practice_examples/05_Sentiment_Deep_Learning_UPDATE23.ipynb

{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ltJPhVtkFsfH",
        "outputId": "b0277141-ec29-4c8d-e188-42073c6a7271"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "--2023-10-20 14:15:45--  https://docs.google.com/uc?export=download&confirm=t&id=1krZbtFXLW2711hZLq7xO0iXsUTBcwNSQ\n",
            "Resolving docs.google.com (docs.google.com)... 108.177.126.139, 108.177.126.101, 108.177.126.113, ...\n",
            "Connecting to docs.google.com (docs.google.com)|108.177.126.139|:443... connected.\n",
            "HTTP request sent, awaiting response... 303 See Other\n",
            "Location: https://doc-0s-2k-docs.googleusercontent.com/docs/securesc/ha0ro937gcuc7l7deffksulhg5h7mbp1/nb4ar9a8nvqdvdt7ahgutg29h1madck8/1697811300000/01308971736394707036/*/1krZbtFXLW2711hZLq7xO0iXsUTBcwNSQ?e=download&uuid=064b4c60-f9e9-4957-b622-c3a168c4a4c8 [following]\n",
            "Warning: wildcards not supported in HTTP.\n",
            "--2023-10-20 14:15:45--  https://doc-0s-2k-docs.googleusercontent.com/docs/securesc/ha0ro937gcuc7l7deffksulhg5h7mbp1/nb4ar9a8nvqdvdt7ahgutg29h1madck8/1697811300000/01308971736394707036/*/1krZbtFXLW2711hZLq7xO0iXsUTBcwNSQ?e=download&uuid=064b4c60-f9e9-4957-b622-c3a168c4a4c8\n",
            "Resolving doc-0s-2k-docs.googleusercontent.com (doc-0s-2k-docs.googleusercontent.com)... 142.250.153.132, 2a00:1450:4013:c16::84\n",
            "Connecting to doc-0s-2k-docs.googleusercontent.com (doc-0s-2k-docs.googleusercontent.com)|142.250.153.132|:443... connected.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 636691366 (607M) [application/rar]\n",
            "Saving to: ‘/content/data.rar’\n",
            "\n",
            "/content/data.rar   100%[===================>] 607.20M  60.8MB/s    in 7.2s    \n",
            "\n",
            "2023-10-20 14:15:53 (84.9 MB/s) - ‘/content/data.rar’ saved [636691366/636691366]\n",
            "\n",
            "\n",
            "UNRAR 6.11 beta 1 freeware      Copyright (c) 1993-2022 Alexander Roshal\n",
            "\n",
            "\n",
            "Extracting from /content/data.rar\n",
            "\n",
            "Creating    /content/data                                             OK\n",
            "Creating    /content/data/03Emoji                                     OK\n",
            "Extracting  /content/data/03Emoji/cc.en.50.magnitude                     \b\b\b\b  0%\b\b\b\b  1%\b\b\b\b  2%\b\b\b\b  3%\b\b\b\b  4%\b\b\b\b  5%\b\b\b\b  6%\b\b\b\b  7%\b\b\b\b  8%\b\b\b\b  9%\b\b\b\b 10%\b\b\b\b 11%\b\b\b\b 12%\b\b\b\b 13%\b\b\b\b 14%\b\b\b\b 15%\b\b\b\b 16%\b\b\b\b 17%\b\b\b\b 18%\b\b\b\b 19%\b\b\b\b 20%\b\b\b\b 21%\b\b\b\b 22%\b\b\b\b 23%\b\b\b\b 24%\b\b\b\b 25%\b\b\b\b 26%\b\b\b\b 27%\b\b\b\b 28%\b\b\b\b 29%\b\b\b\b 30%\b\b\b\b 31%\b\b\b\b 32%\b\b\b\b 33%\b\b\b\b 34%\b\b\b\b 35%\b\b\b\b 36%\b\b\b\b 37%\b\b\b\b 38%\b\b\b\b 39%\b\b\b\b 40%\b\b\b\b 41%\b\b\b\b 42%\b\b\b\b 43%\b\b\b\b 44%\b\b\b\b 45%\b\b\b\b 46%\b\b\b\b 47%\b\b\b\b 48%\b\b\b\b 49%\b\b\b\b 50%\b\b\b\b 51%\b\b\b\b 52%\b\b\b\b 53%\b\b\b\b 54%\b\b\b\b 55%\b\b\b\b 56%\b\b\b\b 57%\b\b\b\b 58%\b\b\b\b 59%\b\b\b\b 60%\b\b\b\b 61%\b\b\b\b 62%\b\b\b\b 63%\b\b\b\b 64%\b\b\b\b 65%\b\b\b\b 66%\b\b\b\b 67%\b\b\b\b 68%\b\b\b\b 69%\b\b\b\b 70%\b\b\b\b 71%\b\b\b\b 72%\b\b\b\b 73%\b\b\b\b 74%\b\b\b\b 75%\b\b\b\b 76%\b\b\b\b 77%\b\b\b\b 78%\b\b\b\b 79%\b\b\b\b 80%\b\b\b\b 81%\b\b\b\b 82%\b\b\b\b 83%\b\b\b\b 84%\b\b\b\b 85%\b\b\b\b 86%\b\b\b\b 87%\b\b\b\b 88%\b\b\b\b 89%\b\b\b\b 90%\b\b\b\b 91%\b\b\b\b 92%\b\b\b\b 93%\b\b\b\b 94%\b\b\b\b 95%\b\b\b\b 96%\b\b\b\b 97%\b\b\b\b 98%\b\b\b\b 99%\b\b\b\b\b  OK \n",
            "Extracting  /content/data/03Emoji/test.txt                               \b\b\b\b 99%\b\b\b\b\b  OK \n",
            "Extracting  /content/data/03Emoji/train.txt                              \b\b\b\b 99%\b\b\b\b\b  OK \n",
            "Extracting  /content/data/03Emoji/val.txt                                \b\b\b\b 99%\b\b\b\b\b  OK \n",
            "All OK\n",
            "--2023-10-20 14:16:16--  https://ikelte-my.sharepoint.com/:u:/g/personal/natabara_inf_elte_hu/Ee9HETZu5o1BkP8EqaKtMigB_KjvuD4vQ5WpoDOG4gLISw?e=oNbmMt&download=1\n",
            "Resolving ikelte-my.sharepoint.com (ikelte-my.sharepoint.com)... 13.107.136.10, 13.107.138.10, 2620:1ec:8f8::10, ...\n",
            "Connecting to ikelte-my.sharepoint.com (ikelte-my.sharepoint.com)|13.107.136.10|:443... connected.\n",
            "HTTP request sent, awaiting response... 302 Found\n",
            "Location: /personal/natabara_inf_elte_hu/Documents/cc.en.64.bin?ga=1 [following]\n",
            "--2023-10-20 14:16:17--  https://ikelte-my.sharepoint.com/personal/natabara_inf_elte_hu/Documents/cc.en.64.bin?ga=1\n",
            "Reusing existing connection to ikelte-my.sharepoint.com:443.\n",
            "HTTP request sent, awaiting response... 200 OK\n",
            "Length: 1573176312 (1.5G) [application/octet-stream]\n",
            "Saving to: ‘cc.en.64.bin’\n",
            "\n",
            "cc.en.64.bin        100%[===================>]   1.46G   199MB/s    in 12s     \n",
            "\n",
            "2023-10-20 14:16:29 (128 MB/s) - ‘cc.en.64.bin’ saved [1573176312/1573176312]\n",
            "\n"
          ]
        }
      ],
      "source": [
        "import locale\n",
        "locale.getpreferredencoding = lambda do_setlocale=False: \"UTF-8\"\n",
        "\n",
        "def download_from_drive(link, filename):\n",
        "  id = link.split(\"/\")[-2]\n",
        "  !wget --load-cookies /tmp/cookies.txt \"https://docs.google.com/uc?export=download&confirm=$(wget --quiet --save-cookies /tmp/cookies.txt --keep-session-cookies --no-check-certificate 'https://docs.google.com/uc?export=download&id={id}' -O- | sed -rn 's/.*confirm=([0-9A-Za-z_]+).*/\\1\\n/p')&id={id}\" -O {\"/content/\"+filename+\".rar\"} && rm -rf /tmp/cookies.txt\n",
        "  !unrar x -Y {\"/content/\"+filename+\".rar\"} {\"/content/\"+filename+\"/\"}\n",
        "\n",
        "download_from_drive(\"https://drive.google.com/file/d/1krZbtFXLW2711hZLq7xO0iXsUTBcwNSQ/view?usp=sharing\", \"data\")\n",
        "\n",
        "!wget --no-check-certificate \"https://ikelte-my.sharepoint.com/:u:/g/personal/natabara_inf_elte_hu/Ee9HETZu5o1BkP8EqaKtMigB_KjvuD4vQ5WpoDOG4gLISw?e=oNbmMt&download=1\" -O cc.en.64.bin"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install gensim"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MRywDKBMMeFW",
        "outputId": "6940f81a-fb12-4668-d426-8c2116f85360"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Requirement already satisfied: gensim in /usr/local/lib/python3.10/dist-packages (4.3.2)\n",
            "Requirement already satisfied: numpy>=1.18.5 in /usr/local/lib/python3.10/dist-packages (from gensim) (1.23.5)\n",
            "Requirement already satisfied: scipy>=1.7.0 in /usr/local/lib/python3.10/dist-packages (from gensim) (1.11.3)\n",
            "Requirement already satisfied: smart-open>=1.8.1 in /usr/local/lib/python3.10/dist-packages (from gensim) (6.4.0)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:40:47.503357Z",
          "start_time": "2021-05-08T09:40:45.731064Z"
        },
        "id": "Nt8U_O0DGDwK"
      },
      "outputs": [],
      "source": [
        "import numpy as np\n",
        "import tensorflow as tf\n",
        "physical_devices = tf.config.list_physical_devices('GPU')\n",
        "tf.config.experimental.set_memory_growth(physical_devices[0], True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "35ZeTEwuGDwK"
      },
      "source": [
        "# Emoji prediction\n",
        "\n",
        "### Plan:\n",
        "- Keras (TensorFlow basics)\n",
        "- Read data\n",
        "- Examine Data\n",
        "- Apply Fasttext embedding (word2vec fundamentals)\n",
        "- Splitting data\n",
        "- Modelling with LSTM"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:40:48.366061Z",
          "start_time": "2021-05-08T09:40:48.236408Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "UVSX2xs4GDwL",
        "outputId": "04b4aa38-2c97-434b-9160-f1377c7c3ae8"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[['i', 'didnt', 'feel', 'humiliated'], ['i', 'can', 'go', 'from', 'feeling', 'so', 'hopeless', 'to', 'so', 'damned', 'hopeful', 'just', 'from', 'being', 'around', 'someone', 'who', 'cares', 'and', 'is', 'awake'], ['im', 'grabbing', 'a', 'minute', 'to', 'post', 'i', 'feel', 'greedy', 'wrong'], ['i', 'am', 'ever', 'feeling', 'nostalgic', 'about', 'the', 'fireplace', 'i', 'will', 'know', 'that', 'it', 'is', 'still', 'on', 'the', 'property'], ['i', 'am', 'feeling', 'grouchy']]\n",
            "['sadness', 'sadness', 'anger', 'love', 'anger']\n",
            "66\n"
          ]
        }
      ],
      "source": [
        "def csvloader(path):\n",
        "    with open(path,\"r\") as f:\n",
        "        lines = f.readlines()\n",
        "\n",
        "    numlines = len(lines)\n",
        "    data = []\n",
        "    label = []\n",
        "    for line in lines:\n",
        "        data.append((line.split(\";\")[0]).split(\" \"))\n",
        "        label.append(line.split(\";\")[-1].replace(\"\\n\",\"\"))\n",
        "    return data, label\n",
        "\n",
        "train_data, train_label = csvloader(\"./data/03Emoji/train.txt\")\n",
        "val_data, val_label = csvloader(\"./data/03Emoji/val.txt\")\n",
        "test_data, test_label = csvloader(\"./data/03Emoji/test.txt\")\n",
        "\n",
        "print(train_data[:5])\n",
        "print(train_label[:5])\n",
        "\n",
        "maxlen = 0\n",
        "for i in range(len(train_data)):\n",
        "    sentlen = len(train_data[i])\n",
        "    if sentlen>maxlen:\n",
        "        maxlen=sentlen\n",
        "\n",
        "print(maxlen)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "w1lNbC2X_-eo"
      },
      "source": [
        "im feeling rather rotten so im not very ambitious right now;sadness\n",
        "\n",
        "\n",
        "im updating my blog because i feel shitty;sadness\n",
        "\n",
        "\n",
        "i never make her separate from me because i don t ever want her to feel like i m ashamed with her;sadness\n",
        "\n",
        "\n",
        "i left with my bouquet of red and yellow tulips under my arm feeling slightly more optimistic than when i arrived;joy\n",
        "\n",
        "\n",
        "i was feeling a little vain when i did this one;sadness"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:40:48.945018Z",
          "start_time": "2021-05-08T09:40:48.917093Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "WP4BGa64GDwL",
        "outputId": "fa55e6f2-b3f9-44ac-cbbb-912002cb7b49"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[('sadness', 4666), ('anger', 2159), ('love', 1304), ('surprise', 572), ('fear', 1937), ('joy', 5362)]\n",
            "{'sadness': 0, 'anger': 1, 'love': 2, 'surprise': 3, 'fear': 4, 'joy': 5}\n"
          ]
        }
      ],
      "source": [
        "from collections import Counter\n",
        "\n",
        "print(list(zip(Counter(train_label).keys(), Counter(train_label).values())))\n",
        "\n",
        "catlist = list(Counter(train_label).keys())\n",
        "catDict = {}\n",
        "catnum = len(catlist)\n",
        "for i in range(catnum):\n",
        "    catDict[catlist[i]]=i\n",
        "print(catDict)\n",
        "\n",
        "y_train = np.array([catDict[emoji] for emoji in train_label])\n",
        "y_val = np.array([catDict[emoji] for emoji in val_label])\n",
        "y_test = np.array([catDict[emoji] for emoji in test_label])"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "y_train[:4]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ksnlFrT0rnOv",
        "outputId": "a5521864-03fe-44f5-efd2-55f72861b832"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "array([4, 0, 5, 5])"
            ]
          },
          "metadata": {},
          "execution_count": 69
        }
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:40:49.374385Z",
          "start_time": "2021-05-08T09:40:49.371393Z"
        },
        "id": "fJL9fK5sGDwL"
      },
      "outputs": [],
      "source": [
        "class_weight = {0: 1.0, 1: 2.0, 2: 3.0, 3: 8.0, 4: 2.3, 5: 1.0}"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:40:59.861017Z",
          "start_time": "2021-05-08T09:40:50.240092Z"
        },
        "id": "cHHEhFL4GDwM",
        "scrolled": true
      },
      "outputs": [],
      "source": [
        "import spacy"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:40:59.871988Z",
          "start_time": "2021-05-08T09:40:59.862015Z"
        },
        "id": "hZIwoVFwGDwM"
      },
      "outputs": [],
      "source": [
        "from gensim.models.fasttext import load_facebook_model\n",
        "\n",
        "vectors = load_facebook_model(\"./cc.en.64.bin\").wv"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "<img src=\"https://miro.medium.com/max/3010/1*OEmWDt4eztOcm5pr2QbxfA.png\" width=90%>"
      ],
      "metadata": {
        "id": "LSZR92LvsJm_"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:41:42.516084Z",
