Speaker_Diarization.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xgeQ7hPvqwks"
      },
      "source": [
        "# 화자 분리(Speaker Diarization)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LMmUI654-3im"
      },
      "source": [
        "* Speaker Diarization(화자 분리)는 오디오에서 각 부분에서의 화자를 인식하는 기술\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eK5Su1qK4UH-"
      },
      "source": [
        "## UIS-RNN\n",
        "\n",
        "* 대표적인 모델 Unbounded Interleaved-State Recurrent Neural Network (UIS-RNN)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9MW5OpV-AYT-"
      },
      "source": [
        "![58151e00-087b-11ea-839c-6ecb32aea6b9.jpg](data:image/jpeg;base64,/9j/4AAQSkZJRgABAQEAYABgAAD/4RDoRXhpZgAATU0AKgAAAAgABAE7AAIAAAAKAAAISodpAAQAAAABAAAIVJydAAEAAAAUAAAQzOocAAcAAAgMAAAAPgAAAAAc6gAAAAgAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAFF1YW4gV2FuZwAABZADAAIAAAAUAAAQopAEAAIAAAAUAAAQtpKRAAIAAAADNDkAAJKSAAIAAAADNDkAAOocAAcAAAgMAAAIlgAAAAAc6gAAAAgAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAADIwMTk6MTE6MTYgMTQ6MDk6NDkAMjAxOToxMToxNiAxNDowOTo0OQAAAFEAdQBhAG4AIABXAGEAbgBnAAAA/+ELHGh0dHA6Ly9ucy5hZG9iZS5jb20veGFwLzEuMC8APD94cGFja2V0IGJlZ2luPSfvu78nIGlkPSdXNU0wTXBDZWhpSHpyZVN6TlRjemtjOWQnPz4NCjx4OnhtcG1ldGEgeG1sbnM6eD0iYWRvYmU6bnM6bWV0YS8iPjxyZGY6UkRGIHhtbG5zOnJkZj0iaHR0cDovL3d3dy53My5vcmcvMTk5OS8wMi8yMi1yZGYtc3ludGF4LW5zIyI+PHJkZjpEZXNjcmlwdGlvbiByZGY6YWJvdXQ9InV1aWQ6ZmFmNWJkZDUtYmEzZC0xMWRhLWFkMzEtZDMzZDc1MTgyZjFiIiB4bWxuczpkYz0iaHR0cDovL3B1cmwub3JnL2RjL2VsZW1lbnRzLzEuMS8iLz48cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0idXVpZDpmYWY1YmRkNS1iYTNkLTExZGEtYWQzMS1kMzNkNzUxODJmMWIiIHhtbG5zOnhtcD0iaHR0cDovL25zLmFkb2JlLmNvbS94YXAvMS4wLyI+PHhtcDpDcmVhdGVEYXRlPjIwMTktMTEtMTZUMTQ6MDk6NDkuNDg3PC94bXA6Q3JlYXRlRGF0ZT48L3JkZjpEZXNjcmlwdGlvbj48cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0idXVpZDpmYWY1YmRkNS1iYTNkLTExZGEtYWQzMS1kMzNkNzUxODJmMWIiIHhtbG5zOmRjPSJodHRwOi8vcHVybC5vcmcvZGMvZWxlbWVudHMvMS4xLyI+PGRjOmNyZWF0b3I+PHJkZjpTZXEgeG1sbnM6cmRmPSJodHRwOi8vd3d3LnczLm9yZy8xOTk5LzAyLzIyLXJkZi1zeW50YXgtbnMjIj48cmRmOmxpPlF1YW4gV2FuZzwvcmRmOmxpPjwvcmRmOlNlcT4NCgkJCTwvZGM6Y3JlYXRvcj48L3JkZjpEZXNjcmlwdGlvbj48L3JkZjpSREY+PC94OnhtcG1ldGE+DQogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgIDw/eHBhY2tldCBlbmQ9J3cnPz7/2wBDAAcFBQYFBAcGBQYIBwcIChELCgkJChUPEAwRGBUaGRgVGBcbHichGx0lHRcYIi4iJSgpKywrGiAvMy8qMicqKyr/2wBDAQcICAoJChQLCxQqHBgcKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKir/wAARCAMWBa4DASIAAhEBAxEB/8QAHwAAAQUBAQEBAQEAAAAAAAAAAAECAwQFBgcICQoL/8QAtRAAAgEDAwIEAwUFBAQAAAF9AQIDAAQRBRIhMUEGE1FhByJxFDKBkaEII0KxwRVS0fAkM2JyggkKFhcYGRolJicoKSo0NTY3ODk6Q0RFRkdISUpTVFVWV1hZWmNkZWZnaGlqc3R1dnd4eXqDhIWGh4iJipKTlJWWl5iZmqKjpKWmp6ipqrKztLW2t7i5usLDxMXGx8jJytLT1NXW19jZ2uHi4+Tl5ufo6erx8vP09fb3+Pn6/8QAHwEAAwEBAQEBAQEBAQAAAAAAAAECAwQFBgcICQoL/8QAtREAAgECBAQDBAcFBAQAAQJ3AAECAxEEBSExBhJBUQdhcRMiMoEIFEKRobHBCSMzUvAVYnLRChYkNOEl8RcYGRomJygpKjU2Nzg5OkNERUZHSElKU1RVVldYWVpjZGVmZ2hpanN0dXZ3eHl6goOEhYaHiImKkpOUlZaXmJmaoqOkpaanqKmqsrO0tba3uLm6wsPExcbHyMnK0tPU1dbX2Nna4uPk5ebn6Onq8vP09fb3+Pn6/9oADAMBAAIRAxEAPwD6RooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAoorK8TeIbLwp4Zvtc1QkW1lEZHC9WPQKPckgD60m0ldjSbdkatFfOvh7xl8Zvir9s1bwdcaXoWlQSmONJ41YSEc7QzI5Y4IyflHNdD8JfjBrWveL7zwV46tIYNatfMCTQrtDtH99GUcZxkgjggdKpJt267ktpK/Q9porwT4NfErxZ4r+LGt6Nr+rfa7C1gneGH7NEm0rMqj5lUE8Ejk0vx/+JPizwT4q0Wz8Mar9hgurYvMn2aKTc2/GcupI49KS15P7239fIq2sl/Ke9UV4Z8efij4l8A+J9FttAvI4LW5gMs6tbpIWw+OCw44qfwV4t+K3jHxrYaxPpA0zwVOXbYViDNHsbYx3HzDk7eVAB+lJO6v6/gKXu7ntlFfKnhv49fEXVNQv9HsbdNa1a5by7CNbZFWHBO5ztxnAx1OB1PTnr9T+LXjP4c/D+KLxzaR3XjDULqQWkD+VsjgAXDsITgjJIAzk+vFPpcbVpcp75RXzzq+tfHnwp4d/wCEt1e80q4sIwstxpggQvChP8WEBwM84ckV2N58Xn1H4C3njnw7DFFfWwVJLecF1il8xVZTgjIw2R04Ioeib7CWrSXU9Vor5o0r4tfF74geHceCdFg+0WrN9tv4Yowuc5VEErFc7eT1PPbvsfBr4weKfFevaj4T8UNEdUW3la1ujCImSVOCjqBjjr07HrTs9V8wut/keg6z8YdA0P4j23gq7s9SfUrmSGNJYoozCDJjbklw3fn5a76vivxVp3jqH472FnrWs2Nx4rae2EF/EgEKuceWSPLA44z8h/GvrbwXaeJbHwvbweNtQttR1hWfzri1UKjDcduAEToMD7oojrC/mxS0nb0N6iuI+LXj6T4c+BZNYtbZLm7kmW3tkkzsDsCctjnACnjvXlWka58dvEPg5fF+jaxpNzZuryJp0NvG0zBScgKIzzx93fn8alO9/Iq1reZ9EXM62trLcSAlIkLsF6kAZ4rivh38WtC+Jc9/FoVpqNu1iqNJ9tjRQdxOMbXb+6fSuT/t34meOfhTJJZWg8Ja9azMt9/aNm8UdzB5ZJMayRseSR9CDzXj3wC07xxfa1qJ8C6xZabFEYG1BbpAxmj3NgLmN8H73p161UVeo4v+vP0Jk7U1Jf15H2NRXifxE+Jfjh/iEngz4daWI5dyxyajcW5ZS5XcdpI2hQOpIPIOPflfE/xE+L3wm1bTm8aX2l61aXm4qsUShWCkbl3KiMrAMOcEc96lNOz7lNNOx9LUV5f8Vfi8ngbwPp2paXbpPqOsIGs45vuxrtDF2A643AY7k1x2n638dtP8L/8ACYalLpuoab5H2l9LeJEnWHGdw2oMEDnBYn1B6UPS9+glra3U+gaK8R+AHxK8ReNIfEU3i/UkuItPWF428iOIRg+YWJ2qM/dHX0rLtvib8Rvit4mv7L4WfYdI0qwPN9eoGLgkhS2VbBbBIULx3NU007fMSaab+R9BUV4b4L+K/irRfiUPAXxTjtnvJmCW1/bqFDMwyucAAq3QEAEHgj09yotomtgvrYKKKKQwooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKK82+OeteKfDXgD+3PB2omyls51+1AQRy74m+XPzq2MMV6epqT4d/Eb+3PgsPFWszrJc2FvL9ucKF3PECScDAGRtOAO9K65ZS7Ds7pdz0WivCPgh8RvE3iVdc8QePPEVvDoNmVhj+0RwW8Syuc/6wKp4GByedwr1MfEXwSzAL4x0AknAA1SHn/x6qaasTfc6SijIxntXnMvx++GcMzxS+JdrxsVYfYLk4I4P/LOlfWw+lz0aivNv+Gg/hj/ANDN/wCSFz/8brS+IHg7UPiFounpoPi288PRqTK0tqjn7QjKMBgHTjvzmh3tdIFa9mM1H4veG9O+JFr4JK3lzqdw6RmS3RGhiduiuS4II6nAPUV3VfGui+Em8D/tOaPoD6gdRa2v4GNyYvLMhdA/3dzf3sda+yqas6al3v8AoJ3U2vJBRRRSGFFFFABRRXm3xz1rxT4a8Af254O1E2UtnOv2oCCOXfE3y5+dWxhivT1NTKXKrsaV3Y9Jorzr4d/Eb+3PgsPFWszrJc2FvL9ucKF3PECScDAGRtOAO9cb8EPiN4m8SrrniDx54it4dBsysMf2iOC3iWVzn/WBVPAwOTzuFXb3nHtr/XqTf3VLvp/Xoe70VzY+IvglmAXxjoBJOABqkPP/AI9XSZGM9qQworzmX4/fDOGZ4pfEu142KsPsFycEcH/lnTP+Gg/hj/0M3/khc/8AxujcNjp/G/jnRvAHh9tW16VhGW2RQxDMkz/3VH9TwK8ctv2tNMfUAl54VuobPdgzR3ivIB67CoH/AI9W78Y/hj4g+KeraFd6HeWK6RbwbmW4ldGJdgSwAQ/wheuKsfHUeFPDvwZl0We3tIZCiRaXbIgDq6kfMo6gAZyffHepbcU5S77FJKTUV956foGvad4n0K21fRblbmyuk3RyAY9iCOxB4IrRryD9mbT76x+EokvldI7q9kmtlb/nnhVyPYsrV6/Ws0lKyM4u6uFFFFQUFFFFABRVe/iuJ9NuYrG4NtcvEywzBQ3luRw2CCDg84Irw/4HfFDxPr/jjWfDHjq/+1XkCEwAwRxFHjbbIvyKueoPP900R1ly/MHpHmPeKK8D8U/EjxfqX7Q1r4K8H6t9jsI5IoboLbRSZwN8rZdCRheOvUV61L8QvBcEzxTeL9BjkjYq6PqcIKkcEEbuDQtYqXcHpJxOioqnpmr6brdn9r0bULXULbcV860nWVMjqNykjNc74q+Kfg3wVqiad4m1j7FdyRCZY/ss0mUJIByiEdQe9D03DfY66orm5hs7WW5u5VhghQySSOcKigZJJ9AK87/4aD+GP/Qzf+SFz/8AG6uaheaP8afhxqun+EPEBit7hxbS3gtJPkIKsy7H2E5Ugdcc0nzWfKNWurnn+tftW6LZ6o8Gi+HrnUrVG2/aZLkQb/dV2scfXB+lemfDv4naF8SdMluNGMkNzb4FxZzgCSLPQ8cFTg8j9K5LTbbwR8APBMeleKL+O7m1FpXab+z2LXmMfIVG4AAFRhmA5+tcR+zVodzdeONf8VWVq1nociSQW8ZPUtIGCgf7KrjPvVRs2476bkyulzba7H0rRRRSGFFFFABRRRQAUUUUAFFFFABRRRQAUVXv4rifTbmKxuDbXLxMsMwUN5bkcNggg4POCK8P+B3xQ8T6/wCONZ8MeOr/AO1XkCEwAwRxFHjbbIvyKueoPP8AdNEdZcvzB6R5j3iivA/FPxI8Xal+0Na+C/B+rfY9PjkihugttFJnA3ytl0JGF469RXrLfEXwSjFX8YaArKcEHU4QQf8AvqhaxUu4PSTidHRVXTdU0/WLFbzSL62v7VyQs9rMsqMQcHDKSOtc34q+Kfg3wVqiad4m1j7FdyRCZY/ss0mUJIByiEdQe9D03DfY66ivNv8AhoP4Y/8AQzf+SFz/APG66TRPiF4W8ReG73X9I1ZJtLsSy3Fy8UkQjKqGPDqD0I7UdLh1sP8AG/jjRvAHh5tY1+STyt4jjhhAaWZj/CoJAPGTyRwKZ4E8c6Z8QvDh1rRYLuC2E7QbLtFV9ygE8KzDHPrXzt4yGqfF3RvEnj6/8608OaLC0OjWx4819yguf5k+uB2Nemfsw/8AJIW/7CM38kpwTfNzdr/iE9LW72/A9iooopAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABXm/x+0271P4MawlirO8JjndFHJRHBb8hz+FekUjosiMjqGVhhlIyCPSplHmViouzueO/s169plx8KE05LmFLvT55TcxM4DAM24OR6YOM+xrc8M/GfSPF3xEufCui2F1P9n8wtqCsphKpxuGDnBOAD7iszW/2bPAOsai93FHf6YZG3NDYzqsZPsrq2PoMCuy8FfDrw38P7OSDw3Y+U82POuJWLyy46ZY9vYYHtWnNzPml2M7cqsu589/AS9t9P8Aj/rlveyrDJcR3UMQc43OJlO0e+FP5Uz9p7VbO++JGkWlpPHNLZWgWcIwOxmkJCn0OMHHuK9D8cfs1aR4r8T3OtabrU2kPeSGW4h+zCZC56svzKVyee/NB/Zd8JHT7GAajqST2xZpbhCgadiR1BU4AxwB6nrUU9qd/s/8E0k/em19o4n9qwBvGfh1T0NowP8A38r6btokh06KKJQkaRBVUDgADgVxXxA+EGgfEfVLK/1y71KCWyjMcYs5Y1UgtnncjV3QQLGEHQDFCX7px63ZMtZqXl/kfLX7L0Mb/EzxBKygvHZMFYjpmVc/yqT9qTTrqLx7oGqPJJBZy2ogW4QE+U6SFiR74cH8K9m8BfB/QPh3rV7qeiXmpTzXsXlSLdyxsoG7dxtRTnI9a6fxJ4Y0fxdosmleIbGO8tJDnY+QVbsysOVPuKctoW+z/wAH/Mf2p9n/AMA8C8S/DrUNP8C3Ot6z8cdWvNFkgJ2ASSrdKR9xVNzhyemPzrO0jR7TTv2VPFt1pd5f3Vle3MbRG9sltmyskakqFlkBGRjORyDxXoVt+zD4Bg1AXEj6tcRhs/ZpbpfLPtlUDY/4FXoOv+CNH1/wNL4SdJLHS5I0iCWW1GjVWDALkEDlR2NKSvGVuv8AncI6Si30PPv2YVVfhExAALajMSfXhK87+HIC/teayAMD7XqHT/gVfQXgTwNpvw98N/2Los93Pbec02+7dWfc2M8qqjHHpWLovwe0DQviPc+NbS81J9SuZJpHilljMIMud2AEDd+Pm/OtG17VS6Wa/BEJP2bj53/FniPxQvbbTf2sNOvb+ZILaC4sZJZXOFRQFySewFfUWm6pYazYpe6RfW99auSFntpVkRiDg4ZSQea4rx78GfCvxE1GPUNZW8tr1IxGbiylCM6joGDKwOM9cZrofBfhCw8C+F7fQdImuZrW3Z2V7plZzuYsclVA6n0qIaQ5X3f4lT1nzLsvwOa+NWreFtM8CCPxzpd/qGmXdwsQ+wqpeGTBKvlmXHQ88+mMGvHh8KLK18DyeNvhn8QbzTrUW5ufJupRE64GfLeSJgA4xjGDzX0rreiab4j0efS9btI7yyuF2yQyDg+hz1BHYjkV5TL+y74CkuzMs+sxJnPkJdJsHtkoW/Wps9fPqVdaFT4F+Pdf8a+AfEcHiSc3kmmxhYbp1AeQOj/KxHUjaOevzc1wv7L3iHR9E1jXotZ1S0sHuo4FgF1MsfmkM2Qu4jJ5HHWvpHwz4S0Twfoa6R4fsUtbQEllyWMjHqzE8sT7155N+zZ4HfXl1S0l1WyZJxOtvb3CeUpDZwAyE4z2zWiaVXm6NJGbTdPl6p3ON1z4geNfHPxuuPAvhrXV8N2ME8sHmpGDI3lg7myeSTg4AIGPzrh/jz4Qk8JXejpqXi7UfEmqXCSNKb6TPkxgrt2qSSoJ3d+1e6eNvgF4Z8ZeIn10Xd9pWoSkNM9oy7ZGH8WCOG9wR+dZ837Mvgq6sEhub3WpLoPvkvTdIZZOMbTuQrgewz7msknyrut/M0bXM+x53+0JpN0/gfwHrEaM9rDp628jAcIzRoy5+uD+Veo3Hxo8FH4UtqA1i0a5fT/LGmiQef5pTbs8vrjPGcYxzmrXxM1lfh78NrK1fws3ivRoo1s7pbmcL5aKoCO+I2BzjrgYOPWvKrTxj8DNL8Ny61pHh4QeITbsYrCVJ5vKmK8YZ8xgA/xDBxzjPFVUakqiW1366igmnBvdL5FX9nixuNU8F/ESwss/aLqwSKLHdmSYD9TWB8FvCWp+KJtU07TPiBf+Er2F1LWdvvBuByC2FlTJUjB4OM16l+y54W1HSPDOq6zqNu9vHqkkYtlkXBdEDfPj0Jfg98V13jD4DeCfGWqSald291p97Md002nyiPzT6lWVlz7gAnvVzVp38kiI6x+d/wBDxS88G2cfxp0fRr34gaz4l16O4hImTTxOsW1t2xpHudy7QCTgHAPrxX1pXE+BfhJ4U+Hsr3Gh2ksl7Iuxry7k8yXb6DAAUfQDNdtRtFRH9psKKKKkYUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAZ3iDR7fxD4d1DSLwZhvrd4H9twxn8OtfE9v4r1Hwd4L8W+ArlXWW7vEjbHSMxsRL/31tUV9m+MPFmn+CPDM+uays7WkDIriBAz/MwUcEjua+V7W0sfjH+0b9r0Ownj0ie4S5uhKoB8uNV3swBIG4jHX+KpjHmqcvR6P8ypPlhd9NV+R7v8PPhnpSfBHT/DXiOx+0RXyLd3sPmNGWkYhwCVIIxhR1/hrzyb4W+Cte+NUPhjwto32PT9DjFzrFyt1NIZXONkA3uQO2SOevpX0eflQ7R0HAFeXfBXwhr/AIdXxLqfi+x+yaprGomcgzRylk5IOUYjqzcVpzXqc23X9F936GdrU+X5fq/68z1HAVcAcAV86XXxg+CkV5NHP8O98iSMrt/YlkdxB5OS9fRlFR1uX0sfN3/C5Pgh/wBE4/8AKHZf/F19E2EsM+m201rH5UEkKNGm0DapAIGBwOKsUVV9LE9T5a8Tf8noWn/X9af+iUr6lrgb74PaBqHxOj8dTXmpLqccscohSWMQ5RQo42bsYH96u+pR0pRi91/wBy1m5eSCiiigAooooAKzvEGj2/iHw7qGkXgzDfW7wP7bhjP4da0aw/GHizT/AAR4Zn1zWVna0gZFcQIGf5mCjgkdzUys1ZlRundHxlb+K9R8HeC/FvgK5V1lu7xI2x0jMbES/wDfW1RX058PPhnpSfBHT/DXiOx+0RXyLd3sPmNGWkYhwCVIIxhR1/hrwi1tLH4x/tG/a9DsJ49InuEuboSqAfLjVd7MASBuIx1/ir7DPyodo6DgCrV/ZXlu7fh/X4EO3tLR2X6nzhN8LfBWvfGqHwx4W0b7Hp+hxi51i5W6mkMrnGyAb3IHbJHPX0r6PwFXAHAFeXfBXwhr/h1fEup+L7H7JqmsaiZyDNHKWTkg5RiOrNxXqVGvIl/V3/lt8hu3M38vu/z3PnO6+MHwUivJo5/h3vkSRldv7EsjuIPJyXqL/hcnwQ/6Jx/5Q7L/AOLr6RoqVohvVnlXxY+L9t8PfClgui26PqmpQB7OGRQEt48D52UemQAo7j0FeXeA/C3hjxTqCeL/AIv+O9Hvby4xImmzarCCB1Al+b5R/wBMxgDv3FewePvgl4d+IviCPV9cv9WhnjgWBUtJo1QKCT0aNjnLHvXMf8MqeCP+gr4g/wDAiD/4zTju5Pf8hPZRWx65oeraJq1gD4b1CwvbS3xEDYTJJHHgDC/ISBgY4rRrlPh98O9J+G+jXGm6HcXtxDcT+e7XjozBtoXA2qoxgDtXV03a+glewUUUUhhRRRQAV8tfFAN8LP2jdP8AFtvG32K+IupFQfeyNkyj3x831avqWvk/49/E3w/4/ttP0bQrG9bUrC9dWkmiC4/hKLgknLAfkKhtqUXHcpJOLT2Oi/Zt0ibxB4t8SePdTXdJLK0UTH/npI2+Qj6DaP8AgVaHxg+Hnw/8N6Or6b4ZW48Sa9d/Z7FTe3HMzn5pCvmYwM9MYyR2r074U+Ez4M+Guk6TNHsuvK866HfzX+Zgfpnb+Fc74l8IeINe/aB8NazJYb/Dmj2zN55mjwJiHP3N27rs5x2rWcY80YLZafLd/eRGT5ZTe71+fQ634feDLXwF4Ls9Cs3aQxDfPKT/AKyVvvN7DPQegFcR8U/H/wAOfC3imGy8beEv7Zv3tVkS4/s23uNsZZgF3SMD1B46c165RUybk7jilFWPm3/hcnwQ/wCicf8AlDsv/i66qL4s6Fpnwj1DxX8N/CcdrZ2moLBc2klvHaglguZMREg/eQZNe0VW1LTrTV9MuNP1KBLi0uYzHNE4yHU9RSd7Ow1a6ucB4U13wn8YvAdvq3iTTNJnmtxItxb3KrIbM55wzcqCApzx+leL/BK5TT/2jL/TvCs7voMzXSYViyNCu4xt74IXB9/euu1T9k7SbjUGl0jxNdWVqzZEE1qJyo9A25f1Br0r4cfCfQPhray/2V5t1fXChZ724xvYddqgcKuecfmTgVcbc3Pto9PUmV+Xl31O4oooqRhRRRQAUUUUAFFFFABRRRQAUUUUAFfLXxQDfCz9o3T/ABbbxt9iviLqRUH3sjZMo98fN9Wr6lr5P+PfxN8P+P7bT9G0KxvW1KwvXVpJoguP4Si4JJywH5CobalFx3KSTi09jov2bdIm8Q+LPEnjzU03STSvDEx/vyNvkI+g2j/gVWfi38NvAPh6xstO8P8AhpX8S+ILoW1juvbgiMkjdKVMmMDI6jGT6CvVfhT4TPgz4a6TpM0ey68rzrod/Nf5mB+mdv4VzupeEPEGrftHaZ4iu7DPh/SrEpb3BmjIMpVs/Ju3dX64/hFayjFTjBbLT5LX8X+ZEZNxlPq9fnsvuO08FeE7PwR4QsdB05meK1T5pHPMjk5ZvbJJ47VwPxT8f/Dnwt4phsvG3hL+2b97VZEuP7Nt7jbGWYBd0jA9QeOnNeuUVMm5O442irHzb/wuT4If9E4/8odl/wDF1c8SeX8WPg69t8GfDb6fbjVgt7ZrHb2QlCx53EB9rDJTvngccV9DUUmk1Z/1qCbTuj5F8Q6V8cNE+Gt1puuWyWvhe1txHNCn2HCxgjH3PnPOORya6L9mj/hOvOg8j/kSvMn87/Uf6/YP+2vXb04r6C8UeHbTxb4YvtC1KSaK1vY/LkeBgHAyDwSCO3oaoeBPA2mfD3w4dF0We7ntjO0++7dWfcwAPKqoxx6VcZNSk31X9X/rcUl7qS7/ANf12OloooqRhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFACModSrgMpGCCMgisRPBHhSO8+1p4Y0ZbnO7zhp8QfPru25rcooAAMDAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAq6jplhrFi9nq1jbX1q5BeC6hWVGwcjKsCDzUWlaFpGhwtFoml2WnRucslpbpEGPuFAq/RQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAVV1HTLDWLF7PVrG2vrVyC8F1CsqNg5GVYEHmrVFAFDStC0jQ4Wi0TS7LTo3OWS0t0iDH3CgVfoooAKKKKACiiigAooooAKKKKACiiigAooooAKxrTwf4Z0/UPt9h4d0m2vNxb7RDYxpJk9TuC5zWzRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFY1p4P8ADOn6h9vsPDuk215uLfaIbGNJMnqdwXOa2aKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAK+TvFPiLW4f2q4tPi1i/SyOu2UZtlunEe0tFlducYOTx719Y15bqfwK0jU/ievjaXVr1Ltb6G9+zqqeXujKkDpnB2CgD1KiiigAooooAKKKKACiiigAooooAKKKKACiiigD5U/aV8Q61pXxRtoNM1e/s4TpkTGO3uXjUnzJOcAgZ4FfU1oSbKAk5JjXJP0rzX4i/A3SfiP4mj1rUdVvbSWO2W3EcCoVIVmOeR1+avTYoxFCkYOQihQT7UAOooooAKKKKACiiigAooooAKKKKACiiigAr55/ao1rVNIfwt/ZOpXlj5ou/M+zTtHvx5OM7SM4yfzr6Grgfid8JtO+KB0w6nqN1Zf2d5uz7Oqnf5mzOcjtsH50AafwuuJrv4VeG7i6mknmk0+JnkkYsznb1JPJNdXWX4a0KHwz4X07RLaV5orC3WBJJANzBRjJxWpQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAV8nfDbxFrd1+0+ljc6xfzWf8AaN+v2eS6do8CObA2k4wMDH0r6xry3w78CtI8OfEoeMrfVr2a6FxPP9nkVNmZVdSOBnjefyoA9SooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA8h/aW1K+0r4X20+mXtxZzHVIlMlvK0bEeXJxkEHHAq5+ztqF5qfwjguNSu57uc3cwMtxKZGIBGBknNdN8RfAFn8R/DUejajeT2kUdytwJIACxKqy457fMam8AeCLT4feFU0KwuprqFJXlEk4AbLHOOOKAOmooooAKKKKACiiigAooooAKKKKACiiigAorN8SXc1h4V1a7tH8ue3sppYnwDtZUJBweOor45t/jr8VbuQpa+IJpnAyVj0+BiB68R00m3ZCbSV2fbFFfF/8Awub4w/8AQVvf/BXD/wDGqP8Ahc3xh/6Ct7/4K4f/AI1V+yqfyv7jP21P+ZfefaFFfF//AAub4w/9BW9/8FcP/wAao/4XN8YB/wAxW9/8FcP/AMao9lU/lf3B7an/ADL7z7Qor5u+AvxT8ZeMviNJpniTWTe2a2EkoiNtEnzhkAOUQHue9fSNZmoUUUUAFFFFABRWZ4lu57Dwnq15aP5c9vZTSxPgHayoSDg8dRXx1b/HX4q3blLXxBNMwGSsenwMQPXiOmk27ITaSuz7Yor4v/4XN8Yf+gre/wDgrh/+NUf8Lm+MP/QVvf8AwVw//Gqv2VT+V/cZ+2p/zL7z7Qor4v8A+FzfGH/oK3v/AIK4f/jVH/C5vjAOuq3v/grh/wDjVHsqn8rD21P+ZfefaFFfN/wE+KfjLxl8RZdM8S6yb2zWwkmERtok+cMgByiA9Ce9fSFZmoUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRWX4nvJ9P8ACOsXtm/l3FvYzyxPgHayxkg4PB5Hevj60+OHxZ1CRo7DXLm6dRuKw6dA5A9cCOk2oq7Czex9q0V8bf8AC2fjV/z+an/4J4v/AI1R/wALZ+NX/P5qf/gni/8AjVZ+3pfzL7yuSXY+yaK+Nv8AhbPxq/5/NT/8E8X/AMapD8WvjUASb3UgB1J0iL/41R7al/MvvDkl2PsqivnD4B/FLxl4z+Ik+m+JdZN7Zrp8kwiNtFHhw6AHKID0Y9+9fR9akhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFZXiq8n07wdrN7ZSeVcW1hPNE+AdrrGxBweDyO9fH1t8c/iveuUs9euLhlGSsWnQMQPXiOgaTbsj7Wor4y/4XD8ZP8AoJah/wCCmH/41R/wuH4yf9BLUP8AwUw//GqdmX7OfZn2bRXxl/wuH4yf9BLUP/BTD/8AGqQ/GL4yAEnU78AdSdKh/wDjVFmL2c+zPs6ivnL4BfFHxj40+IVxpvibWTfWiadJMsZt4o8OHjAOUUHox/Ovo2kQFFFFABRRRQAUVk+K72403wZrV9ZSeVc22nzzQvgHa6xsVODweQOtfIFn8cPizqEjR2GuXN06jcyw6dA5A9cCOk2oq7Gk3sfatFfG3/C2fjV/z+an/wCCeL/41R/wtn41f8/mp/8Agni/+NVn7el/MvvHyS7H2TRXxt/wtn41f8/mp/8Agni/+NUh+LfxpVSzXupAAZJOkRcf+QqPbUv5l94ckux9lUV85fAH4o+MfGnxCudN8Tayb60TTpJljNvFHhxJGAcooPRj+dfRtakhRRRQAUUUUAFFZHiy9uNN8F63fWUnlXNrp880MmAdrrGxU4PBwQOtfINn8b/izqEjJYa5c3TqMssOmwOQPXAjpNqKuxpN7H2rRXxt/wALZ+NX/P5qf/gni/8AjVH/AAtn41f8/mp/+CeL/wCNVn7el/MvvHyS7H2TRXxt/wALZ+NX/P5qf/gni/8AjVI3xb+NKqWa+1IADJJ0iLj/AMhUe2pfzL7w5Jdj7Kor5z+AHxR8Y+NPiDdab4m1k31pHpsk6xm3ijw4kjAOUUHox/OvoytSQooooAK4+7+K/gux8Xjwvda15estcR2wtvssx/ePjau4Jt53DnOOa7Cvj3xb/wAneRf9jBY/+hRUAfYVFFFABRRRQAUV86/Gz4463onig+GfBrz6ffWMwW5uPLilE4ZFZVVWVv73t+NcRofxl+Kx8Q6cNR1C8ms/tUXnxNp0SB4943AsIsjIyM9quFOc3aKuZzqQpq8nY+waK80/4W//ANQP/wAm/wD7CvGPG3xk+Ja+M9QHh26urPTdy+RAllFMEGxc4cx5POTXXWy/FUI81SGnyf5HFh80weJly0p3fo1+aR9ZUV81fC74x+Nn1K//AOEz+0alD5K+QssKW2xs8nKxjPHavTtK+NGgXOtNp2tTWmkOI9yme7BLMSAFxtHXOfwqPqddUvbOPu+q/Lc0+v4Z1/YKV5drP87WPR6KKK5DtCiiigDkPFPxU8G+C9WXTPE2s/Yrx4hMsf2WaTKEkA5RCOqnv2rrkdZI1dDlWGQfUV8iftSf8lZtv+wVD/6Mlr62s/8Ajxg/65r/ACoAmooooAKKK89+MPxJX4c+FhLHBM95qKTQWksW3EEoT5XYNwQCQcYPSgD0KivjG2+MfxjvLdZ7TU7+eF87ZItKhZWwcHBEWOor6B+H3j3WLvwFpk/iW2kn1R0f7RJKBCzHe2MoFAHGO1cOKzDC4SPNWnbW3fX5XNqdCpVdoI9Norw/4xfEXxfp+l6Y3gdLi1neZxOYIFuSy7RjIZDjnvXm/hz4q/F+bxTpUeqXOoyWT3sK3CPpcaBoy43AsIuBjPPanRx+FrUvbQmuX7tvJ6hKhUhLkktT64oridU+JmnaJarc6ukNnCziNZJrjaCxBOPu9cA/lXUaPrem+INNW/0S+hvrR2KrNA+5SQcEZowmOw+MTlQldLya/NIKtGpRdpqxeooortMQrm/F3xC8MeBDaDxVqf2D7bv8j/R5ZN+zbu+4pxjcvX1rpK+bP2uPv+Evpef+0KAPofSdVstc0i11TS5vPs7uMSwy7Su9T0OGAI/EVcrkPhN/ySHwv/2DYf8A0GuvoAKKKKACiqOuapHofh/UdWmjaWOwtZbl0TqwRCxA9+K+TNc+PvxA1vxNezeDr27tdNZg0FmLKGZolwAcnYTycnr3oGk27I+waK+cvhP8WfG9xqOojxr9pv4REnkLNbpbbWyckFYxniu+8QfFG8TwzqbaZprQ3q2cpt5FmDlJNh2kKUw2DjjvVqEmrnbTy/E1Ie0jHT5Hp9FfGX/C4fjJ/wBBLUP/AAUw/wDxqvbvBHxV1RvA+nzeJbJ7nUfLY3E8riEsdzclAgA4xSUW9jKjhK9Z8sI/p+Z6/RXI+CviPoXjdpYNNvbU3sQLm1iuBI3ljaN/QcZYCuuqdjCcXCTi9wooooJCuP0v4r+C9a8Vjw1pmtedq5lkiFt9lmX5owxcbigXgK3ftxXYV8e/C/8A5OuT/sJaj/6LnoA+wqKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAoryT46fFefwBpSaXpKTRavqEPmWt4gRkh2uobKsDnIz2714Qnxq+LskavHrF2yMMqy6ZCQR6/wCqp2bJlKMfidj7SoryLQ/jJdHw9p39paOZ7z7LF58rXAQvJsG5ioTC5OTjtXC/E/4zeNRqNh/whvn6ZD5TeekUKXO9s8HLR8cdq0dKaV2jgp5pg6tT2cJ6+j/ysfS9FfJfgr4z/Eo+MdP/AOEhu7q903c3nwPZRQhxsOPnEeRzg/hXsJ+OWm2t9aQ6vZw6fFcyhPOmveFGQC33OcA5pKnJq9jSpj8NTqKlKXvPayb/ACR6nRVbT9QtNW0+G+024jurWdd8U0TZV19Qas1mdoUUUUAYvinxdofgvSV1PxNffYrN5RCsnlPJlyCQMICeint2qTw14p0fxhoy6r4cvPtlkztGJfKePLL1GHAP6V5b+1L/AMkntf8AsLQ/+i5au/s0/wDJGrf/AK/Z/wD0IUAet0UUUAFFFFABRRRQAUUUUAFFFFABRRRQBj+L/wDkR9d/7B1x/wCi2r4v+FX/ACG73/r3/wDZhX2h4v8A+RH13/sHXH/otq+L/hV/yG73/r3/APZhXpZX/vlP1/Q8rOP9wqen6o9Sooor9DPy4KR/9W30NLSP/q2+hpDW5y/7L3/JXJf+wZN/6HHX1/XyB+y9/wAlcl/7Bk3/AKHHX1/X5WfsoUUUUAFFFFAGN4x/5EXXv+wbcf8Aopq+MPhV/wAhq+/69x/6EK+z/GP/ACIuvf8AYNuP/RTV8YfCr/kNX3/XuP8A0IV6WV/75T9f0PKzj/cKnp+qPUqKKK/Qz8uCmyf6tvoadTZP9W30NJ7DW5zP7Lv/ACVuf/sFzf8AocdfX1fIP7Lv/JW5/wDsFzf+hx19fV+Vn7KFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQBieNP8AkQtf/wCwZc/+imr5P+AH/Iyar/16D/0MV9YeNP8AkQtf/wCwZc/+imr5P+AH/Iyar/16D/0MV42ef8i6r6fqjrwX+8RPeaKKK/Ij6oKjuP8Aj1l/3D/KpKjuP+PWX/cP8qa3EeM/st/8lZuf+wVN/wCjIq+vK+Q/2W/+Ss3P/YKm/wDRkVfXlfup8WFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQBh+N/wDkn3iH/sF3P/opq+P/AINf8hvUf+vdf/Qq+wPG/wDyT7xD/wBgu5/9FNXx/wDBr/kN6j/17r/6FVw+JHoZb/vcP66HsFFFFdZ92FRXH/HrL/uH+VS1Fcf8esv+4f5UA9jhf2Wf+SsXf/YJl/8ARsVfXdfIn7LP/JWLv/sEy/8Ao2KvruuE/NAooooAKKKKAMLxz/yTzxF/2Crn/wBFNXyn8AP+Rg1b/r1X/wBDr6s8c/8AJPPEX/YKuf8A0U1fKfwA/wCRg1b/AK9V/wDQ68XPf+RbV9F+aOvBf7xE93ooor8jPqgqG8/48Z/+ubfyqaobz/jxn/65t/Kqj8SE9jx39lj/AJKvef8AYIl/9GxV9dV8i/ssf8lXvP8AsES/+jYq+uq/dD4sKKKKACiiigDB8d/8k68R/wDYKuv/AEU1fKv7P/8AyHtX/wCvZP8A0Kvqrx3/AMk68R/9gq6/9FNXyr+z/wD8h7V/+vZP/Qq8XPf+RbV9F+aOvBf7xE92ooor8jPqgqC9/wCPC4/65N/Kp6gvf+PC4/65N/Kqj8SE9jx/9lj/AJKtff8AYIl/9Gw19c18jfssf8lWvv8AsES/+jYa+ua/dD4sKKKKACs+TQNHl1D7dLpNi95vD/aGtkMm4dDuxnIwOa0K8N1r4661pnxuTwTFpVg9m2p29n9oYv5m2QoCeuMjee1AHuVFFFABRRRQB8yfFz4XamvxW/4So31p9lv7lJkh+beoiWMEHjHOOOalr0H4v6rp/wBs0q1+3W32mMyq8PnLvUny8ArnIzXn1fdZJThHCKcd3v8Ae0fnPEFWpPGuEto2t80m/wAQooor2z58K8m8Z/8AJSov96D+les10Hw+8JaBr/jCeXWtItL10tC6tPGGIYOgB/DNePnNJ1MI7dNT3chrKljUmviVj3aiiivgj9JCiiigChfaDo+pzifUtKsbuYLtElxbJIwHpkjpyavgAAADAHQCvEfjB8cdZ+HPjSHRtM0uxu4ZLJLgyXBfcCzOMcEDHyiva4JDLbxyEYLoGIHuKAH0UUUAFeXfHL4caj8RPD9hHpl5a2p06SSeQ3G75ht6DAPPFeo1meINSsdO0e5OoXtvaiSF1Qzyqm87TwMnk1lWk405SjukyoJOSTPJPBuhz+G/CFhpF1LHLLbKys8edpy5bjP1rcqOCeK5hWa2lSaJvuvGwZT9CKkr8Tq1JVKkpz3bbfqz7GMVGKS2QUUUVkUeY/Hr/kQ7P/sJR/8AoqWvSv2bv+SK2H/Xzcf+jDVbVNH07W7VbbV7OG8hVxIscy7gGAIz9cE/nXqOj6Jpvh/TVsNEsYbG0RiywwJtUEnJOK/SOFcRGeGlRS1i7/ff/I+fzOm1UU+/6F6iiivrjygqnf6Ppuq+X/amnWl75edn2iBZNmcZxuBxnA/KrleU/Gz4san8MG0QaVYWl5/aIn8z7Tu+Xy/LxjaR13n8qAPUoLeG1t0gtYkhhjXakcahVUegA4AqSsPwXrk3ibwRpGt3UUcM1/aJO8cedqlhnAzzitygAooooAy/E2ly654S1fSbd0jlv7Ga2R3ztVnjKgnHbJr5t8M/DXUfh34g1G21O8tbpmjRQbfdjpu7getfUksscELyzusccalndzhVA5JJPQV414r1Ow1HxhfNp97b3SgRnMEquP8AVqOxrSn8R7GTQjLFXl0V1+BlUUUV1H2YVXvv+Qfcf9cm/kasU1lDoVcZVhgg9xQDV1Y4f9lT/kqepf8AYGl/9Hw19bV518JvCuhaTp9zqOm6Va2160rwGeKMBjHhG259MgGvRa4mrOx+dV6To1HTfQKKKKRiFZ8GgaPbX32220mxhu9xb7RHbIsmTnJ3AZycn860K8N8IfHXWvEfxmXwbc6XYRWhu7qDz4y/mYiSRgeTjJ2DPHegD3KiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA8P/AGh/htqPi6Oy12yvLWCDTIDHJHNu3OXkUDGAR3ridOtms9LtbVyGaGFI2I6EhQP6V7r8T9TsLPwPfQXd7bwTSiMxxyyqrOBKmSATk14irK6BkYMrDIIOQRXfhUuVs+J4jqT9tCn9m1/ndi0UUV1nywV5/wDFT/V6X9Zf/Za9AqbTdF03W/EukQavZQ3kRvIl2TLuGGdQR+IrOrHmg0ehltZUMXCo1/TVv1PY/g7/AMkc8M/9eK/zNdrVbT9PtNJ0+Gw023jtbSBdkUMS4VF9AKs15J+nhRRRQBXvtOstTgEGpWdvdxBtwjuIlkUH1wR15NLZ2Fpp1uLfT7WC1hBJEcEYRcnqcDiuJ+MPj6++HHguHWdMtLe7mkvUtzHcbtoDK5zwQc/KKsfCfxvefEHwJFruo2sFrO88kRjgztwpwDySaAO1ooooAKKKKACiiigAooooAKKKKACiiigDH8X/APIj67/2Drj/ANFtXxf8Kv8AkN3v/Xv/AOzCvtDxf/yI+u/9g64/9FtXxf8ACr/kN3v/AF7/APswr0sr/wB8p+v6HlZx/uFT0/VHqVFFFfoZ+XBSP/q2+hpaR/8AVt9DSGtzl/2Xv+SuS/8AYMm/9Djr6/r5A/Ze/wCSuS/9gyb/ANDjr6/r8rP2UKKKKACiiigDG8Y/8iLr3/YNuP8A0U1fGHwq/wCQ1ff9e4/9CFfZ/jH/AJEXXv8AsG3H/opq+MPhV/yGr7/r3H/oQr0sr/3yn6/oeVnH+4VPT9UepUUUV+hn5cFNk/1bfQ06myf6tvoaT2Gtzmf2Xf8Akrc//YLm/wDQ46+vq+Qf2Xf+Stz/APYLm/8AQ46+vq/Kz9lCiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAxPGn/Iha//ANgy5/8ARTV8n/AD/kZNV/69B/6GK+sPGn/Iha//ANgy5/8ARTV8n/AD/kZNV/69B/6GK8bPP+RdV9P1R14L/eInvNFFFfkR9UFR3H/HrL/uH+VSVHcf8esv+4f5U1uI8Z/Zb/5Kzc/9gqb/ANGRV9eV8h/st/8AJWbn/sFTf+jIq+vK/dT4sKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigDD8b/APJPvEP/AGC7n/0U1fH/AMGv+Q3qP/Xuv/oVfYHjf/kn3iH/ALBdz/6Kavj/AODX/Ib1H/r3X/0Krh8SPQy3/e4f10PYKKKK6z7sKiuP+PWX/cP8qlqK4/49Zf8AcP8AKgHscL+yz/yVi7/7BMv/AKNir67r5E/ZZ/5Kxd/9gmX/ANGxV9d1wn5oFFFFABRRRQBheOf+SeeIv+wVc/8Aopq+U/gB/wAjBq3/AF6r/wCh19WeOf8AknniL/sFXP8A6KavlP4Af8jBq3/Xqv8A6HXi57/yLavovzR14L/eInu9FFFfkZ9UFQ3n/HjP/wBc2/lU1Q3n/HjP/wBc2/lVR+JCex47+yx/yVe8/wCwRL/6Nir66r5F/ZY/5Kvef9giX/0bFX11X7ofFhRRRQAUUUUAYPjv/knXiP8A7BV1/wCimr5V/Z//AOQ9q/8A17J/6FX1V47/AOSdeI/+wVdf+imr5V/Z/wD+Q9q//Xsn/oVeLnv/ACLavovzR14L/eInu1FFFfkZ9UFQXv8Ax4XH/XJv5VPUF7/x4XH/AFyb+VVH4kJ7Hj/7LH/JVr7/ALBEv/o2Gvrmvkb9lj/kq19/2CJf/RsNfXNfuh8WFFFFABXzd4i+E/jS/wD2jI/FNro4fR11i0uTc/aoR+7Qx7m2l93G08YzxX0jRQAUUUUAFFFFAHxb8Wv+TktR/wCv+1/9FxV3ldT8XvAXhtfEtjr407/iZ3kjSzz+fJ87RiMIdu7aMD0H1rlq+1yKjKGHdR7S2+Wh+f8AEdeNTEqmt4rX52YUUUV9AfMhWdpvxcPgDx01sNF/tDzolh3favKxvZTnGxumK0a8/wBb8IeIPEHxESXRdIur1E8p2aFMgKpAJ/DNeLnUpxwj5Or19D3sghTljVz9Fdeuh9rUUUV8IfpAUUUUAfOfx6+FXjLxr8QoNT8NaQLyzTT44WkN1DHhw7kjDuD0YV9EWyNHaxI4wyoAR74qSigAooooAK+fv2s/+Ra8O/8AX3L/AOgCvoGuX8eeCfD/AIz0Qp4ksPti2SSSwDzpI9j7evyMM9O9ROahBzfQcVd2PJvhV/yTDRv+ub/+jGrr6p6VpVnomlw6dpkPkWsAIjj3FtoJJPJJPUmrlfiuKqxrYidWO0m397PsKcXCCi+iCiiiuY0OY8e+Mf8AhCNBh1IWP27zblYPL87y8ZVmznaf7vTHevTvht42/wCFheCoPEB0/wDs/wA6WSPyPO83G1sZ3bV6/SvF/jLo2o654NtbbSLOW8mW/SRo4lyQojkGfzI/OvTPgPoep+HvhRZ6frdlNY3aXEzNDMu1gC5IOK/SuFqdJYSU4/E3r8ttPvPnsylJ1Unt0PR6KKK+sPLCvEv2ifh54n8eN4dPhXTRffYhc+fm4ji2b/K2/fYZztbp6V7bRQBznw90q90L4c6DpeqQ+ReWllHFNHuDbWAwRkEg/ga6OiigAooooA574g/8kz8T/wDYIu//AES1fI/wZ/5Cmqf9cU/9CNfaOoWFtqumXWn38fm2t3C8E0e4rvRlKsMjkZBPIrwrUPA/h7wZ4rv4fDdh9jjZY1I86STjYG/jY9ya0pq8j1cppueKTXTX9P1FooorqPtgqOeXybeSXG7YhbGeuBUlQ3as9jOqDLNGwAHc4oE9i/8AAX4rHxZrF74b/sb7J5cD3/2j7Vvzhoo9m3YP72c57dK9zr5g/Zl8Ka9pPj7UNS1LSrm2sm0yS3E8iYUyebC23PrgE19P1xO7ep+dVpTnUcqm/UKKKKRkFfN3gP4T+NNF/aCTxNqWjiHSBfXkpuPtULfLIkoQ7Q5bksvbvX0jRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAHy7+1n/AMjN4e/685f/AEMVV0L/AJF3Tv8Ar1i/9AFezfGLwR4e8R+GZdW1mw+0Xtiipby+dImxWlUMMKwB69wa8kggjtreOCBdscSBEXOcADAFduFi9ZHx3ElaLcKPVa/fp+g+iiiu0+RCsnV/G3/CEappV+LD7di4EuzzvL/1bKcZ2nrmtauN8f6JqeuSaXb6RZTXkpkdAkS5OW2gD8TWVZtQdj0crhTnjIRq7f8AAdvxPrXwd4i/4SzwbpmvfZfsn2+ATeR5m/y89t2Bn8hW3XJ/C7TrzSPhb4fsNTt5La7t7RUlhkGGRsngiusryj9OCiiigDzL49+ENc8a/D2DTPDNl9svE1COZozKkeECSAnLkDqwq18D/C2seDvhpDpPiK0+yXq3Mshi81JMKx4OVJH616HRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQBj+L/8AkR9d/wCwdcf+i2r4v+FX/Ibvf+vf/wBmFfaHi/8A5EfXf+wdcf8Aotq+L/hV/wAhu9/69/8A2YV6WV/75T9f0PKzj/cKnp+qPUqKKK/Qz8uCkf8A1bfQ0tI/+rb6GkNbnL/svf8AJXJf+wZN/wChx19f18gfsvf8lcl/7Bk3/ocdfX9flZ+yhRRRQAUUUUAY3jH/AJEXXv8AsG3H/opq+MPhV/yGr7/r3H/oQr7P8Y/8iLr3/YNuP/RTV8YfCr/kNX3/AF7j/wBCFellf++U/X9Dys4/3Cp6fqj1Kiiiv0M/Lgpsn+rb6GnU2T/Vt9DSew1ucz+y7/yVuf8A7Bc3/ocdfX1fIP7Lv/JW5/8AsFzf+hx19fV+Vn7KFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQBieNP+RC1//sGXP/opq+T/AIAf8jJqv/XoP/QxX1h40/5ELX/+wZc/+imr5P8AgB/yMmq/9eg/9DFeNnn/ACLqvp+qOvBf7xE95ooor8iPqgqO4/49Zf8AcP8AKpKjuP8Aj1l/3D/KmtxHjP7Lf/JWbn/sFTf+jIq+vK+Q/wBlv/krNz/2Cpv/AEZFX15X7qfFhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAYfjf/AJJ94h/7Bdz/AOimr4/+DX/Ib1H/AK91/wDQq+wPG/8AyT7xD/2C7n/0U1fH/wAGv+Q3qP8A17r/AOhVcPiR6GW/73D+uh7BRRRXWfdhUVx/x6y/7h/lUtRXH/HrL/uH+VAPY4X9ln/krF3/ANgmX/0bFX13XyJ+yz/yVi7/AOwTL/6Nir67rhPzQKKKKACiiigDC8c/8k88Rf8AYKuf/RTV8p/AD/kYNW/69V/9Dr6s8c/8k88Rf9gq5/8ARTV8p/AD/kYNW/69V/8AQ68XPf8AkW1fRfmjrwX+8RPd6KKK/Iz6oKhvP+PGf/rm38qmqG8/48Z/+ubfyqo/EhPY8d/ZY/5Kvef9giX/ANGxV9dV8i/ssf8AJV7z/sES/wDo2Kvrqv3Q+LCiiigAooooAwfHf/JOvEf/AGCrr/0U1fKv7P8A/wAh7V/+vZP/AEKvqrx3/wAk68R/9gq6/wDRTV8q/s//APIe1f8A69k/9Crxc9/5FtX0X5o68F/vET3aiiivyM+qCoL3/jwuP+uTfyqeoL3/AI8Lj/rk38qqPxIT2PH/ANlj/kq19/2CJf8A0bDX1zXyN+yx/wAlWvv+wRL/AOjYa+ua/dD4sKKKKACvkvxVrmrRftXRWUeqXqWn9vWSeQtw4j2losjbnGDk8V9aV5hqXwL0PU/iYvjWbU9QS9W9hvBApTy90ZUgfdzg7B3oA9PooooAKKKKAPnz4u/FLTm8eweFRYXX2mwm8l5srsYyiMgjnPGeazK474taRqQ/aGvbw6fdfZpL62ZJ/JbYwCRZIbGDjvXY19pkNScqDhLZbfmfAcSUqcMSpx3ktflZL8Aooor6E+YCu5+E/wDyNdz/ANeT/wDoaVw1ee694r1/QviKkejaveWKP5SMsEpUFSQSOPXFeTm9VU8JK/XQ9rI6Lq42LT+HU+2KKKK+AP0wKKKKAPlH9pjW9V074pW0On6neWsR0uJikFw6KT5knOAevFfVFoSbKAk5JjXJ/CvOfiH8D9E+I3iSPWdU1PULWaO2W3CWxTaVVmOfmUnPzGvSooxFCkYOQihQT7UAOooooAK4P4rfEiw+HehQSajZXN2NRMkEYgK/IdvU5I45rvK8I/al0u/1Pw3oI02xubsxXUrSC3haTYNg5OBwKmcVKLjLZjTad0aPhzXIfEnh611e2ieGK5UsqSY3DDFecfStSuU+GVvNa/DfSIbmKSGVI33RyKVZf3jdQa6uvxTFwjTxFSENk2l6XPsKUnKnFvewUUUVzGgV6lXzX8Z9X1DRfBlrcaTeTWczX6I0kLlSVMchxkdsgflXpnwF1nUtf+E1nf61fT3129xMrTTuXYgOQBk1+icKYeUKE6zekml91/8AM8DM6ic1Dt+p6RRRRX2R5IV87/tV6rqOmP4V/s2/urPzBd7/ALPM0e7Hk4zg89T+dfRFcH8S/hPpXxPOmnV7+8tP7O83y/su35vM2ZzuB/uD86AND4Wzy3Pwp8NTXMrzSyadEzySMWZjt6knrXWVmeG9Dg8M+GdP0S0lklgsIFgSSXG5goxk44zWnQAUUUUAUdb1SPQ/D+oatPG0kVhay3Lon3mCIWIGe/FfPlr8StP+InibULnTrK5tFVI2xOVz90L2J9K9z8dwyXHw58SQwRtLLJpV0iIikszGFgAAOpr5P+E2l6hpmr6mupWNzaMYUGLiFoz1z3FaU37x6uUzlHFxUeujPUqKKK6j7YKKKhvGKWM7KSGWNiCOxxQDdlc9T+G3/It3H/X23/oCV19fL37MfijXNV+IGoadqWrXd1Zrpck4gllLIJPNhXdj1wSM19Q1xSd3c/PMTVVatKouoUUUUjnCvkv4a65q1x+1Clncapey2v8AaN+vkPcOyYEc2BtJxxgflX1pXmHh74F6H4d+I48ZWup6hLeCeefyJCnl5lV1I4XOBvOOe1AHp9FFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQB5L8dPiVYeEdJj0K8srmebU4hJHJEV2oEkUkHJz2rzWzuVvLGC6RSqzxrIAeoBGf61d/ap0vUL3XdCuLOxubiCK0kEksULMqEuMAkDArL0VWTQNPR1KstrGCpGCDtFduFb1R8fxJTh7lT7W3yLtFFFdp8gFanhn/kbdI/6/of/RgrLrjPiDrOpaK+lz6TfT2cod2DwuVOV2kHj0NRUlywbO3AUXXxUKadtfy1/Q+0aK5L4Wahd6r8K/D99qVxJdXU9mryzStuZzk8k9662vIP1IKKKKAPH/2mL+8074W202n3c9rKdUiUvBIUYjy5OMjtxVz9nS9utQ+EUE9/czXMxvJgZJpC7YyOMmuq+IfgGw+I3huPRtUuri1hjuVuA9tt3FlVhj5gRj5jUvgLwTZfD/wsmhabc3FzAkryiS427ssckcACgDpaKKKACiiigAooooAKKKKACiiigAooooAx/F//ACI+u/8AYOuP/RbV8X/Cr/kN3v8A17/+zCvtDxf/AMiPrv8A2Drj/wBFtXxf8Kv+Q3e/9e//ALMK9LK/98p+v6HlZx/uFT0/VHU+NLTVIrePV9Fu7hJLT5pbdZG2SIOc7c4OO/qPpWPBrWo+ONatoNMludPsLZA908UhRmY9VyPyH4mtLxrrk4aLw/o/zahe4Vip/wBWh/ln9BmsRrK4+HOsWt3HI9xpl0qxXPHRu5/mR+Ir6CvNLEtpv2V1zdr/AOW3MfMYOF8LFSS9q0+S+7XX9eW51HiHxjp3hwG3k3z3aoGSAZGQehLkY/ma3IJ/tOnxz7dvmxB9uc4yM4rA8bzR3HgO8mhcPHIkbKyngguuDWzp3/IDtf8Ar3T/ANBFetTnUeIqQk7pJNad7/ft/wAA8apTpfVYVIxalzNO77Jelt/+CYX7L3/JXJf+wZN/6HHX1/XyB+y9/wAlcl/7Bk3/AKHHX1/X5ufqoUUUUAFFFFAGN4x/5EXXv+wbcf8Aopq+MPhV/wAhq+/69x/6EK+z/GP/ACIuvf8AYNuP/RTV8YfCr/kNX3/XuP8A0IV6WV/75T9f0PKzj/cKnp+qOl8aQarZeVrej3dwBakGe1EjeW6jvtzj6+3Pas601XUPG/iCL7BNdWGlWihpzFIUaRj1XIP4fQE96t+NNYnurmLwzox3Xl2QJmB/1aHsfqOT7fWstYJvhz4ggbzHm0i9UJKxH3XA5P1HX6EjtX0FWaWJbTfsrrm7c3+W1/M+YwsG8KlJL2rT5O9v89+W512r6+2n30OnafYyahfyoZBCsgQKg4yzHpTtG11dZt7pHtpLS6tW8ue3kOShxxz3FVdU1O0s9UibS7Jb7WLuHEWw4HlZzuZugXP51PoejS6ZDeXN9Ms1/fN5lw6DCjA4VfYZr04TqyrtKV1rfsu2u9++55MqdGOHTlG0tLO+rd9dNuW22l7lH9l3/krc/wD2C5v/AEOOvr6vkH9l3/krc/8A2C5v/Q46+vq/Oj9SCiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAxPGn/ACIWv/8AYMuf/RTV8n/AD/kZNV/69B/6GK+sPGn/ACIWv/8AYMuf/RTV8n/AD/kZNV/69B/6GK8bPP8AkXVfT9UdeC/3iJ23xT8Nu9nH4s0eVbXVtH/e+ZkL5kY5IPuO3rkjvWH4IWb4m+LX8Va75QtdLKxWlirbgsmM7iPTPPPU4/u10PjvQ/EPi/WLHRILdrXw8HEl7d+amZcc7Qud3H06nPaqT+DtW8H+PYNW8E2Hn6VdoI76xSZE2AcZXeRn1Hvnsa+Mw9eCwSouovatPlenux/lbvo3rbttpc9etBurzKL5Va/n/nbr3LeteNPEcMmrXujabp40jR3Mcz38jpJcMoy3l449hnr29K7Gyvxqnh2C/SNolurUTBH6ruXOD+deb+NfEHh7Wor3S/F2l3tnq9jJIumxw+Y5mJA2yIVAU5wODnFd54c/tH/hCLH+2wwv/sg88MPmzjv74xn3rgxdCMMNTnycrv8Aerbp3fN5vSzdjenNus1e61+Wv4frueYfst/8lZuf+wVN/wCjIq+vK+Q/2W/+Ss3P/YKm/wDRkVfXlfrZ8sFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQBh+N/wDkn3iH/sF3P/opq+P/AINf8hvUf+vdf/Qq+wPG/wDyT7xD/wBgu5/9FNXx/wDBr/kN6j/17r/6FVw+JHoZb/vcP66HVfELQma2j8R6ZItvqOmfvN+cb0HY+4/XkVk+E1l8e+JG8Q6x5YgsNsdvaK2Qr4zuI/X3P0ra8X6RrXiXVLTSooWt9FDB7q58xcyY52hc5/Tr9Kqt4Y1Hwz4wh1HwpZ+dp1woS7tFkVNgHcbiPqPfPrW0VaV+l/x7/wBep9DiKcniOdRfImuZd33S620v3+Rp+MvFk3h+BY9NgjuLr5ZJRICVijLBcnBHJJwB9fSujnObOQ+sZ/lXmfirS/E1hoerT3jaVLbXdykkkoMhnI3gIvQLgcDH1r0S3+1/2Iv9peT9q8k+b9nzszjtnnH1qoNtO510qtSdecZJpWVvva/H+tjjv2Wf+SsXf/YJl/8ARsVfXdfIn7LP/JWLv/sEy/8Ao2KvruuQ+DCiiigAooooAwvHP/JPPEX/AGCrn/0U1fKfwA/5GDVv+vVf/Q6+rPHP/JPPEX/YKuf/AEU1fKfwA/5GDVv+vVf/AEOvFz3/AJFtX0X5o68F/vETtPin4bd7OPxZo8q2uraP+98zIXzIxyQfcdvXJHesPwQs3xN8Wv4q13yha6WVitLFW3BZMZ3Eemeeepx/drofHeh+IfF+sWOiQW7Wvh4OJL2781My452hc7uPp1Oe1Un8Hat4P8ewat4JsPP0q7QR31ikyJsA4yu8jPqPfPY18bh68FglRdRe1afK9Pdj/K3fRvW3bbS569aDdXmUXyq1/P8Azt17k+peMvEEnirUtP0a68NWlrZypAjavLIjzSFAWC7TzgnHT0ruP9I/sU/bvK+0/Z/33k52b9vO3POM9M15Zrl34Phm8Q2WueGHttbuZZPs+Y5Lh7tmHySRvj5MnsMAe9egeHLW9svh/Y22qbvtcViFlDHJB29D7gcfhXFjqMIUacox5dlqlrotU03dX/M2pzbqSTd9/lr2PNv2WP8Akq95/wBgiX/0bFX11XyL+yx/yVe8/wCwRL/6Nir66r9ZPlwooooAKKKKAMHx3/yTrxH/ANgq6/8ARTV8q/s//wDIe1f/AK9k/wDQq+qvHf8AyTrxH/2Crr/0U1fKv7P/APyHtX/69k/9Crxc9/5FtX0X5o68F/vETs/in4bd7OPxZo8q2uraP+98zIXzIxyQfcdvXJHesPwQs3xN8Wv4q13yha6WVitLFW3BZMZ3Eemeeepx/drofHeh+IfF+sWOiQW7Wvh4OJL2781My452hc7uPp1Oe1Un8Hat4P8AHsGreCbDz9Ku0Ed9YpMibAOMrvIz6j3z2NfG4evBYJUXUXtWnyvT3Y/yt30b1t220uevWg3V5lF8qtfz/wA7de5oz+KvFGraxqcXg7TdPns9LlMEr3krK9xKB8yx44GOmW4rrTLNPoZlurc2s725aSAuGMbbeVyODj1FeLahp+g6VeeIrLxUupLqrXk1xpIh83a+/lWj2/KWzgHd6Adq9Z0IaiPAdmNa3fb/ALEPP3/e3be/v6+9cmYYanSpwlTVlp0ab0Tund3XfRWehrSqSlUkpefy17W08tdTzX9lj/kq19/2CJf/AEbDX1zXyN+yx/yVa+/7BEv/AKNhr65r9XPmAooooAK4y8+LXgyw8ZDwtdaqyaw1xHai3+yykeY+No3BdvO4c5xXZ18e+Lf+TvIv+xgsf/QoqAPsKiiigAooooA80+L/APzB/wDtv/7TryXW9Ht9c0mWxuhw4yj45RuzCqfxb8eeJG+N0+gHUj/ZlpeQxwQeRH8iyJGXG7buOT6mp/EUmqJo8iaFbGe7k+RTvVfLB6t8xH4V9lllanUy905Rb5b3Vt7tvQ+BzmhVp5iqsZJc1rO9rWSWp55b3Oq6vLB4LubiNEgmZJp1fJdE6KPXGD+noa7u6tNZ0+KC28NJpcdnDEFK3fmbgf8AgPb+ua5648Ay23h22l01v+J3bv5xlDY8xj1XJ9O2f61peINR1AeDYYrqD7JqF9Iloyhg20scEggnggH86qhCpQpz9unzWWvdbKN/J799wxVSniKlNYdx5btNW6veVr6prbtsXvC2p6pq1rcT6mlp5SybIJLUMFlAyGb5jnGeh471e0v4Rr4/8dNctrRsPJhWbaLXzM7GUYzvHXNWrW2js7OK2gXbHCgRR7AYro/AHjHw94e8YTRa3q1tZO9qUVZWxlmdCB+OK6MxpRWAcaru118zmyytJ5kpUVZPSyXQ94ooor4M/RwooooA47xX8V/B3gnWF0vxLqjWl28ImWMW0smUJIByqkdVNdgjiSNXQ5VgCD7V8iftSf8AJWbb/sFQ/wDoyWvraz/48YP+ua/yoAmooooAKq6p/wAge8/64P8A+gmrVeN/tGeNPEHg3w7pL+G9Q+xNeTyRTnyY5N67Bx86nHXtWVaDnTlBdUyou0kzSIyMHkV4n4plvPhF4puL7QPKbTtaifbau3EMo/iA9ATkexI9K9H8Bate6z4A07UtSkN1dzRuzsFVS5DsBwMAcADtXLWHgTUPFviTU9Z+IFn5KPGbexsvOV/JQj72VJGRn8yT6V+XZfyYOtVjiWnCOko/zO+llps9b9PmfTV71acXBauzT7eb+X3lzwV4V1HQ/B82p6dLZ3XiTVQtw9zeMzRYYhtpKckYJOR1PtT7XXfHMXjOx0XUU8PTiVfOufsKzloIQcbiWIAJPA680/4eab4k8LpeaHrlv5ulWzM1jfiVDlc/dK7tw9RxxyPSpvhox1Sw1LxNccz6veOyk/wwodiL9Bg1WIn71erU5Zro/wDFt6WSenRqxnTi1CEFdPr8t353/U0PHng4eN9Bh01r42XlXKz+YIvMzhWXGMj+919q9O+HHgkfD7wXB4fW/OoCGWSTzzD5WdzZxtyen1ridW1rTdCtFutYvIrSBnEaySnALEE4/IH8q9R0XXdM8R6Ymo6HexX1m7FVmhOVJBwR+dfQcJ1KrpVIS+FWt8731+44s0jFSi1v1L9FFFfaHkBXMeMfiJ4Z8BGzHinUGs/tu/yMQSSb9m3d9xTjG5evrXT182ftcff8JfS8/wDaFAH0PpGrWeu6Naarpcpms7yISwyFSu5TyDggEfjVyuQ+E3/JIfC//YNh/wDQa6+gAooooAK8i8cf8jlff9s//Ra16J4xv7nS/AuvahYSeVdWmm3E8Mm0HY6xMynB4OCB1r5e+H/jbxD4y1zUpvEmofbZFiRg3kxx89P4FHYCtKbtI9jJ6ihirPqrfr+hY+IWhM1tH4j0yRbfUdM/eb843oOx9x+vIrJ8JrL498SN4h1jyxBYbY7e0VshXxncR+vufpW14v0jWvEuqWmlRQtb6KGD3Vz5i5kxztC5z+nX6VVbwxqPhnxhDqPhSz87TrhQl3aLIqbAO43EfUe+fWtYq0r9L/j3/r1PZxFOTxHOovkTXMu77pdbaX7/ACLOu6r4x0a3nvGOgG2V9sKETmWTJwq4HBY8V01ot5daJGuprHFdyw4mWIHajEc4ye31rn71/wC1viVZWD82+mWxvGXsZGO1c/QHNdW7KiM7nCqMknsKqG39dNDuopupN8zstNe+7/y+8u/Ab4VL4T1a98RjWDdmSB7D7ObbZjLRSb924/3cYx3617jXnHwk8W6Dq9jc6Zpmq29zerI9wYI2ywjwi7vpkgV6PXNKyeh8Ti4whXlGntfQKKKKk5grjNK+LXgzW/Fw8M6bqrS6sZZIhAbWVRujDFxuK7eAjd+1dnXx78L/APk65P8AsJaj/wCi56APsKiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA5j4j/wDJP9S/7Zf+jUrwggEEHkHrW1+0x428Q+Hb3StK0bUPs1lfWzPcReTG29lcFTllJGMdiK5nTLiWfQrO5lzLNJbJI2MAsxUE+g5Nd2GkuVxPi+I6MvawrdGrfc2/1OE1iW58B6xO+llGtNRjYpEx/wBU4749s/lx2rf0DQ7vSfD73Fg9tNq15tlea4JKHJzjK8ngn6mqtp4Xutc1K+1DxVB5bSKYreAOG8pfUEEjI/nk1d8K22r6Nb3Gn6tFutLck29yHU5X0xnI9Rx6+1VTi09dunkYYmtGWH5YyTmuXm/vdlfrbr33G22p+Ix4ig068XS5FK+ZObYSZiT1JY4BJ6da0tW8FDxvqulWDX5ss3Ai3iHzP9YyjOMjpiqXg3N1p1xq83M2oTtISeyA7VX6DFdPp+uaZoXiTSLjWL2KziF5G5eU4GFdSx/AVpZOn7xyc86eOhGirSjo7Lr1/Vemp7/4P8PDwn4P0zQRc/axYQCHzzHs3477cnH51tVW03UrPWNNg1DTLhLm0uE3xTRnKuvqKs15Z+jBRRRQBieK/GGieCNITU/Et2bS0eYQLIInky5BIGFBPRTT/C/irR/GWirq3h26N1ZM7RiQxNH8y9RhgDXl/wC1L/ySe1/7C0P/AKLlq7+zT/yRq3/6/Z//AEIUAet0UUUAFFFFABRRRQAUUUUAFFFFABRRRQBj+L/+RH13/sHXH/otq+L/AIVf8hu9/wCvf/2YV9oeL/8AkR9d/wCwdcf+i2r4v+FX/Ibvf+vf/wBmFellf++U/X9Dys4/3Cp6fqj1Kiiiv0M/Livf2FtqdjJaX0fmwSY3puK5wc9Rz1FSLEkNqIohtSNNqjOcADAqSkf/AFbfQ1PLFNytqWpSso30OX/Ze/5K5L/2DJv/AEOOvr+vkD9l7/krkv8A2DJv/Q46+v6/LD9jCiiigAooooAxvGP/ACIuvf8AYNuP/RTV8YfCr/kNX3/XuP8A0IV9n+Mf+RF17/sG3H/opq+MPhV/yGr7/r3H/oQr0sr/AN8p+v6HlZx/uFT0/VHqVFFFfoZ+XGPqXhPRdYvDdajZ+dMVClvNdeB04BAqXTdA0zQorj+yrbyPOA3/ALxmzjOPvE+prTpsn+rb6GsFh6MZOpGCT72V/vOn6zXlD2Tm+Xtd2+45n9l3/krc/wD2C5v/AEOOvr6vkH9l3/krc/8A2C5v/Q46+vq/MT9dCiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAxPGn/ACIWv/8AYMuf/RTV8n/AD/kZNV/69B/6GK+sPGn/ACIWv/8AYMuf/RTV8n/AD/kZNV/69B/6GK8bPP8AkXVfT9UdeC/3iJ7zRRRX5EfVBUdx/wAesv8AuH+VSVHcf8esv+4f5U1uI8Z/Zb/5Kzc/9gqb/wBGRV9eV8h/st/8lZuf+wVN/wCjIq+vK/dT4sKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigDD8b/wDJPvEP/YLuf/RTV8f/AAa/5Deo/wDXuv8A6FX2B43/AOSfeIf+wXc/+imr4/8Ag1/yG9R/691/9Cq4fEj0Mt/3uH9dD2Ciiius+7Corj/j1l/3D/KpaiuP+PWX/cP8qAexwv7LP/JWLv8A7BMv/o2KvruvkT9ln/krF3/2CZf/AEbFX13XCfmgUUUUAFFFFAGF45/5J54i/wCwVc/+imr5T+AH/Iwat/16r/6HX1Z45/5J54i/7BVz/wCimr5T+AH/ACMGrf8AXqv/AKHXi57/AMi2r6L80deC/wB4ie70UUV+Rn1QVDef8eM//XNv5VNUN5/x4z/9c2/lVR+JCex47+yx/wAlXvP+wRL/AOjYq+uq+Rf2WP8Akq95/wBgiX/0bFX11X7ofFhRRRQAUUUUAYPjv/knXiP/ALBV1/6KavlX9n//AJD2r/8AXsn/AKFX1V47/wCSdeI/+wVdf+imr5V/Z/8A+Q9q/wD17J/6FXi57/yLavovzR14L/eInu1FFFfkZ9UFQXv/AB4XH/XJv5VPUF7/AMeFx/1yb+VVH4kJ7Hj/AOyx/wAlWvv+wRL/AOjYa+ua+Rv2WP8Akq19/wBgiX/0bDX1zX7ofFhRRRQAVSfRdLkvftkmm2b3O4P57QKX3Dod2M5GOtXa8K1v46a9pnxxTwVDpunPYtqdtZmd1k83ZIUBP3sZG844oA91ooooAKKKKAPmb4ufC3UF+Kn/AAlX9oW32e/uUmSHa29REsYIPbnFPrv/AIv6zphvtLshqNp9qiMqyQeeu9CfLwCucjPauAr7rJKcI4RSju9/vZ+c8QVak8a4S2ja3zSb/EKKKK9s+fCvJvGf/JSov96D+les10fw88LaDrnjCeTWdHsr50tC6tcQq5DB0APPcZrx85pOphHbpqe7kNZUsak18Sse50UUV8EfpIUUUUAUrvRtMv5hLf6daXMgG0PNArtj0yRV0AAAAYA6CvDvjH8cNd+HPjaHRtJ07TrmCSyS4L3KuW3Mzgj5WAx8or22CQy28cjcF0DHHuKAJKKKKACvMPjf8N774i+H7GPT763tDp0kk7mcMd429Bj6V6fWV4j1bTtK0a4Op39rZiWF1jNxMse87TwMkZNZVpONOUo7pMqCTkkzyfwdocvhrwjY6RcSpNLaqytJGDtOWLcZ+tbdRW9zBd26z2s0c8L/AHZI2DK30IqWvxOrOdSpKc92236n2MYqMUlsFFFFZFHmPx6/5EOz/wCwlH/6Klr0r9m7/kith/183H/ow1X1LSdP1m2W31aygvIVcOsc6B1DYIzg98E/nXqOk6Ppug6etjotjb2NohLLBbxhEBJyTgV+kcK4iM8NKilrF3++/wDkfP5nTaqKff8AQu0UUV9ceUFVbzTLDUdn9oWNtdeXnZ58Kvtz1xkcdB+VWq8n+N/xX1f4Ytog0aysrr+0RP5n2pXO3y/Lxjaw/vn8hQB6rDDFbwpDbxpFEg2oiKFVR6ADpT6w/BWuT+JvA+j61eRxxT39ok8iRA7VLDJAyScVuUAFFFFAGZ4l0uTXPCer6TDIsUl/YzWyO/RS6FQT7c183eHPhpffDvxDqNtqF9b3bNGi5gVgOm7v9a+oZporeCSe4kSKKNS7yOwVUUDJJJ6ADvXjPinV9N1XxhfvpmoWt4oEZJt5lkH3FHYmtKfxHsZNCMsVeXRXXroZlFFFdR9mFV77/kH3H/XJv5GrFIyh1KsAVYYIPcUA1dWOG/ZU/wCSp6l/2Bpf/R8NfW1eefCfwxoel6bc6hp2k2dreNK0BnihVXMeEbbkdsgHFeh1xNWdj86r0nRqOm+gUUUUjEKpRaLpdvefa4NNs4rnJbzkgUPk9TuAzzk/nV2vCvB/x017xF8aV8HXem6dFZG7uoPOiWTzMRJIynliMnYM8dzQB7rRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAeI/tC/DW+8Xw2eu2l9b28OmQGKSOUMWcvIoBGPrXD6dbNZ6Xa2rsGaCFIyR0JCgf0r3P4o6tp1j4Jvba91C1t55hGY4pplRnAlTJAJya8TR1kRXjYMrDKspyCPWu/CpcrZ8TxHUn7aFP7Nr/O7FooorrPlgrz/4qf6vS/rL/wCy16BU+l6Pp2s+JtIg1axgvIjeRLsnjDjDOoIwfUVnVjzQaPQy2sqGLhUav/wVb9T2L4O/8kc8M/8AXiv8zXa1XsLC00uwhstOtorW1gXZFDEoVUHoAOlWK8k/TwooooAgu7G0v4RFf2sNzGG3BJow6g+uD35pbWztrGAQ2VvFbRA5EcKBFz9BXD/GPx9qHw58FQ6zpNtbXM8l6luUuQxXayuSflIOflFT/CTxtffEDwFFrmqW9vb3D3EkRjtgwTCnA+8Sf1oA7eiiigAooooAKKKKACiiigAooooAKKKKAMfxf/yI+u/9g64/9FtXxf8ACr/kN3v/AF7/APswr7Q8X/8AIj67/wBg64/9FtXxf8Kv+Q3e/wDXv/7MK9LK/wDfKfr+h5Wcf7hU9P1R6lRRRX6GflwUj/6tvoaWkf8A1bfQ0hrc5f8AZe/5K5L/ANgyb/0OOvr+vkD9l7/krkv/AGDJv/Q46+v6/Kz9lCiiigAooooAxvGP/Ii69/2Dbj/0U1fGHwq/5DV9/wBe4/8AQhX2f4x/5EXXv+wbcf8Aopq+MPhV/wAhq+/69x/6EK9LK/8AfKfr+h5Wcf7hU9P1R6lRRRX6GflwU2T/AFbfQ06myf6tvoaT2Gtzmf2Xf+Stz/8AYLm/9Djr6+r5B/Zd/wCStz/9gub/ANDjr6+r8rP2UKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigDE8af8iFr/wD2DLn/ANFNXyf8AP8AkZNV/wCvQf8AoYr6w8af8iFr/wD2DLn/ANFNXyf8AP8AkZNV/wCvQf8AoYrxs8/5F1X0/VHXgv8AeInvNFFFfkR9UFR3H/HrL/uH+VSVHcf8esv+4f5U1uI8Z/Zb/wCSs3P/AGCpv/RkVfXlfIf7Lf8AyVm5/wCwVN/6Mir68r91PiwooