05: Question_Answering_with_a_Fine_Tuned_BERT.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "05: Question Answering with a Fine-Tuned BERT.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "1WThOUtpYvG-",
        "I__1ubvcZYow",
        "-hh6nkIdXq-O",
        "8UyBYNmeegGf"
      ],
      "toc_visible": true,
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "29ac2980dc2349f78db6d734cf17b4fd": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_29964cd8ab02415d831b3914e8794574",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_dbefc058781f46379c2fd792709549dc",
              "IPY_MODEL_c814a6f05226430e99ea50ba2a59f6e1"
            ]
          }
        },
        "29964cd8ab02415d831b3914e8794574": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "dbefc058781f46379c2fd792709549dc": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_885d91cc8ffc441eada95f311ff588fa",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 443,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 443,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_709e4518c5944174924b5b3db0593349"
          }
        },
        "c814a6f05226430e99ea50ba2a59f6e1": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_42db1f25455347dc8610cd3e01264316",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 443/443 [00:32&lt;00:00, 13.6B/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_6c5cdc6b73134ef094963937787ef24d"
          }
        },
        "885d91cc8ffc441eada95f311ff588fa": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "709e4518c5944174924b5b3db0593349": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "42db1f25455347dc8610cd3e01264316": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "6c5cdc6b73134ef094963937787ef24d": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "fa9e0a5f04f945d3a5140feaec6b49b9": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_f88a50e473e149c3a0b8eafaab72aef5",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_63a44d1174f84710b3d4eb9bcc50c507",
              "IPY_MODEL_151feec4b56049959e1d7fa0b29cb317"
            ]
          }
        },
        "f88a50e473e149c3a0b8eafaab72aef5": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "63a44d1174f84710b3d4eb9bcc50c507": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_59fb2af128414f0fad93611962d52295",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 1340675298,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 1340675298,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_6812ae8dd01f4f5a8cc0af6aca6a8c6e"
          }
        },
        "151feec4b56049959e1d7fa0b29cb317": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_dbac267c5cec4320885455cfd8ae3d86",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 1.34G/1.34G [00:21&lt;00:00, 61.5MB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_e4b76e02cfcb401dab35cea9b671acb4"
          }
        },
        "59fb2af128414f0fad93611962d52295": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "6812ae8dd01f4f5a8cc0af6aca6a8c6e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "dbac267c5cec4320885455cfd8ae3d86": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "e4b76e02cfcb401dab35cea9b671acb4": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "9972c013096a483c89f8e5397bebb31b": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "state": {
            "_view_name": "HBoxView",
            "_dom_classes": [],
            "_model_name": "HBoxModel",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "box_style": "",
            "layout": "IPY_MODEL_9e35d916fba846959d0fe0c603254722",
            "_model_module": "@jupyter-widgets/controls",
            "children": [
              "IPY_MODEL_7b927aa0007b44919806b68540277956",
              "IPY_MODEL_b4ee4ebacb1341debc21ff8a65a2a56e"
            ]
          }
        },
        "9e35d916fba846959d0fe0c603254722": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "7b927aa0007b44919806b68540277956": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "state": {
            "_view_name": "ProgressView",
            "style": "IPY_MODEL_c36acbe699534fb98ced22cf336dfb52",
            "_dom_classes": [],
            "description": "Downloading: 100%",
            "_model_name": "FloatProgressModel",
            "bar_style": "success",
            "max": 231508,
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": 231508,
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "orientation": "horizontal",
            "min": 0,
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_5fe3fd0b6ab0432a8edd776b5ca97d5e"
          }
        },
        "b4ee4ebacb1341debc21ff8a65a2a56e": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "state": {
            "_view_name": "HTMLView",
            "style": "IPY_MODEL_2816beedbfca41bdae557a9385a155ea",
            "_dom_classes": [],
            "description": "",
            "_model_name": "HTMLModel",
            "placeholder": "​",
            "_view_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "value": " 232k/232k [00:00&lt;00:00, 1.44MB/s]",
            "_view_count": null,
            "_view_module_version": "1.5.0",
            "description_tooltip": null,
            "_model_module": "@jupyter-widgets/controls",
            "layout": "IPY_MODEL_7e57a24b23af4511bc88d93131c07031"
          }
        },
        "c36acbe699534fb98ced22cf336dfb52": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "ProgressStyleModel",
            "description_width": "initial",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "bar_color": null,
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "5fe3fd0b6ab0432a8edd776b5ca97d5e": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        },
        "2816beedbfca41bdae557a9385a155ea": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "state": {
            "_view_name": "StyleView",
            "_model_name": "DescriptionStyleModel",
            "description_width": "",
            "_view_module": "@jupyter-widgets/base",
            "_model_module_version": "1.5.0",
            "_view_count": null,
            "_view_module_version": "1.2.0",
            "_model_module": "@jupyter-widgets/controls"
          }
        },
        "7e57a24b23af4511bc88d93131c07031": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "state": {
            "_view_name": "LayoutView",
            "grid_template_rows": null,
            "right": null,
            "justify_content": null,
            "_view_module": "@jupyter-widgets/base",
            "overflow": null,
            "_model_module_version": "1.2.0",
            "_view_count": null,
            "flex_flow": null,
            "width": null,
            "min_width": null,
            "border": null,
            "align_items": null,
            "bottom": null,
            "_model_module": "@jupyter-widgets/base",
            "top": null,
            "grid_column": null,
            "overflow_y": null,
            "overflow_x": null,
            "grid_auto_flow": null,
            "grid_area": null,
            "grid_template_columns": null,
            "flex": null,
            "_model_name": "LayoutModel",
            "justify_items": null,
            "grid_row": null,
            "max_height": null,
            "align_content": null,
            "visibility": null,
            "align_self": null,
            "height": null,
            "min_height": null,
            "padding": null,
            "grid_auto_rows": null,
            "grid_gap": null,
            "max_width": null,
            "order": null,
            "_view_module_version": "1.2.0",
            "grid_template_areas": null,
            "object_position": null,
            "object_fit": null,
            "grid_auto_columns": null,
            "margin": null,
            "display": null,
            "left": null
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/gulabpatel/NLP_BERT/blob/main/05%3A%20Question_Answering_with_a_Fine_Tuned_BERT.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "W-1zl5XdYInf"
      },
      "source": [
        "# Question Answering with a Fine-Tuned BERT\n",
        "*by Chris McCormick*"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6_mAnIPKaXyw"
      },
      "source": [
        "What does it mean for BERT to achieve \"human-level performance on Question Answering\"? Is BERT the greatest search engine ever, able to find the answer to any question we pose it?\n",
        "\n",
        "In **Part 1** of this post / notebook, I'll explain what it really means to apply BERT to QA, and illustrate the details.\n",
        "\n",
        "**Part 2** contains example code--we'll be downloading a model that's *already been fine-tuned* for question answering, and try it out on our own text! \n",
        "\n",
        "For something like text classification, you definitely want to fine-tune BERT on your own dataset. For question answering, however, it seems like you may be able to get decent results using a model that's already been fine-tuned on the SQuAD benchmark. In this Notebook, we'll do exactly that, and see that it performs well on text that wasn't in the SQuAD dataset.\n",
        "\n",
        "**Links**\n",
        "\n",
        "* My [video walkthrough](https://youtu.be/l8ZYCvgGu0o) on this topic. \n",
        "* The blog post version.\n",
        "* The [Colab Notebook](https://colab.research.google.com/drive/1uSlWtJdZmLrI3FCNIlUHFxwAJiSu2J0-)."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "X2bUvKUffHNY"
      },
      "source": [
        "# Part 1: How BERT is applied to Question Answering"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Su7fixBdiUex"
      },
      "source": [
        "## The SQuAD v1.1 Benchmark"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "bT5ESKDxfnLf"
      },
      "source": [
        "When someone mentions \"Question Answering\" as an application of BERT, what they are really referring to is applying BERT to the Stanford Question Answering Dataset (SQuAD).\n",
        "\n",
        "The task posed by the SQuAD benchmark is a little different than you might think. Given a question, and *a passage of text containing the answer*, BERT needs to highlight the \"span\" of text corresponding to the correct answer. \n",
        "\n",
        "The SQuAD homepage has a fantastic tool for exploring the questions and reference text for this dataset, and even shows the predictions made by top-performing models.\n",
        "\n",
        "For example, here are some [interesting examples](https://rajpurkar.github.io/SQuAD-explorer/explore/1.1/dev/Super_Bowl_50.html?model=r-net+%20(ensemble)%20(Microsoft%20Research%20Asia)&version=1.1) on the topic of Super Bowl 50.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_xN5f1bxf6K_"
      },
      "source": [
        "## BERT Input Format"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ctum5SK6f9uP"
      },
      "source": [
        "To feed a QA task into BERT, we pack both the question and the reference text into the input.\n",
        "\n",
        "![Input format for QA](http://www.mccormickml.com/assets/BERT/SQuAD/input_formatting.png)\n",
        "\n",
        "The two pieces of text are separated by the special `[SEP]` token. \n",
        "\n",
        "BERT also uses \"Segment Embeddings\" to differentiate the question from the reference text. These are simply two embeddings (for segments \"A\" and \"B\") that BERT learned, and which it adds to the token embeddings before feeding them into the input layer. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Kl6BECEXu4jH"
      },
      "source": [
        "[![BERT eBook Display Ad](http://www.mccormickml.com/assets/BERT/BERT_eBook_Display_Ad.png)](https://bit.ly/3isUyCt)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xs31dcrPg5Tg"
      },
      "source": [
        "## Start & End Token Classifiers"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lvOdUa9Wg-Uv"
      },
      "source": [
        "BERT needs to highlight a \"span\" of text containing the answer--this is represented as simply predicting which token marks the start of the answer, and which token marks the end.\n",
        "\n",
        "![Start token classification](http://www.mccormickml.com/assets/BERT/SQuAD/start_token_classification.png)\n",
        "\n",
