{
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {},
      "outputs": [],
      "source": [
        "import os\n",
        "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '3'  # 设置日志级别为ERROR，以减少警告信息\n",
        "# 禁用 Gemini 的底层库（gRPC 和 Abseil）在初始化日志警告\n",
        "os.environ[\"GRPC_VERBOSITY\"] = \"ERROR\"\n",
        "os.environ[\"GLOG_minloglevel\"] = \"3\"  # 0: INFO, 1: WARNING, 2: ERROR, 3: FATAL\n",
        "os.environ[\"GLOG_minloglevel\"] = \"true\"\n",
        "import logging\n",
        "import tensorflow as tf\n",
        "tf.get_logger().setLevel(logging.ERROR)\n",
        "tf.compat.v1.logging.set_verbosity(tf.compat.v1.logging.ERROR)\n",
        "!export TF_FORCE_GPU_ALLOW_GROWTH=true"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 2,
      "metadata": {
        "cellView": "form",
        "id": "_ckMIh7O7s6D"
      },
      "outputs": [],
      "source": [
        "# Copyright 2018 The TensorFlow Authors.\n",
        "#@title Licensed under the Apache License, Version 2.0 (the \"License\");\n",
        "# you may not use this file except in compliance with the License.\n",
        "# You may obtain a copy of the License at\n",
        "#\n",
        "# https://www.apache.org/licenses/LICENSE-2.0\n",
        "#\n",
        "# Unless required by applicable law or agreed to in writing, software\n",
        "# distributed under the License is distributed on an \"AS IS\" BASIS,\n",
        "# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.\n",
        "# See the License for the specific language governing permissions and\n",
        "# limitations under the License."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 3,
      "metadata": {
        "cellView": "form",
        "id": "vasWnqRgy1H4"
      },
      "outputs": [],
      "source": [
        "#@title MIT License\n",
        "#\n",
        "# Copyright (c) 2017 François Chollet\n",
        "#\n",
        "# Permission is hereby granted, free of charge, to any person obtaining a\n",
        "# copy of this software and associated documentation files (the \"Software\"),\n",
        "# to deal in the Software without restriction, including without limitation\n",
        "# the rights to use, copy, modify, merge, publish, distribute, sublicense,\n",
        "# and/or sell copies of the Software, and to permit persons to whom the\n",
        "# Software is furnished to do so, subject to the following conditions:\n",
        "#\n",
        "# The above copyright notice and this permission notice shall be included in\n",
        "# all copies or substantial portions of the Software.\n",
        "#\n",
        "# THE SOFTWARE IS PROVIDED \"AS IS\", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR\n",
        "# IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,\n",
        "# FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL\n",
        "# THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER\n",
        "# LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING\n",
        "# FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER\n",
        "# DEALINGS IN THE SOFTWARE."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jYysdyb-CaWM"
      },
      "source": [
        "# 基本分类：对服装图像进行分类"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "S5Uhzt6vVIB2"
      },
      "source": [
        "<table class=\"tfo-notebook-buttons\" align=\"left\">\n",
        "  <td><a target=\"_blank\" href=\"https://tensorflow.google.cn/tutorials/keras/classification\"><img src=\"https://tensorflow.google.cn/images/tf_logo_32px.png\">在 TensorFlow.org 上查看</a></td>\n",
        "  <td><a target=\"_blank\" href=\"https://colab.research.google.com/github/tensorflow/docs-l10n/blob/master/site/zh-cn/tutorials/keras/classification.ipynb\"><img src=\"https://tensorflow.google.cn/images/colab_logo_32px.png\">在 Google Colab 中运行</a></td>\n",
        "  <td><a target=\"_blank\" href=\"https://github.com/tensorflow/docs-l10n/blob/master/site/zh-cn/tutorials/keras/classification.ipynb\"><img src=\"https://tensorflow.google.cn/images/GitHub-Mark-32px.png\">在 GitHub 上查看源代码</a></td>\n",
        "  <td><a href=\"https://storage.googleapis.com/tensorflow_docs/docs-l10n/site/zh-cn/tutorials/keras/classification.ipynb\"><img src=\"https://tensorflow.google.cn/images/download_logo_32px.png\">下载笔记本</a></td>\n",
        "</table>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FbVhjPpzn6BM"
      },
      "source": [
        "本指南将训练一个神经网络模型，对运动鞋和衬衫等服装图像进行分类。即使您不理解所有细节也没关系；这只是对完整 TensorFlow 程序的快速概述，详细内容会在您实际操作的同时进行介绍。\n",
        "\n",
        "本指南使用了 [tf.keras](https://tensorflow.google.cn/guide/keras)，它是 TensorFlow 中用来构建和训练模型的高级 API。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 4,
      "metadata": {
        "id": "dzLKpmZICaWN"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "2.17.0\n"
          ]
        }
      ],
      "source": [
        "# TensorFlow and tf.keras\n",
        "import tensorflow as tf\n",
        "\n",
        "# Helper libraries\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "print(tf.__version__)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yR0EdgrLCaWR"
      },
      "source": [
        "## 导入 Fashion MNIST 数据集"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DLdCchMdCaWQ"
      },
      "source": [
        "本指南使用 [Fashion MNIST](https://github.com/zalandoresearch/fashion-mnist) 数据集，该数据集包含 10 个类别的 70,000 个灰度图像。这些图像以低分辨率（28x28 像素）展示了单件衣物，如下所示：\n",
        "\n",
        "<table>\n",
        "  <tr><td>     <img alt=\"Fashion MNIST sprite\" src=\"https://tensorflow.google.cn/images/fashion-mnist-sprite.png\"> </td></tr>\n",
        "  <tr><td align=\"center\">     <b>图 1.</b>  <a href=\"https://github.com/zalandoresearch/fashion-mnist\">Fashion-MNIST 样本</a>（由 Zalando 提供，MIT 许可）。<br> </td></tr>\n",
        "</table>\n",
        "\n",
        "Fashion MNIST 旨在临时替代经典 [MNIST](http://yann.lecun.com/exdb/mnist/) 数据集，后者常被用作计算机视觉机器学习程序的“Hello, World”。MNIST 数据集包含手写数字（0、1、2 等）的图像，其格式与您将使用的衣物图像的格式相同。\n",
        "\n",
        "本指南使用 Fashion MNIST 来实现多样化，因为它比常规 MNIST 更具挑战性。这两个数据集都相对较小，都用于验证某个算法是否按预期工作。对于代码的测试和调试，它们都是很好的起点。\n",
        "\n",
        "在本指南中，我们使用 60,000 张图像来训练网络，使用 10,000 张图像来评估网络学习对图像进行分类的准确程度。您可以直接从 TensorFlow 中访问 Fashion MNIST。直接从 TensorFlow 中导入和加载 Fashion MNIST 数据："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 5,
      "metadata": {
        "id": "7MqDQO0KCaWS"
      },
      "outputs": [],
      "source": [
        "fashion_mnist = tf.keras.datasets.fashion_mnist\n",
        "\n",
        "(train_images, train_labels), (test_images, test_labels) = fashion_mnist.load_data()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "t9FDsUlxCaWW"
      },
      "source": [
        "加载数据集会返回四个 NumPy 数组：\n",
        "\n",
        "- `train_images` 和 `train_labels` 数组是*训练集*，即模型用于学习的数据。\n",
        "- *测试集*、`test_images` 和 `test_labels` 数组会被用来对模型进行测试。\n",
        "\n",
        "图像是 28x28 的 NumPy 数组，像素值介于 0 到 255 之间。*标签*是整数数组，介于 0 到 9 之间。这些标签对应于图像所代表的服装*类*：\n",
        "\n",
        "<table>\n",
        "  <tr>\n",
        "    <th>标签</th>\n",
        "    <th>类</th>\n",
        "  </tr>\n",
        "  <tr>\n",
        "    <td>0</td>\n",
        "    <td>T恤/上衣</td>\n",
        "  </tr>\n",
        "  <tr>\n",
        "    <td>1</td>\n",
        "    <td>裤子</td>\n",
        "  </tr>\n",
        "    <tr>\n",
        "    <td>2</td>\n",
        "    <td>套头衫</td>\n",
        "  </tr>\n",
        "    <tr>\n",
        "    <td>3</td>\n",
        "    <td>连衣裙</td>\n",
        "  </tr>\n",
        "    <tr>\n",
        "    <td>4</td>\n",
        "    <td>外套</td>\n",
        "  </tr>\n",
        "    <tr>\n",
        "    <td>5</td>\n",
        "    <td>凉鞋</td>\n",
        "  </tr>\n",
        "    <tr>\n",
        "    <td>6</td>\n",
        "    <td>衬衫</td>\n",
        "  </tr>\n",
        "    <tr>\n",
        "    <td>7</td>\n",
        "    <td>运动鞋</td>\n",
        "  </tr>\n",
        "    <tr>\n",
        "    <td>8</td>\n",
        "    <td>包</td>\n",
        "  </tr>\n",
        "    <tr>\n",
        "    <td>9</td>\n",
        "    <td>短靴</td>\n",
        "  </tr>\n",
        "</table>\n",
        "\n",
        "每个图像都会被映射到一个标签。由于数据集不包括*类名称*，请将它们存储在下方，供稍后绘制图像时使用："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 6,
      "metadata": {
        "id": "IjnLH5S2CaWx"
      },
      "outputs": [],
      "source": [
        "class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',\n",
        "               'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot']"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Brm0b_KACaWX"
      },
      "source": [
        "## 浏览数据\n",
        "\n",
        "在训练模型之前，我们先浏览一下数据集的格式。以下代码显示训练集中有 60,000 个图像，每个图像由 28 x 28 的像素表示："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 7,
      "metadata": {
        "id": "zW5k_xz1CaWX"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "(60000, 28, 28)"
            ]
          },
          "execution_count": 7,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "train_images.shape"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cIAcvQqMCaWf"
      },
      "source": [
        "同样，训练集中有 60,000 个标签："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 8,
      "metadata": {
        "id": "TRFYHB2mCaWb"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "60000"
            ]
          },
          "execution_count": 8,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "len(train_labels)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YSlYxFuRCaWk"
      },
      "source": [
        "每个标签都是一个 0 到 9 之间的整数："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 9,
      "metadata": {
        "id": "XKnCTHz4CaWg"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "array([9, 0, 0, ..., 3, 0, 5], dtype=uint8)"
            ]
          },
          "execution_count": 9,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "train_labels"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "TMPI88iZpO2T"
      },
      "source": [
        "测试集中有 10,000 个图像。同样，每个图像都由 28x28 个像素表示："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "2KFnYlcwCaWl"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "(10000, 28, 28)"
            ]
          },
          "execution_count": 10,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "test_images.shape"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rd0A0Iu0CaWq"
      },
      "source": [
        "测试集包含 10,000 个图像标签："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 11,
      "metadata": {
        "id": "iJmPr5-ACaWn"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "10000"
            ]
          },
          "execution_count": 11,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "len(test_labels)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ES6uQoLKCaWr"
      },
      "source": [
        "## 预处理数据\n",
        "\n",
        "在训练网络之前，必须对数据进行预处理。如果您检查训练集中的第一个图像，您会看到像素值处于 0 到 255 之间："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "m4VEw8Ud9Quh"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plt.figure()\n",
        "plt.imshow(train_images[0])\n",
        "plt.colorbar()\n",
        "plt.grid(False)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wz7l27Lz9S1P"
      },
      "source": [
        "将这些值缩小至 0 到 1 之间，然后将其馈送到神经网络模型。为此，请将这些值除以 255。请务必以相同的方式对*训练集*和*测试集*进行预处理："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "id": "bW5WzIPlCaWv"
      },
      "outputs": [],
      "source": [
        "train_images = train_images / 255.0\n",
        "\n",
        "test_images = test_images / 255.0"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ee638AlnCaWz"
      },
      "source": [
        "为了验证数据的格式是否正确，以及您是否已准备好构建和训练网络，让我们显示*训练集*中的前 25 个图像，并在每个图像下方显示类名称。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 14,
      "metadata": {
        "id": "oZTImqg_CaW1"
      },
      "outputs": [
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAxoAAAMpCAYAAACDrkVRAAAAP3RFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMS5wb3N0MSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8kixA/AAAACXBIWXMAAA9hAAAPYQGoP6dpAADqh0lEQVR4nOzdd5hV1d3+/88gfRq9DL1IUUBBBCl2USwIRI0hRvHRR6NGxZhYH3tJMbEmahIb0ViIAoqooCgoAiIWOgxFhiJD773t3x/+nK97rRtmMeypvF/XlevK+rDOPvucs87aZzv73isliqLIAAAAACBB5Yp7BwAAAACUPZxoAAAAAEgcJxoAAAAAEseJBgAAAIDEcaIBAAAAIHGcaAAAAABIHCcaAAAAABJXPqTTvn37bPny5Zaenm4pKSmFvU8oBaIoss2bN1tWVpaVK1e456uMP7iKcvyZMQYRx/hDceMYjOJ0MOMv6ERj+fLl1qhRo0R2DmXL0qVLrWHDhoX6HIw/7E9RjD8zxiA0xh+KG8dgFKeQ8Rd0opGenp63wYyMjEPfM5R6mzZtskaNGuWNjcJUFOMviqJYO8n/apOdne3Vfv/733u1/v37e7UOHTrE2hUrVvT6lC/vf43nzJnj1UaOHBlrN23a1OszaNAgr1atWjWvVtyKcvyZlb45cPXq1bH2q6++6vUZMGCAV6tbt26h7ZOZ2fTp02PtefPmeX369u3r1SpUqFBo+1QQjL//Jycnx6tNmDAh1n7vvfe8PtWrV/dqv/jFL7zaMcccE2urMTNixAivNm7cOK9WtWrVWPviiy/2+vzP//yPVyuJytox+HCQm5sba9evX7+Y9uTQHcz4CzrR+PFHV0ZGBoMMMUXxZ9SiGH+FeaKRlpbm1dTJQZUqVfJ9bOiJhntANfN/rFWqVMnro97fkvydL6o/45e2OXDHjh2xduXKlb0+6gBR2K/NHc9qnKp9KGknGj9i/Olx5M5l6vNTc1lqaqpXc1+vmk/VXHbEEUd4NXeuVHNuSXt/81NWjsGHgy1btsTaZeG9DBl/hMEBAAAAJI4TDQAAAACJC7p0Ciit3EuizPSf+kL+/Pftt996tSFDhni1oUOHxtrqT/jun1DNzO68806vtm7dunz3K1SrVq1i7WnTpnl9/vjHP3q1evXqebWzzjor1v7d737n9Wnfvv3B7iIKQI0l95r1l19+2evzxhtveLXatWt7NfcSF3UZjNqHnTt3erWlS5fG2v369fP6qO/LRRdd5NVQuD744AOv9vjjj3s1dfnRrl27Ym116Z7KdqiMxsqVK2NtlS1Tl4+q698zMzNj7bfeesvr88QTT3i1M844w6s99dRTXg1F77TTTvNq69evj7Vr1arl9Xnuuee8mhpbIZYvX+7VTj31VK+2ffv2WLtx48Zen9GjR3s1dUlhacJfNAAAAAAkjhMNAAAAAInjRAMAAABA4shooEwLvfXfpk2bYu3LLrvM66MyDSoD4t5+UV3DrO4hr65N37NnT6y9ceNGr4+6RajaVsh70aVLF6/m3irVzGzixImxtrpnfc+ePb3af/7zn3z3AQdH3e7TvRb9T3/6k9fn4Ycf9mpz5871au418ip7odZaUbc9da91P+ecc7w+Ku+Bwrdw4cJY+7XXXvP6qNyVe9252Q8rSf+UWjlYLQAXcrtPNY+p+S7kNskq29GtWzevtmzZMq/m5tIeffRRf2dR6NyxZma2Zs2aWPv777/3+qixrObSCy+8MNZWx7C9e/d6NZVLcufJzZs3e31Kex5D4S8aAAAAABLHiQYAAACAxHGiAQAAACBxnGgAAAAASBxh8J8IXdzNpQI9n3/+uVc7++yzC7QPKmikQmwFpZ7TFRqqLq369+8fay9ZssTrU7duXa+m3hf381JBRUV9zu5nU7NmzaDHKSGfs6LC7G7QTb0P48eP92pz5szxam3bti3QfmH/3MC2Cmv/5je/8Wp/+9vfvFqlSpUOuO39bf+4447zav/zP/8Ta6tF29SigSh8bpg59HNQYVz3BhJqDlTHsGbNmnk198YG6uYUav5R4zRkH3bv3u3V1EJuM2fOjLVHjhzp9TnvvPPy3Qccmho1ani1RYsWxdrquKkWw12xYoVXc+dEdVOY6dOnezV1wxd3bKn9Kov4iwYAAACAxHGiAQAAACBxnGgAAAAASBwnGgAAAAASRxj8J1SozQ2xLViwwOvz/PPPezUVoHVXfFQrR6qVmUOC3yroq16P6heyfTdwHBpALom+/vprr+aGv2vVquX1cVfp3h93pVy1KmnIarpm/mej3ne16q6ya9euWNtdJddMr+bcsGHDfPdLUfulviusqJs893N0V8o1M2vSpIlXU5+FO35Xr17t9VFhWfUdcvdDfacKetMCHJrLL7881n788ce9Piogrm6S4d4gRc01SsWKFb2aGm8utQp41apVg54zZB82bNjg1dx5keB38WjRooVX++KLL2JtdTMC9yYXodRcp258kpWV5dXc4/62bdsKtA+lDX/RAAAAAJA4TjQAAAAAJI4TDQAAAACJ40QDAAAAQOIIg/+ECtq6IaJPPvnE6/PRRx95tUaNGnk1d6VSFQT68MMPvdpVV13l1dwAnloZNXRF6i1btsTaKsTrButCt10SjR071qu5n41afVa9LyrA7YbMHnnkEa9P/fr1vZoaM8uXL8/3cWofVPjSDYO7n7uZ2TfffOPVnnrqKa/mhkLVarrq/Ro6dKhXIwyevJDv59q1a4O25Ya669Wr5/VRc5m6CYK7X2reUjUUPvdGJN26dfP6vPPOO16ta9euXs0N+avxoVZ0VkFsd65RN1FR21dzkrvK+KpVq7w+irp5x5/+9Kegx6JwtW3b1qu5x0Q1p7g35zHT40+t+u1SY1Ld1MIdk+omBmURf9EAAAAAkDhONAAAAAAkjhMNAAAAAIkjo/ET6vo815QpU7xaTk6OV1PXzbu1M8880+vz7bfferVbb73Vq3Xu3DnWbt++vddHXbv45ZdfejX3NXXv3t3r416vu2nTJq9PafHWW295Nffa8ZDF88z0tcHudcAqY6OyOGohwSuuuCLW/uc//+n1Ofroo72aypi4GaQ6dep4fX772996tWeeecarudeaqudT18DOnTvXq82bNy/WbtWqldcHB8e9Pjg0w6VyamqxssLcr9CFMVG4brzxRq/2xBNPeDW18KObq1BzgVpQL+SadTU+1EKCql/INfIbN270ameffbZXO1yury/pQhaUVfOam1k00xnIjh07xtrqc1f7oH5DuNzfCmUVf9EAAAAAkDhONAAAAAAkjhMNAAAAAInjRAMAAABA4g7bMLhaTEUFE93F+L766iuvjwoHbd261au5oVe3bWZ2/PHHe7WWLVt6NXextYkTJ3p9hg0b5tVUoNldqOm5557z+rhBefX6Sotp06Z5NXexPBUecxf12x8VJnSdddZZXi0tLc2rzZkzJ9b+61//6vXp37+/V3v33Xe9mhuOdENuZnrBvpAQvFqcT9XUooSTJk2KtQmDHzp3flBjVy0ypca9+zmqPmo+VdyApApMqhsLoPC584P63k+YMMGr/d///V++21bBb7WoqFoYr0qVKrG2Gn/qce7CqWZhAV3Vp0+fPvk+DsVDBbjdsaXmJ3UzDDUm3ZutqIUg1ZhRQW93Hg4Zj2UBf9EAAAAAkDhONAAAAAAkjhMNAAAAAInjRAMAAABA4spkGDw0mBji7rvvjrVzc3ODHqdWjHbDRyqs9vnnn3s1FUB3g+udOnXy+hx55JH57oOZ2d///vdY+7vvvvP6DB06NNYuLSuDz5gxw6upVWTd90UFDkNDiDVq1Mh3v2bNmuXV1Hhwx5sKXqrxrkJtbj83hL0/Kmy3fPnyWFuNK3VzBTfYaWb22WefxdoDBw4M2i/snxvsVWNE1VQ40e1X0MeZ+QFj9Tj1PUPhU+Fvl5oLmjdv7tUWLVoUa6sbD6Snp3s1dQMJ97FqzKgbaaxevdqrhYy/xo0bezWUXOp4npOTE2u3adPG66PGpJqzVPjbFXK8NfPHt3uTnbKKv2gAAAAASBwnGgAAAAASx4kGAAAAgMRxogEAAAAgcWUyDK5CqAVVvXr1WFuFwVXAVa3E64aK3NV7zXRASQWO3deoQuRqtXAVUFq5cmWs3bt3b69PafXnP//Zq6n3MzU1NdYOWQ3bTH9ebjBMhfnXrl3r1datW+fV3DHjflbq+fa3X7t27Yq1N2zY4PUZMmSIV1u/fr1Xc8e82pb6Xqhg3ddff+3VcGjckKtamVmFrkNC3Sr4r4TMw+oGCChd1DHFPbapkLc6RqqAuDtvqbktNFQbMnbr1KkTtC2UDPXq1cu3T2jIO2SlbjWvuTff2F/NnXPd35dlFX/RAAAAAJA4TjQAAAAAJI4TDQAAAACJK5MZjSS51+WHXtesrk93ryWsWbOm18ddaMZMX9/qXnMYupic2pZ73eqyZcu8PqVV9+7dvZrKOSxYsCDW3rhxo9dHZTTUoojue9y1a1evj7pWWH02bk2NNXWtachiaWrMZGRkeLVWrVp5ta1bt+a7X2ofsrKyvFq/fv28Gg5NyLXG6vNXY9DtF7Lt/XGvW1YZDfX9RNFTn7MaHw0aNPBq06dPz3db6rNX29+xY8dB9zHTx2A337FmzRqvT8OGDb2a4o7lkAUPUTRUjqeg3EyGymio47kaD+4xUR1vyyL+ogEAAAAgcZxoAAAAAEgcJxoAAAAAEseJBgAAAIDElcn0khu4UUE0Fd5RC+gtX7481lYBNrVYkLvIkHqsu0icmQ4hq9C4G0xWz5eWlubVNm3a5NXat28fa7tBXzN/0Tn1XpVE1113XVDNXZRu/vz5Xp9nn33Wq40bN86r1ahRI9Z2318zs2rVqnk19RkeSvDWFfK9UCE6NSY7dOgQa7/22muHuHcoKLWgohvgVsF8FWpMcryp0K4boFXjTd10QYV9kwx8ouCaNm3q1dzxp+Y2NW6bNGni1dxQrVrsVC18psK47rE65KYZKH0Kumizepw7RlSf0LnU7ad+A5ZF/EUDAAAAQOI40QAAAACQOE40AAAAACSOEw0AAAAAiSuTqSc3cKNWwFVh8CFDhni13NzcWLt27dpeH7UCt9q+G7JesmSJ16dChQpebefOnV7NDayp1aHVfqmVUH/zm9/E2lOnTvX6uCFO9Z6WZm6YsEuXLl4fdSOATz75xKu54099fipw777HZjpQ61KBRlVzt6X2S40/FcRVK66jeKhx6dYKGo4MfawabwVdnTwzM9OrEfwuuapWrerV1PHPpeY2NWZCVgZXYfDVq1d7tZCbmKjgOkoXNR8V9HEhK8CreUyNU7e2atWqg93FUom/aAAAAABIHCcaAAAAABLHiQYAAACAxHGiAQAAACBxZTIM7oZ31MrdSrt27byaG6pUoevQsLkb/FEBR3dVaTMdEnb3Q4WLVUCuUaNGXs1d1fmWW27x+pxwwgmxtlphvLRQgS/3/VRjRoVi09PTvZo7HtRYCA3nhqxKmqTQlaHVyuau0IBcYb+msk69fyX1Zg0hN0pAyRByIwozHY51b5qi5lN1fFLcuUZtS934pG7dul7NDYgfLiszH26SDIO7x8SQ1cPN9O8292YrOTk5B7mHpRN/0QAAAACQOE40AAAAACSOEw0AAAAAiUs0o6GuU1PXCqvrwN3HqoXDDuWa0RBnn322V0tLS4u1q1Sp4vUJXeDHvW5VXcOnFkcLyZio16zeL/V5TJ8+PdZWC2aVJeoaSzXeXC1atPBqGRkZXq2gGaGQaz+TzDOo/QodyyFjRH3PQxbywsEJyWOELo4WIslthY4R1S/0eICCCX3PVV5v/fr1sbY6bq5duzZoP9zj5rZt27w+Gzdu9Goh8656jWohXaWgvzNQ+EIyGiG/Q0O3HZqTc+c2MhoAAAAAUECcaAAAAABIHCcaAAAAABLHiQYAAACAxB1SmilkYbLiCEx99tlnsfbQoUO9Pp9//rlXq1q1qlerWbNmrK0WmFJBIPW63e2rsJDavgqIu88ZuvCQCvu6jx02bJjXp0+fPkHbL63cYJgayyrQ6C7oaOZ/XiporhZ+DAmZqT4hiwwpasFIFbRU2yfUXXKEzA+hi0yFBLEPZTHAkJsbqJqat9T4RXJCw/ZuWNvM7Oijj461Gzdu7PVRc436TFeuXBlrq5B3kyZNgrblBtfr16/v9fn++++9GkquefPmeTV3vlBzSugNLNw5K3QxQNXP/V24Zs2aoG2VdvxFAwAAAEDiONEAAAAAkDhONAAAAAAkjhMNAAAAAIk7pKR2QQOh69at82rLly+PtVXAx+1jpoPL7mNVYFcFgVSg2l29NCsry+ujQmcq7OuG2tR+qYBc9+7dvdrmzZtj7fHjx3t9VJhPrejshpW/+OILr09ZF7Litno/Va2gQdyQ/SpogC30OUNXkw8Jiia5ijn2L+SzDl3NNnT7SQnddkFXHkfhU8eeFi1axNqhYe309HSv5h7rNmzY4PVRN3JRoXH1G8LlHqfNzFatWuXV6tSpE2uzen3xmDNnjldr2LBhrK3GgvqtpbjHv9A5Sx033d98K1as8PpMnDjRq6nfgKUJ3wIAAAAAieNEAwAAAEDiONEAAAAAkDhONAAAAAAk7pDC4JMmTYq177nnHq/P6tWrvZoKc7mhKRWsqlatmldTgXQ3UKZC1yrQo1Z+dkM4Q4YM8focf/zxXs1dgdTMD7/l5OR4fZTp06d7tS1btsTabvjJTIfbVQBq69atBdov6HChO05DV1IuaIC7oNS21Srmqt+ePXsKZZ9w8A5lpe4QIavTKyEBdDWO1OthvBU+95irgsxLly71arNnz/ZqzZs3j7XXr1/v9XFvtGJm1rJlS6/mHp++++47r0/16tW9mjoGh0hLS/Nqr732mle76aabYm2C38Xj448/9mohN1EJDe+781joTTTU9t3HqvH+7LPPejXC4AAAAADg4EQDAAAAQOI40QAAAACQuIPKaOzduzd2/eygQYNi/66uVy9f3n8KdR2cyhO4du7c6dVUrkLVXBs3bvRqixcv9mq33357vttW19TVr1/fq7kZjdNOO83r4y50ZGY2f/58r+Ze36qurVfXNavrBt3PyF2I6HBQ0MXlQhat3LVrl1cLuRZU1UIXXgvpp/ZL5ZnU9kOumWfBvqKhPmt3XIaOkZCF8UI/V9UvZPtqv9R8nZGREbQfCBOSMRg9erRXO+qoo7zajh07Ym31WanjbYMGDbza3LlzY20156qMoso21q1bN9ZWORGV9/j++++9mntcPvLII70+KHxqgWH3N406Xh3Kwnsh1Fznfi/U8VYt2Ffa8RcNAAAAAInjRAMAAABA4jjRAAAAAJA4TjQAAAAAJO6gwuCvvfZaLAzthrncRXrM/MV2zMw2b97s1VQoy6UCPSok6AbDVMBs+/btXs0NipmZDRw4MNZ+++23vT59+vTxaosWLfJq7nvx9ddfe33Gjh3r1VQYyQ0RqaC8CvsqbnBKPc5dqEl9hocjFeZyQ2AqvBi6WFDIAmfqRgAqiOuOI9VH3bxBUYtuonjs3r3bq7njK8lF9pKkxpt6PjdEieKhAtYdOnTwau74U8cUdcxSQm48ETJ3mvk3ZFELEKrgekiYnTB48VALDLuB/kOZ10KOm6Hc74X6HbpixQqvpr4r6rdHScVfNAAAAAAkjhMNAAAAAInjRAMAAABA4jjRAAAAAJC4gwqD165d26pWrZrXdkPXKiCsAiuNGzf2au5jVcBx06ZNXq1GjRperUmTJvnulxsK21/NDfL279/f69O+fXuvpgJKbuBdvTfVqlXzairs6+5XxYoVvT4FXZFaBafmzZsXa6uQ/+EoZGVwpaDhNHVjgNAAt7v90H1Q40+F2EK2heSFrHqrxk1xfD4hY1WNt9AwO5KjbmhSv359r6aC+mlpabG2GqNq7gyZV9QYUse1kLD5T3/P/EiFcdUNZVavXp3v9pGs9evXezX1OdSpUyfWVmNBjRl1kxZ3ngz5DbW/mrsfZ555ptfnv//9r1dTNw7q3r27Vyup+IsGAAAAgMRxogEAAAAgcZxoAAAAAEgcJxoAAAAAEndQYfCsrKxYyMsNxTRq1Mh7jAoNq/COG4KuXbu210fVVMjMDdyoPirAtmXLFq/mhihr1qzp9Zk9e7ZXc8NwZn4I3l29cn/7pV63G4hTAUoVmgtZiTIzM9PrM3Xq1Hz383CkQrYhChrEPZRQrPucIcE3Mx3a3LZtW4H3A8lSN31wqc81NAxZmEJvPsDNJ4qeWjVbjRl1fHXHpDpeqOOTugmMSwWC1bbU3Ozua7Nmzbw+8+fPD9rWxo0bY+1169Z5fdTNalBw3377bVA/dzyo3z2h8587dtV8q46RIXNbdna210eNtTlz5ng1wuAAAAAADmucaAAAAABIHCcaAAAAABJ3UBmNDh06WEZGRl7bXbzupZde8h6TlZXl1Vq0aOHV3MXyVF5CXRunrr1zr/NU15CqxflUP/c6O7XAj1rESF3/517Hp55PLdgXshCiepyqqYX93OsG1UJNdevWjbVDFlYqTZJcvCzJ69xDMhmhOZGQBfvUvode/4zioeZF97NWn2FxLILnji91bbPKaCxcuNCrdezYMbkdg0cdn9T8oI6JboZLZS/UsUiNB/dYqo6HanyrBXG///77WLtz585en88++8yrqWO8+/6o7AgZjWSNHDnSq9WqVcuruXNIyLgy07873XlSfS/U4376W/lH7jhVi0OqfZ0xY4ZXK034iwYAAACAxHGiAQAAACBxnGgAAAAASBwnGgAAAAASd1BhcNedd94Zax977LFen7/+9a9eTYWN3UXpVJBZhc5UOM1dsC9k4R4zHY50Q5UhixOZ6bC0+9jQMKbq574XKiCnFhBSASg3kNShQwevz69+9atYe9OmTXb11VfrHS6F3Pc4NByuAo0FDcqHLCCkgmLqO6C25VKvUY019ZwhYfAkA/bYv+XLl+fbJ3RxRjVu3M869HMNGZdqvKlgrwp8onCtXbvWq6ljnVpQdubMmbG2mhPVwrBq++54CL1RjLrhy/Tp02Ptc8891+ujfnuo7bvhb/XbAMlSN4VQv33c3zTqeKUWX1bh7HfffTfWPu+887w+VapU8WpqUVu1kHPI42bNmpXv40oy/qIBAAAAIHGcaAAAAABIHCcaAAAAABLHiQYAAACAxB1UGHzfvn2xgJ8b7DvnnHO8x6jaJ5984tXcYHlOTo7XZ+PGjV5NhQnd4I9alTR0pdw6derE2ioI2bBhQ6+mgmhuEOhQVld2Q8ihQflevXp5tbZt28ba3bt3L/B+ISzAHboqt1sLDX6H3GhAjeXQVc1ZGbzkUHONO+epz1p9hiE3Awj97NUK3+5jQ1fnbdy4cdBzIjmrV6/2amp+UKHaDRs2xNpqzGRlZXk1FbquXr16rJ2amhq0XyFUONd9PjP9/XH3Izc31+vTunXrAu0XNBXEHjdunFdz5zE1z6jQtRIS4Fa/J9X8F/I4NZ+3b98+322VZPxFAwAAAEDiONEAAAAAkDhONAAAAAAkjhMNAAAAAIk7qDB4uXLlglYczs9pp53m1b744ot8Hzd37lyvpgJrbphr2bJlXp8mTZp4NbXKc4sWLfLdL5R+BV3FWgUa58+fH2urwJf6HqmaG45UfdS+q5q7H+omCaFYGbzk6NKli1ebN29erO2Gc8106FBxg5VqPBf0s1YBWjXGCdUWva1bt3o1ddMRd4VsZceOHV5NHW/V6truMV6tRK72Vf02cGtqpenQm2u4Y16tUI1kXXXVVV7t6quv9mru56VuWKBu0qKE/OatVauWV1NzrjvmN23a5PVRtUGDBuW7DyUZf9EAAAAAkDhONAAAAAAkjhMNAAAAAIk7qIxGcWvTpk1QzdWuXbvC2B1AXofpLjimshBr1671air34C5EdSi5CvfaevV8avHJ7du3ezV1bbMrdHFBHBp13fxll10Wa48dO9brs2bNGq+mrnV3r5sPWYjKTI8vdww2bdrU66MyfOo1onC5WTMzs2bNmnk1lb9wqblALZimckPuArKvvfaa10dlO04//fR890Ptl5rT1fhr3rx5rH3qqad6fVD4pk+f7tU6dOiQ7+MqVaoUtP1Vq1bl22fFihVeTX0v3DlR5XpGjx7t1VSmuDThqA8AAAAgcZxoAAAAAEgcJxoAAAAAEseJBgAAAIDElaowOFBY3AV+Qhcg69Spk1c7+uijY+1q1ap5fUJD3W5YMS0tzeuj9lUtMOUGcVUwWwV9VThSLRLnIvhdNNRn7YZqzz777KBtrVu3zqu5QceNGzd6fdQYrFevXr61gi4auL/nRHKeeeYZr6YWa1SB6osvvjjWVjePUAHXpUuXejU3gN65c2d/ZwNdcMEF+fa56KKLCrx9FL327dt7NXe+GD9+vNdnzpw5Xu2TTz7xaj169Mh3H66//nqvpkLk7vfinHPOyXfbZQG/BAAAAAAkjhMNAAAAAInjRAMAAABA4oIyGj9e77Zp06ZC3RmUHj+OBXXtdNKKYvwVNKOxc+dOr7Zr1658+xQ0o6GukU4yo6EWWVP77y62VdRzQ1GOv58+T0mcA5PML6jX5y5AqRb1U8+nFqNyFz5zvyv7U9IyGofD+FOL4IVmNNy5RY0F9VpC+6HsHYMLk5qz1EK06rjsjsnU1FSvT8jvALPiP24m6WDGX9CJxo9vdKNGjQ5ht1AWbd682TIzMwv9OcwYf/AVxfj78XnMGIOIY/yhuHEMLlzDhg1LbFtvvPFGYtsqKULGX0oUcDqyb98+W758uaWnp3OnD5jZD2exmzdvtqysrEK/uxDjD66iHH9mjEHEMf5Q3DgGozgdzPgLOtEAAAAAgINBGBwAAABA4jjRAAAAAJA4TjQAAAAAJO6wOdG477777Nhjj93vvw8ePNiqVat2SM9x+eWXW79+/Q5pGyj78huLZmannHKK3XTTTUWyPzj8MAYBHK6Y/4pWqTnRmDRpkh1xxBF27rnnFveuFDu+AEUrJSXlgP+77777En/OYcOG2YMPPnjAPjk5OZaSkmJTp06V/37//ffbr371KzP74TW8/fbbCe8ligpjEKXd5ZdfnjdeK1SoYHXr1rVevXrZiy++KNfiAH7E/Fe6Ba2jURK88MILdsMNN9gLL7xgy5cvt6ysrOLeJRwmcnNz8/7/kCFD7J577rHs7Oy8WlpaWuLPWaNGjQP+e8hCZ++8847dfvvtSe0SihFjEGVB79697aWXXrK9e/faypUrbdSoUTZo0CB76623bMSIEXJBwN27d1uFChWKYW9RUjD/lW6l4i8aW7ZssSFDhti1115r5557rg0ePDj27+PGjbOUlBT7+OOPrXPnzla1alXr3r17bCC6Fi5caM2bN7frr79+vysbvvPOO9apUyerXLmyNW/e3O6//365Wqrr/vvvt9q1a1tGRoZdc801sQG5c+dOu/HGG61OnTpWuXJl69mzp02ZMiX2+E8//dS6dOlilSpVsvr169vtt9+e97yXX365ffrpp/bkk0/mnc3n5OTku08ouHr16uX9LzMz01JSUmI1NcmNGzfOunTpYqmpqVatWjXr0aOHLV68ONbnlVdesaZNm1pmZqb94he/iK1A6v7VqmnTpvbggw/aZZddZhkZGXb11Vdbs2bNzMysY8eOlpKSYqecckpe/6VLl9qsWbOsd+/e1rRpUzMz69+/v6WkpOS1zcyeffZZa9GihVWsWNFat25tr7zySmwfU1JS7Nlnn7Wzzz7bqlSpYs2bN7e33nqrgO8kCooxyBgsCypVqmT16tWzBg0aWKdOnezOO++0d955xz744IO84/qPn/f5559vqamp9vDDD5vZgY/HURTZfffdZ40bN7ZKlSpZVlaW3XjjjXnP+8wzz9iRRx5plStXtrp169qFF15Y5K8dBcf8V8rnv6gUeOGFF6LOnTtHURRF7777btSiRYto3759ef8+duzYyMyirl27RuPGjYtmzZoVnXjiiVH37t3z+tx7773RMcccE0VRFE2bNi2qV69e9H//9395//7SSy9FmZmZee3PPvssysjIiAYPHhwtXLgw+vDDD6OmTZtG99133373c+DAgVFaWlp08cUXRzNnzoxGjhwZ1a5dO7rzzjvz+tx4441RVlZW9P7770ezZs2KBg4cGFWvXj1au3ZtFEVRtGzZsqhq1arRddddF82ZMycaPnx4VKtWrejee++NoiiKNmzYEHXr1i266qqrotzc3Cg3Nzfas2dPgd9bHBx3nCi7d++OMjMzo9///vfRggULotmzZ0eDBw+OFi9eHEXRD2MxLS0t+tnPfhbNmDEj+uyzz6J69erFxsnJJ58cDRo0KK/dpEmTKCMjI/rrX/8aLViwIFqwYEH05ZdfRmYWjRkzJsrNzc0bQ1EURX//+9+jM888M4qiKFq1alVkZtFLL70U5ebmRqtWrYqiKIqGDRsWVahQIXr66aej7Ozs6NFHH42OOOKI6JNPPsnbjplFNWvWjJ577rkoOzs7uuuuu6Ijjjgimj179qG+lSggxiBjsDQaOHBg1LdvX/lvxxxzTHT22WdHUfTD512nTp3oxRdfjBYuXBgtXrw43+Pxm2++GWVkZETvv/9+tHjx4mjy5MnRv/71ryiKomjKlCnREUccEb322mtRTk5O9M0330RPPvlkkbxmJI/5r/TNf6XiRKN79+7RE088EUXRDwOoVq1a0dixY/P+/ccTjTFjxuTV3nvvvcjMou3bt0dR9P9ONCZMmBBVr149+utf/xp7Dnfwnn766dEf/vCHWJ9XXnklql+//n73c+DAgVGNGjWirVu35tWeffbZKC0tLdq7d2+0ZcuWqEKFCtGrr76a9++7du2KsrKyokceeSSKoii68847o9atW8dOpJ5++um8bUSR/wVA0QmZ5NauXRuZWTRu3Dj57/fee29UtWrVaNOmTXm1W265JeratWteW01y/fr1i21n0aJFkZlF3377rfccvXr1iv7+97/ntc0sGj58eKxP9+7do6uuuipWu+iii6Jzzjkn9rhrrrkm1qdr167RtddeK18bCh9jkDFYGh3oROPiiy+O2rZtG0XRD5/3TTfdFPv3/I7Hjz76aNSqVato165d3raHDh0aZWRkxMY6Si/mv9I3/5X4S6eys7Ptyy+/tAEDBpiZWfny5e3iiy+2F154wevboUOHvP9fv359MzNbtWpVXm3JkiXWq1cvu+eee+x3v/vdAZ932rRp9sADD1haWlre/6666irLzc21bdu27fdxxxxzjFWtWjWv3a1bN9uyZYstXbrUFi5caLt377YePXrk/XuFChWsS5cuNmfOHDMzmzNnjnXr1s1SUlLy+vTo0cO2bNliy5YtO+A+o+gtWbIkNkb+8Ic/WI0aNezyyy+3s846y/r06WNPPvlk7BpTsx/+DJuenp7Xrl+/fmysKp07dw7ap02bNtmnn35q559//gH7zZkzJzYWzX4Yaz+OxR9169bNa7t9UHwYgyjtoiiKHfPccZbf8fiiiy6y7du3W/Pmze2qq66y4cOH511W1atXL2vSpIk1b97cLr30Unv11VcPeAxH6cL8V/KV+BONF154wfbs2WNZWVlWvnx5K1++vD377LM2dOhQ27hxY6zvTwNjP05aP72bRe3ata1Lly72+uuv26ZNmw74vFu2bLH777/fpk6dmve/GTNm2Pz5861y5coJvkKUZllZWbExcs0115iZ2UsvvWSTJk2y7t2725AhQ6xVq1b2xRdf5D3ODTempKTke+eV1NTUoH364IMP7KijjrJGjRod5KtBacQYRGk3Z86cvOvdzfxxlt/xuFGjRpadnW3PPPOMValSxa677jo76aSTbPfu3Zaenm7ffPONvf7661a/fn2755577JhjjrENGzYU8atEYWD+K/lK9InGnj177OWXX7ZHH300NpCmTZtmWVlZ9vrrrx/U9qpUqWIjR460ypUr21lnnRUL/rg6depk2dnZ1rJlS+9/5crt/22bNm2abd++Pa/9xRdfWFpamjVq1Cgv8DNhwoS8f9+9e7dNmTLFjjrqKDMza9u2rU2aNCkWUJ8wYYKlp6dbw4YNzcysYsWKtnfv3oN67Sgc5cuXj42Nn96pomPHjnbHHXfYxIkTrV27dvbaa68l+twVK1Y0M/PGwjvvvGN9+/aN1SpUqOD1a9u2bWwsmv0w1n4ciz/66eT8Y7tt27aHtO9IDmMQpdknn3xiM2bMsAsuuGC/fUKOx1WqVLE+ffrYU089ZePGjbNJkybZjBkzzOyH78gZZ5xhjzzyiE2fPt1ycnLsk08+KZLXh8LF/Ffylejb244cOdLWr19vV155pWVmZsb+7YILLrAXXngh7+w1VGpqqr333nt29tln29lnn22jRo2Sdyy455577LzzzrPGjRvbhRdeaOXKlbNp06bZzJkz7aGHHtrv9nft2mVXXnml3XXXXZaTk2P33nuvXX/99VauXDlLTU21a6+91m655RarUaOGNW7c2B555BHbtm2bXXnllWZmdt1119kTTzxhN9xwg11//fWWnZ1t9957r9188815E2rTpk1t8uTJlpOTY2lpaVajRo0DnvygaC1atMj+9a9/2fnnn29ZWVmWnZ1t8+fPt8suuyzR56lTp45VqVLFRo0aZQ0bNrTKlStbamqqffDBB/b73/8+1rdp06b28ccfW48ePaxSpUpWvXp1u+WWW+znP/+5dezY0c444wx79913bdiwYTZmzJjYY998803r3Lmz9ezZ01599VX78ssv5aWLKDkYgyiJdu7caStWrIjd3vaPf/yjnXfeeQccm/kdjwcPHmx79+61rl27WtWqVe0///mPValSxZo0aWIjR4607777zk466SSrXr26vf/++7Zv3z5r3bp1Eb5yFCXmvxKmuEMiB3LeeefFQjE/NXny5MjMomnTpuWFwdevX5/3799++21kZtGiRYuiKIrfdSqKomjz5s1R9+7do5NOOinasmWLDBiNGjUq6t69e1SlSpUoIyMj6tKlS96dLJQfw2733HNPVLNmzSgtLS266qqroh07duT12b59e3TDDTdEtWrViipVqhT16NEj+vLLL2PbGTduXHT88cdHFStWjOrVqxfddttt0e7du/P+PTs7OzrhhBOiKlWqxF4jCl9IEG3FihVRv379ovr160cVK1aMmjRpEt1zzz15YX53LEZRFD3++ONRkyZN8toqiPb44497z/Xcc89FjRo1isqVKxedfPLJ0ZgxY6KGDRt6/UaMGBG1bNkyKl++fOx5nnnmmah58+ZRhQoVolatWkUvv/xy7HFmFj399NNRr169okqVKkVNmzaNhgwZcsDXj8LFGGQMlkYDBw6MzCwys6h8+fJR7dq1ozPOOCN68cUX88ZlFOnQbBQd+Hg8fPjwqGvXrlFGRkaUmpoanXDCCXk3hxk/fnx08sknR9WrV4+qVKkSdejQgfFTijH/lb75LyWK9rOIBIBS58Ybb7Q9e/bYM888k8j2UlJSbPjw4davX79EtoeyjzEI4HDF/Ocr0ZdOATg47dq18+5QARQlxiCAwxXzn48TDaAMufrqq4t7F3CYYwwCOFwx//m4dAoAAABA4rhVEQAAAIDEcaIBAAAAIHGcaAAAAABIHCcaAAAAABLHiQYAAACAxAXd3nbfvn22fPlyS09Pt5SUlMLeJ5QCURTZ5s2bLSsry8qVK9zzVcYfXEU5/swYg4hj/KG4cQxGcTqY8Rd0orF8+XJr1KhRIjuHsmXp0qXWsGHDQn0Oxh/2pyjGnxljEBrjD8WNYzCKU8j4CzrRSE9Pz9tgRkbGIe+YWrojybPk1atXx9qffvqp1+ff//63V8vMzPRqrVu3jrUrVqzo9dmwYYNX+/LLL73a8ccfH2vfe++9Xp8qVap4tRCF/Z66Nm3aZI0aNcobG4Up6fGH0q8ox59Z0YzBkCWNkvxOf/75516tWbNmXq1BgwYF2n5OTo5X+/bbb2Pt/v37F2jbxa0sjj+ULhyDUZwOZvwFnWj8eHDLyMgoFScaO3bsiLWrVq3q9Slf3n/pFSpU8GqVKlU6YHt/NbV9t596L0vLiUZxPEdS4w9lR1H9Gb8oxmBRn2ikpqZ6NXXQKOjrVdty5+LS/n0uS+MPpRPHYBSnkPFHGBwAAABA4jjRAAAAAJC4oEunDkVBL+lZs2aNV3vyySe92pgxY7yae+mUukRg165dXm3KlClebdiwYQfcTzN9yZW6rnny5Mmxdvfu3b0+NWrU8Gonn3yyV7vhhhti7erVq+e7nwBKLneuDL2TzLJly7zaiy++GGs/+uijXp9NmzYdxN4lw31Nl156qdfnz3/+s1cbNGhQgZ5v3759+e4DAKDwMOMCAAAASBwnGgAAAAASx4kGAAAAgMQVekYj1MKFC2Pt8847z+tTr149r1atWjWv5mYmjjjiCK+PuiVt586dvdqWLVsKtC2VAXHX99izZ4/XZ+fOnV7to48+8moTJkyItX/96197fX72s595NQDFr6DZgY4dO3q1+fPnezV3HlG3+FbzqZtvM/PzX2rOzc3N9Wrbt2/3au7tu9Xz/f73v/dqf/jDH7za6aefHmu/9tprXh/1npLbKLlUptP9vNRnFXqL16K+hfTEiRO9mspmZmdnx9qtWrUq1P1CuKIeMwX1q1/9yqvdfPPNXq1Tp05ezT1eqN+0h4LZFQAAAEDiONEAAAAAkDhONAAAAAAkjhMNAAAAAIkr9DB4aEjmjjvuiLXr16/v9VGL0qlAtfuc5cv7L1MFfNzgt5kfigkNfm/dutWruSF1tV+VK1f2aiq86D7n008/7fU588wzvVpaWppXA1B41FwTEj7u1q2bV5s5c6ZXq1u3rldz5wc1D6t5S81JK1asiLVV8NsNeZuZVaxY0au54W8136mamudff/31WHvbtm1en7fffturqffe/YxKQrgT2qF8Nkl+ruPGjYu1Z8yY4fVRN2q48847vZo7/j788EOvT9IB3bKioItChz7OranHFXQfdu/e7dXUAtBqbF144YWx9rx587w+6jetmhMLe77jLxoAAAAAEseJBgAAAIDEcaIBAAAAIHGcaAAAAABIXLGsDK7ChG7gMCMjw+ujgjMqvOiGAlUwe+/evV5Nrfrt1lSQUK1uq4KJ7mNV6Eftgwpwu4FJ9RpHjBjh1X75y196NQCFJzRoN3z48Fj7iy++8Po0atTIq6mbRbhzZUjIcX81dy4OWb15f/3cOVDNnWof1FzZuHHjWHv06NFenw8++MCrnX322UHPiYIraLhe9VPHxBAvv/yyVzvhhBNi7fHjx3t9nnrqKa+WlZXl1aZNmxZrq9W81SrMTzzxhFc79thjvRrCqDFT0NW81e9Cl5rr1M0q1A0y3Meqee2zzz7zav379/dq7s022rRp4/VRNwlS1H4kib9oAAAAAEgcJxoAAAAAEseJBgAAAIDEcaIBAAAAIHHFEgZfv369V3PD4CoAtnPnTq+mQtfuY9UKuCGrw5r54R0VIFJBICVkhUkVbl+9erVXq1WrVqytXuOYMWO8GmFwoPCE3mRC+dnPfhZru99xM7PNmzd7tWrVqnk1N9ynbqQROpe5/UJWNd+fkMeGzs3unKfeh3POOcerqZuR1KtXL9ZW74Oam1H05syZ49XU5+Wu3G1m9tVXX8Xa69at8/oMHDjQq5188slezQ16u9veX80N8ZqZLViwINZu2bKl1wfhCnpzh5C5WvUJDVO7c9vSpUu9PmrOSk9P92rusebRRx/1+jRo0MCrFXQV80PBXzQAAAAAJI4TDQAAAACJ40QDAAAAQOKK5aLT6dOnezX3Gks3s2GmF0pRNXcxO7XYTosWLbxa06ZNvVrVqlVjbbUIS2pqqldT1+y5GZMZM2Z4fd59912vpp5zw4YNsfaWLVu8PmoRPwCFJzSP0bdvX6/mZgzUQp05OTn5Ps4sbHFQJWTBqiSpPEboom3u3O/O1Wb+scBMX7v/i1/8It/nQ7iCXvOtMpcTJ06Mtd08jZlZZmamV7viiiu82uOPPx5rq2vYb775Zq+2atUqr+a+RrVg2jfffOPVPvroI6/mjlMyGofGnRsOJVe2cuXKWFvletauXevVvv7663y3pbJFNWrU8GpqzG/cuDHW7ty5s9enpOAvGgAAAAASx4kGAAAAgMRxogEAAAAgcZxoAAAAAEhcsYTB3eCdmdmJJ54Ya7/66qten5kzZ3q1O++806upUFYIFUTbvn37AdtmOnS9Y8cOr+aGxtXieX/84x+92vHHH+/V3LC8CkJ+9913Xg1A8Zs0aVK+fdQCpUpI0FGFc0MDu2qBp6SE7pfaB/d1q0UJ1Tw8ZcoUr+Yekwp7Aauyzr2pQGjoX93UpFKlSrG2+h2gAv7//Oc/vdqoUaNi7bPOOsvro9SpUyffPiowroK933//vVd78cUXY+0ePXp4fdq1a5fvPuAHIeNv4cKFXu2mm27yau6Nd9TiebNmzfJq6iZEs2fPjrVPOeUUr4+6QYE6Frjfi9CFowvKfU8P5sYh/EUDAAAAQOI40QAAAACQOE40AAAAACSOEw0AAAAAiSuWMPitt97q1dywzqmnnur16dixo1fbtGmTV3PD4CpImJGR4dVq1qzp1dxVd9UKu6HhRXclRxVqUyuCqmC8u2qw2nc3LISiERKeVWNGhavc74V6nAqBlS9fsK+2u6Kq2odDoQK77r4eDkHcKlWqeLVdu3bF2qGfoRpv7jwV8r6bhQX8Qlbp3t9+hWxLUWPcXU1ZBSbdG3CYmb322mte7dFHHw3aD4QJmbcU9b1wx9Enn3zi9fnVr37l1f7xj38EPWdS1OrQ6vfJcccd59UqVqwYa6ux7G5/8+bNB7uLhw31O83VokULrzZ48GCvpn5bJaV27dpeTd3AQt0I4OKLL461Vfg85DeF6qfmbvd4ETp3m/EXDQAAAACFgBMNAAAAAInjRAMAAABA4jjRAAAAAJC4YgmDq9U4P/7441h76NChXp8PP/zQqw0cONCrPfPMM7G2G8I2M1uwYIFXU6uSuiE2FUpUQUs33GXmh3BUgE2tOvmnP/3Jq7lB7+rVq3t9hg0b5tUmTpzo1dTqpSi4goaZVQArZFsFDX673xMzs4ceesirLV++vEDbV0JCemXNtGnTvNrq1au9WmZmZqytQoFqXlH93KC0CgWGhrrdfoeymrfbT/VR+6DGuPvY9evXe33UDTEK+n1BuILOger4d9JJJx2wvT/bt2/3au73InQ/Q8Zybm6u10cdl9WNaM4+++x8t7V48eJYW/1ewaFRwW93PlJzaUGPa+qmR+q3rxpHn376aax92223eX1CA9sh/Q7lZgT8RQMAAABA4jjRAAAAAJA4TjQAAAAAJK5YLla9/fbbvZp73axafKRt27ZebcSIEV7tgQceyHcf1DV16nrekOuT1TW/IVmOrVu3en3cBQLNzLp27erV6tWrF2ura/3U4n/kMYpeaPaioNeOqwXIpk6d6tXefPPNWNu9XtlMLyA0YMAAr/b6668fxB7+P+6idGZmjzzySKx91113FWjbJZWaC1QOwaWuwVaLLanx5T5naBZC9XOvSVb7ELqtkGuBQx/n7pea09W+Llu2LN99QMlR0PGnuP0OZtGx/KjclbuwrlnYd1F9993jg5pXcGhCjtWheYyQhXQvu+wyr497nN7ffrk5Y5VJUgtgKrNnz461f/Ob33h9GjRoEGurbPL+8BcNAAAAAInjRAMAAABA4jjRAAAAAJA4TjQAAAAAJK5YwuD9+/f3au6CfV9//bXXx13Uxszs/PPP92qrVq2KtRs3buz1UYuuqHCLG7BRj1NUsLdq1aqxtgoVqUVQ3IV6zMwef/zxfPuMGzfOq3Xs2DGohjAh4bHQRaHmz5/v1dxg2KRJk7w+aiHL5s2be7WGDRvG2mpxrJycHK/2/vvve7WCeuONN7za5MmTE9t+SfTNN994NRWKD1nMTi3YpwJ/7o0mQgOMaqy64duQPmZ6rgxZADV0jnX7qTCkurmBCui6Y1DdgAPFIySwrfqo70XI2Crowqnq5i7//ve/vdp5553n1X75y1/G2mqMuq8n9HuCcAVdaFJRc6JLjQW1ON+GDRu8mrvwo/sb2sysUaNGXk39/napxU/dm85s3rzZ/vvf/+a7LTP+ogEAAACgEHCiAQAAACBxnGgAAAAASBwnGgAAAAASVyxh8Dlz5ng1NyjtrnxtZnbCCSd4tQkTJni1GTNmxNoq4BO6kmhIsFeFx5SQFU7V63aDYmZmxx57bKzdrFkzr48KArVu3Tq/3SxT1Oes3nc3nKtCt0pIeEwFue68806vNmTIEK+Wmpoaa9evX9/r06VLF6+mbmywbdu2WLtNmzZen++//96r3X333V7N5d6AwUy/nptvvtmrzZ07N9ZWN4I47rjj8t2HkkrNDyGrX4cGuEOeU21rx44d+e6DmT9vHcoc6FLb2rlzp1fLzMz0au7qySpYrl632v4TTzwRaxd05fuyrqBB6ZLCHd+hgeqQQHrNmjW9mrrRyldffeXVfv3rX8faCxcu9Pp079491iYMfmgKOpZD5/OCfi/U7zZ1k6B169bF2n369Anaft26db2aO0+eeuqpXh/3t4f72+RA+IsGAAAAgMRxogEAAAAgcZxoAAAAAEgcJxoAAAAAElcsYXAVdHKDTUuXLvX6qKC0GyI384MtapVNFd5Rq3mHBLhDw5FuGFcFFVWoVr1GNwipQrwqhLxixQqvplaRLo1CQ1pKaPjbpVbjHDp0aKztrqhpZlajRg2vdvTRR3s1d0xu3LjR67Np0yavplbFdcNbKpSovmOvvvqqV/vLX/6S7/O1b9/eq6kgrhtKViuWl2Zq/lHcuUbND2qcqjFe0KBo6E0yCsrdV/V61Lyl5lj3Bg7VqlXz+qjXo55TBePhK03B7xAhIe/9mTp1aqx9zDHHeH0GDBjg1UaOHOnVRo8eHWu7Y9vMDwmreR/hinoV8FDTpk3zah06dPBqubm5sfYbb7zh9VFj5J577vFq7u/JXr165bufB4O/aAAAAABIHCcaAAAAABLHiQYAAACAxBVLRkNdU1y5cuVYW+Ul1LXbbu7BzL9eTl2nq65hVvvlPlZdi6cep/q521LXYap9rVWrlldzuYu3mOkFrJYvX+7VykpGQ11zWdBrcJ966imv9uyzz3q1lStXejX3Wtp27dp5fdT4VttyqdcYmhFyx2Tt2rW9PqHX/bqLRw0fPjzocQ899JBXe/rpp2PtJk2aeH3+85//xNpqAaOS6g9/+INXU/kLt6byLOp7rhYKK+gCeoXNnXdVXkJ9Z9V74S5KqbIw6vigMm9vv/12rF3aF6aD5o6/0OPDn//8Z6/mfhevueYar88rr7zi1dT39Zxzzom1c3JyvD7ud6WguUIcHHcuUPOA+q2lxpb7WDXPVKpUyaup374FneMffvhhr+b+7rzooosKtO394S8aAAAAABLHiQYAAACAxHGiAQAAACBxnGgAAAAASFyxhMFV4NkNtqgwdfXq1b3a9u3bvVpIGDw02Of2Cw3eqrCnG2hUASK1r3Xr1vVqbnheBY/U9ktTiDY/33zzTaz90UcfeX2ys7O9mlqcyw3Jq/dJLQjWsGFDr+YuqqeCrGrhPcUNrqrPNPRmBG54VvVRC++5Y83MbPLkybF2/fr1vT5bt271ag0aNPBqrVq1irVVgPe5556LtdV7WlJ99913Xk0F/tzXpG4WoYLy6v0qqWFwV+jcqb6P7nhWc3PojUCaNm2a77ZQ+rnHSRW6vu+++7yamnfr1KkTa7sLtZqZHXnkkV7NHbdm/vHncAx6u3NByO/E/XGPbUkuqBfyfGZhc0jnzp292qmnnurV3AUdQ6ljiJr/3ONKyA2IDgZ/0QAAAACQOE40AAAAACSOEw0AAAAAieNEAwAAAEDiiiUMrrhhKxWkqVevnldTQcgQoQFad79UQCm05gbRVChHUcHRkOCUWn069DlLon/+85+xsPKwYcNi/65uDKA+ZxW0cwN6qampQdvasmWLV3PHkVqJWAXLVeDQ/R6oILvaLxWWdseIer/U9lWgLDMzM9ZWNyNQN29QQV93P0rzDQu+//57r6beZxW2c+cy9V6pOUp9p91+oStdq89Rff4h1L662w9dGVfdPMH9HqubFqixpObFJUuWeLWyTI2Z0FWyi5q7r2rMqDGq5t05c+bE2rfccovXx705hZnZ0qVLvdqjjz4aa4feQGDq1Klezb1hRLdu3YK2VdxCVs0OnXvcWkkdj0po2PxnP/tZrN2hQwevz0svvRS0Lfd4HvL71UzfpKVjx45Bz1lQ/EUDAAAAQOI40QAAAACQOE40AAAAACSOEw0AAAAAiSuWMHhBV11V4VIVdnGpkIwKKKmQoBu4CQkx7Y+7fRXIU/ulwqRumDh0pWQV9i0tfvGLX1hGRkZe+/jjj4/9+4QJE7zHzJw506stXrzYq7mh0fXr13t91IquIWNm1apVXp81a9Z4tZCgrwo9qv0KWUE1LS3Nq6kQvArPu0E99R1Q4dyQ0KYKA5977rmx9tatW+3JJ5/0+hW38ePHB/ULCV2rMLh6T9etW+fV3M8sNPgdMpcV9qrZ6vNX49L9vqgbM6jjg3oP1c00yrKQoG3oKsyFPR5CbqKigt/qxgyPPfZYrH3aaad5fSZPnuzV3nzzzXz3M5R6v9zXpF5PSeS+ltDgd0HNnTvXq7344otezQ35165dO2j7ah5w5xn1G0rNKXfddZdXW716dazt3tDmYIQE0FUf9RpbtGiR77bczzZ0fjDjLxoAAAAACgEnGgAAAAASx4kGAAAAgMSVmAX7CkpdLxeyKFToInuu0OsNQ66NU9cib9iwwaupjMaRRx4Za6tFgNS19QdzXV1JE0VRbP/btWsX+/euXbsGbUflWRYtWhRrL1iwwOuTk5Pj1ZYvX+7V3DEZOv7UmKlZs2asnZ6enm8fM70goLvInuqjrg0OuV5YZS9Cx5q7eJ26Ht/93m3atClo20VN5SoU9d13x4R6/9T8oK5Zd7NDoeMtZF5UrzH0s3b3Vc2nodkUt5/KVYW8N9AKO3uhhFzjH7qQ23333efVsrKyYu3p06d7fYYMGRK0/YJS3zs3s6eO3SXB7t27Y5lA9/NSr01931R+4fnnn4+11QLNinvsNjN75513Yu3s7OygbYXkedVcpBZ0VLme999/P999UL/3frpQ8Y9CFuxTc6L6Xvfs2TPf/SKjAQAAAKBE4UQDAAAAQOI40QAAAACQOE40AAAAACSuWFJxKtDqLrYUuoCSCs64IUEVHgtZ7MQsbEEaVQtZSDA0rK3ei8aNG8faX331lddHBU5VOLK0qFatWmzBvq1bt8b+PTc313tMaGCpRo0asfYpp5zi9VE3HggJ/6r3XAWy1OfsPqfaVugifu621AJn7oJCZv5ihmr76n1Q34Ft27Z5NXc+UOHBJk2axNpq30uCk08+Oaif+vzdOSlkMUgz/d673331OLUP6jNzayqYqMabmnfd8aueT70eNe7d9yt0HxAWulY3Hli5cqVXU/Oumj9DFDSAfu+993o19f1xw9/Dhw8v0POZhR3j1T6osawWcC2JKlSoEHzDiwP55ptvvJo7tkKPkXXq1PFq7iK57777rtenT58++e7n/vbDNWDAAK/Wu3dvrxayMJ6aXwtqxYoVXk3dbKV79+6JPafCXzQAAAAAJI4TDQAAAACJ40QDAAAAQOI40QAAAACQuEIPg6ugqgrXuOG0n4Z+D0QFAENWflX7EBImLOgKuGpbKpAeGtBs2rRprK32XW1f9Sut3FCTCjmFcm8qEBosVaFkd+Xx0PdcjRk3/BYabg0JoKubMjRo0MCrhdzsoKBhYNVPfY7uir4ldWXw9957L6ifuhGEW1PB/Lp16wZty/3MQucH9ZkVNFgeMp5D5zu1Gq+7rZCxtb/a4SYk4Dp79myvplZAVsdq96YPVatWPYi9O7Dvv//eq02cONGrqZt3jB8/PrH9cN/Dgt5gxsxsyZIliexTYZswYUJsfnb3+8ILL/Qeo7676gYCrszMTK9WvXp1r6bC0+4xZNCgQV6f0DC4q2/fvl5t1qxZXs1dnbw4bNy40asV9LvIyuAAAAAAShRONAAAAAAkjhMNAAAAAInjRAMAAABA4go9DK6CTyFBbBVKVUJCr6EhrZBVv1UftX1VCwlCqiC7Wpn5yCOPjLVDw54HE+A5nLiBstDVOVU4DYevUaNGBfVT33M3dK2+988++6xXu+SSS7yaOx+kpaV5fdT8oILlbr/Qle4Vd1sqsKtqKtTorsK+ePFir0+1atWC9sulVsBWQfyiFkVRbA4v6EraISuDF/ZqwQV11VVXebV58+Z5tZEjRxbqfhT0RjHqezd37txE9qmw5eTkxI6Nv/71r2P/fvfdd3uPUXOPCvS7/dQK5OpmBGpb7nusbmBx6623erX//d//9Wq33XZbrD127FivzxlnnOHVatas6dWKmgrdq5vAhHDnh4OZe/iLBgAAAIDEcaIBAAAAIHGcaAAAAABIXKFnNBR1bZd7DZ27QNf+qOuF3evzVH4hZDEptS0l5HpXs4Jf06muTz766KNjbbXvqkZGAyg87mKNZvqaWHdBM7OwuaZ///5e7cYbb/Rqr732Wqyt8h7r1q3zavXr1/dq6jW51CJ4ag50r7tWC16qbXXt2tWruYtwffrpp0H7ELJg34gRI7yaygYUtZSUlALnMtzt5EcdK8455xyvpq6Rv/3222PtX/7ylwexd3EPPPBArK1yUDfddJNXa9++fYGfszCp3x7r168vhj05eJdccklsgcZ//etfsX9Xizyq16bmunr16sXaam7YsGGDV6tVq5ZXc3Neaiz/5S9/CarVrl071lb5zfvvv9+rKe5vstD8cEGp96uguTV3Xw9m3/mLBgAAAIDEcaIBAAAAIHGcaAAAAABIHCcaAAAAABJXYsLgblinSZMmQdtyF7ky88M7KowZEgg08xfWCg1dK+5rVCFLtViVCkWFLGioXuOePXvyfRyAglFzmwpiFzSQp/zpT38KqoVQ84+7/6E3v1A1d0HAnwZLC4PaV3VzkMqVK8fa7777rtenJITBx48fb6mpqXlt9/1Ux7oaNWp4tZ9u40fusdR9T/ZXW7BggVd79NFHY221oFmdOnW82ocffujVnnzyyVj7lFNO8foUdLwnKTSkr34vqN8xpUHTpk1j7S+++MLr07hxY6+2a9cur+YukqneJ7X4n/odFfJZqMV2Qz4HN7RuFn7jgSRu5PAj93WrkLq6kVDIwqPqOKC++6H4iwYAAACAxHGiAQAAACBxnGgAAAAASBwnGgAAAAASV+hhcBXGC1mdWoXalJBAtbsarZnZ2rVrvZob/DYr+GreihtuUkHIrVu3erXc3Fyv5gZz1Puggt8qhAUgGS+88IJXGzZsmFdT3/OiXjVWCQ0Al0RuMNXMbPXq1V5NBfHdY0aPHj2S2q1ELVmyJBb6zMnJif37qlWrvMeomxGoY6IbjlU3E2nUqJFX+9WvfuXVOnToEGuPGTPG6zNx4kSvNmPGDK/Ws2fPWNsNmpv5oXgzfUwsCaFrFdo966yzimFPDt0dd9wRa7/++uten6VLl3o19TvK/c2nfh+pz08FrN3fPuoGEGofVADd/f689tprXh9FbSvJOT3kt6gKdYeEwUNvcBSKv2gAAAAASBwnGgAAAAASx4kGAAAAgMRxogEAAAAgcYUeBt+7d69XU8GtgoauL7zwQq+2adOmWNtdKXx/+xWyWrh6XGjg3Q0CqfB5ZmamV+vcuXO++6XCfer1qP0HkAwVNF68eLFX6969u1dz561f/vKXie2XogJ/IbXQ1W1D+qlwpKqFrEbeu3dvr8/zzz/v1bZs2eLVzj333Fj7tttu83e2BLjkkksSWU1d3Qxl2bJlsfa6devy7WOmPxt3zKvgtzvezczOOeccr+Z+D1QgXSkJwW9FhcEfe+yxWPvuu+8uqt05JO6K2GosjBo1yqvdc889Xm3KlCmxthofxeHEE0+MtU899dRi2pO4kGC5+t5lZWXl+7gkVzA34y8aAAAAAAoBJxoAAAAAEseJBgAAAIDEFXpGY/v27V4t5DrgDRs2BG3fXTDmcKWuqVPvc+j7CiAZjRs39mpq4Ux3YSh1PbyiFv9LTU3N93Gh+YiSQGXL3Izbsccem28fM53RuP766wu+c6VQzZo1g2pInlpYsiyPP5WdUjXXvHnzvNrXX3/t1aZPn+7Vvv/++1hb5Y3Ub6YGDRp4tX/84x8H3E+zsExu0kIySLfeeqtXa926db6PUznqQ1EyjyoAAAAASjVONAAAAAAkjhMNAAAAAInjRAMAAABA4go9DF6jRg2v1qpVK6/mLsLTtWvXoO2HLOyX9OIjJZFa3GvRokVe7bjjjiuK3QHw/1Nz1F/+8hev5s6V9evXD9p+SV2YLEkhc7hamFUtjqber5Iagsfh4cEHHyzuXShx1O9EVRswYEBR7M4BFcdvzJDnPOOMMwq07ZDFqw8GsysAAACAxHGiAQAAACBxnGgAAAAASFxQRuPHa4w3bdqUyJPu3LnTq7kLWG3bts3ro56fjMYP1Hu6e/durxb6vubnx8eEvP+HKunxh9KvKMffT5+nIGNQ7aNayNT9bqqF5dTz79mzx6uphepKM7Vgn3sdsZrb1HuvFjJ1Fz3M73MuTeMPZRPHYBSngxl/KVFAr2XLlnlhbcDMbOnSpdawYcNCfQ7GH/anKMafGWMQGuMPxY1jMIpTyPgLOtHYt2+fLV++3NLT0w+Lvw4gf1EU2ebNmy0rK6vQ79jC+IOrKMefGWMQcYw/FDeOwShOBzP+gk40AAAAAOBgEAYHAAAAkDhONAAAAAAkjhMNAAAAAInjROP/17RpU3viiSfy2ikpKfb2228X2/4ABysnJ8dSUlJs6tSpxb0rKIWYA1FWXH755davX7/g/sydSBLjL67MnGhcfvnllpKSYikpKVaxYkVr2bKlPfDAA/Ie80DSVq9ebddee601btzYKlWqZPXq1bOzzjrLJkyYUNy7hsMEcyBKGuZFFCfGX8lQplZ16t27t7300ku2c+dOe//99+03v/mNVahQwe64447i3rUC2bVrl1WsWLG4dwMBLrjgAtu1a5f9+9//tubNm9vKlSvt448/trVr1xb3rh2S3bt3W4UKFYp7NxCIORAlSVmdF1E6MP5KhjLzFw0zyztjbdKkiV177bV2xhln2IgRI+yUU06xm266Kda3X79+dvnllwdve8aMGXbaaadZlSpVrGbNmnb11Vfnrdz74YcfWuXKlW3Dhg2xxwwaNMhOO+20vPbnn39uJ554olWpUsUaNWpkN954Y2xF2qZNm9qDDz5ol112mWVkZNjVV1990O8Bit6GDRts/Pjx9uc//9lOPfVUa9KkiXXp0sXuuOMOO//8883sh8tQnn/+eevfv79VrVrVjjzySBsxYkRsOzNnzrSzzz7b0tLSrG7dunbppZfamjVr8v591KhR1rNnT6tWrZrVrFnTzjvvPFu4cOF+92vv3r12xRVXWJs2bWzJkiVmZvbOO+9Yp06drHLlyta8eXO7//77Y//FOyUlxZ599lk7//zzLTU11R5++OEk3yoUMuZAlBQh8+Jjjz1m7du3t9TUVGvUqJFdd911eWPKzGzw4MFWrVo1Gz16tLVt29bS0tKsd+/elpubm9dn7969dvPNN+fNi7feequ3WvHBzp0o/Rh/JUeZOtFwValSxXbt2nXI29m6daudddZZVr16dZsyZYq9+eabNmbMGLv++uvNzOz000+3atWq2dChQ/Mes3fvXhsyZIhdcsklZma2cOFC6927t11wwQU2ffp0GzJkiH3++ed52/jRX//6VzvmmGPs22+/tbvvvvuQ9x2FLy0tzdLS0uztt9+2nTt37rff/fffbz//+c9t+vTpds4559gll1xi69atM7MfJsXTTjvNOnbsaF999ZWNGjXKVq5caT//+c/zHr9161a7+eab7auvvrKPP/7YypUrZ/3797d9+/Z5z7Vz50676KKLbOrUqTZ+/Hhr3LixjR8/3i677DIbNGiQzZ492/75z3/a4MGDvZOJ++67z/r3728zZsywK664IqF3CcWBORDFJWReLFeunD311FM2a9Ys+/e//22ffPKJ3XrrrbE+27Zts7/+9a/2yiuv2GeffWZLliyx3//+93n//uijj9rgwYPtxRdftM8//9zWrVtnw4cPj23jYOZOlA2MvxIkKiMGDhwY9e3bN4qiKNq3b1/00UcfRZUqVYp+//vfRyeffHI0aNCgWP++fftGAwcOzGs3adIkevzxx/PaZhYNHz48iqIo+te//hVVr1492rJlS96/v/fee1G5cuWiFStWRFEURYMGDYpOO+20vH8fPXp0VKlSpWj9+vVRFEXRlVdeGV199dWxfRg/fnxUrly5aPv27Xn70K9fv0N4F1Bc3nrrrah69epR5cqVo+7du0d33HFHNG3atLx/N7Porrvuymtv2bIlMrPogw8+iKIoih588MHozDPPjG1z6dKlkZlF2dnZ8jlXr14dmVk0Y8aMKIqiaNGiRZGZRePHj49OP/30qGfPntGGDRvy+p9++unRH/7wh9g2Xnnllah+/fqx/bzpppsK+C6gODEHoqTJb150vfnmm1HNmjXz2i+99FJkZtGCBQvyak8//XRUt27dvHb9+vWjRx55JK+9e/fuqGHDhnnfBWV/c+e3335bgFeJkorxVzKUqb9ojBw50tLS0qxy5cp29tln28UXX2z33XffIW93zpw5dswxx1hqamperUePHrZv3z7Lzs42M7NLLrnExo0bZ8uXLzczs1dffdXOPfdcq1atmpmZTZs2zQYPHpx3lp2WlmZnnXWW7du3zxYtWpS33c6dOx/y/qLoXXDBBbZ8+XIbMWKE9e7d28aNG2edOnWywYMH5/Xp0KFD3v9PTU21jIwMW7VqlZn9MD7Gjh0bGx9t2rQxM8v7E+v8+fNtwIAB1rx5c8vIyLCmTZuameVdFvWjAQMG2NatW+3DDz+0zMzMvPq0adPsgQceiD3HVVddZbm5ubZt27a8fozB0os5ECVJfvPimDFj7PTTT7cGDRpYenq6XXrppbZ27drYfFS1alVr0aJFXrt+/fp58+bGjRstNzfXunbtmvfv5cuX98ZQ6NyJsoXxVzKUqRONU0891aZOnWrz58+37du327///W9LTU21cuXKedfM7d69O9HnPv74461Fixb2xhtv2Pbt22348OF5lwyYmW3ZssV+/etf29SpU/P+N23aNJs/f35sEP/0QI7SpXLlytarVy+7++67beLEiXb55Zfbvffem/fvbqg6JSUl70+nW7ZssT59+sTGx49j+aSTTjIzsz59+ti6devsueees8mTJ9vkyZPNzLxLY8455xybPn26TZo0KVbfsmWL3X///bHtz5gxw+bPn2+VK1fO68cYLL2YA1HS7G9ezMnJsfPOO886dOhgQ4cOta+//tqefvppM4vPaWredMdyfkLnTpQ9jL/iV6buOpWammotW7b06rVr1/bCOzNnzrRTTz01aLtt27a1wYMH29atW/MOghMmTLBy5cpZ69at8/pdcskl9uqrr1rDhg2tXLlydu655+b9W6dOnWz27Nly/1A2HXXUUcHrEHTq1MmGDh1qTZs2tfLl/a/l2rVrLTs725577jk78cQTzeyHYK1y7bXXWrt27ez888+39957z04++eS858jOzmYMlmHMgSjpfpwXv/76a9u3b589+uijVq7cD//N87///e9BbSszM9Pq169vkydPzvsPMnv27LGvv/7aOnXqZGYHN3ei7GP8Fb0y9ReN/TnttNPsvffes/fee8/mzp1r1157rXd3lAO55JJLrHLlyjZw4ECbOXOmjR071m644Qa79NJLrW7durF+33zzjT388MN24YUXWqVKlfL+7bbbbrOJEyfa9ddfn/dfHN955x0vCInSZ+3atXbaaafZf/7zH5s+fbotWrTI3nzzTXvkkUesb9++Qdv4zW9+Y+vWrbMBAwbYlClTbOHChTZ69Gj7n//5H9u7d69Vr17datasaf/6179swYIF9sknn9jNN9+83+3dcMMN9tBDD9l5552XN6ndc8899vLLL9v9999vs2bNsjlz5tgbb7xhd911VyLvA0ou5kAUtfzmxZYtW9ru3bvtb3/7m3333Xf2yiuv2D/+8Y+Dfp5BgwbZn/70J3v77bdt7ty5dt1118XG9sHOnSgbGH8lR5n6i8b+XHHFFTZt2jS77LLLrHz58vbb3/42+L/kmf1wjd7o0aNt0KBBdvzxx1vVqlXtggsusMceeyzWr2XLltalSxf78ssvYyvsmv1wff6nn35q//d//2cnnniiRVFkLVq0sIsvvjiJl4hilJaWZl27drXHH3/cFi5caLt377ZGjRrZVVddZXfeeWfQNrKysmzChAl222232Zlnnmk7d+60Jk2aWO/eva1cuXKWkpJib7zxht14443Wrl07a926tT311FN2yimn7HebN910k+3bt8/OOeccGzVqlJ111lk2cuRIe+CBB+zPf/6zVahQwdq0aWP/+7//m9A7gZKKORBFLb95sUqVKvbYY4/Zn//8Z7vjjjvspJNOsj/+8Y922WWXHdTz/O53v7Pc3FwbOHCglStXzq644grr37+/bdy40cx+uLPQwc6dKP0YfyVHSnSwF5sBAAAAQD4Oi0unAAAAABQtTjQAAAAAJI4TDQAAAACJ40QDAAAAQOI40QAAAACQOE40AAAAACSOEw0AAAAAiQtasG/fvn22fPlyS09Pt5SUlMLeJ5QCURTZ5s2bLSsry8qVK9zzVcYfXEU5/swYg4hj/KG4cQxGcTqY8Rd0orF8+XJr1KhRIjuHsmXp0qXWsGHDQn0Oxh/2pyjGnxljEBrjD8WNYzCKU8j4CzrRSE9Pz9tgRkbGoe9ZAW3dutWrPfTQQ15t8uTJsfaAAQO8PldddVVyO1ZAw4cP92ovv/yyV+vVq5dXu+666wpln0Jt2rTJGjVqlDc2ClNJGX8lwfz5873amDFjvFr16tW9WqVKlWLtrl27en2ysrIOYe/yF0WRVyvIfyEryvFnxhhEHOMPxY1jMIrTwYy/oBONH38IZGRkFOsgO+KII7ya++PJzKx8+fjLqlKlitenJHxZqlat6tXcfTczq1y5slcrCftvVrAfiQV9juIefyVBWlqaV1PjQ415t5+aIAr7/U3qRCOJxxbkeRiD+CnGH4obx2AUp5DxRxgcAAAAQOI40QAAAACQuKBLp4rDNddc49U+/fRTr7Zv3z6vVrdu3Vj77rvv9vo89dRTXk2FnY488shYOzMz0+uzbt06rzZx4kSvtmvXrlh706ZNXp/69et7tWeffdarvfvuu7H2c8895/Vp3ry5V0PJUNBLiK699lqv9uWXX3q1PXv2eLWdO3fmu/3//d//9WrTpk3zatu2bYu1TzrpJK/Po48+6tXUJV179+6NtdUlkgAAoPThLxoAAAAAEseJBgAAAIDEcaIBAAAAIHElJqPxySefxNqLFi3y+nTs2NGrqZyDm9s45phjvD6rV6/2agsXLvRq7todnTt39vpMnz7dq6nb1NaqVSvWVq9n1apVXq1Zs2ZebcOGDbH27373O6+PWqcDJUNBMxorVqzwamrNDDcPZGZWsWLFWNsdQ2Zm//nPf7zajh07vFqFChVi7VmzZnl91HdAZaPcfVU5DgAAUPrwFw0AAAAAieNEAwAAAEDiONEAAAAAkDhONAAAAAAkrsSEwT/66KNYu2nTpl4fteCYG0o1M9u9e3es7YawzXRQVQV03cXEVOhVhVfT0tK8Wnp6eqz9/fffe32qVq0atF8NGzaMtVUo/vPPP/dqPXv29GooemqhyXLl/PN+Nyi9ZMkSr09qaqpXUwv2uTc2UGNUBcvVjRncYLkao7/97W+9mqJeNwAAKP04wgMAAABIHCcaAAAAABLHiQYAAACAxHGiAQAAACBxJSYMvnz58lg7IyPD6xMaBncD3OpxbpjVTIdj1QrLriOOOMKrqXD2tm3bYm0V/Fb7oMKy7mtUq0oTBi8ZVFBahcGVTz75JNZ2A91m/k0GQrevxrbavvr+uDdc6NChQ9C21Mrm9erVi7VDg/IAAKBk4+gNAAAAIHGcaAAAAABIHCcaAAAAABLHiQYAAACAxBVLGFyFPd3wdGZmptdH1Xbs2JHv87nBVTM/TG1mtmXLFq/mrrCsQuRq++o1uttSfdS2Kleu7NVcKgw+b968fB+Hwqc+GzWOlClTpsTabnDazKxatWpeLTs7O9/9UDcjWL16ddB+uTdr6Nu3r9fnww8/9GrHHXecV3NfkwrPAwCA0oe/aAAAAABIHCcaAAAAABLHiQYAAACAxBVLRmPRokVezc0rbN++3eujFvGrXr26V3NzDps3b/b6lC/vv3S1gJl7vbjKhKhrytVCgm5GQz1OXc+vFitT19e7vv/++3z7oPCFfs7K2LFj8+2jMhq9evXyat99912++6AyGscee6xXmzp1aqytvjsXXHCBV2vSpIlXc6kFMFGy5eTkeLVly5Z5NRYMBYDDC3/RAAAAAJA4TjQAAAAAJI4TDQAAAACJ40QDAAAAQOKKJQyem5vr1