Sequence_Classification_RCT_.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "name": "Sequence_Classification_RCT_.ipynb",
      "provenance": [],
      "collapsed_sections": [],
      "toc_visible": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "g3sED57COkjW"
      },
      "source": [
        "# Overview:\n",
        "1. *Classify a Randomized clinical trials (RCTs) abstarct to subclasses for easier to read and understand*.\n",
        "2. *Basically convert a medical abstarct to chunks of sentences of particaular classes like \"Background\", \"Methods\", \"Results\" and \"Conclusion\".* \n",
        "3. *Its a Many to One Text Classification problem. Where we categorize a sequence to a prticular class.*\n",
        "\n",
        "# Dataset Used:\n",
        "\n",
        "[*PubMed 200k RCT: a Dataset for Sequential Sentence Classification in Medical Abstracts*](https://arxiv.org/abs/1710.06071)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "es1rHu7wGgj8",
        "outputId": "c5125d3b-47bc-46f9-dd4d-252786b9c16c"
      },
      "source": [
        "# Getting the dataset from author's Github repo\n",
        "!git clone https://github.com/Franck-Dernoncourt/pubmed-rct.git"
      ],
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Cloning into 'pubmed-rct'...\n",
            "remote: Enumerating objects: 33, done.\u001b[K\n",
            "remote: Counting objects: 100% (3/3), done.\u001b[K\n",
            "remote: Compressing objects: 100% (3/3), done.\u001b[K\n",
            "remote: Total 33 (delta 0), reused 0 (delta 0), pack-reused 30\u001b[K\n",
            "Unpacking objects: 100% (33/33), done.\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vxlC1JHbm8XM",
        "outputId": "4e998e17-55b1-4c59-d749-ad69ad193161"
      },
      "source": [
        "!ls pubmed-rct"
      ],
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "PubMed_200k_RCT\n",
            "PubMed_200k_RCT_numbers_replaced_with_at_sign\n",
            "PubMed_20k_RCT\n",
            "PubMed_20k_RCT_numbers_replaced_with_at_sign\n",
            "README.md\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "W9UuvgpdnGxN",
        "outputId": "77bf30aa-8254-4ae9-defd-37772e655aba"
      },
      "source": [
        "!ls pubmed-rct/PubMed_20k_RCT_numbers_replaced_with_at_sign"
      ],
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "dev.txt  test.txt  train.txt\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "_jQWzeonnZ_C"
      },
      "source": [
        "* train.txt - training samples.\n",
        "* dev.txt - dev is short for development set, which is another name for validation set (in our case, we'll be using and referring to this file as our validation set).\n",
        "* test.txt - test samples."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4wCCvBa2nNv7"
      },
      "source": [
        "data_dir = \"/content/pubmed-rct/PubMed_20k_RCT_numbers_replaced_with_at_sign/\""
      ],
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "A7oDeeiRoKls"
      },
      "source": [
        "import os\n",
        "filenames = [data_dir +filename for filename in os.listdir(data_dir)]"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "mR2fv8w4olCG",
        "outputId": "c91d873d-9b2b-413d-a608-638b224cc3a5"
      },
      "source": [
        "filenames"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['/content/pubmed-rct/PubMed_20k_RCT_numbers_replaced_with_at_sign/train.txt',\n",
              " '/content/pubmed-rct/PubMed_20k_RCT_numbers_replaced_with_at_sign/dev.txt',\n",
              " '/content/pubmed-rct/PubMed_20k_RCT_numbers_replaced_with_at_sign/test.txt']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 8
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "B7tK-Gj-sOdQ"
      },
      "source": [
        "## Preprocessing"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "RS3JMTIlol9l"
      },
      "source": [
        "def get_lines(filename):\n",
        "  \"\"\"\n",
        "  \"\"\"\n",
        "  with open(filename, \"r\") as f:\n",
        "    return f.readlines()"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TN4r2abWsz6j",
        "outputId": "aef90234-bb0b-445e-b0be-65ba8c8ee4ea"
      },
      "source": [
        "train_lines = get_lines('/content/pubmed-rct/PubMed_20k_RCT_numbers_replaced_with_at_sign/train.txt')\n",
        "train_lines[:10]"
      ],
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['###24293578\\n',\n",
              " 'OBJECTIVE\\tTo investigate the efficacy of @ weeks of daily low-dose oral prednisolone in improving pain , mobility , and systemic low-grade inflammation in the short term and whether the effect would be sustained at @ weeks in older adults with moderate to severe knee osteoarthritis ( OA ) .\\n',\n",
              " 'METHODS\\tA total of @ patients with primary knee OA were randomized @:@ ; @ received @ mg/day of prednisolone and @ received placebo for @ weeks .\\n',\n",
              " 'METHODS\\tOutcome measures included pain reduction and improvement in function scores and systemic inflammation markers .\\n',\n",
              " 'METHODS\\tPain was assessed using the visual analog pain scale ( @-@ mm ) .\\n',\n",
              " 'METHODS\\tSecondary outcome measures included the Western Ontario and McMaster Universities Osteoarthritis Index scores , patient global assessment ( PGA ) of the severity of knee OA , and @-min walk distance ( @MWD ) .\\n',\n",
              " 'METHODS\\tSerum levels of interleukin @ ( IL-@ ) , IL-@ , tumor necrosis factor ( TNF ) - , and high-sensitivity C-reactive protein ( hsCRP ) were measured .\\n',\n",
              " 'RESULTS\\tThere was a clinically relevant reduction in the intervention group compared to the placebo group for knee pain , physical function , PGA , and @MWD at @ weeks .\\n',\n",
              " 'RESULTS\\tThe mean difference between treatment arms ( @ % CI ) was @ ( @-@ @ ) , p < @ ; @ ( @-@ @ ) , p < @ ; @ ( @-@ @ ) , p < @ ; and @ ( @-@ @ ) , p < @ , respectively .\\n',\n",
              " 'RESULTS\\tFurther , there was a clinically relevant reduction in the serum levels of IL-@ , IL-@ , TNF - , and hsCRP at @ weeks in the intervention group when compared to the placebo group .\\n']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 10
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "birRd_7mtVfY"
      },
      "source": [
        "Example returned preprocessed sample (a single line from an abstract):\n",
        "> Return all of the lines in the target text file as a list of dictionaries containing the key/value pairs:\n",
        "* \"line_number\" - the position of the line in the abstract (e.g. 3).\n",
        "* \"target\" - the role of the line in the abstract (e.g. OBJECTIVE).\n",
        "* \"text\" - the text of the line in the abstract.\n",
        "* \"total_lines\" - the total lines in an abstract sample (e.g. 14)\n",
        "\n",
        "\n",
        "```\n",
        "[{'line_number': 0,\n",
        "  'target': 'OBJECTIVE',\n",
        "  'text': 'to investigate the efficacy of @ weeks of daily low-dose oral prednisolone in improving pain , mobility , and systemic low-grade inflammation in the short term and whether the effect would be sustained at @ weeks in older adults with moderate to severe knee osteoarthritis ( oa ) .',\n",
        "  'total_lines': 11},\n",
        "  ...]\n",
        "```\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rt2_TsSjtVJb"
      },
      "source": [
        "def preprocess_text_with_line_numbers(filename):\n",
        "  \n",
        "  input_lines = get_lines(filename) # get all lines from filename\n",
        "  abstract_lines = \"\" # create an empty abstract\n",
        "  abstract_samples = [] # create an empty list of abstracts\n",
        "\n",
        "  for line in input_lines:\n",
        "    if line.startswith(\"###\"): # check to see if line is an ID line\n",
        "      abstract_id = line\n",
        "      abstract_lines = \"\"    # reset the abstract string\n",
        "    elif line.isspace():\n",
        "      abstract_line_split = abstract_lines.splitlines() # split the abstract into separate lines\n",
        "\n",
        "      for abstract_line_number, abstract_line in enumerate(abstract_line_split):\n",
        "        line_data = {}\n",
        "        target_text_split = abstract_line.split(\"\\t\")\n",
        "        line_data[\"target\"] = target_text_split[0]\n",
        "        line_data[\"text\"]   = target_text_split[1].lower()\n",
        "        line_data[\"line_number\"] = abstract_line_number\n",
        "        line_data[\"total_lines\"] = len(abstract_line_split) - 1 \n",
        "        abstract_samples.append(line_data)\n",
        "\n",
        "    else:\n",
        "      abstract_lines += line\n",
        "      \n",
        "\n",
        "  return abstract_samples"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XnEeLYZH8Enr"
      },
      "source": [
        ""
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "7mv3XiwhtApD",
        "outputId": "673390db-138f-44cc-c64d-025e203d51ed"
      },
      "source": [
        "# Get data from file and preprocess it\n",
        "%%time\n",
        "train_samples = preprocess_text_with_line_numbers(data_dir + \"train.txt\")\n",
        "val_samples = preprocess_text_with_line_numbers(data_dir + \"dev.txt\") # dev is another name for validation set\n",
        "test_samples = preprocess_text_with_line_numbers(data_dir + \"test.txt\")\n",
        "len(train_samples), len(val_samples), len(test_samples)"
      ],
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CPU times: user 520 ms, sys: 123 ms, total: 643 ms\n",
            "Wall time: 648 ms\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ypsizfh4pJxs"
      },
      "source": [
        " As we are experimenting Some Text Preprocessing are left (like url and special char removal) , we'll do it future and see acuuracy deference."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "e4Eg_9FI5_cI",
        "outputId": "4093cbf6-da80-44db-c9b3-e5acc5ff7c39"
      },
      "source": [
        "train_samples[:10]"
      ],
      "execution_count": 13,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[{'line_number': 0,\n",
              "  'target': 'OBJECTIVE',\n",
              "  'text': 'to investigate the efficacy of @ weeks of daily low-dose oral prednisolone in improving pain , mobility , and systemic low-grade inflammation in the short term and whether the effect would be sustained at @ weeks in older adults with moderate to severe knee osteoarthritis ( oa ) .',\n",
              "  'total_lines': 11},\n",
              " {'line_number': 1,\n",
              "  'target': 'METHODS',\n",
              "  'text': 'a total of @ patients with primary knee oa were randomized @:@ ; @ received @ mg/day of prednisolone and @ received placebo for @ weeks .',\n",
              "  'total_lines': 11},\n",
              " {'line_number': 2,\n",
              "  'target': 'METHODS',\n",
              "  'text': 'outcome measures included pain reduction and improvement in function scores and systemic inflammation markers .',\n",
              "  'total_lines': 11},\n",
              " {'line_number': 3,\n",
              "  'target': 'METHODS',\n",
              "  'text': 'pain was assessed using the visual analog pain scale ( @-@ mm ) .',\n",
              "  'total_lines': 11},\n",
              " {'line_number': 4,\n",
              "  'target': 'METHODS',\n",
              "  'text': 'secondary outcome measures included the western ontario and mcmaster universities osteoarthritis index scores , patient global assessment ( pga ) of the severity of knee oa , and @-min walk distance ( @mwd ) .',\n",
              "  'total_lines': 11},\n",
              " {'line_number': 5,\n",
              "  'target': 'METHODS',\n",
              "  'text': 'serum levels of interleukin @ ( il-@ ) , il-@ , tumor necrosis factor ( tnf ) - , and high-sensitivity c-reactive protein ( hscrp ) were measured .',\n",
              "  'total_lines': 11},\n",
              " {'line_number': 6,\n",
              "  'target': 'RESULTS',\n",
              "  'text': 'there was a clinically relevant reduction in the intervention group compared to the placebo group for knee pain , physical function , pga , and @mwd at @ weeks .',\n",
              "  'total_lines': 11},\n",
              " {'line_number': 7,\n",
              "  'target': 'RESULTS',\n",
              "  'text': 'the mean difference between treatment arms ( @ % ci ) was @ ( @-@ @ ) , p < @ ; @ ( @-@ @ ) , p < @ ; @ ( @-@ @ ) , p < @ ; and @ ( @-@ @ ) , p < @ , respectively .',\n",
              "  'total_lines': 11},\n",
              " {'line_number': 8,\n",
              "  'target': 'RESULTS',\n",
              "  'text': 'further , there was a clinically relevant reduction in the serum levels of il-@ , il-@ , tnf - , and hscrp at @ weeks in the intervention group when compared to the placebo group .',\n",
              "  'total_lines': 11},\n",
              " {'line_number': 9,\n",
              "  'target': 'RESULTS',\n",
              "  'text': 'these differences remained significant at @ weeks .',\n",
              "  'total_lines': 11}]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 13
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 466
        },
        "id": "qYV92INB6B87",
        "outputId": "bc03404a-1ed5-494f-a4a2-31991c106eef"
      },
      "source": [
        "\n",
        "import pandas as pd\n",
        "train_df = pd.DataFrame(train_samples)\n",
        "val_df = pd.DataFrame(val_samples)\n",
        "test_df = pd.DataFrame(test_samples)\n",
        "train_df.head(14)"
      ],
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>target</th>\n",
              "      <th>text</th>\n",
              "      <th>line_number</th>\n",
              "      <th>total_lines</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>OBJECTIVE</td>\n",
              "      <td>to investigate the efficacy of @ weeks of dail...</td>\n",
              "      <td>0</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>a total of @ patients with primary knee oa wer...</td>\n",
              "      <td>1</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>outcome measures included pain reduction and i...</td>\n",
              "      <td>2</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>pain was assessed using the visual analog pain...</td>\n",
              "      <td>3</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>secondary outcome measures included the wester...</td>\n",
              "      <td>4</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>serum levels of interleukin @ ( il-@ ) , il-@ ...</td>\n",
              "      <td>5</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>there was a clinically relevant reduction in t...</td>\n",
              "      <td>6</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>the mean difference between treatment arms ( @...</td>\n",
              "      <td>7</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>further , there was a clinically relevant redu...</td>\n",
              "      <td>8</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>these differences remained significant at @ we...</td>\n",
              "      <td>9</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>the outcome measures in rheumatology clinical ...</td>\n",
              "      <td>10</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>CONCLUSIONS</td>\n",
              "      <td>low-dose oral prednisolone had both a short-te...</td>\n",
              "      <td>11</td>\n",
              "      <td>11</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>BACKGROUND</td>\n",
              "      <td>emotional eating is associated with overeating...</td>\n",
              "      <td>0</td>\n",
              "      <td>10</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>BACKGROUND</td>\n",
              "      <td>yet , empirical evidence for individual ( trai...</td>\n",
              "      <td>1</td>\n",
              "      <td>10</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "         target  ... total_lines\n",
              "0     OBJECTIVE  ...          11\n",
              "1       METHODS  ...          11\n",
              "2       METHODS  ...          11\n",
              "3       METHODS  ...          11\n",
              "4       METHODS  ...          11\n",
              "5       METHODS  ...          11\n",
              "6       RESULTS  ...          11\n",
              "7       RESULTS  ...          11\n",
              "8       RESULTS  ...          11\n",
              "9       RESULTS  ...          11\n",
              "10      RESULTS  ...          11\n",
              "11  CONCLUSIONS  ...          11\n",
              "12   BACKGROUND  ...          10\n",
              "13   BACKGROUND  ...          10\n",
              "\n",
              "[14 rows x 4 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 14
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MoXKfoW0_KyY",
        "outputId": "19c03539-6feb-4cc6-9841-7ed6e50216d9"
      },
      "source": [
        "# Distribution of labels in training dat\n",
        "train_df[\"target\"].value_counts()"
      ],
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "METHODS        59353\n",
              "RESULTS        57953\n",
              "CONCLUSIONS    27168\n",
              "BACKGROUND     21727\n",
              "OBJECTIVE      13839\n",
              "Name: target, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 344
        },
        "id": "Np6r9ycK_Ku9",
        "outputId": "c0c98cd1-52b8-454a-9699-59250c68e4af"
      },
      "source": [
        "train_df[\"target\"].value_counts().plot(kind = 'bar')"
      ],
      "execution_count": 16,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7fefbe5c6d10>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 16
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 283
        },
        "id": "nejRjvdb_Ks8",
        "outputId": "f7022a00-6c2a-4f4f-dc0b-a8a9820dcd4e"
      },
      "source": [
        "train_df.total_lines.plot(kind= \"hist\")"
      ],
      "execution_count": 17,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7fefb31f38d0>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 17
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FI15grdTdXYO"
      },
      "source": [
        " Text Data Visualization Future work.."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uL06aM6K_Krd",
        "outputId": "5d2d8b1e-4ca4-4cb9-df9c-0d957939f829"
      },
      "source": [
        "# Convert abstract text lines into lists \n",
        "train_sentences = train_df[\"text\"].tolist()\n",
        "val_sentences = val_df[\"text\"].tolist()\n",
        "test_sentences = test_df[\"text\"].tolist()\n",
        "len(train_sentences), len(val_sentences), len(test_sentences)"
      ],
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(180040, 30212, 30135)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Dc2VHVqH_Kp-",
        "outputId": "c49f72aa-5b7f-4413-f85f-4d9199ed565f"
      },
      "source": [
        "train_sentences[:10]"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['to investigate the efficacy of @ weeks of daily low-dose oral prednisolone in improving pain , mobility , and systemic low-grade inflammation in the short term and whether the effect would be sustained at @ weeks in older adults with moderate to severe knee osteoarthritis ( oa ) .',\n",
              " 'a total of @ patients with primary knee oa were randomized @:@ ; @ received @ mg/day of prednisolone and @ received placebo for @ weeks .',\n",
              " 'outcome measures included pain reduction and improvement in function scores and systemic inflammation markers .',\n",
              " 'pain was assessed using the visual analog pain scale ( @-@ mm ) .',\n",
              " 'secondary outcome measures included the western ontario and mcmaster universities osteoarthritis index scores , patient global assessment ( pga ) of the severity of knee oa , and @-min walk distance ( @mwd ) .',\n",
              " 'serum levels of interleukin @ ( il-@ ) , il-@ , tumor necrosis factor ( tnf ) - , and high-sensitivity c-reactive protein ( hscrp ) were measured .',\n",
              " 'there was a clinically relevant reduction in the intervention group compared to the placebo group for knee pain , physical function , pga , and @mwd at @ weeks .',\n",
              " 'the mean difference between treatment arms ( @ % ci ) was @ ( @-@ @ ) , p < @ ; @ ( @-@ @ ) , p < @ ; @ ( @-@ @ ) , p < @ ; and @ ( @-@ @ ) , p < @ , respectively .',\n",
              " 'further , there was a clinically relevant reduction in the serum levels of il-@ , il-@ , tnf - , and hscrp at @ weeks in the intervention group when compared to the placebo group .',\n",
              " 'these differences remained significant at @ weeks .']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 19
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nKmhwjnXG_66"
      },
      "source": [
        "### One Hot Encoder\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UNpVR7EFAExP",
        "outputId": "26439e30-e6e5-478d-ee81-faf6a22092f3"
      },
      "source": [
        "# One hot encode labels\n",
        "from sklearn.preprocessing import OneHotEncoder\n",
        "one_hot_encoder = OneHotEncoder(sparse=False)\n",
        "train_labels_one_hot = one_hot_encoder.fit_transform(train_df[\"target\"].to_numpy().reshape(-1, 1))\n",
        "val_labels_one_hot = one_hot_encoder.transform(val_df[\"target\"].to_numpy().reshape(-1, 1))\n",
        "test_labels_one_hot = one_hot_encoder.transform(test_df[\"target\"].to_numpy().reshape(-1, 1))\n",
        "\n",
        "# Check what training labels look like\n",
        "train_labels_one_hot"
      ],
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0., 0., 0., 1., 0.],\n",
              "       [0., 0., 1., 0., 0.],\n",
              "       [0., 0., 1., 0., 0.],\n",
              "       ...,\n",
              "       [0., 0., 0., 0., 1.],\n",
              "       [0., 1., 0., 0., 0.],\n",
              "       [0., 1., 0., 0., 0.]])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 20
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PQ7aB1Q8G_3g"
      },
      "source": [
        "### Label encode labels"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9tHzHEgCG-1y",
        "outputId": "72e8d7a3-8cb5-4a44-e343-291d1c45e636"
      },
      "source": [
        "# Extract labels (\"target\" columns) and encode them into integers \n",
        "from sklearn.preprocessing import LabelEncoder\n",
        "label_encoder = LabelEncoder()\n",
        "train_labels_encoded = label_encoder.fit_transform(train_df[\"target\"].to_numpy())\n",
        "val_labels_encoded = label_encoder.transform(val_df[\"target\"].to_numpy())\n",
        "test_labels_encoded = label_encoder.transform(test_df[\"target\"].to_numpy())\n",
        "\n",
        "# Check what training labels look like\n",
        "train_labels_encoded"
      ],
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([3, 2, 2, ..., 4, 1, 1])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "s_Ipo14nHREA",
        "outputId": "7106a046-23f1-472f-fd71-c87c003bd312"
      },
      "source": [
        "# Get class names and number of classes from LabelEncoder instance \n",
        "num_classes = len(label_encoder.classes_)\n",
        "class_names = label_encoder.classes_\n",
        "num_classes, class_names"
      ],
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(5, array(['BACKGROUND', 'CONCLUSIONS', 'METHODS', 'OBJECTIVE', 'RESULTS'],\n",
              "       dtype=object))"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 22
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "StG7Xd9PHXKe"
      },
      "source": [
        "## Model 0: Getting a baseline"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gve89WqqBSAf"
      },
      "source": [
        "Our first model we'll be a TF-IDF Multinomial Naive Bayes as recommended by [Scikit-Learn's machine learning map](https://scikit-learn.org/stable/tutorial/machine_learning_map/index.html).\n",
        "\n",
        "we'll create a Scikit-Learn `Pipeline` which uses the [`TfidfVectorizer`](https://scikit-learn.org/stable/modules/generated/sklearn.feature_extraction.text.TfidfVectorizer.html) class to convert our abstract sentences to numbers using the TF-IDF (term frequency-inverse document frequecy) algorithm and then learns to classify our sentences using the [`MultinomialNB`](https://scikit-learn.org/stable/modules/generated/sklearn.naive_bayes.MultinomialNB.html) aglorithm."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KWbg4QWUHRi2"
      },
      "source": [
        "from sklearn.feature_extraction.text import TfidfVectorizer\n",
        "from sklearn.naive_bayes import MultinomialNB\n",
        "from sklearn.pipeline import Pipeline\n",
        "\n",
        "# Create a pipeline\n",
        "model_0 = Pipeline([\n",
        "  (\"tf-idf\", TfidfVectorizer()),\n",
        "  (\"clf\", MultinomialNB())\n",
        "])\n",
        "\n",
        "# Fit the pipeline to the training data\n",
        "model_0.fit(X=train_sentences, \n",
        "            y=train_labels_encoded);"
      ],
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rWyUH1WuHcVq",
        "outputId": "5f609cc4-739c-40e9-a1c0-278e325fe5f1"
      },
      "source": [
        "\n",
        "# Evaluate baseline on validation dataset\n",
        "model_0.score(X=val_sentences,\n",
        "              y=val_labels_encoded)"
      ],
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0.7218323844829869"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 24
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rG2OkaQnHfuy",
        "outputId": "a7947f86-8279-43cd-aad6-c84b28ff5ee5"
      },
      "source": [
        "# Make predictions\n",
        "baseline_preds = model_0.predict(val_sentences)\n",
        "baseline_preds"
      ],
      "execution_count": 25,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([4, 1, 3, ..., 4, 4, 1])"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 25
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "udY17G_PHi51",
        "outputId": "bde2d619-cfbd-4a4e-9851-80d3f9d4381f"
      },
      "source": [
        "# Download helper functions script\n",
        "!wget https://raw.githubusercontent.com/mrdbourke/tensorflow-deep-learning/main/extras/helper_functions.py"
      ],
      "execution_count": 26,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "--2021-06-16 16:53:10--  https://raw.githubusercontent.com/mrdbourke/tensorflow-deep-learning/main/extras/helper_functions.py\n",
            "Resolving raw.githubusercontent.com (raw.githubusercontent.com)... 185.199.108.133, 185.199.109.133, 185.199.110.133, ...\n",
            "Connecting to raw.githubusercontent.com (raw.githubusercontent.com)|185.199.108.133|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 10246 (10K) [text/plain]\n",
            "Saving to: ‘helper_functions.py’\n",
            "\n",
            "\rhelper_functions.py   0%[                    ]       0  --.-KB/s               \rhelper_functions.py 100%[===================>]  10.01K  --.-KB/s    in 0s      \n",
            "\n",
            "2021-06-16 16:53:10 (79.1 MB/s) - ‘helper_functions.py’ saved [10246/10246]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Gt6fIiiCHwOq"
      },
      "source": [
        "# Import calculate_results helper function\n",
        "from helper_functions import calculate_results"
      ],
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VFzLlPCsruC6"
      },
      "source": [
        "### Model_0 Results"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XLiIhUTCHosp",
        "outputId": "6f88a1fc-8f80-4440-de61-c9eeef852667"
      },
      "source": [
        "\n",
        "# Calculate baseline results\n",
        "baseline_results = calculate_results(y_true=val_labels_encoded,\n",
        "                                     y_pred=baseline_preds)\n",
        "baseline_results"
      ],
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 72.1832384482987,\n",
              " 'f1': 0.6989250353450294,\n",
              " 'precision': 0.7186466952323352,\n",
              " 'recall': 0.7218323844829869}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 28
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ol2NuEsWgcdQ"
      },
      "source": [
        "# Prepare Data For Deep Neural Network Models\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ugxK-B5MhCas"
      },
      "source": [
        "When our model goes through our sentences, it works best when they're all the same length (this is important for creating batches of the same size tensors)\n",
        "\n",
        "* Finding the average sentence length in the Dataset."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "8FNgfk7Jh_td"
      },
      "source": [
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "from tensorflow.keras import layers"
      ],
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Lxgn6J5eHomX",
        "outputId": "c4fc88d7-794f-457b-f56d-73bf44d3dafb"
      },
      "source": [
        "sen_len = [len(sentences.split()) for sentences in train_sentences]\n",
        "avg_sen_len = np.mean(sen_len)\n",
        "avg_sen_len"