        "For every token in the text, we feed its final embedding into the start token classifier. The start token classifier only has a single set of weights (represented by the blue \"start\" rectangle in the above illustration) which it applies to every word.\n",
        "\n",
        "After taking the dot product between the output embeddings and the 'start' weights, we apply the softmax activation to produce a probability distribution over all of the words. Whichever word has the highest probability of being the start token is the one that we pick.\n",
        "\n",
        "We repeat this process for the end token--we have a separate weight vector this.\n",
        "\n",
        "![End token classification](http://www.mccormickml.com/assets/BERT/SQuAD/end_token_classification.png)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "457VPa20fZzY"
      },
      "source": [
        "# Part 2: Example Code"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9wpCLgVCkki5"
      },
      "source": [
        "In the example code below, we'll be downloading a model that's *already been fine-tuned* for question answering, and try it out on our own text.\n",
        "\n",
        "If you do want to fine-tune on your own dataset, it is possible to fine-tune BERT for question answering yourself. See [run_squad.py](https://github.com/huggingface/transformers/blob/master/examples/run_squad.py) in the `transformers` library. However,you may find that the below \"fine-tuned-on-squad\" model already does a good job, even if your text is from a different domain. \n",
        "\n",
        "> Note: The example code in this Notebook is a commented and expanded version of the short example provided in the `transformers` documentation [here](https://huggingface.co/transformers/model_doc/bert.html?highlight=bertforquestionanswering#transformers.BertForQuestionAnswering)."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gVq-TuylYRDW"
      },
      "source": [
        "## 1. Install huggingface transformers library"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f9nhy3PzGQ44"
      },
      "source": [
        "This example uses the `transformers` [library](https://github.com/huggingface/transformers/) by huggingface. We'll start by installing the package."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "aQl0MMrOGIup",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 360
        },
        "outputId": "c82e42c7-5d33-43d4-83b6-b7427b305c65"
      },
      "source": [
        "!pip install transformers"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Requirement already satisfied: transformers in /usr/local/lib/python3.6/dist-packages (3.3.1)\n",
            "Requirement already satisfied: tqdm>=4.27 in /usr/local/lib/python3.6/dist-packages (from transformers) (4.41.1)\n",
            "Requirement already satisfied: filelock in /usr/local/lib/python3.6/dist-packages (from transformers) (3.0.12)\n",
            "Requirement already satisfied: sentencepiece!=0.1.92 in /usr/local/lib/python3.6/dist-packages (from transformers) (0.1.91)\n",
            "Requirement already satisfied: numpy in /usr/local/lib/python3.6/dist-packages (from transformers) (1.18.5)\n",
            "Requirement already satisfied: packaging in /usr/local/lib/python3.6/dist-packages (from transformers) (20.4)\n",
            "Requirement already satisfied: regex!=2019.12.17 in /usr/local/lib/python3.6/dist-packages (from transformers) (2019.12.20)\n",
            "Requirement already satisfied: dataclasses; python_version < \"3.7\" in /usr/local/lib/python3.6/dist-packages (from transformers) (0.7)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.6/dist-packages (from transformers) (2.23.0)\n",
            "Requirement already satisfied: sacremoses in /usr/local/lib/python3.6/dist-packages (from transformers) (0.0.43)\n",
            "Requirement already satisfied: tokenizers==0.8.1.rc2 in /usr/local/lib/python3.6/dist-packages (from transformers) (0.8.1rc2)\n",
            "Requirement already satisfied: pyparsing>=2.0.2 in /usr/local/lib/python3.6/dist-packages (from packaging->transformers) (2.4.7)\n",
            "Requirement already satisfied: six in /usr/local/lib/python3.6/dist-packages (from packaging->transformers) (1.15.0)\n",
            "Requirement already satisfied: idna<3,>=2.5 in /usr/local/lib/python3.6/dist-packages (from requests->transformers) (2.10)\n",
            "Requirement already satisfied: urllib3!=1.25.0,!=1.25.1,<1.26,>=1.21.1 in /usr/local/lib/python3.6/dist-packages (from requests->transformers) (1.24.3)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.6/dist-packages (from requests->transformers) (2020.6.20)\n",
            "Requirement already satisfied: chardet<4,>=3.0.2 in /usr/local/lib/python3.6/dist-packages (from requests->transformers) (3.0.4)\n",
            "Requirement already satisfied: joblib in /usr/local/lib/python3.6/dist-packages (from sacremoses->transformers) (0.16.0)\n",
            "Requirement already satisfied: click in /usr/local/lib/python3.6/dist-packages (from sacremoses->transformers) (7.1.2)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-ONLrgJK99TQ"
      },
      "source": [
        "import torch"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1WThOUtpYvG-"
      },
      "source": [
        "## 2. Load Fine-Tuned BERT-large"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AaweLnNXGhTY"
      },
      "source": [
        "For Question Answering we use the `BertForQuestionAnswering` class from the `transformers` library.\n",
        "\n",
        "This class supports fine-tuning, but for this example we will keep things simpler and load a BERT model that has already been fine-tuned for the SQuAD benchmark.\n",
        "\n",
        "The `transformers` library has a large collection of pre-trained models which you can reference by name and load easily. The full list is in their documentation [here](https://huggingface.co/transformers/pretrained_models.html).\n",
        "\n",
        "For Question Answering, they have a version of BERT-large that has already been fine-tuned for the SQuAD benchmark. \n",
        "\n",
        "BERT-large is really big... it has 24-layers and an embedding size of 1,024, for a total of 340M parameters! Altogether it is 1.34GB, so expect it to take a couple minutes to download to your Colab instance. \n",
        "\n",
        "(Note that this download is not using your own network bandwidth--it's between the Google instance and wherever the model is stored on the web).\n",
        "\n",
        "Note: I believe this model was trained on version 1 of SQuAD, since it's not outputting whether the question is \"impossible\" to answer from the text (which is part of the task in v2 of SQuAD).\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-Mnv95sX-U9K",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "referenced_widgets": [
            "29ac2980dc2349f78db6d734cf17b4fd",
            "29964cd8ab02415d831b3914e8794574",
            "dbefc058781f46379c2fd792709549dc",
            "c814a6f05226430e99ea50ba2a59f6e1",
            "885d91cc8ffc441eada95f311ff588fa",
            "709e4518c5944174924b5b3db0593349",
            "42db1f25455347dc8610cd3e01264316",
            "6c5cdc6b73134ef094963937787ef24d",
            "fa9e0a5f04f945d3a5140feaec6b49b9",
            "f88a50e473e149c3a0b8eafaab72aef5",
            "63a44d1174f84710b3d4eb9bcc50c507",
            "151feec4b56049959e1d7fa0b29cb317",
            "59fb2af128414f0fad93611962d52295",
            "6812ae8dd01f4f5a8cc0af6aca6a8c6e",
            "dbac267c5cec4320885455cfd8ae3d86",
            "e4b76e02cfcb401dab35cea9b671acb4"
          ]
        },
        "outputId": "38a89f75-60ab-40c5-ed96-2a27db370b08"
      },
      "source": [
        "from transformers import BertForQuestionAnswering\n",
        "\n",
        "model = BertForQuestionAnswering.from_pretrained('bert-large-uncased-whole-word-masking-finetuned-squad')\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "29ac2980dc2349f78db6d734cf17b4fd",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=443.0, style=ProgressStyle(description_…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "fa9e0a5f04f945d3a5140feaec6b49b9",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=1340675298.0, style=ProgressStyle(descr…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8imoOxoqGZ0h"
      },
      "source": [
        "Load the tokenizer as well. \n",
        "\n",
        "Side note: Apparently the vocabulary of this model is identicaly to the one in bert-base-uncased. You can load the tokenizer from `bert-base-uncased` and that works just as well."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SFQ5f7gv-RBH",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "referenced_widgets": [
            "9972c013096a483c89f8e5397bebb31b",
            "9e35d916fba846959d0fe0c603254722",
            "7b927aa0007b44919806b68540277956",
            "b4ee4ebacb1341debc21ff8a65a2a56e",
            "c36acbe699534fb98ced22cf336dfb52",
            "5fe3fd0b6ab0432a8edd776b5ca97d5e",
            "2816beedbfca41bdae557a9385a155ea",
            "7e57a24b23af4511bc88d93131c07031"
          ]
        },
        "outputId": "b12b4f0d-0a8e-435d-82a9-76d018ababce"
      },
      "source": [
        "from transformers import BertTokenizer\n",
        "\n",
        "tokenizer = BertTokenizer.from_pretrained('bert-large-uncased-whole-word-masking-finetuned-squad')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "application/vnd.jupyter.widget-view+json": {
              "model_id": "9972c013096a483c89f8e5397bebb31b",
              "version_minor": 0,
              "version_major": 2
            },
            "text/plain": [
              "HBox(children=(FloatProgress(value=0.0, description='Downloading', max=231508.0, style=ProgressStyle(descripti…"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "stream",
          "text": [
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "I__1ubvcZYow"
      },
      "source": [
        "## 3. Ask a Question"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "o8MQ7b-GJIcM"
      },
      "source": [
        "Now we're ready to feed in an example!\n",
        "\n",
        "A QA example consists of a question and a passage of text containing the answer to that question.\n",
        "\n",
        "Let's try an example using the text in this tutorial!"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "kWzZP4EN-Zxg"
      },
      "source": [
        "question = \"How many parameters does BERT-large have?\"\n",
        "answer_text = \"BERT-large is really big... it has 24-layers and an embedding size of 1,024, for a total of 340M parameters! Altogether it is 1.34GB, so expect it to take a couple minutes to download to your Colab instance.\""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "llLvxhScKLZn"
      },
      "source": [
        "We'll need to run the BERT tokenizer against both the `question` and the `answer_text`. To feed these into BERT, we actually concatenate them together and place the special [SEP] token in between.\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tYoX33CfKGsr",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e54c9130-8d19-4abe-d7d1-2d319692a87d"
      },
      "source": [
        "# Apply the tokenizer to the input text, treating them as a text-pair.\n",
        "input_ids = tokenizer.encode(question, answer_text)\n",
        "\n",
        "print('The input has a total of {:} tokens.'.format(len(input_ids)))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "The input has a total of 70 tokens.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pNRVuaKSNFG8"
      },
      "source": [
        "Just to see exactly what the tokenizer is doing, let's print out the tokens with their IDs."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Iow838yPNDTv",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "00ee7502-5cd9-4685-8dba-a1fa1429bd91"
      },
      "source": [
        "# BERT only needs the token IDs, but for the purpose of inspecting the \n",
        "# tokenizer's behavior, let's also get the token strings and display them.\n",
        "tokens = tokenizer.convert_ids_to_tokens(input_ids)\n",
        "\n",
        "# For each token and its id...\n",
        "for token, id in zip(tokens, input_ids):\n",
        "    \n",
        "    # If this is the [SEP] token, add some space around it to make it stand out.\n",
        "    if id == tokenizer.sep_token_id:\n",
        "        print('')\n",