          "start_time": "2021-05-08T09:40:59.872985Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CHYpkZklGDwN",
        "outputId": "24c7bd99-b386-4a6e-ed36-bb9d089a43e1"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "[('king', 0.865199863910675)]\n",
            "[('puppy', 0.916371762752533)]\n",
            "[('Barcelona', 0.8969185948371887), ('Madrid', 0.8903045654296875), ('Amsterdam', 0.8504062294960022)]\n",
            "[('Thames', 0.8160749673843384), ('Battersea', 0.7844918370246887), ('Castlereagh', 0.774276614189148)]\n",
            "[('divorce', 0.7760174870491028), ('abortion', 0.7623156309127808), ('ostracization', 0.7379628419876099)]\n",
            "[('winter', 0.8505227565765381), ('mid-winter', 0.8278099298477173), ('mid-summer', 0.7899753451347351)]\n"
          ]
        }
      ],
      "source": [
        "result_vec = vectors.get_vector(\"queen\",norm=True)-vectors.get_vector(\"woman\",norm=True)+vectors.get_vector(\"man\",norm=True)\n",
        "print(vectors.most_similar(result_vec, topn = 1))\n",
        "\n",
        "result_vec = vectors.get_vector(\"kitten\",norm=True)-vectors.get_vector(\"cat\",norm=True)+vectors.get_vector(\"dog\",norm=True)\n",
        "result_vec /= np.linalg.norm(result_vec)\n",
        "print(vectors.most_similar(result_vec, topn = 1))\n",
        "\n",
        "print(vectors.most_similar(positive=[\"Budapest\",\"Spain\"], negative=\"Hungary\", topn=3))\n",
        "\n",
        "print(vectors.most_similar(positive=[\"Danube\",\"London\"], negative=\"Budapest\", topn=3))\n",
        "\n",
        "print(vectors.most_similar(positive=[\"marriage\",\"hate\"], negative=\"love\", topn=3))\n",
        "\n",
        "print(vectors.most_similar(positive=[\"summer\",\"cold\"], negative=\"hot\", topn=3))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AEj19nSuGDwN"
      },
      "source": [
        "- Coding the data to vectors"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def query_all_vectors(data, pad_to_length = None):\n",
        "  vecs = []\n",
        "  for datarow in data:\n",
        "    vecrow = []\n",
        "    for word in datarow:\n",
        "      vecrow.append(vectors.get_vector(word, norm=False))\n",
        "    if pad_to_length is not None:\n",
        "      diff = pad_to_length-len(vecrow)\n",
        "      if diff>0:\n",
        "        vecrow += [np.zeros_like(vecrow[-1])]*diff\n",
        "      elif diff < 0:\n",
        "        vecrow[-1] = vecrow[-1][:diff]\n",
        "    vecs.append(np.stack(vecrow))\n",
        "\n",
        "  return np.stack(vecs)"
      ],
      "metadata": {
        "id": "MNKd2zjAO-N2"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:42:34.297966Z",
          "start_time": "2021-05-08T09:41:42.517082Z"
        },
        "id": "juRS-M-OGDwN",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "fb718f2f-7d32-4e87-b9ff-ea0373a3dba5"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:gensim.models.fasttext:could not extract any ngrams from 'zq', returning origin vector\n"
          ]
        }
      ],
      "source": [
        "x_train = query_all_vectors(train_data, pad_to_length=70)\n",
        "x_val = query_all_vectors(val_data, pad_to_length=70)\n",
        "x_test = query_all_vectors(test_data, pad_to_length=70)"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "x_train.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_fjK3Q7rowuB",
        "outputId": "dac129c1-41e4-40dd-eb98-a561be06a7c6"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(16000, 70, 64)"
            ]
          },
          "metadata": {},
          "execution_count": 49
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "<img src=https://www.researchgate.net/profile/Savvas-Varsamopoulos/publication/329362532/figure/fig5/AS:699592479870977@1543807253596/Structure-of-the-LSTM-cell-and-equations-that-describe-the-gates-of-an-LSTM-cell.jpg>\n",
        "\n",
        "<img src=https://www.baeldung.com/wp-content/uploads/sites/4/2022/01/bilstm-1-1024x384.png>"
      ],
      "metadata": {
        "id": "Z51FqxH4pcsm"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:42:34.424628Z",
          "start_time": "2021-05-08T09:42:34.421636Z"
        },
        "id": "MqXMUWZSGDwO"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.models import Model\n",
        "from tensorflow.keras.layers import Input, LSTM, Bidirectional, Dense, GlobalMaxPooling1D\n",
        "from tensorflow.keras.optimizers import Adam\n",
        "from tensorflow.keras.callbacks import History, EarlyStopping"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:42:35.369123Z",
          "start_time": "2021-05-08T09:42:34.425625Z"
        },
        "id": "YH5cdM3aGDwO"
      },
      "outputs": [],
      "source": [
        "inp = Input(shape=(70, 64)) # (batch_size), seq_len, seq_dim\n",
        "x = Bidirectional(LSTM(128, dropout=0.1, return_sequences=True))(inp) # (batch_size), 70, 128\n",
        "x = Bidirectional(LSTM(64, dropout=0.1, return_sequences=True))(x) # (batch_size), 128\n",
        "x = GlobalMaxPooling1D()(x)\n",
        "x = Dense(32, activation='tanh')(x) # (batch_size), 32\n",
        "x = Dense(6, activation='softmax')(x) # (batch_size), 6\n",
        "model = Model([inp],[x])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:42:35.378099Z",
          "start_time": "2021-05-08T09:42:35.370121Z"
        },
        "id": "b-52yz7_GDwO"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer=Adam(learning_rate=1e-4),loss=\"sparse_categorical_crossentropy\",metrics=[\"accuracy\"])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:42:35.382089Z",
          "start_time": "2021-05-08T09:42:35.379097Z"
        },
        "id": "FPDhC3WNGDwO"
      },
      "outputs": [],
      "source": [
        "hist = History()\n",
        "earlystop = EarlyStopping(monitor='val_loss', min_delta=2e-4, patience=20, verbose=1,\n",
        "    mode='min', restore_best_weights=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:53:52.480032Z",
          "start_time": "2021-05-08T09:48:20.885525Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "OTVbbMPZGDwP",
        "outputId": "86bdb309-4dbf-4175-a034-f322ee27474b",
        "scrolled": true
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/100\n",
            "500/500 - 16s - loss: 3.0408 - accuracy: 0.3356 - val_loss: 1.7345 - val_accuracy: 0.3610 - 16s/epoch - 31ms/step\n",
            "Epoch 2/100\n",
            "500/500 - 7s - loss: 2.9658 - accuracy: 0.3684 - val_loss: 1.7226 - val_accuracy: 0.3295 - 7s/epoch - 15ms/step\n",
            "Epoch 3/100\n",
            "500/500 - 7s - loss: 2.8767 - accuracy: 0.3734 - val_loss: 1.6249 - val_accuracy: 0.3900 - 7s/epoch - 14ms/step\n",
            "Epoch 4/100\n",
            "500/500 - 7s - loss: 2.8309 - accuracy: 0.3745 - val_loss: 1.6068 - val_accuracy: 0.3785 - 7s/epoch - 14ms/step\n",
            "Epoch 5/100\n",
            "500/500 - 7s - loss: 2.7943 - accuracy: 0.3791 - val_loss: 1.5792 - val_accuracy: 0.3795 - 7s/epoch - 13ms/step\n",
            "Epoch 6/100\n",
            "500/500 - 7s - loss: 2.7638 - accuracy: 0.3826 - val_loss: 1.5339 - val_accuracy: 0.4210 - 7s/epoch - 15ms/step\n",
            "Epoch 7/100\n",
            "500/500 - 7s - loss: 2.7457 - accuracy: 0.3767 - val_loss: 1.5245 - val_accuracy: 0.4180 - 7s/epoch - 15ms/step\n",
            "Epoch 8/100\n",
            "500/500 - 7s - loss: 2.7087 - accuracy: 0.3851 - val_loss: 1.5419 - val_accuracy: 0.3930 - 7s/epoch - 14ms/step\n",
            "Epoch 9/100\n",
            "500/500 - 7s - loss: 2.6836 - accuracy: 0.3934 - val_loss: 1.5035 - val_accuracy: 0.4105 - 7s/epoch - 14ms/step\n",
            "Epoch 10/100\n",
            "500/500 - 7s - loss: 2.6670 - accuracy: 0.3966 - val_loss: 1.4944 - val_accuracy: 0.4170 - 7s/epoch - 13ms/step\n",
            "Epoch 11/100\n",
            "500/500 - 7s - loss: 2.6525 - accuracy: 0.3952 - val_loss: 1.4506 - val_accuracy: 0.4510 - 7s/epoch - 15ms/step\n",
            "Epoch 12/100\n",
            "500/500 - 7s - loss: 2.6328 - accuracy: 0.4052 - val_loss: 1.4632 - val_accuracy: 0.4320 - 7s/epoch - 14ms/step\n",
            "Epoch 13/100\n",
            "500/500 - 7s - loss: 2.5938 - accuracy: 0.4081 - val_loss: 1.4358 - val_accuracy: 0.4490 - 7s/epoch - 14ms/step\n",
            "Epoch 14/100\n",
            "500/500 - 7s - loss: 2.5813 - accuracy: 0.4164 - val_loss: 1.3902 - val_accuracy: 0.4850 - 7s/epoch - 14ms/step\n",
            "Epoch 15/100\n",
            "500/500 - 7s - loss: 2.5598 - accuracy: 0.4225 - val_loss: 1.4240 - val_accuracy: 0.4495 - 7s/epoch - 14ms/step\n",
            "Epoch 16/100\n",
            "500/500 - 7s - loss: 2.5298 - accuracy: 0.4214 - val_loss: 1.4348 - val_accuracy: 0.4285 - 7s/epoch - 15ms/step\n",
            "Epoch 17/100\n",
            "500/500 - 7s - loss: 2.5189 - accuracy: 0.4226 - val_loss: 1.4141 - val_accuracy: 0.4475 - 7s/epoch - 13ms/step\n",
            "Epoch 18/100\n",
            "500/500 - 7s - loss: 2.4785 - accuracy: 0.4373 - val_loss: 1.3471 - val_accuracy: 0.4835 - 7s/epoch - 14ms/step\n",
            "Epoch 19/100\n",
            "500/500 - 7s - loss: 2.4703 - accuracy: 0.4421 - val_loss: 1.3596 - val_accuracy: 0.4755 - 7s/epoch - 14ms/step\n",
            "Epoch 20/100\n",
            "500/500 - 7s - loss: 2.4319 - accuracy: 0.4436 - val_loss: 1.3233 - val_accuracy: 0.4910 - 7s/epoch - 13ms/step\n",
            "Epoch 21/100\n",
            "500/500 - 9s - loss: 2.4140 - accuracy: 0.4489 - val_loss: 1.3794 - val_accuracy: 0.4625 - 9s/epoch - 18ms/step\n",
            "Epoch 22/100\n",
            "500/500 - 7s - loss: 2.3937 - accuracy: 0.4601 - val_loss: 1.3055 - val_accuracy: 0.4900 - 7s/epoch - 13ms/step\n",
            "Epoch 23/100\n",
            "500/500 - 7s - loss: 2.3594 - accuracy: 0.4641 - val_loss: 1.3439 - val_accuracy: 0.4725 - 7s/epoch - 14ms/step\n",
            "Epoch 24/100\n",
            "500/500 - 7s - loss: 2.3476 - accuracy: 0.4638 - val_loss: 1.3239 - val_accuracy: 0.4745 - 7s/epoch - 14ms/step\n",
            "Epoch 25/100\n",
            "500/500 - 7s - loss: 2.3206 - accuracy: 0.4674 - val_loss: 1.2899 - val_accuracy: 0.5060 - 7s/epoch - 15ms/step\n",
            "Epoch 26/100\n",
            "500/500 - 8s - loss: 2.3069 - accuracy: 0.4726 - val_loss: 1.3060 - val_accuracy: 0.4745 - 8s/epoch - 15ms/step\n",
            "Epoch 27/100\n",
            "500/500 - 9s - loss: 2.2822 - accuracy: 0.4783 - val_loss: 1.3029 - val_accuracy: 0.4885 - 9s/epoch - 19ms/step\n",
            "Epoch 28/100\n",
            "500/500 - 7s - loss: 2.2695 - accuracy: 0.4836 - val_loss: 1.3334 - val_accuracy: 0.4815 - 7s/epoch - 14ms/step\n",
            "Epoch 29/100\n",
            "500/500 - 7s - loss: 2.2419 - accuracy: 0.4888 - val_loss: 1.3722 - val_accuracy: 0.4465 - 7s/epoch - 14ms/step\n",
            "Epoch 30/100\n",
            "500/500 - 7s - loss: 2.2229 - accuracy: 0.4932 - val_loss: 1.3251 - val_accuracy: 0.4735 - 7s/epoch - 13ms/step\n",
            "Epoch 31/100\n",
            "500/500 - 7s - loss: 2.1934 - accuracy: 0.5016 - val_loss: 1.2264 - val_accuracy: 0.5260 - 7s/epoch - 14ms/step\n",
            "Epoch 32/100\n",
            "500/500 - 7s - loss: 2.1718 - accuracy: 0.4953 - val_loss: 1.2093 - val_accuracy: 0.5350 - 7s/epoch - 14ms/step\n",
            "Epoch 33/100\n",
            "500/500 - 7s - loss: 2.1553 - accuracy: 0.5099 - val_loss: 1.3198 - val_accuracy: 0.4780 - 7s/epoch - 14ms/step\n",
            "Epoch 34/100\n",
            "500/500 - 9s - loss: 2.1304 - accuracy: 0.5116 - val_loss: 1.2429 - val_accuracy: 0.5160 - 9s/epoch - 18ms/step\n",
            "Epoch 35/100\n",
            "500/500 - 7s - loss: 2.1159 - accuracy: 0.5147 - val_loss: 1.2844 - val_accuracy: 0.4945 - 7s/epoch - 13ms/step\n",
            "Epoch 36/100\n",
            "500/500 - 7s - loss: 2.1029 - accuracy: 0.5220 - val_loss: 1.2170 - val_accuracy: 0.5260 - 7s/epoch - 15ms/step\n",
            "Epoch 37/100\n",
            "500/500 - 7s - loss: 2.0695 - accuracy: 0.5233 - val_loss: 1.2334 - val_accuracy: 0.5170 - 7s/epoch - 14ms/step\n",
            "Epoch 38/100\n",
            "500/500 - 7s - loss: 2.0559 - accuracy: 0.5264 - val_loss: 1.1991 - val_accuracy: 0.5475 - 7s/epoch - 14ms/step\n",
            "Epoch 39/100\n",
            "500/500 - 7s - loss: 2.0207 - accuracy: 0.5387 - val_loss: 1.1820 - val_accuracy: 0.5465 - 7s/epoch - 14ms/step\n",
            "Epoch 40/100\n",
            "500/500 - 7s - loss: 2.0155 - accuracy: 0.5384 - val_loss: 1.2485 - val_accuracy: 0.5105 - 7s/epoch - 14ms/step\n",
            "Epoch 41/100\n",
            "500/500 - 7s - loss: 1.9951 - accuracy: 0.5459 - val_loss: 1.2142 - val_accuracy: 0.5345 - 7s/epoch - 15ms/step\n",
            "Epoch 42/100\n",
            "500/500 - 7s - loss: 1.9769 - accuracy: 0.5421 - val_loss: 1.1577 - val_accuracy: 0.5610 - 7s/epoch - 14ms/step\n",
            "Epoch 43/100\n",
            "500/500 - 7s - loss: 1.9603 - accuracy: 0.5517 - val_loss: 1.1570 - val_accuracy: 0.5610 - 7s/epoch - 14ms/step\n",
            "Epoch 44/100\n",
