CNN_CIFAR10.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "gpuType": "T4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "# Convolutional Neural Network\n",
        "\n",
        "**Description:**\n",
        "- Train different variations of a CNN on the CIFAR-10 Dataset which contains images of animals and vehicles\n",
        "- Determine the optimal CNN\n",
        "- Enable users to try predicting their own test images\n",
        "\n",
        "**Dataset:**\n",
        "- Found here: https://www.cs.toronto.edu/~kriz/cifar.html\n",
        "- 60,000 32x32 images\n",
        "- 10 classes: airplane, automobile, bird, cat, deer, dog, frog, horse, ship, truck\n",
        "\n",
        "**How To Try It:**\n",
        "- Download CNN_CIFAR10.keras file from this repository's releases\n",
        "- Skip to the \"Try Your Own Images\" section of this notebook and run that code to predict your own 32x32 images\n",
        "\n",
        "  - More details:\n",
        "    - Download this .ipynb file and open in Google Colab\n",
        "    - Upload the .keras file and your image\n",
        "    - Change the file path in the code to the name of your image file\n",
        "\n",
        "**Data:**\n",
        "\n",
        "| Model | Layers | Training Accuracy | Testing Accuracy |\n",
        "|---    |---     |---                | ---              |\n",
        "| 1 | 3 layers of size 32, 64, 128 with 50% dropout | 0.78 | 0.73 |\n",
        "| 2 | 3 layers of size 32, 64, 128 | 0.89 | 0.71 |\n",
        "| 3 | 3 layers of size 32, 64, 64 | 0.79 | 0.70 |\n",
        "| 4 | 3 layers of size 32, 64, 64 with 1 less pooling layer | 0.92 | 0.68 |\n",
        "| 5 | 2 layers of size 32, 64  | 0.90 | 0.66 |\n",
        "\n",
        "**Conclusions:**\n",
        "- Greater than 10 epochs does not substantially improve test accuracy\n",
        "- Dropout of 50% improves generalization; model performs relatively well on test set even though training accuracy is relatively low\n",
        "- 3 layers all with pooling layers generalizes better\n",
        "- A larger 3rd layer improves accuracy slightly\n",
        "\n",
        "\n",
        "\n"
      ],
      "metadata": {
        "id": "q_xtV22qII10"
      }
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Train"
      ],
      "metadata": {
        "id": "jHr0R8oBNgeZ"
      }
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "bu_FGUNwH5Bx"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras import datasets\n",
        "from tensorflow.keras.models import Sequential, load_model\n",
        "from tensorflow.keras.layers import Conv2D, MaxPooling2D, Flatten, Dense, Dropout\n",
        "from tensorflow.keras.losses import SparseCategoricalCrossentropy\n",
        "from tensorflow.keras.preprocessing import image\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib.cm as cm\n",
        "import numpy as np"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Load numpy arrays\n",
        "(train_images, train_labels), (test_images, test_labels) = datasets.cifar10.load_data()\n",
        "\n",
        "img_width = 32\n",
        "img_height = 32\n",
        "img_color_channels = 3\n",
        "class_names = ['airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog',\n",
        "               'horse', 'ship', 'truck']\n",
        "\n",
        "# Normalize RGB\n",
        "train_images = train_images / 255.0\n",
        "test_images = test_images / 255.0\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "hmNs70BePfqn",
        "outputId": "17152e69-a8f1-40e7-9c15-ded7cb7f29b3"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://www.cs.toronto.edu/~kriz/cifar-10-python.tar.gz\n",
            "\u001b[1m170498071/170498071\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m8s\u001b[0m 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "model_list = []"
      ],
      "metadata": {
        "id": "w3xm28WT4Ptg"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Model 1\n",
        "3 layers of size 32, 64, 128 with 50% dropout"
      ],
      "metadata": {
        "id": "xj4y-nR1u1Kw"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model1 = Sequential([\n",
        "    Conv2D(32, (3, 3), activation='relu', input_shape=(img_width, img_height, img_color_channels)),\n",
        "    MaxPooling2D((2, 2)),\n",
        "    Conv2D(64, (3, 3), activation='relu'),\n",
        "    MaxPooling2D((2, 2)),\n",
        "    Conv2D(128, (3, 3), activation='relu'),\n",
        "    Flatten(),\n",
        "    Dense(128, activation='relu'),\n",
        "    Dropout(0.5),\n",
        "    Dense(10)\n",
        "])\n",
        "\n",
        "model1.compile(optimizer='adam',\n",
        "              loss=SparseCategoricalCrossentropy(from_logits=True),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "model1.fit(train_images, train_labels, epochs=10)\n",
        "model_list.append(model1)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ECEcTvbMPtUe",
        "outputId": "47d51e0c-a700-41e0-a037-7bbaabf8dcbf"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/keras/src/layers/convolutional/base_conv.py:107: UserWarning: Do not pass an `input_shape`/`input_dim` argument to a layer. When using Sequential models, prefer using an `Input(shape)` object as the first layer in the model instead.\n",
            "  super().__init__(activity_regularizer=activity_regularizer, **kwargs)\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m87s\u001b[0m 55ms/step - accuracy: 0.3209 - loss: 1.8212\n",
            "Epoch 2/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m136s\u001b[0m 51ms/step - accuracy: 0.5462 - loss: 1.2738\n",
            "Epoch 3/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m87s\u001b[0m 54ms/step - accuracy: 0.6168 - loss: 1.0978\n",
            "Epoch 4/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m81s\u001b[0m 52ms/step - accuracy: 0.6611 - loss: 0.9749\n",
            "Epoch 5/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m79s\u001b[0m 50ms/step - accuracy: 0.6895 - loss: 0.9019\n",
            "Epoch 6/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m81s\u001b[0m 50ms/step - accuracy: 0.7117 - loss: 0.8272\n",
            "Epoch 7/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m83s\u001b[0m 51ms/step - accuracy: 0.7320 - loss: 0.7735\n",
            "Epoch 8/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m81s\u001b[0m 50ms/step - accuracy: 0.7462 - loss: 0.7164\n",
            "Epoch 9/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m82s\u001b[0m 50ms/step - accuracy: 0.7593 - loss: 0.6846\n",
            "Epoch 10/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m85s\u001b[0m 52ms/step - accuracy: 0.7757 - loss: 0.6389\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Model 2\n",
        "3 layers of size 32, 64, 128"
      ],
      "metadata": {
        "id": "ys2AyMkKu3RX"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model2 = Sequential([\n",
        "    Conv2D(32, (3, 3), activation='relu', input_shape=(img_width, img_height, img_color_channels)),\n",
        "    MaxPooling2D((2, 2)),\n",
        "    Conv2D(64, (3, 3), activation='relu'),\n",
        "    MaxPooling2D((2, 2)),\n",
        "    Conv2D(128, (3, 3), activation='relu'),\n",
        "    Flatten(),\n",
        "    Dense(128, activation='relu'),\n",
        "    Dense(10)\n",
        "])\n",