SpVqhRrqwC0CtWqcKm7GF96enrQttSCfW6oW+2XCn6rhfeqVKkSa6vQq1rIrX79+l5t69atsbba95o1a3o1FfatXbu2V0Ny1OKQ6mYEihvg3rZtm9fniy++8Go1atTwau6YVwtgnnLKKV5NhXoHDBgQa//hD3/w+iiHEoxHyfDmm296tbvvvtur9e7d26u5Ny5o165dYvt1KP7zn//E2q1atfL6dOnSpah2BwDKDP6iAQAAACBxnGgAAAAASBwnGgAAAAASx4kGAAAAgMQVSxh87dq1Xs0NPG/cuNHr89lnn3m1Sy65xKtlZWXF2ip8vnPnTq/mhrXNdDjbpYK96nHuyuDqcXXq1PFqKuzrhtLbtm3r9dm0aZNXmzt3rlcjDF64Qle6Hj9+vFdbtWpVrK3Cs+r7tH79eq9WvXr1WFvdGKBevXpebcGCBV5NjTeUXPv27fNq6sYW33//vVe78cYb8+3TvHlzrzZ9+nSvdvXVV8faEydO9Hc2kHsDjBdffNHrs2bNGq+2fft2r5aWlhZru8cQHBz3pg+HcsOHp556Ktbu1KmT1yf0uOke6zp06OD1adCgwcHu4iH74x//GGsfffTRXp/zzz+/qHYHSBR/0QAAAACQOE40AAAAACSOEw0AAAAAieNEAwAAAEDiiiUMrkKomzdvjrXHjh0b9Livv/7aq5100kmxtgoluivUmukAtxuiVKuA79q1y6u5wW8zsx07dsTa7ureZnql86pVq3q1yZMnH3DbZmYNGzb0atOmTfNqJ554oldDckKDkO7qxGZ+qFKNK7UCvLqxgTt21bbU45SLLroo1r755pu9Po899phXU+9FksFRaGpFdmXdunVeLTs7O9Zu2rSp1yc0jOvO4WrMn3rqqV5t5MiRXm348OGxtgp5q7lt4MCBXq2krFBeVuzduzfWVjc+UcaMGePVfvGLX8Ta6uYl7lgwM5s6dapXc4+lzzzzjNdH3djg+OOP92rHHXdcrK1ukJGTk+PVPv74Y6+2ePHiWFuNZcLgJZeaX9VYdsdWixYtgrZV2o+J/EUDAAAAQOI40QAAAACQOE40AAAAACSuWDIa//u//+vVevXqFWtv2LDB6+Mu3GOmF2lyF6WrXLmy10flMVTWwl3Uavfu3V4fdU2d2r57faibSzEz+/LLL73am2++6dXc69/VIlr/+Mc/vFqlSpW8GpLlXp8cumDfhx9+6NXc/IX6nLdt2+bV1DgNWXxSLf6nXHrppbG2eo19+/b1au+8845XK+3XnxYWtcieeq9C3r/QMdi+fXuvVqNGjVh71qxZXh93MUgz/xp2M3983XDDDV4flS075phjvNrvfve7WFvlLNyFYPcnJAul8nmHm9CFH91Mxpw5c7w+6ri2bNkyr/b+++/H2mqsqc+mcePG+e5XZmam10fVli5d6tWmTJkSa6vsiMqm/PznP/dq7qLC8+bN8/pAK+xMw3fffRdrP/DAA14flVv79NNPvVqfPn1ibZVtLI7j4d///vdY+9hjj/X69OzZs8Db5y8aAAAAABLHiQYAAACAxHGiAQAAACBxnGgAAAAASFyxhMEVd6G6YcOGBT1OBQDHjx8fa6twYegCVi4VhlM1NxBsZpaRkRFrq+CtepwbxjQze+ihhw64nyg+IWEutYikWtypWbNmsfbOnTu9PupmB40aNfJqbqitQYMGXh8V7FTc7+uECRO8PpdccknQtg5HIaHa0M+isP3lL3+JtU8//XSvjwr5p6WleTU3oFu3bl2vjxtMNDM7+eST893PQ+F+Z8t68Fsd/9ya6hN6U4FRo0bF2o8//rjX5/rrr/dqarG8kGD0ypUrvZqah90bZ6Smpnp91HdTLWTq9lPj3V3Y1Ex/r92w+fr1670+blBe3UymNAv5TVbQm2Gom6Oom1qMGDHCq7lBfWXGjBleTS2w6H6u7m9Vs2QXUFYLWl933XVezd3/fv36eX0IgwMAAAAoUTjRAAAAAJA4TjQAAAAAJI4TDQAAAACJK5YwuAr9uMEqFchSAT21kq0bylJhIbV9tRqsu7JnaEBTbcvdD3elcDO9AmkIFSJXQsN8KLiQMaJWAVfj213JXYXa1FjbsmWLV3OD5FlZWV6f1atXB+3XkiVLYu27777b66NcfvnlXm3w4MFBjy0JoiiKzV8hQUQ134WMkRUrVni1V155xat98MEHXu2TTz7Jd/uhunbtGmurlY3VPqhVkd15V4Vs1YrRIWFwNQdu3LjRq6nvxvbt22Pt5cuXe31+uiK12kZpEjIm1TEyOzvbq7Vu3dqr3X///bH2iy++6PXZunWrV3NvfmFm9qtf/cqrFdSGDRti7dGjR3t9pk6d6tXcG2mY+UHyFi1aeH3UfKqC625IXc25bhhcvX+FzZ3/3HFU0LD2wfQL4R6f7rzzTq+PGt9qNXl31W91c5709HSvpoLl1apVi7WHDx/u9Zk8ebJXq1mzpldzx8jcuXO9Pu77YGbWo0cPr+beiGbmzJlen0PBXzQAAAAAJI4TDQAAAACJ40QDAAAAQOI40QAAAACQuGIJg6vQjxtSDg1dq9U4XRUrVvRqO3bs8GoqvOgGDEOD5Wr/3edUq5KqfQ2hni/JcBU0FUB1x7Jazfupp57yascee6xXc8OXu3bt8vqoMaPCaa5atWp5tYULF3q1kFXuVaDbXT3czGzcuHFebeTIkbH2eeed5/Upqdzv/qF852666aZY+8svv/T6uO+7mV5F2F399Zlnninwfrn++c9/erXXX3/dq6nP2g0dqtWN//3vf3s1dZOMXr16xdpuWNbMbNOmTV4t5KYfKox75JFH5v1/NzxeUqiQtxqT6njhjjc1rtSq7aeddppXe++992Jt93M30yFvdSMAV8jntz9uGPfiiy/2+qiaCsc+/fTTsfZHH33k9VE371A3GnDn9Z/eeKAkSUlJiY2ngs536jeTe4OFNWvWeH1UuHndunVebf78+bF2o0aNvD7HHHOMV1M3AnCPf2ouVZ/XGWec4dVc6tit5jE1/7ljxr1xjJlZ7dq1vZp74wEzs3POOSfWVjcscG8+cDA3I+AvGgAAAAASx4kGAAAAgMRxogEAAAAgccWS0QihrgtX12GqayBDrrdVCy6pxaPcXIXalrreUO2re12vuj6vVatWXi1E6LW5SFbIAogPPfSQV1PXXLrXD5v515aqBbNUbkPlf0Ko1xOSQVLfHZVNqVy5sld7//33Y211Xf0vf/lLf2eLQVLXKCtHH310rP3qq696fX6aE/hRy5YtvZq7ENTtt9/u9VGLU4VQc6C63lldy+yOCbV4VMeOHb2aWpjVXUirS5cu+T7f/rjz9dq1a70+derUyfv/xbVg3759+2LfR3f8hY7HZ5991qu5OQp3PJqZnXLKKV5NZRPcfp9//rnXx70u3Czs+KdeY+jxL2SBOUVl6tyshfrNojJIan5z536VPXUXWFXbLmrusSF00TiVq3AX11RZApU9VJkd93M+6qijvD6fffaZV1ML49WtWzfW/uk88CP1mTZs2NCruVTOQc2l7kKTZv4xXs1J6j1Ui2JmZmbG2iof6GZmDmb88RcNAAAAAInjRAMAAABA4jjRAAAAAJA4TjQAAAAAJK7EhsFDff/9917NDSuqxfkUFcxRwUeXWvwoJKQeutCfWsDFDRqpMBySFfp5udRidir4rQLi7gJtKhS7YMECr6YW23LDsyo8FjLeFbWAmQruqcW9klxMrrDt2rUrFr53w3ZuqM4sPHB61VVXxdpqETwVxr3nnnu82gknnBBrjx49Ot/nM9Nj8Isvvoi1v/vuO6+PmmM7dOjg1Y4//vhYW920QAW41eKPX331Vayt9l2FKNXiXe53W825Pw0vF3Rh1UNVrly54MVsD0QFVd0QvgrZqpsRtGvXzqu571+nTp3y7WOmFx1zhdyAY39Cvovqu/Lcc895td69e8fa8+bN8/qoRVGrVq3q1dx5Q71GNwyuAsiFbciQIbEbf7g3mbjiiiu8x6hF49TCo24QW713KiS/evXqfJ9Thc/VArlqfLvHtuuvv97ro35/qeOrO7epGwioY7eyatWqWFstcBh6I6Rvvvkm1lYLbB4K/qIBAAAAIHGcaAAAAABIHCcaAAAAABLHiQYAAACAxJXYMHhogHLSpElezQ3AqJWTVbBXBdHcIJDqo4JbKhzrBh9V2E49zg39mPlhPvV6DiU0d7gJWTE2NIT57rvvxtoq9KjC4OqzdwN/ajVOtZKoGvOLFy+OtVUQTe2Xet3ujQ2U5s2be7UXXngh38eVZIsWLYoFEt3gqAoFqu+mWkndDTqqgLW74rd6nJkfVr766qu9PiogqW5i4W6rTZs2Xh8V4HbDsmZmU6ZMibUbNGjg9VHcVYPNzE488cRYe/r06V6f008/3aup76M7F7du3drr89PvQRKB7OKkVv4NCaGuXLnSq1WuXNmrueF9dxVtM7OFCxfm+3yKOm7m5uZ6NTVm3JvHqBvAqH0dOnSoV2vcuHGsXb16da+PutmBCui63zG1orM754bMwUnr1atX7CYf7j6ocTVz5swCPZe6mYg6Ri5atMiruful5ie1LVVzj5Nq/KmxprblzhtqLKg5WIXn3bGljgOhv6Pd37Xqe/7111/H2mqM7k/pni0BAAAAlEicaAAAAABIHCcaAAAAABLHiQYAAACAxJXYMHho2E6tiuyGoFUoR4VlVdDbDeaoAFZo6NoNd6oVQlUQKDs726u5K62Ghn6gJfn+uSs1uytym+kVSNXKxu6YUSt2fv75516tVatWXs39To0dO9bro8a3Ci6rcepSYeYQKlBdUsZ3ampqLHTnBqPV+z537lyvpsJ9bgBPrbKrxo0KQw4aNCjW7tevn9dHrdYcMi/Onz/f66NW7p4xY4ZXc282oEKaah/UeHP3Q93IYPz48V5N3TzBDeyrQHCdOnXy/r/6ThSFSZMmxUKgw4YNi/17/fr1vceo90Ude9xgtPr+qtetVhqeM2dOrK2+02qF9lGjRnk1N3yr5igV6g65iYUKa6ubHahtufP17NmzvT5q3KqaGwpWN4C58sorY211Y5DCVr58+di+/uIXv4j9u9suCur9dD8vNc+oILYapyHHMfUbUG3frZWU41oId7wdzMr0/EUDAAAAQOI40QAAAACQOE40AAAAACSuxGQ03Gtk1TWRaqGU1atXezX3mmJ1HZy6Fk9xrxdWOQ51zajavnsdn1rIS13XpzIartK+gFRJoz4b9R67C5CZmU2dOjXWrl27dtDj1LXUzZo1i7Vbtmzp9VHX6n7zzTdezV3Qp2fPnl6fL774wqupa+bdBabUdywzM9OrhSjJ161WqVIldp26uwCYWgRPXctao0YNr+YumKbGjcrxHHvssV5tyZIlsbbKY6gMhVp0zF04Kysry+ujMg3q2mZ3cTd17bSqqe+j+/6oBSjVGFyxYoVXc48tav7+aa5BHYuKQtu2bWP5Hnf8uW0zs7Vr13q1unXrejU336E+PzWW16xZ49Xcz1DlONR7/NBDD3k1N+OmFgoL/Tzc51T7oL4Xahy5NTVvhWTZzMyOOuqoWFt9jpdddlmBtp2kjIyM2Phzx4MaH+p7qTIN7u+o0Mcp7meh5hS10KTavpp7XGochfzGDF3gVdXc16S+A6Hvl7t9dcz/aUbN7OAymPw6BQAAAJA4TjQAAAAAJI4TDQAAAACJ40QDAAAAQOJKTBg8JDijgkY1a9b0au5iN2pxLBWgVUFsFUhyqRCOej3utlR4TG1LLUroUkHlkrzwWVEJDVa5719ouP62227zam7ITL3nKoimAoDuAn3qca1bt/ZqbrjQzF8QbvHixV6fdu3aeTW14JwbMnPD4WY6WFzapaenx+YTd5youUaNN3VTCTd0reY2FapVC4W5z6kWAFOL/6n5xw0dqtejbmSgFj5zw/JqgTk13tT75e6XCuyq0L0KMTZu3DjfffjpzRRCAqKFoVq1arExdvHFFxdoO+q45r4vahE8Nf7Ue+Eeq9X8oMLMaq7csGFDvs+nblCg5kp3fKtguft86nFm/u8F9d6oMK6aI9wFFBs2bOj1ccfywSyYVljc16JeG8om9f3aH/6iAQAAACBxnGgAAAAASBwnGgAAAAASx4kGAAAAgMSVqjC4WhVXBaJCVoVUKx+GrOwZstrt/ra1ffv2WNsNf5rpVcZDgnQqyK5Cc6ErRZYG7phR4UL13hV0FfW//OUvXk2tpH3yySfH2hMnTvT6qM9BhVndEKJ6jbm5uV5NhX9dzz//vFdTr8dd6dzMD/2p/VIrW5d2FStWjI0p9zPLzs6Wj3G5q4CbmW3cuDHWdldyNwtfSdalPh+1innIKs/qRhpqH9RzhqxmrAKl6jvrjnt1fHBDtmY6WO7O12q19Z/uQ0HnkJJCzT+pqakHbJv5qwMDQH5K92wJAAAAoETiRAMAAABA4jjRAAAAAJA4TjQAAAAAJK7EhMFDqNVaVRjcXUFTBS9DVyp1w7ihYXC1fXfFURXgVttSz+kGR2vVquX1CQnYl2ZugFO954paFXfJkiWx9t/+9jevz+OPP+7VunXr5tVWrFgRa3fv3t3r880333g1FbJ1g6vqJgOhwdQRI0bE2n369PH6vP/++0Hbcp9TjTUVblfcx5am1et/9rOfxdoqFD1//nyv5o4RMz/A/91333l9VEBXzQ/uTSVCbjRgZtasWTOv5q7wrm5iocLFatVvd1uHEqp2v8fqBghqjlU3B3H3P3TsAgAOjL9oAAAAAEgcJxoAAAAAEseJBgAAAIDElaqMhlpgSl1v616f7OYZzMxq1qzp1dS1++714uoabHWts1psy81oqGud1fbVfrnXeKuMxuHmrbfe8mr/8z//49XU56Wuc3ep67ZnzZrl1Y477rhYe/r06V6fFi1aeLWZM2d6NXdf1TXn6nr/4cOHezWVyXCpsRZC5SqysrKCHuuO+dK8qKTKHLRu3TqohoPjjhOVCQEAFC/+ogEAAAAgcZxoAAAAAEgcJxoAAAAAEseJBgAAAIDElZgweMjicosWLfJqKhzr2rJli1dr3ry5V1PBcpcKlruLUJnpxePc/di+fbvXx12gzUwHxNXibq6yvmBfbm5urH3LLbd4fdwbA5jpoH4IFZRWY2bSpEmx9gknnOD1UYuxqf1yFyHbunWr16d///5erV+/fl4tROiih24QV4Wgq1WrFrStsj5OAQA4XPEXDQAAAACJ40QDAAAAQOI40QAAAACQOE40AAAAACSuxITBQ6gVgytXruzV3JC1ClirEPmuXbu8mhu+VauTN2vWLGhbLhUuVq9x9+7dXk2txOxSIfKyZMSIEbG2+mzq1avn1VSg2v0s1Erh6v1UIWg33DxlyhSvT8OGDb1a586dvdo333wTa+fk5Hh9hg0b5tUUN7iuvhepqalB2woZ33Xr1g3aFgAAKJv4iwYAAACAxHGiAQAAACBxnGgAAAAASBwnGgAAAAASV6rC4GrVYhWedoOqderU8fqoEK8Kx7rbUs9Xo0YNr7Zt2zav5gZt1YrIISFvMx2Cd6nXWJZcdtllsfZ///tfr8+cOXO8mlop3n3fVfBbffbqPa5SpUq+21q4cKFXc1cBNzPbsGFDrD127FivTyi1SrpL3SQhZFt79uzx+oSuwO4G8UP2EwAAlHxl+5coAAAAgGLBiQYAAACAxHGiAQAAACBxpepi6Hnz5nk19xp2M/868/Xr13t9VE0tQrZ27dpYe9OmTV6fBQsWeLWVK1d6talTp8ba3bp18/qo/IDKcqi8yuHGzUJ8/PHHXp9ly5Z5tcGDB3u19957L9Z2F8ozC1uk7lCoRQLff//9WPuUU04p1H048sgjg/q537vmzZt7fY4++uigbansCwAAKP34iwYAAACAxHGiAQAAACBxnGgAAAAASBwnGgAAAAASV2LC4CGLy3Xu3NmrrVmzxqu5C/Sphfhq167t1VQodfny5Qdsm5kdd9xxXm3nzp1ebfHixbG2WpyvatWqXs0NkZuZ1atXz6u5yvqCfSEaNmzo1e66666gmkvdjOC7777zau6NBtSCjio8HRrELky33HKLVzv++OO9mvsdU6+xZs2aQc/JAn0AAJRN/BIFAAAAkDhONAAAAAAkjhMNAAAAAIkLujj6xwXj1GJ1Sdm7d2+srfISakEzlYVw++3bt8/rs23bNq+mnnP79u35Pp/aVsh+qYyGylWoheLcz0Jd5+6+p2bJLY724/OrxQSTVhTjL4RaTHHr1q1ezR0PKiOktlXYry/kO6ao1+juv7tIppnOGyWlKMffT5+nuMcgSgbGH4rb4XgMRslxMOMv6ERj8+bNZmbWqFGjQ9gtJOnVV18t7l0wsx/GRmZmZqE/hxnjD76iGH8/Po8ZYxBxjD8UN47BKE4h4y8lCjgd2bdvny1fvtzS09Plf4XH4SeKItu8ebNlZWUV+t2tGH9wFeX4M2MMIo7xh+LGMRjF6WDGX9CJBgAAAAAcDMLgAAAAABLHiQYAAACAxHGiAQAAACBxh+WJxn333WfHHnvsAfuccsopdtNNNxXJ/qDsy2/MDR482KpVq3ZIz3H55Zdbv379DmkbwKE62HGYk5NjKSkpNnXq1ELbJxxeGIMoDowjrVScaKSkpBzwf/fdd1/izzls2DB78MEHD9gnv0F1//33269+9Ssz++E1vP322wnvJYrKpEmT7IgjjrBzzz23uHel2HESXjqsXr3arr32WmvcuLFVqlTJ6tWrZ2eddZZNmDChuHcNhwnGIIoKY63kClpHo7jl5ubm/f8hQ4bYPffcY9nZ2Xm1tLS0xJ+zRo0aB/x3tYie65133rHbb789qV1CMXrhhRfshhtusBdeeMGWL19uWVlZxb1LwAFdcMEFtmvXLvv3v/9tzZs3t5UrV9rHH39sa9euLe5dw2GCMYiiUlbH2u7du+WCuKVJqfiLRr169fL+l5mZaSkpKbGaOtEYN26cdenSxVJTU61atWrWo0cPW7x4cazPK6+8Yk2bNrXMzEz7xS9+kbcojZn/X22bNm1qDz74oF122WWWkZFhV199tTVr1szMzDp27GgpKSl2yimn5PVfunSpzZo1y3r37m1NmzY1M7P+/ftbSkpKXtvM7Nlnn7UWLVpYxYoVrXXr1vbKK6/E9jElJcWeffZZO/vss61KlSrWvHlze+uttwr4TqIgtmzZYkOGDLFrr73Wzj33XBs8eHDs38eNG2cpKSn28ccfW+fOna1q1arWvXv32Mmwa+HChda8eXO7/vrr97uy5jvvvGOdOnWyypUrW/Pmze3++++3PXv25Lu/999/v9WuXdsyMjLsmmuuiZ0U79y502688UarU6eOVa5c2Xr27GlTpkyJPf7TTz+1Ll26WKVKlax+/fp2++235z3v5Zdfbp9++qk9+eSTeX9RzMnJyXefULQ2bNhg48ePtz//+c926qmnWpMmTaxLly52xx132Pnnn29mZo899pi1b9/eUlNTrVGjRnbdddfFVnz/8XK+0aNHW9u2bS0tLc169+4d+w8/e/futZtvvtmqVatmNWvWtFtvvdUbz6NGjbKePXvm9TnvvPNs4cKFRfNGoNgwBlFUQsZaSkqKPf/889a/f3+rWrWqHXnkkTZixIjYdmbOnGlnn322paWlWd26de3SSy+1NWvW5P37wY6jvXv32hVXXGFt2rSxJUuWmFn+x/Uff/Odf/75lpqaag8//HCSb1XxiEqZl156KcrMzDxgn927d0eZmZnR73//+2jBggXR7Nmzo8GDB0eLFy+OoiiK7r333igtLS362c9+Fs2YMSP67LPPonr16kV33nln3jZOPvnkaNCgQXntJk2aRBkZGdFf//rXaMGCBdGCBQuiL7/8MjKzaMyYMVFubm60du3avP5///vfozPPPDOKoihatWpVZGbRSy+9FOXm5karVq2KoiiKhg0bFlWoUCF6+umno+zs7OjRRx+NjjjiiOiTTz7J246ZRTVr1oyee+65KDs7O7rrrruiI444Ipo9e/ahvpUI9MILL0SdO3eOoiiK3n333ahFixbRvn378v597NixkZlFXbt2jcaNGxfNmjUrOvHEE6Pu3bvn9bn33nujY445JoqiKJo2bVpUr1696P/+7//y/t0d15999lmUkZERDR48OFq4cGH04YcfRk2bNo3uu+++/e7nwIEDo7S0tOjiiy+OZs6cGY0cOTKqXbt2bFzfeOONUVZWVvT+++9Hs2bNigYOHBhVr149b+wuW7Ysqlq1anTddddFc+bMiYYPHx7VqlUruvfee6MoiqINGzZE3bp1i6666qooNzc3ys3Njfbs2VPg9xaFY/fu3VFaWlp00003RTt27JB9Hn/88eiTTz6JFi1aFH388cdR69ato2uvvTbv31966aWoQoUK0RlnnBFNmTIl+vrrr6O2bdtGv/zlL/P6/PnPf46qV68eDR06NJo9e3Z05ZVXRunp6VHfvn3z+rz11lvR0KFDo/nz50fffvtt1KdPn6h9+/bR3r17oyiKokWLFkVmFn377beF8l6geDAGUVRCxpqZRQ0bNoxee+21aP78+dGNN94YpaWl5R371q9fH9WuXTu64447ojlz5kTffPNN1KtXr+jUU0/N28bBjKMdO3ZE/fv3jzp27Jj3my/kuG5mUZ06daIXX3wxWrhwYd7v1tKsTJ5orF27NjKzaNy4cfLf77333qhq1arRpk2b8mq33HJL1LVr17y2OtHo169fbDsHmpx69eoV/f3vf89rm1k0fPjwWJ/u3btHV111Vax20UUXReecc07scddcc02sT9euXWOTMQpX9+7doyeeeCKKoh8mtFq1akVjx47N+/cfTzTGjBmTV3vvvfciM4u2b98eRdH/O9GYMGFCVL169eivf/1r7DnccX366adHf/jDH2J9Xnnllah+/fr73c+BAwdGNWrUiLZu3ZpXe/bZZ6O0tLRo79690ZYtW6IKFSpEr776at6/79q1K8rKyooeeeSRKIqi6M4774xat24dO5F6+umn87YRRf53AyXTW2+9FVWvXj2qXLly1L179+iOO+6Ipk2btt/+b775ZlSzZs289ksvvRSZWbRgwYK82tNPPx3VrVs3r12/fv28sRNFP3w/GjZsGPuR51q9enVkZtGMGTOiKOJHXlnGGERRyW+smVl011135bW3bNkSmVn0wQcfRFEURQ8++GDefxz+0dKlSyMzi7Kzs+Vz7m8cjR8/Pjr99NOjnj17Rhs2bMjrH3JcN7PopptuKuC7UDKVikunDmTJkiWWlpaW978//OEPVqNGDbv88svtrLPOsj59+tiTTz4Z+1Or2Q+XQqWnp+e169evb6tWrTrgc3Xu3DlonzZt2mSffvpp3p/s9mfOnDnWo0ePWK1Hjx42Z86cWK1bt25e2+2DwpGdnW1ffvmlDRgwwMzMypcvbxdffLG98MILXt8OHTrk/f/69eubmcXG1JIlS6xXr152zz332O9+97sDPu+0adPsgQceiI3tq666ynJzc23btm37fdwxxxxjVatWzWt369bNtmzZYkuXLrWFCxfa7t27Y2OuQoUK1qVLl7zxNGfOHOvWrZulpKTk9enRo4dt2bLFli1bdsB9RslywQUX2PLly23EiBHWu3dvGzdunHXq1Cnv0r8xY8bY6aefbg0aNLD09HS79NJLbe3atbHxVbVqVWvRokVe+6fz5MaNGy03N9e6du2a9+/ly5f35sn58+fbgAEDrHnz5paRkZF36eiPlxKg7GIMoqjkN9bM4sfo1NRUy8jIyBtL06ZNs7Fjx8aOuW3atDEzy7s8KnQcDRgwwLZu3WoffvihZWZm5tVDj+uhvzVLi1J/opGVlWVTp07N+98111xjZmYvvfSSTZo0ybp3725DhgyxVq1a2RdffJH3ODdck5KSYvv27Tvgc6Wmpgbt0wcffGBHHXWUNWrU6CBfDUqaF154wfbs2WNZWVlWvnx5K1++vD377LM2dOhQ27hxY6zvT8fUjz/UfzqmateubV26dLHXX3/dNm3adMDn3bJli91///2xsT1jxgybP3++Va5cOcFXiLKscuXK1qtXL7v77rtt4sSJdvnll9u9995rOTk5dt5551mHDh1s6NCh9vXXX9vTTz9tZvEbXah5MtpPpmh/+vTpY+vWrbPnnnvOJk+ebJMnT/aeB2UXYxBFZX9j7UcH+t23ZcsW69OnT+yYO3XqVJs/f76ddNJJZhY+js455xybPn26TZo0KVYPPa6H/tYsLUr9iUb58uWtZcuWef/76d2iOnbsaHfccYdNnDjR2rVrZ6+99lqiz12xYkUz+yHw81PvvPOO9e3bN1arUKGC169t27berdcmTJhgRx11VKz20xOkH9tt27Y9pH1H/vbs2WMvv/yyPfroo7GJYdq0aZaVlWWvv/76QW2vSpUqNnLkSKtcubKdddZZsZsPuDp16mTZ2dmxsf3j/8qV2//Xdtq0abZ9+/a89hdffGFpaWnWqFGjvJsO/HTM7d6926ZMmZI35tq2bWuTJk2KHcgnTJhg6enp1rBhQzP7Ydy7Yxmlw1FHHWVbt261r7/+2vbt22ePPvqonXDCCdaqVStbvnz5QW0rMzPT6tevn3ewNfvhO/P111/ntdeuXWvZ2dl211132emnn25t27a19evXJ/Z6UPowBlFUfhxrITp16mSzZs2ypk2besfc1NTUgxpH1157rf3pT3+y888/3z799NPYcxTkuF7alYrb2x6sRYsW2b/+9S87//zzLSsry7Kzs23+/Pl22WWXJfo8derUsSpVqtioUaOsYcOGVrlyZUtNTbUPPvjAfv/738f6Nm3a1D7++GPr0aOHVapUyapXr2633HKL/fznP7eOHTvaGWecYe+++64NGzbMxowZE3vsm2++aZ07d7aePXvaq6++al9++aW8dAfJGjlypK1fv96uvPLK2J8/zX74M+0LL7yQ9xe0UKmpqfbee+/Z2WefbWeffbaNGjVK3jXtnnvusfPOO88aN25sF154oZUrV86mTZtmM2fOtIceemi/29+1a5ddeeWVdtddd1lOTo7de++9dv3111u5cuUsNTXVrr32WrvlllusRo0a1rhxY3vkkUds27ZtduWVV5qZ2XXXXWdPPPGE3XDDDXb99ddbdna23XvvvXbzzTfnTYRNmza1yZMnW05OjqWlpVmNGjXK9CRZGq1du9Yuuugiu+KKK6xDhw6Wnp5uX331lT3yyCPWt29fa9mype3evdv+9re/WZ8+fWzChAn2j3/846CfZ9CgQfanP/3JjjzySGvTpo099thjtmHDhrx/r169utWsWdP+9a9/Wf369W3JkiXc8vswwRhEUclvrIX4zW9+Y88995wNGDDAbr31VqtRo4YtWLDA3njjDXv++ecPehzdcMMNtnfvXjvvvPPsgw8+sJ49exb4uF7qFXNG5KCFhMFXrFgR9evXL6pfv35UsWLFqEmTJtE999yTF2b96R2AfvT4449HTZo0yWurMPjjjz/uPddzzz0XNWrUKCpXrlx08sknR2PGjIkaNmzo9RsxYkTUsmXLqHz58rHneeaZZ6LmzZtHFSpUiFq1ahW9/PLLsceZWfT0009HvXr1iipVqhQ1bdo0GjJkyAFfP5Jx3nnnxYL5PzV58uTIzKJp06blhcHXr1+f9+/ffvttZGbRokWLoijyx9zmzZuj7t27RyeddFK0ZcsWOa5HjRoVde/ePapSpUqUkZERdenSJfrXv/613/0dOHBg1Ldv3+iee+6JatasGaWlpUVXXXVV7C4c27dvj2644YaoVq1aUaVKlaIePXpEX375ZWw748aNi44//vioYsWKUb169aLbbrst2r17d96/Z2dnRyeccEJUpUqV2GtEybFjx47o9ttvjzp16hRlZmZGVatWjVq3bh3ddddd0bZt26IoiqLHHnssql+/flSlSpXorLPOil5++eXYOFZjcvjw4dFPDxu7d++OBg0aFGVkZETVqlWLbr755uiyyy6LBXE/+uijqG3btlGlSpWiDh06ROPGjYvdHIMgbtnEGERRCRlrJm7Ik5mZGb300kt57Xnz5kX9+/ePqlWrFlWpUiVq06ZNdNNNN+XdHKUg4+jRRx+N0tPTowkTJkRRlP9xXe1naZcSRQd5sSMO6MYbb7Q9e/bYM888k8j2UlJSbPjw4davX79EtgcAAAAUhTJ56VRxateunXeXKAAAAOBww4lGwq6++uri3gUAAACg2HGiUcJxZRsAAABKI24VAwAAACBxnGgAAAAASBwnGgAAAAASx4kGAAAAgMQFhcH37dtny5cvt/T0dEtJSSnsfUIpEEWRbd682bKysgp9VWjGH1xFOf7MGIOIY/yhuHEMRnE6mPEXdKKxfPlya9SoUSI7h7Jl6dKl1rBhw0J9DsYf9qcoxp8ZYxAa4w/FjWMwilPI+As60UhPT8/bYEZGxqHvGUq9TZs2WaNGjfLGRmEqbePv66+/jrXfeOMNr0+NGjW8WlpamlcrXz7+FV27dq3XR/0XJvXFnzFjRqy9evVqr8+aNWu82nvvvefViltRjj+z0jcGQ6xbt86rqdfmjsGSwr31t7oVeGH9l97SNP727dvn1dT74vYLfe927drl1ZYuXRprz5071+vTuXNnr1a3bt2g5yyoJUuWxNrZ2dlenzPOOMOrFfS/4oe+9wVxOB6DC/p+btmyxaupMTlnzhyvdvTRR8falSpV8vqsWLHCq9WpU8ertW/f/oD7aabnsZL4V6SDGX9BR5AfX2RGRkaZOcgiGUXxBSht4889YahYsaLXR01WlStX9mrujzz1OPUZVKlSxau5+1GhQoV8n89M//gsKYpqAi5tYzDE7t27vRonGgenNIy/4jjRcH98VK1aNd8+ZoU/14Tsl9qHknii8aPD6Rhc0PdT9UlNTfVq6rjpHs/VMVhtS/2Hw5D3rrScaPwoZN8IgwMAAABIHCcaAAAAABJXMv8mDpRi48aNi7Vnzpzp9VF/bly0aJFXc68tVRmK6tWre7XMzEyvVq1atVi7Vq1aXp+cnByvhpJN/al99OjRsfZ///tfr8/YsWO92sqVK73ajh07Yu1rrrnG6/Ptt996NXWZg3sNdJs2bbw+zz//vFfr0KGDV3O/Q+o7VdouQygM6vUW9BKUX//6115t586dXs29vESNqyeffNKrqX11L/Hr2LGj12f79u1eTV3yN3v27FhbXb41atQor7Zhwwavdv7558faF1xwgdcn5BK1/fWDL/R9crM3mzdv9vrMmzfPq02fPt2rucdSdbxV48OdN838+ejYY4/1+pTF+YnRDQAAACBxnGgAAAAASBwnGgAAAAASR0YDSNjWrVtj7WbNmnl91BoGakEk93re1q1be33UNdLqOmA3o6HW8lDbUrmNpk2bejUka/HixV7t5z//uVdzx5uZ2caNG2NtdW2z+vzVbRrd/XAzSGY6X6S4ayeoa6d/8YtfeDV1vfPVV18da99+++1eH3IbBb/t7x133OHV1q9f79WysrK8mnvLWzW3uWPUzCw3N9eruePh2muv9fp069bNq6k1Odx9VTk1ddtndRtcN/fkrtFhZvbb3/7Wq6nPAwW3cOFCr7Zs2bJYu0mTJl4fNdbU8c8dR+rYd8QRR3i1mjVrejU3y/HVV195fdT6MqUdf9EAAAAAkDhONAAAAAAkjhMNAAAAAInjRAMAAABA4giDAwlzFwJavXq118ddiM9Mh3rdWp06dbw+e/bs8Woq0OgGb1UoUW3rs88+82qEwQvf5Zdf7tVUGFctIOWGulX4VwWg1bbcmxmoRSNPP/10r5aRkeHVNm3aFGunpaV5fULD2u+//36sPWLECK/PxIkTg7ZVloUuEPfdd9/F2mqhURXqVgFa9z1Wz9egQYOgbbkh6zfffNPro8LaKujtjsm9e/d6fdS+qpobLJ8xY4bXR21fBYfdfqoPNLVYnhvgdheQNDNr2LChV3vllVe82vDhw2Ptc845x+tzxhlneLW2bdvmu1/qRitq8ckqVap4tdKEv2gAAAAASBwnGgAAAAASx4kGAAAAgMRxogEAAAAgcYTBgYS5YVm1+nHIas5m/urNKlyowrNq+25AUwUvVRhcBZCRvOeeey7WXrlypddHBVxDA60uNW7UTQS2bdsWa6tgohpvanyFhF5VrXLlyl6tdu3asbYbNDczGzp0qFe74IILvFpZVr582GH+448/jrXVGHLHgpn+bNQ84lLzYv369b2aezONd9991+tz7LHHejV1ww03aKteY4UKFbyaCtS73x/13Rk/frxXO+WUU/LdFvR77t6wwEx/zlOnTo211U0M1M0IFixY4NUqVqwYa7ur3puZLV++3KupG1G4NzZQq5qrkPqAAQOC+pVU/EUDAAAAQOI40QAAAACQOE40AAAAACSOEw0AAAAAiSMMfpDUqqT/+Mc/vNrRRx/t1dzVc/v27ZvcjqHEcEPdKuCoQoizZ8/2am4QWwUvlZBwoVpNVz1O7ReS98wzz8Ta6rNQwW/FDbSGhk3Vqtkhj1WBY7WvbrBSPU6t4qvCxW5YVIXI1Uq/h1sYPJT7PQ+9yYT7mZrpIK9LfV4qaOuOB7WafMjjzPzAtpqH1RyrbtSxY8eOWFt9d9Tq6ioMHhrYP5yo4LcbpjbTx7GWLVvG2tOnT/f6dOnSxavVq1fPq7mrd6uAv9rWl19+6dXcUPppp53m9VHfiwkTJni1Vq1axdodO3b0+pQU/EUDAAAAQOI40QAAAACQOE40AAAAACSOCwMP0hdffOHV1MJDU6ZM8Wp/+9vfYu1BgwZ5fZ544omC75xDXU/70EMPeTV3YbB//vOfXh+1iBH0omTuwmEqr6Oua1bX+G7YsCHW/v77770+asGijIwMr+Zey6oWf6tbt65Xy83N9WoofOo6d3UtuhqD7metrrcPWdTPzB+X6nFq7Krrzt1+IdkLM33dvLtwoHqce321mV5cKysry6sdbtzFw9TnpxalcxfBM/M/LzXfqXGkxqk7RtR+qcepa93dx6ptqe+T2lf3dat9cBcbRDj32GdmVqdOnaB+7jxz5plnen3UMVItBuk+VmXPVNZCjS13LK9bt87rk5qa6tXU9849Lh955JFeH5VnKg78RQMAAABA4jjRAAAAAJA4TjQAAAAAJI4TDQAAAACJO2zD4CrQo8JjLrVwSmZmpldTAXF3oZ4nn3zS63PppZd6teOOOy7f/VKBKLUw0Nq1a73atm3bYu2BAwd6fU4++eR89+FwpMJc6enpsXbt2rW9PiokqIK+7mejQrcqjNmjRw+v5gYa1XhXodvQxd4Q7oorrvBq7ufofvZmZkuXLvVqKtToLjylFi9T402Nr5BxE8p9bOgChCpMvGLFilh7zZo1Xh/3u2hm9umnn3q1AQMGBO1HWaHCpW5w1L2phZn+HNQNKtyFydS8osL76uYALjVuFRXqLujYdRfnM/Pnfvc1m+lF56C585/6nFXAWoWn3W2p4636TJs0aeLV3DGpFudr0KCBV5s1a5ZXc2+qo74Dod8Lt9+yZcu8Pm3atPFqxYG/aAAAAABIHCcaAAAAABLHiQYAAACAxHGiAQAAACBxh20YXIUeFTfwtWjRIq+PCtyoIJob2mzZsqXXp3Pnzl7twgsv9GqNGzeOtR977DGvT7NmzbyaGxI180N/NWvW9PpAW79+vVdzA5NqVVkVxlRBSzcsO3v2bK+PWtV4yZIlXq1p06axtruyspkOFrMqfPJuuOEGr/bhhx/G2mo8qOC/Gktbt26NtVWIUgVjQ+ZF1UfV1E0E3LGkgpwqOOyudG5mNnPmzFhbvTdqvz777DOvdriFwd1Vhc38mw+oeWvLli1eTd0Qo3Xr1rG2Cv2r8aH6ufuhwrKh48+l5jY1L37zzTdezR276nuobtICzb2Zg/qc1dygQt01atSItdXvMTVfqM/r+eefP+C2zfwbU+yPO6erMaPmavV9dbe1cuVKrw9hcAAAAABlFicaAAAAABLHiQYAAACAxHGiAQAAACBxh20YXAXdlNdeey3WrlatmtdHhZZUoMddlVsFHN0QnZnZBx984NXcsGfbtm29Pmo14I0bN3o1NwSoVphs166dV4MOoqmAq0uFwFTQslatWrG2Cj2qMalCczk5ObG2Cv2rcRu6Ei/CdezY0au537sLLrjA66OCt82bN/dq7s0A1Lyi5kA1bkJWa1ZhSzW/udtS3xW1IrUKYDZs2DDfPr/97W+92vHHH+/VDjcq3BzyPVer1avx4c4jar5T40/VQm/cEvK4kJXBVR81L7rBYXWjFTXHuvOwmX+jjsOReyxVx9bNmzd7NXX8C7mxgfp9pOasd955J9Y+5ZRTvD7q81O/tdzvivrtqELqKgx+7LHHxtqhgfTiwF80AAAAACSOEw0AAAAAieNEAwAAAEDiDtuMRqiHH3441s7MzPT6qGuK1TWd7gJC6hpEtchQo0aNvJp7/Wl6errXR13rp64/da+f/eKLL7w+vXv39mrQ1wGrBZ9c6tpMNbbUAn2u6tWre7W0tDSvduSRR8baalE/NSbV2ELhGzp0aFC/X/7yl15t9erVsbbKUKg8hrqW2V1ETc0h6nFqLnOvUVbzpPr+qNzYqFGjvBrCqMW9XOoadjcbaKYX+XSPKepzVnOgGjNuv4JmL8z8BfrU86k8iXovvvvuu1hb5aDU9qdOnerVyGj4mQZ1DFMZDdXPXcxOzX+K+s10xhlnxNrq95h6XMjigmpR1tC8m/vY0N+TBc08HQr+ogEAAAAgcZxoAAAAAEgcJxoAAAAAEseJBgAAAIDEHRZh8NBAzKJFi7yauwiKWpRHhXBU+M3tp/ZLPc4NY5r5iyupxaoUtX03yDlp0qSgbUF/hiGLQao+KgSmFvFztWzZ0qtNmzbNq7lhcBUwU4sMhQYtUTxC5gwVsA5dbFKNcZcaIyrs69bUttV8F7JooKL2Qc39xRGQLE4LFy70am4IWoVZ1YKRrVq18mru/Bb6+YV8XmpbIWPUzH+NaqypcLHq59bUGFKvJzs7O9/9LOvUwo/uTXVUUFr9RlNzlruIX+h3Xi1a6d4MJWReMwubx9TvABUsX7NmjVdzH6tu1OAuEm3mLwJcFPiLBgAAAIDEcaIBAAAAIHGcaAAAAABIHCcaAAAAABJXJsPgbphGreqpgkAPPPCAV6tdu3asrVZBDQ0HhQTWVIBIrVTqBpRUH1VTQSM3/DZu3Lj8dhP/PzWO3HCuCl2rcK67mvP++rlUGHPChAlezQ13qhsb5ObmejU1JlFyuCHKUOpzVaFudx5RIU0117irMJsVPFiuAp8hQm7McDhavny5V3NvDuAGas10iFcdX93Qa2jYvqBzTUE/Z7XvKhBcvXp1r+aOeXXMVzfzUHPs4SZkdXcVblZznfoMQ6ibCoSEs0OOyWb6s3fnP3Ujj3nz5nm1ZcuWeTV3/Kk50r2ZkRlhcAAAAABlBCcaAAAAABLHiQYAAACAxHGiAQAAACBxpT4MrsKFIWGdd99916sNHjzYq7mrLqswkgr0hKxGHvo4tUKrG4BSIT0ValPcoNuCBQu8PqNHj461VVALPwhZyVZ9pqqfCmS6jjrqqKD9clcJVWPNvfmB2eG3anJpo1ZrdufA0OCjCrSGrOocekMMN/Cpwr8qRF7QwCc0NWZUoN+ljlnqZheu0JWTQ24WoPqoY526qYA77+7cudPro+a7kJsRqPdv8+bNXk0F8Q836r1yx5Hqo+aBmjVrejX3WKc+UzWvqXHqfvbq96Uaf2oeC5lL1W8rdVzOzMyMtd2bOeyvVhz4iwYAAACAxHGiAQAAACBxnGgAAAAASFyiGQ117WRozaWuH1bX2YVcP/7HP/7Rqz344INerU2bNl7NvaZOXVcauuhPyOtW1/Cp62LdawnV9aiqFpIdUfmBadOmxdol5dq/4hZyvbB6r9SiZ2rMuwszKscff7xXC7nWXo0Pdc1yyDXYKD5r1qzxau7ComqhTnXduZrf3LEUmtkJySapbJlaFNVdVBSHRo0Hdy5TfdRYU+Mo5Fp0RY0td7/UGFXHLMV9rJpz1bFbzYHuHK7mebUtFkDV33v3fVFjQWWLQnI26jgd8rtK1dQ+qO+A+l3ovm41FtR+qYX3Vq5cGWurrEpJ+Z3GXzQAAAAAJI4TDQAAAACJ40QDAAAAQOI40QAAAACQuETD4AUNax+KESNGeLVbb7011s7Ozvb6HHPMMV5NBcPcQJwKJargmwrSuSGf0PdLhXbdEFFoiFeFj9zAmlqQxg1Aqec7HIUsOqUWHlq/fn2+jzMLW4wvZFE/M398hwY2WbCvaLiff+j7rgKMbjh248aNXh81btS2Qm5iERpqdLelwrKqFhL2DbnRxeEodK52g6Mq+H3sscd6NTWO3KCqCsaqzyYkjKsWQgtZbNAsbCFL9X7VrVvXq7kBYPV+hYaQ3f1Xr7EsUe+V+71Xc0rojUnc30xqfKjfe+omLa6Q+dZML+7sPqea61SoW/2edPdD7cPSpUu9WnHgLxoAAAAAEseJBgAAAIDEcaIBAAAAIHGcaAAAAABIXKJh8FBr1671amPGjIm1p06d6vUZOXKkV5s5c6ZXa9WqVaytVk5W4SAVuHHDQSHByP1xg2cqFKaoVSfdwJoKjKvtq7CTu1/qvUnyfShLQsZRrVq1vD65ubleTYUQGzVqlO8+qNXDVZjQ/ZxVqE2No5CAHIpPyOrDaoVYNUZCVoxWIU31PVDzjzu+1BhU3wMV3EQYdeMJxf1cQ8KsZmFBbPW40M+0oMeakJWf1XdHzXdbt271am4wed68eV4fFZ5Xz7lq1apYu0GDBl6fskSNGfd9Ue+nmi/q1avn1dzfheomPqGrZoeMUzVmNm/e7NWqV68ea3/11Vden8zMTK+mbkbg3nBBfU9U6L44MHsDAAAASBwnGgAAAAASx4kGAAAAgMRxogEAAAAgcYcUBh83blys/cADD3h91MqEbvDJzCwrKyvW3rJli9dHhaJPPPFEr+auOKrCV2pV0pBAT2h4LCMjw6u5ASgVbFKrcqt+7v6r1VlVaFPV3LCTeu+7desWa2/bts3rgx+sXr061g4J85vpsdWyZcsC7YMKv7nPqcaaCsipbSF5BV0ZXH1mbk3NUSoMrr777n6oOURRqy67AV217yrEu27dunyfj1XAtQ0bNng19b67xxk1xzdp0sSrqbnM/QwLuuK8mT/eQj9ndWMLl9qWmofVyubt2rWLtdVvHfW9U98fFTYvy9Tc474voatmq37u2A09hqnPwf3s1RypbpygPnv3d9qiRYu8PkcddZRX69Kli1cbNWpUrN2+fXuvj/qOzZ0716u1adPGqyWJv2gAAAAASBwnGgAAAAASx4kGAAAAgMQdVEZj1apVsQWdrr322ti/q2vS1GJlquZe06kWLVHbV9fuqmviXeqa0dAFzEKoRV7c/VLXh6prCdXiNu6Cb2rf1XWl6rrbkGv3TzrppFhbXZN4OFLjw130bNmyZV4fdf26+pzdxSdDqWtZ3Wu11UJ/akxy7XvJpq6bd7NllStX9vqoz1WNQbefuu5XXXeushbqen6X+m6oGsKE5grdY4P6rM466yyvNn36dK/mXoOvjk/qGKk+Z3c/1LbU+FPbcp8zdCFL9R4eeeSRsfZ///tfr4+6nj90QcCyTC0M6h5L1fjr2bOnVwv5rRWaK1Nzljv/hc5F6neoewx2x9D+qN/M7jFejSs1xxfHIn78RQMAAABA4jjRAAAAAJA4TjQAAAAAJI4TDQAAAACJO6gw+PPPPx8LGLsBKRXgVmErxQ30qAXvVGBKhXHdfioQo0I/KlTkBqPV84UsfmRmVrVq1VhbBczUoj8rVqzwavXq1Yu169ev7/VRgWAVAHZfk1rgqaDhKoQH/VTItkaNGgV6zoYNG3q1OXPmxNoqIKyCbiocieS5c4aaj9QYUTd4cOe3kAWy9idkYTU1l6m50h1fqk9oSDiE2vfD7eYG6linuO+Vepy6wYi6gYA7bx1KGNw99qjHhS6A6lLHSLV9dVx2g8lq0Vz13qgbvhxuN1dRwWX3fVHHJzWPqbEVQv1uC7kJkQqyq9+r33//vVdz97V58+ZBj6tdu7ZXc280oMZ7o0aNvFrIzZKSxl80AAAAACSOEw0AAAAAieNEAwAAAEDiONEAAAAAkLiDCoMfccQRsYCYG3J1w85mOryjQjhuAEsFoEODfW6ISAXYVOArJLAWsu9mOpzmhnxUAOyUU07xag8++KBXGz16dKyt3pvQAKgbKiqOlSPLEnccqbCuCoirz6t69eoF2oc6dep4tblz58baKvSvag0aNCjQPiB56vurvvvunHQoAWu3X2j4MqSfChyr74a6iQXChNwswMw/bqpjWGgY3D1+q3lMhWrXrVvn1dy5TPVR4WI1ZtauXRtrL1myxOujQt1qhW/3t4f6zdK+fXuvpkLO6r0oy9Sc5c4zKmCtgvQhNwlSc5E63qo5MeQGFmr7alvu2FLfsdWrV3s1FfTu0qVLrK2+5+7NjMyKZy7lLxoAAAAAEseJBgAAAIDEcaIBAAAAIHGcaAAAAABI3EGFwW+77bZYQMcNrXzyySfeY1RASq2+6IZpVOhHBdhUONvtp4I6qhayWrh6nBswU48zM7v55ptj7ZtuusnrE+qVV16JtdXK4GpfQ8J8ISuqYv9CgmgquKXCbyrkGEKteOtuS4139dmHriyMwqfmu5Dvecgq3fvjbl8F0tVNOUICkmo+UmNeBT5DsDK4/k6rQPXGjRtjbXUMCwlFm/njNPTGJGpf3d8Z7k0tzMxOOOEEr6ZuiOG+brUPmzdv9mrqvahXr94B22Zmbdq08Wrz58/3aofbMVfNR+5noYLStWrV8mpfffVVgfZBzT1qPLjzkZpT1I19VOhffX9c6revumlB69atY+3PPvvM66Neo7rhS2HjLxoAAAAAEseJBgAAAIDEcaIBAAAAIHGHdPH1U089FWurhXueeOIJr/byyy97NXcxu/Xr13t9UlNTvZpafMS9pk4tWqL2NWSRPbWtu+66y6vdeeedXi1J06dPj7XVNXzqOkiVDahdu3asvXLlSq+Pew3p4XZN6f6oa+bdayzVNZ1qgaasrKzE9qtp06Zezf3M1DWkChmNoqHGiSvJrEJo1sK9Pl1lO9S2QsZNyDXRZnreQhh1XXjIteLqc548ebJXU9fNL1u2LNZWn6naBzVm3DGink9d1662725LZdlmzpzp1dSCgx999FGsrX4/qCyMum5eHXMPd+q3lqKOY+7YVWNZjTX1m8mtqW2pDJI6xrvzmMorq6ym+r3qLv6n5lJFjb/Cxl80AAAAACSOEw0AAAAAieNEAwAAAEDiONEAAAAAkLhDSnm6CzKpQMwtt9wSVHOpxf+++eYbr6aCW4sXL4611QIlKmikQjLXX399rH377bd7fQpKLWilFgZS/vSnP8XaVatW9fqoAJ4KzbmhouOOOy7f5y/oQnJljQppueExFZxXwUH3czgUarEqN+irgr9qX1VADsXDXVTNLCzUHbpoqQqNq3ndFRq2dPc1NMCovmcIs2rVKq/WsmVLr+YeJ9XCdWpROnVDDPdYqoKxalyp8eduXx3D1BwVMpep4K264YIKHLvbV/uVnZ3t1dT34nBbRFJxj4mNGzf2+qiF8WbPnu3V2rdvH2uH3gwj5EYXatyq8aEC/u73Qv3eU9tXvxdCbrYRulBmYeMvGgAAAAASx4kGAAAAgMRxogEAAAAgcZxoAAAAAEjcIYXBQ4PLBXHaaacF1UqzQ3n/Bg4cmOCeoKBUSDUkPKvCaSrQH/I4FSRU4bSQIJoKR4auII5DU9CVwUPGROhcowLirtAwoQrjut8X9V0JDa4jTOjNKNz5Yc2aNV4fNdeoG6u4wWg1h4SMdzM/pN6sWbOgx4XMsWpc1a5d26up74/7GkND6uqmMyHB3rJE3Whg6dKlsfaxxx7r9XFv9GNmlpOT49WOOeaYWFvNWeo9V+PB/QyzsrK8PmvXrs33cWb++FPhdvU7QN3Qwf0uqtezevXqoP0qbPxFAwAAAEDiONEAAAAAkDhONAAAAAAkjhMNAAAAAIk7vBJIQBGoXr16vn1U4EuFKl0hYTUzs5o1a3o1NyymAo6hwXIUDxUGL+iq2RUrVvRqIUFvtXquGiNqrIaMJTUuVQDTDfuyurKWmprq1VQYt2nTprG2WoVehUu3bNni1dz5TT1Ofc5qX92QtQqyq5XHFfd1q8eFzotLliyJtdWNDVRNHR9CA+5lRbt27bya+x5kZmZ6fVToum/fvl5t27Ztsba6MYAKT6t+bnhfzZvqu5Kenu7V3PlbHbvV7wB1Ywb3Jg8/+9nPvD7qex5ys5qk8RcNAAAAAInjRAMAAABA4jjRAAAAAJA4MhrAIVDXHruL69SqVcvrs2PHDq8Wcv16aEZDXYfpXtusrjtW1/ura7CRvJCMgfos1HXF7jW9y5cv9/qoa6DV+HK3rzIa6rp2lfdwvxvq+dT16jNnzvRq7kJuIdmow9HRRx/t1VRGbPr06bH2ww8/7PVR17Wr6+bdOU9lIebPn+/VRowY4dXc7Igaa/PmzfNqajy4Y/fMM8/0+qgx6Y41M/81quv0v/rqK69WrVo1r9ajRw+vVpaphWFVzfXNN98EbV8tiuhSeTfFHW8q96COwWr76rjvUnO8mkvdjFDLli29PionUhz4iwYAAACAxHGiAQAAACBxnGgAAAAASBwnGgAAAAASRxgcOATt27f3an369Im1VXi2Ro0aXu3UU0/N9/lUEFKpV6+eV3PDYirgWLt2ba+mwqRIngraunr37u3VRo8e7dVycnJibbXgkwomqiCiG350F4oy0+NS3VjADaCrceou0GZm1rx5c68WEv5mET+9ONptt93m1T7//PNY+/zzz/f6qMXKknT33XcX6vYLkwqDDxo0yKv17NnTq4V89w836ripQt7qJiru3BayGK6ZvomKO4+p51Ofn7pRjHt8VYFxFYpX+x8Snlc3Ngj9DZEk/qIBAAAAIHGcaAAAAABIHCcaAAAAABIXdGHgjws2bdq0qVB3BqXHj2NBLRaWtJI8/tQ17e71oepaU3Wdu7pe033NauEetViQuv7efU51Db3aV7XYVnF/FkU5/n76PIX5ut3XEpovUGPQHUvbtm3z+rgLOJrpz9odS2rcqH1VY8ndlno+dQ2xeo0hn4UaH0nkNkr7+AsZD+q5CjujUZqp90vN80nNp2X9GKzmDzUPhMwz6riphGQ01PutMhrqeO7OPep7GLotNzunsiOFmdE4mPGXEgX0WrZsmTVq1OjQ9wxlztKlS61hw4aF+hyMP+xPUYw/M8YgNMYfihvHYBSnkPEXdKKxb98+W758uaWnp3MXD5jZD2exmzdvtqysrEK/iwHjD66iHH9mjEHEMf5Q3DgGozgdzPgLOtEAAAAAgINBGBwAAABA4jjRAAAAAJA4TjQAAAAAJO6wOdG477777Nhjj93vvw8ePNiqVat2SM9x+eWXW79+/Q5pGwBQ2PKbD83MTjnlFLvpppuKZH8AV0pKir399tv7/fdx48ZZSkqKbdiwocj2CcDBKzUnGpMmTbIjjjjCzj333OLelWLHD4CSKSUl5YD/u++++4p7F1FKFcfYGjZsmD344IMH7JOTk2MpKSk2depU+e/333+//epXvzKz/H844vCyevVqu/baa61x48ZWqVIlq1evnp111lk2YcKEoMd3797dcnNzLTMz84D9+A+A2J8VK1bYDTfcYM2bN7dKlSpZo0aNrE+fPvbxxx8n9hxNmza1J554IrHtlUZBC/aVBC+88ILdcMMN9sILL9jy5cstKyuruHcJiMnNzc37/0OGDLF77rnHsrOz82ppaWl5/z+KItu7d69cnKe47dq1i4W5SpiDGVtJqVGjxgH/XS1G5nrnnXfs9ttvT2qXUIZccMEFtmvXLvv3v/9tzZs3t5UrV9rHH39sa9euDXp8xYoVrV69evv9971793IrVuxXTk6O9ejRw6pVq2Z/+ctfrH379rZ7924bPXq0/eY3v7G5c+cW9y6WHVEpsHnz5igtLS2aO3dudPHFF0cPP/xw7N/Hjh0bmVk0ZsyY6LjjjouqVKkSdevWLZo7d25en3vvvTc65phj8toLFiyImjVrFv3mN7+J9u3bF7300ktRZmZmbLtvv/121LFjx6hSpUpRs2bNovvuuy/avXv3fvdz4MCBUd++faP77rsvqlWrVpSenh79+te/jnbu3JnXZ8eOHdENN9wQ1a5dO6pUqVLUo0eP6Msvv4xtZ9y4cdHxxx8f/X/t3XtclNW6B/AHFOQyg1cUEEQFNFAxL3nBiixJLbHM3bFOJ3Vn1q68VFt3Wabm7n45VttLW0/J0bKsFCtTPpqXJMQgE0RBBAPBgMy7CCXqc/7wwxzftR7iFV9hBn/fz8fPx/WwZnhnZs16ZzHvsx5PT08OCAjgZ555xvF7x40bx0Rk+FdQUHCZzyhcbep4qh6j69at4969e7OHhwdv2bKl1vEgjcvExES+9K2bkZHBt9xyC9tsNrbb7dy7d29OT093/Dw5OZlvvPFG9vLy4uDgYJ48eTKXl5c7fh4aGspz587lBx98kO12O48bN87y5wOsI40JyZYtW/iGG25gHx8fbt68OcfExHBhYSEz//98uGzZMg4NDWU/Pz8eM2YMnzp1ynH72NhYnjp1qqMtjRN1LoqNjXX0LyoqYk9PTz558iSHhoYa+oWGhjr6LVy4kDt37sweHh7cpUsXXrZsmeFxEBEvXLiQhw0bxl5eXtypUyf+/PPP6/TcgXM4fvw4ExFv3bq1xj5ExEuWLOG7776bvb29OTw8nL/88kvHz6vn1OPHjzPz/78vvvzyS46MjOQmTZqIY3TLli1X+dGBKxg+fDi3b9/ecC6sVj2mDh48yCNHjmRfX1+22+187733cllZmaNffn4+jxw5ktu2bcu+vr7ct29f3rhxo+PnsbGx2vi7FrnEo/7ggw+4b9++zMz89ddfc1hYGF+4cMHx8+oJp3///rx161beu3cv33TTTRwTE+Poc+lCIzMzkwMCAvj55593/Fw9eW/bto39/Pw4ISGBDxw4wBs2bOCOHTvynDlzajzOcePGsc1m4zFjxvCePXt47dq17O/vz88995yjz5QpUzgoKIjXrVvHe/fu5XHjxnHLli356NGjzMx86NAh9vHx4ccff5xzcnI4MTGR27Rpw7Nnz2Zm5hMnTvDAgQN54sSJXFpayqWlpXzu3Lk6P7dwddS00IiOjuYNGzZwfn4+Hz16tNbxYGah0a1bN/6v//ovzsnJ4f379/Nnn33GGRkZzHxxIvT19eV58+bx/v37OSUlhXv16sXjx4933L76g+Zbb73F+fn5nJ+ff/WeGLhiZhYaVVVV3Lx5c542bRrn5+dzdnY2JyQk8MGDB5n54nxos9n4nnvu4aysLN62bRsHBAQY5ippoaGOk7S0NMcfeUpLSx3jlpl5/vz5fPvttzMz8+HDh5mIeOnSpVxaWsqHDx9mZubVq1ezh4cHL1iwgHNzc/ntt9/mJk2a8ObNmx33Q0TcunVrXrJkCefm5vLMmTO5SZMmnJ2dfaVPJTSQqqoqttls/OSTT/Lvv/8u9iEiDg4O5hUrVnBeXh5PmTKFbTabY4xJCw0PDw+OiYnhlJQU3rdvH588eZL/4z/+g4cNG+Y4X176hz+4Nh09epTd3Nz4lVdeqbHP+fPn+frrr+cbb7yRf/zxR96xYwf36dPH8MeUjIwMfv/99zkrK4v379/PM2fOZC8vL8c8e/ToUQ4ODua5c+c6xt+1yCUWGjExMfzOO+8w88UJqk2bNoa/Slz6jUa1b775homIKysrmfn/FxopKSncsmVLfuuttwy/Qz1533bbbdogXL58OQcGBtZ4nOPGjeNWrVrxmTNnHLFFixaxzWbj8+fPc3l5OXt4ePDHH3/s+PnZs2c5KCiI33jjDWZmfu6557hr166GhdSCBQsc98GsfwAA51PTQmPNmjWOmJnxYGahYbfbOSEhQTyOCRMm8COPPGKIJScns7u7u+O9ERoaynfffXedHifUPzMLjaNHj/7pX4xnz57NPj4+hm8wpk+fzv3793e0pYWGOk4KCgqYiHjXrl3a74iLi+P58+c72kTEiYmJhj4xMTE8ceJEQ+zee+/lO+64w3C7v/3tb4Y+/fv358cee0x8bOAavvjiC27ZsiV7eXlxTEwMz5gxgzMzMx0/JyKeOXOmo11eXs5ExOvXr2dmeaFBRI4/slSrvtIAoNoPP/zARMSrV6+usc+GDRu4SZMmXFRU5Ijt3buXiUi7CuVS3bp143/961+OdmhoKM+bN8+S43ZVTp8MnpubS2lpaXT//fcTEVHTpk1pzJgx9MEHH2h9o6OjHf8PDAwkIqLDhw87YkVFRRQXF0ezZs2iv//973/6ezMzM2nu3Llks9kc/yZOnEilpaVUUVFR4+169uxJPj4+jvbAgQOpvLyciouL6cCBA1RVVUWDBg1y/NzDw4P69etHOTk5RESUk5NDAwcONFxbOmjQICovL6dDhw796TGD8+vbt6/j/2bGgxlPP/00PfzwwzRkyBB67bXX6MCBA46fZWZmUkJCgmEcDx06lC5cuEAFBQXicYFrKSoqMry+r7zyCrVq1YrGjx9PQ4cOpfj4eHr33XcNeR5EF5MU7Xa7ox0YGGiYLyVmx8mpU6fou+++o5EjR/5pv5ycHMP4J7o436njf+DAgVr7ct4j4HxGjx5NJSUl9NVXX9GwYcNo69at1Lt3b0pISHD0ufSc7uvrS35+fn86Rj09PQ23AZAwc619cnJyKCQkhEJCQhyxqKgoatGihWPuKS8vp2nTplFkZCS1aNGCbDYb5eTkUFFR0VU7dlfk9AuNDz74gM6dO0dBQUHUtGlTatq0KS1atIhWrVpFJ0+eNPT18PBw/L/6g/qFCxccMX9/f+rXrx998skndOrUqT/9veXl5fTiiy9SRkaG419WVhbl5eWRl5eXhY8QriW+vr6X1d/d3V2bFKuqqgztOXPm0N69e+nOO++kzZs3U1RUFCUmJhLRxXH86KOPGsZxZmYm5eXlUVhYWJ2PC5xHUFCQ4fX929/+RkRES5cupdTUVIqJiaGVK1dSly5daMeOHY7bXTpfEl2cMy+dLyVmx8n69espKirKcJIGUHl5eVFcXBy98MILtH37dho/fjzNnj3b8fPLHaPe3t5IAIdaRUREkJub2xUnfE+bNo0SExPplVdeoeTkZMrIyKAePXqY2ijjWuLUC41z587RsmXL6O2339Y+KAUFBdEnn3xyWffn7e1Na9euJS8vLxo6dCidPn26xr69e/em3NxcCg8P1/65u9f8tGVmZlJlZaWjvWPHDrLZbBQSEkJhYWHk6elp2L6vqqqK0tPTKSoqioiIIiMjKTU11fDhMiUlhex2OwUHBxPRxb/anD9//rIeOzgfM+PB39+fTp8+TWfOnHH0kbYS7dKlCz311FO0YcMGuueee2jp0qVEdHEcZ2dni+MYO0s1Dk2bNjW8rpfuFtWrVy+aMWMGbd++nbp3704rVqyw9HdXjyF1Pvryyy/prrvuMsQ8PDy0fpGRkdp2pikpKY7xX+3SBVJ1OzIy8oqOHZxPVFSUYa6zAs6XoGrVqhUNHTqUFixYII63EydOUGRkJBUXF1NxcbEjnp2dTSdOnHDMTykpKTR+/HgaNWoU9ejRgwICAqiwsNBwXxh/Tr7QWLt2LR0/fpwmTJhA3bt3N/wbPXq0