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAMPxv/yT7xD/ANgu5/8ARTV8f/Br/kN6j/17r/6FX2B43/5J94h/7Bdz/wCimr4/+DX/ACG9R/691/8AQquHxI9DLf8Ae4f10PYKKKK6z7sKiuP+PWX/AHD/ACqWorj/AI9Zf9w/yoB7HC/ss/8AJWLv/sEy/wDo2KvruvkT9ln/AJKxd/8AYJl/9GxV9d1wn5oFFFFABRRRQBheOf8AknniL/sFXP8A6KavlP4Af8jBq3/Xqv8A6HX1Z45/5J54i/7BVz/6KavlP4Af8jBq3/Xqv/odeLnv/Itq+i/NHXgv94ie70UUV+Rn1QVDef8AHjP/ANc2/lU1Q3n/AB4z/wDXNv5VUfiQnseO/ssf8lXvP+wRL/6Nir66r5F/ZY/5Kvef9giX/wBGxV9dV+6HxYUUUUAFFFFAGD47/wCSdeI/+wVdf+imr5V/Z/8A+Q9q/wD17J/6FX1V47/5J14j/wCwVdf+imr5V/Z//wCQ9q//AF7J/wChV4ue/wDItq+i/NHXgv8AeInu1FFFfkZ9UFQXv/Hhcf8AXJv5VPUF7/x4XH/XJv5VUfiQnseP/ssf8lWvv+wRL/6Nhr65r5G/ZY/5Ktff9giX/wBGw19c1+6HxYUUUUAFfOPiH4R+Mr/9omPxVa6ZG+jrq9rdGc3UYPloY9x2lt3G08YzX0dRQAUUUUAFFFFAHxb8Wv8Ak5LUf+v+1/8ARcVd5XV/F7wL4cHiOx14aaP7Tu5HkmuPOk+doxGEON2BgegrlK+1yKjKGHdR7S/TQ/P+I68amJVNbxWvzswooor6A+ZCsyw+LcngHx21tHo633nRLDuNyY8b2U5+6emK064HWfBviHxF8REl0TSbi9RBE7NEBwqkAn8M142dSnHCPk6vX0PeyCFOWNXP0WnrofalFFFfBn6QFFFFAHzt8ePhN4w8b/EGDVPDempdWiafHAztcxx4cO5IwzA9GFfQtsjR2sSOMMqAEe+KkooAKKKKACvn79rP/kWvDv8A19y/+gCvoGuZ8deDNA8Y6IU8SaeL1bNJJYAZXTY23r8pGeneonNQg5voOKu7Hknwq/5Jho3/AFzf/wBGNXX1U0vS7PRdMh0/TIfItYARHGGLbQSSeSSepNW6/FMVVVavOrHaTb+9n2FOLhBRfRBRRRXOaHLeP/GLeCNAh1JLIXpluVg8sy7MZVmznB/u/rXp/wANPGzfEHwRb6/JYiwaaWSPyFl8zG1sZ3YH8q8a+Meialr3g61tdHs5LudL9JGSMchRHIM/mR+del/AnQdU8N/Cmz07XbKWyvEnmZoZRhgC5IP5V+lcLQpLCSnH4m9fltp9589mUpOqk9uh6NRRRX1h5YV4p+0N8OvE3j5vDx8L2KXf2IXPn7p0j27/ACtv3iM/cbp6V7XRQBzvw+0m80H4d6FpWpxCK8s7KOKZAwbawGCMjg/hXRUUUAFFFFAHPfEH/kmfif8A7BF3/wCiWr5H+DP/ACFNU/64p/6Ea+0b+xttU0250+/j821uoXgmjyRvRgVYZHIyCeleGal4K8P+D/Fd/D4b04WUbLGpUSu/GwH+Jj3JrSmryPVymm54pNdNf0/USiiiuo+2Co55PJt5JcZ2IWx64FSVDdqz2U6IMs0bAD1OKBPYu/AP4qyeLNavfDj6QtqsdvJf/aBcbySGij27do/vZzntXutfMf7M/hDX9H8e6hqep6XPbWTaZJbidwNpk82Ftv1wCa+nK4ne+p+dVpVJ1HKpv1CiiikZBXzj4E+EfjLRPj+nifUtMji0kXt5MZxdRsdsiShDtDZ5Lr2719HUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQB8u/tZ/8jN4e/wCvOX/0MVV0L/kXdO/69Yv/AEAV7R8YPBXh/wAQ+F5tV1jThc3tiiJbymV18tWlUMMKwBznuK8ighjtreOCFdscSBEXOcADAFduFi9ZHx3ElaLcKPVa/fp+g+iiiu0+RCsfWfGr+CNT0q/SxW9InEuwy7P9WynGcHrmtiuO8e6FqevS6Xb6PZS3cxkdAkY5y20KPxNZVm1B2PRyuFOeMhGrt/wHb8T608GeIT4s8F6XrzWwtTfwCYwh9+zPbOBn8q265T4X6beaP8L9A0/U7d7a7t7RUlhf7yNk8Gurryj9OCiiigDzT48eDtb8b/D+DS/Ddqt1dpqEc7I0qx4QI4JyxA6sKs/BHwpq/gz4aw6R4htltr1bmWQxrIr/ACseDlSRXoVFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGP4v/5EfXf+wdcf+i2r4v8AhV/yG73/AK9//ZhX2h4uBPgjXAOT/Z1x/wCi2r4L0TVdV8P3Uk+nRkPImxt8RYYzn+ldmCrRoYiNSeyOHMKE8RhZ0obs92oryP8A4WD4n/uR/wDgOaP+Fg+J/wC5H/4DmvrP7dwnn93/AAT4r/VzG/3fv/4B65SP/q2+hryT/hYPif8AuR/+A5oPxA8TkEbI+f8Ap3pf27hPP7v+CC4cxv8Ad+//AIB2f7L3/JXJf+wZN/6HHX1/XyH+zDDLH8WpS8bqP7Mm5Kkfxx19eV8QfoYUUUUAFFFFAGN4x/5EXXv+wbcf+imr4w+FX/Iavv8Ar3H/AKEK+z/GAJ8D66AMk6dcf+i2r4M0TVdV8P3Mk+nRkPImxt8RYYzmuzBVo0MRGpPZHDmFCeJws6UN2e7UV5H/AMLB8T/3I/8AwHNH/CwfE/8Acj/8BzX1n9u4Tz+7/gnxX+rmN/u/f/wD1ymyf6tvoa8l/wCFg+J/7kf/AIDmg/EDxOykFI+Rj/j3pf27hPP7v+CNcOY2/wBn7/8AgHafsu/8lbn/AOwXN/6HHX19XyJ+zBDLH8Wpy8bqP7Lm5Kkfxx19d18QfoQUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGJ40/5ELX/+wZc/+imr5P8AgB/yMmq/9eg/9DFfWPjME+A9fAGSdNucD/tk1fC3hfxJrnhC9mutGTbJNH5biWHcCMg/0rz8zw08VhJ0ae77+pvh6ip1VOWyPrGivnX/AIXJ43/u2/8A4CUf8Lk8b/3bf/wEr4H/AFXx/eP3/wDAPc/tKh5n0VUdx/x6y/7h/lXzz/wuTxv/AHbf/wABKa/xi8bPGyEW+GBBxaU1wvj77x+//gC/tKh5nRfst/8AJWbn/sFTf+jIq+vK+Rv2XoZY/ixcl43Uf2VNyVI/5aR19c1+mnzoUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGH43/5J94h/wCwXc/+imr4/wDg1/yG9R/691/9Cr7A8bAn4f8AiEAZJ0u5wB/1yavhPw/rmseGbqWfS49rypscSQ7gRnNVF2dzrwdaNGvGpLZH0ZRXiP8Aws7xb/ch/wDAaj/hZ3i3+5D/AOA1b+1ifT/21hfP7v8Agnt1RXH/AB6y/wC4f5V4t/ws7xb/AHIf/Aakf4meLHjZCsWGBBxbUe1iJ51hbdfu/wCCdh+yz/yVi7/7BMv/AKNir67r5H/Zchlj+K12ZI3Uf2TLyVI/5aRV9cVzHxwUUUUAFFFFAGF45/5J54i/7BVz/wCimr5T+AH/ACMGrf8AXqv/AKHX1Z44BPw98RADJOl3OAP+uTV8M+F/EuueELya50ZNrzx+W4lh3AjOa87M8NPFYOdGnu+/qb4eoqdVTlsj6xor51/4XJ43/u2//gJR/wALk8b/AN23/wDASvgv9V8f3j9//APc/tKh5n0VUN5/x4z/APXNv5V89/8AC5PG/wDdt/8AwEpsnxh8ayxPGwtwHUqcWnrVLhfHp3vH7/8AgCeZULdTpP2WP+Sr3n/YIl/9GxV9dV8kfstwyx/Fa8MkbqP7Jl5ZSP8AlrFX1vX6YfOhRRRQAUUUUAYPjv8A5J14j/7BV1/6KavlX9n/AP5D2r/9eyf+hV9V+OgT8O/EYAyTpV1gD/rk1fDXhfxNrnhC8nuNGQK88fluJYdwIzmvOzPDTxWDnRp7vv6m+HqKlVU5bI+sKK+df+FyeN/7tv8A+AlH/C5PG/8Adt//AAEr4L/VfH94/f8A8A9z+0qHmfRVQXv/AB4XH/XJv5V8+f8AC5PG/wDdt/8AwEpkvxg8aywvGwtwHUqcWnrVR4Xx6ad4/f8A8ATzKhbqdN+yx/yVa+/7BEv/AKNhr65r5J/Zahkj+Kt6ZI3Uf2RLyykf8tYq+tq/TD50KKKKACvkjxVq2ox/tYRWqahdLb/29ZL5ImYJgtFkbc4xzX1vXmepfAzw/qfxJXxrNqGprfrew3ghR4/K3xlSBjZnHyDPNAHplFFFABRRRQB89fF34pWLeP4fCv8AZ1x5+nziF5967XMqxkEDrxms6uV8f6Fea/8AtT3VhZhUkn1C32vLkIMRRk5IB9K9j/4VPrn/AD96f/38f/4ivq8nx9OnRdOtJK2x8ZnuW1atdVaEG29/lovwOGorzjxR4r8TaF4s1XSVELLY3ctuGWEsDtYjg9+ldz8IrDxF8RLfVZJpLSL7E8SjzVaPO4N0wDn7tejDOcJKfLdrz6HlTyDGwp89k/Jbl6u5+E//ACNdz/15P/6GlYXjTwLr/hjwXqespcWLtZw+YFRnYnkDoVGeteDad408XDxHBc22p6hZNLIkbraO8SldwyMDscVyZlmmHlRlSpvmbXTY7cpyfFRxEa1Rcqi+u5950UUV8cfdhRRRQB8nftNatqNj8VLaKyv7q3jOlxMUimZBnzJOcA19V2hzZQE8ny1/lXnnxA+CGgfEXxHHrOsahqVtPHbrbhLV4wu1WYg/MhOfmPevR4oxFCka5IRQoz7UAOooooAK4H4s/Eiz+HWg28l9YT3g1EyQIIWA2Hb1OfrXfV4B+1hG8nhrw95aM2LuXO0Z/gFTKKnFxezGm07o2fDWuReJfDtpq8ELwx3SlljcglcMV7fStSqXwu8E6lF8LtALzW2ZbUTgbmBCyEuAfl6gMAfeuR+NOveJPh7d6RFpTWzi9SVn/dGTG0rjrjH3q/LZ5BjJ4mVOnC0buze1uh9JHHUlTUpS1O7orxbwF8R/F/ibx5pOjXn2dYLyfy3P2crgYJ69ulfRH/CH6h/z2tv++m/+JrOvw/mFGSXJzempUMdQmr3t6mBXqVfFnxC8eeL4fFGraHDdSWsNhqMsUclmrRuQjMoBYckf4Cvov4C6rqes/CazvNcvLm9vGuJg011IzuQHIAJPPSvsuH8rrYCE5Vmrytp2tf8AzPJx2JhXaUOlz0iiiivpjzgr51/at1G9sH8KfYby4tt4u9/kysm7Hk4zg89TX0VXDfEj4UaP8Tjpp1q8vrX+zvN8r7IyDd5mzOdyn+4OnvQBd+Fksk/wn8NSzyNJI+nRFndiSx29STXWVm+HNDt/DXhvT9FspJJbewgWCN5SC7KowCcADP4VpUAFFFFAFHXNUTQ/D2o6tNG0sdhay3LRqcFgiFiB7nFfPVl8S7L4i+JdQubGwns1VEbEzKf4Qvb6V7l8QAT8NPE4AyTpF3gD/ri1fNH7P3gzU9dl1q7tmghjhWKJhOWUknceMKeOKuDtI9HLa6oYlSbsnozvKKv+OfCmueGPA+qa1BNZvJZw71VSzE/MB02jPWvAv+FneLf7kP8A4DVu6kUfS1M2wtN2vf0PbqK2fC/grW9c8I6Rqss9mr31lDcMCzKQXQN028da4D4yS+I/h/caZb2kluwvopSzRIXK7So6kDH3qPaRsXPNMLGHNzX8up7n8Nv+RbuP+vtv/QErr6+X/wBmXX/EF/8AEHULHVNSv7iyGlyTLBPK7RiTzYRuweM4JGfSvqCuaTu7nx2JqqtWlUXUKKKKRzhXyR8M9W1Gf9qRLWbULqS3/tHUB5TzMUwI5sDGccYFfW9eZ6B8DPD/AId+Ig8Y2eoanJfCeafyZXjMW6VXVhgIDgbzjn0oA9MooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAPI/jr8S7PwhpcehXNhPcS6pEJUljYBUCSKSDn6V5vZ3IvbC3ukUqs8SyBT1AIzj9al/awikk8S+HvLjZsWkudq5/jFdho/we8QQaHYRPeaaWjto1JEknJCgf3K68PUUbqTPmM9wM8Q4VKUW5bP0OQornfinN4j8A+LItJja3lElolxuijZxyzDqQP7tQfDS98R+OPGcWjObeIPDJJukjKD5RnqAa6PrEL2PD/sPGez57L06nU1qeGf+Rt0j/r+h/wDRgrqP+FR69/z96d/39k/+Ir5o1Dxd4oe+QxXVxaSWspKPaBozkHg5HpjilOvBLTUvB5Ni51U5rlSa3PvmiuS+Fl7eaj8K/D93qdxNc3c1mrSyzsWd2yeSTyTXW15p+gBRRRQB49+03eXNj8LLaWyuJbeQ6pEpeJyhx5cnGRVz9nK6uLz4QW8t5PLPIbycb5XLNjI7mut+IHgLTviN4dj0bWLm6toI7hbgPasobcqsAPmBGPmPapfAngmw+H/hhND0m4ubi3SV5Q9yVL5Y5P3QB+lAHR0UUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFeffFf4pWvw0sdKd41uLi+vFUw9xApBlce+CAPdvavQGZUUs5CqoySTgAV8K/F7xw3j34iXuowuWsID9msR28pSfm/wCBHLfjjtQB9zWtzDe2cN1aSLLBPGskUinIdWGQR7EGpa8Q/Zm8df254Pm8MX0u690fmDceXt2PH/fLcewK17fQAUUUUAFFFFABRRRQAVT1bVbPQ9HutU1OZYLS0iaWaRv4VAz+J9u9XK+Yf2mfiT9tvl8EaRNmC2YS6k6Hh5OqxfRep98d1oA+hPB/iez8Z+ErDX9OVkgvY93lscmNgSrKfcMCPwrE+JXxL074b2OmT6gvmvf3iQiMHlYgQZZMd9qnp6kV5R+yp4t8yz1bwpcycwn7dagn+E4WQfQHYf8AgRryn41eOv8AhOviNd3FtLv02x/0SyweGRTy/wDwJsnPpt9KAPuCKVJ4UlhdZI5FDI6nIYHkEGnV43+zd46/4STwGdCvZd1/omIl3Hl7c/6s/wDAcFfYBfWvZKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAriviT8S9O+G9npc2oL5r394kOwHlIgR5smO+0EcepFdqTgZPSvhv40eOv8AhO/iNd3VtLv02y/0WyweCinl/wDgTZP0wO1AH3FHIk0SSxOrxuoZWU5DA9CDTq8d/Zw8df8ACS+Av7EvZd2oaJiIbjy8B/1Z/DBX6BfWvYqACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigArz34rfFS0+GsOj+ZGs89/dqJI+6W6kea4HrggD1J9q9Ad1jjZ5GCIoJZmOAAO5r4S+LPjdvHvxDvtUjcmxiP2exU9oVJwf+BHLf8CoA+67e4iu7aK4tpFlhmQPHIpyGUjIIPoRUleKfs0+Ov7f8FyeHL6XdfaLgRbjy9ux+X/vk5X2G2va6ACiiigAooooAKKKKAIbu8trC1e5vriK2t4xl5ZnCIvbkngVztz8TfA1pnz/ABdooI6hL6NyPwBNXPGOoeHbDwtejxjdW1vpVxE0M32hsCQMMFQOpPoBz6V8Catb21tqlwumySzWBlf7JPNHsaWIMQrY/D880Afb8Pxl8AXOowWFp4jhubq4kWKGK3gllLuxwANqnvTviR8TNN+HFvpL6gvmNqF4sJQHlIQR5kuO+0EcepFeR/szfDYKr+OtZhwPmi0xXHTs8381H/AvavK/jL45Pjz4jXl5byb9OtP9FsQDwY1Jy/8AwJst9CB2oA+5EdJY1kjYOjAMrKcgg9CKdXj/AOzl46/4SfwANGvZd2oaHthO48vAf9W34YK/8BHrXsFABRRRQAUUUUAFeYax+0B4L0HXrzSNW/tK2u7KZoZVa0yMg9QQeQeoPcGtnxh8XfB3gnfFquqpPerwLGz/AH0xPoQOF/4ERXzH8ZBrXirUh47n8IXegaZc+XarJct89w4DFXZSARlQBnGPlHJoA9/j/aO+G7/e1a5j/wB6yl/oprd1D4seGbG08OXIuJHi8RXSwWZeJojtJx5pVwCEBK8453AjjmvmP4O/DW38TXNz4m8VkW/hXRgZbmSTgXDKN3lj2A5bHsByeOY+Inje58deNLjWCDb20eIbC3HAt4F+4ox0Pc47k0AffNee/FX4qWnw1TRlljWea/u1EsfUpbKR5jgevIAHck+la3w38XR+L/hrpWvTSqJGt9t2xOAsqfLIT6DIJ+hFfHfxX8bv49+Id/qqOxsoz9nslP8ADCpODjtuOWPu1AH3bBPFdW8c9vIssMqB43U5DKRkEH0xT68V/Zq8df8ACQeCX8O30u6+0XCx7jy9ufuf98nK+w2+te1UAFFFFABRRRQAVxN78XfB1h40HhS61CVdXa5jtRCLaQjzHxtG7GP4hzmu2r498W/8neRf9jBY/wDoUVAH2FRRRQBHPMLe3kmZXcRoXKxqWY4GcADqfavG9S/aj8E2bMlpZaveupx8sCRr+O5gR+Ven+KfF2ieDNHfU/EV/HaW68KGOXlb+6i9WPsP5V8ufHDwTaT21n8SfCUL/wBia4qy3MZj2m3mb+IjsGOc9t2efmFAHcWn7Tl94i1600bwt4OEl3ezLDC1ze8ZJ6lVTgDqTngA19BR7xGvmlS+BuKjAJ74FfOf7L3w/wBkdx431KLl91tpwYdB0kkH/oAPs9fR1AFLWdXtNB0O81bUpPLtbOFppW/2VGePU9gPWsL4b+Obb4heCrbXLaNYZWZorm3DZ8mVeoz7ghh7MK8i/aj8dfZ9Ps/BlhL+8ucXV/tPSMH92h+rDd/wFfWuH/Zu8df8I347OhXsu2w1vES7jwlwP9Wf+BZK+5K+lAH1/RRRQAUUUUAFFFFAHF+Lvi14Q8D6yml+I9Qlt7t4VnCJbSSDYSQDlQR1U12SOJI1dOVYAj6V8i/tSf8AJWbb/sFQ/wDoyWvraz/48YP+ua/yoAmqC+uhY6fcXZhlnEEbSGKBdzuAM4Udz6Cp65vxt4+0DwBpH27xDeCMuD5FtH80s5HZF/Lk4AzyaAPLdR/as8Lwbl0zQ9VumH/Pby4QfyZj+lZ2hftG+IvGniyx0Hwv4Ws7ee8lCB7q4eYRr1ZyFCcBQT+FcR8dvAENnJaePPDlrLFouuqs00LxFGtpnG7lT90N19myO4r0T9mT4f8A9laBN4v1KHF3qSmKzDDlIAeW/wCBMPyUetAHvQ6c9ayPFfiO08JeFNQ13UT+4soTIVzgu3RUHuzEAfWtevmT9qPx19ov7PwZYS/u7fF1f7T1cj92h+indj/aX0oA90+HXjODx74FsNdhCpLKuy5iU8RTLw6/TPI9iK6evkn9mfx1/YXjKXw1fS7bLWceTuPCXCj5f++hlfcha+tqACiiigAooooAK5bxn8R/DXgA2Q8UXklr9u3+RsgeTds27vug4++tdTXzZ+1x9/wl9Lz/ANoUAfQ2javZ69otpqumSGWzvIlmhcqVLKRkHB5FXa5D4Tf8kh8L/wDYNh/9Brr6ACvMfHfx38PeAdel0bVNM1aa9jRXHlQoI3UjIYMzjI6jIHUEdq9Iu7u3sLOW7vp47e3hQvLLK4VUUdSSeAK8O8d22gfH7wvqkvg1J5tW8OSYtrp4CiXakZaNWPUHBxnBBA6BskAxdR/a1HK6T4SJ9JLm9/8AZVT+te2fD/WNc8Q+C7LWPE1pbWV1fL50dtbqw8uI/c3bicsRz26gY4r5E+DPw9fx18RIbS+gb+zdOP2jUAy4yFPEZ92bjHpu9K+3wAqgKAABgAdqAFrgfDvxW03xB8Vtc8HQbAdOjHkTBv8AXuvEy/8AASQB67WNWfi143XwF8O77VI3AvpR9nsVPeZgcH/gIBb/AIDXxP4b8SX3hnxZY6/YyFru0nE2Wb/Wf3lJ9GBIP1NAH6G0VQ0LWrPxFoFjrGmvvtb2FZoz3AI6H3HQj1FX6ACiiigAooooAK4nSfi74O1zxgPC+nahLJqpllhEJtpFG6MMXG4jHARu9dtXx78L/wDk65P+wlqP/ouegD7CooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigArgfAPxV03x34o8RaRZbFOlz/AOiuGz9pgGFMg/4GD+DJ71Q+PXjr/hC/hxPFaS7NT1bNpbYPzIpH7xx9FOM9iy18nfD3xfP4F8dadrsG5o4JNtxGv/LSFuHX8uR7gGgD7+oqG0u4L+ygu7OVZre4jWWKRTw6sMgj6g1NQAUUUUAFFFFAGF4u8ZaL4H0ZNU8R3L21o8ywK6RNId5BIGFBPRTUnhXxZpHjTQ11fw/cNcWbSNGHaNkO5evDAGvMP2pf+ST2v/YWh/8ARctXf2af+SNW/wD1+z/+hCgD1uiiigAooooAKKKKACiiigAooooAKKKKACiiigCnq+mQa1o13pl20q293E0MphfY21hggEdOOK+b/jL8HPCXgrw7o95oVvcxzXesQ2kpkuGcGNkckAHocqOa+nK8e/aR/wCRO8Of9jFbf+i5aAOi8J/Bfwj4K8QRaz4fivYLuNGTLXTMrKwwQQeo7/UA139FFABRRRQAUUUUAFFFFADZVZ4XRHaNmUgOoBKn1GePzr5U+MnwT0rwT4ck8QQa7qN9fXl6sSRXSozTSOSzEsACTgMelfVteMfEFv8AhNPj14R8IQfvLbRs6vqHdRjBRWH/AAFR/wBtaAOf8N/s06v4d1q3vrXxcFSSJoLtYbcxSCKRCkgRssN2GOCR1ANc38Wfgb4b8B6HpF5pF9qs0l9qsVlILqWNgEZXJI2xjn5R7e1fVdePftIf8il4a/7GK3/9Fy0Aavgf4F+H/h/4lTW9C1bWWnEbRPFcTRNHKjdmAjB6gHqOQK9MoooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAKOt6Wut6He6W9zcWiXkLQvNbMFkRWGDtJBAOO+K+Yvix8DfDfgTSNGutIvtVme/1WKylF1LGwVGViSNsY5+Udcj2r6rrx79o3/kWvC//YxW/wD6BJQBreBvgZ4f+H3iRda0LVdZabymieK4miaOVG7MBGD1APUcgV6XRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAUtZ0q313RLvSr1pVt7yJoZfJcoxRhggEdMjj8a+cPjB8HPCXg3S9Bn0O3uY3vtYitJvMuGfMbKxIGeh4HNfTlePftFf8gLwp/2MVv8A+gvQB0fhL4M+EvBGvprHh+O9gu1RozuumZXVhyCD1HQ/UCu+oooAKKKKACiszXfEui+GLE3niDU7bT4OzTyBS3so6sfYZNeWXPxl8QeM7mTT/hB4Yn1ABtj6xqCGO3j9wDjPryQf9k0Aet6nqthounyX2r3kFlaxDLzTyBFX8TXkuofGfV/Ft9LpHwd0CXV5VOyTVrtDHaw+/OM+vzEdOFapNN+B1zr1/Hq3xa8RXPiS8U7lsYnMdrEfQAYJH0Cj1Br1jT9OstJsYrLS7SGztYhiOGCMIij2A4oA8r0L4HnUdUTXfitrMvinVRytsxItIPYLxuHthV9VNbPxN+EOlfEOw0mFSmnyabOoV4UA/wBHJG+IAcDgZXsCPc16JRQBkX3hyzuvCMvhy0km02ye1+yIbMhXij27cKSCBxxnFfNfxV+BvhvwNp+hz6TfarM2oarFZSi5ljYKjBiSu2Mc8d8j2r6rrx79on/kDeEv+xit/wD0F6ANfwL8DtA+HviQazoWq6y8xiaGSG4miaOVG7MFjB4IBHPUCvSqgvL60061e51C6htbdOWlnkCKv1J4Feb69+0B4L0qf7Ho81z4i1BjtS20uEyBm/3zgH/gO76UAen1S1XWtM0Kya71rULawt16y3Mqxr9Mk9favJf7U+NPjz/kF6ZZ+BtNk/5bXh8y5K/7pBIP1VfrV3Sv2fdCe9Go+ONV1LxbqP8AE97Oyx/goJOPYsR7UAM1P4+2N9fPpnw40HUPFmoDjfBE0cCe5YjOPwA96p/8IP8AFP4gfP478Tp4b02T72laN98jurOD/NnHtXr2maTp+i2KWekWNvY2yfdht4hGo/AVboA4vwf8JfB3gjZLo+kxyXi/8vt3+9mz6hjwv/AQK2fGPhWx8a+E73QdU3LBdoB5iAbo2BBVhnuCBW3RQBxXiH4X6Trfw9t/BlndXWkaTBs+SyKgyBecPuBzlvmPckZzXzvd/Am1n+Mp8DaLrsvlxad9tubua3DmE54XaGGeqc5H3q+u5ZUgheWZ1SONSzMxwFA5JNeO/A5H8TeJPGPxCuEITV737NZFx8wgj/8AreWPqhoA0Ph38Ibzwj4S13wtrWtDUNJ1Q5X7MrQSJuXbIOp4ZQo4PY+teb/F74OeEvB1h4fl0S3uY2v9YitJ/MuGfMbBiQM9Dx1r6crx79oj/kE+EP8AsYrf/wBBagDpPCPwa8J+CNfXWPD0d5BdLG0R33TMrq3UEHqOAfqBXe0UUAFFFFABRRRQAVVbS7B7r7S9jbNPuDeaYVLZHQ5xnNWq8H1z45eI9M+OieCoLDS205tUtrMzPFJ52yQoGOQ+M/OccfhQB7xXC/Ef4gX/AIQaw03w/wCH7rW9b1UstnEiHyRtxkuw9Mg4445JA5qHxzY/ETWvE+n6b4SvrXRdDCrPd6nw825W/wBWEPY8HA4PIJA4PfgcDdgnHXFAGJqPhrSvFljpkvivRbeee1ZblIJiJBBKV+ZcjhgM49DgHHAq/q2j2OuaLdaTqdus9ldRGKWIjgqR29COx7EVdooAp6RpVpoejWml6ZEIbSzhWGFB2VRgfU+/erlFFAHjPj/4B+HNcuNd8Vajq+tyahJFLc7fOi8sFUO1APLztAAAGc4HWuK+E3wG8N+Mvh9pfiW/1PWbW/neUkWk8aKhSVlUrmMkH5QevWvpS9u4rDT7i8uM+VbxNK+0ZO1Rk/oKzfCXijTvGXhm21zRPN+xXJcR+amxvlcqcj6qaANeJDHCiNI0rKoBd8Zb3OABn6CnUUUAFFFFABRRRQBVudMsLyXzLuyt53AxuliVjj0yRVoDAwOBXhfxl+N/iL4deN4dG0ax0u4t5LJLgvdxyM+5mcEZV1GPlHauw8SaP8QPEnjLTF0vW4ND8M26RXUs1sM3E8gOTEQeq8f7uDzu6UATfEPxz4h0DVNO0Hwd4Zn1jVtRUyRzSDFtEithizA9RkdSANwOT0rqr7QNM1uXTrvXNLtbm6sW82AyqJPIkI52kjn8uwPUCtSigCjrei2HiLRLvSdXt1uLK7jMcsbdx6g9iDyD2IBqzaWsFjZw2lnEsNvBGscUaDARVGAB7ACpaKACvE/iH8B/C0ukeJfFMk+q3Oqi1ub4b7hSrShGcDAXpkAYHbivbKpazqkGh6Df6teLI1vYW0lzKsYBYqiliACQM4HqKAPB/hj8B/C+u+B9E8R30mq2mqSr5x8i4CCN1kO0qCpI+6D1r6FAwoBOfc96yfC3iSz8XeGLLXdMSaO1vULxrOoVwAxXkAkdR61rUAFFFFABRRRQAVBdWFne7fttrBcbM7fNjD7c9cZ+lT15J8cvivrXwybQxodpYXP9oifzftiO23y/Lxt2uv8AfOc57UAesxRRwxLHCixxqMKiDAA9AKyvFWvr4W8L32tPZXN+tnH5ht7Vd0j8449hnJPYAmuWfW/GXij4N6Zq/hBdPh1/UreGRvNBEUQf77ICTyuc854B4JwK6LwVomq+HvCttp+v61Lrd+hZpbuUYJLHO0dyBnAJ5+nQAGb4J1DXvF/hC5l8eeHoNMF67pHYsxYvbMOBIp6Hkg5xn0HSul0vSrDRNNh0/SLSGzs4F2xwwoFVR/j796t0UAZek+G9J0O+1O70qyjtp9UuPtN26D/WPtAz7dzj1Zj3NalFFAHGePfhfofxHez/AOEhuL9Usg3lRW0wRctjLEFTk8AV4n4e+CHhnU/jF4o8MXaaimmaXBDJayrNh3LKhOWK4PLHoBX0/XOaZ440zVvHWseE7aK6W/0iOOS4d0URMHCkbSGyfvDqBQBJ4L8Haf4F8Oroujz3ctmkjSRi6kDmPdyQCAOM5OPUmt+iigAooooAKKKKACqsel2ENz9oisbaOfJPmrCobJ6nOM96tV4P4N+OXiPxF8bF8HXlhpcdgbu6g82KKQS7YkkZTkuRk7Bnj1oA94ooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigDg/HXwg8PfETVoL/wAQ3Gpb7eHyYo7edURBkknBU8knk+w9K8c8AfBDw14l8XeM9N1ZNRittF1AW9m0c20um6QZYlSCfkHIxX0/XOeGvHGmeKtc17StOiukn0G5FtdNOihWfLDKEMSR8h6gdqALvhfw5a+EvDdpomnT3M1raKVhNy4dwpJO3IA4GcD24rWoooAKKKKACiiigCK5tbe8jEd3BFOgOQsqBhn1waLe2gtIvKtYY4I852RoFH5CuE+M3j3U/h14Ih1nRre0uLiS9S3KXasybWRyThWU5+Ud6sfCLxrqHj/wDFrmrwW0Fy9xJEUtVZUwpwOGYnP40AdxRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAV49+0j/AMid4c/7GK2/9Fy17DXj37SP/IneHP8AsYrb/wBFy0Aew0UUUAFFFFABRRRQAUUVFc3MFnay3N3MkEEKF5JZGCqigZJJPQUAUPE3iKw8J+Gr3W9Wk8u1s4i7erHoFHuSQB7mvOfgZoV9c2uq/EDxFGV1bxPMZY1b/llbA/IB7HjH+yqVgTS3P7QPjaOC3WaD4f6JNulkIKf2lOOw9sH8FJJwWAHu8caQxLHEipGihVVRgKB0AFADq8e/aQ/5FLw1/wBjFb/+i5a9hrx79pD/AJFLw1/2MVv/AOi5aAPYaKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigArx79o3/kWvC//AGMVv/6BJXsNePftG/8AIteF/wDsYrf/ANAkoA9hooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACvHv2iv+QF4U/7GK3/APQXr2BmCqWYgADJJPSvAv2h/GvhyfTfD9pZaxaXt1ZazFdTwWsolaONVbJOOAeRweeaAPfqCQASTgDqTXjP/C2vHXi/5fhv4AuFt3+5qWsnyoyPULkA/gx+lJ/wqHxl4zbzPij44uJLZuW0vSP3UOPQnAB/FSfegDqPFPxs8D+FGaG41ZdQvAcC004ee5PoSDtB9iQa5X/hKPi18RPl8KaHF4O0l/8AmIaoMzsvqqEcZH+yf96vQPC3w28I+DFU+H9EtoJ1H/Hy6+ZMf+Btkj6DArqaAPKdC+AWgQ3w1TxpfXni7VTy02oSHy8+yZOR7MSPavUba2gsraO3s4I7eCNdqRRIFVB6ADgCpaz9U1/R9Di8zWtVstPTGd11cJED/wB9EUAaFFeY61+0J8P9JfyrXUZ9XuM4EOnW7SFj6Bmwp/A1k/8AC0fiR4l48E/DWe2ib7t3rUnlqR67Ts/RjQB7JWXrXifQ/DkHm69q9lpyYyPtM6oW+gJyfwry/wD4QP4ueKP+Rq8fw6Lbv1ttFhIYD03jYf8Ax5q09F/Z78DaZP8AatTt7rXrwnc0+p3BfcfXauAfxBoAraj+0R4YN0bLwnp+q+J70/djsLVgpP1YbvxCmvNPizr3xE8UafoUuv8Ahy38Maa+rRCxDyeZcCchtpbnoBnqo/GvprTtK0/R7UW2k2FtYwDpFbQrGo/BQBXlP7RP/IG8Jf8AYxW//oL0AOs/2fdP1C6S9+IPiPVvFV2Ods87RQj2ABLAfRh9K9I0HwpoHhi38nw9o9np64wxghCs3+83VvxJrXooAKKKKACiiigAoormPHvj3SPh74cfVNYk3O2VtrVD+8uJOyqPT1PQD8AQDkPjn4puoNEtfBPh3Muv+JnFtHGh5jgJw7H0B5XPpuPau98I+G7Xwh4R07QbDmGxhEe7GPMbqzn3ZiT+Nef/AAo8Gatd6zdfEXx6h/4SDVFItbVxgWFueigH7pI4x1A68s1es0AFePftEf8AIJ8If9jFb/8AoLV7DXj37RH/ACCfCH/YxW//AKC1AHsNFFFABRRRQAUUUUAc1e/EHw1YeMrLwtPqcZ1i9JVLeMbth2kgORwpOOAeTxXi/iP4O+LtT/aCXxRBYxPo51a2uWl+1IreUhj3HbnOflPvXuOm+DPD2keIr/XtP0q3h1TUX33F0FyzHHOM/dzjJxjJ5Oa3KACiiigAooooAKKKKAIL60j1DT7izn3eVcRNE+04O1gQcfnWZ4S8K6f4L8MW2g6MZjZ2pcx+e+5vmcucnA7saua3aS3/AIf1Czt8ebcWskSbjgbmUgZ/E1zXwl8Laj4M+GemaFrXlfbLZpjJ5L7l+aV2GDgdmFAHZ0UUUAFFFFABXKD4leGJPHsPg+2vvtOrSK5ZYF3pEVGSrsOA2AeO2OcZGerrJ0rwtoeh6jf3+k6Xb2t3qMplup40+aVj1yfTPOBxkk9SaAPA/jP8FfGPi34jT6x4fX+0LOeCPH2i6RPIYZBjUMfu/wAX1Y19HWyGO1iR+GVAD9cVJRQAUUUUAFFFFABVPV7C11XRL7TtR/49Lu3kgnw23926lW57cE81crJ8V6XPrngzWtJtGjS4v9PntomlJChnjZQSQCcZPoaAF8M6Jpvhzw3Z6ToefsFqhWHMm/gsT97vyTWrXMfDjw3eeEPh3pGg6nJDJdWUTJI0DFkJLs3BIB6H0rp6ACiiigAooooA5zxn488P+AtJ+3+I70Qhs+Tbp80s59EXv9eAO5FeY/HTwLrvxT0/wtf+DbZLq3jhmlYzSrCQsoiKcNg9FPHavW9V8LaHrupWF/rGl297daczNayTJu8onGSB0PQHnoQD1rWoA5r4daDP4Y+HOiaNewrBc2lqqTxowYCTq3I68kmuloooAKKKKACiiigArn9N8KaHp3jbVvENlu/tfUo0S7zNuG1QoX5e3CiugrgvDvgfU9J+MvinxZczWrWGrwQx28aOxlUoqA7gVAH3T0JoA72iiigAooooAKKKKACvnTwN8IPGGh/HxfFOo2MMelC9vJjKtyjNtkSUKdoOerrX0XRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAVz/AId8KaH4f1nXNQ0bd9r1e5E99mbf+8yx6fw8s3FdBXBfD/wPqfhXxh401XUZrWSDXtQF1arA7FkXdIcOCoAPzjoT3oA72iiigAooooAKKKKAPNfjt4M1nxz4Ag0rw5Ak92l/HOyvKsY2BHBOTx1YVZ+CnhPVvBfw3h0jxBAkF4tzLIUSQONrHjkcV6DRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFePftI/8AIneHP+xitv8A0XLXsNePftI/8id4c/7GK2/9Fy0Aew0UVzvi7x54d8CW1tP4pvzZRXTlImEEkm4gZP3FOOvegDoqK8xl/aI+GkaZTXpZT/dSxnz+qCs6X9pfwQW8vT7XWtQl/hS2sxk/99MKAPX6K8cPxi8aa3+78IfC3VpN33bjUmMEf6gA/wDfVMbwv8ZvGeR4j8T2PhOxk+9a6Sm+YD03A5H4SfhQB3PjT4m+FvAVuW17UkFztyllB+8nk+ijoPdsD3rzhdH8afHK5in8TxTeF/BSuJI9OViLi+A5Bf275IA6YB+9Xa+Dvgv4Q8HXK30Nm+p6pncb/UW82QN6qPuqfcDPvXf0AU9I0iw0HSbfTNHtY7Sytk2RQxjAUf1Pck8k8mrlFFABXj37SH/IpeGv+xit/wD0XLXsNePftIf8il4a/wCxit//AEXLQB7DRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFePftG/wDIteF/+xit/wD0CSvYa8e/aN/5Frwv/wBjFb/+gSUAew0UUA56UAFFFFABRRRQAUUUUAFFFFABRRTXdY1LSMqqOpY4AoAdRWTdeK/D1jn7br2mW2OvnXkaY/M1h3vxd+H+ngmfxdpb4/54Tib/ANAzQB2VFeVX37R/w5tCRb6jd3x9LeykGf8AvsLWTdftJWci50LwXr996NPGsK/mN+BQB7XRXz7cfHHx1fZ+waP4V0WM/wAWra3CWX/gPmI3/jtY95488VX+Rrvxi8N6NA3VNHt2uHH4rHn8noA+mWZUUs5CqBkknAFctrXxO8E+H1b+1fE+mxuvWKOYSyD/AIAmW/SvnO5uPhXdN5ni34i+K/FcgOSqxSRo34SAkfgRW3o3iX4R6aVPhj4Xa3rEw+689l9oDH1G93x+C0Ad3cftE6NeztbeC/DmueJbgHAFtalUP48sP++ajOufHDxX8uleH9K8I2z9J7+USyr+HP6x023+KXjqeBbfwz8HtQtoRxH9rl+zqB/ulFA/OpP7W+PeqcQ+HvDmjIejTzeYw/75kYfpQA1fgTqPiFhL8RvHmsa1k7jaWzeTAD6YORj6Ktc38Xvhz4U8EaD4XPhrSIrSWXX7eOScs0kjrtc4LMScZAOOnFdN/wAIX8adW+XV/iHp+mxnqum2gYj8diH9a8++Lnwwv9A0vQrjXvGmseIZL3VorRlunISMMrEsqlmweP1oA+j9S8T6Do2f7X1vTrHHUXN0kf8A6ERXHat8e/hzpO5W8QJdyD+CzheXP/AgNv61W039nj4cafgyaNLeuP4rq7kb9FIX9K7HSfA/hbQtp0jw7plmy9JIrRA//fWMn86APOf+F/yatx4M8AeItbB+7I0PlRn33KHwPrij/hIPjnr/AMum+FdG8Owv0mv5/MdfwDH9Ur2OigDxz/hVPxE8QfN4x+KF3GjffttIi8lSPTcNn6qav6X+zr4BsZfO1C2vdZnJyZdQu2JJ9SE2g/jmvVKKAMnRvCugeHU26Fothp/GCba3VGP1IGT+Na1FFABRRRQAV49+0T/yBvCX/YxW/wD6C9ew149+0T/yBvCX/YxW/wD6C9AHsNFFcJrvxp8B+GtcudH1rWmt761YLNELOZ9pIB6qhB4I6GgDu6K8ruf2j/hxBnytUurn/rlZSD/0ICqJ/aP0O9+Tw54Z8R6xL2EFmNp/EMT+lAHsVIzKiFnYKqjJJOABXjh8e/F7xJ8vhn4eRaNE3H2jWZ/mUeuw7D+hpv8AwpzxZ4wdZPil45uby3zltM0oeVCfqcAH/vjPvQBp+Lfjjpdjff2H4Ftn8VeIZDsjgswWhjPqzjqB3C/iVqHwZ8KtSvfES+M/irdJqmv5DWtkpzb2A6gAdCR2xwDz8x+au98L+C/D3gyw+yeGtLgsUI+d1GZJP95zlm/E1uUAFFFFABXj37RH/IJ8If8AYxW//oLV7DXj37RH/IJ8If8AYxW//oLUAew0UUUAFFFFABRRRQAV8i+K9QvU/a0it0vJ1h/t+xXyxKwXBaLIxnFfXVebaj8DfDWp/EZfGk99qq6kt5FeCJJoxDvjKlRgx7sfIM/NnrzQB6TRRRQAUUUUAFFFFAFHW7uSw8P6heW+3zbe1klTcMjcqkjP4iua+EvinUfGnwz0zXdZ8n7ZctMJPJTavyyugwMnsorrL67j0/T7i8n3eVbxNK+0ZO1Rk4/Kszwl4q0/xp4Ztte0YTCzui4j85NrfK5Q5GT3U0AbVFFFABRRRQAUUUUAfJf7TuoXlr8VbZLa7nhT+y4jtjlKjPmSc4Br6utDmxgz/wA81/lXn/j74I+G/iL4hj1nW73VYLiO3W3C2ksaptUsQcNGxz8x716JHGIokjXJCKFGfagB1FFFABRRRQAVkeLr+60rwTrmoad/x+Wmn3E8Hy7v3ixsy8d+QOK16pa1qsGhaDf6tdrI9vYW0lzKsQBYqiliACQM4HGSKAMP4aa3qXiL4b6Pq2u/8hC6iZpsR7OQ7D7vbgCuprI8K+JLPxf4Xste0yOeK1vULxpcKFcAMV5AJHUeta9ABRRRQAUUUUAFfOf7WF5dWj+FPslzNBuF5u8uQrnHk9cV9GVxHxF+FOh/E06cdeutQt/7O83yvsUiJu8zbnduRv7gxjHegC18KpHl+EvhmSV2d206IszHJJ2+tdbWb4e0O28NeHLDRbF5ZLawgWCJ5iC5VRgEkADP0ArSoAKKKKACiiigArz7w54q13UPjV4r8P32P7I06CF7P9zt+ZkQt8/8XLGvQa5vS/HOmat481jwnbQXS3+jxxyTySIoiYOFI2kMSfvDqBQB0lFFFABRRRQAUUUUAFfIvwy1C9l/anSCW8neH+0dQHltKxXAjmxxnFfXVebaD8DfDXh34hDxjZX2qyagJ5p/KmmjMW6VWVhgRg4w5xz6daAPSaKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAK8++HfirXdf8ZeN7DWsfZNJ1EQWGIdn7vdIOv8AFwq816DXN+GfHOmeKtc1/StOguo59Buha3TTooV2ywyhDEkfIeoHagDpKKKKACiiigAooooA8c/aduZ7X4VWr200kL/2rEN0blTjy5eMirn7N9xNc/B+3kuZZJX+2TjdIxY9R3Ndh4+8BaX8RfD0eja3PdwW8dwtwGtHVX3KrADLKwx8x7VL4H8Fad4A8MpoejTXU9skryh7p1Z8scnlVUY/CgDoqKKKACiiigAooooAKKKKACiiigAooooAq6pfppekXmoSozpaQPOyr1YKpYgflXh//DWHhv8A6F/Vf++o/wD4qvYvF/8AyI+u/wDYOuP/AEW1fJ/7Pfgnw/448VapZ+KNP+3QW9kJY086SPa29RnKMCeDQB6d/wANYeG/+hf1X/vqP/4qsbxL+0H4B8X2dta+IfCerXcNrcrdQqJxHtlUEBsq4J4Y8HivS/8Ahn34Y/8AQsj/AMD7n/45R/wz78Mf+hZH/gfc/wDxygDjf+GsPDf/AEL+q/8AfUf/AMVVa9/ad8GakirqPhK8u1Q5UTpC4U+2TxXd/wDDPvwx/wChZH/gfc//ABymS/s//DJYXI8M4IUkf6fc/wDxygDH+G/xO8FeP/FLaPo/hBLC4W3a482W2hAwpUY+XnPzV7DHFHCm2JFRfRRgV8hfsvf8lcl/7Bk3/ocdfX9ABRRRQAUUUUAVdUv00rR7zUJUZ47SB52VerBVLED8q8P/AOGsPDf/AEL+q/8AfUf/AMVXsPjH/kRde/7Btx/6KavlH9nvwT4e8ceKdVtPFGn/AG6C3sxLEnnSR7W3gZyjAng0Aem/8NYeG/8AoX9V/wC+o/8A4quf8X/HvwL43sLO01vw/rnlWd2l5F5M0SHzFBAzycj5jxXqX/DPvwx/6Fkf+B9z/wDHKP8Ahn34Y/8AQsj/AMD7n/45QBxv/DWHhv8A6F/Vf++o/wD4qj/hrDw3/wBC/qv/AH1H/wDFV2X/AAz78Mf+hZH/AIH3P/xymTfs/wDwyWCRl8M4IUkf6fc+n/XSgCP4d/HPR/iN4mfRdO0q+tJlt2uDJOUK4UqMcHr81en18g/su/8AJW5/+wXN/wChx19fUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQBU1bUE0nRb3UZkaSOzt5J2RerBFLED34rxD/AIaw8N/9C/qv/fUf/wAVXr/jT/kQtf8A+wZc/wDopq+VP2evBHh7xx4m1a08U6f9ugt7NZIk86SPa28DOUYE8etAHpn/AA1h4b/6F/Vf++o//iq5/wAX/HvwL42s7K21rw/rnl2V2l5F5M0SnzFBAzycj5jxXqX/AAz78Mf+hZH/AIH3P/xyj/hn34Y/9CyP/A+5/wDjlAHG/wDDWHhv/oX9V/76j/8Aiq8f1XWfhVq2sXmpzWXjKK4vJ3nk8q7tgoZ2LHGUJxk+tfSX/DPvwx/6Fkf+B9z/APHKjufgB8M0tZXTwzhlQkH7dc8HH/XSgDwz4eeGfDnxA8TyaN4a8QeK9KkjtmuN9xJGwwrKMfIRz8wr1FfgF4jg4svizr8A7AeZ/SYV5j+y3/yVm5/7BU3/AKMir68oA8T/AOFIeOV+58Zte/4Esx/9uKRvg18RU/1Pxg1Vv99JR/7VNe20UAeHn4RfFFfufFm8P+8Jf/i6P+FRfFI9fizefh5v/wAVXuFFAHhp+DvxOb73xbvx9DN/8XTH+B/xDmGJ/i9qxB6gPPj8vNr3WigDwCT9nLxPc/8AH18UdQkz13QSt/Oaqb/soyXL7r7x1NOfVtOJP5mY19F0UAfPcX7Jelj/AF3iq8f/AHLRV/8AZjVyL9lDwwuPO1/Vm/3BEv8A7Ka94ooA8Zh/Zg8GIuLjU9euB/de7jA/IRitC3/Zu+HMOPM067uP+ul7IM/98kV6tRQB57b/AAI+Gttjy/C8Lf8AXS4mf/0JzWxafDDwLY4+z+EdGyOhkskcj8WBNdVRQBRtNE0rT8fYNMs7bHTybdUx+Qq9RRQBT1fUU0fRL7U5kaSOyt5Lh0TqwRSxA9+K8R/4aw8N/wDQv6r/AN9R/wDxVeveN/8Akn3iH/sF3P8A6Kavlb9nrwP4d8ceI9XtfFOn/bobe0WSJfOkj2sXxnKMCePWgD0v/hrDw3/0L+q/99R//FVjeI/2g/APi23tIdf8J6tdR2dyt1ABOE2SqCA3yuM9TweK9L/4Z9+GP/Qsj/wPuf8A45R/wz78Mf8AoWR/4H3P/wAcoA43/hrDw3/0L+q/99R//FUf8NYeG/8AoX9V/wC+o/8A4quy/wCGffhj/wBCyP8AwPuf/jlQ3fwB+GcdlO6eGsMsbEH7dc8ED/rpQAfDr446R8R/EsujabpV7aSx2zXJkuChUhWVccHr8wr06vkT9ln/AJKxd/8AYJl/9GxV9d0AFFFFABRRRQBT1jUo9H0O+1OZGkjsraS4dE6sEUsQPfivEf8AhrDw3/0L+q/99R//ABVeu+Of+SeeIv8AsFXP/opq+WP2evA/h3xx4i1e28U6d9uhtrVJIl86SPaxfBOUYZ49aAPSv+GsPDf/AEL+q/8AfUf/AMVXP+Lfj34F8aW9hDrXh/XNlhdpeQ+TNEv7xQQM8nI5PFepf8M+/DH/AKFkf+B9z/8AHKP+Gffhj/0LI/8AA+5/+OUAcb/w1h4b/wChf1X/AL6j/wDiqz7n9o/wDeXDz3fgiaeZzlpJYIGZvqT1r0L/AIZ9+GP/AELI/wDA+5/+OVDefAH4ZxWM8kfhrDJGzKft1zwQP+ulAFL4Y/Fbwp488Sy6NoXhhtNmitmuTI8MSrhWVcfL3+cV63XyL+yx/wAlXvP+wRL/AOjYq+uqACiiigAooooAp6xqUejaFf6pMjSR2VtJcOidWCKWIHvxXiP/AA1h4b/6F/Vf++o//iq9c8d/8k68R/8AYKuv/RTV8s/s8+BvDvjjxBrFt4p077dDbWqSRL50ke1i+Ccowzx60Aelf8NYeG/+hf1X/vqP/wCKrG8R/tB+AfFkNnFr/hPVbpLK5W6gHnhNkq5w3yuM9TweK9L/AOGffhj/ANCyP/A+5/8AjlH/AAz78Mf+hZH/AIH3P/xygDjf+GsPDf8A0L+q/wDfUf8A8VR/w1h4b/6F/Vf++o//AIquy/4Z9+GP/Qsj/wAD7n/45UF78AfhnFp9xJH4aw6RMyn7dc8ED/rpQA74c/HDSPiR4km0bTdLvbSWK1a5MlwUKkKyrjgnn5x+Vem18jfssf8AJVr7/sES/wDo2GvrmgAooooAK4a9+MPhDT/G6+E7m7uF1ZrmO1EYtmK+ZJtCjd0/iHNdzXx74t/5O8i/7GCx/wDQoqAPsKiiigAooooA4D4jfF/Qfhytut6p1C4mkKNbWk0ZkiwAcspOQOR+dcjo37Tvh/WtesNLg0LU45L65jtkd2j2qXYKCeenNeX/ABG8L3ms/tMTpc6TeXGmXWp2cczrC+xoysSv846DGeQeK99sPgX8OdM1K2v7Hw75VzaypNDJ9tuDtdSGU4MmDggdaAO5vbSO/wBPuLO43eVcRNE+04O1hg4/A14fL8X/AAl8GLmXwFZaNq1xFpLkCUyRtuMn708kju/pXu9cPrvwb8B+JtcudX1vQftV/dEGab7XOm4hQo4VwBwB0FAFP4afGLS/ibqF9aaZpt5ZtZxLK7XBUhgTjAwTXolcv4T+G/hPwNdXFx4W0r7DLcoElb7RLJuUHIHzscc+leD/ABo8f+K9C+NR0vR9evbOx22x+zxSYX5gM8e9AH1BRRRQAUUUUAcR4w+L3hLwLrSaV4iu54bt4VnCx27ONhJAOR7qa7WNxJGrr91gCPpXyL+1J/yVm2/7BUP/AKMlr62s/wDjxg/65r/KgCaiiigArJ8SeJNO8LaHdanqlxFGsEEkqRPKqNMUUtsTcRljjGPetavB/wBqbTr7UfDmgLp9ncXTJdylhBEzlRsHXAoAb/w1h4b/AOhf1X/vqP8A+Kr1XQtW074l/Dlb1raaLT9atZYXgkbDhCWjYZHQ4B6V5h8Lvgl4J1r4Z6PqHibw3I+qTxubhpbm4iYkSMBlQ4A4A7V7JoWhad4a0S20jRLb7LY2oKww72faCSx5YknknqaAOH8VeL9A+BXhDRrCPTby4sGd4LeOKQMyY+Y5LHn71c3ov7Tnh/W9f0/SoNC1KOW+uo7ZHdo9ql2CgnB6c16Z4s8C+HPHNtb2/inTvt0Vs5eJfPkj2sRgn5GGfxrn9P8AgZ8OtL1K21Cx8O+VdWkyTwyfbbg7HUhlODJg4IHWgD0CivJf2jPEeseGPh1Y3vh/UZ9PuX1WOJpYG2sUMUpK/TKg/hWn8B9c1PxD8KLPUNcvZr67e4mVppmyxAcgDP0oA9HooooAK5Txt8SfDnw+NkPE1zNB9u8zyPKgaTOzbuzjp98V1dfNn7XH3/CX0vP/AGhQB9C6LrFp4g0Oz1bTXZ7S8iWaFmUqSpGRwelXq5D4Tf8AJIfC/wD2DYf/AEGuvoAKKKKAGTTR28LzTyLFFGpd3dgFVQMkknoK8d8S/tK+GfDviS80ldOu9RFq4X7VZyxvFJwD8p3ds4+oNejePI3m+HHiWKFGkkfSbpURRksTC2AB3NfNHwF+F+m+KNY1eLxz4fupIYII2txOZrcBixBwVK54oA9z+Gnxh0v4m319a6Zpt5ZtZRLI5uCpDBiRxgn0rZvND0Pwhf8AiPx5HazNfS2TTXm2UnzEhjyAqngHCCneE/hv4U8DXNzP4W0r7DLdIEmb7RLJuAOQPnY4/Cug1CwttV0y60+/j821u4Xgmj3Fd6MCrDIwRkE8jmgDw3/hrDw3/wBC/qv/AH1H/wDFV7B4P8TW/jLwjYa/ZwS28F6jOkUuNy4YrzjjtXI/8M+/DH/oWR/4H3P/AMcrrrbR7Hwr4NfTNAhNnaWVtJ9njEjP5fDN95iSeSTyaANqivmn9nLx34o8T/ES/svEGuXmoW0elSSrFPJuUOJYgG+uGI/GvpagAooooAK4bR/jD4Q13xmPC2nXdw+qmWWERtbMq7owxb5jx0Q13NfHvwv/AOTrk/7CWo/+i56APsKiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA5vxv440rwJ4dudU1SVHeFA6WiyosswLBfkViM4zXlP/AA1h4b/6F/Vf++o//iqwv2ptJ1HUfEegNp9hdXSpaSBjBCzhTvHXArufCnwH+Ht94N0W71Tw0Tez2EEtyXvLhCZGjUtlRIMHJPGBigD1HRdTj1vQNP1WBGjjvraO5RH+8odQwBx35ryzxb4y8J/A7xNdTro9/cXfiZze3TwzBlLKx7MeOXPSvWLCxt9L022sLGPyrW1hSGGPcTsRQFUZPJwAOtc/4s+G/hTxzdW9x4p0r7dLbIUib7RLHtUnJHyMM8+tAHE+Dv2iND8Z+LrDw/ZaNqFvPesypLKybVwhbnBz/DXr1cPoXwa8BeGdcttX0TQfst/aktDN9rnfaSCp4ZyDwT1FcF+0v4u1/wAK2/hw+HNWudONy9yJjbvt37RHjP0yfzoA91ork/hdqN5q/wALfD9/qdxJc3dxaK8s0hyztk8k11lABRRRQBg+MfGmjeBdFTVfEU0kNq8ywK0cRc7yCQMD2U0/wn4t0nxroK6xoEsktm0jRhpIyh3L14NeZftS/wDJJ7X/ALC0P/ouWrv7NP8AyRq3/wCv2f8A9CFAHrdFFFABRRRQAUUUUAFFFFABRRRQAUUUUAY/i/8A5EfXf+wdcf8Aotq+bf2UP+R41v8A7Bw/9GLX0l4v/wCRH13/ALB1x/6Lavm39lD/AJHjW/8AsHD/ANGLQB9VUUUUAFMm/wCPeT/dP8qfTJv+PeT/AHT/ACoA+Q/2Xv8Akrkv/YMm/wDQ46+v6+QP2Xv+SuS/9gyb/wBDjr6/oAKKKKACiiigDG8Y/wDIi69/2Dbj/wBFNXzf+yf/AMjtrn/YOH/oxa+kPGP/ACIuvf8AYNuP/RTV83/sn/8AI7a5/wBg4f8AoxaAPqmiiigAqO4/49Zf9w/yqSo7j/j1l/3D/KgD5F/Zd/5K3P8A9gub/wBDjr6+r5B/Zd/5K3P/ANgub/0OOvr6gAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAMTxp/wAiFr//AGDLn/0U1fOX7J//ACOWu/8AXgv/AKMFfRvjT/kQtf8A+wZc/wDopq+cv2T/APkctd/68F/9GCgD6oooooAKhu/+PGf/AK5t/Kpqhu/+PGf/AK5t/KgD5J/Zb/5Kzc/9gqb/ANGRV9eV8h/st/8AJWbn/sFTf+jIq+vKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAw/G//JPvEP8A2C7n/wBFNXzr+yd/yN2vf9eKf+jK+ivG/wDyT7xD/wBgu5/9FNXzr+yd/wAjdr3/AF4p/wCjKAPqaiiigAqC+/5B1z/1yb+RqeoL7/kHXP8A1yb+RoA+S/2Wf+SsXf8A2CZf/RsVfXdfIn7LP/JWLv8A7BMv/o2KvrugAooooAKKKKAMLxz/AMk88Rf9gq5/9FNXzv8Asnf8jZr/AP14p/6Mr6I8c/8AJPPEX/YKuf8A0U1fO/7J3/I2a/8A9eKf+jKAPqWiiigAqvqH/IMuv+uL/wAjViq+of8AIMuv+uL/AMjQB8m/ssf8lXvP+wRL/wCjYq+uq+Rf2WP+Sr3n/YIl/wDRsVfXVABRRRQAUUUUAYPjv/knXiP/ALBV1/6Kavnn9k3/AJGrxB/15R/+h19DeO/+SdeI/wDsFXX/AKKavnn9k3/kavEH/XlH/wCh0AfUlFFFABVbUv8AkFXf/XF//QTVmq2pf8gq7/64v/6CaAPk/wDZY/5Ktff9giX/ANGw19c18jfssf8AJVr7/sES/wDo2GvrmgAooooAKgaxtGn85rWEy5B8wxjdn1zU9eCa78cPE2mfHdPBdvaaW2mtqltZmR4ZDNskKBjnfjPzHHH4UAe90UUUAFFFFABRXzh4s+LXjPTP2hx4XstWWPSDqlpb/Z/ssR/dyCPcNxXdzuPOc819H0AFFFFABXjvxA+AjeOfiAfEw8RixBEQ+zfYfM+4APveYOuPSvYq5/VfHnhXQ9V/szV9fsLO9+U/Z5pgr/N0496AOgooooAKKKKAIJrK1uH33FtDK2MbnjDHH41P06V4R8aPjb4l+HfjiHR9EtNMmtnsY7gtdwyM+5ncEZV1GPlHavc7eQy20UjYy6Bjj3FAElFFFABRRXkH7QnjzxF4E0PRrjwvfizluriRJWMCSblCggfOpx+FAHr9Fcf8KNe1HxP8LdF1jWpxcX11G7TShFTcRIyjhQAOAOgrsKACiiigDiPit8Oj8TfCtvow1T+zPIvUuvONv527ajrtxuX+/nOe1XPht4JPw+8FQeHzf/2h5Msknn+T5WdzZxt3N0+tbOueIdI8NWKXmv6jb6dbPIIlluHCqXIJC5PfCk/hUmj61pviDTlv9Evob60diqzQPuUkHBGaAL1FFFABUU9rb3O37TBFNt+75iBsfnUteRfHX4q678M20IaBb2E/9oC4837ZG748vy8bdrrj75z17UAetoiRxhI1VEUYCqMAU6sHwPrdz4l8CaNrV+kSXN9aRzyrCCEDMMnAJJx+JreoAKKKKACisbxhf3OleB9d1Gwk8u6tNOuJ4X2htrpGzKcHg8gcGvIf2e/iV4q8d61rNv4o1JbyO1t43iUW8ce0liD9xRnp3oA93ooooAKhvLf7XYz2+7Z50bR7sZxkYzU1NkkSKJpJGCogLMx6ADqaAPJfhT8DD8MvFVzrJ8QjU/PsmtfJFl5O3c6Nuzvb+5jGO9euVg6H448MeJb57Pw/rtjqNykZlaK3mDMEBALYHbLAfjW9QAUUUUAFQJY2kc3nR2sKy5J3rGA2T15qevBPBnxw8TeIfjcvg+9tNLTTjeXcHmRQyCXbEkjLyXIzlBnj1oA97ooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACivB/2hPiX4q8Ca5o1v4X1JbOK6tpHlU28cm5gwAPzqcfhXsHhG/uNV8E6HqF9J5lzd6dbzzPtC7naNWY4HA5J6UAbFFFFABXnHxd+Ex+KcWkoNZGl/2c0pybXzvM37P9tcY2e/WvR6yNe8V6D4XWA+ItWtNNFwWEJuZAm/bjOM+mR+dAEfg3w7/wifg3TNBNz9r+wQCHz/L2b8d9uTj8zW3VbT9QtNW0+G+024jurWdd8U0TZVx6g1ZoAKKKKAI5oIbhNlxEkq5ztdQwz+NLFDFbx7II0iTOdqKAP0rgfjR481T4d+B4dY0OG0muJL5LcrdozJtZHJOFZTnKjvVj4QeNNS8ffD+LXNZitorp7iSIraoyphTgcMxOfxoA7miiigAooooAKKKKACiiigAooooAKKKKAMfxf/yI+u/9g64/9FtXzb+yh/yPGt/9g4f+jFr6S8X/APIj67/2Drj/ANFtXzb+yh/yPGt/9g4f+jFoA+qqKKKACmTf8e8n+6f5U+mTf8e8n+6f5UAfIf7L3/JXJf8AsGTf+hx19f18gfsvf8lcl/7Bk3/ocdfX9ABRRRQAUUUUAY3jH/kRde/7Btx/6Kavm/8AZP8A+R21z/sHD/0YtfSHjH/kRde/7Btx/wCimr5v/ZP/AOR21z/sHD/0YtAH1TRRRQAVHcf8esv+4f5VJUdx/wAesv8AuH+VAHyL+y7/AMlbn/7Bc3/ocdfX1fIP7Lv/ACVuf/sFzf8AocdfX1ABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAYnjT/kQtf/AOwZc/8Aopq+cv2T/wDkctd/68F/9GCvo3xp/wAiFr//AGDLn/0U1fOX7J//ACOWu/8AXgv/AKMFAH1RRRRQAVDd/wDHjP8A9c2/lU1Q3f8Ax4z/APXNv5UAfJP7Lf8AyVm5/wCwVN/6Mir68r5D/Zb/AOSs3P8A2Cpv/RkVfXlABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAYfjf/AJJ94h/7Bdz/AOimr51/ZO/5G7Xv+vFP/RlfRXjf/kn3iH/sF3P/AKKavnX9k7/kbte/68U/9GUAfU1FFFABUF9/yDrn/rk38jU9QX3/ACDrn/rk38jQB8l/ss/8lYu/+wTL/wCjYq+u6+RP2Wf+SsXf/YJl/wDRsVfXdABRRRQAUUUUAYXjn/knniL/ALBVz/6Kavnf9k7/AJGzX/8ArxT/ANGV9EeOf+SeeIv+wVc/+imr53/ZO/5GzX/+vFP/AEZQB9S0UUUAFV9Q/wCQZdf9cX/kasVX1D/kGXX/AFxf+RoA+Tf2WP8Akq95/wBgiX/0bFX11XyL+yx/yVe8/wCwRL/6Nir66oAKKKKACiiigDB8d/8AJOvEf/YKuv8A0U1fPP7Jv/I1eIP+vKP/ANDr6G8d/wDJOvEf/YKuv/RTV88/sm/8jV4g/wCvKP8A9DoA+pKKKKACq2pf8gq7/wCuL/8AoJqzVbUv+QVd/wDXF/8A0E0AfJ/7LH/JVr7/ALBEv/o2Gvrmvkb9lj/kq19/2CJf/RsNfXNABRRRQAV88a/8HfF2oftBx+Lba2tjpK6ta3Zc3ChvLjMZY7euflPFfQ9FABRRRQAUUUUAeGeI/gXres/GweNINT0+OyGoW115D7/M2xbMjhcZOw45r3OqEuu6RDffYptUso7rcE8h7hBJuPQbc5ycj86v0AFFFFABXx78fv8Ak4A/7lp/IV9hVz+q+AvCuuar/aesaBY3l98o+0TRBn+Xpz7UAdBRRRQAUUUUAfPnxz+EHizx34+g1Xw9bW8tqlhHAWluFQ7w7k8H2YV7/boY7WJG+8qAH64qSigAooooAK8y+NXwy1L4maTpdppN7a2j2c7yObndhgVA42g16bVW+1Ow0xEfUr62s1c4U3Eyxhj7ZPNAGH8OfDFz4N+HulaBfTRT3Fkjq8kOdjZdm4yAejV01RW11b3tulxZzx3EL8pJE4ZW7cEcGpaACiiigDxP9qr/AJJXp3/YZi/9EzVrfs3f8kVsP+vm4/8ARhr0TXPDukeJrFLPxBp1vqFskglWK4TcocAgNj1wxH41Jo+iaZ4f01dP0SyhsbRGLLDAu1QSck4oAvUUUUAFeM/tA/DXxH8Q28Pnw1BBL9hFx53mzCPG/wArbjPX7hr2aigDnvAOj3fh/wCH2h6TqSql3ZWUcMyo24BgMHB710NFFABRRRQBl+JtLl1zwlq+k27pHLf2M1sjvnarPGVBOO2TXmnwV+D+r/DLVdUutW1Cyu1vYUjQW2/KlWJ53AeteuyyxwQvLO6xxxqWd3OFUDkkk9BVWx1jTNTd103UbS8ZBlhbzrIVHvg8UAXKKKKACqesf8gO/wD+vaT/ANBNXKbJGksTRyqGR1Ksp6EHqKAPkv8AZU/5KnqX/YGl/wDR8NfW1YOheBvDHhm+e88P6FZadcyRmJpbeIKxQkErn0yoP4VvUAFFFFABXzx4H+Dvi7Qvj0virULa2XSxe3kxdbhWbbIkoX5evVxX0PRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAHj/xq+D2r/EzVtLu9J1CytEs4HjcXO/LEsDxtBr03w1pkmieE9J0q4dJJbGyhtndM7WZECkjPbIqe+1jTNMdE1LUbSzZxlVuJ1jLD2yeatRSxzwpLC6yRyKGR0OQwPIIPcUAOooooAK+cv2t/+PXwp/v3f8oq+jayNf8ACegeKVgHiLSbXUhbljCLiMNs3Yzj64H5UAYPwd/5I54Z/wCvFf5mu1qtp+nWmk6fDYabbx21pbrsihiXCovoBVmgAooooA83+OfgvWPHngGDSvD0UUt0l/HOVlkCDYEcHk+7CrPwW8I6r4J+HMOj69HHHeLcyyFY5A4wx45Fd/RQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQBj+L/wDkR9d/7B1x/wCi2r5t/ZQ/5HjW/wDsHD/0YtfSXi//AJEfXf8AsHXH/otq+QPhFreoeHNI8bato1x9mvrXR1eGXYr7T58Y6MCDwe4pSlyq7HFczsj7XorwS21j4s6l8KB46TxVZWSQWRuF07+zYnNyiD5neQj5WbBO1RjGOlXtM8Q/Evx58Of+Ew0bXNP8PRw27tHYpYrObtowQ7M752AsrYABwOppy929+m/9f0xR961uu39f0j22mTf8e8n+6f5V5v4A+LdhqvwrtfE3jS+stLl82S3lZn2LK6c/Ip5JIwdoz3rr/DHiix8Z+D7fX9JSeO0vFk8tbhQr/K7IcgEjqp79KbVnYSd1c+XP2Xv+SuS/9gyb/wBDjr6/r5A/Ze/5K5L/ANgyb/0OOvr+kMKKKKACiiigDG8Y/wDIi69/2Dbj/wBFNXzf+yf/AMjtrn/YOH/oxa+kPGP/ACIuvf8AYNuP/RTV8c/C/wAQan4W0HxlrGhXP2W/t9Oh8qby1fbm5jU8MCDwT1FKUuVXY4rmdj7forwW51n4sf8ACq08f/8ACU2NtHHZx3P9kjTo2E0eBl2kIyGYfNtUAc4GKu/8JD8TPFXw0bx3pGu6foMEVo9xHpaWKz/aFjB3s0r5KklWwAOBjJ705e7e/TcUfetbrse21Hcf8esv+4f5V554V+LFpdfB2x8ZeKMW7yMYJIrZCxmmDlQsa9SWxnHbnsK1vDHxG0nxjNqGnW9nqWl6lZw+ZLYaramCYIej7cnI/HuPWm002hJ3Vz5y/Zd/5K3P/wBgub/0OOvr6vkH9l3/AJK3P/2C5v8A0OOvr6kMKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigDE8af8iFr/AP2DLn/0U1fOX7J//I5a7/14L/6MFfRvjT/kQtf/AOwZc/8Aopq+PPhd4g1Pwt4e8Z6xoVz9lv7fT4PKm8tX25uY1PDAg8E9RSlLlV2OK5nY+3qK8FudZ+LH/Cq08f8A/CU2NtHHZx3P9kjTo2E0eBl2kIyGYfNtUAc4GKu/8JD8TPFXw0bx3pGu6foMEVo9xHpaWKz/AGhYwd7NK+SpJVsADgYye9OXu3v03FH3rW67HttQ3f8Ax4z/APXNv5V554X+LdncfB/T/F/idTDNO5tmgtIy7TzhyoWNOpLYzjtzWz4b+Iek+Mk1OytbXUdN1Gyi3T2Gp2xgnRWHDbcng/X+dNpptdhJ6XPnP9lv/krNz/2Cpv8A0ZFX15XyH+y3/wAlZuf+wVN/6Mir68pDCiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAw/G//ACT7xD/2C7n/ANFNXzr+yd/yN2vf9eKf+jK+ivG//JPvEP8A2C7n/wBFNXx/8LvEGp+FvDXjXWNCufst/b2Fv5U3lq+3NwinhgQeCeopSlyq7HFczsfblFeC3Os/Fj/hVaeP/wDhKbG2jjs47n+yRp0bCaPAy7SEZDMPm2qAOcDFXf8AhIfiZ4q+GjeO9I13T9BgitHuI9LSxWf7QsYO9mlfJUkq2ABwMZPenL3b36bij71rddj22oL7/kHXP/XJv5GuW+F/jGbxv8N9P1/UY44LmQOlwE4TcjFSwz0Bxn2zTvDnj/RvHVjrv9g/aGi012t3mlQKkp2khkwSSp9SBTaabQk7q585fss/8lYu/wDsEy/+jYq+u6+RP2Wf+SsXf/YJl/8ARsVfXdIYUUUUAFFFFAGF45/5J54i/wCwVc/+imr53/ZO/wCRs1//AK8U/wDRlfRHjn/knniL/sFXP/opq+Qfhb4g1Pwt4Y8a6xoVz9lv7ext/Km8tX25uEU8MCDwT1FKUuVXY4rmdj7borwW51n4sf8ACq08f/8ACU2NtHHZx3P9kjTo2E0eBl2kIyGYfNtUAc4GKu/8JD8TPFXw0bx3pGu6foMEVo9xHpaWKz/aFjB3s0r5KklWwAOBjJ705e7e/TcUfetbrse21X1D/kGXX/XF/wCRrzvw78XYZPgzYeNPEFnPJNLKLaS20+MO8s3mFBsVmHXrjNbeh+OP+EstdSh/4RjxJovkWzP5ms6f9nWTIIwp3HJptNNrsK+lz52/ZY/5Kvef9giX/wBGxV9dV8i/ssf8lXvP+wRL/wCjYq+uqQwooooAKKKKAMHx3/yTrxH/ANgq6/8ARTV88/sm/wDI1eIP+vKP/wBDr6G8d/8AJOvEf/YKuv8A0U1fIfwt8Qan4W8LeNtY0K5+y39vZW3lTeWr7c3CqeGBB4J6ilKXKrscVzOx9tUV4Lc6z8WP+FVp4/8A+EpsbaOOzjuf7JGnRsJo8DLtIRkMw+baoA5wMVd/4SH4meKvho3jvSNd0/QYIrR7iPS0sVn+0LGDvZpXyVJKtgAcDGT3py929+m4o+9a3XY9tqtqX/IKu/8Ari//AKCa4Dwt8WrK4+Dtj4y8VYtnkYwSRW6FjNMHKhY16ktjOO3PYVp6H8RNK8YQ6pp8Fnqel6la2plksdVtTBN5ZBAcDJyufem002uwk9Lnz3+yx/yVa+/7BEv/AKNhr65r5G/ZY/5Ktff9giX/ANGw19c0hhRRRQAV8heLLu4X9raKJZ5RH/b9iNgc4xui4xX17XnOofBDwxqXxEXxncXOpjUlvIrwIkyCLfGVKjGzOPlGeaAPRqKKKACiiigD498d/wDJ2y/9huw/lDX2FXHX/wAJ/Bep+Lv+EnvdHMmsefHcfaftUw/eJjadofbxtHGMcV5n4M8S/E34geI/E+laf4nt9LtdL1B1F9Jp8U0iLuZUiVMAEYUks2TwKXW39f1qPpc99orwrwb4w+JXivXdc8FPq2nWd/os7C4177EJGKhtqqsPCEkgnJxgds1peBfiD4otPG3irwl4zf8At250O2N3DcWFqEmuFG07BGuAWIdcD1yMmmrP5q/yE9Pvsex1478QPj2fA3xAPhn/AIRv7dxEftP27y/vgH7vlnpn1rUm+OFlazW0eoeCPGtiLqdLeKS60lY1aRjgLkydfYc15Z8afAPivXfjUdT0fQb28sdtsPtEUWV+UDPPtT6XDrY+oaKKKQBRRRQB8kftQXVxD8VrZYZ5Y1/sqI4RyB/rJK+sbP8A48YP+ua/yrgfHfwU8M/EPxAmsa5c6lFcpbrbhbWZFXapYjgoTn5j3r0KNBFEka5wihRn2oAdRRRQAV8/ftZ/8i14d/6+5f8A0AV9A1z3i3wJ4d8dWttb+KdP+3RWrl4l8+SPaxGCfkYZ/GgDn/gT/wAkR8Of9cZf/Rz16DXhfjTWPEHw/wDHnhDwN8NnhtNOurdkjsrhRLGHZ3+dnYGTCk7sBu2Kj8R+LPiX8NfGWiabqWt2PiuPXswwJLZJaeVNkKOU5wC6nnORnp1oj71rdW180EtG12V/ke8UV4f4s8UfEP4Y+JNAute8R2ev6Vq12Lee1TTkt/IyRkIwJY8E4JPbkV2OufFiPQtSvLWbwR4zuo7NmV7y10nfA4HVlfeMr70aWv8A1p/w4a3t/X9aFn4q/EX/AIVl4Vt9ZGl/2n596tr5P2jydu5Hbdu2t/cxjHerfw28bf8ACwvBUHiD+z/7P86WSPyPO83G1sZ3bV6/SvOfjEb74q/BHRNR8IaNqNx9q1JLhbZ4QJljVJk3FQSME4wcnhhXVfAfRNT8PfCez0/W7Kaxu0uJmaGZdrAFyQcU2mnZhuej0UUUgCvnD9rO4mgfwn5E0ke4XmdjEZ/1PpX0fXFfEL4V6F8Szp58QT30X9n+Z5X2SVUzv25zuVs/cHp3oAn+FLs/wk8MM7FmOnREknJPy111Z3h/RLbw34dsdGsGke2sYVgiaVgXKqMDJAAz+FaNABRRRQBz3xB/5Jn4n/7BF3/6Javn/wDZM/5GPxF/16Rf+hmvpnUbC21XS7rT7+PzbW7heCaPcV3IylWGRyMgnkV5R4y8LaP8HPhrr+v/AA5tDo+plIUM/mvcceaoxtlLL0Y9qUpcquxxXM7Hr9FeC3Os/Fj/AIVWnj//AISmxto47OO5/skadGwmjwMu0hGQzD5tqgDnAxV3/hIfiZ4q+GjeO9I13T9BgitHuI9LSxWf7QsYO9mlfJUkq2ABwMZPenL3b36bij71rddj22oby4+yWM9xt3+TG0m3OM4GcZrzLw18Ybi9+HOh67feF9c1a6vxLHMug2H2hY3jbaSw3DaG6gc9/StbQfiPa+OZNb0Wz0DXtMvbOzLyx6pZrCRvBCrgMTk9QCBkU7O9hX0uYPwp+OZ+Jviq50Y+Hv7M8iya6877b527a6Ltx5a/385z2r1uvmj9nHwL4o8MfEW/vfEGh3mn20mlSRLLPHtUuZYiF+uFJ/CvpekMKKKKACvkL4Y3dw/7VSRPPK0f9pagNhckf6ubtX17XnOhfBDwx4e8fjxfY3OptqInmn2SzIYt0qsG4CA4w5xz6UAejUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAHy7+1n/yM3h7/AK85f/QxX0F4A/5Jr4Z/7BFr/wCiVqt4t+G3hTx1dW9x4p0s30tqhSJvtEse1SckfIwz+NeefELxP4o8IfEbwh4O8BTwwWV1ZC3jtLiNXjGMojM5BfCgA8Hnb70uqj3H0b7HtlFeD+I/FnxL+GvjLRNN1LW7HxXHr2YYElsktPKmyFHKc4BdTznIz061Y8WeKPiH8MfEmgXWveI7PX9K1a7FvPappyW/kZIyEYEseCcEntyKpa2t1dvmJ6X8tfke4V5x8XPiz/wqyLSXGi/2r/aLSjH2ryfL2bP9hs53+