAUAAAAAAAAAAAAAAk/gPbl8vXioP4SMZab+AsJ7gC1h7uMvYPVxDxCLeoKk+bLoKiu9Rcj4Ykwsb15CljAXvgQ28Ruwsvx7dMllS8ZAU+8Bly8Q9aNQxclw4omZ7t5prnMy37LCr+8Xd+2KJyYHu5a6ljHM7HsqUIw4Ft3HgymS8facoN4Eru8R/59Em/DtSeU8Pe6Yl4fQS3oR0PolWyLKHtVZak7j2yPF7zEwGgzLd4vo7PKPYR4xto9/mvDYkZ4khZZfUT5dS6Gz61NfJ5FILuhOswLl1cQ9fGSZtIcsZwP75ouRnpRn29i7oIYltS027WFvLb9e8jC/d3MeOJDfto+08XxnKjz9uiziEUO3XW2LqIPsYiykI/WuLLBMujQn/SWXpxPpOzHfNQ3pO1nSh83Pz1Rop0jfKVcifNgS3rrSut8Wp2RL2hN/LufYIimvvh5qft6WsltKyikto3I/0s/1o33hXGWSpR24aHp+O1XNrxfv5PtzHzhkdg4eEgeTvq+47xbxeYr4+K25vOCYx09xakuhWXS1Y7mh617ZPsp7ECWaSz816slEvbM0XQtMHZop5fdK9oEqGszzUvp6s6cfn13Gn51Iuanvc6W8YGaSpDz79NVyUVm3UXMg8f9iwFVeiX+WaCqjxC/JVZaJzrFe4cnVQ7zWetKeXmvS17BEvg+TOt7tgUj81IQ+WeKNF03ztHSZ2c9KW76oj18X6XPRp88Kgu+Wy292hH7i7D2o8DbBj3g3r+tzzBdcemJanmm0v+zSjqap0X0PKtYhDQAAAAAAAAAAAAAACgaDXAAAAAAAAAAAAAAg8AQ+XJFPq5dT97xDXWiIopUqfFo9XfJYTpU/snp2frtvCp9X17mIZ3vhr3vz29GXdvJrlJnlkSNk27mmTabz+kxRp6GLzjnZ8p1izl8xTdUdAe5lsAsLU/Vd9liufkqfLTG11KUDvdliGJlO9bSyMqyNLn/unR/fafTFN8PejV3Yssd0mWM55Z6GLsop9zJ0zS8Eg16/pEnEToiwGbvUzJ1OTS8XGSebZVy8jM902xwNvXGFrqlg4VMfsSnd4jAaWmEJXVlYeco7lCIiQsf9oNPxB64pC7QaMkTDFuFP6cqSocPPxXRsFiruXJBsu5bIVhNPgcuZu9pYV4iMT+hTrxE6V8ZjEppPqcpvt54itSHL1mdk/cmP7J8/Lb8dbzrME3tN+2tZZZ6/yTcMLEz41MOusEMWJizSSBibJftOJCTRTnKxMnUmjq27NsHSyhtMn0sTf2Wv58+wSmnfiYc+2baIfWEZN5s5S4Si09BjFRAKDQGX+pFySutX53u96aGtOXRdOMuU2cOn8j5r50JST6dk6Di/16k6c574/oP8WLpDQtwGzkP7ZyNUyS9OZoKQoVA8VEn5lEvvPlC037vMynaUtpXp6STEWynVscC0jZ2LRcOd5Nc40mraznnKhK9qSp993fO5iuZMP8ui7z6yH+zTJ5b95WKwZPG1bHC1o0OHJ2pifT7vuKT/HO/i+sQ6TZ2aJSGhmoazjc5dC/n3ZmzkfduaHT3mGk3t/CeRdx27nGuXqTDXtDIirJzFp4kwVHKoK6w8RkIZfawkxhW7wLIsNR9hn5hbCXjXkTREeKi80fuXI2Gnmni3eXZyMuyfhBBniT2Acw2ilyuUkd4oGW5+jFIWQXEHAAAAAAAAAAAAAODYwCAXAAAAAAAAAAAAAAg8GOQCAAAAAAAAAAAAAIEncJ5c7qU2be80H28mulQu9eBy9on3gPTl6V42I7994FwRS0yGDCPC+kcu7/rG24xvweIdPCba7uo235Ox5tTTScSuWixN+ErQZZRd8eAklrkYjAgG8bFW8PPW8vUMof5Mchls6tXj5y8kPH1cHhDUw0PELNuJuGfMMvVRo0vsyhhmS9wY5jlQhL4SrjzZhZWTqPDkovvRPv5Fui+9RI6cYDwGWk4Wz0qd8eaJbTNlW7PgF0382ENH8tvJVu7flaupNMfVcP8Bq9JobpWL5ZJpGZY+bMWOXxnxtrly19OkGMryyrzWpObUI6Y35Zm3jPASSdVyD6auWlOPz3jOaOyct8Esdx6p4N9TOXNeK8d9hLIJUhfnpHeU9wLXxeAXIvFrY1nZlp6YvvWwEb31lCksreVM2v6Kdixt7ldJG0+rfl08O9QT6rgFPN90mfR+72fH1W8YDb+fYqiTpd+Yyy+kQK+1/sL7TumZpo48cjz3beqbYW5KuoJfr30RP7a+ndTTr+7hGSdeI9RLy4H2j4TPiKL7sh5m98nHX6+YYL42ahiepdSrVvSP+owXS8cpM1naodPM/ctUct0rd5kbWr1bPC+ire6cZ+rRRNWJLK1ie6vZIXkZFrI/7V0VFwc+5dLXk0v4M1EfLlYvSt+tqbyNaz7J9GXaTxL1pDL7FvFDcrLdxwtRtI+Ub9EHKi0x/SO7r4/njfgeW8ITM0L6y9L3kr0LiTJL33/G02PPz5eU17c8jfrPSi9a6r0m+8ux7syQfSVNy8nGAzFzOSlXSqmV097Ib6dyXMfISfwaLz52XH578V1dPHMtpi8VSfM+WMwu9yyT9N3Hjni/0GTF6AV9D6KST6SPItXc7bulvD30mK7Cz5R60Yp+FS3LVr/wqU2Lh4d+T3h0sTEBUj6lN3FU+tvawhiVQZ4laQtO35kjo9t3KsIuNQAAAAAAAAAAAAAAwwODXAAAAAAAAAAAAAAg8AQiXJFNy/UJfZLhM35LdNLlNV3hZ2QZ2+azallS8yoa5ijX3SSbYgpmSTvfn7qDXHNqDc9bo1nu3M6IqcU0BED+psCsb+2BnMLLwkaGMeWehVNkfZZL5vc2lzTFoW1ZJU8j06FLW/k5E618Gnd8f4vnNFCbLGtulYkQp1Iz1TMiltbOlMY8p2arKJlaKkM0iiFMxnfpXJ4W9Zt+3UtCEnu9p952LCJToUWI4tRlRhtNT7+ZetuzmIdAdB83je2X/6nVu64hU4FpyIezz55V5TktWdYZdLlksar9MS+rOxb4LW9Oy7Lf1GxXWaZLn4sp9zYJMepeNpWl9U0xU6Pbl/K8ZBP8GrlSc43uefw8C79vlrG3O4WuSfPsxIR2VpaEJZd6x6FmS2R5pfPbJ0hkVxVC/iDCB9hzKeoeugy1XNKb7ifbZKHwzlrykLmXM//Mn4fyrY1sv2+xCana+25eJ0RIUZ//kAlZ1sQPtuW3LRFubsVN5iIi5M2m9fBEhrXpbNhHCU8cPG4wzSfUlG67QiJESBsNUeybZ0JiNIdPM+1d7wzbu34QkWiJNtsz5D+S5G2osnOe4ZI0lNFVfwe97yTx+T0sNESUS7pcvPvZNw1S4gjXPT3DfC/ezF8r5vzehHxbHcaKQ9N9cp1nWy2Xrq9PmbDmstfMOeXvsIVVBy17xdhu+oY4+bzvuEKnafsryyztk+R4O9Y7x9SNB87h2lnzTQhaNCusN2hxEiH5la/x88zYbAp18wpeZrveuyS/Xft0O0uLNneYHdE3oOHxIy2/+X6Kj73CmOD3HuvqO6mCQlRpeKIm2mXeSw5cyOvipZe8lt+uLeXWG3u7TTl7tZFbeGQO89Bxe465Rq6Kp0XIe6yE9t1ypE1110/ie9EiL8s+7zquMFTS76X9XBle7HqHpWMXIiTRJiG9bSfw95fO+aS/XM+flZqXeXmte7DB7LS28etXlHuGMloRb0sAZq8UixZc5x1r04yZXAAAAAAAAAAAAAAg8GCQCwAAAAAAAAAAAAAEHgxyAQAAAAAAAAAAAIDAEwhPLl98Yjn50ufeS4GzeHXtLzDHeGQ1v0kEQWdNXGkuzs9Zvd3EmdY+zeNYrf1NbL9n9aL89vbrprO0yp2z8ttz/svETh8LbElV6bkyUd4vR8NnyV26L5dL9lu63ur3Xi754FoTU95TJ/x3mFcUjyeOdfPzzPsD8Wd6fR+/flYEZhPoWbMWj3WOsoBufv0cNZ0QhaAotJVZKHDZbqkdi00Xz0PTahMn3nkaX4a6vMrsN+/knkvl+829rOnkJ+2cK7K91nhHlO3m3hHS42VEFIFUY4VFLADcvmQ+/gPEey1dY/zsNF1zjCdW07myXHl7tinhHxI/Qp6BHaK+rzN1sTpslsR2fkeX8XKKEC8C5xLEz8bKiH9LIt4ELp+kYngIfNtR/2O9vkf9U5w04q9y+GSx7LRl/FtKmnldN+/3xj8k1sCXPrdLuddLC/F+yUwRy2krn/aQePrYQtcJ80krpr5T1tvP1E4aLRvP4Hr0zaKeivy+Vu0029Nf4r5Nsn928Oyq/Hb6ohNZWv1DxsMn2sjL68jbK9q+qkDA/BFH6j8m6yZyH5pXSi80U76