ePlUbXx9fembb76hpk2b0vDhw6m8vFzsN2vWLFq2bBm9+OKLtHfvXsrJyaFPP/2UZs6c+af3f/bsWZowYQJlZ2fTunXraPbs2TRp0iRyd3cnX19feuyxx2j69OmUlJRE2dnZNHHiRKqoqKAJEyYQEdHjjz9OxcXFNHnyZNq3bx99+eWXNHv2bHr66acdC5yOHTvSDz/8QIWFhXTkyJFa/woJzsnMeOjfvz/5+PjQc889RwcOHKAVK1YYLi2orKykSZMm0datW+ngwYOUkpJC6enpjg9hzzzzDG3fvp0mTZpEGRkZlJeXR19++SVNmjSpIR4y1JOCggKaMWMGpaam0sGDB2nDhg2Ul5dn+Yfztm3bkre3NyUlJdGvv/5KJ0+epHPnztH69eu1y6Y6duxImzZtorKyMjp+/DgREU2fPp0SEhJo0aJFlJeXR//93/9Nq1evpmnTphlu+/nnn9OHH35I+/fvp9mzZ1NaWhrGsAs7evQo3XrrrfTRRx/R7t27qaCggD7//HN64403tAXqlerYsSPt3r2bcnNz6ciRI9o3wnBtWrBgAZ0/f5769etHq1atory8PMrJyaH33nuPBg4cSEOGDKEePXrQAw88QD/99BOlpaXR2LFjKTY21nEJaUREBK1evdrxB/D//M//1D6PdezYkbZt20a//PILHTlypCEeasNr4ByRPzVixAhDUuClqpN5MjMztaQwZuZdu3YZtn5Vt7c9ffo0x8TE8M0338zl5eVigmVSUhLHxMSwt7c3+/n5cb9+/Xjx4sU1Hm910tmsWbO4devWbLPZeOLEiYZdNSorK3ny5Mncpk2bOm1vy8ycm5vLAwYMYG9vb2xv66RqSga/dIwymxsPiYmJHB4ezt7e3jxixAhevHixIxn8jz/+4Pvuu49DQkLY09OTg4KCeNKkSY5Eb2bmtLQ0jouLY5vNxr6+vhwdHW3YIhrJaq7FTDJ4WVkZ33333RwYGMienp4cGhrKs2bNcmwooc6HzMzz5s0zbDsrJYNL42TJkiUcEhLC7u7uHBsby99++y0HBwdr/b766isODw/npk2bXvb2tgsWLOC4uDhu1qwZd+zYkVeuXPmnjx+c2++//87PPvss9+7dm5s3b84+Pj7ctWtXnjlzJldUVDCzvHlA8+bNeenSpcxc8/a2qsOHDzvmP8L2tnCJkpISfuKJJzg0NJQ9PT25ffv2PHLkSMcYqW1724KCAh48eDB7e3tzSEgIz58/X5s3U1NTOTo6mps1a3bNbm/rxmwiKwYAAMCEKVOm0Llz52jhwoWW3J+bmxslJiaiujMAgAtyvrLEAADgsrp3767tEgUAANcmLDQAAMAyjzzySEMfAgAAOAksNAAAwGnh6l4AANfl1LtOAQAAAACAa8JCAwAAAAAALIeFBgAAAAAAWA4LDQAAAAAAsBwWGgAAAAAAYDlTu05duHCBSkpKyG63k5ub29U+JnABzEynT5+moKAgcne/uutVjD9Q1ef4I8IYBCOMP2hoOAdDQ7qc8WdqoVFSUkIhISGWHBw0LsXFxRQcHHxVfwfGH9SkPsYfEcYgyDD+oKHhHAwNycz4M7XQsNvtjjv08/O78iMTqHulX8mqedu2bYZ2YWGh1mfs2LF1vn+rLFmyRIt1795dizljld1Tp05RSEiIY2xcTfUx/uqqsrJSi3l7ezfAkVjj3LlzWqxpU+crt1Of44/IecZgXWtKmJ1PS0pKtFhSUpKhfeLECa1PVVWVFrv55pu1mJm5THqM0vFbec64XNfq+APngXOweZ999pkW++6777TY0aNHtZg6t50+fVrr07p1ay02YMAALTZ16tQ/PU5Xcjnjz9QniOoJ3M/PzyUWGr6+voa29MHPGd4s0nGpx07kHMdak/o4udfH+KsrDw8PLYaFRv2prw+XzjIGr/ZCQzqJquP5999/1/o0adJEi9V1LnOFhUZ9/05nGX/gfK71c7AZPj4+WszT01OLSedzlXQ+lG7n5eWlxVzxuauNmfGHZHAAAAAAALCc8/6pUnD8+HEtNnr06Fr7SavN3bt3a7Hz589rMTXJ5cKFC1qfY8eO6QcrKCsrM7QPHz5c6+8jklfGaWlppn4nXF3Stxdnz541tNXXnYioffv2WszMX6ulS7WkvzBL/dSvhVu1aqX1CQ0NrfUYwPmZ+SvT2rVrtdjixYu1mDpO/P39tT7SvLhw4UIttn//fkP7oYce0vrU9S+0Zr8JAQDnJ80pZpPeW7ZsaWifPHlS69O8eXMtFhAQoMXOnDljaEvf1B44cECLbdiwQYu98MILhrZ07pa4+tyGbzQAAAAAAMByWGgAAAAAAIDlsNAAAAAAAADLOU2OhpnrzZ566ikttm/fPi0WERFhaEs7oqSnp2sxaZ9odReW4cOHa31SU1O1mHTtfnl5uaEtbQsmHWteXp4WS0hIMLTHjx+v9YGG8eijjxra6vagREQtWrTQYtJ1mM2aNTO0pW1EpWtZpfeTOpal20nbm4LzkF5XM69/YmKi1mfZsmVaTBpf6nXR6jXLRPL2jmFhYVps8+bNhnafPn20Pj179tRiV3K9NgC4HrPv7/z8fC2mzhfSPCNt092uXbtaj0PK5ZXyaKV8R7XUwowZM7Q+r776qhYzM+8783zovEcGAAAAAAAuCwsNAAAAAACwHBYaAAAAAABgOSw0AAAAAADAck6TDK6Skv9yc3O1mJRw89tvvxnaUoEpKaFHLSZFpBdn2bp1q6nbSWXqVVLyjlrsjYgoMDBQi6kJQ0gGdx579uwxtKUiQJI//vhDi5WWlhra6oYCRPJ7wM/PT4upCWvShgXg3KQNA8wkAUrF+dQCjkT6eCMi6tSpk6EtFZn67rvvtJhUlFLdbOC9997T+ixatEiLeXp6ajFXSoa0CjMbxoCzFu1Sx6l0nGaLkKnnaul1ruv9mz0GVy+Y5mzq+nwWFBRoMbUIHpF+/vvll1+0PufOndNiUqFb9TNZRUWF1kfaSEi6f7WQ4Pr167U+UiHBZ599VouZKSbtLHOicxwFAAAAAAA0KlhoAAAAAACA5bDQAAAAAAAAy2GhAQAAAAAAlnPaZPBnnnlGi0nJslKSoFpNWaq2LSXCSolAp06dMrSlZFwpsUmK+fj4GNpSQrqUaCkdv5qkvmrVKq3P6NGjtRhcfWVlZYZ2q1attD7q60ckJ42rSWydO3fW+khjWXpfqLGUlBStDzi3uiagXnfddVrMw8NDi0lzhppQKFXBHTx4sBaTNrY4fvy4oa1unEBEdPLkSS0mbehxLSaDu7m5/ekYyMrK0mLS6yydx/r27XtlB3cJM+PU7FiWzn/1fQxI/LaWmefzoYce0mIbN27UYm3atKk19uuvv2p9pA17pARudVOLn3/+WesjvZ+kz3Lqed9ms2l9Fi9erMV27NihxdasWWNoS/OfsySIN/6ZGQAAAAAA6h0WGgAAAAAAYDksNAAAAAAAwHJYaAAAAAAAgOWcJhlcTVpJTU3V+phNElSTwSVSsraUoKsm9kqkhJugoKBaf6eUfC7dl5RUpN52wYIFWh8kgzcMNQlWSmY0u7FBu3btar0vKYFNSvhSk3ilhLyDBw9qManyOLiWnJwcLXbs2DEtFh4ersX27t1raEuJ5dJ4liroqnOZ3W7X+qgbcBCZSwa/Fqo3V1RUGJJYP/vsM8PPv/rqK+020dHRWkyaH7Zt22Zod+jQQetz4sQJLSa9XhEREYb2b7/9pvWRXlOJ+jul87v0eKTNVtTjaNGihdZHOgeb+UwhjTVpQwRpvlbfP9LzpSZHnz59utZjciVbtmwxtL///nutjzquiOTXS90AQfpsJ51vpddQfZ4HDRpUax8iokOHDmkxNQFdmv/Ucz6RPH//85//NLSlCunOskGGcxwFAAAAAAA0KlhoAAAAAACA5bDQAAAAAAAAyzlNjoZ6LZl0fd7YsWO1WHp6uhZTr7uUruGTrt+UCrioxdbUglNERIGBgabu68yZM4a2dP2clI8h/U61QJZ6rS7UD+n1Onz4sKEtXess5VpUVVVpMfXaUqk4n3T9sFRASNW6dWstVlJSosWQo1E/1BwDKefA7DW3H3zwgaEdHBys9enWrZsWk+ZKdX6TrkeWrjtXr7kmIoqKijK0pcejFqIiIvr73/+uxdRrrKVjb2w5GuvXrzcUfs3IyDD8/KWXXtJuk5ycrMWSkpK0mJrDdf3112t9CgoKtJhUEFDNsZSKqklF1I4cOaLF1EK3Um7Hvn37tJg0v6m3lQocSnOslMuhzrtqjgsR0dGjR7WY9LyqeU/qZwUiory8vFr7uLLly5cb2tJnKCnnRaK+76VzpHQOlvqpnxWl8S7d11//+lctVlxcbGjv379f6yPltrVs2VKLSXkbzgrfaAAAAAAAgOWw0AAAAAAAAMthoQEAAAAAAJbDQgMAAAAAACznNMngZixbtkyLSUXpNm3aZGhLyVdSsTwpMVFNMJQSzKSEQykZV00clpKdpOJHM2bM0GJPP/20FoP6JxU9U19XKZHLbLElM4Wi1CROInkcqccVEBCg9ZGKYkL9UOcRacMKaY7avHmzFtu5c6ehLSW4SvOPdP9+fn6GtjRG1E0ziIji4+Nr7ScVtZJiU6dO1WLvvvuuoS0de2Mr4hcYGGjYIEJNQv3xxx+126SlpWmx5s2b1xqTkptjY2O12C+//KLF1HP1sGHDtD6FhYVaTEqqHTNmjKGtbrZBJCfQSnOz2k9KqI2JidFi0nlfTeSVNm2R3mPq+4lIL9AnJfCrycVmNvxwJepmKNL8J809YWFhWqyuxQylTS3UmHRc0pwibVCg3pe0IYJUXFBKQFcTy50ZvtEAAAAAAADLYaEBAAAAAACWw0IDAAAAAAAsh4UGAAAAAABYzmmTwaUqr1Ky36pVq7SYmmR2ww03aH2kBKI//vhDi6nJhFIikHSsUhKiKjs7W4tJSUxqZVRwHlLCoZp4K1X8lkhjS2U2uVXqpx6XlHQmVeaFhiElxkq2b9+uxdRKxtKmAlISb/fu3bVYbm5urX2kxFQpgVGtEC1VmlYrkRPJmxuo7z0pIV2am80+r84oLy/P8D5WX0MpQVR6vQ4cOKDF1PPm7t27tT6DBw/WYmVlZVosPDzc0JYqZNtsNi3WoUMHLaZSK8ITEYWEhGgx6fyqPl/SpjCSdu3aabGvv/661j7Sc5+fn6/F0tPTDW3pc4B6rGaP3VWo5x7p856UPB0UFKTF1PlOSvKW5gHpvKmel6U5RRqT0ntR7We327U+e/fu1WJdu3bVYurrr1aOJyKKiIjQYg0B32gAAAAAAIDlsNAAAAAAAADLYaEBAAAAAACWw0IDAAAAAAAs5zTJ4GoSjpQIJCXLSgk9asKhlKgoJf1IMTXxR0q8lZKDpGNV71+6HRK/Gx+1IjyRnCgrUTcokJLapDEjjWX1vSLd19mzZ00dF1hPfc3MVrCWEqWlmEpKxpUSWouKigxtqQqzdKzSxgJq9WRpnpeOXRqXGRkZhvatt96q9WlsyeAtW7Y0vN/VKtkBAQHabaTEb+l5qet9rVmzRov17dvX0JYSY3v27KnFpCr36oYBPXr00PqoydREcoXvrVu3GtrqpglERD/99JMWk8aMeo6XKp2rFb+J5CRu9Tik+VvdVMTsJiOuwkw1b2kekDYjUD8DSsnaZjZfIdI3UZHOm9J9Sb9TjUljQfq8IM2vaj9pcw8kgwMAAAAAQKOFhQYAAAAAAFgOCw0AAAAAALCc0+RomLke2ew1y1JxKpV0faNUsM/Ly8vQNlPQxezvbNq07k9/Xa/nBmtJ14yq151Lr7N0fbx0baZ67a5UuCctLU2L+fn5aTF1jEjXx7vy9euuTr1uXnotpGu+pZyJjh07GtrS9budOnXSYtK17uq4KS0t1fpI181L1+W3bt3a0Jaud5YKVkn5AllZWYa2lKPR2ObFiooKw9yvvoY33XSTdpukpCQtJl0bHhkZaWhLc4hUMO3JJ5/UYmquhZSvs2nTJi02aNAgLaY+Jmks33HHHVosMzNTi+Xk5Bja999/v9Zn2LBhWkzKv1BzTHbs2KH1kQq6SqKiogzt6667Tuuj5k81tnxOtTCov7+/1kf6jCZRPx9Jt5M+A0rzhXqeNJvHKM1t6nFJc7zZArwqdT4kIrrllltqvV19wDcaAAAAAABgOSw0AAAAAADAclhoAAAAAACA5bDQAAAAAAAAyzlNMrgZZhP71GJoUuEUKVHHTBE1KRFIStSREoDVZM/Glsx1LZKKQUrjTSUlgUljUt3YQCq0JSUDS4Wo1PFttigm1A8zCX9fffWVFpOSJtVNA6T5SEqQVJNSifRCbtKYl5JepflN3VxDKtJ15swZLSYlJkuFulRXsuGGMzp8+LDhOVST69UihkRyAUTpXHfy5ElDW3p+pQTr2267rdb7VxN9iYjeeustLSaNmeXLlxvaUjL4X//6Vy0mJcJu2bLF0JY215AS5b/44gstduLECUM7PDxc6yNt8FFSUlLr75Teh+p7RXofugppvlAfT1BQkNZHmrOkc5Y6h0ivgzSnSP3U+5fOm9K5W6ImoJvZTIZI/uyrxnbu3GnqGBoCvtEAAAAAAADLYaEBAAAAAACWw0IDAAAAAAAsh4UGAAAAAABYzmkz5aTEyLpWeZUSvtTENyI5EU1N8pGScaVkJCnZV+3XvHlz/WDBpUjJY2oCqtkK3FJCWZs2bQxtKUFOIlU9VUmJslJiOdQPM/ObVBlcmiu3bt1qaEtjMDQ0VIupCa5EelKw3W7X+khVaaUNCdTHKCWFSvOir6+vFlOTR6VEUXUzBVd3/fXXG56LNWvWGH4uJSQHBgZqse+++06LqUn/UsVvqTL466+/rsXU5/3NN9/U+kjV3t99910tplYVlzbbSE1N1WLx8fFabMqUKYa2+j4hkpPg1SrgRPrniq+//lrrU1xcrMW6d++uxdSkYCnpfsCAAYa2tGmCqygqKtJi6mcrs5/3pHOduhmBdL41u1GEOndK8630udDMfUmk4zKzUYj0nDoLfKMBAAAAAACWw0IDAAAAAAAsh4UGAAAAAABYDgsNAAAAAACwnNMkg6vJLnVN/JZI1Wel5CAp4UZNPJOqcUrJkVJir5psLiUGHT9+XIu1bNmy1mO18vkC86TKnmZIlXnNjD9pLHh7e9fpGCR1fTxw5aQ5SbVnzx4t1rt3by2mJuPu379f6yMl1QYHB2sxdW6REmNtNpt+sIKQkBBD+9ChQ1ofaYMF6blR58+8vDytj5R468p8fHwMyeDr1683/Lxbt27abe6//34tdvTo0Vpj6mtFRLRixQotJlUeP3jwoKGtJjITEYWFhWmxBx98UIutXr3a0JYSb6X3QEFBgRZTNwyQzrfSuVR6vnr16lVrH+n+hw8frsWWLl1qaEvvAfX8YCZB2FlJmwqo50TpdTC7SZAakz5rSZv4SDEzz7N0DNJrqD5G6XwubUYgVatXf6c0nzsLfKMBAAAAAACWw0IDAAAAAAAsh4UGAAAAAABYzmlyNMzkGEjXZkrX3n3wwQeGtnTdnVRMSroOTr1/6fdJBVakIjJqjoZ0Ld6MGTO02Pvvv1/rcUHDkMaWVDhMJY13KT9CvaZTKkAm5QOZKSIpjVEzxw71Q7q+W8qFkK53VgvoSXkVUsG0n3/+WYup1yhLRR3btWunxaTigup181LOmzR29+3bp8XUOTA9PV3r09hyNPLz8w05WWpugnReyM7O1mI33XSTFlPng5SUFK1PdHS0FvPz89NiOTk5hnaHDh20Ph999JEWy83N1WJq4T1pzHz//fdaTMqDu/766w1tKb/N399fi0nFfL/55htDu0uXLlqfp556SotJ+VLqmJfOD2o+kysXV5XmBuk8Zob0OqvPn9miytL7p675r9LnQvU4pHFlJj9HOi6pCLWzwKdVAAAAAACwHBYaAAAAAABgOSw0AAAAAADAclhoAAAAAACA5ZwmGdwMs0k5mzZtMrSlpB8pOUiiJuZIRVGkpFopcV2NXVp4qdrOnTtNHRc4B2kcqa+zlMglJZ1JidhqQR8pgddMEnlNx6GSxjI0DOl1lYqj3X777Vrs8OHDhrY0tqTifNImGWqyeX5+vtZHSmA8cuSIFgsNDTW0zRSiIiKKiorSYmqBNGlzjcYmLCzMcN5Qnz9pfujatasWW758uRZTn+PIyEitz0svvaTFBg4cqMXU12LdunVaHykhuLi4WIupyd9eXl5an48//liL3XXXXbUeV1FRkdZHSm4vLS3VYiNHjjS0pfdYYmKiFuvfv78W69Onj6G9Zs0arY+abC4lxbsKaaMLaeyqpAJ30u3U+chscUPpc5v6+dHs51Cpn3r/0pjp16+fFjt27JgWU88PJ06cMHVcDQHfaAAAAAAAgOWw0AAAAAAAAMthoQEAAAAAAJbDQgMAAAAAACzXKJPB1eRF6XZSEq9UyVFNIpISNKVqzdLvNJNUJCXxmmG2ajpcfeprKL3O0msjJfe1b9/e0A4PD9f6SGNSuv8zZ87oB6uoa3VWsN6qVau0mFQZXHr91df6hx9+0PqsX7++1tsR6Qm6M2bM0PqsXLlSi0lVl9XNLqRqtkOGDNFip0+f1mK//PKLoa0mmjdGVVVVhg0b1ArfUlL+li1btNiPP/6oxYKCggxtKem6c+fOWkyq5q2S5sBbb71Vi0mbHahJ49L5tkePHlpMSqpVk+elZFxpEwPp80JISIihnZeXp/WRksGlJPhRo0YZ2mqiuXQ7M/O5s5I2gVBfC2nMNG/eXItJldzVMSlV4JY+M0lJ42rMbPVwqZ/6GVB6HiIiIrSY9L5Q3+vOfO7GJ1EAAAAAALAcFhoAAAAAAGA5LDQAAAAAAMByWGgAAAAAAIDlnDYZXErKMZsMrlbxlJK8paQ5qVK3mgRmNulHOn71vqSK5VLSDxK9nZeUTKiOEWn8SRXmpWQuNUGzbdu2Wh8pKVDaVEAdb1IfVAZ3HhUVFVpMSgbfs2ePFgsMDDS0d+3apfWR5jIpaVJNCpaq80pjSUp0VOdAaZ6UKo+3a9dOi6kJwFIybmNTVlZmeI3UKtbSeUGq9i4lT6v3tWzZMq2PutEKEVGrVq20mLoRQEpKitZHOv9JVbPVitjSuJo8ebIWUzceINKryffq1UvrIyVrFxYWarHNmzcb2sOHD9f69O7dW4tJFZzVc7yaaE5kvrq1K5A2GlDnBul8eN1112mx1q1bazF1YxUpiVxK8DdTzVt6j5mNqfcvnW/VuZuIKD09XYuZ+Wwq3X9dNxy6Evi0CgAAAAAAlsNCAwAAAAAALIeFBgAAAAAAWK5R5mio16ybvT5PupbezHFJpGM1c/zSddlSUSvpOmmof1KOhjpGzOTrEMmvs91uN7SlHA3pOkwz7x/pGlhp/EHDkF4LqTiflG+2b98+Q1u6Ht7sXKbOi9LtzBa/MjMHSgXTpGvW1Xw2qeBlY2O32w25hIcOHTL8vKysTLtN3759tZia+0VEdODAgVr7dOzYUYtJ+QtqUb3BgwdrfaSxLF2Df+zYMUNbygmRckek+1ev5z948KDWR7p/KUdIzbWQ8lC6du2qxe644w4ttn//fkNbeg/ceeedhrYrj3cz+QRSHynfyExehfTZzmzuq3SuNkO6L/U4pPxKKedEKlKpFjGV8l5KSkq0WEMUNsU3GgAAAAAAYDksNAAAAAAAwHJYaAAAAAAAgOWw0AAAAAAAAMs5bTL4lWjfvr2hrSbNEMnJQVJSkZnEXimp1sx9mS2woibDESEZ3FlI40F9DaXxIZGS+6TkS5VaaItIThBWi7FJyXB1TXwD60mJfDExMVpMKsCUlZVlaEvzitk5UCWNebMJ4mpMSlKXjlUttEakF2STEiulmFSY1VW4u7sbkkzVzShSU1O120iFDKXXRk1uHjVqlNZHmo+2b9+uxdSCgFKBQGkjjSVLlmgxdXy3adNG6yPNncOGDdNiamL866+/rvXZu3evFps4caIW69mzp6H96quvan3U4sFE8ucRNak/IiJC66NufiCNbVdhJjlbOhdJxezMfP6Sfp80z0jzUW3HSVT3In5S8Unps11kZKQWW79+vaGtFskkIjp+/LgWQzI4AAAAAAA0ClhoAAAAAACA5bDQAAAAAAAAy2GhAQAAAAAAlnPaZHCzVWulhCE1wUZKcDSbwG2mmrJZ6rGardYrJZSFhYXV+Tjg6lJfZ2nMSImQUsJa586da/19UiVbKTlSqqoKzkOtDC+9htL8oFZ0JpIrC1vFbDK4RE2klBLZ1aRXIjlpcsiQIYb2hg0btD7Sc+jKyeBt27Ylm83maKsJoFLSqDTXqInfRHrF6tjYWK3Prl27tNjAgQO1mDpvSVXupeOSks3Vqt/Sayrd12+//abF9uzZY2h369ZN6yNVZpYqjxcUFBja0jlZShyWxrf6OePS17im45KqRbsKT09PLaY+B9Jrqm70QyRvdqDev5SYLX3eM9NPOi7pvqS5Tb1/6fOrdF/S+0JNZpeOXT2nNBR8owEAAAAAAJbDQgMAAAAAACyHhQYAAAAAAFgOCw0AAAAAALCc0yaDm6lQSyRXQDZTzVtK0jKb0FjX26n9pOQd6bikREhwDlLCl5r8LSXtma1WbyZxVUrylhIm1cQzqTKqNCahfqjJq1LiY25urhaTNhtQK+ju27dP69O8eXNTx6XOxWaTKM3EpOrGZWVlWkw6fn9/f0NbStLMzs7WYu3atdNiriIvL498fHwc7U8//dTw86CgIO02drtdi0nVtVesWGFoS5sMSBW+1aRoIr3S9e233671kRLLpQrwUmK0SqqAnJ+fr8XUhGqpCrg050oJ4hkZGYb27t27tT5+fn5aTBrz6twvJTjv2LHD0HblzwXSuU6dV86ePav1CQkJ0WLq60Ckb5Igfa4y+xlTJR27REr0Vj8vSNW8pc8GEukxqa7mpiCXA58qAAAAAADAclhoAAAAAACA5bDQAAAAAAAAyzltjoZZZq71k66VM1ssz0z+hdnif+o1dWavdfby8qr1GKBhXHq9dDV1zEjFqqTXXrqWWhqnKvV6fCL5mnb12mBpbJu9/hSsp16fXlxcrPWRCjhGRERoscTERENbyhMyWyzKzO3MXgOtFoqTipxJj0d6b6jXN0v5UnXNu3NWdrvdMOeouQ9SzqJapI5IHg/9+/evtY80l0kF6NTXYufOnVofs7llKmksSIX3pPOyVPxWJRXnKyws1GLq+6BDhw5aHynnRCpWpxZkkwq0de3a1dCWcj1chZRPpZLmGbPzmJnzmHRulcaMOodIt5PmOjNzj5SjId3OzLFKz42Z91N9wDcaAAAAAABgOSw0AAAAAADAclhoAAAAAACA5bDQAAAAAAAAy7l85qeUAKMmAkkJcnUt1iKp632ZTVSUiv7U9b7g6uvUqZOhLRXGk4otSYm+ZkiF3aRCVOo4lRLmsPFAw1EL9knJslKiqjSW1CRDKSnQ7JxhpjCUREpgVO9r/PjxWp8RI0Zosbi4OC0mJd+qpOROV3bq1CnDY1KLFkpzzbfffqvFevXqpcX69etnaEtF/ZKTk7WYVPhRTRqXCuqNGjVKi0lJ40VFRYa2tGGK2UKF6gYL0ucH6TmU3otqYTU1WZtIfm7Wr1+vxW677TZDWypWpyaku3LBPimRXd1AwGwxTzNFbSV13fxH+rxnNhlcjUkbWEjvFWkuVceutMmA9Nm3IeAbDQAAAAAAsBwWGgAAAAAAYDksNAAAAAAAwHJYaAAAAAAAgOVcPhlcoibJnDx5UusjJZTVldlKjmoFS6mipXRcUlKeysrkdjCvoKBAi6nVYFu1aqX1USskExHFxMTU6RikpFhpbKmJYWoyI5FcARfqh5o4Kr2uUnKfNJbU19ZsAqM0/7Rt29bQLikp0fqYrfKszmXz5s3T+jz//PNarGfPnlosPDzc0JaSnqW535Vdd911ZLPZHG01YVba4OHee+/VYtL8kJ2dbWgHBgZqfaSY9NqsXbvW0FaT1onkjQ2kjU+6d+9uaLdu3VrrIyVwS+8VdeMM6fFIxyWdz9UxryaaE+nvHSKiyMhILXbo0CFDWzqvjBkzxtB25crg0mcmNXlf3QSASB63UjL4pe8RInlek15TibqBhXRfZudXlTQ+Tp06pcWkMaMmf0u/z0wF9vqAbzQAAAAAAMByWGgAAAAAAIDlsNAAAAAAAADLYaEBAAAAAACWa5TJ4FKyosrHx0eL1bW6ttnbqck6UlKRlMwnHWtdjwGsJVWDVSuDBwQEaH1+/vlnLXb99dfX6Riio6O1WMuWLbWYmmwsJdYNHTq0TscAV06t9CslBUpVY6WEZzWRXEqslJLIpTGhVik+duyY1kfdAKGmY1XnN6mardkKy7m5uYa2VFG8rlWDnVW3bt0MCdM9evRowKOp2dixYxv6EBo9aX5wZWoyuJqETUQUFhamxTZs2KDF1DlRqgB/7tw5LSbNf6or2XhHTUCXjkH6bBAbG6vF1LlUui+pwnxDwDcaAAAAAABgOSw0AAAAAADAclhoAAAAAACA5Zw2R+NKroNTC/qUlZWZup1UYEqNSQVQpJiUa6Hy9vbWYmauEZSgYF/DkK4Ll2JXk3T95tatW7WY2QJF0DDU64h37typ9ZGKPwYHB2uxFStW1Pr7MjMztZiU36bmX3Tr1k3rEx8fr8WkuUy9Dlu6L7UQX033dc899xja0rH36dNHiwFAw/Lw8NBiBw8eNLSlHA01/5FIzitMTk42tKXPWtL9SzE1/1U6j5otAK32k/LrpMK9ERERWkwtSCnlzh05csTUcV1t+EYDAAAAAAAsh4UGAAAAAABYDgsNAAAAAACwHBYaAAAAAABguUaZDN6+fXtD+/Tp01ofqQielGipFpQ6c+aM1kdKbJIK6KmJQFLBLDXBh4gMBZrAuUgFcaQiZHWljgdpkwEpZibxW0qwlYr+SMWOwHpqUv8777yj9ZHmmjfffLNOv69nz56mYmb06tWrTrczSxrP6jwvzd9xcXFX7ZgAoG6kDVO+/fZbQ1tKzG7btq0We+yxx0zFGpuRI0ca2tL5fPTo0fV1OH8K32gAAAAAAIDlsNAAAAAAAADLYaEBAAAAAACWM5WjUZ1vcOrUqat6MFapqqoytKVr18xen67mipjpQ2QuR0O6L+m4pOddvWZZup7RbBGZuqg+JulxWs2Zx5+z5miY4co5GvU5/i79PfU5BtV5jEgeb874vqgP6nNRn8/NtTD+wLm58jlYyptV5zvpM430+6/m5xxnps530vMlFaG26jW8nPHnxiZ6HTp0iEJCQq78yKDRKS4uFisTWwnjD2pSH+OPCGMQZBh/0NBwDoaGZGb8mVpoXLhwgUpKSshut1/RblDQeDAznT59moKCgq76XxQw/kBVn+OPCGMQjDD+oKHhHAwN6XLGn6mFBgAAAAAAwOW4Ni9uAwAAAACAqwoLDQAAAAAAsBwWGgAAAAAAYDksNAAAAAAAwHJYaAAAAAAAgOWumYXG+PHjyc3Njdzc3MjDw4PatWtHcXFx9OGHH4qFTgAuV/X4qunfnDlzGvoQAQzKyspo8uTJ1LlzZ2rWrBmFhIRQfHw8bdq0ybLf0bFjR3rnnXcsuz9oXC49N7u5uVHr1q1p2LBhtHv37oY+NGjkMP/Vj2tmoUFENGzYMCotLaXCwkJav349DR48mKZOnUojRowQKyITydV5ASSlpaWOf++88w75+fkZYtOmTXP0ZeYax1xDkyosQ+NTWFhIffr0oc2bN9Obb75JWVlZlJSURIMHD6YnnniioQ8PriHV5+bS0lLatGkTNW3alEaMGNHQhwWNGOa/esTXiHHjxvFdd92lxTdt2sRExEuWLGFmZiLihQsXcnx8PPv4+PDs2bOZmXnNmjXcq1cvbtasGXfq1InnzJnDVVVVzMx84cIFnj17NoeEhLCnpycHBgby5MmTHb9jwYIFHB4ezs2aNeO2bdvy6NGjr/rjhYa1dOlSbt68uaO9ZcsWJiJet24d9+7dmz08PHjLli38+++/8+TJk9nf35+bNWvGgwYN4rS0tBrvh5k5MTGRL33rZmRk8C233MI2m43tdjv37t2b09PTHT9PTk7mG2+8kb28vDg4OJgnT57M5eXljp+Hhoby3Llz+cEHH2S73c7jxo2z/PkA5zN8+HBu3769YSxUO378ODMzHzx4kEeOHMm+vr5st9v53nvv5bKyMke//Px8HjlyJLdt25Z9fX25b9++vHHjRsfPY2NjmYgM/wAuJZ2bk5OTmYj48OHDzMz8j3/8gyMiItjb25s7derEM2fO5LNnzxpu889//pP9/f3ZZrPxhAkT+JlnnuGePXvW06MAV4P5r/5cU99oSG699Vbq2bMnrV692hGbM2cOjRo1irKysuihhx6i5ORkGjt2LE2dOpWys7Pp3//+NyUkJNDLL79MRESrVq2iefPm0b///W/Ky8ujNWvWUI8ePYiI6Mcff6QpU6bQ3LlzKTc3l5KSkujmm29ukMcKDe/ZZ5+l1157jXJycig6Opr+8Y9/0KpVq+h///d/6aeffqLw8HAaOnQoHTt2zPR9PvDAAxQcHEzp6em0c+dOevbZZ8nDw4OIiA4cOEDDhg2j0aNH0+7du2nlypX0/fff06RJkwz38dZbb1HPnj1p165d9MILL1j6mMH5HDt2jJKSkuiJJ54gX19f7ectWrSgCxcu0F133UXHjh2j7777jjZu3Eg///wzjRkzxtGvvLyc7rjjDtq0aRPt2rWLhg0bRvHx8VRUVERERKtXr6bg4GCaO3eu4y/WAH+mvLycPvroIwoPD6fWrVsTEZHdbqeEhATKzs6md999l5YsWULz5s1z3Objjz+ml19+mV5//XXauXMndejQgRYtWtRQDwGcHOa/etbQK536