        "\n",
        "model2.compile(optimizer='adam',\n",
        "              loss=SparseCategoricalCrossentropy(from_logits=True),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "model2.fit(train_images, train_labels, epochs=10, validation_data=(test_images, test_labels))\n",
        "model_list.append(model2)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rjWiZ93ik8ln",
        "outputId": "7a937593-e3f2-4bed-b1bb-a7c5ac4a0c5c"
      },
      "execution_count": 5,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m9s\u001b[0m 5ms/step - accuracy: 0.3750 - loss: 1.6897 - val_accuracy: 0.5809 - val_loss: 1.1706\n",
            "Epoch 2/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 3ms/step - accuracy: 0.6158 - loss: 1.0858 - val_accuracy: 0.6626 - val_loss: 0.9663\n",
            "Epoch 3/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.6871 - loss: 0.8899 - val_accuracy: 0.6842 - val_loss: 0.9312\n",
            "Epoch 4/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 3ms/step - accuracy: 0.7346 - loss: 0.7563 - val_accuracy: 0.6823 - val_loss: 0.9341\n",
            "Epoch 5/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m10s\u001b[0m 3ms/step - accuracy: 0.7684 - loss: 0.6713 - val_accuracy: 0.6871 - val_loss: 0.9034\n",
            "Epoch 6/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m9s\u001b[0m 3ms/step - accuracy: 0.7978 - loss: 0.5819 - val_accuracy: 0.7234 - val_loss: 0.8397\n",
            "Epoch 7/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 4ms/step - accuracy: 0.8203 - loss: 0.5091 - val_accuracy: 0.7258 - val_loss: 0.8645\n",
            "Epoch 8/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m9s\u001b[0m 3ms/step - accuracy: 0.8433 - loss: 0.4401 - val_accuracy: 0.7284 - val_loss: 0.8744\n",
            "Epoch 9/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.8654 - loss: 0.3802 - val_accuracy: 0.7230 - val_loss: 0.9450\n",
            "Epoch 10/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.8909 - loss: 0.3065 - val_accuracy: 0.7150 - val_loss: 1.0333\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Model 3\n",
        "3 layers of size 32, 64, 64"
      ],
      "metadata": {
        "id": "3rICdEbbu4dY"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model3 = Sequential([\n",
        "    Conv2D(32, (3, 3), activation='relu', input_shape=(img_width, img_height, img_color_channels)),\n",
        "    MaxPooling2D((2, 2)),\n",
        "    Conv2D(64, (3, 3), activation='relu'),\n",
        "    MaxPooling2D((2, 2)),\n",
        "    Conv2D(64, (3, 3), activation='relu'),\n",
        "    Flatten(),\n",
        "    Dense(64, activation='relu'),\n",
        "    Dense(10)\n",
        "])\n",
        "\n",
        "model3.compile(optimizer='adam',\n",
        "              loss=SparseCategoricalCrossentropy(from_logits=True),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "model3.fit(train_images, train_labels, epochs=10, validation_data=(test_images, test_labels))\n",
        "model_list.append(model3)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oHKlWrack_uR",
        "outputId": "c5a284c4-b6c4-4e1a-b163-6c2e94482809"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m9s\u001b[0m 4ms/step - accuracy: 0.3414 - loss: 1.7784 - val_accuracy: 0.5206 - val_loss: 1.3050\n",
            "Epoch 2/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m7s\u001b[0m 3ms/step - accuracy: 0.5657 - loss: 1.2223 - val_accuracy: 0.5993 - val_loss: 1.1566\n",
            "Epoch 3/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.6293 - loss: 1.0518 - val_accuracy: 0.6215 - val_loss: 1.0677\n",
            "Epoch 4/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.6683 - loss: 0.9387 - val_accuracy: 0.6688 - val_loss: 0.9521\n",
            "Epoch 5/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.7064 - loss: 0.8433 - val_accuracy: 0.6839 - val_loss: 0.8999\n",
            "Epoch 6/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 3ms/step - accuracy: 0.7285 - loss: 0.7836 - val_accuracy: 0.6983 - val_loss: 0.8818\n",
            "Epoch 7/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.7482 - loss: 0.7266 - val_accuracy: 0.7031 - val_loss: 0.8510\n",
            "Epoch 8/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.7634 - loss: 0.6784 - val_accuracy: 0.7010 - val_loss: 0.8770\n",
            "Epoch 9/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.7778 - loss: 0.6353 - val_accuracy: 0.7108 - val_loss: 0.8773\n",
            "Epoch 10/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m11s\u001b[0m 4ms/step - accuracy: 0.7896 - loss: 0.5997 - val_accuracy: 0.6982 - val_loss: 0.9202\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Model 4\n",
        "3 layers of size 32, 32, 64 with 1 less pooling layer"
      ],
      "metadata": {
        "id": "Sj3RPhVk1bXT"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model4 = Sequential([\n",
        "    Conv2D(32, (3, 3), activation='relu', input_shape=(img_width, img_height, img_color_channels)),\n",
        "    Conv2D(32, (3, 3), activation='relu'),\n",
        "    MaxPooling2D((2, 2)),\n",
        "    Conv2D(64, (3, 3), activation='relu'),\n",
        "    Flatten(),\n",
        "    Dense(64, activation='relu'),\n",
        "    Dense(10)\n",
        "])\n",
        "\n",
        "model4.compile(optimizer='adam',\n",
        "              loss=SparseCategoricalCrossentropy(from_logits=True),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "model4.fit(train_images, train_labels, epochs=10, validation_data=(test_images, test_labels))\n",
        "model_list.append(model4)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BDkr7dSI1dGG",
        "outputId": "de21e5e1-fd19-44f1-b9b8-95e029bd6a70"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m11s\u001b[0m 5ms/step - accuracy: 0.3955 - loss: 1.6506 - val_accuracy: 0.6185 - val_loss: 1.0801\n",
            "Epoch 2/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.6417 - loss: 1.0204 - val_accuracy: 0.6674 - val_loss: 0.9489\n",
            "Epoch 3/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.7160 - loss: 0.8197 - val_accuracy: 0.6931 - val_loss: 0.8685\n",
            "Epoch 4/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.7598 - loss: 0.6938 - val_accuracy: 0.6918 - val_loss: 0.9249\n",
            "Epoch 5/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m11s\u001b[0m 4ms/step - accuracy: 0.7885 - loss: 0.5969 - val_accuracy: 0.7095 - val_loss: 0.8587\n",
            "Epoch 6/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.8202 - loss: 0.5132 - val_accuracy: 0.7072 - val_loss: 0.8909\n",
            "Epoch 7/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.8483 - loss: 0.4284 - val_accuracy: 0.7106 - val_loss: 0.9392\n",
            "Epoch 8/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.8783 - loss: 0.3520 - val_accuracy: 0.6922 - val_loss: 1.0310\n",
            "Epoch 9/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m10s\u001b[0m 4ms/step - accuracy: 0.8951 - loss: 0.2966 - val_accuracy: 0.7058 - val_loss: 1.0978\n",
            "Epoch 10/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m10s\u001b[0m 4ms/step - accuracy: 0.9161 - loss: 0.2365 - val_accuracy: 0.6859 - val_loss: 1.2757\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Model 5\n",
        "2 layers of size 32, 64"
      ],
      "metadata": {