            "500/500 - 7s - loss: 1.9341 - accuracy: 0.5536 - val_loss: 1.1237 - val_accuracy: 0.5795 - 7s/epoch - 14ms/step\n",
            "Epoch 45/100\n",
            "500/500 - 7s - loss: 1.9206 - accuracy: 0.5601 - val_loss: 1.2027 - val_accuracy: 0.5340 - 7s/epoch - 14ms/step\n",
            "Epoch 46/100\n",
            "500/500 - 7s - loss: 1.9023 - accuracy: 0.5516 - val_loss: 1.1400 - val_accuracy: 0.5715 - 7s/epoch - 14ms/step\n",
            "Epoch 47/100\n",
            "500/500 - 7s - loss: 1.8844 - accuracy: 0.5654 - val_loss: 1.1584 - val_accuracy: 0.5515 - 7s/epoch - 13ms/step\n",
            "Epoch 48/100\n",
            "500/500 - 7s - loss: 1.8621 - accuracy: 0.5649 - val_loss: 1.1445 - val_accuracy: 0.5600 - 7s/epoch - 14ms/step\n",
            "Epoch 49/100\n",
            "500/500 - 7s - loss: 1.8474 - accuracy: 0.5737 - val_loss: 1.1148 - val_accuracy: 0.5735 - 7s/epoch - 13ms/step\n",
            "Epoch 50/100\n",
            "500/500 - 8s - loss: 1.8329 - accuracy: 0.5691 - val_loss: 1.1323 - val_accuracy: 0.5730 - 8s/epoch - 16ms/step\n",
            "Epoch 51/100\n",
            "500/500 - 7s - loss: 1.8108 - accuracy: 0.5763 - val_loss: 1.1463 - val_accuracy: 0.5540 - 7s/epoch - 14ms/step\n",
            "Epoch 52/100\n",
            "500/500 - 7s - loss: 1.8001 - accuracy: 0.5746 - val_loss: 1.1227 - val_accuracy: 0.5665 - 7s/epoch - 13ms/step\n",
            "Epoch 53/100\n",
            "500/500 - 7s - loss: 1.7865 - accuracy: 0.5820 - val_loss: 1.0890 - val_accuracy: 0.5920 - 7s/epoch - 14ms/step\n",
            "Epoch 54/100\n",
            "500/500 - 7s - loss: 1.7690 - accuracy: 0.5859 - val_loss: 1.1771 - val_accuracy: 0.5495 - 7s/epoch - 13ms/step\n",
            "Epoch 55/100\n",
            "500/500 - 8s - loss: 1.7345 - accuracy: 0.5929 - val_loss: 1.0823 - val_accuracy: 0.5880 - 8s/epoch - 15ms/step\n",
            "Epoch 56/100\n",
            "500/500 - 7s - loss: 1.7286 - accuracy: 0.5897 - val_loss: 1.1240 - val_accuracy: 0.5680 - 7s/epoch - 14ms/step\n",
            "Epoch 57/100\n",
            "500/500 - 7s - loss: 1.7105 - accuracy: 0.5990 - val_loss: 1.0919 - val_accuracy: 0.5815 - 7s/epoch - 13ms/step\n",
            "Epoch 58/100\n",
            "500/500 - 7s - loss: 1.6917 - accuracy: 0.6033 - val_loss: 1.0889 - val_accuracy: 0.5865 - 7s/epoch - 14ms/step\n",
            "Epoch 59/100\n",
            "500/500 - 7s - loss: 1.6829 - accuracy: 0.6023 - val_loss: 1.0655 - val_accuracy: 0.6040 - 7s/epoch - 13ms/step\n",
            "Epoch 60/100\n",
            "500/500 - 7s - loss: 1.6751 - accuracy: 0.5991 - val_loss: 1.1335 - val_accuracy: 0.5585 - 7s/epoch - 14ms/step\n",
            "Epoch 61/100\n",
            "500/500 - 7s - loss: 1.6334 - accuracy: 0.6112 - val_loss: 1.0513 - val_accuracy: 0.6145 - 7s/epoch - 13ms/step\n",
            "Epoch 62/100\n",
            "500/500 - 7s - loss: 1.6271 - accuracy: 0.6137 - val_loss: 1.0963 - val_accuracy: 0.5830 - 7s/epoch - 15ms/step\n",
            "Epoch 63/100\n",
            "500/500 - 7s - loss: 1.6188 - accuracy: 0.6126 - val_loss: 1.0332 - val_accuracy: 0.6175 - 7s/epoch - 14ms/step\n",
            "Epoch 64/100\n",
            "500/500 - 9s - loss: 1.5870 - accuracy: 0.6227 - val_loss: 1.0640 - val_accuracy: 0.5985 - 9s/epoch - 18ms/step\n",
            "Epoch 65/100\n",
            "500/500 - 7s - loss: 1.5937 - accuracy: 0.6184 - val_loss: 1.0434 - val_accuracy: 0.6065 - 7s/epoch - 14ms/step\n",
            "Epoch 66/100\n",
            "500/500 - 7s - loss: 1.5737 - accuracy: 0.6221 - val_loss: 1.1398 - val_accuracy: 0.5605 - 7s/epoch - 14ms/step\n",
            "Epoch 67/100\n",
            "500/500 - 7s - loss: 1.5655 - accuracy: 0.6233 - val_loss: 1.0930 - val_accuracy: 0.5730 - 7s/epoch - 14ms/step\n",
            "Epoch 68/100\n",
            "500/500 - 7s - loss: 1.5475 - accuracy: 0.6305 - val_loss: 1.0481 - val_accuracy: 0.6105 - 7s/epoch - 15ms/step\n",
            "Epoch 69/100\n",
            "500/500 - 7s - loss: 1.5316 - accuracy: 0.6331 - val_loss: 1.0232 - val_accuracy: 0.6140 - 7s/epoch - 14ms/step\n",
            "Epoch 70/100\n",
            "500/500 - 7s - loss: 1.5045 - accuracy: 0.6344 - val_loss: 1.0561 - val_accuracy: 0.5935 - 7s/epoch - 14ms/step\n",
            "Epoch 71/100\n",
            "500/500 - 7s - loss: 1.5039 - accuracy: 0.6403 - val_loss: 1.0680 - val_accuracy: 0.5890 - 7s/epoch - 13ms/step\n",
            "Epoch 72/100\n",
            "500/500 - 7s - loss: 1.4847 - accuracy: 0.6388 - val_loss: 1.0732 - val_accuracy: 0.5950 - 7s/epoch - 14ms/step\n",
            "Epoch 73/100\n",
            "500/500 - 7s - loss: 1.4661 - accuracy: 0.6475 - val_loss: 1.0888 - val_accuracy: 0.5815 - 7s/epoch - 14ms/step\n",
            "Epoch 74/100\n",
            "500/500 - 7s - loss: 1.4445 - accuracy: 0.6496 - val_loss: 1.0195 - val_accuracy: 0.6170 - 7s/epoch - 14ms/step\n",
            "Epoch 75/100\n",
            "500/500 - 7s - loss: 1.4494 - accuracy: 0.6501 - val_loss: 1.0744 - val_accuracy: 0.5820 - 7s/epoch - 14ms/step\n",
            "Epoch 76/100\n",
            "500/500 - 7s - loss: 1.4399 - accuracy: 0.6504 - val_loss: 1.0207 - val_accuracy: 0.6105 - 7s/epoch - 13ms/step\n",
            "Epoch 77/100\n",
            "500/500 - 8s - loss: 1.4102 - accuracy: 0.6604 - val_loss: 1.0605 - val_accuracy: 0.6040 - 8s/epoch - 15ms/step\n",
            "Epoch 78/100\n",
            "500/500 - 7s - loss: 1.3958 - accuracy: 0.6632 - val_loss: 1.0346 - val_accuracy: 0.6000 - 7s/epoch - 13ms/step\n",
            "Epoch 79/100\n",
            "500/500 - 7s - loss: 1.3874 - accuracy: 0.6641 - val_loss: 1.0173 - val_accuracy: 0.6075 - 7s/epoch - 15ms/step\n",
            "Epoch 80/100\n",
            "500/500 - 7s - loss: 1.3851 - accuracy: 0.6618 - val_loss: 1.0063 - val_accuracy: 0.6195 - 7s/epoch - 15ms/step\n",
            "Epoch 81/100\n",
            "500/500 - 7s - loss: 1.3414 - accuracy: 0.6725 - val_loss: 1.0344 - val_accuracy: 0.6075 - 7s/epoch - 13ms/step\n",
            "Epoch 82/100\n",
            "500/500 - 7s - loss: 1.3350 - accuracy: 0.6717 - val_loss: 1.0296 - val_accuracy: 0.6025 - 7s/epoch - 14ms/step\n",
            "Epoch 83/100\n",
            "500/500 - 7s - loss: 1.3314 - accuracy: 0.6735 - val_loss: 1.0721 - val_accuracy: 0.5875 - 7s/epoch - 13ms/step\n",
            "Epoch 84/100\n",
            "500/500 - 7s - loss: 1.3074 - accuracy: 0.6768 - val_loss: 1.0058 - val_accuracy: 0.6140 - 7s/epoch - 14ms/step\n",
            "Epoch 85/100\n",
            "500/500 - 7s - loss: 1.3140 - accuracy: 0.6776 - val_loss: 0.9750 - val_accuracy: 0.6345 - 7s/epoch - 14ms/step\n",
            "Epoch 86/100\n",
            "500/500 - 7s - loss: 1.2799 - accuracy: 0.6829 - val_loss: 0.9997 - val_accuracy: 0.6260 - 7s/epoch - 14ms/step\n",
            "Epoch 87/100\n",
            "500/500 - 7s - loss: 1.2828 - accuracy: 0.6831 - val_loss: 0.9997 - val_accuracy: 0.6250 - 7s/epoch - 14ms/step\n",
            "Epoch 88/100\n",
            "500/500 - 7s - loss: 1.2752 - accuracy: 0.6847 - val_loss: 0.9801 - val_accuracy: 0.6300 - 7s/epoch - 14ms/step\n",
            "Epoch 89/100\n",
            "500/500 - 7s - loss: 1.2539 - accuracy: 0.6890 - val_loss: 1.0039 - val_accuracy: 0.6190 - 7s/epoch - 14ms/step\n",
            "Epoch 90/100\n",
            "500/500 - 7s - loss: 1.2462 - accuracy: 0.6939 - val_loss: 1.0315 - val_accuracy: 0.6070 - 7s/epoch - 13ms/step\n",
            "Epoch 91/100\n",
            "500/500 - 7s - loss: 1.2343 - accuracy: 0.6909 - val_loss: 0.9628 - val_accuracy: 0.6390 - 7s/epoch - 15ms/step\n",
            "Epoch 92/100\n",
            "500/500 - 7s - loss: 1.2126 - accuracy: 0.7028 - val_loss: 1.0118 - val_accuracy: 0.6125 - 7s/epoch - 14ms/step\n",
            "Epoch 93/100\n",
            "500/500 - 7s - loss: 1.2031 - accuracy: 0.7042 - val_loss: 0.9943 - val_accuracy: 0.6255 - 7s/epoch - 13ms/step\n",
            "Epoch 94/100\n",
            "500/500 - 7s - loss: 1.1864 - accuracy: 0.7057 - val_loss: 1.0287 - val_accuracy: 0.6065 - 7s/epoch - 14ms/step\n",
            "Epoch 95/100\n",
            "500/500 - 7s - loss: 1.1853 - accuracy: 0.7036 - val_loss: 0.9820 - val_accuracy: 0.6235 - 7s/epoch - 13ms/step\n",
            "Epoch 96/100\n",
            "500/500 - 7s - loss: 1.1875 - accuracy: 0.7084 - val_loss: 0.9851 - val_accuracy: 0.6250 - 7s/epoch - 15ms/step\n",
            "Epoch 97/100\n",
            "500/500 - 7s - loss: 1.1606 - accuracy: 0.7109 - val_loss: 0.9822 - val_accuracy: 0.6315 - 7s/epoch - 14ms/step\n",
            "Epoch 98/100\n",
            "500/500 - 7s - loss: 1.1376 - accuracy: 0.7198 - val_loss: 1.0279 - val_accuracy: 0.6135 - 7s/epoch - 15ms/step\n",
            "Epoch 99/100\n",
            "500/500 - 7s - loss: 1.1065 - accuracy: 0.7177 - val_loss: 0.9716 - val_accuracy: 0.6405 - 7s/epoch - 15ms/step\n",
            "Epoch 100/100\n",
            "500/500 - 7s - loss: 1.1211 - accuracy: 0.7187 - val_loss: 1.0422 - val_accuracy: 0.6040 - 7s/epoch - 13ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.src.callbacks.History at 0x7fa2da6d2740>"
            ]
          },
          "metadata": {},
          "execution_count": 60
        }
      ],
      "source": [
        "model.fit(x_train, y_train, epochs=100, batch_size=32, validation_data=(x_val, y_val), callbacks=[hist, earlystop],\n",
        "          verbose=2, class_weight=class_weight, shuffle=True)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:53:57.531104Z",
          "start_time": "2021-05-08T09:53:57.270296Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DcicyQFAGDwP",
        "outputId": "88385d10-68de-436b-b3f5-ae920737a0b7"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "63/63 [==============================] - 1s 9ms/step - loss: 1.0547 - accuracy: 0.5905\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[1.054739236831665, 0.590499997138977]"
            ]
          },
          "metadata": {},
          "execution_count": 61
        }
      ],
      "source": [
        "model.evaluate(x_test,y_test)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:53:58.282116Z",
          "start_time": "2021-05-08T09:53:58.278127Z"
        },
        "id": "YtZnlBJJGDwP"
      },
      "outputs": [],
      "source": [
        "from tensorflow.math import confusion_matrix\n",
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:53:58.816686Z",
          "start_time": "2021-05-08T09:53:58.698004Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 466
        },
        "id": "V8bpqu66GDwQ",
        "outputId": "02d904fd-9838-45e6-dd6e-5179e2537f5b"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0, 0.5, 'Loss')"
            ]
          },
          "metadata": {},
          "execution_count": 63
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "plt.plot(hist.history['loss'])\n",
        "plt.plot(hist.history['val_loss'])\n",
        "plt.grid()\n",
        "plt.legend([\"loss\",\"val_loss\"])\n",
        "plt.xlabel(\"Epochs\")\n",
        "plt.ylabel(\"Loss\")"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:53:59.811038Z",
          "start_time": "2021-05-08T09:53:59.353755Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 559
        },
        "id": "xX-w9eudGDwQ",