qRZfVaje+TkqUpwjxnnGOzZpXEruGl9mGc8yztWSPeHWh5w2IRqMN927yLrOZSq5d80pzX+OLiQeW7ocSb1jqweWc85A5z/TN/KZP28Q903Jlpq3un8qvn1li2tGWblO2NTM2pLynZPi900hvWo+2x5b+thOEb5n08Z1mPmxC8+6FFfnt6W85wNIWlhtv2pfbZ7O0PU/V57fnr+d96a45/H41rTPlsEnU93UHzHuRnSjh2SbeUbm48OuM+L2rFvheUSztKB1n8OkDSZ9afpwoeKQ97Knn5aXxDNNfSs8X5pb0Einer+pYyq8xdbvx80ps4u2o3WXaZ6uKe1dbxO/P5Rk4QWAmFwAAAAAAAAAAAAAIPMMe5HriiSfUJZdcourq6pRlWepXv/oVS7dtW331q19Vs2fPVqWlperCCy9Ur702OrORwNjR3rJLvbzxDrXxoRvUn371BdXSsJWlQ9dg0t68S73y9E/Vc7+7QT39v3+vWg5A1zDQ1bBT7XrwdrX1599Qm2/7vGrbvYWlQ9dg0nF4l9r+5E/VpvtvUM/84u9U636U19DVxb9FXRzKvtPvb1BP3f8F1XIQuoau73Tv36NPHKY29k8/Vc8/cIN69m60sWGi68BOtfuBn6hXbv+62vKDz6n2negXT3aGHa7Y3d2tTj75ZPWxj31Mvfvd73alf/e731U/+MEP1M9+9jO1cOFC9ZWvfEVddNFFatu2bSqZ5OEmo8JIlwem0z7FFNDmlSZszY7ysJdIxlyk/kE+xa/sZT4llCF+O1uqN8rP07nYTOvPzSfhMvrQ1i7vZdktn+FLOn13iPuUzaRURdVsNWv+6Wr7xrsmXtejwKaF+kwRpVM7NbkKM222+WS+5G33fO8QpykvmumdORFdY0dEiCC911Gx7HLaTMG3UnL517hniAZdut417ZU9ymJJ8GxKldfUqZkLTlc7npkgXX2W1XWHESvvpYXpPfAJ+1w+/yDbf3m3WaZ8we/4fS19w4Q8tZ3El9w9fCmfqp0uNVPnu37Hj53+QqfnVHnfesmv/vL5Xi6TUqXT6tTUZavVnj/cOX66jkZYpjyH7R2iqsgyxAfO5flOH2fCHFRKVHhkv3QfL7Czn+bTuJOv7M1vt1ywgKW98ndmOvi0DVzzWb/ZZXYy3ss854Zxz3Q97JTXhaerV592l9fvf//7RVUPF4qcxp8tN/VwzxISkqIh4eAzXuDPAw1RlOGJHSfz5c47F9D1u2UGTOGKdfR5hz6NdK67NXSZLa+e4Lp4NMLWqCSi/u5aZMpL7xyxpH2PuZkzXuDfq3nRhM9YHTxcsfukOfwa9SYDtrCMaD/OhOhMae/h+fYJYfKro4/Wr8z3nepPV9ufm3hdnTzT6L1c4b+HPRPie62nmVD/7tX83lqHze+Y8VsSe6qJmb5uZg63C2g8U4SuVZP+US/vO0XS3mHDVr9JjIil62l7HBXPQNaiocnW0OV1/ulqx7PFoWuhyBDVDOn3ti7n97z3RNOOVsR4O5ZKGe2Sr/D+8pxHTTmNN7Z5hnxrWtaYtjO9mLTbOiS9hFgSVAgNkjzMjV2D6iW042Fi/HuZXEqV1dSpGYNtrP6qNc66+oThuSLiCyyvjWu4bUrn8aZMnF5KQoaVUo8eWJrfjv0vL5OLnjN1cc+Capa29NPb2P75SaP744uNnYcm85xpj6PdvI2n766u91garhgpPJQx3y+eXqemnLha7f3tOPaLC7FzkN1e8ttk2GykJ+sZrtg/y+i8/wJezmJ1pkzaPbzfGztkylJZI79eotX2tALJlB7P0sq3H/bs9470/cDPTmncB7kuvvhi5zMUepRUd8K//OUvq0svvdT521133aVmzZrlzPh6//vff+w5BmPC1FnLnI/N+xUO0DW4TKld5nyGAroGl6r6E5zPUEDX4DJl9jLn48Wtt94KXYNeFw/RiUOZDX7fCbqGC/SdJl8bC12DTdX8E5wPHQwbBNpOTkbVk2v37t2qsbHRmQI4SHV1tTrjjDPUhg0bhvxOf3+/6ujoYB9QXOzZs2fYumqgbXEDXcMJdA0vTU1N0DWEjKTMQtfiB3VxOIGu4QS6hhe0sZOTUR3k0g+QRo+MUvT+YJrkxhtvdB60wc+8efNGM0tgFDh06NCwddVA2+IGuoYT6BpuoGv4GEmZha7FD+ricAJdwwl0DS9oYycnww5XHG2++MUvqs997nP5fT1SOloPkk2Dm318JVpP414rbaeR+OE+Pg5YvcOcs+xl7v2jYlHPwOqm8/kyrZkpJg7dSnuPNdrSX4BeQyzHzJY+91vqfJyWXh1bbf0SyRLBCR5/2b7YxDN3zxG+IFtNcaj9E182NdLW5elD0r9kJtvfd5FZVjVxhE+Lnv0QeWbSGV8fhUKhWvstnRwEXf19C7x/W+9sc+/6yTLkmsRe40FRuqtJXIN6A/CkJPHg0iyearwKdlfUeObFpUGhS9BP8GrWx6zrCH8bfe5zor5rOqPc7KwwvmeabKfRtfxV7t0x9/ft+e1oK//Xt9wUvuxx52qzZHbTOuGFFzP7vTOEP0QlyduIfdiEl4hdfLqyPApvK+pBUWidrOlcbLyTlMW9CROHTEGsekr4+5SQpejnTWFJh0/hz06uIuPp2WaRfVv4JlrK25uRYRfuvzEWZfvYdT327GWmlQs/U3ov+X0sbTT3vOZ53ney+kxdm5vOfWBaVnBvkVyS9J2IR6omS5ejt/zq4bEvd8XSxrrqGNfDSQ+mXkY5z3tr5/g5k4dI+arm9WumxvS5mk/iz0v/NH6NCPXhEtksafPrz5P9rKijyL6fB6jLuywS4L6T9KGqNPe1ey6/BzOmmXa1q4/7dUV3GR+u2me4lyX14cpVcr+uthW8f9SyymSorIz3qzIZc6MTbTxvEVovVAjPJOLlRDV29ml7LDx03aZXRdZ38usr+fjrTb2ogaW9c6ZpO3/x0Fksbcl/G+2svsO8bV5u3ocPf4B7711YPjBQNEiadJpPnsavv23ayvx2aSrr+V7r9t2i7zM+03LG4T12uLrKPhAbghDeY+yZ9fFZls/rkePMA52sN/1cTU+7KYeVL/M+8fQXTflN7jHepprOlfwddv9bSAYS/KGb/XszSFjzAn92+I+wCvcFL/D5HwmjWo3X1tbmQyooen8wTZJIJFRVVRX7gOJi5syZw9ZVA22LG+gaTqBruIGu4WMkZRa6Fj+oi8MJdA0n0DW8oI2dnIzqIJderUA/LOvXr2cjn88++6w688wzR/NSYBxZsGABdA0h0DWcQNfwoqfWQ9fwgTIbTqBrOIGu4QS6hhdoOzkZdrhiV1eXev3115nZ/ObNm9XUqVNVfX29uv7669W3vvUttXTp0vwSnXV1deqyyy4bneWQs37T3OQUdLmepxpyKna2RMyPI19LNPNQhtm/2m12SsQ8WHLOQxfMZUktq3jGrb6o928ixPabECmNXVnmHb5V4DS/ocK+spl+1dvdkp9u2dfTqmLtAyEFlmWNia5HxWcKI53OKcP86JT0VI1YZn4BmRrNb62a+byZth1p58uY2xXmvreeysNkDq3m17dLTfhN72ye8amvmOV6E3t5SCSbziqXuo6bfOfEcrNDrYiZ/16mX/V1Nef3+3taVfdE6zpCqM5Na3jYw5vOfiW/vbuDL4m84AEe5kZJzzBhU21L+Zh/ZQkPo+rNmPLeNY/XLelq85zFuvn3rIw5NpIWdRSte4YRMpNN96v+dqNrqrNV9bY0FI+uw5lyLEINKDky4zqd5g96co9JnLWRh1LQEMXsTB460XCOCFdcYb4rq0Y7a/5QfkCW8xLP0HFafocTOu4qr92tqvvIAaUiA7/9k5/85NjrKrPrF91EfpsrDIoWEVFHszAEEfoU7SX3bvZ0ltY33UzHbz2B1+3pSl4mo51Rz7CHkiMRz7Asz98gfsdww8vz2lrFVxfLMAu/Z5b+7t5ZPIQoNdXcsGg3L69z1pvQCqu7l580YcpS4zrevnbP5x2kCLGQkOGK0RTRJ8XrYTteYHd3mKGLg30nRftOHcWh60AmRhhyK+rleI89ZL2oiZDI1D1/xT1v+mqJfqJPHu/gBbOk3Zy3fyq/fo50t10hxtm0d3+Q7vtpK6uvdJ/q7WrhdXHJ+OrqF/7kPpj+Tu8fmmgRVipvmP6SVcZDjGe8araTr/EZMHaFqYvbTuRtbNNafv3YNFPe+/r4exN9liJpu/B7kTbPktaDX5DeC9GXTrvb2J5xqId9w8NJdqO8K6Pi3eZ3Jtq4PrQNPKGch6691D4nv73oXmK34tS/ffntdC0PD2+43ISIvn3Rdpa2u4e3x+Uxk9mGHn6eZKMJdbTjvLzSdxjX+4vPe18uepT32FS/SrU35+91uqNV9TWPcb/YJ78sFFOWyQLLcnc9f9fpWGW0i2f4zSt73bSjc9abkFRNpMPokZ4lLQH4eUpneBvsd882s9mqXk161rW+9hXDwRrnQa7nn39enX/++fn9wXjVj370o+rOO+9UX/jCF1R3d7e65pprVFtbmzrrrLPUQw89pJJJEU8NiorOtv1q69O35ff3bH2ApUPXYNJ1ZL96+Ykf5ff3vPQblg5dg0nP4X1q529uze83bLifpUPXYNLVul9te9yU1zc2D+g6rf5U5/+6k5bNZqFrAHHq4idRF4ey77TB9J12v4y+U1jK69anSJ94C8prGOg6sk+98qiph/f+GX2nsNB7aJ/add8t+f2DT/6apUPbycewB7nOO+88Zfv8K4EeLf3mN7/pfEBwqJm+WJ31zu+yf/HOpPvUs7/9qrMNXYNJ9YzFau3lN7G/aV033v8VZxu6BpPKuiXqlI//M/tbNtWnttzxJWcbugaT6pmL1Zl/dZPrXwF1mW3Z+wJ0DXpd/O6b2L9Moi4OSd/pku+6dH3mQfSdgl5e173ru2zWiFNef4PyGmSqZy5Ra973T2xmUDbdpzbd82VnG7oGl4q5S9RJn/4em0WU7e9T2277B2cb2k4+xmD9EAAAAAAAAAAAAAAAinwm14QzDL8Q3+PYUskyzVykjK90zclwr4i+ZbPz223H80OlP4XrmoRkc4HLYMshSh+fKnYK6Y8yBkvqjjlMP+HdQJan7ajnjziN/57xEl/KONZgllW1y/mSyM1vMr4FLSeL+xcXHhD9dHl6mTdyr2MiGJ34+uQSUe8ld8VStGOxdO5YQv+VRT57NjVbkyHsROeMmF1cHjNaNh/hnktTDxNdy/gXjywzOlsreRx6Msb9D9r7zXftafzZ6ZllYuFrXubLLkeJHwHVUXpQREVtnCX3xs93LWi4lvv2qbeixn5A5VL8JpQ3mGOTO3hFbVcZH4Pmk43vmqZzMddVkSXMXTaOxNtj+tPCk4T4COXKuHjZpNnPlUQCXV4ZfvZifm2s0DzWZ/Yjnfze9dQbffZezL1eeufSNpe3v/G2iKf3TP808VwJ3xPPfMt21Ke/wZ5judR7Mf5zoo+faaEdq0iGHxdJmR9ackT4NnUSjxbhg5ZaaLxeuuYLrbpEnUlkl36msX57ZJ5FPpr7NEnFC21jZfVD9m3h78m8QEVbVbnb+JQ2tnGfmJ464o80VRSuVuMbVL6Pn3PKa7wujnUbQQ+cWyL6TmRH9oEK9KMKAsy7aRg/hbejPK20yXhiTU9zL8Ns0tzn1DLuY9ddR/ogb57H0vqmmbTeWulXyXVNt5F+l+gTR4h3YqLd26A40s/PmSOvrxbxrJW43+/sCWmMmSauvm1hafJdp+oNc0+2twysHjjI6bV789svLVnA0nqnmz5yx3H8np80b09++5V2vupgcxcv90unHTbX2GM8wDTLekx/OjO13Lue8ZHA5evk439VDPjl1+X36tcnIDJ31fJ+76yZxk+ucfc0lrbgj96+0tnpxkvr4DquR//x3COzhPyQrPBfTJBiGEmJMkn6vb4GguNYRRdj1wsAAAAAAAAAAAAAgGGBQS4AAAAAAAAAAAAAEHiCF644nCXM6fTdrJizSnYrDvIpuoezZuyvcz7/WvStCz1DEnMlJPSIRzOpWDefuhcnkVG9s3i+E0fsoad1SnxCHt3HTswU3VFDTuGl2qbE0vE95uZbWR52mEuYL/bM5I9/qmpufruvho//di7wngoe6xDL45LwRUvc69L9fJlfSrYsPvTSsyoAoS9++Ezh9fstrlAhInOMhLFp2tNkOvw+rvnuDxldsyv5UsqJhFlmly9srdTBFr7Mrk2m8FoREYYqwyfo7yBlT043p+E+kbTls1yyT8hJEIqzTygFrZusfj51vnK/2e9cxBVKVZsf3npOPUvrrDc3qG8mrx+sDL9hsXZzozOV/NhYizmPleLthF3Kw2kKJRe3PJ//QGhZIH6hNqWNpgBX7yhjaUdONxr0r+Shv3aHueel+3j9XfM616602bQDDWclPEOfZAixX6h/INDP0OBzRKO4RPgA7S8N67kj96ekjZeJeIe5sakafh8PnWdCX7rm8QumSbmTfaeyBu/M9c4oMDxR7rtCmDy25fdcFXGB5xhjXNrSPPssa++qf0j4ok3C7DXRNlMW657kZTZVYU6UquTtb6Ld5GXa8ybURpPb+QbfX7WM5EXUr7TNI9YOrp/oskCgv9cKT3h4zrsdlWGouZy5X8lGHsZUtduU2aZF/L5mTjPhTy0pXt/S25wlYcqaxF5e35aQ953uuaK/3kdCo5I+YYcZ0Y7Tx0Pci0LDPPPH2UX0PiP6iHKfUtpgtDz8LA9dy1xyIL+d/WALS1s7c19+uyvDy9neTmLN8nAdS+uez/tnHZXm+UjsFh4ikc4RvXIy03j+GPHyK+u1YizazJplGF8j9be0BOjpJ32ggzFvy51SfvPajjMhil3H83Zblrr+XvIu2svbgVi37WnZpCz5FlUgY1j+gva6DAAAAAAAAAAAAACACwxyAQAAAAAAAAAAAIDAg0EuAAAAAAAAAAAAABB4Au/JxXAt/VxYoGdyF49XLt8zO7/dvYQbRDSTZXMtEYce6Yl4xojP/w33Yjqy3Czn2VMnlt/18ffxWx6Z/l7fXx5AyxGpZYT4cMllTK20iROu3s3161xk4pkPnS4vQre5gJGUSazYI9fk5rtdZKXlqVvFNYivT66Ke1f4LrnqnTWVk8uAF9sy2DLNx0/Kpj5Uwj+DlouqPTwW/KUG4x1QsqSD+w/MNUsit/RzL5GGLuO7deiwKZOamX/gXgWHV5ntRD339sqUJgv0grFHZUnwYijC1jCWumbeY9Jbg9wfK8t1TRwx5cXKcj06V5il6juXiTIZJXVCP0+b+oLwGCCrJzefxp+5ZIvPGtD0Z0gfDb8l7elDL4UsvqJcOPLZ9rkHscPGr2PGCzwtU1aR3+6dyT0eKpvMDZqz/gi/3Lad/PonLTVZi/NnJ0J9eqKyPvczwlEjgrXNBdbzRYNPZRNv4f4+1TsTnu1ryxmkbAtfvJJWUyZLicaaOQ80sP3D55D+mbFbHDht0vJ+Hqn33zD6irSeK4Z695hh3mHevpeur5WYslj14mGW5vK8VUO341avqbM1kQT3jbG6THq8ky9zn6o2eU1P4f4/iS5Siee82xdZfrPkebF4N7I4cPnDeadRb0FxC1TMJmUvy79Ytdv4I3a+yvtH9iqjR2kZ70undpj+0szN/JzVO4zXqaZnrqnTu+dEPPt8R47jbXPiSGV+u6SJ97n88PM/Koby7Oc3K/v3lrA9YhBvutnP8LL1XN9J+e3r/uZXLG0TMZve+IcVLG3200bneXsOsbTtn+G+X7GIyazFbZ5Uttq832RL+VBDptTkO5Pgz0M2Eaz20fddh2xbrnaftkfe56/cxyumfV2m7sst4prvf4/RtWMFL69WCdkXYxd2S8LzmaP+v5p4T4Glxh5O++vTJ/aovgt9SjCTCwAAAAAAAAAAAAAEHgxyAQAAAAAAAAAAAIDAg0EuAAAAAAAAAAAAABB4Au/JxfxcZCwnjXO2fYb3RKzsrOdMnOveGu7lkZlpAo8jvfx79Q+btMSGHTwrXTyePHfyGrPNbUdUjnpy+fh3yLjWUPhFeMQFR/qlR5ZPoDq5Z4mD3J9p/u+Mz8POy7mvQ67M3MHqbbxoVO4z16vc2sjSWs+Yyfb7ppvrT3+Ge1eoGPcc8Paw8UkT0ciB83uhWbS9/cVkbL6VMfcueZjHoi/8R6NPzzzj/6DZdu0sk9bPy3Ppf9Xkt5f98VV+vRJ+bPOpC/Lb5UnhT1FKvWD4bxpxwQxagSb5jfTzzEf7jD6R/qynh54l/EKi3aZOLd/Py2vXTHOjo0nuWxB9xZTzeX/kvkHRF19n+z1vPtHz331o3WzHY6PS9kjPjSDjew+4IYX4otmPHeJ19Lz7zL5dKrwiuo33jt3KfV8U8QzSpKvMd7OlwgOie2gvG5k3V33qk+bnAxOIenkkZHhZrn7N9HP6a4yfjqZtmam/Y938fsz+kym/ZRt4+cy0cZ1z5xtPrmySF6ZM0uffbX18t0Krj0bWP8wDsXAfPX4SaaZpez4TFjGIsokn6cDXZOa8vVkylea8qRpeFyf2Ke/fRLxgssL/p2W5Oc/0l4rRlEtAb494e7NZm+OjnY/nb92TPfxyj5sym03yC8aPmHo60sD7uXY7r9PVPOMPlRNlNldN2vF+8TI0ClAP2PHG8vJb8vUzFWl+9RbRsqSll3sZPmwauR/EL2Np5Q3mnIuf2