        "    \n",
        "    # Print the token string and its ID in two columns.\n",
        "    print('{:<12} {:>6,}'.format(token, id))\n",
        "\n",
        "    if id == tokenizer.sep_token_id:\n",
        "        print('')\n",
        "    "
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "[CLS]           101\n",
            "how           2,129\n",
            "many          2,116\n",
            "parameters   11,709\n",
            "does          2,515\n",
            "bert         14,324\n",
            "-             1,011\n",
            "large         2,312\n",
            "have          2,031\n",
            "?             1,029\n",
            "\n",
            "[SEP]           102\n",
            "\n",
            "bert         14,324\n",
            "-             1,011\n",
            "large         2,312\n",
            "is            2,003\n",
            "really        2,428\n",
            "big           2,502\n",
            ".             1,012\n",
            ".             1,012\n",
            ".             1,012\n",
            "it            2,009\n",
            "has           2,038\n",
            "24            2,484\n",
            "-             1,011\n",
            "layers        9,014\n",
            "and           1,998\n",
            "an            2,019\n",
            "em            7,861\n",
            "##bed         8,270\n",
            "##ding        4,667\n",
            "size          2,946\n",
            "of            1,997\n",
            "1             1,015\n",
            ",             1,010\n",
            "02            6,185\n",
            "##4           2,549\n",
            ",             1,010\n",
            "for           2,005\n",
            "a             1,037\n",
            "total         2,561\n",
            "of            1,997\n",
            "340          16,029\n",
            "##m           2,213\n",
            "parameters   11,709\n",
            "!               999\n",
            "altogether   10,462\n",
            "it            2,009\n",
            "is            2,003\n",
            "1             1,015\n",
            ".             1,012\n",
            "34            4,090\n",
            "##gb         18,259\n",
            ",             1,010\n",
            "so            2,061\n",
            "expect        5,987\n",
            "it            2,009\n",
            "to            2,000\n",
            "take          2,202\n",
            "a             1,037\n",
            "couple        3,232\n",
            "minutes       2,781\n",
            "to            2,000\n",
            "download      8,816\n",
            "to            2,000\n",
            "your          2,115\n",
            "cola         15,270\n",
            "##b           2,497\n",
            "instance      6,013\n",
            ".             1,012\n",
            "\n",
            "[SEP]           102\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zm208EApN16k"
      },
      "source": [
        "We've concatenated the `question` and `answer_text` together, but BERT still needs a way to distinguish them. BERT has two special \"Segment\" embeddings, one for segment \"A\" and one for segment \"B\". Before the word embeddings go into the BERT layers, the segment A embedding needs to be added to the `question` tokens, and the segment B embedding needs to be added to each of the `answer_text` tokens. \n",
        "\n",
        "These additions are handled for us by the `transformer` library, and all we need to do is specify a '0' or '1' for each token. \n",
        "\n",
        "Note: In the `transformers` library, huggingface likes to call these `token_type_ids`, but I'm going with `segment_ids` since this seems clearer, and is consistent with the BERT paper."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "uJ7PRx2dKqFN"
      },
      "source": [
        "# Search the input_ids for the first instance of the `[SEP]` token.\n",
        "sep_index = input_ids.index(tokenizer.sep_token_id)\n",
        "\n",
        "# The number of segment A tokens includes the [SEP] token istelf.\n",
        "num_seg_a = sep_index + 1\n",
        "\n",
        "# The remainder are segment B.\n",
        "num_seg_b = len(input_ids) - num_seg_a\n",
        "\n",
        "# Construct the list of 0s and 1s.\n",
        "segment_ids = [0]*num_seg_a + [1]*num_seg_b\n",
        "\n",
        "# There should be a segment_id for every input token.\n",
        "assert len(segment_ids) == len(input_ids)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "a30sBTcqQv6X"
      },
      "source": [
        ">*Side Note: Where's the padding?*\n",
        ">\n",
        "> The original [example code](https://huggingface.co/transformers/model_doc/bert.html?highlight=bertforquestionanswering#transformers.BertForQuestionAnswering) does not perform any padding. I suspect that this is because we are only feeding in a *single example*. If we instead fed in a batch of examples, then we would need to pad or truncate all of the samples in the batch to a single length, and supply an attention mask to tell BERT to ignore the padding tokens. "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "CNwhEw0kQPBN"
      },
      "source": [
        "We're ready to feed our example into the model!\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DQiKr6Aw-YTg"
      },
      "source": [
        "# Run our example through the model.\n",
        "start_scores, end_scores = model(torch.tensor([input_ids]), # The tokens representing our input text.\n",
        "                                 token_type_ids=torch.tensor([segment_ids])) # The segment IDs to differentiate question from answer_text\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "mBdS_QkIbDzh"
      },
      "source": [
        "Now we can highlight the answer just by looking at the most probable start and end words. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "LeUQ44hAJmn9",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ba04b1ea-4b9b-481d-ef5b-3da370368f12"
      },
      "source": [
        "# Find the tokens with the highest `start` and `end` scores.\n",
        "answer_start = torch.argmax(start_scores)\n",
        "answer_end = torch.argmax(end_scores)\n",
        "\n",
        "# Combine the tokens in the answer and print it out.\n",
        "answer = ' '.join(tokens[answer_start:answer_end+1])\n",
        "\n",
        "print('Answer: \"' + answer + '\"')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Answer: \"340 ##m\"\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "twMUWmr2brRw"
      },
      "source": [
        "It got it right! Awesome :)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cERYCGKMbOXX"
      },
      "source": [
        "> *Side Note: It's a little naive to pick the highest scores for start and end--what if it predicts an end word that's before the start word?! The correct implementation is to pick the highest total score for which end >= start.*"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "N6j2znkwXYsn"
      },
      "source": [
        "With a little more effort, we can reconstruct any words that got broken down into subwords."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Khral6HZXCuI",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a80e8bd2-077e-4207-a4aa-ff6256b92010"
      },
      "source": [
        "# Start with the first token.\n",
        "answer = tokens[answer_start]\n",
        "\n",
        "# Select the remaining answer tokens and join them with whitespace.\n",
        "for i in range(answer_start + 1, answer_end + 1):\n",
        "    \n",
        "    # If it's a subword token, then recombine it with the previous token.\n",
        "    if tokens[i][0:2] == '##':\n",
        "        answer += tokens[i][2:]\n",
        "    \n",
        "    # Otherwise, add a space then the token.\n",
        "    else:\n",
        "        answer += ' ' + tokens[i]\n",
        "\n",
        "print('Answer: \"' + answer + '\"')"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Answer: \"340m\"\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-hh6nkIdXq-O"
      },
      "source": [
        "## 4. Visualizing Scores"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-hG2YCHYXtg-"
      },
      "source": [
        "I was curious to see what the scores were for all of the words. The following cells generate bar plots showing the start and end scores for every word in the input."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gkKFa73eJkPE"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "import seaborn as sns\n",
        "\n",
        "# Use plot styling from seaborn.\n",
        "sns.set(style='darkgrid')\n",
        "\n",
        "# Increase the plot size and font size.\n",
        "#sns.set(font_scale=1.5)\n",
        "plt.rcParams[\"figure.figsize\"] = (16,8)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "W7kazkb2iEuQ"
      },
      "source": [
        "Retrieve all of the start and end scores, and use all of the tokens as x-axis labels."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "C56AtMg2UBxN"
      },
      "source": [
        "# Pull the scores out of PyTorch Tensors and convert them to 1D numpy arrays.\n",
        "s_scores = start_scores.detach().numpy().flatten()\n",
        "e_scores = end_scores.detach().numpy().flatten()\n",
        "\n",
        "# We'll use the tokens as the x-axis labels. In order to do that, they all need\n",
        "# to be unique, so we'll add the token index to the end of each one.\n",
        "token_labels = []\n",
        "for (i, token) in enumerate(tokens):\n",
        "    token_labels.append('{:} - {:>2}'.format(token, i))\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "NIaW7RyTiLeu"
      },
      "source": [
        "Create a bar plot showing the score for every input word being the \"start\" word."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y6OAV1dL3-UB",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "18c1f57d-7296-4fbc-bcd3-c74668786fed"
      },
      "source": [
        "# Create a barplot showing the start word score for all of the tokens.\n",
        "ax = sns.barplot(x=token_labels, y=s_scores, ci=None)\n",
        "\n",
        "# Turn the xlabels vertical.\n",
        "ax.set_xticklabels(ax.get_xticklabels(), rotation=90, ha=\"center\")\n",
        "\n",
        "# Turn on the vertical grid to help align words to scores.\n",
        "ax.grid(True)\n",
        "\n",
        "plt.title('Start Word Scores')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA6sAAAI0CAYAAAD7pb+nAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXxU5dk38CsBERVcwKAoFi19FDdEBUGr1q1VEJCtYitWW7VardbW91HrU9fWBbUuT9WqrdrWrVUUcNdWUalbhdelqK3WulWQHQxgyHbeP3yJhCTOZGDCnfD9fj75QM7c17mvc87MZH45ZyYlWZZlAQAAAAkpXdMNAAAAwMqEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqALQS//nPf2K77baL6urqNd0KABSdsApAqzV16tQ44ogjYvfdd4899tgjjjjiiHjttdciIuK+++6Lb33rW6u0/lzhcPbs2bHddtvF3Llz65b9+te/bnTZscceu0q95OOL9gcAtDbCKgCt0uLFi+PEE0+MsWPHxt/+9rd45pln4oc//GF06NBhtaw/n7OX3bp1i549e8ZLL71Ut2zq1Knx5S9/ucGy/v37r/b5V1Ts/bFcTU3Nal0fADRFWAWgVXr33XcjImLIkCHRrl276NixY+y9997Ru3fveOedd+K8886LV155JXbdddfo169fREQ89dRTMXz48Nhtt93ia1/7WvzqV7+qW9/ys6j33HNP7LfffnH00UfH2LFjIyKif//+seuuu8bLL7/coI9+/frF1KlTI+KzIPf666/Hd77znXrLXn755ejXr1/U1tbG9ddfH/vvv3/sueeeccYZZ0R5eXmT89fU1MS4ceNiwIABceCBB8bTTz9d0P5Y7u67745BgwbFrrvuGoMHD47XX389IiLeeeedOOqoo6Jfv35x6KGHxhNPPFFXc9ZZZ8V5550Xxx9/fPTt2zdefPHFmDVrVpxyyikxcODAOOCAA+IPf/hD3fjXXnstRo4cGbvttlvstddecckll+R7SAGgnvZrugEAKMQ222wT7dq1izPPPDMGDx4cffv2jY022igiInr16hUXXHBB3HPPPXHXXXfV1ay33noxbty4+K//+q9466234nvf+15sv/32cdBBB9WNeemll+Lhhx+O0tLSmDt3bhx44IHx0ksvRfv2jf/I7N+/f9x6660REfHGG29Er169Ys8996yb94033ojq6uro06dP3HfffTFhwoT4wx/+EF26dIkzzzwzLrzwwrj88ssbnf/uu++OyZMnx8SJE2O99daLU045paD9ERHxyCOPxK9+9au47rrrYuedd44PPvgg2rdvH1VVVXHiiSfGqFGj4uabb45p06bFSSedFPfee298+ctfjoiIBx98MG666aa48cYbY9myZXHkkUfGAQccEL/85S9j1qxZccwxx8Q222wT++yzT1x00UXxne98J4YPHx5LliyJt99+u7mHFgAiwplVAFqpTp06xZ133hklJSVxzjnnxJ577hknnnhivfeKrmzAgAGx3XbbRWlpafTu3TsOPfTQ+Nvf/lZvzCmnnBLrr79+dOzYMa8++vfvH2+//XZ88sknMW3atOjXr19svfXWMX/+/Lplu+yyS3To0CEeeOCBOOaYY2KrrbaKDTbYIH7yk5/Eww8/XO+S3xXnf+SRR+Loo4+O7t27x8YbbxwnnHBCwftj/Pjxcdxxx0WfPn2ipKQkevbsGVtuuWW8+uqrsXTp0vj+978fHTp0iD333DP233//eOihh+rWfeCBB8buu+8epaWl8dZbb8X8+fPrLjHeaqut4vDDD4+HH344IiLat28fH3zwQcyfPz822GCD6Nu3b177EQBWJqwC0Gr16tUrLr300njmmWfigQceiNmzZ8fFF1/c5PhXX301jjrqqBg4cGDsvvvu8cc//jEWLFhQb8zmm2/erB569OgRm222WUydOjVeeuml2H333SMiYrfddqtbtvz9qrNnz44tt9yyrnbLLbeM6urqmDdvXqPzz549O7p37173/RZbbPGFvXzR/pg5c2Z86UtfalAze/bs2HzzzaO09POXBFtssUXMmjWr7vsVe/joo49i9uzZ0a9fv7qvG264oS4UX3TRRfHee+/FoEGDYtSoUTF58uQv7BkAmiKsAtAm9OrVK0aOHFl32WlJSUmDMaeffnrdez+nTZsWRxxxRGRZVm/MinWNraMx/fr1i5deeileeeWV2G233SIiYvfdd4+XXnoppk2bVhdWu3XrFh999FFd3YwZM6J9+/bRtWvXRucsKyuLmTNn1n2/4v9zWXl/dO/ePT744IMG47p16xYff/xx1NbW1ptns802a3S93bt3jx49esTUqVPrvl5++eX4zW9+ExERW2+9dVx55ZXx/PPPx/HHHx+nnnpqLF26NO++AWA5YRWAVumdd96JW265JT7++OOI+CxgPfjgg7HLLrtERETXrl1j1qxZUVlZWVezZMmS2GijjWLdddeN1157LR588MEvnKNLly5RWloaH3744ReO69+/f0yaNCm6desWnTp1iojPwuqkSZNi8eLFdZfCDhkyJH7/+9/Hhx9+GEuWLImrrroqBg0a1OT7YQcNGhS33XZbfPzxx7Fo0aK46aabCt4fo0ePjltuuSWmT58eWZbF+++/Hx999FH06dMnOnbsGL/97W+jqqoqXnzxxXjyySdj8ODBjc7Tp0+f2GCDDeKmm26KioqKqKmpibfeeqvuT+RMmjQp5s+fH6WlpbHhhhtGRNQ7awsA+fIBSwC0Sp06dYpXX301br311igvL4/OnTvH/vvvH2eccUZERAwcODC+8pWvxN577x0lJSXx4osvxnnnnRfjxo2LCy+8MPbYY48YNGhQfPLJJ03Osd5668WJJ54Y3/rWt6K6ujp++9vfNvoezP79+8e8efPqBbztt98+KioqYscdd4z11lsvIiJGjRoVs2bNirFjx8ayZcti7733jnPOOafJ+Q8//PB477334rDDDosNNtggjj322HjhhRcK2h+DBg2KhQsXxumnn153OfJll10WW265Zdxwww1xwQUXxI033hibbbZZXHbZZdGrV69G52nXrl3ccMMNMW7cuDjwwAOjsrIyttlmmzjttNMiImLKlClx6aWXRkVFRWyxxRZx1VVX5f3+XwBYUUm28vVPAAAAsIa5LgcAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHJaxd9ZXbBgSdTW1v8LO127dop58xY3az2p1qTaVyE1qfZVSE2qfRVSk2pfhdSk2lchNan2VUhNqn0VUpNqX4XUpNpXITWp9tVSNan2VUhNqn0VUpNqX4XUpNpXITWp9lVITap9FVLT1PjS0pLYZJMNGq1pFWG1tjZrEFaXLy9kXSnWpNpXITWp9lVITap9FVKTal+F1KTaVyE1qfZVSE2qfRVSk2pfhdSk2lchNan21VI1qfZVSE2qfRVSk2pfhdSk2lchNan2VUhNqn0VUtPc8S4DBgAAIDnCKgAAAMkRVgEAAEiOsAoAAEByhFUAAACSI6wCAACQHGEVAACA5AirAAAAJEdYBQAAIDnCKgAAAMkRVgEAAEiOsAoAAEByhFUAAACSI6wCAACQHGEVAACA5AirAAAAJEdYBQAAIDnt13QDAACrqvPG60XHdRq+rCkr61zv+4qq6ihf+GlLtQXAKhBWAYBWr+M67WP4+Cdzjps4+oAob4F+AFh1LgMGAAAgOcIqAAAAyRFWAQAASI6wCgAAQHKEVQAAAJLTIp8GvGzZsrj44ovj+eefj3XXXTf69u0bP//5z1tiagAAAFqhFgmrl19+eay77rrx2GOPRUlJScydO7clpgUAAKCVKnpYXbJkSUycODGefvrpKCkpiYiITTfdtNjTAgAA0IoV/T2rH374YWy88cZx7bXXxsiRI+Ooo46KqVOnFntaAAAAWrGSLMuyYk7w+uuvx8iRI+OKK66IoUOHxquvvhonnnhi/PnPf45OnToVc2oAYC0yfPyTOcdMHH1AC3QCwOpQ9MuAu3fvHu3bt48hQ4ZERMQuu+wSm2yySbz77rux884757WOefMWR21t/UxdVtY55swpb1Yvqdak2lchNan2VUhNqn0VUpNqX4XUpNpXITWp9lVITap9FVKTal+F1KTaVyE1ucaXlXXOe1251tMat7811aTaVyE1qfZVSE2qfRVSk2pfhdSk2lchNU2NLy0tia5dGz+JWfTLgLt06RIDBgyIZ599NiIi3n333Zg3b1707Nmz2FMDAADQSrXIpwFfcMEFcfbZZ8e4ceOiffv2cdlll8WGG27YElMDAADQCrVIWN1qq63itttua4mpAAAAaAOKfhkwAAAANJewCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASE6LhtVrr702tttuu3jrrbdacloAAABamRYLq6+//nq88sorseWWW7bUlAAAALRSLRJWKysr48ILL4zzzz+/JaYDAACglWuRsHrNNdfEsGHDokePHi0xHQAAAK1cSZZlWTEnePnll+Pqq6+O3/3ud1FSUhIHHHBA3HDDDbHtttsWc1oAYC0zfPyTOcdMHH1AC3QCwOrQvtgTvPTSS/HOO+/EgQceGBERH3/8cRx77LFxySWXxN57753XOubNWxy1tfUzdVlZ55gzp7xZvaRak2pfhdSk2lchNan2VUhNqn0VUpNqX4XUpNpXITWp9lVITap9FVKTal+F1OQaX1bWOe915VpPa9z+1lSTal+F1KTaVyE1qfZVSE2qfRVSk2pfhdQ0Nb60tCS6du3UaE3Rw+r3v//9+P73v1/3vTOrAAAA5OLvrAIAAJCcop9ZXdmTT+Z+PwkAAABrN2dWAQAASI6wCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHLaF3uCBQsWxBlnnBEffPBBdOjQIXr27BkXXnhhdOnSpdhTAwAA0EoV/cxqSUlJHHfccfHYY4/FAw88EFtttVVcccUVxZ4WAACAVqzoYXXjjTeOAQMG1H3ft2/fmDFjRrGnBQAAoBVr0fes1tbWxl133RUHHHBAS04LAABAK1OSZVnWUpNdcMEFMWvWrLj22mujtNRnOwEAq8/w8U/mHDNxtF+YA7QWRf+ApeXGjRsX77//ftxwww3NDqrz5i2O2tr6mbqsrHPMmVPerPWkWpNqX4XUpNpXITWp9lVITap9FVKTal+F1KTaVyE1qfZVSE2qfRVSk2pfhdTkGl9W1jnvdeVaT2vc/tZUk2pfhdSk2lchNan2VUhNqn0VUpNqX4XUNDW+tLQkunbt1GhNi4TVK6+8MqZPnx433XRTdOjQoSWmBAAAoBUrelh9++2348Ybb4ytt946jjjiiIiI6NGjR1x33XXFnhoAAIBWquhh9b/+67/in//8Z7GnAQAAoA3xKUcAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkR1gFAAAgOcIqAAAAyRFWAQAASI6wCgAAQHKEVQAAAJIjrAIAAJAcYRUAAIDkCKsAAAAkp0XC6rvvvhtjxoyJgw8+OMaMGRPvvfdeS0wLAABAK9UiYfW8886Lb3/72/HYY4/Ft7/97Tj33HNbYloAAABaqaKH1Xnz5sUbb7wRQ4YMiYiIIUOGxBtvvBHz588v9tQAAAC0Uu2LPcHMmTNjs802i3bt2kVERLt27aJbt24xc+bM6NKlS7GnB4Amdd64Y3RcZ50Gy8vKOjdYVlFVFeULK1qiLQAgIkqyLMuKOcH06dPjzDPPjIceeqhu2eDBg+Pyyy+PHXfcMa91ZNU1UdK+XbPGZdXVUdI+dxZfcVxza/IdX7+mKkraN3xh1HD85+MKq6mMkvYd8qj5fFxtdWWU5lGzfFxzxxcyR6E1NdWV0S6PmuXjmju+kDlaqqaQOaprKqN9u9w1K45riZp8x69YU1VTGevkUbPiuJaoqaypjA55bsvysfnWrDiusqYqOrTL/Xyx4rjm1lTWVEeHdvk99y0fm2/NiuNaoqbltqUmOrTL/XNsxXHNrcl3fCE1q9JXITUtty210aFd7ovMVhzX3Jp8xxdSU7+vLDq0K8mj5vNxVTVZrJNHzfJx1TVZtM9jfETUjc23ZsVxNTVZtMujZsVxza3Jd3whNSuOq63JojSPmhXH1VZnUdo+j5r/Py7f8YXUrDiukJqsOouSPGqWj8t3fP2a2ihpn/vxsuK4lqjJd3whNfX7KiQnFb+mkDkaU/Qzq927d49Zs2ZFTU1NtGvXLmpqamL27NnRvXv3vNdR0r5dzPn17TnHlf1gbMyZU/7Z/8s6x5wbbs5dc+KxK9Vcl0fNyTFnTnmUlXWOWb++LOf4iIjNfnBGXc3M68/JOb77ST+v19dH152cs2bLk6+rV/P+/w7PWdPz1In1av5x3WE5a3qfPKluW1759bCc4/v+4P56c/ztxqE5a/Y44YF6NVN+MyRnzT7HP1iv5onfHpqz5sDjHqrblkdvHpxz/CHHPlxvjvtvGZSzZtj3HqlXM/7WQ3LWjP7uo/Vq7vrdwTlrvnXMY3Xb8vvffSPn+KOPebzeHL/9Q+45jvvOY/Vqrr89d81JY+vXXH1n7prTvv35toz7Y+7xERFnHvF5zfl35645//D6fZ0xPvdxuWx0/eNywn25a24c+WhdX8Mn5R4fETHxsM9rBk06Ouf4Rw77fb2+Bk08LXfN8Kvr1Qye+D85ax4eflFdX4Mn/CLn+IiIh0f8rK7m0Am5ny8fGnFGvb4Ove/q3DUjT1upJvfz+EMjT66rWVlZWecmb1t++5B7b8k5x4OjvlevryH3/iGPmu/Urxl/R+6a0UfW7eMh4/+Uc/xnNWPqaoaOH59z/AOjR9fra+j4SXnUHFavZtj4h3JURNw/+tC6vg4b/1jO8RERk0YfvErHcsS9f805x4RRe9fbllH3vpiz5t5RA+q2ZfS9r+QcHxExflTfuprD7/1HzvF3j+pdr68j73s/Z80dI3vWq/nvCf/JWXP5iB51fV0yYWbO8RERPx3Rva7mugmzco4/ecRm9fr6/X1zctYcPbKsXs09987NWfPNUZvW9fXA3bnHR0QMPfzzmj/flbvm69/atF5fz9yee1v2HVt/W1743eycNQOP6VbX18u/zT0+ImLX4z6veeOG3MdlhxPrH5d3/vfjnDW9Tt28/mvYK3LfZ7b8P5/fXz6+PPf9OCJi8//u+XnNL9/KPf70bev19fGVf89d85Od69XMunpazprNTtv985xwzQs5x0dEbPajgZ/X/O8zucefum+9vmb/6omcNd1OObB+zbWP5K754aD6Ndfdn7vm5GF12zL7+tw/X7qdNDrmzVscXbt2avT2or9ntWvXrrH99tvHgw8+GBERDz74YGy//fYuAQYAAKBJRT+zGhFx/vnnx1lnnRXXX399bLjhhjFu3LiWmBYAAIBWqkXCaq9eveKee+5piakAAABoA1okrAIAbUNFVXU8MDr35xtUVFW3QDcAtGXCKgCQt/KFn8bKH1mU64OMKqqq4/7RuT/0TsAFYEXCKgBQVIUEXAAo+qcBAwAAQHMJqwAAACRHWAUAACA5wioAAADJEVYBAABIjrAKAABAcoRVAAAAkiOsAgAAkBxhFQAAgOQIqwAAACRHWAUAACA5wioAAADJEVYBAABIjrAKAABAcoRVAAAAkiOsAgAAkJz2a7oBgDVtWdWyuGz0o3mNAwCgZQirwFrvk4WVEVFZb1lZWeeYM6d8zTQEAIDLgAEAAEiPsAoAAEByhFUAAACSI6wCAACQHGEVAACA5AirAAAAJEdYBQAAIDnCKgAAAMkRVgEAAEiOsAoAAEByhFUAAACSI6wCAACQHGEVAACA5AirAAAAJEdYBQAAIDnCKgAAAMkRVgEAAEiOsAoAAEByhFUAAACSI6wCAACQHGEVAACA5LQv5sovuOCCeP7556NDhw6x/vrrx//8z//EzjvvXMwpAQAAaAOKGlb33XffOPvss2OdddaJyZMnx49//OP4y1/+UswpAQAogsqqmvjpiO55jwVYVUUNq/vvv3/d//v27Rsff/xx1NbWRmmpq48BAFqTRQuXNrq8rKxzzJlT3sLdAGuDoobVFd1xxx2x3377FT2olpV1bjM1qfZVSE2qfRVSk2pfhdSk2ldL1eQaX0gPq2s9a/OxTKGv1nTsW6qmGI+XltjPbf1YptLX6tjPqWzL6qhJta9CalLtq5CaVPsqpCbVvgqp6dq1U5O3rVJYHTFiRMyYMaPR25577rlo165dREQ89NBD8cADD8Qdd9yxKtPlZflv9pqzkwqtae6BaG5NS25LITWp9lVITap9FVKTal+F1LTUY6wxuc4UpLotbem4tFTNylb3sW+pmpQfL82tKWRbCumrLR2XNfV4Wb6e1XEs2+LjxbbYlmLVtKXXcPPmLW4ysK5SWJ0wYULOMX/+85/jqquuit/97nex6aabrsp0AAAArCWKehnw5MmT45JLLolbb701evToUcypAAAAaEOKGlZ/+tOfxjrrrBOnnnpq3bLf/e53sckmmxRzWgAAAFq5oobVF154oZirBwAAoI1qsU8DBgDSUlFVHQ+MHp3XOABoacIqkKzKqoo484jH8h4LNE/5wk9j5c9wLeSTfQGgGIRVIFmLFlZFRFW9ZV5IAwCsHUrXdAMAAACwMmEVAACA5AirAAAAJEdYBQAAIDnCKgAAAMkRVgEAAEiOsAoAAEByhFUAAACS035NNwAAsCZUVFXHhFF75zUOgJYnrAIAa6XyhZ9G+UrLyso6x5w5Ky8FYE0QVmmTqiqXxYHHPZTXOADSUlFVHZNGH5z3WADaJmGVNmnhosqIqKy3zG/LAVoHZzwBiPABSwAAACRIWAUAACA5wioAAADJEVYBAABIjrAKAABAcoRVAAAAkiOsAgAAkBxhFQAAgOQIqwAAACRHWAUAACA5wioAAADJEVYBAABIjrAKAABAcoRVAAAAkiOsAgAAkBxhFQAAgOQIqwAAACRHWAUAACA5wioAAADJEVYBAABIjrAKAABAcoRVAAAAkiOsAgAAkBxhFQAAgOQIqwAAACRHWAUAACA5LRJWX3zxxdh+++3j9ttvb4npAAAAaOWKHlYXL14cV1xxRey7777FngoAAIA2on2xJ7j00kvj2GOPjaeeeqrYUwFQgIqqynh4+EV5jQMAaClFDatPP/10lJeXxyGHHCKsAiSqfOGyKI9l9ZaVlXWOOXPK11BHAACrGFZHjBgRM2bMaPS2Rx99NH75y1/GrbfeuipTNFtZWec2U5NqX4XUpNJXIetcHetItSbVvnLVrI7jmM96Wmqe1VGT6uOlkPWsTfexVGtS6GFNzVFozZqaO5V94udry8/RUjWp9lVITap9FVKTal+F1HTt2qnJ21YprE6YMKHJ26ZOnRpz5syJb37zmxERsWDBgpg8eXIsXLgwfvjDH67KtF9o+ZmA5uykQmuaeyCaW9OS21JITap9NeWLzhSlvC1t6bisjmOZ64xfS9xfCpkn5eeLxqyux8vyedb2+1hqj/0V51lZIWfVW6Imlb7a0mM/teeL1B8vtsW2FKuvQmpSfw3XnPHz5i1uMrAW7TLgfv36xfPPP1/3/VlnnRU77bRTjB07tlhTAgAA0Eb4O6sAAAAkp+ifBrzcpZde2lJTAQDQSlVV1cbRI8vyGge0bS0WVgEAIJeFC5c0WFbI+4+B1s9lwAAAACRHWAUAACA5wioAAADJEVYBAABIjrAKAABAcoRVAAAAkuNP1wC0IRVVlfHI8KvzGgcAkDJhFaANKV+4LMpjWb1l/j4hANAaCasAkKCKqqp4cPSReY2j5VRU1cS9owbkNQ6AVSOsAkCCyhdWRHlU1FvmLPmaV75waax8BBwXgOLwAUsAAAAkR1gFAAAgOS4DBoAiq6iqigdHfSevcQDAZ4RVACgy7z8FgOZzGTAAAADJEVYBAABIjrAKAABAcoRVAAAAkiOsAgAAkByfBgxAkiqqquKhkaflNQ5SVVFVE+NH9c177PJ/7x7VO+/xAG2VsApAkvy5F9qC8oVLo7F77Bfdlxurcd8H1kYuAwYAACA5wioAAADJEVYBAABIjrAKAABAcoRVAAAAkiOsAgAAkBxhFQAAgOQIqwAAACSn/ZpuAGB1qqyqiPMPfyyvcQAApEtYBdqURQurIqKq3rKyss4xZ075mmkIAICCCKsALVHj3B0AACAASURBVKCiallMPOzRvMcCAKzthFWAFlC+sDLKo7LeMmd8AQCa5gOWAAAASI4zqwDQDBVVVfHgqO/lNQ4AKJywCgDNUL6wIsqj/qdJu6QbAFY/lwEDAACQHGdWAWiWiqrKeHjEz/IeCwBQCGEVClRVuSyGfe+RvMZBW1K+cFmUR/37tctgAYDVTViFAi1cVBnhT5EAAEBRFD2s3nbbbXHHHXfEOuusE6WlpTFp0qRiTwkAAEArV9Sw+vjjj8ejjz4a48ePj06dOsXcuXOLOR0AAABtRFHD6i233BI/+tGPolOnThERsemmmxZzOgCAtVJFVU3cMbJnXuMAWouihtV33nknXn311bjmmmuisrIyjjjiiDj88MOLOSUACaqoqoqHRpyR1zig+coXLo2VPzHB5ygArd0qhdURI0bEjBkzGr3tueeei5qampg5c2bceeedsWDBgvjWt74V22yzTfTv339Vpv1CZWWd20xNqn0VUpNKX4Wsc03MUeh6WutxaW5NCj2srvW01P0l1fvTyjqus050LFun4Hkqa6rjoZEn55ynsqa61d7HUr1fpnofa0s1qd5fVtc8fr6mWZNqX4XUpNrX8pqsujY2O3XfnGOz6trkt6U5unbt1ORtqxRWJ0yY8IW3b7HFFjFkyJAoLS2Nrl27xl577RWvvfZaUcPq8t8gNmcnFVrT3APR3JqW3JZCalLtqylf9Bvmlpij0Hna0nFZHfu5kDMFxahpiW1pqfvl6qhJ5bisjppU+8pVU8jPpJboa3XVpNpXS9Wk+jyWa55Uf762pddwtqXtbEtjUnhtWUhNc8bPm7e4ycBamvdaCjBkyJCYMmVKREQsXbo0pk2bFr179y7mlAAAALQBRQ2rxxxzTMycOTMOPfTQ+OY3vxlDhw6Nr371q8WcEgAAgDagqB+w1LFjx7j88suLOQUAAABtUFHDKgDQMiqqquPB0WPyHguwtqqprInNf7JzXuNYs4RVAGgDyhd+6k+XsNaqqqqNb47aNK9xMH/R0gbLPF+mSVgFAKBVW7hwSYNlwge0fkX9gCUAAAAohDOrEBFVlcvikGMfzmtc6qqqKuLoYx7PaxwAAKRKWIWIWLioMiIq6y1rrZcPLVxYFRFV9Za11m0BAGDt1WbDak1lVZSdeGxe4wAAAEhLmw2r8xdVRET9yxydXQIAoKVUV9XGvmPL8hoHNNQqwmpNZVWU/WBsXuNoe6orl8U+xz+Y1zgAgFQs8CnFrKVqKquj20mj8xr3RVpFWF1Yvixqa50lXVstaEPvJwUAgLZu/qJPGyxr6vV7aWlJk+vxp2sAAABIjrAKAABAcoRVAAAAkiOsAgAAkBxhFQAAgOQIqwAAACRHWAUAACA5reLvrAIAwNqguqo2Bh7TLa9xqauprIkt/0/3vMZBY4RVAABIxIKFSxosKyvrHHPmlK+BblbN/EVLGyxrrdvSltRUVke3Hw7Ka9yaJqwCAACsJeYv+rTBslR/ieA9qwAAACRHWAUAACA5LgOGxFVVVcS3jnksr3EAALlUV9bGDidultc4WJOEVUjcwoVVEVFVb1mq7ytg9aqoWhaPHPb7vMYBQL4WLErzQ5xqKmti8//umfdY2j5hFSBR5Qsrozwq6y1L4cUEABRDY58eHOFn39rMe1YBAABIjjOrAAAURWVVbZw8Ivd7Iyur0n9vZFVVbXz9W5vmNQ5YPYRVAACKYtHCNN8bWYiFbWhboLUQVoGCVFZVxEljc39KcaVPKQYAoADCKhCVVRVx3HeaFzwX+ZRiAACKSFgFBM8CLKtaFjeOfDSvcQAANJ+wCrSYyqqKOO3bbePS4U8WVkb4szIAAEUjrAItxhlcAADy5e+sAgAAkBxnVgEAAFajmsqa2OxHA/MeS+OEVQAAgNVo/qKljS739qfmcRkwAAAAyRFWAQAASI7LgClYdeWy6PuD+/MaBwAA0BzCKgVbsMjfmQQAWBtUV9ZGr1M3z2scrC7CKrSgqsplMfq7j+Y1DgAgFQsWLWmwzEkKik1YhRa00NloAADIS1HD6rvvvhvnnntufPLJJ1FZWRmDBw+OU045pZhTAgCQh2VVNXH5iB55jQNYE4oaVi+//PI4+OCDY+zYsbFkyZIYMmRIfO1rX4s+ffoUc9qk1VRWRveTfp7XOACAYvlkYcO/A+lqHyAlRQ2rJSUlUV7+2RNeRUVFlJSURJcuXYo5ZfLmL1oWEfXfj+gHAwAAQH1F/TurZ599djz88MOxzz77xAEHHBDHHnts9OiR+3ITAAAA1m6rdGZ1xIgRMWPGjEZve+655+JPf/pTHHbYYXHcccfF7Nmz46ijjoqddtopdtlll2bN07Vrp0aXl5V1bnbPhdS0xNyp9lVITUttS0tsawr7c3XVpNpXITWp9lVITap9FVKTal+F1KTaVyE1qfZVSE2qfbVUTap9FVKTal+FrMe2pFvT3HWk+ho21b4KWU9z51ilsDphwoQvvP22226Lv/zlLxER0a1btxg4cGC89NJLzQ6r8+YtjtrarN6yQi6dzVXTnJ03Z055s3d2U3Pn6qvLRuvGlidfl3P9NZWV//8y4+ZvS6E1K1vd+/iL1rO6j/+amKOlalLtq5CaVPsqpCbVvgqpSbWvQmpS7auQmlT7KqQm1b5aqibVvgqpSaGvlnoN1xI1bWlbCqlpqdeWa+o1bCp9FTMnlJaWNHlysqjvWe3Ro0dMmTIlhg8fHosXL45p06bFAQccUMwp2yTvcwUAWH2qqmpj6OGb5j02ZdWVtbHrcd3yHkthaiprYrPTds9rXEuqqayObqccmNe41qioYfWSSy6JX/ziF3HLLbdEdXV1DB48OL72ta8Vc8pVUlNZFWUnnpzXOAAAWqeFC5c0urw1ngxYsKjtbEvK5i9K89Oz5y/6tMGyFPpaXYoaVnfaaaf44x//WMwpVqv5iyoioqLesrZ0sAEAAFqLon4aMAAAABSiqGdWAQAAaN1qKquj28nD8hq3OgmrAABAq1RTWRObn75tXuMo3Jp6b6ywCgAAtEqpfvARq4f3rAIAAJAcYRUAAIDkCKsAAAAkx3tW26iaymXR89SJeY0DAABIjbDaRs1fVBkRlfWW5XqzeXXlsuh98qSc664WcAEAgCITVqmzoICACwAAUAzeswoAAEByhFUAAACSI6wCAACQHGEVAACA5AirAAAAJEdYBQAAIDn+dM0qqKmsis1+cEbeYwEAAMiPsLoK5i+qiIiKBsv9bVIAAIBV4zJgAAAAkiOsAgAAkBxhFQAAgOQIqwAAACRHWAUAACA5wioAAADJEVYBAABIjrAKAABActqv6QZYu1RXLos9Tnggr3EAAMDaS1ilRS1YVBkRlfWWlZV1jjlzytdMQwAAQJJcBgwAAEByhFUAAACSI6wCAACQHGEVAACA5AirAAAAJEdYBQAAIDnCKgAAAMkRVgEAAEiOsAoAAEByhFUAAACSI6wCAACQHGEVAACA5AirAAAAJGeVw+qkSZNi6NChscMOO8Ttt99e77ZPP/00TjvttPj6178ehxxySEyePHlVpwMAAGAt0H5VV7D99tvHVVddFTfddFOD226++ebo1KlT/PnPf4733nsvjjzyyHj88cdjgw02WNVpAQAAaMNW+czqtttuG1/5yleitLThqh555JEYM2ZMRERsvfXWsdNOO8UzzzyzqlMCAADQxhX1PaszZsyILbfcsu777t27x8cff1zMKQEAAGgDSrIsy75owIgRI2LGjBmN3vbcc89Fu3btIiLirLPOip122inGjh1bd/uuu+4aTzzxRHTp0iUiIs4///zo2bNnfPe7311d/QMAANAG5XzP6oQJEwpe+RZbbBEfffRRXVidOXNmDBgwoOD1AQAAsHYo6mXAhxxySPzpT3+KiIj33nsv/v73v8c+++xTzCkBAABoA3JeBpzLgw8+GJdddll88sknsc4668R6660Xt9xyS3zlK1+JpUuXxllnnRVvvvlmlJaWxn//93/HQQcdtLp6BwAAoI1a5bAKAAAAq1tRLwMGAACAQgirAAAAJEdYBQAAIDnCKgAAAMkRVgEAAEiOsAoAAEByhNU17J133lnTLbAGOf6wdvLYp61YsGDBmm4BWo0lS5as6RZanXbnn3/++Wu6iXwtWLAg3nvvvZg7d260b98+1ltvvWav45133okuXbo0etuMGTPi7bffjq5du0a7du3qlj/77LPxpS99qdHxZ599djz22GPRr1+/uOKKK+InP/lJPPXUUzFw4MDYcMMN643/9NNPo7q6ut7XEUccEWPGjImqqqpYZ511cva/aNGimD59eqy77rqx/vrrNzrmn//8Z6y33nrRoUOHqKqqihtvvDFuvvnm+Ne//hV9+/aN9u3bN6gZN25cdOnSJcrKynL2EBExZ86c+OUvfxn//Oc/Y5dddomLLroorrjiipg6dWrsuuuuscEGGzRat2zZspg8eXJMmTIlpk+fHkuWLImtttoq53wzZsyIe+65Jx5++OGYMmVK/Otf/4pu3bo12Me5HHfccTFs2LB6yxYsWBC/+MUv4o477ojFixdHnz596m475ZRTYvDgwTnXW11dHf/4xz+iY8eOse666zY6ppDjv2jRoujYsWPd9+PHj4/bb789ZsyYEX369ImSkpIv7OvDDz+MF198MUpLS2OTTTZpdMzkyZOjR48e9e7zudTW1sZjjz0W5eXl0b1793j88cfjzjvvjJkzZ8YOO+zQZF+zZs2K8ePHx0MPPRTPP/98fPzxx9GrV69G75MRq++4RzT9OI6ImDZtWkybNi26dOlS7747YcKE2H777RvtqzmP/caccMIJMXTo0Lz7f/755+P++++PxYsXxzbbbNPomClTpsTGG28cHTt2jE8++SQuuOCCuPrqq+Pll1+Ofv36Nfqcefvtt0ePHj2a9Xz697//PWpqaqJz587x2muvxUMPPRSLFy+Onj17Nlnz2muvxW9+85v44x//GI899li89dZb0atXryafKyIipk6dGjfffHP86U9/ioceeij+7//9v7HBBhvEFltskXevy51zzjlxwAEH1FtW6GN/8uTJ8cEHH8TWW28dU6dOjVtuuSVmzZoVO+64Y6PjW+K5v5Dn/UIe+++8806cddZZ8dJLL8Vuu+0Wp512WpxzzjkxefLk2G233Zr8+drcx/6kSZPi6quvrrvPPP3001FdXR3bbbddzue9lTV27FdFeXl5VFRURMeOHWPRokUxZcqUKCkpafI5tjHXX3999O/fP6+x+Tz258+fH+3atas7lpMmTYrx48d/4f3y1FNPjfbt20fPnj2jtDS/cxdTp06N448/Pp588snYbbfdYuzYsfG///u/cfvtt8euu+4a3bt3b1DT3J/7y5YtixtvvDEuvPDCuPLKK+Omm26KBx54IBYsWBB9+vRp8j7TlC967q+srIy5c+dGp06d6i1f/lpwZW+99VbMnz8/unbtGu+9915MnDgxli5dmtfrmOXOPffc2H///fMen8/rvpUtWbIk3n777Vh//fWjQ4cODW4fOXJkfPrpp/GlL30p7+f+Ql/3Pfzww3HllVfGb37zm7jnnnti2rRpsdVWW8Wmm27a6PjVefyXLFnS6PZHNP/Yryyf132NGTRoUBx99NGN3lbIa75CnvsL0VKv+xpTkmVZtlq2oog++OCDOOecc+KNN96Ibt26RUTE7NmzY4cddogLLrggtt5667zXtd9++8VTTz3VYPn9998fF198cZSVlcXixYvjyiuvjF133TUiIkaMGBETJkxoUHP88cfHPvvsE4sXL45HHnkkhgwZEqNGjYpHHnkknn/++bj++uvrje/du3eUlJREY7u8pKQk3nzzzQbLf/7zn8c555wTERGvvPJKnHTSSbH55pvHjBkz4vLLL4999tmnQc3QoUPj3nvvjQ4dOsS4cePio48+ikGDBsVf//rXqK2tjUsuuaRBze677x4dO3aMsrKyGDlyZAwbNiw23njjxnfi/9/2Xr16xZIlS+K9996L7bbbLoYNGxZPP/10vP7663HDDTc0qHnuuefipz/9aWy00Ubx7rvvRr9+/WL27Nmx7rrrxrXXXtvkC9B77rknrr322jjooIPqfhDOnDkznnjiiTj55JPjm9/8Zr3xn376aZN9H3LIIfH000/XW3bqqadGjx49om/fvnHXXXfFBhtsEFdffXW0b98+hg8fHhMnTmywnueffz5+/OMfR2lpaVxzzTVx2WWXxeLFi2P+/PlxzTXXxMCBAxvUFHL8V7zv3XzzzfH444/H0KFD48knn4wdd9wxTj/99HrjTzzxxLj44oujS5cu8cQTT8R5550XO+ywQ/zzn/+Mn/zkJ3HYYYc1mGP77bePjTbaKIYOHRqjRo2K3r17N7n/lrvwwgtj+vTpUV1dHXvvvXf87W9/iwMPPDCef/75+PKXvxw/+9nPGtTcf//9cdVVV0Xv3r3j5Zdfjr322isWL14cb7/9dtxwww2x3Xbb1Rvf3OOeS1OP/VtvvTXuuuuu+PKXvxyvvfZanH/++fGNb3wjIlbfY/9HP/pRg3U888wzse+++0ZExDXXXNPg9jFjxsSf/vSniIiYOHFiXHfddXHQQQfFX//61xg5cmR897vfbVAzZMiQuP/++6O0tDR+9rOfxfrrrx9DhgyJKVOmxBtvvBHXXXddg5o+ffpEu3btYs8994zRo0fHfvvt94UvXK+99tq47777oqSkJI499tj4wx/+EAMHDoxp06bF4MGD4wc/+EGDmltuuSUmTZoU/fv3j7/+9a8xYMCAKC0tjb/85S/xy1/+MvbYY48GNddff308+uijMXz48HrHf+LEiXHwwQfHySef3GSPjWns+Bfy2L/66qvj2Wefjerq6hg4cGBMnz499tlnn5gyZUrsscceccoppzSoaYnn/kKe9wt57B955JFxyCGHRHl5eTzyyCMxfPjwGD58eDz11FMxceLEuO222xrUNPex/4tf/CLefffdOPzww+sd+7vvvjt69uwZ5557bs4+V9TUY//DDz+Mc845J2bOnBkHHHBAnHbaaXUvOld8/K3o4YcfjnPOOSdKS0vjwgsvjF//+tfRrVu3+Mc//hE/+9nP4pBDDmlQc8cddzRY9qtf/aruvnLkkUfWu62Qx/6wYcPizjvvjE6dOsWvf/3reOaZZ+qek7fddts488wzG9QMHDgwunfvHrNmzaq7D2y77baN7MHPjR49Ok466aT45JNP4uqrr44zzzwzBg0aFC+88EJceeWVcffdd9cbX8jP/Z/85Cex/vrrxxFHHFF3+4wZM+KPf/xjLF68OK6++uov7HFlTR3/v/71r/HjH/84IiK22mqruOqqq+p+4dbYc/9tt90Wt956a1RXV8exxx4bkyZNip133jlefPHFOOqooxocx4iIyy67rMGye+65p+7n1xlnnNHg9kJe95177rlx2mmnRZcuXWLatGlxyimnxCabbBLz58+Pyy+/PPbee+964/fZZ5/o06dPPPvss7H33nvH6NGjY9999/3C5/5CXvdddtll8fbbb8dXv/rV+Mtf/hJ9+vSJLl26xJ133hlnnHFGo4+X1Xn8V9exjyjsdd+ee+7ZYNmiRYtio402qlvnipr7mi+isOf+AQMG1D3mG/uFfGNa4nVfk7JWYMyYMdmkSZOympqaumU1NTXZxIkTs8MPP7zB+Ntvv73Jrz322KPROYYNG5Z9/PHHWZZl2QsvvJDtt99+2ZQpU7Isy7LDDjusyZrlvvrVr9a7bejQoQ3Gn3XWWdnZZ5+dlZeX1y3bf//9m9rsLMuybPjw4XX//973vpc9++yzWZZl2RtvvJGNHj260ZpDDz207v8jRozIqqqqsizLstra2kb7yrLPtrGqqip77LHHshNOOCHr27dvduqpp2bPPPNMVltb22D88vXU1NRke+21V6O3rWzYsGHZe++9l2VZlr322mvZ6aefnmVZlt17773Z8ccf32hNlmXZN77xjWzevHkNls+bNy/7+te/3mD5dtttl/Xu3Tvbbrvt6r6Wf9+7d+8mtyXLPttH559/fva9730vq6ioaPLYjx49OnvzzTezv/3tb9kee+yRvfDCC3XbNWbMmEZrCjn+K84/cuTIuv2wdOnSese5sW0ZM2ZM9v7772dZlmVz5879wmP/5ptvZj//+c+zAQMGZCNGjMhuv/327JNPPmmyr8GDB2fV1dXZ4sWLs1133bVum5YtW9ZoX8trlvf/wQcfZCeeeGKWZVn23HPPZd/+9rcbjG/ucc+yLBs3blyjX5deemm22267NVozZMiQuv7/9a9/ZQcffHB277331u2bxjT3sb/XXntlP/3pT7P77rsvu++++7J77703GzBgQN33jVlx7sMPPzz78MMPsyzLsgULFmRDhgxptGbw4MF1/1/xuWPlnleeZ968edktt9ySDRkyJPvqV7+ajRs3LvvXv/7V5BxLly7N5s6dm/Xt2zebPXt2lmVZVl5e3uSxHzRoULZ06dIsyz47ft/97nezLMuyN998MxsxYkSjNV//+tezioqKBss//fTT7KCDDmq0ZuDAgY1+DRgwINthhx0ajC/ksT9kyJCsuro6Ky8vz3bZZZdswYIFWZZl2ZIlS5o8Li3x3F/o835zH/sr3o/23XffButrTHMf+009vmtqapq8rbnHPss+26+33357Nn369Oyss87KxowZU7ftTW3L8OHDs1mzZmX//ve/s759+2ZvvfVW3XaNHDmy0Zrtt98+O+GEE7Kzzjqr7qtv3751/19ZIY/9FZePGDEiW7x4cZZlWVZZWdlkzfJ5pk+fnl144YXZHnvskY0aNSq78847691Xm+pt5ftwY/uskJ/73/jGNxpd/kW3FfLcP3LkyOzNN9/MsizL7rvvvmz//fev+76xbRk6dGi2ZMmSbM6cOVmfPn2ymTNnZln22XNaU/eXXXbZJTv99NOzX/3qV3Vf/fr1q/t/Ywp53bfiY/yoo47KXn311SzLsuzf//53o8+xy/udO3dudvPNN2eHHnpotvfee2eXX3559u9///sL52jO675DDjmk7rX7kiVLsrFjx2ZZlmUfffRRk/fL5h7/p556qsmvgQMHNrqe5h77LCvsdd/RRx+d/c///E/2/vvvZ//5z3+yDz/8MNt3332z//znP9l//vOfBuOb+5ovywp77t9///2ziy66KBs4cGA2fPjw7LbbbssWLlzY6NjlWuJ1X1NaxXtWFy5cGMOGDav3G5/S0tI47LDDYtGiRQ3GX3zxxfH3v/89pk+f3uCrsrKy0TmyLIvNNtssIj77jcNvfvObOPfcc2Py5MlNntpecfkOO+zQ5G3LXXLJJXHQQQfFMcccE88880yT45oyd+7c2GuvvSLis9+IN7Ut6667brz//vsREdG5c+dYtmxZRHx2yUJNTU2T29K+ffv4xje+ETfccEM8/vjjseOOO8ZFF10U++23X4PxNTU1kWVZVFRUxNKlS+uuwV9+iVtjsiyr+83VzjvvXPeerZEjR8YHH3zQ5HbX1tY2emnZJpts0uiZirKysnj22WfjH//4R93Xm2++Gf/4xz/qzsyvqKqqqt5+OO+882LbbbeN73//+3X7rrGa3r17R//+/WPDDTeMAQMG1G1XRUVFozWFHP8Vby8pKanbD+utt16jl1AsW7YsamtrI+Kz/bb80qeuXbs2uq+Wr7d3797xs5/9LKZMmRLHHXdcTJ48Ofbdd99Gf4sXEXWXm62//vqx3nrr1V1G06FDhyZ/M9uuXbu6/rfaaquYOXNmRHz2m8fG3vPU3OMe8dlvvpdfKrXi1wYb/D/2zjsuiuP//y9AsEFssX2siYmgWBMjdiOWYKEKYkGxUgQkCEZFEQEVEDWWiA3FaCSKCiLYldgVNRoLikoLitJRqtT5/cHv9nvH7d7dzh2nmHs+HvdQ9uZ9MzuzU/ddmkqsa0H5u3XrhgMHDmDfvn04fPiwwvp+dHQ0iouLER8fDyMjI1hYWKBJkyYwNzeHubm51DwqKirQsWNHAEDz5s051WdatGiBR48eAajpB7m5uQBqVKEEzwVbPi1btsScOXMQHR2N4OBgFBcXY+rUqZg6dapYek1NTTRu3BitWrWCjo4OYzqgra3NWS4NDQ1G1eyLL75gyqWnpydxTGarS643lAKZ0NBQHDt2TOzD9izR9P0GDRpAQ0MD2tra6Ny5M6OB0qRJE051WmWM/bTjPt++X1lZibKyMuTl5aGgoIBpy9LSUs4649v3gZp5vzbv379XWNsDQG5uLmbMmAF9fX34+/tj9OjRmDVrFvLz8yW2T5s2bfDVV1+hXbt2+Pbbb5n74qrn0NBQ5OXlYdSoUfD394e/vz9atGjB/L82NH0fqFEFBmqeRcEbYk1NTYntDwD6+vrw8vLCtWvXMGfOHFy8eJH17R1QM/fn5eUhLS0N79+/Z565vLw81ueSZt5XV1fHq1evxK6npaVxtgvN2F9ZWcloE5ibmyMgIACOjo549OgRq4y6ujqaNGmCL7/8Ep06dUK7du0AAC1btuTMIyIiApmZmWjatCmcnJzg7OwMHR0dODs7w9nZmVVGGFnXfcJ9r7i4mDFp+Oqrr0TGOQGC8rZq1Qpz585FTEwMtm3bhvfv32PKlCmsedCs+4SfVzU1NRQWFgKARFMOvu3v4OCAvXv3IiQkROzDZR/Kt+0BunXf/v37oaenh2XLlqGoqAgdO3ZEgwYN0KFDB3To0EEsPd81H0A39jdr1gyenp64evUq7O3tcfXqVfz4449wc3PDjRs3WGWUse7jQjGKzHVM8+bNERMTg4kTJzINSQhBdHQ0q21Yt27dGHWF2ty8eZMzn4KCAub3vvnmG+zbtw/z589n3RADQKNGjVBUVARtbW3s3r2buZ6fn8+5aBk1ahT69esHPz8/mwxJ+QAAIABJREFUnD59mvNBEpCZmYn169eDEIL379+jqqqK+W2uxaeHhwfmzJkDc3Nz6OnpYc6cOfjxxx9x+/ZtmJmZscrUXgC0bt0adnZ2sLOzw/3798XSDxkyBNOmTUN5eTmmTJmCn3/+GYMGDcKtW7fQr18/1jxatGiB6OhoDBs2DDExMSK2LVwLEAAYNmwY5s+fjylTpoiohISHh2Po0KFi6Q0MDPDy5UtmIBFG2CZNQKdOnXD37l0R+6GlS5cyNhZsCNd9bbtDSW3Kt/1fvHiBwYMHgxCC4uJi5OXloWXLlpyD0Pjx4+Hu7g53d3eMGTMGu3fvhomJCa5evco6MAKida+pqYkJEyZgwoQJyMjIYFWDBID27dtj/fr1KC4uRrdu3bBu3ToYGxvj2rVrnHYoHTt2xI4dOzB8+HCcOnWKWeRVVVWx3gvfdgeA7t2746effmJVZzx69CirjKamJrKzs5lNV5s2bfD7779j7ty5zARQG759v2XLltiyZQtOnjwJGxsbLFmyROpmJSUlBZaWliCEIC0tjckPAOviAwBWrFgBFxcXDBgwAK1bt8aUKVMwaNAgPHr0CPPnz2eVqd33+vTpgz59+mD58uW4cOGCWHodHR3GxrN58+bYt28fzMzMcPXqVRFbG2F69uwJLy8vDBs2DOfOncP3338PoGaBxXUvZmZmsLKygpmZmUj7nzhxgnMc69WrF/Lz81nbX3AYKQxt3xdspNetW8dcJ4RwLtiAuh/75R33Ze37xsbGGD9+PCorK+Hi4oJFixZBV1cXf//9N0aPHs0qw7fvz58/H6amphg7dizT9unp6bh06RIWLlzImgfftgcgtrlesGABGjVqhFmzZnGakwiXt7bdGVebGhgYIDQ0FAEBAThz5gy8vLwk9n+avu/o6IhZs2Zh7ty5GDBgABYtWoSffvoJN27cYMwNalO772tpaWHixImYOHEiMjIyWGVsbW0xduxYAICPjw+WLl2KZs2aIT4+nnWMoZn3lyxZgmnTpqFXr17MvJWeno4nT57A19eXVYZm7K+qqkJZWRmzsR84cCA2bdoEV1dX1oMX4X63ePFike+42uXrr7/G77//jt27d2PWrFlYvXq11LGfZt03ePBgBAQEwNXVFQYGBjh9+jQmTJiAGzdusJp0sdV9v3790K9fP1Z1ToBu3Td48GDY2dlhyJAhiI2NZWx1i4qKOPsL3/bv0qUL1q5dy2o3PHLkSNY8+LY9QL/us7GxwbBhw+Dl5YWBAwdKXO/yXfMBdGO/AE1NTRgZGcHIyAiZmZmIjIyEn58fzp49K5ZWGes+TmR+B/sRSUlJIbNmzSI//PADmTRpEpk0aRL54YcfyMyZM0lSUpJY+qioKPLy5UvW34qJiWG9Hh4eTu7cuSN2PS0tjSxatIhVhk09lpAalZDnz59z3Q7D6dOnyerVqyWmEVYd2bZtG8nJySGEEJKRkUGWLl3KKZeRkUE2bdpEnJyciIODA/H19SX379/nTH/gwAGp5RWmurqaXLhwgVy8eJFUV1eTx48fE19fX7Jv3z5SVlbGKpOcnEysra1Jv379iI2NDaPalJubS8LDwznzqqqqIpGRkcTOzo5pfzs7OxIRESGiGk5Lfn4+p/oD13Pk4+PDqiaVlJREZsyYIVO+srS/QFVE8CkvLyeE1NTZ+fPnxdJXV1eTffv2keHDh5PevXsTXV1d8t133xEvLy+Sl5fHmscvv/wiU3mFyc3NJWvXriXr1q0j7969I4cPHyaTJk0iDg4OJC0tjVUmOzubLF68mEycOJEsXbqUUQ3Jz88nV65cEUtP0+7Xr1/nzP/evXus1y9evMioTAmTl5dHAgMDWWXk6fsZGRnEzs6O0yRBQFxcnMhHoNaXlZVF/vjjD065kpISEh4eTtatW0f8/PzI3r17yZs3bzjTb9iwQWI5apOSkkIWLlxInJ2dSXp6Otm0aRPp168fMTY2Jk+ePGGVKS4uJps2bSJ2dnZky5YtpLS0lBBSozosUL1i486dO8THx4fY29sTe3t74uPjQ+Li4jjTl5WVkcrKSpnvhabvX716lVFpFiYlJYWEhITIlO+pU6fqZOznO+7T9H1CatS3Be2Wnp5O9u7dS86dO8eZnm/fJ4SQV69ekZCQEOLn50f8/PxISEgIZ98mhH/bE0KIi4sLiY2NFbv+559/spqMEELIjh07WMf+xMRE4urqKjXPy5cvExMTEzE1SmFo+75AzdbMzIwYGxsTe3t7Eh0dzTlenT17Vmp52cjPz2fmk8LCQnL27FnOvk877xcXF5MzZ86Qffv2kX379pEzZ84w9cAGzdi/fft2cv36dbHrDx8+JJMnTxa7fvz4cdYyJCYmknXr1nGWTcCzZ8+IlZUVp2qqAJq+X1ZWRvz8/MiAAQPImDFjiK6uLtHX1ydz585lrRdJ4wIXtdd9jx49krruq6qqIkeOHCE+Pj7k2LFjzLNYVlbG3BcbfNp/79695PHjx6zfcY3JfNueEPnXfdXV1WT37t3E3t6eM03tNZ+gXrnWfAL4jv1cqs6SqL3u+/PPPxW+7uOiXjhYEpCXl8e8Qm7fvj2nao8KFR8TwqG+qGwEJ5cCQ34VKlRIR9j5hYpPh+LiYomeo9lISkpi1bASIFj+sI3XNPnJSmFhIV68eMFoGMhCXZZHGGl1xsbNmzcZVVUVkqmsrER2djar12RFUFJSgrS0NFRXV6N9+/a8PFSroOdTWffJSnp6Oqe23adIvbBZFdCyZUvo6+tDX19ftVFVAQCcuvVccKlCcsEntIiAkpIS3jJ8ywWA8RbIhba2tsiimya2F9/6VbTMmzdvcP/+fTE7HUl50MicPHkSO3fuREJCgsj1Xbt2sabPz8/HihUrMHfuXDFPn2zeYIGa8DjR0dHIzs4Wuc7mdZC2XLQyfMtWXl6OzMxMsesvX75UWB5ATWgVgffue/fuYc2aNZwqfUDNJiAwMJBRVQoJCYGJiQk8PDxYbSATEhJgYWEBS0tLJCUlwc7ODiNGjMDIkSNZPfTSluvNmzdwcXGBq6srsrOz4ePjg++//x7Tpk3D69evOeVevHjB1Glqair2798v0ZSFb7lo8uBbx7T51KZ2yLHalJaWin0WLFiADx8+cKr0qqmpiS0wBZ68+WwM+Xon1tHRQVRUFC8ZaffPhr29vcTvaeosMTFR7LN8+XIkJSUhMTFRpnJJqy/hcaSiogJbtmyBra0tAgICJHr7FyY5ORknTpwQGwcVKUOTR1paGuLi4uqsXE2aNIGenh569uwpcaMqPB8WFhZiyZIlGDNmDFxcXJCTk8MqU9sk7tixY/D09MQff/zBqdrKJXPw4EFOmePHjzP215mZmbC1tcV3332H6dOnS/RvIuDVq1c4f/48UlJSONPUtrGOioqSabzkyodroyp8LxkZGVLvhW96oCZ0jWB9V1FRgeDgYDg4OGDz5s2ctrS1N6oCnxpFRUWc98tVZ8eOHeMtw6eegXoSukaFCi7Y3JLzDV0jaYKdPXs2rl+/LneZaMpFm4+i0n9sGZpwUjQy69evxz///IOePXvi3LlzmDdvHmbPni1Rhm/IE5rwODTlCgoKwoMHD3jJ8C2bwOU/IQSdO3eWyeU/zf3ThIj5+eef0bZtW5SUlODff/9Ft27dYGZmhnPnziEnJ0csjISNjQ3mzJmDwsJCbNmyBW5ubjAxMUFsbCwOHDiA/fv3K6RcfEMdAfzDZNCUiyYUB986psmHb7gHgC48EN+QMgBdKBK+MoMGDRJb/Eq7f5rwWDR1pqenhw4dOojIZGZmom3btlBTU8OlS5dE0tPUl/CYsGnTJiQmJsLKygrnzp2DpqYm/Pz8xGRcXFywbds2AMCVK1ewYsUKfPfdd3j48CFn2Da+MsrIg1YmMjKScdaXmZkJNzc3PHr0CN27d8emTZvEQjwK17Gvry+qq6sxffp0nDp1Cv/++y9reBiasCo0MpMmTUJMTAyAmvGmX79+jO+NEydOiI3LNCH7aEI90eTD9174pgdEQ9cEBATgzZs3UkPX8A11RFtnNDJs1AsHSyr+27BNdkCN2oXAs5ww/fv3F5uABX+znX5NmjRJbPIVwPWmQNLGkss4n2+5APZFG8B97zTl4lu/ypLZu3cvoqKi0LZtW8TFxWHx4sXw8/PDsGHDOE9kaWSuXr2KyMhIaGpqwtHREQsXLkRRURGcnZ05ZVJTU7F161YAwNixY+Hr6wt7e3vWTQdQ4xEyIiIC2traSEpKgpOTE4qKimBhYaHQcl25coW3DN+y/frrrzh48CD09PQQGRmJOXPmIDg4GHp6egrLAwAuXbqEEydOoLS0FMOGDcPly5fRvHlz2NjYwNramnXzlZSUhM2bN6OqqgpDhgxBaGgoNDQ00KdPH9Y3U8XFxYxToC1btjBpDA0NmfZVRLmysrIwa9YsAEBYWBjzxmvmzJmcJ8xHjx5FTEwMSkpKMHr0aJw7dw7t2rVDXl4e5s6dK7aRoikX3zxo6pgmH11dXXTs2BF2dnbQ0NAAIQQzZsxAWFgY6+8DNYtidXV1LF++nHFGZGhoiNjYWE6ZtWvXYsSIESJvoMrKyvDkyRNOmbCwMIwZM0Zs8d+kSROFyejp6fG+/3v37mHkyJGMszBCCOLi4li9+QugqTNnZ2c8fPgQPj4+jPMrSTI09SU8Jly7dg1//PEHmjZtiuHDh3M6jBH2HhsSEoI9e/agR48eSE9Ph5OTE+tGgq+MMvKglTlw4ACzWd24cSOGDx+OPXv2ICYmBmvXrhVzGCdcx3///TeOHTsGTU1NdO/enVOjTFjm9OnT2LNnD1q2bInJkyfDysqKdeNJIyPsqE5442xmZsa6WXvz5g2jbblnzx6EhYWhc+fOyM3NxZw5c1jrixDCPPMXLlzAwYMH0bRpU9ja2sLCwoJ1E0WTD9974ZtecC9aWloAgDt37iA8PBwNGjSAkZERa5mAmvi9gnvZsmULdu7ciT59+iAlJQXu7u6sm1WaOqORYUO1WVXxyXPw4EHMnz+f1csq2yavdevWiIqKYlUVZ/MM16FDB4SFhbF6jOTyJOfg4IAffviBdaHNpW7Lt1xATUffv38/dHR0xK5PmzZNIeXiW7/KkiEs4aTs7OwketKkkQFqPOIBNW789+7dC0dHR5SVlUl0YS9cdm9vbwQGBkoMeVI7PM7cuXNRXl6u0HLRyvApW22X/x06dICjoyO2bNmisDwAuhAxAtf+GhoaaN++PZNOTU2N1bW+cD+p7WGay+smTbn4hjoC/i9MRpMmTWQKk0FTLr55CPIBZK9jmnz279+PP/74A8uWLYO3tzd0dXWZcA9c+Pv746+//sLs2bOxaNEijBgxQqoNWWhoKDZu3AgLCwvmTX9cXBzrmwgBERER8Pb2hr6+PmbPng01NTVERERIDEPCV4bm/qOjo+Hj44P4+Hi4u7ujcePG+O233zhDYwF0debs7IynT58yb5KmTZsmUYamvsj/D49CCIGGhgajkt2gQQPO8B3CZSgqKkKPHj0AiKs7yiOjjDxoZYTHsoSEBAQGBkJNTQ3W1tasGgTl5eVISkpiDssF8wYAzn5ME1aFRqZz586IjY2FoaEhOnfujNTUVHTt2lXMhESAIGSfurq6zCH7ADCedmUN9USTD9974Zse+L/QNV26dGFC1zRo0ECiB2G+oY4E8K0zWpnaqDarKj55+Lqk5xu6Zty4cUhPT2fdrApc9NeGxlU633IB/MMx0JSLxuW/smT4hpOikdHW1kZaWhoz8Whra2PPnj2wt7fHixcvWGX4hjyhCY9DUy4aGb5lo3H5T3P/NCFi1NXVmbIJq2Jz2ZF36NCBCQuyZs0a5npGRgYTF1YR5aIJc8Y3TAZNuWhCcfCtY9p8+IR7EMA3PBDfkDIAXSgSGhm+908THgvgX2dAzYHLgQMHsHXrVsyePVvi4pbm3p8/f47+/fszz7NAzVg4jnhtXr9+DVdXVxBCkJmZifLycuZtE9fzz1dGGXnQyhQVFeHKlSsghKCqqkpsk1ibDx8+wM7OjnmuBHVcVFTEuVmlCatCI+Pt7Q1nZ2eEhoaiWbNmsLKygr6+Pt6+fcvqq4MmZB9NqCeafPjeC9/0AF3oGr6hjmjrjEaGFZn9BqtQ8ZGgcUlf19C4SqeBbzgGmnLR1K8yZGjCSdHI3L9/nzVMSVlZGWeYCL4hT2jC49CUi0aGb9loXP7T3D9NiJisrCzW/vL27Vty8+ZNVhk2iouLOcMq0JSLJtQR3zAZNOWiCcVBU8fyhPyorq4me/bskRjugQ1ZQoMJI0tImdrIGopEHhma+5c1PFZt+NYZIYQ8ePCA7Nq1S6a0NPUlzPv378mDBw9Yv4uIiBD55OfnE0L+L6SHImSUkQetjI2NjcgnIyODEEJITk4OsbCwYJVho6SkhHOe5htKj1ZGwI0bN8jvv/9OQkNDydmzZ1nHN0LoQvYRwj/UE20+fO6FNj3f0DV8Qx0J4FtntDK1UTlYUjB37tzBwIED6yy9MikpKZFoVyJvehUqVKhQoeJTgSakDE0okroOX/Ip81++949BVVUVysvLOTVFPjeUFbLvcwkNWF9CHdWr0DXCCLxlyUpAQADvPGxsbHjLSLJ1UUR6QFyNShZo7oXNuYYi0wN07aJChQoVKlQoGh0dHV4bVaDGjpLvxotG5nPhv3zvHwMNDY3/zEYVEA/ZV9/zqWtkDXX0sam3m9W9e/fySh8XF8c7D5q4lHxfVNO82JYUO4qLT/VeaNpFhQoVKlSoUKFChQoVnz/1drOqjI2UsHc0WTEyMqrT9IDy7qVjx451mh6guxcVKlSoUKFChQoVKlR8/tTbzaogZp2scMV4lER4eDhvGQcHhzpND4Az1pIkaO7lt99+q9P0AF27ADW2vnzgqzpNozYtyRsmFzQq3XxVp2nKxbd+lSXDV/2fVkYZZgP79u3jnQdNuWhk+JaNpu1p7l8ZbUnT92nKRdP3ldFfaNqS5hnjmw9NHjR1nJqayltGGWMfzf0r67nk22do6oumXMqYL5Q1J9HI8J37acY+Zc1Jn+q6hyYfvveirL5Pg7LWivV2s2phYcErfffu3euoJMqHy610fYS2Xfja+vJVnaZRm6ax2aVR6earOk1TLhpbamXI8FX/p5VRhtlAdHQ07zxoykUjw7dsNG1Pc//KaEuavk9TLpq+r4z+QtOWNM8Y33xo8qCpYzc3N94yyhj7aO5fWc8l3z5DU1805VLGfKGsOYlGhu/cTzP2KWtO+lTXPTT58L0XZfV9ZfjpoZWpt5tVFf9t6loNnEZtmkalWRkyn2q5aGQ+1XIB/J+ZT/lePqd2UfX9us1DWTKfarmUJfOplgtQjX11nYeyZFRj36cpo6xyKcO3Da2MxurVq1fzllKh4iPz7t07DBgwQOb0RkZGaNSokczprayseJfp+vXrmDBhAi8ZvuUCgO+++w6tWrWq03LxrV9lyWhpaaFHjx688qCR4VvHAP9npmPHjujcuTMvGZpy0cjwLRtN29PcvzLakqbv05SLpu8ro7/QtCXNM8Y3H5o8aOr4xYsXGDlyJC8ZZYx9NPevrOeSb5+hqS+aciljvlDWnEQjw3fupxn7lDUnfarrHpp8+N6Lsvp+RESEUvoyjcxnHWf133//RZcuXXjJ7N69G1ZWVrxcOJ8+fRojRoyAtrY2tmzZgkePHsHNzQ29evVSSHoA2LNnDxYsWCD1GhcxMTGYNGmSTGlv3bqFpKQk2NjYICcnB4WFhfjqq68Ull6FChUqVKhQoUKFChUqpNHgYxdAFmRxwqOtrY2FCxeKXFu4cCFKS0thYGCAQYMGYfDgwWjTpo3E38nKysLEiRMxbNgwzJgxA3379pWa944dOzBhwgQ8evQI169fx6xZs7BmzRocPnxYIemBmg1u7Y0p2zUu9u7dK9Nmdffu3bhy5Qqys7NhY2ODyspKeHp64s8//5Q7/ZkzZ7Bq1Spoampi5cqVvE+jVKhQoUKFChUqVKhQ8d+hXmxWT548ialTp0pMExUVJbZZPXXqFLKzs3Hz5k3cvn0bGzduhLa2Nk6fPs35OytXrsTixYsRFRWFlStXQlNTEzNmzMCkSZPQsGFDVpkGDWqq8caNG7CysoKxsbFET5d80t+4cQPXr19HVlaWyKa9qKiIl963rGljYmJw/PhxRhWgXbt2KCoqUkj6LVu24M8//8SHDx/g5uaG1NRUjBkzBv/73/+Qn5+PTp06yXw/KlSoUKFChQoVKlSo+LypFw6Whg4dCmdnZ4mffv36ickRQvD27Vu8efMG6enpaN68OQwMDKTm16RJE1hbW8PFxQV5eXnYvXs3xo4dy7nJVVNTw+nTp3H69GkMHjwYAFBRUcH5+3zSa2pqomnTplBTU0OTJk2Yz9dff80rVIysoX4aNWokZmSvpqamkPRaWlr45ptv0KtXLxw6dAjPnz+Hi4sL7t69izVr1shUPj7s2bNHpmtsKMvtd25uLv755x+p6WRJ87mSlJQk07W6zlMaynpmVHAj8GYYFBT0kUtSf6iursaVK1c+djFEkLcdCwoKUFBQoMgiKZ2qqiqsXLmSt5y846Wsc5KyuXnz5scuglzQtIuy5r6UlBRcvHgRQI2TnXfv3ik8DwF5eXlUcuXl5cjOzlZwaZQDbV9W8X/UizergYGBVGl++OEHfPvtt7C2tkZQUBDatm0r9XdycnJw+PBhREZGolevXggKCsIPP/yAV69eYebMmayqq15eXtizZw8sLS3RqVMnpKamStwU80k/cOBAfP/992jevDmVW2kBsob6adeuHe7duwc1NTVUV1dj586d+PbbbxWSvm3btnjy5Al69eqFNm3aYMuWLcx3o0aN4szj0KFDUsvdpEkTmJubi1yTR3VaVrVpgL/N7vTp07Fr1y4QQmBmZoYvvvgCI0aMwNKlSzllvL29oaGhgenTp8PY2JjzLb+AxMREsWs6Ojoy9QFh4uPjoa+v/1FlPDw8EBkZKfWaMKGhobC0tISOjg6WLFmCx48fY+XKlRg2bJhY2tLSUrFrCxYswOnTp0EIQePGjWW6B1memYCAADg5OaFx48aYNWsWnj59Ch8fH5iamoqlnTx5ssSDomPHjnF+l5ycjB07duDVq1eorKyUSSY3Nxd//PEH0tLSRGSE+ykgvT9KcuW/aNEisXvS0dFBv379YGFhAXV18fPT3Nxc+Pv74+3btzh06BASEhLw4MEDTJs2jTVtfn4+rl+/DhcXFzGNElnbEgCMjY15hde5ePEi2rdvz/oc5+fnY8OGDXj79i1Gjx4tUkcuLi7Ytm2bTHnY29tj165dMpdJWrkAQF1dHZs3b+btVKiwsBB79uzBs2fPUFZWxlw/cOCAWFppm+HaedO0Y15eHjZs2IAzZ84AqDmsVldXh5GRETw8PNCyZUuJZUhLS0NaWhqqqqo4yyUMje8JPmhoaOD58+e85WjGS5o5iQ2+fYZr3Gebv5YvX459+/aBEIJvvvmG8zeLiooQHByM27dvAwAGDRqEhQsXQltbW+ZySeszpaWlCA4OZjbQQ4cOhaOjo8TxhaZdaGSAmvoThDsZNGgQunXrxpk2MjISu3btQkVFBcaMGYPMzEz4+vpi//79nDJ8+r6Ahw8f4ueff2YOxx4/fozw8HD4+flxyri5ucHX1xeampowNTVFfn4+7O3tMW/ePLnvXZ55jE8+AF1fpqlj2mc/JSUFSUlJGDNmDIqLi1FRUcE7RCZXX6Zdv9emXmxWa5OWlobY2Fh07NgRY8aM4Uxnb2+P27dvIyQkBH///TeGDBkCAwMDiZOWmZkZLCwscOjQIbRr14653qlTJ9YNX1VVFa5du4bg4GDmWteuXeHl5cWZR//+/Xml19DQQEREhFybVVnx8vLC0qVL8fLlS/Tt2xcDBgzAhg0bFJI+MDBQZBEsK1u3boWhoaHENHfv3mUedkWoTsuajq+NL1ATRFpHRwdRUVEwNjaGh4cHTE1NJS4MoqKicO/ePYSFhWHz5s0wNjbG9OnTOVWn7ezs8PbtW+jo6ACoGfhatWoFLS0tbNq0iVUTgY0tW7Zg9+7dMqVVtExeXh7y8vJQVlaGpKQkpk0KCwulBuKOiIjAnDlzcPv2beTl5WHdunVYs2YN62a1f//+UFNTE2vzfv36QU1NDc+ePZPpHmR5Zm7evIlly5bh8uXLaNu2LX799VfY2dmxblYFz8Ply5eRnJwMS0tL5t6kOTBbvHgxjIyMYGFhAQ0NDZnKv3DhQvTs2RODBw+WKPPkyRMANRuwO3fuMNoht27dgoGBgcRJvnXr1njy5AkmTpwIoGbBr6OjgzNnziAhIYH19HnlypUYMWIEwsLCAABff/01lixZwrpZHTduHH788UeUl5czz7igbdnakm1RLCA/P5/zOzYuXLiA+Ph4tG3bVizGnbe3Nzp27IiRI0fizz//xK1bt7B582Y0aNAAr169Yv09V1dXsWt37txhrtc+RKAplwA9PT08evQIffr0kek3AcDT0xPdunVDamoqXF1dcfz4cc7FfUhICOfvqKmpiW0K+bYjACxZsgQDBgxAbGws4yAxLy8Phw8fxpIlSyTGHVy/fj1OnDiBr776ijkwYSuXMHx8T7Ad0gjaLygoCEuWLGHNY9CgQfD19YWZmRmaNGnCXGfbrMkzXvKZkxTZZ7jmikmTJqFDhw4iY2pOTg4WLFgANTU1XLp0ifM3PT09oa2tzYwlERER8PT0xNatW2Uul7Q+4+fnh6qqKnh6egKoOQT09fVljR9J0y7ytOWJEyewceNG5tndtWsXPDw8YGJiwpr+999/x/Hjx5lx++uvv0ZOTo7EPPj0fQH+/v7Ys2cPPDw8AAC9e/fGsmXLJMqkpKRAR0cHZ8+ehYGBAZYvX44pU6Zwblb53Ls88xjfOgb49WWAro65Ovs1AAAgAElEQVRpnn2awwo2uPoy3/U7J6QeYGtrS549e0YIIeTt27dkwIABZP78+cTIyIjs2LFDqnxZWRmJjIwko0aNIj169JCali+TJ0/mlT4lJYVMnTqVjBo1ihBCyJMnT8jWrVslygQEBJAzZ87wLhstJSUlpKioqM7S82H27Nm80sTFxZFt27aRoUOHkm3btjGf0NBQ8urVK5nyPH78uEzpjI2NSXl5OTE1NWWuTZo0SaLM+PHjCSGELF++nFy/fp0QQkTkpfH06VMycuRI0qNHD2Jvb08SExPF0qxZs4ZcuHCB+fvChQvEz8+P3L59m1haWsqc18dk//79ZNSoUURfX5+MGjWK+ZiYmJDw8HCJsoL63Lx5M4mMjCSEEGJmZsaadtmyZcTT05MUFhYy1wR9kw+yPDPGxsaEEELWr19PTp06JVJWLiwtLUl1dTXzd2VlpdQ2FOTDB74yCxYsIGlpaczfaWlpxN7eXqLM9OnTRcbYDx8+kKlTp5KysjKmX9TG3NycECJaTyYmJlLzkQVdXV1iaGgo8nwJPvr6+jL9Rm3y8/PFrgnXbXV1NVm9ejWZO3cu+fDhA2f7DxkyhCxfvpxERESQiIgIcvz4cWJgYMD8rYhyCZevZ8+eZPz48WTy5MnMRxKCexKMd2VlZcTa2pp3uSQhazsSQshPP/3E+d24ceMkyo4dO5aUlJTInBch/zeeBAcHkyNHjohcq42gzYQ/Ai5dusSZB9tzaWhoyJpWnvGSz5xUF32mNtu2bSPz588n6enpzDVZx2QjIyOZrskCV5+pPcdXV1dzzvs07SJPWxobG5OsrCzm76ysLIlju2Au4TO+0vR9CwsLsXykzX0TJ04khBDi6+tLLl68KLVsfO+dELp5jCYfPn1ZkAch/OqY5tk3NTUlRUVFIm0hqHdFwHf9zkW9eLOalZUFPT09ADXOlgYPHoytW7eioKAAM2bMgIODA6vcuXPncOvWLdy6dQtVVVUYMmQIc3rCBSEER48eFVOf++WXXzhlfvzxR+zdu1fsxIRLJWT16tVwdHTExo0bAQA9evTAL7/8AhcXF848IiMjERoaikaNGqFx48bMCfOtW7ck3g9f2NS1tLW10b17d+YtnTCzZs2CpaUlxo0bh6ZNmyq0LAJCQ0N5pVGE6rSsatN8bXwF5ZswYQKqqqrg4+ODgoICVvXH2ty7dw+HDh3Cw4cPYWlpCSsrK9y+fRsLFy7EuXPnRNLeuXMHK1asYP4eM2YMtm3bhpUrV+LDhw8y3dvHxtbWFra2tti5cydnH+eiUaNG2L17N06dOoVDhw6BEMJpF+7v74+//voLs2fPxqJFizBixAipbciGLM9Mq1at4O3tjWvXrsHOzg6VlZUiaodsvH//HmVlZUyMwfLycrx//16iTL9+/ZCQkMCMm7LQt29fPH/+HLq6ujKlf/Pmjcib/U6dOuH169cSZXJyckT6S4MGDZCfnw8tLS1oaWmxyggc0gkoKCiQ+hZboHokeAshPC4L06FDB4SFhbGqx/NVixXApj4l/OypqanB29sbgYGBsLOzE1HxEiY6Oho+Pj6Ij4+Hu7s7GjdujN9++036CTSPcgmgsacStJempibevXuHZs2ayWSPVlhYiJSUFJH7/uGHH1jTytqOANCwYUM8ePAA/fv3F7l+//59zmdLQPv27cXGcWkI+54QaEpxjTGS2kzSW4fY2FiZyyPPeMlnTqqLPlMbZ2dnPH36FIsXL4apqSmmTZsm85jcpk0b5OXlMRp0+fn5vM1fBEjqMyUlJczzyGZKIoCmXeRpS6BGg4Xt/2w0b94cKSkpTP1GRUWJaBWyQdP3tbS0UFxczOSTmJgo1ZypW7dumD9/PpKTk+Hu7i7T2oXPvQN08xhNPnz6MkBXxzTPvsAvjjCyamPJAt/1Oxf1YrMq/EDfv3+fUf394osvJFbq+fPnMWjQIMyfPx8dO3aUKS9XV1dUVFSgT58+Uic4AQJHR0FBQVJVlYCayXrEiBHYtGkTgBqbIWkT5fHjx2Uqi7wEBwfj8ePHzIL1xYsX0NXVRWZmJtasWSNmWzp37lxERETA398fo0ePhoWFBb777juFlqmyshKHDx9GSkoKevToIdWWD1Ce6jRfG1+gRiUwISEBnTp1gqamJgoLC6U6mDI2NkbTpk1hY2ODoKAgZgFvamqKkydPiqWvrq7G/fv3mbZ48OABqqurAUCmjfGnQHl5ObS0tGBra8u6GJBkH+Tv74+wsDB4eHigdevWSEtLg7GxMWf6UaNGoV+/fvDz88Pp06elbiBp2bhxI06ePAlzc3M0a9YMr1+/xpw5cyTKjB8/HtbW1oy9/JkzZ6SGfXr06BGjLiw8fkqyWZ06dSpsbGzQrl07mWS+/PJLbN++nfEEfvz4cXz55ZcSyzVw4EARtefo6GgMGDAAxcXFnOPt2LFjsWrVKhQXFyMiIgJhYWGYPHmyxHxevXoFd3d3PHv2DGpqaujZsyeCgoLE1ObHjRuH9PR01gl97NixEvPgQ6dOnXD37l2RTdnSpUuxadMmTqdvLVu2xJYtW3Dy5EnY2NhgyZIlVIcosjBw4EAAsm0KBXTt2hXv3r2DsbExrK2toaOjI1VN7fTp0wgMDERBQQHatGmDtLQ06OnpcdrgydqOAODj44NffvkFDRs2RIcOHQAA6enpKCsrkxr+btmyZXBwcMDQoUNFnkNJqoB8fVXQIuwTITc3FwUFBaxmAPKMl3zmJGX1mZ49e+LAgQPYunUrZs+eLdFpJfB/IQ5btGgBU1NTZq1y+fJlDBgwQGHlAsA888LmDGymHML06dMHhYWFzKF/QUEB4uPjWV+gyNOWnTt3xtatW2FtbQ0AOHr0qMRIC56ennB3d0dKSgoMDQ3RqFEj7Ny5U+K90PR9BwcHzJs3D1lZWVi2bBmuXbsm1YFaYGAgrl+/Dl1dXTRp0gSZmZlwd3fnTM/33gG6eYwmHy71eS41YD51LM+zT3NYwQea9TsbakTaEfUnwJQpU7B27Vq0atUKY8eORVRUFLP5HD9+PONQgQs+E7AsvycvU6ZMwaFDhzBlyhRERkYiMzMTjo6OiIiIkChXVFSEf//9l7fzGj4sWbIEtra2jJOI+Ph4hIaGwtHRkQnpw0Z+fj5OnTqFyMhIFBcX4+zZswor04oVK5CUlITvv/8eN27cgKGhIRYtWiRVLjAwEH379oWRkZHCylKb7OxsLF26FHfu3IGamhpjs9uqVSuxtIIJiOsUVtIE9PjxY/Tu3Vvmct27dw/u7u7M27gPHz5g48aN6NGjB86fP0/9dkaZmJubIzIyEnp6emI2pXxsSfly5swZ3LlzB97e3gr93aqqKixcuJC3gxyg5lT2zp07AGpsX3788UeJ6QVpayPYlLAxYcIETJ48GT179hQ5BOSSyczMxNq1a0WcTHh6eko8ya2oqMDhw4eZ8g0cOBBTp06Velh38uRJxMbGghACQ0NDqQvDOXPmYOLEicymNiIiAjExMTKd4NYF7969g5qaGpo1ayb2XWJiokSHMUBNXa9atQr//PMPU9+KhM+mkI179+6hsLAQw4cPF3sTLoyJiQlCQ0Mxb948nDhxAjdu3MC5c+fg6+vLmp5vOxJC8OTJE7x9+xZAzRvTXr16SV0cubm5ITk5Gbq6uiLPPpsNIlDTl7dv3y7TPCQPwj4Rzp8/j4yMDLi5ubH6RKAZL+WZk5TJgwcPcPfuXdjZ2XGmkRYdwdnZWaFlunr1KqPZNnjwYIwYMUJiejMzM0RGRjLPYnV1NSZPnsx6UCPP3Jebm4s1a9bg5s2bUFNTw5AhQ7BixQrWNQlQs65s3LgxUlNTQQjBV199hczMTPzvf/+TWgeA7H0fqBlnrl27BkIIhg0bhi5dukj9fWHnP0VFRaisrOR848333gG6eYwmH2ENivLycuTk5OB///ufTG9cpdWxPM9+SkoK3N3dkZycjJYtWzKHFZ07d5ZaLlmgXb+LQaeFrFxu3bpFBg4cSHr16kV8fX2Z69evXyeOjo6ccmlpacTKyor06tWL9O7dm1hbW4voprOxYMECEds1WUlOTmZsBIuKiiTaB0VGRhI7OzsycuRIsnXrVmJoaEiio6Ml/v7ly5fJiBEjGLuNR48eSdWrp4HN7kKgOy9JJ//du3fk4MGDxNzcXKp9EF/Gjx/P2LkVFBQwNmzSMDAwILq6uqRv375k0KBBxMDAgAwaNEihZRMgi82uwJ5JV1eX6Onpif3LxsuXLyV+JFFWVkYSEhJIQkIClS12faagoIBs3LiRzJ8/n8ycOZP5fGymTJkiYn/6KcFlb/exSUhIELuWkpIiUYbNrkmaHZYyyMvLI0+fPiVPnz4leXl5H7s4DLNnzyZHjx4l1dXVpLq6mhw7dkwmOyK+CMZu4XlG0nMnbzveuHFDpnRjx47l3S/5+qqggcYnAh9o5qSPxbt37z52EVgpKysTsV/kgu25pfEtIA229aukNe3s2bNJZWUl83dWVhaZMGGCwstFQ0REBPnpp58Y286kpCRia2vLmZ7vvdOiiHxu3rxJAgICFFUkuaisrCSJiYnk5cuXIs+CIqBdv9emXqgBDxo0CDdv3kRxcTG++OIL5nr//v0lnkivWrUKU6ZMETmVXbVqlcTTdR0dHUyePBnDhw8XUQeSZLMaERGB3bt3y+xNy8zMDB07dsRff/2F0tJSBAYGSn1Vv3XrVhw7dowJu9K7d2+kpaVJlKGhcePGiImJYUJwxMTEMG/n2E6nY2NjERkZib///hujR4/GihUr8P333yu0TA0bNmTaQkdHR2ZPvcpQneZj4ys4QU1ISJD59yWdJLN5Rax9Ui44HauqqkJpaeknc1Je1/DxpKeosCKy0LdvXzg5OWHSpEkidiKSbL1owtDQuL0fPnw4rl69KvUNgYDS0lLs2rULr169wsaNG5GUlISUlBSJHtpzc3Nx8OBBsXuR5NnW1NQUP/30E4KCgphxwM3NTWL4BnV1dSQnJ+Prr78GUHN6rEg7HL6kpaXBy8sLT58+RZs2bQDU+GLo2bMnfHx80LVrV07ZHTt2wNHRsU7Ll5eXx3ibBmrCJkl6VmjR0tICIQRdunTBwYMH0aFDB4neTfm0ozzhTrp27YqSkhJefhf4+qqggcYnAh9o5iRlkJCQAE9PT6irqyMwMBCBgYGIi4tD8+bNsXPnTvTo0YNTlk3lWxAeS5rPElmhCanStGlTPHz4EH379gVQE8pFFm0/vsycOVNsbGS7JmD48OFYtmwZgoKCkJeXh7lz57J6IqdFnhBsfD0V8713AcnJyUhISEB5eTlzzczMTOH5CDN48GDWZ3XQoEGs9UVk8FNDM1fcvHkTvXv3ZkLvSFJPp4F2/V6berFZBWpsEIU3qkCNWu+ECRNw+fJlVhmaCfirr76SGhaiNgcOHODt+nvAgAHo2bMnANnUkwFxI25ZbWr54O/vjyVLljAu2b/55husX78eJSUlrBv2gwcPwtzcHEFBQcymVtFkZmaKdOraf3MdJHTo0KHOVaf52vjyha9RvrW1NSIjI1nDsdSl6uynxr///ott27bh0qVLmDRpEsaNG4dZs2axpqUJK0KLoP6F1fikhcigCUND4/Y+PDwcu3fvRtOmTZlNhaTJcfXq1WjdujWz0G3Xrh3c3d0lblZdXFzQrVs3qeFxhNHV1UXXrl0xa9Ys7Ny5E82bN5c64bm5uWHGjBnMojYhIUGq3WJd8ssvv2D69OkIDQ1lbMarq6sRHR2NpUuX4siRI2IyUVFRGDRoEK5cucIsQATqgYpGWZt7V1dXFBUVwcPDA6tXr0ZhYaFEdXs+7ShPuBNtbW1YWFjwOqTm66uCBhqfCJ8Da9asgZOTEwoLCzF//ny4ublh9+7diI2NRWBgoMSwGrm5ubh37x4zDl26dAm9e/fGmTNnMH78eIUc/PANqQLUmFg5OTkxhyaJiYlS1Tf5UFlZiYqKClRXV+PDhw8i4W4kOYCaO3cu1qxZg7Vr1+Lvv/+GnZ0dxo0bp7By8Y3VK4yszn9o7x2oWb8fOXIE2dnZ6N27N+7du4cffviBdbMqTz7Ch2nV1dV4/PixyOZYAM1LFnnmivXr14uk09bWFrsmD7Tr99rUm80qF5IWLTQTMI1dA19vWmlpafDw8GAWebLYBzVt2hQ5OTnMiUtcXByrd155qKqqwr179xAREYGioiIAEAkmPHToUDEZZdiATZ8+XeLfXFy5cgWrVq2ChoYGYmNj8fjxY2zfvl2q8wA+dO7cGV5eXmI2vkFBQVi8eLHcm1VpA2DtU/xP9aRc2fDxpJeamsrEIRs7dix8fX1hb28vEgtZURw8eJC3THV1NW+PkHw26wL4TpLPnz9nHGAANWOUwIkXFwUFBRKDwLOhrq4ONzc3HD16FNOnT8eOHTukvl0aMWIETp06hYcPHwKoeaMtKb52XfPu3TuxGHzq6uowNTXFjh07WGVev34Nd3d3vHjxAl5eXujZs2edefJW1ua+TZs20NHRgY6ODrPhSEpK4kzPpx2dnZ3x8OFD+Pj4MPZ2hoaGMh34ff3118w6QVaUMcbyjXv+uVBcXIzRo0cDqNG6EPQdQ0NDqfFSs7KyEBERwdiHOzk5wcXFBWFhYZgyZYpCNqsCrZC7d+9i5MiRaNy4sVTHhf3798epU6fwzz//AKjx2M5mw07Lzp078dtvv0FNTU0kjrq2tjarEz/hzZOVlRU8PT1hYGCAHj16yGRHLyu1fR4UFxcDgExaDLI6/+F778KEh4fj6NGjmDZtGvbu3YsXL15g+/btrGnlyUdYS65Bgwbo0qULAgICxNIJnMMBNc9ZSkoKgJoXaVw2wfLMFYJDNgHq6uoKdTBJu36vTb3frEpatAhPwIQQPH/+XKYJ+Pr162Lqc5I2sXy9aXl7e/NWT/bw8MCCBQvw+vVrzJw5E6mpqZyLHFo0NDRw5MgRWFtbi2xSJUGjbsgXWscIylCdTkhIYDaqAKCvr48XL16gW7du1OoOwrC9IRUg7RQ/Ly+PWeT169cPLVq0kLs89QU+nvRoworQQgjBsWPH8O+//8LDwwOvX79GVlaWRA/aNGFoaNze89VEqK3ZUVZWJvWZ//bbb5GZmckrlITgN62srNC+fXvMnTtX6iEOUONNV97DIkXRvHlzxMTEYOLEicw8QQhBdHS0mMaQACcnJzg5OcHS0hImJiaIj49HTk4OjIyM0KNHD/z6669