{"cells":[{"metadata":{},"cell_type":"markdown","source":"# About this notebook  \n- PyTorch resnext50_32x4d starter code  \n- StratifiedKFold 5 folds  \n\nIf this notebook is helpful, feel free to upvote :)"},{"metadata":{},"cell_type":"markdown","source":"# 노트북 리뷰\n*  전체 트레이닝 과정이 잘 정리된 노트북이라고 생각합니다. \n*  잘 모르는 개념, 라이브러리, 왜 쓰는지 모르는 코드들을 최대한 이해해보고자 합니다.\n*  항상 해당 코드블럭 위에 설명이 달리는 방식입니다. 설명을 먼저 읽고 -> 코드를 보는 방식\n*  제 배경지식 이외의 정보를 정리하는 것이라 해당 개념에 대한 자세한 소개는 부족할 수 있습니다. <br>아예 뭔지 모르고 쓰는 것 보단 대충 뭔지알고 쓰자라는 취지의 정리입니다."},{"metadata":{},"cell_type":"markdown","source":"# Data Loading"},{"metadata":{"trusted":true},"cell_type":"code","source":"import os\n\nimport pandas as pd\n\nfrom matplotlib import pyplot as plt\nimport seaborn as sns","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"os.listdir('../input/cassava-leaf-disease-classification')","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train = pd.read_csv('../input/cassava-leaf-disease-classification/train.csv')\ntest = pd.read_csv('../input/cassava-leaf-disease-classification/sample_submission.csv')\nlabel_map = pd.read_json('../input/cassava-leaf-disease-classification/label_num_to_disease_map.json', \n                         orient='index')\ndisplay(train.head())\ndisplay(test.head())\ndisplay(label_map)","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"sns.distplot(train['label'], kde=False)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Directory settings"},{"metadata":{"trusted":true},"cell_type":"code","source":"# ====================================================\n# Directory settings\n# ====================================================\nimport os\n\nOUTPUT_DIR = './'\nif not os.path.exists(OUTPUT_DIR):\n    os.makedirs(OUTPUT_DIR)\n\nTRAIN_PATH = '../input/cassava-leaf-disease-classification/train_images'\nTEST_PATH = '../input/cassava-leaf-disease-classification/test_images'","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# CFG"},{"metadata":{},"cell_type":"markdown","source":"* apex : NVIDIA에서 만든 'A Pytorch EXtension'(APEX) 라이브러리. 학습 최적화를 위한 도구이다.<br> mixed precision training과 distributed training 등의 기능을 지원한다.\n* CosineAnnealingWarmRestarts : 학습률의 최솟값, 최대값을 정하여 그 범위 안에서 Cosine 함수를 통해 학습하는 방법(CosineAnnealing). <br>학습률이 작을 때 local optima에 빠질 수 있으므로 다시 학습률을 크게 증가시키는 방법이 WarmRestarts. 두개를 합치면 아래 그림과 같다. \n\n![image.png](attachment:image.png)\n","attachments":{"image.png":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAf0AAAFXCAYAAACoS5cAAAAgAElEQVR4Aeyd9XccZ5b39w/Yc+bsD7tndmD3HcjM7MxkkswkcRgmMEHHQYeTCTODOWZmZmYGMbNkyZJsWWDLsiyywEIzz33P58pPu6TYFnVLVd11z2lXW13w1K2nnu/l+2/ikssBlwMuB1wOuBxwORAQHPi3gLhL9yZdDrgccDngcsDlgMsBcUHfnQQuB1wOuBxwOeByIEA44IJ+gDxo9zZdDrgccDngcsDlgAv67hxwOWBzDtTV1Ul8fLxUVlbK2bNnbT7anh3emTNnpKKiQiIjI6W+vl4HU1NTI0lJScL23LlzPTtA9+ouB3qYAy7o9/ADcC/vXA4AIMeOHZPq6mo5fPiwfvje2Nio4Pyvf/3LKze3c+dOefLJJyU0NNQDZF45cRsnOX/+vDQ1Nel9HT9+XC5cuNDGET3/M8C+Zs0a+e1vfyvJyck6oOjoaHn++eclNjZWjh492i2DRDhjHiCAXOtz4sSJbhmPexGXA4YDLugbTrhblwMd5MD+/ftl2rRp8sADD8jvfvc7+b//+z+5//775fvvv5ddu3YJWqc3CABBa+V8Fy9e9MYp23WOAwcOyJAhQ+RXv/qVzJs3T6qqqtp1XE/udCXQh28NDQ1eFcTausfdu3dL//79lXfw73//93/l5z//uX74v/ksX768rVO5v7sc8CoHXND3KjvdkwUKB9C+BwwYIM8++6xMmjRJNm3apJr42rVrZfr06TJjxgzVkp3Mj8TEROnVq5f8x3/8h7z11ltqIrf7/VwJ9HtizFhI8vLyJDw8XD+zZs2Shx9+WJ555hnP3/ittLS0J4bnXjOAOeCCfgA/fPfWO8cBNMeRI0eqyXjYsGFSWFgoJ0+eVPM3/vecnBz1wZ86dUo1c35ft26djB07Vj+rV6+W/Px83R8XAPtFRUWpoMB5R48eLXPnzpXi4mIdIFsEib179wrm4EOHDnkEjA0bNsiUKVNk3Lhxsn79etXGjRmecaKtr1ixQn8fP368jqOgoKDNG8c0vXTpUvnFL34hL7zwggLWkiVLWlgaioqKZMeOHXr+bdu26b1xjaCgIHUJmIuEhYXJxo0bZevWrTJ79mzl3Zw5cyQzM7OFNeTIkSPqi+deR40apfecnp7uEZ64HzRoxgUv2QcwTUlJ8ZjtrwT63C/83Ldvn/IangQHBwsCGvybOHGiMO4tW7aoq8a4ZXAFpKWlqTVn+PDhsnjxYnUdsB/XNHw293mtbVZWlvzzn/+UL7/88ke78fyZIwsXLtT7YjybN2+WkpISYSzEcvB/3DtmbJykvLxcIiIi9B6sf//RBdw/uBywcMAFfQsz3K8uB9riAOb1srIy6dOnj7z55psKxNc6hoV7/vz58tprr8krr7yiH74DVgcPHhT85gACpuC+ffvKyy+/LK+++qq8++67+ncWc3zRf/7znxWwAUaC+j755BP5+9//Ll999ZVq4fj8OS8ChfETI1gAsoANv3Hed955R4WL2traFgDe+h4A14EDB8pf//pXBesHH3xQBg0aJFzfEICDBeChhx7SfbnG448/Lu+9956CvNnvm2++UX69/fbb8tFHH8mLL74ojzzyiAIcLhLu8fTp0yq0fPfdd/L6668rH+DF4MGD1cIAn4grAKS//vprvR/4xYdzGgHiSqCPQHLrrbeqgIKZHw37gw8+0HFzPZ7jE088ofeCcMK1GBOWDu6ZZ22ug2WHD26djgQFXg30uQ5CCdYihCt489JLL+n4mCPcM8IV9/jpp5+2iOlAmIIX3377bQthwPDd3bocuBIHXNC/Elfcv7kcuAoHWOgB4fvuu0/69evXpra3bNkyBSgWZ6wAfAAyABjtEwsBWvrHH3+soIc2i+aXnZ3tWeCvBPoA4//8z/8oCAJ0CAKcAwDm/8QBzJw5U8EMoQPNnSDDMWPGyHPPPSchISEKtFe6TYAIqwGgi2ZKPAHjRXhgLIYA/aefflruuOMOFQy4F6LmAS2A1GjCgD6gizCAOZv9EHK4B7R2eAqw9e7dW6+HRYN90Ka5H6wfCCnwhfsEJDGfw0s0YKwRWA4Ipmwv6APixGCg8XN/CAKMD/M71+H6AD5CDIIAQklqaqoC88033+w10Oda8IBzcg9o9QhCxFLAsz179uh44AH8T0hIUIBH+Jw8ebIKhwg1PDOXXA60hwMu6LeHS+4+LgcucQAw3b59u5q7Mfm2RV988YWCCWDFQs0HkzFaG5ov4II5FwAEpAmWYwEHCNkXuhLoo+k/9dRTGhnOfhyHZvib3/xGwQ8QNVozwMI52Y9IdgIN+QCcVyLuccSIEWpJAFAAPDRRLBUTJkzwHALoM26sBwgrXINsBgQcNGdAmWsC+uy3atUqj0uDGAiEFFwZmNERgABd/m40bXiAJYOxknLH+Rmb+Z1zI0QgxPAsEJTaC/po+twPPOA8mMpxGWDZ4Jnk5ubqeD788EMFXfNMuH80f29p+pj1GTuWFMaCoMQHoQzBBL7zbHGNwFf4xXixWHAPWFq4Z5dcDrSXAy7ot5dT7n4uB0QUjPG/E6VPIB8L8LUIsGNxRnszhI8eMAM8ADA0V3zKb7zxhpp3Me2j/ZHqBV0J9AEAwBZghQABfMI/+9nP1P2AYMH5iBo3Jmm2WChuu+021Ro55koEeHJutHgAh3tEaOF8mPABdkAQ0OfeAHVD7IsW/49//EPHwf/5HSEFn7QhvgP6aLRo2mjVt9xyi1oNrOP94x//qDwhTgDBAjM51gq0XvZDE//lL38p77//vloG2gv6CGOfffaZx9oBH4mNQPtHCOF+EUK4FystWrRIrRjeAn2yPLh3zPrMBUOkGyJw8JwRbBDiuG8EPYQw46IgrsF6nDne3bocuBoHXNC/Gmfcv7scuAIH0MLQztBkAWf88tciNDGAEc3REAIAZnM0VAiNFqAl2A+TLcIEZmYCxgDEK4E+flz8vPwOYb4n0OynP/2pggRgAhASfY8lwXzQVMksQHPEAnAlQni455575MYbb1TTOtol48UEfdddd6mvG40b0Ae4WwMjsQBEqhNwCL8A/c8//1xN/+Z6gBbjZ1+ED7TdRx99VIUNM1a2CEO4GhCMMH0DxFhJhg4dqlYHhIabbrpJ7xUffHtBH6GLcQGgEHycOnWq/P73v1eNG9Dl+VoFGvYjdRELgbdAn9gJ3D08byt4Y8ZH8GIuEEPCsyK9D2GK+AWsODwTLDEuuRzoCAdc0O8It9x9A54DaLgswIAugVeYpdGGAUG0WgKvWKQJouNvgBKLN8CFpsrfiHhHEAB4OIbgOECb3wjCIx3w+uuvV3MzQkVnQJ8xkFmAZo6WyLkZO2PAjGzM2tYHyu+AIGCMNQCAwa/PBz/9vffeq8IAAM19egv0ORf8wd1BhgC8YCzwhu9o3vCcFLj/+q//0qh9BCfM8AhL1ElgzN4EfXiGdYLYBDRtxoI1AK0cvngL9HlOCGLw1mQXYEkhIJMYB2JCiGeA4uLidC7Bf+IocMGQieCSy4GOcMAF/Y5wy93X5cAlDuAbRxMkCn3lypUacAVAE3yGCRj/MICFRo3GDcgTlMUH07IRBABjorA5H6l+WBH4Duji4wcUOgP6aNhohoA1oMK5cSugWaJFYibHwmAlNE3AjuA5NGr2QSAxH/zMmJzxPxMUiLbuDU2fsSI44T4wFgGsBIwFVwpWC64H6P/6179W6wPuCwCfOAaK3gDO3gR9eIOVARcMz4HngrDG/wmw8xbow3PiGHC7EKyXkZGhzxsLA8IMcwpBDeI7LghcHrg0sOxYrQPWZ+l+dzlwNQ64oH81zrh/dzlwDQ4A1izQmJnRuqiwdt1118ndd9+toA4AoTWjxWKmJyjrD3/4g/qMMeuTI45GhwbL4o1WR1U/zoO5GsEAAQENszOgj6aMhkjQIdf+y1/+oqVp//a3v6kAwvi4BysxVszXjz32mKb6AShc33wwgZMJQJwA5m/uyxugzxi4BuCOZQQXAmV04QdaNdo/94IbABM8rgfAH8EIsMc0Dr+8CfrwD6AnkA6ecc+kRXIthBNSITsCuFdL2ePeEXhIRUSYYo7wrIhZIO7Beg3mE7EGP/nJT3S+IBy65HKgoxxwQb+jHHP3dzlwiQOAJNoz2n1MTIx++G5SygAyA74AOOZZPmio+J4NmBKwZzQ8FnLAC8Ax+faAHSCLtgtQ83/M27gQjLYOOHAezm/81PwNkzTaPWDBGDk3KXGcg+tbCY0bUzYAZYIIrb9zPtLiOBfHcw/cK2O1EiZnAByBhvtnrJiuTfwB+zIuxm81T3O/aPeGnwg7FOchsp775Pp8R8vnXuAJFgx4y/mNiwQQ5T75P4SlgmPYch6uzZj4GB7wd2oqkCXAdSD4iKWF6/BceEb43wnOpL6/2U93buMfBDziOhhna+LaCDXcK88PSwz7wmMrMVbmzn//93+r9QZeueRyoKMccEG/oxxz93c54HIgIDiAkIKQgzAC4ALYxA4Q44BgZASG7mIGwgp1CbCCEE+B0OmSy4GOcsAF/Y5yzN3f5YDLgYDgADEE+NuJ2aBAD1kMxEj88MMPajVACCCuATdH6w+gjCXBW8S5iNEgYh+hw9XyvcXZwDuPC/qB98zdO3Y54HKgHRwgNQ5zPgV7CJ6jNgPBfQgDWAGIfyDYjt9af0g/XLBgQTuu0r5duBZpkAQS4oJwtfz28c3d68cccEH/xzxx/+JywOWAywH12ZMWSGwA/nXSHIlTIPaBWAV8/vyN31p/+LuJrfAGK7ku5yR9kRgAru+Sy4HOcMAF/c5wzT3G5YDLAZcDLgdcDjiQAy7od/KhEW1LWUzyZqlw1pUPOb9UYuvKOQLxWHgG7wLx3jtzz8xVUt46c2ygHuO+mx1b22iLTJoj20CdM5257668m+AQWR/tJRf028upVvuRpkQeMWVWTQOTzm6p/EXuL6U1O3uOQDsOXsEzeBdo997Z+6XiHcV1Ont8IB7nvpvNzZna++ypl0CZYtMoqb3HBfp+vJsU8OoMH26//XYVtFpB1FX/64L+VVlz7R+ojkVRFvJmu0rkFZOj6/rp2s9J/KqUqyV33aX2cYCcdirdudR+DjC/mGfMN5fa5gAxB2idbqBh27yy7kHtDGp+dIYQsGjL3F5yQb+9nGq1nwv6rRjSzf91Qb/jDHdBv+M8c0G/YzxzQb9j/DJ7u6BvOGHjrQv6PftwXNDvOP9d0O84z1zQ7xjPXNDvGL/M3i7oG07YeOuCfs8+HBf0O85/F/Q7zjMX9DvGMxf0O8Yvs7cL+oYTNt66oN+zD8cF/Y7z3wX9jvPMBf2O8cwF/Y7xy+ztgr7hhI23Luj37MNxQb/j/HdBv+M8c0G/YzxzQb9j/DJ7u6BvOGHjrQv6PftwXNDvOP9d0O84z1zQ7xjPXNDvGL/M3n4N+mfOnNEOVQMHDpTXXntNvvnmG9m2bduPenvDDFLYaGTx7bffajvLr7/+WltpUhoTov41jS3eeecdefXVV2XkyJFal9owkpaiEydOlBdeeEEbZdCC0xzLPkzQdevWyWeffaZjGTVqlLYdbU/3LBf0DZd7ZuuCfsf57oJ+x3nmgn7HeOaCfsf4Zfb2a9Cnkh2dooYOHarV1IYPH64FQ8hRpFe4IepL01lqwIABgoBAZSyOGTRokAC4CATktlM8Y+zYsTJz5kwVDsaNG6f9svkdkKcz1bvvvit33XWX9qmmrzVEXWwWQQQPKhpRqWzIkCH6aU/NbBf0zZPqma0L+h3nuwv6HeeZC/od45kL+h3jl9nbb0EfLZ+FBwAOCQkRADghIUErOG3fvl3B2jCBybNs2TKtUkf7SvZNTEyUO++8U7c0wVi/fr3QzQqBgYYUCxcu1EpQdKE6f/68IGDQD5u2mM8995xaAQzoUxAHLf/pp58WLAJcb82aNfLEE09IaWmpNrUwY7nS1gX9K3HFN3+jtUj9ibNSffS0HD3V3GzEBf1r8/r8xX8pzyqbTsmpczSIEX333OI8V+fb6XMXpObYGZ1nZ85fFOadC/pX5xdzCj5VNJ7SuXbuwkVdR93iPFfnGb/wTpbWn5TGk5eVXL8FfUB4x44d8uyzz6qJHwagzaOdjxgxQjV4w64jR47Ip59+qpWG6FsNlZWVSa9evQQBAS2edpOY9k+cOKG/x8bGqtaO1o+AYSgqKupHoA+TuSatMjkvxPFo/tHR0drRyhxvtlyHRYD9k5OT5ZZbbpHU1FQVOBA6Ovs5eKhE4pNT5cSJk50+R2ev7YTjjp84KeG5h2X1zlKJyq+WktrjUt1wXBKSU1RAc8I9dPcYK+uPSVjOYVmafEjSD9VJWd0JSU7Pkvx9+905dpV39UBVo6zPKJNVaaWyu7RBKhpOyP7iUklOSdEKc939DO1+vabjJ+RAVZMsTiqWHbsPS97hJjlQXi1hUTFypLbOnWdXmWdrdx6SeXEHJGl/lYdHVHYF5zrzzD///HP7VuSrqKiQFStWyPvvvy+5ubmKpSUlJTJ//nyhZrMBd35AE3/xxRfVPG9AGa0dTXzt2rUSGRmpDR369eunjOKYXbt2aQMcXAZWE/2VQJ99iRVYsmSJp5QrAI67AAsAAkprwjpBfeR77rlHbrvtNvnZz36mPbOxVnT2E5+QIOuCo2Ty6hCJi4/v9Hk6e30nHBcTlyB9JkfIX34IlRuHhskjEyLkuyVRsmhDsIRFRrs8u8L8m7clVvpOi5DfDwqVG34Ik2emRMqARSGyanuky68r8Iv3YNrGWLl5eJj8YVCo3DoiTF6ZGSljVkXIhm0hwnvqhHelO8cYFBUvQ1dGS68RYfJ/g0PlntHh8vbscJm4IkjCouNcfl1lzrw2M0IemxAmw1ddXrvCwsIU0zrz/OhBYtsyvIA+JnuA0wr6CxYsUNAvKCjw4CygTwAeQkJ5ebn+HdDv3bu3gj4BfHRy6t+/vwf0MzMzFfSHDRvWJuhnZGQIgYFLly5tAfoE9V0N9BEE8vPzhWMJPrzpppskPj5e+23jbujMJ2N/uYzYnCGPTYyQJTG5UlxR06nzdObaTjmmurZenp4ZL/eOj5EnpifIE9Pi5ZZhwfLE+GDZlJwvVUfqXJ61mn/bdx2UV+cnyy0jI+Xtpely37gouWNEiHyxPFmidh90+dWKX7wL61MPyE3DwuXhKfHSZ2ai3D8+Su4dFSwfLYiWgxXVUlff4PLNwrcD5TUyNSxXbh4RIU/OSJQX5iTLHaPC5a4RQTIrfI/kHap0+WXhl1lvP1iWJn1mxMvC2AIPf1BCwUSzT0e2xiLuAc82vnRr7f2amhrZtGmT+vDxo0MHDhxQoCZAj++G2BfTPQJBUVGR/hnwv/fee2Xz5s3aBAPp5uOPP/ZE5OPLx1xPy9W2zPtcf/DgwSo4IIxASFl0O0Kjb2hoMEO54tZbPv2ao6dlccIBuX9suLwwN0Ui86tb+HquePEA+yN+w2dnJyt/JoXvkw27ymXQlr1y+6gI+XJtliQeqJULF/HAumQ4wDx6Y/FO6T0zSefUqrQSeXlugjw2JUaGbs+VQ7UnBL+/S5c5ELSnUm4dGSkfr8pUV9KcuCJ5c3Ga3D0mUubGFcnhxlMaG3H5iMD+VnP0jMyOLVLBckRQvsC/8aH5cvuIEHl+TrJszCyX2uOX3ayBza3Ld//J6kx5bk6yrE1vdivzi9/69Om8hC/8oYce0k5MADMSDj6JxYsXS2VlpYczgC4AThpeSkqKgjh+fHz6mOsRAIjMJz4AAYHAPQL2iPQnzY/of0NXMu8T/Dd37lz56KOP1K2AOwDt/ZlnntGAJ3wr1yJvgT7BMBmFFfLx/HD56/AI+Xr9bkksrJWTZ92uXob/gH6fWUny1pJ0Cd1bJQRc7a8+Jm/Oi5MHxkfJ4C17Jb/yqNnd3YpIeF61vLZop36OnT4v5y9clMURWfLslHB5fFqCzIo5IFVNp+X8BRf4zYTBL91rVKQKRQRa1R0/I6tTD8pD48LkwUlxsiGjXIP8zP6BviWwdkb0AQX9lWmlcvb8RSmubpQvF4bLnaPD5cOVmRKWWyUnz54PdFa1uP+PVjWDPvEjhvwW9EmjA6wffvhhjbTPysqSVatWSZ8+fbRFLcF7mDUImAOEidZHk8fnj+ke1wAgj5ZOLj1R/fj42ZIVQGAebQaJE+B3gBs3AS6CRx55RM32mOeJ4DeZAw888ICCPcIHaYF9+/bV6zPWa5G3QJ9rHK6slA1h8fL+igy5bXSU/LAtV3IPN7na66UHcObcBek9I1FBn0UEOn/hgqwITZFnp0bII1PiZXRwvlpIXI2/mWnw6dWFafL6op0qQDKds3bnyOTNSfLM7GS5a2y0bNtdIUeOnXG110vzbPvuwwpgw4PypKy+uRZIfkmV/LA8Uv42PFxeWZCm2qwLYs0MA/SnRe2Xm0dGqmWEd6+x6aiERMbI5ytS5b4JsfLp6kzZXdYoF9tYTy89goDYfLByl2r6WEIM+S3oc4MUxwkKCtJUueuvv17BmBx5gB6zOml3RM8DumjvU6ZM0X1uuOEG+cc//qHBDk1NTcorBAj2J4qec7300kuyYcMGBXx2wKpAjv4vfvEL+fd//3f55S9/KYD81q1b9fwIGaNHj5b7779fjydyH6sA122LvAn6CCbJqTultP64vLYwTRfkYdtzpfHUZWtFW+Px59/R7B+fnihvL82QiLxqvVVS9lLTdsqiyGzVZllgiFR3edY8E4L3VipIYeJHaGLNRTDOziuUTZnlqtHCs6i8arWc+PP8ae+9bcs+LH8bESGjgvOkvKEZ9KuqqyUoKlHGheTJnWOj5aOVmZJ04MdBvu29hj/th6VoUsQ+uXlEpJqqWbNJfY6Ni5P8shr5av1uuWNMtLpLjp4+5yoxInLx4r/k/eWAfopszgoQ0EcDx3SPZLNz505NvSNAj7/X1dVpXr3Vn46/HbO+2RfAZ8GHKOZDCh1aOr+jxVuj7tmXjADAH389W1IjzD64AMgM4G/meI5pS8vn2t4GfQoNnTl/QWL31cibi3eqD3vjrnJXQqaQ0rkL8ui0BHlnWYam7MF/k6eff7BMVqaVqPn/sWkJklvRpGZGnSAB/E/QnsPy8oJUeXNxupA/jd0K0C86WKw5wosSi3WxJjYCTcwlka3ZFXLT8AgZG1og5Q3N7j3Wl6SUNCk5cky+Wb9bXSMjg/Lk+OnzytNA5hugPzFsn2r66y6Zqk1xnoamo5JcVCtfrs2Wv0+MU0tA40lXiTl/8aK8t3yXxicx3wz5taZvbtLpW1+APsJG06lzMi2qUJ6amaQScsPJswEvIQP6/5gaL+8uy5DoghqdOgb0WZTx748LLZA/DgmVeXFFUnZJS3P6HOvK+DFVvzg/Vf65JF21C84F6FOcB99rYfUxtSr1mZkkK1JLtIBPV67n9GOxhGzJqpAbh0fI+LB9WnCGe2J+oRAw3yLyqpRnL85LVeE80F1JFJnhvbtlVKSs39XsnzagT/wWhbSoeYCVDp7tqzoW0MGjCN68e+8uz5C+81KFd9SQC/qGEzbe+gr0ueWUojr5fE2W3I/5Nb9aXx4bs8LnQwP0H5ocL+8tz9DFlgtaQf/U2QsSv/+I3D8xVl6enyZx+4+oduvzgdn4AmgRLCxvLU33WK4M6DPsE2fPy/KUQ/KPKfHy7Ybdkn84sAMhMbvi9kDTnxSxXyP14ZMV9KkKiRXg8ekJmj1y7PS5gLbEkc1ALE2v0ZGCVRKygj7/31PeKAO35KhAviGjTGNIdMcA/AfQZy1DeUEgJ9vBkAv6hhM23voS9PF/sSDfPa7Zh3ig5nhALy6nzp6XBybFqS8MQIesoM//0e4xu+JfnBVbpJHpNp4+Ph8aWiugj0vEkBX00VIxz6J1kBmxIOFgQGthpC+SCgroT428MujDx/C8KnlzyU55YGKc5JQ3CgJnoBKgzzt3+5goFZjgQ2vQJ3MkPLdKbhgWLh+u2CU7i+sDdi3DmnTizHl9J1FOQnKag5Lhmwv6DniLfAn6SIQZh+o1aIjqYBG51epDdABbvD5EXpTjp88p6LNoJBReGfSRoAuqjqn/kFQ1UvsCmYgMBvTxHxqygj5/g7f49nElvb0kQ6gZEahR1sQ9kEJF2uyM6ELBdA1ZNX3+T945OfsE/FHvwET5684B9g8194dtbw5w3JzV7J9uDfqsZcW1J+TT1Vly66hIVWaIhwhE4n07duq8BiSTCWIykeCFC/oOmBG+BH1uv+7EWfUxAvrfb9wTsMFWgBBxDmj6RE5TiAdqremzH1oF6Y6PTk1QsyPm2EAltFZA/4MVVwd9eEM8BPPr4cnxWkM9UOtDEES7Nr1U/joiwlPDAP60Bn3qHSB4vrZwp9w1NkbiC48