C9OatK2Xz+HJS1bvpftv/aU6RMv/D3XPBePereNlrc+zOOqCL1ox5uyvfy+VjxvfNEyZ3Ov79K3mP1ML+879TQbv72Sw7wsz32Ev880rjHf7ZnH68VsCTVOFoqQutfV7/XxExxLMJMLAAAAAAAAAAAAAAQeDHIBAAAAAAAAAAAAgMAT+HDFgpcwV4Utw6pJvNGa3156uwhJKDchM1YPD5lSJHzCFR1ZapZT1aTLyXljcn5qYdl2h0v4hFmQZZyLCro8qNAvkjb7Ede0erLtF/4jzhlrMtM5l/wP/2K6wkyVTjaIUJiGJnPKXrPksqbzPbVsP9Fitq00n+ppJ8h07BCFLY0ahU5jlc8+CTkq29vJj/2/ZupteQWfwhvtNlOBrbiYKp/g4YoLTz2Q3+5J82P7/WpSv6LntyZukRZZ7yn3flOVbc8yaWVynnV2lCwpP+tZXk9P2W70iXfxchY/RJa+bhKhFOIa6fKIdygFDWUT4XAFh/b4EKbp+K6QhIh3e2SRsG2ph0XCnawuHj7DQqFk+KjY768xemXLuK5WOuq5vHmMtpV+/wwofm+OhFdbIvIpcCHlftD8R/kNirSZEJnZD/Iwi9r1tO0TjV+7CXO0Uzx0Ijp1Ctvvn+atj7R+8Mp34DUYrT6Xb0iJXObdp76j/ZyM6POk057hioqkaXrrjNVA3yz+jMTbTJktaUt52kDkSnh5tknYlKxgIyI7QapjXXUqbVf9GhLxvdiRHp/wI6NBXIShKqKlneVpkekmpErTtthoEqniYXUlCfO8ZFpkG+udb35Bse9zrF9Y+Whjk+sNGbao3Gl+Yf/yHYlZPZBtWSbr72/m3+s0mtu9XI/285fkty8+6wWWtr/H2HtoStpIm9cnGj0S1kbDKp192sYGrd87DrB3eWGnMOs5o93BWDVLO7TQPAPRTv69xb83fen4xm0szSozoYya3NlLzU6SP1fZpLfljm8bMUF9YszkAgAAAAAAAAAAAACBB4NcAAAAAAAAAAAAACDwYJALAAAAAAAAAAAAAASe2GT1F/D1q8qROGMSk+7st5O1x0WMuh01saqW8KpQdbPYbsdSET9NryH9p1iij19IAGOZ2XLWuREugy39Pei+SKN6xg4eYWkxkkZ9JJzv1ZjY58PvW87SeubzWPTqbaRY+S2R7ecj4KdlkI17BL7eTX7lWS47TMqwJf0YiOdTrLVbpJFyGBNeHt3cq2Dn9kX57YXHH2RprWuMD1jqhXKWVtLa67kMd191dGifPv29LjtY2rFlsOUy7t7f8/W2ouX1MPdaizeSk8ql6fuMb55VWcmzMpP7GLQtpb5b3svIS28E5oHh8kAk35NluUBvjvH0DvGF+sCIiilHH+ccvz+5DKmjE9zHwSb1MvXgciHaX5eXEz1nH/fILDtoyl28nZfJ9BSjc+907gOTaPBpi33qpyNLzc2Ytk08j8xMhyaooq+H/ZawH9Y1evu9fZtsb13Ty+ez/VS1twZVb2S8+3U+XmvUMyabEOW8SDSacOgzIcsl0ZPWvQP7RvecSIsIL5i+qcRbq5I/I4kW48EYSec8vU5zoo3NxSOeWtJyynyCigWfLLnKpV+Zle0xOxG5iNTVt23OepbnjtVz2X6KWDnleqU+5jyljcIvrDvt6etE/TKldqztLBZZye2LiG4G9R+O9nMNYr3EZ6mHfzHSY7zprD7hU0fKnfQ5ZJ5PtTNYWstfGc+niHgp29bA/YfrN5HyLH23yDuw25Nr6O2BvA29HTZ833fZgaJMtJl+zbzfdnv6D1r94iFrIx6ZZKzCObaS94/6p5rMRUukT+3I8s3TxL7PWE0kaw3/2vT7w/8KAAAAAAAAAAAAAADFBQa5AAAAAAAAAAAAAEDgCV64os/yqr5T/13Td+m2T0iCDI+g03nFFF06fdcW034b3jaT7dPp3raYQz39JbKkb4SPQ9p035VmFTZ9N+yQe2vRpa01ZN+W+tGQJ7Ek8qG3L85vHznJf55p5X7vMCZf/XyWh/ad+hlk/JZLHmmYiAxRpc+A0JWmyRBV1c/DnyI1Zsp3OivCr0heo70Z7yXMhYwl3Savbcfz56FtujnP9A0xn6W11bjhq0mh9a9fXewTnuYKa/MJL6ZluXu1CTPV7Ps/QrupZsp96atJllax3/Yuy2Spelm/uspvSGEhqn76+7Wx8nmgZbJfhGCkU57hibLNpaHB2YRYer3PPANREirinIdM5c+RZdCdNBL6JMMsupaYZ27qK7L+VpMCFqbsF9ImyivTWdTRPbUmTE2TI5JEUvzGxjtJeFM85lkP01Aa55Jl5tgjx/HrTXlVPIMBgNY/lk/F7K63yI5fHeZnwyAPlc8B/dq8OrZ/ZBnJd5x/r7TJXDNCQ8Ud/eKeZTYXI/1iniT6zKro4RYsPgfmfL4n32louyr7RzRN2rPQ81j85lU++irb75mxLL/dN5dnPJMyokxpEn23flqeRV1MQuByIhyOhr3Geycu3tii994nG+x9zS+Ui1hvuNrKHm6vYXd25bdzPeSd0ilbpo7LrODh4LOntOS3W1M8jC17sJTtR3vNebNlop4mbWUuzn9UutSkZRP8J/GMeicF7p3WL+zfHuF5RN0aIeVFlmXb55Qta3kYarzW6JpJ83KXaPfpE5Ny6Nu25LzDdeU5RxKiyM59bF8HAAAAAAAAAAAAAGDiGdYY2Y033qjuvfdetX37dlVaWqrWrl2rvvOd76jjjz8+f0xfX5/6/Oc/r+6++27V39+vLrroInXLLbeoWbO48TooHva99ohqObhV9XYeUpFoXFVOXaDmn3ixKklW5Y+BrsFk//ZHVEvDFtXbeVhFojFVNXWBmnPCBewYaBs8GjevV+27t6i+9oEyW1Y7X9We9n/YMdA1eBx45RHVcsCU18rpC1T9SW9X8VLUxUFm/w5dD29VPV0D5VXXw/NXvE3F0caGo+/0F10rpyxQC5ZfrEoS0DWMZXbOMvSdwtDGtu7nbWzdigvZMdA1eBx67o+qY+cW1X/kkLJicVU+e4GqXfcOFStDXTyZGdZMrscff1xde+216plnnlEPP/ywSqfT6i1veYvq7jYO/3/7t3+rfvOb36h77rnHOb6hoUG9+93vHou8g1GivXmXmr1grTrp7OvUieuuVradVdue/onKZsw0fegaTDqad6rZi9aqk86/Tp141jUqZ2fVjg0/Y8dA2+DRdXCnmn7iWnXcOz+jFr/j404Iwu7f38mOga7Bo/3wTlW7ZK1a+ebr1AnnflzZuax65fEfoy4OQRtbu2itOvlc3cZeo3K5rHr5qX+HrgGnvWWXmr3wL32nNQN9p5efQd8prGV2xzPoOwWdjsM71awla9WKC65Ty87/uFNmX3sSfaeg031gp5p20jq1+L2fVQsvG+g77f7VbSpHLA6g6+RjWDO5HnroIbZ/5513qpkzZ6pNmzapc845R7W3t6vbb79d/dd//Zd685vf7Bxzxx13qBNOOMEZGFuzZo0qFtgSu34ePnJpehqXTgqP9AixSnh8crpCZIB5+PCxxmhHv6evhCLL5tIldAe+aGJZO+bz61ccJL9JxMquOPNvzPUiSi097b1q44PfVN3tA2upj6WuzLsmYvssM2v5xLD7BGeLmGWb6CmXus51dua3o8uPY2ld9ZZ3PHE/16H8DXMeOxYtWD+b7GfKxBKvdCl7v1h0kbb8rKvZ/tI3vU8998A38vvjUWZlbDb1h3AtH8x0tgv3nCBeMJbwLbCo34v08SFpdi/3NLCqzb8AaZKl5tiqBH92mnaYmPZobztLy1Qa04FsibePXg23sVD9TXHPe7Hk4mvIOZSqP+/9autdX8v/rejqYsu7vI64Lk55+zNR34JMGb/nuYT0/TL5KT8o/BGfbzWnFL4f2fLEkGVX0z3b1L+JDuEl5lN+l59zNXs2Fq9+v9r066+rnrYDxalroYj62xf6PAjfxFyvKXe29MwrK2P73XVGg1wp17ykhXhy9QnvilLi75OIeS9bLzSff793vXbiOtrGWmrp6QP1cHfH2Lexx4qvv5z0WqPl18d3S2qXI+U3Us51zCaER0eJuWbpIX6jI3TZ9JK4t4ee8Peh9E/zq7t40oo1RlfNcae8Vz37h28Wta7MH0d4KbFyKuo72lZFxLNvlRkvQ6uHm+xE6HL1wiemd0EN20+ubMtvlwmTv3hPiXffifn/iPqe+sS4+pH0R/iUWcsa6Dv9dnz7TmMxfYH1JYUmVp94p6FlWLa/pAzbQtfonNlsv3e6ubnRMl6nW/uMz1PZYeGtSXxrXXUx8V6Tmjeda/JT+ygv6yecew3z91u45v3qhXu/Pva6Us9dUf3kSFUl+4jZhNmPEJ9JZ7/ffNGKCQ9CWu4i4n2C1I1Np3Mf0rNrmvLb245wr6bq16QHE/HWSkY98y09ufprvOvUaL/t402sPNEDW5S5F35AvfKTr6rew0Xad7JH6OVFxiAsUSaZR5ssr338nYXSukLomiXapfnzWLWL+LuJ91vpdUlh1bk4LFVp/hBN2aPqvXZMnlz6odFMnTrV+b8e7NKzuy680Ez9XLZsmaqvr1cbNmwY8hx6ymBHRwf7gIklkx4oDLGSgcZn8+bNw9ZVA22LV9tBUGbDQTYFXcNIdrAujpeNuC6GrkXcxsZH3sZC1+Ijkzl2XTXQtvhA3yncbewg0DUcZFMD/2gdTaKNncyMeJArl8up66+/Xq1bt06tWLHC+VtjY6MqKSlRNTX8X2d0vKtO8/L5qq6uzn/mzZs30iyBUcC2c2r3lvsdX66yyoE45UOHDg1bVw20LT5t97x4v6qYWp//G8psOHQ98PSvVNlMs0oOdA1JeX3h145nSGn1yOti6Fqc9XDltAWqrAq6hknXXVvvV1VT0HcKZZl96X5VMQV9p7Dp+samX6vyadA1bLoefOLXqmz2QpWcOjAbDW3s5GTEizNqb66tW7eqp5566pgy8MUvflF97nOfy+/rkdIRP0hyyUo6NZlMe3UtsSynA/osTc+m3MvQJxK+GFmygKWlqsQUvJjZr9ohwnfoNcoSnlN0ZbhEjkwV7LzYLB+rKb8jWVD4wa4Xf6V6OprUynM+qY6VYWsr9aNLBIspvBE6nVIsGR3JxL2nc9Ilo0m4k5OUNPfowEXTWVr/DDIdWzwuFXtF3nqIfsl4wfrR6b7NK3nRnPFS2nsZbL9wEsKuF+5TPR2N6oR1V6nND/+TKsoySyMbSkS5IGFlVkaES5B1Zq0+MR3epzzTpZVlCE1m3YlsvyJpwhBnl/J/0Xmj0xoyPNHZLzd5yyaF5uTxsEQUXUkHmbbt888RB568V/W1NqpFF1+ttt/zXTVhutLp+DIshC3j7r3scESGP7Hw4v6CtYvOmpnf7p4lCkySl3vVZ9JLuuRUaWvIcKeB30Sm6pfyazSea8Sc96B32I8rXJccunvTfaqnvVGdeMG16lgY1fIqYCHlPu2vfB5YeyxD8um+CMGwaAiGSIvM5HV2C5mCb4kQGXU46lnusiQkRLY7NCzGVQ/T3+tTR+964V6nHl5x7qdU0erKbAS8y6slwxN8rQNMG2rTMGRNjoRg1M5gSV1zhAZUS1GAWEiTCMvyC2+iYZe1z4g+X4H/FLxzy69UT2eTOmndBPSdRogsl3y5du+QTvczQUKqSLivxio34WiWCGtLtPJ6u6O1PL+9+60/YWkLD5ows/KGqKdGNDxxIG/e9a1f+NNQZfaEtVepzX+cuL6T7dfG0v6yKJcWaasscX/YntBH0ZBE+b5DQsnptrMv6vRMOXnfqeC2EN2pUs/w52xZfMh6WYbD0f6FpnI76Q/awrKEHPrGc/ep3vZGdfy5f6O2/O4mNdq62uR6NIfyGaUvFf2VMi3mWb9SzWOif0LLXayXa5euN23lzIv2s7QTyg7mtx97YwlLq27n+vRPISHEor+eI5pkRMh5rMceMlRTfs/1flBg6FrDY/eqvpaDavF7Pq2Kth6mv2U4IXl0vEK0o3ZP75D9Y9lHji5dpCiZcl5GYhGzX7KflE/93U4+tuCJz7PaN4XX34dON9t1T6qJH+S67rrr1AMPPKCeeOIJNXfu3Pzfa2trVSqVUm1tbWy0tKmpyUkbikQi4XzAxLPzpV+p1qZX1MqzPqkSpTX56dnad224umqgbfGgB7iONL7ivFjFSoznCcpssNn3p3tV+95taukl16pY0rwgQNdgs/vP96m2hm1q+Zs/pRJlx1YXQ9fiYddmUw9D1/CgB7h030kPcKHvFNIyew76TmFiz6Z7nTb2hAs+pWIJ9J3CwoHH/ld17t6mFl1+rYpX1qhsP9rYycywwhVt23YGuO677z71yCOPqIULF7L0VatWqXg8rtavX5//244dO9TevXvVmWeeOXq5BqOK1lUPcLU0blUr1l2jkuUDHmuDnHLKKdA1wNrqAa7Whq3qxLM/7tIWZTa4ujoDXHu2qCXv+KRKVHG3ZOgaXF31AFfrga3qhPM/oZIVXFfUxQGuhzd718PQNcB9py0DfaeVZ16jkmXQdbKUWbSxwdVVD3C17t+qTnizu42FrsHVVQ9wdezcoha++5OqpBp9JzDMmVw6RFGvTPDrX/9aVVZW5uNYdaxqaWmp8/+rrrrKmd6nzeirqqrUpz/9aecBKtpVn4DTSTu8/wW1fPVHVTSWVKk+szqgBroGF91Ja973glp25hUqGk+oVF+HyqTNtFVoG0z2/+ledWTnn9XCt3zM0TXd06GyKegadPb8+T7VvPcFdfy6K1Q0llCp3g429Ru6Brge3v+CWrbG1MMDQNfA950OvKCWn46+U6jLrK6L0XcKBXqAq+WNF9RxZ1+pIn9pY7MZ6Bp0Gh77X9W2489q/js+piK6T9yNNhYMc5Dr1ltvdf5/3nnnsb/rZTivuOIKZ/tf/uVfVCQSUZdffrmzMsFFF12kbrnlltHLseW9z2J5hYePXHI4QpaQttLCe4DGs8slltlJxPWIR0jDBfxffXJJsUx5hYmljZGlkaUP13CWzaVx6bXEg2sgr94/o3HPwMoSW56+jf190Unvym+Pua5DeUtFSdy2ED5DTE/kQ2yRuH47Fff0Z4pO4QaE9jSz37GUxzrHp5oVWHIH+BLnyRbh40OerZyPJ5drqWuyzG5qivAToZ5GBXpwaZp2DWj78hM/8jxmPLT1xOXjY7azce9ES1hHMN8Niz/79PmIiDJLFbCFx0/7Il4uL5u7Nb+9vul4llbzmslQukJ4RxAfLukzxnT18wsRt6L5laed/7/+wC1jr6v2AJC+hUMdRqpRYYPBnnVLr9tNoPWvXILYEl6KLI36ZYklslPLTRh9+3E8M9EkL9uRQ8ZzIN4pvUWiPnVxxNNjo4p4gthR8bDS8wtdm3YOlNdtj/HyumDVu4ujvPoh/cXosy3bKnJfpUdbhPplirrOIktW2zmua3oOb3PVQuNJUVrCdS1pNHVEvJN7MGVKvX0T2TMuPXGYPxlPa9r9l3r4Sa7rwlPGt431wuWnRrTL5aR/FfG7yXr7mVpCH5omS7VNlrTPzKxiaT3zsp79grJGfv1In9EyV+rTdxJ9Rd5+iMxZ3ro2vvGXvtMG0XdaWRy6DpVni/j/uNoc8uO5P5dOIg8/8Q+Vz0FE+NhZxKOL9s2GYsH/mOuf+fAnWNrcPtIfFNdgmsl6KDoy7xuvMksZE21pHm0/3z/RRyT9JUuUWdrmUs9aVxnO8LLG/IqpH6Lwy5Rea4fXcV+9zBwziNTRxT1+pu8w14j28zojU+bnj+jt81m5L+tZTx96fUDXVx4ZeJ8dijEvs7JKJaJbFtc1XUa91rgG0XJzomgN769GMiYEs3MuL8wf/dvf5bfPLnuVpX3ilQ/mt+N/4nVxNMV17q+JFNQeyrqE7rv8Oll9qwqmdctAn3j3vVynuvMuL5J3HdnmUF8yMT5B+h2yD2JRvzvpZxrx7jvZpI986LyBhVHyx9bwFUanVXfnt/tP5n5uh3ebfta0Fzs821jqmSfLa7SfP+OljaTPRfzABjKnxm+QS08HPBrJZFLdfPPNzgcEg3WXDW1WrX0ldr10n7MNXYPJ2stvGlLXjfd/Jb8PbYPHqdf8s6sjkE31qS13fCn/d+gaPNa8l5RX0knRZVb/C7QGugaPte+muiqm6+7NaGODylmXfNdT111boGtoyiztO/0Gfacgc8b7/8lVZrWum3755fyfoWvwWPmZ7w35d+3JpWd5aaDr5GNYnlwAAAAAAAAAAAAAABQjI1pdsZhgUx3FRDMWQpLwXg6ZTck9ypQ/Fi6R5tNDrVIzDbdrgViSczqfDpg7YI6dslVM+SPTu3MkrGNg33vZXPZ7XUvTey/LOqH4Tc2meZZTs8nUZbEYOVs63I6V88ulRRgnoWNpZX77vNNeZmnVcbM06x83r2ZpFQf4dM5cWUlB+tHpm07eSL5n/0mEaPhpVkx6HiNsieyo9xTnTJL/6CgJc8vFebnMkbIfreQrpUR7KvLbFgl10VTtTrH9D1Y/n99ekmhiaTfNeH9+u7RF5tvyzDcLF/EJ3SxGjd1hPd6hl2x6Og170V8j4RNWWsah2p51sSLhpTER5nhkvtG5esERllaW4Lq2vGbCjyNpXm/TkKdswqcuFuES1XvI8uquaeMqPPgsac9CFGTIm1gOnn2PhIVGSFvoXC7NQ8UpHQt53f6JlX/Ibz/fPp+lvdZbVZCurnY0UmB4saQoy693iAz9nTK0Tylyv0QSDbNgYRV6P2HqZUuERVGLiK65XMeTVu5m+4d7TZuezc5kaVnyvMgQWdo/kuWVPbtFqNWYYfmF9nmHEUlYuRBF24qTZ+so4YqUkk4RtkKQfSeel8LDTYsd2cbSYupb37rOQ8OBRN+JlJOIqJdpfWv1izB/Eq5oJ3mf68hyrvOaJaYMv9hYJ/Jt3oUy5SLEmIa8y5A3Wp596mkXQ4XDjcNj4Rd6R11x5DOaI+XHnUb7lrbns9I7g3/vEzW78ttxi/eJ+x8wdWpZGy+DKRIe6Qrfl7+PhiuKvnxY61tXeaVprj4xaX95N4eF+llZUSbp+22E94csEvYfqRDvvqQ97p3OM7qorpntzyo13pIb3+B9p/rd5n03I96naL7ddkokL6Jqr9mZ8y7Lx/h8hKm7DQAAAAAAAAAAAAAmKRjkAgAAAAAAAAAAAACBB4NcAAAAAAAAAAAAACDwBM6TS8ZrsthOGfPKvDW8PSdcS5iTuHSL+CtJzxi5fHb/TBMDu+Sk/Szt3Bmvsf3/fOkCk8+yeMFeLyxGXcad+8R1c+8UVZzI2Fvb5/f4LL+apcsOC40sZg7ET5mqNNe4o/5JlvbDIyYuefOLad98Z8nzI/1M/Pxe2LHSKyNgXhKj4U3gduTw8eEQ5YSSTZCla0u5OUCkisSwZ/gDkRXLpH9p/yX57f9a+ChL++oCs53o8PZgcntHkB2fpesDgW8dQ5c35ynUW5D62TnHJnz8unLmRJly7g1w5ASz/e0TzHLZmv89vIrt9zTV5rdjvVnvsiyXMPfxC2Hty3B0LfZyLusl5e2bSA+W9476Xrog/l0R4ZdlUU8f+RyJe/e5qcZ35D1HFrG0qjdSnnmjdYn0BfT1JSpG30sfz0t/Dxvi+yHaV9qXci0bT461iLeas58tKcybSWTlV0t/z/bf8erF+e2+/SnPJcylZ2nBZTLEba/Li0360RZYhw+nD2nlCr9/vK0M130fFbwtiJUi9Za85X7vO8z/J8ffRSJp29Ov0sqa/SzxrtS86cxX2T7tLy3efSVLq+mxvfvLJYX5bvn69rjK8xBpE/yoDZmnv5AlvluW9Camuz43IdnCv3fyjz6d386U8rTqXrOfLvV+x3Quyd45h+GF51vfqlAifxf3DBfHRqLe/pG0jU3xMmmXGH87S77PJMw5Z2zm77DVbzOe05r/WPBYfnvhlqtZWrrC8rwG68vL9pf6t8l+Q4HP0Ugo1uEOAAAAAAAAAAAAAAAKBoNcAAAAAAAAAAAAACDwBC5cUcKnsHpPiZTTdy2fpcBzZL3TSFbEK9DZgWKqt0Wm8/7+hAdY2s50F9t/5Jl1nmFRLFxCTuuju66lNn3C3YI4nOkXGkKnzovlV+mNYeEtA180aSKpcp9ZIvn8j11d8P2j072PBtd2hEtdh3Q6rwvX9F7vRFdYkcf35NLjVs77WZHTbXfdenx++wxrGUubQrazJYWHK470nxkmako3va5PdDELnRg4mIQ4iajDLJl+nYvyG0I1sWRMoPIJbZlnpl9fXtHBkm4/YKZ0a+Ld5hoZEc7qFzpOs+NbT0+a6fjeIeXyJrhDkSmkjpQHyvAqQtkhPgX//Cv/hpxS6JPw0S42wlC1YtfVt8CKpIh3SFuOhh7LZztmyo8Vk1YBhVV2pULHCz50lcgb2SmZ4JDEYtf8KMjfzbSW9yg6svsVyL5n0YaHF1icXe2Rd4ixRQqKqw9U4lN+fSrxLY2z2f43Kpbnt0u38PY31pvxCfsnl3P1Kbyvz84RcBsIVu58bDmGQ6LVbCdlfyzpk5dhhB36PqzsnCq0+PaX6bMt+ycsbNy7TyotVvzCvS1at4u2f9Prxo5H838rTslvV23nL9h2xJTXXDIyDGsWn/EJv+foGEHzAwAAAAAAAAAAAAACDwa5AAAAAAAAAAAAAEDgwSAXAAAAAAAAAAAAAAg8gffkosgYULpU/XCWyKZx6dRj4qhLXxPe/BHuI+GKkSex7i6/kEJ9tyS+yz+HDJ/bQJcd9gv4di2fTSlMZo8T+6TZk2ep8pFQaGx2gfYu7hNJXdn+Ue7/aMgTRH8m/VwOPpukzMi8013b79mWHnoWDeRXBXvo8fPz/br/Nk3bwr6/4Y1eMsOPbTGmFJmK6Ig8fvyXMA+CyCOEie6jeUSKV5g3oYthFBje/vsdKC7h6itMPu18fZuIljLN15dvJPkaZvqYt6EhfzTQBylyCmxz3MW50IP9+ss++RJp1f9TyfZ/r87Jb1cJM690mU/lXGgb6/M94MbPu2m08OsPTkZ8+4/yYNJ2ZulYgfyiTCn0vdXmu7V/4P3eJ/6wJr9dNpzyWigjfB6YV1mB3wn7UAgAAAAAAAAAAAAAmARgkAsAAAAAAAAAAAAABJ5QhStK/EMUvOfL2b7zCNXIwt/UyMIOfc+BIcqjM5GhDWDMGPH0Z0h+bNAyI+o7/5A9s5mL8wPlvhfDC1cwB9c9FPFt9jJlwzgv8MZnCXEZEkHbLtlusnrZGp2l18FR8Fv+XSaNkj6+1wcAjA+Fvm+gjIJhgOeliO+5pSYNGCYBAAAAAAAAAAAAAIEHg1wAAAAAAAAAAAAAIPAUXbii/ZfQhUy6b0yvQ1de8s/Q5A1XHNRgUJNjZfA82THWFviTHStdU9B1Ihm8/+NWXkd4nYJXgBmlcMUwMJpldtzq4QKzOlbhipNV11HpO41OFVI4IdMZfadwMmZ9J+g6oaBPHF5Gs1+M8hq8MmvZo1WqR4n9+/erefPmTXQ2AGHfvn1q7ty5x3weaFtcQNdwAl3Dy2hoC12LD+gaTlAXhxPoGk6ga3hBGzs5dS26Qa5cLqcaGhqc0bn6+nrnB1RVVU10toqGjo4Op5CNx33RGnR2dqq6ujoViURGRdsdO3ao5cuXQ9cJ1HYsdEWZ9Qa6hpOg1sXQ1R/oGl6CXBej7xROXVFmvYGu4QRtbDjpKEJdiy5cUWdWj8rpm6XRNwoPkZvxui/V1dWjqu2cOXOcbejqzXjcm9HWFWX26EDXcBK0uhi6FgZ0DS9BrIvRdwqnriizRwe6hhO0seGkqoh0hfE8AAAAAAAAAAAAAAg8GOQCAAAAAAAAAAAAAIGnaAe5EomE+trXvub8H4TnvgQ9/2NJ0O9N0PM/VgT9vgQ9/2NF0O9L0PM/VgT9vgQ9/2NJkO9NkPM+1gT93gQ9/2NF0O9L0PM/VgT9vgQ9/5PpvhSd8TwAAAAAAAAAAAAAAKGZyQUAAAAAAAAAAAAAQKFgkAsAAAAAAAAAAAAABB4McgEAAAAAAAAAAACAwINBLgAAAAAAAAAAAAAQeDDIBQAAAAAAAAAAAAACT9EOct18881qwYIFKplMqjPOOENt3LhRTSZuvPFGdfrpp6vKyko1c+ZMddlll6kdO3awY/r6+tS1116rpk2bpioqKtTll1+umpqaVDEDXaFrGIGu4QS6hhdoG06gaziBruEEuoaTsOqqgbY3Bkdbuwi5++677ZKSEvunP/2p/fLLL9tXX321XVNTYzc1NdmThYsuusi+44477K1bt9qbN2+23/a2t9n19fV2V1dX/phPfOIT9rx58+z169fbzz//vL1mzRp77dq1drECXaFrWIGu4QS6hhdoG06gaziBruEEuoaTMOqqgbZ2oLQtykGu1atX29dee21+P5vN2nV1dfaNN95oT1YOHTpk6zHJxx9/3Nlva2uz4/G4fc899+SPeeWVV5xjNmzYYBcj0NUNdA0n0DWcQNfwAm3DCXQNJ9A1nEDXcBIGXTXQNljaFl24YiqVUps2bVIXXnhh/m+RSMTZ37Bhg5qstLe3O/+fOnWq8399j9LpNLtPy5YtU/X19UV5n6Dr0EDXcAJdwwl0DS/QNpxA13ACXcMJdA0nQddVA22Dp23RDXI1NzerbDarZs2axf6u9xsbG9VkJJfLqeuvv16tW7dOrVixwvmbvhclJSWqpqYmEPcJurqBruEEuoYT6BpeoG04ga7hBLqGE+gaTsKgqwbaBk/b2LheDYwIbd62detW9dRTT010VsAoAl3DCXQNJ9A1vEDbcAJdwwl0DSfQNZxA1/BybZFrW3QzuaZPn66i0ajLhV/v19bWqsnGddddpx544AH16KOPqrlz5+b/ru+FnjrZ1tYWiPsEXTnQNZxA13ACXcMLtA0n0DWcQNdwAl3DSVh01UDb4GlbdINceorbqlWr1Pr169l0OL1/5plnqsmCXhRAP0D33XefeuSRR9TChQtZur5H8Xic3Se9hOfevXuL8j5B1wGgaziBruEEuoYXaBtOoGs4ga7hBLqGk7DpqoG2AdTWLtIlOhOJhH3nnXfa27Zts6+55hpnic7GxkZ7svDJT37Srq6uth977DH74MGD+U9PTw9bolMv2/nII484S3SeeeaZzqdYga7QNaxA13ACXcMLtA0n0DWcQNdwAl3DSRh11UBbO1DaFuUgl+aHP/yhc4NKSkqcJTufeeYZezKhxx+H+txxxx35Y3p7e+1PfepT9pQpU+yysjL7Xe96l/OgFTPQFbqGEegaTqBreIG24QS6hhPoGk6gazgJq64aaKsCo631lwwDAAAAAAAAAAAAABBYis6TCwAAAAAAAAAAAACA4YJBLgAAAAAAAAAAAAAQeDDIBQAAAAAAAAAAAAACDwa5AAAAAAAAAAAAAEDgwSAXAAAAAAAAAAAAAAg8GOQCAAAAAAAAAAAAAIEHg1wAAAAAAAAAAAAAIPBgkAsAAAAAAAAAAAAABB4McgEAAAAAAAAAAACAwINBLgAAAAAAAAAAAAAQeDDIBQAAAAAAAAAAAABU0Pn/ARvQaSJwDIMFAAAAAElFTkSuQmCC", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "transform = transforms.Compose(\n", " [\n", " transforms.RandomAffine(degrees=(-10, 10), translate=(0.05, 0.05), scale=(0.9, 1.1)),\n", " transforms.RandomPerspective(distortion_scale=0.2, p=1.0),\n", " transforms.GaussianBlur(kernel_size=(3, 3), sigma=(0.01, 0.1)),\n", " ]\n", ")\n", "\n", "\n", "# Visualize 30 augmented images.\n", "for i in range(3):\n", " X_aug = transform(torch.from_numpy(X_train[:10]))\n", " fig, axes = plt.subplots(figsize=(15, 5), nrows=1, ncols=10)\n", " for j in range(10):\n", " if j==0:\n", " axes[j].set_title(f'Original', fontsize=FONTSIZE)\n", " axes[j].imshow(X_train[i].reshape(28,28))\n", " else:\n", " axes[j].set_title(f'Augmented', fontsize=FONTSIZE)\n", " axes[j].imshow(transform(torch.from_numpy(X_train[i])).reshape(28,28))\n", " plt.show()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Convolutional Neural Network Ensemble\n", "In the next step, we define a convolutional neural network (CNN) ensemble consisting of ten base CNNs with different initializations." ] }, { "cell_type": "code", "execution_count": 5, "metadata": { "ExecuteTime": { "end_time": "2025-04-07T10:05:01.939355Z", "start_time": "2025-04-07T10:05:01.927485Z" } }, "outputs": [], "source": [ "# Define base module.\n", "class ClassifierModule(nn.Module):\n", " def __init__(self, n_classes, dropout=0.5, transform=None):\n", " super(ClassifierModule, self).__init__()\n", " self.conv1 = nn.Conv2d(1, 32, kernel_size=3)\n", " self.conv2 = nn.Conv2d(32, 64, kernel_size=3)\n", " self.conv2_drop = nn.Dropout2d(p=dropout)\n", " self.fc1 = nn.Linear(1600, 100) # 1600 = number channels * width * height\n", " self.fc2 = nn.Linear(100, n_classes)\n", " self.fc1_drop = nn.Dropout(p=dropout)\n", " self.transform = transform or transforms.Compose([])\n", "\n", " def forward(self, x):\n", " if self.training:\n", " # Usually, transforms are part of the datasets.\n", " with torch.no_grad():\n", " x = self.transform(x)\n", " x = torch.relu(F.max_pool2d(self.conv1(x), 2))\n", " x = torch.relu(F.max_pool2d(self.conv2_drop(self.conv2(x)), 2))\n", "\n", " # flatten over channel, height and width = 1600\n", " x = x.view(x.size(0), -1)\n", "\n", " x = torch.relu(self.fc1_drop(self.fc1(x)))\n", " x = torch.softmax(self.fc2(x), dim=-1)\n", " return x\n", "\n", "def create_ensemble():\n", " # Set seed for consistent initialization.\n", " torch.manual_seed(RANDOM_STATE)\n", " torch.cuda.manual_seed(RANDOM_STATE)\n", "\n", "\n", "\n", " # Create list of ten base CNNs.\n", " estimators = []\n", " for i in range(5):\n", " net = NeuralNetClassifier(\n", " ClassifierModule,\n", " module__n_classes=10,\n", " max_epochs=100,\n", " batch_size=128,\n", " lr=0.01,\n", " verbose=False,\n", " optimizer=torch.optim.SGD,\n", " optimizer__momentum=0.9,\n", " optimizer__weight_decay=3e-3,\n", " train_split=False,\n", " device=device,\n", " torch_load_kwargs={'weights_only': True},\n", " )\n", " net.initialize()\n", " clf = SklearnClassifier(\n", " estimator=net,\n", " missing_label=MISSING_LABEL,\n", " random_state=i,\n", " classes=classes,\n", " )\n", " estimators.append((f'clf {i}', clf))\n", "\n", " # Creat voting ensemble out of given ensemble list.\n", " return SklearnClassifier(\n", " estimator=VotingClassifier(estimators=estimators, voting='soft'),\n", " missing_label=MISSING_LABEL,\n", " random_state=RANDOM_STATE,\n", " classes=classes,\n", " )" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Active Classification\n", "For our ensemble, we evaluate four different query strategies, i.e., random sampling, uncertainty sampling, query-by-committee, and batch BALD, regarding their sample selection. For this purpose, we start with ten labels and make 30 iterations of an active learning cycle with a batch size of 32.