        "id": "fAu5vhpqTDwy"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "model6 = Sequential([\n",
        "    Conv2D(32, (3, 3), activation='relu', input_shape=(img_width, img_height, img_color_channels)),\n",
        "    MaxPooling2D((2, 2)),\n",
        "    Conv2D(64, (3, 3), activation='relu'),\n",
        "    Flatten(),\n",
        "    Dense(64, activation='relu'),\n",
        "    Dense(10)\n",
        "])\n",
        "\n",
        "model6.compile(optimizer='adam',\n",
        "              loss=SparseCategoricalCrossentropy(from_logits=True),\n",
        "              metrics=['accuracy'])\n",
        "\n",
        "model6.fit(train_images, train_labels, epochs=10, validation_data=(test_images, test_labels))\n",
        "model_list.append(model6)"
      ],
      "metadata": {
        "id": "DbYcw1FdTDig",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "000d1702-bbf2-4195-bd58-8b76fe2976cd"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m9s\u001b[0m 4ms/step - accuracy: 0.4016 - loss: 1.6496 - val_accuracy: 0.5950 - val_loss: 1.1496\n",
            "Epoch 2/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m8s\u001b[0m 4ms/step - accuracy: 0.6282 - loss: 1.0675 - val_accuracy: 0.6535 - val_loss: 1.0004\n",
            "Epoch 3/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.6876 - loss: 0.9016 - val_accuracy: 0.6693 - val_loss: 0.9664\n",
            "Epoch 4/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.7302 - loss: 0.7782 - val_accuracy: 0.6712 - val_loss: 0.9784\n",
            "Epoch 5/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.7531 - loss: 0.6998 - val_accuracy: 0.6621 - val_loss: 1.0118\n",
            "Epoch 6/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m4s\u001b[0m 3ms/step - accuracy: 0.7922 - loss: 0.6007 - val_accuracy: 0.6746 - val_loss: 1.0106\n",
            "Epoch 7/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.8201 - loss: 0.5129 - val_accuracy: 0.6772 - val_loss: 1.0527\n",
            "Epoch 8/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.8511 - loss: 0.4270 - val_accuracy: 0.6464 - val_loss: 1.2122\n",
            "Epoch 9/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m5s\u001b[0m 3ms/step - accuracy: 0.8735 - loss: 0.3638 - val_accuracy: 0.6586 - val_loss: 1.2181\n",
            "Epoch 10/10\n",
            "\u001b[1m1563/1563\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m6s\u001b[0m 3ms/step - accuracy: 0.8972 - loss: 0.2991 - val_accuracy: 0.6582 - val_loss: 1.3512\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Test & Compare"
      ],
      "metadata": {
        "id": "JgefWfmfNh8M"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.metrics import f1_score\n",
        "\n",
        "for i in range(len(model_list)):\n",
        "  print(f\"Model {i+1}:\")\n",
        "  model_list[i].evaluate(test_images, test_labels)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QAJA0V_ZBz3t",
        "outputId": "d929edd3-7518-4e69-abe9-01e6e4be48c4"
      },
      "execution_count": 10,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model 1:\n",
            "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 3ms/step - accuracy: 0.7349 - loss: 0.8042\n",
            "Model 2:\n",
            "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 2ms/step - accuracy: 0.7146 - loss: 1.0270\n",
            "Model 3:\n",
            "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 2ms/step - accuracy: 0.6996 - loss: 0.9054\n",
            "Model 4:\n",
            "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 2ms/step - accuracy: 0.6811 - loss: 1.2806\n",
            "Model 5:\n",
            "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 2ms/step - accuracy: 0.6314 - loss: 1.0750\n",
            "Model 6:\n",
            "\u001b[1m313/313\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m1s\u001b[0m 2ms/step - accuracy: 0.6612 - loss: 1.3511\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Save Model"
      ],
      "metadata": {
        "id": "GgIfUlT6onyz"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "best_model = model1\n",
        "best_model.save(\"CNN_CIFAR10.keras\")"
      ],
      "metadata": {
        "id": "WJCWOu3jongR"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Try Your Own Images"
      ],
      "metadata": {
        "id": "VpE3Ttg8Ni0v"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# This cell and below can run independantly of the above cells\n",
        "from tensorflow.keras.models import load_model\n",
        "from tensorflow.keras.preprocessing import image\n",
        "import numpy as np\n",
        "import matplotlib.pyplot as plt\n",
        "import matplotlib.cm as cm\n",
        "\n",
        "class_names = ['airplane', 'automobile', 'bird', 'cat', 'deer', 'dog', 'frog',\n",
        "               'horse', 'ship', 'truck']\n",
        "\n",
        "def preprocess(img_path):\n",
        "  \"\"\" Returns normalized and reshaped image array. \"\"\"\n",
        "\n",
        "  img = image.load_img(img_path, target_size=(32, 32))\n",
        "  img_arr = image.img_to_array(img)\n",
        "  img_arr /= 255.0  # Normalize RGB\n",
        "  img_arr_exp = np.expand_dims(img_arr, axis=0)  # 4D shape for prediction\n",
        "  return img_arr, img_arr_exp\n",
        "\n",
        "def predict(img_path, model, plot=True):\n",
        "  \"\"\" Predict class of image using model. Output is the image and a bar\n",
        "  plot of the prediction values of the image. If plot is false, output\n",
        "  is printed instead of plotted.\"\"\"\n",
        "\n",
        "  img_arr, img_arr_exp = preprocess(img_path)\n",
        "  predictions = model.predict(img_arr_exp)\n",
        "  final_prediction = np.argmax(predictions, axis=1)[0]\n",
        "\n",
        "  if plot:\n",
        "    plt.subplot(1, 2, 1)\n",
        "    plt.imshow(img_arr)\n",
        "    plt.title(f\"Prediction: {class_names[final_prediction]}\")\n",
        "\n",
        "    plt.subplot(1, 2, 2)\n",
        "    colors = cm.plasma(predictions[0] / np.max(predictions[0]))\n",
        "    plt.bar(class_names, predictions[0], color=colors)\n",
        "    plt.title(\"Prediction Values\")\n",
        "    plt.xticks(rotation=90)\n",
        "\n",
        "    plt.show()\n",
        "  else:\n",
        "    print(f\"Predicted class: {class_names[final_prediction]}\")\n",
        "    print(\"Prediction Values:\")\n",
        "    for i in range(len(predictions[0])):\n",
        "      print(f\"\\t{class_names[i]}: \\t{predictions[0][i]}\")\n"
      ],
      "metadata": {
        "id": "XQsrCOzINjT3"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### Predict Image"
      ],
      "metadata": {
        "id": "3_IqHvOdy_4_"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "myModel = load_model(\"CNN_CIFAR10.keras\")\n",
        "\n",
        "# Example below: horse is incorrectly predicted but is very close, and car is correct\n",
        "predict(img_path=\"car32.jpg\", model=myModel, plot=True)\n",
        "predict(img_path=\"horse32.jpg\", model=myModel, plot=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "dXdDfid0g7cV",