EbDxERcMpGbItV+4aF6NrFfxqDfr8jfcyYX+tpvB9sipwsx8AfZrsUG+EdFqTiQSPXNCHCzYnX4M+5kYCX0jhIyqdZjPHzwSehAyYN5w4qxr8J6uzJLmoTmdGa9Dnj5iskw/UysvzU/WlwqwYqLQuvUxBH0HJUGtNn79jISHYitz9lxekaaU+tN5AI/jAO4YGjyaP6wNqDfr8DQ13UVKx/HFwmAayITgFIpHWOGhLjtwzPkYzH+DBlUCf95J3+PM12fLUzESZHl0YsGtZ/fEzGlxL/Yyogub0Y/jmgr4D3iBfgz4sIMVlw64yrTlPalVeAFadY8HApEraz2drsiT14NVBH1MipsMRO/K0FviQrXs1WhYJO9Bozc5S6TsvRdCsDF0J9Pktp7xJ4NXfRkTK9uzD2ibVHBMoWwrukPoJD4hvoPAMdCXQJwI7p6JJhfHn56ZIcE5lQGbYEEfTf/Me1eCpcQBdCfT5O8L7tt2HVbjE+sScCzSCBxTDIo2W+hkx+5ozkeCDC/oOmA3dAfoAHuYgIj0x1ZILG2gAhsWDBZjI/C/WZktacb3Ojitp+mbahORUaiU66vWX19Y9ldgAACAASURBVJ/UErTmt0DZUmsf0EdQMnQ10Mekb6rRUVAFC1OgEQFWy1NLNOccLZ668tCVQJ+/I4hODN+nhbSmRRcKqbWBRsQzfLdxt/x9Yqwn/exqoI/cjcvyq3XN2v7ChIOBxi4VfFjLAPw3l6RrlpFhggv6hhM23nYH6HP7vCyjQ/LlkakJMnR7XsClolEbvrKpGfRZMNIPtQ36LEbDd+SpBoLGS75woBG59wiKn6/N9tz61UCfHbJKG5qrQY6L1t4PgSZc4ndelnJIQZ/KjjXHrg36uAOySxv1vSRIzQijHmYHwBf6E9ArhL4EO/ZcW9OHHaxlc+MPahMotH3SJAOJ0PQrG0+pQoJfP6HwcmVHF/QdMBO6C/RhRWRetfpbKUOLWSyQXhX8y/gOKRnLApNR0jboY37FVEsOOrn9xj/rgGnltSECYID+V+uau1ly4muBPjyeHXtAbhwWLgsSD6qg5bXBOOBEBIsuSW4GfTR+zLDQ1TR9FnCsA2RHYFFanFTsgLv07hBL605qxb2HJsdJUE5zzvnVNH1zZYCOhjNPTE9UFwnCU6AQc4ZKj2QXsZZbyzm7oO+AWdCdoE/60Lcb9mhqFQtTe8oEO4CF7Rri2QsXpaT+hII+RWQyS5tbHl/LvM+JifJ/f8UurZe+q6Qh4OrLL0lqBn1Kxxq6FugDYORQ3zEmqtmNcil2whzr71viZ0hfvGVkpAY2mpawVwN9ww+6FVLn4PtNe9TEH0gCeUndSflsdZYK18Q1QG2BPgI4xY/unRArtC4OpAwb3LWldSc0yJj6GRRhM+SCvuGEjbfdCfpIiEQUk09NBDsFQfhbIBBae3HtcV0kSC3LvlQnvi3Qx/Q4K/aAkPKIudbkXQcCz7hHAAxNH2HR0LVAn+mEH/u1RWlCaV4i2UlPCxQiupwAPkrKkrrXXtAnsJRSxy/NT5O0g3UB814yLw7VndBS4aTIhlyqi9EW6AN85PSj6dMbghz+QFnLuHfWslcWpqmFKK3YBX1HrS/dCfowhkp0aK6Uoy06clzoMR8IxH1SkfDe8bHSf1OOp15BW6DPcZTmJe+atDV81oFECxKaQb/fpvaBPrwhJRThkmpzI4PyPcAXCHyrO35G5sUXya2jorSOfN3x5sC8tjR9qmdSpIeGUFOjCgMq5uZQ7Qmtsvf4tARPMay2QJ+5hOWN+Jw/DQlVE3eg1IYA9A8cOa6uWjrtUYDNkKvpG07YeNvdoI9ZbErkfuk1KkoWJh6U2ks+RxuzyCtDw+dHNDmgP3BzjifVpy3Q5+LkTxMwQ+/4QEurmhd/UDX9/ptzPM/hWpo+OxE/kV95VJ6dk6w1++P2X04p8pzET7/wPiHwAPqbMis8Zue2QB97G5kSz8xKkpcWpKoVLjBscM2V9oiZeXJ6oqe6XHtAn3iJVTtL5Tf9g2V6VKHWhvDTadXitgD9/VXH5KX5zUWzEH4MuaBvOGHjbXeDPgsyubAEp1F8pqjmuI25472hAfp5h49qARBqFeytaM7vbQ/oE4FNQFuv0VEqMGHyDxTCX0rKHsVTDLUF+pj4T527IN9s2KN+aro9BgpRgphARoTqrZZaBW2BPvxBYyN2ggZZmPiJjwgEwjT/9tJ06T0jUcLzmgvNtAf0WcuIH6GS3xuLdrbwbfsz3wB9hGpAH+tjdpkL+o563t0N+jCHIDZy1WkKEruvJiAWF0AIoL9nXIwM2ZqrAgC8aA/oIzDwklFilhQhazEMR022TgyWeAZAf/DWvZ6j2wJ9syMxEBSd+XhVppr4A8HniiVtZswBFRB37D7sybtvD+jj/58ff1DuGBOtufs0ogkEOnjkuDYfIgaEDCOoPaDPfrgGMPGTlbM2vSwg1jJqjuQebtI2w7xbu8sbPdPE1fQ9rLDvl54AffLVl6WUyK/7B8vs2CIJBM0Vf9/uska5e1yMDN2WKwWXqhK2B/Qxs6J18YL1ntHcSQ5pOxBoRjSgn6q9CMz9thf0yT+nkh+LeWJhrVY1NOfw1y2BnpSHvW10lLqCKIoFtQf0EYoi86tV4yWSP/fw0YAookVsEeVkcW3QAhxqL+jDXwpC3T4mWkYF52vcjp7Aj/8B9PeUN+l7SUC2sVpyyy7oO+DB9wToYxbLLGmQG4aGa64rJjJ/J8qjYuEA9IftyPNUi2sP6MMbeEaaIxHGuAcCpXIapnk0fXhmqL2gj6A0IWyfkH89OWJ/QGhh1NMnEA/Qp/tZ46WCTu0BffiLBkfQJNkiMQU1AZEiCujTLY46BdEFzfEf7QV9MkOwrtBXhJx1a/MZM1/9bQvok330wtzmSpm4LQ25oG84YeNtT4A+7Cghz3NBmkYLb79UBcvGbOry0AAgfKb4/0YE5UnhpeYm7QV9NPvs0gaNmH1naYb6Ers8KAecgKBPQB+eGWov6LM/JtenZyVrIRFy2P2dKE4EzwB9tHaK9UDtBX3Kq5Lm+NsBIRoQGAgtd8mqeXFeqjw/J0UFHfjVXtBnXwTyT9dkyVOzsML5f3EjqosSvIfrjEqZuB4NuaBvOGHjbU+BPpGvaCRovgRrGY3Exqzq0tBIIyOflTamVjNge0Gf4LRjp89pLAQghr86EIgCKIA+PDPUEdCnuNEnqzN1nu2rOur3Jn5cZdTSx9wcXVDtKd3cXtAnRZQFHSvcl2uzW6RjGf772xbQf2FOivSdm6oxRtxfR0AfgXx+wkF5fHqixuvwnvozYd2gjPhzc1J0PbJ2Z3RB3wFPvqdAH803pahWu86Rt271CzmAbR0eIqCfXFSrlfXGhBQIwUNQe0GffalgODWyUCPSMcEGQmDahPB90nduisAzQx0BfQqvYNq/cViEFlMhj92ficAyXBqAPjUxyL+H2gv6CJfwDC2OWAiCAfmbPxNWN0z7RKPDM6gjoM97SHEjjn93WYanBoe/8gzLBgoMKbEEMR6oudzYygV9Bzz1ngJ9pGNKV+IXolsTwTD+TDRCoV73nWOjtXc5aUJQR0Cf/WkIQtARCwzWEn8P6BsXVqCa/rjQzoE+wuWWrAq5dVSkDNmWqwWh/HmeMa/gFRH4BC8y76D2gj77Mq/GhhZoRDp1Emjz7M+EpoqA88qCVC2Exb12BPSRiaiPQEbSc3OSheZY/kyAPgrMM7OTtY+IUWC4Zxf0HfDkewr0DWsoSUukMBHH/qy5onHF7Tuii/H4sH0a0wAPOgr6+M/6bcrROgdU6vP3KmBjQ5pBH43fUEc0fY5BK+k9M1Genp2kjY78WXMlKA2rCKCfXFSn1QnhQUdAH0EpOr9GHpgUp6mSVvOteQb+tOX+yNF/dWGaJBR2XNOHF1jhcKs8NTNRhmzb69drGaCPQImbkfLYWJcMuaBvOGHjbU+D/vKUQ/LkjESVGP3ZF0ZbXCKDTQ60CZDqKOjTr4DmKHePi9aGH/7e6ANf/gvzUvRezWvUUdBnUR+wOUf+MjRMQvdW+rVfH/80PMOiZC2w0xHQx3pEdgiLOkVnQi/Vozf897ctlTKJvn99UZo2uOL+OqLpG35szqyQF+enqmWK99RfhUtAHzdIn1nJ8t2GPUKXQkMu6BtO2Hjb06BP6h6NPmjTmF5c77ftdomiJp0HXys+ZqKsoY6CPtYQzNVYRzDx+3sBlZFBebqITom8XFWvo6CPuXpdRpn8ZWi4kPfvz3UhEHBGBDWDPtkiaO1QR0AfczXATwvop2clqZCpJ/HTfwD9R6bGy5uLd3raxHYG9ClSg+ZLtVHSkP3VCkfHUNI5cYnQldGsZUwPF/Qd8JL0NOjTFYxAPjrv0cvbX6VjQD9sb5WCPsF4Bqw7CvpMKTS4z9ZkadU18mX9uWnR8B3NoG8tpdtR0Kei4Z7yRrWy4Hcl6MpfCQCDZ2j61IWgPgTUEdA3vKE3Btr+dxv3eLIAzG/+tC2oOqYNwOhvgUsE6gzos5bhpqQ6H0XH/NUKR8dQ0kFZs3E1mrUMvrmg74A3o6dBH5CfFdPcbhfNgsIP/khU7qJZDvnTNOegKiHUGdCnxgEa6+8HhcqGXWUaeOWPPOOeqF5I9D73a6ijoM8coyY9lhGitDdmlqsP1pzPn7bEfNAtj3oQCDqYmaHOgD590gmyJXA0p7zRb/3U8Ozvk+KE+hemN3xnQB8r3KbMcnUVIECUN/hnGWNAPzy3SuMgcJtRnMiQC/qGEzbe9jTow5qQnEp5c3G6So5Ix/4YkY6PlAwFQB8AowgK1BnQR3uj2tqffwjTvGA0FX8l+hQA+tTgN9RR0Oc4YiqIan9wUpxgNTh1SQM25/SXLRX1CCQD9EmDpecD1BnQxy3y/aYcXdzpwOeP7yW8AfTRzkm3M1agzoA+5yKqndbht4yMUv5TyMbfCNAnNoZYrIFb9qpAbe7RBX3DCRtv7QD6vHQ0VKHtLGlt/pgihOlva3aFgj4AxoIKdQb0OQ4tDt8hKUJJB2ptPMO6NjS665HWSQEnQ50BfUz8mCQJ2CL4yFRENOf0ly1AP2jrXi1GxHvFfUOdAX2OmxFdKI9PT5B+G3O08py/8Ml6H3mVRzUwlt7wpiR4Z0Gf9DVidq4bECLbsi+3NrZez+nfAf2gPZXyxPREnWu1ltoXLug74OnaAfTJJV6YWCy3jWluHWu0YAewr91DxIKB6Y/2uPj7ugr6pMkQC4HlYHNWhRBc44+E+RDQnxffNdBHS2VeUWP9tYU7JSSnyh/ZpY1Q4BmVLul5buI9Ogv6WKdeW5gmz8xO0oqQ/phWS+34O8dEy4crdmkwMROjs6CPwoLl8rcDgmV0SL4U+mHrcOYU88JUILTGLrig74BlxQ6gz0KC5EgzmTcWp4spXOMA9rV7iHUnzsr6XWXa53xuXJHUHW/uftZZTR/pmhrpt4yMVHcB3dX8kag8COhT5tRQZzR9jKykGqHl49e3ugvMef1hSydHhMF7xsdox7eugn5WaYMG8mH+tsYI+AOvzD3gErltVJR2sSTjAeos6CNckpGE0PX20nSP5cBcyx+2zKkt2RVqNfthW65Ye1q4oO+AJ2wH0IdN+NLwqd02Jlr7M/ub/5Dyr2vSS+XWUVHasxxzP9RZ0Le26qXAEQuNPxKR44A+liBDnQF9cyxgD+jDM4QAf/O4ZpU26r0B+sW1xz01CTqr6ROkZVr1ImT6YxljXCIIz/RooO8A1FnQ51hqJbyzLEPdb/i+/c06AuhjtXx0WoIM3Z7XIrPDBX2z0th4axfQJ794dHC+/L9+QZrPTq16fyI0c4KhAP2FCcVi+px3FvRZSAgOpJIY0dVYSvyRvtmwW0GfdE5DXQH9kL1VauKHZzVHz/jdgozwR674veNjpLTuhMcP31nQRzAi24EUwP6bc/yyxgGg/9fhEZoGa4TnroA+mTl0OkSQWJZyyFMV0cxfp28B/fUZZfLI1ARND7XGYLmg74CnaxfQBxTX7ypX0Mfnba3y5AA2tjlEfPhUH2QhAMBMI5TOgj4XJEjrq/XZCvwLLObvNgfjoB2+WtcM+taugl0B/YKqo/LlumwtLELQKEFJ/kSYp0l9JSiWoimANtRZ0OdYykdjHUHALLC0UfUXvuVUNMkNw8K1dj6WEqgroE+mSFRBtdwxNlpGBuWr5u8vvOI+zpy7IGvTS9WSQctrUwuC31zQd8CTtgvoszixYP1xSJh8vW634Ev0JwL0lyQ3gz5bU3K4K6APYM2PP6ixEBRk8TfrCM+fYjqY95enlHimQ1dA//iZczImJF8emhwnc+MO+l3VNKpaItTgg69sPC0mZawroI8vH0GCFFGK1xhBwvNAHP4lp7xJrv8hXDvGUewK6growx8K1jw0OV4+XJnpaeLjcDZ5ho+ygdXy4SnxWvLZZIiwgwv6HjbZ94tdQB/fKr4w0kCen5Oi6VX25VrHR1Zz7IwsSiyWmy+Z/Ex51K6APoWMaLpD2t4nq7MELdbfiMqDgP7KNO+APm4RhC5yjD9fk93CH+kPvKO5EIISoI/7wsTGdAX06RMxN75IfjMgWDvImcwTf+AX94BQ88choSrYEAgJdQX0KQZFUSTKi/edl+p3XfcA+RWpJSrU4JK1Zg65oO+At8IuoA+rkI6JPP77xFiNTPcn0yvpYmjlgD4vjKmU1hXQB8AqGk9pAOTLC9K04p8DplyHhvjJqkwFfYLIDHVF0+ccNJChzjoCJv5Xfwq0IiAWQQnQx2XmDdAnpTYqv0Z+PyhENTt/KwZFzXzu7dsNu1UAYI50BfQ5Hm2fxkcIl5Mi9vuVdQTQx92GtYyOjucvXnaRuaBvVikbb+0E+gSm0YuaoKHJkful5tjl8o42ZmG7hgboU2DmbyMiZVVaqadoSldAnwuzqA/c3Ny7gChrf6OPVjaD/tr0Ms+tdRX0MeGSCkh8BT5cI4B5LuDgL1SEw+pz/8RYDfQ0Ak1XNH3OQU3/XqMjVcBM9LNiUGj31w0M1qwH/PtQV0Gf93LjrnIFfYQJf7KOAPrEJVHdcmxoQQuh2QV9BywedgJ9JhOFMvCF4dc3kbQOYGObQyT1iZa4gD6CjbFidBX0uTCFawi0wqzrb/TBil2q6VPjwFBXQR/tfm78QblxWISsTCsVXC/+QlRn/GhVpoI+AWWmgVVXQB/eUEf+zSU7NX6ELo/+RAiBv+4fLP037dHSudxbV0EfQYmsALJEMPOnHWzO//cHvrFOL0gslgcmxWlp63+ZSebvPv0zZ85Ienq6fPnll9KnTx/58MMPZcOGDcLfWxNMWbdunXz00UfyzDPPyAcffCDZ2dly4sQJ3bW2tlaCg4PlxRdf1HMNGDBA4uLiPKdhvzVr1sibb74pzz77rB6/e/duOXv2rDYO4YUeOXKk/v7cc8/JZ599Jps2bZLz59tOe7MT6POinDxzXv1gry5M07K1HiY4/AtAgyb+1xERgtZqgqG8AfrN5upm/6HVpOtwlunw31veDPqkjRnqKuiTcrRjz2F1tVDmt+jIcXNqx2/JSPhgRaa6yEilMutxV0GfeYX1jaIzCJkmJsXpDIM/BA3/ql+wDNy8V3IPN8fFdBX0iVHCLfLl2mxtT2yNSXE6z+jnwBxA058Qvq/F7fitpg+I19TUKNCPGjVKFi5cKGPHjpV3331XDhw40AL4Aea8vDzp16+fDBs2TBYtWiRjxoyRb775RgoLCxW0U1NT5b333pNp06bJ0qVL5fvvv1cQ5xpcKyoqSkaPHq3HL1myRD7//HOZM2eOFBQUCAJDSEiIvPHGG/o3hIMRI0bIW2+9JYcPH5Zz5861eCit/2Mn0GdsvIQUTqGP98yYy01WWo/baf+nYh65u+QDk+NqfK3eAH1qh8Mz8mbR9EyTFafxqPV4mfsUOSGQj1LDhroK+swxuqn1npmkQZAmYtuc38nbuP1HhBrydI1DgDbUVdAnMyQqv1r+PjFOftie6zeCEopGxqFm0B+8Za9aGuFZV0Gfc3DuCWH7pM+sJBm2Pc88CsdvWV9mxx1Qn/6k8P0t7sdvQf/06dOCpn3PPfdIWFiYauxJSUkKvJs3b1YgNpxg8iAU9O/fXyIjI+X48eOSkpIid9xxh8THx0t9fb1aAR5//HEpKSkRzg2wIxSg7aOtDx06VBAusCxw/LZt2+STTz6RHTt2yKFDh2TevHnywgsvaLoEx69du1buu+8+KSoqUmuAGcuVtnYDfcZIznmfmUkycEuO32gUBClODN+noI+vjwUB8gboU+gHgYIASCRwa1nMKz1zp/wNHr29NENemJcqW7MPe4bdVdDnRAhK1ACgdwE51cby4rmIQ7/E7quR95ZnyIOT41qkI3YV9OEPef+8l7hcuI4/EMI3TXbQ9CkpS5MiyBugz3mw6r00P1XeWpquwjiCrNOJGBgUsocnx2lzIev9+C3oo11v375dnn/+ecnKytJ73r9/v4wfP161cYDU0JEjR+Tjjz+WuXPnyr59zaaQsrIyuf322xW8MfMD2lgJTp48qYcB9lgOZsyYoVaDl19+WaZOnSrHjh1TkKirqxPM+AgTlZWVavrH7I9JH+FjypQp8tJLL6k1AlBpTY2NjSos5OfnS0REhPztb3/T4zh/Vz7FxcUq0PDCdOU8QZnF8vzsRHlrSZrklBzp0rm6Mg5vHru/vFZGbNsjNw4Ll1XJB3RR4fxNTU2SnJysz6Oz14Pfi+P2yYMTYuTjlbvkYGWdX/CssemovL4wVZ6bnSjrU4s895SZmalWrs7yi+P2lR2RqaF75Y+DQ2VRfKGU1jR4zt+V8/b0sUFZJfLPRany4KQYqalv9NwTygHzjPnW2TE2NR2VNxYky7OzEpVnnT2PnY5raDoqMXvL5Vf9gmTgxmzZVVSl/GFdjYmJ0aJGXRlvxO4SeWdJqjw2NV7yy2qlvrFra2NXxuKtY5lXE0L2ykMTY2Rc0N4W8wk8BFc6cy3c0liw20v/1t4dvbFfRUWFLF++XH3rubm5ekq09AULFqjpHbO7oaqqKunbt6+sWLFCysub/ZKY3Xv37q1gDegC7lgCTp1qbprCogZwo+GjuT/xxBMyffp0c0rd8reZM2fKxYsXpbS0VK0M1113nfzkJz9RCwQxBFfz6SOwEB9w8803yw033CA//elPVSiJjY2VrnywZBCb0JVzcOzG0Bh5YWqYPDg2RMaviezy+bo6Hm8cvzEkRj6cGy7XDw6W0ata3hM8g3dduc6UdVHyxIRQ6TUsWNYFR3fpXF0ZhzePjYyOkd4TQ+XRcSEyfvVl/oSGhkp4eHiX7jEkMlZmboySPw4Klvdmh8ny7f7BM+ZBn0mhcvfIEAmPuvw+e+vd/GheuNw3KkQ+mNs1/ntznnTlXFExsTJ7U5RWAn1vTris2NE8D6Kjo9WSyrYr598QEq286jU8RMasjhTmXVfOZ4djw6Ji5fMF4XLXyGD5YkHLedCVd/P111+3N+gvW7bsiqD/xRdfqBZiEBrQx/R+JdDHDA/oA+hXAn1iANoC/YaGBmFi4tPfunWrauxYBRA0MO9fKbAQaR8hBeGExcBumj7S95Ate1Q6HheU0ympsTOSpi+P2VdeK0M275abR0TIGovW6g1Nn3GnFVZKv43Z8uchoRKzt0yOWLQ8X96XL899pKFRXluQIn3nJsnm9GLPPEAoxmrWlWs3HT0mBeW1cs+4GPnnwhQJySrp0vm6MhZvHrtt1yF5Y2GqPDolTo7UX9bqvaHpM85VKQfl6ZmJ8uGynVJaXe94ntU1NknEnlLV9Idu2S3ZxdV6T6zbaPrEVXXl+dQ3HZM50fvk7rFR8sPmLDlU5XwrXE19k4zZkSP/mBwnE0NzW/DHbzV9JgK+e0zo+PYhAvgmT54sgwcPVrA1oM++BOnNnz9f9+HvaPzEA2zZskX99Jg0iOw35v3ExER1FaDtA9oIDQA5v6PZY84lC4BzAtyTJk3SQD+AnMBBtKBXX31Vt5jyr0V29OkzXlqpPnMpDc35XjCRsoaT2pyi16go2WbxT3vDpw+/KNJD8Z8/DApVP6LT09BwfZ44e17eWLxTyOQIy63yTGNv+PQ5GRHpnJ95RpyFP1B4bpX6j3vPSGpRf6CrPn3DG7rQvbU0Q15btFPIb3e6i5r0M4If8emTc15Y05zJ4S2fPvzZml0hT81sboxV0eD8Fth0+Bwftk9b67YOtvZbnz4pdETcP/DAA2qqQRsnyA7fPRYApERDgC7CwPDhw1ULZ1/8+Lfddptq6Jj6V65cKU8//bQeR7Q9pnnS9gjU4//ffvut+vjJAsAFgFDwzjvvyMaNG/VcWAmI7sfMD4gQ7Q/oYzbGEnAtsivok1LFgkxJXiKHTeDbte7Fzr9RypRAodvHRMn23ZeD0rwF+qRQheRUyp+GhGm3QkoaO5kIeKIpEXMAgInIu/xOeQv0m06d06pp1BCncJKpneBkvpG++daSdA24s2ZxeAv0a46e1sp1RKRvzrockOpUnsGjmH01Cvrjwwo8zXG8BfrwJbGwVt5dniF3jY3WIkdOX8sAfQSkx6cnCM3RrOS3oM+CZPzys2fPVhBevHixBtexIBEEQgAfEweQR0AgSGHWrFm6L75/gvPYF80dkAb0icpnX3z53333nRDwx+/r169XoYFgQIJxsCaMGzdOdu3aJYA2vn189IA85hX2wzpAvIGJE7A+GOt3u4I+GgXR1RSA2Fd9zFPBzjp2J30vrT8pg7bu1c5bCDSGvAX6aBQ0LLpjTJQuMJkOb1jEwkhWAsVN3li0U9PFDM+8Bfp0B9uQUaapRzQscrp1BP4E51TKP5fs1GJNvgD9ixf/1bzgT0uQiWH7PKmn5tk4bQuAkYqIpk92janZ4E3Qp+DY0O258ruBIZpSyzWdTLw3lBimjDXCspX8FvS5ScAUfzgBeb/73e/kwQcfVFDn70FBQQq8+OshhAT89g8//LD84Q9/0H0TEhLUF8LvCAkEBt54441CMB5uA0z/hpiA5PCT5venP/1JLQxo+wgUBOvhWiAIolevXvLnP/9Zx4QQ0laOPue3K+ibFDdK8lIBzPSfNzxx2rak7qQM2Jwjd4+LlqA9lZ7hewv0OSFtTwHJf0yJV+3FcxEHfgH0646fldcW7pQ3l6S3uB9vgT6a/Z7yJukzK1k+XZ0lTheUeMxBew5rXwEsZL4Afa5B3XUqQH64ItPTxc+BU0yHjIUsIq8Z9El7La71rnmfi1CYiyZP1w0I0fba/N/JhNuNlrr0FaARk5X8GvTRwAnwADRJUSBojlQ+NUsePSqk1fG7IX5jH7Mv+fYs+BDgTL4+wUn8jm+eAC9D7IflgGI++PABeVwMXIsPfn+OIW2Q3xkT1+e3tsiuoE/PZhrT3DEmWkaH5AtlbJ1MJXUn5PtNe+Se8TEtGuN4E/SxJiCB09QHH/WZ887VKAB96pXjz8dcjd/VkLdAn2vgQkBQenlBqmzLvlwAyFzLaVtcR1hG+s7zHegTX0Fp2SenJwouEvjoVKJqYdjeKtX02FomegAAIABJREFUp0UVyqHa5iqp3tT0iRuILqhRTZ/CRk5vWISgNGx7rvSekaixV9Zn79egb71RJ3+3K+izjmCqxG+Etld86WV0Kq8P1Z2Qbzfs0e5n+N4NeRP0646fkQ27yhX0Z0QXatdCcx2nbTEjI+i9siBNC/QkFHof9OEJgEU1QzTX1kFJTuMZ4yVIFCGGgjCAjSFv+fQ5HwF8/Tbl6FymLa2TGxYdO31O1xnM+zOiDwjCOeRN0GeOUfXx3vGxKsCmFTu7Dj+gP2RbrgYnUkjNSi7oW7lh0+92BX3YlXKwTsuwUjWN5hVO1igQWohRoIwpwVaGvAn6LL4syLhEWJSdXF6WSmm4eGgZTCleygsb8pamb84H2BPBj/vFlEc2vzltiysM0MdC4ivQRxhDK+a9XJdRJvUnzjqNTZ7xYqkgxgbQJygNaxnkTdDnfATWkvWA6w1LiYONI1olddCWvRosuiix2MNLvrig34Id9vyPnUGfkpjUrKZaFuZdq4/Snty8+qgA/c/XZGuTCmv6mTdBH6Go/vhZTaXB/Gq9ztVHZs9fAF/SHF+en6plZVOKfAf623YfVs347aXpGjviZOFyc2aFWsYAfl+BPrEQ69LLNBOF95MOfE4lSlbj1gH058YVCVk2kLdBn5TacaEFcuuoSFm9s9TRaxmafv9NOfL0rGRZknSoxaN3Qb8FO+z5HzuDPhrF8pQS+X/fB2nwEE1rnEoHjxyXT1dnCulh5FIb8iboc04igz9YmamaK7xzKgH6mFoxU7+/YpekHqzz3Iq3NX2sIx+vypTn5qRIVkmDo+vwE8uBlk/wo69AnwcRmV+t0dsvz0/z5LZ7HpCDvjScPKuBwoA+dS7oLwB5G/S5DsIF1pFJEftVoHUQm1oMlRTq77QpWrKuy9YfXdC3csOm3+0M+mj2+HLR9Mlxp0mKU4lUoI9WZsojU+M1Wtjch7dBnwBIUo+4DtvzF50ZZMW4D9Yelxfnp2qUeLrFD+pt0K8/cUbnF6ZXzNVWsDTPySnbDbvKPHEQ1vvwpk8fXqQfqtdufjePiJTs0oZ2BQ3bkYe4JjZlUns/WBYmHhRTPMfboE9Q7f7qY3Lv+Bj5ev1u7exnR360Z0yA/jfrd2sczLIUV9NvD89stY+dQR8za+7hJo3gf2dphiRb/Lq2YmI7BgPo053ssWkJqiWZQ7wN+nRDw69LZC1+faemBwH6+EH7zktVYWnXoctFprwN+ucvXNQgvkenJWj2A+ZLpxJCC5kI7y7LaCG8eBv0C6uPyZiQfI1IJ/qdKHgnEmmh63c1g/6SpGJP8Ku3QZ+1jPgBYkcoOGVN23Ua3wD9r9Zla1tqMqys5Gr6Vm7Y9LudQR+WEfWOf/LxaYmag9yeNEQ7shoAYyF+fHpii0Iz3gZ9zOIEPRIAx/XQyJxI5y/8S6uXvTA3RT5elSVZlmJD3gZ9+ANYYt5/czF+/XNOZJmOeW16qcZBIGD6UtOnkBHX+k3/lmZxpzGOUsy0v0XTR2s1LkRvgz58IRaCuJ6nZyVJ6wA4J/EN0P98bbY8PzdZVqa5oO+kZ6djtTvoVx89LWNDClTbpyiIdSFzErObo3fT5ckZSZqza8bubdAnKhiNAh/183NTBHOvEwnQJ5CTe6BwDn53Q74AfdOH/u8TY9XE69QofoLEiIPAlWR9V7yt6XPulIO18n+DQ2Xg5hzJKb/8fMxzcsL2yLHTGlj3q/7BWhek6mhzPRBfgD5WuOlRhWrtGx2c79iyz1h1eCcRyJlvVnI1fSs3bPrd7qAPgG3bXSF3j4vRJg/G52ZTdl51WIXVx1WLfGpmksQU1Hj28zbomxMP25GnKTX49Z1ILJBYLNC+P1uTJXsqLoOKL0AfNxJxI7eMjBTyqJ1q4kfzenFeqnyy2regz5w6UHNMi01hrrYWT3LSfEOpWJlaIr/uH6wAxv8hX4A+giS1R+hbQPouKalOpGOnz8tHqzIV9Ne4oO+8R2h30D9z/qL6dh+aHC9frs3W+vLO47JoEA+NY0hzQas05CvQX5hYrD43JHInukQAfQq/PDcnWb5Ym62xHYZnvgB9qv8tTCiWG4dFaMaIWfzNNZ2yxcfad26qmpHPnLvoGba3NX1OTKT7e8szNGiUOBInUnOG0CEFfcz8pv+CL0Afvz7V+BCScCMlW9JQncQ7QB/3EVUf4ZmVXE3fyg2bfrc76POikIbGBONlsXaosylLrzgsIndfWZimEa9WrchXoE8BINK2MPVSqc9p5mpAHz8+lfK+XJfdInPDF6CPv5UiLeRRD9maKwePNFdmu+LDtPEfl6WUqAaGgIzAbMgXoI8/fHLEfo1IpzKbExvJ4MMn1xxNf31GmZZ+hme+AH3OSwVAhFjmdWvTuHlWdt8C+u8u36WZNfDMSi7oW7lh0+92B33DNhaxZ+ck/6jWs/nd7tt9Vcf0JaERSryljryvQJ9GMqQGEThI7rlV67M7rxgfoE+nRaKdv1q/WwqqLqdr+gL0uSZV/8iuIM8d/jmRiHvB18qz9zXoA2CUlL5vQqy2czaFbZzEN0zsBNUB+hszy1VAZvy+An1ibvDnY+KnPa0TCdB/e1mGKhSU/baSC/pWbtj0u1NAnw5YvWcmysigPMdprTx6zHoEpfWdm6K1B8x08BXo1xw9I2NCCrQ1MZHpvKhOorMXLsrO4np1h5ATjKXEkK9AH3cCZkuaImF6xcrkNFqcVCzPz0nW4im+Bn3OT+oeqY5or07MFCFGCCsFoI+LwpQU9hXoM5/ouIdg9smqTKHHhNMIYY/2zaSGUuPASi7oW7lh0+9OAf3NmeUqHbO4OLHNLm1vMelhbk+01BvwFehjzmcxo8kHnffMYmbTafijYWFupwofGhGNish+MOQr0KdqIloYEelBOZUtot/Nte2+RWslDqLfpj0+1/S1w+f5C4L1CuvI9j2H7c6eH42PUs/z4ptBn2ZFDZf6CPgS9EP3Vsrri5pdfaS/OU24NJ0paYa1uVUshwv6P5pi9vuDU0A/s6RB+4RTUx6NzGlE+hkARv68tXmMr0Af/mzKrNBrwjOTiuQUvgH6SUW1QrYDJT+tPnZfgX5z7nmZ/PmHMHUjVTY6r50zgh7C5cAtOT4HfeYSwI8gTpbF/FYd15ww12iwMye2SDX9HbsrPTUafAn6pJ/ifnl0aoKuZdbUSifwDNB/bWGaCnpbW7WjdkHfAU/QKaBfc/S0fL4mSxe07budFylMCeHeM5O0GUpy0eU68r4E/dh9R7QtLT5X+oQ7SaMA9CnBDOhTWdD0OeeV8hXoE4iGS+GGoeGavpdb4Ty/PvXjiYMYvHVvt4A+z4NgPp7TiB15jjNXl9SdlFkxBxT0iU8gRRjyJeiTGULzHd5LzOOAqJPo6Klzatp/bVGatnK2jt0FfSs3bPrdKaBPYNfwHXlaXpa+106jvMNHtUEJGQgp3QT6WEQGbtkrfxkarqmOTtIo8BeT5dB7RpL035zjaXnKc/cV6OMSoQIkPn18+9aAS6fMNzrFkRZKzQFf+/QNT2jyg3WBKm1Oc73xvFlP8OnTlRJAg3wJ+tThJ6X2/gmxCv5kQTiJ4BE9MaiU2jqbygV9BzxJp4A+rGRBQ4vBx+s0yj18VM15pNFZO8b5UtMnmpoe4fioiYlw0uICYMUUVMuTMxLVVG26n/HcfQX6nJt8fVIrCbp0Yu45zxvQR0DGWmLIFyl75txYR5jXfChy5KT4R1peT43ar6AfkVetKXXcly9Bn/Mzt5jb7yxLd1yRHkCf9wMFpnUPARf0zVth462TQJ8XheARmrA4LeoVUzFd3N5ami5pxd1j3idICEn894NCZVL4fmGBcwoB+qZ966Cte8XURGf8vgR9NNUh2/bKg5PjtViP0wobzYo9oHEcI4Nalnn1JehT4IaUWgRyKs45yY1E8CbuCTT96IJq4Z2BfA36uN6ItblvfEwL15Ve3Ob/mMZBFBjieVvJBX0rN2z63Umgn1JUqzWfH5maIOUNp4S0LqcQJWUfnBQn7yzLUL+xGbcvNX0EI3hGaVniIZyUe44JNDy3SusMUCzHGojoS9Bn0V+eckifFX5Xp5mrZ8QcUP/6mFa13X0J+rje8OdTEwL/uJMEcrpfUqoa0KdSpim/7GvQJ5iv/6Yc+dOQMK1HYXXFmLXBrltAv8/MJBVaKAJmJRf0rdyw6XcngT4vKIsLaWix+2sclXueU9EkNHOhbKk1n9mXoM+U47poYHysWQM2nY6eYRF/QGAVxXLwT2N2N+RL0D9z7oLm6D8xPVHdSBRVchLR0IW2yggs3WXehz/NsQRJ8t2GPY6qo0Eq6PiwZtAnhsNUFfQ16JM1gFXmdwND1NRP1UynEKBPrM1bS9I1DsI6bhf0rdyw6XcngT6TjZzaO8dGy7z4Ikf5qHPKm7RcKQFiGZbe8L4GfQSlbzbskbvGxqgpjj71TiBAH38hoD90e66nUhpj9yXoE8xHlTYCld5emtGiOZIT+DY1slB9xRPC9nUr6JOvTa7+y/NT5dS5C47x61P0icp4aPqJhbU6dp6zr0EfvziNxAB9Co85yfXGOoxVh/eDOAgruaBv5YZNvzsJ9AkQIsWFVBfyXDHxO4WIpL9rXLS2PKW8rCFfg35l02nVwm4a3txIxkQnm+vbdQvoUyyFXGaC0qzFhXwJ+ohEtPX9cOUuDVaiA5uTaErEfnlyeqJMjuhe0MeKRIW5hybHqdCEyd8JhCWH4lWAPqm0JsPF16CPcElsz43DwjUeYreDao80njonj06NV1clcTdWckHfyg2bfncS6MPCqPwaLWH5xIzEFlXabMpez7B4qe8YEy2frMoSCg0Z8jXoI5VH5VfLraOi1ORL4JITiMUX7ZH4jRFBedJ08qxn2L4EfXMRrvns7CQZF+as+uiTIvbLE9MTNCLdGvPiS58+PMNMjuuNFtiYyZ1S9pmeDowb0E87WOexjvga9OEZGT1E8FOSFyuDU6jx5Dl5eEq8vLssQ4MfreN2Qd/KDZt+dxrooyV/tjpL/jwkTABSp0QKE7jTa3SUfLomS7JKL1cU9DXos/BjOsTKQOU0FjYnECZi8r/JeCAS3Wqh6A7Qp4b9M7OT5Iu1WY7yUU8I3yePT0uU6dGFLQJdfQ36CJdU5EO4JAfdGoNh5/lGpcxh23MV9Im1MYJSd4A+7yUdJBGUcGU5JdUR0H9wcpzGJ8VY2oTznF3Qt/NsvzQ2p4G+Cbz5Vb/mYho0f3ACZZc1ahQ9wMt3Q74GfRYSItAJ5Htj0Y/zas047LYF9GkUhEZBPfzjloZB3QH6EXlVWnyEXOTKxlOOES4JSiMOgih6q4nd16BPrAhlnwmy/X7jnhbFlOw2t6zjoWgWgaJo+ljgzl9yS3QH6JPqyHNCUKI7olWwtY7Rbt8B/fsnxsr7K3ZJ3L4jLYbngn4LdtjzP04DfTSItell8r/fB8mChGKhS5YTCO3+r8MjVLK3+u98DfrwhohkqqU9PStJliQVO4FdcursBVmzs1Qemhyv3QJNVDWD7w7Qp2wyaY4ISxRTsgKonRlIUBpd7+bEFbUYs69BH57EFNR4MkWsXRHtzC+KCVGyGNDPLmvwWHW6A/SxjpCWevuYZtebtdS0nXkG6N87IfaKVStd0Lfzk7s0NqeBPr5eutT9dkCIkL+NX8wJlFXaoDXdCUC0NgzqDtCHZ0R1ExTXOqrbrrwD5FellagZkfQzE2DFeLsD9Om2NmxHnvJs9c7SFte3K88Y15iQfA2yIsvFKqh0B+jjeiNm5W8jIoT57gTXGymtAzbnKOjzXpoxdwfok5tfVHNMy/F+tS67Rf0OO8+xhpNn1SXx4crMH8Ui2Bb0qbLFYnv69Gk5deqUXLx4UbtF8bdz55xhLvbWpHAa6HPf+OF6jYrSlBFr8xpv8cQX58F0eP0PYfLtht2aO2+u0R2gz+K/bfdhIfiR5jU0/LA7AfrLU0vkgUlxmkdtfK2MuztAnzViZsyBSzEFeZ6iLXbnG5Hoj0ylmmD3gz6ut7EhBYLrjVSuE2ebq9vZmWcUrOKdQNOngJbxq3cH6MOXk2fOy1MzaMSVJkEOaU1cf+KMpk1/tDLzR7U/bAv6jY2NEhoaKq+99po8+eSTsmvXLqmtrZXExERZvny5CgJ2nqjeHJsTQZ8AmFcX7tTUJHpTO4HQgv44OFT9nSwuhroD9EkPQot5aX6q+uEySy9nD5hx2G0L6OPnBPQJTjt/8XIKWHeAPvzAjURv+reXpjumE9rIoDwVVAhE7G5Nv+bYaXXJAKCLkoodUVMeVxutmxlzHn0DLr0I3QX6aPsEJj87J1mWJB+y22v4o/EgDNceawZ9rDrW5mHsbFvQDw8Pl6FDh8rkyZOlV69eCvZo/REREfLGG29IU9PlRflHd+1nf3Ai6JN7Pmx7c3oQplfr4mbXxwPo/2FQqJbetLokugP00V6QzqkGSNGZ1p2x7MgzQH9RYrH8fWKc9g2wlnbtLtCPzKtWaxIZBDVHz3hMv3bkF2PiOVPTgPEiMFnfi+4w7xOH0ex6C9aIeCxydieCar/ZsFuuGxCiFkQz3u4CfZ4RZYAfn97sekNAtzPh/sBSSPrxZ2uyWjQPY9y2Bf0FCxbIZ599JllZWXLvvfdKUlKS8jk6OloeeughaWiwvybkrYnhRNCnUMuqnaVyx9homRFd2KJam7f44u3zZByq14Vl4Oa9QsSwoe4Afa7FYkKLWgLTCPKyO2Eapjc8pYtpiGLMroy7u0AfvzSR6LeNjtLYEWtcgR35xzNGGKa2wYrUkm4HfQABoL91VKR8uGLXjwDBjjzjGX+9PlsFcnL2DXUX6JP1sGFXuebrU4vfWoTKjMVOW+YY2SyAvqYAW5qHMU5bg/4333wjpaWlcv/99yvonzhxQrZt26bmfhf0OzfNqqqqJD09XeMjOneG9h0FIOw8VK9pI0TeFthcowCwaD/6m/7BMnjL3hYaRXeBPpydGlWoZkR4Znei8Qn13AH9qZH7Wwy3u0C/tO6klkj964hIzaMmuM/OhNZIyWICNhGKu1vThzcldSe12Azm6tbNWOzIO2JtyJX/45BQKbD0Wegu0EdQSi+uV9cbgXFWK6Ad+QXo0+b69jHRQvChtY8I47Ut6K9evVr69+8vCQkJcvfdd0tUVJT69cePHy/vvPOO1l22I8N9MSYnavpIx3UnzujiRi17KoDZmQD9tEugT06wdXHpTtDHR/3CvBRt72s1l9uRd4A+Ocz49KdFFbYYYneBPh334NkNw8LVomT39NAz5y5qzjl5+mvSSz055zCvO8z7XAfXGxorpbLX7iz1uQLQYmJ04j+7DjXIF2uyNcjW2lypu0CftQFz+XvLdynwh+W27FrXiVvy6SGUqEawA/S/Wb9bMizVRbmwbUG/vLxcVq1aJc8884z853/+p9xwww3y4IMPytdffy3Z2dly/rz9o0699WSdCPp4vZA4X1u0U15ZkKbmMW/xwxfnYay0uEXTH7YjV/ZVX+7c1p2gH7f/iPr1n5uTokFWdm6+A+jjuqEd8fToAy0eS3eBPoIRRXowZX61PlsKLc+txYBs8p9TZ89rzvnj0xJkfUaZ9hAwQ+su0Mc8vSjxoNw1Lka7yNm94AyaKoF01MDvCdDn+VAQiGDCPrOSNI7FPDM7brEeEUgN6NNR0VpSnPHaFvQx3+fm5kpkZKSCP5o/pv2UlBQpKyvTdD47MtwXY3Ii6Bs+oFHgo54d2xIUzO922RJ5TkOS3wwI1jry1sIl3Qn6LGr9Nu3R4i3JRZc7itmFT9ZxoGXjjqA4D6lzVuou0OeaFOZhjmGutlZStI7HLt8RlAZt2Su0Bd6YWd4joM8YsLwRgEl5W7o82plwu32yurm2gPW97C5N3/CGYD7mGTEZdiZSZ0nNJM7l+017WpQUZ9y2BX3Affv27Wp6Ii//zJkzCvRFRUWyaNEiN2Wvk7Ouu3z6ZnizY4u0dziLy7nzl1O6zO922SIdxxceUdAnj9qqMXYn6FNUg/Q3FmQCvexcwhjQJ4CPMryY+a3UnaBPeiVRylRNoymKnd0iPM+BW3I0KGxLVs+APnO9rP6kYG34ZHWmWrisz85u39MO1svHqzK1FG5Pgj7dHPvOS5X3l+9SK6ZdY/jPnr8o8AnQJzCYniJWshXok19IIZ76+nrBd//BBx/IkSNHPB/MX5s2bZJHHnnEjd63PsUOfO9u0N+x57BKx5+vybZ1gw9elNh9NQr6VExDUjbUnaBP0BD50/jJW7erNeOxyxbQR/sh/QzhzkrdCfqH6k7oOAj0ItXxpI0LztDylIW494xE2ZZdIVb3TXeZ9/FRM99fnp8qry5Mk63ZFdZHZ7vvWHIIoAPECqsvv5fdremTHvrWknR5dnayEDCKS9COdOb8BQ1Ehl8Dt+xtUV2U8doK9PHT79+/X4KDg+Xtt9+WJ554QtatW+f5UJRnwIAB8thjj7UL9BEiOOfJkyfl+PHjuj179urRvfxm3ZfFnnNAbLE4kEHAuRBOrHEF/M7x5nfOw/FWoqogx1n3Mee37tf6u5PN+0iZ/1ySLm8uSdeyn63vzS7/P3PugkQXNIP++NCCFibP7gR9+LElq0IFJRrJ2LkyHw12KL9LdTmi+K3UnaBfe/yMbMwskz//ECaUtq1otG+vB/zpVJd7amaS7NhT2SOgb54TQV4UNpoX3/LZmd/tsqW4DMHAd46NlkKLMN7doJ9T0ajVOumbgBvJrumhjAvrFx1DyQKyFhrjmdoK9Cm+Q5T+sGHDNDf/xhtvlI8//tjz+eKLL2T06NESFham5v62JiXnw03w/vvvaxAgRX0IDsRV0JoA3xUrVsibb74pDz/8sLz++uuaLQBAQ1gctm7dqumCDzzwgHz55Zc6VnMe9lu6dKm88MILaokwxxshgy0Czbfffiu9e/eWRx99VN59913NQkAYuBY5GfQpOEMjGXxhW7MPX+s2e/Q3ipYgydMvYGL4fjl4pPm5M6juBn2C+d5ZliH3jI8RUtIuyZ09yp8rXZx+7M115BMUbK37dCfo0+0vo6Re/jI0XIZsy/3RImcdV09/R0DBz9pnZpKE5PQs6E+N2i9PzUxUi1J7lI+e4h2xNhSt4n2wWuC6G/Rrjp2RcWH7NOuBltJHbdo9lPeB6p6APplINCyykq1An8UV8/OePXtk7dq16rvfuXOnmE9GRoYUFBSolt8WUDKJMZf985//lAkTJuj5pk6dqqAO+CIQGAKQuSaAPGbMGN130qRJ8sknn8i+fftUy6c4EELD/PnzZePGjWpxoGIg4+ValAweNWqUjB07VjZs2CDUGJgxY4YGI/L74cOH9XxUGOR34hUQXhhHWy+ck0GfqFfM1L1nJv4o2Mvw3w5bqsuF5zaD/pTI/Rr9asbV3aBP2VHMcqShUSXQrhoFoE9JWdLPFiQcNOzSbXeCPi4RTPy0jEUjTCi0b3ooVQOJAn96VrKE7a3qUU2fQEI0/S/XZmtr5xYP0Eb/oYIgQvD9E2J7FPRxiSxMLNZxYOGqO/5j5dEObGMto2MovU9oSNW66qKtQN/KMKL3AVxq7qP9A6ohISH6iY2NVVO6df/W3zGjZ2Zmyn333aelezG3o/W/9dZbsn79etXczTGU9J07d64MHDhQYmJi1MSPoHHHHXdIXFyc1vxfs2aNavlkDmApwNWAkMDYMPMPHjxYBQauidbPeD/88EPNODh27JgWF0K7p6Igv3MMY2xLeGGMTgZ9xo+/F02fyml2JSKaKVSCpk/OOSBiqLtBv7T+pMyMPaB9APC31tm04AygTyTzE9MTZGFiz4E+zwkt7OX5aVp0xs4+avqzf7thj74PWJZ6wqdv5jV1KXAh4acGGOxqUUoorNXeCg9Mim3hdutuTR++bcos134iWB6od2BHAvRRFm4dFaWZSNaaI4zXtqCPZkz9fTTnV155RT766CM1h9OA5/vvvxeA9FqEOZ4UP8zt5PVDhYWFqvWjoQOkhtiX88+bN09N8PwdcAf0OQelgBEKcBMgPEDx8fE6tmnTpqkQ8NJLLwmWBPz9gARCy7PPPiuUE6aq4JIlS9RSgIaPIINlgXu8mpZPc6G8vDytnodb4aabbtJrct6ufA4cOCDJyckaLNmV83Tk2KVxBfLC7AR5aV6y1NZ3bfwduW5H9i2vrpV1qQfktwOCZXxwjuwprvTwua6uToU2Mkc6cs7O7ltSWSPLEwrkD4NCZELwXsk5VNUt1+3oeMuqaqXf+kx5dEqszIzIbTFGrHLM346es7P7Fx+uka9W75SHJsfK7Kh8qe/ie9LZcbR1XEFptXy6Ml2emh4v2zIOSm1dvYdHzC8sisy3ts7jjd/3Fh+WD5amydMzE2VLerFteRacWSyvzk+WByZES1bRYQ9vqOWCEsU66g1+tOcc29KL5KW5iVqbwrpGtOfY7tqnorpWonNK5OYRETJ4U5ZkFFa04A/4Qzp8Z8aD9XvOnDkGOtvc/lube1h2QLNGK6cq35/+9CdtvkMQH+CKVg24XosqKipUGycDgBuESkpKFIQ///xzdROY4zHR9+3bV336TCSIiYTvnXHQ5Gf69Ok6FrRzCI1+ypQpOi5M9AQdso+V+NvMmTPVYoHbAOGFYkPEDCCMwDzOdyXgx6qBIIKl4vbbb5ef//znOna6DHblw0vCubtyjo4eO2dzrDw/NVzuHhUqQZHxEhfftXvo6PXbs394TIJMWBctv+kfJN8tiZRN4fEeHlEVEp7Bu/acq6v7JCQkytwtsfLnwcHy6vQwWREU1y3X7ei44dnbcyLk/tEhMmh5VIsx4rqixkZHz9nZ/SNiE2TQ8mjpNTxEPl8YKZGxl59fZ8/pi+O2hMfLqzPC5eFxoTJzU6zEJ1x+F8y7yXzzxbVbnzM2PkHenRsh944Klf5Lo4R513ofO/x/1qYY6T0pTO4ZFSLrSycUAAAgAElEQVTrQy8/V6ywQUFBao3trnEu2R4nr82MkN8PCJZlO2IFHnbXtdt7Hd6FeVtj5MYfguXj+ZGyNqTl+oEy3dl3Exe3z0AfzZqGO5j477rrLtX6a2pqNJgObb+tLnuA97JlyzTtrzXoExBIbIAhQB8QJpCvNegTW3At0CfosC3Qz8/Pl0GDBsn111+vrgXuaeXKlSpoMHExU7UmXADcL+PBLXHLLbdIWlqaCgkICp39YHXgPFgsOnuOjh6XU1orgzbvkfvGx0hsXoXUNjVnP3T0PL7cv7rhmKzfeUg1/RlRBXKgssHDH54FzwDrjy/HYD33zqIaLeDy/OxEic8/3G3XtY6hre81Dcfku/XZ8uT0BFkQV9hijFjHsKy1dQ5v/d5w7ISE762QR6bEy7frsyS3tLbbrt2Re9h/mOpyu+TZ2UkStbdCjp+4/B4yv5hnzLeOnLMr+06NKJCnZiRIvw1ZcuJk59eVroyhrWPD9pTLK/NT5bGp8ZJfXufhDRZa3LFYRto6h7d+L6iol4mh+fK7gSGyJeOQVNYf7bZrt/cejjQek5i8Crl5ZKSM2rFXcsta8mf37t1q0W7v+az7oTD7FPQpuQvw4QvHzM53zON02SOX/1rEvlu2bJEXX3yxhXmfAL0ffvhBMKUZYvJguidIj4UKAmwRNjCtYw5BCMFqYMz7SOPjxo1Tkz4+fiwFmPd5YfHTNzY2qnmfc+JKwA2A5QALAal/qampmgGASwGh41rkdJ8+aUozYg5oqVQKztixS1XjybOyOatCffrzEw5KecPltK/u9ukzFyiuQRWye8fHaFaBHf2tlG8lToNI9OUpJS2mcHcG8nHh5nrjJzQwjeIpyQdqW4zHLv85VHtCO5+9MDdFK0Bac727K0/fyot1GWVaTx7fPkG3dqSYghqtJ9B7ZlKLANue8OmTn0/HvesGhmiaKkWO7EYUgKLQ2C0jI2V82D452Kriom19+oAt5nLAm6h4gHrkyJFqYsevcCXt2Mp8wJlgPDr0YTZDWkHDBdxJ2+MFM4TVgGtxDXz17AvQY1ZHkgSUOeapp55Ssz/R/pQF7tevnwbsEZTHd0z4LHZcm2u+9957snnzZpVEqT3Qp08fFUAAEUAfSwY+f38HfSqArd5ZqtHVpJBUNl0GVPMMenrLy0waDoF89Ii35nr3BOgjdEyK2C83j4gUFmY7FpxpOnVO85afnpWk1QOtz7C7QR+hiLTLt5dmaEAfz9KOxAJMoSoqu5F/3tOgD6C+v2KX9oonGt06HrvwLyq/Wl5ZkCpPz04ShCZDPQH6RPBTwvj6H8Jl4OacFi24zbh6eksqIYXGAH2KZ1GH30q2BX3Mg/ge0Jgxzw8ZMkS1aZMfb/LfrTdj/Y6fvLKyUs32ADrR9JgliAkgwAhTGto8FgM0dQKPODd+evadNWuWpvuxL5o7GQO4APDx4w8hroAYA2IH+B2rwogRI1TbZ9zEImBVwJSCUIAFAWsAZn2sBIA9BYjw07QVlOh0TZ/nQmQ8kjpVwKwvrvWZ9eT3uuNntQEKoL8k6ZA2uzHj6QnQxxpCkR5MdDOiD7QQQsy4enoL6NPk5pnZSbIyrWc1fcMLej1QMY1GQHYk8sw/XZ0lL85L1VbOVpDtCU2fwi2UBabjHgVniPy2G0XmV6s1gvRCa1ZNT4A+vKHg2P0TYzXzgQwIuxHvZVRBta4dlMmm456VbAn6ADYR9wBtVwiNnXNg4v/rX/+qmjpgiw+eQKPFixerJs81uCYmfLTxm2++Wfclyt0AMi8kgH/nnXdqJD1BhgSRGGK/2bNnq+uhV69emt6Hf84UAuJ3Sgg/99xzakEgoA9hBFN/W+QPoJ9cVKe5treNipKCqqNitwKWFE2hRSugvzzlUIt0nJ4AfbRWFhc6ZQ3c/ONSmm3Nme74ncXli7XZCrJYcqzU3Zq+uTYCEumhg7bk2G6OMUYaKuG2eWl+qmQcqm+hWfcE6FcdPa1Nk+hSSN4+Fi+7UURetQpJfeelSEndZa21p0Af19tbS9O1/DQCid2o8eQ5rTmCwjA1slBIAbaSLUGfAaIRozl3hQBygB/zOZo9mj8Thb9jguc7vxvCzM8+Zl+EA5NHz8IPcGMd4HfcDvjvDbEfxxP1zz6ch+O5FsTv7M9YzO9YMdpD/gD6ew83aXcxWtemFtXJGZs13zly7IysSiuV6waEqNbKYmioJ0CfgjOMiRr85ATH7Ksxw7HNFtCn0Q0a2BqbgD551PjLMVmfOHvedrnnBZVHtXkMFq+skoYeB33M1cx7CrmMDytoIezaZaKF51bpM0VQsgPo48cfEZSn5nMEJdyXdiJAH8sqrsHp0YVS3uAQ0EeLJjIejdyANWl6fABPA6Z2YravxuIPoM+LQn32//d9kGoUdqtmVXP0tPqlAX20VmvN+54AfeYSxW9eX7RTXpyfqsFDvppfnT0viwvdz56fkyLr0stanKanNP2Ug80WpZcXpGmnRKv5vMUAe+g/lESlecwrC9LUkoNwZ6gnNH0uH5xTqVorAZCt/b9mbD25DcutkufmpGgwn1Vr7SlNH6vgqrQSBf05cUW2aySGtSZoT6X8bUSkVkG1xifxHG2r6ZMq9/zzz2sN/OHDh2tRnYkTJwofIuKtGnpPTsjuuLY/gD5aIR33AH16sNttcQHkl6UcUk1/bUapVngzz7anQB8TP73Xn5yRqAKTBR/M0Hp0C+hT9hbNen2GPUAfTZCKdzS0wfSKJmsnyqlokg9W7lIAyylvlJ4GfXhDO2K67T00Oe5HJVvtwDu01mfnJMsbi3a2MFX/f/bO+ruO89r7779x17o/tPf23vQmTdM20Iaahh1yGBpoHHKYncTM7JiZmRnEzMwsy7ZkW7IlyzJz3P2uz1bGHs0Z2YIDc86ZvZZ0aM6cmT3PPN9n03cHCvTpLknnv/smJggtwwlXOknIB6LbJKAPG6o1cdqxoE/8m3p8uz8S7oxYO25647mTFO/NYwkF0Kekihjm74dG6aRs7fHsTX31ZF9Qai7POKClOJvzD3ZYvQcK9AmBwGnfZ0aqjIuoFBppOEkA/f4r8zUTnTImswTK0qfKgSZAdP6jCsNpfQvoq0Do4Z2l2VLReNIRoM8xDdhcLLcNj5Lc/cc7UAObr2mgnuOJeHFehtIFm0vkAgX6lDYePnFOWfnIz6A3gJME2m6SgP8yJlaNBXOokuN0LOh3VYkQ31CaF8oSCqDP9ak+elqzXnFZO60pSuOJ8woS1N9uLTgsuPAMCRToEyuMqTiqVivWq3nCM44tkI9wG3ywIleT0oilmyVQoM8x0F73mZmp2laUngpOksKGNum/Kl/DNnCiO8HSx+uG9+23AyMkoqRJ4F9wkuwpbZIX5qbLRyvzOtwDgQJ9dMPi8tUFmfLWkmzZXnTYSerSRkDkGgD63AuELs0S9KC/dOlSra83n1SoPQ8V0Mf1SrcsrDBiTk4S4l5Y1YA+N7G5yU2gQJ94dPWRUxr/RW9OKw9qO3dJ3lueI28uzlJiI/P1DCTo43Ug05vFpdPan9IIBfCCDIcscCeAPuWqLNr+a1CELEje1wFYzdc0UM8JC/adnSbkHJiT0gIJ+hev/CLfbyoWOCrovOckwWAh3PaX0bGyJHV/B68lxxn0oA+3Pcx9oSyhAvp0GJsQWSX3T0yQNVn1joq3MpksTKlT0CceZmYNDBToE8PH6jLi5jAGOkkAfYAV4hRrZ7tAgj6EM1jTj01L1oRMM7AGWn+4zyEQ6rcsV+qazzgC9C9e/kVj1FDLDttR5tF/PdA64358dlaafLYG0L9ebRVI0L/0y1XtxklbadrsOonNkKqf9bkH5c+jY2Vp+n5hUWcWF/TN2nDo81ABfUBiU/5BTYChftTqdgqk+nGdz0vaq6CPF4JjNSRQoM/vA1hDtpdp//U5iXuNQ3LEIzoiAYy/ncUdFySBBH1i1IO3lcpfx8dLyaETjsqFyN53XGPT7y3P1WRW84IkENn7DCQqoUhGu3dCvJaHZtQ5K0a9s6hRwzWQGh12COiTo0TcHA/ET1tLPKzpQN6gtJmmugDQJ0/JbMBwXC7oB/LqdPG3QwX0YfsipgmbFUBW3uicrFdCD4Aq7v3IsiYhSc2QQII+xzArYa8S4ABkTsrgpzSI2ul3lmRrtrChLx4DCfp4bahPJnsZwLBOeubj9PdzqHcBfHrYM+acAProANY2riXuasa/kwQv0tMzUuXrDUUdmCkDaekTeiuoP646IzGTxaVTBGNqVdYBBX0qksi9MUvQgz50uT/88IP5nELueaiAPjfKibOX5KmZqep+hR/aKQI1MEAB6MdUHOmQzBRo0Cc+B20rYIEb8Xpld2C1B5jCIY+L35qjEUjQh5SHfgV/GhWrjGRml3BgNSaa6U08nwTIQ8fPdeAbCZSlj04IvQ3bXqYNniBaMi9GAq0zLGoqWL7bWNyBHjuQoM/iu/n0BZ3H4NGILLtx0zR/6pDyYyx8xj8NziiXNkvQgz7tcMePH28+p5B7Hiqgz4UB+AGJfy7O8qjtDuSF0wzmuFptmUmDDzpVGRJo0KfSgRg1zHdYsbgWnSCAPqQpgBhlVWYJJOjjrqZG/4GJCfLthiIl6TEfWyCf06yF8f/hijwFMLPnJpCgz7UEKB6YmChzE/d6AEUgdbat4LA8QcvkzSWOAX30ceXq1fYukyTzpTonmY8FHOWqgD5ufvNcxnE7FvShqqUTnfWPxjg0wWlra9NGNlDi1tTUBHJM+vy3Qwn0gSuaorw4N0PmJjknRk33M5pTkMxEItiZC9dLvQIN+mR5D9xaKiQNURPslHp9gIJmO/2W5Sjtp/lGCCTocxyQp7wyL0PDIk7ihMC79c7SHM3gp37aKaBPaSOLy0d+TpaRu8qFxkBOEcrPIA4auLVEzEQzgbT0Dd1MjanWXg+jdlUYbwX8Ec4R2oMD+utzGwQyIbM4FvRpTTtgwAD56KOPtMMelLyQ8nz55Zfaapde9rS9hQM/1CXUQB+WKBJgmFzIgnWCMMnRhhLQT65pEXN9d6BBn5jclOhqnZBXZbFy73gTB0p/gP7zc9I07ABVqlkCDfp0j/t6faH8bXy89q13irs64dfe8HhuyLJ2CujDCYEXiYXlF+sKJNNByXyUYD46NUlb2QJohjgB9OkuSYiLckK8mE7wwcE5AuX5HaNiNMxFuMssjgV93PbPPvusdqWjzz3Ndz799FPp27ev/r399tvy008/aZta8wmF4vNQAn2uD3W3dEIjMYdMUydIbfMZmRzdDvrQkpr71wca9AEs3HWPTU0WLArA1gkCl8Fzs9I0KS22wlmgT7iGhdL/DYvSJEPz9Qyk7gg7wLsPSNB/wimgz3FAWUx8mmoMawlmIHVGTgvJv5QT4ro2xAmgTwdAcm0IvVHNAvAHWgD9uUntoM+CyeoZdDToDx06VFpbW691uqPRDr3q33vvPWloaBDa2y5cuDDQOvb574ca6Bc1tKlLmJuFbH4nCC70iZGVauln4EI39RUPNOijH2r0WSgRPzc3Awqk7qj/xTL8cGWuxtDNxxJoSx8relPeIblteLS6OpkInSAkiZIlT8Y31Q9OAX1DN99uLFIAW5Syz3gr4I8kZf5jSpKM2FnuONAvOXRSBmwq1vug6OAJR9A+QzRGxc8dI2O0uZmVitqxoE9v+m+++eYa4DPyaE8bFxcnffr00cUAlv/ixYsDPih9fQChBvpYOLRkfXlehuwubvS1+rq0f/qcw2+Pe594sPlGcQLo01qXHt4PTU7swErWpZPz0UaAfp/pKRqfjq/q2Fc80KDPog0inNtHxmgYySnloVFlTVqJAeHSyXOXO7iDA5nIZwwR8loIvY3dU9GhssD4PBCPG3IPyt8nJcro3RUdFrxOsPRZgE+MrFJPBAtzJ4TeCNPMiK+RP46MUS4Ba9Mpx4L+xo0b5cMPP5TJkyerdU/8fsWKFfLjjz+qm7+lpUUXBStXrgzEOPTrb4Ya6BM/HLmzXDnlie87QSAnoUc2oA9YwFJmiBNAHyti0LZSTc4pOtgm0IAGWli8PT4tRQldSH40S6BBHzfr/pYzmo3+2ZoCxzRFgdseimDa65JVbXYGOwH0caXjUfpuU3GHslXztfX3c1pdU1UwLsJ5oE9TLLwieCJ+jqlxROjtYNs5mRpbraBPmIb51iyOBf3q6moF+SFDhsiECRNk4sSJQjIfrXVjY2PlzJkzEhUVJcXFxebzCcnnoQb6XCSIcF6am6HMaU64aBVNpzReDujnH2jrQBHsBNCHyAUegVuHR2tOhBPi+nB8Pzo1WePTSdUtHS5joEGfg8EKo+0vHAfkkThBoJTlmD5bW6B5I04DfciDKCmEKpiQlxNkbXaD3D8hQSm8zSyeTrD00Q9dOQlzsbg05xwESnewi5LPgqXPeLPmGTgW9FEY8fykpCShqQ6x+23btgld9ajDDScJRdAnTvfmoiyN7Tsh6xX3LzFDQL+w4USH1bETQB+GQJJySExjwdRw/FzAbwFAHwsHVzX152ZxAuizMPppS4mSQa3KrHcE4QyWF9wGn68t6BBCQndOsPRh5qNdLAsT+CqcMNWSIX/vhARNtDUn/joF9OMqjior5VMzUqWh9VwH7435nvDXc+YGQg63j4iRPSWNHuPesaB/5coVIXHv2LFjgiuf0jzjz5zc5y9FBvJ3QhH0qVf+aFWePDMrVU6cv+yxGvW3vinxgpHsf4dGCtztV0xZuE4AfRZGKbUtcvvIaAWyyqbAUxgD+n+fnCifrsn3aJXsBNCnwx6lSyxMaB1rztPw9/gyfo+478vzMxRYreWqTgB94r9DtpfqfQlZjxNAH1Y5Si+nxtR04Lh3CujnHTiuJFBY1tyXVsvauPb+emThNj6iUkE/0kKaxTE4FvRra2tlxowZct9998mtt94qv//97+WWW27Rv4cffliOHz/uLx0G/HdCEfTLG0+qax/3cEF9m8DLH0gpPcTxlMktQyOFBYD5xnUC6KObgoY2TZyjpCr3QODHf8uZi9ox0S5m7gTQZ0zBH/Dk9FQZvqPMEd4RiGaImX+5vrDDGOP6OgH0AfnpcTWabzNmd4WHlejvexSvLvzxfx0Xr8dFVYYhTgF9o9z398OitIqFzpiBlAOtZ4VrR51+lA09sGNBf+3atfLxxx8LiXq7d++WPXv2XPuDuOfSJWfUKvvj4oYi6MNbzeRCgg7JQ9amEP7Qq/k3aJgB690tQ6OE+L6ZzMUpoE+yYf9VeRpHtybOmc/FX8+ZgOnMhquaMkezOAH0SWAiLg37I8fohIUSZYQvzM3Q6hVrmNIJoM81bCecydSxZr4PzNfXX89ZfK/IOKAdE2fF1wreJUOcAvrcB2uyGzQ0uCx9fweqYONY/flIHxH4POiyZyXN4jgcC/o00qEkD4ueUr1wllAEfbJe6QT14KRETdAJdO052fE/bilR0AdczROyU0CfWB1uu7vHxsnWgkMBd1eTVEX7WlqekgBmFieAPoBFhjzcBu8syfHoBGg+Xn89p/zshbnp8s2GIo+fdArox5Qf0YZAL8xJV/0FEvjhtwdIce9D2+1E0CdsRP4DSbajdpVLdYATICGmGr6zXO4cEytcS6s4FvRh5Pvuu++kqKhI4NdvamqSI0eO6B8x/nBaCIQi6DMQyahuJ3fJC7jrFZKgAZuLFfSxDs2xTKeAPhYF5Ut/GRMrC1PqxJzJbL2xff0a/ZCpfM+4ePlqfZFk73Me6KMDFm+Un706P1NLq3ytl5vtn+tH+9rvNjoX9LkXfthcrLTPtUdPdyhfvdn5eftz8h5go8SjtCClTlkMjd9wiqXPvUCIkgoDPHH59YElHKOPyNDtpWocwBhoFceCfkREhPTv31/effddzdxfs2aNrFu3Tv927NghFy9ed/NYTyrUXocq6JPx/e7SHHnk56SAd0IjXv79piLNjrc2G3EK6NM4A+IgrOvRu8s1aShQYx3r73DbeQV9rNac/R1zDJxg6Ru6oacCRFDEOQNd94PrHND/fpNnqbFTLH3qvKGkfmBSgkAmZG3NaujVH48Xrvwii1MB/QRZnLpPoH42xCmgz/GQwEc10nOz08Ravmocr78e61rOKKcHeRDQPlvFsaC/fft2pdvt16+fxvY/+eQTMf4GDx6sdfrWkwnV16EK+rjUsa7by+TaOmTM+/tasjqHgvQPw6M9Oow5BfSJb0InS8Y8dd503AuUcCzUA2PpY0lb4+VOAn2sa7jRaSRjJSrxt/5WZR6Ql+amqyVt/W2ngD5shmTu3zMuTmYn7O3Agmc9Zl+/5lggv7lvQoIsTd/fgfzGSaBP8tzgbaWacEhtfCBDInXNZ+SnrSXytwlBBvq48nHt2/3RWpeSvnCRUAV9Yk/T42rltwP3aD1pIC0KQAuLFb5qbhqzOAX0sVLR0asLMrVpy46iwBHOAPpMdHePi9OFm9Wl6STQT65ukX7Lc5V0puH4WblqKsc0X2d/PAdMsfThD7CKU0AfwIJPAA8c90QgOSHodknZ5X0TEzShz0xK5STQJydpSdo+damTg0DznUAJnkrCM/dPTJB4C1Mmx+QoS5+M/JycHGlraxNK9iIjI23/IOxxs/d7NqRYTOXm5nZIVOvZnnr/Lbjbt+Qfkt8M3KM3NpZjoAT3NP0AyHjFPWYWp4A+x8QkiGv4+TnpGus0H6c/n1/55aoQOySpkARIwiNmcRLoE5cGvCiVS997LKDWPvFpjmPQ1lKzuvS5U0Cfg4FHA4Ie3NXoL1BCIiZU3QAY9fo0KTLESaBPmR4VNSQcTois8phDjGP2x2Pt0TPy3cZieXBSgtDK2SqOAn3IeJYvXy719fUSHx8vdNmz+4OKl23DRULV0ieDn0kY9z7EONTuB0qy9x3XLPS7xsQpmJmPw0mgT6YwpUuwf0G1aW2mYT5uXz7HTV7bfFpBf+DWEiFUYxYngT7ekVG7KzRpFEpXxl2ghPg0oD9ke5nHITgJ9PHcGItgri2enUAIYDo3ca+W9q7LbuhgQTsJ9LkP8RBCBEWSpjXHxZ+6IxGZRS7NuWjUZRVHgT7JecnJycrCV15eLps2bbL9o2bfTeSzXsquvXaSpc8RVzSe1NUxXN/Wsq+unZF3tiJBjlpukuSwYM3iJNBnctlZ1CjPzErTGGKgMvg5juojp+SusXF6HMUOBn0y+Gcn1EqfGala8hhIIqiFKfsU9IfucDboY91jseKFo4PiuUuBCafC30FewQOTEmVjbkc+DyeBPvPFmQtXtEMhycl2THjmOcWXz+kY+tW6Qnl4SpJtUqGjQJ+bEzCnHO/y5cty/vx5278LFy44wj3tywtn3neoWvqcIwD75uIsvVnsakrNevDl84y6VvlsTb7cNyFe9h3r6EVyEuhDD4zl9frCLO3UVtTQ0cL2pY7M+8ZahsQI0B+6vUzoK24WJ1n6HJeRzEefAKogAiXzk+u0koA6aqs4ydInRk2lwW8HRigjXtPJC9bD9ctrYvgz42uVz4PGNvSgMMRpoI8XDsOB3gqrsuqNw/T7I6D/xdoCJfFKtvTE4GAcBfpm7XBB8/LylIp3+PDhQsb+oEGD9G/SpEmue9+srG48d5qlf/jEeRm2o0xX8jDzBcqNSJgBQKBMiQRDszgJ9KkJJhcCz8g/F2VJhA23tvnYffWcCQ66YkB/+I5yfW7+LaeBPsxk7y3PlWdmpsnxM5c68DCYj9vXzyGYgXt/5C5ngz6WfVJNs4I+LW0BkkAIZDwz4tpBf1vhoQ7lg04DfRbCk6KqNIw0PbYmYBn8kIvRD+PxackejbC4ho4Ffdz8tNIdMWKEttalva7xB1vfuXOBS/ry9+APZUufnuzQbFKHS8JOoLJeKX/7eFW+xsGcDPqMvSu//Fuzc6k9p3Y5EALoQ10M6I/cWS50KTSL00CfpigDNrVnNANggYrrE2YA9EfvrjCrS587ydJn8V16+KT8ZXSsljoGKkYNIdW02BotU91ZfNjRoE+eCx6lZ2elqSFDI7FACJwBH6/Olyenp3g0wuJ4HAv68+bNkzfffFNiY2OlqqpK9u7dK3V1dfpHoh/WV7hIKIM+rlYA96HJSUphGahM4dTaFum/Mk9ZyOCuNouTLH3juCCcISEM2s9AyPnLvwihhTvHxCmAWbv+OQ30WciR+MgiJbbiSAfw8Kf+cFXDGTB2j7NBH52gMyiDyeKnfWwghDAD3fWYH3aXNIm5mY3TLH0WShl1x1RfUFMHyjvCvfjRyjxN9k2r9eTycCzoL168WAYMGKBufCZd4vzmv0AMwED9ZiiDPqvjxpPndVX65boCvWkCoWfYAblRHpuarPXn5mNwIuhTvsRk/MmagoC4EUmGI8Mb0AfAcCmaxWmgf/LcZVmdWa8lmYtS6qTp5Hnz4frt+bS4GgX9cRGVHr/pJEufgyP0BvESnTAprTVTU3scvI/egOqZxdpDU5IksqxJaJdsiNNAH34Dyo7fX5Er7yzN0RI+41j9+VjReEre11BWqi2Bl2NBHwt/7NixSrsL9z6Nd06cOKF/XGxzQxR/KjQQvxXKoM9EAvADYLSMhRQkEAKBC3FyXGL1rc639LG8iFGT0Efc09+5EGcvXZHc/a0K+jQBokzILE4DfQh5yKi+Z2ycDNpWErA6aqxW+gCQGW8Vp4E+44pyOdjwCCPBjudvgYFyUnSVlsKR6Hv6wvUkTKeBPrphLvt2Q5ESMLEwD4RQ+txvWY4mR6fXBZGlv2TJEnnwwQflsccek9dff13eeust/Xv77bfl66+/ltOnO04ygVCuv34zlEEfHbJChgI3kDFqyEhYHT89I1XqWzvmizjR0ieJDpIeyFNo9uHvGDUkQZQ50vxnYlSV1uyb7wengT7HRhgJfb21ONsj8dB87L58Dqc9oE/Cl1WcBvoQ4wC0uNYnRFaqFWs9Zl+/PtR2Tk9RVhsAACAASURBVMcX9e90snM66MNmwIKO0JvdNfa1vtg/c8O/lubI83PThaokqzjS0seKj46OlmHDhsmsWbM8/lasWKGlfNaTCdXXoQ76WPsk6/SdnSbjIyv9brUybmCuYnVMEk5DEIA+CU5Y2LheKWXydxkaNclUPAD6k6OrPPoVOBH0KXX8eHWeln9l728NiLuaBRKgj8vaKk4DfRaSe4+eUe8Xi/K8AHSP03bSkZXy8M/UnDd3GOdOtPS5ptDw4rmE3CgQIRFA/+3F2brwyLR0v+T4HAv6UMUC/L0RLDTa8GZnZ0tcXJykp6fL/v37O00CBFwzMjKubdva2nqN4x9uAMIMiYmJ+jnlhJS/GUIvABINU1JSlE2Q3+L71hbA8BCUlJTodg0NDV0KU4Q86IsocNG/mwxrCDn8LVgR/1qaLc/PSfPgGneipY8bEU5y+NGxKPytM6xA8iDI7iapcF9Lx5CIE0EfTnLyD24bES2U8FGB4G9hoUbvhKmxNR4/7TTQxwN37uIvWnfeb1luQMpDCbWNjajUcc54My9unQr6lNFC0ENJLfcJevSnlB46qb8NXwDeOKs4EvQ5yGXLlmlNfm+IeM6cOSNbt26VN954Q+6//3554YUXZOrUqWLNCcCzABHQtGnT5MUXX9SwAtvC/EcuAULFwPz58+Xhhx/WfdHyd8OGDdcWBSwApk+fLk888YQ89NBD8swzz+jigN8yCwsDugX+/e9/l6VLlyoJkflzu+fhAPo5+1s1pv/RqjwlfbHTgy/fowXlO0tZHacHBeijC3gN8I7AL4Dl708hixrLC9DHS2Mtc3Qi6B85dUH5228dFq1los2n/Kszrg+LDkB/epzzQd8YTxDOYLlSWutvYVyN2VOhVTV4lggrGeJU0Cfc9vWGIg0lwVrpb6psSmlfX5ip1yzb0vIa3TkW9LHMp0yZolZzTyh3AfJDhw4puK5atUpL/rZs2aJgXFBQ0KHOH8CHF+DTTz+VlStX6rZQAAPchYW4aP6tjX8ef/xx9RawAIAz4KuvvlLPAdY8IYeRI0dq4iHAzue8zszMNMaoPo4fP15effVV+fDDD13QN2kGxi/axTK5BIKZL7biqLy1JFteWZDpEbt0oqWP6kjmw6J4YnqKttw1qdPnT+Gzj69s1mx4AMya/OhE0MdKTK09Jn8YEa2hkUCUVFGfzxindM8qTrP0jeMjRk2Dp0lRniEJYxtfPdL8atSuCg1jZdW1CgmkhjgV9Kl6YKFC6A2r3+ydMI7dl4/Fh07Iq/Mz5I2FWbY9ABwL+nPmzJE//vGPcvfdd8vTTz+tVvpLL70k/PXv31+t9RspjkS/1NRUefLJJyUtLU278gHg33//vSxcuFAaG6+3JaWr37hx4xSo6fJHBz+4///6178qTwAgz2Lgtdde03ABrvzt27drzsHmzZvVWmcBMHnyZMEqxzvBb7333nvqDeA4WViwbwiHSEQcMmSIC/qmC3jxyi/ahITJZUkACGdYaEAHTDb8wTbnJ/KhOkrmCIdQNld++KRfLQroUFko0ZWQBkDW9qtOBH0ojOGB+Ou4eG2uZOf6NA1JnzyFiQ/Qn5UQPKAPHS/1+nRTNFvaPlGQZaeEZCB/AkBpf23um+BU0CdsBNEYCZDcG/4mHCN3hURC5jN0ZhXHgj5APXHiRC3bwzqGehdQ5XHBggU3TeRj1YxlT7Z/cXGxnjcW+MyZM5XSF7IfQ5qbm+WDDz4QKgYAbQQvAa58wgNZWVkCWdCXX355zUNA7B9PBB3/8ES88sorum8AH8uf0AKhAo6V1ywyBg4cKGvXrtUwwZgxY1zQNy7Ar48z4mq01GWEDS+5ZVOvv4wqOyJvLMqSNxdny6G2jjXcTrX06VswNbZaY9QsWsy85F5XkGWH5BCgM0CfhiiHLG2RnQj6nELTifMKYEyIgWiKAuc+rtc5idfnH0O1TrX0iaVT8dB/Vb5HMyrj2H31SCkoNM+PTUvWRa65bNCpoE8If0PuQSXHodsdBEP+FECfRRol0DBRWsWxoM+BApYAKvX5JMUdO3ZM/7DM+exGcvjwYcGtT/wcqx3BYof0B0sblj9DiMdTFrh69WoFe94HpPv27Svr169Xa58qArj/sdgRQgTE8LHcAfpnn31WqwyMffLIe3gsyAuAdwDLnyQ+cgVuBvpNTU2Sn58vSUlJGjL405/+pEmEhg56+lhTU6MhChIce7oPX31vQVyZvDQnRfotyZTmlhZp+fV6++r3zPvdkFEjL81OkdfmpUpJ3eEOumFRyCK0tra2w/vm7wfieU19oyyMK5PfD42UGdHlUrqv0W/Ht/dgk6zPqJY7RkbL5D3FUmrRGYm4TC6B0MuNfrO6vlE+WZ4hfaYlyaKESmlpaZ9TbvQdb372/fo8eWlOqkzZU+KhG8YX44zx5s3f7O2+cqvq5b0lGfLK3DTZnb/PrzrLqmyQAevz5OEpCZJUul8ONR29phvmc1qwk2Dd23P09vc3ZlTLK3NS5NlZ7fOJP+eyxJID8tT0JPnn/DSJL97voRvuTRblPTnnzz//XI1WM87d6Pn/u9GH1s9wsQNMTByRkZGya9cu2bZtm1rexPtvFucHtNesWSMfffSR7oP9HzhwQF37uPirq6/HpwD9f/7zn+rCZwAhfJ+Y/saNGyUhIUFmz54tP/744zXQJ3ufxD8IhAB9FggsDIzFCNbhc889J/QJYFt+E28BoM9x/fDDD7poYEHAtlbhHPkOC4dHH31U/vM//1OTG/E69OaP3AWqIsg16M1+fPHduTtS5eUZsfLwhCiJTU6X9Az/HeP0Lcny5ORoeX5qjOxKSO+gG3QVFRWleR++OO+e7hP9zN+RIrcOjZT358XKxui0Dsfd0/125XsxyRkydXOy/GFYhAxalSi7LTpjkUulS1f25c9t4lMzZeCqJPnr6CgZsDxektMy/HqMH8yPk6emRKvOrOfNvck4c9q9mZyeKR8siJeHJ0TLyLUJkpnZuznIet43er0hOk3Q2QPjomRtVKpwLMb2LJAiIiJ0oWS855TH5XtS5a05cXp/bIhK9etctjIiTf4xPkpemBYjKyNSr+nL0A33JphmvO7OY79+/XwH+rjkR48eLX/729/kt7/9rSbkYe3ecsstwg+fPNmxlacVNPEMYFGTA4BVjgD0RsiA0j1DWFwQkweUDQ8A4H/PPffI7t27ddGwaNEief/9969198MCJ1mP77AAIYyA5c9xAeKs1knYI0Of1ShEQ7fddpvccccd8t///d/yH//xH/qc75tL/4xjYtFDiAAvB7ogv4DFAy2He/NH2MLIW+jNfnzx3aL6VvlxM7SfSZK3v0VOnbvQq3PtzjHuKDwoL81Nl37LsqW+5VSH3+X6UvaJ96g7+/THttl1LfLEtGR5e0mW8Nwfv8lvHDlxVrbkN8gdo2JkXmKtHDx2usNvM2YJp/nreLr6OyfPXpDdxYfliWkpMmx7qexv7nitu7qfnm73w+YieW1BhsxPrPXQDeOLccZ46+n+ffG9S5cvy5ToKs1GH7WrTC5d6t0c1J1jLGpolZ82F2vHuOKGVjltmhMwmJiHmSO7s09/bFt5+IRMiKiQ/xsWJbHljXL89Dm/HSPzQJ/pKfLeshzJ3ec5J2B4Et7uiR6++eYb34E+rnbc8Fj3d911l7rlAWAS7rDeiefcSIxkvPvuu0/Bn+0ZICTj4TkgRGAISX/E2r/77jv9zIgVUeZHvT2fk7BHOR4TGS5+kgFRADcpiX0sAHDZkzzI99etW6efs3JnAcJ21PDzRy4AOQRDhw7V/d3MaxHqJXvGdWimuUZsjZKnwPXtzxj17uJGTX75cGWeRya8U2P66M1orkHPAEro/CVQtG4tOCR/GhWjFK3WuKVTY/rwG5C1T8yTpih5B67PA/7Q3Q+bi+WNhZm23RGdGtNHL/SHJxeB8lB/lp3T6e+nrSVaoVJFd8TL18O6xjyNceQ0oYQWGt5bhkZpqaM/ez2Q4IshwFxW2HDCQzWOjeljAQOquNlJqAMsz549q6BM3T2ruxsJZXbELIjDDx48WAGZRyx6YkFMSli8xNFY8QCsgDBNfnDZ//TTT/odQgLsC28BxwNLIImFPCeJj8UDn+OSA/i//fZb/ZyFCWV8RmKg+Vhxr98spm/ePlxA/+LlX2RlxgFtowlzmRVIzDrx9vOdxY2aRPjxquACfUrlxu2plLvHxsn2wsN+I5xhUtuUd0gt/aVp+8Va8+5U0IcoBcIUaEr5211yvYrH22PKbn9QJ1NKhc6s4mTQh8zogxW58uLcDL8SzlB+RtUAPTFqj57pQDftZNAngx/uj9uGR2sXSmtvCuu19+ZrkvcwAmivS1KfVRwL+rjFyXYnoQ2QJwEP4Ccb/vnnn78p6HOiWPu4MgBpEhCom8fK532S5GJiYq5l9gPcxIfY5osvvpARI0ZIRUXFtRg+4I6nAFD/7LPPNJ6P9W7IuXPnNBTAMbOwGD58uFRWVuqCwtjGeCwqKlLPAXEVu3i+sZ3xGC6gz/kCvk/NTNVVKh2r/CU0+qFc8NM1+Zrhbf5dJ1v6gO+6nHolyVmYsk+aT/snU5jfJUMZ9/6y9AMe5EBOBX2uK/f6dxuL5NUFGbYWt/nae/s5dLYwtS23IbpxMugXNLTJgM0lWjoHgPmLzRDQGrC5WPrMSNXKATPRjZNBH28IJD0PTEyQ/qvy9Lm3x1Jn+6NMD9pi5jIWTVZxLOjv3LlTS+CId8OERxyfsjhi57jh3YY71kvZtdfok+xNJj4nCuVBEM5AL0tNtb+OEisZdrsv1haI1RXnZNCH+COzrlVrz0ftLld3vz+uK5b9uuwGuWNkjKzMPCCtZzqy2zkZ9NEP1MHUMo/ZXeFXd/XX69tBH51ZxcmgzwKcZkEPTErUFreQM/lDWGx8t6lIQR+vVrCAProh9EapI/08kqtb/KEu/Y2c/cflocmJApMi7HxWcSzoE+cmdk42PI9Y1Tt27FBruysle9YTDebX4WTpFx/EnVcs/zskUgob2uTKL/6B/W2Fh