      ],
      "execution_count": 30,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "26.338269273494777"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 30
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 469
        },
        "id": "j_265IaEhwKi",
        "outputId": "0678cfb6-237c-4a29-bdd3-8a2aa5463739"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.hist(sen_len,bins=21) # Checking the Sequence Length Distribution and getting most occurance  sequence length"
      ],
      "execution_count": 31,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([4.2075e+04, 7.9624e+04, 3.8291e+04, 1.2725e+04, 4.3900e+03,\n",
              "        1.6450e+03, 7.2600e+02, 2.8900e+02, 1.3600e+02, 5.5000e+01,\n",
              "        2.9000e+01, 1.5000e+01, 1.1000e+01, 9.0000e+00, 8.0000e+00,\n",
              "        5.0000e+00, 2.0000e+00, 3.0000e+00, 0.0000e+00, 1.0000e+00,\n",
              "        1.0000e+00]),\n",
              " array([  1.        ,  15.04761905,  29.0952381 ,  43.14285714,\n",
              "         57.19047619,  71.23809524,  85.28571429,  99.33333333,\n",
              "        113.38095238, 127.42857143, 141.47619048, 155.52380952,\n",
              "        169.57142857, 183.61904762, 197.66666667, 211.71428571,\n",
              "        225.76190476, 239.80952381, 253.85714286, 267.9047619 ,\n",
              "        281.95238095, 296.        ]),\n",
              " <a list of 21 Patch objects>)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 31
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Vf_zXb9HlR18"
      },
      "source": [
        "Looks like the vast majority of sentences are between 0 and 50 tokens in length.\n",
        "\n",
        "We can use NumPy's [`percentile`](https://numpy.org/doc/stable/reference/generated/numpy.percentile.html) to find the value which covers 95% of the sentence lengthsHow long of a sentesnces cover majority of the data ? (95%)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "V7e7ihftlbDH",
        "outputId": "3323f111-ed86-419b-d0c0-9f1ba5bc7f8d"
      },
      "source": [
        "np.percentile(sen_len,95)"
      ],
      "execution_count": 32,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "55.0"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 32
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "pqKhG2IQmVnp",
        "outputId": "c6c44c42-2ceb-41e7-ea51-97a4d81dd2b4"
      },
      "source": [
        "max(sen_len)   # max length sentence in training set"
      ],
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "296"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 33
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DYnp55rLmNk0"
      },
      "source": [
        "Creaating a text vectorizer layer"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BaKj7OD6EROB"
      },
      "source": [
        "### Create text vectorize\n",
        "\n",
        "Section 3.2 of the [PubMed 200k RCT paper](https://arxiv.org/pdf/1710.06071.pdf) states the vocabulary size of the PubMed 20k dataset as 68,000. So we'll use that as our `max_tokens` parameter.\n",
        "                   "
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SwKpU8WrnEvl"
      },
      "source": [
        "max_tokens = 68000"
      ],
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JlmQV04QleQW"
      },
      "source": [
        "from tensorflow.keras.layers.experimental.preprocessing import TextVectorization\n",
        "text_vectorizer = TextVectorization(max_tokens=max_tokens,standardize='lower_and_strip_punctuation',\n",
        "                                    output_sequence_length=55)\n"
      ],
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "i2JjimMfmwzs"
      },
      "source": [
        "# Adapt text vectorizer to training sentences\n",
        "text_vectorizer.adapt(train_sentences)"
      ],
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ke1VNN9onYhT",
        "outputId": "791de1c5-e1e0-48ef-b032-aeab1971e976"
      },
      "source": [
        "# viewing vectorize training sentences\n",
        "import random\n",
        "target_sentence  = random.choice(train_sentences)\n",
        "print(f\"Text:\\n{target_sentence}\")\n",
        "print(f\"\\nLength of text: {len(target_sentence.split())}\")\n",
        "print(f\"\\nVectorized text:\\n{text_vectorizer([target_sentence])}\")"
      ],
      "execution_count": 37,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Text:\n",
            "after @ months of stable-dose treatment , patients receiving milnacipran @ and/or @ mg/d had significant improvement in mfi total and subscale scores ( p < @ vs placebo ) .\n",
            "\n",
            "Length of text: 31\n",
            "\n",
            "Vectorized text:\n",
            "[[   21    41     4 34172    19    12   245 12540   727  1306    55    37\n",
            "    194     5 18393    76     3  2072   119    14    44    48     0     0\n",
            "      0     0     0     0     0     0     0     0     0     0     0     0\n",
            "      0     0     0     0     0     0     0     0     0     0     0     0\n",
            "      0     0     0     0     0     0     0]]\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wLQMOocDtOiF",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "e63dc024-5545-40e4-e4aa-48d1b6b7f5f5"
      },
      "source": [
        "# Getting the vocabulary and showing most frequent and least frequest words in the vocabulary\n",
        "rct_20k_text_vocab = text_vectorizer.get_vocabulary()\n",
        "most_common = rct_20k_text_vocab[:5]\n",
        "least_common = rct_20k_text_vocab[-5:]\n",
        "print(f\"Number of words in vocabulary: {len(rct_20k_text_vocab)}\"), \n",
        "print(f\"Most common words in the vocabulary: {most_common}\")\n",
        "print(f\"Least common words in the vocabulary: {least_common}\")"
      ],
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Number of words in vocabulary: 64841\n",
            "Most common words in the vocabulary: ['', '[UNK]', 'the', 'and', 'of']\n",
            "Least common words in the vocabulary: ['aainduced', 'aaigroup', 'aachener', 'aachen', 'aaacp']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-SckJNsmF6H4",
        "outputId": "3d7c2837-7ef1-40a0-819f-458f61b7b3f7"
      },
      "source": [
        "# Get the config of our text vectorizer\n",
        "text_vectorizer.get_config()"
      ],
      "execution_count": 39,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'dtype': 'string',\n",
              " 'max_tokens': 68000,\n",
              " 'name': 'text_vectorization',\n",
              " 'ngrams': None,\n",
              " 'output_mode': 'int',\n",
              " 'output_sequence_length': 55,\n",
              " 'pad_to_max_tokens': False,\n",
              " 'split': 'whitespace',\n",
              " 'standardize': 'lower_and_strip_punctuation',\n",
              " 'trainable': True,\n",
              " 'vocabulary_size': 64841}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 39
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DYb-FPIdG7G0"
      },
      "source": [
        "### Create custom text embedding\n",
        "\n",
        "To create a richer numerical representation of our text, we can use an **embedding**.\n",
        "\n",
        "The `input_dim` parameter defines the size of our vocabulary. And the `output_dim` parameter defines the dimension of the embedding output.\n",
        "\n",
        "Once created, our embedding layer will take the integer outputs of our `text_vectorization` layer as inputs and convert them to feature vectors of size `output_dim`."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "u9gtJ6I0F70v",
        "outputId": "1a1daf59-e7dd-432f-d3a9-122fd242fb4f"
      },
      "source": [
        "token_embed = layers.Embedding(input_dim=len(rct_20k_text_vocab), \n",
        "                               output_dim= 128,\n",
        "                               mask_zero=True,\n",
        "                               input_length=55)\n",
        "\n",
        "\n",
        "\n",
        "print(f\"Sentence before Vectorization : \\n{target_sentence}\\n\")\n",
        "vec_sentence = text_vectorizer([target_sentence])\n",
        "print(f\"Sentence After vectorization :\\n {vec_sentence}\\n\")\n",
        "embed_sentence = token_embed(vec_sentence)\n",
        "print(f\"Embedding Sentence :\\n{embed_sentence}\\n\")"
      ],
      "execution_count": 40,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Sentence before Vectorization : \n",
            "after @ months of stable-dose treatment , patients receiving milnacipran @ and/or @ mg/d had significant improvement in mfi total and subscale scores ( p < @ vs placebo ) .\n",
            "\n",
            "Sentence After vectorization :\n",
            " [[   21    41     4 34172    19    12   245 12540   727  1306    55    37\n",
            "    194     5 18393    76     3  2072   119    14    44    48     0     0\n",
            "      0     0     0     0     0     0     0     0     0     0     0     0\n",
            "      0     0     0     0     0     0     0     0     0     0     0     0\n",
            "      0     0     0     0     0     0     0]]\n",
            "\n",
            "Embedding Sentence :\n",
            "[[[ 0.01845442  0.00490598  0.02234909 ... -0.0251475   0.02369246\n",
            "   -0.03116907]\n",
            "  [ 0.03540225 -0.01559786 -0.00104294 ... -0.0325914  -0.01999564\n",
            "   -0.03067011]\n",
            "  [ 0.01375339 -0.0441448  -0.03379129 ... -0.04563415 -0.04586704\n",
            "    0.03887606]\n",
            "  ...\n",
            "  [-0.02985824 -0.01073467 -0.04434704 ...  0.0063103  -0.04246072\n",
            "   -0.02347449]\n",
            "  [-0.02985824 -0.01073467 -0.04434704 ...  0.0063103  -0.04246072\n",
            "   -0.02347449]\n",
            "  [-0.02985824 -0.01073467 -0.04434704 ...  0.0063103  -0.04246072\n",
            "   -0.02347449]]]\n",
            "\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "oZE43CTNiZQ9"
      },
      "source": [
        "# Turn our data into TensorFlow Dataset\n",
        "train_dataset = tf.data.Dataset.from_tensor_slices((train_sentences, train_labels_one_hot))\n",
        "valid_dataset = tf.data.Dataset.from_tensor_slices((val_sentences, val_labels_one_hot))\n",
        "test_dataset = tf.data.Dataset.from_tensor_slices((test_sentences, test_labels_one_hot))"
      ],
      "execution_count": 41,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5VpIopXAl4EV",
        "outputId": "a2157791-5f31-4d23-88b8-e097f6104cbb"
      },
      "source": [
        "len(train_dataset) , train_dataset"
      ],
      "execution_count": 42,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(180040,\n",
              " <TensorSliceDataset shapes: ((), (5,)), types: (tf.string, tf.float64)>)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 42
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eKttoqLWiZNi",
        "outputId": "426d6ef7-3054-423a-a51b-50b2b666e88e"
      },
      "source": [
        "# Take the TensorSliceDataset's and turn them into prefetched batches\n",
        "train_dataset = train_dataset.batch(32).prefetch(tf.data.AUTOTUNE)\n",
        "valid_dataset = valid_dataset.batch(32).prefetch(tf.data.AUTOTUNE)\n",
        "test_dataset = test_dataset.batch(32).prefetch(tf.data.AUTOTUNE)\n",
        "\n",
        "train_dataset"
      ],
      "execution_count": 43,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<PrefetchDataset shapes: ((None,), (None, 5)), types: (tf.string, tf.float64)>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 43
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AByumHtMnfP3"
      },
      "source": [
        "## Model 1: Conv1D with token embedding\n",
        "\n",
        "All of our deep models will follow a similar structure:\n",
        "\n",
        "```\n",
        "Input (text) -> Tokenize -> Embedding -> Layers -> Output (label probability)\n",
        "```"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2t7aSVtqiZLS",
        "outputId": "2b5cc6a5-61d0-446b-da8f-a99009ba3e5c"
      },
      "source": [
        "inputs = layers.Input(shape = (1,),dtype = tf.string)\n",
        "text_vector = text_vectorizer(inputs)\n",
        "embed = token_embed(text_vector)\n",
        "x = layers.Conv1D(filters = 64, kernel_size= 5, padding=\"same\",activation=\"relu\",kernel_regularizer=tf.keras.regularizers.L2(0.01))(embed)\n",
        "x = layers.GlobalMaxPool1D()(x)\n",
        "x = layers.Dropout(0.2)(x)\n",
        "outputs = layers.Dense(num_classes,activation=\"softmax\")(x)\n",
        "\n",
        "model = tf.keras.Model(inputs,outputs)\n",
        "model.summary()"
      ],
      "execution_count": 44,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"model\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_1 (InputLayer)         [(None, 1)]               0         \n",
            "_________________________________________________________________\n",
            "text_vectorization (TextVect (None, 55)                0         \n",
            "_________________________________________________________________\n",
            "embedding (Embedding)        (None, 55, 128)           8299648   \n",
            "_________________________________________________________________\n",
            "conv1d (Conv1D)              (None, 55, 64)            41024     \n",
            "_________________________________________________________________\n",
            "global_max_pooling1d (Global (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dropout (Dropout)            (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dense (Dense)                (None, 5)                 325       \n",
            "=================================================================\n",
            "Total params: 8,340,997\n",
            "Trainable params: 8,340,997\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2ueGzhs-rLnx",
        "outputId": "04b47046-6b7c-476b-c7d6-d03362c85dbf"
      },
      "source": [
        "len(train_dataset)"
      ],
      "execution_count": 45,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "5627"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 45
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1ZkpY3WTiZJ4"
      },
      "source": [
        "model.compile(optimizer='Adam', loss=\"categorical_crossentropy\", metrics=[\"accuracy\"])"
      ],
      "execution_count": 46,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iSP_fPhdiZII",
        "outputId": "cb3b4729-e075-4ee3-816f-61737ac41b93"
      },
      "source": [
        "model_1_history =  model.fit(train_dataset,\n",
        "                             steps_per_epoch=int(0.1 * len(train_dataset)),\n",
        "                             epochs = 10,\n",
        "                             validation_data = valid_dataset,\n",
        "                             validation_steps=int(0.1 * len(valid_dataset)),)"
      ],
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "562/562 [==============================] - 58s 101ms/step - loss: 1.1068 - accuracy: 0.6201 - val_loss: 0.7933 - val_accuracy: 0.7487\n",
            "Epoch 2/10\n",
            "562/562 [==============================] - 56s 100ms/step - loss: 0.7738 - accuracy: 0.7517 - val_loss: 0.7109 - val_accuracy: 0.7736\n",
            "Epoch 3/10\n",
            "562/562 [==============================] - 56s 100ms/step - loss: 0.7241 - accuracy: 0.7625 - val_loss: 0.6671 - val_accuracy: 0.7872\n",
            "Epoch 4/10\n",
            "562/562 [==============================] - 56s 100ms/step - loss: 0.6906 - accuracy: 0.7751 - val_loss: 0.6546 - val_accuracy: 0.7846\n",
            "Epoch 5/10\n",
            "562/562 [==============================] - 56s 100ms/step - loss: 0.6888 - accuracy: 0.7782 - val_loss: 0.6390 - val_accuracy: 0.7902\n",
            "Epoch 6/10\n",
            "562/562 [==============================] - 56s 100ms/step - loss: 0.6796 - accuracy: 0.7763 - val_loss: 0.6246 - val_accuracy: 0.7949\n",
            "Epoch 7/10\n",
            "562/562 [==============================] - 56s 100ms/step - loss: 0.6509 - accuracy: 0.7854 - val_loss: 0.6183 - val_accuracy: 0.7999\n",
            "Epoch 8/10\n",
            "562/562 [==============================] - 56s 100ms/step - loss: 0.6389 - accuracy: 0.7971 - val_loss: 0.6009 - val_accuracy: 0.8122\n",
            "Epoch 9/10\n",
            "562/562 [==============================] - 56s 100ms/step - loss: 0.6379 - accuracy: 0.7930 - val_loss: 0.6430 - val_accuracy: 0.7902\n",
            "Epoch 10/10\n",
            "562/562 [==============================] - 56s 99ms/step - loss: 0.6402 - accuracy: 0.7942 - val_loss: 0.6062 - val_accuracy: 0.8039\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UluIRIHaiZFn",
        "outputId": "3d87f9b6-a1ba-4ebe-b522-519d2fc3ea1a"
      },
      "source": [
        "# Evaluate on whole validation dataset (we only validated on 10% of batches during training)\n",
        "model.evaluate(valid_dataset)"
      ],
      "execution_count": 48,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "945/945 [==============================] - 5s 5ms/step - loss: 0.5962 - accuracy: 0.8083\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.5962496995925903, 0.8082880973815918]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 48
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Q4EFNmNRiZDl",
        "outputId": "bfd8da58-28b0-4df3-90a1-7ca6339ce58a"
      },
      "source": [
        "# Make predictions (our model outputs prediction probabilities for each class)\n",
        "model_1_pred_probs = model.predict(valid_dataset)\n",
        "model_1_pred_probs"
      ],
      "execution_count": 49,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[6.4009082e-01, 1.5787475e-01, 1.1690257e-02, 1.7012680e-01,\n",
              "        2.0217434e-02],\n",
              "       [2.1156156e-01, 7.0904827e-01, 1.1327777e-03, 7.0969328e-02,\n",
              "        7.2881221e-03],\n",
              "       [1.7376225e-01, 8.4552346e-03, 1.4700086e-03, 8.1595337e-01,\n",
              "        3.5914412e-04],\n",
              "       ...,\n",
              "       [2.6505839e-04, 9.3591970e-04, 1.3166884e-02, 1.3237984e-04,\n",
              "        9.8549968e-01],\n",
              "       [2.4481107e-02, 5.5618656e-01, 4.2747390e-02, 1.6767828e-02,\n",
              "        3.5981715e-01],\n",
              "       [9.1729127e-02, 8.6072636e-01, 1.3366650e-02, 1.7221529e-02,\n",
              "        1.6956387e-02]], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "t8M_4eSViZB1",
        "outputId": "3f3f0244-a213-49b6-90cc-53426046f576"
      },
      "source": [
        "# Convert pred probs to classes\n",
        "model_1_preds = tf.argmax(model_1_pred_probs, axis=1)\n",
        "model_1_preds"
      ],
      "execution_count": 50,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(30212,), dtype=int64, numpy=array([0, 1, 3, ..., 4, 1, 1])>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 50
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kyRilMbhrmb7"
      },
      "source": [
        "### Model_1 Results"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DeJI_grvwWio",
        "outputId": "5b7e429f-4caa-48a7-d58a-93b7be8ddce9"
      },
      "source": [
        "# Calculate model_1 results\n",
        "model_1_results = calculate_results(y_true=val_labels_encoded,\n",
        "                                    y_pred=model_1_preds)\n",
        "model_1_results"
      ],
      "execution_count": 51,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 80.8288097444724,\n",
              " 'f1': 0.8054731032957875,\n",
              " 'precision': 0.8047795569207609,\n",
              " 'recall': 0.808288097444724}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 51
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gps3htkSuLP7"
      },
      "source": [
        "## Model 2: Feature extraction with pretrained token embedding (USE)\n",
        "\n",
        "Here We use Universal Sentence Encoder here from TF-HUB.\n",
        "\n",
        "Since we're moving towards replicating the model architecture in [*Neural Networks for Joint Sentence Classification\n",
        "in Medical Paper Abstracts*](https://arxiv.org/pdf/1612.05251.pdf), it mentions they used a [pretrained GloVe embedding](https://nlp.stanford.edu/projects/glove/) as a way to initialise their token embeddings.\n",
        "\n",
        "\n",
        "The model structure will look like:\n",
        "\n",
        "```\n",
        "Inputs (string) -> Pretrained embeddings from TensorFlow Hub (Universal Sentence Encoder) -> Layers -> Output (prediction probabilities)\n",
        "```"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2_Bw0ZL3iY_9"
      },
      "source": [
        "# Download pretrained TensorFlow Hub USE\n",
        "import tensorflow_hub as hub\n",
        "tf_hub_embedding_layer = hub.KerasLayer(\"https://tfhub.dev/google/universal-sentence-encoder/4\",\n",
        "                                        trainable=False,\n",
        "                                        name=\"universal_sentence_encoder\")"
      ],
      "execution_count": 52,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2hOWxaxKwqhz"
      },
      "source": [
        "Beautiful, now our pretrained USE is downloaded and instantiated as a `hub.KerasLayer` instance, let's test it out on a random sentence"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TSFkc6rUwiOk",
        "outputId": "4e21ed8d-f7a4-45ba-b0a2-e6dba9ba63c2"
      },
      "source": [
        "# Test out the embedding on a random sentence\n",
        "random_training_sentence = random.choice(train_sentences)\n",
        "print(f\"Random training sentence:\\n{random_training_sentence}\\n\")\n",
        "use_embedded_sentence = tf_hub_embedding_layer([random_training_sentence])\n",
        "print(f\"Sentence after embedding:\\n{use_embedded_sentence[0][:30]} (truncated output)...\\n\")\n",
        "print(f\"Length of sentence embedding:\\n{len(use_embedded_sentence[0])}\")"
      ],
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Random training sentence:\n",
            "the subjects in the test group practiced cthe , while those in the control group did `` the @th radio calisthenics '' , an official recommended calisthenics for promoting healthcare in china , @ times a week , and @ weeks practicing overall .\n",
            "\n",
            "Sentence after embedding:\n",
            "[-0.04633829 -0.03948136  0.02896223 -0.07942989 -0.01642509  0.03321299\n",
            "  0.04372008 -0.01696645 -0.03029537  0.0055867   0.0817182   0.0511634\n",
            "  0.02764668 -0.04249508  0.02601314  0.02671395 -0.08149037  0.02041833\n",
            " -0.05118315 -0.02549118 -0.02749129 -0.02932397 -0.07956979 -0.02773482\n",
            " -0.03448042  0.03734251 -0.066794    0.01424676 -0.06487625 -0.05911339] (truncated output)...\n",
            "\n",
            "Length of sentence embedding:\n",
            "512\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "H0mzMD6pwt8x"
      },
      "source": [
        "# Define feature extractor model using TF Hub layer\n",
        "inputs = layers.Input(shape=[], dtype=tf.string)\n",
        "pretrained_embedding = tf_hub_embedding_layer(inputs) # tokenize text and create embedding\n",
        "x = layers.Dense(128, activation=\"relu\")(pretrained_embedding) # add a fully connected layer on top of the embedding\n",
        "# x = layers.Dropout(0.2)(x)\n",
        "outputs = layers.Dense(5, activation=\"softmax\",kernel_regularizer=None)(x) # create the output layer\n",
        "model_2 = tf.keras.Model(inputs=inputs,\n",
        "                        outputs=outputs)\n",
        "\n",
        "# Compile the model\n",
        "model_2.compile(loss=\"categorical_crossentropy\",\n",
        "                optimizer=tf.keras.optimizers.Adam(),\n",
        "                metrics=[\"accuracy\"])"
      ],
      "execution_count": 54,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "aNxvlxkAxD-J",
        "outputId": "da137e90-a3e0-4a8b-83b4-5253d74338ce"
      },
      "source": [
        "# Get a summary of the model\n",
        "model_2.summary()"
      ],
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"model_1\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_2 (InputLayer)         [(None,)]                 0         \n",
            "_________________________________________________________________\n",
            "universal_sentence_encoder ( (None, 512)               256797824 \n",
            "_________________________________________________________________\n",
            "dense_1 (Dense)              (None, 128)               65664     \n",
            "_________________________________________________________________\n",
            "dense_2 (Dense)              (None, 5)                 645       \n",
            "=================================================================\n",
            "Total params: 256,864,133\n",
            "Trainable params: 66,309\n",
            "Non-trainable params: 256,797,824\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5WTuR71UxGVG",
        "outputId": "8e6cf99b-8e15-42f4-b721-31c5fa886176"
      },
      "source": [
        "model_2_history =  model_2.fit(train_dataset,\n",
        "                             steps_per_epoch=int(0.1 * len(train_dataset)),\n",
        "                             epochs = 10,\n",
        "                             validation_data = valid_dataset,\n",
        "                             validation_steps=int(0.1 * len(valid_dataset)))"
      ],
      "execution_count": 56,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "562/562 [==============================] - 9s 12ms/step - loss: 0.9225 - accuracy: 0.6502 - val_loss: 0.7968 - val_accuracy: 0.6908\n",
            "Epoch 2/10\n",
            "562/562 [==============================] - 6s 11ms/step - loss: 0.7692 - accuracy: 0.7017 - val_loss: 0.7549 - val_accuracy: 0.7071\n",
            "Epoch 3/10\n",
            "562/562 [==============================] - 6s 11ms/step - loss: 0.7528 - accuracy: 0.7127 - val_loss: 0.7393 - val_accuracy: 0.7128\n",
            "Epoch 4/10\n",
            "562/562 [==============================] - 7s 12ms/step - loss: 0.7198 - accuracy: 0.7244 - val_loss: 0.7123 - val_accuracy: 0.7287\n",
            "Epoch 5/10\n",
            "562/562 [==============================] - 6s 11ms/step - loss: 0.7267 - accuracy: 0.7219 - val_loss: 0.6904 - val_accuracy: 0.7317\n",
            "Epoch 6/10\n",
            "562/562 [==============================] - 6s 12ms/step - loss: 0.7185 - accuracy: 0.7254 - val_loss: 0.6832 - val_accuracy: 0.7350\n",
            "Epoch 7/10\n",
            "562/562 [==============================] - 6s 11ms/step - loss: 0.6862 - accuracy: 0.7390 - val_loss: 0.6673 - val_accuracy: 0.7463\n",
            "Epoch 8/10\n",
            "562/562 [==============================] - 6s 11ms/step - loss: 0.6764 - accuracy: 0.7435 - val_loss: 0.6558 - val_accuracy: 0.7483\n",
            "Epoch 9/10\n",
            "562/562 [==============================] - 6s 11ms/step - loss: 0.6742 - accuracy: 0.7430 - val_loss: 0.6571 - val_accuracy: 0.7480\n",
            "Epoch 10/10\n",
            "562/562 [==============================] - 6s 11ms/step - loss: 0.6690 - accuracy: 0.7479 - val_loss: 0.6518 - val_accuracy: 0.7563\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "x9an0EOqxvYl",
        "outputId": "250212fc-8f6f-4b2e-efe7-3e3d86a25515"
      },
      "source": [
        "# Evaluate on whole validation dataset\n",
        "model_2.evaluate(valid_dataset)"
      ],
      "execution_count": 57,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "945/945 [==============================] - 9s 9ms/step - loss: 0.6559 - accuracy: 0.7511\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.6559162139892578, 0.7510591745376587]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 57
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "48YMmqKbzl6r",
        "outputId": "6fc3f406-c05d-470b-f5cc-503aca13e85d"
      },
      "source": [
        "# Make predictions with feature extraction model\n",
        "model_2_pred_probs = model_2.predict(valid_dataset)\n",
        "model_2_pred_probs"
      ],
      "execution_count": 58,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[4.20623630e-01, 4.38490957e-01, 6.25136832e-04, 1.32154837e-01,\n",
              "        8.10540374e-03],\n",
              "       [3.23061615e-01, 5.82224965e-01, 1.88950659e-03, 9.16711688e-02,\n",
              "        1.15277513e-03],\n",
              "       [4.91145968e-01, 4.19516787e-02, 1.88930240e-02, 4.06143695e-01,\n",
              "        4.18655388e-02],\n",
              "       ...,\n",
              "       [7.41388998e-04, 7.53519998e-04, 2.20944248e-02, 2.34126986e-04,\n",
              "        9.76176500e-01],\n",
              "       [3.42314062e-03, 6.01526648e-02, 1.98868886e-01, 8.81478831e-04,\n",
              "        7.36673832e-01],\n",
              "       [4.39007245e-02, 8.41824710e-01, 1.04148194e-01, 9.37066041e-04,\n",
              "        9.18926578e-03]], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 58