        "outputId": "f05d0afe-6b3e-4528-b682-eb19949b9de9"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 103ms/step\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAh8AAAH0CAYAAACdNk8NAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy81sbWrAAAACXBIWXMAAA9hAAAPYQGoP6dpAABsWklEQVR4nO3deVxU1f8/8NfMAMM+LLImKoLhFmpYSC65JeKetliaiKlZ7ugntUXTLMzKMDPNzydFLbO0T/orzR3N3dx3EsU0FcSFVWWZOb8//DIfx+FemWG4oLyej8c8Hs4995577p1B3px73ueohBACRERERApRV3YDiIiIqHph8EFERESKYvBBREREimLwQURERIpi8EFERESKYvBBREREimLwQURERIpi8EFERESKYvBBREREimLwQURUDdWpUwcDBw40vt+6dStUKhW2bt1qs3OoVCp88MEHNquvslTEvanuGHwQESksKSkJKpXK+HJ0dMTjjz+OESNGICMjo7KbZ5G1a9dWqQAjPDwctWrVgtzKIS1btoSfnx+Ki4sVbBndi8EHEVElmTZtGpYuXYqvvvoKzzzzDObNm4eoqCjcunVL8ba0adMGt2/fRps2bSw6bu3atZg6dWqpZbdv38Z7771ni+aVWb9+/XDx4kVs37691PLz589j9+7dePnll2FnZ6do2+h/GHwQEVWSmJgY9O/fH4MHD0ZSUhLGjBmDtLQ0rF69WvKY/Pz8CmmLWq2Go6Mj1Grb/VpwdHRU/Bf8q6++CpVKhWXLlpVa/sMPP0AIgX79+inaLjLF4IOIqIpo3749ACAtLQ0AMHDgQLi6uuLs2bPo0qUL3NzcjL80DQYDEhMT0ahRIzg6OsLPzw9vvPEGbt68aVKnEALTp09HzZo14ezsjHbt2uHEiRNm55Ya17B371506dIFnp6ecHFxQXh4OGbPnm1s39y5cwHA5DFSidLGfBw6dAgxMTFwd3eHq6srOnTogD179pjsU/JYaufOnYiPj4ePjw9cXFzw/PPPIzMzU/YeBgUFoU2bNli5ciWKiorMypctW4aQkBBERkbi77//xltvvYWwsDA4OTnB29sbL774Is6fPy97DsB8zEyJtm3bom3btibbCgoKMGXKFISGhkKr1SIoKAhvv/02CgoKTPbbuHEjWrVqBQ8PD7i6uiIsLAzvvPPOA9vyMGKfExFRFXH27FkAgLe3t3FbcXExoqOj0apVK3z22WdwdnYGALzxxhtISkpCXFwcRo0ahbS0NHz11Vc4dOgQdu7cCXt7ewDA5MmTMX36dHTp0gVdunTBwYMH0alTJxQWFj6wPRs3bkS3bt0QEBCA0aNHw9/fH6dOncJvv/2G0aNH44033sDly5exceNGLF269IH1nThxAq1bt4a7uzvefvtt2Nvb45tvvkHbtm2xbds2REZGmuw/cuRIeHp6YsqUKTh//jwSExMxYsQI/Pjjj7Ln6devH4YOHYr169ejW7duxu3Hjh3D8ePHMXnyZADAn3/+iV27dqFv376oWbMmzp8/j3nz5qFt27Y4efKk8V6Xh8FgQI8ePbBjxw4MHToUDRo0wLFjx/DFF1/gr7/+wqpVq4z3plu3bggPD8e0adOg1WqRmpqKnTt3lrsNVZIgIiJFLVq0SAAQmzZtEpmZmeLixYti+fLlwtvbWzg5OYl//vlHCCFEbGysACAmTpxocvz27dsFAPH999+bbF+3bp3J9qtXrwoHBwfRtWtXYTAYjPu98847AoCIjY01bktOThYARHJyshBCiOLiYhEcHCxq164tbt68aXKee+saPny4kPpVAkBMmTLF+L5Xr17CwcFBnD171rjt8uXLws3NTbRp08bs/nTs2NHkXGPHjhUajUZkZWWVer4SN27cEFqtVrzyyism2ydOnCgAiJSUFCGEELdu3TI7dvfu3QKAWLJkiXHb/fdGCCFq165tcv9KPPvss+LZZ581vl+6dKlQq9Vi+/btJvvNnz9fABA7d+4UQgjxxRdfCAAiMzNT9toeFXzsQkRUSTp27AgfHx8EBQWhb9++cHV1xS+//ILHHnvMZL8333zT5P2KFSug0+nw3HPP4dq1a8ZXREQEXF1dkZycDADYtGkTCgsLMXLkSJPHIWPGjHlg2w4dOoS0tDSMGTMGHh4eJmX31lVWer0eGzZsQK9evVC3bl3j9oCAALz66qvYsWMHcnJyTI4ZOnSoyblat24NvV6Pv//+W/Zcnp6e6NKlC/7f//t/xjEyQggsX74czZs3x+OPPw4AcHJyMh5TVFSE69evIzQ0FB4eHjh48KDF11iaFStWoEGDBqhfv77JZ1XyiK3ksyq5x6tXr4bBYLDJuasyBh9ERJVk7ty52LhxI5KTk3Hy5EmcO3cO0dHRJvvY2dmhZs2aJtvOnDmD7Oxs+Pr6wsfHx+SVl5eHq1evAoDxl3S9evVMjvfx8YGnp6ds20oeATVu3Lhc11giMzMTt27dQlhYmFlZgwYNYDAYcPHiRZPttWrVMnlf0ub7x7WUpl+/fsjPzzcO3t21axfOnz9vMtD09u3bmDx5MoKCgqDValGjRg34+PggKysL2dnZFl9jac6cOYMTJ06YfU4lAVDJZ/Xyyy+jZcuWGDx4MPz8/NC3b1/89NNPj2wgwjEfRESV5Omnn0bz5s1l99FqtWYZKAaDAb6+vvj+++9LPcbHx8dmbaxMGo2m1O1CZg6PEt26dYNOp8OyZcvw6quvYtmyZdBoNOjbt69xn5EjR2LRokUYM2YMoqKioNPpoFKp0Ldv3wf+0pfq/dHr9SbtNhgMeOKJJzBr1qxS9w8KCgJwtxfmjz/+QHJyMtasWYN169bhxx9/RPv27bFhwwbJe/GwYvBBRPSQCQkJwaZNm9CyZUuTRwf3q127NoC7f33f+6gjMzPzgb0HISEhAIDjx4+jY8eOkvuV9RGMj48PnJ2dkZKSYlZ2+vRpqNVq4y9iW9BqtXjhhRewZMkSZGRkYMWKFWjfvj38/f2N+6xcuRKxsbH4/PPPjdvu3LmDrKysB9bv6elZ6n5///23yb0OCQnBkSNH0KFDhwfeK7VajQ4dOqBDhw6YNWsWPv74Y7z77rtITk6W/QweRnzsQkT0kHnppZeg1+vx4YcfmpUVFxcbfyl27NgR9vb2mDNnjklvQWJi4gPP8eSTTyI4OBiJiYlmv2TvrcvFxQUAHvgLW6PRoFOnTli9erVJKmtGRgaWLVuGVq1awd3d/YHtskS/fv1QVFSEN954A5mZmWZze2g0GrNelDlz5kCv1z+w7pCQEOzZs8cka+i3334ze3T00ksv4dKlS/j3v/9tVsft27eNY1Ju3LhhVt60aVMAMEvJfRSw54OI6CHz7LPP4o033kBCQgIOHz6MTp06wd7eHmfOnMGKFSswe/ZsvPDCC/Dx8cH48eORkJCAbt26oUuXLjh06BB+//131KhRQ/YcarUa8+bNQ/fu3dG0aVPExcUhICAAp0+fxokTJ7B+/XoAQEREBABg1KhRiI6ONnu0ca/p06cb57J46623YGdnh2+++QYFBQWYOXOmbW8S7t6nmjVrYvXq1XByckLv3r1Nyrt164alS5dCp9OhYcOG2L17NzZt2mSS6ixl8ODBWLlyJTp37oyXXnoJZ8+exXfffWfsMSrx2muv4aeffsKwYcOQnJyMli1bQq/X4/Tp0/jpp5+wfv16NG/eHNOmTcMff/yBrl27onbt2rh69Sq+/vpr1KxZE61atbLpfakSKjXXhoioGipJJf3zzz9l94uNjRUuLi6S5QsWLBARERHCyclJuLm5iSeeeEK8/fbb4vLly8Z99Hq9mDp1qggICBBOTk6ibdu24vjx42apoqWlkwohxI4dO8Rzzz0n3NzchIuLiwgPDxdz5swxlhcXF4uRI0cKHx8foVKpTNJucV+qrRBCHDx4UERHRwtXV1fh7Ows2rVrJ3bt2lWm+yPVRjn/+te/BADx0ksvmZXdvHlTxMXFiRo1aghXV1cRHR0tTp8+XeZ78/nnn4vHHntMaLVa0bJlS7F//36zVFshhCgsLBSffPKJaNSokdBqtcLT01NERESIqVOniuzsbCGEEJs3bxY9e/YUgYGBwsHBQQQGBopXXnlF/PXXX2W+1oeJSogyjNwhIiIishGO+SAiIiJFMfggIiIiRTH4ICIiIkUx+CAiIiJFMfggIiIiRTH4ICIiIkVxkjEieiQYDAZcvnwZbm5uVq26SkTlI4RAbm4uAgMDzdYjuh+DDyJ6JFy+fNmma4MQkXUuXrxothLz/Rh8ENEjwc3NDcDd//hsvUYIET1YTk4OgoKCjD+Lchh8ENEjoeRRi7u7O4MPokpUlseeHHBKREREimLwQURERIpi8EFERESKYvBBREREimLwQURERIpi8EFERESKYvBBREREimLwQURERIpi8EFERESKYvBBREREimLwQURERIpi8EFERESKYvBBREREimLwQURERIqyq+wGEBE9CrIL65a7Dp3DORu0hKjqY88HERERKYrBBxERESmKwQcREREpisEHERERKYrBBxERESmKwQcREREpisEHERERKYrBBxERESmKwQcREREpisEHERERKYrBBxERESmKwQcREREpisEHERERKYrBBxERESmKwQcREREpisEHERERKYrBBxERESmKwQcREREpisEHERERKYrBBxERESmKwQcREREpisEHERERKYrBBxFVuD/++APdu3dHYGAgVCoVVq1aZVIuhMDkyZMREBAAJycndOzYEWfOnKmcxhJRhWPwQUQVLj8/H02aNMHcuXNLLZ85cya+/PJLzJ8/H3v37oWLiwuio6Nx584dhVtKREqwq+wGENGjLyYmBjExMaWWCSGQmJiI9957Dz179gQALFmyBH5+fli1ahX69u2rZFOJSAHs+SCiSpWWlob09HR07NjRuE2n0yEyMhK7d++WPK6goAA5OTkmLyJ6ODD4IKJKlZ6eDgDw8/Mz2e7n52csK01CQgJ0Op3xFRQUVKHtJCLbYfBBRA+lSZMmITs72/i6ePFiZTeJiMqIwQcRVSp/f38AQEZGhsn2jIwMY1lptFot3N3dTV5E9HBg8EFElSo4OBj+/v7YvHmzcVtOTg727t2LqKioSmwZEVUUZrsQUYXLy8tDamqq8X1aWhoOHz4MLy8v1KpVC2PGjMH06dNRr149BAcH4/3330dgYCB69epVeY0mogrD4IOIKtz+/fvRrl074/v4+HgAQGxsLJKSkvD2228jPz8fQ4cORVZWFlq1aoV169bB0dGxsppMRBWIj10UUqdOHQwcOND4fuvWrVCpVNi6davNzqFSqfDBBx/YrD6yXsnnu3LlygfuO3DgQNSpU8dk26P2WbZt2xZCCLNXUlISgLvXO23aNKSnp+POnTvYtGkTHn/88cptNBFVmGoRfCQlJUGlUhlfjo6OePzxxzFixAizQW5V3dq1ax+pX0qWuHz5Mj744AMcPny4sptCRETlUK0eu0ybNg3BwcG4c+cOduzYgXnz5mHt2rU4fvw4nJ2dFW1LmzZtcPv2bTg4OFh03Nq1azF37txSA5Dbt2/Dzu7R/UgvX76MqVOnok6dOmjatGllN8dm/v3vf8NgMFR2M4iIFPPo/qYqRUxMDJo3bw4AGDx4MLy9vTFr1iysXr0ar7zySqnH5Ofnw8XFxeZtUavVNn+ezefjDyd7e/vKbgIRkaKqxWMXKe3btwdwd+Q9cPfZu6urK86ePYsuXbrAzc0N/fr1AwAYDAYkJiaiUaNGcHR0hJ+fH9544w3cvHnTpE4hBKZPn46aNWvC2dkZ7dq1w4kTJ8zOLTXmY+/evejSpQs8PT3h4uKC8PBwzJ4929i+koW57n2MVKK0cQKHDh1CTEwM3N3d4erqig4dOmDPnj0m+5Q8ltq5cyfi4+Ph4+MDFxcXPP/888jMzDTZNzs7G6dPn0Z2dvYD7+/q1avRtWtXBAYGQqvVIiQkBB9++CH0er3JfvePhynRtm1btG3b1ni/nnrqKQBAXFyc8dpLxgwAwIoVKxAREQEnJyfUqFED/fv3x6VLl0zqLPmML1y4gG7dusHV1RWPPfaY8b4eO3YM7du3h4uLC2rXro1ly5aZtevcuXN48cUX4eXlBWdnZ7Ro0QJr1qwp9R7o9Xq888478Pf3h4uLC3r06GE2GVZpYz5Kc+nSJQwaNAh+fn7QarVo1KgRFi5c+MDjiIiqmmrV83G/s2fPAgC8vb2N24qLixEdHY1WrVrhs88+Mz6OeeONN5CUlIS4uDiMGjUKaWlp+Oqrr3Do0CHs3LnT+Nfr5MmTMX36dHTp0gVdunTBwYMH0alTJxQWFj6wPRs3bkS3bt0QEBCA0aNHw9/fH6dOncJvv/2G0aNH44033sDly5exceNGLF269IH1nThxAq1bt4a7uzvefvtt2Nvb45tvvkHbtm2xbds2REZGmuw/cuRIeHp6YsqUKTh//jwSExMxYsQI/Pjjj8Z9fvnlF8TFxWHRokWlBgz3SkpKgqurK+Lj4+Hq6ootW7Zg8uTJyMnJwaeffvrA9t+rQYMGmDZtGiZPnoyhQ4eidevWAIBnnnnGeK64uDg89dRTSEhIQEZGBmbPno2dO3fi0KFD8PDwMNal1+sRExODNm3aYObMmfj+++8xYsQIuLi44N1330W/fv3Qu3dvzJ8/HwMGDEBUVBSCg4MB3J346plnnsGtW7cwatQoeHt7Y/HixejRowdWrlyJ559/3qTdH330EVQqFSZMmICrV68iMTERHTt2xOHDh+Hk5FTm68/IyECLFi2gUqkwYsQI+Pj44Pfff8frr7+OnJwcjBkzxqL7SURUqUQ1sGjRIgFAbNq0SWRmZoqLFy+K5cuXC29vb+Hk5CT++ecfIYQQsbGxAoCYOHGiyfHbt28XAMT3339vsn3dunUm269evSocHBxE165dhcFgMO73zjvvCAAiNjbWuC05OVkAEMnJyUIIIYqLi0VwcLCoXbu2uHnzpsl57q1r+PDhQupjAyCmTJlifN+rVy/h4OAgzp49a9x2+fJl4ebmJtq0aWN2fzp27GhyrrFjxwqNRiOysrLM9l20aFGpbbjXrVu3zLa98cYbwtnZWdy5c8e4rXbt2ib3psSzzz4rnn32WeP7P//8s9RzFxYWCl9fX9G4cWNx+/Zt4/bffvtNABCTJ082biv5jD/++GPjtps3bwonJyehUqnE8uXLjdtPnz5tdk/HjBkjAIjt27cbt+Xm5org4GBRp04dodfrhRD/+3wfe+wxkZOTY9z3p59+EgDE7NmzTdpUu3Ztk2u6/7yvv/66CAgIENeuXTPZr2/fvkKn05V6r6ub7OxsAUBkZ2dXyvmzCoLL/SKqCLkJXcr9KgtLfgar1WOXjh07wsfHB0FBQejbty9cXV3xyy+/4LHHHjPZ78033zR5v2LFCuh0Ojz33HO4du2a8RUREQFXV1ckJycDADZt2oTCwkKMHDnS5HFIWf4qPXToENLS0jBmzBiTv9IBmNRVVnq9Hhs2bECvXr1Qt25d4/aAgAC8+uqr2LFjh9kqoEOHDjU5V+vWraHX6/H3338btw0cOBBCiAf2egAw+cs+NzcX165dQ+vWrXHr1i2cPn3a4muSsn//fly9ehVvvfWWybiXrl27on79+qU+Ehk8eLDx3x4eHggLC4OLiwteeukl4/awsDB4eHjg3Llzxm1r167F008/jVatWhm3ubq6YujQoTh//jxOnjxpcp4BAwbAzc3N+P6FF15AQEAA1q5dW+brE0Lg559/Rvfu3SGEMPkORkdHIzs7GwcPHixzfUREla1aPXaZO3cuHn/8cdjZ2cHPzw9hYWFQq03jLzs7O9SsWdNk25kzZ5CdnQ1fX99S67169SoAGH9J16tXz6Tcx8cHnp6esm0reQTUuHHjsl+QjMzMTNy6dQthYWFmZQ0aNIDBYMDFixfRqFEj4/ZatWqZ7FfS5vvHtZTViRMn8N5772HLli1mgU5ZxoyUVcl9L+1a69evjx07dphsc3R0hI+Pj8k2nU6HmjVrmgV6Op3O5Pr//vtvs8dVwN17WlJ+72d4/3dBpVIhNDQU58+fL8OV3ZWZmYmsrCwsWLAACxYsKHWfku8gEdHDoFoFH08//bQx20WKVqs1C0gMBgN8fX3x/fffl3rM/b/IHlYajabU7UIIi+vKysrCs88+C3d3d0ybNg0hISFwdHTEwYMHMWHCBJPUUqmeHb1eL9mm8pCq05bXb0sl96p///6IjY0tdZ/w8HAlm0REVC7VKviwVkhICDZt2oSWLVvKDhKsXbs2gLs9Jfc+6sjMzHxg70FISAgA4Pjx4+jYsaPkfmV9BOPj4wNnZ2ekpKSYlZ0+fRpqtRpBQUFlqssaW7duxfXr1/Hf//4Xbdq0MW4vySy6l6enJ7Kyssy2//333yb3UeraS+57SkqKMYOpREpKirHcFmrXri15T+9tS4kzZ86YvBdCIDU11aJgwcfHB25ubtDr9bLfDSKih0W1GvNhrZdeegl6vR4ffvihWVlxcbHxF2fHjh1hb2+POXPmmPy1nJiY+MBzPPnkkwgODkZiYqLZL+J76yqZc6S0X9b30mg06NSpE1avXm3SxZ+RkYFly5ahVatWVi1BXtZU25JehHvbXlhYiK+//tps35CQEOzZs8ckI+i3334zS0mVuvbmzZvD19cX8+fPR0FBgXH777//jlOnTqFr165lu7gy6NKlC/bt24fdu3cbt+Xn52PBggWoU6cOGjZsaLL/kiVLkJuba3y/cuVKXLlyBTExMWU+p0ajQZ8+ffDzzz/j+PHjZuX3p0MTEVV17Pkog2effRZvvPEGEhIScPjwYXTq1An29vY4c+YMVqxYgdmzZ+OFF16Aj48Pxo8fj4SEBHTr1g1dunTBoUOH8Pvvv6NGjRqy51Cr1Zg3bx66d++Opk2bIi4uDgEBATh9+jROnDiB9evXAwAiIiIAAKNGjUJ0dDQ0Gg369u1bap3Tp0/Hxo0b0apVK7z11luws7PDN998g4KCAsycOdOqe1HWVNtnnnkGnp6eiI2NxahRo6BSqbB06dJSH2EMHjwYK1euROfOnfHSSy/h7Nmz+O6774y9QSVCQkLg4eGB+fPnw83NDS4uLoiMjERwcDA++eQTxMXF4dlnn8Urr7xiTLWtU6cOxo4da9W1lmbixIn44YcfEBMTg1GjRsHLywuLFy9GWloafv75Z7NHdl5eXmjVqhXi4uKQkZGBxMREhIaGYsiQIRadd8aMGUhOTkZkZCSGDBmChg0b4saNGzh48CA2bdqEGzdu2OwaiYgqGoOPMpo/fz4iIiLwzTff4J133oGdnR3q1KmD/v37o2XLlsb9pk+fDkdHR8yfP9/4y2LDhg1l+us7OjoaycnJmDp1Kj7//HMYDAaEhISY/KLq3bs3Ro4cieXLl+O7776DEEIy+GjUqBG2b9+OSZMmISEhAQaDAZGRkfjuu+9KHTRpS97e3vjtt98wbtw4vPfee/D09ET//v3RoUMHREdHm133559/jlmzZmHMmDFo3ry58dh72dvbY/HixZg0aRKGDRuG4uJiLFq0CMHBwRg4cCCcnZ0xY8YMTJgwwThJ2ieffGKWPVQefn5+2LVrFyZMmIA5c+bgzp07CA8Px6+//lrqZ/zOO+/g6NGjSEhIQG5uLjp06ICvv/7a4un8/fz8sG/fPkybNg3//e9/8fXXX8Pb2xuNGjXCJ598YqvLIyJShEpU9mg6IiIbyMnJgU6nQ3Z2tlWPFMsru7Dug3d6AJ3DuQfvRGShvBnlf/TsOrH0WZzvZcnPIMd8EBERkaIYfBAREZGiGHwQERGRohh8EBERkaIYfBAREZGiGHwQERGRoipsno+5c+fi008/RXp6Opo0aYI5c+bg6aeffuBxBoMBly9fhpubm1WruRJR+QkhkJubi8DAQLOJ04iIyqtCgo8ff/wR8fHxmD9/PiIjI5GYmIjo6GikpKRIrgxb4vLlyxW65ggRld3FixfNVnkmIiqvCgk+Zs2ahSFDhiAuLg7A3dlB16xZg4ULF2LixImyx7q5uQEAPBxdSu35EDKdIcUyf6HpZY7Ty/xhJzcDm1xbVDIHah4wrZvGIF2mljlWrszaPiShkm6MTDNhUEnfVINMY+Tujdw9tfba1XIXIUPuvtwtt65M9ntq7XdYri1S24WAPjvf+PNIRGRLNg8+CgsLceDAAUyaNMm4Ta1Wo2PHjiaLcUkpCThUKhXUpQQfcr+45B7TyD3Bsfrpjlyd1h12t7wC6rX6AZbcPZU9zMrPQq7OCihTW3ljxAO+NHLfU9nvTQWUWf/hl30VZSIiS9g8+Lh27Rr0ej38/PxMtvv5+RmXHb9XQUGByUqkOTk5tm4SERERVSGVPpIsISEBOp3O+OJ4DyIiokebzYOPGjVqQKPRICMjw2R7RkYG/P39zfafNGkSsrOzja+LFy/auklERERUhdg8+HBwcEBERAQ2b95s3GYwGLB582ZERUWZ7a/VauHu7m7yIiIiokdXhWS7xMfHIzY2Fs2bN8fTTz+NxMRE5OfnG7NfysIBaqhLGSknl2MgF0npZcqKZY6UzfaQyeiAzHFqIR/zqWWOVckdK6SPE1Zeo+xdtXIsotzVy12fSu7Tt/I4ITuHhbXfNkAlc0/lPguVTPqNWu4zlDlOyHxPpT57ma8SEVG5VUjw8fLLLyMzMxOTJ09Geno6mjZtinXr1pkNQiUiIqLqp8JmOB0xYgRGjBhRUdUTERHRQ6rSs12IiIioemHwQURERIpi8EFERESKYvBBREREiqqwAafl5VBYXOraLnKLZGlkQqkiuYxRuQXpZFIYVbKLp8mkPj4g1bZYLvVV7lgr0ztl22plOq21i8Dp5dJQZVJf5dJQVbJ5o9am2j6IlZ+FlenSMNg2zVqIB6x+SERUDuz5ICIiIkUx+CAiIiJFMfggIiIiRTH4ICIiIkUx+CAiIiJFMfggIiIiRVXZVFuDgwYoJdXWIJP6KVdWSlVGcgubCpXGqjo1kD5Orp0lR0uRS1M1yLRVjrCyrXJtkbs3cmmoatncXtuXqeTOJ38RD2DdZ2HtcXKfoTXnExAosrJGa3zwwQeYOnWqybawsDCcPn1awVYQkVKqbPBBRNVLo0aNsGnTJuN7Ozv+90T0qOJPNxFVCXZ2dvD396/sZhCRAjjmg4iqhDNnziAwMBB169ZFv379cOHCBdn9CwoKkJOTY/IioocDgw8iqnSRkZFISkrCunXrMG/ePKSlpaF169bIzc2VPCYhIQE6nc74CgoKUrDFRFQeDD6IqNLFxMTgxRdfRHh4OKKjo7F27VpkZWXhp59+kjxm0qRJyM7ONr4uXryoYIuJqDw45oOIqhwPDw88/vjjSE1NldxHq9VCq9Uq2CoispUqG3ysO7YPbm5uFh1j7Qqs1q5eKnc62TTUBywYqrHyWLnj5FbDlWurtQwyqwED0icUcndV7uJlU2atPE5GeZJw5Vj7WTzoO2WpnNxcBDdsZNtKLZCXl4ezZ8/itddeq7Q2EFHF4WMXIqp048ePx7Zt23D+/Hns2rULzz//PDQaDV555ZXKbhoRVYAq2/NBRNXHP//8g1deeQXXr1+Hj48PWrVqhT179sDHx6eym0ZEFYDBBxFVuuXLl1d2E4hIQXzsQkRERIpizwcREVEVsTBgYbnrGHRlkA1aUrHY80FERESKsnnPh61Wp7RzcYadq0spJdIpnCq5dFLZ46TboZJdgVXmOJmyB6VTypXLlWkM0tdvbWqo3D0VKul7I1dmkLk51qbaGmSOE7J5qFam2pYjtVUtc09lv2+yn691PxdSP0/FRcUyxxARlU+FPHbh6pREREQkpUKiAq5OSURERFIqZMyHJatTcmVKIiKi6sXmwYelq1NyZUoiIqLqxebBh6WrU3JlSiIiouqlwkeCPmh1Sq5MSUREVL1UePBh7eqUjnoNHPUas+0qmVVI5VMYpY+TX0XW9uuXlqdKmWxLFKvkUo2tI59qK30hBtkyuTRc61J7DTJXKHdcRU11I59qbeX3VDbV1rqfC6lU42INpwAioopj8/9huDolERERybF5zwdXpyQiIiI5Ng8+uDolERERyeGDXSIiIlIUgw8iIiJSFIMPIiIiUlTVXfFNqEtN87R6xVeZdEONQWZFUNn4TG5VV+mj9DI1PuhYIZdSKtNUg2y6pTSZjFlYG7vKpqHKHKeXK7WynXL3szyETMqwHLm0WL3MNcodJ/d9kkqlNlTh/xqI6OHHng8iIiJSFIMPIiIiUhSDDyIiIlIUgw8iIiJSFIMPIiIiUhSDDyIiIlJUlc2nK4RAAcxzMp2dnSSPybt2U7JM5yR9nFotsxqsQTovNC8vX7LMw99Xsuz2nduSZQBQIKSTcVV25iv9ltDIlBXJXIdabrVUmVzb4oJCyTKtZAng5CD9WeTekr6nTloHyTK9zPXJrQTs4uYuWZaZeU3mOFfpSgHAIH3fRCnf6/8dJ5O+rZdLCZemkSmV+nw1kP4uERGVF3s+iIiISFEMPoiIiEhRDD6IiIhIUQw+iIiISFEMPoiIiEhRDD6IiIhIUQw+iIiISFFVdp4PtUYFtcZ8DgI7tfT8Az41vCTL7GWOy714WbLMzcVFsszD11+y7E5GhmSZk6+PZBkAOOmLJMtuy8ytoZcpc3VwlizTaKTvjaFAes4RfbH0/BEOMnOOqIuLJct0OunPEEXS15d/+5ZkmZ299PwghvwCybIa7h6SZXLznwBAgcw1FhVJl8nN2GEvcx12dtI/ynLzg+j1Ep+vzM8LEVF5seeDiIiIFMXgg4iIiBTF4IOIiIgUVWXHfBARVZQW9gvKXceeoqE2aAlR9cSeDyIiIlKUxcHHH3/8ge7duyMwMBAqlQqrVq0yKRdCYPLkyQgICICTkxM6duyIM2fO2Kq9RERE9JCz+LFLfn4+mjRpgkGDBqF3795m5TNnzsSXX36JxYsXIzg4GO+//z6io6Nx8uRJODo6lvk8+jt3oLe3N9teVCSdhqqRWVa9OFd6qXY3Z+ll1WEv3eaioyclyxwD/aTrvHZDugwAHKQXpHe6fUeyrPiOdJkG0qmoQirdEoC9Rrot9j41JMsgnRUKw61syTK1VuY7IpNq66SWjqPVMimqeTnSbXFxdZUsKzDIpcsCapm4Xusg/WNnZy99vzX20mm4xdJffRQVSn++hfrSr6MA8tdHRFQeFgcfMTExiImJKbVMCIHExES899576NmzJwBgyZIl8PPzw6pVq9C3b9/ytZaIiIgeejYd85GWlob09HR07NjRuE2n0yEyMhK7d+8u9ZiCggLk5OSYvIioepo7dy7q1KkDR0dHREZGYt++fZXdJCKqADYNPtLT0wEAfn6mjxz8/PyMZfdLSEiATqczvoKCgmzZJCJ6SPz444+Ij4/HlClTcPDgQTRp0gTR0dG4evVqZTeNiGys0rNdJk2ahOzsbOPr4sWLld0kIqoEs2bNwpAhQxAXF4eGDRti/vz5cHZ2xsKFCyu7aURkYzYNPvz97651knHfuiYZGRnGsvtptVq4u7ubvIioeiksLMSBAwdMHtmq1Wp07NhR8pEtET28bBp8BAcHw9/fH5s3bzZuy8nJwd69exEVFWXLUxHRI+TatWvQ6/UWPbLleDGih5fF2S55eXlITU01vk9LS8Phw4fh5eWFWrVqYcyYMZg+fTrq1atnTLUNDAxEr169LDqPn6sT3N3MV2JV25mn35YQGdIprCqDdLph9vEjkmWHdu+RLDt/NlWyLO38ecmyGoHSq+ECQJFaOm/yRp50yrBKZnVaJ0fp1XlvXL0uWZafc1uyzNvfV7LM4Cz9OQWG1pIse/bZZyXLGjZsKFmm8ZJJ+5XhYicdf6tkFnZ1lEntvXuwdHmhkP4u3irIlSy7ky+dZg6N9PkcnKRXNHZwdyp9O2TOVUUkJCRg6tSpVh1bEbOT6hzO2bxOd6cZ5a4j5/bEalnn/fU9LHUOujKoXPWVxnXiGpvXWV4WBx/79+9Hu3btjO/j4+MBALGxsUhKSsLbb7+N/Px8DB06FFlZWWjVqhXWrVtn0RwfRFS91KhRAxqNxqJHtpMmTTL+/wPc7WXlgHWih4PFwUfbtm0hhPRf5iqVCtOmTcO0adPK1TAiqj4cHBwQERGBzZs3G3tJDQYDNm/ejBEjRpR6jFarhVYrPSkbEVVdXFiOiKqE+Ph4xMbGonnz5nj66aeRmJiI/Px8xMXFVXbTiMjGGHwQUZXw8ssvIzMzE5MnT0Z6ejqaNm2KdevWmQ1CJaKHH4MPIqoyRowYIfmYhYgeHZU+yRgRERFVL1W35yM3C0ApKYmlrHRbQlUknRa6Okl6lsRLKWckyzQyK76eOZMiWeZTUzqddu/h7ZJlAOARGCBZ5uLnI1nmqJNJp83Lkyw7lXVeujHF0iupenpKt7NQ5rM4fvSAZNmuHcmSZc2bN5cs69Gth2RZ6OOPS5ap3HSSZSiUTmvWS6wGa6xXZiCkWiv9Y+ekkV6BV6OS/u4XwSBTJp02ewelrxScB+lrJyIqL/Z8EBERkaIYfBAREZGiGHwQERGRohh8EBERkaIYfBAREZGiGHwQERGRoqpsqq26sBjqQvN0xrwjJyWP+WXp95JlN/++IFnWvW17ybLsa1cly26fOi1ZduFP6XTSgpvSq8gCwO3b0mmOdnnSq57edpBOxbySkSlZln72omSZp7tMKqqv9Kq2cqu+2hVLp37+c1I6fXn7GelVQ28ePi5ZFh7eVLKse/fukmWO4U0kyzQyKd8AgCLpVFz1Hem0Z2ikU5vt7KWX2S2yky67Y5BuS4FEmSafqbZEVHHY80FERESKYvBBREREimLwQURERIpi8EFERESKYvBBREREimLwQURERIqqsqm2uJAJuJivjPrDhCmSh+iEdCwV5V1DsszhoHR6Z22