Basic_Neural_Network_from_Scratch_(ML,_Difficulty_5).ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "toc_visible": true,
      "gpuType": "T4"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "source": [
        "---\n",
        "---"
      ],
      "metadata": {
        "id": "IzTxuHT4SfNZ"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "import warnings"
      ],
      "metadata": {
        "id": "1I3X9W1-LMJz"
      },
      "execution_count": 1,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "warnings.filterwarnings('ignore')"
      ],
      "metadata": {
        "id": "QwQ_6OBsLPlf"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "996V_wnBLPno"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "!pip install dash"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "X1Pkes18LPp7",
        "outputId": "585681a2-bc87-415f-9b81-0a0321318c60"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting dash\n",
            "  Downloading dash-2.18.2-py3-none-any.whl.metadata (10 kB)\n",
            "Collecting Flask<3.1,>=1.0.4 (from dash)\n",
            "  Downloading flask-3.0.3-py3-none-any.whl.metadata (3.2 kB)\n",
            "Collecting Werkzeug<3.1 (from dash)\n",
            "  Downloading werkzeug-3.0.6-py3-none-any.whl.metadata (3.7 kB)\n",
            "Requirement already satisfied: plotly>=5.0.0 in /usr/local/lib/python3.10/dist-packages (from dash) (5.24.1)\n",
            "Collecting dash-html-components==2.0.0 (from dash)\n",
            "  Downloading dash_html_components-2.0.0-py3-none-any.whl.metadata (3.8 kB)\n",
            "Collecting dash-core-components==2.0.0 (from dash)\n",
            "  Downloading dash_core_components-2.0.0-py3-none-any.whl.metadata (2.9 kB)\n",
            "Collecting dash-table==5.0.0 (from dash)\n",
            "  Downloading dash_table-5.0.0-py3-none-any.whl.metadata (2.4 kB)\n",
            "Requirement already satisfied: importlib-metadata in /usr/local/lib/python3.10/dist-packages (from dash) (8.5.0)\n",
            "Requirement already satisfied: typing-extensions>=4.1.1 in /usr/local/lib/python3.10/dist-packages (from dash) (4.12.2)\n",
            "Requirement already satisfied: requests in /usr/local/lib/python3.10/dist-packages (from dash) (2.32.3)\n",
            "Collecting retrying (from dash)\n",
            "  Downloading retrying-1.3.4-py3-none-any.whl.metadata (6.9 kB)\n",
            "Requirement already satisfied: nest-asyncio in /usr/local/lib/python3.10/dist-packages (from dash) (1.6.0)\n",
            "Requirement already satisfied: setuptools in /usr/local/lib/python3.10/dist-packages (from dash) (75.1.0)\n",
            "Requirement already satisfied: Jinja2>=3.1.2 in /usr/local/lib/python3.10/dist-packages (from Flask<3.1,>=1.0.4->dash) (3.1.4)\n",
            "Requirement already satisfied: itsdangerous>=2.1.2 in /usr/local/lib/python3.10/dist-packages (from Flask<3.1,>=1.0.4->dash) (2.2.0)\n",
            "Requirement already satisfied: click>=8.1.3 in /usr/local/lib/python3.10/dist-packages (from Flask<3.1,>=1.0.4->dash) (8.1.7)\n",
            "Requirement already satisfied: blinker>=1.6.2 in /usr/local/lib/python3.10/dist-packages (from Flask<3.1,>=1.0.4->dash) (1.9.0)\n",
            "Requirement already satisfied: tenacity>=6.2.0 in /usr/local/lib/python3.10/dist-packages (from plotly>=5.0.0->dash) (9.0.0)\n",
            "Requirement already satisfied: packaging in /usr/local/lib/python3.10/dist-packages (from plotly>=5.0.0->dash) (24.2)\n",
            "Requirement already satisfied: MarkupSafe>=2.1.1 in /usr/local/lib/python3.10/dist-packages (from Werkzeug<3.1->dash) (3.0.2)\n",
            "Requirement already satisfied: zipp>=3.20 in /usr/local/lib/python3.10/dist-packages (from importlib-metadata->dash) (3.21.0)\n",
            "Requirement already satisfied: charset-normalizer<4,>=2 in /usr/local/lib/python3.10/dist-packages (from requests->dash) (3.4.0)\n",
            "Requirement already satisfied: idna<4,>=2.5 in /usr/local/lib/python3.10/dist-packages (from requests->dash) (3.10)\n",
            "Requirement already satisfied: urllib3<3,>=1.21.1 in /usr/local/lib/python3.10/dist-packages (from requests->dash) (2.2.3)\n",
            "Requirement already satisfied: certifi>=2017.4.17 in /usr/local/lib/python3.10/dist-packages (from requests->dash) (2024.12.14)\n",
            "Requirement already satisfied: six>=1.7.0 in /usr/local/lib/python3.10/dist-packages (from retrying->dash) (1.17.0)\n",
            "Downloading dash-2.18.2-py3-none-any.whl (7.8 MB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m7.8/7.8 MB\u001b[0m \u001b[31m27.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading dash_core_components-2.0.0-py3-none-any.whl (3.8 kB)\n",
            "Downloading dash_html_components-2.0.0-py3-none-any.whl (4.1 kB)\n",
            "Downloading dash_table-5.0.0-py3-none-any.whl (3.9 kB)\n",
            "Downloading flask-3.0.3-py3-none-any.whl (101 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m101.7/101.7 kB\u001b[0m \u001b[31m8.2 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading werkzeug-3.0.6-py3-none-any.whl (227 kB)\n",
            "\u001b[2K   \u001b[90m━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━\u001b[0m \u001b[32m228.0/228.0 kB\u001b[0m \u001b[31m18.0 MB/s\u001b[0m eta \u001b[36m0:00:00\u001b[0m\n",
            "\u001b[?25hDownloading retrying-1.3.4-py3-none-any.whl (11 kB)\n",
            "Installing collected packages: dash-table, dash-html-components, dash-core-components, Werkzeug, retrying, Flask, dash\n",
            "  Attempting uninstall: Werkzeug\n",
            "    Found existing installation: Werkzeug 3.1.3\n",
            "    Uninstalling Werkzeug-3.1.3:\n",
            "      Successfully uninstalled Werkzeug-3.1.3\n",
            "  Attempting uninstall: Flask\n",
            "    Found existing installation: Flask 3.1.0\n",
            "    Uninstalling Flask-3.1.0:\n",
            "      Successfully uninstalled Flask-3.1.0\n",