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9HPWWEEwz6Ku",
        "outputId": "95522658-b10e-45ee-cdd1-bdc15ea87c60"
      },
      "source": [
        "# Convert the predictions with feature extraction model to classes\n",
        "model_2_preds = tf.argmax(model_2_pred_probs, axis=1)\n",
        "model_2_preds"
      ],
      "execution_count": 59,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(30212,), dtype=int64, numpy=array([1, 1, 0, ..., 4, 4, 1])>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 59
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "59ezmA13r0Pq"
      },
      "source": [
        "### Model_2 Results"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "LBrAxS52zpfn",
        "outputId": "4ced0392-3cda-49c1-85ae-e55e21df6fe9"
      },
      "source": [
        "# Calculate results from TF Hub pretrained embeddings results on validation set\n",
        "model_2_results = calculate_results(y_true=val_labels_encoded,\n",
        "                                    y_pred=model_2_preds)\n",
        "model_2_results"
      ],
      "execution_count": 60,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 75.10591817820733,\n",
              " 'f1': 0.746383502412015,\n",
              " 'precision': 0.7457372543188239,\n",
              " 'recall': 0.7510591817820733}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 60
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "04R1gQu35pVb"
      },
      "source": [
        "## Model 3: Conv1D with character embedding"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "IEIASb3e5uZl"
      },
      "source": [
        "### Creating a character-level tokenizer\n",
        "\n",
        "The [*Neural Networks for Joint Sentence Classification\n",
        "in Medical Paper Abstracts*](https://arxiv.org/pdf/1612.05251.pdf) paper mentions their model uses a hybrid of token and character embeddings.\n",
        "\n",
        "The difference between a character and token embedding is that the **character embedding** is created using sequences split into characters (e.g. `hello` -> [`h`, `e`, `l`, `l`, `o`]) where as a **token embedding** is created on sequences split into tokens.\n",
        "\n",
        "*Token level embeddings split sequences into tokens (words) and embeddings each of them, character embeddings split sequences into characters and creates a feature vector for each.*\n",
        "\n",
        "Before we can vectorize our sequences on a character-level we'll need to split them into characters. Let's write a function to do so"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 89
        },
        "id": "QtC6PRpezt98",
        "outputId": "f03cde21-9a40-488f-b744-4212d2f2abfd"
      },
      "source": [
        "\" \".join(list(train_sentences[0]))"
      ],
      "execution_count": 61,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'t o   i n v e s t i g a t e   t h e   e f f i c a c y   o f   @   w e e k s   o f   d a i l y   l o w - d o s e   o r a l   p r e d n i s o l o n e   i n   i m p r o v i n g   p a i n   ,   m o b i l i t y   ,   a n d   s y s t e m i c   l o w - g r a d e   i n f l a m m a t i o n   i n   t h e   s h o r t   t e r m   a n d   w h e t h e r   t h e   e f f e c t   w o u l d   b e   s u s t a i n e d   a t   @   w e e k s   i n   o l d e r   a d u l t s   w i t h   m o d e r a t e   t o   s e v e r e   k n e e   o s t e o a r t h r i t i s   (   o a   )   .'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 61
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 71
        },
        "id": "we8ivMCg6HGF",
        "outputId": "c0f705da-0b04-4c48-e111-2a9757b9c36c"
      },
      "source": [
        "def split_chars(text):\n",
        "  return \" \".join(list(text))\n",
        "\n",
        "split_chars(random_training_sentence)"
      ],
      "execution_count": 62,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "\"t h e   s u b j e c t s   i n   t h e   t e s t   g r o u p   p r a c t i c e d   c t h e   ,   w h i l e   t h o s e   i n   t h e   c o n t r o l   g r o u p   d i d   ` `   t h e   @ t h   r a d i o   c a l i s t h e n i c s   ' '   ,   a n   o f f i c i a l   r e c o m m e n d e d   c a l i s t h e n i c s   f o r   p r o m o t i n g   h e a l t h c a r e   i n   c h i n a   ,   @   t i m e s   a   w e e k   ,   a n d   @   w e e k s   p r a c t i c i n g   o v e r a l l   .\""
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 62
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ECeI8HeD60oB",
        "outputId": "be614f8e-4adc-4d20-e77e-62df5016afec"
      },
      "source": [
        "# Split sequence-level data splits into character-level data splits\n",
        "train_chars = [split_chars(sentence) for sentence in train_sentences]\n",
        "val_chars = [split_chars(sentence) for sentence in val_sentences]\n",
        "test_chars = [split_chars(sentence) for sentence in test_sentences]\n",
        "print(train_chars[0])"
      ],
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "t o   i n v e s t i g a t e   t h e   e f f i c a c y   o f   @   w e e k s   o f   d a i l y   l o w - d o s e   o r a l   p r e d n i s o l o n e   i n   i m p r o v i n g   p a i n   ,   m o b i l i t y   ,   a n d   s y s t e m i c   l o w - g r a d e   i n f l a m m a t i o n   i n   t h e   s h o r t   t e r m   a n d   w h e t h e r   t h e   e f f e c t   w o u l d   b e   s u s t a i n e d   a t   @   w e e k s   i n   o l d e r   a d u l t s   w i t h   m o d e r a t e   t o   s e v e r e   k n e e   o s t e o a r t h r i t i s   (   o a   )   .\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nLd_rXYo61a1",
        "outputId": "bf80ba7e-ebe0-44c9-8b05-a7a651e86927"
      },
      "source": [
        "train_chars[:5]"
      ],
      "execution_count": 64,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['t o   i n v e s t i g a t e   t h e   e f f i c a c y   o f   @   w e e k s   o f   d a i l y   l o w - d o s e   o r a l   p r e d n i s o l o n e   i n   i m p r o v i n g   p a i n   ,   m o b i l i t y   ,   a n d   s y s t e m i c   l o w - g r a d e   i n f l a m m a t i o n   i n   t h e   s h o r t   t e r m   a n d   w h e t h e r   t h e   e f f e c t   w o u l d   b e   s u s t a i n e d   a t   @   w e e k s   i n   o l d e r   a d u l t s   w i t h   m o d e r a t e   t o   s e v e r e   k n e e   o s t e o a r t h r i t i s   (   o a   )   .',\n",
              " 'a   t o t a l   o f   @   p a t i e n t s   w i t h   p r i m a r y   k n e e   o a   w e r e   r a n d o m i z e d   @ : @   ;   @   r e c e i v e d   @   m g / d a y   o f   p r e d n i s o l o n e   a n d   @   r e c e i v e d   p l a c e b o   f o r   @   w e e k s   .',\n",
              " 'o u t c o m e   m e a s u r e s   i n c l u d e d   p a i n   r e d u c t i o n   a n d   i m p r o v e m e n t   i n   f u n c t i o n   s c o r e s   a n d   s y s t e m i c   i n f l a m m a t i o n   m a r k e r s   .',\n",
              " 'p a i n   w a s   a s s e s s e d   u s i n g   t h e   v i s u a l   a n a l o g   p a i n   s c a l e   (   @ - @   m m   )   .',\n",
              " 's e c o n d a r y   o u t c o m e   m e a s u r e s   i n c l u d e d   t h e   w e s t e r n   o n t a r i o   a n d   m c m a s t e r   u n i v e r s i t i e s   o s t e o a r t h r i t i s   i n d e x   s c o r e s   ,   p a t i e n t   g l o b a l   a s s e s s m e n t   (   p g a   )   o f   t h e   s e v e r i t y   o f   k n e e   o a   ,   a n d   @ - m i n   w a l k   d i s t a n c e   (   @ m w d   )   .']"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 64
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dmYHDLM47R-P",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5adf75ca-6856-43d3-c85f-d6001cac63ac"
      },
      "source": [
        "# Check the Avrage Char Length in the training Sentences\n",
        "char_lens = [len(sentence) for sentence in train_sentences]\n",
        "avg_char_lens = sum(char_lens)/len(char_lens)\n",
        "avg_char_lens"
      ],
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "149.3662574983337"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 65
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 478
        },
        "id": "TKZ62YGKh_t5",
        "outputId": "6eab3468-eef6-4d0f-e009-c4ddace36854"
      },
      "source": [
        "import matplotlib.pyplot as plt\n",
        "plt.hist(char_lens,bins =25)"
      ],
      "execution_count": 66,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(array([1.2341e+04, 5.0061e+04, 5.6508e+04, 3.4345e+04, 1.5779e+04,\n",
              "        6.5770e+03, 2.4690e+03, 1.0890e+03, 4.4400e+02, 1.9900e+02,\n",
              "        1.0500e+02, 5.0000e+01, 2.4000e+01, 1.9000e+01, 1.1000e+01,\n",
              "        7.0000e+00, 4.0000e+00, 3.0000e+00, 0.0000e+00, 2.0000e+00,\n",
              "        1.0000e+00, 1.0000e+00, 0.0000e+00, 0.0000e+00, 1.0000e+00]),\n",
              " array([1.0000e+00, 5.6400e+01, 1.1180e+02, 1.6720e+02, 2.2260e+02,\n",
              "        2.7800e+02, 3.3340e+02, 3.8880e+02, 4.4420e+02, 4.9960e+02,\n",
              "        5.5500e+02, 6.1040e+02, 6.6580e+02, 7.2120e+02, 7.7660e+02,\n",
              "        8.3200e+02, 8.8740e+02, 9.4280e+02, 9.9820e+02, 1.0536e+03,\n",
              "        1.1090e+03, 1.1644e+03, 1.2198e+03, 1.2752e+03, 1.3306e+03,\n",
              "        1.3860e+03]),\n",
              " <a list of 25 Patch objects>)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 66
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYMAAAD4CAYAAAAO9oqkAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAASW0lEQVR4nO3deYyc9X3H8fcndjhygE3YuK7tdJ3GauVEChALjBJVKbTGQBRTiUSgqDiUxlIhUtJGSk0iFeVAgrYKCVIuFNyYiAQoSYrFUdcFoqp/cCzhhhBvOIotwA7maIpykHz7x/wWJs6uPWt2Z8fs+yWN9nm+v9/MfOfR7nz2OXY2VYUkaXZ7zUw3IEmaeYaBJMkwkCQZBpIkDANJEjB3phvYV4cffngNDw/PdBuStN+44447flpVQ+ON7bdhMDw8zMjIyEy3IUn7jSSPTTTmYSJJkmEgSTIMJEkYBpIkDANJEoaBJAnDQJKEYSBJwjCQJLEf/wXyIBtef92k5j96wcnT1Ikk9cY9A0mSYSBJMgwkSRgGkiQMA0kShoEkCcNAkoRhIEnCMJAkYRhIkjAMJEkYBpIkDANJEoaBJAnDQJKEYSBJwjCQJGEYSJIwDCRJ9BgGSR5Ncm+Su5KMtNphSbYk2dq+zm/1JLk4yWiSe5Ic1fU4a9v8rUnWdtXf1R5/tN03U/1CJUkTm8yewZ9W1RFVtaKtrwdurKplwI1tHeBEYFm7rQO+Cp3wAM4DjgGOBs4bC5A25yNd91u9z69IkjRpr+Qw0RpgY1veCJzSVb+sOm4B5iVZCJwAbKmqXVX1DLAFWN3GDqmqW6qqgMu6HkuS1Adze5xXwH8kKeDrVXUJsKCqnmjjTwIL2vIi4PGu+25rtT3Vt41THxjD66+b6RYkaVr1GgbvqartSd4MbEnyo+7BqqoWFNMqyTo6h554y1veMt1PJ0mzRk+Hiapqe/u6A/g+nWP+T7VDPLSvO9r07cCSrrsvbrU91RePUx+vj0uqakVVrRgaGuqldUlSD/YaBklen+SNY8vAKuA+YBMwdkXQWuCatrwJOKNdVbQSeK4dTtoMrEoyv504XgVsbmPPJ1nZriI6o+uxJEl90MthogXA99vVnnOBb1fVvye5HbgqyVnAY8AH2/zrgZOAUeAF4EyAqtqV5HPA7W3eZ6tqV1s+G/gmcDBwQ7tJkvpkr2FQVQ8D7xyn/jRw/Dj1As6Z4LE2ABvGqY8A7+ihX0nSNPAvkCVJhoEkyTCQJGEYSJIwDCRJGAaSJAwDSRKGgSQJw0CShGEgScIwkCRhGEiSMAwkSRgGkiQMA0kShoEkCcNAkoRhIEnCMJAkYRhIkjAMJEkYBpIkDANJEoaBJAnDQJKEYSBJwjCQJGEYSJKYRBgkmZPkziTXtvWlSW5NMprkyiQHtPqBbX20jQ93Pca5rf5QkhO66qtbbTTJ+ql7eZKkXkxmz+BjwINd6xcCF1XV24BngLNa/SzgmVa/qM0jyXLgNODtwGrgKy1g5gBfBk4ElgOnt7mSpD7pKQySLAZOBr7R1gMcB1zdpmwETmnLa9o6bfz4Nn8NcEVV/aKqHgFGgaPbbbSqHq6qXwJXtLmSpD7pdc/gi8Angd+09TcBz1bVi219G7CoLS8CHgdo48+1+S/Vd7vPRPXfkWRdkpEkIzt37uyxdUnS3uw1DJK8D9hRVXf0oZ89qqpLqmpFVa0YGhqa6XYk6VVjbg9z3g28P8lJwEHAIcCXgHlJ5rbf/hcD29v87cASYFuSucChwNNd9THd95moLknqg73uGVTVuVW1uKqG6ZwAvqmqPgTcDJzapq0FrmnLm9o6bfymqqpWP61dbbQUWAbcBtwOLGtXJx3QnmPTlLw6SVJPetkzmMjfA1ck+TxwJ3Bpq18KfCvJKLCLzps7VXV/kquAB4AXgXOq6tcAST4KbAbmABuq6v5X0JckaZImFQZV9QPgB235YTpXAu0+5+fABya4//nA+ePUrweun0wvkqSp418gS5IMA0mSYSBJwjCQJGEYSJIwDCRJGAaSJAwDSRKGgSQJw0CShGEgScIwkCRhGEiSMAwkSRgGkiRe2T+30RQZXn/dpOY/esHJ09SJpNnKPQNJkmEgSTIMJEkYBpIkDANJEoaBJAnDQJKEYSBJwjCQJGEYSJIwDCRJGAaSJHoIgyQHJbktyd1J7k/ymVZfmuTWJKNJrkxyQKsf2NZH2/hw12Od2+oPJTmhq7661UaTrJ/6lylJ2pNe9gx+ARxXVe8EjgBWJ1kJXAhcVFVvA54BzmrzzwKeafWL2jySLAdOA94OrAa+kmROkjnAl4ETgeXA6W2uJKlP9hoG1fGztvradivgOODqVt8InNKW17R12vjxSdLqV1TVL6rqEWAUOLrdRqvq4ar6JXBFmytJ6pOezhm03+DvAnYAW4CfAM9W1YttyjZgUVteBDwO0MafA97UXd/tPhPVx+tjXZKRJCM7d+7spXVJUg96CoOq+nVVHQEspvOb/B9Pa1cT93FJVa2oqhVDQ0Mz0YIkvSpN6mqiqnoWuBk4FpiXZOw/pS0Gtrfl7cASgDZ+KPB0d323+0xUlyT1SS9XEw0lmdeWDwb+HHiQTiic2qatBa5py5vaOm38pqqqVj+tXW20FFgG3AbcDixrVycdQOck86apeHGSpN708j+QFwIb21U/rwGuqqprkzwAXJHk88CdwKVt/qXAt5KMArvovLlTVfcnuQp4AHgROKeqfg2Q5KPAZmAOsKGq7p+yVyhJ2qu9hkFV3QMcOU79YTrnD3av/xz4wASPdT5w/jj164Hre+hXkjQN/AtkSZJhIEkyDCRJGAaSJAwDSRKGgSQJw0CShGEgScIwkCRhGEiSMAwkSRgGkiQMA0kShoEkCcNAkoRhIEnCMJAkYRhIkjAMJEkYBpIkDANJEoaBJAnDQJKEYSBJwjCQJGEYSJIwDCRJ9BAGSZYkuTnJA0nuT/KxVj8syZYkW9vX+a2eJBcnGU1yT5Kjuh5rbZu/Ncnarvq7ktzb7nNxkkzHi5Ukja+XPYMXgU9U1XJgJXBOkuXAeuDGqloG3NjWAU4ElrXbOuCr0AkP4DzgGOBo4LyxAGlzPtJ1v9Wv/KVJknq11zCoqieq6odt+X+BB4FFwBpgY5u2ETilLa8BLquOW4B5SRYCJwBbqmpXVT0DbAFWt7FDquqWqirgsq7HkiT1waTOGSQZBo4EbgUWVNUTbehJYEFbXgQ83nW3ba22p/q2ceqSpD7pOQySvAH4LvDxqnq+e6z9Rl9T3Nt4PaxLMpJkZOfOndP9dJI0a/QUBkleSycILq+q77XyU+0QD+3rjlbfDizpuvviVttTffE49d9RVZdU1YqqWjE0NNRL65KkHvRyNVGAS4EHq+oLXUObgLErgtYC13TVz2hXFa0EnmuHkzYDq5LMbyeOVwGb29jzSVa25zqj67EkSX0wt4c57wb+Erg3yV2t9ingAuCqJGcBjwEfbGPXAycBo8ALwJkAVbUryeeA29u8z1bVrrZ8NvBN4GDghnaTJPXJXsOgqv4bmOi6/+PHmV/AORM81gZgwzj1EeAde+tFkjQ9/AtkSZJhIEkyDCRJGAaSJAwDSRKGgSQJw0CShGEgScIwkCRhGEiS6O2ziTRghtdfN+n7PHrBydPQiaRXC/cMJEmGgSTJMJAkYRhIkjAMJEkYBpIkDANJEoaBJAnDQJKEYSBJwjCQJGEYSJKYpR9Uty8f9CZJr2buGUiSDANJkmEgScIwkCTRQxgk2ZBkR5L7umqHJdmSZGv7Or/Vk+TiJKNJ7klyVNd91rb5W5Os7aq/K8m97T4XJ8lUv0hJ0p71smfwTWD1brX1wI1VtQy4sa0DnAgsa7d1wFehEx7AecAxwNHAeWMB0uZ8pOt+uz+XJGma7TUMquq/gF27ldcAG9vyRuCUrvpl1XELMC/JQuAEYEtV7aqqZ4AtwOo2dkhV3VJVBVzW9ViSpD7Z13MGC6rqibb8JLCgLS8CHu+at63V9lTfNk59XEnWJRlJMrJz5859bF2StLtXfAK5/UZfU9BLL891SVWtqKoVQ0ND/XhKSZoV9jUMnmqHeGhfd7T6dmBJ17zFrban+uJx6pKkPtrXMNgEjF0RtBa4pqt+RruqaCXwXDuctBlYlWR+O3G8Ctjcxp5PsrJdRXRG12NJkvpkr59NlOQ7wHuBw5Nso3NV0AXAVUnOAh4DPtimXw+cBIwCLwBnAlTVriSfA25v8z5bVWMnpc+mc8XSwcAN7SZJ6qO9hkFVnT7B0PHjzC3gnAkeZwOwYZz6CPCOvfUhSZo+/gWyJMkwkCQZBpIkDANJEoaBJAnDQJKEYSBJooe/M9Crw/D66yY1/9ELTp6mTiQNIvcMJEmGgSTJMJAkYRhIkjAMJEkYBpIkDANJEoaBJAnDQJKEYSBJwjCQJGEYSJIwDCRJGAaSJPwIa03Aj7yWZhf3DCRJhoEkyTCQJGEYSJLwBLKmiCecpf3bwOwZJFmd5KEko0nWz3Q/kjSbDMSeQZI5wJeBPwe2Abcn2VRVD8xsZ5ou7klIg2UgwgA4GhitqocBklwBrAEMAwGTD499YeBoNhuUMFgEPN61vg04ZvdJSdYB69rqz5I8tI/Pdzjw032870zYn/rdn3qFrn5z4Qx30pv9dvvuJ17t/f7BRAODEgY9qapLgEte6eMkGamqFVPQUl/sT/3uT72C/U43+51eU9nvoJxA3g4s6Vpf3GqSpD4YlDC4HViWZGmSA4DTgE0z3JMkzRoDcZioql5M8lFgMzAH2FBV90/jU77iQ019tj/1uz/1CvY73ex3ek1Zv6mqqXosSdJ+alAOE0mSZpBhIEmaXWEwiB95kWRJkpuTPJDk/iQfa/XDkmxJsrV9nd/qSXJxew33JDlqhvqek+TOJNe29aVJbm19XdkuBCDJgW19tI0Pz0Cv85JcneRHSR5Mcuwgb98kf9u+F+5L8p0kBw3S9k2yIcmOJPd11Sa9PZOsbfO3Jlnb537/qX0/3JPk+0nmdY2d2/p9KMkJXfW+vH+M12/X2CeSVJLD2/rUbd+qmhU3OiemfwK8FTgAuBtYPgB9LQSOastvBH4MLAf+EVjf6uuBC9vyScANQICVwK0z1PffAd8Grm3rVwGnteWvAX/Tls8GvtaWTwOunIFeNwJ/3ZYPAOYN6val8weYjwAHd23XDw/S9gX+BDgKuK+rNqntCRwGPNy+zm/L8/vY7ypgblu+sKvf5e294UBgaXvPmNPP94/x+m31JXQusnkMOHyqt29ffyhn8gYcC2zuWj8XOHem+xqnz2vofEbTQ8DCVlsIPNSWvw6c3jX/pXl97HExcCNwHHBt+0b8adcP10vbun3zHtuW57Z56WOvh7Y31+xWH8jty8t/jX9Y217XAicM2vYFhnd7c53U9gROB77eVf+tedPd725jfwFc3pZ/631hbPv2+/1jvH6Bq4F3Ao/ychhM2fadTYeJxvvIi0Uz1Mu42i7+kcCtwIKqeqINPQksaMuD8Dq+CHwS+E1bfxPwbFW9OE5PL/Xbxp9r8/tlKbAT+Jd2WOsbSV7PgG7fqtoO/DPwP8ATdLbXHQzu9h0z2e05CN/HY/6Kzm/XMKD9JlkDbK+qu3cbmrJ+Z1MYDLQkbwC+C3y8qp7vHqtOtA/ENcBJ3gfsqKo7ZrqXHs2ls8v91ao6Evg/OocxXjJg23c+nQ9pXAr8PvB6YPWMNjVJg7Q99ybJp4EXgctnupeJJHkd8CngH6bzeWZTGAzsR14keS2dILi8qr7Xyk8lWdjGFwI7Wn2mX8e7gfcneRS4gs6hoi8B85KM/RFjd08v9dvGDwWe7mO/24BtVXVrW7+aTjgM6vb9M+CRqtpZVb8Cvkdnmw/q9h0z2e0509uZJB8G3gd8qAUYe+hrJvv9Qzq/HNzdfu4WAz9M8nt76GvS/c6mMBjIj7xIEuBS4MGq+kLX0CZg7AqAtXTOJYzVz2hXEawEnuvaPZ92VXVuVS2uqmE62/CmqvoQcDNw6gT9jr2OU9v8vv3WWFVPAo8n+aNWOp7OR6MP5Palc3hoZZLXte+NsX4Hcvt2mez23AysSjK/7Q2tarW+SLKazqHO91fVC11Dm4DT2lVaS4FlwG3M4PtHVd1bVW+uquH2c7eNzkUnTzKV23e6ToAM4o3Omfcf07kq4NMz3U/r6T10dqnvAe5qt5PoHPe9EdgK/CdwWJsfOv8I6CfAvcCKGez9vbx8NdFb6fzQjAL/ChzY6ge19dE2/tYZ6PMIYKRt43+jc3XFwG5f4DPAj4D7gG/RubJlYLYv8B065zN+1d6YztqX7UnnWP1ou53Z535H6RxTH/uZ+1rX/E+3fh8CTuyq9+X9Y7x+dxt/lJdPIE/Z9vXjKCRJs+owkSRpAoaBJMkwkCQZBpIkDANJEoaBJAnDQJIE/D92oqatmBaBkgAAAABJRU5ErkJggg==\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6Yimd67Si2Po"
      },
      "source": [
        "Okay, looks like most of our sequences are between 0 and 200 characters long.\n",
        "\n",
        "Let's use NumPy's percentile to figure out what length covers 95% of our sequences"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sewz5tu3jlzO",
        "outputId": "db6b8d01-531c-4a84-df46-4f374ed1c044"
      },
      "source": [
        "output_seq_char_len = int(np.percentile(char_lens, 95))\n",
        "output_seq_char_len"
      ],
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "290"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 67
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 53
        },
        "id": "fNt3Bpf0jwZH",
        "outputId": "8d1176b4-dde3-449d-8fb3-2af161c47347"
      },
      "source": [
        "random.choice(train_sentences)"
      ],
      "execution_count": 68,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'in patients with her@-positive metastatic breast cancer , the addition of pertuzumab to trastuzumab and docetaxel , as compared with the addition of placebo , significantly improved the median overall survival to @ months and extended the results of previous analyses showing the efficacy of this drug combination .'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 68
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 36
        },
        "id": "OgDL467dl4as",
        "outputId": "29f081db-dde4-4784-bcad-636c06adbbe0"
      },
      "source": [
        "# Get all keyboard characters for char-level embedding\n",
        "import string\n",
        "alphabet = string.ascii_lowercase + string.digits + string.punctuation\n",
        "alphabet"
      ],
      "execution_count": 69,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            },
            "text/plain": [
              "'abcdefghijklmnopqrstuvwxyz0123456789!\"#$%&\\'()*+,-./:;<=>?@[\\\\]^_`{|}~'"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 69
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5xqQYGL2k2V8"
      },
      "source": [
        "# Create char-level token vectorizer instance\n",
        "NUM_CHAR_TOKENS = len(alphabet) + 2 # num characters in alphabet + space + OOV token\n",
        "char_vectorizer = TextVectorization(max_tokens=NUM_CHAR_TOKENS,  \n",
        "                                    output_sequence_length=output_seq_char_len,\n",
        "                                    standardize=\"lower_and_strip_punctuation\",\n",
        "                                    name=\"char_vectorizer\")\n",
        "\n",
        "# Adapt character vectorizer to training characters\n",
        "char_vectorizer.adapt(train_chars)"
      ],
      "execution_count": 70,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "eOqYGJiOl8ty",
        "outputId": "a1770561-8229-4401-fc1a-630f3f2d833a"
      },
      "source": [
        "# Check character vocabulary characteristics\n",
        "char_vocab = char_vectorizer.get_vocabulary()\n",
        "print(f\"Number of different characters in character vocab: {len(char_vocab)}\")\n",
        "print(f\"5 most common characters: {char_vocab[:5]}\")\n",
        "print(f\"5 least common characters: {char_vocab[-5:]}\")"
      ],
      "execution_count": 71,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Number of different characters in character vocab: 28\n",
            "5 most common characters: ['', '[UNK]', 'e', 't', 'i']\n",
            "5 least common characters: ['k', 'x', 'z', 'q', 'j']\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "seeTxZK5mE0G",
        "outputId": "9a1cd49c-8d98-4791-aab7-ab0449133395"
      },
      "source": [
        "# Test out character vectorizer\n",
        "random_train_chars = random.choice(train_chars)\n",
        "print(f\"Charified text:\\n{random_train_chars}\")\n",
        "print(f\"\\nLength of chars: {len(random_train_chars.split())}\")\n",
        "vectorized_chars = char_vectorizer([random_train_chars])\n",
        "print(f\"\\nVectorized chars:\\n{vectorized_chars}\")\n",
        "print(f\"\\nLength of vectorized chars: {len(vectorized_chars[0])}\")"
      ],
      "execution_count": 72,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Charified text:\n",
            "a   c h e s t   x - r a y   w a s   t a k e n   a t   p o s t o p e r a t i v e   d a y   @   ,   a n d   t h e   r e s i d u a l   i n t r a a b d o m i n a l   g a s   v o l u m e   w a s   m e a s u r e d   .\n",
            "\n",
            "Length of chars: 88\n",
            "\n",
            "Vectorized chars:\n",
            "[[ 5 11 13  2  9  3 24  8  5 19 20  5  9  3  5 23  2  6  5  3 14  7  9  3\n",
            "   7 14  2  8  5  3  4 21  2 10  5 19  5  6 10  3 13  2  8  2  9  4 10 16\n",
            "   5 12  4  6  3  8  5  5 22 10  7 15  4  6  5 12 18  5  9 21  7 12 16 15\n",
            "   2 20  5  9 15  2  5  9 16  8  2 10  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0  0\n",
            "   0  0]]\n",
            "\n",
            "Length of vectorized chars: 290\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rFd8w_PBnGDm"
      },
      "source": [
        "### Creating a character-level embedding\n",
        "We've got a way to vectorize our character-level sequences, now's time to create a character-level embedding.\n",
        "\n",
        "The input dimension (`input_dim`) will be equal to the number of different characters in our `char_vocab` (28). And since we're following the structure of the model in Figure 1 of [*Neural Networks for Joint Sentence Classification\n",
        "in Medical Paper Abstracts*](https://arxiv.org/pdf/1612.05251.pdf), the output dimension of the character embedding (`output_dim`) will be 25."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GLFNqMdumt0A",
        "outputId": "4f284d60-7eaa-447a-cb3a-f01c65e5d9fe"
      },
      "source": [
        "char_embed = layers.Embedding(input_dim=NUM_CHAR_TOKENS,\n",
        "                              output_dim= 25,\n",
        "                              mask_zero= True, \n",
        "                              name= \"char_embed\")\n",
        "\n",
        "# Test out character embedding layer\n",
        "print(f\"Charified text (before vectorization and embedding):\\n{random_train_chars}\\n\")\n",
        "char_embed_example = char_embed(char_vectorizer([random_train_chars]))\n",
        "print(f\"Embedded chars (after vectorization and embedding):\\n{char_embed_example}\\n\")\n",
        "print(f\"Character embedding shape: {char_embed_example.shape}\")"
      ],
      "execution_count": 73,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Charified text (before vectorization and embedding):\n",
            "a   c h e s t   x - r a y   w a s   t a k e n   a t   p o s t o p e r a t i v e   d a y   @   ,   a n d   t h e   r e s i d u a l   i n t r a a b d o m i n a l   g a s   v o l u m e   w a s   m e a s u r e d   .\n",
            "\n",
            "Embedded chars (after vectorization and embedding):\n",
            "[[[ 0.01198739  0.01475773  0.00853588 ...  0.04154284 -0.00139939\n",
            "   -0.04978269]\n",
            "  [ 0.03823442  0.02977749 -0.03549073 ... -0.03217635 -0.00892054\n",
            "    0.03190038]\n",
            "  [ 0.00683529  0.01736682  0.00283471 ... -0.00423067 -0.00172087\n",
            "    0.01615829]\n",
            "  ...\n",
            "  [ 0.00047475  0.01790795 -0.0195184  ... -0.03690598  0.01174641\n",
            "    0.02456704]\n",
            "  [ 0.00047475  0.01790795 -0.0195184  ... -0.03690598  0.01174641\n",
            "    0.02456704]\n",
            "  [ 0.00047475  0.01790795 -0.0195184  ... -0.03690598  0.01174641\n",
            "    0.02456704]]]\n",
            "\n",
            "Character embedding shape: (1, 290, 25)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "63-7tbibqYQ5"
      },
      "source": [
        "Before fitting our model on the data, we'll create char-level batched `PrefetchedDataset`'s."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GRCYFl-Mowsc",
        "outputId": "24c34af2-7fde-4f55-cabf-e0159894541d"
      },
      "source": [
        "train_char_dataset = tf.data.Dataset.from_tensor_slices((train_chars, train_labels_one_hot)).batch(32).prefetch(tf.data.AUTOTUNE)\n",
        "val_char_dataset = tf.data.Dataset.from_tensor_slices((val_chars, val_labels_one_hot)).batch(32).prefetch(tf.data.AUTOTUNE)\n",
        "\n",
        "train_char_dataset"
      ],
      "execution_count": 74,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<PrefetchDataset shapes: ((None,), (None, 5)), types: (tf.string, tf.float64)>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 74
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ofw3Ow8sqdRC"
      },
      "source": [
        "### Building a Conv1D model to fit on character embeddings\n",
        "Now we've got a way to turn our character-level sequences into numbers (`char_vectorizer`) as well as numerically represent them as an embedding (`char_embed`) let's test how effective they are at encoding the information in our sequences by creating a character-level sequence model.\n",
        "\n",
        "The model will have the same structure as our custom token embedding model (`model_1`) except it'll take character-level sequences as input instead of token-level sequences.\n",
        "\n",
        "```\n",
        "Input (character-level text) -> Tokenize -> Embedding -> Layers (Conv1D, GlobalMaxPool1D) -> Output (label probability)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9B4TadikoJ_L",
        "outputId": "b7c0d8c8-28e3-4ae8-d0bd-692f907a805e"
      },
      "source": [
        "inputs = layers.Input(shape=(1,), dtype=\"string\")\n",
        "char_vectors = char_vectorizer(inputs)\n",
        "char_embeddings = char_embed(char_vectors)\n",
        "x = layers.Conv1D(64, kernel_size=5, padding=\"same\", activation=\"relu\",kernel_regularizer=tf.keras.regularizers.L2(0.01))(char_embeddings)\n",
        "x = layers.GlobalMaxPool1D()(x)\n",
        "outputs = layers.Dense(num_classes, activation=\"softmax\")(x)\n",
        "model_3 = tf.keras.Model(inputs=inputs,\n",
        "                         outputs=outputs,\n",
        "                         name=\"model_3_conv1D_char_embedding\")\n",
        "\n",
        "# Compile model\n",
        "model_3.compile(loss=\"categorical_crossentropy\",\n",
        "                optimizer=tf.keras.optimizers.Adam(),\n",
        "                metrics=[\"accuracy\"])\n",
        "\n",
        "model_3.summary()"
      ],
      "execution_count": 75,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"model_3_conv1D_char_embedding\"\n",
            "_________________________________________________________________\n",
            "Layer (type)                 Output Shape              Param #   \n",
            "=================================================================\n",
            "input_3 (InputLayer)         [(None, 1)]               0         \n",
            "_________________________________________________________________\n",
            "char_vectorizer (TextVectori (None, 290)               0         \n",
            "_________________________________________________________________\n",
            "char_embed (Embedding)       (None, 290, 25)           1750      \n",
            "_________________________________________________________________\n",
            "conv1d_1 (Conv1D)            (None, 290, 64)           8064      \n",
            "_________________________________________________________________\n",
            "global_max_pooling1d_1 (Glob (None, 64)                0         \n",
            "_________________________________________________________________\n",
            "dense_3 (Dense)              (None, 5)                 325       \n",
            "=================================================================\n",
            "Total params: 10,139\n",
            "Trainable params: 10,139\n",
            "Non-trainable params: 0\n",
            "_________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hDE27dS3pL1Q",
        "outputId": "89a37619-fb91-499e-e776-f62a22454033"
      },
      "source": [
        "# Fit the model on chars only\n",
        "model_3_history = model_3.fit(train_char_dataset,\n",
        "                              steps_per_epoch=int(0.1 * len(train_char_dataset)),\n",
        "                              epochs=10,\n",
        "                              validation_data=val_char_dataset,\n",
        "                              validation_steps=int(0.1 * len(val_char_dataset)))"
      ],
      "execution_count": 76,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "562/562 [==============================] - 14s 23ms/step - loss: 1.4009 - accuracy: 0.4576 - val_loss: 1.2213 - val_accuracy: 0.5432\n",
            "Epoch 2/10\n",
            "562/562 [==============================] - 13s 23ms/step - loss: 1.1807 - accuracy: 0.5530 - val_loss: 1.1203 - val_accuracy: 0.6001\n",
            "Epoch 3/10\n",
            "562/562 [==============================] - 13s 23ms/step - loss: 1.1121 - accuracy: 0.5955 - val_loss: 1.0601 - val_accuracy: 0.6303\n",
            "Epoch 4/10\n",
            "562/562 [==============================] - 13s 23ms/step - loss: 1.0576 - accuracy: 0.6207 - val_loss: 1.0272 - val_accuracy: 0.6400\n",
            "Epoch 5/10\n",
            "562/562 [==============================] - 13s 23ms/step - loss: 1.0327 - accuracy: 0.6365 - val_loss: 1.0117 - val_accuracy: 0.6592\n",