3SrGufFiPQtSvLWbwR4zuo7NmV7y10nfA4HVlfeMr715l8bbXVvix4O8Ha54N0HU7m3k+1SGJ4AJI1JRVLAEgZ2Ejk5FJaq6B6Ox7j4O8Rf8JZ4N0zXvsv2T7fAJvI8zf5ee27Az+Qrbrk/hdp15pPwt8P2Gp28ltd29oqSwyDDI2TwRXWUAFFFFAHjX7UE0kPwptWhkeNv7ViGUYg/6uWrv7Nssk3wdt3mdpG+2T/Mxyeors/HfgTSviH4fj0fXJbqK2juFuA1q6q25QwHJUjGGPapfBPgvTfAXhtNE0WS5ktUleUNcuGfLHJ5AA/SgDoKKKKACiiigAooooAKKKKACiiigAooooAx/F//ACI+u/8AYOuP/RbV8Y/D3/kTvH//AGBV/wDSiOvs7xf/AMiPrv8A2Drj/wBFtXzb+yh/yPGt/wDYOH/oxamUeaLRUXyyTPT9J/5NL/7l2X/0BqX4Qf8AJtFv/wBed5/6HJXrVFVU9/n/AL3/AAf8yYe6oL+U8K+BWiaVrfwFkj1rTLPUUhvLmSNbu3SUI20fMAwOD711vwJ/5IPoX/XO5/8ASiWvRpI1mieOQZR1KsM9QazdG0DTfC/huPR9CtvsthbI/lQ+Yz7dxLHliSeWJ5NNu7b9PwuC0jb1/E+VP2Xv+SuS/wDYMm/9Djr6/r5A/Ze/5K5L/wBgyb/0OOvr+kAUUUUAFFFFAGN4x/5EXXv+wbcf+imr4n8H/wDIi+Ov+wdb/wDpXFX2x4x/5EXXv+wbcf8Aopq+b/2T/wDkdtc/7Bw/9GLUyjzKxUXyu56dqf8AyaUP+xbi/wDRa0vgj/k1NP8AsCXf8pK9aoqqnv8AP/e/4P8AmTT9zl/u/wDA/wAjwD4ZSeG7L4A+HdV8VQy3H9n6pJNYQQFjJLdeYwRUQEb2OTgHjvXoHhnw/rF54q1Lxx4nt49Ou7uwFjaabG4dre3Db/3jDguWwcDgdOa6DxZ4H8PeObOC18U6eb6C3kMkSefJFtYjGfkYZ49azfDnwu8H+CLq61DwvpH2G6mtmgeT7TNJlCQxGHcjqo568VXNeTb/AK0t9/6CSsrL+tbnzl+y7/yVuf8A7Bc3/ocdfX1fIP7Lv/JW5/8AsFzf+hx19fVIwooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAMTxp/yIWv/wDYMuf/AEU1fFXg7/kQvHf/AGD7b/0qjr7V8af8iFr/AP2DLn/0U1fOX7J//I5a7/14L/6MFTKPMrFRfK7npup/8mlD/sW4v/Ra0vgj/k1NP+wJd/ykr1qiqqe/z/3v+D/mTT9zl/u/8D/I8H+D1z4Yh+EnhGPxNJFFdyalcf2VJNGzKtwHbByBtB543EZrQ8GwahoXxm8Q6V4g1WLxHqd5o63J1VI/KeFFbb5LRqSq9d3Ht616p4h8NaP4r0ltN8Q6fFf2jMG8uUHhh0IIwQfcEGs3QfAnhnwTpd5H4X0mGwE0Z811Znd8A4BdiWI9s4qua8+Z/wBaW/rutCbWjb+t7/19580/st/8lZuf+wVN/wCjIq+vK+Q/2W/+Ss3P/YKm/wDRkVfXlSUFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQBh+N/8Akn3iH/sF3P8A6Kavi3wd/wAk+8ef9eFt/wClUdfaXjf/AJJ94h/7Bdz/AOimr51/ZO/5G7Xv+vFP/RlTKPMrFRfK7npep/8AJpQ/7FuL/wBFrS+CP+TU0/7Al3/KSvWqKqp7/P8A3v8Ag/5k0/c5f7v/AAP8j5p8H6tqVr+z/wCHdLs9G1rUbDUr25TUn0a0aeZLdZCWQAYCl87ckjA3Yya6b4Sa/DdeKPHVla+HtX0yCaXzo1urPyktVSJUEL8/I+OQvoK9wrGj8P6Zo661e6dbeTcamTPdv5jN5rhMA4JIHHYYFVze833X+X+X9XZNvdS8/wDP/M+Xf2Wf+SsXf/YJl/8ARsVfXdfIn7LP/JWLv/sEy/8Ao2KvrupKCiiigAooooAwvHP/ACTzxF/2Crn/ANFNXxf4N/5J549/68bX/wBKkr7Q8c/8k88Rf9gq5/8ARTV87/snf8jZr/8A14p/6MqZR5lYqL5Xc9K1P/k0of8AYtxf+i1pfBH/ACamn/YEu/5SV61RVVPf5/73/B/zJp+5y/3f+B/kfNfguw07WPgb4L03U/EzeHjJrryW8q2pkM0qO+1A33YzzkFsjI6Gu38NSap4d+Jmt+D28Sah4i0z+xTfeZqMwmmtJd23YXA7g5xx249fStf8OaR4p0l9M8QWEV9ZuQxilHQjoQRyD7gg1l6N4I8OeCdCv7fwvpUNgk0TGUqzO74U4yzEsQPTNVzXk3/W1v8Agk29239b3/4B82fssf8AJV7z/sES/wDo2KvrqvkX9lj/AJKvef8AYIl/9GxV9dVJQUUUUAFFFFAGD47/AOSdeI/+wVdf+imr4x8G/wDJOvHv/Xja/wDpUlfZ3jv/AJJ14j/7BV1/6Kavnn9k3/kavEH/AF5R/wDodTKPMrFRfK7npOp/8mlD/sW4v/Ra0vgj/k1NP+wJd/ykr1qiqqe/z/3v+D/mTT9zl/u/8D/I+UtCW+/4VT8NHsZreADxBOomuxmGKUs3luw74OSB3xXrHh7X/E1j451rwX4o1q18QhNIN/FfwWq27xc7TG6LwOuR3x9a9E13QNL8TaRLpmvWMV9ZS43xSjjI6EEcg+45rH0fwJ4Z8E6JqMfhfSYbATwsZWVmd3wpwCzktj2ziq5tW3/WiX6X/Am2ll/Wrf62Pm/9lj/kq19/2CJf/RsNfXNfI37LH/JVr7/sES/+jYa+uakoKKKKACuDvvjJ4S07x0vhG5muhqrXUdoFW3JTzJCoX5vT5hzXeV8e+Lf+TvIv+xgsf/QoqAPsKiiigAooooA8u1j466No3xMHgufSr6S7N3Da/aEKeXul24PXOBvH5Vgfs9/8jL8RP+wv/wCzy157428P6zP+1Ot9BpN9Jaf2xYv9oS2cx7QIsndjGBg8+1fWVC0d/K34p/oD1VvO/wCf+Z4f8Iv+S8fEv/r5H/oxqPB//J3HjL/sHr/KCvcKKI+7y+SaCXvc3m7nmuiS/wDCbfGPV9RuP3mm+EyLGxjP3TdMMzS/UD5R9a9KrL0Pw3pPhtb1dFtfswvrp7u4/eO/mSvjc3zE4zgcDj2r5w+NXjrxTonxsOm6Rr+oWdlttj9nhnZU+YDPA9aF8KX9X6/iD3b/AKt0PqOiiigAooooA4Xxn8YfCngPXE0nxDNdJdPAs4EVuXG0kgcj3U13EbiSNXX7rAEfSvkX9qT/AJKzbf8AYKh/9GS19bWf/HjB/wBc1/lQBNRRRQAVxPxK+J2nfDLT7G71Sxurxb2Vo0W3K5UgZydxFdtXhP7Umlajqvh3QE0uwur1kupC628LSFRsHJ2g4oAytV8T2/jP43fDDxBZQS28F7A7JFNjcuJJF5xx/DW38cP+SqfDH/sJ/wDtWGu1+CdpcWPwa8P217BLbzxxSB4pkKMv71zyDyK7uiPu8vk2wl7zk+6seH/tK/8AHr4P/wCwwP5Cup+Jt3JrWu+HvANtIyLrkzTaiyHBFnENzrntvPy/nXo9ZcnhvSZfFMPiOS03atDbG1juPMf5YicldudvU9cZ96Fpo9r3/BfqvuG+63tb8X/n95owQxW1vHBbxrFFEoREQYCqBgAD0p9eSftH+INX8N/DixvNA1G5065fVY4mltpCjFDFKSuR2yoP4Vp/AXWtS1/4TWd/rV9PfXb3EytNO5ZiA5AGTQLY9IooooAK5Lxz8TPDvw7NiPEstxH9v8zyfJhMmdm3dnHT74rra+bP2uPv+Evpef8AtCgD6E0PWLTxDoVnq+mlmtL2FZoS67SVIyMjtV+uQ+E3/JIfC/8A2DYf/Qa6+gAooooAo63qkeh+H9Q1aeNpIrC1kuXRMbmVELEDPfivAvGvxh0r4m/B7xZa6Xp15ZtZRW0jm4K4YNOg42k+le2+O4pLj4c+JIYI2llk0q6VERSWZjCwAAHU14V+y3ouq6V4g199U0y8slktYgjXFu0YY7zwNwGamUeZWKi+V3O41P8A5NKH/Ytxf+i1pfBH/Jqaf9gS7/lJXrVFVU9/n/vf8H/Mmn7nL/d/4H+R4h8JvEkfhH9mBtdmAb7ELl0Q/wAb+YQi/ixAr0H4aaBJofgu2mv2Muran/p+ozt96SaQbjn2UEKB6CtvxD4d0vxVosuk69a/arGVlZ4vMdNxUhhypB6gd6sakxt9Fu2gJQx27lCp+7hTiqcrtvvb7v8Ag/oiUrJL1/r5a/eW6K+Zv2b/ABt4n8SfEa/s9f12+1G2TSpJViuZy6hxNEA2D3wxH419M1JQUUUUAFcHo3xk8Ja942HhTT5rptUM0sIV7cqm6MMW+b6Ia7yvj34X/wDJ1yf9hLUf/Rc9AH2FRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAeefEr4xaV8MtQsbTVNOvLxr2JpEa3K4UA4wdxFcL4s1OPW/j98L9VgRo4r6xS5RH+8ocMwBx35rI/aj0XVdV8RaA+l6ZeXqpaSB2t7dpAp3jg7QcV7v4Gikg+HfhyGeNo5Y9KtVdHXDKREoIIPQ0LScZdncHrGUe6seWfHD/kqnwx/7Cf/ALVho/aV/wCPXwf/ANhgfyFe4UUR0SXZ3/FP9Aerb7q35/5nnHxNu5Na13w94BtpGRdcmabUWQ4Is4hudc9t5+X869EghitreOC3jWKKJQiIgwFUDAAHpWdJ4b0mXxTD4jktN2rQ2xtY7jzH+WInJXbnb1PXGfevHP2mvFWveGLfw2fD2r3emm4e5ExtpSm/aIsZx1xk/nQtI2/r+rfjcHq/6/r/AIFj3eiuS+FmoXeq/Cvw/falcSXV1PZq8s0rbmc5PJPeutoAKKKKAOf8Z+NdH8BaGmreIXmS1edYAYYy53kEjgeympPCHi/SvG+gLrOgvK9o8jRgyxlGyvXivM/2pf8Akk9r/wBhaH/0XLV39mn/AJI1b/8AX7P/AOhCgD1uiiigAooooAKKKKACiiigAooooAKKKKAMfxf/AMiPrv8A2Drj/wBFtXzb+yh/yPGt/wDYOH/oxa+kvF//ACI+u/8AYOuP/RbV82/sof8AI8a3/wBg4f8AoxaAPqqiiigApk3/AB7yf7p/lT6ZN/x7yf7p/lQB8h/svf8AJXJf+wZN/wChx19f18gfsvf8lcl/7Bk3/ocdfX9ABRRRQAUUUUAY3jH/AJEXXv8AsG3H/opq+b/2T/8Akdtc/wCwcP8A0YtfSHjH/kRde/7Btx/6Kavm/wDZP/5HbXP+wcP/AEYtAH1TRRRQAVHcf8esv+4f5VJUdx/x6y/7h/lQB8i/su/8lbn/AOwXN/6HHX19XyD+y7/yVuf/ALBc3/ocdfX1ABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAYnjT/kQtf/7Blz/6KavnL9k//kctd/68F/8ARgr6N8af8iFr/wD2DLn/ANFNXzl+yf8A8jlrv/Xgv/owUAfVFFFFABUN3/x4z/8AXNv5VNUN3/x4z/8AXNv5UAfJP7Lf/JWbn/sFTf8AoyKvryvkP9lv/krNz/2Cpv8A0ZFX15QAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGH43/5J94h/wCwXc/+imr51/ZO/wCRu17/AK8U/wDRlfRXjf8A5J94h/7Bdz/6KavnX9k7/kbte/68U/8ARlAH1NRRRQAVBff8g65/65N/I1PUF9/yDrn/AK5N/I0AfJf7LP8AyVi7/wCwTL/6Nir67r5E/ZZ/5Kxd/wDYJl/9GxV9d0AFFFFABRRRQBheOf8AknniL/sFXP8A6Kavnf8AZO/5GzX/APrxT/0ZX0R45/5J54i/7BVz/wCimr53/ZO/5GzX/wDrxT/0ZQB9S0UUUAFV9Q/5Bl1/1xf+RqxVfUP+QZdf9cX/AJGgD5N/ZY/5Kvef9giX/wBGxV9dV8i/ssf8lXvP+wRL/wCjYq+uqACiiigAooooAwfHf/JOvEf/AGCrr/0U1fPP7Jv/ACNXiD/ryj/9Dr6G8d/8k68R/wDYKuv/AEU1fPP7Jv8AyNXiD/ryj/8AQ6APqSiiigAqtqX/ACCrv/ri/wD6Cas1W1L/AJBV3/1xf/0E0AfJ/wCyx/yVa+/7BEv/AKNhr65r5G/ZY/5Ktff9giX/ANGw19c0AFFFFABUZt4TJ5hhjL5zuKjOfrUleA678bfFGm/HqPwbbw6cdMbVbazLPCxl2SFA3zbsZ+Y44oA9+ooooAKKKKACivm3xZ8VvGmmftFDwzZa0YtH/tS0t/s32WE/u3Ee4bim7ncec55r6SoAKKKKACvIPH3wETxz4/PiZvETWJIiH2YWXmfcAH3t4649K9frnNX+IPhPQdX/ALL1jXrOzvsKfIlfDfN0496AOjooooAKKKKAI3t4ZW3SRRuemWUGpK8F+NXxp8T/AA98dQ6PoUOnPbPYx3BNzCztuZ3B5DDjCivdrdzLbRSN1ZAxx7igCSiiigAoorx79ofxz4i8D6Ho1x4W1H7BLc3MiSt5Mcm5QoIHzqcfhQB7DRXHfCbXdR8S/CzRdX1u5+1X11G7TTbFTcRI6jhQAOAOgrsaACiiigDivin8O1+Jvha30ZtTOmiC8W684Qebu2o67cbl/v5zntVv4ceCR8PvBcHh9b86gIZZJPPMPlZ3NnG3cen1rX1/xJo/hewS98Q6jBp9tJKIllnbClyCQv1wpP4VJouu6Z4i0xNR0O9ivrN2KrNCcqSDgj86AL9FFFABTJIYpcebGj46blBxT68g+PHxS174atoQ8PxWUn9oC4877VEz42eXjGGGPvn9KAPXlUIoVQFA6ADpS1geBtauvEfgLRdZ1ARrc31nHPKIlIUMwycAk8fjW/QAUUUUAFFY3jG/udL8C69qFhJ5V1aabcTwybQdjrEzKcHIOCB1rx/9nj4j+K/HGta1B4p1U38VrbxvCv2eKPaSxBPyKM/jQB7zRRRQAVDeW/2uxnt92zzo2j3YzjIxmpqbJIkMTySsFRFLMx7AdTQB5R8K/gYnwy8U3Osr4gbUjPZta+SbPytu50bdne39zGMd69ZrntA8eeFvFN+9l4e1y01C5jiMrRQPkhAQC30ywH410NABRRRQAVGtvCsm9YYw/wDeCjNSV4D4L+NvijxB8cV8I30OnLppvLuDdFCwk2xJIV5LEZygzx60Ae/UUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRXgv7Q/wASPFfgfXNGt/C2qmxiubaR5V+zxSbmDAA/Opx+FexeEL651TwPoV/fSebdXWnW800m0De7RqzHA4GST0oA2KKKKACvOvi18Jl+KUWlI2snS/7OaU5Ft53mb9n+2uMbPfrXotY3iHxf4f8ACi258R6tbacLksITO2N+3GcfTI/OgBPB/h4eE/B+maCLn7WLCAQ+eY9m/Hfbk4/Otqq2m6lZ6vpsGoaZcJc2lwm+KaM5V19RVmgAooooAa8aSrtlRXGc4YZoSNI12xoqL6KMCvPvjX461X4e+BYNY0JLZ7mS+jtyLlC67WRyeARzlRVj4O+M9T8efD2LW9aS3S6e4liIt0KphTgcEn+dAHd0UUUAFFFFABRRRQAUUUUAFFFFABRRRQBj+L/+RH13/sHXH/otq+bf2UP+R41v/sHD/wBGLX0l4v8A+RH13/sHXH/otq+bf2UP+R41v/sHD/0YtAH1VRRRQAUyb/j3k/3T/Kn0yb/j3k/3T/KgD5D/AGXv+SuS/wDYMm/9Djr6/r5A/Ze/5K5L/wBgyb/0OOvr+gAooooAKKKKAMbxj/yIuvf9g24/9FNXzf8Asn/8jtrn/YOH/oxa+kPGP/Ii69/2Dbj/ANFNXzf+yf8A8jtrn/YOH/oxaAPqmiiigAqO4/49Zf8AcP8AKpKjuP8Aj1l/3D/KgD5F/Zd/5K3P/wBgub/0OOvr6vkH9l3/AJK3P/2C5v8A0OOvr6gAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAMTxp/yIWv/APYMuf8A0U1fOX7J/wDyOWu/9eC/+jBX0b40/wCRC1//ALBlz/6KavnL9k//AJHLXf8ArwX/ANGCgD6oooooAKhu/wDjxn/65t/Kpqhu/wDjxn/65t/KgD5J/Zb/AOSs3P8A2Cpv/RkVfXlfIf7Lf/JWbn/sFTf+jIq+vKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAw/G/8AyT7xD/2C7n/0U1fOv7J3/I3a9/14p/6Mr6K8b/8AJPvEP/YLuf8A0U1fOv7J3/I3a9/14p/6MoA+pqKKKACoL7/kHXP/AFyb+RqeoL7/AJB1z/1yb+RoA+S/2Wf+SsXf/YJl/wDRsVfXdfIn7LP/ACVi7/7BMv8A6Nir67oAKKKKACiiigDC8c/8k88Rf9gq5/8ARTV87/snf8jZr/8A14p/6Mr6I8c/8k88Rf8AYKuf/RTV87/snf8AI2a//wBeKf8AoygD6looooAKr6h/yDLr/ri/8jViq+of8gy6/wCuL/yNAHyb+yx/yVe8/wCwRL/6Nir66r5F/ZY/5Kvef9giX/0bFX11QAUUUUAFFFFAGD47/wCSdeI/+wVdf+imr55/ZN/5GrxB/wBeUf8A6HX0N47/AOSdeI/+wVdf+imr55/ZN/5GrxB/15R/+h0AfUlFFFABVbUv+QVd/wDXF/8A0E1ZqtqX/IKu/wDri/8A6CaAPk/9lj/kq19/2CJf/RsNfXNfI37LH/JVr7/sES/+jYa+uaACiiigAr561/4NeLNR/aAj8X20dodKXVrW7LNcAP5cZjLfLjr8p4r6FooAKKKKACiiigDw/wARfArV9Z+NQ8aQ6tZR2gv7a6+zur+Zti2ZHTGTsP517hWbL4j0S31H+z59Z0+K93hPsz3SCTccYG0nOTkYHvWlQAUUUUAFfHvx+/5OAP8AuWn8hX2FWFqfgfwvrWp/2jq3h/Tr29+X/SJ7dXf5enJGeKAN2iiigAooooA+f/jh8HfFfj3x7Dq3h+O0a1SwjgJmuNjbg7k8Y9GFe+W6GO1iRvvKgB/AVJRQAUUUUAFeafGf4Y3/AMTdJ0y002/trJrKd5Ga4DEMCoGBgV6XVPUdX03SI0fVtQtbFJDhGuZ1jDH0G4jNAGN8O/DFx4M+H+l+H7yeO4mskdXliBCtl2bjPP8AFXS1DaXltf2qXNjcRXNvIMpLC4dG5xwRweamoAKKKKAPE/2qv+SV6d/2GYv/AETNWt+zd/yRWw/6+bj/ANGGvRta0DSfEdklpr2m22o2ySCVYrmIOocAgNg98Ej8afpOjaboOnrY6LYwWNohLLBbxhEBJyTgUAXaKKKACvG/j98MvEPxFbQD4bS2f7ALjzvPm2Y3+Xtxxz9w17JRQBz/AIC0a78PfD/RNI1EILqys44ZQjbl3AYOD3roKKKACiiigDM8S6XJrnhPV9JgkWKS/sZrZHf7ql0KgnHbmvNfgx8HtT+GWqapdalqVperewpGq26sCpViecj3r1qaaK3gkmuJEiijUu8jsFVFAySSegA71T07XdI1d3TSdUsr5oxl1trhJCo9TtJxQBfooooAKp6x/wAgO/8A+vaT/wBBNXKbJGssbRyKGRwVZSOCD2oA+S/2VP8Akqepf9gaX/0fDX1tWJovgvwz4cvHu9B0HT9OuHjMTS21usbFCQSuQOmQD+ArboAKKKKACvnrwR8GvFmg/HdfFeoR2g0wXl3MSlwGfbIkoX5cerivoWigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAPIvjP8HdT+JuraZd6bqdpZLZQPEy3CsSxLA5GBXpfhvTJNE8KaTpU0iySWNlDbO6dGKIFJHtxT9R13SNIkRNW1SysXkGUW5uEjLD1G4jNXIZo7iCOa3kSWKRQ6SIwZWUjIII6gigB9FFFABXzl+1v/wAevhT/AH7v+UVfRtZWueF9C8SrAPEOkWepC3LGIXUKyeXuxnGemcD8qAOf+Dv/ACRzwz/14r/M12tV7CwtNLsIbLTraK1tYF2RQxKFVB6ADpVigAooooA84+OHgnV/HvgODSfD6wtdJfxzkTSbF2hHB59csKs/Bjwhqvgf4dRaNrqxLdpcyyEQyb1wx45rvqKACiiigAooooAKKKKACiiigAooooAKKKKAMfxf/wAiPrv/AGDrj/0W1fIfwM+IejfDrxNqV94gS6aG6s/JT7NGHIbeDyCR2FfZl/ZRalptzY3O7ybmF4ZNpwdrAg4P0NeU/wDDMvw+/wCeep/+Bf8A9agCH/hqDwF/zy1j/wABV/8Ai6P+GoPAX/PLWP8AwFX/AOLqb/hmX4ff889T/wDAv/61Zut/AX4V+Hba3n1ibULaO5uo7SJmvPvSSNhR0+pJ7AE0AXP+GoPAX/PLWP8AwFX/AOLpsv7T/gJoXUQ6wSVIH+ip/wDF1Y/4Zl+H3/PPU/8AwL/+tR/wzL8Pv+eep/8AgX/9agDx79l7/krkv/YMm/8AQ46+v64HwX8GvCngLXm1fQFvBdNC0B8+feu1iCeMf7IrvqACiiigAooooAxvGP8AyIuvf9g24/8ARTV8i/Az4iaN8OvEmpX3iBLpobq0EKfZow5DbweQSOwr7L1Cyh1LTbqxud3k3ULwybTg7WBBwfoa8p/4Zl+H3/PPU/8AwL/+tQBD/wANQeAv+eWsf+Aq/wDxdH/DUHgL/nlrH/gKv/xdTf8ADMvw+/556n/4F/8A1qzdb+Avwr8O29tPrE2oW0d1dR2kRa8+9JIcKPu/Uk9gCaALn/DUHgL/AJ5ax/4Cr/8AF0yb9p/wG8EirDrBJUgf6Knp/v1Z/wCGZfh9/wA89T/8C/8A61H/AAzL8Pv+eep/+Bf/ANagDx/9l3/krc//AGC5v/Q46+vq4LwX8GvCvgLXW1fQFvBdNC0B8+feu1iCeMdflFd7QAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGJ40/wCRC1//ALBlz/6Kavkj4F/EXRfh14i1O98QJdNDdWoij+zRhyGDg85I7V9lajYw6ppd1YXW7yLqF4ZNpwdrKVOD9DXlP/DMvw+/556n/wCBf/1qAIf+GoPAX/PLWP8AwFX/AOLo/wCGoPAX/PLWP/AVf/i6m/4Zl+H3/PPU/wDwL/8ArVm638BfhX4dgtptYm1C2S6uo7SEtefelkOFHT6nPYA0AXP+GoPAX/PLWP8AwFX/AOLqO5/ae8ByWsqJDrBZkIA+yp1x/v1a/wCGZfh9/wA89T/8C/8A61H/AAzL8Pv+eep/+Bf/ANagDyH9lv8A5Kzc/wDYKm/9GRV9eVwXgr4N+FfAOuvq2gLeC6eBoD58+9dpIJ4x1yorvaACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAw/G/8AyT7xD/2C7n/0U1fJXwL+I2i/DnxBql34gS7aK7tVij+zRhyGDA85I7V9kalYQ6rpV3p93u8i7heCXacHaylTg+uDXlX/AAzL8Pv+eep/+Bf/ANagCH/hqDwF/wA8tY/8BV/+Lo/4ag8Bf88tY/8AAVf/AIupv+GZfh9/zz1P/wAC/wD61ZutfAX4V+HobWXV5tQtku7qO0hLXn35XOFX7v1P0BoAuf8ADUHgL/nlrH/gKv8A8XUV3+074Elsp40h1gs0bKB9lTkkf79W/wDhmX4ff889T/8AAv8A+tR/wzL8Pv8Annqf/gX/APWoA8i/ZZ/5Kxd/9gmX/wBGxV9d1wfgr4OeFfAGuSat4fW8F1JA1u3nz712EqTxjrlRXeUAFFFFABRRRQBheOf+SeeIv+wVc/8Aopq+S/gX8RtF+HOvapd+IEu2iu7ZYo/s0Ychg2eckV9kalYQ6rpN3p13u8i7geCXacHa6lTg+uDXlX/DMvw+/wCeep/+Bf8A9agCH/hqDwF/zy1j/wABV/8Ai6P+GoPAX/PLWP8AwFX/AOLqb/hmX4ff889T/wDAv/61ZutfAX4V+HorSTV5tQtlvLqOzgLXn35XOFX7v1P0FAFz/hqDwF/zy1j/AMBV/wDi6hvP2nfAktjPHHDrBZ42VR9lTkkf79XP+GZfh9/zz1P/AMC//rUf8My/D7/nnqf/AIF//WoA8j/ZY/5Kvef9giX/ANGxV9dVwfgn4O+FvAGuSat4fW8F1JA1u3nz712FlY8Y65UV3lABRRRQAUUUUAYPjv8A5J14j/7BV1/6Kavk34FfEfRfhzruqXXiBLtoru2WOM20Ychg2eckV9janp8GraTd6dd7vs95A8Eu04O11KnB7HBryv8A4Zl+H3/PPU//AAL/APrUAQ/8NQeAv+eWsf8AgKv/AMXR/wANQeAv+eWsf+Aq/wDxdTf8My/D7/nnqf8A4F//AFqzda+Avwr8PR2j6vNqFst5dR2cBa8+/K5wq/d+v5UAXP8AhqDwF/zy1j/wFX/4uoL39pzwJNp9xHHBrBd4mVR9lTkkf79Xf+GZfh9/zz1P/wAC/wD61H/DMvw+/wCeep/+Bf8A9agDyT9lj/kq19/2CJf/AEbDX1zXB+Cfg74W8Aa5Jq3h9bwXMlu1u3nz712FlY8Y65UV3lABRRRQAV8f+LZZB+1zEodtv/CQWIxnj70VfYFeeah8FPC2pfEBfGNxJqA1NbuK7AWdRHvjKlfl25x8o70Aeh0UUUAFFFFAHx945Vn/AGuFCgsf7asDgD0ENfYNcRL8JfDM3xFHjZ1u/wC1xKs2RP8Au9yoEHy49BXb0AeU63+0V4K0DXr7SL6LVDc2Nw9vKY7ZSu5GIODv5GRXT+AfiZoXxIhvpPDyXarYsiy/aYgnL5xjBOfumub139n7wJrOsahrWqC+Wa7mkubhxd7UDMSzHpwOTWt8J/CXhDw/4dfU/Ac1xPYawEkMk0u/OzcMdBgglgR6igDqvEev2fhbw5e63qYkNpZR+ZKIl3NjOOBketeH3/7VVkdfS00Lw5Jd2UjIi3Fzc+S5Jxn5ArdCfXmvcfEOg2fifw9eaLqgkNpeR+XL5bbWxnPB7dK82h/Zq8AQXEc0cepb42DLm77g59KAPW6KKKACiiigD5F/ailkT4sWwR2Uf2VFwDj/AJaS19aWf/HjB/1zX+VcJ44+C/hj4g68mr69JqC3KQLbgW06ou1SxHBU85Y138aCKJY16KoUZ9qAHUUUUAFfP37Wf/IteHf+vuX/ANAFfQNcr46+HehfEOztLbxEtwY7SRpIvIl2HJGDng0AZHwKDD4I+HNysp8mQ4YY4858H8etXPH3xV8P/Diaxi8QpeM18rtF9miD8KQDnLDH3hXUaNpVtoWh2Wk2G8WtjAlvDvbLbEUKMnucCuU+I/w58K+NLeLUfGD3EcOkwyv5kU3lqiEBnJ4PQLQBj+G/2gfB3inxHZaJpkepi7vZPLiMtuqrnBPJ3H0r1CvJPAHwd+Hsb6N4y8LG/k+Vbm1aW5yBkYwy46jJBHYivW6APFvHH7SeieGL+70vStKutR1OxvGtbiOdhDGAhYMyuNxPzAYGBwa734aeNm+IPgi3197EWBmlkj8gS+Zja2M7sDr9K5rWf2evA+u65fatfpqBub64e4m2XWF3uxY4GOBk12vhDwlpngjw7FouhiYWcTs6iZ97ZY5PP1oA3KKKKACvm79raR0fwlsZlyLzODj/AJ419I1xvj/4XaB8STp58QveL/Z/meT9llCZ37d2cqc/cFAD/hQS3wj8MEnJOmw5J/3a6+s/QdFtfDnh+x0bTzIbWxhWGIytltqjAyeMmtCgAooooA574g/8kz8T/wDYIu//AES1eAfsmK3/AAkHiJsHb9liGccZ3mvpXVdNg1nRr3TLzcbe9t3t5dhwdjqVOD2ODXPeBPhvoPw7gvYvDi3IW9ZGl8+XfyoIGOBj7xoA2PEniCz8K+G73W9TEptLKPzJREu5sZA4BI9a8v8A+GoPAX/PLWP/AAFX/wCLr1HxFoFl4o8O3mi6oJDaXieXKI22tjIPB7dK8es/gX8KdQ8TajoFrNqL6jpscUlzF9r+6JASvbngDPpuHrQB7Pouq2+u6DY6tYhxbX1vHcRCQYbY6hhkdjg1ynxE+K+hfDU2setw3k015HI8CW0asGKY4JLDHJFdZo+lW2h6HY6TY7hbWNulvDvbLbEUKMnucCuZ8dfC3w58RJ7KbxGt0WsldYvIm2cMQTng56CgDk/hN8cpfib4rudGk0BNNEFk115q3Zl3bXRduNg/v5zntXr1cJ4I+D/hb4fa3Nqvh5bwXM1u1s3nz712FlY8Y65QV3dABRRRQAV8f/DCWQ/tWIpdiv8AaWocE8f6uevsCvPND+Cnhbw/48Hi6xk1A6kJpp8STqY90qsG425xhzjmgD0OiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA+Xf2s/+Rm8Pf8AXnL/AOhivoLwCGHw18Mh0ZGGkWuVYYKnyV4I9ay/HXwp8N/EO8tLnxEt0ZLSNo4vIm2DBOTng12UUaxQpGn3UUKM+goA4Hxz8Z/DHw+16PSNeS/a5kt1uAbaBXXaWYDksOcqai8GfG/wr478RpomiR6gt28bSA3ECquFGTyGNWfGvwc8K+P9cj1bxAt4bqOBbdfIn2LtBYjjHXLGsT4b/Dv4e6H4z1O78IT3cup6LKbK6WW43qjOgJ4xz3GfVWHagD1evBfEv7VGi2EkUfhzRLrUXDOs/wBrkFuExjG3Afdnn0xj3r3qvI5P2afAEsryPHqW52LHF33P4UAeg+DPEJ8WeDNL15rYWpv4BMYQ+/ZntnAz+VbdZ3h/Q7Pw14fs9G0wOLSyjEUXmNubaPU960aACiiigDxn9qN2T4T2pRip/tWLkHH/ACzlq7+zWzP8G7cuSx+2T8k57iu18c+BtJ+IOgppGvNcLbRzrcA20gRtyhgOSDxhjUvgvwbpngTw4mi6I1w1qkjSA3DhmyxyeQBQBv0UUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFfIv7Snjs+IPHCeH7CbNjomVfaeHuD98/8BGF9ju9a+rdZ/tH+xLz+wxCdSMLC1+0MVjEmPlLEA8A89K+MfHXwY8WeDrW31TxDd6fcf2jfrbBoZ3d2lcM25soOPlOTQB9O/BnxyPHfw5s7u4k36lZ/wCi3uTyZFAw/wDwJcH6kjtXfV4V8F/hR44+GviuefUbnS5dJvofLuooLh2YMuSjgFACQcjr0Y17rQAUUUUAFFFFABRRRQAUUUUAFfIn7Sfjs+IPHKaBYTZsdDyrlTw9wfvn/gPC+xDetfV2tf2l/Yd6NCEJ1IwsLX7QxEYkI+UsQDwDz0r4x8c/BjxZ4Ot7bU/EF3p9wdSvltQ0M7uxlcM25soOPlOTQB9PfBrxyPHnw5s7y4k36jaf6LfAnkyKBh/+BLhvqSO1d7XhXwW+FHjj4a+Kp5tSudLm0i+h8u6iguHZgy5KOAUAJByOvRjXutABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAV8h/tI+Oz4h8dpoNhMTY6GSjFTw9wfvn/gOAvsQ3rX1drf9pnQr0aCIf7SMLC1NwxWMSEYUsQDwDz07V8ZeOPgv4r8HQ2moa/d6fOdSvltVaGd3YyuGbLZQccHJoA+n/g545Hjz4c2d9cSb9Rtf9FvhnkyKB8//AAIYb6kjtXeV4X8FvhT43+Gvii4l1K50ubSL6HZcxQXDswdclHAKAEgkjr0Y+le6UAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABXyD+0h48PiLx6uh2ExNjoZMZKnh7g/fP/AcBfYhvWvq/XRqZ0G9XQPI/tNoWW1NwxWNZCMAsQDwDz05xXxl42+C3ivwdHZ3uvXenznU75bVGhnd2Mr5OWyg44OTzQB9RfB7xyPHnw6sr+eQNqFsPst8M8+aoHzf8CGG+pI7V3VeG/BX4VeN/hr4nuZNTudMm0i+h2XEUFw7MHXlHAKAEjJHXox9K9yoAKKKKACiiigAooooAKKKKACvkD9o/x4fEXj5dE0+Ymx0ImMlDw9wfvn/gOAvsQ3rX1hro1Q6BfL4f8j+02hZbU3DFY1kIwGYgHgHnpzivjLxr8FvFfg6OyvNeutPnOp3y2qNDO7sZXyctlBxwcnmgD6i+D/jgePfh1ZahPIG1C3H2W+HfzVA+b/gQw34kdq7qvDfgr8KvG/w18TXMmp3OmTaRfQ7LiKC4dmV15RwCgB6kdeje1e5UAFFFFABRRRQAUUUUAFFFFABXyB+0d48PiPx+ui6fMfsOhEx5Q8NcH77f8BwF9iretfV+vDVG8P3y+H/I/tNoWW1NwxWNZCMBmIB4HXGOcV8Z+Nfgt4r8HLY3Wu3Wnztql8tqjQzu7GV8nLZQccHJ5oA+o/hB44Hj34dWWoTSBtQtx9mvh381QPm/4EMN+JHau5rw74KfCvxv8NfEt0+qXOmTaRfQ7biOC4dmV15RwCgB6kdeje1e40AFFFFABRRRQAVwF/8AGfwpp3j1fCFy17/ajXUVoAsGY/MkKhfmz0+Yc139fHvi3/k7yL/sYLH/ANCioA+wqKKKACiiigAooooA8f8A2jvHX/CMeAf7Fspduoa5uh+U8pAP9Y345C/8CPpXD/st+OvKurzwXfy/LNm7sNx6MB+8QfUAMB7N61L8YPhB478ZeLNU8TTXWlLp9rCRawtcvvSCME4xsxuPLEZ6t1rivhn8H/G+q2ukeNvCt5pcAScy2/2id1bMchUhgEIwSpGM8g0AfZFFNiLtChmUJIVBdVbIB7gHvTqACiiigAooooA4Lxr8ZPCvgHXU0nX2vBdPAs48iDeu0kgc5HOVNd3G4ljV1+6wBGfevkX9qT/krNt/2Cof/RktfW1n/wAeMH/XNf5UATUUUUAFFFFABXhf7Tnjr+x/CsHhWxlxd6t+8udp5S3U9P8AgTDH0Vh3r3SvmT4s/BbxlruseIPGuqatpJtoIZbhIEklLpBEhKoBsxnavqASSe9AFz9lvx1ujvPBd/Lyu670/ce3/LRB+OGA93r6Pr5K+GHwV8Yzx6F458OatpUA8wXEMdxJKrEBirK21CMEAg4PQ19ajO0buD3waACiiigAooooAK4/x58TvD3w5NgPEZuh9v8AM8n7PDv+5t3Z5GPviuwr5s/a4+/4S+l5/wC0KAPoTQtZtfEWg2Wsadv+y3sKzReYu1tpGRkdqv1yHwm/5JD4X/7BsP8A6DXX0AFFFFABRRRQBjeLvEtp4P8ACWo69qB/c2UJcJnBkboqD3ZiB+NfGPgX4lX/AIf+Lcfi3Upml+23Df2lj+OKRvnwP9nhgP8AZAr6Y+NPw/8AE/xH0uw0rQb/AE+zsIZDPci6kcNK4GEHyqeACx+pHpXzxpfwM1/VvHus+ErfUtNS+0eOOSeV2kEThwpG0hcn7w6gUAfacUsc8KTQOskcihkdTkMCMgg0+uT+Gega94W8C2eieJ7q1vLmxzFDNauzAwj7gO5RyPu/QCusoAKKKKACiiigArgNF+M/hTX/AByPCdg17/aZmlhAeDCbowxb5s+iGu/r49+F/wDydcn/AGEtR/8ARc9AH2FRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAHK/ErxnF4D8A6hrblTcInlWkbf8tJm4QY7gfePspr5Q+C3xCl8JfFCK71O5ZrLV38i/kkbu7ZWU+4Y5J9C1e9fGr4X+LPiZe6fBpOpabaaVZIX8q5kkDPM3BYhUIwFAA57t614R4Y+Bmv+K9d8QaVp+o6bFPoN0LW5adpArtlhlMKSR8h6gdqAPtaisPwXYa1pXg7TtP8AE1xb3WpWsIhlnt2ZllC8K2WAOduM++a3KACiiigAooooA53xt430jwDoSatr5nFq86wDyI97biGI4yOMKal8HeMNL8c+Hk1nQjMbR5GjHnR7GyvXjNeaftS/8kntf+wtD/6Llq7+zT/yRq3/AOv2f/0IUAet0UUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFePftI/8id4c/7GK2/9Fy17DXj37SP/ACJ3hz/sYrb/ANFy0Aew0UUUAFFFFABRRRQAUUUUAFFFFABXj37SH/IpeGv+xit//Rctew149+0h/wAil4a/7GK3/wDRctAHsNFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAV49+0b/yLXhf/sYrf/0CSvYa8e/aN/5Frwv/ANjFb/8AoElAHsNFFFABRRRQAUUUUAFFFZ8HiDRrnUBYW+r2Mt4c4to7lGkOOvyg54oA0KK5jxP8R/CXg28S08S61DY3EkQmWJkd2KEkBsKp4ypH4VY8U+M9J8IeFW8Q6q0zWClPmgj3Md5AU449RQBv0Vzfgbx1pfxA0ObVdEiu4reK4a3IuogjFgqtkAE8YYc/WsLwr8T73xP4r/shvBWuabbYkzqF3AyxAr0GcY5xxzQB6DRXBeOPGPjTQdcisfCngaTX7Z7dZWuxdCNVcswKYx1AAPX+KtzxXf8Aia08K/afCGk2+oawSmLS5lCKAfvcllHH1oA6Giuc8D3viy+0KSXx5plnpmpC4ZY4bOTehi2rgk7m5yWHXsKwfCcnxUfxT/xWMWhx6Ltf/jzLebn+DqT+NAHoNFcD43svijda7GfAOqaJZ6b9nUONQRjJ5u5skYRuMbf1re8WWnim78L+T4P1CzsNZ3IftF0m6PH8Yxtbr24oA6Ciuc8EWfiyx0KSLx5qdnqWpG4Zkms49iCLauFxtXnO7t3FYPhLRfifZ+KPtHi/xRpmo6PtcfZbe2VHyfunIjXp9aAPQaK4HxvpHxNvtdjm8CeI9K0zThbqrwXkAdjLubLZ8tuMFR17VveLbXxVc+F/J8HX9lZ6zuT9/eLmLA+9wFbr9KAOgrx79or/AJAXhT/sYrf/ANBeu+8EWvi200KSPx9fWF9qZuGKS2CkRiLauAcqvOd3buK8I+Kq/Elf+EeHjt9GfTf7ch+zmwz5nmfNjOQONu78aAPpqiuB8b6l8UbPXY18BaLouoaZ5Cl2vpCsnm5bIH7xRjG3t61veLtT8RaV4Y+1+GNFj1jVA6A2hlEakH7xySOlAHQUVzvgjWfEGuaC9z4s0H+wr5Z2jFr5wk3IApD5HqSRj2rA8IfEbXvEXib+ytW8AaxoUWx3F7chjF8vQZKKMn60Aeg0VwHjb4w6J4C15dL1jTNYnLQLN9otLZXiAJIwSWBz8vp3FdJ4n8XaP4O0Iax4huGtbIyLGXETOQzdBhQT2oA26Kx/C/izRfGekHVPDd79tsxIYjJ5Tx4YAEjDgHuO1Gk+LvDmu3HkaLr2m6hPgkxW12kjgDqdoOaANiijIyRnkdRRQAV49+0T/wAgbwl/2MVv/wCgvXsNePftE/8AIG8Jf9jFb/8AoL0Aew0UUUAFFFFABRRRQAUUUUAFFFFABXj37RH/ACCfCH/YxW//AKC1ew149+0R/wAgnwh/2MVv/wCgtQB7DRRTXkSJC8jKijqzHAFADqKq3+qWGl6a+oane29nZRgF7meVUjUEgDLE45JAH1FVdD8S6L4lhmm8P6pa6jFA+yR7aQOqtjOMj2oA1KK4zTvi14M1bxivhfTtVafVmlkh8n7NKoDxhiwLMoHG096i8d/FnQfh7qNrY6zbajcT3URljWygEnyg47sOcigDuKTauc4GfXFc/rXi1NK8CHxPb6ZfahGYYZks7ePM7CQqANvqN2T9DXhVv+0J4r1H4sWGiQ6dDp2m3mp29sba9tiLiNHdFYMc8HkkcdxQB9LUV59onxG1/VfHQ0K5+H+r6fYedNGdWn3CIBAxVuUAIYqAOf4h1qz478b+JPC2pW1t4f8AAt74limh8x57ecxiJskbT+7bnHP40AdxRXP63r2rad4FOtad4em1DVPJhkGkrLtfc5UMm7aeVDE9P4e1VvAnibXPE+m3Nx4i8LXHhuaGbZHBPN5hlXAO4HavfigDqaK8+0T4g+JtU8dDQ774e6lp2nmaaP8AtaWYmIKgYq2PLHDbQBz/ABDrVnx38QtR8HalbW1j4M1nX4pofMe40+JnWI5I2nCkZ4z1oA7G9ggutPuLe8x9nliZJctt+QjB57cd6zfCWiaJ4d8M22meFwg0uEuYdkxlHzOWb5iTn5ie9ZfjPXzZ/Cq/1eTTbyR57FR9ihj3TKZgExt9VL5PsDXLfs63wk+FMGltaXlvNplxLFKbmHy95d2kBUZJIw4HOOaAPVqK4LQvjL4S8ReMv+EX06a6Op+bLEEe3IUtGGLfN0xhDWz4j+IPhXwjqENj4k1mHT7iePzY0lVvmXJGcgEdQaAOkoqnLq+nW+krqlzfW8FgyI4uZpQke1sbTuOAM5GPqKkstQstSh87TryC7izjfBKrrn6g0AWKKKKAEKg9QD+FLXgXxt+M3ij4f+O4dI0FbE2z2MdwftEBdtzO4PIYcYUV7zbuZbaKRurIGOPcUASUUUUAFFFFABVPV7exvNEvrbWPL/s+a3kjuvMfYvlFSHy2Rgbc85GKuVl+J9Kk17wjrGkQyLFJqFjNapI4yELxlQT7DNACeGbDRtM8N2dn4X8n+yYkItvImMqbdxJw5JzyT3rVrnPh/wCGZvB3gLS9AuriO5lsY2RpYwQrZdm4z9a6OgAooooAKKKKACkKg9QD9RS1498evih4g+G7aCPDotD9vFx532mIv9zy9uMEY++aAPYenSisDwJrV14i8A6LrGoeX9qvrOOaXy12ruYZOB2Fb9ABRRRQAUUUUAFc/pukeF7bxtq2paZ9l/4SC6jRdQ2XJaTaAoXdHuO3gL2Ga6CuI0HwJdaR8W/Evi6W8hkt9Yhijjt1U74yioCSen8NAHb0UUUAFFFFABRRRQAUm1QcgDP0pa+fvBXxq8Va/wDHRfCV+th/ZpvLuHMcBEm2JJCvzbuuUGeKAPoGiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACue8O6R4X07WdcuPDf2X7feXIk1Xybkyt5uW++u47Dktxgd/SuhriPAvgS68J+K/F+rXN5DcR+IL4XUUcakGIbpDhs9T846elAHb0UUUAFFFFABRRRQAEA9Rn60AAdBivOvjb441b4f8AgSDV9BFubl76O3P2iMuu0o5PAI5yoqz8G/GOp+Ovh5FrWuCAXb3EsZ8hCi4U8cZNAHeUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFePftI/8id4c/wCxitv/AEXLXsNePftI/wDIneHP+xitv/RctAHsNFFFABRRRQAUUUUAFFFFABRRRQAV49+0h/yKXhr/ALGK3/8ARctew149+0h/yKXhr/sYrf8A9Fy0Aew0UUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUVzV98RPCWm+IodCu9etF1SadLdLRGLuJGICqQoO0kkdcdaAOlorkPH3ji88GQ2I07wvqfiG4vWdUisEJ2FcfeIBIzu44PQ1dXUPEOsfD37dp9gmk6/cWheK0vSWWCb+65wMj3x+FAHRV49+0b/yLXhf/sYrf/0CSux8A6d48sVvpPiHrOn6jJMUNrHYxbVgA3bhnYuc5XrnpXkHxt+HlpoGpaZ4nGraneXWpeIo18i5mDQwI+99qLjIwVAHOMdqAPc/EPjTw34T8v8A4SPWrTT2kUtGk0mHcDqQvUj6Ck1vxfpuieCX8Ulbi+05YY50+xx7nlSQqFKqSP7wPOOKfrfg/wAO+Jbq3ufEGjWepS2qssJuohIEBxkYPHYVrW9vDaWsVtaQxwQQoEjiiUKqKBgAAcAAdqAOZ8B+OU8d6fdXkWi6lpUUEoRBqEPltMCM7l7Y7cE1kaJr3xLvvHf2XVfClhYeG0mlR7z7SHmdQG2Mq788kL1XoTXoNFAHC+O/D3j3XNUtv+EO8WwaDpwh2zo1sskjSbj8wJXIGMD7w6Vs+IPDFx4j8E/2Fc6zd2dw8cQk1Cyby5CylSxHPG7BGPeuhooA5nwL4HtvAejz2Fnqeo6iLic3Dy6hKJHDFQvBAGBhR61V0H4U+DPDPiA65o2jiDUyzt9oNxKxBfIb5SxUZye1dhRQBlal4W8P61fJe6xoenahdRoI0murVJWVck4BYHAySfxrTWNFjWNUUIoAVQOAB0wKdRQAUVma54l0Xw1afafEGq2mnRH7rXEoTd7KDyx9hXmt/wDtC6Nc3bWPgXQ9W8V3o6C0t2SP8SQWH12496APXaCcda8aH/C9PFvzqdH8GWjdFYCafb+TjP8A3zTv+FB3WtfN478f69rYPJgjk8mL6bWLjH0xQB6HqXj7wjpDMmp+JtJtpF6xveR7x/wHOf0rl7/4/wDw3sMj/hIPtLj+G2tZXz+O3b+tTaZ8CfhzpaKI/DcNw46vdSySlvqGbH6V1Nh4O8M6Vj+zPD2lWhHQwWUaH8wKAPNz+014JY4trLXLn/rjZqc/m4o/4aN0eX/jz8IeKp/T/QUH8nNewgAAADAHQCigDx7/AIaQ0GH/AI/vC3im2PfdYpgfnIKdH+0z4CZws66tbHv5tmOPyY16/TJYo5oyk0ayIequoINAHB6f8cvhxqWBD4nt4mPa5jkhx+LqB+tdVpninQNaYLo+uadfsei212kh/JSao6h8PfB2q5OoeF9ImY9XNlGG/wC+gM/rXKap+zz8OtSUmHR5dPlPSWzupFI+gYlf0oA9Orx79or/AJAXhT/sYrf/ANBemf8ACm/GHh/5vA3xN1S3RPuWmpr58Y/9lH/fFef/ABbufifpum6HB49t9L1Kzg1aKa2u9O4eaVQ2IyvHUZ6J1FAH1JRXl2h/tBeC9SuPsesyXfh2/U7Xg1SAoFb03DIH/Attek2N/Z6naJdabdwXdu/3JreQSI30I4NAFiiiigAqG6tLa9hMN5bxXEROSkqB1P4GpqKAK1jptjplubfTbO3s4SxYx28SxqSepwAOfesDQfhr4Q8Ma0dW0DQ4LG9MbRebEz/dOMjaTjsO1dRRQB5/4z+DXh7xv4hXW9RvNUtL9YliEllchBtXOBgqfU1veNvD2reJdAWx0DxFP4dulnWX7Xbxl2KgEbMbl4OQevauiooA57wjpGvaF4YNn4g1v+3dRV3ZbuSPy9wP3VIHp+NeCfE+4+Jm3w1bfEG00f7CutwNDd6ex3PJzwQW6YyfujpX05Xj37RP/IG8Jf8AYxW//oL0AdB4o+JeqeGvF39lL4G1rUrA+WF1OzQtGS2Mj7uBjP8AerovF3jbQvAumwX/AImvGtLaecW6OsTSfMVLdFBOMKecVv1BeWNpqEBhv7WG6iPWOeMOv5GgCnoXiPSfEuhx6xol6l1p8m7bOFKg7Tg/eAIwQavW9zBdwiW1njnjPR43DA/iKpnSNIsfD8+mx2ttY6X5UiyQwKII0Rslz8uNuckkjHUmvOvhhonwv0fxlqK/DvVBdag1pieGO4aaNIt45ViMH5tv8RoA9WorwfxJqHxQi+PEkHhs60+gfa7YTj7GptkhKx79jvkHjcTjBzmuz+L8/wARoNKsT8MIt8haT7ayLCzquF27RJ/wLoCaAPRaiiureaaWGGeKSWHHmIjgsmemR26H8q4LT5Lx/gDK/wATHu4rg6ZP/abKAs6r8/I7B9uMds4rL+CUXw2jt9SPw1nuppysX2/7V5u8D5tmQwC/3vu+9AHRSfFvwcnjOPwquptJrD3P2UwLbyYSTOMFioXr6E0vj7xxq3hGSxh0XwfqXiSW8Dn/AEPIWLbt++Qrbc7uMjsa6O203Rl1Ce6s7KxF4ZMzzRRJ5hf/AGmAzn61oUAc7dX3iS/+Hy3ujWENj4hntEkSzviWSGU4LIxGM45Gf/1V4n8S7fx1DoHh1viHd6ZcTSeJbU2q6eCBGu19wbKjvjua+jq8e/aI/wCQT4Q/7GK3/wDQWoA6DTvh1r9t4/PiG+8fateWa3Us0OkEMsCo24LGRvIbaG64HSrnjj4U+G/iFqFnd+Ixds1nGY40gm2KQTnnjP612tFAGTP4X0i78KR+G720F1pUcEdv5EzE7kjxtBPU42jn2o8P+FtD8K20lv4d0u306KVg0iwJjeQMZPrWtRQBRt9E0q0u2urXTLOG4Zi7TR26q5Y9SSBnJyeavUUUAFfP2vfBbxVqPx9j8YW/2H+y11W2uzunIk8uMxlvl29flPGa+gaMj1oAKKKKACiuL8ffE/Sfh89lb39lqN/e34b7LbWUG8ybcZ5JA7jgZPPStDVrvxDrHw/W98IJHp2tXcEUsEWpxkCEsVLK4xkMFLduo6UAdJRmuX8B6V4s0rSLhPHOuQazfyz+YkkEIjWJdoGwYC5GQT0HWsfw18HdE8N+Mj4nTUtWvtS3SFPtdyHRA4IKgbc4we5NAHTeJ9etNF0S+kfUbW1vEtJJIFmlVWLBTtIU9eRWD8K/Emqax8LNN1nxjOI7+ZpRNJPGsHSVlXjAA+UD60ePPhx4N8TyyeIPFmkyX81hZso2XMkeY03PjCsozkt+dN0seG/jF8K7Uz2Fymi3Z2pbSyFJF8qQqMlWPdM9aAOyh+x3LC5t/IlYHAlTDEfiKo6x4V8P+IJFk13Q9O1KRF2I93apKyr1wCwJA5NVPBngfRPAWkTaZ4chkhtprg3DrJKZDvKqp5PbCiuf8K/DPU/C/iw6mfHGtalYESZ0y8cvGWbofvYGM/3aAOs1rwzo/iHw8+havYpcaY6optgzIuFIKj5SCMEDp6VW8JeCdB8DafPY+GLI2dtcTee6GV5MttC9XJPRRxmud8aa38StG8Rxy+E/Den63oflKJEafy5xJk7iCWHGCOzdK6TxP4x0PwXp9te+J75bGC4nFujlWcbyCf4QTjg84wKAOa8OfCKx8MeMx4gsfEOuT5eV3srm5DQsXBB+UKOm7I69ql8deG/H2q61Be+B/GEOjQx24je0ntRIsj7mJckhuoIGMdq7HTNVsNa0+O+0i9gvbSUfJPbyB0b8RVugD53+N3wd8ZeNPFVrrGiRW18IdOitpQZ1id5FZyxAbC4+YdxXsPgHUfE2o+HWbxnokWi6hDMYVgimEgeMKuHyCRySwxk9K6fIHU0UAefeEPi/pfi3xOdATRda02/CPJi9tgiFV6nO7PcdR3rpdS8aeGdG1gaVq+vafYXzRiQQ3VwsRKkkAjcQOxrcrmvEXw78JeLLoXXiHQbS9udoXz2UrIQOg3KQcfjQB0oORkciiuc8b+D08baCmmPquoaUY51nS4sJAj7lBABOOnzZwMcgc0vhXw7qPhbwk2mSa3da7eoZGju9RdizE/dUkljgcD+lAHRVkeLri/tPBOuXGj+Z/aEOn3Elr5Sb280RsUwuDk7scYOa5rwJe/Et9YnsviDpelR2UcBaG/sJOZZNwG0ruJ6bjnaOgqj4u+OfhnwZ4g1DRdXttSF5aRB42EH7q4YoGCq2c5yQuSMA96AOh+Gl9rWpfDfR7vxR539rSxMbn7RD5T7t7AZXAxxjtXU1znhfxrp3iTwlpeuuyadHqgJhhuZlDEhiMD1Py54ro6ACiiigAooooAK8d+Pfww8QfEdtBPh37J/oAuPO+0SlPv8Al7ccHP3DXsVGcdaAMDwJo114d8A6Jo+obPtVjZxwy+W25dyjBwe4rfoooAKKKKACiiigArz7w5q3ii4+NXivTtT+1f8ACP20ELaf5lsFi3FELbZNo3clu5xXoNcrpHju21f4ja54Rjs5Y7jRoo5JLhmBWQOqkADqPvUAdVRRRQAUUUUAFFFFABXz94J+C3irQPjovi2/+w/2YLy7mPlzkybZUkC/Lt65cZ5r6BoyPWgAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigArz74d6t4o1Dxl43t/En2r7BZ6iI9K8+2Ea+Vuk+420bxgLzk9vWvQa5Xwl47tvFniHxJpNvZy28nh+7FrLJIwIlOXGVx0HyHr60AdVRRRQAUUUUAFFFFAHnXxu8D6t4/8CQaRoH2f7Sl9HcH7RJsXaEcHnB5ywqz8G/B2qeBfh5FouueR9rS4lkPkPvXDHjnArvM460daACimySJFG0krKiICzMxwFA6kmsr/hLfDn/Qf0v/AMDY/wDGgDXorI/4S3w5/wBB/S//AANj/wAaP+Et8Of9B/S//A2P/GgDXorI/wCEt8Of9B/S/wDwNj/xo/4S3w5/0H9L/wDA2P8AxoA16KoWWu6RqU/kadqtldy7d3lwXCO2PXAPSr9ABRRRQAUUUUAFFNkkSKNpJXVEQFmZjgKB1JNZX/CW+HP+g/pf/gbH/jQBr149+0j/AMid4c/7GK2/9Fy16V/wlvhz/oP6X/4Gx/41wvxXstK8faFpVjpvinQ7aSy1SK9dri9XDIiuCBjPPzCgD1Gisj/hLfDn/Qf0v/wNj/xo/wCEt8Of9B/S/wDwNj/xoA16KoWWu6RqU/kadqtldy7d3lwXCO2PXAPSr9ABRRRQAUUUUAFFNkkSKNpJXVEQFmZjgKB1JNZX/CW+HP8AoP6X/wCBsf8AjQBr149+0h/yKXhr/sYrf/0XLXpX/CW+HP8AoP6X/wCBsf8AjXC/Fay0rx7omk2Wm+KdDtnsdUivXa4vVwyIrggYzz8woA9RorI/4S3w5/0H9L/8DY/8aP8AhLfDn/Qf0v8A8DY/8aANeiqFlrukalOYNO1Syu5Qu4xwXCO2PXAPSr9ABRRRQAUUUUAFFNkkSKJpJXVEQFmZjgKB1JNZX/CW+HP+g/pf/gbH/jQBr0Vkf8Jb4c/6D+l/+Bsf+NYPib4v+CPCcSHUtchlkkUtHFZgzs2P9zIH4kUAdrRXD698RbZvA51bwZeaTf6lNFFLbWd5fRxbg5XIbLjBCkkgkdMVD4E8X6hJo08nxB1zw1HfyTloY7C+QrHFtHyn5jyDu5yetAHel1DhSwDEZAzya4fxf8X/AAl4J1lNI1i6uG1Fgh+zW9szEK3Q7jhcfQ5rK8B/Dzwfofi6TW9L8T3Ouay0bgtNqSTFY2PI2ryQOOSTXo32Cz+3/bvskH2vYE+0eWPM2jOBu645PHvQBg+OtS8V6ZokMngXRrbWNQknEbRXEojWNCrHecsuQCAMZHWn6BF4m1HwOYPF7W+n63cRyxySacx2w7iwRl54IBHfqOtdHRQBw3gL4at4L1C61G78T6xr19dxCKV7+XcmAc5AOSD/AMCPWt0+CvDL+IJdck0HT5dUldXa8kt1eTcoABBIOCAB0x0rcooAKKbJIkMTSSuqRopZnY4CgdST2FZX/CW+HP8AoP6X/wCBsf8AjQBr149+0b/yLXhf/sYrf/0CSvSv+Et8Of8AQf0v/wADY/8AGuF+KllpXjzSdHtdN8U6HbvYapFeyG4vVwyKrAgYzz81AHqNFZH/AAlvhz/oP6X/AOBsf+NH/CW+HP8AoP6X/wCBsf8AjQBr0VQstd0jUpzDp2qWV3KF3GOC4R2A9cA9Oav0AFFFFABRXI+Ofid4Z+H9rnW73feOuYbC3w88vp8vYe7YFcAF+K/xWALN/wAIF4cl6AZa8mT9GH/jnX+KgDtvGvxc8J+BiYNSvvtWo9F06yAlmJ7AjOF/4ER7Zrjf7d+MHxD40DSYPBGkyf8AL3f/AD3LL7KRkcf7I/3q7Hwd8KfCHw/hN3Y2ay3qKWl1O+YPL7tuPCD/AHQPeuh/4S3w5/0H9L/8DY/8aAOB0L4AeG7a8/tLxfd3nizVW5e41KUlCfZMnI9mLV6ZYadZaVaLa6ZZ29nbp92G3iWNB9AABVH/AIS3w5/0H9L/APA2P/Gj/hLfDn/Qf0v/AMDY/wDGgDXorI/4S3w5/wBB/S//AANj/wAaP+Et8Of9B/S//A2P/GgDXoqhZa7pOpTmHTtUsruULuMcFwjsB64B6cir9ABRRRQAUUUUAFFNkkSGJ5ZnWONFLM7HAUDqSewrK/4S3w5/0H9L/wDA2P8AxoA168e/aK/5AXhT/sYrf/0F69K/4S3w5/0H9L/8DY/8a4X4p2WleO9O0W307xTodu2n6pFeyGe9XDIoYEDGeeaAO81zwvoXia38jxBpFnqKAYX7RCrFf90nkfhXmd/8Bv7FvH1L4W+Jb/wveE5NsZGltpPQEHnH+9uHtXpX/CW+HP8AoP6X/wCBsf8AjR/wlvhz/oP6X/4Gx/40AeWj4neP/AP7v4neEmv7BODrOi/MuP7zL0H47PpXpHhPx34b8b2X2jw3qkN2VGZIc7ZYv95DyPr0PY1pWeuaRqcxg0/VLK8l2ljHBcJI231wD05FcF4s+Bvh3Xb3+1vD8k3hjXEO+O+0w7Bu9WQED8VKk9yaAPTKK8S/4Tn4i/C4iL4jaR/wkWhocf23pg/eRr6yLwPTqF/3mr1Twx4u0LxlpQ1Dw5qMN7BxvCHDxn0dTyp9iKANmiio5riG3CG4ljiDuEUuwXcxOABnqSe1AElFcf8AEL4j2Hw7s7SW+07UNQlvWdYIrKINkqATuJIxwffofSqb+OYvEvw6+16Jrem+H9bvLUPDHqFxGTayf3XU/iPu++KAO8rwX48+L/D+rP4Z0fStYtL2/t9eglmht5BIY1G5TuI4ByQMZzXa+AbhvD1neHxd8SNP8QXV3IsgzPGiW+AQVT5uh47L06VzHjf4f+CNYNhP4T1Lwzpd7Hq66heXMl2u+ZcsWUNkkZY528Lx7UAdb448Y+NdI16LSfBngiTWjLAsv2+Sfy4EJYjYcgDIxn7w4Irb8Z6FrXibw2lloOvT+HLtpUeS5hTe4TB3JwR6jkEdKu/8Jb4c/wCg/pf/AIGx/wCNH/CW+HP+g/pf/gbH/jQBV8IeFX8M+Fho9/q95rhZnaS5v23O+/qD1OOvUnrU3h/wb4c8KqR4d0Sy09mXY0kEIEjD0L/ePTuauWWu6TqU5h07VLK7lC7jHBcI7AeuAenIq/QAUUUUAHXrVW10ywsZ5prGxtraW4x5zwwqjSYzjcQOcZPX1NWqKAOC/wCFOeGE+ICeMbZtQt9T+0m6kVLkmOZzydykE4yc4BA49KsePtP+IV1cWVz8PNY06zWBH+0Wt9FkXBJGPm2tjGD0x1612csscMLyzOscaKWd3OAoHUk9hWV/wlvhz/oP6X/4Gx/40AU9V13VfDfw/XVr/S5dX1W3t4jcWWmRkmSUlQ+wcnaCSfoK8Y+Jnjy08e+FfDN5Z6bqGnG18UW8E0V/EEYPsLEDBOeCPTrXuf8Awlvhz/oP6X/4Gx/41wvxRstK8dWWhw6d4p0O3bTtUivZDPer8yKCCBjPPNAHoOna9pGrySx6VqlnevCxWRLedXZCDgggHIIPrWhXm+k+FfhXoPi1PEmjXWlWeoqHG5NSBQlxhjtLkA4J6Y61Q8a3niO78RR6n4F+JXh+xtlhWN9NvJImjYgklt3zcnPYDoOaAPV6K4vxX8StP8O6PHeaTA/iaXzQs1tpMizSRR4O6Uhc4UEAc4GWHNa3hPxfp/jDwrF4gsYri0s5C4xeoI2XYSrZ5IwCDzntQBvUVBZX9pqVqtzp11DdwN92WCQOp+hHFT0AFfHvi7j9ruL/ALGCx/8AQoq+s7++n/sW9uNCSDULyGOQQw+cAryqD8hYZwcjB9K8g8HfBjUNY8WSeOviWY4dalvIb23s7F8LbtGQQHPIP3VGBnp945oA7Xwx8WtA8Y+MrjQfDsV9dpbQs8moC3YW4YEDZk8gnnkgA44zVfUPhfdaz8RU8Sar4u1aSytbiO4sdKicJFC64PPUMCQewODgk13VnY2mnwtFYWsNtGztIyQxhAWY5LYHcnkmp6AEKKXDlQWUEBscgHr/ACFLRRQAUVVvtUsNLRH1K+trNXOFa4mWMMfQZPNVYvFGgTzJFDrmmySSMFREu4yWJ4AAzyaALt7BBc6fcQXuPs8sTJLlto2EYPPbjPNZvhLRtD8P+GbbTfCojGlQlzCI5zMvLlm+ckk/MT3q7rFm+o6Hf2UTKr3NtJCrN0BZSAT+dcZ8OLTT/hz4DsfC2u+INJ+32LS+btulUDfIzjhsEcMOooA7+iqVjrOmao7ppupWl4yDLLbzrIVHqcHirtABVTU9J07WrFrPWLG3vrZ/vQ3MQkU/gat0UAYX/CNR6T4LuNB8HGLRcQSpZuqF1t3fJ34J5+Ziev8AhWT8PB4+gt76z+In9nztbMi2l9aHm6XB3MwGAMcfwr347ns6KAPjT9o3XNM1/wCKEVxo17Deww2CW8jwtkLIksoZT7j+RB6GvsWz/wCPGD/rmv8AKvNde/Z68Ba/rcmpSWdxYtIuHgsZFiiLZ+9t2nB+mB7Zq+nxN/s34ljwZrfh7UbBLhxFpV/s8yO7AUZJ2/d5zzzgfe20Aeg0UUUAFFFNlljgheWZ1jjjUs7ucBQOSSewoAdVDXLHS9R0O8ttfjgk054XFz552ose07iW424GecjHXNV/+Et8Of8AQf0v/wADY/8AGm67Zx+K/BOqWGnXkLJqdjPbRXKMHQF0ZN2R1AJ7elAHP33wz8CeLvBumaUlqlxo9kHNg1neOVj3H5irBiG5z1zWtb+FpPD/AMPD4b8IXjWcsFq8NldXP7wxMckOeOSCc9Kq+DtNtfh14A0jQtb1azR7VGjE8kgiWQ7i3G4+9bMXijQJ5kih1zTZJJGCoiXcZLE8AAZ5NAGD4At/H9ot9b/ES80y+EZQWdzYoVaX724uMKB/DwFHeqWnfGDQ734gSeDriw1Sy1P7RJBA09riO42Z+ZTnO0hSckAY7139NaNHdGdFZkOVJGSpxjj04oAabiFbhYGljEzKWWMsNxHqB1xUlcR48+E/h74g3dve6u99bX1rH5cF1Z3BRkGSehyvUnnGffpWl4ntvEtj4MS28BG0fU7cRJGdRclXRcZBPdiBjnHUnNAHS182ftcff8JfS8/9oV7d4N1bxHeeF5Lzx5pdtot/DI4kjhl3oY1A/eZycA88ZPA61xuo23w0+PpslXWJLufTUkdLeCUwSqrlAxZGXJAKqMjjn3oA6j4Tf8kh8L/9g2H/ANBrr6oaFo1r4d0Cy0fTt/2WyhWGLzG3NtAwMnvV+gAooooAKKKzbnxHollcPb3ms6fbzIcPFLdIrL9QTkUAaVc/pul+Frfxtq2oaYbT/hIbiNBqHl3O6XaAu3dHuO3gL2FadjrOmao7rpuo2l4yDLi3nWQqPfB4rk9G8Ey6D8VPE/jS71CD7HqtvEoiIKmERooLMx4x8hNAHc0Vkf8ACW+HP+g/pf8A4Gx/41pW11b3tulxZzx3ELjKSxOGVvoRwaAJaKKKACiiigAr49+F/wDydcn/AGEtR/8ARc9fYVef6L8F/CmgeOh4ssBe/wBpiaWYeZPlN0gYN8uPRzQB6BRRRQAUUUUAFFVb7VNP0tUbU762s1c4Q3Eyxhj7ZIzVWLxPoE8yQwa5psksjBURLuMsxPAAAPJoA1KKKzbnxHolncPb3ms6fbzIcPHLdIrL9QTkUAaVFU7HWdM1RnXTNRtLxkGXFvOshUe+CcVcoAKKKKACiiigAooooAKKiubq3srZ7i8njt4E5eWVwqr25J4FZv8Awlvhz/oP6X/4Gx/40Aa9c94d0vwtYazrk/ho2n2+7uQ+q+Rc+Y3m5b767jsOS3GB39K34pY54UlhdZI5FDI6HIYHkEHuK8+8KeHLT4f+LfFmp61r2nxjxHei7giklETRqGc4O48/fHT0oA9DorNtvEeiXtylvZ6xp9xO/CRRXSMzd+ADk1pUAFFFFABRRRQB4v8AtS/8kntf+wtD/wCi5au/s0/8kat/+v2f/wBCFdz428EaR4/0JNI1/wA/7Kk63A8iTY24BgOcHjDGpPB3g7S/Avh5NF0PzvsiSNIPPfe2W684FAE/i3/kSdc/7B9x/wCi2r4k+G/w31H4l6zd6dpd5a2j2tv57Pc7sEbguBtB9a+2/Fv/ACJOuf8AYPuP/RbV81/sof8AI9a1/wBg0f8Ao1aAD/hlDxP/ANB/Sfyl/wDiaP8AhlDxP/0H9J/KX/4mvquigD5U/wCGUPE//Qf0n8pf/iaRv2UvE6oW/t7STgZ6S/8AxNfVlNl/1L/7poA+Qf2X/wDkrkn/AGDJv/Qkr7Ar4/8A2X/+SuSf9gyb/wBCSvsCgAooooAKKKKAMfxf/wAiPrv/AGDrj/0W1fEvw3+G2o/EvWLvT9LvLW0e1g893ud2CNwXA2g+tfbXi/8A5EfXf+wdcf8Aotq+bf2UP+R41v8A7Bw/9GLQAn/DKHif/oP6T+Uv/wATR/wyh4n/AOg/pP5S/wDxNfVdFAHyp/wyh4n/AOg/pP5S/wDxNNf9lLxOkbN/b2knaCcYl/8Aia+raZN/x7yf7p/lQB8h/svf8lcl/wCwZN/6HHX1/XyB+y9/yVyX/sGTf+hx19f0AFFFFABRRRQBjeMf+RF17/sG3H/opq+J/hv8NtR+Jer3en6XeWto9rB57vc7sEbguBtB9a+2PGP/ACIuvf8AYNuP/RTV83/sn/8AI7a5/wBg4f8AoxaAG/8ADKHif/oP6T+Uv/xNH/DKHif/AKD+k/lL/wDE19V0UAfKn/DKHif/AKD+k/lL/wDE02T9lPxPHEz/ANvaSdoJxiX/AOJr6tqO4/49Zf8AcP8AKgD5F/Zd/wCStz/9gub/ANDjr6+r5B/Zd/5K3P8A9gub/wBDjr6+oAKKK4bTfi94W1jx6vhLSZLu7v8AdIkkqWzLFCyAllZmwc8EcDGe9AHc1U1LVdP0aza71e+trG2XrNcyrGo/EkCuK8caR8StY8QQ23g/XtO0TRfJUzXDw77gybjlQCCCMbT1Xqea6TxN4P0XxnpdvYeKLJL+CCZZ1TcyDzACM/Kc4+Y8ZoAoa9r2n678Kdd1bQbmPULR9Mu/LkhPEhVHUgfiCK+OPhv8NtS+JesXen6Xd21o1rb+e8lxuwRuC4GATnn9K+y9c0bTdB+GWt2Gi2NvY2kem3O2G3jCKCY2ycDufWvn79k//kdNc/7B6/8AowUAM/4ZQ8T/APQf0n8pf/iaP+GUPE//AEH9J/KX/wCJr6rooA+VP+GUPE//AEH9J/KX/wCJpsv7KfieOF3/ALe0k7VJx+95/wDHa+raiuv+POb/AK5t/KgD5H/Zc/5K1cf9gqb/ANDjr69r5C/Zc/5K1cf9gqb/ANDjr69oAKKKKACiiigDE8af8iFr/wD2DLn/ANFNXxT8NvhrqPxM1a8sNLvLW0a0gEztc7sEFgMDaD619reNP+RC1/8A7Blz/wCimr5y/ZP/AORy13/rwX/0YKAGf8MoeJ/+g/pP5S//ABNH/DKHif8A6D+k/lL/APE19V0UAfKn/DKHif8A6D+k/lL/APE0yb9lTxPFA8n9vaSdilsfvecD/dr6uqG7/wCPGf8A65t/KgD5J/Zb/wCSs3P/AGCpv/RkVfXlfIf7Lf8AyVm5/wCwVN/6Mir6A+IHxZ0XwKyaeiSatr8+BbaVafNIxPTdjO0H6EnsDQB2Gq6tp+h6ZNqOsXkNnZwLukmmcKq//X9uprx+5+IvjP4oXUunfCawOnaQrGObxFfptB9fLUjr+BbkcL1p2lfDDxJ8RdTh8QfGO6K2yNvtPDts5WKEdvMwevtkn1b+GvZLSzttPs4rSxt4ra2hUJHDEgVEUdgBwBQBwfgb4N6B4Quv7VvWk13X3O+XVL/5339ygOdv15b3r0KiigDE8bf8iB4g/wCwZc/+imr4q+G3w01H4mareWOl3traNaQCZ2ud2CCwGBtBr7V8bf8AIgeIP+wZc/8Aopq+c/2Tv+Rw17/rwT/0YKAI/wDhlDxP/wBB/Sfyl/8AiaP+GUPE/wD0H9J/KX/4mvquigD5U/4ZQ8T/APQf0n8pf/iaZN+yp4nht5JP7d0k7FLY/e84H+7X1dUN7/x4XH/XJv5UAfJP7LX/ACVi6/7BM3/oyKvryvkP9lr/AJKxdf8AYJm/9GRV9eUAFFFFABRRRQBh+N/+SfeIf+wXc/8Aopq+K/ht8NNS+JmqXllpd7a2jWkImdrndggtjA2g19qeN/8Akn3iH/sF3P8A6KavnX9k7/kbte/68U/9GUARf8MoeJ/+g/pP5S//ABNH/DKHif8A6D+k/lL/APE19V0UAfKn/DKHif8A6D+k/lL/APE1HP8Asq+J4LeSX+3dJby0LY/ejOBn+7X1fUF9/wAg65/65N/I0AfJf7LP/JWLv/sEy/8Ao2KvruvjP9nPxBpnh74p+ZrF0tsl5ZtZwMykh5Xli2rwOM4PJ4GK+mfiDceP1WxtPh1Z6c7XJcXV5evxagY2kLnnOT2bp0oA665lggtZJLySOOBVPmPKwCgd8k8Yrxu++GWheILy48U/BbxHDo2r20rRSPp8ga0mkADFGUZA6r0yv+ya9GvPCy+KPAcOgeOfLv5JYYhevbkxrLIhDFlIwVyy54x1q74c8K6H4S042PhvTINPty25liHLnGMsx5Y47kmgDxzwt8cPEum+KrLSfi1o8Xh+0ngZUvZbaSLzJQRhiSSoXGQcDAJByBXZ698MvCmr+Kn8eeINUvJ4bdEuokN6VtbcRqD5ikcgfKG4IH512mv+HtJ8UaRLpmv2EN9Zy9Y5V6H1B6qfcYNfHfxOvpfBWp6p4A8K+J72+8O71ae0lIIgkBJMQbuAcE4wM8EEjNAH1LqviHTfFnwb1fW9Fm86zutIumjYjBGI3BBHYgggj2r5A+G3wz1L4mapeWWlXtraNZwiV2ud2CC2ABtBrsfgx49+w+FPFng3UJcQX2l3dxY7j92YQNvT/gSjP1U+tbX7J3/I2a//ANeKf+jKAIf+GUPE/wD0H9J/KX/4mj/hlDxP/wBB/Sfyl/8Aia+q6KAPlT/hlDxP/wBB/Sfyl/8AiajuP2VfE8FrLN/bukt5aFsfvRnAz/dr6vqvqH/IMuv+uL/yNAHyb+yx/wAlXvP+wRL/AOjYq+uq+Rf2WP8Akq95/wBgiX/0bFX11QAUUUUAFFFFAGD47/5J14j/AOwVdf8Aopq+Lvht8M9S+Jmp3tnpV7a2jWcIldrndggtgAbQa+0fHf8AyTrxH/2Crr/0U1fPP7Jv/I1eIP8Aryj/APQ6AIP+GUPE/wD0H9J/KX/4mj/hlDxP/wBB/Sfyl/8Aia+q6KAPlT/hlDxP/wBB/Sfyl/8Aiaiuf2VvE9vayzf27pLeWhfH70ZwM/3a+sKral/yCrv/AK4v/wCgmgD5P/ZY/wCSrX3/AGCJf/RsNfWd5aQX9jPZ3cfmW9xG0UqZI3KwwRkc9DXyZ+yx/wAlWvv+wRL/AOjYa+nvFvi7R/BPh+bWPEFz5FtGdqgDLyuQcIg7scH8iTgAmgDA8P8Aw+8F/CyfU9f00tpdvLAFuDc3RaGJAc8FuRk46k9sVxVzf/EX4jfapfBOv6c3hHVrgwR3sCmO70+NSA3BCnJxnGSSG4I4I7u3g0f4w/Dmzm8Q6JcwWd4yzi0uiUkUq3DAjBwR0I6q3vXWWNhaaXYQ2WnW0VrawKEihhQKqD0AHSgDzSDVfA3wA0vTfD97cX0aai0k73bo026QbQzuB93PGAo/hOfU2f8AhfPg1/F1p4dga/nubyaCGCaO3HlOZgpRslgcEOvavQ7yytdQtWtr+2huoH4aKaMOrfUHg18bePtFW+/aWl0i38yxgn1W0tUktl2mBSIlBTHA2g8emKAPtCiuX8C+GNW8JaHPY614ku/Eb+eXgnuhh0j2gBOSSTkHknv2rH8F/ETXNf8AEsug+JPBOpaDdxxPMtw53wOqkD7+AM/MPu7qAPQKK5bxv4403wp4X1u8+3Wpv9PtDKtr5qmQO3EeVzkAsVGcVx3wV+M1v4+sF0jW3jg8RW6fMOFW8UfxqP7395fxHHAAPHfjdrOv+LfjM/gqW/Q2MN/BDYxPEqrC8qRgksq7iMt3Jro/Df7MfiPRfFWk6pPrelyRWN7DcuiCTcwRwxAyvXivZdQ+Efg3VPGI8U3umSPq/nx3Hni6kA3x7dp2htvG0cYrtKACvnv4jfs7694z+IOqeILHWNOt4L10ZI5hJuXEarzhcdVr6EooA8g+C3wd1b4Z6vqd3qmo2d4l5AkSLbb8qQ2cncBS+Pfj1H4G8fHwy3h1r04iP2gXnl/fAP3dh6Z9a9er5b+NfgbxTrXxsOpaRoGoXtltth9ogt2ZPlAzyPSgD6kooooAKKKKAPP/ABx8Z/C/w/15NI15L9rl4FuAbeAOu1iwHJYc5U13qFJ0il2g8b0JHK5H6cGvkj9qON2+LFsVViP7Kh6D/ppLX1pZ/wDHjB/1zX+VAHCeJT8R9M+IdlqHh37Hq/hy6Mdtc6dJiJ7Xn5pt/JPUnIzxxt4zXoORnGefSivL/iv4Zkt7y18fad4rl8O3ejoEneUmSCSAtlk8vuxzwvRjgcHBAB6ZLcwwywxzTRxvOxSJWYAyMFLEKO5wpPHYGvGv2lfFmteGvCNhb6JefZotUea1u18pH82Ix4K/MDjqeRg14b4q+M+teIvibp/iiMtBbaTcB7Cy3fKiA/MG9WccMffHQCvq7XPCnhX4qeHtLutZtmv7FkF1aFZnj4dQc/KR2xQB8zeDv2d9f8Z+EbHxBZaxptvBeqzJFL5m5QHK84XH8NfUHw68MXPgz4faVoF7PFcT2SOryQ52tl2bjPP8Vavh/QNO8L6DbaNosJgsbVSsUZcuVBYseSSTyTWjQB5n8afhjqPxN0nS7XS761s2sp3kc3O7DAqBxtB9K8z8Nfsx+ItE8WaTqs+t6XJFY3sNy6IJNzKjhiBlevFfTFFAHFfFP4iL8M/C1vrL6YdSE94tr5In8rG5Hbdna39zGMd6t/DjxsvxB8GQeIEsDYCaSSPyDL5mNrYzuwOv0rjf2kPD+r+I/hxY2egabc6jcpqscrRW0ZdggilBbA7ZIH41qfATRtS0H4TWdhrVjPY3aXEzNBcIUYAuSDg0AekUUUUABAZSGGQeCD3rzbxf4l8A/CDWbbVL3RBaX2rxyR+fptogLqhQsGwV7svPfFek183ftbozv4S2qW4vOg/64UAegWekeLfEPiPT/GvhPxtNFoGqtBcS6Re24ISDChlT7wDEA9AOTnNb3j/4kW3w8+wTalo+p3tjcl/PurKHetrjGN5OBzk9/wCE1J8JwR8IvDAIwf7Nh/8AQa68gEEEZB6g0AU9I1W21vRbPVLAubW8hWaEyIUJVhkEg8jg1zHxK+JOn/DXS9PvdShe4F5eLB5UZw4TBLuB32jHHckCtnxV4WsvF3hW50C+luLa0uAoLWjiN12sGGDgjqo4xXxd8V/tOm+MJvDD+Ir7XrTRXMUE165Zo2YKXQcnoQF7cqeKAPsjXfEccvwx1XxF4cvY5VXSp7uzuYwGXcsTMrYIxwR0I7YIr5E8J+DvEnxy8WavezaraLfoiTXM9ymzzCflGFjXA4X0FWfhN8UV8K/afDXifzLnwpqyPBdRBjutt42s6Y5wQfmA57jkc/UvgX4eeD/B4k1HwZa+WmoQpmZbp5llT7ykFmIxznIoA5H4LfB7VfhlqmqXWqajZ3i3sKRoLbflSrE87gPWvSvE2lya54S1fSYHWOW/sZrZHfO1WeMqCcdua1KKAPlT/hlDxP8A9B/Sfyl/+Jr6A8D6DceA/hjY6RfSRXU+mW8hd4SQr/Mz8ZGe+OldXVXVI3l0e8jjUs7wOqqBySVPFAHmPws+OUfxN8UXOjJoDaaYLNrrzTeebu2ui7cbF/v5zntXrFfMv7N/grxN4b+I1/d6/oV/p1s+kyRLLcwMilzNEQuT3wpP4V9NUAFFFFABXnuifGvwt4g8eDwjYJqA1IzTQ5kgUR7ogxb5t3TCHHFehV8f/C+Nx+1ajFGA/tLUOcf9M56APsCiiigAooooA+QfjPq+v+MvjSfBU+oIbKLUorawjeJVWFpVjXJZV3EZbuTXT+GP2ZfEWheLtH1a41vS5IrC+guXRBJuZUkDEDK9cCvY734R+DdQ8ZDxVdaZI+ri4juhOLmQDzExtO0Nt42jjGK7SgAr55+If7O2veMfiBquv2WsabBBeyK6RzCTcuEVecLjtX0NRQB5D8Fvg/qvwy1PVbnVNRs7xb2GONBbb8qVJPO4D1o8c/HqPwX8Qz4Wbw614QYR9pF5s/1gB+7sPTPrXr1fLfxi8DeKdY+OranpWgaheWO61/0mG3Zk+VV3cj0xQB9SUUUUAFFFFABRRRQB87ftR+LNa0v+z/D9jeeVpmp2rNdweUh8wrIpX5iNw6DoRXIaL+zH4j1vQNP1WDW9LjivraO5RHEm5Q6hgDhevNfRnjD4Z+FvHl1bXHifT3u5bVDHEVuJI9oJyfukZro9OsLfStLtdOsUMdtaQpBChYnaiqFUZPJ4AoAq+GtMk0TwppOlTusktjZQ2zumdrMiBSRntxXmfxp+DurfE3VtLu9L1Gzs1s4HidbnfliWzkbQa9fooA+evh1+zvr3gz4g6X4gvdY064gsndnjhEm5sxsvGVx/FXoHxZ+LKfC2LSnfR21P+0WlAAufJ8vZs/2Wznf7dK9Frwj9pvwtr3ia38Njw9pF5qRt3uTKLWEv5e4RYzjpnB/KgD13wf4hHizwfpmvLbG1F/AJhAX37M9t2Bn8q2q5L4WWF3pXwr8P2OpW0lrdQWapLDKpVkOTwQeldbQAUUUUAc5448c6T8PtBTV9eW4a2edbcC3jDtuYMRwSOMKak8GeMtM8d+HU1rRBOLV5GjAuECNlTg8AmvN/2o1LfCe1Cgk/2rF0H/TOWrv7NSlfg3bhgQftk/BH+0KAPQ/Fv/Ik65/2D7j/ANFtXzX+yh/yPWtf9g0f+jVr6U8W/wDIk65/2D7j/wBFtXzX+yh/yPWtf9g0f+jVoA+q6KKKACmy/wCpf/dNOpsv+pf/AHTQB8g/sv8A/JXJP+wZN/6ElfYFfH/7L/8AyVyT/sGTf+hJX2BQAUUUUAFFFFAGP4v/AORH13/sHXH/AKLavm39lD/keNb/AOwcP/Ri19JeL/8AkR9d/wCwdcf+i2r5t/ZQ/wCR41v/ALBw/wDRi0AfVVFFFABTJv8Aj3k/3T/Kn0yb/j3k/wB0/wAqAPkP9l7/AJK5L/2DJv8A0OOvr+vkD9l7/krkv/YMm/8AQ46+v6ACiiigAooooAxvGP8AyIuvf9g24/8ARTV83/sn/wDI7a5/2Dh/6MWvpDxj/wAiLr3/AGDbj/0U1fN/7J//ACO2uf8AYOH/AKMWgD6pooooAKjuP+PWX/cP8qkqO4/49Zf9w/yoA+Ifgt4xj8EeO5NSfTLzVJZbKS3gtbJNzySMyEfh8p6A/Svry1udZ8X/AA5MohufCur31s4RJRuks5MkKTkDPQHoDg9jXzP+y7/yVuf/ALBc3/ocdfX1AHH/AA/8B3HgqG+e/wDEmpa/eX7I8818+VBUH7gJJGc85Y9BXUWthZ2TTGytYLczyGWUxRhPMcnJZsdSSc5NWKKACiiigDF8Zf8AIia9/wBg24/9FNXzj+yf/wAjprn/AGD1/wDRgr6O8Zf8iJr3/YNuP/RTV84/sn/8jprn/YPX/wBGCgD6oooooAKiuv8Ajzm/65t/Kpaiuv8Ajzm/65t/KgD5H/Zc/wCStXH/AGCpv/Q46+va+Qv2XP8AkrVx/wBgqb/0OOvr2gAooooAKKKKAMTxp/yIWv8A/YMuf/RTV85fsn/8jlrv/Xgv/owV9G+NP+RC1/8A7Blz/wCimr5y/ZP/AORy13/rwX/0YKAPqiiiigAqG7/48Z/+ubfyqaobv/jxn/65t/KgD4h+DieLZvGF1b+AWt49UuLB4mnuCAIIjJHukGe447HjOATivqj4efCjSfAqvfzSPq3iC5y11qt180jMeoXOdoP1ye5Pb5//AGW/+Ss3P/YKm/8ARkVfXlABRRRQAUUUUAYnjb/kQPEH/YMuf/RTV85/snf8jhr3/Xgn/owV9GeNv+RA8Qf9gy5/9FNXzn+yd/yOGvf9eCf+jBQB9T0UUUAFQ3v/AB4XH/XJv5VNUN7/AMeFx/1yb+VAHyT+y1/yVi6/7BM3/oyKvryvkP8AZa/5Kxdf9gmb/wBGRV9eUAFFFFABRRRQBh+N/wDkn3iH/sF3P/opq+df2Tv+Ru17/rxT/wBGV9FeN/8Akn3iH/sF3P8A6KavnX9k7/kbte/68U/9GUAfU1FFFABUF9/yDrn/AK5N/I1PUF9/yDrn/rk38jQB8h/sy2Fnf/FeYX1rDciDTZJYhNGH8txJGAwz0IBPPvX2HXyJ+yz/AMlYu/8AsEy/+jYq+u6ACiiigBGUOhU5wwwcEg/mOleQ6v8Asz+BNTkklt21Swkcli0N3vyT3PmBifzr1+igD5k8T/swjQdHv9Z0bxTIRYW8lz5M9pywRSxG9W7gY+7VX9k7/kbNf/68U/8ARlfRHjn/AJJ54i/7BVz/AOimr53/AGTv+Rs1/wD68U/9GUAfUtFFFABVfUP+QZdf9cX/AJGrFV9Q/wCQZdf9cX/kaAPk39lj/kq95/2CJf8A0bFX11XyL+yx/wAlXvP+wRL/AOjYq+uqACiiigAooooAwfHf/JOvEf8A2Crr/wBFNXzz+yb/AMjV4g/68o//AEOvobx3/wAk68R/9gq6/wDRTV88/sm/8jV4g/68o/8A0OgD6kooooAKral/yCrv/ri//oJqzVbUv+QVd/8AXF//AEE0AfFnwT8Z2fgPxVq2uX1vNciPSZI4oIVJMkjTRBVJ6KCe5/U4B+tvD/2jxl4HsZvHHh6C1uZws0unzqJVRlbKNg9DwDg8joa+av2W40k+Kl6siq6/2RIcMM8iaEg/nX13QAdOlFFFABRRXz3r/wAZvFmnftAR+ELZ7P8AsptWtbQhrfL+XIYw3zZ6/MeaAPoSiiigD5d/aosvD9nrely2lmq69fgzXdwJGyYUARAVzjkg8gZ+SvFtC03xIdQt73w5Yak91A4khmsrd2ZWHQgqK/QOTTrKW7F1LZ273AUKJmiUuAOg3Yzjk1YoA8N8K/tD/wCnaZ4Z8a+HtStPEcs8VpNtiVE3uVCuysQy53AkY+npXuVfJXjfRtUm/aqW7i028e2/tmxbzlgYpgCLJ3YxgYNfWtABRRRQAUUV8v8Axo8f+K9C+NJ0vR9evbOx22x+zxSYX5gM8e9AH1BRRRQAUUUUAFFfP/xw+MfirwD48h0nw+9mtq9hHORNBvbcXcHnPTCivfLdzJaxO33mQE/iKAJK4L4o/C2P4m2FtbT69e6clqSyQxKrws5/jZOCSBwPm459TXe0UAfImv8A7L/jPTdz6LcWGsxj7qpJ5Mp/4C/y/wDj1dp4J+Jt78Ifhzb6N8Q/D2rxT2t20NrtiGHiYFxhmIDYO4cE4G2voevBv2ptOvdR8OaAun2dxdMl3KWEETOVGwdcCgD1/wAIeJ7Xxl4Tsdf0+GaC3vVZkjnA3rhivOCR1X1rarg/gjbz2nwY8PQXUMkEyRSbo5FKsv71+oNd5QAUUUUAFFeTftGeI9Y8L/DqxvfD+oz6fcvqscTSwNtYoYpSV+mVB/CtL4D65qfiL4UWeoa5ezX129xMrTTNliA5AGfpQB6PRRRQAUUV438ffib4h+HTaAPDbWy/bxced58O/wC55e3HPH3zQB7JRXP+A9Zu/EPw/wBE1fUShur2zjmlKLtXcRk4HaugoAK8+PwM+Hkt9Pe3egfa7q4laWWW4u5nLuxySRvxyT6V6DRQBw9/4B+H/hbQb/Vj4O0qSLT7aS5cGzjkcqiliAX74HrVH4XfFXw94+kutM8NaTdaZFp0Sv5cscaIFYkYUIxxzXT+PY3m+HHiWKFGkkfSbpVRRksTC2AB3NeEfssaVqOneIPEDahYXVqr2sQUzwsgY7z0yKAPpaiiigAooqrqkjw6PeSRMVdIHZWHYhTg0AWqK+af2cvHfijxR8RL+y8Q65eahbR6VJKsU8m5Q4liAb64Yj8a+lqACiiigAoor578EfGbxZr3x2XwnqD2Z0w3l3CQlvh9saSlfmz6oKAPoSiiigAooooA8y1P46aDpfxKXwVNpuovfNew2YnRY/K3SFcH72cDeM8V6bXyV4s0bVJP2roryPTbx7b+3bJ/PWBim0GLJ3YxgYPNfWtABRRRQAUUV8v/ABg8f+K9E+OTaVpOvXtpYbrX/R4pMJ8yru498mgD6gooooAKKKKACiiigDz74j/GHRvhnf2VprFhfXT3kTSIbUIQoBxzuYV2mjanFreg2Gq26PHDfW0dzGkmNyq6hgDjvg186/tTaVqOo+I9AbT7C6ulS0kDGCFnCneOuBXvPgWN4fh14cimRo5E0q1VkYYKkRLkEdjQBvUUUUAFFFeF/tL+Ltf8KW/hw+HNWudONy9yJjbvt37RHjP0yfzoA90ork/hdqN5q/wt8P3+p3Elzd3ForyzSHLO2Tya6ygAooooAKK85+OHjfV/APgODVvD7QrdPfxwEzR712lHJ49cqKs/Bnxhqnjj4dxazrpha7e5ljJhj2LhTxxQB03i3/kSdc/7B9x/6Lavmv8AZQ/5HrWv+waP/Rq19KeLf+RJ1z/sH3H/AKLavmv9lD/keta/7Bo/9GrQB9V0UUUAFNl/1L/7pp1Nl/1L/wC6aAPkH9l//krkn/YMm/8AQkr7Ar4//Zf/AOSuSf8AYMm/9CSvsCgAooooAKKKKAMfxf8A8iPrv/YOuP8A0W1fNv7KH/I8a3/2Dh/6MWvpLxf/AMiPrv8A2Drj/wBFtXzb+yh/yPGt/wDYOH/oxaAPqqiiigApk3/HvJ/un+VPpk3/AB7yf7p/lQB8h/svf8lcl/7Bk3/ocdfX9fIH7L3/ACVyX/sGTf8AocdfX9ABRRRQAUUUUAY3jH/kRde/7Btx/wCimr5v/ZP/AOR21z/sHD/0YtfSHjH/AJEXXv8AsG3H/opq+b/2T/8Akdtc/wCwcP8A0YtAH1TRRRQAVHcf8esv+4f5VJUdx/x6y/7h/lQB8i/su/8AJW5/+wXN/wChx19fV8g/su/8lbn/AOwXN/6HHX19QAUUUUAFFFFAGL4y/wCRE17/ALBtx/6KavnH9k//AJHTXP8AsHr/AOjBX0d4y/5ETXv+wbcf+imr5x/ZP/5HTXP+wev/AKMFAH1RRRRQAVFdf8ec3/XNv5VLUV1/x5zf9c2/lQB8j/suf8lauP8AsFTf+hx19e18hfsuf8lauP8AsFTf+hx19e0AFFFFABRRRQBieNP+RC1//sGXP/opq+cv2T/+Ry13/rwX/wBGCvo3xp/yIWv/APYMuf8A0U1fOX7J/wDyOWu/9eC/+jBQB9UUUUUAFQ3f/HjP/wBc2/lU1Q3f/HjP/wBc2/lQB8k/st/8lZuf+wVN/wCjIq+vK+Q/2W/+Ss3P/YKm/wDRkVfXlABRRRQAUUUUAYnjb/kQPEH/AGDLn/0U1fOf7J3/ACOGvf8AXgn/AKMFfRnjb/kQPEH/AGDLn/0U1fOf7J3/ACOGvf8AXgn/AKMFAH1PRRRQAVDe/wDHhcf9cm/lU1Q3v/Hhcf8AXJv5UAfJP7LX/JWLr/sEzf8AoyKvryvkP9lr/krF1/2CZv8A0ZFX15QAUUUUAFFFFAGH43/5J94h/wCwXc/+imr51/ZO/wCRu17/AK8U/wDRlfRXjf8A5J94h/7Bdz/6KavnX9k7/kbte/68U/8ARlAH1NRRRQAVBff8g65/65N/I1PUF9/yDrn/AK5N/I0AfJf7LP8AyVi7/wCwTL/6Nir67r5E/ZZ/5Kxd/wDYJl/9GxV9d0AFFFFABRRRQBheOf8AknniL/sFXP8A6Kavnf8AZO/5GzX/APrxT/0ZX0R45/5J54i/7BVz/wCimr53/ZO/5GzX/wDrxT/0ZQB9S0UUUAFV9Q/5Bl1/1xf+RqxVfUP+QZdf9cX/AJGgD5N/ZY/5Kvef9giX/wBGxV9dV8i/ssf8lXvP+wRL/wCjYq+uqACiiigAooooAwfHf/JOvEf/AGCrr/0U1fPP7Jv/ACNXiD/ryj/9Dr6G8d/8k68R/wDYKuv/AEU1fPP7Jv8AyNXiD/ryj/8AQ6APqSiiigAqtqX/ACCrv/ri/wD6Cas1W1L/AJBV3/1xf/0E0AfJ/wCyx/yVa+/7BEv/AKNhr65r5G/ZY/5Ktff9giX/ANGw19c0AFFFFABXgOu/BPxRqXx6j8ZW8unDTF1W2vCrTsJdkZQt8u3GflOOa9+phniD7DKgbONu4ZoAfRRRQAUUUUAFFfOHiv4teM9M/aGHhey1ZY9IOqWlv9n+yxE+XII9w3Fd3O485719H0AFFFFABXjvxA+AjeOfiCfEw8RixBEQ+zfYvM+4APveYOuPSvYq5/VfHfhXQ9V/s3V9fsLO9+U/Z5pgr/N0496AOgooooAKKKKAPBfjX8FvE/xB8dQ6xoUunpbJYx25FzMyNuV3J4CnjDCvdrdDFbRRt1VApx7CleaKNsSSIp9GYCn0AFFFFABRRXkH7QnjzxF4E0PRrjwvfCzluriRJWMCSblCggfODj8KAPX6K4/4Ua9qPif4W6LrGtTi4vrqN2mlCKm4iRlHCgAcAdBXYUAFFFFAHEfFb4dH4m+FbfRhqY0zyL1Lrzjb+du2o67cbl/v5zntVz4beCT8PvBMHh83/wDaHkyySeeIfKzubONu5un1rZ1zxDpHhqxS81/UbfTrZ5BEstw4RS5BIXJ74Un8Kk0fWtN8Qact/ol9DfWjsVWaB9ykg4IzQBeooooAK8f+PHwt174ktoJ8PSWSf2eLjzvtUrJnf5e3GFOfuH9K9gpryxxY8x1TPTccZoAwvAui3XhzwFoujagY2ubGzjhlMTEqWUYOCQOK36RWDKCpBB6Ed6WgAooooAKKxvGF/c6V4H13UbCTy7q0064nhfaG2ukbMpweDyBwa8h/Z7+JXirx3rWs2/ijUlvI7W3jeJRbxx7SWIP3FGenegD3eiiigAqG8t/tdjPb7tnnRtHuxnGRjNTU2SRIomkkYKiAszHoAOpoA8l+FHwMb4ZeKrnWT4hGp+fZNa+SLLydu50bdne39zGMd69crB0Pxx4Y8S3z2ega7Y6jcpGZWit5g7BAQC2B2ywH41vUAFFFFABXgPgv4J+KPD/xyXxdfS6cdNF5dz7Y52Mm2VJAvG0DOXGefWvfqYJ4mfYsqFv7oYZoAfRRRQAUUUUAFFfOHiP4teM9P/aKj8LWmrKmjnV7S2Nv9liJ8tzHuG4ru53HnOa+j6ACiiigArx3x38BG8a/EY+Kh4jFkCYT9m+xeZ/qwB97zB1x6V7FXP6n478K6Nq39l6rr9haX/y/6PNMFf5vu8e+aAOgooooAKKKKACiiigAorwj9oT4l+KvAmuaNb+F9SWziuraR5VNvHJuYMAD86nH4V6/4Rv7jVfBOh6hfP5lzd6dbzzPtA3O0asxwOByT0oA2KKKKACvOPi78Jj8U4tJRdZGl/2c0pybXzvM37P9tcY2e/WvR6yNe8V6D4XWA+ItWtNNFwWEJuZQm/bjOM9cZH50AR+DfDv/AAifg3TNBNz9r+wQCHz/AC9m/Hfbk4/M1t1W0/ULTVdPhvtNuI7q1nXfFNE2VceoNWaACiiigDzz41+BdV+IPgWDSNCe2S5jvo7gm5kKLtVHB5APOWFWPg74N1PwH8PYtE1trd7pLiWUm3csuGORyQP5V3LyJGuZGVR6scUqOsi7kYMPUHNAGT4t/wCRJ1z/ALB9x/6Lavmv9lD/AJHrWv8AsGj/ANGrX0p4t/5EnXP+wfcf+i2r5r/ZQ/5HrWv+waP/AEatAHR/DbXtL8N/Gb4m6lrt9DY2cdywaWZsDPnNgDuT7DmvYvC3xG8JeNJ5IPDOtwXs8Q3NDteN8eu1wCR7gV4j8O/CGjeKv2g/HT+ILOO/hsbuWSO2nG6NnaVhuZTw2Bng8c1a1bQNM8J/ta+GIvDlpFp0F5beZJBbqEjyVlU4UcAEKOBRT1hTi+q/zCppKpJdH/kfRVNl/wBS/wDumvF7S18SfFLXvFd/beLtV0K00a9k0/TLbTpvLRpIxy8o/jBJHHvXafCvxVeeMfhlZapqu036iS3uWUYDujFS2O2QAfxojqr+j+TB6O3y+f8AX5Hzr+y//wAlck/7Bk3/AKElfYFfG/7N1s958Sr61iuJLZ59HuY1njOGiJKAMPcZzXs//CkvEv8A0V3xV/4ESf8Ax2gD2GivHv8AhSXiX/orvir/AMCJP/jtH/CkvEv/AEV3xV/4ESf/AB2gDo/jF451H4eeDLfW9Jhgnk+3xwyRTglXRlckcEEHgc1leBf2gfCPjDy7W+m/sPUmwPs9448tz/sS8A/Q7SfSvLvjX8N9Y8KeAk1DUvHuua/Cb2OL7Jfyu0YJVjuwXIyMenevKvBvw18U+PLgJ4e0ySS3DYe8l+SCP6uep9hk+1AH3D4uOfA+uEcj+zrj/wBFtXzb+yh/yPGt/wDYOH/oxa9U8M/DvU/h/wDC3XbTVPE13qxbTJwLU/8AHvb/ALtuIw2W/UA/3a8r/ZQ/5HjW/wDsHD/0YtAHRfDbXtL8N/Gb4m6lrt9DY2cdywaWZsDPnNgDuT7DmvYvC3xG8JeNJ5IPDOtwXs8Q3NDteN8eu1wCR7gV4j8O/CGjeKv2g/HT+ILOO/hsbuWSO2nG6NnaVhuZTw2Bng8c1a1bQNM8J/ta+GIvDlpFp0F5beZJBbqEjyVlU4UcAEKOBRT1hTi+q/zCppKpJdH/AJH0VTJv+PeT/dP8q8R+I+q/EXSvH/hhb7VtPtNAvNfiht4NLklSaWPeMCckYI2nBAO0ntXt03/HvJ/un+VC1jzf10/zB6St/XX/ACPkP9l7/krkv/YMm/8AQ46+v6+OP2brZ734k39rFcSW0k+jXMazxnDRElAGHuM5r2b/AIUl4l/6K74q/wDAiT/47QB7DRXj3/CkvEv/AEV3xV/4ESf/AB2j/hSXiX/orvir/wACJP8A47QB0fxj8daj8PPBltrekwwTyfb44ZIpwSroyuSOCCDwOf51leBf2gPCPjDy7W9m/sPUmwPs944Ebn0SXgH6HaT6V5d8bPhvrHhTwFHqGpePdc8QQm9jiFpfyu0YJVzvwXIyMY6dzXlXg34a+KfHlwF8PaZI9uGw95N+7gj+rnqfYZPtQB9weMDnwLrpHI/s24/9FNXzf+yf/wAjtrn/AGDh/wCjFr1Lw38PNT+H/wAK9etNU8TXmrFtLnAtm/497f8AdtxGGy36gH+7Xlv7J/8AyO2uf9g4f+jFoA3vhB4j0fwt41+Jeo+IdQhsLRdRC+ZKep82bgAck+wBNez+FfiD4V8amVfDGswX0kI3SRBWjkUeuxwGx74xXhvwg8G6J4o+LXjq58QWMWoJYX7+Tb3C74tzyyZYqeGIC4GfU1dh0LT/AAr+2Dptn4fto7C0urJpHt4BtQExPkBRwBlQcetENYwT6r9L/oE9JTa6P/I+iajuP+PWX/cP8q8q1PQvFXizxB4kv9b8S6z4R0fS2MemLYziBJVVSWnkP8a5HQ44re+E/iHVfFHwnstT15vMvHWWMz7dvnqrFVfHuBRHVfc/vB6M+fP2Xf8Akrc//YLm/wDQ46+vq+Qf2Xf+Stz/APYLm/8AQ46+vqACiiigAooooAxfGX/Iia9/2Dbj/wBFNXzj+yf/AMjprn/YPX/0YK+jvGX/ACImvf8AYNuP/RTV84/sn/8AI6a5/wBg9f8A0YKAN34QeI9H8LeNfiXqPiHUIbC0XUQvmSnqfNm4AHJPsATXs/hX4g+FfGplXwxrMF9JCN0kQVo5FHrscBse+MV4b8IPBuieKPi146ufEFjFqCWF+/k29wu+Lc8smWKnhiAuBn1NXYdC0/wr+2Dptn4fto7C0urJpHt4BtQExPkBRwBlQcetENYwT6r9L/oE9JTa6P8AyPomorr/AI85v+ubfyrxq0tfEnxS17xXf23i7VdCtNGvZNP0y206by0aSMcvKP4wSRx712Xww8VXnjH4WWuqartN+qy29yyjAd0JUtjtkAH8aI6q/o/kwejt8vn/AF+R89fsuf8AJWrj/sFTf+hx19e18c/s22r33xG1K0huJLWSfRLmNZ4jhoizRgMPcZyPpXsv/CkvEv8A0V3xV/4ESf8Ax2gD2GivHv8AhSXiX/orvir/AMCJP/jtH/CkvEv/AEV3xV/4ESf/AB2gDovjH471H4d+DrXW9KhguJDqEcEkU4O10ZHJGQQQflHP86zPAvx/8IeMfLtbyf8AsPUmwPs964COfRJeh+h2k+leWfG34cax4T8Bw6hqXjzXPEELX0cQtL+V2jBKud+C55GMdO5ryzwZ8NPFPjy4C+H9Md7fdh7yb93An1c9T7DJ9qAPt3xmc+AdfI/6Blz/AOimr5y/ZP8A+Ry13/rwX/0YK9R8P/D3Uvh/8JvEFpqfia81ctpVwBbN/wAe9v8Aum4jByw/MA/3RXl37J//ACOWu/8AXgv/AKMFAG78IPEej+FvGvxL1HxDqENhaLqIXzJT1PmzcADkn2AJr2fwr8QfCvjUyr4Y1mC+khG6SIK0cij12OA2PfGK8N+EHg3RPFHxa8dXPiCxi1BLC/fybe4XfFueWTLFTwxAXAz6mrsOhaf4V/bB02z8P20dhaXVk0j28A2oCYnyAo4AyoOPWiGsYJ9V+l/0CekptdH/AJH0TUN3/wAeM/8A1zb+VeO+MrfXRda9rvjPxtfeDrG3kaLQrWxvURbgKpId1GTIWP8AB1rufh/qms618KdN1DxNE0epz2jNNuTYWHIVivYlQD+NEdVf0/EHo7f1ofOX7Lf/ACVm5/7BU3/oyKvryvkP9lv/AJKzc/8AYKm/9GRV9eUAFFFFABRRRQBieNv+RA8Qf9gy5/8ARTV85/snf8jhr3/Xgn/owV9GeNv+RA8Qf9gy5/8ARTV85/snf8jhr3/Xgn/owUAbnwg8R6P4W8a/EvUfEOoQ2FouohfMlPU+bNwAOSfYAmvZ/CvxB8K+NTKvhjWYL6SEbpIgrRyKPXY4DY98Yrw34QeDdE8UfFrx1c+ILGLUEsL9/Jt7hd8W55ZMsVPDEBcDPqauw6Fp/hX9sHTbPw/bR2FpdWTSPbwDagJifICjgDKg49aIaxgn1X6X/QJ6Sm10f+R9E1De/wDHhcf9cm/lXjdpa+JPilr3iu/tvF2q6FaaNeyafpltp03lo0kY5eUfxgkjj3rsPhn4qvPGPwpt9U1Xab9Y5be5ZRgO6EqWx2yAD+NEdVf0fyYPR2+Xz/r8j59/Za/5Kxdf9gmb/wBGRV9eV8dfs2Wsl/8AEPVbSG5ktJLjQ7mJLiI4eIs0YDr7jOR9K9k/4Ul4l/6K74q/8CJP/jtAHsNFePf8KS8S/wDRXfFX/gRJ/wDHaP8AhSXiX/orvir/AMCJP/jtAHQ/GTx5qPw68H2etaVBBcO2oxwSxTg7XjZHJGQQQflHP86zvAvx+8IeMfLtruf+xNSfA+zXrgI59El6H6HaT6V5T8bvhzrHhPwJBqGpePNb8QQtfxxC0v5XaNSUc78FzyMEdO5ry/wZ8M/FXjy4C+H9Mdrbdh72b93An1c9T7Lk+1AH21435+HviHH/AEC7n/0U1fO37J3/ACN2vf8AXin/AKMr07RPh9qXgD4ReIrPU/E15q5bSbgC3f8A497f903EYOWH5gH+6K8x/ZO/5G7Xv+vFP/RlAG38IPEej+FvGvxL1HxDqENhaLqIXzJT1PmzcADkn2AJr2fwr8QfCvjUyr4Y1mC+khG6SIK0cij12OA2PfGK8N+EHg3RPFHxa8dXPiCxi1BLC/fybe4XfFueWTLFTwxAXAz6mrsOhaf4V/bB02z8P20dhaXVk0j28A2oCYnyAo4AyoOPWiGsYJ9V+l/0CekptdH/AJH0TUF9/wAg65/65N/I15P8afF+sHSdU0TwbdyWk+mWn27VdQgcq1sn/LOJWHIdzz7KD616DoU0tz8O9OnuJHllk0qJ3kdizOxiBJJPUk96I6pv+uv+QPRpf10/zPmL9ln/AJKxd/8AYJl/9GxV9d18ifss/wDJWLv/ALBMv/o2KvrugAooooAKKKKAMLxz/wAk88Rf9gq5/wDRTV87/snf8jZr/wD14p/6Mr6I8c/8k88Rf9gq5/8ARTV87/snf8jZr/8A14p/6MoA2vhB4j0fwt41+Jeo+IdQhsLRdRC+ZKep82bgAck+wBNez+FfiD4V8amVfDGswX0kI3SRBWjkUeuxwGx74xXhvwg8G6J4o+LXjq58QWMWoJYX7+Tb3C74tzyyZYqeGIC4GfU1dh0LT/Cv7YOm2fh+2jsLS6smke3gG1ATE+QFHAGVBx60Q1jBPqv0v+gT0lNro/8AI+iar6h/yDLr/ri/8jXz14q8TrqPxP8AFY1DW/HkNhpbxQxQeFZWCQBU/eSS/wAKjd+PWvZPCdxp118NLabRdYvdasntZDHfX8vmTyfezvYgfMDkdOMUR1jzf1qEtJWPm79lj/kq95/2CJf/AEbFX11XyL+yx/yVe8/7BEv/AKNir1p/gn4lZ2YfFzxUoJzgXEnH/kWgD2CivHv+FJeJf+iu+Kv/AAIk/wDjtH/CkvEv/RXfFX/gRJ/8doA7P4p+KL7wX8NdU8QaSsL3dmYCizqWRg06IwIBB6Me9cZ4F/aO8LeJvLtNf/4kGoNgfv3zbufaT+H/AIEB9TXH/Fb4W654c+GOrarffEbxBrMFv5O+xvJnaKXdMijcC5HBIYcdQK8N8J+BfEfja9+zeG9Lmu9pxJNjbFF/vOeB9M5PYUAfcfjeRJfhr4hkidXR9JuWVlOQwMLcg189/sm/8jV4g/68o/8A0Ou28NfCvWPh58LfEn9reJ7m88zSLrOmwnNrETE3I3DJPuNvuDXE/sm/8jV4g/68o/8A0OgDZ+EHiPR/C3jX4l6j4h1CGwtF1EL5kp6nzZuAByT7AE17P4V+IPhXxqZV8MazBfSQjdJEFaORR67HAbHvjFeG/CDwbonij4teOrnxBYxaglhfv5NvcLvi3PLJlip4YgLgZ9TV2HQtP8K/tg6bZ+H7aOwtLqyaR7eAbUBMT5AUcAZUHHrRDWME+q/S/wCgT0lNro/8j6JqtqX/ACCrv/ri/wD6Ca+f9c8dWviHxVr8niD4iah4RstPvHsNKtNNdgZXj+9LKEViy5x1wO2eK9f8K3Gp3fwyt5tc1Gx1S8ezcte6e5aG4GDtcEqvJGM8dc0R1jzf1qEtJWPnD9lj/kq19/2CJf8A0bDX1zXyN+yx/wAlWvv+wRL/AOjYa+uaACiiigAr498W/wDJ3kX/AGMFj/6FFX2FXBX3wa8J6j47XxdcxXZ1VbqO7DLcEJ5kZUr8vp8o4oA72iiigAooooA8M8R/AvXNZ+Ng8aQanp8dkNQtrryHL+Zti2ZHC4ydhxzUHxgdY/2hPh27sFVZEJYnAA87rXtcmu6TDffYptUso7rcE8hrhA+49BtznJyPzrwf476TBrvxr8C6VeFxb3m2GXY2CUaXBAPbikr+1hbv+jG7eznft/kepWvxi+H95rg0m28T2b3bP5ajDiNm6YEhGw/ga7Wvnv8AaI8A+F9A+GlrqGhaJZ6bdW95HEstrCI2dGVshiPvdAcnJ4rv73wgniXwzpGt6h4y8TaGkOkwmcaXqf2eIgJuaRxtOW5OT6AU01yt9nb8LiafMl3X62PRa+Pfj9/ycAf9y0/kK9p+CmmarOdS8S3Gv69f6LesYdKt9YvWuHMStzM2cAFiOABwPWu41XwF4V1zVv7T1jQLG8vsKPtE0QZ/l6c+1Nqwr3OhooopDCiiigD5D/ak/wCSs23/AGCof/RktfW1n/x4wf8AXNf5VxXjP4PeFPHuuJq3iCK6e6SBYAYbgoNoJI4+rGu5jQRxqi/dUAD8KAHUUUUAFeZfGr4Zal8TNJ0u00m9tLR7Od5HNzuwwKgcbQa9NqrfanYaYiPqV7bWiucKbiVYwx9snmgDx/xp4YufBv7J95oF9NFPcWUCq8kOdjZuQ3GQD0arfgT4m+DvB3wr8LWXiPXYLO6ewRhCEeRwDnBYICVH1xWt8cbqC8+BOvXFnPHcQvHEVkicMrfvkHBHBrB+F/ws8Iah8GrC41TRLW+vNSszLNd3EYeVSc42MeUwMAbcdKlNqM36fkymk3H5/oet6Tq+n69pcOo6NeQ3tnMMxzQtuVv/AK/t2q5Xzx+z1o7+IvhTr2jHV9T0yMaoClxptz5M0eFQkK2DgHHPFaI8J37fFmy8O+GvHfjG7j0zZea1Nfaw0kUaZykIUBcs/fJwB2NaNLnUe/8Alf8AAzu+Vvt/nYtftVf8kr07/sMxf+iZq1v2bv8Akith/wBfNx/6MNeia74d0fxNYpZ+INOt9QtkkEqxXCblDgEBseuGI/GpNH0TTPD2mrp+iWUNjaIxZYYF2qCTknFSUXqKKKACvmz9rj7/AIS+l5/7Qr6TrkfHXwz8O/EQ2J8SR3D/AGDzPJ8mYx437d2fX7goAT4Tf8kh8L/9g2H/ANBrr6oaHo9p4e0Kz0jTg62llCsMIdtzBQMDJ71foAKKKKAMvxPpcuueEtX0m3dI5b+xmtkd87VZ4yoJx2ya80+Cvwf1f4Z6rql1q2oWV2t7CkaC235UqxPO4D1r12WWOCF5ZnWOONSzu5wFA5JJ7CqtjrGmam7rpuo2l2yDLC3nWQqPfB4oA8A+EHiPR/C3jX4l6j4h1CGwtF1EL5kp6nzZuAByT7AE17P4V+IPhXxqZV8MazBfSQjdJEFaORR67HAbHvjFeG/CDwbonij4teOrnxBYxaglhfv5NvcLvi3PLJlip4YgLgZ9TV2HQtP8K/tg6bZ+H7aOwtLqyaR7eAbUBMT5AUcAZUHHrRDWME+q/S/6BPSU2uj/AMj6JqnrH/IDv/8Ar2k/9BNePfEbwj/wi2kteWvjvx5carqVyLfTdPj1vass7n5VA2cIOp9hivT/AAhod7ovgmw0jXtRm1e8jhK3V1cyNI0rMSWBZskgbtoz2AoWquD0dj5n/ZU/5KnqX/YGl/8AR8NfW1YOheBvC/hm+e88P6FZafcyRmJpbeIKxQkErn0yoP4VvUAFFFFABXx78L/+Trk/7CWo/wDouevsKuC0b4NeE9B8bjxXp8V2NUE0swZ7gsm6QMG+X6OaAO9ooooAKKKKAPDNc+Beuap8b08aw6np6WS6nbXnkMX8zbGUyPu4ydh71B8YHWP9oT4du7BVWRCWJwAPO617XJrukxX32KXVLJLvcE8hrhBJuPQbc5ycjivB/jvpMGu/GvwLpV4XFvebYZdjYJRpcEA9uKSv7WFu/wCjG7eznft/kepWvxi+H95rg0m28T2b3bP5ajDiNm6YEhGw/ga7Wvnv9ojwD4X0D4aWuoaFolnpt1b3kcSy2sIjZ0ZWyGI+90Bycniuh1nVdb8Sa14K8E6ZrV3pEd9o66jqV9ZvtuGQIAFV/wCElgcn3/CnH3lpve34X/IT0eu1r/jY9jr49+On/JxTf71n/wCgpXunw81HWNF+IniDwJrOr3Wsw2UEV7YXl626fynwGR2/iwSOfr9B1uqeAvCutav/AGrq2gWN3f8Ay/6RLEC/y/d59sCn0TXUOrR0NFefeM/htrPinxD/AGjp3j7XNBh8pY/sdjM6x5GctgOBk59O1dB418NXvivQBp2na/faBMJlk+12DlZMAH5chgcHPr2pAdDRXP8Agzw1eeFvD39m6jr99r03mtJ9svnLSYOMLksTgY9e9c/4K+G+s+FNfOo6j4+1zX4TC0f2S/ldo8kj5sFyMjHp3oA8+0H9pu3tPEN7pHjbT/KigupIY9QslJAVXIG+Pr0HJUn/AHa9w0TxBpPiXTUv9B1G3v7VukkDhsH0I6g+x5r4K1fTr3VvHmp2Wl2k15dS38wjhgjLux8w9AOa9y+E3wD8XaTqUOt6vrtx4b6E2thIGnkH91zygHsQ3uBQB2nxq+D2r/EzVtLu9J1CytEs4HjcXO/LEsDxtBqz8VdMl0T9mW60q4dJJbHTrS2d0ztZkaJSRntkV6VeavpullE1LUrW1dhlRcTpGWHryRXDfHWWOf4F69LC6yRyRQsjochgZUIIPcVnV+A0p/GjG8CfE3wd4O+Ffhay8R67BZ3T2CMIQjyOAc4LBASo+uK9T0nV9P17S4dR0a8hvbOYZjmhbcrf/X9u1eSfC/4WeENQ+DVhcapolrfXmpWZlmu7iMPKpOcbGPKYGANuOlc/+z1o7+IvhTr2jHV9T0yMaoClxptz5M0eFQkK2DgHHPFdE9ZzT3Wv42MIaQi1tt+B9D185ftb/wDHr4U/37v+UVa48J37fFmy8O+GvHfjG7j0zZea1Nfaw0kUaZykIUBcs/fJwB2Neva/4T0DxSsA8RaTa6kLcsYRcRhtm7GcfXA/Ko6Jl9bGF8Hf+SOeGf8ArxX+Zrtaradp1ppOnQ2Gm28dtaW67IoYxhUX0AqzSAKKKKAPF/2pf+ST2v8A2Fof/RctXf2af+SNW/8A1+z/APoQru/GngnR/HuhppPiBJntUnWcCGTYd4BA5+jGpPCHhDSvA/h9NG0FJUtEkaQCWTe2W680AS+Lf+RJ1z/sH3H/AKLavmv9lD/keta/7Bo/9GrX0p4t/wCRJ1z/ALB9x/6Lavmv9lD/AJHrWv8AsGj/ANGrQB7B4A+HGr+FfiZ4t8RahcWUlprcrPbpA7mRAZC3zgqAOD2Jo8R/DjV9X+OmgeNLa4sl07TbcRTRSO4mY/vPugKVI+cdWHevS6KFpy2+zsD15r/a3PFbS68R/C3XvFdhbeEdW1201i9fUNMudOh8xFkkHKSn+AAgc+1dp8K/Ct54O+GVlpmqbRfsJLi5VTkI7sWK574BA/Cu2psv+pf/AHTRHRW8kvkgerv8/n/X5nyD+y//AMlck/7Bk3/oSV9gV8f/ALL/APyVyT/sGTf+hJX2BQAUUUUAZev+GtI8U2UNn4gsY762hnWdYZc7S6ggEjuOTweK0Le3htLeOC1hjhhjUKkcahVUegA4AqSigDH8X/8AIj67/wBg64/9FtXzb+yh/wAjxrf/AGDh/wCjFr6S8X/8iPrv/YOuP/RbV82/sof8jxrf/YOH/oxaAPX/AAB8ONX8K/Ezxb4i1C4spLTW5We3SB3MiAyFvnBUAcHsTR4j+HGr6v8AHTQPGltcWS6dptuIpopHcTMf3n3QFKkfOOrDvXpdFC05bfZ2B681/tbnmHxh0nUdT1nwI+m6fdXiWuvwzXDW8LSCFARlmwPlX3PFemTf8e8n+6f5U+mTf8e8n+6f5ULRW87/AIJfoD1d/wCuv+Z8h/svf8lcl/7Bk3/ocdfX9fIH7L3/ACVyX/sGTf8AocdfX9ABRRRQBl6/4a0fxTZQ2fiCxjvraGdbhYZc7C6ggEjuPmPB4rQt7eG1t0gtYo4YY1CpHGoVVHoAOAKkooAxvGP/ACIuvf8AYNuP/RTV83/sn/8AI7a5/wBg4f8Aoxa+kPGP/Ii69/2Dbj/0U1fN/wCyf/yO2uf9g4f+jFoA9e+GHw41fwV4u8XapqtxZSwa3dCa3W2d2ZF3yN84ZQAcOOhPejUfhxq93+0Bp3jmO4shplraeQ8LO/nFtjrkDbtxlx/F616XRQtLeX+VgevN5ngPizXtV8SeOr+z8XeEvGt14WsJvLtNP0jS3MV8VPMkzkqWUkZCjjGPx9f8M6rBrHhJLi00TUNDgRGijsdQtBbSRqowP3YJwvp7Vv1Hcf8AHrL/ALh/lRHSNgerufIv7Lv/ACVuf/sFzf8AocdfX1fIP7Lv/JW5/wDsFzf+hx19fUAFFFFABRRRQBi+Mv8AkRNe/wCwbcf+imr5x/ZP/wCR01z/ALB6/wDowV9HeMv+RE17/sG3H/opq+cf2T/+R01z/sHr/wCjBQB678MPhxq/grxd4u1TVbiylg1u6E1uts7syLvkb5wygA4cdCe9Go/DjV7v9oDTvHMdxZDTLW08h4Wd/OLbHXIG3bjLj+L1r0uihaW8v8rA9ebzPFbS68R/C3XvFdhbeEdW1201i9fUNMudOh8xFkkHKSn+AAgc+1dl8MPCt54O+FlrpmqbRfsktxcqpyEdyWK574BA/Cu4qK6/485v+ubfyojoreSXyQPV3+fz/r8z5H/Zc/5K1cf9gqb/ANDjr69r5C/Zc/5K1cf9gqb/ANDjr69oAKKKKAMrX/DOj+KbOC08QWMd/bQTrcJDLnYXUEAkfxDDHg8VowQQ2tukFtEkMMahUjjUKqgdgBwBUlFAGJ40/wCRC1//ALBlz/6KavnL9k//AJHLXf8ArwX/ANGCvo3xp/yIWv8A/YMuf/RTV85fsn/8jlrv/Xgv/owUAeu/DD4cav4K8XeLtU1W4spYNbuhNbrbO7Mi75G+cMoAOHHQnvRqPw41e7/aA07xzHcWQ0y1tPIeFnfzi2x1yBt24y4/i9a9LooWlvL/ACsD15vM8Y1vUtctr7WdE+IXgTUPG1m1282jXVnp0U0aRMMKjEAeWR03fe5PtXVfCvQdZ8N/Ce307xHuS7VZXW3aTebeNiSse7vgfl0rvahu/wDjxn/65t/KiOit/Wn6+YS1Z8k/st/8lZuf+wVN/wCjIq+vK+Q/2W/+Ss3P/YKm/wDRkVfXlABRRRQAUUUUAYnjb/kQPEH/AGDLn/0U1fOf7J3/ACOGvf8AXgn/AKMFfRnjb/kQPEH/AGDLn/0U1fOf7J3/ACOGvf8AXgn/AKMFAHrnww+HGr+CvF3i7VNVuLKWDW7oTW62zuzIu+RvnDKADhx0J70aj8ONXu/2gNO8cx3FkNMtbTyHhZ384tsdcgbduMuP4vWvS6KFpby/ysD15vM8VtLrxH8Lde8V2Ft4R1bXbTWL19Q0y506HzEWSQcpKf4ACBz7V2Hwz8K3ng74UW+maptF+0ctxcqpyEdyWK574BA/Cu6qG9/48Lj/AK5N/KiOit5JfJA9Xf5/P+vzPkn9lr/krF1/2CZv/RkVfXlfIf7LX/JWLr/sEzf+jIq+vKACiiigDK1/wzo/im0gtfEFhHf21vOtwkMuSm8AgEjowwx4ORWjBBFa26QW0SQwxqFSONQqqB2AHQVJRQBh+N/+SfeIf+wXc/8Aopq+df2Tv+Ru17/rxT/0ZX0V43/5J94h/wCwXc/+imr51/ZO/wCRu17/AK8U/wDRlAHrfww+HGr+CvF3i7VNVuLKWDW7oTW62zuzIu+RvnDKADhx0J70aj8ONXu/2gNO8cx3FkNMtbTyHhZ384tsdcgbduMuP4vWvS6KFpby/wArA9ebzPCfF/wq8ZaV4N8UtpHjObULfUXlup9Ki0ON5btmP3PN3NITjAGPTgV6R4J03VtJ+GNna6/qD312tkDl7VbdoV8sYiKjuvTJ5OOa66oL7/kHXP8A1yb+Roj7qt6fgD1d/X8T5L/ZZ/5Kxd/9gmX/ANGxV9d18ifss/8AJWLv/sEy/wDo2KvrugAooooAKKKKAMLxz/yTzxF/2Crn/wBFNXzv+yd/yNmv/wDXin/oyvojxz/yTzxF/wBgq5/9FNXzv+yd/wAjZr//AF4p/wCjKAPWvhh8ONX8FeLvF2qarcWUsGt3Qmt1tndmRd8jfOGUAHDjoT3o1H4cavd/tAad45juLIaZa2nkPCzv5xbY65A27cZcfxetel0ULS3l/lYHrzeZ5O/ivX/B+t6/ZXfw2uLw3ly0tle+HbAGK7DdDcNuJD+rH349dn4Y+Fb/AMH/AAlTTdXVY711nuJoVORCXJbYCOOBjp3zXf1X1D/kGXX/AFxf+Rojovw+4Hq/xPk39lj/AJKvef8AYIl/9GxV9dV8i/ssf8lXvP8AsES/+jYq+uqACiiigDP13QdN8TaNNpOt2wurGdkMsLMQH2OHXOCDjco+tWLGws9Lso7PTbWG0tohiOGCMIiD2A4FWKKAMHx3/wAk68R/9gq6/wDRTV88/sm/8jV4g/68o/8A0Ovobx3/AMk68R/9gq6/9FNXzz+yb/yNXiD/AK8o/wD0OgD1n4YfDjV/BXi7xdqmq3FlLBrd0JrdbZ3ZkXfI3zhlABw46E96NR+HGr3f7QGneOY7iyGmWtp5Dws7+cW2OuQNu3GXH8XrXpdFC0t5f5WB683meGWNvqHwz8T+L7afwJqXiK2166e5s7vTrYTq6vk+TL/cUEnrnucV2Pwv8M6l4S+D8em62giu9k8zW4bcIA5LBM+wP5k16FVbUv8AkFXf/XF//QTRH3VbyS+SCWrv53+Z8n/ssf8AJVr7/sES/wDo2Gvrmvkb9lj/AJKtff8AYIl/9Gw19c0AFFFFABXnWofG7wxpvxDXwbcW+pHU2u4rMOkKGLfIVCnO/OPmGeK9Fr5C8WW07ftbxSCCQx/2/YncEOMbou9AH17RRRQAUUUUAfHvjv8A5O2X/sN2H8oa938dfDjV/E/xU8J+JbC4so7LRXVriOZ3Ej4k3fIApB49SK2r/wCE/gvU/F3/AAk97o5k1jz47j7T9qmH7xMbTtD7eNo4xjiuxoWklLsD1TXc4L4xeBtT+IXgQaLos9pBc/ao5t927Km1Q2eVVjnn0rL+KWna7B8F7Pw3otpcXl7dC1064NlE8uyMAeY/AyF+XGSBwa9RopWVmujaf3Du7p9kVtNsINK0q10+zQJBawpDGo7KoAH8q8n8f/Hv/hBviAfDP/CN/buIj9p+3eX98A/d8s9M+tew18vfGnwD4r1341HU9H0G+vLHbbD7RDEWT5QM8+1U227slJJWR9Q0UUUhhRRRQB5547+NXhn4e+IE0fXLfUpLl7dbgG1hRl2sWA5Ljn5T2r0GOQSxJIucMoYZ96+Sv2oLaeb4r2zRQyOv9lRDKoSP9ZJX1jZ/8eMH/XNf5UATUUUUAFfP37Wf/IteHf8Ar7l/9AFfQNc94u8CeHfHVrbW/inTzexWrl4l8+SPaxGCfkYZ/GgDzXwr4XvfGf7JlhoGlywRXV5bkRvcMyoNt0WOSoJ6A9q9I8D+Hrvwz8PNK0G/khkurO1EMjwsShbnoSAcfhWnoOhad4Z0O20fRLf7NY2oKwxb2faCSx5Yknknqa0KVtGu4dU+39foeZ/Bz4fav8NPDOrWutzWl1LcXRuYxYs7/LsAx8yqc8dBTvglpOo23hfUtZ8Q2VxZ6vrepTXdxHdRNHIq5wilWAIAAJH1r0qiqvrfyt/X3IHtbzv+f+ZxPxV+Iv8AwrLwrb6z/Zf9p+fera+T9o8nbuR23btrf3MYx3q38NvG3/CwvBUHiD+z/wCz/Olkj8jzvNxtbGd21ev0rkP2jfDuseJ/hzY2Xh/TrjULlNVjlaK3TcwQRSgtj0ywH41p/AfRNT8PfCez0/W7Kaxu0uJmaGZdrAFyQcUgPR6KKKACuL+IXxT0H4aHTx4ghvpf7Q8zyfskSvjZtzncwx98frXaV84ftaQTTP4T8mJ5MC8zsUnH+poA9+8P63beJPDtjrNgsi219Cs8QlUBgrDIyATz+NaNcj8KVZPhJ4YV1KsNOiBBGCPlrrqACiiigDnviD/yTPxP/wBgi7/9EtXz/wDsmf8AIx+Iv+vSL/0M19M6jYW2q6Xdaffx+ba3cLwTR7iu5GUqwyORkE8iue8I/Dbwp4FubmfwtpZsZbpAkzfaJZNwByB87HH4UAc78MPhxq/grxd4u1TVbiylg1u6E1uts7syLvkb5wygA4cdCe9Go/DjV7v9oDTvHMdxZDTLW08h4Wd/OLbHXIG3bjLj+L1r0uihaW8v8rA9ebzPMr3TNS139ouwuLywuho2g6W8ltcPCwhe5kIBw+NpYKRxnI216PeXH2SxnuNu/wAmNpNucZwM4zU1VdUjeXR7yOJSzvA6qo6klTgULRJf1vcHq7nl3wp+Of8Aws3xVc6N/wAI9/ZnkWTXXnfbfO3bXRdu3y1/v5zntXrdfNH7OPgXxR4Y+It/eeINDvdPtpNKkiWW4iKqXMsRC59cKT+FfS9ABRRRQAV51oXxu8MeIfHw8IWNvqS6iZ5oN8sKCLdErFuQ5OMIccelei18hfDG2nT9qpJHgkVP7S1A7ihA/wBXN3oA+vaKKKACiiigD498X/8AJ3cX/Yfsf5xV7v46+HGr+J/ip4T8S2FxZR2WiurXEczuJHxJu+QBSDx6kVtXnwn8F3/jAeKLvRzJrAuI7kXP2qYfvExtbaH28bRxjHFdjQtJKXYHqmu5wXxi8Dan8QvAg0XRZ7SC5+1Rzb7t2VNqhs8qrHPPpXP+JNC1zwh4o8IeLdO0m41xNL0z+ytStdOXfNt28PGpwW+bPH06Z49dopLTbvf8LfkD139Pxv8AmeY/DzT9X1r4ieIPHes6RdaNDewRWVhZ3q7Z/KTBZ3X+HJA4+v1OZ46+Pf8AwhXxGPhX/hG/tuDCPtX27y/9YAfueWemfXmvYa+XvjD4B8V618c21XSdBvruw3Wv+kRREp8qru59sGn0SXQOrZ9Q0UUUAFFFFAGPoPhLQfDHnnQtLt7SS5dpJ5VXMkrE5JZzyeT0zgdq2KKKAPl39rP/AJGbw9/15y/+hivV73wve+M/2cNN0DS5YIrq80axEb3DMqDasbHJUE9Ae1dF4t+GvhTx1dW1x4p0s3stqhSJvtEse1SckfIwz+NdDp9hb6Xplrp9jH5VraQpBCm4ttRQFUZPJwAOTSaTVmNNp3RieB/D134Z+HmlaDfyQyXVnaiGR4WJQtz0JAOPwrlvg58PtX+GnhnVrXW5rS6luLo3MYsWd/l2AY+ZVOeOgr0yiqbbcpdX/wAOSkkkuiPNfglpOo23hfUtZ8Q2VxZ6vrepTXdxHdRNHIq5wilWAIAAJH1p/wAXPi1/wqyLSX/sX+1f7RaUY+1eT5ezZ/sNnO/26V6PXhX7THhLX/FNv4bHh3SbrUjbvcmYW8e7ZuEeM/XB/Kh+X9WH3Z614O8Rf8JZ4N0zXvsv2T7fAJvI8zf5ee27Az+Qrbrk/hbp15pPwt8P2GpW8ltd29oqSwyDDI2TwRXWUgCiiigDmvHfjvSvh54fj1jXY7qS2kuFtwLVFdtzBiOCRxhT3qXwT4z03x74bTW9FjuI7V5WiC3KBXypweASP1rzr9qGKSb4U2qxIzt/asRwoyf9XLV39myN4vg7brKjI32yfhhg9RQB6D4t/wCRJ1z/ALB9x/6Lavmv9lD/AJHrWv8AsGj/ANGrX0p4t/5EnXP+wfcf+i2r5r/ZQ/5HrWv+waP/AEatAH1XRRRQAU2X/Uv/ALpp1Nl/1L/7poA+Qf2X/wDkrkn/AGDJv/Qkr7Ar4/8A2X/+SuSf9gyb/wBCSvsCgAooooAKKKKAMfxf/wAiPrv/AGDrj/0W1fNv7KH/ACPGt/8AYOH/AKMWvpLxf/yI+u/9g64/9FtXzb+yh/yPGt/9g4f+jFoA+qqKKKACmTf8e8n+6f5U+mTf8e8n+6f5UAfIf7L3/JXJf+wZN/6HHX1/XyB+y9/yVyX/ALBk3/ocdfX9ABRRRQAUUUUAY3jH/kRde/7Btx/6Kavm/wDZP/5HbXP+wcP/AEYtfSHjH/kRde/7Btx/6Kavm/8AZP8A+R21z/sHD/0YtAH1TRRRQAVHcf8AHrL/ALh/lUlR3H/HrL/uH+VAHyL+y7/yVuf/ALBc3/ocdfX1fIP7Lv8AyVuf/sFzf+hx19fUAFFFFABRRRQBi+Mv+RE17/sG3H/opq+cf2T/APkdNc/7B6/+jBX0d4y/5ETXv+wbcf8Aopq+cf2T/wDkdNc/7B6/+jBQB9UUUUUAFRXX/HnN/wBc2/lUtRXX/HnN/wBc2/lQB8j/ALLn/JWrj/sFTf8AocdfXtfIX7Ln/JWrj/sFTf8AocdfXtABRRRQAUUUUAYnjT/kQtf/AOwZc/8Aopq+cv2T/wDkctd/68F/9GCvo3xp/wAiFr//AGDLn/0U1fOX7J//ACOWu/8AXgv/AKMFAH1RRRRQAVDd/wDHjP8A9c2/lU1Q3f8Ax4z/APXNv5UAfJP7Lf8AyVm5/wCwVN/6Mir68r5D/Zb/AOSs3P8A2Cpv/RkVfXlABRRRQAUUUUAYnjb/AJEDxB/2DLn/ANFNXzn+yd/yOGvf9eCf+jBX0Z42/wCRA8Qf9gy5/wDRTV85/snf8jhr3/Xgn/owUAfU9FFFABUN7/x4XH/XJv5VNUN7/wAeFx/1yb+VAHyT+y1/yVi6/wCwTN/6Mir68r5D/Za/5Kxdf9gmb/0ZFX15QAUUUUAFFFFAGH43/wCSfeIf+wXc/wDopq+df2Tv+Ru17/rxT/0ZX0V43/5J94h/7Bdz/wCimr51/ZO/5G7Xv+vFP/RlAH1NRRRQAVBff8g65/65N/I1PUF9/wAg65/65N/I0AfJf7LP/JWLv/sEy/8Ao2KvruvkT9ln/krF3/2CZf8A0bFX13QAUUUUAFFFFAGF45/5J54i/wCwVc/+imr53/ZO/wCRs1//AK8U/wDRlfRHjn/knniL/sFXP/opq+d/2Tv+Rs1//rxT/wBGUAfUtFFFABVfUP8AkGXX/XF/5GrFV9Q/5Bl1/wBcX/kaAPk39lj/AJKvef8AYIl/9GxV9dV8i/ssf8lXvP8AsES/+jYq+uqACiiigAooooAwfHf/ACTrxH/2Crr/ANFNXzz+yb/yNXiD/ryj/wDQ6+hvHf8AyTrxH/2Crr/0U1fPP7Jv/I1eIP8Aryj/APQ6APqSiiigAqtqX/IKu/8Ari//AKCas1W1L/kFXf8A1xf/ANBNAHyf+yx/yVa+/wCwRL/6Nhr65r5G/ZY/5Ktff9giX/0bDX1zQAUUUUAFFFfPGv8Axj8Xaf8AtBR+Era4thpLata2hQ26l/LkMYYbuufmPNAH0PRRRQAUUUUAeX6x8ddG0b4mDwXPpV/JeG7htfPQp5e6Xbg8nOBvHbtXqFfJvjbQNZn/AGp1vodJvpLT+2LF/tCWzmPaBFk7sYwMHn2r6yoAKKKKACiivlz41eOvFOifGs6bpGv39nZbbY/Z4ZyqfMBngetAH1HRRRQAUUUUAFFfPnxy+MHizwH4+g0rw9cW0dq9hHORLbq53l3B5Psor3+3cyWsTt95kBP1xQBJRRRQAVxXxK+J2nfDLT7G71Sxurxb2Vo0W2K5UgZ53EV2teE/tSaVqGq+HdATTLC6vGS7kLrbwtIVGwcnAOKAPW/B3ie38Z+EbHxBZQS28F6rMkc2Ny4Yrzjj+GtuuE+CdrcWPwa8P297BLbzxxSB4pUKMv71zyDyK7ugAooooAKK8k/aP8Q6v4b+HNjeaBqNzp1y+qxxNLbSFGKGKUlcjtlQfwrT+Autalr/AMJ7O/1q+nvrt7iZWmncsxAcgDJoA9IooooAKKK8Z/aA+JfiL4eN4fHhqaCL7cLjzvOhEmdnlbcZ6ffNAHs1Fc94B1i78QfD7Q9X1Jla7vbOOaZkXaCxGTgdq6GgAooooAo63qkeh+H9Q1aeNpIrC1kuXRMbmVFLEDPfiuK+Gvxh0r4m319a6Xp15ZtZRrI5uSuGDEjjaT6V0vjyKS4+HPiSGCNpZZNKulREUlmJhYAADqa8K/Zb0XVNK8Qa++p6beWSyWsQRriBowx3npkDNAH0nRRRQAUUVV1SR4tHvJI2KukDsrDqCFPNAFqivmb9m/xv4n8S/Ea/s9f12+1G2TSpJViuZi6hxLEA2D3wxH419M0AFFFFABRRXzx4H+Mfi7XfjyvhXULi2bSze3kJRbdVbbGkpX5vqgoA+h6KKKACiiigDy/VPjro2l/E5fBUulX73jXsNn9oUp5e6Qrg9c4G8dq9Qr5N8V6BrMv7Vcd9FpN89n/btk/2hbZzHtBiyd2MYGDz7V9ZUAFFFFABRRXy58YvHXinRvjo2l6Vr9/aWO61/wBHhnKp8yru4980AfUdFFFABRRRQAUUUUAeefEr4w6V8MtQsbTVNOvLxr2JpEa2KYUA4wdxFdtoupx63oOn6rAjRxX1tHcoj/eUOoYA478188ftSaLquq+ItAfTNNvLxUtJA7W8DSBTvHBwDivd/A0UkHw78OQzxtHLHpVqro64ZSIlBBB6GgDdooooAKKK8I/aa8Va94Yt/DZ8Pavd6abh7kTG2lKb9oixnHXGT+dAHu9Fcl8LNQu9V+Fnh++1K4kurqezV5ZpW3M5yeSa62gAooooAKK83+OXjXWPAfgGDVfD0kUd09/HATLGHGwo5PB91FWfgt4u1Xxv8OYdY16SOS8a5ljLRRhBhTxwKAOo8W/8iTrn/YPuP/RbV81/sof8j1rX/YNH/o1a+lPFv/Ik65/2D7j/ANFtXzX+ygQPHWtZ/wCgb/7VWgD6ropNw9R+dG4eo/OgBabL/qX/AN00u4eo/OmysPJfkfdPegD5C/Zf/wCSuSf9gyb/ANCSvsCvj/8AZf8A+SuSf9gyb/0JK+wKACiiigAooooAx/F//Ij67/2Drj/0W1fNv7KH/I8a3/2Dh/6MWvpLxf8A8iPrv/YOuP8A0W1fNv7KBA8ca3n/AKBw/wDRi0AfVVFJuHqPzo3D1H50ALTJv+PeT/dP8qduHqPzpkzD7PJyPunv7UAfIn7L3/JXJf8AsGTf+hx19f18gfsvf8lcl/7Bk3/ocdfX9ABRRRQAUUUUAY3jH/kRde/7Btx/6Kavm/8AZP8A+R21z/sHD/0YtfSHjH/kRde/7Btx/wCimr5u/ZQIHjbXM/8AQOH/AKMWgD6qopNw9R+dG4eo/OgBajuP+PWX/cP8qfuHqPzqO4YfZZeR9w9/agD5G/Zd/wCStz/9gub/ANDjr6+r5B/Zd/5K3P8A9gub/wBDjr6+oAKKKKACiiigDF8Zf8iJr3/YNuP/AEU1fOP7J/8AyOmuf9g9f/Rgr6O8Zf8AIia9/wBg24/9FNXzh+ygQPGmuZ/6B6/+jBQB9U0Um4eo/OjcPUfnQAtRXX/HnN/1zb+VSbh6j86iumH2Obkf6tu/tQB8kfsuf8lauP8AsFTf+hx19e18hfsuf8lauP8AsFTf+hx19e0AFFFFABRRRQBieNP+RC1//sGXP/opq+cv2T/+Ry13/rwX/wBGCvo3xp/yIWv/APYMuf8A0U1fOX7J5A8Za7n/AJ8F/wDRgoA+qKKTcPUfnRuHqPzoAWobv/jxn/65t/Kpdw9R+dQ3bD7DPyP9W3f2oA+Sv2W/+Ss3P/YKm/8ARkVfXlfIf7Lf/JWbn/sFTf8AoyKvrygAooooAKKKKAMTxt/yIHiD/sGXP/opq+c/2Tv+Rw17/rwT/wBGCvozxt/yIHiD/sGXP/opq+c/2TyB4w17P/Pgv/owUAfU9FJuHqPzo3D1H50ALUN7/wAeFx/1yb+VS7h6j86hvWH2C45H+qbv7UAfJX7LX/JWLr/sEzf+jIq+vK+Q/wBlr/krF1/2CZv/AEZFX15QAUUUUAFFFFAGH43/AOSfeIf+wXc/+imr51/ZO/5G7Xv+vFP/AEZX0V43/wCSfeIf+wXc/wDopq+df2TyB4u17P8Az4p/6MFAH1NRSbh6j86Nw9R+dAC1Bff8g65/65N/I1NuHqPzqC+Yf2dc8j/VN39jQB8mfss/8lYu/wDsEy/+jYq+u6+RP2Wf+SsXf/YJl/8ARsVfXdABRRRQAUUUUAYXjn/knniL/sFXP/opq+d/2Tv+Rs1//rxT/wBGV9EeOf8AknniL/sFXP8A6Kavnf8AZOIHizX8/wDPin/oygD6lopNw9R+dG4eo/OgBar6h/yDLr/ri/8AI1PuHqPzqvqDD+zLrkf6l+/saAPk79lj/kq95/2CJf8A0bFX11XyL+yx/wAlXvP+wRL/AOjYq+uqACiiigAooooAwfHf/JOvEf8A2Crr/wBFNXzz+yb/AMjV4g/68o//AEOvobx3/wAk68R/9gq6/wDRTV88fsnEDxV4gz/z5R/+h0AfUtFJuHqPzo3D1H50ALVbUv8AkFXf/XF//QTVjcPUfnVbUmH9lXfI/wBS/f8A2TQB8ofssf8AJVr7/sES/wDo2Gvrmvkb9lj/AJKtff8AYIl/9Gw19c0AFFFFABXgmu/BDxNqXx4Txnb3WmDTV1S2vDG8ziXZGULDGzGflOOa97qFr21WbyWuYRLnGwyDdn0xQBNRRRQAUUUUAFFfNviz4reNNM/aKHhmx1rytH/tS0t/s32WE/u3Ee4bim7ncec55r6SoAKKTI9RSgg9KACvIPH3wETxz4/PiY+ImsSREPswsvM+4APveYOuPSvX65zVviD4T0LVv7L1jXrOzvvl/cSvhvm6ce9AHR0UUUAFFFFAHg/xo+CfiX4h+OIdY0O60yK2Sxjtyt1M6tuV3J4VCMfMO9e6W6GK2ijbGUQKcewpst5bW77J7mGJsZ2u4Bx+NTdelABRRRQAUUV49+0N458R+BtD0a48Laj9hlubmRJW8mOTcoUED51OPwoA9horjvhNruo+JfhZour63c/ar66jdpptipuIkZRwoAHAHQV2JIHU0AFFJkeopaAOK+Kfw7HxN8LW+jNqZ00QXi3XnCDzd21HXbjcv9/Oc9qt/DjwSPh94Lg8PrfnUBDLJJ55h8rO5s427j0+ta2veJdG8L2KXviDUYNPtnlESyztgFyCQv1wpP4VLouuaZ4i0xNR0O9ivrN2KrNCcqSDgj86AL9FFFABXkXx1+Feu/ExtCOgXFhD/Z4uPN+2Sumd/l4xtVs/cPp2r12oprqC22/aJ44t3Te4XP50AY3gfRLnw34D0bRr9onubG0jglaEkoWUYOCQDj8K3qRHWRA8bBlYZDKcg0tABRRRQAUVjeMb+50rwLr2oWEnlXVpptxPDJtDbHWJmU4OQcEDg14/+zx8R/FfjnWtag8U6r9uitbeN4V+zxR7SWIJ+RRn8aAPeaKKTcPUfnQAtQ3lv9rsZ7fds86No92M4yMZqamSypDE8krBERSzMewHU0AeU/Cv4GL8MvFNzrK+IDqZnsmtfJNn5W3c6Nuzvb+5jGO9es1z+g+PPC3ii/ey8Pa5aahcxxGVooHywQEAt9MsB+NdBQAUUUUAFeCeDPgh4m8PfG9fGF9daY+nC8u5/Limcy7ZUkVeCgGcuM8+te91Cl7ayTeUlzC0mSNgkBPHtQBNRRRQAUUUUAFFfNviT4reNLD9o2Pwxaa15ejHWLS2Nt9lhP7tzHuXcU3c7jznPNfSVABRSZHqKUEHpQAV5B45+AieNPiIfFJ8RNZEmE/ZhZeZ/qwB97zB1x6V6/XOap8QfCeiax/ZWra9Z2l/8v8Ao8j4b5vu8e+RQB0dFFFABRRRQAUUUUAFFeC/tDfEjxX4G1zRrfwtqv2GK5tpHlX7PFJuYMAD86nH4V7F4QvrnVPA+hahfSebdXenW880m0De7RqzHA4GST0oA2KKMgdTRkHoaACvOvi18Jl+KcWlI2snS/7OaU5Fr53mb9n+2uMbPfrXotYviDxh4f8ACi258R6tbacLksIfPfG/bjOPpkfnQAeD/Do8J+DtM0EXP2sWEAh8/wAvZvx325OPzNbVVtN1Kz1fTYNQ0y4S5tLhN8U0Zyrr6irNABRRRQB578afAmqfEPwPDo+hy2sVzHfR3Ba6dlXaqODyqk5yw7VY+D/gvUvAPw/i0TWZbaW6S4klLWzsyYY5HJAP6V28s8Vum+eVIlzjc7ADP40RTRTpvgkSROm5GBH6UAU9espdS8N6lY2+3zrm0lhj3HA3MhAz+Jr5N/4Zg8ff39J/8Cm/+Ir7BooA+Pv+GYPH39/Sf/Apv/iKP+GYPH39/Sf/AAKb/wCIr7BooA+Pv+GYPH39/Sf/AAKb/wCIo/4Zg8ff39J/8Cm/+Ir7BooA8A+CvwX8U+AfHr6xrrWBtWs5IB9nnLtuZlI4Kjj5TXv9FFABRRRQAUUUUAZ+v2UupeGtTsbbb511aSwx7jgbmQgZ9smvk7/hmDx9/f0n/wACm/8AiK+waKAPj7/hmDx9/f0n/wACm/8AiKP+GYPH39/Sf/Apv/iK+wa8D/aM+KF34bvdH0Dw9c+VfRzR6jdOp6BGzGh9QWBYj0C+tAHnX/DMHj7+/pP/AIFN/wDEUf8ADMHj7+/pP/gU3/xFfVHhHxLaeMPCWna9p5/c3sIcpnJjboyH3VgR+FbNAHgPwU+C/inwB49k1jXmsDatZSQD7POXbczKRwVHHymvfqKKACiiigAooooAzvEFjLqfhnVLC22+ddWcsMe44G5kIGT6ZNfJ/wDwzB4+/v6T/wCBTf8AxFfYNFAHx9/wzB4+/v6T/wCBTf8AxFH/AAzB4+/v6T/4FN/8RX2DXgX7RfxQu/Dd/o+g+Hrnyr2KaPUbplPQI2Y4z6gkFiPQL60Aed/8MwePv7+k/wDgU3/xFH/DMHj7+/pP/gU3/wARX1T4S8SWni/wnp2vaef3N7CJNucmNujIfdWBH4VsUAeA/BT4MeKfAHjyTV9eaxNq1lJAPs85dtzMhHBUcfKa9+oooAKKKKACiiigDO8Q2MuqeGNUsLbb511ZzQx7jgbmQqMn0ya+T/8AhmDx9/f0n/wKb/4ivsGigD4+/wCGYPH39/Sf/Apv/iKP+GYPH39/Sf8AwKb/AOIr7BooA+Pv+GYPH39/Sf8AwKb/AOIo/wCGYPH39/Sf/Apv/iK+waKAPAvgn8GPFPgDx5Lq+vNYm2exkgH2ecu25mQjgqOPlNe+0UUAFFFFABRRRQBm+IrCbVPC+q6fa7fOu7OaCPecDcyFRk+mTXyh/wAMwePv7+k/+BTf/EV9g0UAfH3/AAzB4+/v6T/4FN/8RR/wzB4+/v6T/wCBTf8AxFfYNeA/tFfFG78N6lo+g+Hrnyr2CaPUbplPQK2Y4z6gkFiPQL60Aeef8MwePv7+k/8AgU3/AMRR/wAMwePv7+k/+BTf/EV9VeFPEdp4u8Kadrunn9xewiTbnJRujIfdWBB+la9AHgXwS+DHij4f+O5tX15rE2z2MkA+zzl23M6EcFRxhTXvtFFABRRRQAUUUUAZviSwm1Xwrq2n2u3z7uymgj3nA3MhUZPpk18of8MwePv7+k/+BTf/ABFfYNFAHx9/wzB4+/v6T/4FN/8AEUf8MwePv7+k/wDgU3/xFfYNFAHx9/wzB4+/v6T/AOBTf/EUf8MwePv7+k/+BTf/ABFfYNFAHgfwS+DPij4f+O59X15rE2z2MluPs85dtxdCOCo4wpr3yiigAooooAKKKKAMzxLYTar4U1bTrXb593ZTQR7zgbnQqMn0ya+Uf+GYPH39/Sf/AAKb/wCIr7BooA+Pv+GYPH39/Sf/AAKb/wCIo/4Zg8ff39J/8Cm/+Ir7BrwD9on4pXfhvVtG0Lw9c+XeW00eo3bKePlOY429QSCxHoF9aAPPf+GYPH39/Sf/AAKb/wCIo/4Zg8ff39J/8Cm/+Ir6r8K+IrTxb4V0/XdOP7i+hEgXOSjdGQ+6sCD9K1qAPA/gj8GfFHw/8dT6vrzWJtpLCS3H2ecu24uhHBUcYU175RRQAUUUUAFFFFAGZ4m0+bVvCer6da7fPvLGaCPecDc8ZUZPpk18o/8ADMHj7+/pP/gU3/xFfYNFAHx9/wAMwePv7+k/+BTf/EUf8MwePv7+k/8AgU3/AMRX2DXgH7RHxTvPDesaNoXh658u8tZo9Ru2U8fKcxxt6g4LEem31oA89/4Zg8ff39J/8Cm/+Io/4Zg8ff39J/8AApv/AIivqvwt4itPFnhXT9d045t76ESBc5KHoyn3VgQfcVrUAeCfBH4NeKPh946uNW19rE20lhJbr9nnLtuLow4KjjCmve6KKACiiigAooooAy/E+nzav4R1fTbTb595YzQRbzhdzxlRk+mTXyl/wzB4+/v6T/4FN/8AEV9g0UAfH3/DMHj7+/pP/gU3/wARR/wzB4+/v6T/AOBTf/EV9g18/wD7RHxTvPDetaNoXh258u7tJo9Ruyp4+U5jjb1B5Yj02+tAHn3/AAzB4+/v6T/4FN/8RR/wzB4+/v6T/wCBTf8AxFfVfhfxDZ+K/C2n65pxzb30IkAzkoejKfdSCD7itagDwT4IfBrxR8PvHNzq2vtYm2l097dfs85dt5eNhwVHGFNe90UUAFFFFABXx74t/wCTvIv+xgsf/Qoq+wq4W9+D3hHUPHC+Lbm1uG1ZbqO6EguGC+ZGVKnb0x8o4oA7qiiigAooooA8I1f4L69qv7Qg8YTfY20X7dDc488iXEaKB8uP7y+te70UUAfLPjT9nfxtr/jrXNXsW0z7NfX81xF5lywbazkjI28HBr0r4DfDTXvhxaa3H4hNoWvnhaL7NKX4QPnOQMfeFdX8VPGyeAvh7f6urKLxl8iyU/xTMDt4745Y+ymuM/Zx8fv4p8Fy6Lqdw0up6O20NI2WlgYnY2T1wcr9AvrQB3/xE0C88UfDvWdF0zyxd3tv5cXmttXOQeTz6V81WX7M3jy3v7eaR9J2xyq7Yum6A5/uV9dUUAFFFFABRRRQB8h/tSf8lZtv+wVD/wCjJa+trP8A48YP+ua/yrjfGPwh8JeO9aTVfEVrcS3SQrAGjuGQbASRwPdjXbRoI41RfuqAB9KAHUUUUAFeU/Hb4c658RdG0m18PG1ElpcPJJ9plKDBUAYwDXq1FAHOfD/w23hH4faNoUiqs1naqs+xyymU/NIQT2Lsxrzz48fCvxD8R73RZfDxswtjHMsv2mYpyxTGMA5+6a9mrzz42+Ov+EF+HF1Nay7NSv8ANpZYPKsw+Zx/urk59dvrQB5N8OfgB4y8L/EPRtb1J9NNpZz+ZJ5VyWbG0jgbeetfTleRfs6eOv8AhKfh+NJvZd+o6JtgbceXhP8Aq2/AAr/wEeteu0AfLHjP9nbxtr3jnXNWsW0z7NfX89xF5lywba7lhkbeDg17d8H/AAhqXgf4c2uia2YTdxTSu3kOWXDMSOcCu5ooAKKKKACvmz9rj7/hL6Xn/tCvpOuT8b/DXw58QjZHxNbzTfYfM8jypmjxv27s46/cFADPhN/ySHwv/wBg2H/0Guvqjouj2nh/Q7PSdNRktLOJYYVZixCgYGSetXqACiiigDJ8V6bPrPgzWtLs9v2i90+e3i3nC73jZRk9hk15p8CPhTq/w5bWLjxF9lNzeCNIWtpi42DcSDkDuRXsNFAHM/Efw/eeKfh1rGiaX5Yu7yDy4vNbaudwPJwfSvmT/hmHx9nG/Sc/9fbf/EV9e3NxDaWstzdSLFDChkkkY4CKBkk+wFfM/wAPPjZNqfx9v5NRmZNH191tLaNzxBsJEB9ickH3fPagD6C8G6VcaF4G0PSb7Z9psdPgt5fLOV3pGFOD3GRXmXx4+FXiH4jX+izeHjZhbKKVZftMxTlipGMA5+6a9nooA8H+B3wc8T/D3xvearr7WJtptOe2X7POXbeZI2HBUcYQ17xRRQAUUUUAFfHvwv8A+Trk/wCwlqP/AKLnr7CrhdH+D3hHQvGg8Vada3CaoJZZhI1wzLukDBvlPHRzQB3VFFFABRRRQB4RqHwX17Uf2hl8Yz/Y20Zb+K6AE5Ev7tF2/Lj++o4z0r3eiigD5X8Yfs6+Ntc8ca5q1k2mfZr7UJ7mHzLlg2x5GYZG3g4Nen/Af4a678OdP1mHxCbUveyxPF9mlL8KGBzkDHUV6zRQBzHxI8P3vir4c6xoml+WLu8hCRea21c7geTg+lfN2nfs0ePLXVLW4lfStkUyO2Lps4DAn+CvriigAooooAKKKKACiiigDxT47fCfxH8RdZ0m68PNZiO0t3jk+0zFDksCMYBr17RtNh0bQbDS7ZdkNlbR28a7icKihQMnrwKu0UAeB/G74M+KPiB46g1fQWsRbR2EduftE5RtwdyeAp4wwqr8HPgl4r8C/ESHWtbbTzax28sTCCcu2WHHG0V7zrOrWmg6JearqUnl2tnC00reiqM8ep7AetfPPwN+L13rPxR1mw1+bEfiGZrm1VmysMqjAjHsY1C/8AX1oA+k6+Q7n9mTx7LdSyI+k7XdmGbpuhP+5X15RQBznw+0K78M/D3RtF1LyzdWVsIpfKbcuRnoe9dHRRQAUUUUAeL/ALUv/JJ7X/sLQ/8AouWrv7NP/JGrf/r9n/8AQhXeeMfBejeO9FTSvEUMktqkyzhY5Sh3gEDkezGpPCXhLSfBOgro+gRSRWayNIFkkLnc3Xk0AbdFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAFLWdWtdC0S81XUGKW1nC00hAycKM4A7k9APWvgLxZrmoeKfFOoa9qiOs99OZMEHCDoqD2CgAewr9Cq8e/aR/5E7w5/wBjFbf+i5aAOF/Ze8Zy2WpXng7Ud6w3WbqxLg4WQD50/FRu/wCAn1r6boooAKKKKACiiigAooooAKKKKAKWtata6Dod7quoMVtrOFppCBkkKM4A7k9APWvgLxXrmoeKfFOoa7qiOJ76cyEEHCDoqD2CgAewr9Cq8e/aQ/5FLw1/2MVv/wCi5aAOG/Ze8Zy2eoXng7Ud6w3WbqxLg4EgH7xPxUBv+At619NUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAFHW9XtdA0K91bUGK21nC00mBkkAZwB3J6Aepr4C8Va3qHijxRf67qiOLi+nMhBBwo/hQewAAHsK/QuvHv2jf+Ra8L/8AYxW//oElAHEfsveM5bS+vPB2o71iuM3ViXBwHA/eIPqoDD/db1r6YoooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAKGu6xa+HtBvdX1BittZQtNJjqQB0HqT0A9TXwF4o1rUPE/ia/1zVVcXF9OZGBBwufuqPYAAD2FfoZXj37RX/IC8Kf9jFb/APoL0AcV+y74zltby88G6jvWOfN1YlwcBwP3iD6gBgP9lvWvpaiigAooooAKKKKACiiigAooooAoa7rNr4e0C+1fUCRbWULTPjqQB0HuTwPc18BeJ9Z1DxN4mv8AXNVV/tF9OZGyDhc9FHsBgD2FfoZXj37RP/IG8Jf9jFb/APoL0AcV+y74zltrq88G6jvWObN3YFwcBwP3iD6gBgP9lvWvpaiigAooooAKKKKACiiigAooooAz9e1m18O+H77WNQJFvZQtM+OrYHQe5PA9zXwF4m1nUPEviW/1vVVf7TfztI2QcLnoo9gMAewr9Da8e/aI/wCQT4Q/7GK3/wDQWoA4v9l3xnLb3N74N1Lesc2buwLg8MB+8QfUAMB7N619K0UUAFFFFABRRRQAV5vqPxx8NaZ8Rl8Fz2WqtqTXkVmJUhjMO+QqFOTJnHzjPHrxXpFfIviuwvH/AGtIp1tJ2h/t+xbzBGduA0WTnpQB9dUUUUAFFFFABRRRQB8hftH+MpvE3jwaLY+Y2naJmIlQdrzn/WN+GAv4N61xfwv8WXXgPx/p+tLHKbTd5N4iqTvgbhvqRww91Ffcms2cmo6Ff2ULKslzbSQoX6AspAz7c1zvws8JXvgb4c6d4f1Sa3nurVpS8luWKHfKzjBYA9GHagDrYpUnhSWF1eORQyMpyGB5BFOoooAKKKKACiiigDznx78bvDfw78Qx6Nrdnqk9xJbrcBrSGNk2sWAGWdTn5T2r0SOQSxJIucOoYZ96+Tf2nbC8ufirbPbWs8yf2XEN0cZYZ8yT0r6utOLGDP8AzzX+VAE1FFFABRRRQAV8cfH3xBqnjH4hXCW1leNpOjhra3byG2MQf3kmcY5Ixn0Va+x6y/E+lSa94R1jSIJFil1CxntUkf7ql4yoJx2GaAPjL4Qa9q3gj4gWGriwvW06c/Z7zZA5DQsQCeBztIDf8BxX3CCGAIOQeQR3rm/h94an8HeAdK0C7njuJrGNkeWIEK2XZuM8966SgAooooAKKKKACuI+IvxW0P4ZHThr1rqE/wDaHm+V9ijR9vl7c7tzr/fGMZ7129fOX7WNnc3T+FPstvLNtF5u8tC2P9T1xQB754e1y28S+HLDWrBJY7a+gWeJZgA4VhkZAJGfxNaNcl8Ko3i+EvhlJUZHXTogysMEHb6V1tABRRRQAUUUUAeOftHeKL7TvBKeHtFt7ia61glZ2gjZvLgX7wJA43HA9xur5Mi0vVBdFYbG8E8RDEJC25PQ9Miv0Wrh9A8CXekfFzxL4ulu4ZLbWIYY4oFB3xlFQEnt/D2oAs/C7xXP4w+H2n6jqEMkOoRr5F4kiFD5q8FsHsww3/Asdq6+iigAooooAKKKKACvN9B+OPhrxD8QR4PsrLVU1AzzQeZLDGIt0SszciQnGEOOPTpXpFfIvwysLyL9qdJ5LSdIv7S1A+Y0ZC4Mc2OaAPrqiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAPA/2m/EuoNpNn4R0a1uZvtRFzfPDEzDYp+RMgd2BYj/ZX1r5r0y21qx1SK80y0vEu7KZZEeOFi0Tqcg9ODkV+iFcP4E8CXfhPxZ4w1a6u4Z49fvxdQpGCDEN0hw2e/wA46elAG54L8RjxZ4O07WTA9tLcwgzQOpUxSDh1wecBgceowa3KKKACiiigAooooA5nx7480v4deHo9Y1uC7nt5LhbcLaIrPuZWIOGZRj5T3qTwP4107x/4ZTXNGhuobV5XiCXSKr5U4PCswx+NefftPW81z8KrVLaGSZ/7ViO2NSxx5cvpVz9m+Ca3+D1vHcRPE/2yc7XUqeo7GgD1eiiigAooooAKKKKACiiigAooooAKKKKACiiigArx79pH/kTvDn/YxW3/AKLlr2GvHv2kf+RO8Of9jFbf+i5aAPYaKKKACiiigAooooAKKKKACiiigArx79pD/kUvDX/YxW//AKLlr2GvHv2kP+RS8Nf9jFb/APouWgD2GiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAK8e/aN/5Frwv/ANjFb/8AoElew149+0b/AMi14X/7GK3/APQJKAPYaKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigArx79or/kBeFP+xit//QXr2GvHv2iv+QF4U/7GK3/9BegD2GiiigAooooAKKKKACiiigAooooAK8e/aJ/5A3hL/sYrf/0F69hrx79on/kDeEv+xit//QXoA9hooooAKKKKACiiigAooooAKKKKACvHv2iP+QT4Q/7GK3/9BavYa8e/aI/5BPhD/sYrf/0FqAPYaKKKACiiigAooooAKKK+dfEHxg8X6f8AtCx+E7a9gXSW1e1tTEbZC3lyGMMN2M/xHmgD6KooooAKKKKACiiigCjrdxPaeH9QuLP/AI+IbWR4sLn5gpI4781zXwl1zWvEfwz0zVPE5Y6nM0wm3QiI/LK6r8oAx8oHaur1C8TTtNur2ZWaO2heZlXqQoJIHvxWT4K8W2XjnwlaeINMgngtrouEjuAA42OUOcEjqp70Ab1FFFABRRRQAUUUUAFFfPXx1+Lvi7wL4/g0rw7ewwWj2Ec5WS2SQ7y7gnJHoor6Bt3MlrE78syAn64oAkooooAKKKKACsjxdc31n4J1y60ff/aEGn3Elr5ab280RsUwuDk7gOMHNa9UNd1WPQfDupavNG0sen2kt08aHlwiFiB7nFAGL8NNQ1nVPhvo974n87+1ZomNz50IifdvYDKADHAHYV1NYvg/xNB4x8I2Gv2kElvDfIXWKUgsuGK8447VtUAFFFFABRRRQAUUV4v+0H8SPEnw+bw+PDF1Fb/bhc+f5kCybtnlbcZHH3zQB7RRXPeANXu9f+Huh6tqbrJd3llHNMyqFBYjJ4HSuhoAKKKKACiiigArz/w5rPie6+NXivTNSNz/AMI/awQtp++2Cx7iiFtsm0FuS3c4r0CuW0jx3aav8RNc8IxWc0dzo8Uckk7MNkgdVIAHX+KgDqaKKKACiiigAooooAKKK+dfA3xg8X658fF8LajewPpRvbyExrbIrbY0lKjcBnqgoA+iqKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAK8++Hes+J9S8ZeN7bxGbn7BZaiI9L862ES+Vuk+6wUbxgLySe3rXoNcr4T8d2nizxB4j0m1s5oJNAuxazPIwIlOXGVx2+Q9fWgDqqKKKACiiigAooooAKK81+OvjPWfAvgCDVfDs8cF29/HAzSRLINhRyRg+6irXwU8W6t41+G8Osa/Mk141zLGXSMINqnjgcUAegUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFePftI/wDIneHP+xitv/Rctew149+0j/yJ3hz/ALGK2/8ARctAHsNFFFABRRRQAUUUUAFFFFABRRRQAV49+0h/yKXhr/sYrf8A9Fy17DXj37SH/IpeGv8AsYrf/wBFy0Aew0UUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABXj37Rv8AyLXhf/sYrf8A9Akr2GvHv2jf+Ra8L/8AYxW//oElAHsNFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAV49+0V/wAgLwp/2MVv/wCgvXsNePftFf8AIC8Kf9jFb/8AoL0Aew0UUUAFFFFABRRRQAUUUUAFFFFABXj37RP/ACBvCX/YxW//AKC9ew149+0T/wAgbwl/2MVv/wCgvQB7DRRRQAUUUUAFFFFABRRRQAUUUUAFePftEf8AIJ8If9jFb/8AoLV7DXj37RH/ACCfCH/YxW//AKC1AHsNFFFABRRRQAUUUUAFeDa58DfEmp/HRPGkF9pS6cuqW14YnlkE2yMoWGBHjPyHHPpzXvNVm1OxS5+zve26z7gvlGVQ2T0GM5zQBZooooAKKKKACiiigCC9ggutPuLe8x9nliZJctt+QjB57cd6zfCWiaL4d8M22meFwo0yEuYdkxlHzOWb5iTn5ie9XdYs31HQr+yhZVkubaSFWboCykAn25rnfhb4SvPA3w607w/qc8E9zatKXktySh3ys4xkA9GHagDrqKKKACiiigAooooA8K+M3wQ8R/EXxvDrOi32l29vHZJblLuWRX3KzknCowx8w717jBGYraONsEogU49hUVxqVjaSeXd3tvA+M7ZJVU49cE1ZByMjpQAUUUUAFFFFABVPV7exvNEvrbWNn9nzW8kd15j7F8oqQ+WyMDaTzkYq5WX4n0qTXvCOsaRDIsUmoWM1qkjjIQvGVBPsM0AJ4ZsNG0zw3Z2fhfyf7JiQi28iYypt3EnDknPJPetWuc+H/hmbwd4C0vQLq4juJrGNkaWMEK2XZuM/WujoAKKKKACiiigAryT45fCnW/ia2hnQrvT7f+zhceb9skdd3meXjbtRv7hznHavW6gub60stv2y6ht9+dvmyBd2OuM/WgDJ8EaHc+GvAujaLfPFJcWFokEjwklGZRgkEgHH4Ct2mxypNGskLrIjDKspyCPY06gAooooAKKKKACuf03SPC9t421bUtM+zf8ACQXUaLqGy5LSbQFC7o9xC8BewzXQVxGg+BLrSPi34l8XS3kMlvrEMUccCqd8ZRUBJPT+GgDt6KKKACiiigAooooAK8G8G/A3xJ4d+Ni+Mb2+0qTTxd3U/lRSyGXbKkiqMGMDOXGefXrXvNVo9TsZbj7PFe27zZI8tZVLZHUYzntQBZooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigArnvDukeF9O1rXLnw39m+33lyJNU8m5MreblvvqWOw5LcYHf0roa4jwL4EuvCfivxfq1zeQ3EfiC+F1FHGpBiG6Q4bPU/OOnpQB29FFFABRRRQAUUUUAef/ABm8Ban8RfBEOjaLcWlvcR3qXBe7dlTaqOCMqrHPzDtVj4ReCtR8AeAItD1ie1nuUuJJS9qzMmGORyyqc/hXZ3F1b2kfmXc8cCE4DSOFGfTJpYLmC6i8y1mjmjzjfGwYfmKAJKKKKACiiigAooooAKKKKACiiigAooooAKKzfEd3Np/hbVby0fZPb2U0sTYB2sqEg4PB5FfIVl8d/izqMxh0/VpLuVV3FINMhdgOmcCPpyPzoA+z68k/aIs7m98I+Hks7ea4ZPEFu7LEhYqojl5OO3PWvGP+Fu/G3/npqH/gkj/+NUf8Ld+Nv/PTUP8AwSR//GqAPsKivj3/AIW78bf+emof+CSP/wCNUH4u/G0DJk1AAf8AUFj/APjVAH2FRXzl8CPix4y8a/ESTS/EmrLd2a2MkwjFrFH84ZADlVB7mvo2gAooooAKKKKACis3xJeT6f4V1a9tH2T29lNLE2AdrKhIODweRXx9D8f/AIpXDlLfXDKwGSqadAxx68R00m9EJtJXZ9p15J+0NZ3N74V8OpZ281wyeILd2WJCxVQknJx2968N/wCF5fFz/oJzf+CqH/43R/wvL4uf9BOb/wAFUP8A8bqvZz7Mj2kO6Ps2ivjL/heXxc/6Cc3/AIKof/jdB+OXxcHXU5v/AAVw/wDxuj2c+zD2kO6Ps2ivnP4D/Fjxl41+Icul+JNWW7s1sJJhGLWKP5wyAHKqD0Jr6MqDQKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAorL8TXk+neEtXvbN/LuLaxmmifAO1ljJBweDyK+RLL47/FnUZjDp+rSXcqruKQaZC7AdM4EfTkfnQB9n15J+0JZ3N74d8NLZ281wyeILd2ESFiq7JOTjoPevGP8Ahbvxt/56ah/4JI//AI1R/wALd+Nv/PTUP/BJH/8AGqAPsKivj3/hbvxt/wCemof+CSP/AONUh+L3xsVSWlvwAMknRY+P/IVAH2HRXzp8Bviv4x8bfEKfTPEuqrd2iafJMIxaxR4cOgByqg9GNfRdABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFZXim9n03wfrN9Zv5dxa2E80T4B2usbEHB4PI718i2Xx2+LOozGLT9Wku5FXcUg0yFyB0zgR9ORQB9n15J+0DZ3N5onhZbO2mnKeIIHcRRltq7X5OOg968Y/wCFu/G3/npqH/gkj/8AjVH/AAt342/89NQ/8Ekf/wAaoA+wqK+Pf+Fu/G3/AJ6ah/4JI/8A41SN8XvjYilmlv1UDJJ0WPj/AMhUAfYdFfOvwF+K3jHxt8QbjTPEuqreWiafJOsYtoo8OHjAOVUHoxr6KoAKKKKACiiigAorK8VX1xpng3Wb+yfy7m1sJ5on2g7XWNiDg8HkDrXyHbfHv4q3shjs9Za4cDcVi02BiB64EdJtJXYbn2jXkn7QFnc3mj+FRaW805j8QQO4iQttXa3Jx0HvXiX/AAuj4yf8/l3/AOCiL/43R/wuj4yf8/l3/wCCiL/43Ue1p/zL7yuWXY+yKK+N/wDhdHxk/wCfy7/8FEX/AMbpD8avjGqkte3QAGSTpEXH/kOj2tP+ZfeHLLsfZNFfO3wE+K3jHxv8QLnTPEuqreWkenSTrGLaKPDiSMA5RQejGvomtCQooooAKKKKACisnxXfXGmeDNav7J/LubXT55oX2g7XWNmU4PB5A618iW3x7+K17IUs9Za4cDJWLTYGIHrgR0m0ldhufaNeSfH+zubzS/CYtLeacx+IIHcRIW2rtbk46D3rxL/hdHxk/wCfy7/8FEX/AMbo/wCF0fGT/n8u/wDwURf/ABuo9rT/AJl95XLLsfZFFfG//C6PjJ/z+Xf/AIKIv/jdI3xq+MaqWa9ugAMknSIuP/IdHtaf8y+8OWXY+yaK+dvgH8VfGPjf4gXWmeJtVW8tI9OknWMW0UeHEkYByig9GPHvX0TWhIUUUUAFfHvi3/k7yL/sYLH/ANCir7CriL34Q+D9Q8ar4rurCZtXW5juhMLlwvmJjaducfwjigDt6KKKACiiigAor55+NXxz1bQPEjeGvBrz2OoWMwW6neCKVJgyKyhQwY/xegriNA+N/wAT5PEmmJq2oXElg13ELlBpkKlot43jOwY4zzkVUYSk7RVyJTjBXk7H1frc89r4f1C4s8/aIrWR4sLu+cKSOO/Paua+Euta34h+Geman4pLnVJmmE2+ERHiV1X5QBj5QO1Rf8LY0P8A59NQ/wC/af8AxdeGePfjZ8QbfxzqUXhW8uYNIV0+zRvp0TlRsXPzFGz82e5rpq4LEUlzTg0jlo4/C15ctOabPquivnT4SfGbxPd6pqK+P5ru7t1hU26x2UUW193PIC54+tep2fxf8IXOqHT7rUU02cJuAvpI492SAAPmJyc/oaj6rWVP2rj7vcv65QdX2KkubsdxRRRXOdQUUUUAfIf7Un/JWbb/ALBUP/oyWvraz/48YP8Armv8q5Dxd8JPCHjnWU1TxHYS3F2kKwB0uXjGwEkDCkDqxrs0QRxqicKoAH0oAdRRRQAUUV5/8XviTF8OfC4mEM73uoJNDZSRKpEMwTKswbqASOx+lAHoFZHi64v7TwTrlzo/mf2hDp9xJa+Wm9vNEbFMLg5O7HGDmvkqD44fF66hWa21GeaJvuvHpULKe3UR19EfDvx7dah8P9LuvFC3MmqyI5uH8hI8newHyjGOAOwrkr4zD4dc1WaXQ1hRqVHaKubfw0vta1L4b6Pd+KPO/taWJjc+fD5T7t7AZXAxxjtXU14r8aPiT4k0bSdLfwE11BPJO4uCLNJsrtGOCGxz9K808NfGL4r3PivSYNVurprGW9hS5X+yolzEXAcZ8sY4zzkU6eMw9Wn7SE00EqVSMuVrU+taK5XUviN4f0e2W41WdrOFnCLJOURS2CcZLdcA/lW9pWsadrunrfaNfW9/aMSqz28gdCQcEZHFPD4uhiU5UZcyQqlKdN2mrFyiiiukzCvmz9rj7/hL6Xn/ALQr6TrlvGnw48NfEA2R8T2clz9h3+RsnePbv27vukZ+4tAEPwm/5JD4X/7BsP8A6DXX1S0bSLPQNFtNK0yMx2lnEsMKMxYqoGAMnk1doAKKKKACiqWtapFoegahq1wjyRWFrJcuiY3MqKWIGe+BXyfr/wC0L471bxNey+Drye20tmBt7VrGGV4lwAcnYx65PU9aBpN6I+va8/8ADmreKLn41eK9O1P7V/wj9tBC2n77YLFuKIW2ybRu5LdzivOvg/8AF7xVqGp6kvj+W6uoEhQ24Wyji2tk55AXPH1r0TxH8TbWHwtqsukQ3qX6WUzWrmFG2yhDsONxz82OMGq5Wzphg8ROPPGDseg0V8af8Lv+L/8A0ELj/wAFUP8A8br3nwH8UzceAtNufFa3kuqGNjdSiCOMEh2wcAjHy47Ckot7EU8PWqu0ItnqVFc14V+IPhvxlJJDoWpwT3MQLNbiVDIEG0F9qk/LlgM+tdLSMpRcXZhRRRQSFfHvwv8A+Trk/wCwlqP/AKLnr7CriNI+EPg/Q/GI8UadYTR6qJZZhKbl2G6QMHO0nHIdqAO3ooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACivJ/jj8WJPh9pK6bpSTx6zqEPmWl0qI0cO11Dbg2c/Lnsa8Si+MXxpnhSaCe+kikUMjpo0RVgeQQRFyKAPsWvPvh3q3ijUPGXje38SfavsFnqIj0rz7YRr5W6T7jbRvGAvOT29a6zwrdXl94O0a71Td9tuLCCW53JsPmNGpbK4GOSeMcV5R8d/Gnj3wvrWkw+BmuRBPbu1x5NgtwNwYAZJRscUAe2UV80/DD4kfFXW/iVpGneJnvDpc8ji4EmlJEuBGxGWEYI5A719A634n0Lw2sJ8QavZaaLgsIjdzrH5mMZxk84yPzoA1aKr2N/aapYw3unXMV1azruimhcMjj1BHWrFABRRRQB4v8AtS/8kntf+wtD/wCi5au/s0/8kat/+v2f/wBCFd94u8G6L450ZNL8R273FokyzqiStGd4BAOVIPRjUnhXwnpPgvQ10jw/A8FmsjSBHkZzubryTmgDZooooAKKKKACiiigAooooAKKKKACiiigDH8X/wDIj67/ANg64/8ARbV82/sof8jxrf8A2Dh/6MWvpLxf/wAiPrv/AGDrj/0W1fNv7KH/ACPGt/8AYOH/AKMWgD6qooooAKZN/wAe8n+6f5U+mTf8e8n+6f5UAfIf7L3/ACVyX/sGTf8AocdfX9fIH7L3/JXJf+wZN/6HHX1/QAUUUUAFFFFAGN4x/wCRF17/ALBtx/6KavjD4Vf8hq+/69x/6EK+z/GP/Ii69/2Dbj/0U1fGHwq/5DV9/wBe4/8AQhXpZX/vlP1/Q8rOP9wqen6o9Sooor9DPy4KbJ/q2+hp1Nk/1bfQ0nsNbnM/su/8lbn/AOwXN/6HHX19XyD+y7/yVuf/ALBc3/ocdfX1flZ+yhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAYnjT/kQtf/7Blz/6KavnL9k//kctd/68F/8ARgr6N8af8iFr/wD2DLn/ANFNXzl+yf8A8jlrv/Xgv/owUAfVFFFFABUN3/x4z/8AXNv5VNUN3/x4z/8AXNv5UAfJP7Lf/JWbn/sFTf8AoyKvryvkP9lv/krNz/2Cpv8A0ZFX15QAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGH43/5J94h/wCwXc/+imr51/ZO/wCRu17/AK8U/wDRlfRXjf8A5J94h/7Bdz/6KavnX9k7/kbte/68U/8ARlAH1NRRRQAVBff8g65/65N/I1PUF9/yDrn/AK5N/I0AfJf7LP8AyVi7/wCwTL/6Nir67r5E/ZZ/5Kxd/wDYJl/9GxV9d0AFFFFABRRRQBheOf8AknniL/sFXP8A6KavlP4Af8jBq3/Xqv8A6HX1Z45/5J54i/7BVz/6KavlP4Af8jBq3/Xqv/odeLnv/Itq+i/NHXgv94ie70UUV+Rn1QVDef8AHjP/ANc2/lU1Q3n/AB4z/wDXNv5VUfiQnseO/ssf8lXvP+wRL/6Nir66r5F/ZY/5Kvef9giX/wBGxV9dV+6HxYUUUUAFFFFAGD47/wCSdeI/+wVdf+imr5V/Z/8A+Q9q/wD17J/6FX1V47/5J14j/wCwVdf+imr5V/Z//wCQ9q//AF7J/wChV4ue/wDItq+i/NHXgv8AeInu1FFFfkZ9UFQXv/Hhcf8AXJv5VPUF7/x4XH/XJv5VUfiQnseP/ssf8lWvv+wRL/6Nhr65r5G/ZY/5Ktff9giX/wBGw19c1+6HxYUUUUAFeaaj8c/D2mfElfBU+n6m2oNeQ2YmSOPyt8hUA5L5x84zxXpdfJHivStRk/axiuksLprf+3rJvOELFMBosnOMYoA+t6KKKACiiigD5i+Lnwv1dfiyfFJu7L7Hf3Mc0cW5/MURLGGBG3GT25qevQ/i/qFn9p0q3+1wechmDReYNyk+XgEZzzXnlfdZJShHCKcd5b/JtH5zxBWnPGuEto2t80mFFFFe2fPhXk3jP/kpUX+9B/SvWa3vAPg3w94i8YTS63pNveyJal1aVScMroAfwzXj5zSdTCO3TU93IaypY1J/aVj3iiiivgj9JCiiigDzb4gfG/w/8OvEUejaxYalcTyW63Ae1jjK7WZgB8zg5+U9q9HikEsKSLkB1DDPvXyh+03pWoXvxUtpbOxubiMaXEpeKFmGfMk4yBX1XaDFlADwfLX+VAE1FFFABXlnx1+HOqfEPw/p8ekXVnbnTpJJ5ftTMNy7Og2qeeO9ep1na7fWllo9yby6htw8LhPNkC7jtPAz1rKtJwpylHdJlQSckmeQ+C9DuPDfg+w0m8kjkntlYO8RJU5ctxkA963aZFLHPEJIJFkjboyNkH8RT6/E6tSVWpKpPdtt+rPsYxUYqK2QUUUVkUeY/Hr/AJEOz/7CUf8A6Klr0r9m7/kith/183H/AKMNVNX0TTdetFtdYs47uBHEipIMgMARn8ifzr1HRNC0vw5paadodlFZWaMWWGIYUEnJP51+kcK4iM8LKilrF3++/wDkfPZnTaqKff8AQ0KKKK+uPLCuG+I/xX0b4YnTRrdnfXP9o+b5X2NEbb5ezOdzL/fHTPeu5r50/au0+9vn8KfYbSe52C73eTEX258nGcDjoaAPefDuuW/iXw3Ya1ZRyx29/As8aTAB1VhkA4JGfxrSrk/hXFJB8J/DMU8bRyJp0QZHXBU7ehFdZQAUUUUAZfijS5tb8IaxpVs6JNfWM9tG0hO1WeMqCcZOMmvmrwt8NtV+HniDUbfV7qzuGaNFBtWYjON38Sjsa+p5JEijaSV1REBZmY4CgdSTXjni6+tL3xhfNZ3UNwP3fMUgb/lmvpWlNe8exk0IyxV30V1+C/UyKKKK6j7MKr33/IPuP+uTfyNWKa6q6MjjKsMEHuKAaurHD/sqf8lT1L/sDS/+j4a+tq83+EnhHQNHsbnU9M0q3tr1pHtzPGPmMeEbb9MgGvSK4mrOx+dV6To1HTe6CiiikYhXmmgfHPw94i+Ig8HWen6nHfmeaDzZY4xFuiV2Y5Dk4Ow449K9Lr5I+GelajD+1IlzNYXUcH9o6gfNeFguDHNg5xjnIoA+t6KKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigDxz42/B/W/iXq+l3Wi32n2yWcDxuLt3BJLA8bVNen+GdMl0TwlpGlXLo81jYw20jRk7WZIwpIzg4yKt3ep2FgyrfXtvbMwyomlVCR7ZNTxyJLGskTq6OAyspyGB6EGgB1FFFABXzl+1v/AMevhT/fu/5RV9G1i+IvB/h/xatuviTSrfURbFjCJ1zs3Yzj64H5UAYnwd/5I54Z/wCvFf5mu1qrpum2ej6bBp+mW6W1pbpsihjHyovoKtUAFFFFAHL/ABA8e6b8OvDses6xb3VxBJcLbhLVVLbmViD8zAY+U96l8C+NrDx/4YTXNJgube2eV4glyqh8qcH7pI/WuA/abtLm9+FdtFZ28txINUiYpEhY48uTnAq7+zja3Fn8ILeK7gkgk+2TnZKhU4yOxoA9VooooAKKKKACiiigAooooAKKKKACiiigDH8X/wDIj67/ANg64/8ARbV82/sof8jxrf8A2Dh/6MWvpLxf/wAiPrv/AGDrj/0W1fNv7KH/ACPGt/8AYOH/AKMWgD6qooooAKZN/wAe8n+6f5U+mTf8e8n+6f5UAfIf7L3/ACVyX/sGTf8AocdfX9fIH7L3/JXJf+wZN/6HHX1/QAUUUUAFFFFAGN4x/wCRF17/ALBtx/6KavjD4Vf8hq+/69x/6EK+z/GP/Ii69/2Dbj/0U1fGHwq/5DV9/wBe4/8AQhXpZX/vlP1/Q8rOP9wqen6o9Sooor9DPy4KbJ/q2+hp1Nk/1bfQ0nsNbnM/su/8lbn/AOwXN/6HHX19XyD+y7/yVuf/ALBc3/ocdfX1flZ+yhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAYnjT/kQtf/7Blz/6KavnL9k//kctd/68F/8ARgr6N8af8iFr/wD2DLn/ANFNXzl+yf8A8jlrv/Xgv/owUAfVFFFFABUN3/x4z/8AXNv5VNUN3/x4z/8AXNv5UAfJP7Lf/JWbn/sFTf8AoyKvryvkP9lv/krNz/2Cpv8A0ZFX15QAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGH43/5J94h/wCwXc/+imr51/ZO/wCRu17/AK8U/wDRlfRXjf8A5J94h/7Bdz/6KavnX9k7/kbte/68U/8ARlAH1NRRRQAVBff8g65/65N/I1PUF9/yDrn/AK5N/I0AfJf7LP8AyVi7/wCwTL/6Nir67r5E/ZZ/5Kxd/wDYJl/9GxV9d0AFFFFABRRRQBheOf8AknniL/sFXP8A6KavlP4Af8jBq3/Xqv8A6HX1Z45/5J54i/7BVz/6KavlP4Af8jBq3/Xqv/odeLnv/Itq+i/NHXgv94ie70UUV+Rn1QVDef8AHjP/ANc2/lU1Q3n/AB4z/wDXNv5VUfiQnseO/ssf8lXvP+wRL/6Nir66r5F/ZY/5Kvef9giX/wBGxV9dV+6HxYUUUUAFFFFAGD47/wCSdeI/+wVdf+imr5V/Z/8A+Q9q/wD17J/6FX1V47/5J14j/wCwVdf+imr5V/Z//wCQ9q//AF7J/wChV4ue/wDItq+i/NHXgv8AeInu1FFFfkZ9UFQXv/Hhcf8AXJv5VPUF7/x4XH/XJv5VUfiQnseP/ssf8lWvv+wRL/6Nhr65r5G/ZY/5Ktff9giX/wBGw19c1+6HxYUUUUAFFFfOXiH4ueMbD9oiPwra6lGmkNq9ramA20ZPluY9w3Fc87jzmgD6NooooAKKKKAPi34tf8nJaj/1/wBr/wCi4q7yun+L3w/8OL4osvEIsn/tK8kaWaXz3wzRiMIducDA9BXMV9rkVGUMO6j2lt8tD8/4jrxqYlU1vFa/OzCiiivoD5kKo6X8W08A+Omtn0Zr/wA6FYdwufLxvZTn7p6Yq9XnmveFde134ipJo2j3l8ieU7NBCXAUEAnjsM142dTnHCPk6vX0PeyCFOeNXP0V166H2zRRRXwZ+kBRRRQAUV87/Hf4s+MPA/xAg0vw3qMdtaPp8c5RraOQ7y7gnLKT0UV9CWztJaxO/LMgJPvigCSiiigAr5+/az/5Frw7/wBfcv8A6AK+ga5Tx/4H0HxtomzxHaNcrYrJNAFmePa23r8pGeneonNQi5PoOKu7HlPwq/5Jho3/AFzf/wBGNXX1S0jSbPQtJg03TIzFawAiNCxbAJJ6nnqTV2vxXFVY1sROpHaTb+9n2FOLhCMX0QUUUVzGhzPjvxivgjQodSayN6JblYPLEvl4yrNnOD/d/WvTvhz42X4g+DINfSwNgJpJI/IMvmY2tjO7A6/SvFfjPpGoaz4MtbfSbKe8mW/R2jgQsQvlyDOB2yR+del/ATRtS0H4TWdjrVjPY3aXEzNBcRlHALkg4NfpXC1OlHCSnH4m9fltp9589mUpOqk9lsek0UUV9YeWFFFeKftDfEbxL4Abw8PC98lr9uFz5+6BJN2zytv3gcffbpQB7XRXO/D/AFa8174d6FqupyCW8vLKOaZwoUMxGScDgV0VABRRRQBz3xB/5Jn4n/7BF3/6Javkf4M/8hTVP+uKf+hGvtDUtPt9W0q706+QyWt5C8EyBiNyOpVhkcjgmvCb3wLoHgrxVqEHhyza2jZY1IaZ5ONob+InuTWlNXkerlNNzxSa6a/p+o+iiiuo+2Co55fJt5JcZ2KWx64FSVDeKXsZ1UEs0bAAdzigT2NL4D/FVPFmrXvhxdIa1McD3/2g3G/OGij2bdo/vZznt0r3Cvl79mPwxrml/EHUNQ1HSby1s20uSATywlUMnmwttye+ATivqGuJ3b1PzqrOc6jlU3e4UUUUjIKKK+cvAvxc8Y638f08MalqUcukm9vITCLaNTtjSUoNwXPBRe/agD6NooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAPl39rP/kZvD3/XnL/6GK+gvAH/ACTXwz/2CLX/ANErVPxh8MvC3jy6trjxPp73ctqhjiK3Eke0E5P3SM10enWFvpWl2unWKGO2tIUghQsTtRVCqMnk8AUAWaKKKACvO/iz8WU+FsWlPJo7an/aLSgAXPleXs2f7LZzv9uleiV4R+034W13xNb+Gx4e0i81I273JlFrCZPL3CLGcdM4P5UAeu+D/EI8WeD9M15bY2ov4BMIS+/ZntuwM/lW1XJfCywu9L+Ffh+x1K2ktbqCzVJYZVKshyeCD0rraACiiigAorzT48eMda8D/D+DVPDdyttdvqEcDO0SyfIUckYYEdVFWfgl4r1fxn8NodX8Q3C3N61zLGZFjVBtU8cKAKAPQqKKKACiiigAooooAKKKKACiiigAooooAx/F/wDyI+u/9g64/wDRbV82/sof8jxrf/YOH/oxa+kvF/8AyI+u/wDYOuP/AEW1fNv7KH/I8a3/ANg4f+jFoA+qqKKKACmTf8e8n+6f5U+mTf8AHvJ/un+VAHyH+y9/yVyX/sGTf+hx19f18gfsvf8AJXJf+wZN/wChx19f0AFFFFABRRRQBjeMf+RF17/sG3H/AKKavjD4Vf8AIavv+vcf+hCvs/xj/wAiLr3/AGDbj/0U1fGHwq/5DV9/17j/ANCFellf++U/X9Dys4/3Cp6fqj1Kiiiv0M/Lgpsn+rb6GnU2T/Vt9DSew1ucz+y7/wAlbn/7Bc3/AKHHX19XyD+y7/yVuf8A7Bc3/ocdfX1flZ+yhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAYnjT/AJELX/8AsGXP/opq+cv2T/8Akctd/wCvBf8A0YK+jfGn/Iha/wD9gy5/9FNXzl+yf/yOWu/9eC/+jBQB9UUUUUAFQ3f/AB4z/wDXNv5VNUN3/wAeM/8A1zb+VAHyT+y3/wAlZuf+wVN/6Mir68r5D/Zb/wCSs3P/AGCpv/RkVfXlABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAYfjf/kn3iH/sF3P/AKKavnX9k7/kbte/68U/9GV9FeN/+SfeIf8AsF3P/opq+df2Tv8Akbte/wCvFP8A0ZQB9TUUUUAFQX3/ACDrn/rk38jU9QX3/IOuf+uTfyNAHyX+yz/yVi7/AOwTL/6Nir67r5E/ZZ/5Kxd/9gmX/wBGxV9d0AFFFFABRRRQBheOf+SeeIv+wVc/+imr5T+AH/Iwat/16r/6HX1Z45/5J54i/wCwVc/+imr5T+AH/Iwat/16r/6HXi57/wAi2r6L80deC/3iJ7vRRRX5GfVBUN5/x4z/APXNv5VNUN5/x4z/APXNv5VUfiQnseO/ssf8lXvP+wRL/wCjYq+uq+Rf2WP+Sr3n/YIl/wDRsVfXVfuh8WFFFFABRRRQBg+O/wDknXiP/sFXX/opq+Vf2f8A/kPav/17J/6FX1V47/5J14j/AOwVdf8Aopq+Vf2f/wDkPav/ANeyf+hV4ue/8i2r6L80deC/3iJ7tRRRX5GfVBUF7/x4XH/XJv5VPUF7/wAeFx/1yb+VVH4kJ7Hj/wCyx/yVa+/7BEv/AKNhr65r5G/ZY/5Ktff9giX/ANGw19c1+6HxYUUUUAFeFa38DNf1P45J41h1HTVsF1O2vDC7SebsjKEjGzGfkOOa91qm+saZHefZJNRtFudwTyWnUPuPQbc5zzQBcooooAKKKKAPn74u/FHSm8dW/hYWd59rsJvJklwvlsZRGVI+bOBnnisquM+LWnX3/DRF9c/Y7jyHvrYpL5TbWASLJBxg4rs6+0yGrOVBwltHb56nwHElGEMSpx3ktflZfkFFFFfQnzAV3Pwn/wCRruf+vJ//AENK4auA1vxh4g8PfERItE1a5skfykZYXwCrEEj8cV5Ob1VTwkr9dD2sjourjYtP4dT7Vooor4A/TAooooA8N+MfwP174jeNodZ0jUdOtoI7JLcpdNIG3Kzkn5VIx8wr26CMxW8cbYJRApx7Cq93rGm2EwivtQtLaQjcEmnVDj1wTVwEEZHINABRRRQAVwvxT+I+m/DzQoZNVtLu5Go+ZBH9mCnadvU7iOOa7qvBv2qLC8v/AA3oAsbSe5Md1Kz+TGX2jYOTjpUzipRcZbMabTujU8O65B4k8P2urWkckUNypZUlA3DDFecEjtWnXJ/DCGW3+GukRTxvFIsb7kdSCP3jdjXWV+KYunGniKkIbJtL0TPsKUnKnGT3aCiiiuY0CvUq+bfjLrWo6F4NtbnSLyWzna/SNpImwSpjkOPzA/KvS/gRrmp+IvhRZ6hrl7NfXj3EytNM2WIDkAflX6Jwph5QoTrN6SaX3X/zPAzOonOMO36no9FFFfZHkhXk3xw+FGsfE5tEOjXtja/2cJ/M+1s43eZ5eMbVP9w/pXrNVrzUrHT9n2+9t7XzM7PPlVN2OuMnnqKAMzwTodx4Z8DaPot5JHLPYWiQSPESVYqMEjIBx+FblMhmiuIUlgkSWNxlXRgVYeoI60+gAooooApa1qkWh6DqGrXKPJDYW0lzIkeNzKiliBnjOBXz9B8SdM+IfibULjSrS7tlVI2IuQoP3Qv8JPcV7f49jeb4ceJYoUaSR9JulVFGSxMLYAHc18m/CPT7yw1bU1vrSe2byUGJoinc+orSm/ePVymco4uKj10Z6rRRRXUfbBRRUN2zJYzshwyxsQR2OKAbsrnqfw2/5Fu4/wCvtv8A0BK6+vmH9mXxZr2r+PtQ03U9VubmyXTJLgQSPlRJ5sK7seuCRX09XHJ3dz88xVVVq0qi6hRRRUnOFeFeDvgZr/h341L4xu9R02SxF3dT+TE0nm4lSRVGCgGRvGefWvdapxaxpk939lg1G0kuMkeSk6l8jqNoOeMGgC5RRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAef/ABH+MGjfDO/srTWLC+unvImkQ2qoQoBxzuYV2ejanFreg2Gq26PHDfW0dzGkmNyq6hgDjvg186/tT6VqGo+I9AbT7C6ulS0kDGCFnCneOuBXvPgSN4fh14cjlRkkTSrVWRhgqREuQR2NAG9RRRQAUUV4X+0v4u1/wpb+HD4c1a5043L3ImMD7d+0R4z9Mn86APdKK5P4Xajeav8AC3QL/U7h7m7uLRXlmkOWdsnk11lABRRRQBwHxk8A6j8RvBMOjaRc2ttPHepcF7osF2qjgj5QTn5hVj4SeCb/AOH/AICi0PVbi3uLhLiSUvbFimGOR94A/pXYXd7a2EIlvrmG2jJ2h5pAgz6ZNLa3dtewiayuIriInAeJw65+ooAmooooAKKKKACiiigAooooAKKKKACiiigDH8X/APIj67/2Drj/ANFtXzb+yh/yPGt/9g4f+jFr6S8X/wDIj67/ANg64/8ARbV82/sof8jxrf8A2Dh/6MWgD1X43/DuTxLosfiXQG+zeItDHnwTIdrSop3Fc+oxlffI715x4cv9V/aM8Y6dHryJa6BoEMct7bRv/wAfMx4Jx1+Yg/7qg85Nd/8AGm78V681n4G8H6VqGzVGUahqgtnFvDET9wyY2+7c9OO+K5XXPhxq/wAHvFGh+KPhtY32rWqxraapYwI0skwx8z7VycNjPorAetKnZO72vp69/Tb56jndqy3t+Hb13+R6J4r8Y65Z+MLLwT8PdN06bU/sf2qaXUGZba1gB2qMJzknjjpx+FzwF42vPFVprOna9YR6fruizm2voIXLRtkEq6E87SAevpXIvr9h4c/aBXXfEEjaVpuv6Ai20+oKYBHIjAmN92NjYHQ46j1q38LZ013xz8QvFGnkvpV9cw29rOB8sxijIZl9RyOfeqjd79n96lb8iZWW3lb5q/8AXoeNfsvf8lcl/wCwZN/6HHX1/XyB+y9/yVyX/sGTf+hx19f0hhRRRQAUUUUAY3jH/kRde/7Btx/6KavjD4Vf8hq+/wCvcf8AoQr7P8Y/8iLr3/YNuP8A0U1fGHwq/wCQ1ff9e4/9CFellf8AvlP1/Q8rOP8AcKnp+qOj8caFJJEmvaWfKv7D5yw43oOfzH8sisfSTP8AEHX0vdRRY9O09V/0cNkO5/oSPyAFa3jD+1NZvoNA022uI7aVgbq7MTCPHXG7ofX3OBVG50S88G+Iba/8P2txd2EyiO5t4lLsMdTgfmD659a+gqx/2lzUX7K65vOXe3ZaX7vufMYWdsKqcpL2rT5PJdr9G9bdja8VeKbjQIJFtNLuJmVARcNEfITJxyw/lxW7bTNcaXFO4AaSEOQOmSuaxPHLiTwLeuAwDKhAYEEfOvUHpWvp/wDyA7X/AK9k/wDQRXq05T+sVIyldWTXlfm/yPGqRp/VITjGz5mn52S/zMP9l3/krc//AGC5v/Q46+vq+Qf2Xf8Akrc//YLm/wDQ46+vq/OD9VCiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAxPGn/Iha/8A9gy5/wDRTV85fsn/API5a7/14L/6MFfRvjT/AJELX/8AsGXP/opq+cv2T/8Akctd/wCvBf8A0YKAPS/jl4Bn1Kwh8beGX+y+IdBHn+YhwZok+Yg+68keoyO4rh/Cs2oftEePLXUfEMSW3h7w/FG0lij5E07DnI9GKn6KAOpJrs/jJJ4q8XapY+AfC+mahDY3rqdT1Y2ri3VOuzzMbSMDJweThfUVzuq+BNY+DPjzSvEXw70zUNW0e4jW11PT7aN5pDgDLYAJ5xuB6BgR0OKKeju9r6evf0/XUdS7VlvbX07ev6aHoXi7xpr1v4zs/BngHTtPuNWa0N5cTaizrb20IO0ZCfMST6dOPXiz4H8bXnirTtc07XrCPT9d0WVra+ghctG2VJV0J52kA9fSuWuNbsPC/wC0MdZ8QTjTdN17QUS2ub0GFVkRgSjFsbDgdDjqKm+GU6a740+IvijTyX0q+uIre1nA+WYxRkMy+o5HPvRDXfs/vUrfkTLTby/FXf8AXkeP/st/8lZuf+wVN/6Mir68r5D/AGW/+Ss3P/YKm/8ARkVfXlAwooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKAMPxv8A8k+8Q/8AYLuf/RTV86/snf8AI3a9/wBeKf8Aoyvorxv/AMk+8Q/9gu5/9FNXzr+yd/yN2vf9eKf+jKAPSfjl4Bn1Kwh8beGX+y+IdBHn+YhwZok+Yg+68keoyO4rh/Cs2oftEePLXUfEMSW3h7w/FG0lij5E07DnI9GKn6KAOpJrs/jJJ4q8XapY+AfC+mahDY3rqdT1Y2ri3VOuzzMbSMDJweThfUVzuq+BNY+DPjzSvEXw70zUNW0e4jW11PT7aN5pDgDLYAJ5xuB6BgR0OKKeju9r6evf0/XUdS7VlvbX07ev6aHf+J/GviM+OF8G/DzTdNn1C3tBd3lxqbOsFuhOFQBOdx4/Me+LvgjxrceMPDerx6tYrp+s6TNLZ39sjbkVwPvKf7p7fQ9etcnba/pvhT9oHWNQ8R3K6XY+INIt5rO4vh5K5QAMjFsbW68HB/OrHwocatefEHxNaBjpuq6iws5CCBMsaMC49jmnDVa9m/nzW/ImWm3dfjG/5nkH7LP/ACVi7/7BMv8A6Nir67r5E/ZZ/wCSsXf/AGCZf/RsVfXdIYUUUUAFFFFAGF45/wCSeeIv+wVc/wDopq+U/gB/yMGrf9eq/wDodfVnjn/knniL/sFXP/opq+U/gB/yMGrf9eq/+h14ue/8i2r6L80deC/3iJ2vxS8Iy6hZR+JNEbyNY0keaHU4Mka84+o6j8R3rldCmvPjH4qt7nWEWHR9IjRpLVWyJZSOePQkH6AY7103xJk17X7y18J6DY3aW92wN7qHkOIVTrt34x7nnngetY2o+E9R+G/ijTta8G2V3qFhIi299aQo0jtgctgeuM57MPQ4r5PBVOXBxpyklVal7N9l1V+l9eXseriI3qNxXu6c3n/w3U6/W/Eetf8ACTL4c8IWNnJdQ24uLie+LLDEucBQF5JNTeF/E8/iTRNUXULRLW/0+aS1uUjbchZR1U+lZ+t+K7C91O58N63cXPh6C4tY7i21L7T9neQEglfmUbCDkEE889Kz/hjIsWl+ItOsZFvNLtLhha34jCm4LKS2SPvEYHPv9K854aKwblKnZrld9dbvV321v8O6t5M2dT96kpaPT8O343MH9lj/AJKvef8AYIl/9GxV9dV8i/ssf8lXvP8AsES/+jYq+uq/Vz5kKKKKACiiigDB8d/8k68R/wDYKuv/AEU1fKv7P/8AyHtX/wCvZP8A0Kvqrx3/AMk68R/9gq6/9FNXyr+z/wD8h7V/+vZP/Qq8XPf+RbV9F+aOvBf7xE7T4peEZdQso/EmiN5GsaSPNDqcGSNecfUdR+I71yuhTXnxj8VW9zrCLDo+kRo0lqrZEspHPHoSD9AMd66b4kya9r95a+E9BsbtLe7YG91DyHEKp1278Y9zzzwPWsbUfCeo/DfxRp2teDbK71CwkRbe+tIUaR2wOWwPXGc9mHocV8ngqnLg405SSqtS9m+y6q/S+vL2PVxEb1G4r3dObz/4bqdP4u8V6rpHiKz0bQX0a23WjXM82rMyRRqGCqMqRjJzWxol1rF54cuZ9dn0md3VjDJpLu0TJt9W6nOenFcxqj+Crr4ifaPFKSxXTWCGFdWREtSh5+UNzvGSCG6HNN+HMcKx+LG0fP8AYTXTfYcZ2H5Tv257dK4KlCmsImo2aUW21vd9Hf8ApJm3tH7XfRu1r+Xb+tzB/ZY/5Ktff9giX/0bDX1zXyN+yx/yVa+/7BEv/o2Gvrmv1U+ZCiiigAr498W/8neRf9jBY/8AoUVfYVcXefCTwbf+Mh4qutLd9YW5juhP9qlA8xMbTtDbeNo4xigDtKKKKACiiigDzT4v/wDMH/7b/wDtOvJda0i31zSpbG6HyuMq2OUbswqt8W/H/iNvjZN4fN+v9mWd3DFDD5EfyrIkZcbtuTk+pqTxFqVzpmjyS6faTXd03yxJFEz4J7nA6CvsssrUZ5e4TV1G9/O7b0Pgc5w9aGZKpB6ytbysktTzoajrf2ceCSVFx5/k+fv/AOWXXb9Mc/Tiu3ne28E+F4obKDzn3rFEmcGaVu5P+fSuZPgS7Phn+0d8v9vb/tROTu9dn+93+vFbv/CUi38JR3+u6fLFdxuIxBPEU3ydmXcOB3z25owynRjP27cZWVm9bR7eq692XjHCu4KglKPN7yWl5d/R9OxNaa5rFpr1rpviC2tF+2qxgltGYgFRkqwatTTvhG3j/wAdNcjWhYeTEs202vm52MoxneOuaxNFNtf61FqWraxYXOo7SttaW06ssAPUDnLNjqa9N+H3izQNB8YTx6zq9pYu9oUVZ5QpLF0IHPrirxaVTATdR3s7xu9fK9uu5GCcqWY01SVm1aVlpv0v02177HutFFFfFH3wUUUUAfIf7Un/ACVm2/7BUP8A6Mlr62s/+PGD/rmv8q5PxZ8J/B3jfWE1TxJpb3V2kIhWRbmWPCAkgYVgOrGuxRBHGqIMKoAA9qAFooooAKq6p/yB7z/rg/8A6CatV47+0T431/wX4d0l/Dl6LRr2eSGcmFJN67Bx8wOOvasq0HOnKC6plRdpJmgRkYNeL+I0uvhF4xfW9GhWXR9VVle03bQkmCQPYA8j2JFeieA9Yu9Z8A6dqmqSma5mjdpXWMAth2HCqPQDoK4mHw9qPxL8bXWoeKbC9sNDskaKztbiNoXkJ/iwcH/aJ/3R2r8vy6P1atWjiGvZxuprvrZJba32fQ+lxDVSnFx3dreX9dTS+GHhuURTeM/EEiz6pqgMqOTkQxH09CR+QAHrT4/G3izVbKfXfD2h2d1okMrJHE7Obq6VTgsgXgd+CCfrWf4KtfEHhq81TwZqVneT6c6yfYNQWBjEu4HgsBgA5z14OR3rkdPsvDFv8N559T1Oa28T6eZI4YvtTJLBIrHaiR56E9TjueRXovDwrYipUqWndxUbJtcrvbRNW2S8u12cylKFOMVo9b621Xn5nq3jTwqfiD4Vs7UXMmmfvkuv3tuWZfkYbCuRg/P+leqfDXwSfh94Jg8Pm/8A7QMMsknniHys7mzjbuPT615zYa7Hpfg/S77xXdxWU00ESyvcMEzKUyR9eCcfWvY9I1rTdf05b/RL6C+tHYqs0DhlJBwRkV6nC8qqjVpS+GLVu3W9n12Ry5hyvkmt2tfwL1FFFfYHlhXzZ+1x9/wl9Lz/ANoV9J1zHjH4deGfHpsz4psGvPsW/wAjbPJHt37d33CM52r19KAK/wAJv+SQ+F/+wbD/AOg119U9H0mz0HRrTStLiMNnZxCKGMsW2qOAMnk/jVygAooooAK8i8cf8jlff9s//Ra16L4w1C50rwPruo2Enl3Vpp1xPC+0NtdI2ZTg8HkDg18v+AvHPiDxprmpT+I70XcixIwIhSPn7v8ACB2ArSm7SPYyeooYqz6q36/oT/EHw3JeWqa7pR8rU9OHmB14LovOPqOo/EVzukSXXxP8RQ3GposWmaail7dWyJJD/iR+QxW/46fWNZurfw5o9pcpBcMPtd75TCNV/u7sY9z+ArLvfDl94F8QWWqeF7W5vbKRRDd20SmRjgcnA9euex+taR0lfpf8e/oeri481dyjF8itz+fb1t1O317WI9C0kzLH5szERW1uvWRz0Ue39Kq+F9UufEvhGO8vViimuBIjCJTtGCV6En+dYmpN4gfxQ+qQ+Hn1G0igC2W+7SDyty5dijc7u3OMYqX4Y3NyfB6rdWnkW0LOYpzKG83LMW+Ucrjpz1q4ttu52e3csSoaqNmtnra2t7W/rzO++AnwqPhPWL3xIdYF2JLd7D7P9m2Yy0Um/duP93GMd+te51538JvFOharp9zp2m6ta3V4srzmCKUM4jwi7semSBmvRK5pWT0Pj8XCEK8o09r6BRRRUnMFfHvwv/5OuT/sJaj/AOi56+wq4vSvhJ4N0TxcPE+m6W8WrCWSYTm6lYbpAwc7S23kO3bvQB2lFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAVheMvCWneN/Ct3oerJmG4X5JAPmhcfddfcH+o715L+0J8S/FXgTXNGt/C+pLZxXVtI8qm3jk3MGAB+dTiu41PxlrWm/BnTNdsNLvNb12+063aOKztGlzNJEGLsqD5VBJPYdh1qZ25dSo3voeDjxJ43sdPf4IgxjUGvRZpfGXGLZhnZn+6Qc+u0lcV7bevp3wM+FdpZ6JZfbrtpktreLO1ry6k/iY/gT9ABXmg+A2sXnwzk8SXM1z/wnss/9pglyHHfyvZ/4s9mwK1fFPiHW9e+FvhnxD4m0PUdOvfDuuWs+qRXNm8W9FyDMoYDKnIzjoSavW1pPW6u/LRfhr89SNL3itLOy89X+Ony0O20Lxz4t07x1p/hn4i6ZpdvJq8LyWF1pUjmPegy0Th8ndjuOOnXPDvi78Jm+KcWkousjS/7OaU5Nr53mb9n+2uMbPfrWJr2v6V46+NPgWDwpfwamml/aL68ntXDpChQBQzDgEkYx15Fema94r0HwusB8RataaaLgsITcyhN+3GcZ64yPzo3im/P8/6XyD7Vl5f1+vzI/Bvh0+E/BumaCbn7WbCAQ+f5ezfjvtycfma26r6fqFpqunw32m3Ed1azrvimibcrj1BqxSGFFFFAHi/7Uv8AySe1/wCwtD/6Llq7+zT/AMkat/8Ar9n/APQhXf8AizwfonjfR00vxLaNdWiTCdY1lePDgEA5Ug9GNSeF/CukeDdEXSfD1sbWyV2kEZlaT5m6nLEmgDYooooAKKKKACiiigAooooAKKKKACiiigDH8X/8iPrv/YOuP/RbV82/sof8jxrf/YOH/oxa+kvF/wDyI+u/9g64/wDRbV82/sof8jxrf/YOH/oxaAPqqiiigCnqej6ZrdqLbWdOtNQgDbhFdwLKufXDAjNPis7aw037LYW8VtbxIRHDCgREHoAOBVmmTf8AHvJ/un+VAHyH+y9/yVyX/sGTf+hx19f18gfsvf8AJXJf+wZN/wChx19f0AFFFFABRRRQBjeMf+RF17/sG3H/AKKavjD4Vf8AIavv+vcf+hCvs/xj/wAiLr3/AGDbj/0U1fGHwq/5DV9/17j/ANCFellf++U/X9Dys4/3Cp6fqj1Kiiiv0M/LiO4t4buBobqGOaJvvRyKGU/UGlKrHblEUKqrgKBgAY6U+myf6tvoamy3KTexzP7Lv/JW5/8AsFzf+hx19fV8g/su/wDJW5/+wXN/6HHX19X5YfsgUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGJ40/5ELX/APsGXP8A6KavnL9k/wD5HLXf+vBf/Rgr6N8af8iFr/8A2DLn/wBFNXzl+yf/AMjlrv8A14L/AOjBQB9UUUUUAU9T0fTNbtRbazp1pqEAbcIruBZVz64YEZo+x21ho72thbxW1vFEwjhhQIiDB4AHAq5UN3/x4z/9c2/lQB8k/st/8lZuf+wVN/6Mir68r5D/AGW/+Ss3P/YKm/8ARkVfXlABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAYfjf8A5J94h/7Bdz/6KavnX9k7/kbte/68U/8ARlfRXjf/AJJ94h/7Bdz/AOimr51/ZO/5G7Xv+vFP/RlAH1NRRRQBR1TRNK1yBYNa0yz1GJG3LHd26yqp9QGB5p01tBZ6LLb2kMcEEcDKkUSBVQYPAA4Aq5UF9/yDrn/rk38jQB8l/ss/8lYu/wDsEy/+jYq+u6+RP2Wf+SsXf/YJl/8ARsVfXdABRRRQAUUUUAYXjn/knniL/sFXP/opq+U/gB/yMGrf9eq/+h19WeOf+SeeIv8AsFXP/opq+U/gB/yMGrf9eq/+h14ue/8AItq+i/NHXgv94ie70UUV+Rn1RUv9L0/VI1j1OxtrxFOVW4hWQA+wINEltBaaVLBaQxwQpEwWONAqrx2Aq3UN5/x4z/8AXNv5VpGcnaN9CWlueO/ssf8AJV7z/sES/wDo2KvrqvkX9lj/AJKvef8AYIl/9GxV9dV+5HxgUUUUAFFFFAGD47/5J14j/wCwVdf+imr5V/Z//wCQ9q//AF7J/wChV9VeO/8AknXiP/sFXX/opq+Vf2f/APkPav8A9eyf+hV4ue/8i2r6L80deC/3iJ7tRRRX5GfVFS/0rT9VjWPU7C2vEU5VbiFZAPoCDRNBDa6TLDbRJDEkLBI41CqoweAB0q3UF7/x4XH/AFyb+VaRlJ2jfQlpbnj/AOyx/wAlWvv+wRL/AOjYa+ua+Rv2WP8Akq19/wBgiX/0bDX1zX7kfGBRRRQAV5jqXx00LTPiYvgqbTNQe+a9hsxOoTyt0hUA/ezgbxnivTq+S/FWiarJ+1dFeR6ZePa/29ZP5627lNoaLJ3YxgYPNAH1pRRRQAUUUUAfMnxc+F2qL8V/+EpN7afZb+5SaOLLeYoiWMEHjGTjjmpa9B+L+qWH2vSrX7dbfaIzKHh85d6k+XgEZyM159X3WSU4RwinHd7/AHtH5zxBVqTxrhLaNrfNJv8AEKgu7G0v41S+tYblFOQs0YcA+uDU9Fey4qSszwYycXdMpQaJpVrMs1tplnDKnKvHborL9CBXmfjP/kpUX+9B/SvWa6D4f+EfD+v+MJ5da0i1vXS0Lq00e4hg6AH8M14+b4dSwb5NLO572R4lxxy57vmVj3aiiivhD9GCiiigDzP4hfHDQ/hz4kj0bVdN1C6mktluA9sE27WZhj5mBz8pr0qKQSwpIBgOoYA+9fKX7TGi6pqHxStprDTby6iGlxKXggZ1B8yTjIHXmvqm0BFlACMERrkfhQBNRRRQAV5d8cvhxqPxE8P2Eel3lranTpJJ5DcFvmGzoNoPPFeo1meINRsdO0e5OoXlvaiSF1Qzyqm47TwMnmsq0nGnKUd0mVBJySZ5J4N0Ofw34QsNJu5I5ZrZWVnjztOXLcZ+tblRwzxXMKy28qSxt9142DKfoRUlfidWpKpUlOe7bb9WfYxioxUVsgqjLoekzagL6bTLOS8Ugi4e3QyD/gWM1eoqIylH4XYbSe55j8ev+RDs/wDsJR/+ipa9K/Zu/wCSK2H/AF83H/ow1W1XRtO1y1W21ezhvIVcSLHKuQGAIz9cE/nXqOjaJpnh7TV0/RLKGxtEYssMC7VBJyTiv0bhXERnhZUUtYu/33/yPAzOm1UU+/6F6iiivrjygrhPiV8WNJ+GB00avY3l3/aPm+X9lCfL5ezOdzD++PyNd3Xzv+1Xpeoak/hX+zrG5u/LF3v8iFn258nGcDjofyoA918N65B4m8M6frdpHJFBfwLPGkuNyhhnBxkZrTrk/hbBLbfCjw1DcRPFLHp0SukilWU7ehB6V1lABRRRQBl+J9Ll1zwjq+k27pHLf2M1sjvnarPGVBOO2TXzZ4Z+GupfDvxBqNvql5a3TNGig25bHTd/EB619SyyxwQvLM6xxxqWd3OAoHJJPYV414s1Kx1DxhfNp97b3S4jOYJVcf6tfQ1pT+I9jJoRliry6K6/AyqKKK6j7MQjIweRVOW1t7LR54LOCK3iWJ8RxIFUZBJ4FXaayh0KuMqwwQe4pCaTOH/ZU/5KnqX/AGBpf/R8NfW1edfCXwpoWk6fc6lpulWttetK8Bnjjwxjwjbc+mQDXotcbVnY/O69J0ajpvoFFFFIxCvMfD/x00LxF8Rh4OtdM1CK9M88HnSBPLzErsTw2cHYccdxXp1fJfw00TVbf9qJLu40y8itv7Rv2897d1TBjmwdxGOcj86APrSiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA8f+NXwe1f4matpd3pOoWNolnA8bi5L5YlgeNqmvTfDWmS6J4T0jSrh0klsbKG2d0ztZkQKSM9sip73WNN0x0TUdRtLRnGVE86xlh7ZPNWopY54UlhdZI5FDI6HIYHkEHuKAHU2WKOeF4pkWSN1KsjjIYHqCO4p1FAGfpWgaPoayDRNJsdOEpzILS2SLefU7QM14F+1v/wAevhT/AH7v+UVfRtY+v+EtA8UrAPEWk2upC3LGEXEe7ZuxnH1wPyoAwvg7/wAkc8M/9eK/zNdrVbTtOs9J06Gw023jtrS3XZFDGMKi+gFWaACiiigDlfiF4+sPhz4bj1nVbW5uoZLlbcJbBdwZlY5+YgY+U1N4C8bWXxA8LJrumW1xbQPK8QjuAu7KnB6EiuD/AGmbG71D4W20NhazXUo1SJikEZdgPLk5wO3NXP2c7K6sPhDBBfW01tKLyYmOaMo2MjsaAPVKKKKACiiigAooooAKKKKACiiigAooooAx/F//ACI+u/8AYOuP/RbV82/sof8AI8a3/wBg4f8Aoxa+kvF//Ij67/2Drj/0W1fNv7KH/I8a3/2Dh/6MWgD6qooooAKZN/x7yf7p/lT6ZN/x7yf7p/lQB8h/svf8lcl/7Bk3/ocdfX9fIH7L3/JXJf8AsGTf+hx19f0AFFFFABRRRQBjeMf+RF17/sG3H/opq+MPhV/yGr7/AK9x/wChCvs/xj/yIuvf9g24/wDRTV8YfCr/AJDV9/17j/0IV6WV/wC+U/X9Dys4/wBwqen6o9Sooor9DPy4KbJ/q2+hp1Nk/wBW30NJ7DW5zP7Lv/JW5/8AsFzf+hx19fV8g/su/wDJW5/+wXN/6HHX19X5WfsoUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGJ40/5ELX/APsGXP8A6KavnL9k/wD5HLXf+vBf/Rgr6N8af8iFr/8A2DLn/wBFNXzl+yf/AMjlrv8A14L/AOjBQB9UUUUUAFQ3f/HjP/1zb+VTVDd/8eM//XNv5UAfJP7Lf/JWbn/sFTf+jIq+vK+Q/wBlv/krNz/2Cpv/AEZFX15QAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGH43/AOSfeIf+wXc/+imr51/ZO/5G7Xv+vFP/AEZX0V43/wCSfeIf+wXc/wDopq+df2Tv+Ru17/rxT/0ZQB9TUUUUAFQX3/IOuf8Ark38jU9QX3/IOuf+uTfyNAHyX+yz/wAlYu/+wTL/AOjYq+u6+RP2Wf8AkrF3/wBgmX/0bFX13QAUUUUAFFFFAGF45/5J54i/7BVz/wCimr5T+AH/ACMGrf8AXqv/AKHX1Z45/wCSeeIv+wVc/wDopq+U/gB/yMGrf9eq/wDodeLnv/Itq+i/NHXgv94ie70UUV+Rn1QVDef8eM//AFzb+VTVDef8eM//AFzb+VVH4kJ7Hjv7LH/JV7z/ALBEv/o2KvrqvkX9lj/kq95/2CJf/RsVfXVfuh8WFFFFABRRRQBg+O/+SdeI/wDsFXX/AKKavlX9n/8A5D2r/wDXsn/oVfVXjv8A5J14j/7BV1/6KavlX9n/AP5D2r/9eyf+hV4ue/8AItq+i/NHXgv94ie7UUUV+Rn1QVBe/wDHhcf9cm/lU9QXv/Hhcf8AXJv5VUfiQnseP/ssf8lWvv8AsES/+jYa+ua+Rv2WP+SrX3/YIl/9Gw19c1+6HxYUUUUAFFFfN/iH4s+M7D9ouPwta6uE0dtYtLU2/wBliP7tzHuXcV3c7jznPNAH0hRRRQAUUUUAfFvxa/5OS1H/AK/7X/0XFXeV1Pxe8A+G18S2XiAaef7TvJGlnn8+T52jEYQ7d20YHoK5avtcioyhh3Ue0tvlofn/ABHXjUxKpreK1+dmFFFFfQHzIVn6b8XP+EA8dNbf2J/aHnRLDu+1+VjeynONjdMVoV5/rnhHxBr/AMRUl0XSLq9RPKdmhj3AKCAT+Ga8bOpTjhHydXr6HvZBCnLGrn6K69dD7Wooor4M/SAooooAKK+dPj18VvGXgn4hQaZ4a1YWdm+nxzNGbWKTLl3BOXUnoor6HtnaS1idzlmQEn3xQBJRRRQAV8/ftZ/8i14d/wCvuX/0AV9A1y3j3wT4f8aaJs8SWJvFslklgAmePY23r8jDPTvUTmoQc30HFXdjyf4Vf8kw0b/rm/8A6MauvqnpOlWeh6XDp2mQ+TawAiOPcW2gkk8kk9SauV+K4qrGtiJ1Y7Sbf3s+wpxcIKL6IKKKK5jQ5jx74x/4QjQYdS+w/bvNuVg8vzvLxlWbOdp/u9Md69O+G3jb/hYXgqDxB/Z/9n+dLJH5Hnebja2M7tq9fpXi/wAZdG1HW/BtrbaRZzXky36SNHCu4hRHIM/TJH516Z8B9E1Pw98J7PT9bsprG7S4mZoZ12sAXJBxX6VwtTpRwkpx+JvX5bafefPZlKTqpPboej0UUV9YeWFFFeJftE/EPxP4Dbw6PCuoiy+2i58/NvHJv2eVt++pxjc3T1oA9tornPh7q15rvw50HVNUl868u7KOWaTaF3MRknAAA/CujoAKKKKAOe+IP/JM/E//AGCLv/0S1fI/wZ/5Cmqf9cU/9CNfaOo2Ftqul3Wn38fm2t3C8E0e4ruRlKsMjkZBPIrwq/8AA/h7wZ4rv4fDdgbONljUgzSScbA38bHuTWlNXkerlNNzxSa6a/p+otFFFdR9sFRzy+TbyS43bELYz1wKkqG8UvYzqgyzRsAB3OKBPY0PgL8Vv+Es1i98N/2N9k8uB7/7R9q35w0UezbsH97Oc9ule5V8wfsy+FNd0nx/qGpalpV1bWTaZJAJ5I8KZPNhbbn1wCa+n64ndvU/OqspzqOVTfqFFFFIyCiivm/wH8WfGetftBJ4Z1LVxNpBvryI2/2WJfljSUoNwUNwVXv2oA+kKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigD5d/az/wCRm8Pf9ecv/oYr6C8Af8k18M/9gi1/9ErVbxd8NfCnjq6trjxTpZvZbVCkTC4lj2qTkj5GGfxrodPsLfS9MtdPsY/KtbSFIIU3FtqKAqjJ5PAHJoAsUUUUAFecfFz4tf8ACrItKf8AsX+1f7RaUY+1eT5ezZ/sNnO/26V6PXhX7THhLX/FNv4bHh3SbrUjbvcmYW8e7ZuEeM/XB/KgD1vwd4i/4SzwbpmvfZfsn2+ATeR5m/y89t2Bn8hW1XJ/C3TrzSfhb4fsNSt5La7t7NUlhkGGRsngiusoAKKKKACivMvj14w1vwT8PYNT8NXgs7t9QjhaQxJJlCkhIw4I6qKtfA/xVrHjH4aQ6t4iuxd3rXMsZlESR5VTwMKAP0oA9DooooAKKKKACiiigAooooAKKKKACiiigDH8X/8AIj67/wBg64/9FtXzb+yh/wAjxrf/AGDh/wCjFr6S8X/8iPrv/YOuP/RbV8OeBPiDrfw71W41Dw/9nMtzD5Ei3EW9Su4N6jnIoA++6K+Pv+Gn/H39zSf/AAEb/wCLo/4af8ff3NJ/8BG/+LoA+waZN/x7yf7p/lXyD/w0/wCPv7mk/wDgI3/xdI37Tvj50KldJwRj/j1b/wCLoAf+y9/yVyX/ALBk3/ocdfX9fIH7L3/JXJf+wZN/6HHX1/QAUUUUAFFFFAGN4x/5EXXv+wbcf+imr4w+FX/Iavv+vcf+hCvs/wAY/wDIi69/2Dbj/wBFNXwRoXiC98PXUk+n+XukTYwkXIxnP9K7MFWjQxEak9kcOYUJ4jCzpQ3Z71RXkP8Aws3XvS0/79H/ABo/4Wbr3paf9+j/AI19b/bmE8/uPif9Xcb5ff8A8A9epsn+rb6GvI/+Fm696Wn/AH6P+NB+JmvMpBFryMf6o/40v7cwnn9w1w7jb9Pv/wCAdt+y7/yVuf8A7Bc3/ocdfX1fIP7Lv/JW5/8AsFzf+hx19fV8OfoYUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAGJ40/5ELX/APsGXP8A6KavnL9k/wD5HLXf+vBf/Rgr6N8af8iFr/8A2DLn/wBFNXxB4E+IWt/DvU7m+8PfZjLcw+TILiLeuMg8cjnigD76or4+/wCGn/H39zSf/ARv/i6P+Gn/AB9/c0n/AMBG/wDi6APsGobv/jxn/wCubfyr5E/4af8AH39zSf8AwEb/AOLpsn7Tfj2WJ42XScMpU4tW7/8AAqAJ/wBlv/krNz/2Cpv/AEZFX15XyH+y3/yVm5/7BU3/AKMir68oAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigDD8b/APJPvEP/AGC7n/0U1fOv7J3/ACN2vf8AXin/AKMr6K8b/wDJPvEP/YLuf/RTV8Q+BPiHrfw61K6vfD32YyXUQikFxFvGAcjGCOaAPvqivj7/AIaf8ff3NJ/8BG/+Lo/4af8AH39zSf8AwEb/AOLoA+wagvv+Qdc/9cm/ka+Rf+Gn/H39zSf/AAEb/wCLpkv7TXj2aF4mXSgrqVOLVu4/3qALP7LP/JWLv/sEy/8Ao2KvruvkT9ln/krF3/2CZf8A0bFX13QAUUUUAFFFFAGF45/5J54i/wCwVc/+imr5T+AH/Iwat/16r/6HX1Z45/5J54i/7BVz/wCimr4X8KeMdU8HXk9zo/k754xG4mTcMZz6ivOzPDTxWDnRp7v/ADOjD1I06qnLZH1nRXzp/wALy8W/3dP/APAc/wDxVH/C8vFv93T/APwHP/xVfA/6r4/+79//AAD2/wC0qHmfRdQ3n/HjP/1zb+VfPX/C8vFv93T/APwHP/xVNk+N3iyWJ42FgA6lTiA9/wDgVNcMY9O/u/f/AMAHmVC3U6X9lj/kq95/2CJf/RsVfXVfIv7LH/JV7z/sES/+jYq+uq/Tj5wKKKKACiiigDB8d/8AJOvEf/YKuv8A0U1fKv7P/wDyHtX/AOvZP/Qq+qvHf/JOvEf/AGCrr/0U1fDHhTxjqng68nuNH8nfcRiNxMm4YBz6ivOzTDTxWDnRp7v/ADOjDVI06qnLZH1nRXzp/wALy8W/3dP/APAc/wDxVH/C8vFv93T/APwHP/xVfA/6r4/+79//AAD2/wC0qHmfRdQXv/Hhcf8AXJv5V89/8Ly8W/3dP/8AAc//ABVNl+N3iyWF42FgA6lTiA9/+BVUeGMemn7v3/8AAB5jQt1Om/ZY/wCSrX3/AGCJf/RsNfXNfI37LH/JVr7/ALBEv/o2Gvrmv00+cCiiigArwzW/gXrep/G9PGsWqWCWS6nb3nkNv8zbGUJH3cZOw969zqhJr2jxX/2GXVbFLveE+ztcoJNx6DbnOTkcUAX6KKKACiiigD58+LvxR01vHkHhUWN19psJvJeb5djGURkEc54zzWZXHfEXS7rWf2pLyw09BJcTX9vsUsFziKNup46CvV/+FeeKP+gX/wCTEX/xVfXZLjIRoOnVmkltdpeZ8Rn+BqTxCqUYNtrWyb20RzNFeaeI/HPiDQ/FGp6U0dsjWV1JAVZNxG1iMZBweldl8LT4k+IFvqUkVpFP9jaNT5bLHjcG/vNz92vQp5vhJ1PZ3t5vb7zy6mR42nS9py38ldv7rGzXc/Cf/ka7n/ryf/0NK5Pxf4Y8TeG/COo6w+nrGtnF5hZpY2A5A6BsnrXimmfELxQniSC5s9YuLFpXSN1tXKKV3DIx74rlzTMcP7GVGD5nJdLNHbk+VYn28a81yqL63TPvSiiiviz74KKKKAPEPjD8Dta+I3jSHWdM1OwtYY7JLcpcb9xKs5zwpGPmFe2QRmK3jjJyUQKSPYVTvdd0jTZxDqOq2VpKV3COe4RGI9cE9ODV8EEAg5B6EUAFFFFABXCfFb4kaf8ADvQoJNSsrm6GomSCMW+35Tt6nJHHNd3Xz9+1n/yLXh3/AK+5f/QBUzipxcXsxptO6Nzw5rkPiTw9a6vaxSRRXKllSTG4YYrzj6Vp1kfC/wAK6vF8L9Bzbq4ltROjLIuCshLr1PXDCuW+MfivxF8PLrSYrOG3QXqSs3nKJM7SvTDcfer8olkuKqYuVKlTajd2bTSsr21sfTrF0o0lKUleyPQKK8T8DfFTxR4n8caXo08dk0d5N5bCOLax4J4JbjpX0B/wjWrf8+n/AJET/Gs8TkmOw8lF03L/AAptfkVTxlGor81vXQy69Sr4q8f/ABK8VQeJdV0OK6WxSw1CWJXtspJhGZdpYHkf1Ar6N+Autalr/wAJ7O/1q+nvrt7iZWmncsxAcgDJr7bh/LK2BhOVa3v207Wvv9542OxMKzSh0uekUUUV9OecFeUfG34T6p8T20Q6Tf2dn/Zwn8z7Tu+bzPLxjaD/AHD+der1Tv8AV9N0vy/7T1C1s/Mzs+0TLHvxjOMkZxkfnQBn+C9Dm8M+CNH0S6ljmmsLRIHkjztYqMZGecVuVHBPDdW6T20qTQyDckkbBlYeoI61JQAUUUUAUdb1SPQ/D+oatOjSRWFrJcuiY3MqKWIGe/FfPtr8StP+InibULnTbK5tVVI2xcbc/dC9ifSvb/iD/wAkz8T/APYIu/8A0S1fLvwF8M6rrM2s3em24mjhWON/3iqQTuI6kehq6btI9LLK3ssSm3ZPR3PR6Kd400TxF4Z8G6lrIslT7HFv3PIjAcgcgNk9a8O/4W34k9LL/vyf8a3dSKPp62aYak7c1/TU9voq54b8N+Itb8K6Vqn2FW+22cVxuWVFB3oG4BbI61wvxb1TxH4Bn06AQRW/26OQ5fDn5SoyNrcfepucUrmlTMMNCnz8yfkmr/ce9/Db/kW7j/r7b/0BK6+vl/8AZk8U67qvxA1DTtS1a6urNdLknEEspZBJ5sK7seuCRmvqCuWTu7nxWKqqtWlUXUKKKKRzhXhnhD4F634d+M6+MrnVLCWzF3dT+RHv8zEqSKByuMjeM89q9zqhDr2j3N99it9VsZbrcV8iO5RpMjORtBzkYP5UAX6KKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigDzz4lfGHSvhlqFjaapp15eNeRNIjWxTCgHGDuIrttF1OPW9B0/VYEaOK+to7lEf7yq6hgDjvzXzV+1n/AMjN4e/685f/AEMV9FeELCbS/A+hafdsjT2mnW8ErRn5SyxqpI9sigDYorwP43fGbxR8P/HUGkaCtibaSwjuD9ogLtuLuDyGHGFFVfg58bvFfjr4iQ6Lra2AtHt5ZD5EBRsqMjncaAPoaiivinxT8ePHut3SxLq39mLaySAf2duhMmSPvcnOMcfU0Afa1Fcn8LdQu9V+Fnh++1K4kurqezV5ZpW3M5yeSa6ygAooooA4L4w+Ab74j+C4dG0y7t7WaO9S4L3G7aQquMcAnPzCrHwm8E3nw/8AAcWhalcwXM6TySmS3ztwxyByAa6291Cy02ATajdwWkRbaJJ5Qik+mSevBpbO+tNRtxPp91DdQkkCSCQOuR2yOKAJ6KKKACiiigAooooAKKKKACiiigAooooAKMD0oooAMD0owPSiigAwPSjA9KKKADFFFFABRRRQAUUUUAFGB6UUUAGB6UYHpRRQAYHpRgelFFABiiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACjA9KKKADA9KMD0oooAMD0owPSiigAxRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRgelFFABgelGB6UUUAGB6UYHpRRQAYooooAKKKKACiiigAowPSiigAwPSjA9KKKADA9KMD0oooAMUUUUAFFFFABRRRQAUYHpRRQAYHpRgelFFABgelGB6UUUAGKKKKACiiigAr498W/8AJ3kX/YwWP/oUVfYVcdd/CjwXf+Lx4outG8zWVuI7kXP2qYfvExtbaH28bRxjHFAHY0UUUAFFFFABiiiigAwPSjGKKKACjA9KKKACiiigAooooA+Q/wBqT/krNt/2Cof/AEZLX1tZ/wDHjB/1zX+Vct4p+FXg3xrqy6n4m0f7beJEIVk+1TR4QEkDCOB1Y9u9dciLHGqIMKoAA9qAFooooAKKKKACjFFFABgelFFFABgelFFFABRRRQAV82ftcff8JfS8/wDaFfSdc14u+Hvhjx2bQ+K9M+3/AGLf5H+kSxbN+3d9xhnO1evpQBV+E3/JIfC//YNh/wDQa6+qek6VZaHo9rpelw+RZ2kQihi3ltijoMsST+Jq5QAUUUUAFFFFABRgelFFABRiiigAxRRRQAUUUUAFfHvwv/5OuT/sJaj/AOi56+wq47S/hR4L0XxYPEumaN5OriWSUXP2qZvmkDBztLleQzdu/FAHY0UUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABiiiigAxRgelFFABRgelFFABRRRQAUUUUAeL/ALUv/JJ7X/sLQ/8AouWrv7NP/JGrf/r9n/8AQhXoPinwhofjXSU0zxNZfbbNJRMsfmvHhwCAcoQejHv3p/hnwto/g/Rl0nw5Z/Y7JXaQRea8mGbqcuSf1oA16KKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAry7U/jro+mfE5fBMulXz3bXsNl9oUp5e6QqAeucDeK9Rr5O8U+Htal/ariv4tIv3sxrtlIblbZzHtDRZbdjGBg8+1AH1jRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQB5h8RfjjpHw48Sx6NqWlX13LJbLcCSAoFAZmGOT1+U16bFIJYUkAwHUMAfevlX9pbw/rOqfFG2n0zSb68hGmRKZLe2eRQfMk4yARnkV9T2gIsoARgiNcg/SgCWiiigAooooAKKKKACiiigAooooAKKKKACuC+JvxZ034YHTBqmn3V5/aPm7Ps5UbPL2ZzuI67x+Vd7Xzz+1Ro2p6s/hb+ytNu77yhd+Z9mgaTZnycZ2g4zg/lQB7l4a12HxN4Y07W7aJ4Yb+3WdI5MblDDODjjNalcp8Lrea0+FPhuC6hkhmj0+JXjkUqynb0IPINdXQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAV5d4d+Ouj+I/iSPBttpV9Fdm4ng8+Qp5eYldieDnB2H869Rr5O+G3h7WrX9p9L650i/htP7Rv2+0SWzrHgxzYO4jGDkY+tAH1jRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQBynxF8f2Xw48NR6zqVpPdxSXK24jgKhgWVjnk9PlNS+APG9p8QfCqa7p9rNawvK8QjnILZU4J44rhv2ltNvtU+F1tBplncXkw1SJjHbxNIwHlyc4AJxyKufs7WF5pvwjgt9RtJ7ScXcxMU8ZRgCRg4PNAHqdFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAH/9k=)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WD9SW7L24gxo"