NdJ09atSRLLvjIb2q7dXb0umyAHBNpjzzknTZHb30yqZ1NdKpoZFedaTrzJNenfbWRul04mJn6dVZ3bzcJcuaFEofV5Rb+gqsAOB4LE2y7Mwx6TTrL5b9KlnmqvOQLAtt3ECyDABcfaW/bwEhdaTLQutKlrnUekyyTOvtJV3m6iRZBvvSyxxlVsIlIiov9nwQERGRohh8EBERkaIYfBAREZGiGHwQERGRohh8EBERkaIsDj7++OMPdO/eHYGBgVCpVFi1apVJ+cCBA6FSqUxenTt3tlV7iYiI6CFncaptfn4+mjRpgkGDBqF3796l7tO5c2csWrTI+F6r1VresosZgLN5WqlL3h3JQ1zV0ucJ9pBOfawRECjdDiGdvur5mHTqoyEnS7LM7dIl6fMBiKgTKlnm6uEpfU4IyTK9Xi9Z5qCWSW/NkV7d9EpGhmSZi4906qeTu6tkmZ1Gui03btyQLHN2k07fhYN0nX9fvix9mKf0ir7CyVH6fAAyr2dJlu06tVa6PTL31GAn/bdCg/DGkmWdoqMlyzxaRJZekCufDk5EVB4WBx8xMTGIiYmR3Uer1cLfX3qeCyIiIqq+KmTMx9atW+Hr64uwsDC8+eabuH79uuS+BQUFyMnJMXkRERHRo8vmwUfnzp2xZMkSbN68GZ988gm2bduGmJgYyW7/hIQE6HQ64ysoKMjWTSIiIqIqxObTq/ft29f47yeeeALh4eEICQnB1q1b0aFDB7P9J02ahPj4eOP7nJwcBiBERESPsApPta1bty5q1KiB1NTUUsu1Wi3c3d1NXkRERPToqvDg459//sH169cREBBQ0aciIiKih4DFj13y8vJMejHS0tJw+PBheHl5wcvLC1OnTkWfPn3g7++Ps2fP4u2330ZoaCiiZdL9SmWvBezN0xnv5N6SPCRfJvsxA9Ips7tPSq/OejXnpmRZzu08ybLa3j6SZQ7X5NMY7c6elSwTt6VXmVXppVd9tbNXSZY5qaRjUOdi6ftmXyydvpurkj4f1NJfOyHzlSyQSXu+pdFIluldXCTLVDKrweqLpb9r8JZOwwWA8zevSZYVaqVTokOfqCdZ9pi3r2TZpVTpVX03J30vWdanWKLgtsy1ExGVk8XBx/79+9GuXTvj+5LxGrGxsZg3bx6OHj2KxYsXIysrC4GBgejUqRM+/PBD6+b6ICIiokeOxcFH27ZtIYT0X27r168vV4OIiIjo0ca1XYiIiEhRDD6IiIhIUQw+iIiISFEMPoiIiEhRNp/h1GZu3QaEebqm2sND8pDrhQWSZYvW/y5Zlm4nnaJa4Cidwlkss4psccpJybIn3eUX3fMt5bpLeGukPzJnrb1kWaFKKqcSuF0gnTKcd0t6FWFNgUxqr5C+b2q1TMxrJ7PCrkb6+vLV0ufLKpBOG72RI50Se+0fySKk5WVLFwLQuzhJFzpIf4aO9tLX7y6kj2sYVFuyrCBdejXgG+dKT9HNKZD+3ImIyos9H0RERKQoBh9ERESkKAYfREREpCgGH0RERKQoBh9ERESkKAYfRFTp6tSpA5VKZfKaMWNGZTeLiCpIlU21vZ6fhUKDeersBbX0Sqo5eun0wMO5mZJljdq1lCxzD5JOi5VOtAXSjhyTLNtx5JTMkUCInfQqrKH+fpJlAd4ekmWOLtIpnI4q6SuRW7nWQWblWo3MyrXFxdJpv4VF0m25JbPCbtadIsmyC7nSabFnZVYt/ue6zKrFEU0lywDAp05NyTI7R+lFFg2F0texa/0W6To9pFfZvXH1smRZL7fSv2t59sr/XTJt2jQMGTLE+N7NzU3xNhCRMqps8EFE1Yubmxv8/eXnwCGiRwMfuxBRlTBjxgx4e3ujWbNm+PTTT2V7yACgoKAAOTk5Ji8iejiw54OIKt2oUaPw5JNPwsvLC7t27cKkSZNw5coVzJo1S/KYhIQETJ06VcFWEpGtsOeDiCrExIkTzQaR3v86ffo0ACA+Ph5t27ZFeHg4hg0bhs8//xxz5sxBQYH0kgmTJk1Cdna28XXx4kWlLo2Iyok9H0RUIcaNG4eBAwfK7lO3bt1St0dGRqK4uBjnz59HWFhYqftotVpotdKDd4mo6mLwQUQVwsfHBz4+PlYde/jwYajVavj6+tq4VURUFVTZ4MO7ZSTcXV3NtmcvWyJ5zLFLFyTLinxrSJapaz4mWXbosnRXbrGQTv0MCg2RLDNclE77BYA8e+m/5v6xl1699dot6dTQ4jzp7uviwtuSZQ4F0qmfdjLJxnl3pOvUy6zaq1JJX5/GwVGyTNhLl93SSD9dvOEundZ8x1kmJVbm+wQAafnSK+lmp6dLltWtFSRZlu8kfY0X7kinmRc6SK8GXODpWfoxt6XPZWu7d+/G3r170a5dO7i5uWH37t0YO3Ys+vfvD0+J9hHRw63KBh9EVD1otVosX74cH3zwAQoKChAcHIyxY8ciPj6+sptGRBWEwQcRVaonn3wSe/bsqexmEJGCmO1CREREimLwQURERIpi8EFERESKsij4SEhIwFNPPQU3Nzf4+vqiV69eSElJMdnnzp07GD58OLy9veHq6oo+ffogIyPDpo0mIiKih5dFA063bduG4cOH46mnnkJxcTHeeecddOrUCSdPnoSLy92UxbFjx2LNmjVYsWIFdDodRowYgd69e2Pnzp0WNUzv7gx9KStu9hkzQvKYmvv2S5atTd4kWXbwonSKrr1MKmberXzJssy/pVN0cx6wZkVekXQK76UC6ZRKrcxKpMJOuk4HtfRxjvbSaZoajXTKrH0ND8kyvXRTUCyzcm1xkfQKuwWyq+FKpwvnCOnj8mVW7b1zRTpdFgBqBEovkHbLUTqN9Wqh9DXekPlbQcik9rZt2Uqy7PHWpa/onJOXK3kMEVF5WRR8rFu3zuR9UlISfH19ceDAAbRp0wbZ2dn49ttvsWzZMrRv3x4AsGjRIjRo0AB79uxBixYtbNdyIiIieiiVa8xHdnY2AMDLywsAcODAARQVFaFjx47GferXr49atWph9+7d5TkVERERPSKsnufDYDBgzJgxaNmyJRo3bgwASE9Ph4ODAzw8PEz29fPzQ7rErI4FBQUmi0dxWWwiIqJHm9U9H8OHD8fx48exfPnycjUgISEBOp3O+AoKkp5emoiIiB5+VgUfI0aMwG+//Ybk5GTUrFnTuN3f3x+FhYXIysoy2T8jIwP+/qUPwOOy2ERERNWLRcGHEAIjRozAL7/8gi1btiA4ONikPCIiAvb29ti8ebNxW0pKCi5cuICoqKhS69RqtXB3dzd5ERER0aPLojEfw4cPx7Jly7B69Wq4ubkZx3HodDo4OTlBp9Ph9ddfR3x8PLy8vODu7o6RI0ciKirK4kwXTUAANKUEIkFPPCF5zNkbNyTLgm80lSy7kCW9yqydi3Sq7W0hvQKrq5P0apx3rkun6AKAQSbd9LZeOhWzUCV9nEEjnVIKlXSdGpnzGQzS5yvMlV5hV2Unfd/sVDJfSbX0ccUa6bYUycTYhTLptHqZ1XC9ataWLAOAgNq1JMuuXbsmWZaVKz3mKazZk5JlEQ0aSpb1eK6zZBn8/UrfnuMkfQwRUTlZFHzMmzcPANC2bVuT7YsWLcLAgQMBAF988QXUajX69OmDgoICREdH4+uvv7ZJY4mIiOjhZ1HwIWQmZCrh6OiIuXPnYu7cuVY3ioiIiB5dXNuFiIiIFMXgg4iIiBTF4IOIiIgUxeCDiIiIFGX19OoV7ebVa9DfLjDbnnFdOp32/KVLkmVn085Lll3Oui5ZFhAknTKpvyO9Ou2N7CzJMqGSXikWAPQyaaMqjfRHJlTSA4JlU20hXVakkb5Gg8wAZL2ddAqrSiOdMquSuTdqmVjZziBzPkiX2amk69TLlJ1LOStZdvec0tfo5qGTLMsvzpYsc3JxlS5zdZMsM8ikKN/JLX312jt50qnSRETlxZ4PIiIiUhSDDyIiIlIUgw8iIiJSFIMPIiIiUhSDDyIiIlIUgw8iIiJSVJVNtfX09YV7KavahmkdJI/xC/CXLOvb/zXJsgIhvXJrQZF0WX5O6WmKAIBi6eNu35Q5DvIfip1MWqx0gi5ggMzqtDJlQi2dpgq1dFuKH7wMUOlV6qUPVAvpWFktk9qqklm5VsjE33LX7unnK1kGALcLbkuWeXh7SZbdklnV1slBK1nm7CCdouygkr43jjVKb0uhQ5X9r4GIHgHs+SAiIiJFMfggIiIiRTH4ICIiIkUx+CAiIiJFMfggIiIiRTH4ICIiIkVV2Xy6/JvXoSkuNNteVGi+rYRGI50a6SCTiumglr4N7g7S8ZmPo7NkmUEvk9rq6SNZBgAqmaRZubRRuTLZRFyZlNJimXsqVDJ1yqwyKxfxyizoC5lFe6GSScO1lkGmSrWddMo3AOjV0hciZD4mN6cakmUamVV27WU+JzuZVW1vZGWWuj1HYrVbIiJbYM8HERERKYrBBxERESmKwQcREREpisEHERERKYrBBxERESmKwQcREREpyqLgIyEhAU899RTc3Nzg6+uLXr16ISUlxWSftm3bQqVSmbyGDRtm00YTERHRw8uieT62bduG4cOH46mnnkJxcTHeeecddOrUCSdPnoSLi4txvyFDhmDatGnG987O0vNhSDbMXg07B/P5CQwa6aXD9cXSsZRGyCzVLjM/hkZmfhC55d/1cvN8uErXCQCQmT9Dbj4LlcxEGBpIX6PMlBwotpObO0RGsfQ1WDvPhxy5OUDkSbdGbh4TlZ309xAAiiH9+QuZ76LsOWW+pzJTeUCtlr5GqTJNkfR8OkRE5WVR8LFu3TqT90lJSfD19cWBAwfQpk0b43ZnZ2f4+/vbpoVERET0SCnXmI/s7GwAgJeXl8n277//HjVq1EDjxo0xadIk3Lp1S7KOgoIC5OTkmLyI6NHx0Ucf4ZlnnoGzszM8PDxK3efChQvo2rUrnJ2d4evri3/9618oLi5WtqFEpBirp1c3GAwYM2YMWrZsicaNGxu3v/rqq6hduzYCAwNx9OhRTJgwASkpKfjvf/9baj0JCQmYOnWqtc0goiqusLAQL774IqKiovDtt9+alev1enTt2hX+/v7YtWsXrly5ggEDBsDe3h4ff/xxJbSYiCqa1cHH8OHDcfz4cezYscNk+9ChQ43/fuKJJxAQEIAOHTrg7NmzCAkJMatn0qRJiI+PN77PyclBUFCQtc0ioiqm5I+LpKSkUss3bNiAkydPYtOmTfDz80PTpk3x4YcfYsKECfjggw/g4CC/jg4RPXyseuwyYsQI/Pbbb0hOTkbNmjVl942MjAQApKamllqu1Wrh7u5u8iKi6mP37t144okn4OfnZ9wWHR2NnJwcnDhxQvI4PrIlenhZFHwIITBixAj88ssv2LJlC4KDgx94zOHDhwEAAQEBVjWQiB5t6enpJoEHAOP79PR0yeMSEhKg0+mML/aYEj08LHrsMnz4cCxbtgyrV6+Gm5ub8T8GnU4HJycnnD17FsuWLUOXLl3g7e2No0ePYuzYsWjTpg3Cw8MtalihvhiF+iLz7QbpFMYimTRFg0wuplYlnaeokkmLlEtthb3MUvT28qm2Qi6lUi71V8ilt8qkzMoUqWWWapdL0ZU5DBqr02Irglyqrdxh8inIKrmbKnP9cpnGct/hYpn2yH5P1aX/PN2R2F5i4sSJ+OSTT2T3OXXqFOrXry+7T3nwkS3Rw8ui4GPevHkA7k4kdq9FixZh4MCBcHBwwKZNm5CYmIj8/HwEBQWhT58+eO+992zWYCKqfOPGjcPAgQNl96lbt26Z6vL398e+fftMtmVkZBjLpGi1Wmi12jKdg4iqFouCD7nJkQAgKCgI27ZtK1eDiKjq8/HxgY+Pj03qioqKwkcffYSrV6/C19cXALBx40a4u7ujYcOGNjkHEVUtVme7EBGVxYULF3Djxg1cuHABer3eOA4sNDQUrq6u6NSpExo2bIjXXnsNM2fORHp6Ot577z0MHz6cPRtEjygGH0RUoSZPnozFixcb3zdr1gwAkJycjLZt20Kj0eC3337Dm2++iaioKLi4uCA2NtZkiQYierQw+CCiCpWUlCQ5x0eJ2rVrY+3atco0iIgqXbmmVyciIiKyVJXt+VBrNFBrzJun0cit6iqdqCiX3Gons1KsvUy+pcpOJnaTSW80OMjHfAaZ1F+5NE21kEnDlcnhlEsaNcisait3FXL3VC7VVi5919oMXdmUWRmyKxrLrBQLAHrZuyOXEi19TqvvjUwarlT6ruoB6eBEROXBng8iIiJSFIMPIiIiUhSDDyIiIlIUgw8iIiJSFIMPIiIiUhSDDyIiIlJUlU21NajUMKhKi41kUl9l0nDlVgsVMjmMBpkD5VJG5VIfi2RWnwUAvUxKpUEu1VbmOuxkFimVy0SVuw7ZVW1lDpRbL1Vu+SC5lNmKSNGVJbO6MiB/T+UImSPVMqsdy12/3I1TS/xkqK3NTyYiKgP2fBAREZGiGHwQERGRohh8EBERkaIYfBAREZGiGHwQERGRoqpstgsREdlezu2Jld0EoqobfIgiA0SheRqgQaavRj6BVea4YukjDXK5tnrpJE69zKKgxQ/IYiyWuUa5LF21TPanfZF0mWyWptzipnKprzJfLan0TkA+zVgu+1M2PbkiskaL5VNt1XIp0XLp4jIrGsulDKvkVq6VO04ifVd1R+YgIqJyqrLBBxFRdcdeCnpUccwHERERKYrBBxERESmKwQcREREpisEHERERKcqi4GPevHkIDw+Hu7s73N3dERUVhd9//91YfufOHQwfPhze3t5wdXVFnz59kJGRYfNGExER0cPLomyXmjVrYsaMGahXrx6EEFi8eDF69uyJQ4cOoVGjRhg7dizWrFmDFStWQKfTYcSIEejduzd27txpccOaNWxa6iqexXbSKYW37S0+DQDASSYNVaOXTqmUW7m1UCZFNV8rn/tZJBMS2smk9zoWSx/nUihdJheBFshch17mQI1MTrRBpWyqrbULtKpkPl+VXj6x204unVamXg2kb7hUWmx5SKUhG+SWFyYiKieLgo/u3bubvP/oo48wb9487NmzBzVr1sS3336LZcuWoX379gCARYsWoUGDBtizZw9atGhhu1YTERHRQ8vqMR96vR7Lly9Hfn4+oqKicODAARQVFaFjx47GferXr49atWph9+7dkvUUFBQgJyfH5EVERESPLouDj2PHjsHV1RVarRbDhg3DL7/8goYNGyI9PR0ODg7w8PAw2d/Pzw/p6emS9SUkJECn0xlfQUFBFl8EERERPTwsDj7CwsJw+PBh7N27F2+++SZiY2Nx8uRJqxswadIkZGdnG18XL160ui4iIiKq+iyeXt3BwQGhoaEAgIiICPz555+YPXs2Xn75ZRQWFiIrK8uk9yMjIwP+/v6S9Wm1Wmi1WstbTkRERA+lcs/zYTAYUFBQgIiICNjb22Pz5s3GspSUFFy4cAFRUVHlPQ0RERE9Iizq+Zg0aRJiYmJQq1Yt5ObmYtmyZdi6dSvWr18PnU6H119/HfHx8fDy8oK7uztGjhyJqKgoqzJdRLEoNUVSJZPCqNZYF0upZNb91AjpOtUy64zayaRFqmTXJwVUMmmOcsmWsimccmmjMmV2KunrF3LpmHLXIHdv1DIrDEufTfaOyq4GK3dfZI6TW5n2QeXy3ymZ77dM+rLcZyG3qq9a6jhm2hJRBbIo+Lh69SoGDBiAK1euQKfTITw8HOvXr8dzzz0HAPjiiy+gVqvRp08fFBQUIDo6Gl9//XWFNJyIiIgeThYFH99++61suaOjI+bOnYu5c+eWq1FERET06OLaLkRERKQoBh9ERESkKAYfREREpCiL5/moaCWj9qUWtpJb8EoI+cW+pMjVKb/AltxxMkfJFQIQMikYViaYyLZHLuND9t7IXL9cnbJ3VO7zlTvOyjK5Qvl2ylUqf6zVi7bJfL9lP3u5Kh/wcyab0UREZKUqF3zk5uYCAG7qc0vfQWblVtyx7py3rDusUsgsTitbZvWKOTIr/tKjLzc3FzqdrrKbQUSPmCoXfAQGBuLixYtwc3ODSqVCTk4OgoKCcPHiRbi7u1d286oM3hdpvDels+S+CCGQm5uLwMBAhVpHRNVJlQs+1Go1atasabbd3d2dv0hKwfsijfemdGW9L+zxIKKKUuWCDyIiooqQc3tiZTeB/g+zXYiIiEhRVT740Gq1mDJlCle+vQ/vizTem9LxvhBRVaESzKUjogr00UcfYc2aNTh8+DAcHByQlZVltk9piw3+8MMP6Nu3b5nPk5OTA51Oh+zsbI71UZi704xy13H/I5Hy1qnUI5aHpZ1KsORnkGM+iKhCFRYW4sUXX0RUVJTs+lCLFi1C586dje89PDwUaB0RVQYGH0RUoaZOnQoASEpKkt3Pw8MD/v7+CrSIiCpblR/zQUTVw/Dhw1GjRg08/fTTWLhw4QNnVy0oKEBOTo7Ji4geDuz5IKJKN23aNLRv3x7Ozs7YsGED3nrrLeTl5WHUqFGSxyQkJBh7VYjo4VKlez7mzp2LOnXqwNHREZGRkdi3b19lN0lxf/zxB7p3747AwECoVCqsWrXKpFwIgcmTJyMgIABOTk7o2LEjzpw5UzmNVVBCQgKeeuopuLm5wdfXF7169UJKSorJPnfu3MHw4cPh7e0NV1dX9OnTBxkZGZXUYuXMmzcP4eHhxsnEoqKi8PvvvxvLbXFfJk6cCJVKJfs6ffp0met7//330bJlSzRr1gwTJkzA22+/jU8//VT2mEmTJiE7O9v4unjxokXXQESVp8oGHz/++CPi4+MxZcoUHDx4EE2aNEF0dDSuXr1a2U1TVH5+Ppo0aYK5c+eWWj5z5kx8+eWXmD9/Pvbu3QsXFxdER0fjzh0rF7p5SGzbtg3Dhw/Hnj17sHHjRhQVFaFTp07Iz8837jN27Fj8+uuvWLFiBbZt24bLly+jd+/eldhqZdSsWRMzZszAgQMHsH//frRv3x49e/bEiRMnANjmvowbNw6nTp2SfdWtW9fqa4iMjMQ///yDgoICyX20Wq0xwOJstkQPlyqbahsZGYmnnnoKX331FQDAYDAgKCgII0eOxMSJj05qkiVUKhV++eUX9OrVC8DdXo/AwECMGzcO48ePBwBkZ2fDz88PSUlJFqUpPuwyMzPh6+uLbdu2oU2bNsjOzoaPjw+WLVuGF154AQBw+vRpNGjQALt370aLFi0qucXK8vLywqeffooXXnih0u5LUlISxowZU2qq7f0++ugjfP7557hx40aZ62eqbeVhqq31qmuqbZXs+SgsLMSBAwfQsWNH4za1Wo2OHTti9+7dldiyqiUtLQ3p6ekm90mn0yEyMrLa3afs7GwAd3/JAsCBAwdQVFRkcm/q16+PWrVqVat7o9frsXz5cuTn5yMqKqpS7suFCxdw+PBhXLhwAXq9HocPH8bhw4eRl5cHAPj111/xn//8B8ePH0dqairmzZuHjz/+GCNHjqyQ9hBR5auSA06vXbsGvV4PPz8/k+1+fn4WPUd+1KWnpwNAqfeppKw6MBgMGDNmDFq2bInGjRsDuHtvHBwczOaKqC735tixY4iKisKdO3fg6uqKX375BQ0bNjRO9KXkfZk8eTIWL15sfN+sWTMAQHJyMtq2bQt7e3vMnTsXY8eOhRACoaGhmDVrFoYMGVIh7SGiylclgw8iSwwfPhzHjx/Hjh07KrspVUZYWBgOHz6M7OxsrFy5ErGxsdi2bVultCUpKUl2jo/OnTubTC5GRI++KvnYpUaNGtBoNGYj8DMyMjgJ0T1K7kV1vk8jRozAb7/9huTkZNSsWdO43d/fH4WFhWbjC6rLvXFwcEBoaCgiIiKQkJCAJk2aYPbs2dX+vhBR1VAlgw8HBwdERERg8+bNxm0GgwGbN29GVFRUJbasagkODoa/v7/JfcrJycHevXsf+fskhMCIESPwyy+/YMuWLQgODjYpj4iIgL29vcm9SUlJwYULFx75e1Mag8GAgoIC3hciqhKq7GOX+Ph4xMbGonnz5nj66aeRmJiI/Px8xMXFVXbTFJWXl4fU1FTj+7S0NBw+fBheXl6oVasWxowZg+nTp6NevXoIDg7G+++/j8DAQGNGzKNq+PDhWLZsGVavXg03NzfjeAWdTgcnJyfodDq8/vrriI+Ph5eXF9zd3TFy5EhERUU98pkukyZNQkxMDGrVqoXc3FwsW7YMW7duxfr166v1fSGiqqPKBh8vv/wyMjMzMXnyZKSnp6Np06ZYt26d2eDKR93+/fvRrl074/v4+HgAQGxsLJKSkvD2228jPz8fQ4cORVZWFlq1aoV169bB0dGxspqsiHnz5gEA2rZta7J90aJFGDhwIADgiy++gFqtRp8+fVBQUIDo6Gh8/fXXCrdUeVevXsWAAQNw5coV6HQ6hIeHY/369XjuuecAVN/7QkRVR5Wd54OIyBKc56PycJ4P63GeDyIiIiIFMPggIiIiRTH4ICIiIkUx+CAiIiJFMfggIiIiRTH4ICIiIkUx+CAiIiJFMfggIiIiRTH4ICIiIkUx+CAiIiJFMfggIiIiRTH4ICIiIkUx+CAiIiJF2VV2A4iIiB5Wj9KqtEpi8EFEROXCX8BkKT52ISIiIkUx+CAiIiJFMfggIiIiRTH4ICIiIkUx+CAiIiJFMfggIiIiRTH4ICIiIkUx+CAiIiJFMfggIiIiRTH4ICIiIkUx+CAiIiJFMfggIiIiRTH4ICIiIkUx+CAiIiJFMfggIiIiRTH4ICIiIkUx+CAiIiJFMfggIiIiRdlVdgOIiIjul3N7YmU3gSoQez6IiIhIUQw+iIiISFEMPoioQp0/fx6vv/46goOD4eTkhJCQEEyZMgWFhYUm+x09ehStW7eGo6MjgoKCMHPmzEpqMRFVNI75IKIKdfr0aRgMBnzzzTcIDQ3F8ePHMWTIEOTn5+Ozzz4DAOTk5KBTp07o2LEj5s+fj2PHjmHQoEHw8PDA0KFDK/kKiMjWVEIIUdmNIKLq5dNPP8W8efNw7tw5AMC8efPw7rvvIj09HQ4ODgCAiRMnYtWqVTh9+nSZ6szJyYFOp0N2djbc3d0rrO1EVDpLfgb52IWIFJednQ0vLy/j+927d6NNmzbGwAMAoqOjkZKSgps3b1ZGE4moAjH4ICJFpaamYs6cOXjjjTeM29LT0+Hn52eyX8n79PT0UuspKChATk6OyYuIHg4MPojIKhMnToRKpZJ93f/I5NKlS+jcuTNefPFFDBkypFznT0hIgE6nM76CgoLKVR8RKYdjPojIKpmZmbh+/brsPnXr1jU+Srl8+TLatm2LFi1aICkpCWr1//72GTBgAHJycrBq1SrjtuTkZLRv3x43btyAp6enWd0FBQUoKCgwvs/JyUFQUBDHfBBVEkvGfDDbhYis4uPjAx8fnzLte+nSJbRr1w4RERFYtGiRSeABAFFRUXj33XdRVFQEe3t7AMDGjRsRFhZWauABAFqtFlqttnwXQUSVgo9diKhCXbp0CW3btkWtWrXw2WefITMzE+np6SZjOV599VU4ODjg9ddfx4kTJ/Djjz9i9uzZiI+Pr8SWE1FFYc8HEVWojRs3IjU1FampqahZs6ZJWclTX51Ohw0bNmD48OGIiIhAjRo1MHnyZM7xQfSI4pgPInokcJ4PosrFeT6IiIioymLwQURERIrimA8ieiSUPEHmZGNElaPkZ68sozkYfBDRIyE3NxcAONkYUSXLzc2FTqeT3YcDTonokWAwGHD58mW4ublBpVKVq66SCcsuXrxok8Grtq6PdbLOqlinEAK5ubkIDAw0m8vnfuz5IKJHglqtNkvlLS93d3ebZs7Yuj7WyTqrWp0P6vEowQGnREREpCgGH0RERKQoBh9ERPfRarWYMmWKzdaOsXV9rJN1VvU6H4QDTomIiEhR7PkgIiIiRTH4ICIiIkUx+CAiIiJFMfggIiIiRTH4ICKqxpYsWYKCggKz7YWFhViyZInF9RUVFWHQoEFIS0uzRfOoAvzzzz+SZXv27FGkDQw+iIj+T2FhIVJSUlBcXGyT+rZv347+/fsjKioKly5dAgAsXboUO3bssEn9thAXF4fs7Gyz7bm5uYiLi7O4Pnt7e/z888+2aBrdIyUlBSNGjECHDh3QoUMHjBgxAikpKVbV1alTJ9y4ccNs+86dO9G5c+fyNrVMOL06EVV7t27dwsiRI7F48WIAwF9//YW6deti5MiReOyxxzBx4kSL6/z555/x2muvoV+/fjh06JCxdyE7Oxsff/wx1q5da9NrsJYQotS1cP75558yT5V9v169emHVqlUYO3ZseZtn5OnpWWo7VSoVHB0dERoaioEDB5Y5YIqPjy91+7319ezZE15eXha3denSpZg/fz7S0tKwe/du1K5dG4mJiQgODkbPnj0tru/nn39G37590bx5c0RFRQG420PRuHFjLF++HH369LGovhYtWqBTp05ITk6Gm5sbAOCPP/5A9+7d8cEHH1jcPmtwng8iqvZGjx6NnTt3IjExEZ07d8bRo0dRt25drF69Gh988AEOHTpkcZ3NmjXD2LFjMWDAALi5ueHIkSOoW7cuDh06hJiYGKSnp1tUV1kXyzt48KBFdR45cgSNGjWCnd3//hbV6/VIS0tD586d8dNPP5W5nSWmT5+Ozz//HB06dEBERARcXFxMykeNGmVxnV988QU++ugjxMTE4OmnnwYA7Nu3D+vWrcPYsWORlpaGpUuXYs6cORgyZMgD62vXrh0OHjwIvV6PsLAwAHeDTo1Gg/r16yMlJQUqlQo7duxAw4YNy9zOefPmYfLkyRgzZgw++ugjHD9+HHXr1kVSUhIWL16M5ORki689JCQE/fr1w7Rp00y2T5kyBd999x3Onj1rUX0GgwEvvPACbty4gfXr12PXrl3o0aMHpk+fjtGjR1vcPmsw+CCiaq927dr48ccf0aJFC5NAITU1FU8++SRycnIsrtPZ2RknT55EnTp1TOo8d+4cGjZsiDt37pS5rqlTpxr/fefOHXz99ddo2LChyV/BJ06cwFtvvYWEhASL6pw6dSrGjRsHV1dXY5mDgwPq1KmDPn36wMHBocztLBEcHCxZplKpcO7cOYvr7NOnD5577jkMGzbMZPs333yDDRs24Oeff8acOXOwYMECHDt27IH1JSYmYvv27Vi0aJFxMbXs7GwMHjwYrVq1wpAhQ/Dqq6/i9u3bWL9+fZnb2bBhQ3z88cfo1auXyed+/PhxtG3bFteuXbPswnH3u3T06FGEhoaabD9z5gyaNGmCW7duWVxnYWEhunbtilu3buHo0aNISEjAiBEjLK7HaoKIqJpzcnISZ8+eFUII4erqavz34cOHhbu7u1V1BgcHi40bN5rVuXjxYtGgQQOr2/r666+L9957z2z75MmTRVxcnMX1JSUlidu3b1vdHqW4uLiIM2fOmG0/c+aMcHFxEUIIkZqaKpydnctUX2BgoDhx4oTZ9uPHj4vAwEAhhBAHDhwQ3t7eFrXT0dFRnD9/Xghh+rn/9ddfwtHR0aK6SsTExIiFCxeabV+4cKHo1KlTmeo4cuSI2WvHjh0iKChIDBs2zGS7Ejjmg4iqvebNm2PNmjUYOXIkABgfcfznP/8x9i5YasiQIRg9ejQWLlwIlUqFy5cvY/fu3Rg/fjzef/99q9u6YsUK7N+/32x7//790bx5cyxcuNCi+mJjY61uy4MUFhYiLS0NISEhJo91rOHl5YVff/3VbBzJr7/+ahyXkZ+fbxzD8CDZ2dm4evWq2SOVzMxMY0+Xh4cHCgsLLWpncHAwDh8+jNq1a5tsX7duHRo0aGBRXSV69OiBCRMm4MCBA2jRogWAu71dK1aswNSpU/H//t//M9m3NE2bNoVKpYK452FHyftvvvkGCxYsMI7/0ev1VrXTEgw+iKja+/jjjxETE4OTJ0+iuLgYs2fPxsmTJ7Fr1y5s27bNqjonTpwIg8GADh064NatW2jTpg20Wi3Gjx9vDHKs4eTkhJ07d6JevXom23fu3AlHR0eL69Pr9fjiiy/w008/4cKFC2a/bEvLiniQihjA+/777+PNN99EcnKycczHn3/+ibVr12L+/PkAgI0bN+LZZ58tU309e/bEoEGD8Pnnn+Opp54y1jd+/Hj06tULwN0xJY8//rhF7YyPj8fw4cNx584dCCGwb98+/PDDD0hISMB//vMfi+oq8dZbbwEAvv76a3z99dellgGQDRyqXOqzIv0rRERVXGpqqhg8eLB46qmnRIMGDUS/fv3E0aNHy11vQUGBOHHihNi7d6/Izc0td30JCQnC0dFRjBw5UixdulQsXbpUjBgxQjg7O4uEhASL63v//fdFQECA+Oyzz4Sjo6P48MMPxeuvvy68vb3F7NmzrWrjqFGjREREhNi+fbtwcXExPnpYtWqVaNq0qVV1CiHEjh07RN++fUWzZs1Es2bNRN++fcXOnTutqis3N1cMHjxYODg4CLVaLdRqtXBwcBBDhgwReXl5QgghDh06JA4dOmRx3d99950IDQ0VKpVKqFQq8dhjj4n//Oc/VrXzUcUBp0RED5mffvoJs2fPxqlTpwAADRo0wOjRo/HSSy9ZXFdISAi+/PJLdO3aFW5ubjh8+LBx2549e7Bs2TKL66yIAbwVJS8vzzgAtm7duiYDb8vr1q1byMvLg6+vr83qtIWEhAT4+flh0KBBJtsXLlyIzMxMTJgwocLbwMcuRES4m36YmpqKq1evwmAwmJS1adOmTHX07t27zOf773//a1H7AKC4uBgff/wxBg0aZFWgUZr09HQ88cQTAABXV1fjhGPdunWzemxKZmZmqb9w8/Pzy5wyXBq9Xo9Vq1YZg65GjRqhR48e0Gg0Vtfp6upqHDNii8Dj9u3bEELA2dkZzs7OyMzMRGJiIho2bIhOnTqVuZ4vv/wSQ4cOhaOjI7788kvZfS1NXf7mm29KDSobNWqEvn37MvggIlLCnj178Oqrr+Lvv//G/Z3BlgzAs3ZSrrKys7PDzJkzMWDAAJvVWbNmTVy5cgW1atVCSEgINmzYgCeffBJ//vkntFqtVXVWxADe1NRUdOnSBZcuXTLOy5GQkICgoCCsWbMGISEhFtVnMBiM85Hk5eUBANzc3DBu3Di8++67UKutmwC8Z8+e6N27N4YNG4asrCw8/fTTcHBwwLVr1zBr1iy8+eabZarniy++QL9+/eDo6IgvvvhCcj+VSmVx8JGeno6AgACz7T4+Prhy5YpFdVmtcp/6EBFVviZNmogXX3xRnDx5Uty8eVNkZWWZvKqSHj16iKSkJJvVN2HCBPHRRx8JIYRYvny5sLOzE6GhocLBwUFMmDDBqjq3b98uXF1dxbBhw4Sjo6MYPXq0eO6554SLi4vYv3+/VXXGxMSIzp07i+vXrxu3Xbt2TXTu3Fl06dLF4vomTpwofHx8xNdff21MMZ07d67w8fER77zzjlVtFEIIb29vcfz4cSGEEP/+979FeHi40Ov14qeffhL169e3ul5bCg0NFUuXLjXbvmTJEhEcHKxIGzjmg4iqPRcXFxw5csRsEqeqaP78+Zg6dSr69etX6uyhUqmWZbVnzx7s2rUL9erVQ/fu3a2u5+zZs5gxYwaOHDmCvLw8PPnkk5gwYYLxEY+lXFxcsGfPHrPjjxw5gpYtWxp7L8oqMDAQ8+fPN7tfq1evxltvvWVci8dSzs7OOH36NGrVqoWXXnoJjRo1wpQpU3Dx4kWEhYVZNSGYrc2cORMzZ87Ep59+ivbt2wMANm/ejLfffhvjxo3DpEmTKrwNfOxCRNVeZGQkUlNTyx18PPnkk9i8eTM8PT0fOCV6WadBv19JauWsWbPMyqyZo+H+wYctWrRAixYtsHDhQnzyySdWP/8PCQnBv//9b6uOLY1Wq0Vubq7Z9ry8PKtmYb1x4wbq169vtr1+/fpWpReXCA0NxapVq/D8889j/fr1xnlJrl69apxJ1VJ6vR5JSUnYvHlzqWOStmzZYlF9//rXv3D9+nW89dZbxtRqR0dHTJgwQZHAA2DwQUSEkSNHYty4ccbBl/b29ibl4eHhZaqnZ8+exnESJXNF2Nr9v3jKq6IGH549exaLFi3CuXPnkJiYCF9fX/z++++oVasWGjVqZHF93bp1w9ChQ/Htt98a5/nYu3cvhg0bZlVvT5MmTfDVV1+ZDeb86quv0KRJE4vrKzF58mS8+uqrGDt2LDp06GAc47JhwwY0a9bMqjpHjx6NpKQkdO3aFY0bNy7XoF3gbpD6ySef4P3338epU6fg5OSEevXqWT3Gx6o28LELEVV3pQ0uLJn9UakZHyuLo6MjTp06ZbYeizVr0JTYtm0bYmJi0LJlS/zxxx84deoU6tatixkzZmD//v1YuXKlxXVmZWUhNjYWv/76qzE4LCoqQs+ePbFo0SJ4eHhY3MauXbuiVq1axgBh9+7duHjxItauXYvWrVtb3MYS6enpuHLlCpo0aWL8bu3btw/u7u6l9rY8SI0aNbBkyRJ06dLF6jZVNez5IKJqryJnf9y/f78xNbRhw4aIiIiwuI6KTLsMCgrCzp07zYKPnTt3IjAw0OK2Andnd50+fTri4+NNpjtv3749vvrqK6vq9PDwwOrVq5Gammoyv4m1j8qeffZZ/PXXX5g7dy5Onz4N4G6q9FtvvWX1dRcVFcHJyQmHDx826+Uo6a2xhoODg03HI7Vr106298TSxzjWYPBBRNXe/etw2MI///yDV155BTt37jT+VZ6VlYVnnnkGy5cvR82aNctcV0WmXQ4ZMgRjxoxBUVFRqYMPrXHs2LFSH+X4+vpatKprfHy8bPm9y9OXNgZGSlFRETp37oz58+fjo48+KvNxD2Jvb49atWrZvKds3LhxmD17Nr766qtyP3IB7q7zcq+ioiIcPnwYx48fr9C1fu7F4IOI6P+cPHmy1PVNrBlTMHjwYBQVFeHUqVPGeSlSUlIQFxeHwYMHY926dWWu696emXv/XfLUvDy/kCpi8KGHhweuXLli1pty6NAhPPbYY2Wu59ChQybvDx48iOLiYuP9/Ouvv6DRaCzuTbK3t8fRo0ctOqas3n33XbzzzjtYunSpcfIya9w/Yd2WLVvw+++/o1GjRmZjkiydsE4qgP3ggw8szhqyFsd8EFG1d+7cOTz//PM4duyYycqfJb/UrflL1snJCbt27TLrfj9w4ABat25drpTLb7/9Fl988QXOnDkDAKhXrx7GjBmDwYMHW11nXl6ezQYfjh8/Hnv37sWKFSvw+OOP4+DBg8jIyMCAAQMwYMAATJkyxeI6Z82aha1bt2Lx4sXw9PQEANy8eRNxcXFo3bq1xb00Y8eOhVarxYwZMyxui5xmzZohNTUVRUVFqF27tlkqdFmznOLi4sp8zkWLFlnURimpqal4+umny5XtU1bs+SCiam/06NEIDg7G5s2bERwcjH379uH69esYN24cPvvsM6vqDAoKQlFRkdl2vV5v9ZgC4G42xaxZszBy5EiTgZJjx47FhQsXMG3aNKvqdXV1Na7uWl4ff/wxhg8fjqCgIOj1ejRs2BDFxcXo168f3nvvPavq/Pzzz7FhwwZj4AEAnp6emD59Ojp16mRx8FFcXIyFCxdi06ZNpc6XYsljnHvZKsvp3oDi9u3bMBgMxjaeP38eq1atQoMGDRAdHW2T8wF3v0fWrIxsDfZ8EFG1V6NGDWzZsgXh4eHQ6XTYt28fwsLCsGXLFowbN86s+78sVq9ejY8//hhz585F8+bNAdwdfDpy5EhMmDDB6l9SPj4++PLLL/HKK6+YbP/hhx8wcuRIi8ZUVLSLFy/i2LFjyMvLQ7NmzVCvXj2r63Jzc8Ovv/6Ktm3bmmxPTk5Gjx49Sp0D5H5Hjx5F48aNoVar0a5dO8n9VCqVIoMuy6pTp04mU7bXr18f9vb2Fk/ZXuL+RzpCCFy5cgX79+/H+++/b1XPlKXY80FE1Z5erzdmZdSoUQOXL19GWFgYateujZSUlDLX4+npaTL+Ij8/H5GRkbCzu/tfbXFxMezs7DBo0CCrg4+ioiJjMHOviIgIFBcXW1WnLTxocOiePXuM/7amV+H5559HXFwcPv/8c5N5Pv71r3+VeUG/Zs2a4cqVK/D19cXff/+NP//8E97e3ha3pSwOHDhgsgCetXN8AHcf1ZSM01i5ciX8/Pxw6NAh/Pzzz5g8ebLFwcf9axCp1WqEhYVh2rRpFi1+Vx4MPoio2mvcuDGOHDmC4OBgREZGYubMmXBwcMCCBQtQt27dMteTmJhYcY38P6+99hrmzZtn9gt8wYIF6NevX4WfX0pFDQ4tMX/+fIwfPx6vvvqq8XGWnZ0dXn/9dXz66adlqsPDwwNpaWnw9fXF+fPnbT5hG3B3JtO+ffti69atJllO7dq1w/Lly+Hj42Nxnbdu3TIGxxs2bEDv3r2hVqvRokUL/P333xbVpdfrERcXhyeeeMLkEZbS+NiFiKq99evXIz8/H71790Zqaiq6deuGv/76C97e3vjxxx+NKaiV5d5eheLiYiQlJaFWrVpo0aIFgLs9ABcuXMCAAQMwZ86cymqmka0Hh94rPz8fZ8+eBXB3Cvf7x2rIGTp0KJYsWYKAgABcuHABNWvWhEajKXXfc+fOWdW+l19+GefOncOSJUvQoEEDAHezqGJjYxEaGooffvjB4jrDw8MxePBgPP/882jcuDHWrVuHqKgoHDhwAF27dkV6erpF9UlNLKckBh9ERKW4ceOG2WMUS+n1eqxatcqk+71Hjx6Sv/CkyI1PuFdVGavw2GOPYcOGDWbTqB8/fhydOnXC5cuXK6llwLp165CamopRo0Zh2rRpJpOg3Wv06NFW1a/T6bBp0yazwbv79u1Dp06dkJWVZXGdK1euxKuvvgq9Xo8OHTpgw4YNAO6uy/PHH3/g999/t6i+5s2b45NPPkGHDh0sbout8LELEVEpyjNHA3A3bbFLly64dOmS8dFDQkICgoKCsGbNGoSEhJS5rnsn03oY5OTkIDMz02x7ZmZmmQaGVqTOnTsDuDsmY/To0ZLBh7UMBoPZPBzA3blFrH3M88ILL6BVq1bGKdtLdOjQAc8//7zF9U2fPh3jx4/Hhx9+WGqmj7UL4FmCPR9EVC2VdZAiYPkkTgDQpUsXCCHw/fffGwOZ69evo3///lCr1VizZo3FdT4sBgwYgO3bt5c6OLR169ZYvHhxJbew4vTs2RNZWVn44YcfjCnVly5dQr9+/eDp6YlffvmlkltoupbRvT17Sq5lxJ4PIqqW7h/xb2vbtm3Dnj17THpQvL29MWPGDLRs2bJCz13ZbDE49GH11VdfoUePHqhTpw6CgoIAABcuXMATTzyB7777rpJbd9eiRYsQFBRk9vjPYDDgwoULirSBPR9ERBXAy8sLv/32G5555hmT7Tt37kT37t0VmUWyspVncOjDTAiBzZs3myyA17Fjx0pu1f9oNBpjyvG9rl+/Dl9fX0V6Phh8EBH9n6tXrxrn9QgLCzP7z9kSAwYMwMGDB/Htt9+aPHoYMmQIIiIikJSUZIsmUxW0efNmbN68GVevXjUb57Fw4cJKatX/qNVqZGRkmKX9/v3332jYsCHy8/MrvA187EJE1V5OTg6GDx+O5cuXG//q02g0ePnllzF37lyrHtF8+eWXiI2NRVRUlHEAYnFxMXr06IHZs2fbtP1UdUydOhXTpk1D8+bNERAQYJNVaG2lJGVbpVLh/fffh7Ozs7FMr9dj7969ZiveVhT2fBBRtffyyy/j0KFDmDNnjsl6KaNHj0bTpk2xfPlyq+s+c+YMTp8+DeBu93toaKhN2kxVU0BAAGbOnInXXnutsptipiRle9u2bYiKioKDg4OxzMHBAXXq1MH48ePLNQ1+WTH4IKJqz8XFBevXr0erVq1Mtm/fvh2dO3dWpBuaHg3e3t7Yt2+fRanUSouLi8Ps2bMVSamVwscuRFTteXt7l/poRafTWT0FtRACK1euRHJycqnP/q1J36Wqb/DgwVi2bBnef//9ym6KpHtXzK0sDD6IqNp77733EB8fj6VLl8Lf3x8AkJ6ejn/9619W/xIZM2YMvvnmG7Rr1w5+fn5V6tk/2da9098bDAYsWLAAmzZtQnh4uNmEY9Ysqvco4mMXIqr2mjVrhtTUVBQUFKBWrVoA7s7NoNVqzZ5/Hzx4sEx1enl54bvvvkOXLl1s3l6qWh626e+rAvZ8EFG1Z+3y9nJ0Op1FK+LSw+thm/6+KmDPBxFVa3q9Hjt37kR4eLhxCXRbWLx4MdatW4eFCxfCycnJZvUSPQoYfBBRtVcRS4zfvn0bzz//PHbu3Ik6deqYPfsv6+MbokcRH7sQUbXXuHFjnDt3zqbBR2xsLA4cOID+/ftzwCnRfdjzQUTV3rp16zBp0iSbLjEuNXcIETH4ICKqkCXG69evj59++gnh4eE2aSPRo4SPXYio2quIbIXPP/8cb7/9NubPn486derYvH6ihxl7PoiIKoCnpydu3bqF4uJiODs7mw04vXHjRiW1jKjyseeDiKqlo0ePonHjxlCr1Th69KjsvtY8OklMTLSyZUSPPvZ8EFG1pFarkZ6eDl9fX6jVaqhUKpT236G1Yz6ISBp7PoioWkpLS4OPj4/x3xVBr9dj1apVOHXqFACgUaNG6NGjBzQaTYWcj+hhwZ4PIqL/c/LkSVy4cAGFhYXGbSqVCt27d7e4rtTUVHTp0gWXLl1CWFgYACAlJQVBQUFYs2ZNlV5ynaiiMfggomrv3LlzeP7553Hs2DGTxy8labfWPHbp0qULhBD4/vvv4eXlBQC4fv06+vfvD7VajTVr1tjuAogeMuoH70JE9GgbPXo0goODcfXqVTg7O+P48eP4448/0Lx5c2zdutWqOrdt24aZM2caAw8A8Pb2xowZM7Bt2zYbtZzo4cQxH0RU7e3evRtbtmxBjRo1oFarodFo0KpVKyQkJGDUqFE4dOiQxXVqtVrk5uaabc/Ly4ODg4Mtmk300GLPBxFVe3q9Hm5ubgCAGjVq4PLlywCA2rVrIyUlxao6u3XrhqFDh2Lv3r0QQkAIgT179mDYsGHo0aOHzdpO9DBizwcRVXuNGzfGkSNHEBwcjMjISMycORMODg5YsGAB6tata1WdX375JWJjYxEVFWWcYKy4uBg9evTgHCBU7XHAKRFVe+vXr0d+fj569+6N1NRUdOvWDX/99Re8vb3x448/on379lbXnZqaaky1bdCgAUJDQ23VbKKHFoMPIqJS3LhxA56eniYLzVli2rRpGD9+PJydnU223759G59++ikmT55si2YSPZQYfBARVQCNRoMrV67A19fXZPv169fh6+vLWVOpWuOAUyKiCiCEKLXX5MiRIybpt0TVEQecEhHZUMmjGpVKhccff9wkANHr9cjLy8OwYcMqsYVElY+PXYiIbGjx4sUQQmDQoEFITEyETqczljk4OKBOnTqIioqqxBYSVT4GH0REFWDbtm145plnjGm2RPQ/DD6IiCrAhQsXZMtr1aqlUEuIqh4GH0REFUCtVsum6TLbhaozDjglIqoA968HU1RUhEOHDmHWrFn46KOPKqlVRFUDez6IiBS0Zs0afPrpp1avlkv0KOA8H0RECgoLC8Off/5Z2c0gqlR87EJEVAFycnJM3gshcOXKFXzwwQeoV69eJbWKqGpg8EFEVAE8PDzMBpwKIRAUFITly5dXUquIqgaO+SAiqgDbtm0zea9Wq+Hj44PQ0FDY2fHvPqreGHwQEVWgkydP4sKFCygsLDTZ3qNHj0pqEVHlY/hNRFQBzp07h969e+Po0aNQqVQo+Tuv5FEM5/mg6ozZLkREFWD06NGoU6cOrl69CmdnZxw/fhx//PEHmjdvzjRbqvb42IWIqALUqFEDW7ZsQXh4OHQ6Hfbt24ewsDBs2bIF48aNM5uEjKg6Yc8HEVEF0Ov1cHNzA3A3ELl8+TIAoHbt2khJSanMphFVOo75ICKqAI0bN8aRI0cQHByMyMhIzJw5Ew4ODliwYAHq1q1b2c0jqlR87EJEVAHWr1+P/Px89O7dG6mpqejWrRv++usveHt748cff0T79u0ru4lElYbBBxGRQm7cuAFPT0/Z1W6JqgMGH0RERKQoDjglIiIiRTH4ICIiIkUx+CAiIiJFMfggIiIiRTH4ICIiIkUx+CAiIiJFMfggIiIiRTH4ICIiIkX9f+o9/ur2JynbAAAAAElFTkSuQmCC\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 33ms/step\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    }
  ]
}