            "Epoch 6/10\n",
            "562/562 [==============================] - 13s 24ms/step - loss: 1.0137 - accuracy: 0.6437 - val_loss: 0.9768 - val_accuracy: 0.6576\n",
            "Epoch 7/10\n",
            "562/562 [==============================] - 13s 24ms/step - loss: 0.9968 - accuracy: 0.6497 - val_loss: 0.9487 - val_accuracy: 0.6729\n",
            "Epoch 8/10\n",
            "562/562 [==============================] - 13s 24ms/step - loss: 0.9578 - accuracy: 0.6701 - val_loss: 0.9473 - val_accuracy: 0.6712\n",
            "Epoch 9/10\n",
            "562/562 [==============================] - 13s 23ms/step - loss: 0.9539 - accuracy: 0.6697 - val_loss: 0.9498 - val_accuracy: 0.6659\n",
            "Epoch 10/10\n",
            "562/562 [==============================] - 13s 23ms/step - loss: 0.9476 - accuracy: 0.6668 - val_loss: 0.9276 - val_accuracy: 0.6689\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "o-pTFsdJpNn6",
        "outputId": "49612a00-4123-4e01-9058-042b988d6f6e"
      },
      "source": [
        "# Evaluate model_3 on whole validation char dataset\n",
        "model_3.evaluate(val_char_dataset)"
      ],
      "execution_count": 77,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "945/945 [==============================] - 6s 7ms/step - loss: 0.9418 - accuracy: 0.6693\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.941790759563446, 0.6693366765975952]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 77
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2CvDMAnJpNkZ",
        "outputId": "350ff442-3cfc-441e-f9c0-990ef04d8b5b"
      },
      "source": [
        "# Make predictions with character model only\n",
        "model_3_pred_probs = model_3.predict(val_char_dataset)\n",
        "model_3_pred_probs"
      ],
      "execution_count": 78,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.21880805, 0.47745466, 0.06887063, 0.17308971, 0.06177694],\n",
              "       [0.23530294, 0.43719485, 0.00282957, 0.3047214 , 0.01995131],\n",
              "       [0.15760659, 0.33806348, 0.04908546, 0.3893838 , 0.06586069],\n",
              "       ...,\n",
              "       [0.01113645, 0.02624117, 0.0150859 , 0.00669696, 0.9408395 ],\n",
              "       [0.01904003, 0.16237463, 0.07217802, 0.01535511, 0.7310522 ],\n",
              "       [0.17195038, 0.6881715 , 0.10410943, 0.01864745, 0.01712121]],\n",
              "      dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 78
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "5IfbmitrpWJA",
        "outputId": "ab6fcd0b-7577-4295-d595-7de113219f75"
      },
      "source": [
        "# Convert predictions to classes\n",
        "model_3_preds = tf.argmax(model_3_pred_probs, axis=1)\n",
        "model_3_preds"
      ],
      "execution_count": 79,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(30212,), dtype=int64, numpy=array([1, 1, 3, ..., 4, 4, 1])>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 79
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YcCeIIFRsiUE"
      },
      "source": [
        "### Model_3 Results"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "vBPJAulEpNjS",
        "outputId": "6f2ce553-2a82-43bb-d65e-980d9ffc290f"
      },
      "source": [
        "# Calculate Conv1D char only model results\n",
        "model_3_results = calculate_results(y_true=val_labels_encoded,\n",
        "                                        y_pred=model_3_preds)\n",
        "model_3_results"
      ],
      "execution_count": 80,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 66.93366874089766,\n",
              " 'f1': 0.6597343453210113,\n",
              " 'precision': 0.6638601787800059,\n",
              " 'recall': 0.6693366874089766}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 80
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rZKQgGfesqbJ"
      },
      "source": [
        "## Model 4: Combining pretrained token embeddings + character embeddings (hybrid embedding layer)\n",
        "\n",
        "In moving closer to build a model similar to the one in Figure 1 of [*Neural Networks for Joint Sentence Classification\n",
        "in Medical Paper Abstracts*](https://arxiv.org/pdf/1612.05251.pdf), it's time we tackled the hybrid token embedding layer they speak of."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "l68PUtwitOlG"
      },
      "source": [
        "![nnn.JPG](data:image/jpeg;base64,/9j/4AAQSkZJRgABAQEAeAB4AAD/4RDSRXhpZgAATU0AKgAAAAgABAE7AAIAAAAERDN2AIdpAAQAAAABAAAISpydAAEAAAAIAAAQwuocAAcAAAgMAAAAPgAAAAAc6gAAAAgAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAFkAMAAgAAABQAABCYkAQAAgAAABQAABCskpEAAgAAAAM3NgAAkpIAAgAAAAM3NgAA6hwABwAACAwAAAiMAAAAABzqAAAACAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAMjAyMTowNjowNiAxOTo1NzowMQAyMDIxOjA2OjA2IDE5OjU3OjAxAAAARAAzAHYAAAD/4QsWaHR0cDovL25zLmFkb2JlLmNvbS94YXAvMS4wLwA8P3hwYWNrZXQgYmVnaW49J++7vycgaWQ9J1c1TTBNcENlaGlIenJlU3pOVGN6a2M5ZCc/Pg0KPHg6eG1wbWV0YSB4bWxuczp4PSJhZG9iZTpuczptZXRhLyI+PHJkZjpSREYgeG1sbnM6cmRmPSJodHRwOi8vd3d3LnczLm9yZy8xOTk5LzAyLzIyLXJkZi1zeW50YXgtbnMjIj48cmRmOkRlc2NyaXB0aW9uIHJkZjphYm91dD0idXVpZDpmYWY1YmRkNS1iYTNkLTExZGEtYWQzMS1kMzNkNzUxODJmMWIiIHhtbG5zOmRjPSJodHRwOi8vcHVybC5vcmcvZGMvZWxlbWVudHMvMS4xLyIvPjxyZGY6RGVzY3JpcHRpb24gcmRmOmFib3V0PSJ1dWlkOmZhZjViZGQ1LWJhM2QtMTFkYS1hZDMxLWQzM2Q3NTE4MmYxYiIgeG1sbnM6eG1wPSJodHRwOi8vbnMuYWRvYmUuY29tL3hhcC8xLjAvIj48eG1wOkNyZWF0ZURhdGU+MjAyMS0wNi0wNlQxOTo1NzowMS43NTY8L3htcDpDcmVhdGVEYXRlPjwvcmRmOkRlc2NyaXB0aW9uPjxyZGY6RGVzY3JpcHRpb24gcmRmOmFib3V0PSJ1dWlkOmZhZjViZGQ1LWJhM2QtMTFkYS1hZDMxLWQzM2Q3NTE4MmYxYiIgeG1sbnM6ZGM9Imh0dHA6Ly9wdXJsLm9yZy9kYy9lbGVtZW50cy8xLjEvIj48ZGM6Y3JlYXRvcj48cmRmOlNlcSB4bWxuczpyZGY9Imh0dHA6Ly93d3cudzMub3JnLzE5OTkvMDIvMjItcmRmLXN5bnRheC1ucyMiPjxyZGY6bGk+RDN2PC9yZGY6bGk+PC9yZGY6U2VxPg0KCQkJPC9kYzpjcmVhdG9yPjwvcmRmOkRlc2NyaXB0aW9uPjwvcmRmOlJERj48L3g6eG1wbWV0YT4NCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgCiAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAKICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgICAgIAogICAgICAgICAgICAgICAgICAgICAgICAgICAgPD94cGFja2V0IGVuZD0ndyc/Pv/bAEMABwUFBgUEBwYFBggHBwgKEQsKCQkKFQ8QDBEYFRoZGBUYFxseJyEbHSUdFxgiLiIlKCkrLCsaIC8zLyoyJyorKv/bAEMBBwgICgkKFAsLFCocGBwqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKioqKv/AABEIAxwCJgMBIgACEQEDEQH/xAAfAAABBQEBAQEBAQAAAAAAAAAAAQIDBAUGBwgJCgv/xAC1EAACAQMDAgQDBQUEBAAAAX0BAgMABBEFEiExQQYTUWEHInEUMoGRoQgjQrHBFVLR8CQzYnKCCQoWFxgZGiUmJygpKjQ1Njc4OTpDREVGR0hJSlNUVVZXWFlaY2RlZmdoaWpzdHV2d3h5eoOEhYaHiImKkpOUlZaXmJmaoqOkpaanqKmqsrO0tba3uLm6wsPExcbHyMnK0tPU1dbX2Nna4eLj5OXm5+jp6vHy8/T19vf4+fr/xAAfAQADAQEBAQEBAQEBAAAAAAAAAQIDBAUGBwgJCgv/xAC1EQACAQIEBAMEBwUEBAABAncAAQIDEQQFITEGEkFRB2FxEyIygQgUQpGhscEJIzNS8BVictEKFiQ04SXxFxgZGiYnKCkqNTY3ODk6Q0RFRkdISUpTVFVWV1hZWmNkZWZnaGlqc3R1dnd4eXqCg4SFhoeIiYqSk5SVlpeYmZqio6Slpqeoqaqys7S1tre4ubrCw8TFxsfIycrS09TV1tfY2dri4+Tl5ufo6ery8/T19vf4+fr/2gAMAwEAAhEDEQA/APpGiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAPTiuB0LWdWHjzV4NSvS9ppumRzaguQY4rmRi4ROOAsYwfXIJ5rvq8b8Qm4u9A16XwfdMJfGN79lit7m0ZZyygQyuh3D92EUtuI459Ridb6b/rt+v4DVra7f0/0O9+HWo6nrPgSw1XWpTJc3++5UFQuyN3JjXAHZCtdIkscjOscis0bbXCtkqcA4PocEH8aoGyt9L8MLZpNJa2tnahPMibayIi9j24Fc58JbN4fh5aXs/mGfVZZdQkMrl2PmuWXLEknCbBk+laOzk7bL+l+pKvypvd/0ztaKKKkYUUUUAFFFFABRRRQBm6+9ymlM1rcLaoGDXFwSMxQjl2XIPzYGB6Zz2rz6HxR4gb4Y+G5Uuni1TxBqkcNrI6hnjtnlZxnI5IgXqfX15rtvGlzZQ+EL6HVPtRt71PsRFpEZJSZj5Y2qO+WrjdJ0rUbz4k6LYTXNvd6b4RsCrSW8BjT7S6CNFOWOXEeWIH3dw9aI/F81+GrX5feEvh+//Jfn+B6TdXltYw+be3EVvHnG+Vwi59MmnxTR3EKSwSLLG4DI6MCrD1BHWuX1e4l03xG2rap+5sLVVWGUsrbwVbciICWMjOVGAOQgwecUfDjTdR03wmf7XhNtPd3lxeLaE5NsksjOsZ9CAeR2JNC1B6HV0UUUAFFFFABRRRQAUUUUAFFFFABRRRQAV59rqXr/ABV8P2UN5J50s8t9KUYgQ2cUWzysZwd8kmSe/H90V6DXnfiTUIR4u1C8gjvrPWNMsxaWhgMbvqQlw+xImBPysF+fGAd2eAaV7NP+v6vYdrpr+v6saXgi+utY8SeLtTluJZLMakLC0iZyURYECuVHQZdmzjriuosdTs9TFwbC4ScW07W8pT+CRfvL9RmuZ8I6FqfhL4f6XpMYjm1JiXvLiQ7kjkkZpJHbkFsEkDHU46DmtLwTrlx4k8JWuq3aRq87y7DGCFdFkZUfBJxuUA4z3qrW07afPuTe+vf8jeooopDCiiigAooooAKKKKACiiigArgPiuLwaHGtpdOk97JFp9hFGxUrcyyAeaSD/AgJHocn0rv647xvPYzapo1hqYltUjlbUIdTEqKltLDjauG++W3kbQCcZPuF1Xr/AMP+A+j/AK9PxI/td3ffGO206K7m+x6Po5muUVyFlmmcKm8DgkKjEZ6Zrta4DwTper2Np4k8S3cNxcalrFwZLWC7CpIYY12wq4AATPJx2Dc85qofHPiFPCGqa4y6e0dhqLWVsFhfN+ROIsj5/kySVH3uRnpxVdl1/wA3+l7E7arbb7l+trnpVISFUliAAMkntWRqPi/w5pGoLYatr+mWN4wBFvcXkcchz0+UnPNZHxUlhh+GurPLEJpWi8q3jbOGmkPlpx3+Zxwal3tdFJJuzOnsL+11SxjvNPnW4tpQTHKnKuAcZB7jjr3qxVDQtLj0Tw9p+lwACOyto4Fx/sqB/Sr9VKybSJi21dhRRRSGFFFFABRRRQBk+JhjQppZbk21pAPPu2UkM0KAsygjpnGCfQmvN4brVx8MfCGmtd3EF/4j1SJ/3chV4LZpGuGRSOQFjAX6HFegeNbuG08I3v2qwk1C3uFFrNBHIsfySEIzF2ICgBiST0xXJ+H9Bmv/AIg6dew39zf6H4csGtrOeZUCSzOApKbVG8Ki4L9CTgdDRH4vmvw1f36L5hL4fv8Ax0X3as9Korh7PxN4gvfFWraZItjZWun2sdzNPJGXa33mTbGQGwx2KrE5GMkYORje8H6zceIfBulavewrBPe2yzOiAhRkdQDyARzj3oWqv/XX/IHo7f1/WptUUUUAFFFFABRRRQAUUUUAFFFFABVWDTLC1upLm1sraGeXPmSxxKrPk5OSBk81aooA5zxzp2ua14Uv9J8PLZLLfW7wNPdXDx+UGGCQFjbPBPpWto1tJZaLaWk0ENu1vEsQihlMiKqjAwxVSeAOwq7RQtAetgooooAKKKKACiiigAooooAjuLaC7t3gu4Y54XGHjkUMrD3B4NJbWtvZW6wWcEVvCv3Y4kCqPwFS0UAcXe2Xi221STVrPTtJ1a4kkeOCK5uXgNlFnC7WCuGyPmbAUnOMkAY6jSre6tdLgi1G5+1XYXM0oGAzk5O0dlBOAPQCrlFC0Vg63CiiigAooooAKKKKACiiigAooooAKKKKACiiigDkfG3iq2stI1bSNOmMmvNZhba0VG3O825I8HGDyCSAeApJwK3tA0qPQ/DmnaVDjZZW0cAx32qBn9Ktwvb3Kx3UBjlVl+SVMHKn0PpUtC0T8/8Ag2/MHr/Xp/kFFFFABRRRQAUUUUAFFFFABRRRQAUUUUANlDtC6xsEcqQrEZwfXFcangB4PB2gaLb6gpk0e5ju3lkhLJcyruJLKGB++2/GeoFdpTJZooFDTyLGrMqAscZZjgD6kkCj/gfgH9feQWdhHa2cEDs1w8KBDNN8zuQOpPqetYXi3wxqfiaSwSDVbS0tLO8hvPJksWlMrxnIVmEq/LnBxjsOa6eijrcOlhsYcRqJWVnwNzKuAT3IGTj8zTqKKACiiigAooooAKKKKACiimTTRW8LS3EixRqMs7nAH40ActN4JefS/Flu2pMtz4jkcm4SPBgQxLGiYzzgL7Zyelbeh6Y+k6cLeSYStkEBAVSNQoVURSThQqj6nJ71o0ULT8F9wPX+u4UUUUAFFFFABRRRQAUUUUAFFFFABWV4n1638MeFdS1u75isbd5iucbyBwv1JwPxrVryz406brHjSHSvAfh4FDqUv2nUbogmO2tozwX/AN58FR3MZ+tAEf7PXji48Y+AZodVuWudU066dJ5HPzSLIS6McfVl/wCAV6vXzZ8JvD2p/Dn4mXE1qZLvw1fajc6HNOfvQTRSYhZwB/EcKD0y5HHGfpOgAooooAKKKKACiiigAooooAK8i/aF8f3PgvwrpsGkT+VqV7epIhB5EcLB2/Ntgx3BNeu18z/FnwtqfxA1fxT4tvGltNA8NWj2un5HN5LGTvKj+4HLAt3wAOh2gH0Toer22v6BYavYnNvfW6Tx+wZQcH3GcVfry/4J2+seGtH1DwR4jjYXGjS+ZaXA/wBXc20pLKyHqcMHz6ZAr1CgAooooAKKKKACiiigAooooAKKKKACiiigArk/if4qHgz4b6vrKPtuI4DHbf8AXZ/lT64Jz9Aa6yvNviV4JvfiT4h0bQp2e28N2DG91GUHDTyHKxxJ7gbiT2DjvigDA/Zk8Wf218O5tFuJC91o05UbjkmGQllP/fW8ewAr2ivAvhv4MvfC/wATNQ8QeGkX+w5NYvdHv7FQS0Eav+5kT1AbaDnlRk8gnb77QAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAV8/ftN+Op9Fm8PaLpU5ju4rhdVkKnG3yziLPqC2449UFfQNfOnjn4bTa/Nq3jfxlvEmo6rZ2em2SMVNvZm4jjDP6OyEnb23Enk4UA948N65b+JfDGna1Z/6m+t0nUZztyMlT7g5B+ladcD8KvDWq+BdOv/AAnqLm6sbO4M+l3naSCQklD6OrZJH+2MV31ABRRRQAUUUUAFFFFABRRRQAV4p+074tOi+AbbQ7WUpdaxON204IhjIZj7ZbYPcZr2uvn34teBb3xNfeIfGfiEmPTdLSGy0a0BIaQeaqySuPQsz7cckYPQDcAes/DXxUPGfw60jWiwaeaAJc47TJ8r8dssCR7EV1NebfDPwXf/AA313WtAjZrnw7eOL3TJictC2AskUnv9wj1Ck9cgek0AFFFFABRRRQAUUUUAFFFFABRRRQAVyOp/Ejw1purTWFtJcatqkfyzWukWj3cqY7OUBC4yeGI71T8V3V94n8VReCdHupbK3W3F3rV7A22SOBiQkCN/C8hVsnqFB9a6zRtE0zw9pcWnaJZQ2VpEMLFCuB9T6k9yeTQB5h4N8b6Z4dg1qLxdp2raHFqGt3V5DNqemSxwlJHyu59pVT9T+NesW11Be2sdzZzx3EEqh45YnDK6noQRwRUjosiMkihlYYZWGQR6V5vqGmp8LvEEGt6KWg8LahcLb6ppqn9zZySMFS5iXog3EK4GBhgcccAHpNFFFABRRRQAUUUUAFFFISFUliAAMkntQBW1PVLHRtOlv9Wu4bO0hG6SadwqqPqa8v8AG/jAeL/h7q+jeD/DHiDUo720aG3u4tP8q3PoQZCpK/7qmtTw1pyfEXVf+Ey1+P7RpUUzLoGnyjMaIp2/amXo0jkErn7q4xyc16NQBxmmfEfQZ9RttP1e3vtA1KUeXBDrNobcyk4yEk5RskDgNzxXZ1T1bSNP13TJtO1izhvbOddskMyblP8AgfQ9RXHeC3u/CfiOfwJql5cX1ulub3Rr26k3yyW+7Dwux6tGxGP9lh0xigDvaKKKACiiigAooooAKKKKACiiigAooooAZPPFbW8k9zKkMMal3kkYKqKOSST0FcSPiRLq7MPBHhnUvEMKkj7cGS1tWwcHZJKRv57qpFV9Yt/+E/8AH03h65Jbw7oKxTajCDxe3TjdHC3qiLhyO5ZQRxXfxxpFGscSKiIAqqowFA6ACgDyvwxrXiLwXHq7+KPBWpx2l/qtxfm406WK8EKyHOGRG34GOoU/SvR9F1zTPEWlRalod7De2kv3ZYmyM9wR1BHcHkVfrzrxNp8PgHxPH430iMW9lfTxW3iGBeEkR2CJc47MjMM46qzd+aAPRaKKKACiiigAooooAKKKKACiiigAooqvqF9BpemXV/eNst7WF5pW/uoqlifyFAGT4m8Y6X4WSBL3zrm+uyVtNPs4/NuLkjqEQdh3Y4A7muO8S2nxE8b6PHbQ+H9H0S3W6gulXUdSeSZvKlWQAiKNlXO3BGT1rY+HOjy3Fm/jLXYw2ua8gmJbn7JbHmK3T0ULgn1Ykmu3oA4R/H2reHWX/hP/AA2+mWZIU6tp8/2u1Q+snyq8Y9ypHvXb29xDdW0dxaypNDKoeOSNgyupGQQRwQR3pzoksbRyKrowKsrDIIPYiuB8J2w8D+Obrwbb/Lot9A+paRGeluQ4E8A/2Qzq4HbeaAPQKKKKACiiigAooooAKKKCcDJ4FAGP4k8U6V4U09bvWJ2XzX8u3giQyTXEh6JGg5Zj6D8cVw/iiPx78QfDF1pdj4asNDs7rYVk1jUD55CurjMcSOFJ29CxrR8A2o8U6lcfEDU182S8d4NGRxxaWasVDKOzyYLMeuCB0rv6AOEl8aeJfDi+b418K7bBf9bqWiXBu44h3Z4iqyBR1JAbFdlp2pWWr6dBf6XdRXdpcLvimhYMrj2IqzXn1jZx+AviZDp+nqItC8UmWRLdR8lrfIu9tvosiBjj+8nGM4oA9BooooAKKKKACiiigAooooAKKKKAOI8CAS+L/HlzN/x9NrKQt6+UlrD5f4fM35mu3rz/AFW5HgT4kPrl4fL0HxGkVveXB+7aXcY2xO5/hR0ITPQFVz1r0AHIyORQAVzPxJghuPhb4njuQDH/AGVctyehETEH8CAa6auA8f3n/CTXkPw/0l/MuNQKyatJGeLOyDAvuI6NJjYo/wBonoKAI9F8T/EKXQbCQ+BbSYvbRsZX11UL5UfMV8o4J647Vd/4SL4hf9CBYfj4hX/4zXaoqoiogCqowAOwpaAOJ/4SH4h/9CFpw/7mEf8Axij+3/iJ/wBCLpg+viD/AO0V21FAHE/298Rf+hH0ofXX/wD7RR/bnxG/6ErSB9deP/xiu2ooA4n+2viOf+ZO0YfXXG/+MVk+K9Z+Iw8Ga0ZPC2kQINPn3Sxawzug8tvmVfJG4jqBkZ9a9MpksUc8LwzIHjkUq6sOGBGCKAMfwVHDF4B0BLYKIV0y3CBTkbfKXGDW3XAeAdQPhq5b4f63JsutPB/smaQ4F9ZZ+TaT1eMfIy9RtB6Gu/oAK87+I95c6b428D3ml6bNqd8l1dotpA6I8kZtm3jLkAYIU8n+H6V6DLLHBC808ixxRqWd3bCqB1JJ6CuD8NSt438dSeL0Vv7E02B7HRmYEfaWcjzrkD+6doRT3AY0AWf+E08Vf9E21f8A8D7P/wCO0f8ACaeKv+ibav8A+B9n/wDHa7aigDif+E08Vf8ARNtX/wDA+z/+O12NtLJNaQyzQPbySIrPC5BaMkZKkgkEjpwSKlooAK5/QbfxVD4g1qTxDeWE+lySqdKitlIkiTnIkyo5+70J710FFHUOljn9At/FUOua0/iO8sJ9NkmB0uO2UiSKPnIkyo5+70J70mhW/iyHUtcfxBeafPaST50hLdTuij+biX5Rk/d6E9DzXQ0UAcTs+Kf/AD8eEP8Avxdf/F0bPin/AM/Hg/8A78XX/wAXXbUUAcTs+Kf/AD8eD/8Avxdf/F0bPin/AM/Hg/8A78XX/wAXXbUUAef/AAo+3L/wlya2bc6qPEMxuTbhghzDDt2hiTt24xmvQK8/8QSSeA/G0vi3y3fQNVijg1kxqWNpInEVzgclNp2PjoAp7V3drdW99axXVlPHcW8yh45YnDK6noQRwRQBLXI/Fcwj4R+KPtONh02YLn++Vwv/AI9iuurzrWb6P4keJLfw5ozC40HTbpLjWr1OYpXjYNHaKejEsAz46AAZycUAWbTXfiPDYwJP4H0+d1jUM414KScdSDD1qb/hI/iD/wBE/svw8QL/APGa7aigDif+Eg+Ib8J4D06M+smvgj9Ia6+xe6ksIH1CGOC6aNTNFFIZFR8cgMQMgHvgVPRQBz9vb+Kh48up7m8sG8MNahbe2VT9oWb5cknbjb97+LuOKLW38VL47vJ7u8sG8MtbKLW2RT9oWb5cljtxt+9/Eeo4roKKFoBgWdv4pXxzfz315Yv4aa3UWdvGD9oSX5dxY7cY+9/Eeorfooo6WAx/EKeJJLeEeFLjS4Jt5846lBJKpXHG0I64OfXNYP2L4nt97W/C6f7umTn+c1dtRQBxP9lfEpvveKdAT/d0aQ/zmrn/AB5oHxBl+HniAXXizTbmEadOZLeDRjG0qhCWUP5xwSMjODXq1I6LIjI6hlYYKkZBHpQBR0K4gu/Dum3NoQYJrSKSIr02lAR+lX6848P6ovw1vh4T8SSGDRWlP9hapKf3Xlsci1lborqSQpPDLjuMV6MCGUFSCCMgjvQAtecePrTUNU+J3g2z0HVP7Mv4oL+d7kW6zGOLbGvKtxgsQP17V2PiLxPpHhXTTfa5eJbR52xp1kmbsiIOWY+gFYXgrS9SvtYv/GXiS2azv9SjW3s7GTG6ys1JKo3+2zEuw7HA7UAM/wCEW8df9FGb/wAElv8A40f8Ir45/wCijP8A+CS3rtqKAOJ/4RXxx/0UaT/wS29H/CK+OP8Aoo8v/gltq7aigDif+EV8cf8ARR5f/BLbUf8ACK+Of+ijSf8Aglt67aigDif+EW8c/wDRRm/8ElvVe/8ACXjuXTbmNPiE0jPC6hP7GgXcSDxkHj6131FAHJ/CyeG4+EvhdrfG1dMgjYDs6IFf/wAeBrrK85gvR8LfEF3a6ruj8Japctc2l9jKadPIcyQyY+7GzEsrdASQfWvQ4pY54UlgkWSNwGV0bIYeoI60APriPiERNrvge0i5uX8QJMgHXy44JjIfpg/rXUazremeHtLl1HW76GxtIhlpZmwPoPU+gHJrk/C1tfeKvFbeNdXtZrKzhga10SyuF2yLExBkuHX+Fn2qAOyjnrQB3dFFFABRRRQAUUUUAFFFFABRRRQBDeWdtqNlNZ39vHc206FJYZVDK6nqCD1FcVH4F1zw6uzwP4rksrBeU0zVbf7ZBEP7qPuWRF9txFd3XAfGzxZ/wiPwp1S5ik2Xd4v2K25wd8gIJHuE3t+FAGf4al8XfETw/DqkXjays9LuHdM6PpZjmcK5Q4eV32HK8EL0wa7fw34V0rwpp72ujwMplfzLi4lcyTXEnd5HPLMff8MV4T+yv4oeNtZ8IXxZJIz9tt43GCOiSrz0/gOP96vo6gAooooAKKKKACiiigAooooAyfEXhfSfFWniz1u0Eyo3mQyKxSSBx0eNxyrD1BrgvFS+IPh5oa3x+Ikv9nfaYrdF1XTI7qVd7hc+YrIW2gljnJwpr1OvmP8Aak8STajrmm+FNPWSZLCFr+7WNS2GIIXOOm1Axz6PQB7H/wAK5uNZkRvHfiW88QwowYWCxLa2bEcjfEnMnPZmI9q7eKKOCFIoUWOONQqIgwFA6ADsK4j4N+LP+Ew+Fmk30sm+7t4/sl0ScnzI+Mn3Zdrf8CruaACiiigAooooAKKKKACiiigAooooAKKKKAEZQ6lXAZWGCCMgiuFv/hzpGkJdal4f1nVPCkaK086abcAW3AyzGF1ZBxn7oFd3XlX7Q3iWXRvhq2kWG99R1+YWMMUYy7IeZMAcnIwn/AxQBH4H0O3+Jngyy1zWfFPiLU7O63LLp0tylvHlWKlXECqW5GeTyCOK9Q07TbLSNPhsdLtYbS0gXbHDCgVUHsBXz7+zLrV3o2s6/wCBNbRre5hc3MULn7rofLmXPQ9E6ejV9F0AFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAQXtjaalZS2eo20N3bTLtkhnjDo49CDwa8c+JGl+F/hnpul3Fhea3pEN/qMVs1tp2sTwxRxE5lkCbiBtUHAGBkivaq+UP2gb3U/HPxCnsNCt3u7Dw2kVrK8ZyguZpFUrnpuLFEx1yjehoA+iND+HvhrQtQGp2tk13qOONQv53upx9HkJK/wDAcV09cB8FPE8nib4X6ebzcNQ03On3iSAhlkiwBuzzkrtJ9ya7+gAooooAKKKKACiiigAooooAZNDFcwPDcRpLFIpV43UMrA9QQeoryH4m+HvDHw58F3viDRzqejyeaiR2mk6pNaxTSOw42K20cbm4A+7XsNfNH7TGr3viHxDZ+FtGRp4dIhF7qBTlYnkIVN57YBGP+uooA9j8O+AfCX+heIIIp9ZuJYkntr/VLqS7kCsAysvmEhTgg5UCu0ryj9nvXbq68BzeHNZjkg1Xw5cG0nglGHSM5aPI/wC+lHsler0AFFFFABRRRQAUUUUAFFFFABRRRQAVyniPwLa+LfFmkajrrefp+jBpYLDqk07cb5AeCFCjaPUnPHB6uigDyDwh4JttT8S6h4qsX+y6zpfijUE8wfduLdpCHif14Zip7GvX6jht4bcOLeGOISOZHCKF3MeSxx1J9akoAKKKKACiiigAooooAKKKKACvJvHXgqy0H4b+PddlP2rWtYtpJLm7fkqnGyFPRFAA9TjJ7Aes1HcW8N3bvBdQxzQyLteORQysPQg8EUAc34Y8FWXhbXNUvtFPkWerhJp7If6uOdcgvGOg3A/MPVR9B1FFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABWJc+FbC+8X2viK/X7RdWMBhskcfLbbj87gf324GewUY7526KAPM/Cfgi3vNZv9cvbSWz1XTvE19PZXRQqzwyNhlwfvIylgD2zkd8+mUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFeeeMvCkGleBEstBsXklm1uyu7lo03STyG7jaSVyBknuT2A7AV6HRQBm2OgWGm61f6nYQiCbUdhulThZXXOHI/vYOCe+BnpWlRRQAUUUUAFFFFABRRRQAUUUUAFeefELwpb23w98SyaPYvcanqs8M1zKke+aciZNq8DJVF4A6AA+pJ9DooAzI/D+nw+JZddt4RDf3EHkXLxjH2hQRtL+rLjAPUAkemNOiigAooooAKKKKACiiigAooooAKKKKAMjxXquoaJ4WvtR0fSpNXvYI90NlEfmlOQMcAnjOeOeK4TV/iVr3g9PDl34ssbBrXWZI4Z4LYPFcWDuMgMrMwcDkH7vSt74saN4n17wFcWPgq6NvqDyIXCzeU0sXO5Ff8AhJ47joR3rzzxp8PNS13wno9v4R8CjSZ9Nu4rm4S4mtlmuSvBUOrnf1JLOwzxjPYjvr3X3dQltp2f39D3eiuB+IeqeJLL4b3+v6XdyaDPa2LyvZy28U0ocdPnDMg/DP1rd8AX1zqXw58P3t/M09zcadBJLK5yXYoCSffNC6+X63/yDt5/8D/M6GiiigAooooAKKKKACiiigDkvFGveJbLxRo+keHdFE1vfbzc6rNG0kNptHAZVK9fdh7Zqh4R+IVzq3j3WfB2s21suo6ZEs6XVk5MNxGdvODkqRuHGT39KyPGfhjxBqXxU0/U9Q0ybxF4Shttv9lwzxqI5+fneKRlWQfUn6cc0vD3hLxBoXxwu/E0XhcQaNqlklqsVrPbqbLBQZkTeB0TJ2buvGaIbq/n/wAAJdbeX/BPYKK8b1lfHVn8P/EvibxPrN1oOq6fNLNYR2txE9u0K42KY8EHd0+Y7uR9K9A+Huu3/ib4e6NrGsQC3vbu2DyoF2gnJG4DsCAG/Ghaq/p+IPR/f+B0lFFFABRRRQAUUUUAFFFFABRRRQAUUUUAB6cda8W1Dxp4t0nwzrmoa54n0iw1+wllkttEhaCeOSFcEBgP3gJ5Gdw7cCvX9UtJNQ0i7s4bh7WS4geJZ4/vRFlIDD3Gc15b4d+GviXR/hnf+B3TRUjvFmSXWIppGllV88tEYxlsHbnfgDHXGKl31t20KVtL99TvPBnihfFHgPTPEVzGloLu2EsqlvljPRuT2yD1rfVg6hlIZSMgg8GvJX8M3vhv4A3fhzxpp0eo2NhbEO2jXuJHhUmQyHzVQKRgcDdn9K7rwFPYXPw90OXRoZoLBrKP7PFO+51TbgBj3NaSs5Stt/nczjdRjff/ACsdDRRRUlBRRRQAUUUUAFFFFABRRRQAV5LqHifxRZ3HiU654o0rQ7q0d20fT4mgmNxEoJBkQ5k5wOhXqeK9YkUvGyhipYEBh1HvXlPgz4deJvCGh61oQOkXi6rNK76088guNrjHzx7DuYckDeBz19Zd9bdvxKVtL9zq/hh4xm8efD7T9eu7dLe5m3pMkedu5WKkrnnBxnFdbXknhuLWvgh8LbiLXrC31XT9NkaRZ9Nuf3hR3JLOkiqBjIHysx56cV0Nz8U7Gys9E1G90m/g0rWp47e3vW8vAeQZUlN27b1+bHbOMYNaSs5e75fiZq6Xvef4HdUUV5x8boLu2+Gus6vY6xqVlLbWyqsVrP5aNlwCTgbicMR1x7VnKXKrmkVzOx6PRWL4MdpPAegvIxZ2023LMxySfLXmtqtJR5ZNGcXzRTCiiipKCiiigAooooA4fx1q3iHTvEGjJZ6lp2jeHpS/9o6jcTwpLGQPlCCXKnP0P4VkeAviFfat8S9b8IXt7b6vb2dut1Z6nDGI2kQ7cq4Hykjf1AHQ+tWPEPgPWp/i5YeNdONhqcNra/Zxp2oSvF5Lc/vI2COAee49efSDTfA3inTvjFc+NmbSZ4tStVtru1E8itbgFf8AVnyyJCFQfe2ZJPSiGjV/P/gBLrby/wCCenUVwejfE/8At3UdZ0yy8N6j/aOjy+Xc27SQ4Xg8l9+3nGAAST6YBNb/AIN8Xaf448L2+uaSs0cExZTHMoDxspwVOCR1HY0LXYHobtFFFABRRRQAUUUUAFFFFABRRRQAUUUUAcN8XbhpfhzrOkWtlqF5fX9m6W8dnYTThm44LIhVf+BEVa+GF1n4f6Pp01pf2t3p9hBDcR3ljNb7XCYIUyKofBB5XIHHqK6+ihaX87fhf/MHrby/W3+QUUUUAFFFFABRRRQAUUUUAFFFFAHg3jDxvq2u+KGt9R+GvirVfD9hKDbWq2Escd3Ip/1soKHcoP3U6dznoPXvBt5qWoeD9OvNbtWs76eMySWzx+W0ALEqhXAwQuB07Vt0ULSNgesrhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQByPxF1KxPgPxPpwvbf7cujXExtfNXzQnlsN2zOcZ4z0pPhOCPhH4Yzx/wAS6L+VW9b+H3hrxFqr6lqunGS7lt/s0skc8kXnRZz5bhGAZcgcH6dK6C3t4bS1itrWJIYIUCRxouFRQMAAdgBRHRPzt+F/8werXlf8bf5ElFFFABRRRQAUUUUAFFFFABRRRQAUUUUAcd8V9M1PXfhrquj6Hp0t/eX8YhRI5I0Ccg7mMjKMcY4yeRxXE+MPDXinWvhz4M0ix8MXpvNJurWa7Rrm1AQQqVYA+dgk9Rjt1weK9nooWjv5p/dsD1VvX8TgNT+EfhrxJqkmt6nHq1veXZWWWEagyeW2B8uEYqMY7Ej3p/xa07WNd8Aaj4c0DQ7vUJ76BVSdJoEijIcHDGSRW6DsprvKKTSceXoNNqXN1Oe8EG/h8K6fp+qaPd6ZPYWkNu32iSFxKyoASpjkfjI/iweeldDRRVSbk7smKsrIKKKKQwooooAKKKKACgnAJAz7DvRRQB5T4G0nxHofi/xxquo+F79INanFzaKtxaMxChhsOJsBju45xwckVofBHQNb8K+AzoviPSprC5huZJQzTRSJIHYkbSjseO+QOvGa9GooWm3ZL7gev33CiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigArG8X+IoPCfg/VNdusFLG3aUKT99+iL+LED8a2a8c+O0Gr+L5NK8A+HEJluj9v1GYg7Le3TIUufQtkgdSUFAHS/BfxnN43+GNjqF/MJtQt2a1vG7l0PBPuUKE+5Nd9Xzj+z5Bq3grXbXTtVQtpXi2wS9064UEp5yKWaM/wB1thbOeuxa+jqACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKw/FPiuz8K2MMtxFNd3d1KILKxtl3TXUp/hUe3Uk8AcmsCPQfG/iZfO8R6//AMI5auMjTdD2mVR6PcuCd3Y7FA9D3oA7umpIkgJjdXAOCVOa4ofCDwdL82p2Fzq8x+9Lqd/PcM313OR+lc3458EeA/CNhZ3dr4DstQudQv4bGK3Sc24LvkKc8gdPTvQB63RXkOk/DjTbvVIYdV+E+n6faMT5lyusecY+Dj5AATzgde9bq/DL4XPqLaemj6U16oy1sLgmQD127s0Aeg0V59efDH4XadJFHqGj6VavMcRLPcFC59AC3P4U6++Fvwx0y3E+paJplpCTtElxOY1z6ZLYoA7+qd5bww2t9cRxKsssJ8yQDltqkDJ9v6mvOPDngP4deIr3WoIfCNpGul3v2QSCZnE37tH3jngfP79KpTeE/h1/wsiHwjZ+DrG4k+zfabq5N3sEIO4BQm7LtleQMYBzR1S7/wCV/wAg6N9v+GOt+F9pBc/CLwmLiJZPLsLeVNw+6wUYI9DXZ155B8MfhZNfPp9to+lSXUI+e3S4LSJj1XdkUT/DX4V2r7brSdJhbeUxJclTuAyV5brjnFAHodFcIfhL8OBcLbnw9YCZ1LrGZG3MvqBuzioZvhh8L7a0kurjRtLit4mKSTPcFURhwQSWwD7UAeg0V5/F8L/hhPYi9h0XS5LQ9LhJyYzzj727HWo5fht8KoLWW5m0vSI4IZDFJK9yQqP/AHSd+AfbrQB6JRXleg+BPhzr+ta3YW/hKzCaVNFGJ1mZ1nEkQkDDB4HzY6mt3/hUHg6L5tPsbvTZR92Ww1K4hZfptfH6UAdvRXCv4W8Y+Hx5vhXxXJqkS8/2d4hUSh/YXCASKfTduFa/hXxfF4ja6sruym0rWrAgXum3DAvFno6sOHQ9mHX2oA6OiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigDgfEM8Oj/GPw/quskLp1zp82n2k7/ct7tpEbBPRTIi7Qe+3Heu+qpqmlWGt6ZPp2r2kV5Z3C7ZYZl3Kw/wA857Vx6eEvFnhldngvxHHd2KjEema+jTLGP7qTqfMUDoAwfFAHd15x8aIHutF8NwQ3ElrJL4js0WeLG+Mkt8wyCMjryCK0B4p8cWfy6j8PmuNvWXTNWhkU+4WTy2/CoLvxfd36xLqHww1+58iVZohNFaSCORfuuuZeGHYjkUdV6p/c7h0fo/yL8XhzXNEs7+8j8V6vrE62cogtrtINnmbcqfkjU5yPXvXi9hpl3N8NdEv1fwjZXEtzBJFqitO+pNdGQFg22Mszlshl5GM9hXsP/Cf6t/0TzxP+Vt/8erFhvrG311tag+DurR6mxLG7W0sxJk9Tu83OT69aI6Sv6fg7g9Y29Tl7iyv9W8c+OX1Kz8KXRguBEx8QvIJILTywUMeFIVDknIwd2faneCrOzvfE/hew8Z3Frq1pD4bZ9Le6RjDM5nIZlWVQSwi8scjOOfeun1fULPxBeQ3WufB/V9QuIeI5bm1s5GUemTL09qm1jWovEFnFa658Jdav4ITmOO4t7R1Q+wMvH4Uo6JeX+TX63HLVvz/zT/Swnwei0qCfxlF4f8v+zk151gERygHlR5C/7IOQMcYrJ8SadHbfFHxbf6Rp1v8A2tD4V+020yQL5izlpV3qcZ3kADPUjit/T/FU2kxvHpfwt12ySQhnW2gtIwxACgkLKM8AD6ACn/8ACZXUd8+of8Ky8QLdNEI3ufKtBIYwSQpbzs7QSTjpRJXSXZW/8lt/wQi7X83f8b/8A4rUNP8ADFn8LvCF94Uishrsl3YmyubcL9pmmZ183cw+ZsgvuB/GtHS/A+k+K9T+JRvLKGa/m1CW0t7iVAzQZt0wUz905bkjk4FXtK1TS7e/Gu6N8IdTjurlfMW9t7WyDOGGdwYS9weo65rXtvGN3ZyTyWnwx8QQPcyebO0UNoplfAG5iJuTgAZPpTl71/O/yvb/AC/ESvG3lb52v/meWWnia4kudO+JV55nleHxa6PdLzj5oHE5I9pZIx+FdAbBbPRPAuhXWk2F/rV+LjVCdXnKWkcrjfIXQA+Y48zCjHGM11J16NtOn09vhJrBsrmQyz2xtLPy5XJyWZfNwxJAOTzkU3Wdcj8RWkdtrvwn1vUIIm3Rx3NvaOEPqMzcU276vv8A8P8AjqFktv67fhoeczFT4C+KNmsum7FvrPKaOpjtkY+WG8sZOORgnuQa7W88OeHNO+LHh3RL3TbGDQ10qeaytpYlEM15vUMzA8PJ5eOTk96trfWKNMU+Duqjz4UglUWdlteNCCqkebggEDA7YHpVjWfEA8Q2QtNd+FOuahbq25Y7mC0cKfUZm4NLaz/ra3/BDo1/W9/+ARfCmDSLbxX47h8OCFdPTU4hGsGPLVvJG4LjjAbcMDgV6ZXnmneJpNIR00j4V65Yq4UOLa3s4w20bVztlGcAYHoKuf8ACXeMrrjTvh3dRg9JNR1S3hUfUIXb9KfRLyX4IDt64GSeHV/jnZtoxDnRtMnh1a4T7uZWQwwE/wB4FWfHYH3qR/D/AI58RjZ4j8Q22h2TfftNARjM49Dcycj/AIAgPvXUaD4e0vwxpKabodnHaWyEttXkux6szHlmPckkmkBpUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABXnPx38Vjwr8J9SMb7bvUh9gt8HnMgO8/ggY59cV6NXkfxR8C3XjnXJ77WWMXh3w7ps0sEAOGu7opuJP8A0zUBB65DDoTQBN+zr4q/4SL4U21nO+660ZzZvk8mMcxn6bTt/wCAGvVq8Y+E3gO58FalpGtaKWl0TxHo1ub+AnLW10Iw6ye6HLj2L+nT1m+1zSdMnSHUtUsrSWT7kdxcJGzfQE80AXqKjt7iG7t1ntZo54XGVkjYMrfQipKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKxvGJ2+BtdPpptwf/ACE1bNYfjc7fh/4hPppdyf8AyE1ACeBht+HnhwemlWw/8hLXmN1a+O9M8U+JtR8Cw6H4r0u+vnF5aXny3EMiqFaHJKggAAAZIx29fUfBg2+A9AHpptuP/IS1U/4V/oCXt5d2sd7Zz30zT3T2mo3EHnu3UsEcA/07Ura38v8AId9LFT4VajY6l8NtMl0vT5dMgjEkJs5W3GF1dgy57gMDj2rsKq6Zpllo+mw2Gl26W1rAu2OKMcKOv4nPOe9WqqTu7kpWQUUUUhhRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABXL/ErWb7w/wDDXW9V0ibyL21ti8MpQNsbIGcEEHr3FdRXE/GT/kjviT/r0/8AZloAP+EP8Yf9FL1L/wAFdn/8bo/4Q/xh/wBFL1L/AMFdn/8AG67aigDif+EP8Yf9FL1L/wAFdn/8bo/4Q/xh/wBFL1L/AMFdn/8AG67aigDif+EP8Yf9FL1L/wAFdn/8bo/4Q/xh/wBFL1L/AMFdn/8AG67aigDif+EP8Yf9FL1L/wAFdn/8bo/4Q/xh/wBFL1L/AMFdn/8AG67aigDif+EP8Yf9FL1L/wAFdn/8bo/4Q/xh/wBFL1L/AMFdn/8AG67aigDif+EP8Yf9FL1L/wAFdn/8bo/4Q/xh/wBFL1L/AMFdn/8AG67aigDif+EP8Yf9FL1L/wAFdn/8bo/4Q/xh/wBFL1L/AMFdn/8AG67aigDif+EP8Yf9FL1L/wAFdn/8bo/4Q/xh/wBFL1L/AMFdn/8AG67aigDif+EP8Yf9FL1L/wAFdn/8bo/4Q/xh/wBFL1L/AMFdn/8AG67aigDif+EP8Yf9FL1L/wAFdn/8bo/4Q/xh/wBFL1L/AMFdn/8AG67asTxN4u0jwnaxyarMxnnbZbWduhkuLp/7scY5Y/oO5FAGJ/wh/jD/AKKXqX/grs//AI3R/wAIf4w/6KXqX/grs/8A43TU8XeNZEDp8NroKwyok1e2Vse4zwfanf8ACV+N/wDom9x/4ObX/GgA/wCEP8Yf9FL1L/wV2f8A8bo/4Q/xh/0UvUv/AAV2f/xuj/hK/G//AETe4/8ABza/40f8JX43/wCib3H/AIObX/GgA/4Q/wAYf9FL1L/wV2f/AMbrG8ZeFfFdt4F16e5+IeoXMMWm3DyQNptoolURMSpIjyARxkc1s/8ACV+N/wDom9x/4ObX/GszxJq/jrXPCmraTF8PJonv7Ka2WQ6xbEIXQqDjPOM0AHhvwl4sk8KaS8HxF1CCNrKEpEumWjCMbBhQTHk46c1p/wDCH+MP+il6l/4K7P8A+N1T0nX/ABxpui2Vi3w6nc21vHCWGsWw3bVAz19quf8ACV+N/wDom9x/4ObX/GgA/wCEP8Yf9FL1L/wV2f8A8bo/4Q/xh/0UvUv/AAV2f/xuj/hK/G//AETe4/8ABza/40f8JX43/wCib3H/AIObX/GgA/4Q/wAYf9FL1L/wV2f/AMbo/wCEP8Yf9FL1L/wV2f8A8bo/4Svxv/0Te4/8HNr/AI02Txl4wtozNdfDe+8iP5pPs+p20sgXvtQNlj7DrQA7/hD/ABh/0UvUv/BXZ/8Axuj/AIQ/xh/0UvUv/BXZ/wDxuug8O+J9I8Vab9t0S7W4jVtkqEFZIXHVHQ8qw9CK1qAOJ/4Q/wAYf9FL1L/wV2f/AMbo/wCEP8Yf9FL1L/wV2f8A8brtqKAOJ/4Q/wAYf9FL1L/wV2f/AMbo/wCEP8Yf9FL1L/wV2f8A8brtqKAOJ/4Q/wAYf9FL1L/wV2f/AMbo/wCEP8Yf9FL1L/wV2f8A8brtqKAOJ/4Q/wAYf9FL1L/wV2f/AMbo/wCEP8Yf9FL1L/wV2f8A8brtqKAOJ/4Q/wAYf9FL1L/wV2f/AMbo/wCEP8Yf9FL1L/wV2f8A8brtqKAOJ/4Q/wAYf9FL1L/wV2f/AMbo/wCEP8Yf9FL1L/wV2f8A8brtqKAOJ/4Q/wAYf9FL1L/wV2f/AMbo/wCEP8Yf9FL1L/wV2f8A8bron8QWaeLI/DxWX7bJZteq20bPLDhDznOcn0qxqepf2ZBDJ9ju7vzZ0h22kXmMm443sOyjqT2FH9foByv/AAh/jD/opepf+Cuz/wDjdH/CH+MP+il6l/4K7P8A+N111/eR6dptze3AYxW0TTOFGTtUEnHvxUGhaxbeINAsdXsRIttfQLPEJQAwVhkZAJ5oA5j/AIQ/xh/0UvUv/BXZ/wDxuj/hD/GH/RS9S/8ABXZ//G6n1H4iWtrrF5pulaHrWuy2BC3j6Zbo6QMRnYWd1y2CDhcnmnJ8Q7C70C01fRdK1jWILl3jaOxtN0lu6HDLIrMpUg8fhR0uHWxW/wCEP8Yf9FL1L/wV2f8A8bo/4Q/xh/0UvUv/AAV2f/xum6B8UrTxJMg07w34i8hp2t3upLNBFE6nDbmDnGD1rT8D+PNH+IGkT6hoXnrHBOYJI7hAjqwAOcAngg8GjcDO/wCEP8Yf9FL1L/wV2f8A8brK1WLxb4U1rw1LceNrrVrbUNYisZ7WfT7aNWRo5GJ3IgYHKDoa6bQPHuj+JPFOs6Bpn2hrrRmCXLugEbEkj5TnJwQQeB0rO+I//H54L/7Ga3/9FTUdE+4dbHbUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAVxPxk/wCSO+JP+vT/ANmWu2rifjJ/yR3xJ/16f+zLQB21FFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUEhVJY4A5JPasTxN4u0jwnaRy6rOxmnbZbWcCGSe5f+7HGOWP6DuRXMjw74g8fETeNi+kaG3Mfh+1m/eTjt9qlXr/1zTj1JxQBNeeN7/xHeS6V8NreK9aNjHc63cA/YrU9wpH+ucf3V4HGTWr4Z8DWHh+6k1O6nm1fXbhcXGq3uGlYf3UHSNPRVwPrXQWdnbafZxWlhbxW1tCoSOGFAqIo7ADgCpqAOV8QeM59O8QRaBoOiT65qzW/2qWGOdIUgiztDO78ZJ4A6nBrPPxNSbTtNXTtBv7jW9Rlmhj0h2SKSJojiUu5O0KvHzc5yMVFq9nrfhv4k3PifStEn12y1LT47WeG1ljSaCSNmKth2UFSGwcHIIrPnsPGtvrGheMr/SI9Qv4Irm2vNKs5UV4YJWDR7GYhXZdoDcjOeKUdlfz/AFt9+n53Q3vp/Wmv9fI0p/inb6f4d1y91jRruy1HQmiF7prSK7bZGAR0deGU56+xGKnsfiHN/wAJFa6Z4h8PXWhx39vLcWVxcTxuJFjUM4dVJ8tgpzg5rlNb8IeJvFGj+L9Xl0hrO/1lbO3stMkuIzIsMMgYs7BtgY5Y4zwBjrXT+N/Cl/4j8UeHmt4iLKG2v4LucOo8nzoNinBOTz6A0O6V+v8AwP8AP/IFZu3T+v8Agken/E24vn0+9bwtqEPh/U7hYLTVDIjFtxwjtCDuRGPRj6jI5o1f4lX2lLqN/wD8IjqEuhaZM0VzqLTxxt8pwzpEx3OoJ68Z7VT8Mz+ObDSdC8Lr4cFidOMUF7qss8b28lvGMZiUNvLOAOCo255ri9d8CeI9Y0bX7LUvCd1q/iCeWd4NZutTU2wiLExiKMv8rbcKF2DB5JFN2T021+e3/B7fgKOu/l+v9dfxO8Xxhrz/ABqXQ49PlfRG0kXIYSQ45f8A1/8Afx/Bs655xjmnr8TbyNrK91HwlqFjod9dpaQX808Yk3O2xGaDO5VJ/HHOKoiw8SDx9pmr2mizxRXnhz+z3lkljzYThi4Mi7jkdB8ua4ibwR4gm0PTpJPBF/P4g068t7nUNTvdUSZ7opIGfyAZCPm68hABxyaaSTSff/25/p/T6Tq031t/7av1/pdfQ7z4nXgvNdg0bwlfammgTNHezLcRxIFVQ2U3HLNjPygduvIrSu/HU01no0nhfQLzW5dXtftcSh1gjiiwDmSRuAfmACjJPNUtB8PatbWfjn7VZGF9YvZp7NDIhMitAijoTj5gRziuSm8I+IbbT/CVprXh/UNa0my0VLW40ux1FYfKvBj55P3ih1xxncQOalbJPy/J3/Gxb3uvP81b8Lmn4q+Jurt4AfU/D2lXNpqNrq8enX0UkkLG2cSIGTJJVwwYKGHTdnjFdHfeO9QXVBpOi+F7rVNVgto7nULdbqKNbPePlQyE4Zzg8DsM5rz+DwX4ltPhz4l0iDwx9muP7bh1OztYLiNo5Yt8bmNGJ+8oQg5Ayemava54Mnk8aX3iPUvBl/rcGtW1u4t7TUhBNYzom0xuBKisp4+YE4INPp/X8q/W/wDTF1/rv/kdHDo+mePrKHxh4TuLnw7r3zRNdJGA+9GKtDcxfdlAIxg88DBFXNL8d3GnalDonxBtI9G1KVtltexsTY3x/wCmch+43/TN8H0zV/4e+H5fDfg+GzubC106eSSSeW1tZXkSIu2cb3ZizYxk5wTnArc1TSrDWtNm0/VrSG8tJ12yQzIGVh9D/Om9xIt0V59/ZviX4efNoP2jxL4bTrpksm69s1/6YO3+tUD+Bju4ABNdX4d8TaR4q037bod4txGrbJEwVkhcdUdDyrD0IpDNaiiigAooooAKKKKACiiigAooooA8v8U+H9P8SfHPTLHV5Zvs39gyu1vFO0QuMTr8rFSCQM5xnsK5i7mudHh1TQrC9uX0zS/F2nQ2bPOztGkhRnh3E5KqTjBPevVfEHgXQPE+oRahqtrMb+CLyoLuC6lhkhXOflKMMHPfrjjpRb+A/DltodvpMWnf6Jb3a3qgzSFmnVtwkZ925myM/MTn6UQ921+n/wAlf8tAlrfz/wDkbfmZOt6z4luLXxBY33hT7HpSWN0I9T/tKOTzAEbafKA3Dd+lXfhd/wAkn8Mf9gyD/wBAFdNd2sN9ZzWl0m+CeNo5EyRuVhgjI5HBrkLX4R+DrHyRaWV/EsOPLRdYvNq46Db5uMe1EdLp+X4X/wAwlrZrz/G3+Ri2Oi311rWvat8M/FiWvmajIuo6bqNj5sP2pQFcg5V0zgdMg9q2/hprDapperRXWlWOm6hY6pNbX408fuJ5xgtKvf5sjOecjmrOrfDfw5rGrTanJDd2l5cAC4lsL6a2NwBwN4jYBvqea2tE0HTPDelpp2iWcdnaoSwjTJyT1Yk8sT6kk0R0Wva35fh5f5BLV6d/8/6v/mcV8Hf+RF1P/sM3/wD6OavOvAOuJ8OvDVhrc3FprehzsF7NeW0j7F+rI2P+A17zo+gaboFjLZ6RbfZ7eWaSd08xmy7ncxyxJ5J6dKy5fh54Wm0DTdFm0lJNP0ucXFnC0sh8qQEkHO7J5Y8EkH0pW0t5Jfcmn+d0O+rfm3+N1/wThPhToknh/wCIGqWN0c3f9h2U10x6tNI8ryH/AL6Y11XxH/4/PBf/AGM1v/6Kmrp4dD0631661qG326hdwpBNNvY70QkqNucDG48gZrmPiP8A8fngv/sZrf8A9FTVTd19/wCbZNtW/T8kjtqKKKQwooooAKKKKACiiigAooooAKKKKACuJ+Mn/JHfEn/Xp/7MtdtXE/GT/kjviT/r0/8AZloA7aiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA4O++EekX3i268Sf2z4gt9Sucgy2+olPLT+4nGVX2zU3/AArKL/ocPGH/AIOX/wAK7SSRIo2eVlRFGSzHAA+tctqXxP8ABelz+RP4is5rjp9ns2NzLn02RBjn8KAKn/Csov8AocPGH/g5f/Cj/hWUX/Q4eMP/AAcv/hTf+E91rVOPC3gbV7odrjVCunw/X58yEfRKP7I+Ims/8hTxHpugQnrDo9oZ5SPTzpuAfcJQBBqHw9srKxknvPHniuyiUfNPLrhVU98sMfnXzx4613xR4ZuXm8OfE7VPEWmeb5f2mG4uVWJjkhC+fLY4B+6xJwTgV9KWnwr8MR3SXesQXPiG9XkXOt3DXZH0RvkX8FFU/E/w2g8a+LNMfXdi+HNFj3W+lxjC3M7dWfHARQAAvUnd0HDAHkXwj8P/ABJ8fOmra54s17TvD6n5ZFvpRLdn0jy3C+rEY7DPOPZf+FZRf9Dh4w/8HL/4V2kUUcMKRQoscaKFREGAoHQAdhTqAOJ/4VlF/wBDh4w/8HL/AOFH/Csov+hw8Yf+Dl/8K7aigDif+FZRf9Dh4w/8HL/4Uf8ACsov+hw8Yf8Ag5f/AArtqKAOJ/4VlF/0OHjD/wAHL/4V5H8XPC/xH8EJLrXhrxbr+o6EBmVWvpTNaepbDfMv+0AMdxxk/SVIyh1KsAykYII4IoA+NPBXiHxb4mut+ufEvVPD2mrJ5ZvZ5rmRC/XbuB2KcH+Jl696+i9M+H1lfafFPZ+P/FWoRMOLiHXCyv7gqMU7QvhpB4Q8eXup+GvLj0LWYTHqGkMo8uOUZKypnt1Up0G8noABevfhZ4YmumvNJt7jw/fN1utFna0Y/VV+Rv8AgSmgCL/hWUX/AEOHjD/wcv8A4Uf8Kyi/6HDxh/4OX/wpn9mfETQv+Qbrmm+JLYdIdWgNtOB6CWIFSfcpS/8ACxrrS+PGHhDWtHA+9c28QvrdR6l4csB9UFADv+FZRf8AQ4eMP/By/wDhTdD+FOkaD4wXxLb6rrlxqOwpK11fmRbhSpAEgxlwM5GTwQPQVu6H4z8N+JQP7C1yxvX7xRTDzF+qH5h+IrboAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAK4n4j/8fngv/sZrf/0VNXbVxPxH/wCPzwX/ANjNb/8AoqagDtqKKKACiiigAooooAKKKKACiiigAooooAK4n4yf8kd8Sf8AXp/7MtdtXE/GT/kjviT/AK9P/ZloA7aiiigAJwMngVx178T9BjvZLHQ0vfEd9EcSQaLbm4EZ9Gk4jX8Wql4qhl8XfEC28GXFzLb6NFp39pajHBIUe9DSGNISw5CZVi2DzwK7XT9OstJsY7LS7SGztYhhIYIwiKPYDigDkf7d+Ieocaf4N0/TFPSTVdWDH8UhRv8A0Kl/4um3bwfH+N0/+FdtRQBxOfimn8Hg+f8A4HdRf0aj+1fiXb/67wvoF5/17azJHn/vuGu2ooA4n/hK/HEf+v8AhxK/r9m1q3f/AND2Uf8ACd69F/x9fDnxCvr5MtrL/Kau2ooA4n/hY9yOD4B8Xg+n2KE/r52KP+Fi3p+78PvFx+trbj/2tXbUUAcT/wAJz4gn4sPhzrzn/p5ntYB+spo/4SP4gT8W/gK0t89Gu9dTj6hI2/nXbUUAcR5fxRu+Gn8KaWp7xxXF06/mYx+lH/CFeKL7jW/iHqZTvHpVnBZj/voq7f8Aj1dvRQBxSfCXwpJIsusQXuuzKc+Zq9/Ndf8AjrNt/Sup03R9M0eDydI060sIv7lrAsa/koFXKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAwtc8EeGPEjFtc0GwvZT/y2kgXzB9HHzD8DWJ/wrT+zufCvirX9E2/cg+1/a4F/7Zzh/wBCK7iigDh/L+Juk8JP4f8AEcK/89UksJ2/EeYn6Cl/4T/V9P48ReAtetfWTTxHfx/X922//wAdrt6KAOMg+LngeWUQ3GvRafN3i1GKS0YfXzVWuk0/XdI1ZQ2lapZXoPQ21wkn/oJNW57eG6iMVzDHNGeqSKGB/A1zeofDTwTqjFr3wppLuerraIjH/gSgGgDqKK4n/hUXhGLmwttQ05uzWWq3MWPwEmP0o/4Vjbp/x7+LPF0A7Aa3K4H/AH3mgDtqK4n/AIVvMOnjvxgP+3+M/wA4qP8AhXFz/wBD94wx6fbYf/jNAHbUVxP/AArTd/rfGnjB/wDuLbf/AEFRR/wq3Tn/AOPjxD4ruB6Sa9cY/RhQB21FcT/wqLwk/wDx8w6pcn1m1m7b/wBq0f8ACnfAvfQ93u15OT+r0AdtRXE/8Ke8DD7mivGfWO+uEP5iQGkPwt0+3/5Auv8AibSPRbbWJXQH/dlLg/lQB29FcM1j8RfD/wC8stVsPFlsvW2v4RZ3OPRZY8oT/vIPrWl4d8eabrmoNpN3FPo+uRjMmlagoSXH95D92RePvKT+FAHT0UUUAFcT8R/+PzwX/wBjNb/+ipq7auJ+I/8Ax+eC/wDsZrf/ANFTUAdtRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABXE/GT/kjviT/AK9P/Zlrtq4n4yf8kd8Sf9en/sy0AdtRRRQBxnjXS9Ss9W0/xj4ct2u9Q0tHgurFMbr20cgui/7alQ6+pBHeuh0LxDpniPQ4dX0m6WazlUneflKEfeVgfusDwQelaVeS654QsvEfxan0jSJLiw09rVbrxPFaylIr4s37mJlHR2CMWYYJTg9aAOhf4g3mu3Elt8PNEbXFjYpJqlxN9nsUbOMLJgtLjvsBHvThY/E+f55Nb8M2h/55RabPKB/wJpQT+Qrsra2gs7WK2tIY4IIVCRxRqFVFHAAA4AqWgDhX1v4gaADLrfh/T9es15eXQZWSdB3PkS/f+ivn2rpPDvijSPFWntd6HeLcJG2yaMgpJA/dJEOGRvYitauA8c6U3hm8b4geHYdl7ZqDq9vEMDUbQfe3DvJGvzK3XCkcjigDv6KitrmG8tIbq1kWWCZFkjdejKRkEfUGotS1Sw0exe91e9t7G1TAee5lWNFycDLMQOtAFqiq9jfW+pWMV5ZSeZbzLujk2lQ69iMjoex7jkVYoAKx/EninSfClgl1rNyY/Nfy4II0Mktw/ZI0XJZvp+NaF/fW+l6bc399IIra1iaaaQ9FRQST+QrjPAejTavKPHXiWHdq2px7rKCQZGnWh5SJB2Yj5nPUk46CgByav8Q9dAl0jQdM8P2jDKNrcrTXDA9CYYiAh9i+acbL4n2/zxaz4YvT/wA8ptPnhB/4EsrY/I129FAHDRfEOfRbqO0+Iejt4eMjBItRSYT2MjdMGUAeWT2DgfWu4BDKCpBBGQR3qG9srXUrGay1C3jubWdDHLDKoZXU9QQeorifB5m8IeKZ/At1K8unfZze6FLIxLLAG2yW5J6+WSu3vtYelAHe0VHcXENpbyXF1NHBDGpZ5JGCqgHUkngCq2lazp2uWZu9Hu47y13FFnhO6NyOu1ujDtkEjORQBdooooAKKKKACiiigAooooAKKKiubmCytZLm8njt4IlLySyuFVFHUkngCgCWsnxF4n0nwrpwvNbuhCjuI4o1UvJO56IiDJZj6AVJa+ItJvtGm1ezvop9OgV2e7jOYiqDLMrdGAweRkcH0rlPAmlSeIbkePvEURa+v0J0u3kGRp1ofuBR2dxhmbrzjgCgB6a18QdeAl0Xw/p2gWjcpJrsrSTuD0Pkxfc+jPmnmx+J8Hzx634ZvD/zym06eIH/AIEsrEfka7eigDhk+IV1oVxHbfETRW0FZCETVIZvtFi7dPmkwDFntvAHvXcI6yIrxsGVhlWU5BHrUV3aW9/ZzWl9BHcW06GOWGVAyupGCCDwQa4fwoJfBfi+TwTNK8mk3MDXmhPKxLRIpAltdx6hNysvfa2O1AHfUU2WWOGJpJnWONBlnc4Cj1JrF0bxloXiLU7my0C9XUmtP+Pia2UvDGT0XzPukn0Uk+uKANyiiigAooooAKKKKACis/Vtf0nQo4m1jUbez85tkKyyANM3ZUXqzewBNY3jrxDeaRoltbaEFOtaxcJZaeJFOEdgS0rDrhEDOfoB3oATX/HlppWq/wBi6RZXOva6VDf2dYgZiB6NLIfliX3Y59AaoqPifqP7wyeGtDjbkQmOa9kX2Zt0a/kK3fCnhWw8JaMtjYBpJXPmXV3Kcy3cp+9JI3UsT+XQcVt0AcOzfE7TP3hXw3r8S8mGMS2UzeysTImfritDw9480/WtSOj31vcaLrqKWfS79QsjAdWjYZWVeDypPviuorn/ABf4Qs/FulrHKxtdQtT5un6jEP3tnL2dT1xkDK9CODQB0FY3ibwppXizTha6tC2+Nt9vcwtsmtpB0eNxyrD/APXkVU8C+IrnxH4cL6pEkOrWM8ljqMKfdSeM4Yj/AGWG1x7MK6SgDjvBut6nFq174R8VSrPq+nRrPBeBQo1C1YkLLt7MCNrjpnBHWuxriPFAEPxZ8DTRcSzfb7eTH8UXkb8H2Don5129ABXE/Ef/AI/PBf8A2M1v/wCipq7auJ+I/wDx+eC/+xmt/wD0VNQB21FFFABRRRQAUUUUAFFFFABRRRQAUUUUAFcT8ZP+SO+JP+vT/wBmWu2rifjJ/wAkd8Sf9en/ALMtAHbUUUUAFcR4Gw3jPx683/H1/a8Stnr5QtYfL/DBbFdvXAeIZj4F8fDxVKCNB1iKOz1aQdLWZDiG4b/YIYox7fKaAO/opEdZI1eNg6MMqynII9RS0AFVtRWB9Lu1vNv2doXEu7ps2nOfwqzXC/ELVptSVfA3h+XOsazGUuHQ5+w2h4lmf0JUlVHdjx0oA4CLUfi3b/B7w03gbT9Pmt/7KhBl3eZeAbBghHwmMYwBvJ9q8h8O3eta38VLWD4iKL/UAx8iz8TXMsEHmE/KrLsbAPZcBSeOeh+0LGyg03T7exs08u3tolhiQfwooAA/IVT1zw3oviaz+y+INLtdRhH3VuIg20+qnqp9xQBzn234nDpofhf/AMGk/wD8Zo+3fE//AKAfhf8A8Gk//wAZrrrCyj06xitIGlaKFdiGaQyMFHQFmyTjpkkn3qxQB4/8Tbz4hN8NNbXVdJ0GCxa3xcyWmoTPIsZYBsKYgDxnPPTNevRKiwosIURhQEC9MdsVW1XTbbWdHvNMv03215A8Eyg9UZSp/Q1ynw+1yeOB/B/iKQLr+iRrE244+2244juE9QRgN6MCD2oA7aiiigAryT4xz+K7bxZ4Rl8D2lpcasrXgthLJ8zZh/eDacLtAAOS33gox6+sSyxwQvNPIscUalnd2wqgckknoK4TwtI/jXxtN4y2MNGsoHsNFLcefuYGa5A7KxVVX1Ck96APl3XNU8Yal43srT4ry3wiacbrbVJJLW269fkUhVzjLIvTuOo+otPn+IdvpttDpegeEY7KOJVt0g1OYRrHj5QuIcYxjGK7TUtLsNYsns9Wsre+tn+9DcRLIh/AjFQ6Jodj4d01dP0mN4bOMkxQGRnWIf3V3EkL7dB2xQAaJJq8ulo3iK2s7a+3HdHZTtLHjPHzMqnOPatCiigAqjJa6i2vQ3UeoqmnLAySWP2cEySE8P5mcjA4245q9RQBlwWOrpFqaz6wkr3EjtYsLQL9jUrhVI3fvMHnJxnpVvTobq3023h1G7F5dRxqs1wsQiErActsBO3PpmrNFABRRRQBxfxFuviFbacD8OtP0u6fafNN1KfOH+4pwn4sx+lfKmr6p4u1Lx3Y2vxWlvPLM4JttWlktLb6/IhCoTgFlXp3HUfb9VNT0nT9asXs9Xsbe+tn+9DcxCRT+BoA8n8dT+PLf4R6zbx6L4btNJTTHjzp1/Kxjg2YPlr5QXGztkDFeu2SwJYW62e37OsSiLb0244x+FZtr4W0qy8Nz+H7eBxpU0Tw/ZmlZ1SNxhkUsSQuCcDOB2wOK574e6vNYI3gjxBJt1rRYwkLucfbrQcRTp6/KArejA560AdzRRRQAV538T4tRk8QeDB4cuILbWm1C4S2luIzIiqbWXeWUEZHC/Tg89K9CeRIo2klZURQWZmOAAOpJrgvDkx8c+PG8WRgnQtJhkstHc9LqRiBPcL/ALOFCKe+GNAHz98UvB3xiklkn8WPea5YqdwfT38y3X38pQNv1KD616p8J9f1+/8AAltH4B0HwtFYW52SQPqU6zRyd/NHk/ePXPOe3Fe2VQh0PS7fV5dUtrC3gvp12zXEUYR5h/tkfex2znHagDmVvfiduG7Q/C+M8/8AE0n/APjNdTqtvfXekzwaVfLp9664iumgEwiPrsJAb8xVyigDOubPVJLvTXttUSGG3JN5EbUN9rG3AAbP7vB54z6U+3tdRj1m8uLjUVmsZVQW9mLcKYCAdxMmcvu44IGMVeooAK8O+K2q/Gy0NwPDmnWsek5O2bRwbi5C+jbxuz7onHrXuNFAHyH8FLrU7/x7d3c0Wnar4rQEQf8ACQ6hMkqgA7jGPLbLjnOTkDOBjJr2pp/E8/xh8IDxfY6XaRrb6g1p9gupJg0nlxg7tyLghS2PYt6V3Gs+ENB1+6hutU0yGS8t2DwXiAxzxEdCsi4YfnWZ4+0TUL3SbHVdBTz9a0G5W9tIy203AAKywk9t8bMPrigDraKyvDfiPTvFWhQarpEu+GUYZGGHhcfejdf4WU8EVq0AFFFY3irxPY+EtCk1LUCznIjt7aPmS5lPCxIvdieP16CgDB8H4X4lePkh/wBR9qs2IHTzTarv/HATNdvXL+ANBvdF8Pyz63tOsardSahqGxsqksmMID6IgVP+A0zxv4mutKitdG8PIk/iPVyYrGJuVhUffnk9EQc+5wB1oAz9OceJ/i/ealF+807w1atp0Un8LXkpVptv+6iop92Iru6yPCvhy28KeGrTSLR2lECkyTyffnkY7nkb3ZiT+Na9ABXE/Ef/AI/PBf8A2M1v/wCipq7auJ+I/wDx+eC/+xmt/wD0VNQB21FFFABRRRQAUUUUAFFFFABRRRQAUUUUAFcT8ZP+SO+JP+vT/wBmWu2rifjJ/wAkd8Sf9en/ALMtAHbUUUUAFRzwQ3VvJb3MSTQyqUkjkUMrqRggg9QakrA8XeKovC+nwmO2fUNTvpRb6fp8TAPcyntk/dUDlmPAA+goAwP+EO13wiHfwBrUEemqC39jayGkt4vXypQd8Q/2fmUegrFuPjNeadMbbUdK8PeevUw+LLQKfwfaw/EVuW/w7l8QFb74lX7a1cNhl0yF2jsLbvtWMEeYR/ekzn0FdNbeFfD1nCIbTQtMgiHRIrONV/ICgDkLLU/G/ja3J02/8P6BYNw89jdDU7kD/ZICxqcdzu+ldV4Z8JaX4Utpk05ZZbm6fzLu9upDLcXT/wB6RzyfYcAdgKy9T+F/hW/l+1WWnjRdQUfu7/SD9kmQ+uUwG+jAiqmj+Idb8OeJIPDPjiSO7jvTt0nXEQRrdsBkwyoOElxyCOHHTBGKAO6ooooAKKKKACsHxP4P0zxVFA1559te2jF7PULOTyri2Y9Sjjse6nIPcVvVwmpa7rXizX7vw/4KuVsLOwfytU1woJDHJjJggU8NIARuY8L6E0AVL7WfG3gmAf2veeHdesgMJcXd4NLuGA/vbg0bH3G36Cs62+Md9qcwttM0vw8J26GfxZalR+EYZj+VdVpfww8J6bJ9on0tNVv2x5l/qx+1zufXdJnH/AQB7Vr3XhTw7fQmG90HTLiI9Uls42H5EUAcwPBOs+K2SX4h6zDdWIIddF0pWitHPUea5O+YdODtXjpXdxRRwwpFCixxooVEQYCgdAB2FcHdfD+78Mq998Mr19NlQFm0a4kaSxuu+3axzCx/vIQPUV0HhDxTD4r0Y3H2d7G/tpDb39hK2XtJ1+8hPcdw3Qgg+1AG9RRRQAUUUUAFFFFABRRRQAUUUUAFFFFABWH4m8I6Z4rt4V1ATQ3VqxktL61kMVxav/eRxyPccg9wa3K4bVtf1nxN4huvDfgidLOKxYJqmtvGJBbuRnyYVPDy4wSTwmecnigCne6p438E24Op33h/X7FeEnvroaZcn/eJDRsfcbfpWZb/ABlvNRmFtp2leHvPboZvFlqVH4JuY/lXVaX8L/Cuny/abvThrOoMP3l/q5+1zOfXL5C/RQB7VsXPhXw9eQmG70HTJ4j1SWzjYfkRQBzH/CG654t2SfEDWoJtOOGGjaOGjtpe482UnfKPb5V9jXdQQRW1vHBbRJFDEoRI41CqigYAAHQCuFuvh5P4dD33wzvm0e4XLHSpnaSwuj12mMn90T03JjHoa3fB/iuPxVpcjy2r6fqdnIbfUdOlYF7WYdVyPvKeqsOGBB9qAOgooooAKKKKACiiigAooooAKKKr6hqFrpWm3F/qM6W9rbRtLNK5wEUDJJoA5fWPAZfWJdd8I6pJ4f1mYg3DxxiS3vMf89oSQGPbcCGGetYGo/EjxF4Vk+y+JrLwvcyqcebbeIY7Qt/2ynAI/wC+jVu0sdd+JUa6hrNzeaD4amw1rpdq5iubyPs88g+ZAw58tSDg8muk0vwF4T0WER6Z4d02Djl/sys7f7zkFm/EmgDkdN+IPibxa/2fwxaeF7Rzx51xrqXrL64igHP/AH0K6HQvAiWmsJrviXUpvEGuoCIrm4QJFag9RBCPljyOCeWPc1b1X4f+EtbjK6l4d06Vu0q26xyL/uuuGX8CK5DXj4s+FmmT6hoUkvifw9FG7SWeoTE3Gn8cOsuC0kQ/iU5YADBIzgA7Dxd4utvC1nCiQvf6tesYtP02A/vbqT0/2VHVmPCj8BVbwd4UuNKludb8RXCX3iTUlH2u4UfJAg5WCEH7sa/mx5PtF4K8MR22fE2qahHreuanErSaig/dJEfmWKAfwxDP1bqfbr6ACiiigArifiP/AMfngv8A7Ga3/wDRU1dtXE/Ef/j88F/9jNb/APoqagDtqKKKACiiigAooooAKKKKACiiigAooooAK4n4yf8AJHfEn/Xp/wCzLXbVxPxk/wCSO+JP+vT/ANmWgDtqKKKACuF0GMa98WvEWsXPzx6GkWk2II4jZkWWdh/tEui59F967quG8LP/AGT8T/F+iz/K1+8OsWuf+WkbRrFJj/deMD/gQoA7miiigArn/HPh9fE3gvUdOGVuDEZbSVfvRTp80bg9iGA/WugrH8Xa5D4a8H6pq9wfltbZnUYzvfGEUe5YgD3NADfB2tnxL4J0bWXAEl9ZRTSADADlRuA/HNbVeZeFPg9pVl4P0m31K51qO9jtIxcJDrFxGiybRuAVXAABzwK1/wDhVHh//n91/wD8Ht3/APHKAOs1KK9m06ZNLuY7a725hklj8xA3YMuQSvrgg+hFeSRfH5PDevSeH/ihoc2iajCQGurTM1vKvaRR94Ke2N3vzkDpdS+GeiWOnzXFufEt7Mi/u7eDXbndI3ZQTIAPqSAO9eaWf7Oes+KtUOp+PNbksoCf3OnW9y93LDH/AHPOlJwfXAYE+nSgD1XXviNo7/DbWtf8KataajLa2jNF5EgYpKw2x7l6j5iOCBW74R8PQ+FfCWn6PBybaECWQ9ZZTy7k9yzEn8a4PWfgt4Z0P4b63aeEtGT+1XsnMNzKTLPI6/MFDN03FQMLgc16H4e1q28R+G9P1ixYNBe26TLjtkcg+4OQfcUAaVFFFABXA61JB4X+MGkap5sdta+IrWWxvi7BUMsK+ZDIT/e2+YuT2wK76vPfF1hY+LPin4b0G9s7fULbTLe41K+huIlkjAZfKhVgQQcsWbH+xmgCl8QfiDrHg9k1rw/PpHiHR1/4/dONwqXMPo0TKeV9QVYj6E7WeFf2hvAviTZFd3r6JdN1j1EBEz7SDK4+pH0qn4/+Fx8QqNC8FeFPDmiW0g/0zWpbCFZAP7kKqu7Pqxx6A5yQnhL9m3wZ4f2Taws2v3a8k3XyQg+0a9foxagD1u3uIbu3Se1mjmhkG5JI2DKw9QRwakqCzsrXTrOO00+2htbaIbY4YIwiIPQKOBU9ABRRVG+0Wx1K/sLy8iZ59PkaS2YSuoRiMEkAgNx2YEUAXqKoQaLYW2tXerQxMt7eIkc8hlchlT7oCk7R17AZ70uj6NY6Dpq2GlRNDbK7OEaV5DlmLH5mJPUnvQBeooooAK818YfFO/8AhxrSJ4u0CW40K4bFvrGmtna3/POSJvusBk53cjkDqBu6j8LvB+ralPf6hpLS3Nw5klf7XMu5j1OA4A/AV5R47+E03ifVBoXgDwxFpenxEG71y/uZdrsDzHEjElgODvAIJ4BAGSAegXnxm8JXngbV9X8N65aXF5aWUs0VrIfLlLhTtHlthiN2MkAium8E+Hk8L+DdO0wZadIg91I3LSzt80jk9yWJrx4fsv6LpngzVGku7rV9dNnIbVs+TEswUlMIDk84HzEj2r2bwhrFrr/g3SdTsFRILi1RhGi4EZAwyY7bSCuPagDZooooAK4XW4xoPxd8P6tbfJHr0Uul3wA4d0RpYHP+0Nsi59G9q7qvOvHml2PjH4h+F/DN5Cbm3tVn1S+RXZdkYQxR5III3O//AI4aAOq8VXev6fpDXvheyt9TuIPmewmcxtcL3CSdFb0yCD04rhfDf7Q/gzWJ/setPceHdQVjHJBqMeFVxwRvHAwf722pfFXw18KaPozyaH4HfWtTk+S2tFuZQrN6u7PtRR1ySM9BXn/hz9l2e/vG1Dxxqsdv5rbzYaUvC5/h8xhxjpgA/XvQB9FWl7a6hapc2FzDdW8gyksMgdGHsRwamrn/AAj4F8O+BrF7XwzpqWaSkGV9zO8h9WZiT+HT2rS1TRrHWRajUYml+x3KXUO2V02yLnaTtIyOehyD6UAXqKo/2LY/8JAdb8pvt/2b7L5vmvjy927GzO3Oe+M+9Gl6NY6KtyunRNELq4e5m3Su+6R/vH5icDjoMAdhQBeqnqo1E6bKdFa3F6ozELpSY3P91tpBAPTI6dcHoTVtKstc0ufTtUhM1pOAJIw7JuAIPVSCOQOhritT+FngPTdOmuj4YmvGjX5YIJp5JJW7KBv7+pwB1JA5oAw9P/aD0az1eXRfH+k3vhbVYG2SpKPPhB7EOgzg9c7cYPUjmtXxPrGleOtZ8J6BpWo2uo6VqF1Je3ptpVkWSK2UOI2x2MjR5B9K8wsv2ctX8W65JqviVbPwpYSN+60qxb7RLGg6KXJ2577st9B0HZxfDjw38J/GXhHV9EtpUt5biXTb67uJTIxaZMQsf4V+dduVAHz80AeyUUUUAFIyh1KsAykYII4IpaKAOG+G6/2PdeJPCan/AEfRNRzZj+5bToJkT/gJZ1HsBXc1w3w/f+1vEHi/xJFza6hqS21q3aSO2jERceoLiQZ9q7mgAooooAK4n4j/APH54L/7Ga3/APRU1dtXE/Ef/j88F/8AYzW//oqagDtqKKKACiiigAooooAKKKKACiiigAooooAK4n4yf8kd8Sf9en/sy121cT8ZP+SO+JP+vT/2ZaAO2ooooAK5jxl4WuNbWz1TQrmOy8QaU5ksbmQEowPDwyAcmNxwe44I6V09FAHHaL8R9OubtdJ8Txnw3roGGsb9gqynpmGT7sq56YOfUCux69Kpapo2m65ZNaazp9tf27dYrmJZF+uCOteMfEDSfAHgfxV4U0ldKkt49YvSlykGpXMKRQ42BtqSAD53U5x0RhQB654g8W6D4WtvO1/VLezyPkjd8ySH0RB8zH2ANcvaWOqfELXLPVtfsJtL8OadKJ7DTLobZ7yYfdmmX+BV6qh5zyewrf0LwD4V8NXBuNF0Kztrk9bgpvl/7+Nlv1roaACiiigAooooAK88ltdT+G2rXV5o9hPqnhS+lae4sbRd0+myty8kSfxxMeSg5UkkDGRXodFAGToPinQ/FFp9o0DVLa+TGWEUgLp7Mv3lPsQDWsSACScAdSa87+K+h+D7Hwdq3ifW9Csp7yzt2aKcL5Uryn5YwZEw3LFR171n+B/AfgPxj4H0jXm0iS5F5bq8kNxqNzPGkg4ddryEHDBh07UAdDrHxHslu30jwfF/wkuukYFtZOGitz03Tyj5Y1Hpncewq94N8LS+H7a7vNVuVv8AXdUlE+o3irhXYDCxoDyI0HyqPqe9bWm6Vp+jWa2mkWNvY2y/dhtoljUfgBVugAooooAKKKKACiiigAooooAKKKKACiiigArz66stU+Het3mqaFYTap4a1CUz3um2q7p7GY/emhX+NG6sg5B5HcV6DRQBj+H/ABZoXim18/QNUtr0AfOkb/vI/Z0PzKfYgVsE4GTXO654A8KeJLgXOs6FZ3FyP+XkJ5cv/fxcN+teW/DzSfAHjrxJ4q0xtJkuE0i+2WyT6lczJLBjbu2vIQfnRz06MooA9C1r4j6dBdtpPhaM+JNdIwtlYOGSE9N00o+WJc9cnPoDVzwb4Wn0OO81LW7lL7X9VcS390ikIMDCRRg8iNBwO55J61taXo+m6JZLaaNYW1hbr0itoljX8gKuUAFFFFABRRRQAUUUUAFFFFABWfr2h2PiTQrvSNWi820u49kig4I7hgexBAIPYgVoUUAee6f4uvvA/l6P8Ry/2eMiO08SBCbe5XoonP8Ayyk9SflPJBrvLS8tr+1S5sLiG5gkGUlhcOrD2I4NSSRpNE0cyLJG4wysMhh6EV5v438DeAvDvhjV/Es2gQ2z2lu8xFlcS2YlfHyqfKZfvNgfjQB6FfahZ6ZaPdaldwWdunLzXEgjRfqTwK4K/wDEmofENX0fwGZrfSZSY73xGyFECdGS2zzI55G/7q9ck4rO+F3hLwN4o8E6N4mGg291eSwjzmvZJLoxzKdr481mx8wJHtivVURY0VEUKqjAUDAAoAqaTpVloej2ul6XAtvZ2kYihjX+FR/M+p7mrlFFABRRRQAVxPxH/wCPzwX/ANjNb/8Aoqau2rifiP8A8fngv/sZrf8A9FTUAdtRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABXE/GT/kjviT/r0/8AZlrtq474t21xefCXxDBZwS3Ez2h2RRIWZvmB4A5NAHY0VxP/AAtrwx/c1n/wR3f/AMbo/wCFteGP7ms/+CO7/wDjdAHbUVxP/C2vDH9zWf8AwR3f/wAbo/4W14Y/uaz/AOCO7/8AjdAHbV8dfE+3174p/FO8vNAtHurKOWTTtNYHCz/Z42kkCH+Ikh2GP7yjvX0ZL8VvDEsLx/8AE8TepXcmiXYIz3B8rrXGal428JWHizwJFo1pqNrp+lzXQ8ldHuU2q1s6japjy5ycnGT1J7mgDvvhd4qPjD4c6XqUzZvVj+z3qt95Z4/lfcOxON2PRhXXV55Y/ETwXpt1ez2NtrEDX0omuFTQ7sK8mMF8eV94gDJ74FXf+FteGP7ms/8Agju//jdAHbUVxP8Awtrwx/c1n/wR3f8A8bo/4W14Y/uaz/4I7v8A+N0AdtRXE/8AC2vDH9zWf/BHd/8Axuj/AIW14Y/uaz/4I7v/AON0AdtRXE/8La8Mf3NZ/wDBHd//ABuj/hbXhj+5rP8A4I7v/wCN0AcJ8fzq3jHVtC+HPhiIzXl2TqF3kkJFEpKIzt2XJcn3VcAkgVR/Zh125trDUfCOqI8ckYGpWO8HEkLHY5U91DgfizehrvLX4ieC7TU7zUYbbWBeXpXz5zod2WYKMKufK4UDt0ySepNcn8N/GfhKz8D+G5NQttTOpabFKqTwaRcyYDs25d6xkMpyDjJGQD1FAHt9FcT/AMLa8Mf3NZ/8Ed3/APG6P+FteGP7ms/+CO7/APjdAHbUVxP/AAtrwx/c1n/wR3f/AMbo/wCFteGP7ms/+CO7/wDjdAHbUVxP/C2vDH9zWf8AwR3f/wAbo/4W14Y/uaz/AOCO7/8AjdAHbUVxP/C2vDH9zWf/AAR3f/xuj/hbXhj+5rP/AII7v/43QB21FcT/AMLa8Mf3NZ/8Ed3/APG6P+FteGP7ms/+CO7/APjdAHbUVxP/AAtrwx/c1n/wR3f/AMbo/wCFteGP7ms/+CO7/wDjdAHbUVxP/C2vDH9zWf8AwR3f/wAbo/4W14Y/uaz/AOCO7/8AjdAHbUVxP/C2vDH9zWf/AAR3f/xuj/hbXhj+5rP/AII7v/43QBN8VPFJ8IfDnU9Qt2IvpU+y2KL9555PlTaO5GS2PRTXzZ8JrfXvhp8RrHVNbtHttMubw6JfuTxDK6JIqv8A3cF4zk/3W9K+gL34h+C9RvLK5vrbWJ3sZDNbh9DuyqSYwHx5X3gCQD2ya5Xw3428JXa+NbLXLTUbqz1PW5JfK/si5kBQwQgbtseUYFc4OGHB9DQB7bRXDx/FfwxHEif8Tt9qgbm0S7JPuf3XWnf8La8Mf3NZ/wDBHd//ABugDtqK4n/hbXhj+5rP/gju/wD43R/wtrwx/c1n/wAEd3/8boA7aiuJ/wCFteGP7ms/+CO7/wDjdH/C2vDH9zWf/BHd/wDxugDtqK4n/hbXhj+5rP8A4I7v/wCN0f8AC2vDH9zWf/BHd/8AxugDtqK4n/hbXhj+5rP/AII7v/43R/wtrwx/c1n/AMEd3/8AG6AO2orif+FteGP7ms/+CO7/APjdH/C2vDH9zWf/AAR3f/xugDtq8W/aCuNT8Rrovw88MRm41HVpftVyinAjgQ4VnPZC/Of+mf4V2f8Awtrwx/c1n/wR3f8A8bqlD8RPBcGsXOqJbawb26RI5J20O7LbF+6gPlcLkk49STQB51+zJrN3pVxqXhDVlaPz4/7T08t92Vd3lSFSeoyq/irehr6IrxD4d+M/CVv4G8PHU7bU21DTHneGeHSLmTy98jhgHWMggq2CM46dwMdz/wALa8Mf3NZ/8Ed3/wDG6AO2orif+FteGP7ms/8Agju//jdH/C2vDH9zWf8AwR3f/wAboA7aiuJ/4W14Y/uaz/4I7v8A+N0f8La8Mf3NZ/8ABHd//G6AO2rifiP/AMfngv8A7Ga3/wDRU1H/AAtrwx/c1n/wR3f/AMbrB8Q+L9O8Xa54RstDttVllt9ehuZTLpdxCqRrFKCxZ0A6sO/egD1SiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACuJ8Xf8lO+H//AF93v/pHJXbVkap4eh1TxDoerSTvHJo8s0kcagYk8yJoyD9A2aANeiiigAooooAKKKKACiiigArifg5/ySHQP+uL/wDoxq7asjwp4eh8KeF7LRLad54rNCqySABmyxbnH1oA16KKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigArifh3/yFvHH/AGMkv/pPb121ZGheHotCutYmhneU6rqDXzhwB5bGNE2j2/dg/jQBr0UUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAcT8Hv+ST6N/wBtv/R8ldtWR4V8PQ+FfDNpo1tO88drv2ySABm3Ozdv96tegAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigDkviN4+tvh94ejvpbZr27up1trO1V9vmyH1bnAHc4rO1bxl4g8HXWiSeLbbTZrDVbtLKSSw8xWs5XHy53E+YvBBI2/So/jF8Pbnx/4YtY9LeFdR025F1bpcf6ubjDI3pn19veo7bRheGzSz+GGm6PeLKjT3d7b2bRQAEFjH5TF3bg7eF7EkdKI769/w0/4IS207fj/Vj0aiuKn1/WrX4w6d4fmubWTTbvT7i6CR2xR1KsoUMxY56noF+ldrR0T/AK3sHVr+u4UUUUAFFFFABSMwVSzHAAySe1LTZI1liaNxlXBUj2NJ3toCPO/DXj3XfiANVvfB1rp0Gl2Fw1tBLqHmM946gEkBSPLXkcnceeldD4C8YxeOPCyaqls1pMkr29zbM27ypUOGXPcdCD71514N+HD+AZNU0rVfBNv4qsZbpp7DUIY7V5VUgDy5BMykYx1BI5Ndlaad4msrKwttM07SdAs7i7llvlsFjVrOHb8mMqUdyQNxxjsP71UrW+S+/wDq/wDW49/m/u/r+u3cUVx3wz8Rat4l8OXV1rKpIIb+a3tbyOPy1vYUOFmC9s89ODjiuxpAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAHm/xQ+I8/hTWdB8P6ZNaWl7rU217+9/1VnCCAXIJAJ5OMnHFQW3jybTfiJomiR+I7LxRp2sLJG00HlGazmUZBPlfLsbpyM8dTWr8Qfh9P4p1TRNf0W7htNc0KfzbZrhSYplyCY3xyBx1GcZPHNa9qvi2/1G0fU107SbOBt80dnctdPdHBAXLxIEXJzxknAHFEPPu/u/r8Ql5dvxOgSaKR3SORHaM4dVYEqeuD6U+vM/hv8A2EvxF8bx6ZpWo6fqouIpNR+13CSIzPvZdm0kAdTye+PavTKOifdA/ia7BRRRQAUUUUAFFFFABRRRQAUUUUAc58QfE0ng7wBq2vQRLNNZwbokboXJCrn2yQTXmFhq2t6h8LYdestZ8U3Pi2a3F3EI9OuWtXcnIiCCLyihHGfxzXsus6RZa/ot3pWqQiazvIjFKmcZU+/Y981yWh+CfEnhzw/H4f0rxXCNKhBjglm03fdwxk/dEgkCEjPDFDj0NLXX5W/Efb+u3/BOr0G9utR8O6feajavZ3dxbRyT27qVMTlQWUg8jByKv1574t8ZX/w+1Pwvo9rpq6nZalItjHLLdsbkyAAc5XBzkfMW6k5p0nxA1q0+Isfg270O1kv7uyF5aywXjeUq7iGEhKAjG08gHPAwM1Tak9O7/wA/yJSaWvZf5fmegUVx/g3xvP4i8R+ItB1Gxjtb7QZ0jkeCUvHMrglWGQCDxyK7Cl0TH1a7BRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFAHI33hDU7r4h2fimPWLSP7HbSWsdq2ns2Y3YE5fzh83HXAHtXXUUUdLB1uFFFFABRRRQAUUUUAFc1448Oax4o0ZNO0bX49FRnzcO1iLnzk/uYLqAp7jnI46Zz0tFJq407GB4Q0HU/D+lzQa3rf9tXUs2/7QLRbZUUKqqixqSFAC9vWt+iiqbuSlYKKKKQwooooAKKKKACiiigAooooAKKKKACiiigAoorzn4Z/E+Lx34m8WacGTbpt7/oRX/lpb42bh6/Mhb/ALaAUAR+BtG16w+K3jDWNT0K6srDWmga2mknt2x5SlSGVJGIznjAPvivSqKKOiXYHq2+4UUUUAFFFFABRRRQAUUUUAFFFFABRRRQB578RfCHiHxP4l8M32ix6Z5GhXgvW+13ckbTNkfJhYmCjAHzZPXpxy268I+JZ/jVZeMVh0kWNtYfYDCb6TzSpYsZMeTjI3Ebc846jPHolFC0/H8Vb8gev9dnf8zzzwX4Q8R6H8RvE+v6rFpYtNeeN9lteSSSQGNSAMNCobOfUY969Dooo6Jdg6t9wooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooA4v4teJ5PC3w41C4sg7aleAWOnxxKS7zy/Ku0DqwGWA/wBmvnP4ZaV4h+F3xEtNZ1e28vTFvRo2qPG24W7SxxuA5HQBnQ55GUI9M/W9xp1rd3lrdXMKyy2jM0Bbny2IwWA9cZGeuCfU1xfg6xtdS1P4gWWoQR3FtPrjRyxSLlXU2sGQRQB3tFNjQRRLGpYhVCgsck49T3p1ABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFR3FxDaW0lxdzRwQRKWklkYKqAdSSeAK5rxH48sdF1AaRptvNrevyLmPS7LBdR/elY/LEnT5m/AGsy38Dah4muY9Q+JV3HfBGDwaFakixgPYvnmdh6t8vXC0ANfxhrXjORrX4dW6w2GSsviO+iPkjsfs8ZwZj/tHCcdTUsXwn0V4w+qanr2o3jczXUusXEZkb12xuqqPQAAAV3CIscapGoRFACqowAPQCqGvSXcXhvUpNNBN4lpK0AAyTIEO39cVMnyxbHFXaRxqfDzwPLqL6fFqd+96gy1sviK5Mi/VfNyKJfh54Hg1BLCbU7+O8k+5bv4iuRI30Xzcmud8L6f4Cj8AeDL+4EQ1ae4t3iurUA3ct6Tlw7AbiN24MDxjrWXZWfhi7+DHibVfEsVm/iHzr1ry4nC/aorpXYRKrH5lIwm0Cql7t79L/p/n+XcUfetbrb8b/wCX9WO3u/h14I09XN/qWoWwjAZzN4iuU2gnAJzLxmnv8NfBken/AG6S/wBSWz27/tDeILoR7fXd5uMe9c94e0O3174qQP4ssYr67g8J2LyR3cYcLMzMHYqeN2cjPbJrmtAt7N9T8M6LrCR/8I3H4h1eNLebHkGVGPkRsDxgZYqD3p215fX8Hb/hvuFfS/p+KueiwfDfwXc2P2221DUprTBPnx+Ibpo8DqdwlxVS3+Hvw+axu9Q0/ULyaCPMlxNbeIbhhkDqxEuM4HesPxQ/gvTdO1rTtA09LwX2s2drc2Xn/Z7BbkjK7mUYC4UFwM5OAetVdDsoh8WtW0vUbfw3ibw24ubLRof3BIlG0SBuGfHsOMVDeja8/wAI3KS1V/L87G/4f8JeA/EfhWDxDaXWsQWM0e8m5165Uxez4mIU/jWjD8OfBNxYi9g1HUJbQ9LhPENyY+uPvCXHWvL7G3D+BfhlZaXb6K1ldSzPeR352Ws12qfuxPsBy3XAbqQKf4g06e00Lx9C02iQRyQWIudP0LzfJhm89cPlkChyvUKc8AkVpJe80u/9f16dyVsmz0y2+HXgi9upraz1LULieA4lii8Q3LNH/vAS5H403/hX3gU6l/Zw1W++3f8APr/wkdz5v/fPm5qjqmh6R4W+KXgh9D023sFe2vopvssQQyosIYBsfeORnnnNecatJBL8PrLXLDT/AA3pVrd6pFPau0rT6rI5uQSTLxhupPXA4qVq0v63sH2W/wCtj0ax8KeBNQ8Xah4bt7jW/t2nxxySZ1252vv3cLibJI2nIwMcVeT4eeE7i5mtdE8Q6pa6pb8h7bX5pZbdh3MbSMPqGXFcvrlvb6f46+JF3olraxayuhQz2ckUSibe0cvmOhAzuPBJHXjNZGm6RNDZ+B7m0Pg/Ti15bNaXWnNO95dAj94jYQlty7txbgHqRRHVpen4tr9By0v/AFsk/wBT0RfFuueCnFt8Q4FutNztj8R2ER8v2+0RDJiP+0Mpz2rura6gvbWO5s547iCVQ0csThlcHoQRwRUjKroUdQysMEEZBFcLc+Bb7w5dSaj8NbuLTmdi8+i3OTY3J7lQOYWP95OOmVoA7uiuV8O+PbLV9ROjatazaFr6Ll9MvSAzj+9E4+WVevK/iBXVUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFeOeIvH3ifW9buNJs/D/irQNEhYpLqNto0st3dYOCIuNsan+/8x5GAK9jooA8y8OeJfD/hTTzaaH4H8WwK7b5pToc7STv3eRz8zt7k1r/8LOtv+hT8Yf8Agim/wrtqKAOJ/wCFnW3/AEKfjD/wRTf4Uf8ACzrb/oU/GH/gim/wrtqKAPMLTX/CdhrcmsWXw21+31KQktdx+GXWQk9TuC5ye571lav8Qfhpb62NZ17wfew6lGQ32u80PbICOhLN3Hr1FeheMPEfhnRdO+zeKJIJ1uxsj05ofPluj/dWEAl+fbHrXzN8V/AMENlaato/hFPDJ1K9W2stOM7y3l475OTEGKQqOBtGTkgcUbAewad8ePAGpa8p0nT9Ru9WuU8kG10zzJ5UXLbflyxA5OO3JrRuvFHhm90ubTbr4d+IZbKeUzS27eGX2PITkuV243E8561B8F/hFb/DvRft2pok3iC8QfaJRyLdevlIf5nufYCvUaAPMl8ReFk0BtDT4ceIF0pvvWQ8MyeUT1yVxjOec9aqaR4s8EQ3SHRPh9rUc+kSPAptfDjBrZ2UFl+UZUlWGQecH3r1ioYLO2tZJ3treKF7iTzZmjQKZXwF3NjqcKBk9gKAPNU13wlHo8+kp8NtfXTriQyy2v8AwjL+W7n+IrjGeBz7VJaeJPC9jozaRafDnxBFpztue0Xwy/lOcg5K7cE5A5PoK9MooA8n1j45eBNM1mBfEGl6pZ6laKWh+26UY5oQ4wSu7lQRxx1rHs/iD8ItRkvH0/whJdPfAi5aDRFcygnJBx78/Xmux+LPwssPiV4e2DZbaxaqTZXZHQ/8839UP6Hkdwfn74X/AA3nfVNZGseFrbXb/Q7hY7zRrm4e3mRWGVkjYN5cmcN8rcEYIPNAHu0Pi3w2up2+qweAPEYvYIPs8N0nh2TekWMBA2OmO1Q6drvhLSNVk1PS/hrr1nfS533EHhh0c568heM9/Wuk8G+KfC1/GdH0KBNGu7cfvNFmthazw/8AbLjI/wBpcj3rrKPMDif+FnW3/Qp+MP8AwRTf4Uf8LOtv+hT8Yf8Agim/wrtqKAPM/EXifw/4q04WeueBvFlzGrb4n/sKdZIX7Ojj5kb3BFY2g+PvE3h/WYNMufD/AIs8Q6HKwSK9utGljvLTnAEhxtlX/a+Vuuc17LVPVNWstFshd6lN5MJljhDbS2XdwijA55ZgKALlFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAeW+H/ABjr3xF8YeI7TQtQTRtI0KX7MjpAkk11LlgSS4KqvyngLnnrXX+BJfFcvhlf+E9gtYdWWV1P2ZgVeMH5WOCQCeent06VzGjeA9d8DeNdc1Xwl/Z9/peuSCeexvbh7d4Jck5R1jcFfmPBA7enPcaJBq0NrK+vXUM9zNMZBHbriO3XAAjUkAsBjO48kk8AYAI/CvT8ev8AW1gl8T9fw/r53NKis7W9f0nw3prX+u6hBYWq8eZM+Mn0A6sfYZNcn/b/AIu8ZfJ4S08+H9Lb/mMavDmaQesNsefcNJgexoA6fxD4p0XwrYi617UIrNGO2NWOXlb+6iDLOfYA1zH9oeNfGfGj2zeEdIb/AJfb+MPfTL/sQ/di+r5PfbWt4e8AaPoN8dTk8/VdZcYk1XUX86c+yk8Iv+ygArp6AOd8OeBtE8MTSXVnBJc6lMP3+p3shmupv96RuQPYYHtWhdaDYXviCy1i7hE11p8ciWhcZEJkxvZR/eIUDPYZA6nOlRQAUUUUAFFFFABRRRQAVnjQ7FfEX9uRRCO/a3NtLInHmx5BAb1KkcHtk+taFFAGL4j8I6H4rt0j1ywSd4jmG4UlJoG9UkXDKfoa5z7N448Gc2Ux8ZaOv/LC5dYtQhX/AGZOEm/4FtY+prvaKAOf8OeN9D8UPJb6fctDfw/8fGnXcZhuYD/tRtz+IyPeugrC8R+DdD8VJGdXsg1zDzb3kLGK4tz6pKuGX6Zx6iufx448GdCfGmjp2O2LUoV/RJ8D/dY+9AHe14n8fW1hNW8IQ22sNBZXmsQoLZYFIWRWBDk9WwT93gV6f4b8Z6H4qWRdJvP9Kg4uLKdDFcW59HibDL9cY9DWB8UvAl/40ttFuNGuLeK+0bUEvI47osscwHVSygleg5waPtRfmvzH0a8n+R12j2moWdj5Wral/aU+8nz/ALOsPHYbV4/Gr1Ynhuw1S1Oo3euSxm5v7rzhbwzNLHbIEVAiswXP3dx+UcseO9bdMkKKKKQwooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACuQ+J+oahp3glpNIvZLC5mvrS3FxEFLoslxGjEbgRnDHtXX1xPxZ/5EiL/ALC2n/8ApXFQAf8ACAav/wBFE8T/AJ23/wAZo/4QDV/+iieJ/wA7b/4zW34v8Rr4T8Mz6w9sboQyRJ5QfZnfIqZzg9N2fwrQtdV0+/lmisL62uZYDiVIZlcxn0YA8fjQByn/AAgGr/8ARRPE/wCdt/8AGaP+EA1f/oonif8AO2/+M102h6hPqujQ3l3aJZzSbt0KXCzhcMQPnXg5Azx0ziqGneKVv/HOs+HBaFG0uCCYz+ZkSeaGONuOMbfU0dbB0uZH/CAav/0UTxP+dt/8Zo/4QDV/+iieJ/ztv/jNaXivxbJoF3pmm6Zpj6rq+qu62tqJhCu1F3O7uQdqgEdieelZ2n+OdWfW59A1vw19g1v7E95ZQx3yywXiqcFRLtG05IyCvAOaNwE/4QDV/wDoonif87b/AOM0f8IBq/8A0UTxP+dt/wDGazX+IvixPFUfh1vAcf8AaMlo14qf21Ht8oMFJ3eXjOT0q1N8Tm0/4g6J4T1rRGs7nVLVZXmW6EiW8jFwsZwvzZKYzkdelC1tbr/X6A9L+X9fqWP+EA1f/oonif8AO2/+M0f8IBq//RRPE/523/xmnePPiNb+CNS0Sxawe+n1a5EW1JdnkpuVPMPByNzqMcV2lG6uGzscT/wgGr/9FE8T/nbf/GaP+EA1f/oonif87b/4zXbUUAcT/wAIBq//AEUTxP8Anbf/ABmj/hANX/6KJ4n/ADtv/jNdtRQBxP8AwgGr/wDRRPE/523/AMZo/wCEA1f/AKKJ4n/O2/8AjNdtRQBxP/CAav8A9FE8T/nbf/GaP+EA1f8A6KJ4n/O2/wDjNdtRQBxP/CAav/0UTxP+dt/8Zo/4QDV/+iieJ/ztv/jNdtRQBxP/AAgGr/8ARRPE/wCdt/8AGaP+EA1f/oonif8AO2/+M121FAHE/wDCAav/ANFE8T/nbf8Axmj/AIQDV/8Aoonif87b/wCM121FAHE/8IBq/wD0UTxP+dt/8Zo/4QDV/wDoonif87b/AOM12F5e2unWcl3f3MVrbRLukmmcIiD1JPAriW8dat4pYwfDjSRdQE4OuakrRWa+8a/fm/4CAv8AtUARap4Vn0XT5L7V/ih4gsrSIZeaeW1RR+Jh/SuRt7Hxr4quEHgfxV4li0zILazrKwRxSL/0xhEIkkz6koPeu90z4c2Y1CPVvFt7N4n1eM7o575QIbc/9MYB8ifXlveuyoA8uj+CYfWItZv/ABr4gvNWjUBbubyH8s/9Mw8beWPZTW3/AMIBq/8A0UTxP+dt/wDGa7aigDif+EA1f/oonif87b/4zR/wgGr/APRRPE/523/xmu2ooA4n/hANX/6KJ4n/ADtv/jNH/CAav/0UTxP+dt/8ZrtqKAOJ/wCEA1f/AKKJ4n/O2/8AjNH/AAgGr/8ARRPE/wCdt/8AGa7aigDif+EA1f8A6KJ4n/O2/wDjNH/CAav/ANFE8T/nbf8Axmu2ooA4n/hANX/6KJ4n/O2/+M0f8IBq/wD0UTxP+dt/8ZrtqKAOJ/4QDV/+iieJ/wA7b/4zR/wgGr/9FE8T/nbf/Ga7aigDif8AhANX/wCiieJ/ztv/AIzUXghtVsfHfinQtS1291m2sYLGa3e9Ee9DKJt4yirkfIvbtXd1xPh3/ksnjX/ry0v+VxQB21FFFAHP+JPBOh+KWjm1G1aO+g/499QtXMNzAfVJF5H0OR7Vhef448Gf8fUR8Z6On/LaBVi1GFfdOEmx/s7WPoa72igDF8OeL9D8VwO+iXyTSQnE9s4Mc0B9HjbDKfqK2q53xH4F0TxNOl5dQyWmpwj9xqdjIYLqL6SLyR7Nke1Yn2/xt4N41a1Pi/SF/wCXywjEd/Ev+3D92X6oQf8AZoA72isjw94r0TxVZtcaDqEV0IztljGVkhb+66HDIfYgVr0AFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAVxPxZ/wCRIi/7C2n/APpXFXbVxPxZ/wCRIi/7C2n/APpXFQBX+N3k/wDCo9V+1ErB5lt5hB5C/aI8/pWNf6b4d034jeB/+EIgsbeaVbgTjTwoEln5JOX2/eG7bgnvXqF5Y2mo2rW2oWsN1bsQWinjDqSCCCQeOCAfqKoaR4U0DQLm4uNE0Ww0+a5/10ltbrGz98EgdPalbf8ArpYHqjzr4ff8Jl/wqnw//wAIb/Yez/SPtH9q+dnPnvjb5f45z7Vo6FfW9l8cvF51G6t7dm0/T+XkCAna+cZr0OysbTTbRLXTrWG0t487IYIwiLk5OFHA5JNZupeDfDGs3rXmseHNJv7pgFM91YxSuQOg3MpNVfW4bpr+t0zlfG83hXXdf0DTtVu7qzuJjM+m63YXYh8iVQA0YkB+8wP3SCDiqujahrPh74oad4Yl8T/8JPY31pPKwuIo/tNhsAKlnQDcrE4+Yda7U+EPDZ0VtH/sDTBprPvNmtnGIt397YBjPv1p+ieF9C8Nxunh/R7LTRJ9/wCywLGX+pAyfxpR0YPVHJ3P/Jx1j/2LMv8A6ULXM/EPSJ9Y+J2riwH+n2fhiG+syOomhuy64+u3H416+dPszqS6gbSA3qxGFbkxDzBGTkpuxnbnnHTNH2Cz/tA3/wBkg+2NF5JufLHmGPOdm7rtzzjpmlqlG26v+N7fdcfV362/C3+R4FrmqR+OLebxrED9k/tPStNsc9gJUkm/8fcL/wAAr6FrOj8PaLFpsenxaRYJZRyiZLZbZBGsgbcHC4wG3c5655rRqtLWXe/4L9U2Tre7/rf9AooopDCiisPxN4w0nwpbxNqUryXVwdlrY2yGS4um/uxxjk/XoO5FAG2zKilnIVVGSScACuEuvGupeKbqTTPhrDFcIjGO41+5Umztz3EY/wCW7j0X5Rxk9qjXwzr3jxhcePGOm6MTui8O2k2TKOxupV+//wBc1+Xpkmu7tbW3srWK2s4I7e3hUJHFEgVUUdAAOAKAONHg/wAY4G74l6iT3xpVmP8A2nS/8If4w/6KXqX/AIK7P/43VG9/tLxj8StX0JNc1DRtN0S1gbbpsixyzzTBm3M5BO1QMbe561nW3/CR614qs/BmpeKZkTT9Na8u9Q0plimvGMzRxgtg7cKvzberULW3n/wf8gel/L9bf5m//wAIf4w/6KXqX/grs/8A43R/wh/jD/opepf+Cuz/APjdcTqviPxNYaRqmgW2vTPeaZ4jsrC31SRFMjwzbWCyAABiu7B4GeK07/VNV+Hni7UIP7a1HW7M+HbnU/K1KRZCs8LADaQowpB5UcUrq1+n/wBrzfkOzvb+t7fmdH/wh/jD/opepf8Agrs//jdH/CH+MP8Aopepf+Cuz/8AjdZvh7RPEMdjoHiSbxndTXF55c2o2t66/ZZY5F3FIkCjYwyNpB7c1xOo+LNUtrEeIdK17xRqk/8AaMYN2tqItIeJpwhRVcAkbTtDLkk85qrPm5XvsTdcvMj0GPw74llvprOL4q3b3UCq0sC6fYl4w2dpZfLyAcHGeuKn/wCEP8Yf9FL1L/wV2f8A8brm7O0Ol/GLx1qsmvXdulrpkE5NyytAm5ZSN6hclUxkAHPXrWPpPiLVLLxB4UuLTVfFl/Hql8lteXGrW6xWVyroxzCjAMvIyuBjHWlHVpd/1dhy0v5f5J/qdDrHwe1PxBqVtfa3481K/kteYYriwtnhU/3vJKeWW9yua2l8G+L1UKvxJ1EKBgAaVZcf+Q65Bk8RazonjnWW8XataNomoXq6dBauixoIl3APlSXHQYJwB9am1XxHPrlxoqTaz4h82XRoLuXTPDFt++Esgz5ksmMKp6BSR60k7pP0/FN/oN6P7/waX6nR3Xh3xLY+T9t+Kt3b+fKIYvN0+xTzJD0Rcx8sccAc1JL4V8d20Zm0/wCIUt1cR/MkF/pdsIJSP4XMaK4B9VORXnN5Nqni7wJ4Ju9W1fU4LpPE/wBhZ0aNJCBLIqu2FI81QoGRxkng1pal4l1TWvFev2xu/GMEGkXH2GyGgWXmJvVAWkmbad7En7vAx9afT+trL/MXX+t7tfod/oHjyO71RdB8UWTaD4gx8trO+YroD+KCXpIPbhh3HFddXGaJp3/CffDLTU8f6Tm8mizcRTxGKSORWKiRehjY43AjGM1R+0+J/h3xqH2rxT4aXpdIu/ULFf8Apoo/16D+8PnHOQabVnYSd1c9Boqlo+s6br+lxajot7De2cwyk0LZB9vYjuDyKu0hhRRRQAUUUUAFFFFABRRRQAUUUUAFcT4d/wCSyeNf+vLS/wCVxXbVxPh3/ksnjX/ry0v+VxQB21FFFABRRRQAUUUUAc14h8BaL4hvF1Bkm07V4xiLVdOk8i5T2LD74/2WBHtWP/a/jPwd8viKxPinSl/5iWlwhLuMestv0f6xnt92u9ooAytA8TaN4psPtmgajDewg4fy2+aM/wB11PzKfYgGtWuX1/4f6Prd/wD2pbmfSNaUfJqmmv5M30fHEg9nBFZX9v8Ai/wd8vizTv8AhIdLX/mL6RDiaMes1t19y0ZI9qAO9orN0PxDpHiXTlvtB1CC/tjwXhfO0+jDqp9jg1pUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFcT8Wf+RIi/7C2n/wDpXFXbVxPxZ/5EiL/sLaf/AOlcVAHbUUUUAFFFFABRRRQAUUUUAFFFFABRRRQB5P4m+NWltrk/hzwvq2l2t3Cxju9W1SUR29oQSCEUkGZxg8D5emTU3hnXfhl4duJdRm8aabqut3Ixc6re3qNNJ/sr2jT0RcDp1r0prCzdiz2kDMxySYwSTTWsNPQZe1tlHqY1FAHOf8LV8Bf9Dfo//gYn+NH/AAtXwF/0N+j/APgYn+NbMj6DD/rW05P94xiqk+teD7ZS1zqWhwqOpknhUD8zQBw+va18PtW1061pnxJg0LUZIBbXE1hfQ/6REDkBlcMMjJwwwRmqTN8LrW30z/hHviDb6Je6bC8Ed9a38TSSo7bnEocMr5YluR1PFdXc/Ef4aWr7H17RJW7LbFZyfwQGoP8AhYfhq5/5AXhzWdbPrZaHIF/77kVF/WjYNzChuvhVHocenv45tZn/ALTj1We8k1CIzXNwjBgXOMY4AwAOBxitS/8AFHww1PxIms3vi/S5JVsJdPaA3sflPFIwLZHXPGOvTtSagniTxVZPZJ8NdHtrSX+LxBdRsPY+VErnP/Aga+eZvhDqHiP4rX3hvww1m8Vm4Go3drG62lk5+8gLuzMR0xnJIIwACaP6/C35aB/X43/PU9q01vhpZXVgLz4kpqenaW2/T9Nu9QhMNucEA5Chn2gkLuJxVCXTvhrNoR0R/ipIdLhYPY2n9pQbLRg25SPly+D0DkgfXmvR/A/wz8OeA9DWw0uyjmmYA3F5PGGlnb1J7D0UcD65J6X+zbH/AJ87f/v0v+FH9f15+YHlV5e/DTUNcvNRvPiDbSLqOmjT7+2+3RCO6UKVDtgZDDcTxgZ7VnlPh/Kumtd/FuW6m0mdJdPkl1C3xbheMbQgDEjgswJ9Mc17L/Ztj/z52/8A36X/AAo/s2x/58rf/v0v+FGzv/XcN1b+ux51beIfhja6Nr2mR+NrAw67PcT3LNfRbkaZdrBOMAAdMg/jWNLdfDe1vLe50f4mtpLrYw2FwbO+hJuoohhN2UOGAz8y469qyfi7+zzb6wJ9d8BwR2t/y8+mrhY5/ePsje3Q+x64vwQ8L38OgTap4f0zw9qGqWly9vqFlrNs8V3ayLxtSb5toIP9wc5Bzg0f192gf1+p1SQ/DGPwquiWfxIjt4IdSGp2Mov4vNtJMk4UkZYZZj82Tz1q7qN/8PrrVLq+0v4oDRn1BVXUFsb+AC6Krt3ncp2PjgsuK3/+Ey/sz5fFfw/1TTMfeuLS1S/gA9d0OWA+qCr+meOvh7q8nlWWsaOJs48ifbBJn/ckCt+lH9fp+gf1+pX0fx98NdB0a10rS/FOjQ2lpGI4k+2IcAepzye5NXf+Fq+Av+hv0f8A8DE/xroYrLTZoxJDbWsiN0ZI1IP40/8As2x/587f/v0v+FNtt3YLTQ8n1jWfA9vqkuveBfHmiaHrEh3TxG5VrO/PpNED1/21wwyetdF4B+Lei+Nr6XRy8VtrlupaW2imE0UqjGXilX5XXkcHDDnI4Ndt/Ztj/wA+dv8A9+l/wp8VnbQPvgt4o2xjciAHH4UgJqKKKACiiigAooooAKKKKACiiigArifDv/JZPGv/AF5aX/K4rtq4nw7/AMlk8a/9eWl/yuKAO2ooooAKKKbLII42c9hQA7OOtYer+NfDGgkrrOv6bZOOqTXSK3/fOc15Ro/iPRviB4i1a1+IXiS40tra/ltLbQBeG0t5I0ONxkUqZiTnjcMY+7jBr1XRfBXhXQ0RtD0DTLU4yJYbZNx/4HjJ/OgDF/4XB4Qm40y5vtVfsunabcT5+hVMfrR/wsmeX/j08CeLpgehawSIH/v5IpH5V3FFAHD/APCbeKpv+PP4bau3p9ovbSL+cho/4Sb4gSf6r4ewp/1312Jf/QUau4ooA8c1zw14x1XUjq+meDLHQNb/AOglpviDy5H9pF8jbKPZgfwrd0fU/i3a6ekWteHdB1G4Xgzxam0G4e6+Wwz9MD2FejUUAcR/b3xGHXwPpbfTX8f+0aT/AISjx7H/AK34dB/+uGtwN/6EFruKKAOH/wCE58SQ/wDH78N9cX1+z3FrN/KWg/Ewxc3vgrxdbL3Y6X5oH/fpmruKKAOHX4w+CVYJe6rJpz9CuoWc1vg/V0A/Wuj0nxRoOuqDous2F/kZ/wBGuUkP5A1qModSrAMDwQR1rl9e8B+CNRhe513QtKQL8zXRiWF19/NXDD65oA6mivELPxhY+GfiBo2k+C/Ed94h0W+uGtryK4kN1BZttLJ5dyed3yn5MtwD04r21G3oGHcUAOooooAKKKKACuJ+LP8AyJEX/YW0/wD9K4q7auJ+LP8AyJEX/YW0/wD9K4qAO2ooooAKKKKACoLy9tdOspby/uIra2hUvJNM4VEHqSeBU9cB4jtIfEnxc0nQdbAl0m00x9TjtH/1d1cCUIN4/iCKc49WB7UASL4+1XxD/wAiD4Zn1G2b7uqak5s7Q/7SZBkkHuq496ePD3j7U/m1bxpa6Wrfet9F01eB6CWYufx2iu4AAAAGAOgFcnqfxO8I6XfNYvqy3l6vDWunQvdyKfQiJW2n64oAq/8ACsbeX/j/APFfiy8HcPrDxg/hGFFH/CoPBjcz6fd3Dd3n1S6kJ/OSkPxZ8Ow/NqNrrmnRf89rzRbpE/PZx+NdPo2v6R4ishd6FqVrqFv3ktpQ4U+hx0PsaAOZ/wCFOeAT97w5C/8A10mlb+bUf8Kc+H//AEK9n+b/APxVdtRQBxP/AApz4f8A/Qr2f5v/APFUq/B34fKc/wDCKaex/wBtC38zXa0UAcnH8LPAcX3fB+jH/eskb+Yq5D4A8HWzBrfwnocTDoU06EH/ANBpNe8eeGPDNwttrOs28N033bWPMs5/7ZIC/wClZX/C19CA3yad4ijg6m4bQrrYB652Zx+FAHX21la2SbLO2ht1/uxRhR+lT1ieH/Gfh3xTvGgavbXkkf34VbbKn+9G2GX8RW3QAEZBAOPf0rO0TQNN8O2LWuk2ywJJI00z4y88rfekdurMe5P9K0aKACiiigAooooAKow6PYW+tXGq29usV5dRrHcSRjHnBful/wC8RyATyAcVeooAKoanoWk61H5esaXZagmMbbq3SUf+PA1fooA4uX4ReCDIZbTRf7OlPPmabcy2pB/7Zsopn/Ctnh/48PG/i22A+6p1FZwP+/qMf1q/rHxI8J6HfGxvdYjkvlyDaWcb3Myn0KRBiPxxVD/hbGgRfNe2Ov2UPee40O6VB+Ow0AL/AMIh4yg4s/iPeFR0W80q2l/VVQ/rR/ZnxMtP9R4k8Pah/wBfekyxE/8AfExrotC8T6J4ntTceH9VtdQjX73kShinsy9VPsQK1KAOJ+0fFGDl9O8KXgHaK8uISfwMbD9aP+Ek8fw8T+ALabHe112M5/B41/nXbUUAcT/wnWv2/wDyEPhzr6ev2Wa1uP5Sik/4Wtotv/yGdL8Q6OP717o0+3/vpFYfrXb1i+IPGPh7wsqHX9XtbJ5P9XE75kk/3Yxlm/AGgCjpvxM8FatKIrHxRpjTE4EUlwI3J9Nr4P6V06sHUMpDKRkEHrXnep+OvBmuQldS8N6rqtoR/rZvDdxLGR6/NH0/CsrRNG+HOtXbQ+AvEd74f1EHJtdPvZbZ89ebWbgj/gGKAPWqK4Rrj4g+Fvmu4rXxlp6/ee1jFpfIPXYT5cn0BQn0rf8ADXi/R/FcMp0q4YXFu225s7hDFcWzf3ZI25X69D2JoA3KKKKACuJ8O/8AJZPGv/Xlpf8AK4rtq4nw7/yWTxr/ANeWl/yuKAO2ooooAKhu+bWT/dqaobr/AI9ZPpQB578NNK0/V/C/iC01axtr63bxDfZhuYlkU/vPQgitJvhJ4WhYvoseoaDKTnfpGozWw/75Vtv6VB8JP+QHr3/Yw33/AKMrvaAOI/4QPXbT/kD/ABD1+L/ZvUt7sf8Aj0YP60v9ifEe3/1PjHR7sDp9q0QqT9Skw/lXbUUAcT5fxTj6XPhC4Hvb3UWf/H2o+2fFCH/WaN4Wuh/0x1K4jP8A49Ca6bUtf0bRnjTWNWsbBpASi3VykRcDqRuIzUNl4s8O6ldpa6dr+l3dxJnZDBeRu7YGThQcngUbhsc//bHxIHXwhojfTW3H/tCj+2/iP/0Jmj/X+3W/+MV2kkiRRtJK6oiAszMcBQOpJpsE8N1bxz20qTQyqHjkjYMrqehBHBFAHG/2x8ST08JaGv11tz/7Qo+0/