        "outputId": "8179552b-afd3-4bde-9433-9f7423d72d4c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "63/63 [==============================] - 2s 8ms/step\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x600 with 2 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAwgAAAINCAYAAACav1gQAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAACw6klEQVR4nOzddVxUWR8G8GdoUOmyuxUMFFy7WzFWXXOxExXbVWxR1147ULE7sBVFxUSQUBGlDVJKQXLm/YN9R0YG1wAvMM/3/ZzP65x77r2/w+zcmXNPXJFEIpGAiIiIiIgIgJLQARARERERUcHBBgIREREREUmxgUBERERERFJsIBARERERkRQbCEREREREJMUGAhERERERSbGBQEREREREUmwgEBERERGRFBsIREREREQkpSJ0APkh5c5+oUOgrxgy+ITQIVAuHn0MFjoEykX4x1ihQ6Cv6GxST+gQKBdeSWFCh0C5CH3vI3QIuUqPCcq3Y6saVsq3Y+cV9iAQEREREZFUkexBICIiIiL6YeJMoSMQFBsIRERERETZScRCRyAoDjEiIiIiIiIp9iAQEREREWUnZg8CERERERERAPYgEBERERHJkHAOAhERERERURb2IBARERERZcc5CERERERERFnYg0BERERElJ2Cz0FgA4GIiIiIKDsFf5Ky4EOMPn36hOTkZOnr0NBQrF+/HlevXhUwKiIiIiIixSR4A6Fnz55wcnICAMTHx8PS0hJr1qxBz549sXXrVoGjIyIiIiKFIxHnXyoEBG8geHp6onnz5gCAEydOwMTEBKGhoXBycsLGjRsFjo6IiIiISLEIPgchOTkZJUqUAABcvXoVvXv3hpKSEqysrBAaGipwdERERESkcLjMqbCqVKmCM2fO4PXr17hy5Qo6dOgAAIiKioK2trbA0RERERERKRbBGwj29vaYPn06KlSoAEtLSzRp0gRAVm9C/fr1BY6OiIiIiBSNRCLOt1QYCD7EqG/fvmjWrBnCw8Nhbm4uzW/bti169eolYGRERERERIpH8AYCAJiamsLU1BQAkJiYiBs3bqB69eqoUaOGwJERERERkcLhHARh9evXD5s2bQKQ9UwECwsL9OvXD2ZmZjh58qTA0RERERGRwuEyp8K6ffu2dJnT06dPQyKRID4+Hhs3bsTSpUsFjo6IiIiISLEI3kBISEiAvr4+AODy5cvo06cPtLS00LVrV7x69Urg6IiIiIhI4Ygz8y8VAoI3EMqWLYv79+8jKSkJly9fli5zGhcXBw0NDYGjIyIiIiJSLIJPUp4yZQoGDRqE4sWLo1y5cmjVqhWArKFHdevWFTY4IiIiIlI8hWSuQH4RvIEwfvx4NG7cGK9fv0b79u2hpJTVqVGpUiXOQSAiIiIi+sUEbyAAgIWFBczMzBAcHIzKlStDRUUFXbt2FTosIiIiIlJEXOZUWMnJyRgxYgS0tLRQu3ZthIWFAQAmTZqEFStWCBwdEREREZFiEbyBMGfOHHh7e8PV1VVmUnK7du1w9OhRASMjIiIiIoWk4M9BEHyI0ZkzZ3D06FFYWVlBJBJJ82vXro3AwEABIyMiIiIihcQhRsKKjo6GsbFxjvykpCSZBgMREREREeU/wRsIFhYWuHDhgvT1/xsFu3btQpMmTYQKK98cufEYnWf9g0ZjHTBomSN8g97mWvbsXW+Yj1wqkxqNdZApk5yShuUHL6P9jA1oPG4Fes3fhmOuHvldjSKr49Au2Oy2Awf9j2P5mb9RxbxqrmXLVC2LadtmYbPbDhwPPYsuw7vnKPP7lAE4HnpWJq132ZyfVSgyho7oD7cnl+D/1h1nrh6EeYM6Xy3fpUd7uDw4C/+37rhy5yRat2sms331piUIfe8jk/Yd2yrdbtXUIsf2/yez+rXzpY6F1dixw/DS/z4SEwLgdscZFhb1vlq+T++u8PVxRWJCADw9rqNTpzYy2617dsaFCwcR/s4XaalvYG5W66vHO3duP9JS36BHj44/WxWF0GVoV+y4uxvHX57C32fXoKp5tVzLlq1WDrO2zcGOu7txNuw8uo/o8dVj9xnfF2fDzmPEglF5HXaR9KuvawBQx6wmDpzcDp8gN3i9ug2HtfbQKqaZ53UraiSSzHxLhYHgDYTly5dj7ty5GDduHDIyMrBhwwZ06NABe/bswbJly4QOL09dfvQMq49dw5juzXHEfiSqlzXBuPWH8T4xKdd9imuqw2XNFGm6vHKSzPbVx67h3tNALB/RE6eXjMWgdo2x4tBluHq9zO/qFDm/dWuGYfOG4/iGo5jVzQ6hfsH4a/9CaBvoyC2vrqmOqLBIHFy5H3FRsbkeN8w/FKMshknT/L6z86sKRUY3646Yt2QGNvy9Dd3a9IffU3/sP74NBob6css3bGSOf3auxLEDp9G1dT9cvXgDO/ZvQLUaVWTKuV53g0XN1tI0adRM6TaPR14y2yxqtsZhp5MIC3kDnyfP8rW+hcnvfbvj71X2WLpsHSwtO8PH9zkunD8AIyMDueWtrBpi//7N2LP3CBpbdsK5c5dx4vgu1K5VXVqmWDEt3Lvrjrl/Lf/P89vajoREIsmz+hR1zbo3x/D5I3F0/WHYdZ2MYL9gLDywGDq5Xdc01BEZFoH9K/Yh9ivXNQCoYlYVHQd2QvDz4PwIvcgR4rpmbGqEg6d2ICT4Naw7DMbQfuNQrUZlrNnEZeQLm82bN6NChQrQ0NCApaUlHj169NXy8fHxmDBhAkqWLAl1dXVUq1YNFy9e/ObzCd5AaNasGby8vJCRkYG6devi6tWrMDY2xv3799GwYUOhw8tT+689RO/m9WHdrB4qlzLCvMFdoKGmijNuXrnuIwJgqFNcmgx0ists9wp4g+6/maFRjQoobaiLvi0boFoZEzwNzr1nguTrNrInXI5chetxF7x59Ro75m5F2qdUtOnXTm75QJ8A7F++F/ec7yA9NT3X44ozMhEfHS9NH+I+5FcVioyR44fiyP6TOH7oLF75B2HutCX49OkT+g2yllveZswg3HK5i+2b9iLgZTDWOGzGUx8/DBs5QKZcaloaoqPeS1Niwuf3Ij09Q2ZbXGwC2ndujeOHzuRjTQufyZNHY7fjYTg5HYPfi1eYMGE2kpNT8OewAXLLT5o4AleuumLt2m148SIACxetxpMnTzFu/J/SMgcPncSy5etx48adr57b3KwWpkweg9Gjp+VllYq0niOtcfXwFbgcv47Xr15j65zNSP2Uinb928stH+DzCnuX78Ed59tfva5paGnAbuN0bJ79Dz4mfMyv8IsUIa5rbTu0QHp6BubPWIaggBD4PHmGudOWokuP9ihfsWx+VrfwK0CTlI8ePQo7OzssWLAAnp6eMDc3R8eOHREVFSW3fFpaGtq3b4+QkBCcOHEC/v7+2LlzJ0qXLv3N5xS8gQAAlStXxs6dO/Ho0SM8f/4cBw4cKHJPUU7PyIRfaDisalWU5ikpiWBVswJ8vjLMKDk1DZ1mbkSHGRswedMxBLyNltler0oZ3PJ+ici4REgkEjx6EYLQyFg0qV0p3+pSFKmoqqBS3crwcfOW5kkkEvi4eaNag+pf2fO/mVYshe2P9mDTne2w3WAHw1KGPxtukaaqqoK65jXhduuBNE8ikcDt1kM0aGQud58GjczhduuhTN7tG/dylLdqagGPF6648fAclq6eB109+XdRAaB951bQ09fBscNnf6I2RYuqqioaNKgr80NeIpHgxo07sLJqIHcfS8uGOX74X7t2C1aW33cDSFNTA05OmzB5yl+IjIz+7x0IKqoqqFy3Cryz3YSSSCTwdvNC9QY1furYY5aOg8cNd3hnu2ZS7oS6rqmrqyE9PV2m1y0lJQUA0Miq/k/Xi36NtWvXYtSoUbCxsUGtWrWwbds2aGlpwdHRUW55R0dHxMbG4syZM2jatCkqVKiAli1bwtxc/n9r8gi+ihEAiMViBAQEICoqCuIvZo23aNFCoKjyVtzHZGSKJTDQLiaTb6BdHMER7+XuU8HUAIv+7I6qZYzx8VMq9l15gGEr9uLUojEw0dcGAMz+oyMWO11AhxkboaKsBJFIhAVDu6JhtfL5XqeipISeNpRVlJEQEy+TnxATj9KVy/zwcV95vcTmaRvwLugt9Iz18fuUAVh83AF2HWyRkvTpJ6MumvQM9KCiooKYKNnPRUzUe1SuWlHuPkbGhoiJ/qJ89HsYGX9ujN1yuYvL513wOvQtylcsg5nzbLHv2Bb06jgkx3UHAPoP6oXbN+4h4l1kHtSqaDA01IeKikqOH+hRUTGoXr2K3H1MTY0QFRkjkxcZFQ0TE6PvOvfq1Qtx/74HnJ2vfl/QCkxbP+u6Fv/FdS0+Jh5lfuK61rx7C1SqUxnTu0/9yQgVh1DXtbu3H2HekukYM/FPOG4/AE0tTcy2nwIAMP7Oz6DCycdVjFJTU5GamiqTp66uDnV19Rxl09LS4OHhgTlz5kjzlJSU0K5dO9y/f1/u8c+dO4cmTZpgwoQJOHv2LIyMjDBw4EDMmjULysrK3xSj4A2EBw8eYODAgQgNDc0xrlQkEiEz8+uTOeT9kSVp6VBXU83zWH8188plYJ7tIm5euQx6zd+G47c9MdG6FQDg8A13+AS9xYaJ/VDKQAcer8Kw/OBlGOkWh1Ut9iIIzcvVU/rvsBeheOX1Elvv7sRv3ZrixtHrAkameJxPX5b+29/vFfyevYSb5yU0adYId2/L3qUzLWWCFm1+w4QRM351mCRHt27t0apVUzRuzEnJQjMsaYiRC0fBftD8rw5Bol/jv65rr/wDMW3CfMxbMh0z59siM1OMvTsOISoyRu6NEfo1HBwcsGjRIpm8BQsWYOHChTnKxsTEIDMzEyYmJjL5JiYmePHihdzjBwUF4caNGxg0aBAuXryIgIAAjB8/Hunp6ViwYME3xSh4A2Hs2LHSlYxKliz53Uubyvsj//WnNeYN752XYf40veJaUFYS5ZiQ/D7xIwy/mFeQG1UVZdQoZ4rX/04cS0lLx8ZTN7Fuwu9oYZa12k61sibwD4vEvisP2ED4Dh/iEpGZkQkdQ12ZfB1DXcRHx+XZeZITk/Au+B1My5fMs2MWNXHv45CRkQFDY9lJr4bGBoiOipG7T3RUDAy/mCRraJR7eQB4HfoW72NiUb5i2RwNhH5/9ERcbAKuXXL9sUoUUTExscjIyMhx99/Y2BCRkfLHwkZERMPYRHZYnYmx0XcNE2rVqikqVyqP6KjnMvlHj+yAm9sjtO/w+zcfS5EkxmZd13S/uK7pGuoi7geva5XrVoGukR7WXdwgzVNWUUZty9roOqwb+lbpxR+ecgh5XTt78iLOnrwIQyN9JCd/gkQCjBw/BGEhb36yVkVcPj7QbM6cObCzs5PJk9d78KPEYjGMjY2xY8cOKCsro2HDhnj79i3+/vvvb24gCD4H4dWrV1i+fDlq1qwJXV1d6OjoyKT/MmfOHCQkJMikGYNzLjcpNFUVZdQsXxIP/T6v9iAWS/DwRQjMKn3bpJFMsRiv3kbBUKcEACAjU4yMTDGUvmhUKSmJIOYqH98lIz0DQb6BqNvUTJonEolQt6kZXnr659l5NLQ0YFreFHFRedfoKGrS0zPg6+2Hpi0spXkikQhNW1jC013+eGdPd2+Z8gDQvJVVruWBrF4CPX3dHMNfAOD3gdY4ddQZGRkZP1iLoik9PR2enr5o3frzUosikQitWzfDgweecvd5+NADbVrLLs3Ytm1zPHj47csx//33ZjRs2B6NGnWUJgCYPmMRRo22+4+9FVdGegYCfQNg1vTzuGORSASzpubw95R/5/G/+Nz1xqR2EzClk600vfJ+iVtnXDGlky0bB7koCNe1mOhYJCd9QnfrjkhNSYOb6wM5RyApcWa+JXV1dWhra8uk3BoIhoaGUFZWRmSk7HDXyMhImJqayt2nZMmSqFatmsxwopo1ayIiIgJpaWnfVH3BexAsLS0REBCAKlXkj1/9L/LGbKUU0OFFQ9pbYr7jOdQuXxJ1KpbGgesP8Sk1Hdb/Xrz/2n0WxrolMLlP1hrh25xvw6xSaZQz1seH5BTsvXIf4e8T0Lt5PQBZS6BaVCuHtcddoK6qgpIGOvB4GYbz930xvZ/8FSood+d3ncWENZMR6BOAAO9X6Dq8O9S1NHDzeNZQoIlrpyA24j0OrdoPIGsCYJmqWatAqKipwsDUABVqVURK0idEhEYAAIb89Sc8rrsj+m009Ez00X/qHxBninH33G1hKllI7NrihDWbl8LH6zm8PX0xfMxgaGlpSlcUWrtlGSLCI7FqyUYAwJ7tB3HU2RGjxg/FjWu30b1XZ9StVxuzpy4GAGgV08SUGeNw6fx1REfGoHzFspizYCpCgsJw+8ZdmXM3bWGJchXK4MiBk7+0zoXFhg07sHv3Onh6eMP9sRcmTRqJYsU0sc/pKADAcfd6vHsXgXnzVwAA/tm0Gy7XT2DKlNG4dMkF/X7viYYNzTB+/CzpMfX0dFGubCmULJX1ZVetWmUAQERkNCKzpS+9fv0WISGv87vKhdrZXWcwec1UBPi+wiuvl+g+oic0tDRw/VjWdW3KOju8j3iP/Sv3Aci6rpX997qmqqYCAxMDVKxVEZ+SUhARGo5PSZ8Q9jJU5hwpyan4EPchRz7JEuq6NmzkAHg88kZSUjKat7LC3IV2WLFkAxITuaJeYaCmpoaGDRvCxcUF1tbWALJ6CFxcXDBx4kS5+zRt2hSHDh2CWCyGklJWX8DLly9RsmRJqKmpfdN5BW8gTJo0CdOmTUNERATq1q0LVVXZH/dmZma57Fn4dGpcG3Efk7Hl7C3EJCahelkTbJnyh3Tp0oj3CTK9AR+SUrB43wXEJCZBW0sDtcqXxL45f6Jyqc/d+yvH9MaGkzcwZ9dZJCZ9QkkDHUzs1Qq/t5K/ogjl7t55N2gbaKO/3UDoGukh5Hkwlg1dhISYBACAYSlDSLLdHdMz0cffl9ZLX/cY0ws9xvTCs/u+WDhgHgDAwNQQk/+ZjhK6JZAYm4AX7n6Yaz0TibGJv7Ruhc35M1dgYKgHu9njYWRsiOdP/TG03zjERGcNrytV2lTmTqWHuzdsR8/G9L8mYcY8W4QEhWH0kMl4+SIAAJCZKUaN2lXRZ0APaOuUQGREFO7cvI81DpuQliY7jrr/oF54/PAJAl+F/LL6FibHTzjD0MgA9vbTYWpqBG/v5+jWfQii/h32ULZsaZn35sEDDwwdOhGLFs3EksWzEBAQjL6/j8Sz55975rp1a4/du9ZJXx88mPWgpyVL1mLJ0rW/qGZFk5vzHWjr62Cg3WDoGekh+HkQFg2xly7IYFjKSOb90jfRx/rL/0hf9xrbB73G9oHvfV/M6z/ny8PTdxDqumbeoC6mzhoPrWJaCHwVjDnTluD0sfO/tvKFUT4OMfpednZ2GDZsGCwsLNC4cWOsX78eSUlJsLGxAQAMHToUpUuXhoND1sN0x40bh02bNmHy5MmYNGmSdLSOra3tN59TJBH4iTP/b9lkJxKJIJFIvmmSsjwpd/bnRWiUT4YMPiF0CJSLRx/5wKOCKvzj1x9aRcLqbFJP6BAoF15JYUKHQLkIfe8jdAi5Snl0PN+OrdH4++dNbdq0CX///TciIiJQr149bNy4EZaWWUPQWrVqhQoVKmDv3r3S8vfv38fUqVPh5eWF0qVLY8SIEYVrFaPgYP4gISIiIqICpIDNp5k4cWKuQ4pcXV1z5DVp0gQPHvz4PBPBGwjly3O9fiIiIiKigkKQBsK5c+e+uWyPHj3yMRIiIiIioi8UoDkIQhCkgfD/Wdj/9/85B9lf/9+PzEEgIiIiIqIfI8hzEMRisTRdvXoV9erVw6VLlxAfH4/4+HhcvHgRDRo0wOXLl//7YEREREREeUkszr9UCAg+B2HKlCnYtm0bmjX7/CCdjh07QktLC6NHj4afn5+A0RERERGRwikkP+Tzi+BPUg4MDISurm6OfB0dHYSEhPzyeIiIiIiIFJngDYRGjRrBzs5O5hHSkZGRmDFjBho3bixgZERERESkiCSSzHxLhYHgDQRHR0eEh4ejXLlyqFKlCqpUqYJy5crh7du32L17t9DhEREREREpFMHnIFSpUgU+Pj64du0aXrx4AQCoWbMm2rVrJ7OaERERERHRL6HgcxAEbyAAWcuadujQAR06dBA6FCIiIiIihVYgGghJSUm4desWwsLCkJaWJrPN1tZWoKiIiIiISCHxQWnCevLkCbp06YLk5GQkJSVBX18fMTEx0NLSgrGxMRsIRERERES/kOCTlKdOnYru3bsjLi4OmpqaePDgAUJDQ9GwYUOsXr1a6PCIiIiISNEo+IPSBG8geHl5Ydq0aVBSUoKysjJSU1NRtmxZrFq1CnPnzhU6PCIiIiJSNBJx/qVCQPAGgqqqKpSUssIwNjZGWFgYgKwHpb1+/VrI0IiIiIiIFI7gcxDq168Pd3d3VK1aFS1btoS9vT1iYmKwf/9+1KlTR+jwiIiIiEjRFJKhQPlF8B6E5cuXo2TJkgCAZcuWQU9PD+PGjUNMTAy2b98ucHRERERERIpF8B6E2rVrQyKRAMgaYrRt2zacPn0atWrVQr169YQNjoiIiIgUTyGZK5BfBO9B6NmzJ5ycnAAA8fHxsLKywtq1a2FtbY2tW7cKHB0RERERkWIRvIHg6emJ5s2bAwBOnDgBExMThIaGwsnJCRs3bhQ4OiIiIiJSOFzmVFjJyckoUaIEAODq1avo3bs3lJSUYGVlhdDQUIGjIyIiIiJSLII3EKpUqYIzZ87g9evXuHLlCjp06AAAiIqKgra2tsDREREREZHCYQ+CsOzt7TF9+nRUqFABlpaWaNKkCYCs3oT69esLHB0RERERKRwFf1Ca4KsY9e3bF82aNUN4eDjMzc2l+W3btkWvXr0EjIyIiIiISPEI3kAAAFNTU5iamsrkNW7cWKBoiIiIiEihFZKhQPlF8CFGRERERERUcBSIHgQiIiIiogKjkMwVyC/sQSAiIiIiIin2IBARERERZcc5CERERERERFnYg0BERERElB3nIBAREREREWVhDwIRERERUXYKPgehSDYQ/hh0TOgQ6CsOLTcTOgTKRXXbIKFDoFwoidjhW5BZiXSFDoFy4ZrqJ3QIVBgpeAOB3zhERERERCRVJHsQiIiIiIh+mEQidASCYg8CERERERFJsQeBiIiIiCg7zkEgIiIiIiLKwh4EIiIiIqLs2INARERERESUhT0IRERERETZSRS7B4ENBCIiIiKi7DjEiIiIiIiIKAt7EIiIiIiIsuOD0oiIiIiIiLKwB4GIiIiIKDvOQSAiIiIiIsrCHgQiIiIiouzYg0BERERERJSFPQhERERERNnxQWlERERERPR/EjGXOSUiIiIiIgIgcAMhIyMDixcvxps3b4QMg4iIiIjoM7E4/1IhIGgDQUVFBX///TcyMjKEDIOIiIiIiP4l+BCjNm3a4NatW0KHQURERESURSLOv1QICD5JuXPnzpg9ezZ8fX3RsGFDFCtWTGZ7jx49BIqMiIiIiEjxCN5AGD9+PABg7dq1ObaJRCJkZmb+6pCIiIiISJEp+CpGgjcQxIVksgYRERERkSIQvIGQXUpKCjQ0NIQOg4iIiIgUmYLfwBZ8knJmZiaWLFmC0qVLo3jx4ggKCgIAzJ8/H7t37xY4OiIiIiJSOAVsmdPNmzejQoUK0NDQgKWlJR49epRr2b1790IkEsmk770BL3gDYdmyZdi7dy9WrVoFNTU1aX6dOnWwa9cuASMjIiIiIhLW0aNHYWdnhwULFsDT0xPm5ubo2LEjoqKict1HW1sb4eHh0hQaGvpd5xS8geDk5IQdO3Zg0KBBUFZWluabm5vjxYsXAkZGRERERApJIsm/9J3Wrl2LUaNGwcbGBrVq1cK2bdugpaUFR0fHXPcRiUQwNTWVJhMTk+86p+ANhLdv36JKlSo58sViMdLT0wWIiIiIiIgof6SmpiIxMVEmpaamyi2blpYGDw8PtGvXTpqnpKSEdu3a4f79+7me4+PHjyhfvjzKli2Lnj174tmzZ98Vo+ANhFq1auHOnTs58k+cOIH69esLEBERERERKbR8nIPg4OAAHR0dmeTg4CA3jJiYGGRmZuboATAxMUFERITcfapXrw5HR0ecPXsWBw4cgFgsxm+//YY3b958c/UFX8XI3t4ew4YNw9u3byEWi3Hq1Cn4+/vDyckJ58+fFzo8IiIiIqI8M2fOHNjZ2cnkqaur59nxmzRpgiZNmkhf//bbb6hZsya2b9+OJUuWfNMxBO9B6NmzJ5ydnXH9+nUUK1YM9vb28PPzg7OzM9q3by90eHmu89Au2H53F46+PImVZ1ejqnnVXMuWrVYOM7fNwfa7u3A6zBndRnz9qdK9x/fF6TBnDF8wMq/DVhhH3APQeeNFNF5+CoN3u8D3bWyuZc96h6DekhMyqfHyUzJltt56BustV2C14jSa/30WYw7chu/b9/ldjSJh6IgBuOt1GS/fPcbZawdh3qDOV8t37dkBNx6cw8t3j3HV7RRat2sus33NpqUIi/WVSU7Ht8qUuet1OUeZ8ZNH5HndCrsxY4bC3/8u4uNf4vbts7CwMP9q+d69u8Lb+wbi41/i8eOr6Nixtcz2nj074fz5A3j71hspKWEwM6uV4xgmJkZwdFyPkJDHeP/+Be7fvwBr6855Wq+iqsHQdhjntg4z/B0x7MxClDSvlGtZ8wGtMPj4fEzx2Y4pPtsx4ODsHOWbTemN0S6rMM1vl7RMqXqV87saRcLI0YPh8+wWImOew+XmSTRoaPbV8ta9OsPd8yoiY57j3sOLaN+hlXSbiooKFi2eiXsPL+JdpC9evLqHbTtWw9TUWOYY5ua1cebcPoS+eYLg0MfY8M8yFCumlR/VK1rEknxL6urq0NbWlkm5NRAMDQ2hrKyMyMhImfzIyEiYmpp+U1VUVVVRv359BAQEfHP1BW8gAEDz5s1x7do1REVFITk5GW5ubujQoYPQYeW5pt2bwWb+SBxdfxjTuk5BiF8w7A8sho6Bjtzy6hrqiAyLwP4V+xAblfsPVQCoYlYVHQZ2QvDz4PwIXSFcefYaa675YEyLWjg8qh2qmehi/KE7iE1KyXWf4uoquD61mzRdsu0is728fgnM7lQPJ8a0x55hrVBKRwvjDt5BbJL8sYaUpXuvjpi/dAbWr9qGrq37we/pSxw4sR0GhvpyyzdsbI5/dq7E0YOn0KXV77hy8QZ2HtiAajVl5zfdvO6GhjVaSdOkkbNyHGv18k0yZfbsPJQvdSys+vbtjlWr5mPZsvWwsuoKX18/ODsfgJGRgdzyVlYN4eT0D/buPQpLyy5wdr6C48d3olatatIyxYpp4d49d8ybJ7+LHQB2716HqlUroW/fEbCw6ICzZy/j4MEtMDevned1LEpqdrNE23mD4LbhNBy7zUOkXxj6758FLQNtueXLN6mJ5+fu49CAZXDqtRAf3sViwP5ZKG6iJy0TGxyOq/b7sLvDHBzosxgJb2LQf/8saOqX+FXVKpR69+mK5Q5zsdJhI1o064GnT1/g9Jm9MMzls9PYsgF271mP/fuOo3nT7rhw/hoOHdmKmv9+drS0NGBerzb+XrkJLZr1wOCB41G1akUcObZDegxTU2OcdXZCUFAo2rbujT69bFCjRlVs3b7ql9SZfp6amhoaNmwIFxcXaZ5YLIaLi4tML8HXZGZmwtfXFyVLlvzm8xaIBoKi6DHSGtcOX8GN4y548+o1ts3ZgtRPqWjbX35PSYDPK+xbvgduzneQkZr7hG0NLQ1M3TgNW2b/g6SEj/kVfpG3/8FL9K5fEdb1KqCykTbmdW0ADVVlnPEK+cpeIhgW15Amg+Ky6wx3qVsOVpVMUEavOKoY62BaB3N8TM3Aq6j4/KxKoTdy/FAcdjqJ44fO4JV/EObYLcan5E/oP6iX3PLDxwzGLZe72P7PXgS8DMaa5Zvw1Oc5/hz5h0y5tNQ0REe9l6aEhMQcx0r6mCRT5lPyp3ypY2FlazsSjo6H4eR0HC9evMLEiXOQnPwJw4b1l1t+woThuHr1Ftat2w5//wAsWrQGT548xbhxf0rLHDp0CsuXb8CNG265ntfKqiG2bt2Lx4+9ERwchhUr/kF8fCIaNKib11UsUhqP7AzvIzfhe/w23r96h8tz9yDjUyrM+rWUW/7c5K3w3H8dUc/DEBsYjouzdkKkpIQKTT83xJ6fvY+Qu88Q/zoaMa/ewmXJQWhoa8G4ZrlfVa1CacLE4di39ygOHjgJ/xcBmGI7D8mfPmHIkL5yy48b/yeuX7uNjRt24qV/IJYtWQdvr2cYPWYIACAx8SOsewzD6VMXEfAqGI/dvTBj2kLUb1AXZcpk/RDs1LkN0jMyMG3qAgS8Coanpy+mTpmHntadUalS+V9W90JJIs6/9J3s7Oywc+dO7Nu3D35+fhg3bhySkpJgY2MDABg6dCjmzJkjLb948WJcvXoVQUFB8PT0xODBgxEaGoqRI799hIngDQQ9PT3o6+vnSAYGBihdujRatmyJPXv2CB3mT1NRVUHlulXg7eYtzZNIJPBx80L1BtV/6tijl47F4xuP4ZPt2PR90jPF8AuPh2XFz12zSiIRLCuawOdN7kOCPqVloPPGi+i44QKmHL2LgKiEr57jpGcQiquropqJbl6GX6SoqqqgrnktuN16IM2TSCRwu/UADRrJH8rSoJG5THkAuH3jXo7yVs0s4OnvipsPz2HZ6nnQ1cvZezdu8gh4B9zBRddjGDPpT5nllxWdqqoqGjSoK/NDXiKR4OZNN1haNpC7j5VVgxw//K9fv51r+dw8eOCBvn27Q09PByKRCL//3h0aGuq4dSv3VTwUnZKqMkzrVkSwW7bVSyQShLg9Q+kGOVcPlEdVUx1KqspIiZd/80lJVRn1BrZGSkISop5/3zrrikRVVRX16teB68170jyJRALXm/fQqLH8BVkaNa4P15t3ZfJcXO7kWh4AtLVLQCwWIyHhAwBATV0NaWnpkGRbWjPlU1YPtlUTix+uj0LIxyFG36t///5YvXo17O3tUa9ePXh5eeHy5cvSicthYWEIDw+Xlo+Li8OoUaNQs2ZNdOnSBYmJibh37x5q1co5fDM3BWKS8rJly9C5c2c0btwYAPDo0SNcvnwZEyZMQHBwMMaNG4eMjAyMGjUqx/6pqak5lobKlGRCWVSwvtRL6GtDWUUZCTFxMvnxMfEoXbnMDx+3WffmqFSnMmZ0t/vvwpSruORUZEokOXoADIqpIyQm511mAKhgUAILu1ugqokOPqamw+n+S/y59yZOju0AE+3P4ztvv3yHWaceIiU9E4YlNLBtcHPoaeXdZKSiRt9ADyoqKoiJlm2YxUS/R+VqFeXuY2RsiOgo2fLRUe9hZGwofe16ww2Xz19HWOhblK9YFrPm28Lp2FZYdxwM8b9Pttyz4xCeej9HfFwiLBqbY5b9FBibGGHJvL/zuJaFk6GhPlRUVBAVFSOTHxkZg2rV5I9BNzExQlRU9Bflo2FiYvRd5x40aDwOHNiM8HBfpKenIzn5E/r3H4WgIP4ozY2WXgkoqSgjOUb2xkVSTAIMKn/bUIPWcwbgY2Qcgu/KLpFYpU099Nw0EaqaavgYFY8jg1fiUxx7sHNj8O917cvPTnRUDKpVkz8nxMTEEFHRX17XYnL97Kirq2HRklk4cdwZHz5kvRe3b93Hcoe5sJ08Clu37EWxYppYuHgGAMDU9Ps+gySsiRMnYuLEiXK3ubq6yrxet24d1q1b91PnE7yB4ObmhqVLl2Ls2LEy+du3b8fVq1dx8uRJmJmZYePGjXIbCA4ODli0aJFMXnXtqqip83N35QsDg5KGGLFwFBYOskf6V4YgUf4wL2MA8zIGMq97b72CEx5BmND684TaRhWMcXR0e8Qnp+LUk2DMPPkAB4a3gX6x73vsOf0c51OXpf/293uFF89ewu3JJTRp1gh3bz8EAOza4iQt8+L5S6Slp8NhrT1WLl6PtDR+xoS0YME06Ohoo3PnPxATE4sePTriwIEtaNu2L5498xc6vCLJalx31OxuhYP9lyHzi++Y0Pt+cOz8FzT1i6PeH61hvWUi9vVciOT38m+oUP5SUVHBXqd/IBIBdlPspfkv/F5h7OgZWL7iLyxYNB2ZmZnYvtUJkZHR0hsjJJ9Ewf8+gg8xunLliszDH/6vbdu2uHLlCgCgS5cuCAoKkrv/nDlzkJCQIJOqaX9b1+mv9CE2EZkZmdAx1JPJ1zXURXx0XC57fV3lulWga6SHNRfX40TQGZwIOoM6Teqiq013nAg6AyUlwd/eQkNPSx3KIhHef5SdkPw+KRWGxb/th7yqshKqm+ridVySTL6mmgrK6ReHWRkDLOxuAWUlJZx+EpJXoRc5se/jkJGRkWPinqGRAaIj5Q/3io6KgZGxbHkjYwNEf3G3Lruw0Dd4HxOLChVzHzft5eELVVVVlClX+jtqUHTFxMQiIyMDxtl6ZoCsO52RkdFy94mMjIaxsdEX5Y1yLS9PpUrlMX68DcaMmYGbN+/C19cPy5ath6enL8aOHfb9FVEQyXEfIM7IhJah7FC6YoY6+Bid+3BIAGg8uguajOuGI4NXIvrF6xzb0z+lIi40Eu+eBOLizF0QZ4hh3l/+vAYC3v97Xfvys2Nk/LXPTgyMjb68ruUsr6Kigr37/0HZcqXRs8cwae/B/5047oxqla1Qo9pvqFjOAg7LN8DQUB8hITnfV6L/E/wXpL6+PpydnXPkOzs7Q18/a8WSpKQklCghf3UEeUtFFbThRQCQkZ6BQN8AmDX9vKSZSCRC3abm8Pf8sbtfPne9MbndBNh1spWmV96vcPvMLdh1suXdge+gqqyEmiV18SgkSponlkjwKDgKZmXkrzDxpUyxBAFRif/ZoJBIJEjLzPypeIuy9PQM+Ho/R