/XmZFKGHdAsTgZ9WurSfvTvkxKV0TBtr38mF0IvxCz/ODJG473Hz3bsl+B00F+X06BdHVnkAST+GWWieQRY+nZtV50M+lj2eCfuGRcnMxNqhfwbfwiualrVPjUjVZkyyckwxMmWPsd4oPWsDN1eKnePi1MPpr/aXwP6D05KlC/XFQg5EVZxLOhzoFxU3POUrxFbJ7uVrHtuDqM0znpCofg6nED/wLGzylb2X4MiNN7qr0YyJKU9MytNW/weCSLQB6xIeIRJEKuCMIU/BB3BxQ7ow9hmvU5OB308Su8tz9W4PpOzv5qiYH29uShLdWa9Tk4GffTD2CJW/NW6Qg8GRuu5eOt17v7jwkKckB+L22ACfRbGdNy7a2ycLEnbL9aFsbd0ZN0PHSTvm5CgeqPjnlUcC/oAHRnzxNgNUhuyNEm+I8mPGHq4SDiBPjfGtsJD8tuBEUplaaV39dU135x/SCcWaG1pzGIWJ1v6HCfUqD9uLtHwBJOLPwRvCBMaoI/VbHX3Oh304XQnvv7ivAztFmgGE1/qjzgrbIDozCpOBn2OleoQWCufnpkq/kpMy97Xqt4RfhMQvRJElr7RifJv4+O110Nt82nrJffJa+il+U3Gd3mjZ5WbY0GfmD4Z8mTU//nPf9ZSONz8lO1BenOz7H2faDNAOw0n0MfVmll3TH4/LEoGbysVu5WqLy4DHev6zEiRn7aUeFQNOB30z1/+ReYm7dXjp7c9VRC+lsaT5zUJ7vaRMZrQd+pCxxiv00Gf4yWeD5gQX6d3vD+E6hBAnyRIqzgd9ImvQyvLQo8kNX+4q8lX+WJtoTwzK1UoE6V/giFOd++zkMRz+fCUJO26569eD5n7WrWah0qRimACfaNkjxvhkUce0Xp5ks8guqFe3lxnbwyCUH0MJ9DnGpIAQyIK5BIZdcf8clk35h6UJ6elyKCtpUEH+iyU6Ob13Ow0Gbi11INnwBcKxNVKtQCgT+metZOY00GfuWRu4l4Ffbq4mRu5+EJfxj4/WtkO+iwyreJ00K9rPiNToqvlNwP3aCfMMxeud7yznou3XtNOl+6bhN7azl7qsNBwOuijA7xwcELQtnuXn8pDM+pa5c7RsfLD5hLbxF7HWvqU5lEzz40ADS0kOdTTU3Lngn7PbymnZ+9zZvTQfn95rk7INHXxh6zPbVDWr6E7yjySlJxu6WNxsaInQezjVfl+IZyhfei85Dq5fUSMQKRk7b7mdNBnTGFt0/GOsWZdtPhqzL2/IlfeWpylOrP+htNB3yjTBPQXp+6Tw5bGVNbz8cZrWiF/sqZAk2wJIZm9C8EA+njhvtlQqJUiy9M9Kza8oSPrPjL2HpM/j4qRgVtKpKopiNz7ENjg2ofEhha7EOFAc0sNPCQ4EPWEi4SbpU8ntPERlZqMQuzzkimO56trzu/QvnPEzjJpPt2x/MzpoE/15anzl7ULGsQv/kjmO9R2TjvF3T4iWqh8MHc/4xoFA+jHVx5VPognpqdIy+kLfknm67cM0M+WrQWeCZdOB30y+LG8/3tQhNAJ08517O37k4RLSGbwYrEwMydcBgPoEzaaHlej1Qd4Sa7cpFGcN/THNSIEM2hbqcBiaBXHWvpk6gP8UNa+/PLL2qCGjnbU7DOhQH0bLhJuoI8bD/cnGagz42v9QjizNrtBuQFG767wIJpxOuhzH1y9+m8ZvK1MLQq7tq3evleo2+ba/GFEtC4yrGQtwQD6lIeSN3LvhHitZ8Yq87XAAgjowwthFaeDPoAL0JMkRpiCeLuvheTBj1blKeizsAw20Ceub3CADNxa4pe5LG3vMfnD8GgtF6w+GkSgz2BiJQdzHu1taZNLdyAodMNNwg30ia/CxPWPKUl+S+aj7IzfG7unMihBn3uCvvYvz8vQvARfcy9RVTEttkZBf1dxYwfSFI4lGEAf9/SshFq5c0w7hXGrhWvAF/MMPc75g3nSKk4HfY6X8sY3F2Ure2VkaZP1FLz+OqGqWb0xEGddvPJLByKlYLD0CUcUHWxrD71Bi9vg+14PqbUt2uhn+M4y2yoLR1n6WO8NDQ1ai089fmd/dMDD+goXCTfQ50YhfkdmNRZFfJXvG8lAlvLQ5CSZEFmlWcLmsRUMlj7HC9cAcf1+y3M1g9+XwN/Qek5Z7bD095Q2dqif5liCAfSxHDflH1QKY7rIHWzzfRnwm4vbQd+O8z8YQJ9SzZG7ypVedm12fYcSOvM9463nhGDIueg7J11/yzymgwH0Od7jZy9qiILQGwtkXwq/R0iERj9cp9pmz2ZEjgJ9ksxon0vv+Bv90WaWznXhIuEG+lxXgJ8GH9wodjXN3r72lG3BYjU5KnhBP3PfMeXbhl6WOnRf1p5DiTopqkpuGxGttKzWWGUwgD7FX1iSeHg+W1Og3fe8Pa6s+2M8Y+lTbWGVYAB9Qm/rcxrkgYmJGt6hjM6XEltxVMiDoC+GlTo8GEAf3QDEg7eXahb/vKS9vlSX/haUv4A+ocq9Tgd9qHczMjIkOTm50z+a3mRlZdk2svGpNgO483AEfWJ3lFKxwicRxnrDe/tyrMg8IA9MSlDr1TqRBYulX996TobtKNNe2jHlR3xahoabF68IXcTowGaOtXJtggH0OU6akxBjp6GLXRtSb44zxjDVApRvRZYFJ+gTeqNGn6RX8iHsyF+8qTPGFqBPdz+rBAvoc9yEkYzQm/U8vPkaY4mF7O+HRsm4iEpd/Fv37yhL33pw7ut2DYQj6LM6ps0uK/wh28s8SsK8PTbgFb9vYoIuMFrPdOSRDxbQp2yO5DomZCwKMvp9JZRVjo2oVNCPqTjisSgLFtCHEwJuhj+PjpXE6mafVYrgVWBCfnVBplL/AmZWCQZLH3IcmisRY6d7HADjS4ksOyLvLsuRl+cHN+jD+AkpE9TPJL0yv/lC8LjR0hfQZ1G+r8Xh7n1fKCEU9hmuoJ9c3ayc8pBzADK+FChl752QILPia8Wa0BUsoI9+yE2Ap5zOZNbFizf1t+/YWXUfEtOPqzjqsetgAX1KDxcm1ykD5IbcBp/xozPJQ6IE6PdbliO4ra0SDKDPMePVIc7uj9AbYRAqHtCbVYLJ0qeMDgrmF+akC6ExX4Xe2C+LcECfUKXdvOla+taR5MDXYQn6IrKv5azWnuN+9WVrSiYxOOspEaTczdoYI5hAn4xqksWYXEi68pFBoRYEiULUA5NsZZVgAX0SRgmFEKag3S65EL4QQB8L75X5mWrtxVd6WsjBAvroZ+TOcnlxrhF684XG2vdJwiOgT0MpqwQT6GNxE3qDE4L7xVcMkPACRJUf0UUs4xkaYKu4oG/ViANfhyPocxkYwD9uKdG4vl3/cW9dKlyWMIzh3l+QXBfUoE+DErqg3TkmVmqPnu7AYOYtfbEfwHHYzjL506gYWxdvsIA+rZurj5yWP42KVV55Yvy+EBaWhF9wU3+wIk8SbdziwQT6y9IPKEveoK0lHuWa3tQfpY3kQFCVYpVgAv1zl65oS2KShRck79POmNbz8cZrFpZ4R+hdMj22Rr0K1v26oG/ViANfhyvoYx39HFutFsWYPRU+uzK4xOCRv39igixK3Sdt54Izpo+CAHqy6mHKS6g86rO4PlnBQ7aXKse3HYAFC+gzxqjXf3ZWmryzJFuTEn0x0AB9cixI5vpwZZ4kVbd4/EwwgT6eCoAYlzVJnb4S2CUBfSoerBJMoM/1p8SRpl508oTS2BdCGerukiat0ye/x65LqQv6vtC8l/cZrqCPGolR49rrvypfWee8rFrdHV3psPAfmJggS9P2BzXok2QF0dCtw6JlSeo+nzXfYXFBAtzdY+O05ar1ugQL6HPcx05flK/XFyovBOWhLAS8LSTxnTh7SRkTGcvUUlslmECfRd8nq9vbBEMG4yuB4pnFGC5+qwQT6HPsVGyQB0EDHl/1LcCbhHfk/4ZFaajSBX3rqAmS1+EM+mSifrgiT2PU1i5b3rp8rI7Jdn9gUqKQxX8iiC19ksXQGW53LHFc174Q+qkTevnruHhbqzWYQP/E+ctaUvXIz0nKash48LYA+izIKD0DLKFKtUowgT48+Cz6yB1hYe4r2VJwSK18yvasEmygT2vdr9YXyj3j4rURDqElbwvXBdpfQH9eUp0t4ZRr6Xtb6z7YXziDftnhkwJn9ePTUqSo4YRHYxdvqJvVMQl8xNtWZQL6HUvdgimRD33k17fJ49OT1S3Kc18Ii4kBm4u14iHZxmoNJtA/e+mKxkGfmJaiLGa+YOYD9JtPXdBQFS5xu5bRwQT6jKmpMTXSd3a61oN7H77aR+3mvEPy9uJsrRawjuNgA30WypTRAchJVS1y2getiU9fuKzMnPzGgpQ6OWTTCdEFfetIcuDrcAZ9Yl9wvAPINOGxArI3Lhc3H7z1f5+UqHE3MrrNEmygT+05LuSHpyTZWuHmc+vpc7p3fbexWPMg7FzVwQT65HQQrsDt+vnaQsnZ7/1kPpJFm06e//U3CgSrzyrBBvoaeluQqfX6vgL9jbkHlTyJPAirBBvoM5ehM9jySIRsPHHeekq9fs3ctTn/oNw6LKq9/bHNb7ig32s1+34H4Qz6l69cVev775MTtRnO0VMXvK5wEqxIeuE31uccVN5/848EG+jDzIdFQQY/fPy+6B5XdeSUZruzULKL6QYT6MOUh7fnWge8Is8OeObx0JPnuHKJ40I29cW6Qsm2WVgEG+jDNUAlAufEPYQ3w9sC5S8lu7TXtUqwgT7VSAlVR+W24VEyfEe5T0JvhCY35gH60bI0fb/twsIFfetIcuDrcAZ9Lge1uvTTfndpjm02am8vGd6DGXG12nBnU94hj4z3YAN9aIRZ7f9ldKzmKjTZrPZ7qzO8CV+uL1Te+rRaz/h0MIE+uiB578fNJfLK/AyNhfZWP9bv400gqQpaaeK6dqWBwQb6tCb+aUuJPPJzshCG88XikpbXMNkRErFKsIE+YwwK44d/TlKuhrwD3g+9UXm0PrdBbh0eLSsyDihXh1VvLuhbNeLA1+EO+iQ9kcz34KQEwa1M+Ys3BTIeQgg0XsEyxmoxS7CBvvKjN7Rp3/Oh28u0V7z5fLzxvKLplHy+tkBIfoNtzCrBBvocP94emhUN31Hm9UoRQB+iFGLgX28o0rwLq86CDfSPnLyg9w2lrtw31gRY6/n15PWarHoFfcaaVYIN9Dl+cmFgM3xyeoryW3h5KlM2URZKgP6qzHrhGlnFBX2rRhz4OtxBHysC8Prd4AjJ3NeqpD3evExkVZOUBOiT+XrqQnCDPosiksbg4Kc1sS/40Wm0gvX12NRkyQgR0Kc8DLpXXMkkRHlzQr70y1WpazmtHqtvNxZJoU1f9WADfSpFAOV7J8TLhMhKW4Dp7X0KcMEH8OW6Qo9dBSPo4+0Zt6dS7hoTJ1vyD3l9LsPLR94AoL8mu0HswqEu6HsMJee9Ee6gTwYqjHn/NShCXVfNXia2aDl9QaZEV6vbjVCCNas22Cx9RjALFywK3NUbcg96fVCXNZ6U/qvz5YnpybaZ6MFo6eNuJQESkCk9dNKrMWoAkuztZ2enyfcbiwXXuFWCDfQ5ftjfnp6ZqvkQdjzv1nPs7mtc1G8sylLviPW7wQj6gDIhxLvGxmnFEFTZ3hSSBVdm1CutNPkQLP6t4oK+VSMOfB3uoE9JVWzlUQV92kUyeXpTWA3DYkesjUksFEAfF//4iEp1I+K29naSFd4XMqqfmpEqmXWemejBCPrNpy7KoG2l8szMNMHqB6i9JSRxkQcB8x+MbCWHQgP0ufaMA/gaKhpPedU7gu6Xp7eDPpUiVglG0IcDgmtPnw94DuwWf9bz7M5r5rLl6fsV9Enos2P+c0G/OxoN0LbhDvq//PvfUnTwhE4suF69XVJFKdXEyCpNSIoubxIIw/W2sgAAIABJREFULswSjJY+IINFgRVGo4/m054rfvM5dvc5ljCeBADSrvwsGEEfkCe3A51Njq72amIa14OFEvoiYZDnVglGS7+q6ZR2W/yfwZHamhiOeW8KDJmw2LFQskowgj6Lb4CYxd8HK3K12ZP1vHrzmhg+OqOBFMRGeBas4oK+VSMOfB3uoM8lqWs+I28uytKYaEzFUa92jzt84rxaxcTAaRMbCqBPiRhkRq8tyJTP1hTYxpB7M9SxVmBJo6rCrvwsGEEffRCjhh//41X5HuOgN/qC6hmdsaD4aWuJkAhplWAEfUBmVdYB+c1Pe4SmWCygvSmE9aDhhv3RKsEI+pwDXrhP1uRr0ig5C94UKnXQGaBPfhL5SlZxQd+qEQe+dkFfBGAevbtc3WLrcw8KiVHeEnqqEzYA9El6o2bbLMFo6VN7TjY1iXzERHd6ufYct+S/lmZrjbad5yVYQZ/acyO7Gk5+b1WK0P2ssOGEPDUzVUMIVKFYJRhBH3c15Ewk2Y7aVa4hDOt59eY1jbBeX5gpA7eWeuwmWEGfBSBtbwmN8XjFi3MZ8yQ6u21EtOwsPqzZ/FbFuaBv1YgDX7ugL9rudnPeQaV9nRFf49VMYTJq6eJHJnpydUtIgD7DmMLGwdtLlfoVxkFvCuEWSFNenJduW3MerKDPYoa4/r0TEjTe6i0efkAfSmQm+sHbSm3zUoIR9FlckqtATB+ingyb/I7ejLv5yXXqrUJnVglW0Kd8EyscFz+LGbtkO+u5dvU1Sc/o7A8jomVPSZNHm3D244J+V7UZwO1c0BeNrxIHhVoWQhBAx1tCa9BRuyuU3x92OdxvZglGS984fppuUHs+YFOJMEF7SwgdkOVOdUCeTQ/6YAV93NNzE/fKnWPilNnMLibaEx1CXJOzv1X6zEjV8tPa5jMeuwlG0OckuH/+tSxHQxcAjTeF8UuIirwUqwQr6BPXLz18Ut5cnK2eOLv7x3quXX198Pg5mZu0V0E/svSIR8dQ9uOCfle1GcDtXNAXzT7HXQ0/+vsrciWq7IjXrsi+ljMycme5gj4156EE+ruKG+XtJTlaUgVbl7fc1dSZ43ZlQs6r9+SqD1bQxyKHaAbLdczuCoHS2BuCxyBz3zF5ckaKAlhdS+iAPhnjE6OqtD8GcX1vMvPRCAvuhBE7yz0uQ7CCPmtv5rLP1hbo/UO9vreE8To7oVZBP7r8iG2vEhf0vaVtH+7HBf125V65elVvFKxLWuB6S+gNDgvbE9NTJHtfq0cnv2C29AFnaF/hRyeRzFtlaAUNbfLK/EzNrLbr5BesoM+YSq5ulj7TU3ShZJdw15NxB+jDXAgTGwBmV9MerJY+TV5YXNIUi6oH4sreklkJ7aCPJ84qwQr6nAdet5G7ytWImRpbYz21Hr/G6zIzvkZBn6Rka/MwduyCfo/V678vuqDfrmus1J9jqzVjnBI7b9We02Ft2PYynZBxtVnjuMEM+sQLx+6pVC8GORHWJMWejmKAnrABLkoWAFYJZtBnodR/VTvtM0mK3vCO4D0i4Y2FJZN9fetZq8okWEGfxDTuIc6Nevo8L7ZzhmOCxSU5N1YJZtDnXJak7ZNXF2Ro4ypvzWUsJqfHtoM+SclWSnF+1wV960hy4GsX9NsvCpMvpCkvzEmX7zcV29ag9uTywYc9BNCfkaKlbbh4zRLMoH/16r9lYXKdcuSP3VPhNX50JnZCLW8vydasdLO+eB7MoI/rHVpZWqBGljZ5LAKt59qV19SvJ1U3yxPTUrSundirVYIV9HFX0w2TUsd3lmSr1W89t56+nh7XDvpU11gl2EE/qqxJK2AIkWGRe2NxSahyaky13D4iWpJrmpVO2qo3F/StGnHgaxf02y8KkwvZ1e8uy2nvUuUli4LsYzK2SbIqPnRCQgn00RwxaqxySuzsGLp6MuTpEtd3TnvnQ7ukymAGfXS0Ke+g/N+wKFmQXCeUdPZW8LDEVx5VjwueF7KsrRKsoM95cG9+t6lIgZ+SMW8Jru+X52coeZZ1n8EO+hBcYbw8MzNV8uuPe8w71vPtymtClYRYbh8RI2m1LbZcEy7od0WTAd7GBf3rF4COeN9sKNKJYLOXEmBoHkPpDOVUVAiEGujjVobJ8O+TE2X/sTNeCYsA+s/MSpX3luXaUokGM+iTiMb53TEqRsvrvEGVCujDAfDYtGTlhGi0iXsHM+hzh1IWykKQ+Lu3eDSoY8eDMCm66vok8OuzYAf91jOXZHJUtXrh1mbXezT68jjhLrxBmAVKcUA/o+6YbTgvpEEft2xTU5MkJCTItm3bJDY2Vqqrq4X37aSqqkq32b59u8TExEhzc7Ncvtzece38+fMC+O7atUv3lZqaKgcPXm9kwnaVlZUSGRkp5u/zW3x3//79EhERITt37tR9ZGRkSEtLS5fKqFzQv361cFfj6nt+Tpo2ybn+Sc+fAfQwfkGcgtUPZapZgtm9z3mUN55U4pQ/joyRzE4mAvP5duU5sW4WSVCJ4h2xSjCDPklWxNwha3pvea5XuhTC8kjFCVwQ4yOrbHkmgh30KdcjyZaOeHYtXa1jpCuvJ0dXqZdqSky1x+bBDvpXrv5bewswJkgktmPP8zjpm7xBqHJCZJXcPjJG6bGtlUh8PaRB//Tp07JmzRr517/+JU8++aS88cYbMnHiRDlx4oRcvXp9Yuc52/7888/y+uuvS58+feS1116TrVu3yrFj7b3CAe2ZM2fK008/rft6//33dd+XLl3Sy8ACYNq0afLcc8/pNs8//7xERUVJW1ubJuiwL/b5wgsvCJ99/vnnsmPHDrlypSP7m901dUG/o1aWpe/XMp4v1hZ4JQ5GVvsPm4u1zpibJtRAn5KqZekHlJoTylRvTMhUOZCJ/tGqPCk55MkjH8ygz2gjAfLjVXmaNIqrH/d1bwTQ31PapAsJklC5JlYJdtBncUnzHfI87PoxWM+3K68pBSQ0RU8EqwQ76HM+VD2wUIJq2M77Yz3nm72G6ZHwEQt8vFXWpGS+H7Kgz2q9oaFB7rnnHlm/fr1a/Lt375YnnnhCcnNz5ezZ69mz586dk/j4eOnfv7+sW7dOtwWQH330UcnPz9cFAt996KGHpKCgQFpbW2Xy5Mny2WefSV1dnX6+aNEiGTFihHoBuHmnTJkiQ4cOlbS0NAX+4uJiOXDggJw5c0Y9BNOnT5dHHnlE+O2biQv6HTUUXXZE+i3L0UQy3Ka9TYAhJm3E1vY2n/Yoawt2Sx8GMFzL8HFjUeAC7K0wqT8+LVnDBhCNWCXYQR9OA+LJhETmJe31GBPW873Za0B/V/FhbeqE+9UutyLYQZ9zHLC5WBdK67IbbqaSLn0+PqJKGSWnx4Um6GfsbZVPVxfIXWPitL9Ib+cyPJXwSwD6VNhYQ5UoPWRB31gFPvXUU4IrHYsa4P3pp59k7ty5cujQdUKE48ePy6hRoxTIAXm2xdXPggE3P1b+ihUr1FMA4AMCuPkBeRYUWPssAAB6FhoXL15UxeJhYBHB/gB3tmMxwuPatWvlb3/7m5w8efKmLn4X9DvOD6WHTigrH6BDAoydC6vjN278ipKzbzcWyTOz0mR/yxmPCT7YQR8rlVU/lvlbS7J1MrixRm7+aea+VgUwmvnYdYwLdtBnMUnmPmVo1NXbldjdXEvXt6BdM9SrtG+eEl1t68oNdtAHsIjBE3obbVNXf10bXX82bk+lgv6shFqPLxlzPIZUsEq7O75SF+QketqV2HXn3PC2UBIK6Bc1tAnllFYJWdA/cuSIbNq0SV37JSXtHZoAz1mzZsmgQYOktvb6ICJ2j7t+6dKlGrdHSYcPH1ZLf8uWLbpomDdvnnz11VfXLPPMzEwFeSx+QP7ll19W9z/PCRcA8rjyFyxYYNW57puFR9++fdXyZyFgFSbN1atX60KE473llls0V4Dj7s0fnorExMRe7aM3v++N72YXlcuQ9Rly77hYmb6nUPLLqnt1PtvTSuSDxany+JR4Sc4vl8rqmg77q6mp0byQwsLCDu9741z8tY/o7FJ5e26C3D8hXlYlFEt1Te/G0aaUYrl/XKy8uzBFIrPKPPRCzktOTo7H+/46397+DmMgLrtY+vwcJ/9amCqbUkt7dS5F5dUyNyJf7h8fK4PXZ0p+aZXH/hhf5B8x3np7/IH6/rSd2fLc9Hj557xUqerlGOMcvl+dIX1+jpdRm7I8dAJ4YZSVl5d7fBao8+/u7+YUV8i0Hdnyv0MiZPz2PEkrquzVuURklclXK9PlD8MjZXdGqVRUeY4lvM/Z2dk9+h284fPnz7fCVaev/1+nn/jgA0B71apV8sknn+ig4Cfq6+tl8eLF8vXXX6slb/wsCwRi+YCs4QFobGxUUMaSJwHQWCyQlIcAnrjo8RBcuHBBnn32Wd3G2CePvDdnzhzzWwryxPcBcnIEjJyADhuJCBMAxzpy5Eg93t/97neaY4AHojd/eXl5OrGQdNib/QTyu+WVVTJzd548NDFO+s1PkJS80l6dy+bkQum3IFkenxInyXmlUlbRUTfoivAPXqBAnndvfju1oFzGbs2VP46IkolbMyWrsHc6W59YKH8bGy39FibLnoxiD72kpKRIVlaWx/u9OQd/freyqkpKKyrl7QUp8uy0eJm8NbNX55JfUiEzduXIfeNiZPD6DMktKffYH+OLcRbM9+b6pCJdCP5jYowkZBdJaXmFx3l25zp+vTJNnpwSp6Bv/R6e2+joaMGos34WLK/LK6pkXWKR3Dk6Rt6dlyA7Ugt7dS670orl82Wp8odhkbI7vVhKyzvOZeiFBXlP782PPvrIuaAPaONC//DDD9XVDqgSU2eVMmDAAF1NG0CLW+2tt95SFz7ueYTvExrAW4BlPHv2bPnhhx80E5/PyQsgcW/cuHEK+iTnsTDAFYzljksfSx6LHuE9MvzxEHz//fcK+iw27Kx8/YLpn+veNynj16ewTb25KEsJdahNtXGWeH6pk3ey9x3XjGN6w9NwhV70Zgl29z7ngtsQcpi7x8bJ2IhKwa3YG4FS9oGJCVo+SRzRKsHu3jfOZ3xEu3uZx97EW9H/+pwGeWhKkhCfPnWhvSrI+B0eg929zznACMf4IhcituKI7Xmaz/lmz+mJAYX0wmTP2v9QcO9z/oTHYObrOztNaPjVm7mM8lI6EuLer2g6KeTzWCVk3fvE3imfw8WOVY5Qrjd+/HgFauL0hpCh/8033yhAs8pGAP+77rpL9uzZo54CEvXee++9awmALATYF4sIXPrvvPOOWv5UBgD43MBk6xMyQHiPhQQehWHDhulqviuAz3dd0FcVdvjH4Kbb3v8MidR4dW845TPrWuXztQXSd066Zm1TSmOWUAB9aD7JDv77pESlmIVfvjeStveY3DchXr7bWGTbQz1UQH9D7kHNHid3gTh/x5HRdQ3CuLYmu14empwkUMuS9GaVUAB9Esfoi0HDIsrt7KoUrOd9o9cknnJfLkoNXdCHL3/4znLt7Ejehx1Q30hH5s9ISoZz5I6RMVJ95JRcCSfQx6rGlUGyHEl3JMwB1C+99JK66wFnQ0gE2bhxo7rRydJnW2JrDzzwgMbzyfQntv/YY49pHIR4vRHjxyUHKJDEN3r0aElKStLvr1y5Ur777juNOQH4e/fu1YUB4QCsfUICXRUX9D01dbjtvCxO3Sf/NShC1uXU92pyobMekzoWBa1UrTzYoQD6gNWJc5fl/eW5Smy0Ibd32dWptcfkr+PjtdSx6kjoWvr0Yui/Kl/bCMNHYB0bniPT/h1Af2VmvVrAkNjYJZ+GAujjDaE0kV7xlO7ZNRay15D9u0O3l6kFvDTtupFmbBkqlj5zzsa8g+qFo0NebxoWFTacUM4RiKVqj9oTcYWspc/AMLLyceeTtY9bHSAmbk/MnNgGlj2gjGVPNj7JegMHDpRvv/32WjY+Fjnxox9//PGaa/7LL7/UmLyRfU+sHA4AsviHDBki/fr10zp+8gj4vYULF8pvfvMbefHFF3U/hpeAOn5A5Ubigr6ndrC64iqPyi1DI5V4xg54PL9l/w50lZ+szlfQp1TLOrGHAuhz5vC/T46qkqdmpKiL2S6z115Dnu+m1LboJIW3xS5UECqWPqV19GWgsoOJuaceJUB/ecZ+7URHu1i7UqpQAH1GCuWcMEDeNTZOiZus95PnaOr8ncHbSxX00Z1VQgX0YYCk7JVw2Q+bS3pVXVNQ3yYDNhUrmyQ8/JCZWSWkQZ8kuYqKCgVngBxrHAseLwAgjlVO5icCsMfFxek2JNmRlU8WrWGR4xmg9G/48OG6gCAr3wgb8H0S/Ej4Gzt2rIL+pEmT9Pv8FkmFJARCyMO+WVzwRyKfUQJovTDm1y7om7XR/hyLAu5qYofU7GfUtXpu1MV3oKnFmqOBDLFX640SKqAP6dDu4kalz8UF2BtOeXR255hY7VlQY1P3HyqgT6iHcjH40Uftquhxr3i8LHRVo/3svKQ628VDqIA+PBCQ6vxm4B5lIbTLX+jiranji1yblZmerbRDBfSZy6AXf3FuuryzNEd2lzR2VT0e2+UdaNNOh1j6B46dtc0ZC2nQ99BIkL7hgr79hcN1iIUOnzk3Sk8TrUhw+2hlnt50Zy56drsKFdDH4mJCph3uByvytCGHvWZv/m5yTYv8eXSs4H7FjWiVUAF9zoseD28szJK3FmfrorAniVYnzl0SGtE8MClR5ifX2cZtQwX04ZTfkn9IQX9Owt5ecRwM3FqiZD/kQ1glVECf84I5Dwu97+z2/IWezmXwcXy7oUj+NCpG6ORoN1Zd0LeOJAe+dkHf/qKQJDQ3ca+6mZek7bdtI2n/zY7vJlY1y4crAP0MdYFbb7hQAX3OmiQh6ItfmpuhMeaOmuj6q6TqFvnjqBgZBsNfc2iDPpUKJHpipTccP2cL2DfTHGGj+Ul