FGXlNN8J2w9JL65kI/KIfzrtqKAOJ8j4pS8Nf8AhK2H+xZ3MpH5yLR/YHxDuP8Aj48cafajuLPQwT+ckrD9K7aigDiP+Ffaldc6v4/8TXDdxayw2in8I4wR+dSW/wAJvByTrcX+lvq9wpz52rXMl4xPriRiP0rs6KAPNvifbw2lx4IgtYY4IY9dUJHGoVVH2ebgAdK9Dtf+PWP6VwHxX/4//Bf/AGHh/wCk81d/a/8AHrH9KAJqKKKACiiigArifiz/AMiRF/2FtP8A/SuKu2rifiz/AMiRF/2FtP8A/SuKgDtqKKKACiiigArmvGfhebXra1vtHulsde0qQz6ddsCVDEYaNwOsbgYYfQ9q6WuS+KeoXOm/C/W5rFzHcSwrbRyA4KGV1i3fhvz+FAHMaJPrPxhtRc6ss2ieGIiYntLS4O/VJV4c+cuD9nByBtwX5ye1ej6To2m6FYJZaLYW9jbIOIreIIv146n3p2laZbaNo9pplhGI7azhSCJB2VRgfyq3QAVx/iL4eWWpXD6t4dnbw74hVT5ep2Khd564mj+7KueoYH2NdhRQBzPgrxPca7a3Vjrdulnr+lSCDUbZM7NxGVljzyY3HzL+I7V01ef+KLu18M/Fnw/rU0y20GpWF3ZXzscKViUTRsfcYkHr82K4zxV+1J4a01Wi8K2NxrU2OJpAbeEf99DefptH1oA9zrhdd1TVfFfiafwn4WvH0+1s1U6zq8Q+eHcMrbwk8eaRyW/gBHciuK8A/Ez+1438S/EHxxo1kGUvZ6BZ3ccQiGPvS/NuZj2RiQOpGeF7v4SW2z4a6dqExD3esb9Tu5O8kszFyT9AVX6KKANrw74R0PwrbGLRNPjgd+ZbhvnmmJ6s8hyzEn1NbVFFAHP+JPBGh+KNsuoWphv4ubfUbRvJuoG7FJF5H0OR6isjwjresaZrTeDvGlyt5qkcRnsNTWIRrqUAOGO0cLImQGXPQg88mu3rzv4xalY+HtF0bxJczrBc6Rq0E0RGd8kbMI5kGOxjdsj2HtQB6JRXzt4m/aniab7H4I0YSSO2xbzVJBGgJOM7AenuzDHcV3XgvxxoenaX9o8YfEnR9U1i5G6YR30SQW47RxopAwP7xGTQB6fRWXovibQ/EizN4f1ey1MQECU2k6yeXnOM4PGcH8jWpQAUUUUAFFcv40+I3hnwDbxyeJL8wyTAmGCOJpJJMegAwPqSB714rq37S+o+IdUj0jwVZ2WipcNs/tTWp1AiH98rnauOvJfPTB6UAfSVee3NxqHxI1y803TL2fTvCunSm3vLy1YpNqMy/fijccrGp4ZhyxyAcZNZl58RfDvhv4a6mYfHtjruvJYyyLObyPfPcbDt2RqcIu7GFUcD1OSe78G6ND4e8FaRpVsBstrRFYj+NyMs31LEk/WgCxofh3R/DVgtloOnW9hbqPuwoAW92PVj7nJrSoooA5TxN8P9M12VtR09m0XX41P2fV7EbJUbtvAwJUz1RuCM9OtJ4L8SX97NeeHvFCpF4i0oL55jGEvIj9y4j/2WwQR/CwI9K6yuD8cyRaF448H+Iy4hDXcul3TZxvhlheQbvZXhDfn60Ad5RXiniv8Aad8J6OHh8OW9xrtwOA6gwQg/7zDcfwXB9ap+APidceKr/wDt/wAeeNtF0PTlyLXQba6jiZ2B4klYtvx/sE4bgkAcMAei+K9f1S61yDwj4QkSLVbiLz7y/dN6abb5xvx0aRiCEU+hJ4FaHhvwNonhhmns7drnUZebjU7xvOup27lpDz+AwPasX4VyQ6vZ694oSRZ31rV7gpMrbswQsYYVB9AqZH+8a7ygArH8Q+E9D8VWog13TobnbzHLjbLCezJIMMp9wRWxRQBwOhahq/g3xLB4X8VahJqlhqDN/Y2rTLiTcBn7LMR1cKMq5+/g9+K2PFHgew8RzR6hBLJpeu2w/wBE1a0+WaL/AGW7SJ6o2QQT061U+K9l9q+GWsXMTeXdaXD/AGlay94pYP3qsPf5SPxNdTYXQvtNtrtVKi4iWUKe24A4/WgDnPCHie+vry78PeKIY7XxFpyh5RFnyruEnC3EWf4SeCOqtwe1dXXDeP4xp3iLwf4gt/luodXj06Qj/lpBcgxsp9cNsYehWu5oAK4nw7/yWTxr/wBeWl/yuK7auJ8O/wDJZPGv/Xlpf8rigDtqKKKACobr/j1k+lTVDdf8esn0oA4j4Sf8gPXv+xhvv/Rld7XBfCT/AJAevf8AYw33/oyu9oAKKKKAM/VtI03UoGfUdPtbto42CNPArleO2RxXjvgy6j0D4MeEJ9D0zTv+Eg1e7+wWt5PbqfKZ5JMyMRhjhFPGeeBXuEieZE6ZxuUjPpXBw/DAW/w30nw1FrDpe6NOLqy1NLcDZMHZgxjLHIwxBGeRSWjfy+67v+A919/5aGfqOreItB1K68NeJdTg1mDVNHu7i1vI7QW8kTxJ8yMqkgqQ2Qevasv4X+N5/EGmaRpeizxWumeH9Jt31KSWPM103l42RKeiAqQX7kYHrXU23gTVb7VLrVvFmtwX2oNp8un2gtLMww2qSfffaXYsxwO44GKLf4arZN4XuLDUxBeaHaLYXEv2bK39ttAaNl3DbyNwOTtPrTV0n/X81vuur/qJ2f4/+2/nZ/8AAPO/+F03z6R/wksXiXTSxl3r4ZFgxYw78Y8/r5m35v7ueMV2jat4w8SeOfEekaFrNppNhpkVtJFM9iJ5WaWLdtwSBtzkk8noBUtn8O/EGn2EXh+x8XNbeGoZd0cUNqUvFi3bvJE4fhe24LuxxXS6V4Y/svxZr+tfa/N/tgW48ny8eT5SFPvZO7Oc9Bj3oe2n9bBrch+HviG68VeAdL1jUUjS7uI2EwiGFLq5QkDsDtz+NdLWF4L8Nf8ACIeErPRPtf2z7MX/AH3l+Xu3Oz/dycY3Y69q3acrN6CWiCiiikM86+K//H/4L/7Dw/8ASeau/tf+PWP6VwHxX/4//Bf/AGHh/wCk81d/a/8AHrH9KAJqKKKACiiigArifiz/AMiRF/2FtP8A/SuKu2rifiz/AMiRF/2FtP8A/SuKgDtqKKKACiiigArnfH+hT+JfAOr6VZnF1PblrYk4/eoQ8f8A48q10VFAGN4S8RW/ivwrY6xa/L9oiBliPWGQcPGw7FWBB+lbNcNqnh3WvDOvXXiHwNFFdx3z+ZqehzSCNbh8Y86F+iSnAyD8rd8HmprP4r+FZJRbazeP4dvsfNZ63GbR1+jP8rD3ViKAOzormrv4jeCrGEy3PizRlUf3b6NifoASTWLN4x1vxeps/h7p01vbyDD+INTt2ihiU/xQxMA0zYzjgLnGTQBJLKPEfxotUtSHtfCtnK1xIDx9quAFWP6rGrE+m8VP4r+Efgrxlvk1bRYY7p+Td2n7mXPqSvDH/eBrc8L+GrLwpoiadYGSUl2mnuZm3S3MrHLyO3dif6DoK2KAPNPBPwuu/AtzNpqX9vrvhufOy3v4FFxZk/3XAIdT3B246jnO7S+FNwbTwzN4VvGxqPhqdrCZCeWiBJhkA/utGVwfUH0rua5PxT4UvbrVoPEnhS6isfEFrH5R84EwXsOc+TMBzjPIYcqfWgDrKK4aD4pabYOtr44s7nwrfZ2kXyFraQ+sdwo2MPqQfatdviD4NSDzm8WaGI8Z3f2jFj/0KgDoq4P4iumualoHgyELLJqN9HeXsZ5CWdu4kYsOwZ1RB6kn0pZviWmsk2vw80yfxFdMdv2vY0NjCfV52GGx6JuJrV8I+E5NDku9U1m8/tPX9SKm9vdm1QB92KJf4Y1zwOpOSeTQBy/iv9n/AMC+Ji80Fg2i3jcifTSI1z7x8p+QB963fBXhPUtC0r+yvE8mn62LcAW+pC2Ec0qf3ZUOfmH94McjryMnsaKAIoLS3tQRawRQhuSI0C5/Kq19a6jPf2EtjqK2tvDIzXUBtxIblSMBQxI2YPORn0q9RQBQgtdRj1q7uJ9SWWwlRBb2YtwpgYfeYyZy2eOMDFLo9rqFnpqw6xqK6ldB2LXC24hBBYlRtBPQYGc84zV6igCvfWFnqdm9pqVpBd20gw8M8YdG+oPBryzxF+zr4U1C7GoeGpbjw3qUbeZFLaHfEH7Exsf0UqK9booA4w+D7nVvh9faF4jg0yTUZ7OS1N9awhVlJQqsu3aNjZ5IGQD0PYXfh5rx1/wTYy3A2X9ov2O/hY/NDcxfJIp9ORkexBrpq4vXfDOr6Z4hl8UeBjCb64VV1HS7h9kOoqowrBsfu5QOA3QjANAHaUVxNr8VvDqSLa+JXn8Mah0a21mMwDPfbKf3bj3DVp3HxD8GWsJln8WaKqDuNQiOfoA2TQB0dcJr0o8Q/Ffw/otsRJDoO/V9QIPEbsjRQIf9ol3bHovvTJvHuo+J1Nr8N9LlvPMGDrV/C8FlCP7y7gGmPsox74rofCfhW38K6ZLEtxJfX13Kbi+v5wPMupj1Y46DsFHAAAoAz/Ffwv8AB/jMO2u6JbyXLf8AL3CPKmz671wT9DkVzng34TXPw/1Zo9H1GLV/D1w2ZLDU4VMtuf78cgGD7qQoPXOa9QooA4PwA6aF4i8SeD5wsT296+p2KdBJa3Db/lHokhdT6cetd5XN+LvCbeIBaX+l3h0zXdNYvY3yruC5GGjdf4o2HBH0I6VkQ/Ek6Hi1+I2lzaBcqQv26NGmsJj6rMo+TP8AdcKR70Ad3RXOp8QfBskHnJ4s0Qx4zu/tGLj/AMerJuPinpV47Wvgq1ufFV/naE09D9nQ+slwR5aD8SfagBfirePJ4RPhyxbOp+JJBplsgPO1/wDWuR/dWPcSfp612dvBHa20VvCu2OJAiD0AGBXK+GPCt/FrMvibxfcw3mvTx+THHAD5Gnw5z5UWeSSeWc8n2ArW8UeJrHwnoUup6iWbBEcFvGN0lzK3CRIvdmPAH49BQBzvjB/7c8f+FfDdv85tLn+274j/AJZxQhliz/vSsMf7h9K7quT8DeHr7T4rzXPEZVvEGtMs14EOVtkUYjt0/wBlAcZ7sWPeusoAK4nw7/yWTxr/ANeWl/yuK7auJ8O/8lk8a/8AXlpf8rigDtqKKKACobr/AI9ZPpU1Q3X/AB6yfSgDiPhJ/wAgPXv+xhvv/Rldvd3trp9s1xf3MNrAv3pZpAij6k8VxHwk/wCQHr3/AGMN9/6MrC+JUeoX3xU8NWKwaRcWhsriW3g1tmFtJcBlB4UHc4Q8AjuTS6pD6NnqS6hZPZpdpdwNbSY2TCVSjZOBhs4PNV/+Eg0byZ5v7XsfKtm2zv8AaU2xN6Mc/KfrXg2o6a0fgfxhZyXGlG0l1vT1az0Uy/Z7WQyIJVRmVRk/KSFJwc9K7GTwV4ZX48WVmuhaetonh5pRbC2Xyi4mChymMFgrEZIzzTWtvP8A+R5hPRf13sdvq3jC00vWNBtMRzW+sNNtuxOBHEscZk3Z6EED1FbNlf2epWwuNOu4LuAnAlgkDqT9RxXz7ptrosreD9P19IP7Ji8T6tBHBNgQgAt5aEHjbuwMdO1XvGa/2PefEKPwIsdtZLpFob1LIbYopmlIcgLwreTknHOOaV9L9/8AJPUdtbf1u0e42eradqE0sVhf2t1JCcSpBMrlD7gHj8ayPFfjTTfC+ialdtNb3V5YW5uGsFuVWVwPbkj64rzCw0i5sfGXhCWyHg/TZQX+zroZmaW8g8o7lbCYK9DucgZHXJrEu7Hwrcfs36jqmrxWbeIneZrm5lC/ahe+aQVJ+9ntt/u+1N7P+v6/z08wjq15/wBf1/wD36XWtPtVtvt17bWsl0oMUc0yozk9lBPP4VheHPGk3iOw0W9t9LWK31M3AkL3iboPKYqMKQDJuK9hx3ritCsfDmq+LfGreOYLGe5t1gSIX4U+TZeQpVk3fdBYsSR3rG8BeX5Xwq8gkxY1bYW6lfmxQ9Nf62bITfKj2sa3pRuYbcanZmecZhi+0Juk7fKM5PQ9KkutUsLGaGG+vra2lnOIkmmVGkPooJ5/CvAofDmkW/7NH9txafbjVRci4W+8secri82gh+owBjAOK6q1Twtd+OPHb/EAac9xC8SQjUduY7PyQQY93QFixJXv70PS/lf9P8y/6/r+vI9RuNY0y0slvLvUbSC1Y7VnlnVUJ9NxOO1WYZoriFJreRJYnG5HRgysPUEda+e/DGmQ6p4Z8Ef2be6RLqNvY3j2+l6/bu9tcxNOQXDAYWQYHqcHpivTvhNdWk3hW9gstNj0wWep3EEtvBc+fAJAQzGFsD93luAAMcinbVr+t7f19xPb+vMp/Ff/AI//AAX/ANh4f+k81d/a/wDHrH9K4D4r/wDH/wCC/wDsPD/0nmrv7X/j1j+lIZNRRRQAUUUUAFcT8Wf+RIi/7C2n/wDpXFXbVxPxZ/5EiL/sLaf/AOlcVAHbUUUUAFFFFABRRWd4g1q28OeHNQ1m+z9nsbd53A6sFGcD3PT8aAM3xP40s/Dk9vYRW1xqms3gJtdLswGlkA6sc8Ig7uxA+tYc2g+PfFEH/E81bStBtnHFjZWK3rgHs8s3y5H+ymK0Ph/4dn0/TZNc14CXxHrQW4v5SOYsj5IF9EQYXHqCe9dfQB5tbfCvUtKmFxoni77PcD+KbRLJ1/8AHI0b8mFW5fGPiHwdz8QtNgl0vgHXNHV2ii7ZmhOXQf7Slhz2rvqbJGk0TRyorxuCrKwyGB6gigBtvcQ3drFc2k0c8EyB45Y2DK6kZBBHBBHepK8/8NI/gjxpeeEIleTSLy2fUtFiGMw4bE9spJAwCysoOMBiM8Vu6B4/8M+Jryey0vVYjf20jRT2M4MNxG6/eBjcBjg8EjIz3oA6OsnxJ4m0zwrpf27V5mRWcRwwxoXluJD92ONByzH0H8q1q8/8GW48YeIbvx1qI82FJZbPQomHywW6NseYDs8jK3PXaFHQ0APz8QPFkZZVsPCOmyfcjuYBe3jr23LkRpn0+YiqC/CK6imFzD4sdbsHd5h0SwK59dvkj+demUUAcE+oeO/CKh9Ws7XxVpUf359LhNveRL/e8gkrIB6IQfaur0DxBpfijRINW0K7S8srgZSRMj6gg8gjuDzWlXn+qWy+BviFYaxp48rSfEl0tlqduo+RbpgfJuAOzMRsY98qTzQB6BRWB4n8Z6T4OSCfxE09rYzuIxeiEvFG56K5XJXPqRj3rU0zVtO1qyW80e+tr62bpNbSrIp/EGgC3RRRQAUUUUAFFFRwXMFyrNbTRzBGKMY2DbWHUHHcelAElYPijxhp3haO3S5Se8v7xilnp1mnmT3LDrtX0HdjgDua1768g07T7i9u32W9tE00r/3VUEk/kK434daVLfW8njfXY92s66gkjD8/Y7Q8wwJ6DaQzerE56UARSaZ8QPFMBOp3ul+GLRxkWcNqt/Pj0d5P3YP+6pHvVOD4TX1hMLjS/FrQ3K9Hk0SxZfyWJT+RFelUUAcBP4o8V+C4zN420+31TR4wTJrGjxsHt1H8c1uSSFxyWQtjHSu3sL+01TT4L7TriO5tbhBJFNG2VdT0INWOvWvP9Ktx4E+JCaHZjZoPiNJbiyt1Hy2l4g3SxoP4UdMuB0BVsYzQB6BRXNWHxB8NX3iC50FtSSz1m1YJLYXg8mUE8jbu4fIwcoW4IPeuloAo61rWneHtIn1PWruO0s7dcySyHp6ADqSTwAOSa5BNU8c+L1MmiWNt4X0mT7l1q0Jmu5V/vLbghUHs7E+1MsrceOfiNe6hfjzdG8L3P2Wwt2HyS3oUGWZh3Kbgi+h3Ec16BQB5o/wju55jc3Pix3uydxkXRLALn6GIn9avbPiD4UiDL/Z3i3To+WhhgFjeKvfYATE+OuPlzXe0UAcZ/wALV8Lt4ZbV4rmZ5BObQab5RF4bkcfZ/JPPmZ4x075xzTPDvhrU9V1yPxb44SNdRRSNO0tG3xaWjdeejzEfeft0HFZ3j7SLPwx4ksPiTZ2sQnsnW21Y+WCZbSQhDJ7OhIORyVBB4wK9HoAKKKKACuJ8O/8AJZPGv/Xlpf8AK4rtq4nw7/yWTxr/ANeWl/yuKAO2ooooAKhuv+PWT6VNUN3/AMesn0oA4j4Sf8gPXv8AsYb7/wBGV12saFpXiCx+x65p1rqFtu3eVcxCRQfUZ6H3rzb4WeLNI0xdY0TXr2DSdYl1m6uVsruTyyySPlSjNgPn/ZzXrAORkcigDNh8OaJb6RHpUGkWKafGwdLQW6eUrA5DbcYyCAc9c81bNhaHUBfm1hN4IvJFz5Y8wR5zs3dduecdM1PRQBkz+FfD11bfZ7nQtNlh3yP5b2kZXdJ99sY6t3PU96m0rQNI0LTjYaNplpY2jEloLeFUViepIA5/GtCigDH0fwj4d8P3U1zoeh6fp883EkltbLGzD0yB09ulRz+CfC91qk+o3Xh3S57y4UpNPLZozSAjB3EjnI4NblFAGRqfhPw9rN5bXWraJp97cWoAglntkdowOgBI6e1WYdD0m3NqbfS7KI2e/wCzFLdF8jf9/ZgfLu74696vUUAUf7D0n+yf7K/suy/s7/nz+zp5P3t33MY+9z0681x3ijwn4k1LxFLe2EXhjUrcov2Q61Y7ptOYDkxsqksCfmwSMHvXf0UmrjOT034daDD4K0zw7rdha6zFp6cSXcCtlycs4BztySeldHp+m2Wk2MdlpdpBZ2sQxHBbxhEX6AcVZpk00dvC0s8iRRoMs7sAFHqSapu7bJPPfiv/AMf/AIL/AOw8P/Seau/tf+PWP6V5R428V6N4t8WeGNK8M3qapLp2p/a7ua1O+GBBFImDIPl3EsMAEnr0r1i1/wCPWP6UhktFFFABRRRQAVxPxZ/5EiL/ALC2n/8ApXFXbVxPxZ/5EiL/ALC2n/8ApXFQB21FFFABRRRQAVxPxhUn4T6y2CyRiGSUD/nmsyM//joau2qvqFhbarplzp99EJba6iaGaM9GRhgj8jQBOrBlDKQVIyCO9LXB+B9am0S4TwJ4om26pYJt0+5k4XU7VR8rqT1kUYDr1yM8g13lABRRVPVtXsNC0qfUtXu4rOzt13yzStgKP6n0A5JoA4X4g22pX/xH8E22gajHp2oKL+X7S9sJxHEIlDZTcuQSVHXqQe1eCfFbStU8VfEVdP0K7/4SzXbMeXe3Ol6T9nETA4VXdXYMVwQWONuMZ4IH0P4PsLvxHrd/401y1ktRfW4stLs5QVkgswSxZv7ryNhiOoAUZrsbDTbLSrRbXTLOCzt16RW8QjUfgOKAPHPhp4O+L+k2Y/4SPxPFFZeWQNPuv9Mm6cDzMjZ7YdgPSu7+EjxP8IvDXkjASwSNxjGHX5XH13A12Ned2F2vw28WXOk6oRD4b1u6a5028biO0uJDmS2c9FDNl0PA5YUAeiUUUUAFcL8Yleb4f/ZbSQR311qdhFZPt3FZjdRkED2AJruXdY42eRgiKCWZjgAepNee2Nx/wsfxtaapajd4X8Pys9rMQduo3mCvmJ6xxgsA3dicZAoAxviAmu+GvB15deLviBaTWEyGH7IfD0TtdMw4jVTJyT+nU4AzXiPw/wDhJ8Tb3UV1LQEuvC8LHIu7iZ7c7c9Ng+dhj1GD619hSWFpNexXk1rDJcwqVimdAWjB67Seme+OuBVigDE8KaXrekaGlt4l17+3b0Hm6+yrBx2G1euPU8mrt8dWF/Yf2atkbQyN9tNwziQJjjywBgnPXOOKvUUAUIDq39tXYulshpmxPsrRs5mL/wAe8EbQOmMH60ujnVjpq/8ACQLZre723CyZ2j27jtwWAOduM++avUUAcnqfirxFZ6nPbWPgPU9Qhjbalyl7aoko9QGkBA+oFeK/Ge71fTmh8ZW3h/U/Bmtblga9i1W2IvF7I8cchZiPXB4GDkAY+lq5HUfhp4e17xL/AG54mgbWrqMbbeG8O6C2T+6sQ+U+pLbiT+AAB896X8WviZ4u8A65pdxoba3YyafPHNqccBhaFdh3FnHyNgZ+XAJr6h8PXEd14Z0y4gQRxS2kTooIIClAQOParYtLdbP7IsEa22zy/JCAJtxjbjpjHauD8HagfBOpr4D16Ty4lLHQLyQ/Ld2+ciDcePNjztx3XaQKAPQqKKKACvPvilDeXupeC7LRbxbLVZNb8yC4aHzfLRLaYyHbkZGCMjIrur29tdNsZrzULiO2toELyzSsFVFHUknpXE+FVn8Z+Lm8bXcMkGl28DWmhQzKVd0YgyXRB5XfgKvfaM96APEPjlYalq2v2OgHVovFniWEZaHTdE8qW2iPOHdZGzkkHaRwDnjIz1fwp8DfGTSPIbUPESaTpikE2Oof6axX0CZ+QfR1I9K92sNKsNLWRdOs4LbzWLymKMKZGPJZj1Ykk8nnmrdAHEfCj5PDOpwS/wDH1BruoJc8cmT7Q7Z/75ZT+VdvXnuqTn4deNrrXZ1I8Ma8yf2hIoJGn3YG1Zmx0jdQqsezKCetegRypNEksLrJG6hldTkMD0IPcUAOooooA5T4pSwxfCXxQ1zjY2lzouRn52Qqv/jxFdBpEc0Oi2MV1nzkt41kycncFGf1rhtXu1+I/iiDw/pRE/h/SbpLjWLxeY7iaM7ktFPRsMAz44GAOpr0SgAooooAK4nw7/yWTxr/ANeWl/yuK7auJ8O/8lk8a/8AXlpf8rigDtqKKKACmyIJIyp6EU6igDgfE3w7s9b3fabeK4Rv4ZUDD9a4sfC250dj/YF/qmke1hfSxKf+A52/pXuVIVU9QD+FAHiItviJphxZeMtQZR/DeWsFxn8SgP61Ovij4pWox9s0S7x3n02RSf8AviUV7G1rC33o1P4VE2m2rdYl/KgDyQfEL4lRf67TPDk3+6LiP+rVIvxU8bw/6/wlpc//AFx1GRP/AEKI16i+iWT9Yl/Koz4esD/yyX8qAPNv+Fv+Kx18BW5+mtf/AGmoY/jb4hlvZrOPwLbtcQqryRjXBlQ2duf3Pfaa9Ll0DTIYnllREjRSzMwACgdSa+WPh746tr/9oie8ugF03Xp2tEVhjYpwIOPX5UX/AIEaAPYj8XvFh6eA7YfXWc/+0aib4o+Opf8AUeF9Hg/6638r/wAoxXpv/COWH/PJfypR4esB/wAsl/KgDy0+P/iZN/qrDw3D/vRXEn/s61G/iL4o3YwdU0qzz3ttMLH/AMfkP8q9aXRbJekS/lU66dbL0iX8qAPFzpnj7U/+Qh4z1g57WqQ2w/8AHEz+tNj+Eg1OZZdba81WUHO/UbqS4P5MSP0r25baFPuxr+VSBQOgAoA4vw74FttI2bIkjVOiooAH4Cu0VQqhR0FLRQAUUUUAFFFFABXE/Fn/AJEiL/sLaf8A+lcVdtXE/Fn/AJEiL/sLaf8A+lcVAHbUUUUAFFFFABRRRQBleIfDOkeKtN+xa5ZpcxK2+NslXicdHRxgqw9QRXLT6L408LWU02k+L7TUdOt4y5j8RWpd4kUZJ8+IqW4HVlJ+td9XlP7RPio+H/hbNp9sxF5rcgs41X73lnmQ49No2/8AAxQBc8O658RfGHhux1jTLjwpZ2d9CJEl8u4mkTPUFCVG4HIPJGRWvYfDuKXU4dV8Y6rc+J9Qt2DwfakWO2t2/vRwL8ob/abceOteb/sueJ5JvD+q+Er/AHJc6XMZ4Y5Bhljc4dcdtrgk+7173QAUUUUAFVtQ0+z1bT5rHU7WK7tJ12SwzIGVx6EGrNFAHCR+B9f8O/L4F8UPb2S/c0vWITeQRj+6j7lkRR6biKy9L8VfEDWfEWs6DYyeEftujNEl25+1EL5i7lIXjPQgjdwRivQdc1e20DQL/V744t7G3eeT3CqTge5xivkn4LePLyy+OAv9YkZU8TySRTs2QpeRyUI/7aAKPQE0AfRR+H2oa+6t8QPEU2sW4O7+yrOL7JZk+jqCXkA9GbHtXbQQQ2tvHb20SQwxKEjjjUKqKOAABwBUlFABRRRQAUUUUAFFFFABRRRQAVn65oOl+JNKk03XLKK9tJCCY5B0I6MCOVI7EYIrQooA4VPCnjHQAE8K+LUvLNeEs/EFubgxj0E6FXx2+bdWV4c8UeP/ABfFqB0ifwlEun30thcS7bmUrJHjdtX5QRyCDmuo+I/ilfBnw71fWwwWaCArb57zN8qcd/mIJ9ga+fP2YvFE2leOL7w3qLOi6xALiES5BMqDeDz/AHo2Y577RQB7nb/Do6leRXvjzWp/Es0LB4rSSJYLKJuxEC8MR6uWrtwMDA4FFFABRRRQAyaGO4heGeNZYpFKujqCrA8EEHqK4j/hXt7oMjSfD7xBNokLNuOl3MQurLPfahIaPP8AsMB7V3VFAHmFx4p+IFn4wsvC80vhB9SvbWS5ib/SVBRCAcrzycsQM9EbkcZ1ZPBPiLxENnjfxS0lk339M0WE2kMg/uvIWaVlPcBlr5r8YfEu4k/aEbxdZM8tppV4sEGzo1vGdjgH0fc5/wCB19k21zDeWkNzayCSGZFkjdejKRkEfgaAItN0yx0bTYdP0q0hs7SBdsUMKBVUewFWqKKACiiigArifDv/ACWTxr/15aX/ACuK7auJ8O/8lk8a/wDXlpf8rigDtqKKKACiiigAooooAKKKKACiiigDnPHukap4h8HXmiaJcrZz6li2ku2G4QQsf3jbe5KblAGOWHI6jyW/+Cnh688Z3HhfSN+nS6f4ftbiyv15lS4+0T/vXIxuLYGenQYxgY99ribb/kvWpf8AYt2v/pTPQB1elPevpFqdVjWO+8pRcKhyokx820+mc49sdKt0UUAFFFFABRRRQAUUUUAFFFFABRRRQAVxPxZ/5EiL/sLaf/6VxV21ct8RtF1LXfBz2uiQxT3sd3a3McMsvlrJ5U6SFd2DjIU84oA6miuJ/wCEj+IP/QgWP/hQr/8AGaP+Ej+IP/QgWP8A4UK//GaAO2orif8AhI/iD/0IFj/4UK//ABmj/hI/iD/0IFj/AOFCv/xmgDtqK4n/AISP4g/9CBY/+FCv/wAZo/4SP4g/9CBY/wDhQr/8ZoA7auWv/A9nrPxAs/E2st9p/suDy9NtDzHDITl5iO79AOw2g8nG2l/wkfxB/wChAsf/AAoV/wDjNH/CR/EH/oQLH/woV/8AjNAHL/DPwRaTLpnjHTnNpqkF9qNtd7ThL2BruYbXA/iU4ZT/ALIBzwV9erx34aa541t/BSR6b4MtLyD7denzX1tYjuN1KWXb5R6MSM55xnjOK6z/AISP4g/9CBY/+FCv/wAZoA7aiuJ/4SP4g/8AQgWP/hQr/wDGaP8AhI/iD/0IFj/4UK//ABmgDtqK4n/hI/iD/wBCBY/+FCv/AMZo/wCEj+IP/QgWP/hQr/8AGaANXxp4VTxpoaaJeXDw6dNOj3qxnDzRodwjB7ZYLk+gIHXI4Gf4e6H4p8beMtBmtxZRWtppRsJrRQj2Tqs21oz2xjp3FdT/AMJH8Qf+hAsf/ChX/wCM1yeh6541T4oeLJYPBlpLdyW1gJ7c62qiEBZdhD+V827J4wMY75oA9dtBcLZQi9ZHuBGolaMYUvjkj2zUtcT/AMJH8Qf+hAsf/ChX/wCM0f8ACR/EH/oQLH/woV/+M0AdtRXE/wDCR/EH/oQLH/woV/8AjNH/AAkfxB/6ECx/8KFf/jNAHbUVxP8AwkfxB/6ECx/8KFf/AIzR/wAJH8Qf+hAsf/ChX/4zQB21FcT/AMJH8Qf+hAsf/ChX/wCM0f8ACR/EH/oQLH/woV/+M0AdtRXE/wDCR/EH/oQLH/woV/8AjNH/AAkfxB/6ECx/8KFf/jNAHbUVxP8AwkfxB/6ECx/8KFf/AIzR/wAJH8Qf+hAsf/ChX/4zQBoeLvBVp41utKh1pvM0qwuPtclnji5lAwgf/YALEjuSOwOeHsvA9l4m8YeKb6Iiz1bRvEUVxp90gxsK20B8tsdY2xgiuo/4SP4g/wDQgWP/AIUK/wDxmuT8H6541i8SeMWtPBlpcSyaqrXEba2qCB/s8Q2g+Ud3ABzx1x2oA9iorif+Ej+IP/QgWP8A4UK//GaP+Ej+IP8A0IFj/wCFCv8A8ZoA7aiuJ/4SP4g/9CBY/wDhQr/8Zo/4SP4g/wDQgWP/AIUK/wDxmgDtqo63Z3OoaFe2Vhc/ZLi5haFLgZzDuGN4x3AOR7gVy/8AwkfxB/6ECx/8KFf/AIzR/wAJH8Qf+hAsf/ChX/4zQBy9v8O/DWk+O9J8Hxaek2kv4avEnSXlp2M9vmRmGPnJGcjGMDGMDHofg/w+/hXwva6GbuS8hscxW00p/eGHOUVu2VB28cYUHA6Dzi61zxqfjDps7eDLRb1dEuUS1/ttSrRmaEs+/wArgggDbjnOc8V1n/CR/EH/AKECx/8AChX/AOM0AdtRXE/8JH8Qf+hAsf8AwoV/+M0f8JH8Qf8AoQLH/wAKFf8A4zQB21FcT/wkfxB/6ECx/wDChX/4zR/wkfxB/wChAsf/AAoV/wDjNAHbVxPh3/ksnjX/AK8tL/lcUf8ACR/EH/oQLH/woV/+M0eDNN8Q/wDCY+I9f8SaVb6V/aUNnDDbw3guT+5Eu4lgq4z5g7etAHbUUUUAFFFFABRRRQAUUUUAFFFFABXE23/JetS/7Fu1/wDSmeu2ribb/kvWpf8AYt2v/pTPQB21FFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFYur+MNB0O/jsdR1FEvZV3paxI0sxX+95aAtj3xitqvFPhe0lp8eviDba+duqXMiSWnm/ektgzY2Z6jaY+np7ULWXL5N/cD0jzen4nrWia/pXiTT/ALdoV9De228oXiP3WHVSOoI9DzWjXI6LL4P8IaVqyeHbZoLS1u2a9FlbTTgTFQzfdDZwMZxwvQ4re0LWrTxFoNnrGnbza3sQliMi7W2npkdqe+3kHqcz8I/+SfJ/2EdQ/wDSyau2rifhH/yT5P8AsI6h/wClk1dtSAKKKKAMXV/GGg6Hfx2Oo6iiXsq70tYkaWYr/e8tAWx74xVnRNf0rxJp/wBu0K+hvbbeULxH7rDqpHUEeh5ryX4XtJafHr4g22vnbqlzIklp5v3pLYM2Nmeo2mPp6e1drYal4I8B6XfRaJHHawf2mLeWO3DESXkgHyBmO0noDzhcYOMUR1im+qv+P9L1B6Sa7P8AT+vkdtXE+HP+SxeNv+vTTP8A0GetvQPFeneIbvULO1We3vtNkEd5Z3KbZISwypOCQQR0IJFYnhz/AJLF42/69NM/9BnoA7aiiigAooooAKKKKACiiigAooooAKKKKACuJ8Cf8jb48/7DSf8ApLDXbVxHgfd/wlXj3Z97+2Vx9fskNJuyuBral478NaTqE1jeaon2q3TfPDBG8zQr6uEB2D/exWvpmqWOtabDqGk3cV5ZzruimhbcrD2NeNfACCKXR/GOjeJEVtYfU5f7ShuPvyRsoGW7lSd/Pv713Nv4i8I+B/Atq+gQOdJW7+xWlvaIzGWUyEHYXPzDO47s4wCcmqS0Xnb8f6QPd+Tf4HbUVFdXKWdq88yysiDJEMTSsfoqgsfwFZ3hrxLp/ivSTqOkGVrcTSQZljMbbkYq3ynkcjvSA1qKKKAOJvf+S76T/wBi9d/+lEFdNrGv6R4etln1zUrWwidtqNcShN59FB6n2Fcze/8AJd9J/wCxeu//AEogrhGkluv2wxDrfzQ2+mZ0tJPug+WCSo9c+Zz7e1C1ko+v4K4bRb/rex67o3ifRPELTLouqW148GPNjikBePPTcvUZ9xWpXM2Hhnw1pHxAvtXtAsev6vb7plMxJeNCoLBCeBnbk1vWeoWeoCY2N1DciCVoZTE4bZIv3lOOhHcUdEHUsUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFcTbf8l61L/sW7X/0pnrtq4m2/5L1qX/Yt2v8A6Uz0AdtRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFcX8SPF+seFtN05PDGlQ6rqupXy2dvbzSiNclHfOSQP4MYyOvWgDtKzNW8N6HrxjOt6PY6i0X+ra6tkkKfQsDivJP+Ew+LUnGr6PdaS/cafoAvVH/AAL7Vz+VH/CT+K/+Yh4k8VWY7lPAxUD8fnoA9R1yC00jwZe2mn2LRQ/ZpIYbawtGfBZSAAkanAz7YrL+FIlt/hjoljd2t3aXVnaRxTw3VrJCysB0+dRn6jIrgv8AhJbN/wDkKfFLxZAe4/4R/wCz/wDttij+2/B7f6/4weK1+pWP/wBthQuvnb8L/wCYPp5X/G3+R3Hwj/5J8n/YR1D/ANLJq7bNeE6dL8M9NtRaWXxa1y3h3u/ljVBGNzMWY4EY6sSfxrQSf4cS/e+LesN/veJnT+RFAHs2RRketeQpafDCb/WfEy6l/wCuvi9xn/yIKnTQPhTN97xeLj/e8WSnP5TUAeiat4c0LXmjOt6RYag0X+ra6t0kKfQsDis7xJdWPhPwi0lh4dk1BLYj7Lpum2PmHfnK4RFO0A87scfWuSHhD4TP/wAxeF/r4nuD/wC16ePBPwnbpeW7fXxFOf8A2tS6WQ+upP8ACuT7XqGvanc6LrFjfXzxS3d3q9mbVriTDAJHGScRooUDknk5rQ8OEf8AC4vG3I/49NM/9BnrMHgH4Ut0a0b663Mf/atOT4b/AAp3s62tgzNjc39pyEnHTP7yqZJ6PuHqPzo3L6j8689Hw0+FzdNO05vrfOf/AGenj4W/DBumiaY31nJ/9mpDO/LqOrAfjTfOj/56L/31XDL8J/hofu+G9Kb8c/1p/wDwqb4b/wDQsaT/AN8//XoA7bz4v+eqf99Ck+0Q/wDPVP8AvoVxf/Cpvhx/0LGk/wDfH/16X/hU3w5/6FfSf+/Y/wAaAOz+0wf89o/++hR9oh/57R/99CuN/wCFTfDr/oVtJ/79j/Gj/hU3w5/6FbSf+/Y/xoA7Lz4f+eqf99Cl86L/AJ6J/wB9CuL/AOFTfDj/AKFfSf8Avgf40n/Cpvhv/wBCxpP/AHz/APXoA7YSIejr+dLuX1H51wx+E3w1xz4Z0kfh/wDXqM/Cr4Yr10DS1+khH/s1AHe7h6j864nwKR/wlvjzkf8AIaT/ANJYaqH4Y/C9eulaav0u2H/s9R/8K1+FcZJFlp8ZY5JGouuf/IlAHXan4V8Oa1drdaxoemX1wowJrm1jkfHpkjOPauZ8Xabdv4x8Nta6RNc6XYRSmFbVUKR3J2CIyKSu1FAJyD7fWofh98Kl6rZr9NYlH/tWmHwL8KF63Fqv01+Yf+1qAOr0eXxNYwXc/jC50u5jVVMI0iynDjrnKlnLdsbR61z3wcWez8JXNjfWV9Z3A1C6n2XVnLDlHmYqQXUA5Bzgc1RPg34TL/zEoE+niW4H/teo28MfCeLkeIo4fceKph/Oemt7+Vv6+4OlvO/9feeqZHrRmvJH0v4WQ/d+IMkH+54vkH/tWq7j4bRfc+K2pp7J4rdv/ZjSA628/wCS76T/ANi9d/8ApRBW14h8HaD4pe2l1qwE1xaNut7mKR4ZoT/syIQw/A15RKfhgupJqH/C19YF5HE0KTjW97qjEEqCVPBKg49hUv8AbPgcf6n4xeJz/uXayf8AtA0Ad14gttM+H3g3XNd08SfbxaFVur26eeaR8YjTzJWJxuIwucc1d+HnhhPCXg220/H+kOzXFyd24mVzk5Pc9BnvjNeY3up+Fb6xmtH+Kfii8gmQpJDLpouVdT1BX7LzUOlX8mhaelj4U8XeJzaoSUhh8Fbxk9TxCpJ9zQtLg+h73RXiH/CVfEJObK41+99Fn8F+Vn8TOv8AKrFh8R/iZpus6Wni7wjaw6TqF/FYrdlhbyhnzyY/NkPAUnHQ4xkZoA9nopFYOoYdDS0AFFFFABRRRQAUUUUAFFFFABRRRQAVxNt/yXrUv+xbtf8A0pnrtq4m2/5L1qX/AGLdr/6Uz0AdtRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFeefEP/kbvA//AGH4/wD0RNXodeefEP8A5G7wP/2H4/8A0RNQB6HRRRQAUUUUAIyqwwwBHuKgews5P9ZaQP8A70YNWKKAKD6Do8n+s0qxf/etkP8ASq7+EvDkn+s8P6W/+9ZRn+la9cF8T728lfw74bsbqayXX9RFtc3EDbZFgVS7qrdiwGM/WjXRLr+oeZ0B8EeE2znwxop9f+JfF/8AE0w+AfBzfe8J6GfrpsP/AMTWFP8ACXRLCS0vPBajw7qlrMji7g3P5yA/PHKpYeYGGRyc55rmNJ1jxbYa14/n8OWOm3FrY6tJcTPqE8imTEKHy4wo4OFzknHzDjrRddfN/l/n+AWfT+t/8j0E/DrwS33vB+gH66ZD/wDE0w/DXwM3XwdoP4abCP8A2Wsm/wDHGsXeg6FqPh+z020t9UsxdzX2s3XlwWwKgiPjlmOT6DCmslfizfN4Dl1mDTLO+v7bWk0l4rO53wXDF1G+JyOhDDGadne3X/g2/MV7q/8AW1/yOoPww8CN18H6J+FhGP6Uw/CrwE3Xwho/4WaD+lZ2qeJPGumwWUF1aeHLC4kjea6v7y9dbOL5sJEuQGZ8ck8AYNUbP4rXOoeD7K6sdOtbnW7/AFR9Jt4YrktbSSoTulEmM+VtG7pnt70t9v61t+Y/6/X8jcPwl8AHr4R0n8LYCm/8Ki+H/wD0KOl/9+BXLePPEPj6y+H/AIkhvbG0sLu0t0mj1XT7iVInjY4YRnG5ZVIHBOCD1rT1Dx7rVle6d4di/sGDWzYC8vrm/vHS1iUkqqrwHdmxntjFH9fn/kH9fl/ma3/Cofh//wBClpf/AH4FJ/wqH4ff9Clpn/fmufPxa1KbRbE6do9pd6vJrR0ae3ju8wmTYWEiSY5Q/KckcDPcV6HaXV5aeH1uvEf2WK6hhMl2bQs0S4BJ2lgCRgd6Hor9P+Bf8mG7t/W9jnP+FQfD7/oUtM/780f8Kg+H3/QpaZ/35rl4/jBqK6fa+IrqDQxoNzMi/ZI9Q3ahFE7BVlZMbT1BKDkA9a0dU8eeLG1bxVb+H9I0t7bw3teWe8nkBnXyhIUVVH3uvJOOnXPA9N/6tb/Ma12Nj/hUPw//AOhS0v8A780f8Ki+H/8A0KOl/wDfgVkaV8QfEc194Zu9Y0jT7bRvEz+XaLDcO9zATGZEMmVCkEDoOme9c14O8T+MNA+GT6zb6Rp9zoem3F08yy3Li6njE7l3QY2gKCcAnnaenFD03EtVp/W/+R3o+Efw/H/Mo6V/4Dinj4UeAV6eENI/G1U1j6r8Sbq68QSaV4Uk0SL7Naw3Nxda1dmFG81dyRoq8k7cEnoMio7f4pXmuaPokPhzTLZte1a4ntzBcXBMFsYP9a7Ooyy9MY67hRqGn9feb4+FvgNeng/RfxsYz/Snj4Z+BV6eDtC/HToj/wCy15/D4w1jwz4z8Y6j4j06F9Ths9Ptre1s5yYrqR3kWMoWGVBLDORkYPWtG5vPFh+K/gyDxZZ6dCGW9kjl02Z2Q/uOUYOAcjg56HPbFHS4PRtHZD4ceCF6eDtA/wDBZD/8TUg+H3gxfu+EdCH00yH/AOJrg9B8cahbeG/B1h4V0CzEmuNeIkM11J5duYnJ3FjuYjqSOvYVqf8ACzNT0jTvEUHiPS7WTWdGuLe3iisJW8q7a4A8rBcZXk85zjFH9foB1Y8CeEV+74W0UfTTov8A4mpF8GeF0+54b0hfpYRD/wBlrl18Y+LdH8V6To/irSNMZNRhuJluNMmkfHlR7zHtcA7unPQ57Yqt4O+Imv8AiifTr2Kz0W60y+fbLb2F6XvNOBzhplYAHGMEDBGe9C1dkGyudwnhrQo/9Xounp/u2qD+lWE0nTo/9XYWqf7sKj+lW6KAGJBFH/q4kT/dUCn0UUAFFFFABXnvxd/48/Cn/YyWv/oEtehV578Xv+PLwr/2Mlr/AOgS0Ad5a/8AHrH/ALtTVDaf8esf+7U1ABRRRQAUUUUAFFFFABRRRQAUUUUAFcTbf8l61L/sW7X/ANKZ67auJtv+S9al/wBi3a/+lM9AHbUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABXnnxD/5G7wP/wBh+P8A9ETV6HXm/wAVf7Qtrvw5qumac+pHS9VjuprdJVjZoxHIpILcZ+ccd6APSKK8yj+OGmji98L+Jbd++LSOVfzSQ/yqwnxv8Kf8vFvrdt/110mb+imgD0WiuBT43eAG+/rUsJ9JtOuUx+cdTD4z/D0/8zPbL/vxyL/NaAO4orjE+MHw+k+74t00f70uP51OnxT8Byfd8X6MP969QfzNAHWVzPjjwpN4o02zbTbtbHVtMukvbC4dNyLIuRtcd1YEg/WnJ8R/BEn3PGGgn2/tOH/4qp08ceE5f9X4o0Z/93UIj/7NQBzV/ofjbxktrpviiPSdJ0mKdJrs6ddSTS3mw7gi7kXYhIGckmtHSvCN/YweMkmltyddvJp7bazfIrwqgD/LwcqemeK2h4t8ON93xBpZ+l7H/jTx4n0Fvu63px+l3H/jSaTTXf8AW3+SGm00+3/B/wAzzd/hrr1i3hedLDQtfOlaOumy2epSuIYpQQfPjPltnpg5UHAFLY/DbxMmk32nXj6SPM8RW+sxzQO6IVDK0kYTaduNmF5Oe+K9LHiDRm+7q9ifpcp/jTxrWlt93UrM/Sdf8aq7vzf1vf8AMmytb+trfkcR4s8GavffECHxDYaZomuQfYRaLbaw7Kto4ct5qAIwOc4PAPHWsbTvhr4osdJBWbS/7V0zXZNWsJVZ1huRICJI3ULmMYJAxur1MapYN92+tj9Jl/xp4vbVvu3MJ+kgpLTb+tb/AJjeu/8AWlvyOA1fwv4z8W+FfEtvrt1YWcupWa29jp1vK0kMBU5LvIUBLMcDgYAHeqWp+BNbu9XsfEs+heHdW1F9OWyv9Mv2LQqVYlJIpDG2DzggryO9eoCeJvuyofowpfMT++v50f1+f+Yf1+X+SPOLL4d6rFbaC8raPBPZ65/ad1DY24ghjj8tkEUe1AXIyPmfBPPPQV6FfWcWo6dc2VxkxXMTRPjrtYEH9DU29f7w/OjcPUfnQ9Vyvb/gJfoC0d/63b/U8g0v4X6zZ2ljoTaN4TjtbORA+uizSS7nhU5A8p4iokIGCxYjqRzXVxeDdQjm8dMZrbHiL/j0wzfJ/o4j+f5ePmHbPH5V2uRSZHqKH7yafX9bf5DWjTRwjeB9SOleArbz7Xf4bmie7O9sSBYGjPl/Lzye+OKwLTwR4+tfBVx4LaXRDpt88wk1FZpPNt4ZZGZ0EZTDthjg5A59sn1rcv8AeH50nmJ/eX86bd279f8Agf5CWiSXT9L/AOZ5hq3w2udP8UXOqaDoXh/Xbe9toIHttaGGtniTYro3lvlSoGV45HWpf+Ffa7pNl4e1PRZdKk17R5J2ltxALW1uEn+/Gvlr8uMLtYgk45616UZYx1dR+NMN3br96eIfVxS1DQ8tvfhv4n8RXWv6nrd7ptnfaglnJYJas8iWktu7MqsSo3KcjJH95uOBnVh0DxprXjbQNc8Rpo9jbaStwrWtncSStI0ke3fuKAdcfL2GeT0ruTqNkv3ry3H1lX/GmHV9NX72oWo+s6/40dLA9dTgvDXw61bRm8Fm6uLJ/wCwDffavLdzv8/OzZlRnGec49s1T8beEBbx+MNd1fVrfTbW9m0+5s7rY8v2eW3wAZFC/dLYHBPBJOK9GOvaQv3tVsh9bhP8ahm1/wAPyxPFcatprxuCrI9zGQwPUEE0O+6/rW4K3U8tivtf8SfFrw2l5rGhyy2tleuBoUjXC2waIIJnZx1LEYXGODyc1ft/APiXUfEeiXmuaZ4dtLrSrtLm412wZlur4J/CUEagb/4ssR1wK7TTr7wRoayDSLrw/pwkOZBayQRbz77cZq03jLwwn3/EekL9b6If+zU9rNdP87/13DWzRtUVzz/EHwZH/rPF2hL/AL2pQj/2aoH+J3gWP73jHQz/ALuoRH+TUgOoorj3+LXgCP73i7ST/u3IP8qgb4y/D1f+ZqsW/wB0s38hQB29FcI3xr+Hy/8AMwq//XO1nf8AkhqB/jf4K/5d7jUrr/rjpNzz+aCgD0KvPfi9/wAeXhX/ALGS1/8AQJagf44aAP8AU6L4kn9NumFc/wDfRFc94g8YXfxCvtBstM8NalY29lq0N7Lc37RJ8iK4wEV2JJ3UAeyWn/HrH/u1NUVqMWsefSpaACiiigAooooAKKKKACiiigAooooAK4m2/wCS9al/2Ldr/wClM9dtXE23/JetS/7Fu1/9KZ6AO2ooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACq15Yw3se2ZcirNFAHPv4QsHP3B+VQP4Jsj0UV09FAHIv4EtG6AVg+LPD2n+F/Cepa5dY8qxt2l2n+NgPlX8WwPxr0yvG/j2dV8U/wBjfDzwzGZb7Vpftd0eQkNvGcBpD2Xec59UA6kAgEPwx0u08e/D6w1iTY1zhoboYHEqHB+mRhv+BV1DfDCxb70ER+qCvM/2dbq/8JeKL/wtqqyCy1YzT6bcFcJNLbyNFKF9yFyQeQEHqM/R1AHm7/CjTG+9ZW5+sK/4VA/wh0dvvaZZn626/wCFen0UAeVt8HNEP/MJsT/27J/hUTfBjRD/AMwbTz/26p/hXrNFAHkLfBXQz/zBdP8A/AVP8K878HaB4Y8V/EbxL4dh0rTiunMDalbdcuqHZKc45G/BHsa97+I3iY+Efh/qurQ5a6SHyrRFGS07nZGAO/zMDj0Br5a8K6H4l+FPjy18SavbsLfS2tm1kI29reK63qNwHU4B+jbR6UAezn4IaGf+YPZf9+R/hTD8DdDP/MHtP+/Qr2cEMoKkEHkEd6WgDxQ/AvQz/wAwi1/74pv/AAonQ/8AoEW3/fNe20UAeJH4E6H/ANAi2/75NJ/wonQ/+gTb/wDfJr26igDxH/hROh/9Am3/ACNH/CidD/6BFv8A98mvbqKAPEf+FE6H/wBAm3/75NL/AMKJ0P8A6BFt/wB8mvbaKAPEx8CtD/6BFt/3xTh8DNDH/MItf+/de1UUAeLP8E9AhjaSXSrJERSzM0QwAOpNcN8K/Dfhv4hQ6yYtIsN9leMI1+zqD5D5MZPHXhh+Fev/ABt1q7074dTaXo0Tz6tr0q6ZZwxDLMZM7/8AxwMM9iRXifwr0/WPhP8AFyOHWAG028uV0S7niy0aXMkSSxr/AN9MAG9Nx9aAPT1+Cmhj/mDWH/gMn+FSr8F9DH/MG0//AMBE/wAK9booA8oX4NaIP+YRYD/t1T/CpV+D2jDppdkPpbJ/hXqVFAHmifCXSl+7p9qPpAv+FTL8LdPX7ttCv0jFei1Be3kGnafcXt5IIre2iaaVz/CijJP5CgDwPVZ9K0n43aV4LxEI7m2ImcKPknf5o1+uFx/20Felr8Pbcf8A6q+YdR8PeNvFvi2bxrZ2Usd7erLrdjAf9c0EMsaqUXvgMu0dWCZGcjP194T8Q2/ivwjpmu2nEd9brKV/uNj5l/BgR+FAGKngG1HWp08D2a9VFdXRQBza+DLBeqCr1p4dsrRw0aDI9q1qKAADAAHQUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAVxNt/wAl61L/ALFu1/8ASmeu2ribb/kvWpf9i3a/+lM9AHbUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFV4rC1hv572OFRdXCqkkuMsyrnaufQZJx0ySe5qxRQBwPwtsLW98A2j3UKyPbatfzQuR80bi7mGQe3BIPqCR0Nd9XIfDCxutO8Dpb39tLbTC/vn8uVCrbWu5WU4PYggj1BBrr6ACiiigAooooAr3Nha3k1tJdQrK1rJ50O8ZCPgjcB64J59647QoYrj4t+OobiNJYpLLTUeN1BVlKTggg9RXc1yGg2N1B8VfF95NbSpbXNtp6wzMhCSFVm3BT0OMjPpkUAdVa20VnaQ21uu2GFBHGuScKBgDJ9qloooAKKKKACiiigAooooAKKKKACiiigCvJYWsuoQ30sKvcwIyRSMMmMNjdj0ztGfpXE+E7C11PX/iBZ38Kz28usoHRxwf9FgIPsQcEHqCM131ch4Msbq08T+NJbq2lhjudXSSB3QgSr9mhG5SeoyCMjuCO1AHX0UUUAFFFFABUF9Y22pWUtnfRLNbzDbJE33XHoR3B7joRxU9FAHEXYC/HbSAowB4euwAO37+Cuvs7C10+OSOyhWBJJWmZEGF3scsQO2Tkn1JJ6muYu7G6b4z6XfLbSm0TQrqJ5wh2K5nhIUt0yQCcex9K6+gAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACuJtv+S9al/2Ldr/6Uz121YUXh54/iFdeJDcKY59MhsRBt5BSWRy2fQ+YBj2oA3aKKKACiiigAooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACsHxNqmuafFEPD2l2t4+1pZ5b268iGJFxkbgCdxzxxgYJJ9d6uN8VPr814Xg0A6rpFoy7rKO6WOW8fg7sMNrIvHyllyQTyAAUNG54e1a51zTzfz2D2MEpH2eOUgyMmPvHBIAJzj1GD3wNauWvdZ1zSPAWoa1qdvbQ6hHHJOlmX3rAMfJGWX77dM4wCTgHGK6HTxcrptsL9xJdCJfOcDAZ8Dccduc1TJLFFFFIYUUUUAc94n1fXNMCHQtKtbuONDNdT3t39njjjHVQcHLkZPOAMcmr2g6lc6vpovrmxksY5m3QQzY8zy8DBcAkAk5OM9Md8iue8Sy+IX1H7RH4dbVtKtJVEdlFdpHLPJ180qw2sinAClhyCxBwMXW1DW7fTLQ6vts5p2ea6khAZLOMMMJvIK5CnJYjB2tjqMC2u/6/r+tQe/8AX9f12OnorH8KX17qfhi0vNTjZJ5d5G9NjMm9hGxXsWQKxHHJ6CtigAooooAKKKKACiiigAooooAKKKKAOc8T63rekyK+kaTbXNpAnn3txd3fkKqZwVj4OXwCecAcc81p6LfXWp6at5eWT2PnEtFBL/rFj/h3gcBj1x2yB1rmfEE/iM6sLxfDTavpttMEtrKK8SORpB0ndW+UqD90bsj72CcbbNvda2/ibQbLULkR3Btrm8v4LfBjC5VY484ycF/vcZ2E96I6r+vX+v8AMJbnW0VT1a9OnaPd3o8sfZ4WlJkJCgKMknHsKg8NXF/d+F9OudX2/bp7dJZwi7QrMNxUD0GcfhRvf+v62A06KKKACiiigDmtZ8b2mja0untpuq3gQK11dWdoZYbMN0MhBz05wASBycCtjSNUh1rTUvrVJFt5STC0i7fNTPDgddp6jPbBrh9f8daBNfGDxDfPYaErskZkgk26pIv3lVguGQHjaDlz/sj5ugi8S3kmp+HrQaYLYarFNNLFM5EtvGigglQMZyyAjsWx2ojqgejOlooooAKKKKACiiigAooooAKKKKACiiigAooooAKKKKACs1/EWix3E0EmrWKTQAtKjXCAoAQDkZ4xkZ9Mirt0VFpMZJvs6bDulDAeWMfeyeBjrzXnN9YWF58UvBug2ISWz0Swm1NjkNuztjiYnuSxLZ7kZoWrt/XV/oD0Vz0SyvrTUbUXFhcxXMLEgSQuGXIOCMjuDxU9ZXhuy0ax0VV8NwxxWMkskg8sHDuXO5ueTls8/lxWrQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAYvifRLvX7GGzt72G1hW4immElsZfNEciuF4dcAlcHrxW0M4GeT3xRRQAUUUUAFFFFABWFrmk6rf3Uc+n3lkghUeXBe2rTR788uQrryOMdcc+vG7RQBDZwyW9lFFPO1xKigPMwwXbucdue3apqKKACiiigAooooAKKKKACiiigAooooAKwp/DS3Xia51K6nWW1urNLWS0aPIKqXON2eh8zJGOoHNbtFAHPaz4UW88Jz6BoklvpVrcApKBbGQFCfmAAdcZGRnnrW/ErJCiuVLKoBKrtBPsMnH506igAooooAKKKKAGGKMqoMakIcqNo+U+3pXMIty/xUlmms5zBDpqQW82w+WNzs8h3dM5WJcdfwrqqKOt/wCuwdLBRRRQAUUUUAFFFFABRRRQAUUUUAFFFFABRRRQAUUUUAVdT0201jS7nTtShE9pdRNFNESRvUjBGRzWJD4A0C3lSa3huIZ1h+zmaO6kR3hwAIyQR8g2jA7Y4xk56WigBkEEVtbxwW8axRRKEREGAqgYAA9KfRRQAUUUUAFFFFABRRRQAUUUUAf/2Q==)\n",
        "* This hybrid token embedding layer is a combination of token embeddings and character embeddings. In other words, they create a stacked embedding to represent sequences before passing them to the sequence label prediction layer\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "N4lW_EfluZB9"
      },
      "source": [
        "To start replicating (or getting close to replicating) the model in Figure 1, we're going to go through the following steps:\n",
        "1. Create a token-level model (similar to `model_1`)\n",
        "2. Create a character-level model (similar to `model_3` with a slight modification to reflect the paper)\n",
        "3. Combine (using [`layers.Concatenate`](https://www.tensorflow.org/api_docs/python/tf/keras/layers/Concatenate)) the outputs of 1 and 2\n",
        "4. Build a series of output layers on top of 3 similar to Figure 1 and section 4.2 of [*Neural Networks for Joint Sentence Classification\n",
        "in Medical Paper Abstracts*](https://arxiv.org/pdf/1612.05251.pdf)\n",
        "5. Construct a model which takes token and character-level sequences as input and produces sequence label probabilities as output"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "9xs5EFrvpNg7"
      },
      "source": [
        "1 # Token_level Model (using Pretrained -- Universal Sentence Encoder)\n",
        "token_inputs = layers.Input(shape = [], dtype= tf.string, name = \"token_input\")\n",
        "token_embedding = tf_hub_embedding_layer(token_inputs)\n",
        "token_dense = layers.Dense(128,activation=\"relu\")(token_embedding)\n",
        "token_model = tf.keras.Model(inputs = token_inputs,\n",
        "                             outputs = token_dense)\n",
        "\n",
        "2 # char_level Model\n",
        "char_inputs = layers.Input(shape=(1,), dtype= tf.string, name=\"char_input\")\n",
        "char_vectors = char_vectorizer(char_inputs)\n",
        "char_embedding = char_embed(char_vectors)\n",
        "char_bi_lstm = layers.Bidirectional(layers.LSTM(25,activation=\"relu\"))(char_embedding)\n",
        "char_model = tf.keras.Model(inputs= char_inputs,    # char_dense = layers.Dense(128,activation=\"relu\")(char_bilstm)\n",
        "                            outputs =char_bi_lstm)\n",
        "\n",
        "\n",
        "3 # Now Concatenate token_model and char_model \n",
        "concat_layer = layers.Concatenate(name = \"token_char_hybrid\")([token_model.output, \n",
        "                                                               char_model.output])\n",
        "\n",
        "4 # Add Some Layer on top of concat_layer\n",
        "concat_dropout = layers.Dropout(0.5)(concat_layer)\n",
        "concat_dense = layers.Dense(256,activation=\"relu\")(concat_dropout)\n",
        "final_dropout = layers.Dropout(0.2)(concat_dense)\n",
        "output_layer = layers.Dense(num_classes,activation=\"softmax\")(final_dropout)\n",
        "\n",
        "model_4 = tf.keras.Model(inputs = [token_model.input, char_model.input], \n",
        "                         outputs = output_layer,\n",
        "                         name=\"model_4_token_and_char_embeddings\")\n",
        "\n"
      ],
      "execution_count": 81,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nXSF7qgTlUDT",