UtM3GszMY8aM4cjISGa++I3Gk08+afj5bbfdxq+88oohtnz5cg4MDGRm5rfffpu7dOmi/YWFmXnVqlXs5+fHp06dsuBRgKuo6RuNNWvWOGLl5eXs4eHBH3/8sSN29uxZDgoK4jfeeEO8H2b9Gw273c4JCQnicUyYMIEfeeQRQyw5OZnd3d25srKSmS9+o3H33XfX6XGCa/rhhx+YiHj16tU19tmwYQM3adKEi4qKHLG9e/cyERm+dVN169aN//WvfznaoaGhPG/ePEuOGxqfcePGcZMmTdjX15d9fX2ZiDgwMJB37txZ423efPNN7tOnj6Pdv39/fuKJJwx9Bg0ahG80QIT5r35d899oEF28Xt7Nzc3R7tu3r+HnmZmZNHfuXLLZbI5/EydOpNLSUqqoqKB7772XKisrqXPnzjRx4kRKTEx0XH8fFxdHoaGh1LlzZ3rwwQfp448/poqKinp9fOA8Lh1bBw4coKqqKho0aJAj5uHhQf369aOcnBzT9/n000/Tww8/TEOGDKHXXnuNDhw44PhZZmYmJSQkGMbu0KFD6cKFC1RQUCAeFzR+zFxrn5ycHAoJCaGQkBBHLCoqilq0aOEYn+Xl5TRt2jSKjIykFi1akM1mo5ycHMdf9ADMGDx4MGVkZFBGRgalpaXR0KFDafjw4XTw4EEiIlq5ciUNGjSIAgICyGaz0cyZMw1jLDc3l/r162e4T7UNUA3zX/3CQoMuDqhOnTo52upXaeXl5fTiiy86JsKMjAzKysqivLw88vLyopCQEMrNzaWFCxeSt7c3Pf7443TzzTdTVVUV2e12+umnn+iTTz6hwMBAmjVrFvXs2ZNOnDhRz48SnIH0Ne2fcXd31yZFdYOCOXPm0N69e+nOO++kzZs3U1RUFCUmJhLRxbH76KOPGsZuZmYm5eXlUVhYWJ2PC1xbREQEubm50b59+67ofqZNm0aJiYn0yiuvUHJyMmVkZFCPHj2woQBcFl9fXwoPD6fw8HC64YYb6H/+53/ozJkztGTJEkpNTaUHHniA7rjjDlq7di3t2rWLnn/+eYwxqDPMf/Xrml9obN68mbKysmj06NE19unduzfl5uY6JsJL/7m7X3wKvb29KT4+nt577z3aunUrpaamUlZWFhERNW3alIYMGUJvvPEG7d69mwoLC2nz5s318vjAeYWFhZGnpyelpKQ4YlVVVZSenk5RUVFEROTv70+nT5+mM2fOOPpkZGRo99WlSxd66qmnaMOGDXTPPffQ0qVLieji2M3OzhbHrqen59V9gOC0WrVqRUOHDqUFCxYYxla1EydOUGRkJBUXF1NxcbEjnp2dTSdOnHCMz5SUFBo/fjyNGjWKevToQQEBAVRYWGi4L09PTzp//vxVfTzQuLi5uZG7uztVVlbS9u3bKTQ0lJ5//nnq27cvRUREOL7pqNa1a1dKT083xNQ2QDXMf/WraUMfQH36448/qKysjM6fP0+//vorJSUl0auvvkojRoygsWPH1ni7WbNm0YgRI6hDhw70l7/8hdzd3SkzM5P27NlDL730EiUkJND58+epf//+5OPjQx999BF5e3tTaGgorV27ln7++We6+eabqWXLlrRu3Tq6cOECde3atR4fOTgjX19feuyxx2j69OnUqlUr6tChA73xxhtUUVFBEyZMICJyjKnnnnuOpkyZQj/88AMlJCQ47qOyspKmT59Of/nLX6hTp0506NAhSk9Pdyycn3nmGRowYABNmjSJHn74YfL19aXs7GzauHEjzZ8/vyEeNjiJBQsW0KBBg6hfv340d+5cio6OpnPnztHGjRtp0aJFlJ2dTT169KAHHniA3nnnHTp37hw9/vjjFBsb67jULiIiglavXk3x8fHk5uZGL7zwglaXqGPHjrRt2za67777qFmzZtSmTZuGeLjgxKrPzUREx48fp/nz51N5eTnFx8fTqVOnqKioiD799FO64YYb6JtvvnF8Y1tt8uTJNHHiROrbty/FxMTQypUraffu3dS5c+eGeDjgAjD/1aOGTRGpP+PGjXNsL9a0aVP29/fnIUOG8Icffsjnz5939CMiTkxM1G6flJTEMTEx7O3tzX5+ftyvXz9evHgxM19Mzu3fvz/7+fmxr68vDxgwgL/99ltmvph4Gxsbyy1btmRvb2+Ojo7mlStX1stjhoZTUzJ49bZ51SorK3ny5Mncpk0bcXtb5ovjKzw8nL29vXnEiBG8ePFiRzL4H3/8wffdd59ja+WgoCCeNGmSI9GbmTktLY3j4uLYZrOxr68vR0dH88svv+z4OZLVrl0lJSX8xBNPcGhoKHt6enL79u155MiRvGXLFmaufXvHgoICHjx4MHt7e3NISAjPnz+fY2NjeerUqY4+qampHB0dzc2aNbtmt3eEml16biYittvtfMMNN/AXX3zh6DN9+nRu3bo122w2HjNmDM+bN0/bJGPu3Lncpk0bttls/NBDD/GUKVN4wIAB9fxowJVg/qsfbswmsmIAAAAAXERcXBwFBATQ8uXLG/pQAK5p19SlUwAAANC4VFRU0Pvvv09Dhw6lJk2a0CeffELffvstbdy4saEPDeCah280AAAAwGVVVlZSfHw87dq1i37//Xfq2rUrzZw5k+65556GPjSAax4WGgAAAAAAYLlrfntbAAAAAACwHhYaAAAAAABgOSw0AAAAAADAclhoAAAAAACA5bDQAAAAAAAAy2GhAQAAAAAAlsNCAwAAAAAALIeFBgAAAAAAWO7/APuGlHx1tnh4AAAAAElFTkSuQmCC",
            "text/plain": [
              "<Figure size 1000x1000 with 25 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plt.figure(figsize=(10, 10))\n",
        "for i in range(25):\n",
        "    plt.subplot(5,5,i+1)\n",
        "    plt.xticks([])\n",
        "    plt.yticks([])\n",
        "    plt.grid(False)\n",
        "    plt.imshow(train_images[i], cmap=plt.cm.binary)\n",
        "    plt.xlabel(class_names[train_labels[i]])\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "59veuiEZCaW4"
      },
      "source": [
        "## 构建模型\n",
        "\n",
        "构建神经网络需要先配置模型的层，然后再编译模型。"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Gxg1XGm0eOBy"
      },
      "source": [
        "### 设置层\n",
        "\n",
        "神经网络的基本组成部分是<em>层</em>。层会从向其馈送的数据中提取表示形式。希望这些表示形式有助于解决手头上的问题。\n",
        "\n",
        "大多数深度学习都包括将简单的层链接在一起。大多数层（如 {class}`tensorflow.keras.layers.Dense`）都具有在训练期间才会学习的参数。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 15,
      "metadata": {
        "id": "9ODch-OFCaW4"
      },
      "outputs": [],
      "source": [
        "model = tf.keras.Sequential([\n",
        "    tf.keras.layers.Flatten(),\n",
        "    tf.keras.layers.Dense(128, activation='relu'),\n",
        "    tf.keras.layers.Dense(10)\n",
        "])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "gut8A_7rCaW6"
      },
      "source": [
        "该网络的第一层 {class}`tensorflow.keras.layers.Flatten` 将图像格式从二维数组（28 x 28 像素）转换成一维数组（28 x 28 = 784 像素）。将该层视为图像中未堆叠的像素行并将其排列起来。该层没有要学习的参数，它只会重新格式化数据。\n",
        "\n",
        "展平像素后，网络会包括两个 {class}`tensorflow.keras.layers.Dense` 层的序列。它们是密集连接或全连接神经层。第一个 `Dense` 层有 128 个节点（或神经元）。第二个（也是最后一个）层会返回一个长度为 10 的 logits 数组。每个节点都包含一个得分，用来表示当前图像属于 10 个类中的哪一类。\n",
        "\n",
        "### 编译模型\n",
        "\n",
        "在准备对模型进行训练之前，还需要再对其进行一些设置。以下内容是在模型的<em>编译</em>步骤中添加的：\n",
        "\n",
        "- <em>损失函数</em> - 测量模型在训练期间的准确程度。你希望最小化此函数，以便将模型“引导”到正确的方向上。\n",
        "- <em>优化器</em> - 决定模型如何根据其看到的数据和自身的损失函数进行更新。\n",
        "- <em>指标</em> - 用于监控训练和测试步骤。以下示例使用了*准确率*，即被正确分类的图像的比率。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 16,
      "metadata": {
        "id": "Lhan11blCaW7"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer='adam',\n",
        "              loss=tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True),\n",
        "              metrics=['accuracy'])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "qKF6uW-BCaW-"
      },
      "source": [
        "## 训练模型\n",
        "\n",
        "训练神经网络模型需要执行以下步骤：\n",
        "\n",
        "1. 将训练数据馈送给模型。在本例中，训练数据位于 `train_images` 和 `train_labels` 数组中。\n",
        "2. 模型学习将图像和标签关联起来。\n",
        "3. 要求模型对测试集（在本例中为 `test_images` 数组）进行预测。\n",
        "4. 验证预测是否与 `test_labels` 数组中的标签相匹配。\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Z4P4zIV7E28Z"
      },
      "source": [
        "### 向模型馈送数据\n",
        "\n",
        "要开始训练，请调用 <code>model.fit</code> 方法，这样命名是因为该方法会将模型与训练数据进行“拟合”："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 17,
      "metadata": {
        "id": "xvwvpA64CaW_"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/10\n"
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "WARNING: All log messages before absl::InitializeLog() is called are written to STDERR\n",
            "I0000 00:00:1729736509.779215 3208907 service.cc:146] XLA service 0x7f4dfc01b500 initialized for platform CUDA (this does not guarantee that XLA will be used). Devices:\n",
            "I0000 00:00:1729736509.779287 3208907 service.cc:154]   StreamExecutor device (0): NVIDIA GeForce RTX 3090, Compute Capability 8.6\n",
            "I0000 00:00:1729736509.779300 3208907 service.cc:154]   StreamExecutor device (1): NVIDIA GeForce RTX 2080 Ti, Compute Capability 7.5\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m  74/1875\u001b[0m \u001b[37m━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[1m3s\u001b[0m 2ms/step - accuracy: 0.4871 - loss: 1.4441    "
          ]
        },
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "I0000 00:00:1729736513.497273 3208907 device_compiler.h:188] Compiled cluster using XLA!  This line is logged at most once for the lifetime of the process.\n"
          ]
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m1875/1875\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m9s\u001b[0m 2ms/step - accuracy: 0.7771 - loss: 0.6364\n",
            "Epoch 2/10\n",
            "\u001b[1m1875/1875\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 2ms/step - accuracy: 0.8618 - loss: 0.3829\n",
            "Epoch 3/10\n",
            "\u001b[1m1875/1875\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 2ms/step - accuracy: 0.8774 - loss: 0.3358\n",
            "Epoch 4/10\n",
            "\u001b[1m1875/1875\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 2ms/step - accuracy: 0.8860 - loss: 0.3124\n",
            "Epoch 5/10\n",
            "\u001b[1m1875/1875\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 2ms/step - accuracy: 0.8899 - loss: 0.2986\n",
            "Epoch 6/10\n",
            "\u001b[1m1875/1875\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 2ms/step - accuracy: 0.8965 - loss: 0.2785\n",
            "Epoch 7/10\n",
            "\u001b[1m1875/1875\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 2ms/step - accuracy: 0.9030 - loss: 0.2657\n",
            "Epoch 8/10\n",
            "\u001b[1m1875/1875\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 2ms/step - accuracy: 0.9058 - loss: 0.2536\n",
            "Epoch 9/10\n",
            "\u001b[1m1875/1875\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 2ms/step - accuracy: 0.9098 - loss: 0.2427\n",
            "Epoch 10/10\n",
            "\u001b[1m1875/1875\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 2ms/step - accuracy: 0.9118 - loss: 0.2340\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "<keras.src.callbacks.history.History at 0x7f518f49ec30>"
            ]
          },
          "execution_count": 17,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "model.fit(train_images, train_labels, epochs=10)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "W3ZVOhugCaXA"
      },
      "source": [
        "在模型训练期间，会显示损失和准确率指标。此模型在训练数据上的准确率达到了 0.91（或 91%）左右。"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wCpr6DGyE28h"
      },
      "source": [
        "### 评估准确率\n",
        "\n",
        "接下来，比较模型在测试数据集上的表现："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 18,
      "metadata": {
        "id": "VflXLEeECaXC"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "313/313 - 2s - 6ms/step - accuracy: 0.8796 - loss: 0.3494\n",
            "\n",
            "Test accuracy: 0.8795999884605408\n"
          ]
        }
      ],
      "source": [
        "test_loss, test_acc = model.evaluate(test_images,  test_labels, verbose=2)\n",
        "\n",
        "print('\\nTest accuracy:', test_acc)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "yWfgsmVXCaXG"
      },
      "source": [
        "结果表明，模型在测试数据集上的准确率略低于训练数据集。训练准确率和测试准确率之间的差距代表*过拟合*。过拟合是指机器学习模型在新的、以前未曾见过的输入上的表现不如在训练数据上的表现。过拟合的模型会“记住”训练数据集中的噪声和细节，从而对模型在新数据上的表现产生负面影响。有关更多信息，请参阅以下内容：\n",
        "\n",
        "- [演示过拟合](https://tensorflow.google.cn/tutorials/keras/overfit_and_underfit#demonstrate_overfitting)\n",
        "- [防止过拟合的策略](https://tensorflow.google.cn/tutorials/keras/overfit_and_underfit#strategies_to_prevent_overfitting)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "v-PyD1SYE28q"
      },
      "source": [
        "### 进行预测\n",
        "\n",
        "模型经过训练后，您可以使用它对一些图像进行预测。附加一个 Softmax 层，将模型的线性输出 [logits](https://developers.google.com/machine-learning/glossary#logits) 转换成更容易理解的概率。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "id": "DnfNA0CrQLSD"
      },
      "outputs": [],
      "source": [
        "probability_model = tf.keras.Sequential([model, \n",
        "                                         tf.keras.layers.Softmax()])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 20,
      "metadata": {
        "id": "Gl91RPhdCaXI"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 3ms/step\n"
          ]
        }
      ],
      "source": [
        "predictions = probability_model.predict(test_images)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "x9Kk1voUCaXJ"
      },
      "source": [
        "在上例中，模型预测了测试集中每个图像的标签。我们来看看第一个预测结果："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 21,
      "metadata": {
        "id": "3DmJEUinCaXK"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "array([1.6527409e-08, 1.4129248e-09, 2.2416434e-10, 8.0054169e-10,\n",
              "       3.7958323e-08, 8.9384953e-04, 2.1727022e-08, 1.2796247e-02,\n",
              "       3.7120977e-08, 9.8630977e-01], dtype=float32)"
            ]
          },
          "execution_count": 21,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "predictions[0]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-hw1hgeSCaXN"
      },
      "source": [
        "预测结果是一个包含 10 个数字的数组。它们代表模型对 10 种不同服装中每种服装的“置信度”。您可以看到哪个标签的置信度值最大："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 22,
      "metadata": {
        "id": "qsqenuPnCaXO"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "9"
            ]
          },
          "execution_count": 22,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "np.argmax(predictions[0])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E51yS7iCCaXO"
      },
      "source": [
        "因此，该模型非常确信这个图像是短靴，或 `class_names[9]`。通过检查测试标签发现这个分类是正确的："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 23,
      "metadata": {
        "id": "Sd7Pgsu6CaXP"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "9"
            ]
          },
          "execution_count": 23,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "test_labels[0]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ygh2yYC972ne"
      },
      "source": [
        "您可以将其绘制成图表，看看模型对于全部 10 个类的预测。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 24,
      "metadata": {
        "id": "DvYmmrpIy6Y1"
      },
      "outputs": [],
      "source": [
        "def plot_image(i, predictions_array, true_label, img):\n",
        "  true_label, img = true_label[i], img[i]\n",
        "  plt.grid(False)\n",
        "  plt.xticks([])\n",
        "  plt.yticks([])\n",
        "\n",
        "  plt.imshow(img, cmap=plt.cm.binary)\n",
        "\n",
        "  predicted_label = np.argmax(predictions_array)\n",
        "  if predicted_label == true_label:\n",
        "    color = 'blue'\n",
        "  else:\n",
        "    color = 'red'\n",
        "\n",
        "  plt.xlabel(\"{} {:2.0f}% ({})\".format(class_names[predicted_label],\n",
        "                                100*np.max(predictions_array),\n",
        "                                class_names[true_label]),\n",
        "                                color=color)\n",
        "\n",
        "def plot_value_array(i, predictions_array, true_label):\n",
        "  true_label = true_label[i]\n",
        "  plt.grid(False)\n",
        "  plt.xticks(range(10))\n",
        "  plt.yticks([])\n",
        "  thisplot = plt.bar(range(10), predictions_array, color=\"#777777\")\n",
        "  plt.ylim([0, 1])\n",
        "  predicted_label = np.argmax(predictions_array)\n",
        "\n",
        "  thisplot[predicted_label].set_color('red')\n",
        "  thisplot[true_label].set_color('blue')"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Zh9yABaME29S"
      },
      "source": [
        "### 验证预测结果\n",
        "\n",
        "在模型经过训练后，您可以使用它对一些图像进行预测。"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "d4Ov9OFDMmOD"
      },
      "source": [
        "我们来看看第 0 个图像、预测结果和预测数组。正确的预测标签为蓝色，错误的预测标签为红色。数字表示预测标签的百分比（总计为 100）。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 25,
      "metadata": {
        "id": "HV5jw-5HwSmO"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 600x300 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "i = 0\n",
        "plt.figure(figsize=(6,3))\n",
        "plt.subplot(1,2,1)\n",
        "plot_image(i, predictions[i], test_labels, test_images)\n",
        "plt.subplot(1,2,2)\n",