1cv/EBFmYUucRQuL7oQL6eEdwM//vkEj5Zn2RQH7VE8H7irsbS39tjmcOSiiBPl64Jan75IlpyRquJHzZE6EnxtfrC+VPo2NtOzmyTxf0e6JZP3/HBX17hZ+5cEXSao/pZAwLlV1s2f6bHd+FCYsENzi+L15pL7k0bxFKoM/KH0Y4XInorKeSWN0sfxgRrWx1lExaJZQsfWKjUOfeMy5eosuPaEKk9Xxv9hrQJ5YP6JOJble5EyqgTxSZMQEzH/dUXMXRm6nH9nMW34A+1i9VFFYJJdBnocTi8pX5mfLFukKpaPRMjrWev91remJ8ta5QQ2927Zv5jgv6dppz2Hsu6NtfEOrpaR7zzMw0+XhVfo+7ocFL3+9X0McStqa+hBLoo0kayPxzUZY2R4Ee1urZsNd2x3fhSbh1eLTGuetaQhv00dHa7Ab1KEEz25NciLazl2RWfK2CPl4pOwkV0OfcKA8l9PPQ5ES10q1NrOzO3/oe5Wa4vKmqYcxaJZRAn3sQzyUNrLg3exrXp/z4y3UFcs+4uE6rAFzQt44kB752Qb/ziwJIE49nhbzKJtmn829e/wTr7d1lOdq325rEx1ahBvq5+4/LZ2sLNJxR0Xiy273PsVLxjtw6LErG7qkQLGGrhJKlz4TM+dICFY9QTypFSNSClIeY/tL00Ad9PBub8g/qImdqbLVAetVdwev23aZiTbDdUnC9V4qxn1ACfc7pytWrMnRHmXrh4HLoiWTUtYei/jY+3gX9nijQKd9xQb/zK0FyGu5q6oLpI90TYovIsiOaNUuLSzu3a6iBPt4RGqLQTIaJubvxQxZGsK1h6cNUZ1eLHUqgz+gjLs3iEnIj4qbd9Y4A+ngJqDahzbGdhJKlT0kiTI2PT0tR4IbvoLtCKRuNsMi1gbTGKqEG+pwf3CMYMN9sKBJc/lavo1UH1teECOAcIYzUWate19K3as2Br13Q7/yiMPnuLG5Ua4AJ4kgPLIqI0iZ5Z2m2vLYg0/aHQg30Ae1FKfvkkalJGpMns7w7wmSEdwTQpzSL0iCrhBroG3F9eCFwvXZ3odR69pIy1D00JVFWZIQ+6JM7wgIcAKNaxA60rWPG+hodA37kBewoCg/Qjyk/Iv9alqN6a7UhCrPqyPoa0qxP17Qnnbqgb9VOEL12Qb/zi8XkQp04JWiUVCVXt3S+cSefMIm/syRb3liUZbtFqIE+J8lC6dUFmWpFYfnblfbYKuPXCgDazgL6U2KqbUuyQg30sdR3lTTK/w2LkqmxNbYhjc70xftM4BMiK+UfU5Jsa87ZJpQsfc6HMTVoW6lm38+Iq7H1ot1IZ6cvXJavNxQJHjj4JawSipZ+zZHTMmhrqbbBJm/GjrnRqgfza/gzIEZinLmgb9ZMkD13Qf/GFwyLANY0yHUgPumuAIBQhrJosJNQBH1IPH7YXKxUuvn1bVonbHfudu+RlLWnpEl7gE+LrZH61nMem4Ua6F/65ao2RsFtiiWFG7U7AtUq+Q+P/JwkqzI9a87ZV6iBPue0JqteF5aMNdgNuyMQGn21vlAXp3jjrBKKoE/SKFUej05N1rAluRHdkeSaZu2t8ejPydJ0wj6PwnXvd0ejAdrWBf2bK54yqJfnZegkoRn43QiG4Xp8e3G2Wvt2vxSKoI8VAFkMZXers+rlyKmu94IgXruruFFBf0Z8rdauW/UWaqDP+RE6gv2RuvHNeQc9mButOjC/BvTJo2AyR992EoqgX9RwQitjYJqDnrc7gnfly3WFGnaLKj/i8dVQBH1OkvkIA4Ysfnj4u+OFo5SW3JPHpyVL00n7e9oFfY+h5Lw3XNC/+TUhuxqCnb5zjPa4ni0lO9vLtoLDauW/uzTHdpNQBH045ClVvHNMnJZF2bXHtVXGr+xhxFhvGx4tNI+B+csqoQj6WJ6cL53yZsTVKjuf9bw7e91y5qLmT5DYtibbk2iG74Ui6EPBa7DqdVaq2JnOjp2+qDXrry/M0rFq3S5UQV9r7ddTax8nJYdO2FI2W3VhvCYk8OGKXOkzPcUFfUMpwfjogv7Nr9q+lrNaM04sK7L0iOAm66pAGYpr/73lubZfCUXQ50Rpw/nq/Ax5akaqUOrTVTl76YpsKzykoD83aa8t81cogj4eDuqg0df3m4q1SUpXdYZrm7r1J6alyDobdjn2E4qgz3nNS9qrRD0k2nZHqFuHCZFcG5prWSVUQR/WR1z85I/Q5Km5G2ERg2js6VmpymFi1RmvXUvfTisOe88F/ZtfELirl6XvVzIQGqN0p4/3prx20CcZ0E5CFfTJuifOjEUBv7xdj3c7fbAdddOEBhYk1wltjq0SiqBPpQjWPqGgNxdlCx6irgqg356glSrrwwz0KYmF/AqXNbz8VH90Rajtp/wMNzdU2VYJVdBnLqM6hvtryPbSbvFC4L3DeCEERYKunbigb6cVh73ngv7NLwj19VFlRzTTlwFPiVVXY2FQfGLpw4ZlJ6EK+iQJkSD1lzGxSjNrR6drpw9ctjCkMSlR+meXJRyKoI8uKHccvqNMXpiTLpOjq7uckY61RhtivAR2lLLsO1Qt/aojp7VHA144mlt11QtHPJukSa3KqQkf0GcuK2xoU+CmXLE7XriY8qPSb1n7AquzPB0X9O1mNYe954J+1y4I7uoft5TI7SOiJWPvsS5npGN5vbk4Szv22f1SqII+Gel1zWc0uezT1fmSXO05sdrpg97wABegvzRtn22WcKiCPvqAkvfV+ZmaZEayWVcWl82nLmjuBC167Shl2W+ogj7tqmkrDDHR5OiqLnvhyBWhiyaVNam1nqW4oWrpMxYgvBq6vVTunZAgWwsOaSMwu3vR+p6yiy7NUW6DzrydLuhbtebA1y7od+2iYHHi4v/d4EhZnr7f1gK129Pa7HoFfVyJdhKqoM+5MiF/uDJPXa+0L+0K0xxkPsSlAf3lGfttE4ZCGfTpIEevh9cXZkrO/lbbjnnWccQETEz72VmpGkqxfs7rUAV9zg1eh2dnp2omPr0auuLgP9B6VvX8zpIcydjrmfkfyqBPGASwv2tsnHRWFms3hiLL2onG4OBwQd9OQ0Hyngv6XbtQlOpl1LXKLUOi5PtNRZr52pVvUkL15qIsbTlrt30ogz7kH3OSqAtOkolRlXL+0s0TILFuWSDcPiJGVnVS7hfKoE9OAx0KoTGGNpVKiJsJEzAtT/vOTtN8CLvtQxn04YKg5p52uzyHZ/5mQnLuRyvz5V9LczSB0rp9KIM+cxnW/gOTEjSvIa2LvBCE6/CMsCB1Qd86YoLotQv6XbtYuFlphkIcjM5clK90RVZmHtBkISYlOwll0G9fKB1TnXH+pYdO2qmgw3tYIRDMAPqAv93kEsqgjzdkadp+HWPUkZN4dTMhvkrLVMbl1k4SAEMZ9HHVk/T524ERgmfNbsyYdYgngBwTvFA0w6LfgVVCGfSZywijkQAJI2FnFR9WnUCa9dbiLPVcElKyE9e9b6cVh73ngn7XLwjJP2P3VMoDExPVCu1K0hBc6GQIw/NtJ6EM+gAYxDH9V+WrDsjiv5mwPTq7fWSMZqLblRSFMuijH7KkAaSnZ6Zqq13awN5IyKQmfEQGe2c89KEM+rBmksRH74Ih20qV3fBG+gL0q4+002sDfHYNe0IZ9NENJaJ0KHxqZqrmQnSl3wNU0cxl0Irb3Zfs1wX9G408h3zmgn7XLwSr45iKI9plitI9Jo6bCXkAbyzMku82FttuGsqgzwkD/GMjKhWQxuyuuGlJVfPpC5o7AehTR2xHrxrqoM+4Qld/HRev441xdyMB9OFE76x5DN8NZdA3vHBPTk9RyzWu4ugN80cAfQij3l+RqyVodDm0SqiDPl64+Kqjel/iUeoKgRaU4sxleEdc0LeOmCB67YJ+1y8WFhd0qTDzMfBxd91MYArjRhmwKTxBH/1Qd4+FACshrldaFncmfI7O/jgyRiA2wvK3SqiDPh4kXK6A/vAd5ULS2Y0ESlQ8A1BF7yzybB7Dd0MZ9Dk/Em1H7ixeRuOrAAAgAElEQVSX+yck6KKRhjqdCaOvvPGkAv4HK3Kl+OAJj01DHfRZjBMWYrFIYl5npZ5mxcCU+fqCLHl/ea7tYpxtXUvfrDGHPndBv3sXBvKPgb92qqIjGrXVNxJ4+99YmKl11HbbhbqlzznXHj2t5Y5YYnQdvFGcGgBDZ4A+zHx0kLNKqIM+50tGOSD+8JQkwRKlvrozgWgGqxUGRPoW2Emog/6ZC1e09A4aY+5PKGY7EzTJ59Scs1iy2zbUQd/QDXwQlHr+uKX4hmOM7Qkd0SIcndl54NjGBX1Dsw5+dEG/exeHFfLOosM6IX+5vlBqm2/s4l+QUqfZrrQBtZNwAH3ihyRawQ0P18GNunthsaEzQB+rlV7xVgkH0Kf0bGJUldw2Ilo25jXYLn4MvaAzvChMyCyq7CTUQR/vEWEQPErw6RMa6kwAfXg38NbRQKa88ZTHpuEC+tTew7LH2GHM4fbvTCjzI/EP74CdB47vuaDfmfYc9L4L+t27GEwY8FeTOAWJyqb8zicX9jw/uR30oby0k3AAfc6b5DQsK4CfmDXkPXYC7S4tjO8YGSO7S5qEEj6rhAPow0xIhcgdo2KULtXOBW3oBdA3Sqns2sSyXaiDPufIgpyFEqyZY/ZUdEr9zD2M94RyPZJM7eLZ4QL6zGXkJz02NVlbFd8oOZlwG94n5j4X9I27LwgfXdDv/kW78su/ZWpMjfSdnS5Y8KyOmUjshKYxry3MVKpQu8/DBfRrm88oHS8JetsKD3c6aVB+RUMQwA7SFTuvQDiAPgBG/4IX56ardYXXo7NQEqCPhcsfBCp2Eg6gz3nDo08ODbH6kk5KRLlXydinJS9Wq11CbriAPjwQcIlQKQIpFOG1ziJJm/MPykvz0uWLdS7o291jQfOeC/o9u1R023t3Wa68ODdD+1J35habnbhXXWcjdpbb/lC4gD5EPXtKm+TOMbHyw+YSW+sKBWF5zEpoB/2Y8iPahMaquHAAfc4ZL8eU6GqBV35GfK2tLtiOUlJc2m8uzha73vBsEy6gT8vcAZuLpe+cdOU7sI4dXgP6sB3iHfl0dYHW7Fu3CxfQ57wh50EPfx0fr2GPi5ftuSEImVAWChGUXa4N+3Ld+9aR5MDXLuj37KKQUY0LEc5v+M47K6uamVCroI8LzU7CBfQ59+JDJ+SNRZny0JRE5eK3S05DrzPia9TSp+WpnV7DBfTJhcAixQqjrCrngCeJDHolJPLK/EwFMRZKdhIuoM+5L0zeJ8/NSpNPVuXblu5hyWbWHdNmO7iqiWVbJZxAH+8aITXa7VJijOfITsjwhwAKymcX9O00FCTvuaDfswt18covsjanQR6blqyrZKwtO5keX6slMSwQ7CScQP9Q23l13dN5D4pZO+Y06EHhA//TqBh11cLfb5VwAX1c/GcvXFG3K7FUJmQ7Qa8vzstQdzW5E3YSTqCfUtuicWcWS8TrWTyZBdBP33tM6bE/X1tgWxIZTqCPix9yIyx9wh0kOdoJZaSENCk/tkuw5TuupW+nOYe954J+zy8IE0f/VXly7/gE5fy2c4sBYNTBjo+otP2hcAJ9XPx5B9qU3AhrIXufp+VK2+KfY6rlT6NiJbmmWUhos0q4gD7nDUCR44BblcoHPB/WmOuhtnNqgZGYhnfETsIJ9IlLc989MClRFqbUeeSPoL/UmhZ5g54Y6wql4bgnD0I4gT7hjqqmUwr4/5iSqBUg1oUSYwojhyRJ2jjbJdiyjQv6dnefw95zQb/nF4T487ykvfI/gyNlSep+dbNa9waAUT89MbLK+pG+DifQZ7Il5tpvWY72jF+ZWS8XLQ1l4ESndhjQp+WpXTZxOIE+gySzrlUT00gIZaFpJTc62HZOnp2VpnqNd0FfWR93FDVq/TnNYeiZAb+GIYzDpOoWzYOg7PZw2znjo2uP4QT6nDQMe/S6IOeG+4/Ft1X4nHE2aGupC/pW5QTTaxf0e361WA0zCcOcRjlaWu2xDpMLe+YGoraVXt92Ek6gz/lDzENDGbrIQaJinVzI8qfsCtCnqyFd56wSbqBPu+HxkZXSZ0aqjN5doQshs7VPTBYOdeqtO2sEFU6WPuOFEscfN5coH//O4sMdQArdoSdq079aX2TbvjncQJ/FNyRajKO3l2YLlLvmhRI6peslIZOh28s66NN8f7qWvlkbDn3ugn7vLgzW/uBtpbpChknO6vaaFFWloI/FbyfhBvpYqU0nzst7y3Pk+Tlp2lznug0mUnP0tIyPrJI/j46VrH2tYtcIJNxAn3FD/f2bi7Pk4Z+TlBDK7CEB9GE7pEyN2KydhBvoExaKKmuS3w2O1CRIWu4aAujDz0/YjUZYdrkl4Qb66Ibw29g9FYKLf9Tuco+5jI6hT81IkeE7yqTNhj+Dfbigb4wyBz+6oN+7i8ONUtjQJo9OTZYPV+Z6xFQnRLaD/rS4GtsfCjfQRwmUN8LQR5z6k9X56lo0XNa4YsdGVCjo5+4/rhORVXHhCPokis5OqJW7xsaqJ4Q4viEsPCFX+XBFniTXtBhvd3gMN9BnPMGW+dGqPHlocqISzxheI0CfKgcqHsgtsaOUDUfQx7KHkvitxdka+oBr3yx0v+wzPUUXBHb8GWzrgr5ZYw597oJ+7y4MGdZYFbiqiXfhxscda8i4iEp5eX6G1lkb75kfwxH0Of/Swyfl+03F0mdGirbQPXep3Y1PtjUubCx9rDM7nv5wBH0jw5oY9RPTkjWUZCRbAfpw9PdfmS8pLuhfu71IeowuO6K6AdypzUcAfUiMuC/pfnncpr9DOII+Hjc8a4Qkn/81Sx8dMschy9MPaFgOb4AL+qqS4Pzngn7vrxs3Be5CCFJIUjNPvGP2VCp1JYQzdhKuoI+HBMuBEqB+y3O0bIouhhWNJ2XkrnIFfUqHDGAz6y4cQZ/zB9znJtXJbcOjZUZczbV8iPrWc0KjGRjVyCuxk3Cz9NEBlivhtk9X56tXiaRbxh0Ytqe0UVsRs/A0L9IN3YUj6BvnTgItHjiMGHIfjIU3uTiPT0vW/BI7nfF919I3tOjgRxf0vXNxsPZh3XtmZpqC1vnLv+gKGauVGmvKruwkXEEfXRTUt2n5z19Gx2p3OCbossMnVY+8V3r4hMCHYJVwBX2s/dJDJzWUhMVPa2dCJYD+AxMT5ZM1+ZpYatUXr8MR9DlvAN4glQHIaK7De3QjhE0T9j47LohwBn2sexZIcJCQ8wBZD4YNLa8JI5GndOKcZykt+nZB3+7uc9h7Luh774KQbAW1JxmuFU0n1UrFaoWven5Sne0PhTPoY3XRVIdMfdp1knFNTJHsYECfnufmhDVDgeEK+pw/CyOoee+bEK8WF211YTHkNexyGXWupW+ME+MRnWHRPzEtRev3iffTJvaFOema4U9bXquEM+ijCzxG/VfmKUlWxt5jWkUDodajPyfLlJhqF/StAyaYXrug772r1XzqgsyIq9USPjKp9zaf1sx+QB+SEDsJZ9DHeiCGz4R899g4nZBp3zlkW6nA2gdhiF03vnAGfax9EtQoNyOTmoZOJD/eMy5OYJfL3Ncet7aOtXC19NEDIA9VNuE3dVlXNsvKjANKaATRDItPq4Q76NNFb31Og9w+IlreWZqtlTSMNapHID7CG2AnrqVvpxWHveeCvvcuCN33cg8cv+ayHrunUsv1yFJnlWwn4Qz66IOwCFwHZFKjp0/X5Gu9+V/GxGn5nl0jo3AGfVzTeD8gSoFRjvbOVIj8cWSMdj/LdkHf7jbTfAhCbFj7dCP8fmOxPPJzkibg2jFphjvokw9B90FClvdPTFBCHqpDKOebGV8bnqB/5coVOXjwoOzevVtWrlwp27dvFyYj3reTkpIS3WbVqlWybds2aWxslEuX2rO8z507J9XV1bJ27VrdV3R0tADGhrBdcXGxbNq0SVavXq3fb2pqEgADuXDhghQUFMjmzZuvfZ/92zU0MfZpPLqgb2jCO4/EB5Orm7XsBQIastChriQJxk7CHfTJC6aUCl5vPCIPTErQ+PSdY+K0+5mVIAQdhjPoG2NIS/gS92oyGl34/mdIez16zn5PamO+E86WPuePV4mKkdG7KjSchJuahZLRGtvQq/EY7qCPHqioIeQG1TgeknvGxmmDMUpHw9LSZ1AsWbJEPvjgA3nppZfk3XfflTFjxkhra+s1MNbBdvWqtLW1yaRJk+Sdd96RV155RR8B+ObmdiKN2tpamTJlirz66qvy8ssvy0cffSTLly9XMGcfAPPPP/+sn7/22mu6DxYbx44d099qaGiQgQMHyltvvaXH8umnn8q6devk8mV7F4wxsI1933fffVJUVGR+u0fPjxw5Irm5uV1abPToB4LkSyS5QAwC2P9+aJTG+Cl3sZNwB310gvWKi3VWQq1mB//XoAgB9Gm+Y9Tvm3Xngn67NnDrT46q0vyH3wzcI1+tL1RPk1lXxvNwB330gNeo/PBJeX95rgI+OoNYy26MuaDfPnIIr0Ht/M6SHLl1eLQ8OClRk/zskh/5Rsi6969evSoHDhyQO++8U61rgB7r/IknnpCsrCw5c+Y6bzFWPJ99/PHH17aNiIiQhx56SAGSSX/nzp3y4IMPqsJOnz6tAP/JJ59ITU2N8Fvz5s2TESNGqFeBBcSMGTNk0KBBkpycLCdOnJDY2Fi599579bdbWlp0MfL4448LA/dm1r5r6RvToncfsVCXpO3TOvS+s9NkdZYL+jfTMCQp7c12YuSecfHqksVCs4oL+tc1QrUDXc9uGRol320qUm6D659ef+aCfrsuCI3Q3+GleRny+2FRMmxHmVw18fIbGnNB39AEi/J//+qJy1DOA/KTCMvZSciCPkCbkJAgzzzzjAItwM1ENGTIEJk5c6a6/Q2FHD9+XIYNGyZTp06VwsJCtcyx7O+++25dDNTV1cmyZcvkzTffVI8AIM+iYNSoUerKx7XPggFL//DhwxoSIBTw9ttv6+csDKZPn66LAACc8AKLDDwCOTk5HRYgxjGZH13QN2vDe8+BKnpOr86q13h+6SH7dpWupX9d50YcEZ0Rp+7MheiC/nWdURpa2XhKRu4slz0ljUI2v524oN+uFSMnArriWfG1Wj1ipy8X9DtqhQoIehhMj6uRooNtQkKpnYQs6OPG3rhxo7r0idUjgOfs2bPlp59+EkDdEG62fv36KbDv398e1wW8H3vsMdmyZYukp6fL3Llz5ZtvvhG8AgjeAkCekADxesIHLCZYALDqOn/+vDz//PMyf/58dcsPHjxYnxPnR9hn//791YPAosMqLD4WLVokw4cPl6+//lp+97vf6bFUVlZKb/7y8vJ0MdSbfYTad+MzCyU6PV9yi0ptdVtRUSHx8fGC7kLt3HtyPoUl5ZKYVSg7k3OltLzCVicpKSl6j/Rk/6H4naLSctmWmCPJOUXCc7tzZHwxzhhvdp+H23tZBSUSmZYvSdlFtvoghwrjifk93HRjd74VlZWSklsku1NyJaewVCoq7LGCezMzM7NHOiOsDaZ1Vf5fVzf0xnaANgl5uODLy8t1l/X19bJ48WIF0aqq6x3VWCC8/vrrapUfOnRItyXJrm/fvrJ+/Xp1zc+aNUstdcAcISkP6x1rH9B/9tlnhW3Mwntz5sxRsPj+++9lxYoVmqzDNiwavvrqK9mwYYPG/c3f4zmrMXIKyCPAO3HLLbfIjh07ZO/evb36YzGRmJioi57e7itcvo+nBp2hu3A5596eZ2pqqobGerufcPq+cW8y3sLpvHt6rsyRMTExOr/3dB/h+L20tDT1MPfk3DFUHQv6WNQkyr3//vsKoAApVjyx9x9++EFj8QbQYunjigeUWRggLBr69OmjMX7i8oA3wG1Y+rjlCQeMHz9eQf+FF15QSx/XPZY+Fj+LBn6PBQLeBSx3FhhIRkaGkMxHlQD5BjcS171/I+34/jPXvd99Hbvu/e7rzHXvd09nrnu/e/oytg5Z9z4uc1w/uNhxmwHEuETGjh0rEydO1CQ/Qwlk2APouP7xCrAt4H/XXXdp7J7v4SEg+99IAMQNN27cOAXyixcv6mfTpk0TfhfgZ9GA94AMf9x1/CaVA+yX/fP9f/7zn4JFRGLgjcQF/Rtpx/efuaDffR27oN99nbmg3z2duaDfPX0ZW4cs6AO8uMnuuecedYuT2AfQsgjAVXvy5ElDBwrkW7dulS+//FJj7GwbFxen2frEPnDp8/kjjzyitfpnz57VBcIXX3yh8XpAAcDH1c/3+D6lggMGDNDXLCp27doljz76qC5AWBjgVXjyySeFTH6+fyNxQf9G2vH9Zy7od1/HLuh3X2cu6HdPZy7od09fxtYhC/qcoFF7TyIcAM0fbnZc/9Sqk92PAlggEMMHtHG5A/6ff/65xuiJ8WOZM4kNHTpUPvvsM43F8zkuf6x0PiephKQ+OAFI+KPen3g938fVT+Ig32HfHMe3336r3+ezm4kL+jfTkG8/d0G/+/p1Qb/7OnNBv3s6c0G/e/oytg5p0If4BmufWDqudTLwSWLgfdz4xNUprTOErEZi8Ljt2ZZSPZL0EAYYCwWAnX3B8GdUBfA52xH7J7mPOD8LAsDaIN/BO8Aig/p9QgwkGbJ/yv9uJi7o30xDvv3cBf3u69cF/e7rzAX97unMBf3u6cvYOqRB3zjJYH90QT+wV9AF/e7r3wX97uvMBf3u6cwF/e7py9jaBX1DEw5+dEE/sBfHBf3u698F/e7rzAX97unMBf3u6cvY2gV9QxMOfnRBP7AXxwX97uvfBf3u68wF/e7pzAX97unL2NoFfUMTDn50QT+wF8cF/e7r3wX97uvMBf3u6cwF/e7py9jaBX1DEw5+JOHv9ttv1wTBNWvWSG/+4CKgMVBv9hFu3yXpEp2R3Blu597T8yUZljLWnn4/HL/H+GKcMd7C8fy7e85wp9DUjPLo7n43nLcn0RxiuZ7o4LnnntNk9K7CpV9peLt6UMGw3cGDB7VVL/0B4D7uzR/UwHf8//bu/VWqsosDuH9EBUVUEIXQ7YcSxUqzm4Z0oyhvXYmsLKKbRjcqo9SSLnSHyijMMgypTLtSBBlYGplCQdYPFZVhGUVQ9Lx81uv2neN7xnN05pw5Z/Z6YBzP7Nl7Zn9nPc933Z61jjiiXHLJJS1dp5XvMNzOtQ3z8MMPjwqLw+27d+r7jhkzpugi2anPH46fq4InOSNvw/H7D/Z3VvzsoIMOikZog/3Zw/nzdHtVM2Z37kFreeXh+zuS9PuL1AC+T5GgKVOmRG2CAfyYrrq07Zhnn312WbFiRVfd10DejFoVTzzxxEB+RNddm3yRs2qbcNfdYJtvSAhp9OjRPZqntfkjuvJyisbxKg3GSNIfDJT7+Iwk/T4A6uVwkn4voPTxUpJ+HwD1cjhJvxdQdvJSkv5OwNnJoST9nYDTjYeS9Hf9V03S33XMkvR3HbMk/V3DLEl/1/Cq3p2kXyFRk2cVCW+99dY+6/3XBI5+3aZSyTfddFP0oO7XCfmmSPbRCjpH/xHQ54Oc9ac0d/+v2r3vlOAsN6lqh969d9reO5PMrZfMYIx07w8Gyn18hqRA5YL7U/63j0vV5rAte++99150TqzNTbd4o0pWK4Gdo/8I6MxJzshbjr4R0Fvl9ddfjxLpfb8731EhoOtsY/n56vWBeE7SHwhU85qJQCKQCCQCicAQRCBJfwj+KPmVEoFEIBFIBBKBgUAgSX8gUM1rJgKJQCKQCCQCQxCBJP0O/yibN2+OdsIqWKlm9e677xalP3OUSJ769ttvywsvvFCeeeaZ8tRTTxU7HarY119//RXYPffcc+Xpp5+ORJiff/4546+llH///besXbs2Kny99tprsc9cMtq6deuikAd5W7JkSfnmm2+yPkQpgcHXX38dMkTOPLT5Jk8//fRTvE7GyJrkvkzsK0VrcrHoxYsXx/wzTyUl//3338Xc/OijjwIvuL388svlxx9/rM3c1L5dnpZ1S22MVatWbV/SyY5dDnCr5uGmTZu2z0Pnfvrpp+XFF18MTnjllVeKdRCu7RhJ+u1AcTevYWGWJCQ7+JxzzokqVrNmzSrvvPNOTJrdvGzXnPb777/HwisbWOXD888/v8ycObPIdHVs/fr15fbbb4/CRtOnTw/83n777SKZqO7DAqvk7vjx48t5551Xfvnll6JfxEMPPRR4zZgxI2Ru6dKl5Ycffqg7XLGoWoTJ2gUXXBCyhqwoTm+99VbIFhmbOnVqufPOO8vGjRtDsaorcBIbv/rqq3LNNdeUiy++OOYnObvhhhtCkfzss8/K3Llzt89NxcfefPPNkMM6YPbnn3+We++9N+aYansq7VWDcvnwww8X1QureUgB//7770MpMneVfr7ooosCv0svvTTIf+vWrdUlWnpO0m8JvtZOptHddtttBdHLEqYBWnDuv//+IqO/7uOPP/4IHCwg9uV7PPnkk+WMM86IxRjhW1goTrYImWQWnU8++aTW0FEmWVb2/iKpivR5kiwmLAdW2uOPPx510hutkDoCBy8W/HXXXVdsaWRRecDogw8+CJkiW1UmP5mbN29eMX/rOmCDxA888MCwSuG1cuXKILdFixaFXFGOGDDfffddue+++8rs2bOLHSR1GPBYs2ZNeELuuOOOHqTP+rfum6OUA2saw8+uB8TOW6JcdqUkvfTSS1HlkLXfjh1eSfodlEDbp2655ZZyzz33xCJj8Xnsscfib9ZF3QcBN3m4CmHjsWzZsrDGLDCTJk0KNyz3q8nDZTZ58uRal+aFGUuBXKn98OCDD24n/auvvrpouoO84MryV1O+7qV5