        "outputId": "b8073640-2ff8-41c1-8e18-7dc45a0128c7"
      },
      "source": [
        "model_4.summary()"
      ],
      "execution_count": 82,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"model_4_token_and_char_embeddings\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "char_input (InputLayer)         [(None, 1)]          0                                            \n",
            "__________________________________________________________________________________________________\n",
            "token_input (InputLayer)        [(None,)]            0                                            \n",
            "__________________________________________________________________________________________________\n",
            "char_vectorizer (TextVectorizat (None, 290)          0           char_input[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "universal_sentence_encoder (Ker (None, 512)          256797824   token_input[0][0]                \n",
            "__________________________________________________________________________________________________\n",
            "char_embed (Embedding)          (None, 290, 25)      1750        char_vectorizer[1][0]            \n",
            "__________________________________________________________________________________________________\n",
            "dense_4 (Dense)                 (None, 128)          65664       universal_sentence_encoder[1][0] \n",
            "__________________________________________________________________________________________________\n",
            "bidirectional (Bidirectional)   (None, 50)           10200       char_embed[1][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "token_char_hybrid (Concatenate) (None, 178)          0           dense_4[0][0]                    \n",
            "                                                                 bidirectional[0][0]              \n",
            "__________________________________________________________________________________________________\n",
            "dropout_1 (Dropout)             (None, 178)          0           token_char_hybrid[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "dense_5 (Dense)                 (None, 256)          45824       dropout_1[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "dropout_2 (Dropout)             (None, 256)          0           dense_5[0][0]                    \n",
            "__________________________________________________________________________________________________\n",
            "dense_6 (Dense)                 (None, 5)            1285        dropout_2[0][0]                  \n",
            "==================================================================================================\n",
            "Total params: 256,922,547\n",
            "Trainable params: 124,723\n",
            "Non-trainable params: 256,797,824\n",
            "__________________________________________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WX8LZdC1lh4P"
      },
      "source": [
        "### Visualize the Hybrid Model"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 856
        },
        "id": "j7z82GivlaYB",
        "outputId": "77d0fc61-59b1-4b27-ba7e-212c2ee28f40"
      },
      "source": [
        "tf.keras.utils.plot_model(\n",
        "    model_4, to_file='model.png', show_shapes=False, show_dtype=False,\n",
        "    show_layer_names=True, rankdir='TB', expand_nested=False, dpi=96\n",
        ")"
      ],
      "execution_count": 83,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 83
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "HIMyXb_kl4Mr"
      },
      "source": [
        "# Compile token char model\n",
        "model_4.compile(loss=\"categorical_crossentropy\",\n",
        "                optimizer=tf.keras.optimizers.Adam(), # section 4.2 of https://arxiv.org/pdf/1612.05251.pdf mentions using SGD but we'll stick with Adam\n",
        "                metrics=[\"accuracy\"])"
      ],
      "execution_count": 84,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "CHNadX_yoKKX"
      },
      "source": [
        "# Combine chars and tokens into a dataset\n",
        "train_char_token_data = tf.data.Dataset.from_tensor_slices((train_sentences, train_chars)) # make data\n",
        "train_char_token_labels = tf.data.Dataset.from_tensor_slices(train_labels_one_hot) # make labels\n",
        "train_char_token_dataset = tf.data.Dataset.zip((train_char_token_data, train_char_token_labels)) # combine data and labels\n",
        "\n",
        "# Prefetch and batch train data\n",
        "train_char_token_dataset = train_char_token_dataset.batch(32).prefetch(tf.data.AUTOTUNE) \n",
        "\n",
        "# Repeat same steps validation data\n",
        "val_char_token_data = tf.data.Dataset.from_tensor_slices((val_sentences, val_chars))\n",
        "val_char_token_labels = tf.data.Dataset.from_tensor_slices(val_labels_one_hot)\n",
        "val_char_token_dataset = tf.data.Dataset.zip((val_char_token_data, val_char_token_labels))\n",
        "val_char_token_dataset = val_char_token_dataset.batch(32).prefetch(tf.data.AUTOTUNE)"
      ],
      "execution_count": 85,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "JIeQA_xdoPzJ",
        "outputId": "9c787545-6977-49c8-d86a-d19d3fc51b5b"
      },
      "source": [
        "# Check out training char and token embedding dataset\n",
        "train_char_token_dataset, val_char_token_dataset"
      ],
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(<PrefetchDataset shapes: (((None,), (None,)), (None, 5)), types: ((tf.string, tf.string), tf.float64)>,\n",
              " <PrefetchDataset shapes: (((None,), (None,)), (None, 5)), types: ((tf.string, tf.string), tf.float64)>)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 86
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1eBAI8lNoShY",
        "outputId": "5acf89ec-9544-4cb0-b1f0-f535e882b6e4"
      },
      "source": [
        "# Fit the model on tokens and chars\n",
        "model_4_history = model_4.fit(train_char_token_dataset, # train on dataset of token and characters\n",
        "                              steps_per_epoch=int(0.1 * len(train_char_token_dataset)),\n",
        "                              epochs=10,\n",
        "                              validation_data=val_char_token_dataset,\n",
        "                              validation_steps=int(0.1 * len(val_char_token_dataset)))"
      ],
      "execution_count": 87,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "562/562 [==============================] - 167s 289ms/step - loss: 0.9316 - accuracy: 0.6286 - val_loss: 0.7693 - val_accuracy: 0.7048\n",
            "Epoch 2/10\n",
            "562/562 [==============================] - 162s 288ms/step - loss: 0.7773 - accuracy: 0.6966 - val_loss: 0.7133 - val_accuracy: 0.7271\n",
            "Epoch 3/10\n",
            "562/562 [==============================] - 163s 290ms/step - loss: 0.7541 - accuracy: 0.7097 - val_loss: 0.6872 - val_accuracy: 0.7424\n",
            "Epoch 4/10\n",
            "562/562 [==============================] - 163s 290ms/step - loss: 11.8752 - accuracy: 0.7228 - val_loss: 0.6679 - val_accuracy: 0.7463\n",
            "Epoch 5/10\n",
            "562/562 [==============================] - 163s 291ms/step - loss: 0.7405 - accuracy: 0.7171 - val_loss: 0.6611 - val_accuracy: 0.7473\n",
            "Epoch 6/10\n",
            "562/562 [==============================] - 164s 291ms/step - loss: 0.7305 - accuracy: 0.7208 - val_loss: 0.6454 - val_accuracy: 0.7580\n",
            "Epoch 7/10\n",
            "562/562 [==============================] - 164s 292ms/step - loss: 0.7063 - accuracy: 0.7299 - val_loss: 0.6443 - val_accuracy: 0.7600\n",
            "Epoch 8/10\n",
            "562/562 [==============================] - 164s 292ms/step - loss: 0.6984 - accuracy: 0.7340 - val_loss: 0.6335 - val_accuracy: 0.7557\n",
            "Epoch 9/10\n",
            "562/562 [==============================] - 162s 289ms/step - loss: 0.6996 - accuracy: 0.7311 - val_loss: 0.6434 - val_accuracy: 0.7557\n",
            "Epoch 10/10\n",
            "562/562 [==============================] - 162s 289ms/step - loss: 0.7018 - accuracy: 0.7309 - val_loss: 0.6321 - val_accuracy: 0.7603\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "_0AOOtFK1a6p",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8fba8500-6b66-4547-bb09-67d9be6413f8"
      },
      "source": [
        "# Evaluate on the whole validation dataset\n",
        "model_4.evaluate(val_char_token_dataset)"
      ],
      "execution_count": 88,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "945/945 [==============================] - 58s 61ms/step - loss: 23792656.0000 - accuracy: 0.7598\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[23792656.0, 0.7598305344581604]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 88
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QPLhvd_F1eIm",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5c6f95d5-5fa7-47a2-b365-416b12340db3"
      },
      "source": [
        "# Make predictions using the token-character model hybrid\n",
        "model_4_pred_probs = model_4.predict(val_char_token_dataset)\n",
        "model_4_pred_probs"
      ],
      "execution_count": 89,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[4.81065422e-01, 3.40999186e-01, 1.79997389e-03, 1.66568533e-01,\n",
              "        9.56688821e-03],\n",
              "       [2.74215668e-01, 6.57577336e-01, 1.48074096e-03, 6.47516549e-02,\n",
              "        1.97461597e-03],\n",
              "       [4.36756134e-01, 5.47802337e-02, 5.64954877e-02, 4.24435914e-01,\n",
              "        2.75322516e-02],\n",
              "       ...,\n",
              "       [1.14926996e-04, 4.47407336e-04, 1.09152636e-02, 4.81208008e-05,\n",
              "        9.88474309e-01],\n",
              "       [9.39613860e-03, 6.19878434e-02, 1.74138308e-01, 2.82176491e-03,\n",
              "        7.51655936e-01],\n",
              "       [6.84660450e-02, 8.80190194e-01, 4.18375172e-02, 2.41007935e-03,\n",
              "        7.09615275e-03]], dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 89
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "QvEiqpcG1hXO",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5d26705f-4e78-48ee-ca48-f182843460c7"
      },
      "source": [
        "# Turn prediction probabilities into prediction classes\n",
        "model_4_preds = tf.argmax(model_4_pred_probs, axis=1)\n",
        "model_4_preds"
      ],
      "execution_count": 90,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(30212,), dtype=int64, numpy=array([0, 1, 0, ..., 4, 4, 1])>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 90
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "J2RykWDb1kRe",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "47db6dcd-c958-4a1a-d947-7b5241aec563"
      },
      "source": [
        "# Get results of token-char-hybrid model\n",
        "model_4_results = calculate_results(y_true=val_labels_encoded,\n",
        "                                    y_pred=model_4_preds)\n",
        "model_4_results"
      ],
      "execution_count": 91,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 75.98305309148682,\n",
              " 'f1': 0.7545412454754323,\n",
              " 'precision': 0.7561886438045121,\n",
              " 'recall': 0.7598305309148683}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 91
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ejfoplj8a-ws"
      },
      "source": [
        "## Model 5: Transfer Learning with pretrained token embeddings + character embeddings + positional embeddings"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OoH1EhpzhzgF"
      },
      "source": [
        "As it's a Sequential classification problem the sequences come in a particular order. Like *OBJECTIVE* comes first rather then *CONCLUSION*.\n",
        "\n",
        "Abstracts typically come in a sequential order, such as:\n",
        "\n",
        "* `OBJECTIVE` ...\n",
        "* `METHODS` ...\n",
        "* `METHODS` ...\n",
        "* `METHODS` ...\n",
        "* `RESULTS` ...\n",
        "* `CONCLUSIONS` ...\n",
        "\n",
        "Or\n",
        "\n",
        "* `BACKGROUND` ...\n",
        "* `OBJECTIVE` ...\n",
        "* `METHODS` ...\n",
        "* `METHODS` ...\n",
        "* `RESULTS` ...\n",
        "* `RESULTS` ...\n",
        "* `CONCLUSIONS` ...\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MWPF0-s2i5uK"
      },
      "source": [
        "Here we do some Feature Engineering so that our model can learn the order sentences in the Abstract and know where the sentence appear in the Abstract. The `\"line_number\"` and `\"total_lines\"` columns are features which didn't necessarily come with the training data but can be passed to our model as a **positional embedding**.\n",
        "\n",
        "***Positional Embedding make your model include the information about order of the input. The positional encoding step allows the model to recognize which part of the sequence an input belongs to.***"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "4KNucgpElCiO"
      },
      "source": [
        "But to avoid our model thinking a line with \"line_number\"=5 is five times greater than a line with \"line_number\"=1, we'll use one-hot-encoding to encode our \"line_number\" and \"total_lines\" features.\n",
        "\n",
        "That is why we have to use one-hot encoding. We use tf.one_hot for it."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SiNBKWP21kq2",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "d401e86e-b054-48c6-dfb7-72df08bed1a5"
      },
      "source": [
        "# Checking the Distribution of line Number in The data.\n",
        "train_df[\"line_number\"].value_counts()"
      ],
      "execution_count": 92,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "0     15000\n",
              "1     15000\n",
              "2     15000\n",
              "3     15000\n",
              "4     14992\n",
              "5     14949\n",
              "6     14758\n",
              "7     14279\n",
              "8     13346\n",
              "9     11981\n",
              "10    10041\n",
              "11     7892\n",
              "12     5853\n",
              "13     4152\n",
              "14     2835\n",
              "15     1861\n",
              "16     1188\n",
              "17      751\n",
              "18      462\n",
              "19      286\n",
              "20      162\n",
              "21      101\n",
              "22       66\n",
              "23       33\n",
              "24       22\n",
              "25       14\n",
              "26        7\n",
              "27        4\n",
              "28        3\n",
              "29        1\n",
              "30        1\n",
              "Name: line_number, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 92
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 281
        },
        "id": "7JHRdo92mxBJ",
        "outputId": "48ed7443-a215-4298-cf6e-33282c59b5c8"
      },
      "source": [
        "# Check the distribution of line numbers\n",
        "train_df.line_number.plot.hist()"
      ],
      "execution_count": 93,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.axes._subplots.AxesSubplot at 0x7fef52380690>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 93
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gpCOBf3xnKMf"
      },
      "source": [
        "Looking at the distribution of the \"line_number\" column, it looks like the majority of lines have a position of 15 or less.\n",
        "\n",
        "Knowing this, let's set the depth parameter of tf.one_hot to 15."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "SdL4m9tenC4U"
      },
      "source": [
        "train_line_numbers_one_hot = tf.one_hot(train_df[\"line_number\"].to_numpy(),depth= 15)\n",
        "val_line_numbers_one_hot = tf.one_hot(val_df[\"line_number\"].to_numpy(),depth= 15)\n",
        "test_line_numbers_one_hot = tf.one_hot(test_df[\"line_number\"].to_numpy(),depth= 15)"
      ],
      "execution_count": 94,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tVEL44K7ngg8",
        "outputId": "192dd1c9-1103-4538-ccd3-a738fc977d36"
      },
      "source": [
        "# Check one-hot encoded \"line_number\" feature samples\n",
        "train_line_numbers_one_hot.shape, train_line_numbers_one_hot[:20]"
      ],
      "execution_count": 96,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(TensorShape([180040, 15]), <tf.Tensor: shape=(20, 15), dtype=float32, numpy=\n",
              " array([[1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0.],\n",
              "        [1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0., 0., 0., 0.]],\n",
              "       dtype=float32)>)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 96
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sUTx71vW-L6V"
      },
      "source": [
        "We could create a one-hot tensor which has room for all of the potential values of \"line_number\" (depth=30), however, this would end up in a tensor of double the size of our current one (depth=15) where the vast majority of values are 0. Plus, only ~2,000/180,000 samples have a \"line_number\" value of over 15. So we would not be gaining much information about our data for doubling our feature space. This kind of problem is called the curse of dimensionality. However, since this we're working with deep models, it might be worth trying to throw as much information at the model as possible and seeing what happens. I'll leave exploring values of the depth parameter as an extension."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rsl9FWp39ji3"
      },
      "source": [
        "We can do the same above process for the total line also in data."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YPa_BUQ2nidf",
        "outputId": "b581d7e4-7418-4251-ae03-1aa84d00abe0"
      },
      "source": [
        "# How many different numbers of lines are there?\n",
        "train_df[\"total_lines\"].value_counts()"
      ],
      "execution_count": 97,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "11    24468\n",
              "10    23639\n",
              "12    22113\n",
              "9     19400\n",
              "13    18438\n",
              "14    14610\n",
              "8     12285\n",
              "15    10768\n",
              "7      7464\n",
              "16     7429\n",
              "17     5202\n",
              "6      3353\n",
              "18     3344\n",
              "19     2480\n",
              "20     1281\n",
              "5      1146\n",
              "21      770\n",
              "22      759\n",
              "23      264\n",
              "4       215\n",
              "24      200\n",
              "25      182\n",
              "26       81\n",
              "28       58\n",
              "3        32\n",
              "30       31\n",
              "27       28\n",
              "Name: total_lines, dtype: int64"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 97
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 266
        },
        "id": "TXRiQjp5-V_b",
        "outputId": "07dfbc2e-c9a3-448d-8497-254cd73ab0a6"
      },
      "source": [
        "# Check the distribution of total lines\n",
        "train_df.total_lines.plot.hist();"
      ],
      "execution_count": 98,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "JfwxK17a-p-5"
      },
      "source": [
        "It shows that majority of data has line number below 20. We can perform numpy percentile to check this."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UsgZg3Wl-Yje",
        "outputId": "2942bb06-71c7-401c-a52f-749e7349dbfc"
      },
      "source": [
        "# Check the coverage of a \"total_lines\" value of 20\n",
        "np.percentile(train_df.total_lines, 98) # a value of 20 covers 98% of samples"
      ],
      "execution_count": 101,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "20.0"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 101
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TXQGj_Vy_Hwn",
        "outputId": "f1782ceb-dd4f-440d-9ed9-d9d43b7e3f0e"
      },
      "source": [
        "# Use TensorFlow to create one-hot-encoded tensors of our \"total_lines\" column \n",
        "train_total_lines_one_hot = tf.one_hot(train_df[\"total_lines\"].to_numpy(), depth=20)\n",
        "val_total_lines_one_hot = tf.one_hot(val_df[\"total_lines\"].to_numpy(), depth=20)\n",
        "test_total_lines_one_hot = tf.one_hot(test_df[\"total_lines\"].to_numpy(), depth=20)\n",
        "\n",
        "# Check shape and samples of total lines one-hot tensor\n",
        "train_total_lines_one_hot.shape, train_total_lines_one_hot[:10]"
      ],
      "execution_count": 102,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(TensorShape([180040, 20]), <tf.Tensor: shape=(10, 20), dtype=float32, numpy=\n",
              " array([[0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.,\n",
              "         0., 0., 0., 0.],\n",
              "        [0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 0., 1., 0., 0., 0., 0.,\n",
              "         0., 0., 0., 0.]], dtype=float32)>)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 102
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "a_rVHJOd_p0-"
      },
      "source": [
        "### Creating The Beast (tribrid embedding model)\n",
        "\n",
        "Steps for Creating the Model:\n",
        "1. Create a token-level model (similar to `model_1`)\n",
        "2. Create a character-level model (similar to `model_3` with a slight modification to reflect the paper)\n",
        "3. Create a `\"line_number\"` model (takes in one-hot-encoded `\"line_number\"` tensor and passes it through a non-linear layer)\n",
        "4. Create a `\"total_lines\"` model (takes in one-hot-encoded `\"total_lines\"` tensor and passes it through a non-linear layer)\n",
        "5. Combine (using [`layers.Concatenate`](https://www.tensorflow.org/api_docs/python/tf/keras/layers/Concatenate)) the outputs of 1 and 2 into a token-character-hybrid embedding and pass it series of output to Figure 1 and section 4.2 of [*Neural Networks for Joint Sentence Classification\n",
        "in Medical Paper Abstracts*](https://arxiv.org/pdf/1612.05251.pdf)\n",
        "6. Combine (using [`layers.Concatenate`](https://www.tensorflow.org/api_docs/python/tf/keras/layers/Concatenate)) the outputs of 3, 4 and 5 into a token-character-positional tribrid embedding \n",
        "7. Create an output layer to accept the tribrid embedding and output predicted label probabilities\n",
        "8. Combine the inputs of 1, 2, 3, 4 and outputs of 7 into a [`tf.keras.Model`](https://www.tensorflow.org/api_docs/python/tf/keras/Model)"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IpvrmD4I_IBp"
      },
      "source": [
        "\n",
        "# 1. Token Model\n",
        "token_inputs = layers.Input(shape=[], dtype=\"string\", name=\"token_inputs\")\n",
        "token_embeddings = tf_hub_embedding_layer(token_inputs)\n",
        "token_outputs = layers.Dense(128, activation=\"relu\")(token_embeddings)\n",
        "token_model = tf.keras.Model(inputs=token_inputs,\n",
        "                             outputs=token_outputs)\n",
        "\n",
        "\n",
        "\n",
        "# 2. Char Model\n",
        "char_inputs = layers.Input(shape=(1,), dtype= tf.string, name=\"char_input\")\n",
        "char_vectors = char_vectorizer(char_inputs)\n",
        "char_embedding = char_embed(char_vectors)\n",
        "char_bi_lstm = layers.Bidirectional(layers.LSTM(25,activation=\"relu\"))(char_embedding)\n",
        "char_model = tf.keras.Model(inputs= char_inputs,    # char_dense = layers.Dense(128,activation=\"relu\")(char_bilstm)\n",
        "                            outputs =char_bi_lstm)\n",
        "\n",
        "\n",
        "# 3. Line numbers inputs\n",
        "line_number_inputs = layers.Input(shape=(15,), dtype=tf.int32, name=\"line_number_input\")\n",
        "x = layers.Dense(32, activation=\"relu\")(line_number_inputs)\n",
        "line_number_model = tf.keras.Model(inputs=line_number_inputs,\n",
        "                                   outputs=x)\n",
        "\n",
        "# 4. Total lines inputs\n",
        "total_lines_inputs = layers.Input(shape=(20,), dtype=tf.int32, name=\"total_lines_input\")\n",
        "y = layers.Dense(32, activation=\"relu\")(total_lines_inputs)\n",
        "total_line_model = tf.keras.Model(inputs=total_lines_inputs,\n",
        "                                  outputs=y)\n",
        "\n",
        "\n",
        "# 5. Combine token and char embeddings into a hybrid embedding\n",
        "combined_embeddings = layers.Concatenate(name=\"token_char_hybrid_embedding\")([token_model.output, \n",
        "                                                                              char_model.output])\n",
        "z = layers.Dense(256, activation=\"relu\")(combined_embeddings)\n",
        "z = layers.Dropout(0.5)(z)\n",
        "\n",
        "# 6. Combine positional embeddings with combined token and char embeddings into a tribrid embedding\n",
        "z = layers.Concatenate(name=\"token_char_positional_embedding\")([line_number_model.output,\n",
        "                                                                total_line_model.output,\n",
        "                                                                z])\n",
        "\n",
        "# 7. Create output layer\n",
        "output_layer = layers.Dense(5, activation=\"softmax\", name=\"output_layer\")(z)\n",
        "\n",
        "# 8. Put together model\n",
        "model_5 = tf.keras.Model(inputs=[line_number_model.input,\n",
        "                                 total_line_model.input,\n",
        "                                 token_model.input, \n",
        "                                 char_model.input],\n",
        "                         outputs=output_layer)"
      ],
      "execution_count": 103,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MH1k7xn0MEDj",
        "outputId": "04d5c42f-563e-4704-c770-d2971fa52de6"
      },
      "source": [
        "model_5.summary()"
      ],
      "execution_count": 104,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Model: \"model_8\"\n",
            "__________________________________________________________________________________________________\n",
            "Layer (type)                    Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            "char_input (InputLayer)         [(None, 1)]          0                                            \n",
            "__________________________________________________________________________________________________\n",
            "token_inputs (InputLayer)       [(None,)]            0                                            \n",
            "__________________________________________________________________________________________________\n",
            "char_vectorizer (TextVectorizat (None, 290)          0           char_input[0][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "universal_sentence_encoder (Ker (None, 512)          256797824   token_inputs[0][0]               \n",
            "__________________________________________________________________________________________________\n",
            "char_embed (Embedding)          (None, 290, 25)      1750        char_vectorizer[2][0]            \n",
            "__________________________________________________________________________________________________\n",
            "dense_7 (Dense)                 (None, 128)          65664       universal_sentence_encoder[2][0] \n",
            "__________________________________________________________________________________________________\n",
            "bidirectional_1 (Bidirectional) (None, 50)           10200       char_embed[2][0]                 \n",
            "__________________________________________________________________________________________________\n",
            "token_char_hybrid_embedding (Co (None, 178)          0           dense_7[0][0]                    \n",
            "                                                                 bidirectional_1[0][0]            \n",
            "__________________________________________________________________________________________________\n",
            "line_number_input (InputLayer)  [(None, 15)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "total_lines_input (InputLayer)  [(None, 20)]         0                                            \n",
            "__________________________________________________________________________________________________\n",
            "dense_10 (Dense)                (None, 256)          45824       token_char_hybrid_embedding[0][0]\n",
            "__________________________________________________________________________________________________\n",
            "dense_8 (Dense)                 (None, 32)           512         line_number_input[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "dense_9 (Dense)                 (None, 32)           672         total_lines_input[0][0]          \n",
            "__________________________________________________________________________________________________\n",