ylys1NRVdu3ZV2ubew8ND7MSe7ZowsrajPOFOaOcYYecvxcXFqKiokBjuhC+tW7fGvn37UFpaiurq6joLD/epITyO1D4kl3YglpmZKbIJ/OKLL5CdnQ1tbW2FaaPRhFQBauZkgblYcXEx3r17p7DnxdnZGc7OzvD19cWqVaukpmczMTp37hzOnTsnVROBhqSkJPzyyy948eIF1NTU0L17dwQGBjLqp2wsX75cJk/FfO9dGC0tLTRp0gTV1dUghKB79+5ITU1lTStPPny15B4/foxFixYxWk6VlZXYtm0b69wsz1xR1+rpinJsVi82q1wunwGI2D3VZsSIEYiJicGjR48AyDYBb9iwAY8fP0ZiYiJGjx6NS5cuSdXdlrVDCaBRT+7Tpw8OHDiA+/fvA6jZxHAtcuTBwMAAZ8+eldmDLo26IV/++OMPJgRNbZu6bdu2SYxPW9eq03xtfPlCe3p//vx5eHl5MW3h6ekJPz8/iSqanxOCtw9z5sxB7969GU96bNCEFaHF398fubm5iI+Ph4eHB5o2bYp169ZJtNuhCUNDE1qArybCgAEDsHPnTpSXlyMuLg6hoaESY0YCNW9WTUxM0L9/f5F7kWSzKmxDPGzYMAQHB0tUA/wUCQgIgLe3N3x9fZmNemZmJvT09FhP14Eaz/Tdu3dnDhtmzpyJqKgoREZGIj4+XiHlWrx4MSIiImBra4vff/+9zjb3eXl5yMvLQ1lZGZKSkkTU5yTZrPKFb7gTATS2Xnx9VdCSlpaGtLQ0kbcdiopn+qkiODjT1tYWCaGTkZEh1Sb4m2++gZeXFxP3OjIyEt26dUN5ebnCwrbxDakiKMeuXbvq/HmRdRPFd/MkL8uXL8fMmTMZT+4nT57E8uXLER4ezinz9ddf4+jRoyKeiiVpLfLdQAI1a7iKigro6ekhKCgI7du3l3ogQpMPAFy7dg03b94EUDOXsWkrCli7di3WrVvH7D9u3boFPz8/HD58WCytPHOFsHo6IQRJSUkKVU+XZ/0uApVbJiUzatQoiR9plJWVkZKSEuYjiUmTJpGKigrGS1tGRgaZP3++1Dz4eNMyMzMjSUlJzN/JyckyecjKzc0lsbGxJDY2luTm5kpNTwNfD7qCehJ4KCwrKyPW1tYKLZOwt8janiMleZKcOXMmyc7OZtLcvn2b2NjYKLRsiYmJxNzcnPTu3Zv07t2bmJubk5cvX5Li4mJy/fp1hebFByMjI5KcnMz8nZKSQoyMjD5aeT4WsvT9/Px8Tk+T0jwu88XExIRUV1fz8u4ZFxfH+pGVu3fvktjYWFJRUSExnYWFBcnKyhIp2/jx4znTl5eXk+DgYGJpaUkmT55Mtm/fLjWPiIgI1s9/hdzcXPLkyRPy5MkTmcbwp0+fklGjRhEnJycybtw40r9/f7J9+3by4MEDhZRn/Pjx5OzZs2TMmDHk8uXLYh9FsX//fjJq1Ciir68vMnebmJiQ8PBwheUjzP3798muXbskpjlx4gTJyMgQmbNk9YptampKioqKRPrLxIkT6QrLwYYNG4iBgQGZPn06sbGxITY2Np+EV/OPRXFxMcnJyZGYprCwkAQEBBBzc3Nibm5O/P39SUFBAamoqKizdZMsKON5kYecnBySnp7OfBQN271Km/sWLVok0zV5eP78OfNceXp6EhcXF/L06VOF5kEIIXv27CETJ04kwcHBJDg4mEyaNImEhIRwpufrCV2eueLdu3fMmK9or9u06/fa1Is3q7QnQOfPn8eaNWuQnZ0NADI5QNDS0kKDBg2gpqaGiooKtG3bFhkZGRLzcXV1xZYtW0TUGQTX2KBRTxa8KevVqxcIIXX2poyv3RqNuiFfiJBKEKmlZlj7b2HqWnWaxsaXloSEBCZwu7BRPtez3LBhQxFHYV27dq0zB1ifInz6viQ1LEXZ7Qho2LChyBt3aSe4gOTYqLIgzdO4MHw0ETQ1NeHo6MjrjRSf+L5LlixBUFAQpzdJSW+WP1VatmzJS722R48eaN26NXPSLXAidPjwYRGbKVoWL16MI0eOICcnByEhISLfSXP8xQdbW1vY2tpi586dvO2vaenfv79UO1R5bL34+qqg4ezZs7h48aLMZjmfO02aNJGqoqitrc3p1Odj2q0r43mh4datW1i2bBlyc3Ohrq7OqLJL8jpLg76+Pu7du8fMR3///beICRUbbGZbycnJCi1X9+7dAdQ8W2vXrlXobwsTFRWFw4cPM3155syZmDZtGqdTrsaNGyMuLg4GBgYAamKnS9IqkGeuaNasGQYPHsxobygycgTt+r029WKz+vLlS2RnZ2PIkCEi12/evIk2bdpwLiiDgoKwefNm9OvXT2b1j6ZNm6K0tBT9+/fHsmXL0Lp1a6mLfL4disY+6Ndff8Xhw4eZDUhqaiocHR0VvlkVNu6WBRp1Q74IL1RrL1olqdrWteo0jY0vLatXr8bPP/8Mf39/hISE4NChQ6z2SwJbvtGjR2PHjh2wtLQEIQQRERGMw4r/AjR9Xxl0794dJ0+eBCEEr1+/xu7du6WGelKGXTjA34lbUVERgoODcfv2bQA1rvgdHR1Z+8Lvv/8OW1tbzkM5No+Atra2AOTzJvk5IPDMDtQcJpiamjKqdPIyZswYjBkzBv7+/li+fLlCflMSDg4OSExMRFxcHICaMA2SbNb4QBPuRB5bL76+Kmho3bq1aqPKE2WEeqJBGc8LDUFBQdi/fz8TDuzYsWN4/fq1wvNJSEjAzJkzmXB6r169Qvfu3RmTOOHDx/DwcBw5cgSpqakiJnOFhYW8o3VIIy8vD35+frh58ybU1NQwdOhQrFixok4ONoT7srR+7enpCVdXV+bAuKKiQqpzMZq5gualHh9o1+9iv0P4bG0/Eg4ODvj555/FHIw8f/4cv/76K6dNlaWlJe/T95ycHHzxxReoqqpCaGgoCgsLMXPmTMa7oDCCDpWcnCwy4Qo6VO1y8fXsKoyZmRlOnDghcq0uwhi8ffsWQUFBSEhIEFkYy2Jsf+/ePcY2kMtrGQ1Dhgxh3IifOHGC+T8hBCdPnsSNGzc4ZYWdDNWF05DAwED07dtXZhtfWiwsLBAREQFjY2NER0cDqLF1rv0mXE9Pj9oh0+cETd9XBkVFRQgICGC0RQwNDeHp6SnxbYEg3MupU6dE7MI9PDwUWrZHjx7B29sbr1+/hp6eHqOJwHX6vWjRImhrazOLiYiICBQUFLBOqIcPH8bUqVM5bWFkdcJQXl6O9+/fi70BVlE/OHHiBDZu3Mi8sb169So8PDzEPCXTYGNjgzlz5qCwsBBbtmyBm5sbTExMEBsbiwMHDrDaBgpsveLj4xEcHIxu3brB0tKSsfWSdPCanJwMDw8PJCcno2XLloyvCsFiXB4E8bvj4uKQkZEBIyMjERvvz91mlQZB+A5XV1fGrq+uQj3RkJKSAnd39zp5XuRBsLaYNGkSYmJiRK4pkjt37kj8XliDKD09Ha9fv4afn5+Ifai2tjZ0dXUV+kbaxcUF33zzDaZOnQpCCMLDw/HixQuF2m0CYA4DraysANRszgkh8Pf355SpqKgQ8QZcO+ayIhg7diwCAwPr7GBfnvW7MPVisypp4Sm8eBcg2BQeOHAAOjo6mDBhgshAr6jX23w7lGAjwQXbRkJwLyEhIWjQoIHIm7LKykqFedoSMGfOHEyYMAH79u3DunXr8Oeff6Jz56ehuQ4AABzSSURBVM5S8ykvLxdxAKHIwOjSBg2ustVWnX727JnCVacHDRqEd+/eoVGjRmjcuLFMgZtpsLKywtGjR2FjY4OVK1eibdu2mDx5stKdJHzqKKvvKxMTExOcPHmSGevKy8sxa9YsVkcL8lJYWCizJsL48eNx5swZqdfkxc3NDb6+vtDU1ISpqSny8/Nhb28vMabh58SVK1c+m82JiYkJ9u7dyxw2ZGdnY968eTh58qTcvy28MRk1ahT++usv5ju2w14Bz549g5OTE3r27Mlocc2fPx9DhgyRqj5XVVUls/MXPsycOZPzOzU1NYVrVXwObN++Hbdu3cLTp08xceJExtmWoscjeair50UepkyZgvDwcMybNw8zZsxAhw4dsHDhQoV7A/5UMTU1RVRUlNRr8lJSUsI8o0DNJm7hwoWcB9U3b95E7969Ge2mgoICxMfHS3T4SjNX1PXBPu36vTb1Qg24sLCQ8zs2b3+1w334+vrKHLQ7OTkZO3fuRFpamoinYbbG7NChAzp06MCcRuXl5Ul8cyfw7BocHAwtLS1YW1uDEIKjR49yei2sfS/CdrBqamoK36zm5+fDysoKBw4cQP/+/dG3b19YW1tL3BDWpQoBQO/6Whmq0zQBnGmYMGEC8vPzYWdnh2nTpqG6uhqLFi1SSt71CXn6vjIwMjKCjY0NzM3NZQ5DoQy7cAE6OjoyT3Zt2rQRGfPy8/M5Q9JIs8mXFBg8JSUFOjo6OHv2LAwMDLB8+XJMmTLls9+sZmRkoF27dtixYwfTJl5eXrzj1H5qCL8VV+QbcuFzdz7hTmhtvfj6quADTTzm/zrKCPUkL0lJSSJvGBXtE4GGWbNm4f3793B1dYW7uzsKCwvrxCSAxveAMvwVVFdXIzc3F61atQIA5ObmyuRLgi8aGhpYsmSJzOnXr18vohWgra0tdk2APHPF2LFjERYWVmcH+/+p0DUtW7bE06dP0bNnT5HrT58+ZXWOIk+wbldXV5iamsLc3FzmU69Hjx7B1dUV1dXVuHLlCh4/fozw8HDOB+XChQsiD9y8efNgYWHB6nhCGYHHhRGoGTRp0gRv3rzBl19+KXFhrAzbwMTERBHHJKdPn0ZMTAw6deoEFxcXTt1/ZTgZ4mvjS4sg4PSIESNw584dlJWVqWyZWFB2f+HLpk2bcOjQIWzfvh3jxo3DjBkzGAcPXCjDLpwPgo1nixYtYGpqyoQ7uXz5MqczJ8HpcVpaGu7evYuxY8cCAC5evCgSMogNwaHh3bt3MXLkSDRu3PiTskOuK5YuXYq8vDzk5+cjOjoa+vr6ePz48ccullx07twZW7duhbW1NQDg6NGj6NSpk0J+W55wJzS2Xspw/jJt2jT8+eefUq+pUE6oJ3k4dOgQdu7ciR9//BEAsHv3bjg4OGD69OkftVw//vgjtLW10adPH1y4cAEAGIeRikTY90BZWRlOnTqFNm3aKFyGL/PmzYOZmRnTLleuXJEahogGQ0NDmJiYYPr06TKNeYIDdgHq6uoi2ovCyDNXCA5x6upgn3b9Xpt6oQZ87do1eHl5wcnJCb179wZQEzA3ODgYPj4+InF75IXGxmHq1KlYs2YNPDw8GFWjiRMn4tSpU6zpx48fj507d6JLly4AaiY9e3v7T0JdJTAwEPb29rh8+TICAgKgpaUFIyMjkclcGGXYBlpaWiI4OBht2rTB48ePYWtrC0dHRzx//hwNGjQQi1OoTNVpeWx8+UAIwbFjx5CamoolS5bg9evXyMrKwnfffafQfFQoh4KCAhw/fhz79+9Hx44dYWtri3HjxkmVqyu7cD7Io9Yza9YsbNmyBS1atABQ8zbW1dVVolqjq6sriouLkZycjJiYGKirq8Pa2lrhalqfIqWlpZg8eTImTJiA+Ph43Lx5EyNHjsTgwYMxbdq0j1083uTm5mLNmjWMM5MhQ4ZgxYoVzFuNuqCkpASlpaUKy4Ovrwp5qL0eqaqqgrGxMU6fPq2wPD4n5FHprmt++uknhIWFMc9hXl4epk2bhnPnzn3UcrGteZVh60sIwbRp03iZs9DIyMKLFy+YN94GBgb49ttvFfr7QM3Yd+TIERw9ehTdunXDjBkzJMa1nj59OpYuXYq+ffsCAB4+fAh/f3/Oe/9U5wq+63cu6sWb1eHDh2PNmjUIDg5mXEvr6+vD19cXw4cPV3hefPW+KyoqxNQ5JBlCu7m5YcqUKYzjkqdPn34yql2CkywzMzMMHDgQRUVFrG9+BBvCulYhAGpO1ASnaefOnYO5uTkWLFiAqqoq1tNvZapOe3p6YsKECXj27Bk2bNjA2PgqGn9/f+Tm5iI+Ph5LlixB06ZNsW7duk/SiZAK6Tx8+BBxcXFo1KgRhg8fjsOHD+P06dPYvHmzRDk+YWjqCnn6T05ODrNRBWrezubk5EiUCQwMxPXr16Grq4smTZogMzOzTk6+PzUWLVoEAwMDaGlpMXVuamqKhQsXKtwmXlk0bNhQTB2zLt7iCCNLuBM+c/7QoUPRpUsX+Pn5iaivC3xVKIKQkBCEhISgqKhIxEbtw4cPMDY2VkgenyN1HepJHpo2bSpyYNKyZUuZTUHqgsrKSlRUVKC6uhofPnxg1kuFhYVSnYEqgqKiIqljvyJkZKF79+5SNZzkpVWrVli4cCHs7e1x6dIl+Pj4wM/PDzY2NpgxY4bI+hmoCd3m5OTE7C0SExM5D4o/5bmC7/qdi3qxWQWAYcOGYdiwYXWez+DBg7Fw4UKoq6tDS0tLJoc5WlpaKC4uZl7ZJyYmij14wowbNw7ff/8946W2X79+HzX+F8DuqbhFixZo0aIFa8wlZdoGCqv8PXz4EDNmzABQYwPApqqtTFVQvja+tMTFxeHEiRNMnMoWLVqIvMlVUT/Yu3cvjhw5gk6dOmHmzJkYOXIk1NTU4ODgwKjG1hdoQkR88803WLFihYgHYWl2W40aNRKxM2/bti2nbeznhLOzM27duoU3b95g3Lhx6NSpE3Jzc1FcXIxZs2Z97OJRMXPmTLE3NmzXlAWNrRdfXxU0WFtbw8jIiNV5Y7NmzRSa1+dGXYZ6kgdBSBTB2BcZGYnhw4cjMTERgPLtV3fu3InffvsNampqIht5bW1txuxIkQjbn1ZXV+P169dS86GRoSkPG3XxIqC0tBRRUVEICwtD586dYWVlhbi4OCxYsEBMu6h///44deoU/vnnHwA1+wSuvi/PXJGQkABvb28kJCSgvLycua6oNTzf9TsX9WKzeuvWLYkesGRNIwurVq2Cv78/9PX1ZbaLcnBwwLx585CVlYVly5bh2rVrCAoKkijTqlUrGBoayl1eRVF78wlA4uZTmRvCdu3a4dChQ2jbtq2IN7Ty8nJOx1TKgq+NLy0NGzYUGVjrwgGAironPT0dO3bsYI0t+Sk4AJEFQYiIK1euMJtVWdXG1q1bh+3btzObAQMDg/98HFUuBKf9p0+fZmIOzp07FxEREfD29mY2S/WBj/0Whwt5bL34+qrgg46ODnR0dLBr1y4UFRXh33///ah26vUJgdokAISGhn7EkogiMAur/eIjOjoaampqSve+6+zsDGdnZ/j6+oociNQVwuO8hoYGOnXqxMtmVVYZmvIoA19fX5w/fx6GhobYsGED8ybX2NiYM/Rhs2bNZNL4kGeuWL16NX7++Wf4+/sjJCQEhw4dUugbf4Wt30k9wMTEhJSWlpKSkhLOj6mpqULymjx5MpVcWloaOXToEPnjjz9IamqqQsqioob09HRiZ2dHTE1NycmTJ5nrly9fJoGBgR+xZIQEBASQ/Px8EhkZSQwMDMjw4cPJ2rVrFZ7PihUrSFRUFDExMSGvXr0iXl5exNvbW+H5qFAhjd9++43MmDGD9O/fn6xcuZKEhYURIyOjj12sz5ZDhw4x/1fUPKdstm3bRnR1dYmenh7R1dVlPt9//z357bffPmrZSkpKyPjx48m2bduIg4MD6dOnD3FxcSFhYWES5aytrcnLly9F2mTChAkKLdvly5fJiBEjyKhRowghhDx69IjY29srNI/PjcuXL3/sIqjgoKKigrx48YK8ePGCVFRUfOziKJWQkBDy7t071u8yMzMVkgfNXGFubk4IIWTSpEnMNQsLC4WUhxDFrd/rhYMlQXxStqIKrn/55Ze4fv263Hnt3LkTzZo1w/jx4+t9fEYVyuXNmzecNr7yUlRUhICAACauqqGhITw9PaXaYqn4tFCWQy5lYGlpiaVLlyI+Ph7bt29Hq1atpIaIoFEdVvF/pKenK80DeV2grLc4siKw9Tp69CjjHNHU1BSBgYG4deuWRJXDyZMn4/j/a+/+Y6Ku/ziAP7kUdfxMl24VCVIOCQVXKdLUjbXkhwcEGGFFcxN/horKVv5qerOsS02MHy6y1EydduTPpivL2Dhh5iLHr2b8VGNLUDzFEzju+we7z9dTFE8+v+7u+fgrPuK93znxPq/7vN7P1w8/2M1wfdg818eRkpKCwsJCZGZmCq8bFxfHgKU+2Fq633rrLSGERk2jnn7//ff7wkAPHDggJGO7ugsXLmDJkiXC8bru7m5s3779oR0D58+fh16vR3NzMywWiyRz7Ovq6lBQUIDm5uZ+x1U6k0d9r5g1axYOHjyId955B2vWrMGoUaOQkpIi3GuqhVO0AcvZcmoLOFm/fv0jn8E8d+4ctmzZgqamJsl+oNyZnG3gj8rRM74DYbFYsHPnTrtxDOScbIFcNTU1kgZySelxRkQMpHWY/s+ZC1UAqipUgYGd9XI0q+Jx3TuL1jZ3meypfdSTXq9HeXk5li9fDrPZjDVr1uDq1atuU6xu3LgRH3/8sXCfZjQaodPpHprsu3r1aixatEjS0YjLly9HTEwMkpOTHTpDqXaP+l4RFxeHa9euYd68eUhPT0dPTw+ysrJE24dY9+9O8WRV7WbMmIFly5YhLCzM7gfK2W8s1CIxMREHDhzo88m6TXp6uqifaPenr6f9UgVMAfKMCCLp2Z68aLVaHD16FD09PUhLS8PBgweV3ppDHB0RkZeXB6PRiKqqKsTHxyM0NBS7d+9WxbguorS0NLuzXlOmTEFFRcVDz3qdOXNGeCIzdepUIasiKipKtH1lZGRgy5YtyMzMRHFxMcrKyvDll19iz549oq3hStQ6vsO2N51Oh7q6Oty4cQNxcXFCmKc7SExMvG/cWF/X7ibHB5oJCQk4cuSIpGs4i66uLty5c+eRZ58+CrHu353iyara+fr6IjY2VultuKza2lpMnDix3zZwOcn5tB/oHdz99ddfIykpya71l+3pzkWuQC6pOToiYvHixVi8eDFSU1ORkJCAyspKXL16FTExMf22DhNJzZYWGxgYCF9fX2FE3sNMnz4dY8aMQUlJCaxWKxYuXCjMThfLypUrkZmZiUuXLuHdd99FQ0MDCgoKRF3DVah5fAfQ+14dGhqKkpISaDQaREVFuU2hCvT+/5eVlWHy5MkAgPLy8n7vX6ZNm+bwKElHRUREoKamBiEhIZKtoWbp6enYt28fgN77k8GDB9tdGyix7t/5ZFUEu3btgqenJ8+5kmTu/odUyie4JK1PP/0U8+fPx2+//YZNmzbB09MTMTExduMWnEVFRYWQvDlnzpyHJm/aWocrKyuRn5+P4OBgpKamCq3DTDoltVDbuWCTyYTz588D6E3t9/X1VXhH6vT333/DaDQiLy8P/v7+CAgIQG1tLbZt24aIiAjFWzyzsrLQ0dEhnMHMyclBamoq5s2bp+i+5GJLz7a1sXd1dSE3NxdhYWEP/D2RkZG4fv06vLy8HnmUpKOSkpJw8eJFBAUF2d2/u0sn271Pry0WC7RarerOxbNYFcGxY8ewdu1amM1mAGAhQUT9kjKQS40cbR0mUjM5sir6ykbgh+AP9zgt3XIoKCjAggULhDPO7e3tWL16tdCd4g66urpQX18PAAgKChI6jR7k8uXLfV4X88Ok8vLyPq9PmjRJtDXUqKioCEVFRbh58yZ8fHyE62azGVqtFhs2bFBwd/djsSqC6OhobNu2zaHZrETkPvqbJemMN6CP055lu5EEeluHMzMzYTQasWnTJim2SCQZObIqbNkIdxs0aBAmTJgAnU6HMWPGiLaWq/j+++8xe/ZsAOKnMw/UvTNzbR9wuIPS0lKMHz9eKIxu3LhhN3dTKY2NjaK37zsDk8mE9vZ26HQ6u+A7b29v+Pn5KbizvrFYFcHdUelERPeSO5BLSgMZEeFI6zCRmtlGPkhpx44dGDJkCFJTU2G1WlFcXIxr164hICAAxcXFDFrqh5paus+cOYN169bhiSeewOnTp3HhwgXk5eWhsLBQ6a3JIikpCcXFxUJx3tPTg5SUlD4DlHJycqDX65GSktJnMS9mi258fDxu376NyZMnIzIyElOmTMHIkSNFe31n0traiubmZlV2OjFgSQSRkZHQ6/WIi4uz63l//vnnFdwVEamF3IFcUhrIiAhboQqAhSo5tZkzZ2Lfvn2SZlWcPHkSBoNB+DojIwPJyckwGAz8+XkEailUASA3NxeHDh1CZmYmAGD8+PFoampSeFfyufcpskajgcVi6fN733vvPQC97zVSO378OP777z+Ulpbi7Nmz2Lx5M7y9vVV3ZlMqs2fPxo4dO2C1WpGUlARfX19MmzZNlj97R7BYFYEt9vruMQweHh745ZdflNoSEZEkdu3aJYyIaGxsxIkTJ1BfX48lS5b0OyJC6mRHIrmMGDECa9euFc52SdElcfv2bTQ3NyMgIAAA0NzcjI6ODgBQPDCIHOfOM3O9vLzsOmsqKirsJhvczRa6JMe5UavVin///RdXrlzB5cuX4e/vj5deeknyddWio6MDPj4+OHz4MLRaLVauXInExEQWq67o9OnTSm+BiEgWjzMiwtY6XFBQIBSrj9o6TKRGW7Zswe7duyXNqli2bBlSU1MRFhYGq9WKqqoqrF+/Hrdu3UJMTIwka5I0vLy8cPXqVeHpYllZmV2wjavLycnB4sWLhY7Dixcv9hsuVVdXh8LCQjQ1NaG7u1u4LmYb8CuvvIIXXngBaWlp0Ov1GDVqlGiv7Qw6OzsB9P59jI+Ph0ajUeUHYSxWRdTa2oo7d+4IXz/99NMK7oaISHzvv/8+jEYjrly5gtdffx0BAQFobW3FrVu3kJGR0efvGUjrMJEajRw5EuPHj5d0jRkzZuDll19GRUUFgN42+hEjRgAAFixYIOnaJC53n5k7ceJEHD9+HH/++SeA3vmm/QX5LF26FImJiXjjjTckK6Dmz5+Ps2fPoqioCH/88QeioqIwefJkDB8+XJL11GbSpEmIi4uDxWLB+vXrcePGDVUGxTJgSQRGoxEffPABWltbodFo0NXVBX9/f1UMoiYikoKjIyJsrcNxcXGorKxEaWkppk+f3m/rMJEaffHFF+jq6pI8q+LatWt2xeqTTz4p6uuTfDgz1zH3zgCVUmdnJ06cOIHc3Fy0tLSgqqpKlnWVZrVaUVNTg4CAAHh7e6OtrQ0tLS0IDQ1Vemt2+GRVBHq9Ht9++y2ys7NRXFyMQ4cO4dKlS0pvi4hIMomJiQCAwMBA+Pr6YuPGjQ/83sdpHSZSMzmyKkpKSpCTk4PQ0FBYrVbU1tZCr9fj1VdfFW0Nko+Pjw/P7Dtg6tSpkuccnDx5EkajEUajERaLBVFRUYqP05GTh4cHAgMD0dLSgpaWFgDqPEvNYlUkQUFB6O7uhoeHB2bNmoXk5GRkZ2crvS0iIknYZhkCQF5e3kO/93Fah4nUTI6siq1bt2Lv3r0IDg4GAPzzzz/IyclhsUpuYcqUKVi0aBE0Gg08PT2FEDMxP+A8deoUIiMjMXfuXDz77LOiva6z2Lt3Lz7//HP4+/sL56nVGBDLYlUEgwb1/jGOGjUKp0+fxjPPPIP29naFd0VEJI/+RkSMHTsWY8eOxYkTJ+xahw0GAz766KMHtg4TqZ2UWRXd3d1CoQoAwcHBdkEzRK5s3bp1+OSTTyQNMdu8eTOA3lTcjo6OByYUu6qdO3fi2LFjqhrz1BcWqyLIyMhAe3s7li5dihUrVsBkMmHVqlVKb4uISFUcaR0mUjM5siqGDx8Og8GA5ORkAEBxcbHbBL8Q+fn5SZ563dzcjBUrVqC6uhoeHh4IDQ2FXq8XxkW5uqeeekr1hSrAgKUB6+npQUlJCc8hEBE54PLly07xJknUl+TkZGzevPm+rAoxj/80NjYiJycHNTU1AIBx48ZBr9fjueeeE20NIrUqLCyEn58fYmNj7ULMhg0bJtoac+bMQXx8PFJSUgAABoMBx44dwzfffCPaGmqWm5sLs9mM+Ph4SYPiBorFqgjkTCwjIiIiZSUnJ8NgMGDmzJlCG7vtmthu3boFoHdWJ5G7CAkJEf7bw8NDOLNaXV0t2hqJiYk4fPhwv9dcVXR09H3XeGbVRYWEhOCvv/7ChAkTlN4KERERSUyOrIr09HTs27fPrki1XSNydbaOAilpNBrU1dVhzJgxAID6+nrJZrqqkRxBcWJgsSqCyspKpKenY/To0XaHsw8dOqTgroiIiEgKcmRVmM1mu68tFgvDG4lElJ2djbfffhvjxo0TxkN99tlnSm9Lcp2dnfD09MTt27f7/HUxW63FwDZgEZSXl/d5fdKkSTLvhIiIiKQkdVZFUVERioqKcPPmTfj4+AjXzWYztFotNmzYIMm6RO6ora0NFRUVAIDw8HC3CDGzHV8MCQkRWqxtxG61FgOLVSIiIiIHSJlVYTKZ0N7eDp1Oh3Xr1gnXvb294efnJ8maRO7kQU8UbdT2ZNHdsVgVgclkwldffYXq6mq7eWu7d+9WcFdEREQkhQ8//BDp6enMqiByQrYnig+itieL7o5nVkWwatUqBAcHo6GhAUuXLsUPP/yAF198UeltERERkQSkzKpISUl56I008zCIBsYW3pSfnw9PT0+kpaXBarXi4MGD6OrqUnh3dC8+WRVBQkICjhw5Aq1Wi6NHj6KzsxMZGRnYv3+/0lsjIiIikUmZVfGg1xZzDSLqu51fqhFU9Pj4ZFUEnp6eAIDBgwfj+vXr8PPzQ1tbm8K7IiIiIilIWTCyGCWSh9lsRmNjI0aPHg0AaGpq6vc8K8mPxaoIAgMDcf36dWi1WqSlpcHHx4dtwERERC5KjqwK5mEQSSs7OxtvvvkmwsLCAABVVVXQ6XQK74ruxTZgkZ07dw4mkwlTp04VhoYTERGR68jKykJwcDCOHz9ul1WxcuVKp1qDyN21trYKo2siIiLcYnSNs9EovQFX0dbWhl9//RUmkwnh4eEsVImIiFxUY2Mjli1bhqFDh2LmzJnYsWMHzp0753RrELm7ESNGIDo6GtHR0SxUVYrFqghOnTqF2NhYfPfdd9izZw/i4+Px888/K70tIiIiksC9WRWDBw8WPatCjjWIiNSOj/9EsHXrVuzfvx9BQUEAgIaGBixcuBCvvfaawjsjIiIiscmRVcE8DCIinlkVRVJSEn788Ue7a33FYRMREZFrkSOrgnkYROSuWKyKYPv27Rg0aBBSU1NhtVphMBjQ3d2NuXPnwmq1YtiwYUpvkYiIiETU1tYmBLOEh4fzvBsRkQRYrIogJCTkgb/m4eGB6upqGXdDREREUjp16hTWrl2LsLAwWK1WVFdXQ6fT8fgPEZHIWKwSEREROSA2Nhb5+fn3ZVX89NNPCu+MiMi1MA2YiIiIyAFDhgwRClWgNwxp6NChCu6IiMg18ckqERERkQOYVUFEJA8Wq0REREQOYFYFEZE8WKwSERERERGR6vDMKhEREREREakOi1UiIiIiIiJSHRarREREREREpDosVomIiIiIiEh1WKwSERERERGR6rBYJSIiIiIiItX5H5ynHj47jCqlAAAAAElFTkSuQmCC\n",