yQ4iV7yIB6mSNVgKjeiNQba8j6xxyZ511lm19sYhfcriAQccUFavXh0KudAbeXr11VfLKaecEnJFFuH2xRdflNNPPz3Cc3VY18iQdYvCM3/+/B6kTxmnCDBUeEzMQx5M1r9wG6PP/OXJdFw4SVMdod9ff/21ZfiS9FuGcPcvIGbIWmVxVYMVZgESz8nREwFaMKy4Cjds2BDaL5csi8sCzYU9duzYQSty0fPbDY2/kBaif/TRR0NBQlCVpe+ZgsmjZFHiOZk2bVpYYUPj23fmW1CCbr755iDyBx54oNx4441RyIjHSJz66KOPjpARGSNrZG7cuHGBX2e+cec/ldIoxHHZZZcVYQ+ht8q9j8yOOeaY6ExYzU0td4WauLHrNCiJCxYs6EH6wiG8INU8pBzw8Jqb8pXmzJkTdfjF+Q0hJuFNFn878r2S9DsogVxfND6xw2oo+0kgtFjM8T8ELB6wsTjfddddsQgfddRRPZQj4YBRo0bF5PjfmfX5n/uXqCdcxFMkLCIhrSJ9MUQLS+PwGld1nQeP25VXXllOPvnksLJYWtdee2143bhhyRRrtRqIS1MZSlNdBwu0wo2HRDvdq666KmL85O7II4/ssYYhOAr5rnSD6wZseyN989G63zi8xtjjEeGRo6xXIV5eJsqVEBTPQasjSb9VBFs4X+yGpd/oXl2+fHn8+HXTiHcGowWGtksDFpO2qFiEWWAsMVYHK2zLli1F0oxYWR0Hi7Wy7PV/h4PFmKv1ww8/DPcq0qdAsfQtxGL+4tV1HlWYg8ve/8kS/Fj8XNJkilvV62SNpc+SrTPps9wZLUIfVc6NJFqExdN06KGHBobwIms8ULwjdVvXeiN9CXpI3zw0zEOW/t133x3eE3lJwrxc/QZFnneA97cdSbdJ+gFrZ/6h1YndcP9YUAyLtoX5jTfe6MyXGmKfatEwcWRVIzCKkmHBFeeSFCNb399cYxMnTow47BC7jUH5Ol9++WVgxMo67LDDysiRI8u+++5b9tprr7D2jz322Ijpi7NSpLhhLUBCAXUe5EciFcKq5iHXPit/woQJkTvCquWGFULimTvttNNqHdOnHEkMJT8SzAwhN7tDrrjiivCEOA4vuMn0l28jfFmn0Rvp84wgffOQvJmHsvspS1z6lHDHqx0iciYonxL72rEzKUm/gxKIqCRbcYvR4Gh8yG3hwoXFto4cJSwvrlYua1n6lXUFKxNDYhpFgNvLgmMLUV0yhHeUD5aDug/cghZiOSMUShYshUi8mqdEcppELOTFeyJruM7DwsulL3vfAswy5YamCJibFmmyRcbIGpnjiq0stTpiZ70SghRi436mnJufFCeZ+jAThnv//fdjbbPN1jy2d79OozfSl61PoZT4aD179tlnw6vkbx4lCXsUJJ4UJK8Rz5gxY2ItpKy3OpL0W0WwhfMtLix6Cy+yv/zyyyN247XGGGILHzGsTyXwXNQyhFmpXFwyXxEZS0xWKxLzupiXuBhCU1AlRwmroTGmzxKzd58rkRIlAQu5VbHDOmP28ccfh4UFF+RkPgq7UZxko5MtMsayFZITbqq8AnXEjfUOA3PPuoXs4WMtI2fIXW4SvLxO1pAa67YOA5lTquWKIGxeNwokDGwDZdU3zkPhpKoAD+VSaAlu9ug7j1Lajsx92Cfpd1gCZWNy23DpIDNCITabo0SnLouHCcB6YKmaLDRj7laTQMKarS4mBVc/C6Qd2nA34E9pUjFNLFCSn4Waq5A7H55Ijdu1zhZr9TuLUcOGRU/OKEtV7oiiKRZwChNZI3OVx6k6v47P5p+165FHHglsyBNsWP2MFpYqvOAGT2TmWB0GT6S1vPKu8eha4xG+NYo3snEe8sRV89AzZZNHwDkUAi2L24Vdkn4dJDDvMRFIBBKBRCARSEs/ZSARSAQSgUQgEagPAmnp1+e3zjtNBBKBRCARqDkCSfo1F4C8/UQgEUgEEoH6IJCkX5/fOu80EUgEEoFEoOYIJOnXXADy9hOBRCARSATqg0CSfn1+67zTRGDYIKCGha1Lyrfa/pQjEUgE2oNAkn57cMyrJAKJQBsRUFNAjQFlquteMbCNsOalEoEszpMykAgkAkMPAcVvlLxVx33p0qVD7wvmN0oEhikCabuVM7sAAAS2SURBVOkP0x8uv3YiMNgIqB6popjqdJqpLFu2rGjyozqiiojKRyvru2jRoqgfriobN73KgM5TlU1DKX3Blf5VYYzrXgVKrznuuq5TfZYypKx9160+Uz1zQ0nXVatWReUy1cuUbP78888HG5b8vERgWCGQpD+sfq78solAZxBgeWsEop66euIscLXBtTZG1jr5zZ49OxoeqSmudLL+COLySrNqMKIO+8yZM6OOvSYiytsifH3Ckbtj6rgrtazhlGYt5557brT/1QznwgsvLGeeeWa0CdYwSO8F19ShzLkaMCkLmyMRSASaI5Ck3xybPJIIJALbEEDC6oDrPMfSZr37u1IC9thjjyB/dcWRtVrjWiFv3bo1CJ21rr64eu1qseuGuHLlyqIH+0knnRRWukZJrH8KxpYtW8K9f9xxx0VHOy1adcCbNGlSeBp4F3gHTjjhhKjp7hyKgB4DORKBRKA5Akn6zbHJI4lAIrANAQQ9Y8aMcsghh5SJEyfGY9SoUWXatGnx2G+//aL1J9LmmufiP/XUU4Oop06dGo1HWP3c+evXr4/ubJW7fuzYsWH1V13rhASEBsT0Wfc8AY5x5/MyaPDCja+j4ujRo6ON6+LFiyPU4DNyJAKJQHMEkvSbY5NHEoFEYBsCOoZx37PedU/z4LbXklav9P333z+6hyFsVjkSphxoFzplypR4j0tRCjZt2hRkrsuYJD2tR1n2jaNK5BNCYNEbPAHc/DrhIX3hAcQvT0C7W99tyZIljZfJ/ycCicAOCCTp7wBI/pkIJAL/jwByR6xz584NKxy5s9rto3/++efL3nvvHW597vyNGzdGXF6vdaEA1vq8efOCtLnfbcGTE4Cgkfb48ePLmjVrwjXPoqcYuM6O2ftI//rrr49rr1u3Lr4HBUMy4fLlyyO+TynIkQgkAs0RSNJvjk0eSQQSgW0IIG8W/axZs8qKFSsiG7+y9Lnb99xzz4jVy6bXV13sX4IfN/38+fMjxo/kHZfkt2DBgkj0k+znmgsXLgxlQD7A2rVrt2fvN27ZayR9/cj1GBd2sDPAdRG+hMEciUAi0ByBJP3m2OSRRCARaEBgw4YNQc4TJkwI6/z4448P8rZdb5999glil20/efLk8AhI9uMR4M6X9CdhT+Kd3IDVq1cXBXgk3ynCM3369HLiiScW16YUIHTZ+XPmzClCCwarnrdBLgDSZ927pmS/cePGxXnOyZEIJALNEUjSb45NHkkEEoEGBCTJ/fbbb2GFS9bz8Lf9+gcffHBss2ONS7iTpV8l5nHXe5/XnbN58+YgfApBFSaoznNcfN85lAL/5y0wXM91q9K8QgXVNavznJMjEUgEmiOQpN8cmzySCCQC/UCAJT5y5MiI2ffj7fmWRCAR6CACSfodBD8/OhHoBgTsv1dQJ/fId8OvmffQ7Qgk6Xf7L5z3lwgkAolAIpAIbEMgST9FIRFIBBKBRCARqAkCSfo1+aHzNhOBRCARSAQSgRFVBm0+/zeTOHFIHFIGUgZSBlIGulUGRtgGk4/EIGUgZSBlIGUgZaD7ZWDEP//8U/KRGKQMpAykDKQMpAx0vwyMUD87H4lBykDKQMpAykDKQPfLwH8AoHrLYxFwSgsAAAAASUVORK5CYII="}}},{"metadata":{},"cell_type":"markdown","source":"* max_grad_norm : 그래디언트 클리핑에서 쓰이는 변수. gradient explosion을 방지하기 위해 사용한다. gradient가 일정 범위를 범어나면 max_grad_norm으로 나눠준다.\n* seed : 난수를 생성할 때 사용하는 시드. 딥러닝에선 weight initialization 등에서 난수를 사용한다. \n* target_size : ???\n* n_fold : K-fold Cross Validation(딥러닝 k겹 교차검증)에서 사용되는 변수. <br>데이터셋의 크기가 작을 경우 테스트셋에 대한 성능 평가의 신뢰성이 떨어질 수 있다. <br>이를 해결하기 위해 모든 데이터가 최소 한 번은 테스트 데이터로 쓰이게 함. 각 결과를 평균내서 사용\n![image.png](attachment:image.png)","attachments":{"image.png":{"image/png":"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"}}},{"metadata":{},"cell_type":"markdown","source":""},{"metadata":{"trusted":true},"cell_type":"code","source":"# ====================================================\n# CFG\n# ====================================================\nclass CFG:\n    debug=False\n    apex=False\n    print_freq=100\n    num_workers=4\n    model_name='resnext50_32x4d'\n    size=256\n    scheduler='CosineAnnealingWarmRestarts' # ['ReduceLROnPlateau', 'CosineAnnealingLR', 'CosineAnnealingWarmRestarts']\n    epochs=10\n    #factor=0.2 # ReduceLROnPlateau\n    #patience=4 # ReduceLROnPlateau\n    #eps=1e-6 # ReduceLROnPlateau\n    #T_max=10 # CosineAnnealingLR\n    T_0=10 # CosineAnnealingWarmRestarts\n    lr=1e-4\n    min_lr=1e-6\n    batch_size=32\n    weight_decay=1e-6\n    gradient_accumulation_steps=1\n    max_grad_norm=1000\n    seed=42\n    target_size=5\n    target_col='label'\n    n_fold=5\n    trn_fold=[0, 1, 2, 3, 4]\n    train=True\n    inference=True\n    \nif CFG.debug:\n    CFG.epochs = 1\n    train = train.sample(n=1000, random_state=CFG.seed).reset_index(drop=True)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Library"},{"metadata":{},"cell_type":"markdown","source":"* timm : PyTorch Image Models 관련. 노트북에서 사용하는 pytorch-image-models 폴더에 들어있다."},{"metadata":{"trusted":true},"cell_type":"code","source":"# ====================================================\n# Library\n# ====================================================\nimport sys\nsys.path.append('../input/pytorch-image-models/pytorch-image-models-master')\nsys.path.append('../input/adamp-optimizer/AdamP-master')\nimport os\nimport math\nimport time\nimport random\nimport shutil\nfrom pathlib import Path\nfrom contextlib import contextmanager\nfrom collections import defaultdict, Counter\n\nimport scipy as sp\nimport numpy as np\nimport pandas as pd\n\nfrom sklearn import preprocessing\nfrom sklearn.metrics import accuracy_score\nfrom sklearn.model_selection import StratifiedKFold\n\nfrom tqdm.auto import tqdm\nfrom functools import partial\n\nimport cv2\nfrom PIL import Image\n\nimport torch\nimport torch.nn as nn\nimport torch.nn.functional as F\nfrom torch.optim import Adam, SGD\nimport torchvision.models as models\nfrom torch.nn.parameter import Parameter\nfrom torch.utils.data import DataLoader, Dataset\nfrom torch.optim.lr_scheduler import CosineAnnealingWarmRestarts, CosineAnnealingLR, ReduceLROnPlateau\n#!pip install adamp\nimport adamp\n\nfrom albumentations import (\n    Compose, OneOf, Normalize, Resize, RandomResizedCrop, RandomCrop, HorizontalFlip, VerticalFlip, \n    RandomBrightness, RandomContrast, RandomBrightnessContrast, Rotate, ShiftScaleRotate, Cutout, \n    IAAAdditiveGaussianNoise, Transpose\n    )\nfrom albumentations.pytorch import ToTensorV2\nfrom albumentations import ImageOnlyTransform\n\nimport timm\n\nimport warnings \nwarnings.filterwarnings('ignore')\n\nif CFG.apex:\n    from apex import amp\n\ndevice = torch.device('cuda' if torch.cuda.is_available() else 'cpu')","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Utils"},{"metadata":{"trusted":true},"cell_type":"code","source":"# ====================================================\n# Utils\n# ====================================================\ndef get_score(y_true, y_pred):\n    return accuracy_score(y_true, y_pred)\n\n\n@contextmanager\ndef timer(name):\n    t0 = time.time()\n    LOGGER.info(f'[{name}] start')\n    yield\n    LOGGER.info(f'[{name}] done in {time.time() - t0:.0f} s.')\n\n\ndef init_logger(log_file=OUTPUT_DIR+'train.log'):\n    from logging import getLogger, INFO, FileHandler,  Formatter,  StreamHandler\n    logger = getLogger(__name__)\n    logger.setLevel(INFO)\n    handler1 = StreamHandler()\n    handler1.setFormatter(Formatter(\"%(message)s\"))\n    handler2 = FileHandler(filename=log_file)\n    handler2.setFormatter(Formatter(\"%(message)s\"))\n    logger.addHandler(handler1)\n    logger.addHandler(handler2)\n    return logger\n\nLOGGER = init_logger()\n\n\ndef seed_torch(seed=42):\n    random.seed(seed)\n    os.environ['PYTHONHASHSEED'] = str(seed)\n    np.random.seed(seed)\n    torch.manual_seed(seed)\n    torch.cuda.manual_seed(seed)\n    torch.backends.cudnn.deterministic = True\n\nseed_torch(seed=CFG.seed)","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# CV split"},{"metadata":{"trusted":true},"cell_type":"code","source":"folds = train.copy()\nFold = StratifiedKFold(n_splits=CFG.n_fold, shuffle=True, random_state=CFG.seed)\nfor n, (train_index, val_index) in enumerate(Fold.split(folds, folds[CFG.target_col])):\n    folds.loc[val_index, 'fold'] = int(n)\nfolds['fold'] = folds['fold'].astype(int)\nprint(folds.groupby(['fold', CFG.target_col]).size())","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Dataset"},{"metadata":{"trusted":true},"cell_type":"code","source":"# ====================================================\n# Dataset\n# ====================================================\nclass TrainDataset(Dataset):\n    def __init__(self, df, transform=None):\n        self.df = df\n        self.file_names = df['image_id'].values\n        self.labels = df['label'].values\n        self.transform = transform\n        \n    def __len__(self):\n        return len(self.df)\n\n    def __getitem__(self, idx):\n        file_name = self.file_names[idx]\n        file_path = f'{TRAIN_PATH}/{file_name}'\n        image = cv2.imread(file_path)\n        image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)\n        if self.transform:\n            augmented = self.transform(image=image)\n            image = augmented['image']\n        label = torch.tensor(self.labels[idx]).long()\n        return image, label\n    \n\nclass TestDataset(Dataset):\n    def __init__(self, df, transform=None):\n        self.df = df\n        self.file_names = df['image_id'].values\n        self.transform = transform\n        \n    def __len__(self):\n        return len(self.df)\n\n    def __getitem__(self, idx):\n        file_name = self.file_names[idx]\n        file_path = f'{TEST_PATH}/{file_name}'\n        image = cv2.imread(file_path)\n        image = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)\n        if self.transform:\n            augmented = self.transform(image=image)\n            image = augmented['image']\n        return image","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_dataset = TrainDataset(train, transform=None)\n\nfor i in range(1):\n    image, label = train_dataset[i]\n    plt.imshow(image)\n    plt.title(f'label: {label}')\n    plt.show() ","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Transforms"},{"metadata":{"trusted":true},"cell_type":"code","source":"# ====================================================\n# Transforms\n# ====================================================\ndef get_transforms(*, data):\n    \n    if data == 'train':\n        return Compose([\n            #Resize(CFG.size, CFG.size),\n            RandomResizedCrop(CFG.size, CFG.size),\n            Transpose(p=0.5),\n            HorizontalFlip(p=0.5),\n            VerticalFlip(p=0.5),\n            ShiftScaleRotate(p=0.5),\n            Normalize(\n                mean=[0.485, 0.456, 0.406],\n                std=[0.229, 0.224, 0.225],\n            ),\n            ToTensorV2(),\n        ])\n\n    elif data == 'valid':\n        return Compose([\n            Resize(CFG.size, CFG.size),\n            Normalize(\n                mean=[0.485, 0.456, 0.406],\n                std=[0.229, 0.224, 0.225],\n            ),\n            ToTensorV2(),\n        ])","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"train_dataset = TrainDataset(train, transform=get_transforms(data='train'))\n\nfor i in range(1):\n    image, label = train_dataset[i]\n    plt.imshow(image[0])\n    plt.title(f'label: {label}')\n    plt.show() ","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# MODEL"},{"metadata":{},"cell_type":"markdown","source":""},{"metadata":{"trusted":true},"cell_type":"code","source":"# ====================================================\n# MODEL\n# ====================================================\nclass CustomResNext(nn.Module):\n    def __init__(self, model_name='resnext50_32x4d', pretrained=False):\n        super().