      },
      "source": [
        "## 라이브러리 설치"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ei2VWoib-yIz"
      },
      "source": [
        "\n",
        "*   uisrnn을 간편하게 구현할 수 있는 uisrnn 라이브러리가 존재\n",
        "*   여기서는 uisrnn 라이브러리를 통해 uisrnn을 구현하고 학습, 평가\n",
        "*   실습을 위해 uisrnn과 easydict 라이브러리를 설치\n",
        "\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lMoTU9k-q3qX",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "440614fb-58bb-41ac-c935-04c97eff9ff5"
      },
      "source": [
        "!pip install uisrnn easydict"
      ],
      "execution_count": 1,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting uisrnn\n",
            "  Downloading uisrnn-0.1.2-py3-none-any.whl.metadata (12 kB)\n",
            "Requirement already satisfied: easydict in /usr/local/lib/python3.10/dist-packages (1.13)\n",
            "Requirement already satisfied: numpy>=1.18.0 in /usr/local/lib/python3.10/dist-packages (from uisrnn) (1.26.4)\n",
            "Requirement already satisfied: scipy>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from uisrnn) (1.13.1)\n",
            "Requirement already satisfied: torch>=1.4.0 in /usr/local/lib/python3.10/dist-packages (from uisrnn) (2.5.0+cu121)\n",
            "Collecting codecov==2.1.13 (from uisrnn)\n",
            "  Downloading codecov-2.1.13-py2.py3-none-any.whl.metadata (1.3 kB)\n",
            "Collecting coverage==7.6.0 (from uisrnn)\n",
            "  Downloading coverage-7.6.0-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl.metadata (8.2 kB)\n",
            "Collecting colortimelog (from uisrnn)\n",
            "  Downloading colortimelog-0.0.9-py3-none-any.whl.metadata (1.5 kB)\n",
            "Requirement already satisfied: requests>=2.7.9 in /usr/local/lib/python3.10/dist-packages (from codecov==2.1.13->uisrnn) (2.32.3)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.10/dist-packages (from torch>=1.4.0->uisrnn) (3.16.1)\n",
            "Requirement already satisfied: typing-extensions>=4.8.0 in /usr/local/lib/python3.10/dist-packages (from torch>=1.4.0->uisrnn) (4.12.2)\n",
            "Requirement already satisfied: networkx in /usr/local/lib/python3.10/dist-packages (from torch>=1.4.0->uisrnn) (3.4.2)\n",
            "Requirement already satisfied: jinja2 in /usr/local/lib/python3.10/dist-packages (from torch>=1.4.0->uisrnn) (3.1.4)\n",
            "Requirement already satisfied: fsspec in /usr/local/lib/python3.10/dist-packages (from torch>=1.4.0->uisrnn) (2024.10.0)\n",
            "Requirement already satisfied: sympy==1.13.1 in /usr/local/lib/python3.10/dist-packages (from torch>=1.4.0->uisrnn) (1.13.1)\n",
            "Requirement already satisfied: mpmath<1.4,>=1.1.0 in /usr/local/lib/python3.10/dist-packages (from sympy==1.13.1->torch>=1.4.0->uisrnn) (1.3.0)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests>=2.7.9->codecov==2.1.13->uisrnn) (3.4.0)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests>=2.7.9->codecov==2.1.13->uisrnn) (3.10)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests>=2.7.9->codecov==2.1.13->uisrnn) (2.2.3)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests>=2.7.9->codecov==2.1.13->uisrnn) (2024.8.30)\n",
            "Requirement already satisfied: MarkupSafe>=2.0 in /usr/local/lib/python3.10/dist-packages (from jinja2->torch>=1.4.0->uisrnn) (3.0.2)\n",
            "Downloading uisrnn-0.1.2-py3-none-any.whl (28 kB)\n",
            "Downloading codecov-2.1.13-py2.py3-none-any.whl (16 kB)\n",
            "Downloading coverage-7.6.0-cp310-cp310-manylinux_2_5_x86_64.manylinux1_x86_64.manylinux_2_17_x86_64.manylinux2014_x86_64.whl (233 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m233.8/233.8 kB\u001b[0m \u001b[31m5.7 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading colortimelog-0.0.9-py3-none-any.whl (7.9 kB)\n",
            "Installing collected packages: colortimelog, coverage, codecov, uisrnn\n",
            "Successfully installed codecov-2.1.13 colortimelog-0.0.9 coverage-7.6.0 uisrnn-0.1.2\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yb4Th_h64eoX"
      },
      "source": [
        "## 데이터 다운로드"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WD8-pfRw-mcs"
      },
      "source": [
        "* 학습 및 평가에는 uisrnn 라이브러리에서 제공하는 sample dataset을 사용\n",
        "* urlretrieve를 통해 url에서 데이터를 받아옴\n",
        "* https://github.com/google/uis-rnn/blob/master/data/toy_training_data.npz?raw=True\n",
        "* https://github.com/google/uis-rnn/blob/master/data/toy_testing_data.npz?raw=True\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3a7r3t1Dt-RD",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ce26a9e6-db34-4b78-af3a-823f26ea49f1"
      },
      "source": [
        "import urllib.request\n",
        "\n",
        "training_url = 'https://github.com/google/uis-rnn/blob/master/data/toy_training_data.npz?raw=True'\n",
        "urllib.request.urlretrieve(training_url, './toy_training_data.npz')\n",
        "\n",
        "testing_url = 'https://github.com/google/uis-rnn/blob/master/data/toy_testing_data.npz?raw=True'\n",
        "urllib.request.urlretrieve(testing_url, './toy_testing_data.npz')"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "('./toy_testing_data.npz', <http.client.HTTPMessage at 0x7d04c3f2c0d0>)"
            ]
          },
          "metadata": {},
          "execution_count": 3
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ECZnZSqb_8oT"
      },
      "source": [
        "\n",
        "\n",
        "*   다운로드한 데이터에서 데이터를 받아오고 sequence와 label을 분리\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RtKo3Ncz4Z42"
      },
      "source": [
        "import numpy as np\n",
        "import uisrnn\n",
        "\n",
        "train_data = np.load('./toy_training_data.npz', allow_pickle=True)\n",
        "test_data = np.load('./toy_testing_data.npz', allow_pickle=True)\n",
        "\n",
        "train_sequence = train_data['train_sequence']\n",
        "train_cluster_id = train_data['train_cluster_id']\n",
        "\n",
        "test_sequences = test_data['test_sequences'].tolist()\n",
        "test_cluster_ids = test_data['test_cluster_ids'].tolist()\n"
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Njq-3a7d4k4O"
      },
      "source": [
        "## 파라미터 설정"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xIHC1c0U-lq7"
      },
      "source": [
        "*   model, training, inference에 필요한 인자들은 uisrnn.parse_argement()를 통해 얻을 수 있음\n",
        "*   colab 환경에서는 argument를 사용할 수 없기 때문에 easydict로 대체\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "G66tYcOA4VAi"
      },
      "source": [
        "import easydict\n",
        "\n",
        "model_args = easydict.EasyDict({\"crp_alpha\":1.0,\n",
        "                                \"enable_cuda\": True,\n",
        "                                \"observation_dim\":256,\n",
        "                                \"rnn_depth\": 1,\n",
        "                                \"rnn_dropout\": 0.2,\n",
        "                                \"rnn_hidden_size\":512,\n",
        "                                \"sigma2\": None,\n",
        "                                \"transition_bias\": None,\n",
        "                                \"verbosity\":2})\n",
        "\n",
        "training_args = easydict.EasyDict({\"batch_size\":10,\n",
        "                                   \"enforce_cluster_id_uniqueness\": True,\n",
        "                                   \"grad_max_norm\": 5.0,\n",
        "                                   \"learning_rate\":0.001,\n",
        "                                   \"num_permutations\": 10,\n",
        "                                   \"optimizer\": 'adam',\n",
        "                                   \"regularization_weight\":1e-05,\n",
        "                                   \"sigma_alpha\":1.0,\n",
        "                                   \"sigma_beta\":1.0,\n",
        "                                   \"train_iteration\":5000})\n",
        "\n",
        "inference_args = easydict.EasyDict({\"batchsize\":100,\n",
        "                                    \"look_ahead\":1,\n",
        "                                    \"test_iteration\":2,\n",
        "                                    \"beam_size\":10})\n"
      ],
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "L4tag-_q4q6n"
      },
      "source": [
        "## UISRNN 모델 학습"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UOBG5e-8AHJM"
      },
      "source": [
        "\n",
        "\n",
        "*   앞서 구성한 argment를 사용해 모델을 구성\n",
        "*   데이터를 입력해 모델을 훈련\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TKow4oSh4e2-",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "7c3a7cd4-210e-43d5-b634-706137a37e89"
      },
      "source": [
        "model = uisrnn.UISRNN(model_args)\n",
        "\n",
        "model.fit(train_sequence, train_cluster_id, training_args)"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[93m[WARNING]\u001b[0m Warning: transition_bias cannot be correctly estimated from a concatenated sequence; train_sequences will be treated as a single sequence. This can lead to inaccurate estimation of transition_bias. Please, consider estimating transition_bias before concatenating the sequences and passing it as argument.\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 0  \tTraining Loss: -284.0649    \n",
            "    Negative Log Likelihood: 5.9837\tSigma2 Prior: -290.0493\tRegularization: 0.0006\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 10  \tTraining Loss: -298.1677    \n",
            "    Negative Log Likelihood: 5.6163\tSigma2 Prior: -303.7846\tRegularization: 0.0006\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 20  \tTraining Loss: -312.1277    \n",
            "    Negative Log Likelihood: 6.2553\tSigma2 Prior: -318.3837\tRegularization: 0.0006\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 30  \tTraining Loss: -329.4703    \n",
            "    Negative Log Likelihood: 7.2099\tSigma2 Prior: -336.6808\tRegularization: 0.0006\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 40  \tTraining Loss: -347.2236    \n",
            "    Negative Log Likelihood: 8.3147\tSigma2 Prior: -355.5390\tRegularization: 0.0006\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 50  \tTraining Loss: -372.5184    \n",
            "    Negative Log Likelihood: 10.3497\tSigma2 Prior: -382.8687\tRegularization: 0.0007\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 60  \tTraining Loss: -401.8133    \n",
            "    Negative Log Likelihood: 13.7288\tSigma2 Prior: -415.5428\tRegularization: 0.0007\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 70  \tTraining Loss: -444.8376    \n",
            "    Negative Log Likelihood: 22.1058\tSigma2 Prior: -466.9440\tRegularization: 0.0007\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 80  \tTraining Loss: -487.4564    \n",
            "    Negative Log Likelihood: 54.8064\tSigma2 Prior: -542.2634\tRegularization: 0.0007\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 90  \tTraining Loss: -444.7010    \n",
            "    Negative Log Likelihood: 43.4134\tSigma2 Prior: -488.1151\tRegularization: 0.0007\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 100  \tTraining Loss: -479.4600    \n",
            "    Negative Log Likelihood: 39.4503\tSigma2 Prior: -518.9110\tRegularization: 0.0007\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 110  \tTraining Loss: -477.9497    \n",
            "    Negative Log Likelihood: 49.4004\tSigma2 Prior: -527.3509\tRegularization: 0.0007\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 120  \tTraining Loss: -491.2445    \n",
            "    Negative Log Likelihood: 45.3157\tSigma2 Prior: -536.5609\tRegularization: 0.0007\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 130  \tTraining Loss: -503.3567    \n",
            "    Negative Log Likelihood: 46.5000\tSigma2 Prior: -549.8574\tRegularization: 0.0008\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 140  \tTraining Loss: -496.2961    \n",
            "    Negative Log Likelihood: 42.8306\tSigma2 Prior: -539.1274\tRegularization: 0.0008\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 150  \tTraining Loss: -470.6193    \n",
            "    Negative Log Likelihood: 39.7434\tSigma2 Prior: -510.3635\tRegularization: 0.0008\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 160  \tTraining Loss: -484.6760    \n",
            "    Negative Log Likelihood: 48.9224\tSigma2 Prior: -533.5992\tRegularization: 0.0008\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 170  \tTraining Loss: -444.9995    \n",
            "    Negative Log Likelihood: 42.4189\tSigma2 Prior: -487.4192\tRegularization: 0.0008\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 180  \tTraining Loss: -495.2631    \n",
            "    Negative Log Likelihood: 40.3207\tSigma2 Prior: -535.5847\tRegularization: 0.0008\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 190  \tTraining Loss: -479.8751    \n",
            "    Negative Log Likelihood: 48.4802\tSigma2 Prior: -528.3561\tRegularization: 0.0009\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 200  \tTraining Loss: -497.6149    \n",
            "    Negative Log Likelihood: 40.6761\tSigma2 Prior: -538.2919\tRegularization: 0.0009\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 210  \tTraining Loss: -470.7069    \n",
            "    Negative Log Likelihood: 43.2235\tSigma2 Prior: -513.9313\tRegularization: 0.0009\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 220  \tTraining Loss: -463.0341    \n",
            "    Negative Log Likelihood: 42.2585\tSigma2 Prior: -505.2936\tRegularization: 0.0009\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 230  \tTraining Loss: -482.9796    \n",
            "    Negative Log Likelihood: 45.7997\tSigma2 Prior: -528.7802\tRegularization: 0.0009\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 240  \tTraining Loss: -473.5541    \n",
            "    Negative Log Likelihood: 47.9356\tSigma2 Prior: -521.4907\tRegularization: 0.0009\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 250  \tTraining Loss: -481.1813    \n",
            "    Negative Log Likelihood: 39.2583\tSigma2 Prior: -520.4406\tRegularization: 0.0009\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 260  \tTraining Loss: -438.3551    \n",
            "    Negative Log Likelihood: 57.0326\tSigma2 Prior: -495.3887\tRegularization: 0.0010\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 270  \tTraining Loss: -492.2722    \n",
            "    Negative Log Likelihood: 35.6691\tSigma2 Prior: -527.9423\tRegularization: 0.0010\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 280  \tTraining Loss: -484.7509    \n",
            "    Negative Log Likelihood: 43.1248\tSigma2 Prior: -527.8768\tRegularization: 0.0010\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 290  \tTraining Loss: -502.1585    \n",
            "    Negative Log Likelihood: 33.6348\tSigma2 Prior: -535.7943\tRegularization: 0.0010\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 300  \tTraining Loss: -493.3909    \n",
            "    Negative Log Likelihood: 42.2006\tSigma2 Prior: -535.5925\tRegularization: 0.0010\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 310  \tTraining Loss: -505.4479    \n",
            "    Negative Log Likelihood: 40.9044\tSigma2 Prior: -546.3534\tRegularization: 0.0010\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 320  \tTraining Loss: -492.6924    \n",
            "    Negative Log Likelihood: 43.5700\tSigma2 Prior: -536.2635\tRegularization: 0.0011\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 330  \tTraining Loss: -489.1778    \n",
            "    Negative Log Likelihood: 32.7927\tSigma2 Prior: -521.9715\tRegularization: 0.0011\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 340  \tTraining Loss: -475.4410    \n",
            "    Negative Log Likelihood: 44.2944\tSigma2 Prior: -519.7365\tRegularization: 0.0011\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 350  \tTraining Loss: -485.6389    \n",
            "    Negative Log Likelihood: 37.4211\tSigma2 Prior: -523.0611\tRegularization: 0.0011\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 360  \tTraining Loss: -516.8461    \n",
            "    Negative Log Likelihood: 42.6602\tSigma2 Prior: -559.5073\tRegularization: 0.0011\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 370  \tTraining Loss: -487.8629    \n",
            "    Negative Log Likelihood: 40.6611\tSigma2 Prior: -528.5251\tRegularization: 0.0011\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 380  \tTraining Loss: -493.9244    \n",
            "    Negative Log Likelihood: 40.9778\tSigma2 Prior: -534.9033\tRegularization: 0.0011\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 390  \tTraining Loss: -503.7721    \n",
            "    Negative Log Likelihood: 42.0257\tSigma2 Prior: -545.7990\tRegularization: 0.0011\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 400  \tTraining Loss: -441.9872    \n",
            "    Negative Log Likelihood: 42.5759\tSigma2 Prior: -484.5642\tRegularization: 0.0011\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 410  \tTraining Loss: -495.1097    \n",
            "    Negative Log Likelihood: 36.2376\tSigma2 Prior: -531.3484\tRegularization: 0.0011\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 420  \tTraining Loss: -479.8339    \n",
            "    Negative Log Likelihood: 37.4820\tSigma2 Prior: -517.3171\tRegularization: 0.0012\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 430  \tTraining Loss: -483.3629    \n",
            "    Negative Log Likelihood: 36.3639\tSigma2 Prior: -519.7279\tRegularization: 0.0012\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 440  \tTraining Loss: -480.2392    \n",
            "    Negative Log Likelihood: 37.6422\tSigma2 Prior: -517.8825\tRegularization: 0.0012\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 450  \tTraining Loss: -491.1364    \n",
            "    Negative Log Likelihood: 37.2780\tSigma2 Prior: -528.4155\tRegularization: 0.0012\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 460  \tTraining Loss: -472.6053    \n",
            "    Negative Log Likelihood: 44.5287\tSigma2 Prior: -517.1351\tRegularization: 0.0012\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 470  \tTraining Loss: -518.2779    \n",
            "    Negative Log Likelihood: 37.8305\tSigma2 Prior: -556.1096\tRegularization: 0.0012\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 480  \tTraining Loss: -491.8395    \n",
            "    Negative Log Likelihood: 43.5003\tSigma2 Prior: -535.3410\tRegularization: 0.0012\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 490  \tTraining Loss: -485.2963    \n",
            "    Negative Log Likelihood: 41.1834\tSigma2 Prior: -526.4810\tRegularization: 0.0012\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 500  \tTraining Loss: -483.6018    \n",
            "    Negative Log Likelihood: 36.8163\tSigma2 Prior: -520.4193\tRegularization: 0.0012\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 510  \tTraining Loss: -500.5812    \n",
            "    Negative Log Likelihood: 37.3790\tSigma2 Prior: -537.9614\tRegularization: 0.0012\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 520  \tTraining Loss: -469.6735    \n",
            "    Negative Log Likelihood: 40.0836\tSigma2 Prior: -509.7582\tRegularization: 0.0012\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 530  \tTraining Loss: -473.1138    \n",
            "    Negative Log Likelihood: 36.9701\tSigma2 Prior: -510.0851\tRegularization: 0.0012\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 540  \tTraining Loss: -466.0745    \n",
            "    Negative Log Likelihood: 38.6727\tSigma2 Prior: -504.7484\tRegularization: 0.0012\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 550  \tTraining Loss: -508.2016    \n",
            "    Negative Log Likelihood: 35.6243\tSigma2 Prior: -543.8271\tRegularization: 0.0012\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 560  \tTraining Loss: -496.8326    \n",
            "    Negative Log Likelihood: 39.9188\tSigma2 Prior: -536.7526\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 570  \tTraining Loss: -490.8741    \n",
            "    Negative Log Likelihood: 40.9815\tSigma2 Prior: -531.8568\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 580  \tTraining Loss: -506.7660    \n",
            "    Negative Log Likelihood: 36.3272\tSigma2 Prior: -543.0944\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 590  \tTraining Loss: -506.1915    \n",
            "    Negative Log Likelihood: 42.0349\tSigma2 Prior: -548.2277\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 600  \tTraining Loss: -454.1206    \n",
            "    Negative Log Likelihood: 37.0407\tSigma2 Prior: -491.1625\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 610  \tTraining Loss: -504.1299    \n",
            "    Negative Log Likelihood: 37.1050\tSigma2 Prior: -541.2361\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 620  \tTraining Loss: -448.0758    \n",
            "    Negative Log Likelihood: 40.0505\tSigma2 Prior: -488.1277\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 630  \tTraining Loss: -487.3239    \n",
            "    Negative Log Likelihood: 29.0132\tSigma2 Prior: -516.3384\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 640  \tTraining Loss: -468.2510    \n",
            "    Negative Log Likelihood: 38.0787\tSigma2 Prior: -506.3310\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 650  \tTraining Loss: -497.4399    \n",
            "    Negative Log Likelihood: 35.5327\tSigma2 Prior: -532.9739\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 660  \tTraining Loss: -482.3793    \n",
            "    Negative Log Likelihood: 37.8734\tSigma2 Prior: -520.2540\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 670  \tTraining Loss: -495.3937    \n",
            "    Negative Log Likelihood: 36.2846\tSigma2 Prior: -531.6796\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 680  \tTraining Loss: -504.1350    \n",
            "    Negative Log Likelihood: 33.5595\tSigma2 Prior: -537.6958\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 690  \tTraining Loss: -507.0430    \n",
            "    Negative Log Likelihood: 42.7079\tSigma2 Prior: -549.7522\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 700  \tTraining Loss: -498.9112    \n",
            "    Negative Log Likelihood: 38.6746\tSigma2 Prior: -537.5872\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 710  \tTraining Loss: -468.5883    \n",
            "    Negative Log Likelihood: 38.9754\tSigma2 Prior: -507.5651\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 720  \tTraining Loss: -517.2393    \n",
            "    Negative Log Likelihood: 39.3394\tSigma2 Prior: -556.5800\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 730  \tTraining Loss: -501.3972    \n",
            "    Negative Log Likelihood: 45.3473\tSigma2 Prior: -546.7459\tRegularization: 0.0013\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 740  \tTraining Loss: -503.7172    \n",
            "    Negative Log Likelihood: 34.3793\tSigma2 Prior: -538.0978\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 750  \tTraining Loss: -504.4537    \n",
            "    Negative Log Likelihood: 34.1459\tSigma2 Prior: -538.6009\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 760  \tTraining Loss: -472.4145    \n",
            "    Negative Log Likelihood: 42.2604\tSigma2 Prior: -514.6763\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 770  \tTraining Loss: -434.3964    \n",
            "    Negative Log Likelihood: 39.1138\tSigma2 Prior: -473.5116\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 780  \tTraining Loss: -511.9073    \n",
            "    Negative Log Likelihood: 35.9305\tSigma2 Prior: -547.8392\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 790  \tTraining Loss: -456.6973    \n",
            "    Negative Log Likelihood: 35.3299\tSigma2 Prior: -492.0286\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 800  \tTraining Loss: -477.6501    \n",
            "    Negative Log Likelihood: 31.8222\tSigma2 Prior: -509.4737\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 810  \tTraining Loss: -505.5540    \n",
            "    Negative Log Likelihood: 30.4308\tSigma2 Prior: -535.9862\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 820  \tTraining Loss: -485.7673    \n",
            "    Negative Log Likelihood: 40.5883\tSigma2 Prior: -526.3570\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 830  \tTraining Loss: -489.6903    \n",
            "    Negative Log Likelihood: 32.5660\tSigma2 Prior: -522.2578\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 840  \tTraining Loss: -485.2715    \n",
            "    Negative Log Likelihood: 38.6429\tSigma2 Prior: -523.9158\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 850  \tTraining Loss: -472.8717    \n",
            "    Negative Log Likelihood: 40.6779\tSigma2 Prior: -513.5510\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 860  \tTraining Loss: -431.0152    \n",
            "    Negative Log Likelihood: 37.1972\tSigma2 Prior: -468.2138\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 870  \tTraining Loss: -496.6141    \n",
            "    Negative Log Likelihood: 31.6501\tSigma2 Prior: -528.2656\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 880  \tTraining Loss: -470.4486    \n",
            "    Negative Log Likelihood: 41.5854\tSigma2 Prior: -512.0354\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 890  \tTraining Loss: -489.3028    \n",
            "    Negative Log Likelihood: 32.9444\tSigma2 Prior: -522.2486\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 900  \tTraining Loss: -510.9068    \n",
            "    Negative Log Likelihood: 38.7717\tSigma2 Prior: -549.6799\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 910  \tTraining Loss: -480.1278    \n",
            "    Negative Log Likelihood: 33.3082\tSigma2 Prior: -513.4374\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 920  \tTraining Loss: -487.7626    \n",
            "    Negative Log Likelihood: 33.4320\tSigma2 Prior: -521.1960\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 930  \tTraining Loss: -451.6906    \n",
            "    Negative Log Likelihood: 36.0909\tSigma2 Prior: -487.7830\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 940  \tTraining Loss: -505.1353    \n",
            "    Negative Log Likelihood: 35.0362\tSigma2 Prior: -540.1730\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 950  \tTraining Loss: -404.1749    \n",
            "    Negative Log Likelihood: 39.7028\tSigma2 Prior: -443.8791\tRegularization: 0.0014\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 960  \tTraining Loss: -502.5347    \n",
            "    Negative Log Likelihood: 32.2507\tSigma2 Prior: -534.7869\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 970  \tTraining Loss: -474.5847    \n",
            "    Negative Log Likelihood: 39.5373\tSigma2 Prior: -514.1235\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 980  \tTraining Loss: -478.5890    \n",
            "    Negative Log Likelihood: 35.5600\tSigma2 Prior: -514.1505\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 990  \tTraining Loss: -510.8695    \n",
            "    Negative Log Likelihood: 39.8621\tSigma2 Prior: -550.7331\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1000  \tTraining Loss: -475.8330    \n",
            "    Negative Log Likelihood: 34.5642\tSigma2 Prior: -510.3987\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1010  \tTraining Loss: -391.8126    \n",
            "    Negative Log Likelihood: 37.3186\tSigma2 Prior: -429.1326\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1020  \tTraining Loss: -473.6771    \n",
            "    Negative Log Likelihood: 36.6902\tSigma2 Prior: -510.3688\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1030  \tTraining Loss: -453.4580    \n",
            "    Negative Log Likelihood: 37.4912\tSigma2 Prior: -490.9507\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1040  \tTraining Loss: -484.4218    \n",
            "    Negative Log Likelihood: 34.5560\tSigma2 Prior: -518.9792\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1050  \tTraining Loss: -491.4339    \n",
            "    Negative Log Likelihood: 41.7165\tSigma2 Prior: -533.1519\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1060  \tTraining Loss: -498.9981    \n",
            "    Negative Log Likelihood: 33.0470\tSigma2 Prior: -532.0466\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1070  \tTraining Loss: -495.8535    \n",
            "    Negative Log Likelihood: 37.6266\tSigma2 Prior: -533.4816\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1080  \tTraining Loss: -500.3042    \n",
            "    Negative Log Likelihood: 33.6787\tSigma2 Prior: -533.9844\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1090  \tTraining Loss: -519.5334    \n",
            "    Negative Log Likelihood: 28.2138\tSigma2 Prior: -547.7487\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1100  \tTraining Loss: -494.2901    \n",
            "    Negative Log Likelihood: 39.2644\tSigma2 Prior: -533.5560\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1110  \tTraining Loss: -495.1631    \n",
            "    Negative Log Likelihood: 33.2424\tSigma2 Prior: -528.4070\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1120  \tTraining Loss: -462.0648    \n",
            "    Negative Log Likelihood: 33.4621\tSigma2 Prior: -495.5284\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1130  \tTraining Loss: -506.0272    \n",
            "    Negative Log Likelihood: 39.5221\tSigma2 Prior: -545.5508\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1140  \tTraining Loss: -482.8120    \n",
            "    Negative Log Likelihood: 33.5977\tSigma2 Prior: -516.4113\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1150  \tTraining Loss: -453.9818    \n",
            "    Negative Log Likelihood: 36.9326\tSigma2 Prior: -490.9160\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1160  \tTraining Loss: -476.9201    \n",
            "    Negative Log Likelihood: 36.6885\tSigma2 Prior: -513.6102\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1170  \tTraining Loss: -481.1871    \n",
            "    Negative Log Likelihood: 35.7476\tSigma2 Prior: -516.9363\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1180  \tTraining Loss: -468.2264    \n",
            "    Negative Log Likelihood: 35.0478\tSigma2 Prior: -503.2757\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1190  \tTraining Loss: -506.9666    \n",
            "    Negative Log Likelihood: 38.9858\tSigma2 Prior: -545.9540\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1200  \tTraining Loss: -473.7388    \n",
            "    Negative Log Likelihood: 33.6426\tSigma2 Prior: -507.3830\tRegularization: 0.0015\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1210  \tTraining Loss: -476.3891    \n",
            "    Negative Log Likelihood: 30.1959\tSigma2 Prior: -506.5865\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1220  \tTraining Loss: -513.1602    \n",
            "    Negative Log Likelihood: 33.2262\tSigma2 Prior: -546.3879\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1230  \tTraining Loss: -463.3015    \n",
            "    Negative Log Likelihood: 40.8643\tSigma2 Prior: -504.1674\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1240  \tTraining Loss: -488.2820    \n",
            "    Negative Log Likelihood: 27.6899\tSigma2 Prior: -515.9734\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1250  \tTraining Loss: -481.7791    \n",
            "    Negative Log Likelihood: 40.9687\tSigma2 Prior: -522.7494\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1260  \tTraining Loss: -486.2877    \n",
            "    Negative Log Likelihood: 35.4649\tSigma2 Prior: -521.7541\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1270  \tTraining Loss: -476.9932    \n",
            "    Negative Log Likelihood: 35.9515\tSigma2 Prior: -512.9463\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1280  \tTraining Loss: -519.2979    \n",
            "    Negative Log Likelihood: 35.5360\tSigma2 Prior: -554.8354\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1290  \tTraining Loss: -495.6243    \n",
            "    Negative Log Likelihood: 31.3592\tSigma2 Prior: -526.9852\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1300  \tTraining Loss: -497.6327    \n",
            "    Negative Log Likelihood: 35.5586\tSigma2 Prior: -533.1929\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1310  \tTraining Loss: -412.2866    \n",
            "    Negative Log Likelihood: 36.6280\tSigma2 Prior: -448.9162\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1320  \tTraining Loss: -496.0127    \n",
            "    Negative Log Likelihood: 29.1797\tSigma2 Prior: -525.1940\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1330  \tTraining Loss: -491.3557    \n",
            "    Negative Log Likelihood: 38.8282\tSigma2 Prior: -530.1854\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1340  \tTraining Loss: -492.1563    \n",
            "    Negative Log Likelihood: 37.4484\tSigma2 Prior: -529.6063\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1350  \tTraining Loss: -507.3717    \n",
            "    Negative Log Likelihood: 36.9192\tSigma2 Prior: -544.2925\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1360  \tTraining Loss: -491.0928    \n",
            "    Negative Log Likelihood: 32.7853\tSigma2 Prior: -523.8797\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1370  \tTraining Loss: -507.9426    \n",
            "    Negative Log Likelihood: 33.8750\tSigma2 Prior: -541.8193\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1380  \tTraining Loss: -475.8621    \n",
            "    Negative Log Likelihood: 36.7840\tSigma2 Prior: -512.6477\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1390  \tTraining Loss: -509.8405    \n",
            "    Negative Log Likelihood: 32.9614\tSigma2 Prior: -542.8035\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1400  \tTraining Loss: -491.9915    \n",
            "    Negative Log Likelihood: 35.2142\tSigma2 Prior: -527.2073\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1410  \tTraining Loss: -500.1005    \n",
            "    Negative Log Likelihood: 37.7040\tSigma2 Prior: -537.8062\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1420  \tTraining Loss: -503.6717    \n",
            "    Negative Log Likelihood: 34.5729\tSigma2 Prior: -538.2462\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1430  \tTraining Loss: -495.7433    \n",
            "    Negative Log Likelihood: 31.8655\tSigma2 Prior: -527.6104\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1440  \tTraining Loss: -464.7394    \n",
            "    Negative Log Likelihood: 31.1734\tSigma2 Prior: -495.9145\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1450  \tTraining Loss: -513.2775    \n",
            "    Negative Log Likelihood: 37.5264\tSigma2 Prior: -550.8055\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1460  \tTraining Loss: -496.5870    \n",
            "    Negative Log Likelihood: 33.9376\tSigma2 Prior: -530.5262\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1470  \tTraining Loss: -511.3659    \n",
            "    Negative Log Likelihood: 31.1588\tSigma2 Prior: -542.5263\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1480  \tTraining Loss: -507.7761    \n",
            "    Negative Log Likelihood: 34.8750\tSigma2 Prior: -542.6528\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1490  \tTraining Loss: -478.4828    \n",
            "    Negative Log Likelihood: 42.1178\tSigma2 Prior: -520.6022\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1500  \tTraining Loss: -493.1504    \n",
            "    Negative Log Likelihood: 27.8505\tSigma2 Prior: -521.0026\tRegularization: 0.0016\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1510  \tTraining Loss: -491.8486    \n",