\n", "\n", "> **Note:** The execution time strongly depends on whether a GPU or CPU will be used. For example, the below reported times will be much faster when using an advanced GPU." ] }, { "cell_type": "code", "execution_count": 6, "metadata": { "ExecuteTime": { "end_time": "2025-04-07T14:23:51.953307Z", "start_time": "2025-04-07T10:05:02.072808Z" } }, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Execute active learning using random sampling.\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 30/30 [02:09<00:00, 4.32s/it]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Execute active learning using uncertainty sampling.\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 30/30 [02:10<00:00, 4.34s/it]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Execute active learning using query-by-committee.\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 30/30 [02:09<00:00, 4.33s/it]\n" ] }, { "name": "stdout", "output_type": "stream", "text": [ "Execute active learning using batch-bald.\n" ] }, { "name": "stderr", "output_type": "stream", "text": [ "100%|██████████| 30/30 [02:12<00:00, 4.42s/it]\n" ] } ], "source": [ "# Define setup.\n", "n_cycles = 30\n", "batch_size = 32\n", "n_sub_set = 1000\n", "qs_dict = {\n", " 'random sampling': RandomSampling(random_state=RANDOM_STATE, missing_label=MISSING_LABEL),\n", " 'uncertainty sampling': UncertaintySampling(random_state=RANDOM_STATE, missing_label=MISSING_LABEL),\n", " 'query-by-committee': QueryByCommittee(random_state=RANDOM_STATE, missing_label=MISSING_LABEL),\n", " 'batch-bald': BatchBALD(random_state=RANDOM_STATE, missing_label=MISSING_LABEL),\n", "}\n", "acc_dict = {key: np.zeros(n_cycles + 1) for key in qs_dict}\n", "\n", "# Perform active learning with each query strategy.\n", "for qs_name, qs in qs_dict.items():\n", " print(f'Execute active learning using {qs_name}.')\n", "\n", " # Copy initial ensemble model.\n", " ensemble = create_ensemble()\n", "\n", " # Create array of missing labels as initial labels.\n", " y = np.full_like(y_train, fill_value=MISSING_LABEL, dtype=np.int64)\n", " X = X_train\n", "\n", " # Label one sample from each class for warm start.\n", " for i in range(10):\n", " y[np.argwhere(y_train==i)[0]] = i\n", "\n", " # Execute active learning cycle.\n", " for c in tqdm(range(n_cycles)):\n", " # Fit and evaluate ensemble.\n", " acc = ensemble.fit(X, y).score(X_test, y_test)\n", " acc_dict[qs_name][c] = acc\n", "\n", " ulbd_indices = np.random.choice(unlabeled_indices(y, MISSING_LABEL), n_sub_set)\n", "\n", " # Select and update training data.\n", " query_idx = call_func(\n", " qs.query,\n", " X=X[ulbd_indices],\n", " y=y[ulbd_indices],\n", " clf=ensemble,\n", " fit_clf=False,\n", " ensemble=ensemble,\n", " fit_ensemble=False,\n", " batch_size=batch_size,\n", " )\n", " query_idx = ulbd_indices[query_idx]\n", " y[query_idx] = y_train[query_idx]\n", "\n", " # Fit and evaluate ensemble.\n", " ensemble.fit(X, y)\n", " acc_dict[qs_name][n_cycles] = ensemble.score(X_test, y_test)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "## Visualize Results\n", "In the following, we plot the obtained learning curves including the area under learning curve (AULC) scores per query strategy." ] }, { "cell_type": "code", "execution_count": 7, "metadata": { "ExecuteTime": { "end_time": "2025-04-07T14:23:52.683443Z", "start_time": "2025-04-07T14:23:52.250599Z" } }, "outputs": [ { "data": { "image/png": "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", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "cycles = np.arange(n_cycles + 1, dtype=int)\n", "plt.figure(figsize=(16, 9))\n", "for qs_name, acc in acc_dict.items():\n", " plt.plot(cycles, acc, label=f'{qs_name}: AULC={round(acc.mean(), 2)}')\n", "plt.xticks(cycles, fontsize=FONTSIZE)\n", "plt.yticks(fontsize=FONTSIZE)\n", "plt.xlabel('# cycle', fontsize=FONTSIZE)\n", "plt.ylabel('test accuracy', fontsize=FONTSIZE)\n", "plt.legend(loc='lower right', fontsize='x-large')\n", "plt.show()" ] } ], "metadata": { "kernelspec": { "display_name": "skaml_skorch", "language": "python", "name": "python3" }, "language_info": { "codemirror_mode": { "name": "ipython", "version": 3 }, "file_extension": ".py", "mimetype": "text/x-python", "name": "python", "nbconvert_exporter": "python", "pygments_lexer": "ipython3", "version": "3.11.11" } }, "nbformat": 4, "nbformat_minor": 4 }