        "plot_value_array(i, predictions[i],  test_labels)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 26,
      "metadata": {
        "id": "Ko-uzOufSCSe"
      },
      "outputs": [
        {
          "data": {
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeQAAAESCAYAAAAsZab9AAAAP3RFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjkuMS5wb3N0MSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8kixA/AAAACXBIWXMAAA9hAAAPYQGoP6dpAAAdgklEQVR4nO3deXzURZ7/8XcIBCIkXOFIIAmGUUAQlkNYQEdXUWRZFo8HIgsriDozTtwBGdlBnVnGVQ7Z1dXxQFEXVwQBDzxQREQBdUAjGCWiIF6gqIgcCQlHSD6/P+rXJoH0t9PdYVLK6/l49AO661vV1Z3Au6vrW99KMDMTAACoU/XqugMAAIBABgDACwQyAAAeIJABAPAAgQwAgAcIZAAAPEAgAwDggfp13QEAfiovL9eOHTuUkpKihISEuu4O8JNlZioqKlJGRobq1Qs/DiaQAVRrx44dyszMrOtuAD8b27dvV/v27cOWE8gAqpWSkiLJ/SeSmppax70BfroKCwuVmZn547+pcAhkANUKfU2dmppKIAO1INLUDyd1AQDgAQIZAAAPEMgAAHiAQAYAwAMEMgAAHiCQAQDwQI2WPXHFHqB21PSKPQBOPDUKZK7YA9SuSFfsAXDiqVEgc8UeoHbU9Io9AE48NQpkrtgD1C6mfgAcjUksAAA8QCADAOABAhkAAA8QyAAAeIBABgDAAwQyAAAeIJABAPAAgQwAgAcIZAAAPEAgAwDgAQIZAAAPEMgAAHiAQAYAwAMEMgAAHiCQAQDwAIEMAIAHCGQAADxAIAMA4AECGQAADxDIAAB4gEAGAMADBDIAAB4gkAEA8ACBDACABwhkAAA8QCADAOABAhkAAA8QyAAAeIBABgDAAwQyAAAeIJABAPAAgQwAgAcIZAAAPEAgAwDgAQIZAAAPEMgAAHiAQAYAwAMEMgAAHiCQAQDwAIEMAIAHCGQAADxAIAMA4AECGQAADxDIAAB4gEAGAMADBDIAAB4gkAEA8ACBDACABwhkAAA8QCADAOABAhkAAA8QyAAAeIBABgDAAwQyAAAeIJABAPAAgQwAgAcIZAAAPEAgAwDgAQIZAAAPEMgAAHiAQAYAwAMEMgAAHiCQAQDwAIEMAIAH6td1B05E999/f2B5QUFBzHXjYWaB5QkJCcftuQHgRMcIGQAADxDIAAB4gEAGAMADBDIAAB4gkAEA8ACBDACABwhkAAA8UGvrkA8cOBC2LDk5+bi0K0lJSUkxtx1JYmJizHWXLl0atmzHjh2BdVu3bh227IorrgisO23atMDyzMzMsGXxrDMuKyuLuW487zMA/FwwQgYAwAMEMgAAHiCQAQDwAIEMAIAHCGQAADxAIAMA4IFaW/YUtBznuuuuC6x79tlnhy2LZ8lUXQraJrFv376BdYOWcrVv3z6w7qJFiwLLg5ZUXXzxxYF1U1JSwpZFWroUtCwq0raPxxNbSgLwBSNkAAA8QCADAOABAhkAAA8QyAAAeIBABgDAAwQyAAAeIJABAPBAVOuQi4uLw643/eqrr8LWe/755wPbLSkpCVvWrVu3wLotWrQIW3bSSScF1i0vLw8s37ZtW9iyuXPnBtZt27Zt2LK0tLTAui+88ELYsuHDhwfW3bt3b2D5Sy+9FLbs448/Dqybk5MTtuz8888PrJudnR1YfrxE2hYy6HegXr3gz6tsGwmgNtXahUEAAHVo2zZp167Y66elSVlZtdcfRI1ABoCfum3bpE6dpIMHY2+jUSNp82ZCuQ4xhwwAP3W7dsUXxpKrH88IG3EjkAEA8ACBDACABwhkAAA8ENVJXR999JEaN24c9ZMUFxcHls+fPz9sWffu3QPrBm1VGFQmSVu3bg0s37hxY9iyw4cPB9Y966yzwpZt2LAhsO7gwYPDlkVayhXpNV944YVhy3bu3BlYd8uWLWHL1q5dG1i3S5cuYcu6du0aWLdPnz6B5a1atQpbFmlpEkuXAPiCETIAAB4gkAEA8ACBDACABwhkAAA8QCADAOABAhkAAA8QyAAAeCCqdch79+5VaWlptWW7d+8O/yT1g59m3759YcuWLFkSWLd58+Zhy8L1NSQlJSWwvH///mHLTj311MC6QVv3RdpSclfA9WSDtqqUgrejlIJ/TpHWOGcFXHQ+qEySCgsLw5a98cYbgXXz8vICy4P63axZs8C6QdtCtm7dOrBu586dw5Y1bNgwsC4AHI0RMgAAHiCQAQDwAIEMAIAHCGQAADxAIAMA4AECGQAADxDIAAB4IKp1yI0bNw67H3LQXrlXXnllYLsdOnQIWxa0blaSDh48GLYs0hrURo0axdz2Bx98EFg3SJMmTQLLg9bVRtqH+dtvvw0sD9ovOTU1NbBu0HMHrTOWpLS0tLBlkdZORxL0c4q0x/OOHTvClkVal33bbbeFLRszZky1j0daRw7gxMUIGQAADxDIAAB4gEAGAMADBDIAAB4gkAEA8ACBDACAB6Ja9vTSSy+F3VYuPT09bL2gpSVS8JKZnJycwLpB2/4dOXIksG6kfh06dChsWVlZWWDdIHv37g0sD9qOskGDBoF1I20ZGM+ypyCRlgi1adMmbFmk9zLSUq+g5WuRlrYF/e5F+v1ISEgIW3bnnXdW+3ik30kAJy5GyAAAeIBABgDAAwQyAAAeIJABAPAAgQwAgAcIZAAAPEAgAwDggajWIX/22Wdh18F27NgxbL1u3boFtltQUBC27KuvvgqsG8/2eeXl5YHl8dQNWjsbaV1t0PrWcOvAQ77//vvA8qD6ycnJgXUjrYEOsmvXrrBlkd7LoqKiwPKgdd2R6gZthRm0raMkffLJJ1H3KZ7fOQA/b4yQAQDwAIEMAIAHCGQAADxAIAMA4AECGQAAD0R1ljUAAHVtypQpcdWfOXNmLfWkdkUVyCeddFLYpS/r1q0LWy+e7fMi1S0pKQlbFmk7wbS0tMDy/fv3hy2LZ/vFxMTEwPL69cP/WILKJKleveAvPYK2X4wkaNlT0PIhSdq5c2fYsqCfoRS8RaIUvDyptLQ0sG7Q+xlpq8Sg573llluqffzAgQP6zW9+E9gugBMTX1kDAOABAhkAAA8QyAAAeIBABgDAAwQyAAAeIJABAPAAgQwAgAeiWoc8a9assGt7s7KywtZr0aJFYLtBW/NFWocctK420vrW3bt3B5anpKSELYu0RjVoPXCktcRBW/QdOHAgsG7Q1o1S8PsV6b2O5zXFUzfS70+zZs3ClgWtcY/UdqdOnQLrnn/++YHl1SksLGQdMoBqMUIGAMADBDIAAB4gkAEA8ACBDACABwhkAAA8QCADAOCBqJY9NW/ePOyyp+nTp9dKhwAAOBExQgYAwAMEMgAAHiCQAQDwAIEMAIAHCGQAADxAIAMA4AECGQAADxDIAAB4gEAGAMADBDIAAB4gkAEA8ACBDACABwhkAAA8QCADAOABAhkAAA8QyAAAeIBABgDAAwQyAAAeIJABAPAAgQwAgAcIZAAAPEAgAwDgAQIZAAAPEMgAAHiAQAYAwAMEMgAAHiCQAQDwAIEMAIAHCGQAADxAIAMA4AECGQAADxDIAAB4gEAGAMADBDIAAB4gkAEA8ACBDACABwhkAAA8QCADAOABAhkAAA8QyAAAeIBABgDAAwQyAAAeIJABAPAAgQwAgAcIZAAAPEAgAwDgAQIZAAAPEMgAAHiAQAYAwAMEMgAAHiCQAQDwAIEMAIAHCGQAADxAIAMA4IH6dd0BAMDP27Zt0q5dsddPS5OysmqvP74ikAEAx822bVKnTtLBg7G30aiRtHnzzz+U+coaAHDc7NoVXxhLrn48I+yfCgIZAAAPEMgAAHigRnPIZiZJKiwsPK6dAX7uQv+GQv+mACCkRoFcVFQkScrMzDyunQFOFEVFRWratGlddwOAR2oUyBkZGdq+fbtSUlKUkJBwvPsE/GyZmYqKipSRkVHXXQHgmRoFcr169dS+ffvj3RfghMDIGEB1OKkLAAAPEMgAAHiAQAYAwAMEMgAAHiCQwxk3Trroopof/8UXUkKClJ9/fPpzPPzwg9S6teu7b6J9/2vq8sulO+6o/XYBIE5+B/L330vXXuuuKN6wodS2rTR4sPTWW3Xds9iMG+dCu/LtwguPPe7FF6V+/aTkZKl586rBtHu3NGyY1KSJ1LOn9N57Vevm5tY8cKZNk4YPlzp0qHhsyRLp7/9eatpUSkmRunaVJk6M7nX67I9/dK9737667gkAVOH3bk+XXiodPiz93/9JOTnSd99JK1e6kd1P1YUXSnPnVtxv2LBq+dNPS9dcI02fLp17rnTkiFRQUFE+bZpUVCRt2CDNnu2OffddV7ZunfT229Jf/hK5HyUl0iOPSMuXVzy2cqU0cqR7jn/+Z/eBYdMmacWK2F+vLw4flpKSpG7dpI4dpccfdx9eEBZX6PsJ2b+/9tqp5Z/38ejaoUOH4mrrb/07XeMr9Jmv9uwxk8xWrQo+7o47zLp1MzvpJLP27c2uvdasqKiifO5cs6ZNzV5+2axzZ7PGjc0GDzbbsaPimCNHzK6/3h3XooXZ5MlmV1xhNnx4xTHLlpkNHFhxzNChZlu3VpR//rnr73vvhe/r2LFV2zxaaalZu3ZmDz8c/pghQ8xmz3Z/37TJvW4zs8OHzXr0MMvLC1+3siefNGvVqupjEyaYnXNOcL2pU93zPPaYWXa2WWqq2ciRZoWFFceUlZlNn27WoYNZo0Zm3bu75ws5csRs/PiK8lNPNbvrrqrPc/R79c47ZmlpZjNnuvt79phddZV7LCXF7B/+wSw//9h+PvSQe56EhIqyW24xO/PM4NcJ2759u0nixo1bLd22b98e+G/O3xFykybu9uyz7ivUo0eSIfXquRHhySdLn30m/fa30r//u3T//RXHlJRI//3f0rx57vgxY6QbbpDmz3fld9whPfqo9L//K3Xp4u4vWeJGqCHFxdKkSVL37u6j2n/8h3TxxW7OuF4U3/yvWuXmbZs3d+3fdpvUsqUr27BB+vpr117PntK330p/93fSf/2XG9lJUo8e0muvSVdf7Ua33bu7x2fNks45R+rTp2b9eOMNqXfvqo+1bSstWOBG5KHnq86nn7qfy9Kl0p490mWXSTNnupG1JM2Y4UagDzwgnXKKtGaNe89btZLOPlsqL5fat5eefNK99r/+VfrVr6T0dNfW0V57TbrkEvcaf/Ur99iIEe4r/WXL3NfrDz4onXeetGWL1KKFO2brVveNwzPPSImJFe317ev6euhQ+N8r1MoV+goLC5WZmant27crNTU17j7VZns+962226Nvddue1fQKfX+bz9oxeuops+bN3ShqwACzG280e//94DpPPmnWsmXF/blz3ci18mj2vvvM2rSpuJ+ebjZrVsX90lI32g4azX7/vWt340Z3vyYj5CeeMHvuObMPPjBbssSsSxezM85wI8ZQuWSWleVe+7vvmo0a5V7PDz+4Y/budY9lZZn98pdmH35otmWL2SmnmO3aZfbrX5udfLLZiBHu2HCGD3ej1Mr27zf7x390fcjOdiPfRx4xO3iw4pipU92ovPKIePJks3793N8PHnTlf/1r1bavusr1O5zcXLNLL624HxohP/OMWZMmZgsXVpS98YYbmVful5lZx45mDz5Y0c8GDcx27jz2ud5/373GL74I3x/Uin379pkk27dvn3ft+dy32m6PvvnTXhC/T+q69FJpxw7p+efd3OuqVVKvXm40G/Lqq25k1K6dOwnpX//VzTGXlFQcc9JJbt4wJD1d2rnT/X3fPumbb9xJVCH16x870vzkE2nUKDeXnZpacSLUtm01fz2XX+7mZk8/3Z2otXSplJfnXpfkRo6SdPPN7rX37u3mmxMS3GhScqPBBQukL7+UVq+WTjtN+vWv3Sh6/nz3LcHmze41/+d/hu/LgQNSo0ZVH2vc2J1QtnWrO/mpSRPp9793I8rK72eHDu69Dqn8fm7d6o49//yKbzmaNJEee8yNrEPuu8+9vlatXPmcOce+l2+/7UbC8+a5ue2Q999331K0bFn1OT7/vOpzZGe79o+WnOz+rPyaAKCO+R3IkguN88+X/vQn99XmuHHS1Kmu7IsvpH/6J/e17dNPS+vXu//oJXcST0iDBlXbTEiQot3+btgwd4bzQw+5oHj77WOfJ1o5OVJamgsxyQWb5EI2pGFDd1y44J87V2rWzJ0tvWqVC/oGDVyQhYK+Omlp7uvm6nTs6L4Sf/hh9zX6pk3SokUV5dW9n6EPE6EzOF580X2dH7pt2iQ99ZQrW7jQTRlcdZX0yiuu/Morj30vO3aUOnd2UwmlpRWP79/v3qvK7efnuw8ikydXHNe4cfWvb/du92d1YQ0AdcTfOeRwTjvNzV9KLoDLy92cb2ged/Hi6Npr2tT95/7229Ivf+keO3LEtd2rl7v/ww/uP/uHHpLOOss99uabcb8UffWVazsUxL17uwDevFk680z3WGmp++CRnX1s/e+/d6PgUF/KyiqCq7TU3Q+nZ083zxtJhw5utF1cXLPXdNpp7jVs2+bmi6vz1lvSgAFuvj+k8sg2JC3Nzf+ec46bW1682H0Y6NXLza/Xr191yVZNFRS4Oey0tOjrIioNGzbU1KlT1bCW5uprsz2f+1bb7dE3f9oLdNy/FI/Vrl3uzNl589yc32efmS1e7OZ+Q3Of+fluLvCuu8w+/dSd+duunXtszx53TOgs68qWLHHHhMyc6c6cXrLE7KOPzK65xp25G5pDLitz87hjxph98onZypVu7ldydcwizyEXFZndcIPZ2rXu2FdfNevVy839Vp4LnTDBvYbly80+/tjNvbZubbZ797Ft/su/mN1zT8X92283693bnX09ZIjZb38b/v394AOz+vWrtjt1qpsPfv11935v2GA2bpxZcrLrS+iYHj2qtvU//+PmnENuvtm9X48+6ubu1683+8tf3H0zs7vvdnPAL79stnmz2R//6O5XbrfyWdbffOPOkL/0Uje/X17uzpLu0cO9T59/bvbWW2Y33VRxlnl1/azc9tHz5wBQx/wN5IMHzaZMcaHVtKk7UahTJ/efd0lJxXF33ulOykpOdsuZHnss+kAuLXVBmJpq1qyZ2aRJxy57WrHCnYTVsKFbxrNqVXSBXFJidsEFbqlRgwYuwK65xuzbb6sed/iw2e9/70I4JcVs0CCzgoJj23v5ZbO+fd2HhZDiYncyV0qK2XnnmX33XfV9Cenb1+yBByruv/aaC73MTLOkJPfh58IL3UlUITUJ5PJy9yGpUyf3Wlu1cj+b1atd+cGDLuibNnXv97XXup91uEA2c8vUTj3V7LLL3ElwhYVm//ZvZhkZ7jkyM81Gjzbbti18P83MDhxwz7t2bfB7AwB/Ywlm0U6m4mfjxRfdnGtBQXRLt37KZs92S9peeaWuewIAVfz05pBRe4YOdWePf/21lJlZ173522jQQLrnnrruBQAcgxEyAAAeOEG+pwQAwG8EMoDj5r777lOHDh3UqFEj9evXT++8805M7axZs0bDhg1TRkaGEhIS9Gxo6WMMZsyYoTPOOEMpKSlq3bq1LrroIm3evDnm9mbPnq3u3bsrNTVVqamp6t+/v5YtWxZze5XNnDlTCQkJmhjjjmt//vOflZCQUOXWuXPnmPvz9ddfa8yYMWrZsqWSk5N1+umn693Q5jZR6tChwzF9S0hIUG6Mm76UlZXpT3/6k04++WQlJyerY8eOuvXWWyNv6BBGUVGRJk6cqOzsbCUnJ2vAgAHKy8uLqa2aIpABHBeLFi3SpEmTNHXqVG3YsEE9evTQ4MGDtTN0VbcoFBcXq0ePHrovdOGfOKxevVq5ublat26dVqxYodLSUl1wwQUqrula+6O0b99eM2fO1Pr16/Xuu+/q3HPP1fDhw/Xhhx/G1c+8vDw9+OCD6h66Xn2Munbtqm+++ebH25sxXkNhz549GjhwoBo0aKBly5Zp06ZNuuOOO9S8efOY2svLy6vSrxX/f1e5ESNGxNTe7bffrtmzZ+vee+/VRx99pNtvv12zZs3SPTGeM3L11VdrxYoVmjdvnjZu3KgLLrhAgwYN0tdffx1TezVSp+d4A/jZ6tu3r+Xm5v54v6yszDIyMmzGjBlxtSvJloSWG9aCnTt3miRbHVqWVwuaN29uDwft2hZBUVGRnXLKKbZixQo7++yzbcKECTG1M3XqVOsRbj1+lP7whz/Ymcdxl7QJEyZYx44drby8PKb6Q4cOtfFHXV/gkksusdGjR0fdVklJiSUmJtrSpUurPN6rVy+7+eabY+pfTTBCBlDrDh8+rPXr12vQoEE/PlavXj0NGjRIa9eurcOeHWvfvn2SpBahXcLiUFZWpoULF6q4uFj9+/ePuZ3c3FwNHTq0yvsXq08++UQZGRnKycnR6NGjtS2a6+9X8vzzz6tPnz4aMWKEWrdurZ49e+qhhx6Ku3+S+315/PHHNX78+Jh3FhswYIBWrlypLVu2SJLef/99vfnmmxoyZEjUbR05ckRlZWVqdNT1/pOTk2P+hqEmWPYEoNbt2rVLZWVlatOmTZXH27Rpo48//riOenWs8vJyTZw4UQMHDlS3oC1HI9i4caP69++vgwcPqkmTJlqyZIlOq3xN+igsXLhQGzZsqJX5yn79+unRRx9Vp06d9M033+iWW27RWWedpYKCAqVU3iCmBj777DPNnj1bkyZN0k033aS8vDz97ne/U1JSksaOHRtXP5999lnt3btX48aNi7mNKVOmqLCwUJ07d1ZiYqLKyso0bdo0jR49Ouq2UlJS1L9/f916663q0qWL2rRpoyeeeEJr167VL37xi5j7GAmBDOCElZubq4KCgrhHPZ06dVJ+fr727dunp556SmPHjtXq1aujDuXt27drwoQJWrFixTGjs1hUHh12795d/fr1U3Z2thYvXqyrrroqqrbKy8vVp08fTZ8+XZLUs2dPFRQU6IEHHog7kB955BENGTIk8n7BARYvXqz58+drwYIF6tq1q/Lz8zVx4kRlZGTE1L958+Zp/PjxateunRITE9WrVy+NGjVK69evj7mPkRDIAGpdWlqaEhMT9d1331V5/LvvvlPbtm3rqFdVXXfddVq6dKnWrFmj9u3bx9VWUlLSjyOn3r17Ky8vT3fffbcefPDBqNpZv369du7cqV6hjW3kvgZfs2aN7r33Xh06dEiJiYkx97NZs2Y69dRTtTW0w1wU0tPTj/mA0aVLFz399NMx90eSvvzyS7366qt65pln4mpn8uTJmjJlii6//HJJ0umnn64vv/xSM2bMiCmQO3bsqNWrV6u4uFiFhYVKT0/XyJEjlZOTE1c/gzCHDKDWJSUlqXfv3lq5cuWPj5WXl2vlypVxza3WBjPTddddpyVLlui1117TySefXOvPUV5erkOHDkVd77zzztPGjRuVn5//461Pnz4aPXq08vPz4wpjSdq/f78+/fRTpYd2mIvCwIEDj1ketmXLFmVXtxNdFObOnavWrVtr6NChcbVTUlKiekddAjgxMVHloa1hY9S4cWOlp6drz549Wr58uYYPHx5Xe0EYIQM4LiZNmqSxY8eqT58+6tu3r+666y4VFxfryiuvjLqt/fv3VxnVff7558rPz1eLFi2UlZUVVVu5ublasGCBnnvuOaWkpOjbb7+VJDVt2lTJyclR9+3GG2/UkCFDlJWVpaKiIi1YsECrVq3S8uXLo24rJSXlmLnsxo0bq2XLljHNcd9www0aNmyYsrOztWPHDk2dOlWJiYkaNWpU1G1df/31GjBggKZPn67LLrtM77zzjubMmaM5c+ZE3VZIeXm55s6dq7Fjx6p+/fjiaNiwYZo2bZqysrLUtWtXvffee7rzzjs1fvz4mNpbvny5zEydOnXS1q1bNXnyZHXu3Dmm398aO27nbwM44d1zzz2WlZVlSUlJ1rdvX1u3bl1M7bz++usm6Zjb2LFjo26runYk2dy5c2Pq2/jx4y07O9uSkpKsVatWdt5559krr7wSU1vViWfZ08iRIy09Pd2SkpKsXbt2NnLkSNu6dWvMfXnhhResW7du1rBhQ+vcubPNmTMn5rbMzJYvX26SbPPmzXG1Y2ZWWFhoEyZMsKysLGvUqJHl5OTYzTffbIcOHYqpvUWLFllOTo4lJSVZ27ZtLTc31/bu3Rt3P4NwLWsAADzAHDIAAB4gkAEA8ACBDACABwhkAAA8QCADAOABAhkAAA8QyAAAeIBABgDAAwQyAAAeIJABAPAAgQwAgAf+H2WkpSZy1ZdHAAAAAElFTkSuQmCC",
            "text/plain": [
              "<Figure size 600x300 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "i = 12\n",
        "plt.figure(figsize=(6,3))\n",
        "plt.subplot(1,2,1)\n",
        "plot_image(i, predictions[i], test_labels, test_images)\n",
        "plt.subplot(1,2,2)\n",
        "plot_value_array(i, predictions[i],  test_labels)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kgdvGD52CaXR"
      },
      "source": [
        "让我们用模型的预测绘制几张图像。请注意，即使置信度很高，模型也可能出错。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 27,
      "metadata": {
        "id": "hQlnbqaw2Qu_"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1200x1000 with 30 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "# Plot the first X test images, their predicted labels, and the true labels.\n",
        "# Color correct predictions in blue and incorrect predictions in red.\n",
        "num_rows = 5\n",
        "num_cols = 3\n",
        "num_images = num_rows*num_cols\n",
        "plt.figure(figsize=(2*2*num_cols, 2*num_rows))\n",
        "for i in range(num_images):\n",
        "  plt.subplot(num_rows, 2*num_cols, 2*i+1)\n",
        "  plot_image(i, predictions[i], test_labels, test_images)\n",
        "  plt.subplot(num_rows, 2*num_cols, 2*i+2)\n",
        "  plot_value_array(i, predictions[i], test_labels)\n",
        "plt.tight_layout()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "R32zteKHCaXT"
      },
      "source": [
        "## 使用训练好的模型\n",
        "\n",
        "最后，使用训练好的模型对单个图像进行预测。"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 28,
      "metadata": {
        "id": "yRJ7JU7JCaXT"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "(28, 28)\n"
          ]
        }
      ],
      "source": [
        "# Grab an image from the test dataset.\n",
        "img = test_images[1]\n",
        "\n",
        "print(img.shape)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vz3bVp21CaXV"
      },
      "source": [
        "`tf.keras` 模型经过了优化，可同时对一个*批*或一组样本进行预测。因此，即便您只使用一个图像，您也需要将其添加到列表中："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 29,
      "metadata": {
        "id": "lDFh5yF_CaXW"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "(1, 28, 28)\n"
          ]
        }
      ],
      "source": [
        "# Add the image to a batch where it's the only member.\n",
        "img = (np.expand_dims(img,0))\n",
        "\n",
        "print(img.shape)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EQ5wLTkcCaXY"
      },
      "source": [
        "现在预测这个图像的正确标签："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 30,
      "metadata": {
        "id": "o_rzNSdrCaXY"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 143ms/step\n",
            "[[3.28207592e-04 2.89583964e-11 9.99377131e-01 1.29967814e-10\n",
            "  1.93019659e-04 2.80547027e-11 1.01659716e-04 2.54715407e-16\n",
            "  1.74937620e-09 1.95157475e-12]]\n"
          ]
        }
      ],
      "source": [
        "predictions_single = probability_model.predict(img)\n",
        "\n",
        "print(predictions_single)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 31,
      "metadata": {
        "id": "6Ai-cpLjO-3A"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 640x480 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plot_value_array(1, predictions_single[0], test_labels)\n",
        "_ = plt.xticks(range(10), class_names, rotation=45)\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cU1Y2OAMCaXb"
      },
      "source": [
        "`keras.Model.predict` 会返回一组列表，每个列表对应一批数据中的每个图像。在批次中获取对我们（唯一）图像的预测："
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 32,
      "metadata": {
        "id": "2tRmdq_8CaXb"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "2"
            ]
          },
          "execution_count": 32,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "np.argmax(predictions_single[0])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YFc2HbEVCaXd"
      },
      "source": [
        "该模型会按照预期预测标签。"
      ]
    }
  ],
  "metadata": {
    "colab": {
      "collapsed_sections": [],
      "name": "classification.ipynb",
      "toc_visible": true
    },
    "kernelspec": {
      "display_name": "xxx",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.12.2"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}