            "    Negative Log Likelihood: 37.8407\tSigma2 Prior: -529.6910\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1520  \tTraining Loss: -499.5078    \n",
            "    Negative Log Likelihood: 39.6902\tSigma2 Prior: -539.1997\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1530  \tTraining Loss: -502.1729    \n",
            "    Negative Log Likelihood: 32.0364\tSigma2 Prior: -534.2109\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1540  \tTraining Loss: -471.7366    \n",
            "    Negative Log Likelihood: 32.7244\tSigma2 Prior: -504.4626\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1550  \tTraining Loss: -480.2736    \n",
            "    Negative Log Likelihood: 33.2825\tSigma2 Prior: -513.5577\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1560  \tTraining Loss: -441.8167    \n",
            "    Negative Log Likelihood: 39.5917\tSigma2 Prior: -481.4100\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1570  \tTraining Loss: -483.2900    \n",
            "    Negative Log Likelihood: 38.9801\tSigma2 Prior: -522.2718\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1580  \tTraining Loss: -512.3768    \n",
            "    Negative Log Likelihood: 34.6635\tSigma2 Prior: -547.0419\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1590  \tTraining Loss: -465.6804    \n",
            "    Negative Log Likelihood: 29.8436\tSigma2 Prior: -495.5256\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1600  \tTraining Loss: -514.4561    \n",
            "    Negative Log Likelihood: 34.5893\tSigma2 Prior: -549.0471\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1610  \tTraining Loss: -475.0172    \n",
            "    Negative Log Likelihood: 36.8209\tSigma2 Prior: -511.8398\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1620  \tTraining Loss: -492.1468    \n",
            "    Negative Log Likelihood: 38.6925\tSigma2 Prior: -530.8409\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1630  \tTraining Loss: -502.3964    \n",
            "    Negative Log Likelihood: 31.4723\tSigma2 Prior: -533.8704\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1640  \tTraining Loss: -437.9215    \n",
            "    Negative Log Likelihood: 39.0760\tSigma2 Prior: -476.9993\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1650  \tTraining Loss: -495.4244    \n",
            "    Negative Log Likelihood: 30.0587\tSigma2 Prior: -525.4849\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1660  \tTraining Loss: -498.5816    \n",
            "    Negative Log Likelihood: 42.3595\tSigma2 Prior: -540.9428\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1670  \tTraining Loss: -498.2380    \n",
            "    Negative Log Likelihood: 31.4633\tSigma2 Prior: -529.7031\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1680  \tTraining Loss: -493.3757    \n",
            "    Negative Log Likelihood: 33.3210\tSigma2 Prior: -526.6984\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1690  \tTraining Loss: -492.2413    \n",
            "    Negative Log Likelihood: 35.7301\tSigma2 Prior: -527.9731\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1700  \tTraining Loss: -474.2402    \n",
            "    Negative Log Likelihood: 38.1863\tSigma2 Prior: -512.4282\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1710  \tTraining Loss: -510.0632    \n",
            "    Negative Log Likelihood: 29.4719\tSigma2 Prior: -539.5368\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1720  \tTraining Loss: -478.7060    \n",
            "    Negative Log Likelihood: 35.9149\tSigma2 Prior: -514.6226\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1730  \tTraining Loss: -497.6178    \n",
            "    Negative Log Likelihood: 37.9728\tSigma2 Prior: -535.5923\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1740  \tTraining Loss: -510.2890    \n",
            "    Negative Log Likelihood: 39.2530\tSigma2 Prior: -549.5438\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1750  \tTraining Loss: -489.2424    \n",
            "    Negative Log Likelihood: 34.9825\tSigma2 Prior: -524.2266\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1760  \tTraining Loss: -469.4942    \n",
            "    Negative Log Likelihood: 32.8957\tSigma2 Prior: -502.3917\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1770  \tTraining Loss: -508.0334    \n",
            "    Negative Log Likelihood: 33.9795\tSigma2 Prior: -542.0146\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1780  \tTraining Loss: -488.8563    \n",
            "    Negative Log Likelihood: 37.2755\tSigma2 Prior: -526.1335\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1790  \tTraining Loss: -483.9385    \n",
            "    Negative Log Likelihood: 29.4622\tSigma2 Prior: -513.4025\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1800  \tTraining Loss: -489.5865    \n",
            "    Negative Log Likelihood: 41.5396\tSigma2 Prior: -531.1279\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1810  \tTraining Loss: -484.8945    \n",
            "    Negative Log Likelihood: 32.3508\tSigma2 Prior: -517.2471\tRegularization: 0.0017\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1820  \tTraining Loss: -482.3726    \n",
            "    Negative Log Likelihood: 29.9731\tSigma2 Prior: -512.3474\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1830  \tTraining Loss: -491.0302    \n",
            "    Negative Log Likelihood: 42.6323\tSigma2 Prior: -533.6642\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1840  \tTraining Loss: -490.5290    \n",
            "    Negative Log Likelihood: 35.6418\tSigma2 Prior: -526.1725\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1850  \tTraining Loss: -485.3666    \n",
            "    Negative Log Likelihood: 37.1627\tSigma2 Prior: -522.5310\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1860  \tTraining Loss: -473.8346    \n",
            "    Negative Log Likelihood: 35.8532\tSigma2 Prior: -509.6896\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1870  \tTraining Loss: -514.9271    \n",
            "    Negative Log Likelihood: 37.0601\tSigma2 Prior: -551.9890\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1880  \tTraining Loss: -491.0806    \n",
            "    Negative Log Likelihood: 31.0093\tSigma2 Prior: -522.0917\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1890  \tTraining Loss: -478.5331    \n",
            "    Negative Log Likelihood: 33.7220\tSigma2 Prior: -512.2568\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1900  \tTraining Loss: -496.7316    \n",
            "    Negative Log Likelihood: 38.2777\tSigma2 Prior: -535.0111\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1910  \tTraining Loss: -482.8228    \n",
            "    Negative Log Likelihood: 33.9585\tSigma2 Prior: -516.7830\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1920  \tTraining Loss: -477.9197    \n",
            "    Negative Log Likelihood: 38.5963\tSigma2 Prior: -516.5178\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1930  \tTraining Loss: -441.1372    \n",
            "    Negative Log Likelihood: 33.6481\tSigma2 Prior: -474.7870\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1940  \tTraining Loss: -469.5066    \n",
            "    Negative Log Likelihood: 31.1769\tSigma2 Prior: -500.6853\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1950  \tTraining Loss: -487.5472    \n",
            "    Negative Log Likelihood: 42.8307\tSigma2 Prior: -530.3798\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1960  \tTraining Loss: -498.4918    \n",
            "    Negative Log Likelihood: 31.5405\tSigma2 Prior: -530.0341\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1970  \tTraining Loss: -479.2064    \n",
            "    Negative Log Likelihood: 35.7867\tSigma2 Prior: -514.9949\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1980  \tTraining Loss: -485.2961    \n",
            "    Negative Log Likelihood: 33.4706\tSigma2 Prior: -518.7686\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 1990  \tTraining Loss: -490.3880    \n",
            "    Negative Log Likelihood: 35.8887\tSigma2 Prior: -526.2785\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2000  \tTraining Loss: -509.1720    \n",
            "    Negative Log Likelihood: 33.2083\tSigma2 Prior: -542.3821\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2010  \tTraining Loss: -495.1037    \n",
            "    Negative Log Likelihood: 34.3719\tSigma2 Prior: -529.4774\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2020  \tTraining Loss: -498.0098    \n",
            "    Negative Log Likelihood: 37.4180\tSigma2 Prior: -535.4296\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2030  \tTraining Loss: -502.1792    \n",
            "    Negative Log Likelihood: 27.4945\tSigma2 Prior: -529.6755\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2040  \tTraining Loss: -520.6959    \n",
            "    Negative Log Likelihood: 37.5041\tSigma2 Prior: -558.2019\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2050  \tTraining Loss: -506.7650    \n",
            "    Negative Log Likelihood: 33.3042\tSigma2 Prior: -540.0711\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2060  \tTraining Loss: -495.8998    \n",
            "    Negative Log Likelihood: 31.8530\tSigma2 Prior: -527.7547\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2070  \tTraining Loss: -511.7425    \n",
            "    Negative Log Likelihood: 35.0451\tSigma2 Prior: -546.7894\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2080  \tTraining Loss: -489.7748    \n",
            "    Negative Log Likelihood: 33.2897\tSigma2 Prior: -523.0663\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2090  \tTraining Loss: -509.1682    \n",
            "    Negative Log Likelihood: 34.7149\tSigma2 Prior: -543.8849\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2100  \tTraining Loss: -475.1032    \n",
            "    Negative Log Likelihood: 30.5353\tSigma2 Prior: -505.6404\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2110  \tTraining Loss: -517.1083    \n",
            "    Negative Log Likelihood: 28.5729\tSigma2 Prior: -545.6830\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2120  \tTraining Loss: -478.0606    \n",
            "    Negative Log Likelihood: 36.1721\tSigma2 Prior: -514.2346\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2130  \tTraining Loss: -498.7550    \n",
            "    Negative Log Likelihood: 29.8360\tSigma2 Prior: -528.5928\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2140  \tTraining Loss: -509.3271    \n",
            "    Negative Log Likelihood: 36.0123\tSigma2 Prior: -545.3413\tRegularization: 0.0018\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2150  \tTraining Loss: -506.0025    \n",
            "    Negative Log Likelihood: 34.9918\tSigma2 Prior: -540.9962\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2160  \tTraining Loss: -494.3886    \n",
            "    Negative Log Likelihood: 33.3283\tSigma2 Prior: -527.7188\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2170  \tTraining Loss: -461.1842    \n",
            "    Negative Log Likelihood: 30.4556\tSigma2 Prior: -491.6417\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2180  \tTraining Loss: -485.3532    \n",
            "    Negative Log Likelihood: 37.6373\tSigma2 Prior: -522.9924\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2190  \tTraining Loss: -432.9035    \n",
            "    Negative Log Likelihood: 39.6627\tSigma2 Prior: -472.5681\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2200  \tTraining Loss: -483.4579    \n",
            "    Negative Log Likelihood: 36.8254\tSigma2 Prior: -520.2852\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2210  \tTraining Loss: -501.7830    \n",
            "    Negative Log Likelihood: 35.6115\tSigma2 Prior: -537.3964\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2220  \tTraining Loss: -496.8990    \n",
            "    Negative Log Likelihood: 42.4272\tSigma2 Prior: -539.3281\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2230  \tTraining Loss: -479.3708    \n",
            "    Negative Log Likelihood: 30.6960\tSigma2 Prior: -510.0687\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2240  \tTraining Loss: -489.7040    \n",
            "    Negative Log Likelihood: 31.3732\tSigma2 Prior: -521.0790\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2250  \tTraining Loss: -510.2989    \n",
            "    Negative Log Likelihood: 38.8120\tSigma2 Prior: -549.1128\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2260  \tTraining Loss: -467.1608    \n",
            "    Negative Log Likelihood: 41.5587\tSigma2 Prior: -508.7213\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2270  \tTraining Loss: -509.3349    \n",
            "    Negative Log Likelihood: 29.5466\tSigma2 Prior: -538.8834\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2280  \tTraining Loss: -476.5529    \n",
            "    Negative Log Likelihood: 34.3559\tSigma2 Prior: -510.9107\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2290  \tTraining Loss: -486.2805    \n",
            "    Negative Log Likelihood: 33.5914\tSigma2 Prior: -519.8738\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2300  \tTraining Loss: -499.8540    \n",
            "    Negative Log Likelihood: 31.5975\tSigma2 Prior: -531.4534\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2310  \tTraining Loss: -510.5157    \n",
            "    Negative Log Likelihood: 36.9299\tSigma2 Prior: -547.4474\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2320  \tTraining Loss: -407.9319    \n",
            "    Negative Log Likelihood: 40.2850\tSigma2 Prior: -448.2188\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2330  \tTraining Loss: -506.8140    \n",
            "    Negative Log Likelihood: 30.2212\tSigma2 Prior: -537.0370\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2340  \tTraining Loss: -490.3896    \n",
            "    Negative Log Likelihood: 37.4169\tSigma2 Prior: -527.8083\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2350  \tTraining Loss: -474.1174    \n",
            "    Negative Log Likelihood: 31.6870\tSigma2 Prior: -505.8064\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2360  \tTraining Loss: -491.6279    \n",
            "    Negative Log Likelihood: 32.1437\tSigma2 Prior: -523.7736\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2370  \tTraining Loss: -504.1161    \n",
            "    Negative Log Likelihood: 35.0875\tSigma2 Prior: -539.2054\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2380  \tTraining Loss: -460.9409    \n",
            "    Negative Log Likelihood: 37.0316\tSigma2 Prior: -497.9744\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2390  \tTraining Loss: -508.3423    \n",
            "    Negative Log Likelihood: 40.1773\tSigma2 Prior: -548.5216\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2400  \tTraining Loss: -487.9811    \n",
            "    Negative Log Likelihood: 33.5009\tSigma2 Prior: -521.4839\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2410  \tTraining Loss: -511.6748    \n",
            "    Negative Log Likelihood: 36.8882\tSigma2 Prior: -548.5650\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2420  \tTraining Loss: -473.8750    \n",
            "    Negative Log Likelihood: 31.6101\tSigma2 Prior: -505.4869\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2430  \tTraining Loss: -475.3194    \n",
            "    Negative Log Likelihood: 32.7017\tSigma2 Prior: -508.0231\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2440  \tTraining Loss: -504.9854    \n",
            "    Negative Log Likelihood: 31.7076\tSigma2 Prior: -536.6949\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2450  \tTraining Loss: -466.4778    \n",
            "    Negative Log Likelihood: 39.0737\tSigma2 Prior: -505.5535\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2460  \tTraining Loss: -500.5873    \n",
            "    Negative Log Likelihood: 28.9292\tSigma2 Prior: -529.5185\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2470  \tTraining Loss: -483.1190    \n",
            "    Negative Log Likelihood: 34.3697\tSigma2 Prior: -517.4907\tRegularization: 0.0019\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2480  \tTraining Loss: -502.9335    \n",
            "    Negative Log Likelihood: 32.8264\tSigma2 Prior: -535.7618\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2490  \tTraining Loss: -510.5772    \n",
            "    Negative Log Likelihood: 34.6561\tSigma2 Prior: -545.2352\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2500  \tTraining Loss: -483.4310    \n",
            "    Negative Log Likelihood: 36.1546\tSigma2 Prior: -519.5876\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2510  \tTraining Loss: -518.8179    \n",
            "    Negative Log Likelihood: 33.9973\tSigma2 Prior: -552.8171\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2520  \tTraining Loss: -474.5899    \n",
            "    Negative Log Likelihood: 33.5152\tSigma2 Prior: -508.1071\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2530  \tTraining Loss: -477.9283    \n",
            "    Negative Log Likelihood: 35.6954\tSigma2 Prior: -513.6256\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2540  \tTraining Loss: -480.9093    \n",
            "    Negative Log Likelihood: 35.6920\tSigma2 Prior: -516.6033\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2550  \tTraining Loss: -522.0373    \n",
            "    Negative Log Likelihood: 33.9662\tSigma2 Prior: -556.0054\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2560  \tTraining Loss: -463.8127    \n",
            "    Negative Log Likelihood: 32.1580\tSigma2 Prior: -495.9726\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2570  \tTraining Loss: -487.9629    \n",
            "    Negative Log Likelihood: 37.3301\tSigma2 Prior: -525.2950\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2580  \tTraining Loss: -503.9924    \n",
            "    Negative Log Likelihood: 33.8594\tSigma2 Prior: -537.8538\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2590  \tTraining Loss: -508.7873    \n",
            "    Negative Log Likelihood: 34.1229\tSigma2 Prior: -542.9122\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2600  \tTraining Loss: -500.0277    \n",
            "    Negative Log Likelihood: 35.1777\tSigma2 Prior: -535.2073\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2610  \tTraining Loss: -506.7677    \n",
            "    Negative Log Likelihood: 32.9409\tSigma2 Prior: -539.7106\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2620  \tTraining Loss: -510.3966    \n",
            "    Negative Log Likelihood: 34.1050\tSigma2 Prior: -544.5037\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2630  \tTraining Loss: -485.5011    \n",
            "    Negative Log Likelihood: 43.3908\tSigma2 Prior: -528.8938\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2640  \tTraining Loss: -477.4621    \n",
            "    Negative Log Likelihood: 30.7789\tSigma2 Prior: -508.2430\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2650  \tTraining Loss: -501.8073    \n",
            "    Negative Log Likelihood: 28.8487\tSigma2 Prior: -530.6580\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2660  \tTraining Loss: -490.7578    \n",
            "    Negative Log Likelihood: 38.9522\tSigma2 Prior: -529.7120\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2670  \tTraining Loss: -478.5486    \n",
            "    Negative Log Likelihood: 33.3858\tSigma2 Prior: -511.9364\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2680  \tTraining Loss: -503.6440    \n",
            "    Negative Log Likelihood: 34.6013\tSigma2 Prior: -538.2473\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2690  \tTraining Loss: -517.4777    \n",
            "    Negative Log Likelihood: 40.4851\tSigma2 Prior: -557.9649\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2700  \tTraining Loss: -495.5968    \n",
            "    Negative Log Likelihood: 30.5426\tSigma2 Prior: -526.1414\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2710  \tTraining Loss: -492.6135    \n",
            "    Negative Log Likelihood: 36.7480\tSigma2 Prior: -529.3635\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2720  \tTraining Loss: -495.6584    \n",
            "    Negative Log Likelihood: 31.9879\tSigma2 Prior: -527.6483\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2730  \tTraining Loss: -489.0216    \n",
            "    Negative Log Likelihood: 33.9577\tSigma2 Prior: -522.9813\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2740  \tTraining Loss: -456.4890    \n",
            "    Negative Log Likelihood: 37.9366\tSigma2 Prior: -494.4277\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2750  \tTraining Loss: -490.1767    \n",
            "    Negative Log Likelihood: 32.3423\tSigma2 Prior: -522.5211\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2760  \tTraining Loss: -461.4238    \n",
            "    Negative Log Likelihood: 36.0659\tSigma2 Prior: -497.4918\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2770  \tTraining Loss: -494.5065    \n",
            "    Negative Log Likelihood: 35.6541\tSigma2 Prior: -530.1627\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2780  \tTraining Loss: -516.5157    \n",
            "    Negative Log Likelihood: 35.5791\tSigma2 Prior: -552.0968\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2790  \tTraining Loss: -510.3008    \n",
            "    Negative Log Likelihood: 37.0164\tSigma2 Prior: -547.3193\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2800  \tTraining Loss: -457.9164    \n",
            "    Negative Log Likelihood: 33.7083\tSigma2 Prior: -491.6267\tRegularization: 0.0020\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2810  \tTraining Loss: -488.1208    \n",
            "    Negative Log Likelihood: 31.6426\tSigma2 Prior: -519.7654\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2820  \tTraining Loss: -500.1155    \n",
            "    Negative Log Likelihood: 34.3129\tSigma2 Prior: -534.4304\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2830  \tTraining Loss: -473.5716    \n",
            "    Negative Log Likelihood: 32.7763\tSigma2 Prior: -506.3500\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2840  \tTraining Loss: -513.9589    \n",
            "    Negative Log Likelihood: 29.8295\tSigma2 Prior: -543.7905\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2850  \tTraining Loss: -496.8755    \n",
            "    Negative Log Likelihood: 35.7023\tSigma2 Prior: -532.5798\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2860  \tTraining Loss: -504.6918    \n",
            "    Negative Log Likelihood: 33.0376\tSigma2 Prior: -537.7314\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2870  \tTraining Loss: -512.7689    \n",
            "    Negative Log Likelihood: 36.0286\tSigma2 Prior: -548.7995\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2880  \tTraining Loss: -495.9415    \n",
            "    Negative Log Likelihood: 39.3418\tSigma2 Prior: -535.2853\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2890  \tTraining Loss: -487.4877    \n",
            "    Negative Log Likelihood: 31.8837\tSigma2 Prior: -519.3734\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2900  \tTraining Loss: -393.5889    \n",
            "    Negative Log Likelihood: 41.5425\tSigma2 Prior: -435.1335\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2910  \tTraining Loss: -481.3545    \n",
            "    Negative Log Likelihood: 33.9091\tSigma2 Prior: -515.2656\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2920  \tTraining Loss: -504.5936    \n",
            "    Negative Log Likelihood: 31.8577\tSigma2 Prior: -536.4534\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2930  \tTraining Loss: -497.4902    \n",
            "    Negative Log Likelihood: 34.3441\tSigma2 Prior: -531.8364\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2940  \tTraining Loss: -516.6004    \n",
            "    Negative Log Likelihood: 31.4842\tSigma2 Prior: -548.0867\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2950  \tTraining Loss: -499.4868    \n",
            "    Negative Log Likelihood: 32.3843\tSigma2 Prior: -531.8732\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2960  \tTraining Loss: -523.1426    \n",
            "    Negative Log Likelihood: 38.7787\tSigma2 Prior: -561.9233\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2970  \tTraining Loss: -501.6057    \n",
            "    Negative Log Likelihood: 29.1275\tSigma2 Prior: -530.7354\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2980  \tTraining Loss: -497.9651    \n",
            "    Negative Log Likelihood: 31.3428\tSigma2 Prior: -529.3101\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 2990  \tTraining Loss: -520.0834    \n",
            "    Negative Log Likelihood: 40.2789\tSigma2 Prior: -560.3643\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3000  \tTraining Loss: -463.0237    \n",
            "    Negative Log Likelihood: 31.2321\tSigma2 Prior: -494.2580\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3010  \tTraining Loss: -487.4254    \n",
            "    Negative Log Likelihood: 33.8224\tSigma2 Prior: -521.2499\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3020  \tTraining Loss: -492.8383    \n",
            "    Negative Log Likelihood: 36.1163\tSigma2 Prior: -528.9567\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3030  \tTraining Loss: -515.2045    \n",
            "    Negative Log Likelihood: 30.5936\tSigma2 Prior: -545.8002\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3040  \tTraining Loss: -477.1449    \n",
            "    Negative Log Likelihood: 38.8436\tSigma2 Prior: -515.9906\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3050  \tTraining Loss: -488.1847    \n",
            "    Negative Log Likelihood: 32.8587\tSigma2 Prior: -521.0455\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3060  \tTraining Loss: -464.1572    \n",
            "    Negative Log Likelihood: 39.8996\tSigma2 Prior: -504.0589\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3070  \tTraining Loss: -492.8294    \n",
            "    Negative Log Likelihood: 25.6684\tSigma2 Prior: -518.4999\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3080  \tTraining Loss: -469.0231    \n",
            "    Negative Log Likelihood: 30.7112\tSigma2 Prior: -499.7365\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3090  \tTraining Loss: -496.4466    \n",
            "    Negative Log Likelihood: 33.8906\tSigma2 Prior: -530.3394\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3100  \tTraining Loss: -491.6753    \n",
            "    Negative Log Likelihood: 37.7857\tSigma2 Prior: -529.4632\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3110  \tTraining Loss: -501.0298    \n",
            "    Negative Log Likelihood: 35.8312\tSigma2 Prior: -536.8632\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3120  \tTraining Loss: -511.5127    \n",
            "    Negative Log Likelihood: 34.2338\tSigma2 Prior: -545.7486\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3130  \tTraining Loss: -486.0852    \n",
            "    Negative Log Likelihood: 29.3458\tSigma2 Prior: -515.4332\tRegularization: 0.0021\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3140  \tTraining Loss: -461.9216    \n",
            "    Negative Log Likelihood: 37.9881\tSigma2 Prior: -499.9118\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3150  \tTraining Loss: -516.0745    \n",
            "    Negative Log Likelihood: 30.3310\tSigma2 Prior: -546.4076\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3160  \tTraining Loss: -485.3160    \n",
            "    Negative Log Likelihood: 32.0514\tSigma2 Prior: -517.3696\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3170  \tTraining Loss: -502.3676    \n",
            "    Negative Log Likelihood: 32.1189\tSigma2 Prior: -534.4886\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3180  \tTraining Loss: -467.7535    \n",
            "    Negative Log Likelihood: 34.3104\tSigma2 Prior: -502.0660\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3190  \tTraining Loss: -503.2776    \n",
            "    Negative Log Likelihood: 29.8165\tSigma2 Prior: -533.0962\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3200  \tTraining Loss: -510.6639    \n",
            "    Negative Log Likelihood: 36.4866\tSigma2 Prior: -547.1527\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3210  \tTraining Loss: -501.9718    \n",
            "    Negative Log Likelihood: 31.4928\tSigma2 Prior: -533.4668\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3220  \tTraining Loss: -518.6299    \n",
            "    Negative Log Likelihood: 33.1786\tSigma2 Prior: -551.8108\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3230  \tTraining Loss: -478.3269    \n",
            "    Negative Log Likelihood: 34.3587\tSigma2 Prior: -512.6877\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3240  \tTraining Loss: -492.3845    \n",
            "    Negative Log Likelihood: 35.3398\tSigma2 Prior: -527.7264\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3250  \tTraining Loss: -468.4552    \n",
            "    Negative Log Likelihood: 35.5410\tSigma2 Prior: -503.9984\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3260  \tTraining Loss: -508.1178    \n",
            "    Negative Log Likelihood: 36.0974\tSigma2 Prior: -544.2173\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3270  \tTraining Loss: -487.2239    \n",
            "    Negative Log Likelihood: 37.3610\tSigma2 Prior: -524.5870\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3280  \tTraining Loss: -514.3523    \n",
            "    Negative Log Likelihood: 29.5988\tSigma2 Prior: -543.9532\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3290  \tTraining Loss: -484.0908    \n",
            "    Negative Log Likelihood: 42.5220\tSigma2 Prior: -526.6149\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3300  \tTraining Loss: -472.8216    \n",
            "    Negative Log Likelihood: 31.2458\tSigma2 Prior: -504.0695\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3310  \tTraining Loss: -445.2252    \n",
            "    Negative Log Likelihood: 33.5344\tSigma2 Prior: -478.7618\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3320  \tTraining Loss: -486.2830    \n",
            "    Negative Log Likelihood: 42.9804\tSigma2 Prior: -529.2656\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3330  \tTraining Loss: -468.1328    \n",
            "    Negative Log Likelihood: 30.6909\tSigma2 Prior: -498.8259\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3340  \tTraining Loss: -482.9976    \n",
            "    Negative Log Likelihood: 33.7698\tSigma2 Prior: -516.7695\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3350  \tTraining Loss: -504.9796    \n",
            "    Negative Log Likelihood: 31.6785\tSigma2 Prior: -536.6604\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3360  \tTraining Loss: -500.9280    \n",
            "    Negative Log Likelihood: 38.4962\tSigma2 Prior: -539.4265\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3370  \tTraining Loss: -524.4235    \n",
            "    Negative Log Likelihood: 36.2744\tSigma2 Prior: -560.7001\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3380  \tTraining Loss: -529.0453    \n",
            "    Negative Log Likelihood: 42.8243\tSigma2 Prior: -571.8718\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3390  \tTraining Loss: -488.8758    \n",
            "    Negative Log Likelihood: 29.1526\tSigma2 Prior: -518.0306\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3400  \tTraining Loss: -490.9486    \n",
            "    Negative Log Likelihood: 31.0389\tSigma2 Prior: -521.9897\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3410  \tTraining Loss: -471.2770    \n",
            "    Negative Log Likelihood: 41.4233\tSigma2 Prior: -512.7025\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3420  \tTraining Loss: -493.7870    \n",
            "    Negative Log Likelihood: 31.3547\tSigma2 Prior: -525.1439\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3430  \tTraining Loss: -495.8432    \n",
            "    Negative Log Likelihood: 32.4001\tSigma2 Prior: -528.2455\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3440  \tTraining Loss: -473.4520    \n",
            "    Negative Log Likelihood: 33.8253\tSigma2 Prior: -507.2795\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3450  \tTraining Loss: -465.2935    \n",
            "    Negative Log Likelihood: 35.2631\tSigma2 Prior: -500.5588\tRegularization: 0.0022\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3460  \tTraining Loss: -465.1237    \n",
            "    Negative Log Likelihood: 33.8901\tSigma2 Prior: -499.0161\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3470  \tTraining Loss: -507.9213    \n",
            "    Negative Log Likelihood: 35.8046\tSigma2 Prior: -543.7281\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3480  \tTraining Loss: -479.7274    \n",
            "    Negative Log Likelihood: 32.8588\tSigma2 Prior: -512.5885\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3490  \tTraining Loss: -506.4076    \n",
            "    Negative Log Likelihood: 30.8057\tSigma2 Prior: -537.2155\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3500  \tTraining Loss: -485.2296    \n",
            "    Negative Log Likelihood: 39.1575\tSigma2 Prior: -524.3894\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3510  \tTraining Loss: -491.4626    \n",
            "    Negative Log Likelihood: 29.3497\tSigma2 Prior: -520.8145\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3520  \tTraining Loss: -487.8568    \n",
            "    Negative Log Likelihood: 36.1034\tSigma2 Prior: -523.9625\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3530  \tTraining Loss: -480.1962    \n",
            "    Negative Log Likelihood: 34.3666\tSigma2 Prior: -514.5651\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3540  \tTraining Loss: -499.7120    \n",
            "    Negative Log Likelihood: 27.7731\tSigma2 Prior: -527.4874\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3550  \tTraining Loss: -485.5559    \n",
            "    Negative Log Likelihood: 42.4311\tSigma2 Prior: -527.9893\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3560  \tTraining Loss: -508.7781    \n",
            "    Negative Log Likelihood: 36.3953\tSigma2 Prior: -545.1757\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3570  \tTraining Loss: -518.3438    \n",
            "    Negative Log Likelihood: 30.8182\tSigma2 Prior: -549.1642\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3580  \tTraining Loss: -478.1681    \n",
            "    Negative Log Likelihood: 39.0553\tSigma2 Prior: -517.2256\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3590  \tTraining Loss: -503.1750    \n",
            "    Negative Log Likelihood: 27.6221\tSigma2 Prior: -530.7994\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3600  \tTraining Loss: -470.8617    \n",
            "    Negative Log Likelihood: 37.2905\tSigma2 Prior: -508.1545\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3610  \tTraining Loss: -506.9646    \n",
            "    Negative Log Likelihood: 30.1281\tSigma2 Prior: -537.0950\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3620  \tTraining Loss: -518.5574    \n",
            "    Negative Log Likelihood: 34.7330\tSigma2 Prior: -553.2927\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3630  \tTraining Loss: -497.3142    \n",
            "    Negative Log Likelihood: 39.6893\tSigma2 Prior: -537.0059\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3640  \tTraining Loss: -512.0290    \n",
            "    Negative Log Likelihood: 30.4539\tSigma2 Prior: -542.4852\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3650  \tTraining Loss: -468.6902    \n",
            "    Negative Log Likelihood: 40.8947\tSigma2 Prior: -509.5873\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3660  \tTraining Loss: -483.8046    \n",
            "    Negative Log Likelihood: 28.8735\tSigma2 Prior: -512.6804\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3670  \tTraining Loss: -484.3726    \n",