9MWltI8kUiEpi2t4Okuf56Np7u3THkAaNaqSa7lAcC0lAn09HUR9ZUfqrXq1EBmZibeR399FTFFkZ6eDk9PX7Ru3VSaJxKJ0KpVUzx86Cl3nwcPPGXKA0CbNs1yLS+PpmbWZ+rLa1pmZiZvhHyFOD0TEb7BMhOMIRKhfNPaeOuZ+3KHlmO6oukkaxwdtgoRvt+2Mp5ISQRlNdWfDbnISk9Ph9eTp2jZ6jdpnkgkQstWTeD+6IncfdwfPZEpDwCtWzeTKf//xkHlyhXQs/tQxMXG5xpDdNR7JCUlo3efrkhJScXNrywKQChQcxCEIPgQo/nz52PcuHG4efOmdA6Cu7s7Ll68iG3btgEArl27hpYtC/+diXO7zsB2zVQE+gbglddLdBvRExpaGnA5dh0AYLtuKmIj3uPAyqxhDiqqKihTtWzWv9VUYGBigAq1KiIlKQURoeFISfqEsJdhMudITU7Bh7jEHPn034ZYVcP8s+6oVVIPdUrp4+CjV/iUnoGe5hUAAPPOPIJxCU3Yts1aNWX77eeoW1of5fSL40NKOvbdf4nwhCT0qp81Tv5TWgZ2uvmhVbVSMCyugfhPaTjqHoioxE9oX/PH550ogl1bnLBm8zL4ej2Dl6cvRowdAi0tTRw7dAYAsG7LMkSER2Hlkg0AAMftB3DMeQ9GTRiKG1fvoEfvTjCrVxuzp2YNP9QqpokpM8fhkvN1REfGoHzFspi70A4hQWG4dSNrEmCDRuao37Au7t15hKSPyWjQyBz2y2bg9LHzclc7UlQbN+7Crl1r4OnpC3d3L0yaNALFimnByekYgKzlSN+9i8D8+SsBAJs3O+LatWOYPHkULl26gX79eqBhQzNMmDBbekw9PR2ULVsaJUtmTbj7/3yGyMhoREZGw98/EAEBwdi82QGzZy9FbGw8unfvgLZtm6NXL5tf/BcoXB7tuoRua8YgwicY77wD0Wh4J6hqqcPn+C0AQLe1Y/AhIg63VmW9f1Zju6G5XR+cm7wFCW9iUMwoq/chLSkF6cmpUNVUx28Te+LVdQ98jIqHll4JNBjWHiVM9PDiwkPB6lkYbN7kiK3b/8YTT194eHhj/AQbFNPSwoEDJwAA23asRvi7CCxauBoAsHXLXly8fAgTJ43AlSs30advN9RvUAeTbf8CkNU4cDqwCeb16qB/35FQVlKS9lDExSUgPT1rWNioMUPw6IEnPiYloXWbZliydDYWLvhbOpGZSB7BGwijRo1CrVq1sGnTJpw6lfWQqerVq+PWrVv47beslvO0adOEDDHP3HV2g7a+DgbYDYKekR6Cnwdh8ZAFSIiJBwAYlTKCJFvLUs9EH+sub5S+th7bG9Zje+PpfV/M7z/3V4df5HWsXRZxyanYeus5Yj6moLqJDrYMbCaduByemAyRSCQtn5iShiUXPBHzMQXaGqqoWVIP+/5sjcpGWeuLKymJEBLzAdN87iM+OQ26mmqoXUoPjn+2QhVj+c++oCzOp69A30AfdnMmwMjYEM+fvsCQ38dKJy6XKlMS4myfFY9H3rAdPRvT507EzHmTERIUilGDJ+OlX9Zd0sxMMWrWroa+A3pAW0cbkRFRuHPzPlYv3ySdW5CWmobuvTthyqxxUFdTw+uwt9i9dT92ZpuXQMCJE84wNNSHvb0dTEyM4O39HD16DJFOvixbtpTMnf4HDzwwbJgtFi6cjsWLZyIgIAS//z4Kz5+/lJbp1q09du5cK3194MBmAMDSpeuwdOk6ZGRkoGfPYVi6dDZOnnRE8eLFEBgYgpEj7XDlys1fVPPCye/8Q2gZaKO5XR8UM9JB1PNQHBu6Csn/Lr6gXcpQ5nun/uC2UFFXRe9tk2WOc2fdKbitPwWxWAyDKiVRt+9kaOqVwKf4jwj3DsKB35ci5tXbX1q3wubUyQswMNTH3HlTYGJiCF8fP/TuZSNdYKFM2ZIyn51HDz0xcvhUzJtvB/uF0xAYGIqBA8bB79/PTqlSJujaLWuZ9LsPLsicq2vngXC7k9Vga9jQHHPnTkax4lp4+TIIU2zn4eiRM7+gxoXcDyxHWpSIJNnXvioiepXrLnQI9BWHln/9yZEknOq254QOgXIRlfz1MeMkrAUmLYQOgXKx4j2Xwi2oEj4GCh1CrpKWDs63YxebdyDfjp1XBO9BALLGlQYEBCAqKirHGNMWLXjRJSIiIqJfqJDMFcgvgjcQHjx4gIEDByI0NBRfdmaIRCJkcjInEREREf1KCr7Qi+ANhLFjx8LCwgIXLlxAyZIlZcZ4ExERERHRryV4A+HVq1c4ceIEqlQpeM8uICIiIiIFpOBDjARfQNrS0hIBAbmvx0xERERERL+O4D0IkyZNwrRp0xAREYG6detCVVX2QStmZlzxhoiIiIh+IQVf5lTwBkKfPn0AAMOHD8+xjZOUiYiIiIh+LcEbCMHB3/YYdyIiIiKiX0LB5yAI3kAoX748AOD58+cICwtDWlqadJtIJJJuJyIiIiKi/Cd4AyEoKAi9evWCr68vRCKR9FkI/1/ulEOMiIiIiOhXkij4cxAEX8Vo8uTJqFixIqKioqClpYWnT5/i9u3bsLCwgKurq9DhEREREZGiEUvyLxUCgvcg3L9/Hzdu3IChoSGUlJSgrKyMZs2awcHBAba2tnjy5InQIRIRERERKQzBexAyMzNRokQJAIChoSHevXsHIGtugr+/v5ChEREREZEiYg+CsOrUqQNvb29UrFgRlpaWWLVqFdTU1LBjxw5UqlRJ6PCIiIiIiBSK4A2EefPmISkpCQCwePFidOvWDc2bN4eBgQGOHj0qcHREREREpHD4oDRhdezYUfrvKlWq4MWLF4iNjYWenp50JSMiIiIiIvo1BG8gyKOvry90CERERESkqArJXIH8IvgkZSIiIiIiKjgKZA8CEREREZFQJAreg8AGAhERERFRdgreQOAQIyIiIiIikmIPAhERERFRdmLFXuaUPQhERERERCTFHgQiIiIiouw4B4GIiIiIiCgLexCIiIiIiLJjDwIREREREVEW9iAQEREREWUjkbAHgYiIiIiICAB7EIiIiIiIZCn4HAQ2EIiIiIiIslPwBgKHGBERERERkRR7EIiIiIiIspEoeA9CkWwgeHwMEToE+ooSwz2FDoFy8eHSAqFDoFyYDdgudAj0FWviHgkdAuVirW4ToUMgKnSKZAOBiIiIiOiHKXgPAucgEBERERGRFHsQiIiIiIiyEwsdgLDYg0BERERERFLsQSAiIiIiyoarGBERERER0WcK3kDgECMiIiIiIpJiDwIRERERUXacpExERERERJSFPQhERERERNko+iRl9iAQEREREZEUexCIiIiIiLLjHAQiIiIiIqIs7EEgIiIiIspG0ecgsIFARERERJQdhxgRERERERFlYQ8CEREREVE2EvYgEBERERERZWEPAhERERFRduxBICIiIiIiysIeBCIiIiKibDgHgYiIiIiICqzNmzejQoUK0NDQgKWlJR49evRN+x05cgQikQjW1tbfdT42EIiIiIiIshPnY/pOR48ehZ2dHRYsWABPT0+Ym5ujY8eOiIqK+up+ISEhmD59Opo3b/7d52QDgYiIiIgoG4k4/9L3Wrt2LUaNGgUbGxvUqlUL27Ztg5aWFhwdHXPdJzMzE4MGDcKiRYtQqVKl7z4nGwhERERERL9IamoqEhMTZVJqaqrcsmlpafDw8EC7du2keUpKSmjXrh3u37+f6zkWL14MY2NjjBgx4odiZAOBiIiIiCib/OxBcHBwgI6OjkxycHCQG0dMTAwyMzNhYmIik29iYoKIiAi5+7i5uWH37t3YuXPnD9e/wKxilJGRAVdXVwQGBmLgwIEoUaIE3r17B21tbRQvXlzo8IiIiIiIftqcOXNgZ2cnk6eurp4nx/7w4QOGDBmCnTt3wtDQ8IePUyAaCKGhoejUqRPCwsKQmpqK9u3bo0SJEli5ciVSU1Oxbds2oUMkIiIiIgWRn8ucqqurf3ODwNDQEMrKyoiMjJTJj4yMhKmpaY7ygYGBCAkJQffu3aV5YnFWZVRUVODv74/KlSv/53kLxBCjyZMnw8LCAnFxcdDU1JTm9+rVCy4uLgJGRkREREQkDDU1NTRs2FDm97BYLIaLiwuaNGmSo3yNGjXg6+sLLy8vaerRowdat24NLy8vlC1b9pvOWyB6EO7cuYN79+5BTU1NJr9ChQp4+/atQFERERERkUKSiISOQMrOzg7Dhg2DhYUFGjdujPXr1yMpKQk2NjYAgKFDh6J06dJwcHCAhoYG6tSpI7O/rq4uAOTI/5oC0UAQi8XIzMzMkf/mzRuUKFFCgIiIiIiIiITXv39/REdHw97eHhEREahXrx4uX74snbgcFhYGJaW8HRRUIIYYdejQAevXr5e+FolE+PjxIxYsWIAuXboIF1geGDpiAO56XcbLd49x9tpBmDf4euuta88OuPHgHF6+e4yrbqfQup3swy3WbFqKsFhfmeR0fKt0u1VTixzb/5/M6tfOlzoWFuPGDkPAywf4mBiIe27OaGRR76vl+/Tphqe+t/AxMRBPPK+jc6c2OcosXDAdr0M98SEhAFcuHUGVKhVzlOnSuS3uuTnjQ0IAoiOf4eSJ3dJt+vp6uOB8AGEhHkj6EITgQHdsWL8UJUpwYv6Xjtx6gs7zdqKx7XoMXnUQviHhuZY9e/8p6o1fI5Ma266XKfPl9v+nvdfc87kmhd+g4b/jpocznr6+hxOX9/3ntaVTj3a4fO8knr6+h/O3jqJlu6Y5ylSuWgHb9q+FZ+AteIe44eRVJ5Qs/Xl8bbkKZbB572o89LuOJ0G3sGHXChgY6ed53Qq74aMGwdP3Bt5E+eLKjeOo39Dsq+V7WHfC/ceX8SbKF7fvO6Ndh5Yy22fOmYT7jy8jNNwLAaHuOHl2LxpYyB6zcpUK2H94C/yDHyL4jSfOXzmMZs0t87xuRVHNYe3Q7/46DAtwRHfnhTCsl/t69dUHtkLXk/Mx+Ol2DH66HZ0Oz5ZbXqdKKbRztMOQ5zsw9OUu9Di/GMVKGeRnNYqkgvQcBACYOHEiQkNDkZqaiocPH8LS8vNnzNXVFXv37s1137179+LMmTPfdb4C0UBYs2YN7t69i1q1aiElJQUDBw6UDi9auXKl0OH9sO69OmL+0hlYv2oburbuB7+nL3HgxHYYGMr/UmvY2Bz/7FyJowdPoUur33Hl4g3sPLAB1WpWkSl387obGtZoJU2TRs6SbvN45CWzrWGNVjjsdAJhIW/g8+RZvta3IPv99x5Y/fcCLFm6Fo0sO8Hb5zkuXjgIIyP5F80mVhY4uH8z9uw5DIvGHXHu3BWcPLEbtWtXl5aZMX08Jk4YjvETZ+O3Zt2RlJyMi+cPykw86tWrC/bu2YC9+46hgUUHtGhljcNHzki3i8VinHO+il69bVCzdnOMGDkVbds2x5bNK/Ltb1EYXXn8AmtO3sKYrk1weM4QVCtthPH/nETsh+Rc9ymuoYbrDmOl6dLSUTLbs2+77jAWC4d0hEgEtKtfNb+rU6h1sW6PuYvtsGn1Dli3HQS/Zy/heGwT9A315Jav38gM67Yvw4mDZ9CzzUBcv+SKLfvWoGqNz5PkylUog8PndyPoVQgGW49G91YDsHnNLum64JpaGthzbDMgkWBI77Ho33UEVFVVsf3AOohEBWcYgNCse3fBkuVz8PeKTWjT3BrPfF/g+KndMMzlO6dR4/rY4bgWB52Oo3Uza1y8cB1OhzajRs3Pn4HAgGDMmr4YLZp0R9eOf+B12FucOL0HBgaf3+9Dx7ZDRUUFvboNRduWvfDs6QscPLYdxsY/voKKIqjY3RKW9oPwZN1pnO08D7HPw9DpwCxoGGjLLW/apCaCzt7HxX7L4NxzIZLexaLTwVnQMv38XpQob4xup+cjIfAdLv6+DKfbz4XXhjPITE3/VdUqMiRiUb6lwkAkkUgkQgcBZC1zeuTIEfj4+ODjx49o0KABBg0aJDNp+VuV06+bDxF+v7PXDsLb8xnsZy0HkNUz8tD3GvbuPIwtG3bnKL9599/Q0tKEzR8TpXlnrh7Ac19/zJ22BEBWD4K2TgmMGjL5m2JQUVHBo2fXsXfnYWxcvT0PavXz3n2M/eXnvOfmDPfH3pg8ZR6ArPciJMgdm7fswaq/N+cof+jgVhTT0kLPXsOkeXfvOMPL+xkmTJwNAHgd6ol167dj7bqsv6u2dgm8e+OF4SOn4tixc1BWVkbgq4dYtHg19uw98s2xTpwwHNPsxqFi5UY/U+Uf8uHSgl9+zm8xeNVB1C5vijn92wIAxGIJOv61A3+0qofhHXPeqTx7/yn+PuEKtzUTc2zLzZRtZ5Ccmo4dk3/Ps7jzktmAgvH5PXF5H3y8nmHx7FUAsj5Lt70vYv+uo9ixcW+O8ut3OkBLSxOjB02R5h2/tBd+T/1hPyNr3e91O5YjIz0DMybYyz1ns1ZW2HVkIyyqtMbHj0kAgOIlisMj4CZsfp+Ae7cf5W0lf0Bc6gehQ8CVG8fxxNMXs6cvBpD13vj43cbO7fuxcd2OHOV37VkPrWKaGNhvjDTvsssxPPXxw/Sp8q8FxUsUQ8jbJ+jVfRju3LoPfX09vAx5iG4dB+LB/cdZZYoXQ8i7J+jd40/cdr2XDzX9Piu1GwsdglzdnRcixjsI9+c5ZWWIRBjgvgHP91yDz2bn/9xfpCTC4Gc7cH/ePgScdAMAtN48AeKMTNyaXDhWfxzx5oDQIeQqvFnrfDt2Sbeb+XbsvFIgehBSUlKgoqKCwYMHY9WqVdiyZQtGjhz5Q42DgkJVVQV1zWvB7dYDaZ5EIoHbrQdo0Mhc7j4NGpnLlAeA2zfu5Shv1cwCnv6uuPnwHJatngddPZ1c42jfuRX09HVx7NCZH69MIaeqqooGDczgcuOONE8ikcDlhhusrBrK3cfKsqFMeQC4es1VWr5ixXIoWdIELjfcpNsTEz/g0aMnsLLMKtOgfl2UKVMSYrEY7o+u4HWoJ86f2y/TC/GlkiVN0Mu6M27fyf3piIomPSMTfmGRsKxeTpqnpCSCZY1y8AnOfZjRp9Q0dJ63Ax3nbseUbWcQ8C4m17LvE5Pg9jQY1r99+wQuRaSqqoLa5jVw79bnH+QSiQT3bj9CfQv5N2bqW5jh3u2HMnl3bt5HvX+HqYhEIrRq3wwhgWFwPLYJD55fw4nL+9CucytpeTU1VUgkEqSlpUnz0lJTIRaL0dCyXt5VsBBTVVWFeb3auHXz8w9yiUSCW6730KhxPbn7WDSuh1tf/IC/6eIGi8b1cz3HsD/7IyE+Ec98XwAAYmPj8OplEPr/YQ0tLU0oKytjmE1/REXFwNvrad5UrghSUlWGYd2KeHcnW8++RIJ3d57BuEGV3HfMRkVTHUqqykiN/5iVIRKhTNt6SAiKQMcDMzHQazO6Oy9E+Y7yv+fo6wraEKNfrUA0EIyNjTFs2DBcu3ZNulZrYadvoAcVFRXERL+XyY+Jfg8jE/nDWoyMDREdJVs+Ouo9jLJ107recIPduL/wh/UoOCxaD6umFnA6tjXXySn9B/fGrRv3EPEuUu52RWBoqA8VFRVERcr+QIyKioapiZHcfUxNjRAZFS2TFxkZIy1vamL8b94XZaJiYGqata1ipawftPbzp2G5wwb0tB6GuPgEuFw7AT09XZn9DuzfjMT4ALwO9UTih48YPWbGj1W2CIr7+AmZYgkMtIvJ5BuU0EJMYpLcfSqY6GPh4I5YN8Yay/7sArFYgj9XH0ZknPy7vOcePIOWhhra1uPwoq/R09eVe117/8V1KjtDYwPERMn2GsZEx8LIOOs6aGCkj+LFi2G07Z+47XIPNv0m4OrFm9i89280/q0BAMDLwxefklMww94WGpoa0NTSwKxFU6CiogJjEw5jAQCDf79zoqNlr3PRUTEwzuU6Z2xiiOioL6+LMTn+ph06tULIuyd4G+2LsRNs0NfaBrGxcdLtvXsMQ12zmtIy4ybaoH/vEUiIT8yj2hU9GvoloKSijE/RCTL5n2ISoGmc+02/7BrNHYDkiDi8c8tqZGgaakOtuCbMJnTDG1cfXB64EqGXPdB252SYWtXI8zpQ0VYgGgj79u1DcnIyevbsidKlS2PKlCl4/PjxN+2bmpqKxMREmSQpLM2zH+B86jKuXXaFv98rXL14AzYDJqJew7po0izncBTTUiZo2eY3HD1wSoBI6f+NNocVG3H69EV4PvHFiJF2kEgk6Nunm0zZadMXopFlR1j3/hOVKpXH6r8L5lCfwsK8Uil0t6qNGmWNYVGtLNaM6QG9Elo44eYjt/zZ+0/RpVENqKsWiIXdFIrSv3MIXC7fwt7th+D39CV2bNyLm1fv4I9hfQAAse/jYTtiFtp0aAHvkDvwDLwFbe0SeOrtB7G4QIySLdLcbj9E62Y90bl9f7hcv41de9fLzGtYtWYBomNi0a3jQHRo3RcXL1zHwaPbYZJLw4R+ntmE7qjU0wrXR62Xzi8QKWV9lsKueuLZrsuIfR4Gn83OCLvuhRqD2woZbqEkkYjyLRUGBaKB0KtXLxw/fhyRkZFYvnw5nj9/DisrK1SrVg2LFy/+6r4ODg7Q0dGRSYkp0V/d51eIfR+HjIwMGH4xCdbQyADRke/l7hMdFSO9q/Z/RsYGOe7wZBcW+gbvY2JRoWK5HNv6DbRGXGw8rl1y/f4KFCExMbHIyMjIcVfM2NgIEZHy/1uJiIiGibHsl5uJiaG0fERk1L95X5QxNkRERNa2iPCs//fzeyndnpaWhuDgUJQrV1pmv8jIaPj7B+L8+WsYP34Wxo0dJu2JUHR6xTWhrCTC+y96C95/SIbhF70KuVFVVkb1MsZ4HR2XY5tnwBuERMahV9OCMXepIIuLjZd7XTP4ynUqJuo9DI1lJ8kaGulLe0vjYuORnp6BgJdBMmUCXwajZJnPqxi5uT5A28Y9YVWzPRpXb4sZE+xhUtIIr0Pf5EXVCr33/37nGBnJXueMjA0Rlct1LioyJkfPj7GxYY7e1uTkTwgOCoOHuzemTPwLmZmZGDQ0a65O85ZN0KFTa4yymYJHDz3h4/0cM+0WISUlBf0H9srDGhYtKbEfIM7IhKaRbG+BpqEOPkUl5LJXljpjusBsfDdcHrgScX6vZY+ZnoH4l7LPj0oIeIvipbmKEX2fAtFA+L8SJUrAxsYGV69ehY+PD4oVK4ZFixZ9dZ85c+YgISFBJmlrCH/XIj09A77ez9G0xecJlCKRCE1bWsHT3VvuPp7u3jLlAaBZqya5lgeyegn09HXlfgH0G2iNk0edkZGR8YO1KBrS09Ph6emDNq2bSfNEIhHatG6GBw885O7z4KEH2rRpJpPXrm0Lafng4DCEh0fKHLNEieJo3Lg+HjzMKuPh6YOUlBRUq/Z5tRYVFRWUL18WoV/5UfP/ngd1dbVcyygSVRVl1Cxngkf+YdI8sViCR/5hMKtY8puOkSkWI+BdNAy1cy4fe/reU9QqZ4LqZdgg+y/p6Rl45v0CTVp87rEUiUT4rXkjPHnsK3efJ4990KS57CTRpi0t4fXYR3pM3yfPULFyeZkyFSqXx7vXETmOFxcbjw+JH2HVrBEMDPXhcvn2z1arSEhPT4e31zO0aPX5yaoikQgtWjaB+yMvufs8fuSFFi1ln8TasvVvePzoyVfPJVJSkl6ftDQ1AACSL3pyxGJxnq/LXpSI0zMR4xuMks2yLREsEqFUs9qI8gzIdb+647qi/mRrXBmyCjE+wTmOGe0dBJ3KstdF7Uol8fFt7jcaST5Fn4NQoPrTU1JScO7cORw6dEj6AIgZM74+FltdXV1mWUkAEIkKxkVp1xYnrNm8DL5ez+Dl6YsRY4dAS0tTOmF43ZZliAiPwsolGwAAjtsP4JjzHoyaMBQ3rt5Bj96dYFavNmZPzWokaRXTxJSZ43DJ+TqiI2NQvmJZzF1oh5CgMNy6cVfm3E1bWKJchTI4sp/DiwBg3Yad2LN7HTw8feDu/gS2k0ahWDFN7N13FACwx3ED3r0Lx1/zspYX/eef3bjhcgJTp4zBxUvX0b9fTzRsaIax42dKj7nxn12YO8cWrwKCEBLyGosWzsC7d5E4e/YKAODDh4/YvuMAFthPx5s37xAa9hbT7MYCAE6cPA8A6NypDYyNDfHYwxsfPyahdq3qWLFiHu7effTVRoSiGdKmIeY7XUat8qaoU94UB2964lNqOno2yZpUPG/vJRjrFoetddZzQ7ZfvI+6FUqinLEuPiSnYt91d4THfsjRS/DxUyquefpjWu9Wv7pKhZbjtgNY9c8iPPXyg4/nU/w5ZiA0tTRx8vA5AMCqTYsQGRGNNUs3AQD27TiMg2d3Yvi4wXC95oauvTqgTr1amDdtmfSYuzbvx/qdDnC//wQP7rqjRZvf0KZjcwy2/ry6Tp8/uiPwZTBi38ejnkVdzFs2HXu2HUJwYOiv/QMUYFs37cGmbSvh9eQpPB/7YOz4YdDS0sThAycBAJu3r0L4u0gsXbQGALB96z6cu3QA4ycOx9Urrujdtyvq1a8DO9v5AAAtLU1MnT4Oly+5IDIiGgYGehg+ahBKljTB2dOXAADuj7wQH5+ITdtWYvXKzUhJScGQYf1QrnwZXLviKsjfobB4uuMSWqwbgxjvYER7BaLOyE5Q0VTHy6O3AAAt1o9BckQcHq84BgAwG98NDab1geukLfj4Okba+5CelIKM5KwlgX23XUTrLRMR8fAF3t3zQ5lWZijXrj4u/r5MfhBEuSgQDYQrV67g0KFDOHPmDFRUVNC3b19cvXoVLVq0EDq0n+J8+gr0DfRhN2cCjIwN8fzpCwz5fax0gl+pMiVlxs96PPKG7ejZmD53ImbOm4yQoFCMGjwZL/2y7iZkZopRs3Y19B3QA9o62oiMiMKdm/exevkmpKXJrnHcf3BvPH74BIGvZO8wKKrjx8/ByFAfC+2nw9TUCN7ez9C122BE/TssolzZUjIT5O8/eIzBQydi8aKZWLpkFl4FBKNP3xF49sxfWubv1VtQrJgWtm1ZBV1dbdy9646u3QdL124HgFmzlyAzIwN792yEpqYGHj16gvYd+yE+PqsL+dOnFIwcMQhrVi+EuroaXr8Jx5kzF7FyVc6lVxVZR4saiPv4CVvP30VMYjKqlzHClol9pBOXw+MSpeNvASAxOQVLDl1FTGIytLXUUbOsCfZNH4DKJWW72S97+AMSoFMjTuD7VhfPXIO+gR4mzxoLI2MD+D19iRH9J+F9dNZE5FJlTJF99ewn7j6wG/sXps4Zh2l/TUBIUBjGD5uGVy8CpWWuXbyJBTOWY8xkG8xfPh3BgaGYaDMTHg+9pGUqVqmAafMmQkdXB29fv8PWdY7Ys+3gL6t3YXDm1EUYGOpj9lxbGJsY4amvH/r1GYHof79z/r+q2v+5P3qCMSOmYe78KfhrgR2CAkMwdOAEvPB7BQDIzMxE1WqVMGBgL+gb6CEuNg5PPH3RvdNA+L/I+l6KjY1D/94jMNd+Kk6f3wdVFVW8ePEKQ/4Yj2dPX/z6P0IhEuz8EBoG2mg4vQ80jXTw/nkorgxZhZSYrMndxUsbyvTM1BjSFsrqqmi7Q3aZc8+1p/BkbdbNwNDLj3F3jiPMJ/aA1eKhSAgMh8voDYh0fwn6PoXleQX5pUA8B0FLSwvdunXDoEGD0KVLF6iqqv7U8QrKcxBIPiGeg0DfpqA+B4EKznMQSL6C8BwEkq+gPgeBCvZzEMIs8m9id7nHLvl27LxSIHoQIiMjUaJECaHDICIiIiJSeAWigVCiRAlkZmbizJkz8PPzAwDUqlULPXv2hLKyssDREREREZEiUfQhRgWigRAQEIAuXbrg7du3qF496ymzDg4OKFu2LC5cuIDKlSv/xxGIiIiIiCgvFIjlfmxtbVG5cmW8fv0anp6e8PT0RFhYGCpWrAhbW1uhwyMiIiIiBSIRi/ItFQYFogfh1q1bePDgAfT1Pz9Mx8DAACtWrEDTpk0FjIyIiIiISLEUiAaCuro6PnzIuQLEx48foabGh0URERER0a8j/BqfwioQQ4y6deuG0aNH4+HDh5BIJJBIJHjw4AHGjh2LHj16CB0eEREREZHCKBANhI0bN6Jy5cpo0qQJNDQ0oKGhgd9++w1VqlTB+vXrhQ6PiIiIiBQI5yAUALq6ujh79iwCAgKky5zWrFkTVapUETgyIiIiIlI0Eknh+CGfXwRrINjZ2X11+82bN6X/Xrt2bX6HQ0REREREELCB8OTJk28qJxIpdguOiIiIiH4tiVjoCIQlWAMhew8BEREREREVDAViDgIRERERUUEhVvA5CAViFSMiIiIiIioY2INARERERJSNoq9ixB4EIiIiIiKSYg8CEREREVE2heWBZvmFDQQiIiIiomwkEqEjEBaHGBERERERkdQ39yCcO3fumw/ao0ePHwqGiIiIiEhoHGL0jaytrb+pnEgkQmZm5o/GQ0REREREAvrmBoJYrODPnCYiIiIihcAHpf2klJSUvIiDiIiIiIgKgB9qIGRmZmLJkiUoXbo0ihcvjqCgIADA/PnzsXv37jwNkIiIiIjoV5JIRPmWCoMfaiAsW7YMe/fuxapVq6CmpibNr1OnDnbt2pVnwRERERER0a/1Qw0EJycn7NixA4MGDYKysrI039zcHC9evMiz4IiIiIiIfjWJJP9SYfBDDYS3b9+iSpUqOfLFYjHS09N/OigiIiIiIhLGDzUQatWqhTt37uTIP3HiBOrXr//TQRERERERCUUsEeVbKgy+eZnT7Ozt7TFs2DC8ffsWYrEYp06dgr+/P5ycnHD+/Pm8jpGIiIiI6JcpLJOJ88sP9SD07NkTzs7OuH79OooVKwZ7e3v4+fnB2dkZ7du3z+sYiYiIiIjoF/mhHgQAaN68Oa5du5aXsRARERERCa6wTCbOLz/cQACAx48fw8/PD0DWvISGDRvmSVBERERERCSMH2ogvHnzBn/88Qfu3r0LXV1dAEB8fDx+++03HDlyBGXKlMnLGImIiIiIfpnCMpk4v/zQHISRI0ciPT0dfn5+iI2NRWxsLPz8/CAWizFy5Mi8jpGIiIiIiH4RkUTy/aOsNDU1ce/evRxLmnp4eKB58+ZITk7OswB/RPdy3QQ9P33d7Vg/oUOgXGiqqP13IRJE4HQO4SzIhu/8KHQIRIXO0dAzQoeQK/fSvfLt2I3ens63Y+eVH+pBKFu2rNwHomVmZqJUqVI/HRQREREREQnjhxoIf//9NyZNmoTHjx9L8x4/fozJkydj9erVeRYcEREREdGvxgelfSM9PT2IRJ8rlZSUBEtLS6ioZB0iIyMDKioqGD58OKytrfM8UCIiIiKiX0HBVzn99gbC+vXr8zEMIiIiIiIqCL65gTBs2LD8jIOIiIiIqEAoLEOB8stPPSgNAFJSUpCWliaTp62t/bOHJSIiIiIiAfxQAyEpKQmzZs3CsWPH8P79+xzbMzMzfzowIiIiIiIhSBS8B+GHVjGaOXMmbty4ga1bt0JdXR27du3CokWLUKpUKTg5OeV1jERERERE9Iv8UA+Cs7MznJyc0KpVK9jY2KB58+aoUqUKypcvj4MHD2LQoEF5HScRERER0S8hFjoAgf1QD0JsbCwqVaoEIGu+QWxsLACgWbNmuH37dt5FR0REREREv9QPNRAqVaqE4OBgAECNGjVw7NgxAFk9Czo6OnkXHRERERHRLyaBKN9SYfBDDQQbGxt4e3sDAGbPno3NmzdDQ0MDU6dOxcyZM384mJSUlB/el4iIiIgoL4gl+ZcKgx9qIEydOhW2trYAgHbt2uHFixc4dOgQbt68iWfPnn3XscRiMZYsWYLSpUujePHiCAoKAgDMnz8fu3fv/pHwiIiIiIjoB/1QA+FL5cuXR+/evaGjo/PdP+qXLl2KvXv3YtWqVVBTU5Pm16lTB7t27cqL8IiIiIiIvpkYonxLhUGeNBB+hpOTE3bs2IFBgwZBWVlZmm9ubo4XL14IGBkRERERkeIRvIHw9u1bVKlSJUe+WCxGenq6ABERERERkSIraJOUN2/ejAoVKkBDQwOWlpZ49OhRrmVPnToFCwsL6OrqolixYqhXrx7279//XecTvIFQq1Yt3LlzJ0f+iRMnUL9+fQEiIiIiIiIqGI4ePQo7OzssWLAAnp6eMDc3R8eOHREVFSW3vL6+Pv766y/cv38fPj4+sLGxgY2NDa5cufLN5/yuB6X17t37q9vj4+O/53AAAHt7ewwbNgxv376FWCzGqVOn4O/vDycnJ5w/f/67j0dERERE9DMK0oPS1q5di1GjRsHGxgYAsG3bNly4cAGOjo6YPXt2jvKtWrWSeT158mTs27cPbm5u6Nix4zed87t6EHR0dL6aypcvj6FDh37PIdGzZ084Ozvj+vXrKFasGOzt7eHn5wdnZ2e0b9/+u45FRERERFRUpKWlwcPDA+3atZPmKSkpoV27drh///5/7i+RSODi4gJ/f3+0aNHim8/7XT0Ie/bs+Z7i36x58+a4du1avhybiIiIiOh75OcDzVJTU5GamiqTp66uDnV19RxlY2JikJmZCRMTE5l8ExOTry7mk5CQgNKlSyM1NRXKysrYsmXLd914F3wOwuvXr/HmzRvp60ePHmHKlCnYsWOHgFERERERkaIS52NycHDIMQrHwcEhT+MvUaIEvLy84O7ujmXLlsHOzg6urq7fvP939SDkh4EDB2L06NEYMmQIIiIi0K5dO9SpUwcHDx5EREQE7O3thQ6RiIiIiChPzJkzB3Z2djJ58noPAMDQ0BDKysqIjIyUyY+MjISpqWmu51BSUpKuElqvXj34+fnBwcEhx/yEXPf/plL56OnTp2jcuDEA4NixY6hbty7u3buHgwcPYu/evcIGR0REREQKJz97ENTV1aGtrS2TcmsgqKmpoWHDhnBxcfkcm1gMFxcXNGnS5NvrIxbnGNb0NYL3IKSnp0v/KNevX0ePHj0AADVq1EB4eLiQoRERERERCcrOzg7Dhg2DhYUFGjdujPXr1yMpKUm6qtHQoUNRunRp6TAlBwcHWFhYoHLlykhNTcXFixexf/9+bN269ZvPKXgDoXbt2ti2bRu6du2Ka9euYcmSJQCAd+/ewcDAQODoiIiIiEjR5Ock5e/Vv39/REdHw97eHhEREahXrx4uX74snbgcFhYGJaXPg4KSkpIwfvx4vHnzBpqamqhRowYOHDiA/v37f/M5RRKJRJLnNfkOrq6u6NWrFxITEzFs2DA4OjoCAObOnYsXL17g1KlT333M7uW65XWYlIdux/oJHQLlQlNFTegQKBeB0xsKHQJ9xfCdH4UOgajQORp6RugQcnXB5I98O3bXyMP5duy8IngPQqtWrRATE4PExETo6elJ80ePHg0tLS0BIyMiIiIiRSQuOB0IghC8gQAAysrKMo0DAKhQoYIwwRARERERKTBBGggNGjSAi4sL9PT0UL9+fYhEuTfTPD09f2FkRERERKToxAVoDoIQBGkg9OzZU7pykbW1tRAhEBERERHJJegE3QJAkAbCggULAACZmZlo3bo1zMzMoKurK0QoRERERESUjaAPSlNWVkaHDh0QFxcnZBhERERERFL5+aC0wkDwJynXqVMHQUFBQofxy3QZ2hW77u7GyZensPrsGlQ1r5Zr2XLVymHOtjnYdXc3nMPOo8eIHl89dt/xfeEcdh4jF4zK67AVxsjRg+Hz7BYiY57D5eZJNGho9tXy1r06w93zKiJjnuPew4to36GVzPbZc23h7nkV7yJ9EfraE2edndDQwjwfa1B02IwcCHef6wiJ8MLF60dQv0Hdr5bv3rMj7jy6gJAIL9y8exZt27fItezKtQsQEe+HUeOGyuS7+1xHRLyfTJo4ZWSe1KcoU2nYDpoT10FrtiM0bBZCqVSlb9pPuZYVis07APXfp8huUFWHWseh0LTdCK1ZjtAcsxIqDdrkfeAKosPQzvjHbQf2+x/D0jOrUNm8aq5ly1QtC7tts/CP2w4cDT2DLsO75yjTd8oAHA09I5PWumzKzyoUWXxvqKASvIGwdOlSTJ8+HefPn0d4eDgSExNlUlHSrHtzjJw/EofXH8aUrpMR7BeMxQcWQ8dAR255dQ11RIRFYN+KfYiNiv3qsauaVUWngZ0Q/Dw4P0JXCL37dMVyh7lY6bARLZr1wNOnL3D6zF4YGsl/YF9jywbYvWc99u87juZNu+PC+Ws4dGQratb63OgLeBWMGXYL8ZtlF3Ts0B9hoW9w+uw+GBjq/6JaFU49e3XGwmWzsGblZnRo2QfPnvrj8KmdMMzl72bRuB627l6Nw/tPon2L3rh00QV7Dv6DGjVzftl27tYODRuZI/xdpNxjrVy2EXWrNZcmxx0H87RuRY1yLUuotR+E9Dun8WnXPIgjw6DxxyxAS/ur+4l0DKHWbiAyw17k2KbWfhCUK5sj9exWfNo2E+mPLkOt0zAoV22QX9Uospp0a4qh84bj5IYjmN3NDqF+IZi7fwG0c/ve0VRHZFgEDq90QtxXvnde+4ditMWf0rSg75z8qkKRxfemYBOLRPmWCgPBGwhdunSBt7c3evTogTJlykBPTw96enrQ1dXNsfRpYWc90hpXDl+By/HreP3qNbbM2YzUT6lo37+93PKvfF5hz/I9uON8G+mp6bkeV0NLA9M2Tsc/s//BxwQ+rOdHTZg4HPv2HsXBAyfh/yIAU2znIfnTJwwZ0ldu+XHj/8T1a7exccNOvPQPxLIl6+Dt9QyjxwyRljlx3BmurvcQEvIaL/xeYe6c5dDRKYE6dWr8qmoVSmMmDMPBfcdx5OBpvPQPxMypC/EpOQUDBveWW37U2KG4ed0NW/5xxKuXQVi1bCN8vf1gM2qgTDnTksZYtvIvTBg1ExkZGXKP9fFjEqKjYqQpOflTntevKFG17IyMJzeR4X0bkph3SLu4B5L0VKjWa5n7TiIR1K3HI/32SYjjonJsVi5TFRk+dyAO9YMkIQYZT25CHBkGpdLf1jNBn3Ud2RMuR67C9fgNvH31BrvmbkXap1S07tdWbvlAnwAcXL4P95zdkJ4q/zMCAJkZYiREx0vTh7gP+VWFIovvDRVkgjcQbt68KU03btyQpv+/LipUVFVQpW4VeLt5SfMkEgm83LxQvcHP/Vgcu3QcHt9wh7eb909GqbhUVVVRr34duN68J82TSCRwvXkPjRrXl7tPo8b14Xrzrkyei8udXMurqqriT5sBiI9PhK8vnyadG1VVVZjVq43bt+5L8yQSCe7cug+LxvXk7tOwkblMeQBwveEmU14kEmHT9pXY8o8j/F8E5Hr+SVNG4nnQfVy7fRLjJw2HsrLyT9WnSFNShlLJisgMfpYtU4LMkGdQKl0l191Um/eCJCkRGV635G7PfPMKytUaQFQi6yaRUvmaUNI3RWaQb15GX+Qpq6qgUt3K8HXzkeZJJBL4unmjaoPqP3Vs04olsfWRIzbe2YZJG6bCoJThz4arUPjeFHySfEyFgeAPSmvZ8it3mb5BamoqUlNTZfIyJZlQFhWsL3VtfW0oqygjLiZeJj8+Jh5lKpf54eM2794CletUhl33qT8ZoWIzMNCDiooKoqJiZPKjo2JQrZr8u5YmJoaIin6fo7yJiZFMXsdOreG4dwO0tDQRERGFXj2GIvY9J+bnRt9AFyoqKoiO+vJv+x5VqlaUu4+xiSGic7x372Fs/PmLceKUkcjIyMSubftzPfeu7fvh6/0ccXEJaNS4PuYumApjUyMs/GvlT9So6BJplYBISRmSpASZfMnHBCgZlJS7j1LZalCp1wqfds7N9bhpV5yg1nUEtCb/A0lmBiCRIO3CbojD/PM0/qJOW68ElFWUkfDF905CTAJK/cT3ToDXS2ydthHvgt5Cz1gPfaYMwKLjyzG9gy1SklJ+MmrFwPeGCjrBGwgAEBcXh927d8PPL+uuaq1atWBjYwN9/f8ep+3g4IBFixbJ5FXVrorqOrlP/i0qDEsaYtTCUbAfNP+rQ5BIWHduP0Dz37pD30APf/7ZH3ud/kGb1n0Q80XjgvKPmXktjBo7BO1b9vlque2b90n/7ffsJdLT07Fq3UIsX7QWaWn8jP00NQ2o9xyLtAu7gE+5D4dUadQByqWrIOXoGogTYqBcrgbUOg2D+GMcxDK9FSQEL9fPDzANexGKV16vsPnuDjTp1gw3j14XMDLie5N3CstqQ/lF8CFGt2/fRoUKFbBx40bExcUhLi4OGzduRMWKFXH79u3/3H/OnDlISEiQSVW0K/+CyL9PYmwiMjMyoWeoK5Ova6iLuOgfu5tcpW4V6BnpYf3FDTgTdBZngs6ibpO66G7THWeCzkJJSfC3t9B4/z4OGRkZMnecAcDI2BCRkdFy94mMjIHxFxOY5ZVPTv6EoKBQPHb3wsQJc5CRkYmhQ3/P2woUIbHv45GRkQEj4y//tgY5enj+LyoyBkY53rvP5S1/s4ChkQE8nt7AmxhfvInxRdlypbFw6Uy4++T+pen52AeqqqooW670T9aqaJIkf4BEnAlRMdlJlaLiOpB8TMhRXknPGEq6xlDvPw1ac/dBa+4+qJg1g3K1BtCauw8iPWNARRVqrfsh7dpBZL56AknUa2Q8voaM5w+hatX1V1WtSEiM+4DMjEzofPG9o2Oog/gf/N6RJzkxCeHB72Ba3jTPjlnU8b0p+MSi/EuFgeC/ICdMmID+/fsjODgYp06dwqlTpxAUFIQBAwZgwoQJ/7m/uro6tLW1ZVJBG14EABnpGQjwDYBZ089LXIpEIpg3NYe/Z85VPL6F911vTGg3AbadbKXplfdL3DrjCttOthCLFb39++3S09Ph9eQpWrb6TZonEonQslUTuD96Incf90dPZMoDQOvWzXIt/39KSiKoqav9fNBFVHp6Ony8nqF5SytpnkgkQrMWVnj8yEvuPh7u3jLlAaBFq9+k5U8cOYc2Ta3RrnlvaQp/F4ktGx0xoHfuy5jWrlsDmZmZiIn++ipiCkucCXF4MJQr1s6WKYJyhdoQv805z0McE47k7bPxaedf0pT50hPiED982vkXJAnvASUViJRVAMkX1y+JGKJCsvpHQZGZnoEg30DUbfp5uWaRSIQ6Tc3wyjPvhmupa2nApLwp4qI4dPJb8b2hgk7wIUYBAQE4ceKEzERAZWVl2NnZwcnJScDI8t6ZXWcwdc1UBPi+wkuvl+g5oic0tDRw/VjWHcyp6+zwPuI9nFZmDXNQUVVB2apls/6tpgIDEwNUrFURKUkpCA8Nx6ekTwh7GSpzjpTkVCTGfciRT/9t8yZHbN3+N554+sLDwxvjJ9igmJYWDhw4AQDYtmM1wt9FYNHC1QCArVv24uLlQ5g4aQSuXLmJPn27oX6DOphs+xcAQEtLE9NnjMfFiy6IjIiCgYE+Ro4ejJKlTHHm9CXB6lkYbN+8Dxu2OsD7yVM88fDFqHFDoVVME0cOngYA/LNtBcLfRWL54nUAgJ3bnHD6ghPGTvwT16/cgnWfLjCvXxszpmQ9tT0uLh5xcfEy58jIyEBUVAwCA0IAAA0b1UMDCzPcvfMQHz8kwaJxPSxePhsnjzkjIaFoLbmcl9IfXoJ6jzEQhwcj820gVC07QaSqjnTvrAnIaj3GQPIhDuk3jwGZ6ZBEv5HZX5KSnPX//89P+4TMUD+otf0DqRnpkPw7xEilbjOkXeOSs9/rwq6zGL9mMgJ9AhDo/QpdhneHupYGXI+7AAAmrJ2M2Ij3OLzqAICsybNlsn3v6Jnqo3ytikhJ+oTI0AgAwOC//oTHdXfEvI2Gnokefp/6B8SZYtw9d0eYShZSfG8KNjEU+4aE4A2EBg0awM/PD9Wry87a9/Pzg7l50XqglJvzHejo62CQ3WDoGekh6HkQFgyxR/y/k5SMShlBku2uv76JPjZe/kf6uvfYPug9tg987/tibn+ua5zXTp28AANDfcydNwUmJobw9fFD71420smyZcqWlOmVefTQEyOHT8W8+XawXzgNgYGhGDhgHPyevwQAZGZmolr1yvhjUG8YGOghNjYenh4+6NyhP174vRKkjoXF2dOXYGCoh5lzbWFkbIhnvn74o89o6byN0mVk34vHj7wwfuQMzJo3GXPmT0VwYChsBk36rr9zWloarHt3wfTZE6CmpobXoW+wfcs+bN+8N6+rV6RkPn+INC1tqLbsA7ViOhBHhiLl8CogKatRpaRjCLHk+9btSD21Capt+kO95ziINItDkhCDNNfjyPB0yY8qFGn3z9+FtoEO+tn9AV0jPYQ8D4bD0EVIiMkaAmZQyghi8ef3R99EH6surZO+7jGmF3qM6YVn959i8YB5WfuYGsD2n2kooVsCibEJ8Hf3wzzrWfgQy4b09+B7QwWZSCL5zit3Hjt69ChmzpyJSZMmwcoqa4jAgwcPsHnzZqxYsQI1a9aUljUz+/pTbf+ve7lu+RIr5Y3bsVzis6DSVOHQp4IqcHpDoUOgrxi+k8+gIfpeR0PPCB1Crg6UGpxvxx787kC+HTuvCN6D8McffwAAZs6cKXebSCSCRCKBSCRCZmbmrw6PiIiIiEihCN5ACA4OFjoEIiIiIiKpwrLaUH4RtIGQnp6ORYsWYf78+ahYUf4DkIiIiIiI6NcRdJlTVVVVnDx5UsgQiIiIiIhkiPMxFQaCPwfB2toaZ86cEToMIiIiIiIAgCQfU2Eg+ByEqlWrYvHixbh79y4aNmyIYsWKyWy3tbUVKDIiIiIiIsUjeANh9+7d0NXVhYeHBzw8PGS2iUQiNhCIiIiI6JfiJGWBcRUjIiIiIqKCQ/AGAhERERFRQVJYJhPnF8EbCMOHD//qdkdHx18UCRERERERCd5AiIuLk3mdnp6Op0+fIj4+Hm3atBEoKiIiIiJSVOxBENjp06dz5InFYowbNw6VK1cWICIiIiIiIsUl+HMQ5FFSUoKdnR3WrVsndChEREREpGAkovxLhYHgPQi5CQwMREZGhtBhEBEREZGC4RAjgdnZ2cm8lkgkCA8Px4ULFzBs2DCBoiIiIiIiUkyCNxCePHki81pJSQlGRkZYs2bNf65wRERERESU19iDILALFy5AIpGgWLFiAICQkBCcOXMG5cuXh4qK4OERERERESkUwScpW1tbY//+/QCA+Ph4WFlZYc2aNbC2tsbWrVsFjo6IiIiIFI0kH1NhIHgDwdPTE82bNwcAnDhxAiYmJggNDYWTkxM2btwocHRERERERIpF8DE8ycnJKFGiBADg6tWr6N27N5SUlGBlZYXQ0FCBoyMiIiIiRSMuJMuR5hfBexCqVKmCM2fO4PXr17hy5Qo6dOgAAIiKioK2trbA0RERERERKRbBGwj29vaYPn06KlSoAEtLSzRp0gRAVm9C/fr1BY6OiIiIiBSNOB9TYSD4EKO+ffuiWbNmCA8Ph7m5uTS/bdu26NWrl4CREREREZEiKiw/5POL4A0EADA1NYWpqalMXuPGjQWKhoiIiIhIcRWIBgIRERERUUFRWJYjzS+Cz0EgIiIiIqKCgz0IRERERETZcJlTIiIiIiKif7EHgYiIiIgoG0VfxYg9CEREREREJMUeBCIiIiKibLiKERERERER0b/Yg0BERERElI1YwfsQimQDITYzWegQ6Cuqa5cROgTKhU9ssNAhUC5+2xQkdAj0FY+fHhA6BMpFqcqdhQ6BCiFOUiYiIiIiIvpXkexBICIiIiL6UYo9wIg9CERERERElA17EIiIiIiIsuEcBCIiIiIion+xB4GIiIiIKBuxSOgIhMUeBCIiIiIikmIPAhERERFRNnxQGhERERERSSl284BDjIiIiIiIKBs2EIiIiIiIshHnY/oRmzdvRoUKFaChoQFLS0s8evQo17I7d+5E8+bNoaenBz09PbRr1+6r5eVhA4GIiIiIqIA6evQo7OzssGDBAnh6esLc3BwdO3ZEVFSU3PKurq74448/cPPmTdy/fx9ly5ZFhw4d8Pbt228+JxsIRERERETZiCHJt/S91q5di1GjRsHGxga1atXCtm3boKWlBUdHR7nlDx48iPHjx6NevXqoUaMGdu3aBbFYDBcXl28+JxsIRERERES/SGpqKhITE2VSamqq3LJpaWnw8PBAu3btpHlKSkpo164d7t+//03nS05ORnp6OvT19b85RjYQiIiIiIiykeRjcnBwgI6OjkxycHCQG0dMTAwyMzNhYmIik29iYoKIiIhvqsusWbNQqlQpmUbGf+Eyp0REREREv8icOXNgZ2cnk6eurp4v51qxYgWOHDkCV1dXaGhofPN+bCAQEREREWXzo6sNfQt1dfVvbhAYGhpCWVkZkZGRMvmRkZEwNTX96r6rV6/GihUrcP36dZiZmX1XjBxiRERERESUTUGZpKympoaGDRvKTDD+/4TjJk2a5LrfqlWrsGTJEly+fBkWFhbfXX/2IBARERERFVB2dnYYNmwYLCws0LhxY6xfvx5JSUmwsbEBAAwdOhSlS5eWzmNYuXIl7O3tcejQIVSoUEE6V6F48eIoXrz4N52TDQQiIiIiomy+fzHS/NO/f39ER0fD3t4eERERqFevHi5fviyduBwWFgYlpc+DgrZu3Yq0tDT07dtX5jgLFizAwoULv+mcbCAQERERERVgEydOxMSJE+Vuc3V1lXkdEhLy0+djA4GIiIiIKJv8nKRcGHCSMhERERERSbEHgYiIiIgoG0mBmoXw67EHgYiIiIiIpARtIKSnp6Ny5crw8/MTMgwiIiIiIilxPqbCQNAhRqqqqkhJSREyBCIiIiIiGd/7QLOiRvAhRhMmTMDKlSuRkZEhdChERERERApP8EnK7u7ucHFxwdWrV1G3bl0UK1ZMZvupU6cEioyIiIiIFJFi9x8UgAaCrq4u+vTpI3QYRERERESEAtBA2LNnj9AhEBERERFJcQ4C/VK9h/XEiQeHcCPwMnY4b0bNejVyLVuxWgUs27EQJx4cwt23N9BvZM6eFnNLM6zcuwxnPY7h7tsbaN6xaX6GX6T0/dMaZx4ewZ2gq3A8vxW1vvJeAEDbbq1w7LYT7gRdxSGXPfitjaXMdn1DPdivm40LnidxO/AKNhxchbIVS8uUMTDSx8KNf+GS1yncCrgMpys70bpLizyvW1EwZsxQ+PvfRXz8S9y+fRYWFuZfLd+7d1d4e99AfPxLPH58FR07tpbZ3rNnJ5w/fwBv33ojJSUMZma1chzDxMQIjo7rERLyGO/fv8D9+xdgbd05T+tVFPS36Y2L7ifxKOQmDlzciTr1a361fPvurXHmzmE8CrmJEzf3o1nbJjLbNbU0MWe5Ha56nsHD4Js4dfsgfh9qLVNm16lN8I64J5PmrZyR11Urkg6fdEaHPsPQoHUP/DFqCnyf+3+1fOKHj1i6ZjNa9RiI+q26o+uAkbh975F0++bdB1CnaWeZ1P2PUfldjSJp+MiB8PBxwetIH1x2OYb6Dep+tXwP6064534JryN9cOveObRrn/v3x9/rFiE6wR9jxg3L67BJARSIBsKJEyfQr18/WFlZoUGDBjKpKGnboxUmLRgHx7VOGN5pDAKeB2LtwZXQNdCVW15dUx3vwsKxdflOxES+l1tGU0sDAc8DseavjfkYedHTrkdrTFkwAbvW7sPQjqPw6nkgNh5aDb1c3ou6FrWxZMt8nDt8EUM6jMKty3fwt+MyVKpeUVrmb8dlKF2+FKbb/IXBHUYi/E0kNh1dCw1NDWmZBRvnonzlspj251z80cYGrhdvY/n2hahWp2p+V7lQ6du3O1atmo9ly9bDyqorfH394Ox8AEZGBnLLW1k1hJPTP9i79ygsLbvA2fkKjh/fiVq1qknLFCumhXv33DFvnkOu5929ex2qVq2Evn1HwMKiA86evYyDB7fA3Lx2ntexsOrYsy2mL7TF9jWOGNDBBv7PArD18DroG+rJLW9uUQcrti7C6cPO6N/+T9y8dBvr96xAlRqVpGWmL7LFb62tMHfiIvRq8QcO7jiG2cvt0LJDM5ljndh/Fm3qdpOmdUs252tdi4JL129h1T87MG74IBx3/AfVq1TEGLt5eB8XL7d8eno6Rk2Zi7fhkVi79C+cP7wLC2fZwtjIUKZclYrl4XruoDQ5bV39C2pTtFj37ozFy+dg9crNaNuiF549fYFjp3fD0FBfbvlGjetj++41OLj/BNo0t8alCy7Yd2gzatTM+f3RpVs7WFiYI/xdZH5Xo8hS9GVOBW8gbNy4ETY2NjAxMcGTJ0/QuHFjGBgYICgoCJ07F607d/1H/Q7nQxdx8dhlhLwKxd+z1yH1Uyq6DZBfzxfe/ti8dDtczt1Eelq63DIPbj7CzlWOuH3ZLT9DL3IGju6HM4fO4/zRSwh+FYoVs9Yg5VMKuv/RRW75ASP74sHNRziw9QhCAkKx/W9HvPB9iX42vQAA5SqVQV2L2lg5ey38vF8gLPA1Vs5eC3UNdXTs1VZ6HDOL2jjmeArPvV7gXVg4HDfsx8eEj6hpVk3ueRWVre1IODoehpPTcbx48QoTJ85BcvInDBvWX275CROG4+rVW1i3bjv8/QOwaNEaPHnyFOPG/Sktc+jQKSxfvgE3buT+WbGyaoitW/fi8WNvBAeHYcWKfxAfn4gG/3FXT5EMGTMApw6ew9kjFxD0MgRLZ65CyqdUWA/oJrf8oFH9cO/mQ+zbcgjBr0KxedVO+Pn6Y4DN5x7Reo3qwvnYRTy+9wTvXkfg5IGzePksAHXqy/bypHxKwfvoWGlK+picr3UtCpyOnkbf7p3Rq2sHVK5YHvYzJkFDXR2nz1+VW/7U+atISPyAjSvs0cCsNkqXNEGj+maoUbWSTDllZWUYGuhLk56uzq+oTpEydoINDuw7hsMHT+GlfyCmT1mAT8kpGDhE/rzM0eOG4sb1O9i8cTdevQzCimUb4OP9HCNGD5YpZ1rSGA6r5mPsqOlIT5f/24HovwjeQNiyZQt27NiBf/75B2pqapg5cyauXbsGW1tbJCQkCB1enlFRVUF1s2pwv+MhzZNIJHjs5oE6DXMOdaD8o6Kqghpy3gv3Ox6o21D+neK6DWvjUbbyAPDglru0vKqaGgAgNTVN5pjpaekwb/T5x6XP42do36M1tHVLQCQSoX3PNlDTUIPHPa+8ql6hp6qqigYN6sr8kJdIJLh50w2WlvJ7Fa2sGuT44X/9+u1cy+fmwQMP9O3bHXp6OhCJRPj99+7Q0FDHrVv3v78iRZCKqgpqmlXHg9uPpXkSiQQP7rjDzKKO3H3MGtbBg9vuMnn3XB/KlPdy90XLjs1hbJp1l7pR0wYoX7ks7t96JLNflz4d4PrsIk66HoDt3LHQ0FTPq6oVSenp6Xju/wpWjepJ85SUlGBlUQ/eT+U/oNTV7QHM69TEsjWb0aLbH7AePBY79h1BZmamTLmwN2/RuscgdPrdBrMWrkR4RFR+VqXIUVVVhXm92rjlek+aJ5FIcNv1Hiwa1Ze7j0WjerjtKnstuuniBots769IJMKWHX9j88bd8H8RkC+xKwpJPv6vMBB8knJYWBh+++03AICmpiY+fPgAABgyZAisrKywadMmIcPLM7r6OlBRUUZsTJxMfmx0HMpVLidQVIop671QQWz0F+9FTBzKV5H/XhgY6ct97/SNs7qCQwJCEf4mAhPmjIbDrNVZd4FG/w6TUsYwNPk8LGbumIVYvm0Brj8/j4z0DKR8SsHMEfPwJuRtHtey8DI01IeKigqiomJk8iMjY1CtWmW5+5iYGCEqKvqL8tEwMTH6rnMPGjQeBw5sRni4L9LT05Gc/An9+49CUFDo91WiiNLT14WKigreR8fK5L+PjkXFKuXl7mNobID3X3zW3kfHwdD48+dixV9rYb96Fq55nUN6egYkYjEWTV8Bzwde0jKXTl1D+JsIREVEo1qtKpgybzwqVC4HuxFz866CRUxcfCIyM8Uw0Jcd/mWgr4fgsDdy93nzLgJvPb3RtUNrbF29GGFv3mHpms3IyMzE+OGDAABmtapj6V/TUKFcGcS8j8UWx4MYOn4GzuzfimLFtPK9XkWBvoEeVFRUEB0lO3w4Kvo9qlSrJHcfYxPDHNfF6Oj3MDb5PPzLduooZGRkYMc2p7wPWsEUlqFA+UXwBoKpqSliY2NRvnx5lCtXDg8ePIC5uTmCg4Mhkfx3Kys1NRWpqakyeWKJGEoiwTtHSIFkZmRi1oj5mLd2Jlz8LiAjIwPudzxw1+UBRCKRtNzYmSNQXLs4JvSbivjYBLTs1AzLty3E6F62CHwRJGANCAAWLJgGHR1tdO78B2JiYtGjR0ccOLAFbdv2xbNnX5/YST/ujxF9YdagNmyHzMC7NxFo2KQe5jpMQ3REDB7eyeqtOHngrLR8wIsgxES+x86T/6BM+dJ4E8oGdl4RSyTQ19PFwpm2UFZWRu0aVREV8x57Dp2QNhCaN2kkLV+9SkXUrVUdHfoMw+Ubd9Cne0ehQld4ZvVqY/TYoWjTorfQoVARIHgDoU2bNjh37hzq168PGxsbTJ06FSdOnMDjx4/Ru/d//0fu4OCARYsWyeSVKV4B5bQr5rKHMOJjE5CRkZljIp++kR5iv7gbR/kr673IgL7RF++FoV6OO6P/9z46Vv57F/W5/AvflxjcfiSKlSgGVVUVxMcmwPH8Vvj5ZP2wLF2+FPoN740BrYYh6GUIAODV80DUszTD739aY8XstXlYy8IrJiYWGRkZMDaWnRRpYmKIyMhouftERkbD2Njoi/JGuZaXp1Kl8hg/3gb167eDn99LAICvrx+aNm2MsWOHYdIk3qmOi41HRkYGDIxkJ1EaGOkjJkr+Zycm6j0MvvisGRjpIebfO6fqGmqwnTMWU4fPwZ3rWcMtXvkFonrtqhg2bqC0gfAl3yfPAADlKpZhAyEXerraUFZWwvvYL3pwYuNgqC9/UrnRv3e2lZWVpXmVypdFzPs4pKenQ1VVNcc+2iWKo3zZ0gh78y5vK1CExb6PQ0ZGBoyMZRdeMDYyQFRkjNx9oiJjclwXjbKVb9LEAoZGBvB6dlO6XUVFBYuWzcLocUPR0Kwt6NsVlqFA+UXw2+w7duzAX3/9BQCYMGECHB0dUbNmTSxevBhbt279z/3nzJmDhIQEmVSmhPyubiFlpGfA3+clLJp9HhMtEonQsFkDPPV4LmBkiicjPQMvfF6iUbOG0jyRSASLZg3g6/FM7j6+Hs/QqHlDmTzLFhZyyyd9SEJ8bALKViyNmubVcftK1tj4/69mJBbLXnTEmWKIlAT/KBYY6enp8PT0RevWn5fsFYlEaNWqKR4+9JS7z4MHnjLlAaBNm2a5lpdHU/r+yHYsZ2ZmQonvD4Csz46fjz8sm8t+diybWcDn8VO5+/h4PIVlcwuZPKsWjaXlVVRUoKqmmuPvLs4Uf/XvXr121sot0bn8mKKsce61qlfFw8de0jyxWIyHHl4wryN/adp6dWsj7M07mfcj5PVbGBnoy20cAEBy8ie8fhsOo1xW36Gc0tPT4e31DC1afl7yVyQSoXnLJnjs/kTuPo/dvdC8pZVMXsvWv+GxuxcA4NiRs2j5Ww+0bmYtTeHvIrF542706z0y3+pCRZPgPQhKSkoyXwIDBgzAgAEDvnl/dXV1qKvLTlQrqMOLju48jr/WzcYLH388f/IC/Ub1gYamBi4cvQwAmLdhNmLCY7BtxS4AWRMCK1bLauyoqqrAyNQQVWtXRnLSJ7wNybpTo6mlgTLZ1tovVa4kqtaujMS4D4h8x0ljuTm04xgWrJ8DP+8XePbkBQaM6gtNLU2cP3IJALBww1xERURji8NOAMCRXSew/eRGDBzTD3ddHqBDzzaoaVYdy2d8XtqvbbdWiHsfj4i3kahSsxLsFk/CrctueHgr6w5oSEAowoLeYM6qadiweAsS4hLRslMzNG5hAbuhs3/9H6EA27hxF3btWgNPT1+4u3th0qQRKFZMC05OxwBkLUf67l0E5s9fCQDYvNkR164dw+TJo3Dp0g3069cDDRuaYcKEz39XPT0dlC1bGiVLmgCAdD5DZGQ0IiOj4e8fiICAYGze7IDZs5ciNjYe3bt3QNu2zdGrl80v/gsUXPu3H8GSDfPwzPsFnj55jsGj+kNTSwNnjpwHACz9Zz6iwqOxcfk2AMDBncew+/QWDB37B25fv4dO1u1Q27wGlszIeu+SPibD/Z4n7OwnIjUlFeFvItCwSX10+70zVi/MWr65TPnS6NK7Pe643EdCXAKq1qyCGYsn4/H9J3jlFyjMH6KQGNq/F/5atga1a1RFnVrVceDYGXxKSYV11/YAgDlLVsPY0ABTx2X9N96/V1ccPnkOK9Zvw8C+PRD65h12Oh3FoN97SI/596adaNXUEqVMTRAV8x6bdx2AsrISurRrKUgdC6ttm/fgn60r4fXkKTw9fDBm/DBoFdPE4QOnAACbtq1ERHgkli7K6l3esdUJZy/ux7iJNrh25RZ69emCevXrYNpkewBAXFw84r5YvjY9PR1RkTEIDAj+pXUrCjgHoQC4c+cOtm/fjsDAQJw4cQKlS5fG/v37UbFiRTRr1uy/D1BIuJxzha6+LkZOt4G+kR5ePQvEtMGzEPfv5FeTUsaQZLtrY2higL1Xd0pfDxzXHwPH9YfnPS9M+t0OAFDDvDo2nVgnLWO7cDwA4OKxy1g2ddWvqFahdP3cTegZ6GL0jOEwMNLHy2cBmDxohnQisklpY5k7aL6Pn2H+hCUYO2sExs8ehdfBbzBj+F8I8v980TUwMcCUhROgb5g1fOLi8SvYvf7zRLHMjExMHTITE+aOwZp9DtAqpok3wW+xaLID7t14+OsqXwicOOEMQ0N92NvbwcTECN7ez9GjxxDpBL2yZUvJvD8PHnhg2DBbLFw4HYsXz0RAQAh+/30Unj9/KS3TrVt77Nz5eRjXgQNZa+gvXboOS5euQ0ZGBnr2HIalS2fj5ElHFC9eDIGBIRg50g5Xrnzusld0V866QM9AF+NnjoKhkT78n73C+D/spJ8d09ImMu+N9+OnmDN+ASbOGo1Jc8YgLPgNptjMRkC2OTezxthj8l/j4LB5IbR1tRH+JgKbVmzH8X2nAWT9yLFs0QiD/m2MRLyLwvULN7Fz3d5fWvfCqHO7loiLT8CmXQcQExuLGlUrY9uaJdIhRuGRUVDKNk+qpIkRtq9bhlUbtqP3sPEwNjTA4N97YsTg36VlIqNiMHPBSsQnJkJfVwf1zWrj4PZ10NfT/dXVK9TOnLoEAwN9zJprC2MTIzz19UP/3iMRHZ01/K5MmZIyvwncHz3B2JHTMWfeFPxlb4egwBAMGzgBL/xeCVUFKsJEkm+ZCZyPTp48iSFDhmDQoEHYv38/nj9/jkqVKmHTpk24ePEiLl68+N3HbFq6TT5ESnklXZL534VIED6xvMtUUFXXLSN0CPQVj58eEDoEykWpykXrmUpFSXRCwV38YUj5/JvsvT/0VL4dO68IPhZn6dKl2LZtG3bu3CkzvrFp06bw9Pz28cNERERERPTzBB9i5O/vjxYtWuTI19HRQXx8/K8PiIiIiIgUmmKvYVQAehBMTU0REJDzaX9ubm6oVEn+w0KIiIiIiPKLGJJ8S4WB4A2EUaNGYfLkyXj48CFEIhHevXuHgwcPYvr06Rg3bpzQ4RERERERKRRBhhj5+PigTp06UFJSwpw5cyAWi9G2bVskJyejRYsWUFdXx/Tp0zFp0iQhwiMiIiIiBaboD0oTpIFQv359hIeHw9jYGJUqVYK7uztmzJiBgIAAfPz4EbVq1ULx4sWFCI2IiIiISKEJ0kDQ1dVFcHAwjI2NERISArFYDDU1NdSqVUuIcIiIiIiIpPigNAH06dMHLVu2RMmSJSESiWBhYQFlZWW5ZYOCguTmExERERFR3hOkgbBjxw707t0bAQEBsLW1xahRo1CiRAkhQiEiIiIiklFYVhvKL4I9B6FTp04AAA8PD0yePJkNBCIiIiKiAkDwB6Xt2bNH6BCIiIiIiKS4ihEREREREUkp+iRlwR+URkREREREBQd7EIiIiIiIspFIFHuIEXsQiIiIiIhIij0IRERERETZKPoyp+xBICIiIiIiKfYgEBERERFlw1WMiIiIiIiI/sUeBCIiIiKibPigNCIiIiIikuIkZSIiIiIion+xB4GIiIiIKBs+KI2IiIiIiOhf7EEgIiIi+l97dx5XU/7/Afx1W7XvpTCFkgyyZDeWERlb2cZeWbJv02A0S5bhx2CimbEbNWMbawZD+CLUkCVFSipRiIposrTdz+8PHF3KxKhbej097uOhcz7nnPc5n+6n+76fz+ccokJ4m1MiIiIiIqLn2INARERERFRIZb/NKXsQiIiIiIhIwh4EIiIiIqJC+BwEIiIiIiKi59iDQERERERUSGV/DgITBCIiIiKiQjjEiIiIiIiI6Dn2IBARERERFVLZb3P6QSYIsVkpyg6BqEL6zuITZYdAxViTfUnZIdAbmNt0UXYIVIw7Z9cqOwSiCueDTBCIiIiIiN6VvJJPUuYcBCIiIiIikrAHgYiIiIiokMrdf8AeBCIiIiIiKoQJAhERERFRIXKIUnu9i+XLl8PGxgZVqlRBixYtcObMmWLLXr58GX379oWNjQ1kMhmWLVv21sdjgkBEREREVEh5ShC2bt0Kb29vzJo1CxEREXB0dISLiwvS0tKKLP/48WPUqlULCxcuRNWqVd/p/JkgEBERERGVU35+fvDy8sLw4cNRr149rFq1Ctra2li/fn2R5Zs1a4bFixdj4MCB0NTUfKdjMkEgIiIiIipECFFqr7eRm5uL8+fPw9nZWVqmoqICZ2dnnDp16n2ftoR3MSIiIiIiKiM5OTnIyclRWKapqVnkt/0ZGRkoKCiAhYWFwnILCwtcuXKl1GJkDwIRERERUSGlOQdhwYIFMDAwUHgtWLBA2aesgD0IRERERERlxMfHB97e3grLipsrYGpqClVVVdy9e1dh+d27d995AnJJsAeBiIiIiKgQUYr/NDU1oa+vr/AqLkHQ0NBA06ZNceTIEWmZXC7HkSNH0KpVq1I7f/YgEBERERGVU97e3vDw8ICTkxOaN2+OZcuW4dGjRxg+fDgAwN3dHdWqVZOGKeXm5iImJkb6/61btxAZGQldXV3Y2tqW6JhMEIiIiIiICnnbuw2VpgEDBiA9PR2+vr64c+cOGjVqhODgYGnicnJyMlRUXg4Kun37Nho3biz9vGTJEixZsgTt27dHSEhIiY7JBIGIiIiIqJB3feJxaZk4cSImTpxY5LpXP/Tb2Nj85wSHcxCIiIiIiEjCHgQiIiIiokLK0xAjZWAPAhERERERSdiDQERERERUSHmbg1DW2INAREREREQS9iAQERERERUi2INARERERET0DHsQiIiIiIgKkVfyuxgxQSAiIiIiKoRDjIiIiIiIiJ5jDwIRERERUSGVfYgRexCIiIiIiEjCHgQiIiIiokI4B0GJrl27pszDExERERHRK5SaINja2qJjx47YuHEjnj59qsxQiIiIiIgAPJuDUFqvikCpCUJERAQaNmwIb29vVK1aFWPGjMGZM2eUGRIRERERUaWm1AShUaNG8Pf3x+3bt7F+/Xqkpqaibdu2qF+/Pvz8/JCenq7M8IiIiIioEhKl+K8iKBd3MVJTU0OfPn2wfft2/PDDD0hISMC0adNQo0YNuLu7IzU1VdkhvrORXkMQGX0Mt9OjcfjoDjRp2vCN5V3duuL0+WDcTo9G6Ol9cO7SXmH9Vz6TcPp8MFLuROFa8jns2hOIpk6Or+2ns0sHHD66A7fSLuFa8jls2LLivZ7Xh4B1U7E0de+MCaHL8FVcADx3z4GVY61iyzYa2BHDtn8H74tr4H1xDQZv8lEor6Kmio4zB8Lr4EJMj/0Vk8/8gp5+Y6FrblgGZ1LxuY8cgNALBxB36yx2H9oExyb131i+W6/OOHL6T8TdOouDJ3eio3NbhfVLfvkeN+5dVHj9tm2lQpn6DR2wcedqXLwWisj4E1jg5wttHa33fm4V3ajRQxF1OQSpGZdx+FgJ2rXenyE84iBSMy4jLPwvdH61Xft6MsIjDuLm3YtISjmPoL2/KbRrNT6qhp+WL5Da0oiLRzHzmylQV1cvlfP70PyxPwRdvb6GU/+JGDx9IS5dTXpj+azsx5i/egs+HT4DTftNRM/xvjh57pK0/tzleEyctxydhn+Fhm5jcfR0ZCmfwYeLQ4zKgXPnzmH8+PGwtLSEn58fpk2bhsTERBw+fBi3b9+Gq6urskN8J737dMO8BV9j0cJf0LGtG6KjY7EjaD1MTY2LLN+8RWOsDViKTb/vQIe2rti/73/YuGUFHBzspDIJCdfx1Zdz0bZlD3TrMhApybewc3cATArts2cvF6xasxibNu5Eu9Y98Vnngdi5bW+pn29FwrqpWBx6tITzt0Nw0n8Xfu3xLdJikzFww0xom+gXWd66lQNi9pzCpoHz8VvvWci6fQ+DNsyEnoURAEBdSwNV69sg9Kcg/Nr9W+wYswwmtSzx+a9fluVpVUg93Fzw7ffT4b94FXp8OgCx0XHYsH2Vwu95YU2bOeLntT9g28YgdO/4OQ7tP4o1G/xRp66tQrmQ/4XCyaGj9JrkNUNaZ17VDJt2rcH1pBS4dRkK98/HoU7d2vjxl3mleq4VTe++z9q1Hxb8jA5tXREdfQU7dwfA1Kz4dm1dwFJs/G072rfphb/2HcbGP1bCod7Ldi0xPgkzvOegTYvu+KzLQCTfuIVdfwZK9V2nTm2oqMjwxeTv0KrZZ/hm5nwMHzkI383me+nfBIeew+L1OzB2YA9s9fsa9jbVMXbOz7j3IKvI8nl5+Rgz2x+30+7hxxmjsWf5bMwaPxTmJkZSmSdPc2Bfszq+HjOwrE6DPlAyIZSXyvj5+SEgIABxcXHo1q0bRo0ahW7dukFF5WXecvPmTdjY2CA/P7/E+zXWs/v3QmXg8NEdiIi4iK+mzQUAyGQyXLpyAmtXb4C/35rXyv8auAzaOtoY1H+0tOzQ0e24dDEWX071LfIYenq6uHH7Atx6uOPE8VNQVVVF1OUQLPw/f2z8fUfpnNgHgHVTtC+Nmys7hCJ57p6D1IvXcND3t2cLZDJMPv0TzgYewqmV/55gyVRk+PLiWhz0DcSlXaFFlrFsWAsj9n6Pn1tNRtbte+8z/PdiTfalfy9UBnYf2oSLF6Lh+9UCAM/eO6cvHULg2i1Y6b/+tfK/rFsEbW0tjBg8SVoWdHAjYi5dwTfTnn3AX/LL99A30MPoYVOLPOYg97740mcimtX7FC/+ZNk72OFQ6E60c+qOG0kp7/ks315WzmNlh4DDx3bgQsQlzPhyDoBndRMddxJrV23AMr/Vr5X/9Td/6GhrYaBCu7YD0Zdi4D2l+HYtOTUSrj2G4UTIqSLLTJoyCiNGDUbjBp++h7P67+6cXavsEIo0ePpC1LezxtejBwEA5HI5uozywaDuHTGyb9fXym8LPoHAoEP4c/kcqKup/uv+G7qNxbKZY/Fpy0bvO/T3RtOho7JDKFYt08altu9rGRdKbd/vi1J7EFauXInBgwfjxo0b2L17N3r06KGQHACAubk5fv31VyVF+O7U1dXh2PhjHA/5W1omhMDxkL/RrHnRv3TNmjfG8WN/Kyw7+r+TaNa8UbHH8Bg+AA8fZCE6+goAwLHRx7CqVhVyuUBI6J+IiQ/Dtp3rFL7pruxYNxWLiroqLBvURFJo9MuFQiApNBrVm5Ts2qlraUJFXRVPHjwqtoymnhaEXI6nWcr/oFdeqauroYGjA0KPn5aWCSEQejwcTZq9PpwOAJo0c0To8XCFZSeO/v1a+ZZtnHD+SgiOhu/BvCXfwtDIQFqnqamBvLw8FP4+68Wd75q1LL0/4hWJuro6GjWuj5BjYdIyIQSOHyu+XWvevDFCXm3XjpwstrxCu3bpSrGx6BvoITPz4TucReWRl5eP2MRktGzoIC1TUVFBC0cHRMUVfQv4kDNRcKxbC/+3egs6eExH78lzsXb7ARQUyMsqbKpElJogxMfHw8fHB5aWlsWW0dDQgIeHR7Hrc3JykJWVpfBSYqeIxMTECGpqakhPy1BYnp52DxbmZkVuY25hirRXyqelZcDcQrF8l64dkZwaidSMaIyd4Ik+rp64fy8TAGBjUwPAs/HwPy5agUH9R+PBgyzsObBR4Q9uZca6qVi0jfSgoqaKRxmKHzgeZWRBx6xk1+1Tn4HIvpuJpLDoIteraqrjU59BuLznFHKzn/znmD9URs/fOxlpij0sGWn3YGZuWuQ2ZuamyEh/pXy6YvnjR8LgPf5bDO7thYVzlqJl66b4bdsK6QujsBNnYGZugjETPaGurgZ9Az3M9J0KAK+9Byurl+2a4rVOT8uAuUXRdWNuYYr09FfbwdfbNZeuHZFyJwp37l3GuInD0buXh9SuvapmLWuMHuOOwPV//Iez+fBl/pONArkcJoaKwyRNDPSQkVn0EKObdzNw+O8IyOVyrPhuIsZ83g2/7/kf1mzfXxYhVzpCyEvtVREofQ7CgwcP8OOPP2LUqFEYNWoUli5diocPS/7Nw4IFC2BgYKDwepp7vxQjVr7QE6fRvk0vdHUegKP/O4n1v/lLY+dlz/+g+i1Zib17DiIq8jImjpsJIQBXt8+UGXalwLopf1qN64l6PVthx+ilKMjJe229ipoq+iyfBJkMOPBNgBIipL1BwfhfcAjiYuNxaP8xDB80EY2aNECrts0AAPFxifhywncYNd4dV26ewbnYY0i5cQtpdzMgl1eMP7YV2ckTp9GudS+4dPocRw6fRMDvPxU5r8HS0gI7gtZjd9AB/B64VQmRftiEEDA20IPv+KGoZ2uNrm2d4NXvM2w/eELZodEHSKkJwrlz51C7dm0sXboU9+/fx/379+Hn54fatWsjIiKiRPvw8fHBw4cPFV5VNIqekFWW7t3LRH5+/mvfqpmZm+BuWtG3b027mwHzV8qbm5si7a5i+cePnyDpWjLOnY3E5AlfIz+/AEM9+gMA7t5JAwBcuZIglc/NzcWNpBRUr1F8T01lwrqpWB5n/gN5fgF0TBV7C3RM9fEo/c1fJrQY3Q2tx/XElqELkXbl9XHqL5IDg2qm2DxkIXsP/kXm8/eOqbmJwnJTc5PXeuReSE/LgKnZK+XNii8PACk3buFexn1Y16whLftz5340q/cpWtR3RiO7T7B00UqYmBoh+frN/3BGH46X7ZritTYzN0Xa3aKvddrdDJiZvdoOFteu3XjervkgP78Aw9w/VyhTtao59uzfiDPhEZg66Zv3cEYfNiM9XaiqqLw2Ifnew39galT0zRdMjQxgbWUBVdWXH91qVq+KjMws5OWVfJ4mlYwcotReFYFSE4QvvvgCvXr1wvXr17Fr1y7s2rULSUlJ6NGjB6ZOnVqifWhqakJfX1/hJZPJSjfwEsjLy0PUhcto176VtEwmk6F9+9Y4e6boySlnz1xAuw6tFJZ1+LQNzp6JfOOxVFRUoKmhAQCIiryMp09zYGdXU1qvpqaGGtbVcDPl9juezYeFdVOxyPMKkHopCTZtPn65UCaDTZv6uBkRX+x2Lcf0QNtJvbHFYxFSL71+68AXyYFRzarYPGQBnjzILo3wPyh5efm4FBWLNu1aSMtkMhnatGuBiLNRRW4TcTZKoTwAfNKhZbHlAaCqlQWMjA2L/GCbkX4fjx89QU83F+Q8zUVoyOki9lD55OXlIfJCNNp3aC0tk8lkaNeh+HbtzJkLCuUBoGPHNsWWf0FFRQUamhrSz5aWFth7YBOiIqMxYexX5WKYb3mnrq4Gh9ofIfziy7kccrkc4RevwNG+6Fs4N6pbGympaQq9Zjdu34WZkQHU1dVKPWaqXJT6G3Xu3DmsXbsWamovw1BTU8OMGTPg5OSkxMjejxW/rMfy1YsQeSEaEecvYux4T2hra2Hzhp3P1q9ehNTUu/h+9o8AgNUrf8PeA5swYdIIHDoYgj59u6NR4/r4YtK3AABtbS14Tx+H4P1HcedOGkxMjDBq9FBYWlngz6ADAIB//slG4K9bMPPrKbh18w5SUm5h0pRRAIDdz8sQ66aiCV93AL1+HIPUi0m4HZWI5iO6Ql1bExe3HwcA9PQbi3/uZCJk0bNhDa3G9kA7737YPWU5Ht5Ml+Yq5D56irzHOVBRU0XflVNQtb4Nto5YApmqilTmyYNsyPMKlHOiFcC6Fb/jx+XzcDEyBlERlzBizFBoa2th++bdAAC/FfNxJ/UuFn3/EwAgYPUmbN27Hl7j3XH08An07P0ZGjT6GDO/eHYHMW0dLUydPg4H9v0P6XczYF2zBnxmfYHr15Jx4ujLCbceowbi/JkoPHr0GJ90aImvZ3tj4ff+yMr6p8yvQXm14pf1WLF6MS5EXELE+YsYN8ETOtpa2LTx2V3TVq5ZjNTbdzF39hIAwOoVgdgXvBkTJo3EoYPH0KdfDzRqUh9TJz/rAdDW1sKX08fjwP4juHsnDcZFtGsvkoOUlFv47uuFCreKfnXeFilyd3XGt/6BqGdrjQZ2Nti49yiePM2FW6dnSdvXywJgYWKIKcN6AwAGdG2HP/aH4Id12zCoe0ckp6Zh3Y5gDO7x8k5Aj588RXLqyx6gW2kZuHItBQZ6OrAs5na3VLTKnugqNUHQ19dHcnIy6tatq7A8JSUFenp6Sorq/QnatR8mpsbw+WYKzC3MEH0xFv37jET68wl71WtYKTww40z4BYwe4Y2vfb/At7O+xLXE6xg6aDxiY599S1pQUAC7OrUxcHBvmJgY4/79TFyIuITuLoMUhq34fvsD8gvysXLtYmhVqYLz56Lg1n0YHhZzb+XKiHVTscTuOw0dEz209+4HHTMD3I25gT/cf8CjjGfXzcDKBEL+sr6aDHWGmqY6+q2aqrCfE0t34uSyXdCraoQ6XZoCALyCFyiU2TBgHpJPx5buCVVg+3YfhImpEbxnjoeZuSliouPg/vk4ZKQ/m/v17E5dL7/hPH82CpNHz8S0byZh+reTcf1aMkYPm4Krz98XBQVy1P3YDn0H9oK+gR7u3knDyWOn8OOCX5Cb+3LOiGOTBvjiq/HQ1tFGYnwSfL78HkHb9pXtyZdzQTv3w9TUBF9/OxXmFma4dDEG/XqPkCYuV69hpVA3Z8IvwGuEN7757gt8N/t5uzZwHGJjCrVr9rUwcEihdu38JXTrMhBXnrd9HT5tg9q2Nqhta4OY+DCFeIx0FZ91QYq6tnVC5sN/sGLLXmRkZsG+ZnWsnDVJmrh8J/0+VAqNiKhqZoxVsyZj0frt6Df1e5gbG2JIj08xoo+LVOZywg2M/G6p9PPi9c+Sw14dW2LeFM+yObEPREUZClRalPochMmTJyMoKAhLlixB69bPMuawsDBMnz4dffv2xbJly95pv+XlOQhEFU15fQ4ClZ/nIFDRysNzEKho5fU5CFS+n4NQ3fjNT4j/L27eL/qOeuWJUnsQlixZAplMBnd3d+Tn50MIAQ0NDYwbNw4LFy5UZmhEREREVElxiJESaWhowN/fHwsWLEBiYiIAoHbt2tDW1lZmWERERERElVaZJwh9+vRBYGAg9PX10adPnzeW1dXVxccff4yxY8fCwIAPkiIiIiKi0idnD0LZMjAwkG5D+m8f+nNycrBq1SqEhYVhz549ZREeEREREVGlVuYJQkBAQJH/L05MTAyaNWtWmiEREREREUlEJb+LkVIflFYS9vb2+Pvvv5UdBhERERFRpVDuH72nqqoKR0dHZYdBRERERJUE72JERERERESSyv6gtHI/xIiIiIiIiMoOexCIiIiIiAqp7EOM2INAREREREQS9iAQERERERVS2R+Uxh4EIiIiIiKSsAeBiIiIiKgQzkEgIiIiIiJ6jj0IRERERESFVPbnIDBBICIiIiIqhEOMiIiIiIiInmMPAhERERFRIbzNKRERERER0XPsQSAiIiIiKkRU8knK7EEgIiIiIiIJexCIiIiIiArhHAQiIiIiIqLn2INARERERFQIn4NARERERET0HHsQiIiIiIgKqex3MWKCQERERERUCIcYERERERERPcceBCIiIiKiQtiDQERERERE9Bx7EIiIiIiICqnc/QfsQSAiIiIiokJkorIPsirncnJysGDBAvj4+EBTU1PZ4VAhrJvyjfVTfrFuyi/WTfnG+qGywgShnMvKyoKBgQEePnwIfX19ZYdDhbBuyjfWT/nFuim/WDflG+uHygqHGBERERERkYQJAhERERERSZggEBERERGRhAlCOaepqYlZs2ZxMlI5xLop31g/5Rfrpvxi3ZRvrB8qK5ykTEREREREEvYgEBERERGRhAkCERERERFJmCAQEREREZGECUI5IpPJsHv3bmWHQVTudOjQAVOnTlV2GFTGZs+ejUaNGik7jApPCIHRo0fD2NgYMpkMkZGRyg6JiuHp6Qk3Nzdlh0EENWUHQEREVJRp06Zh0qRJyg6jwgsODkZgYCBCQkJQq1YtmJqaKjskKoa/vz947xgqD5ggEJUjeXl5UFdXV3YYRO9Fbm4uNDQ03no7IQQKCgqgq6sLXV3dUoiscklMTISlpSVat25dasd417omRQYGBsoOgQgAhxj9Jzt27ECDBg2gpaUFExMTODs749GjRzh79iw6d+4MU1NTGBgYoH379oiIiFDYNj4+Hu3atUOVKlVQr149HD58WGH99evXIZPJsGvXLnTs2BHa2tpwdHTEqVOnFMqFhobik08+gZaWFmrUqIHJkyfj0aNH0voVK1bAzs4OVapUgYWFBfr16/ev8VcGwcHBaNu2LQwNDWFiYoIePXogMTERQMmv/dq1a1GjRg1oa2ujd+/e8PPzg6GhoUKZP//8E02aNEGVKlVQq1YtzJkzB/n5+dJ6mUyGlStXolevXtDR0cH8+fNL/dwruszMTLi7u8PIyAja2tr47LPPEB8fDwDIysqClpYWDhw4oLBNUFAQ9PT08PjxYwBASkoKPv/8cxgaGsLY2Biurq64fv16WZ9KuVRcu1DUMC83Nzd4enpKP9vY2OD777+Hu7s79PX1MXr0aOn99Mcff6B169aoUqUK6tevj+PHj0vbhYSEQCaT4cCBA2jatCk0NTURGhr62hCjkJAQNG/eHDo6OjA0NESbNm1w48YNaf2/vd8qI09PT0yaNAnJycmQyWSwsbGBXC7HggULULNmTWhpacHR0RE7duyQtikoKMDIkSOl9fb29vD3939tv25ubpg/fz6srKxgb29f1qf2QSo8xCgnJweTJ0+Gubk5qlSpgrZt2+Ls2bMAniXRtra2WLJkicL2kZGRkMlkSEhIKOvQ6UMj6J3cvn1bqKmpCT8/P5GUlCQuXrwoli9fLv755x9x5MgRsWHDBhEbGytiYmLEyJEjhYWFhcjKyhJCCFFQUCDq168vOnXqJCIjI8Xx48dF48aNBQARFBQkhBAiKSlJABB169YV+/btE3FxcaJfv37C2tpa5OXlCSGESEhIEDo6OmLp0qXi6tWrIiwsTDRu3Fh4enoKIYQ4e/asUFVVFZs3bxbXr18XERERwt/f/1/jrwx27Nghdu7cKeLj48WFCxdEz549RYMGDURBQUGJrn1oaKhQUVERixcvFnFxcWL58