            "text/plain": [
              "<Figure size 1152x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "zIwrF7y6iS1l"
      },
      "source": [
        "Create a second bar plot showing the score for every input word being the \"end\" word."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "6tXEqIp-Tzou",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "6f4f140d-cd1d-415d-b922-c8da26573cf1"
      },
      "source": [
        "# Create a barplot showing the end word score for all of the tokens.\n",
        "ax = sns.barplot(x=token_labels, y=e_scores, ci=None)\n",
        "\n",
        "# Turn the xlabels vertical.\n",
        "ax.set_xticklabels(ax.get_xticklabels(), rotation=90, ha=\"center\")\n",
        "\n",
        "# Turn on the vertical grid to help align words to scores.\n",
        "ax.grid(True)\n",
        "\n",
        "plt.title('End Word Scores')\n",
        "\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1152x576 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "awgi7Z_a9KSq"
      },
      "source": [
        "**Alternate View**\n",
        "\n",
        "I also tried visualizing both the start and end scores on a single bar plot, but I think it may actually be more confusing then seeing them separately. "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "m4VUk6R05uXS"
      },
      "source": [
        "import pandas as pd\n",
        "\n",
        "# Store the tokens and scores in a DataFrame. \n",
        "# Each token will have two rows, one for its start score and one for its end\n",
        "# score. The \"marker\" column will differentiate them. A little wacky, I know.\n",
        "scores = []\n",
        "for (i, token_label) in enumerate(token_labels):\n",
        "\n",
        "    # Add the token's start score as one row.\n",
        "    scores.append({'token_label': token_label, \n",
        "                   'score': s_scores[i],\n",
        "                   'marker': 'start'})\n",
        "    \n",
        "    # Add  the token's end score as another row.\n",
        "    scores.append({'token_label': token_label, \n",
        "                   'score': e_scores[i],\n",
        "                   'marker': 'end'})\n",
        "    \n",
        "df = pd.DataFrame(scores)\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "07xyo-I97Ntt",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "b9573949-4e42-4b5a-f3b7-1eeaea771688"
      },
      "source": [
        "# Draw a grouped barplot to show start and end scores for each word.\n",
        "# The \"hue\" parameter is where we tell it which datapoints belong to which\n",
        "# of the two series.\n",
        "g = sns.catplot(x=\"token_label\", y=\"score\", hue=\"marker\", data=df,\n",
        "                kind=\"bar\", height=6, aspect=4)\n",
        "\n",
        "# Turn the xlabels vertical.\n",
        "g.set_xticklabels(g.ax.get_xticklabels(), rotation=90, ha=\"center\")\n",
        "\n",
        "# Turn on the vertical grid to help align words to scores.\n",
        "g.ax.grid(True)\n"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1793.47x432 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8UyBYNmeegGf"
      },
      "source": [
        "## 5. More Examples"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MWtcRpPef-Ce"
      },
      "source": [
        "Turn the QA process into a function so we can easily try out other examples."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rH8NbBlsfxZ_"
      },
      "source": [
        "def answer_question(question, answer_text):\n",
        "    '''\n",
        "    Takes a `question` string and an `answer_text` string (which contains the\n",
        "    answer), and identifies the words within the `answer_text` that are the\n",
        "    answer. Prints them out.\n",
        "    '''\n",
        "    # ======== Tokenize ========\n",
        "    # Apply the tokenizer to the input text, treating them as a text-pair.\n",
        "    input_ids = tokenizer.encode(question, answer_text)\n",
        "\n",
        "    # Report how long the input sequence is.\n",
        "    print('Query has {:,} tokens.\\n'.format(len(input_ids)))\n",
        "\n",
        "    # ======== Set Segment IDs ========\n",
        "    # Search the input_ids for the first instance of the `[SEP]` token.\n",
        "    sep_index = input_ids.index(tokenizer.sep_token_id)\n",
        "\n",
        "    # The number of segment A tokens includes the [SEP] token istelf.\n",
        "    num_seg_a = sep_index + 1\n",
        "\n",
        "    # The remainder are segment B.\n",
        "    num_seg_b = len(input_ids) - num_seg_a\n",
        "\n",
        "    # Construct the list of 0s and 1s.\n",
        "    segment_ids = [0]*num_seg_a + [1]*num_seg_b\n",
        "\n",
        "    # There should be a segment_id for every input token.\n",
        "    assert len(segment_ids) == len(input_ids)\n",
        "\n",
        "    # ======== Evaluate ========\n",
        "    # Run our example question through the model.\n",
        "    start_scores, end_scores = model(torch.tensor([input_ids]), # The tokens representing our input text.\n",
        "                                    token_type_ids=torch.tensor([segment_ids])) # The segment IDs to differentiate question from answer_text\n",
        "\n",
        "    # ======== Reconstruct Answer ========\n",
        "    # Find the tokens with the highest `start` and `end` scores.\n",
        "    answer_start = torch.argmax(start_scores)\n",
        "    answer_end = torch.argmax(end_scores)\n",
        "\n",
        "    # Get the string versions of the input tokens.\n",
        "    tokens = tokenizer.convert_ids_to_tokens(input_ids)\n",
        "\n",
        "    # Start with the first token.\n",
        "    answer = tokens[answer_start]\n",
        "\n",
        "    # Select the remaining answer tokens and join them with whitespace.\n",
        "    for i in range(answer_start + 1, answer_end + 1):\n",
        "        \n",
        "        # If it's a subword token, then recombine it with the previous token.\n",
        "        if tokens[i][0:2] == '##':\n",
        "            answer += tokens[i][2:]\n",
        "        \n",
        "        # Otherwise, add a space then the token.\n",
        "        else:\n",
        "            answer += ' ' + tokens[i]\n",
        "\n",
        "    print('Answer: \"' + answer + '\"')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DVlKTK-njWrX"
      },
      "source": [
        "As our reference text, I've taken the Abstract of the [BERT paper](https://arxiv.org/pdf/1810.04805.pdf).\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "y4VPq6FdjxyX",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "bfd5f9e3-eac4-460c-ca5f-1bf3624c732a"
      },
      "source": [
        "import textwrap\n",
        "\n",
        "# Wrap text to 80 characters.\n",
        "wrapper = textwrap.TextWrapper(width=80) \n",
        "\n",
        "bert_abstract = \"We introduce a new language representation model called BERT, which stands for Bidirectional Encoder Representations from Transformers. Unlike recent language representation models (Peters et al., 2018a; Radford et al., 2018), BERT is designed to pretrain deep bidirectional representations from unlabeled text by jointly conditioning on both left and right context in all layers. As a result, the pre-trained BERT model can be finetuned with just one additional output layer to create state-of-the-art models for a wide range of tasks, such as question answering and language inference, without substantial taskspecific architecture modifications. BERT is conceptually simple and empirically powerful. It obtains new state-of-the-art results on eleven natural language processing tasks, including pushing the GLUE score to 80.5% (7.7% point absolute improvement), MultiNLI accuracy to 86.7% (4.6% absolute improvement), SQuAD v1.1 question answering Test F1 to 93.2 (1.5 point absolute improvement) and SQuAD v2.0 Test F1 to 83.1 (5.1 point absolute improvement).\"\n",
        "\n",
        "print(wrapper.fill(bert_abstract))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "We introduce a new language representation model called BERT, which stands for\n",
            "Bidirectional Encoder Representations from Transformers. Unlike recent language\n",
            "representation models (Peters et al., 2018a; Radford et al., 2018), BERT is\n",
            "designed to pretrain deep bidirectional representations from unlabeled text by\n",
            "jointly conditioning on both left and right context in all layers. As a result,\n",
            "the pre-trained BERT model can be finetuned with just one additional output\n",
            "layer to create state-of-the-art models for a wide range of tasks, such as\n",
            "question answering and language inference, without substantial taskspecific\n",
            "architecture modifications. BERT is conceptually simple and empirically\n",
            "powerful. It obtains new state-of-the-art results on eleven natural language\n",
            "processing tasks, including pushing the GLUE score to 80.5% (7.7% point absolute\n",
            "improvement), MultiNLI accuracy to 86.7% (4.6% absolute improvement), SQuAD v1.1\n",
            "question answering Test F1 to 93.2 (1.5 point absolute improvement) and SQuAD\n",
            "v2.0 Test F1 to 83.1 (5.1 point absolute improvement).\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tEB654YCknYv"
      },
      "source": [
        "-----------------------------\n",
        "Ask BERT what its name stands for (the answer is in the first sentence of the abstract)."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wfntqRCBegGj",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5e212582-8b5d-40cc-f7fb-206fe24ca013"
      },
      "source": [
        "question = \"What does the 'B' in BERT stand for?\"\n",
        "\n",
        "answer_question(question, bert_abstract)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Query has 258 tokens.\n",
            "\n",
            "Answer: \"bidirectional encoder representations from transformers\"\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "B6HcijzxkTO9"
      },
      "source": [
        "---------------------\n",
        "Ask BERT about example applications of itself :)\n",
        "\n",
        "The answer to the question comes from this passage from the abstract: \n",
        "\n",
        "> \"...BERT model can be finetuned with just one additional output\n",
        "layer to create state-of-the-art models for **a wide range of tasks, such as\n",
        "question answering and language inference,** without substantial taskspecific\n",
        "architecture modifications.\""
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MVNVGN5-gI06",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8094e8dd-2079-49ac-cd1d-52f43276861b"
      },
      "source": [
        "question = \"What are some example applications of BERT?\"\n",
        "\n",
        "answer_question(question, bert_abstract)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Query has 255 tokens.\n",
            "\n",
            "Answer: \"question answering and language inference\"\n"
          ],
          "name": "stdout"
        }
      ]
    }
  ]
}