            "    Negative Log Likelihood: 37.0560\tSigma2 Prior: -521.4308\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3680  \tTraining Loss: -495.5125    \n",
            "    Negative Log Likelihood: 33.7101\tSigma2 Prior: -529.2250\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3690  \tTraining Loss: -505.7149    \n",
            "    Negative Log Likelihood: 31.5579\tSigma2 Prior: -537.2751\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3700  \tTraining Loss: -472.2559    \n",
            "    Negative Log Likelihood: 39.0799\tSigma2 Prior: -511.3382\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3710  \tTraining Loss: -500.9451    \n",
            "    Negative Log Likelihood: 33.9630\tSigma2 Prior: -534.9105\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3720  \tTraining Loss: -487.1030    \n",
            "    Negative Log Likelihood: 35.0748\tSigma2 Prior: -522.1801\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3730  \tTraining Loss: -519.4761    \n",
            "    Negative Log Likelihood: 32.4447\tSigma2 Prior: -551.9232\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3740  \tTraining Loss: -455.4962    \n",
            "    Negative Log Likelihood: 34.9369\tSigma2 Prior: -490.4354\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3750  \tTraining Loss: -427.9712    \n",
            "    Negative Log Likelihood: 40.2278\tSigma2 Prior: -468.2013\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3760  \tTraining Loss: -510.6772    \n",
            "    Negative Log Likelihood: 30.6334\tSigma2 Prior: -541.3129\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3770  \tTraining Loss: -466.3221    \n",
            "    Negative Log Likelihood: 40.5797\tSigma2 Prior: -506.9041\tRegularization: 0.0023\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3780  \tTraining Loss: -485.2470    \n",
            "    Negative Log Likelihood: 27.6019\tSigma2 Prior: -512.8513\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3790  \tTraining Loss: -427.8607    \n",
            "    Negative Log Likelihood: 34.7515\tSigma2 Prior: -462.6145\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3800  \tTraining Loss: -484.2589    \n",
            "    Negative Log Likelihood: 37.1351\tSigma2 Prior: -521.3964\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3810  \tTraining Loss: -480.0855    \n",
            "    Negative Log Likelihood: 34.5453\tSigma2 Prior: -514.6332\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3820  \tTraining Loss: -493.5281    \n",
            "    Negative Log Likelihood: 39.8497\tSigma2 Prior: -533.3802\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3830  \tTraining Loss: -520.8099    \n",
            "    Negative Log Likelihood: 32.7386\tSigma2 Prior: -553.5510\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3840  \tTraining Loss: -485.4052    \n",
            "    Negative Log Likelihood: 36.1796\tSigma2 Prior: -521.5872\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3850  \tTraining Loss: -482.3124    \n",
            "    Negative Log Likelihood: 33.1686\tSigma2 Prior: -515.4834\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3860  \tTraining Loss: -470.8318    \n",
            "    Negative Log Likelihood: 38.1062\tSigma2 Prior: -508.9404\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3870  \tTraining Loss: -486.2938    \n",
            "    Negative Log Likelihood: 30.9893\tSigma2 Prior: -517.2855\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3880  \tTraining Loss: -479.2754    \n",
            "    Negative Log Likelihood: 38.6831\tSigma2 Prior: -517.9609\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3890  \tTraining Loss: -498.8934    \n",
            "    Negative Log Likelihood: 33.2176\tSigma2 Prior: -532.1134\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3900  \tTraining Loss: -495.8357    \n",
            "    Negative Log Likelihood: 34.4872\tSigma2 Prior: -530.3253\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3910  \tTraining Loss: -485.4881    \n",
            "    Negative Log Likelihood: 32.0291\tSigma2 Prior: -517.5197\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3920  \tTraining Loss: -489.3958    \n",
            "    Negative Log Likelihood: 28.3216\tSigma2 Prior: -517.7197\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3930  \tTraining Loss: -516.8892    \n",
            "    Negative Log Likelihood: 39.3572\tSigma2 Prior: -556.2488\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3940  \tTraining Loss: -485.5028    \n",
            "    Negative Log Likelihood: 31.5923\tSigma2 Prior: -517.0975\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3950  \tTraining Loss: -458.0181    \n",
            "    Negative Log Likelihood: 38.7703\tSigma2 Prior: -496.7908\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3960  \tTraining Loss: -487.5587    \n",
            "    Negative Log Likelihood: 36.1299\tSigma2 Prior: -523.6910\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3970  \tTraining Loss: -508.0588    \n",
            "    Negative Log Likelihood: 36.7713\tSigma2 Prior: -544.8325\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3980  \tTraining Loss: -467.4016    \n",
            "    Negative Log Likelihood: 35.0144\tSigma2 Prior: -502.4185\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 3990  \tTraining Loss: -522.0236    \n",
            "    Negative Log Likelihood: 33.1638\tSigma2 Prior: -555.1899\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4000  \tTraining Loss: -525.3102    \n",
            "    Negative Log Likelihood: 37.4280\tSigma2 Prior: -562.7407\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4010  \tTraining Loss: -483.5207    \n",
            "    Negative Log Likelihood: 31.4368\tSigma2 Prior: -514.9599\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4020  \tTraining Loss: -509.4027    \n",
            "    Negative Log Likelihood: 35.5135\tSigma2 Prior: -544.9187\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4030  \tTraining Loss: -525.7238    \n",
            "    Negative Log Likelihood: 34.1727\tSigma2 Prior: -559.8989\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4040  \tTraining Loss: -491.4323    \n",
            "    Negative Log Likelihood: 33.9223\tSigma2 Prior: -525.3571\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4050  \tTraining Loss: -488.8731    \n",
            "    Negative Log Likelihood: 34.6491\tSigma2 Prior: -523.5247\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4060  \tTraining Loss: -490.8303    \n",
            "    Negative Log Likelihood: 34.9857\tSigma2 Prior: -525.8185\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4070  \tTraining Loss: -516.9814    \n",
            "    Negative Log Likelihood: 36.4179\tSigma2 Prior: -553.4017\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4080  \tTraining Loss: -467.0759    \n",
            "    Negative Log Likelihood: 38.9337\tSigma2 Prior: -506.0121\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4090  \tTraining Loss: -489.5915    \n",
            "    Negative Log Likelihood: 37.7960\tSigma2 Prior: -527.3899\tRegularization: 0.0024\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4100  \tTraining Loss: -441.4938    \n",
            "    Negative Log Likelihood: 40.4955\tSigma2 Prior: -481.9918\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4110  \tTraining Loss: -482.6979    \n",
            "    Negative Log Likelihood: 31.2004\tSigma2 Prior: -513.9008\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4120  \tTraining Loss: -493.8609    \n",
            "    Negative Log Likelihood: 41.6407\tSigma2 Prior: -535.5041\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4130  \tTraining Loss: -489.1706    \n",
            "    Negative Log Likelihood: 30.2881\tSigma2 Prior: -519.4612\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4140  \tTraining Loss: -484.5410    \n",
            "    Negative Log Likelihood: 33.7958\tSigma2 Prior: -518.3393\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4150  \tTraining Loss: -513.8804    \n",
            "    Negative Log Likelihood: 35.2302\tSigma2 Prior: -549.1130\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4160  \tTraining Loss: -503.5212    \n",
            "    Negative Log Likelihood: 31.6067\tSigma2 Prior: -535.1304\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4170  \tTraining Loss: -466.3871    \n",
            "    Negative Log Likelihood: 35.1492\tSigma2 Prior: -501.5387\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4180  \tTraining Loss: -482.5613    \n",
            "    Negative Log Likelihood: 31.1785\tSigma2 Prior: -513.7423\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4190  \tTraining Loss: -501.4665    \n",
            "    Negative Log Likelihood: 31.7778\tSigma2 Prior: -533.2467\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4200  \tTraining Loss: -481.5886    \n",
            "    Negative Log Likelihood: 39.0488\tSigma2 Prior: -520.6399\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4210  \tTraining Loss: -487.3419    \n",
            "    Negative Log Likelihood: 25.3550\tSigma2 Prior: -512.6993\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4220  \tTraining Loss: -496.5957    \n",
            "    Negative Log Likelihood: 35.5066\tSigma2 Prior: -532.1048\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4230  \tTraining Loss: -503.4322    \n",
            "    Negative Log Likelihood: 30.0322\tSigma2 Prior: -533.4670\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4240  \tTraining Loss: -425.7310    \n",
            "    Negative Log Likelihood: 31.8831\tSigma2 Prior: -457.6166\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4250  \tTraining Loss: -493.2435    \n",
            "    Negative Log Likelihood: 33.3021\tSigma2 Prior: -526.5481\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4260  \tTraining Loss: -467.5518    \n",
            "    Negative Log Likelihood: 42.6964\tSigma2 Prior: -510.2507\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4270  \tTraining Loss: -502.7360    \n",
            "    Negative Log Likelihood: 28.2772\tSigma2 Prior: -531.0157\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4280  \tTraining Loss: -490.0279    \n",
            "    Negative Log Likelihood: 30.9981\tSigma2 Prior: -521.0285\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4290  \tTraining Loss: -501.7215    \n",
            "    Negative Log Likelihood: 41.6612\tSigma2 Prior: -543.3852\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4300  \tTraining Loss: -489.8723    \n",
            "    Negative Log Likelihood: 34.2461\tSigma2 Prior: -524.1210\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4310  \tTraining Loss: -453.9871    \n",
            "    Negative Log Likelihood: 33.2990\tSigma2 Prior: -487.2886\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4320  \tTraining Loss: -499.0924    \n",
            "    Negative Log Likelihood: 32.3444\tSigma2 Prior: -531.4393\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4330  \tTraining Loss: -478.3274    \n",
            "    Negative Log Likelihood: 33.5187\tSigma2 Prior: -511.8486\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4340  \tTraining Loss: -478.5274    \n",
            "    Negative Log Likelihood: 35.5358\tSigma2 Prior: -514.0657\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4350  \tTraining Loss: -466.4791    \n",
            "    Negative Log Likelihood: 36.7337\tSigma2 Prior: -503.2153\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4360  \tTraining Loss: -495.0162    \n",
            "    Negative Log Likelihood: 32.8205\tSigma2 Prior: -527.8392\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4370  \tTraining Loss: -488.0616    \n",
            "    Negative Log Likelihood: 39.3460\tSigma2 Prior: -527.4102\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4380  \tTraining Loss: -496.8979    \n",
            "    Negative Log Likelihood: 33.0710\tSigma2 Prior: -529.9714\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4390  \tTraining Loss: -482.2334    \n",
            "    Negative Log Likelihood: 30.5983\tSigma2 Prior: -512.8342\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4400  \tTraining Loss: -490.9920    \n",
            "    Negative Log Likelihood: 32.2814\tSigma2 Prior: -523.2759\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4410  \tTraining Loss: -449.6325    \n",
            "    Negative Log Likelihood: 35.3293\tSigma2 Prior: -484.9644\tRegularization: 0.0025\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4420  \tTraining Loss: -510.1934    \n",
            "    Negative Log Likelihood: 33.4805\tSigma2 Prior: -543.6765\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4430  \tTraining Loss: -494.8135    \n",
            "    Negative Log Likelihood: 38.3058\tSigma2 Prior: -533.1219\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4440  \tTraining Loss: -477.6885    \n",
            "    Negative Log Likelihood: 32.6050\tSigma2 Prior: -510.2961\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4450  \tTraining Loss: -472.7169    \n",
            "    Negative Log Likelihood: 37.5189\tSigma2 Prior: -510.2384\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4460  \tTraining Loss: -513.7725    \n",
            "    Negative Log Likelihood: 29.0962\tSigma2 Prior: -542.8713\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4470  \tTraining Loss: -501.9709    \n",
            "    Negative Log Likelihood: 35.1866\tSigma2 Prior: -537.1600\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4480  \tTraining Loss: -493.0634    \n",
            "    Negative Log Likelihood: 34.3409\tSigma2 Prior: -527.4069\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4490  \tTraining Loss: -506.2090    \n",
            "    Negative Log Likelihood: 32.5378\tSigma2 Prior: -538.7493\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4500  \tTraining Loss: -479.1488    \n",
            "    Negative Log Likelihood: 36.4411\tSigma2 Prior: -515.5925\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4510  \tTraining Loss: -490.6097    \n",
            "    Negative Log Likelihood: 35.9942\tSigma2 Prior: -526.6064\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4520  \tTraining Loss: -481.6920    \n",
            "    Negative Log Likelihood: 35.8000\tSigma2 Prior: -517.4946\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4530  \tTraining Loss: -488.0507    \n",
            "    Negative Log Likelihood: 31.7410\tSigma2 Prior: -519.7943\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4540  \tTraining Loss: -503.8962    \n",
            "    Negative Log Likelihood: 37.2952\tSigma2 Prior: -541.1940\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4550  \tTraining Loss: -507.8309    \n",
            "    Negative Log Likelihood: 34.0615\tSigma2 Prior: -541.8950\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4560  \tTraining Loss: -505.5293    \n",
            "    Negative Log Likelihood: 33.1395\tSigma2 Prior: -538.6713\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4570  \tTraining Loss: -489.2876    \n",
            "    Negative Log Likelihood: 32.9105\tSigma2 Prior: -522.2006\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4580  \tTraining Loss: -491.5067    \n",
            "    Negative Log Likelihood: 39.3997\tSigma2 Prior: -530.9090\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4590  \tTraining Loss: -510.4799    \n",
            "    Negative Log Likelihood: 28.8819\tSigma2 Prior: -539.3644\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4600  \tTraining Loss: -496.4625    \n",
            "    Negative Log Likelihood: 41.0807\tSigma2 Prior: -537.5458\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4610  \tTraining Loss: -449.7040    \n",
            "    Negative Log Likelihood: 32.1432\tSigma2 Prior: -481.8499\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4620  \tTraining Loss: -482.9901    \n",
            "    Negative Log Likelihood: 30.0797\tSigma2 Prior: -513.0724\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4630  \tTraining Loss: -512.6724    \n",
            "    Negative Log Likelihood: 39.4908\tSigma2 Prior: -552.1658\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4640  \tTraining Loss: -454.2380    \n",
            "    Negative Log Likelihood: 32.9094\tSigma2 Prior: -487.1500\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4650  \tTraining Loss: -496.8252    \n",
            "    Negative Log Likelihood: 38.5658\tSigma2 Prior: -535.3937\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4660  \tTraining Loss: -505.7715    \n",
            "    Negative Log Likelihood: 28.4083\tSigma2 Prior: -534.1824\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4670  \tTraining Loss: -506.0402    \n",
            "    Negative Log Likelihood: 31.9863\tSigma2 Prior: -538.0292\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4680  \tTraining Loss: -498.3374    \n",
            "    Negative Log Likelihood: 35.9325\tSigma2 Prior: -534.2725\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4690  \tTraining Loss: -494.6064    \n",
            "    Negative Log Likelihood: 27.3501\tSigma2 Prior: -521.9591\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4700  \tTraining Loss: -523.3301    \n",
            "    Negative Log Likelihood: 37.5444\tSigma2 Prior: -560.8772\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4710  \tTraining Loss: -518.4112    \n",
            "    Negative Log Likelihood: 33.1856\tSigma2 Prior: -551.5994\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4720  \tTraining Loss: -476.7811    \n",
            "    Negative Log Likelihood: 27.9172\tSigma2 Prior: -504.7009\tRegularization: 0.0026\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4730  \tTraining Loss: -520.6341    \n",
            "    Negative Log Likelihood: 32.6173\tSigma2 Prior: -553.2541\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4740  \tTraining Loss: -459.6581    \n",
            "    Negative Log Likelihood: 36.7985\tSigma2 Prior: -496.4594\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4750  \tTraining Loss: -468.9312    \n",
            "    Negative Log Likelihood: 30.0658\tSigma2 Prior: -498.9996\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4760  \tTraining Loss: -481.1051    \n",
            "    Negative Log Likelihood: 30.5548\tSigma2 Prior: -511.6626\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4770  \tTraining Loss: -495.1274    \n",
            "    Negative Log Likelihood: 37.2674\tSigma2 Prior: -532.3975\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4780  \tTraining Loss: -499.1108    \n",
            "    Negative Log Likelihood: 33.5893\tSigma2 Prior: -532.7028\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4790  \tTraining Loss: -490.5771    \n",
            "    Negative Log Likelihood: 34.8220\tSigma2 Prior: -525.4017\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4800  \tTraining Loss: -489.4675    \n",
            "    Negative Log Likelihood: 36.5103\tSigma2 Prior: -525.9805\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4810  \tTraining Loss: -498.7029    \n",
            "    Negative Log Likelihood: 32.0573\tSigma2 Prior: -530.7629\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4820  \tTraining Loss: -463.7112    \n",
            "    Negative Log Likelihood: 31.8483\tSigma2 Prior: -495.5622\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4830  \tTraining Loss: -490.1550    \n",
            "    Negative Log Likelihood: 34.4426\tSigma2 Prior: -524.6002\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4840  \tTraining Loss: -502.3331    \n",
            "    Negative Log Likelihood: 31.8856\tSigma2 Prior: -534.2214\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4850  \tTraining Loss: -503.2429    \n",
            "    Negative Log Likelihood: 35.7253\tSigma2 Prior: -538.9709\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4860  \tTraining Loss: -477.9863    \n",
            "    Negative Log Likelihood: 34.3668\tSigma2 Prior: -512.3558\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4870  \tTraining Loss: -490.7841    \n",
            "    Negative Log Likelihood: 31.4858\tSigma2 Prior: -522.2726\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4880  \tTraining Loss: -475.5663    \n",
            "    Negative Log Likelihood: 40.1345\tSigma2 Prior: -515.7035\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4890  \tTraining Loss: -497.7273    \n",
            "    Negative Log Likelihood: 29.5652\tSigma2 Prior: -527.2953\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4900  \tTraining Loss: -467.7719    \n",
            "    Negative Log Likelihood: 35.7611\tSigma2 Prior: -503.5358\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4910  \tTraining Loss: -493.9006    \n",
            "    Negative Log Likelihood: 37.0967\tSigma2 Prior: -531.0000\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4920  \tTraining Loss: -455.2005    \n",
            "    Negative Log Likelihood: 36.7447\tSigma2 Prior: -491.9480\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4930  \tTraining Loss: -478.9658    \n",
            "    Negative Log Likelihood: 29.1987\tSigma2 Prior: -508.1672\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4940  \tTraining Loss: -443.1084    \n",
            "    Negative Log Likelihood: 37.2927\tSigma2 Prior: -480.4038\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4950  \tTraining Loss: -484.7807    \n",
            "    Negative Log Likelihood: 32.2222\tSigma2 Prior: -517.0056\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4960  \tTraining Loss: -512.9415    \n",
            "    Negative Log Likelihood: 36.9494\tSigma2 Prior: -549.8936\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4970  \tTraining Loss: -523.3010    \n",
            "    Negative Log Likelihood: 33.1592\tSigma2 Prior: -556.4630\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4980  \tTraining Loss: -511.4314    \n",
            "    Negative Log Likelihood: 38.8163\tSigma2 Prior: -550.2504\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4990  \tTraining Loss: -510.4946    \n",
            "    Negative Log Likelihood: 32.0143\tSigma2 Prior: -542.5117\tRegularization: 0.0027\n",
            "\u001b[93m[WARNING]\u001b[0m Iter: 4999  \tTraining Loss: -497.0769    \n",
            "    Negative Log Likelihood: 30.7117\tSigma2 Prior: -527.7914\tRegularization: 0.0027\n",
            "\u001b[94m[ERROR]\u001b[0m Done training with 5000 iterations\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "iQ4iuIT34wmI"
      },
      "source": [
        "## 모델 평가"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wcct4Q_7APML"
      },
      "source": [
        "\n",
        "\n",
        "*   평가 데이터를 사용해 각 sequence에 해당하는 화자 id를 얻음\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "vuZtArnpqyTz",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a480d747-8a1e-485c-ceb9-dcd539a4e25c"
      },
      "source": [
        "predicted_cluster_ids=[]\n",
        "test_record = []\n",
        "\n",
        "for(test_sequence, test_cluster_id) in zip(test_sequences, test_cluster_ids):\n",
        "  predicted_cluster_id = model.predict(test_sequence, inference_args)\n",
        "  predicted_cluster_ids.append(predicted_cluster_id)\n",
        "  accuracy = uisrnn.compute_sequence_match_accuracy(test_cluster_id, predicted_cluster_id)\n",
        "  test_record.append((accuracy, len(test_cluster_id)))\n",
        "\n",
        "  print('Ground truth labels: ')\n",
        "  print(test_cluster_id)\n",
        "  print('Predicted labels:')\n",
        "  print(predicted_cluster_id)\n",
        "  print('-' * 100)\n",
        "\n",
        "output_result = uisrnn.output_result(model_args, training_args, test_record)\n",
        "print(output_result)"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Ground truth labels: \n",
            "['15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_2', '15_2', '15_2', '15_2', '15_2', '15_2', '15_2', '15_2', '15_2', '15_2', '15_2', '15_2', '15_0', '15_0', '15_0', '15_0', '15_0', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_2', '15_2', '15_2', '15_2', '15_2', '15_2', '15_2', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_0', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1', '15_1']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_1', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_0', '30_1']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_0', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_2', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1', '52_1']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_2', '71_2', '71_2', '71_2', '71_2', '71_2', '71_2', '71_2', '71_2', '71_2', '71_2', '71_1', '71_1', '71_1', '71_1', '71_1', '71_1', '71_1', '71_1', '71_1', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_1', '71_1', '71_1', '71_1', '71_1', '71_1', '71_1', '71_0', '71_0', '71_0', '71_0', '71_0', '71_0', '71_1', '71_1', '71_1', '71_1', '71_1', '71_1', '71_1', '71_1', '71_1', '71_1', '71_2', '71_2', '71_2', '71_2', '71_2', '71_2', '71_1', '71_1', '71_1', '71_1', '71_1', '71_1']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_2', '75_2', '75_2', '75_2', '75_2', '75_2', '75_2', '75_2', '75_2', '75_2', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_0', '75_3', '75_3', '75_3', '75_3', '75_3', '75_3', '75_3', '75_3', '75_4', '75_4', '75_4', '75_4', '75_4', '75_4', '75_4', '75_4', '75_2', '75_2', '75_2', '75_2', '75_2', '75_4', '75_4', '75_4', '75_4', '75_4', '75_4', '75_4', '75_4', '75_4', '75_4', '75_4', '75_4', '75_4', '75_3', '75_3', '75_3', '75_3', '75_3', '75_1', '75_1', '75_1', '75_1', '75_1', '75_1', '75_1', '75_1', '75_2', '75_2', '75_2']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 2, 2, 2, 2, 2, 4, 4, 4, 4, 4, 4, 4, 4, 1, 1, 1]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_5', '80_5', '80_5', '80_5', '80_5', '80_5', '80_5', '80_5', '80_5', '80_5', '80_5', '80_5', '80_5', '80_4', '80_4', '80_4', '80_4', '80_4', '80_4', '80_4', '80_4', '80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_0', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_3', '80_4', '80_4', '80_4', '80_4', '80_4', '80_4', '80_4', '80_4', '80_4', '80_4', '80_4', '80_4', '80_4', '80_4']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_1', '140_1', '140_1', '140_1', '140_1', '140_1', '140_1', '140_1', '140_1', '140_1', '140_1', '140_1', '140_1', '140_1', '140_1', '140_1', '140_1', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_2', '140_2', '140_2', '140_2', '140_2', '140_5', '140_5', '140_5', '140_5', '140_5', '140_5', '140_5', '140_5', '140_5', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_0', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4', '140_4']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 4, 4, 4, 4, 4, 4, 4, 4, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['146_0', '146_0', '146_0', '146_0', '146_0', '146_0', '146_0', '146_0', '146_0', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_0', '146_0', '146_0', '146_0', '146_0', '146_0', '146_0', '146_0', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_0', '146_0', '146_0', '146_0', '146_0', '146_0', '146_0', '146_0', '146_0', '146_0', '146_0', '146_0', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1', '146_1']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_4', '152_4', '152_4', '152_4', '152_4', '152_4', '152_4', '152_4', '152_4', '152_4', '152_4', '152_4', '152_4', '152_4', '152_4', '152_4', '152_2', '152_2', '152_2', '152_2', '152_2', '152_2', '152_2', '152_2', '152_2', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_1', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_2', '152_2', '152_2', '152_2', '152_2', '152_2', '152_2', '152_2', '152_2', '152_2', '152_2', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_0', '152_2', '152_2', '152_2', '152_2']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['175_0', '175_0', '175_0', '175_0', '175_0', '175_0', '175_0', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_1', '175_1', '175_1', '175_1', '175_1', '175_1', '175_1', '175_1', '175_1', '175_1', '175_1', '175_1', '175_1', '175_1', '175_1', '175_1', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_1', '175_1', '175_1', '175_1', '175_1', '175_1', '175_1', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_2', '175_0', '175_0', '175_0', '175_0', '175_0', '175_0', '175_0', '175_0', '175_0', '175_0', '175_1', '175_1', '175_1', '175_1']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 3, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['204_0', '204_0', '204_0', '204_0', '204_0', '204_0', '204_0', '204_0', '204_0', '204_0', '204_0', '204_0', '204_0', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_1', '204_1', '204_1', '204_1', '204_1', '204_1', '204_1', '204_1', '204_1', '204_1', '204_1', '204_1', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_1', '204_1', '204_1', '204_1', '204_1', '204_1', '204_1', '204_1', '204_1', '204_1', '204_0', '204_0', '204_0', '204_0', '204_0', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_3', '204_2', '204_2', '204_2', '204_2', '204_2', '204_2', '204_2', '204_2', '204_2', '204_2', '204_2', '204_0', '204_0', '204_0', '204_0', '204_0', '204_0', '204_0', '204_0', '204_0', '204_0']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 4, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['315_0', '315_0', '315_0', '315_0', '315_0', '315_0', '315_0', '315_0', '315_0', '315_0', '315_0', '315_0', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_4', '315_4', '315_4', '315_4', '315_4', '315_4', '315_4', '315_4', '315_4', '315_2', '315_2', '315_2', '315_2', '315_2', '315_2', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_1', '315_5', '315_5', '315_5', '315_5']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 4, 4, 4, 5]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_0', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1', '333_1']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0', '347_0']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['361_0', '361_0', '361_0', '361_0', '361_0', '361_0', '361_0', '361_0', '361_0', '361_0', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_0', '361_0', '361_0', '361_0', '361_0', '361_0', '361_0', '361_0', '361_0', '361_0', '361_0', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_2', '361_3', '361_3', '361_3', '361_3', '361_3', '361_3', '361_3', '361_3', '361_3', '361_3', '361_3', '361_3', '361_3', '361_3', '361_3', '361_3', '361_3', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1', '361_1']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_1', '374_1', '374_1', '374_1', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_0', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_1', '374_0', '374_0', '374_0']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_1', '405_1', '405_1', '405_1', '405_1', '405_1', '405_1', '405_1', '405_2', '405_2', '405_2', '405_2', '405_2', '405_2', '405_2', '405_2', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_1', '405_1', '405_1', '405_1', '405_1', '405_1', '405_1', '405_1', '405_1', '405_1', '405_1', '405_1', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_2', '405_2', '405_2', '405_2', '405_2', '405_2', '405_2', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_0', '405_1', '405_1', '405_1', '405_1']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 2, 3, 3, 3, 3, 3, 3, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 3, 3, 3, 3, 3, 3, 3, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['419_0', '419_0', '419_0', '419_0', '419_0', '419_0', '419_0', '419_0', '419_0', '419_0', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_4', '419_4', '419_4', '419_4', '419_4', '419_4', '419_4', '419_4', '419_4', '419_4', '419_4', '419_3', '419_3', '419_3', '419_3', '419_3', '419_3', '419_3', '419_3', '419_3', '419_3', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_2', '419_2', '419_2', '419_2', '419_2', '419_2', '419_2', '419_0', '419_0', '419_0', '419_0', '419_0', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_1', '419_6', '419_6', '419_6', '419_6', '419_6', '419_6', '419_6', '419_6', '419_6', '419_6', '419_6', '419_6', '419_6', '419_6', '419_4', '419_4', '419_4', '419_4', '419_4', '419_4']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 4, 4, 4, 4, 4, 4, 4, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 5, 2, 2, 2, 2, 2, 2]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0', '430_0']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['439_0', '439_0', '439_0', '439_0', '439_0', '439_0', '439_0', '439_0', '439_0', '439_2', '439_2', '439_2', '439_2', '439_2', '439_2', '439_2', '439_2', '439_2', '439_2', '439_4', '439_4', '439_4', '439_4', '439_4', '439_4', '439_4', '439_4', '439_4', '439_4', '439_4', '439_2', '439_2', '439_2', '439_2', '439_2', '439_3', '439_3', '439_3', '439_3', '439_3', '439_3', '439_3', '439_3', '439_3', '439_4', '439_4', '439_4', '439_4', '439_4', '439_4', '439_4', '439_4', '439_1', '439_1', '439_1', '439_1', '439_1', '439_1', '439_1', '439_1', '439_1', '439_1', '439_1', '439_1', '439_1', '439_1', '439_3', '439_3', '439_3', '439_3', '439_3', '439_3', '439_3', '439_3', '439_3', '439_3', '439_2', '439_2', '439_2', '439_2', '439_2', '439_1', '439_1', '439_1', '439_1', '439_1', '439_1', '439_1', '439_1', '439_1', '439_4', '439_4', '439_4', '439_4', '439_4', '439_4', '439_4', '439_4', '439_4']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 3, 3, 3, 3, 3, 3, 3, 3, 3, 2, 2, 2, 2, 2, 2, 2, 2, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 4, 3, 3, 3, 3, 3, 3, 3, 3, 3, 3, 1, 1, 1, 1, 1, 4, 4, 4, 4, 4, 4, 4, 4, 4, 2, 2, 2, 2, 2, 2, 2, 2, 2]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_1', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0', '441_0']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0', '442_0']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_0', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1', '459_1']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_0', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_3', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1', '491_1']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Ground truth labels: \n",
            "['494_0', '494_0', '494_0', '494_0', '494_0', '494_0', '494_0', '494_0', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_0', '494_0', '494_0', '494_0', '494_0', '494_0', '494_0', '494_0', '494_0', '494_0', '494_2', '494_2', '494_2', '494_2', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_2', '494_2', '494_2', '494_2', '494_2', '494_2', '494_2', '494_2', '494_2', '494_2', '494_0', '494_0', '494_0', '494_0', '494_0', '494_0', '494_0', '494_0', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_1', '494_0', '494_0']\n",
            "Predicted labels:\n",
            "[0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 2, 2, 2, 2, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 0, 0, 0, 0, 0, 0, 0, 0, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0]\n",
            "----------------------------------------------------------------------------------------------------\n",
            "Config:\n",
            "  sigma_alpha: 1.0\n",
            "  sigma_beta: 1.0\n",
            "  crp_alpha: 1.0\n",
            "  learning rate: 0.001\n",
            "  regularization: 1e-05\n",
            "  batch size: 10\n",
            "\n",
            "Performance:\n",
            "  averaged accuracy: 0.997492\n",
            "  accuracy numbers for all testing sequences:\n",
            "    1.000000\n",
            "    0.988764\n",
            "    0.989362\n",
            "    1.000000\n",
            "    1.000000\n",
            "    1.000000\n",
            "    1.000000\n",
            "    1.000000\n",
            "    1.000000\n",
            "    0.989362\n",
            "    0.989362\n",
            "    0.989796\n",
            "    1.000000\n",
            "    1.000000\n",
            "    1.000000\n",
            "    1.000000\n",
            "    0.990654\n",
            "    1.000000\n",
            "    1.000000\n",
            "    1.000000\n",
            "    1.000000\n",
            "    1.000000\n",
            "    1.000000\n",
            "    1.000000\n",
            "    1.000000\n",
            "================================================================================\n",
            "\n"
          ]
        }
      ]
    }
  ]
}