uXC2NhYGBgYSMc4ceKE0NfXF4GBgSIxMVEcOnRI2NjYiNmzZ0tlAAhzc3Oxfv16kZiYKG7cuFHWl6JCaN++vZgyZYoQQohevXoJBwcHceLECREZGSlcXFyEra2tyM3NFUII0a9fPzF06FCF7fv27Ssty83NFQ4ODmLEiBHi4sWLIiYmRgwePFjY29uLnJycMj2v8uZN7ULhOnjB1dVVeHh4SD9bW1sLfX19sWTJEpGQkCASEhKk91P16tXFjh07RExMjBg1apTQ09MTGRkZQgghjh07JgCIhg0bikOHDomEhARx7949MWvWLOHo6CiEECIvL08YGBiIadOmiYSEBBETEyMCAwOl90xJ3m+V0YMHD8TcuXNF9erVRWpqqkhLSxPz5s0TdevWFcHBwSIxMVEEBAQITU1NERISIoR49h7x9fUVZ8+eFdeuXRMbN24U2traYuvWrdJ+PTw8hK6urhg2bJiIjo4W0dHRyjrFD4qHh4dwdXUVQggxefJkYWVlJfbv3y8uX74sPDw8hJGRkbh3754QQoj58+eLevXqKWw/efJk0a5du7IOmz5ATBDe0fnz5wUAcf369X8tW1BQIPT09MTevXuFEEIcPHhQqKmpiVu3bkllDhw4UGSCsG7dOqnM5cuXBQARGxsrhBBi5MiRYvTo0QrHOnnypFBRURFPnjwRO3fuFPr6+lJi8q7xVwbp6ekCgLh06VKJrv2AAQNE9+7dFfYxZMgQhQShU6dO4v/+7/8UymzYsEFYWlpKPwMQU6dOLYUz+rC8+HB69epVAUCEhYVJ6zIyMoSWlpbYtm2bEEKIoKAgoaurKx49eiSEEOLhw4eiSpUq4sCBA0KIZ3Vgb28v5HK5tI+cnByhpaUlDh48WIZnVf68qV0oaYLg5uamUObF+2nhwoXSsry8PFG9enXxww8/CCFeJgi7d+9W2LZwgnDv3j0BQPoQ+6qSvN8qq6VLlwpra2shhBBPnz4V2tra4u+//1YoM3LkSDFo0KBi9zFhwgTRt29f6WcPDw9hYWFR6ZPq9+1FgpCdnS3U1dXFpk2bpHW5ubnCyspKLFq0SAghxK1bt4SqqqoIDw+X1puamorAwEClxE4fFg4xekeOjo7o1KkTGjRogP79+2Pt2rXIzMwEANy9exdeXl6ws7ODgYEB9PX1kZ2djeTkZABAbGwsatSoASsrK2l/rVq1KvI4DRs2lP5vaWkJAEhLSwMAREVFITAwUBqnq6urCxcXF8jlciQlJaFz586wtrZGrVq1MGzYMGzatEkaYvGm+CuD+Ph4DBo0CLVq1YK+vj5sbGwAQKoj4M3XPi4uDs2bN1fY56s/R0VFYe7cuQr14+XlhdTUVKkeAMDJyem9ntuHLDY2FmpqamjRooW0zMTEBPb29oiNjQUAdOvWDerq6tizZw8AYOfOndDX14ezszOAZ/WSkJAAPT09qV6MjY3x9OlTaZhZZfU+2oXifp8Lt3FqampwcnKS6uzftgUAY2NjeHp6wsXFBT179oS/vz9SU1Ol9SV9v1V2CQkJePz4MTp37qxwrX7//XeF3//ly5ejadOmMDMzg66uLtasWaPQPgJAgwYNOO+glCQmJiIvLw9t2rSRlqmrq6N58+bS+8bKygrdu3fH+vXrAQB79+5FTk4O+vfvr5SY6cPCBOEdqaqq4vDhwzhw4ADq1auHn3/+Gfb29khKSoKHhwciIyPh7++Pv//+G5GRkTAxMUFubu5bH6fwhFWZTAYAkMvlAIDs7GyMGTMGkZGR0isqKgrx8fGoXbs29PT0EBERgS1btsDS0hK+vr5wdHTEgwcP3hh/ZdCzZ0/cv38fa9euRXh4OMLDwwFAoY7edO1LIjs7G3PmzFGon0uXLiE+Ph5VqlSRyuno6PzX06FCNDQ00K9fP2zevBkAsHnzZgwYMABqas/uyZCdnY2mTZsq1EtkZCSuXr2KwYMHKzN0pXtTu6CiovLa3VXy8vJe28d/+X3+t20DAgJw6tQptG7dGlu3bkWdOnVw+vRpACV/v1V22dnZAIC//vpL4VrFxMRI8xD++OMPTJs2DSNHjsShQ4cQGRmJ4cOHv/Y3jG2X8o0aNQp//PEHnjx5goCAAAwYMADa2trKDos+ALyL0X8gk8nQpk0btGnTBr6+vrC2tkZQUBDCwsKwYsUKdOvWDcCzCZEZGRnSdg4ODkhJSUFqaqr0zfSLP3Jvo0mTJoiJiYGtrW2xZdTU1ODs7AxnZ2fMmjULhoaGOHr0KPr06VNs/N7e3m8dS0Vy7949xMXFYe3atfjkk08APJvs/Tbs7e2lyWIvvPpzkyZNEBcX98b6obfj4OCA/Px8hIeHS3dkeVGf9erVk8oNGTIEnTt3xuXLl3H06FHMmzdPWtekSRNs3boV5ubm0NfXL/NzKO+KaxfMzMwUvrEvKChAdHQ0OnbsWKL9nj59Gu3atQMA5Ofn4/z585g4ceJbx9e4cWM0btwYPj4+aNWqFTZv3oyWLVvy/VZC9erVg6amJpKTk9G+ffsiy4SFhaF169YYP368tKyy966Vtdq1a0NDQwNhYWGwtrYG8CwhP3v2rMLNArp16wYdHR2sXLkSwcHBOHHihJIipg8NE4R3FB4ejiNHjqBLly4wNzdHeHg40tPT4eDgADs7O2zYsAFOTk7IysrC9OnToaWlJW3r7OyMOnXqwMPDA4sXL0ZWVha++eabt47hq6++QsuWLTFx4kSMGjUKOjo6iImJweHDh/HLL79g3759uHbtGtq1awcjIyPs378fcrkc9vb2b4z/Q2dkZAQTExOsWbMGlpaWSE5OxsyZM99qH5MmTUK7du3g5+eHnj174ujRozhw4IDU0wAAvr6+6NGjBz766CP069cPKioqiIqKQnR0tMIHVio5Ozs7uLq6wsvLC6tXr4aenh5mzpyJatWqwdXVVSrXrl07VK1aFUOGDEHNmjUVhiQNGTIEixcvhqurK+bOnYvq1avjxo0b2LVrF2bMmIHq1asr49TKhTe1Czo6OvD29sZff/2F2rVrw8/PDw8ePCjxvpcvXw47Ozs4ODhg6dKlyMzMxIgRI0q8fVJSEtasWYNevXrBysoKcXFxiI+Ph7u7OwC+30pKT08P06ZNwxdffAG5XI62bdvi4cOHCAsLg76+Pjw8PGBnZ4fff/8dBw8eRM2aNbFhwwacPXsWNWvWVHb4lYaOjg7GjRuH6dOnw9jYGB999BEWLVqEx48fY+TIkVI5VVVVeHp6wsfHB3Z2dsUOVyZ6a8qeBFFRxcTECBcXF2FmZiY0NTVFnTp1xM8//yyEECIiIkI4OTmJKlWqCDs7O7F9+3ZhbW0tli5dKm0fFxcn2rZtKzQ0NESdOnVEcHBwkZOUL1y4IG2TmZkpAIhjx45Jy86cOSM6d+4sdHV1hY6OjmjYsKGYP3++EOLZhOX27dsLIyMjoaWlJRo2bCjdheJN8VcGhw8fFg4ODkJTU1M0bNhQhISESNe/pNd+zZo1olq1akJLS0u4ubmJefPmiapVqyocJzg4WLRu3VpoaWkJfX190bx5c7FmzRppfeE6p+IVniB7//59MWzYMGFgYCC0tLSEi4uLuHr16mvbzJgxQwAQvr6+r61LTU0V7u7uwtTUVGhqaopatWoJLy8v8fDhw9I+lXLtTe1Cbm6uGDdunDA2Nhbm5uZiwYIFRU5SLtzOCfGyLdu8ebNo3ry50NDQEPXq1RNHjx6VyryYpJyZmamwbeFJynfu3BFubm7C0tJSaGhoCGtra+Hr6ysKCgqk8v/2fqusCk9SFkIIuVwuli1bJuzt7YW6urowMzMTLi4u4vjx40KIZxOZPT09hYGBgTA0NBTjxo0TM2fOlOpCCMW77dD7U/i6PnnyREyaNElqp9q0aSPOnDnz2jaJiYkCgDR5meh9kAnBR/YRvQ9eXl64cuUKTp48qexQiMqN69evo2bNmrhw4YLCMw2I6HWDBg2CqqoqNm7cWOJtTp48iU6dOiElJQUWFhalGB1VJpykTPSOlixZIt0R5+eff8Zvv/0GDw8PZYdFREQVTH5+PmJiYnDq1Cl8/PHHJdomJycHN2/exOzZs9G/f38mB/ReMUEgekdnzpxB586d0aBBA6xatQo//fQTRo0apeywiIiogomOjoaTkxM+/vhjjB07tkTbbNmyBdbW1njw4AEWLVpUyhFSZcMhRkREREREJGEPAhERERERSZggEBERERGRhAkCERERERFJmCAQEREREZGECQIRUTnh6ekJNzc36ecOHTpg6tSp/2mf72MfRERUuTBBICL6F56enpDJZJDJZNDQ0ICtrS3mzp2L/Pz8Uj3url278P3335eobEhICGQyGR48ePDO+yAiIgIANWUHQERUEXTt2hUBAQHIycnB/v37MWHCBKirq8PHx0ehXG5uLjQ0NN7LMY2NjcvFPoiIqHJhDwIRUQloamqiatWqsLa2xrhx4+Ds7Iw9e/ZIw4Lmz58PKysr2NvbAwBSUlLw+eefw9DQEMbGxnB1dcX169el/RUUFMDb2xuGhoYwMTHBjBkz8OpjaV4dHpSTk4OvvvoKNWrUgKamJmxtbfHrr7/i+vXr6NixIwDAyMgIMpkMnp6eRe4jMzMT7u7uMDIygra2Nj777DPEx8dL6wMDA2FoaIiDBw/CwcEBurq66Nq1K1JTU9/vBSUionKLCQIR0TvQ0tJCbm4uAODIkSOIi4vD4cOHsW/fPuTl5cHFxQV6eno4efIkwsLCpA/aL7b58ccfERgYiPXr1yM0NBT3799HUFDQG4/p7u6OLVu24KeffkJsbCxWr14NXV1d1KhRAzt37gQAxMXFITU1Ff7+/kXuw9PTE+fOncOePXtw6tQpCCHQrVs35OXlSWUeP36MJUuWYMOGDThx4gSSk5Mxbdq093HZiIioAuAQIyKityCEwJEjR3Dw4EFMmjQJ6enp0NHRwbp166ShRRs3boRcLse6desgk8kAAAEBATA0NERISAi6dOmCZcuWwcfHB3369AEArFq1CgcPHiz2uFevXsW2bdtw+PBhODs7AwBq1aolrX8xlMjc3ByGhoZF7iM+Ph579uxBWFgYWrduDQDYtGkTatSogd27d6N///4AgLy8PKxatQq1a9cGAEycOBFz585910tGREQVDBMEIqIS2LdvH3R1dZGXlwe5XI7Bgwdj9uzZmDBhAho0aKAw7yAqKgoJCQnQ09NT2MfTp0+RmJiIhw8fIjU1FS1atJDWqampwcnJ6bVhRi9ERkZCVVUV7du3f+dziI2NhZqamsJxTUxMYG9vj9jYWGmZtra2lBwAgKWlJdLS0t75uEREVLEwQSAiKoGOHTti5cqV0NDQgJWVFdTUXjafOjo6CmWzs7PRtGlTbNq06bX9mJmZvdPxtbS03mm7d6Gurq7ws0wmKzZxISKiDw/nIBARlYCOjg5sbW3x0UcfKSQHRWnSpAni4+Nhbm4OW1tbhZeBgQEMDAxgaWmJ8PBwaZv8/HycP3++2H02aNAAcrkcx48fL3L9ix6MgoKCYvfh4OCA/Px8hePeu3cPcXFxqFev3hvPiYiIKg8mCERE79mQIUNgamoKV1dXnDx5EklJSQgJCcHkyZNx8+ZNAMCUKVOwcOFC7N69G1euXMH48eNfe4ZBYTY2NvDw8MCIESOwe/duaZ/btm0DAFhbW0Mmk2Hfvn1IT09Hdnb2a/uws7ODq6srvLy8EBoaiqioKAwdOhTVqlWDq6trqVwLIiKqeJggEBG9Z9ra2jhx4gQ++ugj9OnTBw4ODhg5ciSePn0KfX19AMCXX36JYcOGwcPDA61atYKenh569+79xv2uXLkS/fr1w/jx41G3bl14eXnh0aNHAIBq1aphzpw5mDlzJiwsLDBx4sQi9xEQEICmTZuiR48eaNWqFYQQ2L9//2vDioiIqPKSCQ4sJSIiIiKi59iDQEREREREEiYIREREREQkYYJAREREREQSJghERERERCRhgkBERERERBImCEREREREJGGCQEREREREEiYIREREREQkYYJAREREREQSJghERERERCRhgkBERERERBImCEREREREJPl/yExiUY73xm0AAAAASUVORK5CYII=\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "conf_mat = confusion_matrix(\n",
        "    y_test,\n",
        "    np.argmax(model.predict(x_test), axis=1).flatten())\n",
        "\n",
        "conf_mat /= np.sum(conf_mat, axis=1).reshape(-1,1)\n",
        "\n",
        "plt.figure(0,figsize=(10,6))\n",
        "sns.heatmap(conf_mat, xticklabels=catDict.keys(), yticklabels=catDict.keys(), annot=True)\n",
        "plt.ylabel(\"Label\")\n",
        "plt.xlabel(\"Prediction\");"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:54:02.924279Z",
          "start_time": "2021-05-08T09:54:02.920777Z"
        },
        "id": "PbhT3twdGDwQ"
      },
      "outputs": [],
      "source": [
        "def emoji_tester(model, text):\n",
        "    catDictInverse = {0:\"😓\",1:\"👿\",2:\"🥰\",3:\"😲\",4:\"😱\",5:\"😃\"}\n",
        "    net_input = query_all_vectors([text.lower().split(\" \")], pad_to_length=70)\n",
        "    net_output = model.predict(net_input)\n",
        "    emoji = catDictInverse[np.argmax(net_output)]\n",
        "    print(emoji)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "ExecuteTime": {
          "end_time": "2021-05-08T09:55:36.199882Z",
          "start_time": "2021-05-08T09:55:36.162981Z"
        },
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HJZ8tu9zGDwQ",
        "outputId": "9078ae0a-1230-42af-f07d-f9d64348ab24"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 64ms/step\n",
            "😃\n"
          ]
        }
      ],
      "source": [
        "emoji_tester(model, \"I love NLP\")"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "for idx, label in enumerate(test_label):\n",
        "  if label == \"love\":\n",
        "    print(idx)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ypUJl0VBiKjY",
        "outputId": "5c310c59-6564-4e6b-fd06-323f013eb3e7"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "14\n",
            "71\n",
            "74\n",
            "79\n",
            "96\n",
            "111\n",
            "125\n",
            "131\n",
            "149\n",
            "151\n",
            "166\n",
            "167\n",
            "168\n",
            "170\n",
            "183\n",
            "201\n",
            "202\n",
            "206\n",
            "228\n",
            "241\n",
            "254\n",
            "265\n",
            "286\n",
            "292\n",
            "294\n",
            "302\n",
            "303\n",
            "307\n",
            "316\n",
            "326\n",
            "333\n",
            "341\n",
            "355\n",
            "356\n",
            "366\n",
            "369\n",
            "391\n",
            "404\n",
            "466\n",
            "513\n",
            "520\n",
            "529\n",
            "535\n",
            "574\n",
            "575\n",
            "593\n",
            "595\n",
            "608\n",
            "614\n",
            "615\n",
            "621\n",
            "627\n",
            "629\n",
            "630\n",
            "643\n",
            "650\n",
            "670\n",
            "713\n",
            "715\n",
            "735\n",
            "746\n",
            "765\n",
            "766\n",
            "802\n",
            "821\n",
            "833\n",
            "869\n",
            "886\n",
            "888\n",
            "890\n",
            "894\n",
            "895\n",
            "909\n",
            "923\n",
            "936\n",
            "957\n",
            "959\n",
            "960\n",
            "978\n",
            "982\n",
            "985\n",
            "992\n",
            "994\n",
            "1013\n",
            "1015\n",
            "1017\n",
            "1059\n",
            "1061\n",
            "1064\n",
            "1072\n",
            "1081\n",
            "1083\n",
            "1092\n",
            "1111\n",
            "1117\n",
            "1123\n",
            "1146\n",
            "1178\n",
            "1194\n",
            "1207\n",
            "1216\n",
            "1245\n",
            "1301\n",
            "1333\n",
            "1340\n",
            "1346\n",
            "1347\n",
            "1358\n",
            "1370\n",
            "1377\n",
            "1381\n",
            "1386\n",
            "1387\n",
            "1395\n",
            "1402\n",
            "1436\n",
            "1458\n",
            "1463\n",
            "1476\n",
            "1478\n",
            "1488\n",
            "1497\n",
            "1513\n",
            "1524\n",
            "1541\n",
            "1545\n",
            "1571\n",
            "1588\n",
            "1592\n",
            "1642\n",
            "1650\n",
            "1680\n",
            "1684\n",
            "1685\n",
            "1701\n",
            "1709\n",
            "1714\n",
            "1731\n",
            "1732\n",
            "1739\n",
            "1745\n",
            "1756\n",
            "1765\n",
            "1818\n",
            "1833\n",
            "1836\n",
            "1844\n",
            "1849\n",
            "1851\n",
            "1889\n",
            "1899\n",
            "1902\n",
            "1936\n",
            "1945\n",
            "1961\n",
            "1969\n",
            "1970\n",
            "1990\n",
            "1991\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "test_data[821]"
      ],
      "metadata": {
        "id": "TuwtzXZN3gMX",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "a23ad791-b751-4ecb-b1ac-99d570c4a3ec"
      },
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "['i',\n",
              " 'don',\n",
              " 't',\n",
              " 'know',\n",
              " 'how',\n",
              " 'i',\n",
              " 'feel',\n",
              " 'i',\n",
              " 'guess',\n",
              " 'it',\n",
              " 's',\n",
              " 'one',\n",
              " 'of',\n",
              " 'those',\n",
              " 'moments',\n",
              " 'where',\n",
              " 'you',\n",
              " 'want',\n",
              " 'to',\n",
              " 'feel',\n",
              " 'like',\n",
              " 'you',\n",
              " 're',\n",
              " 'accepted',\n",
              " 'even',\n",
              " 'though',\n",
              " 'whatever',\n",
              " 'you',\n",
              " 'did',\n",
              " 'or',\n",
              " 'did',\n",
              " 'not',\n",
              " 'get',\n",
              " 'mattered',\n",
              " 'to',\n",
              " 'you',\n",
              " 'the',\n",
              " 'most']"
            ]
          },
          "metadata": {},
          "execution_count": 70
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "5nGY9pyZ3rNJ"
      },
      "execution_count": null,
      "outputs": []
    }
  ],
  "metadata": {
    "language_info": {
      "name": "python"
    },
    "orig_nbformat": 4,
    "colab": {
      "provenance": []
    },
    "accelerator": "GPU",
    "gpuClass": "standard",
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}