            "dropout_3 (Dropout)             (None, 256)          0           dense_10[0][0]                   \n",
            "__________________________________________________________________________________________________\n",
            "token_char_positional_embedding (None, 320)          0           dense_8[0][0]                    \n",
            "                                                                 dense_9[0][0]                    \n",
            "                                                                 dropout_3[0][0]                  \n",
            "__________________________________________________________________________________________________\n",
            "output_layer (Dense)            (None, 5)            1605        token_char_positional_embedding[0\n",
            "==================================================================================================\n",
            "Total params: 256,924,051\n",
            "Trainable params: 126,227\n",
            "Non-trainable params: 256,797,824\n",
            "__________________________________________________________________________________________________\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 856
        },
        "id": "HgqY5xkJMGIs",
        "outputId": "d47f6b75-eb5c-4c53-a693-31ee7bc956d1"
      },
      "source": [
        "# Plot the token, char, positional embedding model\n",
        "from tensorflow.keras.utils import plot_model\n",
        "plot_model(model_5)"
      ],
      "execution_count": 105,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<IPython.core.display.Image object>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 105
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tpUCGqGtNLJA"
      },
      "source": [
        "model_5.compile(loss =tf.keras.losses.CategoricalCrossentropy(label_smoothing=0.2), optimizer=\"Adam\", metrics= [\"accuracy\"])"
      ],
      "execution_count": 106,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "gt-ZUm80MJWL",
        "outputId": "f07a0350-1eb5-4e60-82f6-14fbbb4e11bd"
      },
      "source": [
        "# Create training and validation datasets (all four kinds of inputs)\n",
        "train_pos_char_token_data = tf.data.Dataset.from_tensor_slices((train_line_numbers_one_hot, # line numbers\n",
        "                                                                train_total_lines_one_hot, # total lines\n",
        "                                                                train_sentences, # train tokens\n",
        "                                                                train_chars)) # train chars\n",
        "train_pos_char_token_labels = tf.data.Dataset.from_tensor_slices(train_labels_one_hot) # train labels\n",
        "train_pos_char_token_dataset = tf.data.Dataset.zip((train_pos_char_token_data, train_pos_char_token_labels)) # combine data and labels\n",
        "train_pos_char_token_dataset = train_pos_char_token_dataset.batch(32).prefetch(tf.data.AUTOTUNE) # turn into batches and prefetch appropriately\n",
        "\n",
        "# Validation dataset\n",
        "val_pos_char_token_data = tf.data.Dataset.from_tensor_slices((val_line_numbers_one_hot,\n",
        "                                                              val_total_lines_one_hot,\n",
        "                                                              val_sentences,\n",
        "                                                              val_chars))\n",
        "val_pos_char_token_labels = tf.data.Dataset.from_tensor_slices(val_labels_one_hot)\n",
        "val_pos_char_token_dataset = tf.data.Dataset.zip((val_pos_char_token_data, val_pos_char_token_labels))\n",
        "val_pos_char_token_dataset = val_pos_char_token_dataset.batch(32).prefetch(tf.data.AUTOTUNE) # turn into batches and prefetch appropriately\n",
        "\n",
        "# Check input shapes\n",
        "train_pos_char_token_dataset, val_pos_char_token_dataset"
      ],
      "execution_count": 107,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(<PrefetchDataset shapes: (((None, 15), (None, 20), (None,), (None,)), (None, 5)), types: ((tf.float32, tf.float32, tf.string, tf.string), tf.float64)>,\n",
              " <PrefetchDataset shapes: (((None, 15), (None, 20), (None,), (None,)), (None, 5)), types: ((tf.float32, tf.float32, tf.string, tf.string), tf.float64)>)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 107
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "-OtxLQ25OG3M",
        "outputId": "c9a12bd7-4585-4e53-d737-caf160d83726"
      },
      "source": [
        "# Fit the token, char and positional embedding model\n",
        "history_model_5 = model_5.fit(train_pos_char_token_dataset,\n",
        "                              steps_per_epoch=int(0.1 * len(train_pos_char_token_dataset)),\n",
        "                              epochs=10,\n",
        "                              validation_data=val_pos_char_token_dataset,\n",
        "                              validation_steps=int(0.1 * len(val_pos_char_token_dataset)))"
      ],
      "execution_count": 109,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n",
            "562/562 [==============================] - 166s 296ms/step - loss: 1.0464 - accuracy: 0.7606 - val_loss: 0.9803 - val_accuracy: 0.8042\n",
            "Epoch 2/10\n",
            "562/562 [==============================] - 166s 296ms/step - loss: 0.9671 - accuracy: 0.8126 - val_loss: 0.9522 - val_accuracy: 0.8235\n",
            "Epoch 3/10\n",
            "562/562 [==============================] - 167s 297ms/step - loss: 0.9502 - accuracy: 0.8240 - val_loss: 0.9400 - val_accuracy: 0.8301\n",
            "Epoch 4/10\n",
            "562/562 [==============================] - 166s 296ms/step - loss: 0.9417 - accuracy: 0.8320 - val_loss: 0.9343 - val_accuracy: 0.8314\n",
            "Epoch 5/10\n",
            "562/562 [==============================] - 166s 296ms/step - loss: 0.9384 - accuracy: 0.8335 - val_loss: 0.9268 - val_accuracy: 0.8378\n",
            "Epoch 6/10\n",
            "562/562 [==============================] - 167s 297ms/step - loss: 0.9414 - accuracy: 0.8290 - val_loss: 0.9242 - val_accuracy: 0.8391\n",
            "Epoch 7/10\n",
            "562/562 [==============================] - 167s 298ms/step - loss: 0.9288 - accuracy: 0.8379 - val_loss: 0.9248 - val_accuracy: 0.8388\n",
            "Epoch 8/10\n",
            "562/562 [==============================] - 166s 296ms/step - loss: 0.9249 - accuracy: 0.8428 - val_loss: 0.9119 - val_accuracy: 0.8507\n",
            "Epoch 9/10\n",
            "562/562 [==============================] - 166s 296ms/step - loss: 0.9270 - accuracy: 0.8385 - val_loss: 0.9236 - val_accuracy: 0.8411\n",
            "Epoch 10/10\n",
            "562/562 [==============================] - 167s 297ms/step - loss: 0.9226 - accuracy: 0.8458 - val_loss: 0.9117 - val_accuracy: 0.8411\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Q_el2kkMOHFI",
        "outputId": "e4fab881-4083-43d4-fcad-eebcd1436fbe"
      },
      "source": [
        "# Make predictions with token-char-positional hybrid model\n",
        "model_5_pred_probs = model_5.predict(val_pos_char_token_dataset, verbose=1)\n",
        "model_5_pred_probs"
      ],
      "execution_count": 110,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "945/945 [==============================] - 60s 62ms/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([[0.55797863, 0.11753111, 0.01437984, 0.28599715, 0.02411333],\n",
              "       [0.5643995 , 0.12752993, 0.05248916, 0.24036951, 0.01521185],\n",
              "       [0.35432863, 0.09137847, 0.1489166 , 0.3426122 , 0.06276409],\n",
              "       ...,\n",
              "       [0.02403956, 0.04742315, 0.01662572, 0.02715172, 0.88475984],\n",
              "       [0.02105152, 0.34086674, 0.04562447, 0.02258216, 0.56987506],\n",
              "       [0.10035124, 0.76991165, 0.05545753, 0.04023053, 0.03404894]],\n",
              "      dtype=float32)"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 110
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qujy9B7EU_M5",
        "outputId": "e458c3d3-83d4-42b2-fedf-9a90cfcae738"
      },
      "source": [
        "# Turn prediction probabilities into prediction classes\n",
        "model_5_preds = tf.argmax(model_5_pred_probs, axis=1)\n",
        "model_5_preds"
      ],
      "execution_count": 111,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(30212,), dtype=int64, numpy=array([0, 0, 0, ..., 4, 4, 1])>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 111
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2t5VGwB8VBZY",
        "outputId": "974650af-67d1-4143-d901-97297fb18056"
      },
      "source": [
        "# Calculate results of token-char-positional hybrid model\n",
        "model_5_results = calculate_results(y_true=val_labels_encoded,\n",
        "                                    y_pred=model_5_preds)\n",
        "model_5_results"
      ],
      "execution_count": 112,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 84.80074142724744,\n",
              " 'f1': 0.8445848608307035,\n",
              " 'precision': 0.8499109192748401,\n",
              " 'recall': 0.8480074142724745}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 112
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 223
        },
        "id": "ymq0rtJ_VMUM",
        "outputId": "1f25a437-6bff-4813-bc97-671a2980a29a"
      },
      "source": [
        "# Combine model results into a DataFrame\n",
        "all_model_results = pd.DataFrame({\"baseline\": baseline_results,\n",
        "                                  \"custom_token_embed_conv1d\": model_1_results,\n",
        "                                  \"pretrained_token_embed\": model_2_results,\n",
        "                                  \"custom_char_embed_conv1d\": model_3_results,\n",
        "                                  \"hybrid_char_token_embed\": model_4_results,\n",
        "                                  \"tribrid_pos_char_token_embed\": model_5_results})\n",
        "all_model_results = all_model_results.transpose()\n",
        "all_model_results"
      ],
      "execution_count": 113,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>accuracy</th>\n",
              "      <th>precision</th>\n",
              "      <th>recall</th>\n",
              "      <th>f1</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>baseline</th>\n",
              "      <td>72.183238</td>\n",
              "      <td>0.718647</td>\n",
              "      <td>0.721832</td>\n",
              "      <td>0.698925</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>custom_token_embed_conv1d</th>\n",
              "      <td>80.828810</td>\n",
              "      <td>0.804780</td>\n",
              "      <td>0.808288</td>\n",
              "      <td>0.805473</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>pretrained_token_embed</th>\n",
              "      <td>75.105918</td>\n",
              "      <td>0.745737</td>\n",
              "      <td>0.751059</td>\n",
              "      <td>0.746384</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>custom_char_embed_conv1d</th>\n",
              "      <td>66.933669</td>\n",
              "      <td>0.663860</td>\n",
              "      <td>0.669337</td>\n",
              "      <td>0.659734</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>hybrid_char_token_embed</th>\n",
              "      <td>75.983053</td>\n",
              "      <td>0.756189</td>\n",
              "      <td>0.759831</td>\n",
              "      <td>0.754541</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>tribrid_pos_char_token_embed</th>\n",
              "      <td>84.800741</td>\n",
              "      <td>0.849911</td>\n",
              "      <td>0.848007</td>\n",
              "      <td>0.844585</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "                               accuracy  precision    recall        f1\n",
              "baseline                      72.183238   0.718647  0.721832  0.698925\n",
              "custom_token_embed_conv1d     80.828810   0.804780  0.808288  0.805473\n",
              "pretrained_token_embed        75.105918   0.745737  0.751059  0.746384\n",
              "custom_char_embed_conv1d      66.933669   0.663860  0.669337  0.659734\n",
              "hybrid_char_token_embed       75.983053   0.756189  0.759831  0.754541\n",
              "tribrid_pos_char_token_embed  84.800741   0.849911  0.848007  0.844585"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 113
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7P2HD5dwVVNM"
      },
      "source": [
        "# Reduce the accuracy to same scale as other metrics\n",
        "all_model_results[\"accuracy\"] = all_model_results[\"accuracy\"]/100"
      ],
      "execution_count": 114,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 571
        },
        "id": "MQ1L_wCtVbd8",
        "outputId": "b536edd4-6c6d-4db3-9da5-946efbe3a8a9"
      },
      "source": [
        "# Plot and compare all of the model results\n",
        "all_model_results.plot(kind=\"bar\", figsize=(10, 7)).legend(bbox_to_anchor=(1.0, 1.0));"
      ],
      "execution_count": 115,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 571
        },
        "id": "KDL8UCgCWSAx",
        "outputId": "b5cb294f-20b7-4486-e088-748013b23adc"
      },
      "source": [
        "# Sort model results by f1-score\n",
        "all_model_results.sort_values(\"f1\", ascending=False)[\"f1\"].plot(kind=\"bar\", figsize=(10, 7));"
      ],
      "execution_count": 116,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 720x504 with 1 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "X7lhYQE_WSP6",
        "outputId": "1c93b3c3-caf9-49e4-f0b7-fc8a3ca494f6"
      },
      "source": [
        "# Save best performing model to SavedModel format (default)\n",
        "model_5.save(\"tribrid_model\") "
      ],
      "execution_count": 118,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "INFO:tensorflow:Assets written to: tribrid_model/assets\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "stream",
          "text": [
            "INFO:tensorflow:Assets written to: tribrid_model/assets\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "-_VSn9I9YOAD"
      },
      "source": [
        "model_path = \"/content/tribrid_model\""
      ],
      "execution_count": 119,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Lxc6CbGJYH_-"
      },
      "source": [
        "loaded_model = tf.keras.models.load_model(model_path)"
      ],
      "execution_count": 120,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "jCnGiPpuXqO0",
        "outputId": "dd193eb9-fe59-4813-b5f9-05ccad11be8f"
      },
      "source": [
        "loaded_pred_probs = loaded_model.predict(val_pos_char_token_dataset, verbose=1)\n",
        "loaded_preds = tf.argmax(loaded_pred_probs, axis=1)\n",
        "loaded_preds[:10]"
      ],
      "execution_count": 121,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "945/945 [==============================] - 61s 63ms/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(10,), dtype=int64, numpy=array([0, 0, 0, 2, 2, 4, 4, 4, 4, 1])>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 121
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "GNZRIYb0YsHk",
        "outputId": "c3d3466f-e554-48f4-fdc4-6bce8075d5a0"
      },
      "source": [
        "# Evaluate loaded model's predictions\n",
        "loaded_model_results = calculate_results(val_labels_encoded,\n",
        "                                         loaded_preds)\n",
        "loaded_model_results"
      ],
      "execution_count": 122,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 84.80074142724744,\n",
              " 'f1': 0.8445848608307035,\n",
              " 'precision': 0.8499109192748401,\n",
              " 'recall': 0.8480074142724745}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 122
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Yj1DhGDhZAJo"
      },
      "source": [
        "### Evaluate model on test dataset\n",
        "\n",
        "To make our model's performance more comparable with the results reported in Table 3 of the [*PubMed 200k RCT:\n",
        "a Dataset for Sequential Sentence Classification in Medical Abstracts*](https://arxiv.org/pdf/1710.06071.pdf) paper, let's make predictions on the test dataset and evaluate them."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "AB3tHOk3YzLH",
        "outputId": "f884ad3f-55e5-4bc9-facc-22b5caf1701a"
      },
      "source": [
        "# Create test dataset batch and prefetched\n",
        "test_pos_char_token_data = tf.data.Dataset.from_tensor_slices((test_line_numbers_one_hot,\n",
        "                                                               test_total_lines_one_hot,\n",
        "                                                               test_sentences,\n",
        "                                                               test_chars))\n",
        "test_pos_char_token_labels = tf.data.Dataset.from_tensor_slices(test_labels_one_hot)\n",
        "test_pos_char_token_dataset = tf.data.Dataset.zip((test_pos_char_token_data, test_pos_char_token_labels))\n",
        "test_pos_char_token_dataset = test_pos_char_token_dataset.batch(32).prefetch(tf.data.AUTOTUNE)\n",
        "\n",
        "# Check shapes\n",
        "test_pos_char_token_dataset"
      ],
      "execution_count": 123,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<PrefetchDataset shapes: (((None, 15), (None, 20), (None,), (None,)), (None, 5)), types: ((tf.float32, tf.float32, tf.string, tf.string), tf.float64)>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 123
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Tp5QqjxsZVBu",
        "outputId": "70ce4738-9c30-4ad2-853e-37a7a319215e"
      },
      "source": [
        "# Make predictions on the test dataset\n",
        "test_pred_probs = loaded_model.predict(test_pos_char_token_dataset,\n",
        "                                       verbose=1)\n",
        "test_preds = tf.argmax(test_pred_probs, axis=1)\n",
        "test_preds[:10]"
      ],
      "execution_count": 124,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "942/942 [==============================] - 58s 62ms/step\n"
          ],
          "name": "stdout"
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<tf.Tensor: shape=(10,), dtype=int64, numpy=array([3, 2, 2, 2, 4, 4, 4, 1, 1, 0])>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 124
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dcMxok1vZVQ0",
        "outputId": "ad17d2f9-8021-4171-c457-f78f729f7fce"
      },
      "source": [
        "# Evaluate loaded model test predictions\n",
        "loaded_model_test_results = calculate_results(y_true=test_labels_encoded,\n",
        "                                              y_pred=test_preds)\n",
        "loaded_model_test_results"
      ],
      "execution_count": 125,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "{'accuracy': 84.48979591836735,\n",
              " 'f1': 0.8415058973152829,\n",
              " 'precision': 0.8456318536204341,\n",
              " 'recall': 0.8448979591836735}"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 125
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kPYtdB-WaHqQ",
        "outputId": "2c99e02a-b394-49e6-cf55-6221a7b23b27"
      },
      "source": [
        "%%time\n",
        "# Get list of class names of test predictions\n",
        "test_pred_classes = [label_encoder.classes_[pred] for pred in test_preds]\n",
        "test_pred_classes"
      ],
      "execution_count": 127,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "CPU times: user 3.43 s, sys: 10 ms, total: 3.44 s\n",
            "Wall time: 3.44 s\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 634
        },
        "id": "yfoqq3fzZ0p-",
        "outputId": "40e2836f-6905-4e8c-b01b-a16133abdcdd"
      },
      "source": [
        "# Create prediction-enriched test dataframe\n",
        "test_df[\"prediction\"] = test_pred_classes # create column with test prediction class names\n",
        "test_df[\"pred_prob\"] = tf.reduce_max(test_pred_probs, axis=1).numpy() # get the maximum prediction probability\n",
        "test_df[\"correct\"] = test_df[\"prediction\"] == test_df[\"target\"] # create binary column for whether the prediction is right or not\n",
        "test_df.head(20)"
      ],
      "execution_count": 128,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/html": [
              "<div>\n",
              "<style scoped>\n",
              "    .dataframe tbody tr th:only-of-type {\n",
              "        vertical-align: middle;\n",
              "    }\n",
              "\n",
              "    .dataframe tbody tr th {\n",
              "        vertical-align: top;\n",
              "    }\n",
              "\n",
              "    .dataframe thead th {\n",
              "        text-align: right;\n",
              "    }\n",
              "</style>\n",
              "<table border=\"1\" class=\"dataframe\">\n",
              "  <thead>\n",
              "    <tr style=\"text-align: right;\">\n",
              "      <th></th>\n",
              "      <th>target</th>\n",
              "      <th>text</th>\n",
              "      <th>line_number</th>\n",
              "      <th>total_lines</th>\n",
              "      <th>prediction</th>\n",
              "      <th>pred_prob</th>\n",
              "      <th>correct</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>0</th>\n",
              "      <td>BACKGROUND</td>\n",
              "      <td>this study analyzed liver function abnormaliti...</td>\n",
              "      <td>0</td>\n",
              "      <td>8</td>\n",
              "      <td>OBJECTIVE</td>\n",
              "      <td>0.359128</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>1</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>a post hoc analysis was conducted with the use...</td>\n",
              "      <td>1</td>\n",
              "      <td>8</td>\n",
              "      <td>METHODS</td>\n",
              "      <td>0.395141</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>2</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>liver function tests ( lfts ) were measured at...</td>\n",
              "      <td>2</td>\n",
              "      <td>8</td>\n",
              "      <td>METHODS</td>\n",
              "      <td>0.857965</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>3</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>survival analyses were used to assess the asso...</td>\n",
              "      <td>3</td>\n",
              "      <td>8</td>\n",
              "      <td>METHODS</td>\n",
              "      <td>0.733716</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>4</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>the percentage of patients with abnormal lfts ...</td>\n",
              "      <td>4</td>\n",
              "      <td>8</td>\n",
              "      <td>RESULTS</td>\n",
              "      <td>0.734855</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>5</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>when mean hemodynamic profiles were compared i...</td>\n",
              "      <td>5</td>\n",
              "      <td>8</td>\n",
              "      <td>RESULTS</td>\n",
              "      <td>0.888223</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>6</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>multivariable analyses revealed that patients ...</td>\n",
              "      <td>6</td>\n",
              "      <td>8</td>\n",
              "      <td>RESULTS</td>\n",
              "      <td>0.569608</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>7</th>\n",
              "      <td>CONCLUSIONS</td>\n",
              "      <td>abnormal lfts are common in the adhf populatio...</td>\n",
              "      <td>7</td>\n",
              "      <td>8</td>\n",
              "      <td>CONCLUSIONS</td>\n",
              "      <td>0.476394</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>8</th>\n",
              "      <td>CONCLUSIONS</td>\n",
              "      <td>elevated meld-xi scores are associated with po...</td>\n",
              "      <td>8</td>\n",
              "      <td>8</td>\n",
              "      <td>CONCLUSIONS</td>\n",
              "      <td>0.583600</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>9</th>\n",
              "      <td>BACKGROUND</td>\n",
              "      <td>minimally invasive endovascular aneurysm repai...</td>\n",
              "      <td>0</td>\n",
              "      <td>12</td>\n",
              "      <td>BACKGROUND</td>\n",
              "      <td>0.633004</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>10</th>\n",
              "      <td>BACKGROUND</td>\n",
              "      <td>the aim of this study was to analyse the cost-...</td>\n",
              "      <td>1</td>\n",
              "      <td>12</td>\n",
              "      <td>OBJECTIVE</td>\n",
              "      <td>0.445351</td>\n",
              "      <td>False</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>11</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>resource use was determined from the amsterdam...</td>\n",
              "      <td>2</td>\n",
              "      <td>12</td>\n",
              "      <td>METHODS</td>\n",
              "      <td>0.773279</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>12</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>the analysis was performed from a provider per...</td>\n",
              "      <td>3</td>\n",
              "      <td>12</td>\n",
              "      <td>METHODS</td>\n",
              "      <td>0.865364</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>13</th>\n",
              "      <td>METHODS</td>\n",
              "      <td>all costs were calculated as if all patients h...</td>\n",
              "      <td>4</td>\n",
              "      <td>12</td>\n",
              "      <td>METHODS</td>\n",
              "      <td>0.490311</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>14</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>a total of @ patients were randomized .</td>\n",
              "      <td>5</td>\n",
              "      <td>12</td>\n",
              "      <td>RESULTS</td>\n",
              "      <td>0.788061</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>15</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>the @-day mortality rate was @ per cent after ...</td>\n",
              "      <td>6</td>\n",
              "      <td>12</td>\n",
              "      <td>RESULTS</td>\n",
              "      <td>0.816789</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>16</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>at @months , the total mortality rate for evar...</td>\n",
              "      <td>7</td>\n",
              "      <td>12</td>\n",
              "      <td>RESULTS</td>\n",
              "      <td>0.890838</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>17</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>the mean cost difference between evar and or w...</td>\n",
              "      <td>8</td>\n",
              "      <td>12</td>\n",
              "      <td>RESULTS</td>\n",
              "      <td>0.881075</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>18</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>the incremental cost-effectiveness ratio per p...</td>\n",
              "      <td>9</td>\n",
              "      <td>12</td>\n",
              "      <td>RESULTS</td>\n",
              "      <td>0.832961</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>19</th>\n",
              "      <td>RESULTS</td>\n",
              "      <td>there was no significant difference in quality...</td>\n",
              "      <td>10</td>\n",
              "      <td>12</td>\n",
              "      <td>RESULTS</td>\n",
              "      <td>0.817365</td>\n",
              "      <td>True</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>"
            ],
            "text/plain": [
              "         target  ... correct\n",
              "0    BACKGROUND  ...   False\n",
              "1       RESULTS  ...   False\n",
              "2       RESULTS  ...   False\n",
              "3       RESULTS  ...   False\n",
              "4       RESULTS  ...    True\n",
              "5       RESULTS  ...    True\n",
              "6       RESULTS  ...    True\n",
              "7   CONCLUSIONS  ...    True\n",
              "8   CONCLUSIONS  ...    True\n",
              "9    BACKGROUND  ...    True\n",
              "10   BACKGROUND  ...   False\n",
              "11      METHODS  ...    True\n",
              "12      METHODS  ...    True\n",
              "13      METHODS  ...    True\n",
              "14      RESULTS  ...    True\n",
              "15      RESULTS  ...    True\n",
              "16      RESULTS  ...    True\n",
              "17      RESULTS  ...    True\n",
              "18      RESULTS  ...    True\n",
              "19      RESULTS  ...    True\n",
              "\n",
              "[20 rows x 7 columns]"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 128
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2Z0VT8c3Z_uM"
      },
      "source": [
        ""
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}