__init__()\n        self.model = timm.create_model(model_name, pretrained=pretrained)\n        n_features = self.model.fc.in_features\n        self.model.fc = nn.Linear(n_features, CFG.target_size)\n\n    def forward(self, x):\n        x = self.model(x)\n        return x","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"model = CustomResNext(model_name=CFG.model_name, pretrained=False)\ntrain_dataset = TrainDataset(train, transform=get_transforms(data='train'))\ntrain_loader = DataLoader(train_dataset, batch_size=4, shuffle=True,\n                          num_workers=4, pin_memory=True, drop_last=True)\n\nfor image, label in train_loader:\n    output = model(image)\n    print(output)\n    break","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Helper functions"},{"metadata":{"trusted":true},"cell_type":"code","source":"# ====================================================\n# Helper functions\n# ====================================================\nclass AverageMeter(object):\n    \"\"\"Computes and stores the average and current value\"\"\"\n    def __init__(self):\n        self.reset()\n\n    def reset(self):\n        self.val = 0\n        self.avg = 0\n        self.sum = 0\n        self.count = 0\n\n    def update(self, val, n=1):\n        self.val = val\n        self.sum += val * n\n        self.count += n\n        self.avg = self.sum / self.count\n\n\ndef asMinutes(s):\n    m = math.floor(s / 60)\n    s -= m * 60\n    return '%dm %ds' % (m, s)\n\n\ndef timeSince(since, percent):\n    now = time.time()\n    s = now - since\n    es = s / (percent)\n    rs = es - s\n    return '%s (remain %s)' % (asMinutes(s), asMinutes(rs))\n\n\ndef train_fn(train_loader, model, criterion, optimizer, epoch, scheduler, device):\n    batch_time = AverageMeter()\n    data_time = AverageMeter()\n    losses = AverageMeter()\n    scores = AverageMeter()\n    # switch to train mode\n    model.train()\n    start = end = time.time()\n    global_step = 0\n    for step, (images, labels) in enumerate(train_loader):\n        # measure data loading time\n        data_time.update(time.time() - end)\n        images = images.to(device)\n        labels = labels.to(device)\n        batch_size = labels.size(0)\n        y_preds = model(images)\n        loss = criterion(y_preds, labels)\n        # record loss\n        losses.update(loss.item(), batch_size)\n        if CFG.gradient_accumulation_steps > 1:\n            loss = loss / CFG.gradient_accumulation_steps\n        if CFG.apex:\n            with amp.scale_loss(loss, optimizer) as scaled_loss:\n                scaled_loss.backward()\n        else:\n            loss.backward()\n        grad_norm = torch.nn.utils.clip_grad_norm_(model.parameters(), CFG.max_grad_norm)\n        if (step + 1) % CFG.gradient_accumulation_steps == 0:\n            optimizer.step()\n            optimizer.zero_grad()\n            global_step += 1\n        # measure elapsed time\n        batch_time.update(time.time() - end)\n        end = time.time()\n        if step % CFG.print_freq == 0 or step == (len(train_loader)-1):\n            print('Epoch: [{0}][{1}/{2}] '\n                  'Data {data_time.val:.3f} ({data_time.avg:.3f}) '\n                  'Elapsed {remain:s} '\n                  'Loss: {loss.val:.4f}({loss.avg:.4f}) '\n                  'Grad: {grad_norm:.4f}  '\n                  #'LR: {lr:.6f}  '\n                  .format(\n                   epoch+1, step, len(train_loader), batch_time=batch_time,\n                   data_time=data_time, loss=losses,\n                   remain=timeSince(start, float(step+1)/len(train_loader)),\n                   grad_norm=grad_norm,\n                   #lr=scheduler.get_lr()[0],\n                   ))\n    return losses.avg\n\n\ndef valid_fn(valid_loader, model, criterion, device):\n    batch_time = AverageMeter()\n    data_time = AverageMeter()\n    losses = AverageMeter()\n    scores = AverageMeter()\n    # switch to evaluation mode\n    model.eval()\n    preds = []\n    start = end = time.time()\n    for step, (images, labels) in enumerate(valid_loader):\n        # measure data loading time\n        data_time.update(time.time() - end)\n        images = images.to(device)\n        labels = labels.to(device)\n        batch_size = labels.size(0)\n        # compute loss\n        with torch.no_grad():\n            y_preds = model(images)\n        loss = criterion(y_preds, labels)\n        losses.update(loss.item(), batch_size)\n        # record accuracy\n        preds.append(y_preds.softmax(1).to('cpu').numpy())\n        if CFG.gradient_accumulation_steps > 1:\n            loss = loss / CFG.gradient_accumulation_steps\n        # measure elapsed time\n        batch_time.update(time.time() - end)\n        end = time.time()\n        if step % CFG.print_freq == 0 or step == (len(valid_loader)-1):\n            print('EVAL: [{0}/{1}] '\n                  'Data {data_time.val:.3f} ({data_time.avg:.3f}) '\n                  'Elapsed {remain:s} '\n                  'Loss: {loss.val:.4f}({loss.avg:.4f}) '\n                  .format(\n                   step, len(valid_loader), batch_time=batch_time,\n                   data_time=data_time, loss=losses,\n                   remain=timeSince(start, float(step+1)/len(valid_loader)),\n                   ))\n    predictions = np.concatenate(preds)\n    return losses.avg, predictions\n\n\ndef inference(model, states, test_loader, device):\n    model.to(device)\n    tk0 = tqdm(enumerate(test_loader), total=len(test_loader))\n    probs = []\n    for i, (images) in tk0:\n        images = images.to(device)\n        avg_preds = []\n        for state in states:\n            model.load_state_dict(state['model'])\n            model.eval()\n            with torch.no_grad():\n                y_preds = model(images)\n            avg_preds.append(y_preds.softmax(1).to('cpu').numpy())\n        avg_preds = np.mean(avg_preds, axis=0)\n        probs.append(avg_preds)\n    probs = np.concatenate(probs)\n    return probs","execution_count":null,"outputs":[]},{"metadata":{},"cell_type":"markdown","source":"# Train loop"},{"metadata":{},"cell_type":"markdown","source":"* 원본 노트북에서는 Adam optimizer로 학습을 진행하였는데, 찬란님의 Adamp 사용 노트북을 보고 Adamp를 적용해보았다. https://www.kaggle.com/seriousran/adam-vs-adamp-iclr-2021"},{"metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true},"cell_type":"code","source":"# ====================================================\n# Train loop\n# ====================================================\ndef train_loop(folds, fold):\n\n    LOGGER.info(f\"========== fold: {fold} training ==========\")\n\n    # ====================================================\n    # loader\n    # ====================================================\n    trn_idx = folds[folds['fold'] != fold].index\n    val_idx = folds[folds['fold'] == fold].index\n\n    train_folds = folds.loc[trn_idx].reset_index(drop=True)\n    valid_folds = folds.loc[val_idx].reset_index(drop=True)\n\n    train_dataset = TrainDataset(train_folds, \n                                 transform=get_transforms(data='train'))\n    valid_dataset = TrainDataset(valid_folds, \n                                 transform=get_transforms(data='valid'))\n\n    train_loader = DataLoader(train_dataset, \n                              batch_size=CFG.batch_size, \n                              shuffle=True, \n                              num_workers=CFG.num_workers, pin_memory=True, drop_last=True)\n    valid_loader = DataLoader(valid_dataset, \n                              batch_size=CFG.batch_size, \n                              shuffle=False, \n                              num_workers=CFG.num_workers, pin_memory=True, drop_last=False)\n    \n    # ====================================================\n    # scheduler \n    # ====================================================\n    def get_scheduler(optimizer):\n        if CFG.scheduler=='ReduceLROnPlateau':\n            scheduler = ReduceLROnPlateau(optimizer, mode='min', factor=CFG.factor, patience=CFG.patience, verbose=True, eps=CFG.eps)\n        elif CFG.scheduler=='CosineAnnealingLR':\n            scheduler = CosineAnnealingLR(optimizer, T_max=CFG.T_max, eta_min=CFG.min_lr, last_epoch=-1)\n        elif CFG.scheduler=='CosineAnnealingWarmRestarts':\n            scheduler = CosineAnnealingWarmRestarts(optimizer, T_0=CFG.T_0, T_mult=1, eta_min=CFG.min_lr, last_epoch=-1)\n        return scheduler\n\n    # ====================================================\n    # model & optimizer\n    # ====================================================\n    model = CustomResNext(CFG.model_name, pretrained=True)\n    model.to(device)\n    !pip install adamp\n    #optimizer = Adam(model.parameters(), lr=CFG.lr, weight_decay=CFG.weight_decay, amsgrad=False)\n    optimizer = AdamP(model.parameters(), lr=CFG.lr, weight_decay=CFG.weight_decay)\n    scheduler = get_scheduler(optimizer)\n\n    # ====================================================\n    # apex\n    # ====================================================\n    if CFG.apex:\n        model, optimizer = amp.initialize(model, optimizer, opt_level='O1', verbosity=0)\n\n    # ====================================================\n    # loop\n    # ====================================================\n    criterion = nn.CrossEntropyLoss()\n\n    best_score = 0.\n    best_loss = np.inf\n    \n    for epoch in range(CFG.epochs):\n        \n        start_time = time.time()\n        \n        # train\n        avg_loss = train_fn(train_loader, model, criterion, optimizer, epoch, scheduler, device)\n\n        # eval\n        avg_val_loss, preds = valid_fn(valid_loader, model, criterion, device)\n        valid_labels = valid_folds[CFG.target_col].values\n        \n        if isinstance(scheduler, ReduceLROnPlateau):\n            scheduler.step(avg_val_loss)\n        elif isinstance(scheduler, CosineAnnealingLR):\n            scheduler.step()\n        elif isinstance(scheduler, CosineAnnealingWarmRestarts):\n            scheduler.step()\n\n        # scoring\n        score = get_score(valid_labels, preds.argmax(1))\n\n        elapsed = time.time() - start_time\n\n        LOGGER.info(f'Epoch {epoch+1} - avg_train_loss: {avg_loss:.4f}  avg_val_loss: {avg_val_loss:.4f}  time: {elapsed:.0f}s')\n        LOGGER.info(f'Epoch {epoch+1} - Accuracy: {score}')\n\n        if score > best_score:\n            best_score = score\n            LOGGER.info(f'Epoch {epoch+1} - Save Best Score: {best_score:.4f} Model')\n            torch.save({'model': model.state_dict(), \n                        'preds': preds},\n                        OUTPUT_DIR+f'{CFG.model_name}_fold{fold}_best.pth')\n    \n    check_point = torch.load(OUTPUT_DIR+f'{CFG.model_name}_fold{fold}_best.pth')\n    valid_folds[[str(c) for c in range(5)]] = check_point['preds']\n    valid_folds['preds'] = check_point['preds'].argmax(1)\n\n    return valid_folds","execution_count":null,"outputs":[]},{"metadata":{"trusted":true},"cell_type":"code","source":"# ====================================================\n# main\n# ====================================================\ndef main():\n\n    \"\"\"\n    Prepare: 1.train  2.test  3.submission  4.folds\n    \"\"\"\n\n    def get_result(result_df):\n        preds = result_df['preds'].values\n        labels = result_df[CFG.target_col].values\n        score = get_score(labels, preds)\n        LOGGER.info(f'Score: {score:<.5f}')\n    \n    if CFG.train:\n        # train \n        oof_df = pd.DataFrame()\n        for fold in range(CFG.n_fold):\n            if fold in CFG.trn_fold:\n                _oof_df = train_loop(folds, fold)\n                oof_df = pd.concat([oof_df, _oof_df])\n                LOGGER.info(f\"========== fold: {fold} result ==========\")\n                get_result(_oof_df)\n        # CV result\n        LOGGER.info(f\"========== CV ==========\")\n        get_result(oof_df)\n        # save result\n        oof_df.to_csv(OUTPUT_DIR+'oof_df.csv', index=False)\n    \n    if CFG.inference:\n        # inference\n        model = CustomResNext(CFG.model_name, pretrained=False)\n        states = [torch.load(OUTPUT_DIR+f'{CFG.model_name}_fold{fold}_best.pth') for fold in CFG.trn_fold]\n        test_dataset = TestDataset(test, transform=get_transforms(data='valid'))\n        test_loader = DataLoader(test_dataset, batch_size=CFG.batch_size, shuffle=False, \n                                 num_workers=CFG.num_workers, pin_memory=True)\n        predictions = inference(model, states, test_loader, device)\n        # submission\n        test['label'] = predictions.argmax(1)\n        test[['image_id', 'label']].to_csv(OUTPUT_DIR+'submission.csv', index=False)","execution_count":null,"outputs":[]},{"metadata":{"_uuid":"d629ff2d2480ee46fbb7e2d37f6b5fab8052498a","_cell_guid":"79c7e3d0-c299-4dcb-8224-4455121ee9b0","trusted":true},"cell_type":"code","source":"if __name__ == '__main__':\n    main()","execution_count":null,"outputs":[]}],"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat":4,"nbformat_minor":4}