            "Successfully installed Flask-3.0.3 Werkzeug-3.0.6 dash-2.18.2 dash-core-components-2.0.0 dash-html-components-2.0.0 dash-table-5.0.0 retrying-1.3.4\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "XTO8u5RLL1Dw"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Ob71eokOL1JC"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Import Required Libraries"
      ],
      "metadata": {
        "id": "T7L6gs7aL1c5"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np"
      ],
      "metadata": {
        "id": "SuzZiIKSL7BF"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf"
      ],
      "metadata": {
        "id": "yvroRhD-L7E1"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt"
      ],
      "metadata": {
        "id": "j8TRtk1cMESA"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import seaborn as sns"
      ],
      "metadata": {
        "id": "px4IsX4IMEVX"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from sklearn.metrics import confusion_matrix"
      ],
      "metadata": {
        "id": "gOKgLI7fMEY-"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.utils import to_categorical"
      ],
      "metadata": {
        "id": "35nqXSgRMXXi"
      },
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.datasets import mnist"
      ],
      "metadata": {
        "id": "TkuY_WZ8MEcu"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import dash"
      ],
      "metadata": {
        "id": "_igNlmuvL7N1"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from dash import dcc, html"
      ],
      "metadata": {
        "id": "ovkX6VKeL7S4"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from dash.dependencies import Input, Output"
      ],
      "metadata": {
        "id": "GLQRxrz3L7YT"
      },
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import base64"
      ],
      "metadata": {
        "id": "meWGDpBfL7be"
      },
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import io"
      ],
      "metadata": {
        "id": "dIfxZamtL7eS"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from PIL import Image"
      ],
      "metadata": {
        "id": "O-zmgnDuL7hK"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "----\n",
        "---"
      ],
      "metadata": {
        "id": "R8xloeOJMlHT"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "3X10DlpOMmz1"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "6cLgUdASMm3V"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "knzDblTAMnKl"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Step 1: Load MNIST Dataset"
      ],
      "metadata": {
        "id": "19CtyixGMxnU"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "(x_train, y_train), (x_test, y_test) = mnist.load_data()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xcWzwQAbM1bZ",
        "outputId": "9094220b-a169-4a95-f3a1-cfc66c15f7a8"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/mnist.npz\n",
            "\u001b[1m11490434/11490434\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 0us/step\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "lp3rBIxBMorB"
      },
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "bQH9NSPZL7kc"
      },
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "fSz2WYIwLurN"
      },
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Step 2: Preprocess Images"
      ],
      "metadata": {
        "id": "5OyR140tM9GS"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "x_train = x_train.reshape(-1, 28*28).astype('float32') / 255.0  # Normalize to [0,1]\n"
      ],
      "metadata": {
        "id": "VRYlxF12M_Ry"
      },
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "x_test = x_test.reshape(-1, 28*28).astype('float32') / 255.0"
      ],
      "metadata": {
        "id": "ijpIsftWLuu0"
      },
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "blppXKqILux3"
      },
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "tw_SbljaLu0_"
      },
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Step 3: One-Hot Encode Labels"
      ],
      "metadata": {
        "id": "s39XH3foNK3V"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "y_train = to_categorical(y_train, num_classes=10)"
      ],
      "metadata": {
        "id": "IC8nFevGNW-R"
      },
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "y_test = to_categorical(y_test, num_classes=10)"
      ],
      "metadata": {
        "id": "a-CHdCnQNOgZ"
      },
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "vnr0j-n0Lu4P"
      },
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Print Shape Confirmations"
      ],
      "metadata": {
        "id": "IC-9vtiJLu7U"
      },
      "execution_count": 23,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(f'Training data shape: {x_train.shape}, Labels shape: {y_train.shape}')\n",
        "print(f'Testing data shape: {x_test.shape}, Labels shape: {y_test.shape}')\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Qk3AQ8vpNSEU",
        "outputId": "43795ce5-34c0-4482-bf51-97f37ab31108"
      },
      "execution_count": 24,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Training data shape: (60000, 784), Labels shape: (60000, 10)\n",
            "Testing data shape: (10000, 784), Labels shape: (10000, 10)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "F2i98umdNgcc"
      },
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "6drg52FJNggt"
      },
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "bVYvKlrDNglK"
      },
      "execution_count": 24,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Step 4: Define Neural Network Class"
      ],
      "metadata": {
        "id": "R3BrWZmuNi92"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "class NeuralNetwork:\n",
        "    def __init__(self, input_size, hidden_size, output_size):\n",
        "        self.input_size = input_size\n",
        "        self.hidden_size = hidden_size\n",
        "        self.output_size = output_size\n",
        "\n",
        "        # Initialize Weights and Biases\n",
        "        self.W1 = np.random.randn(self.input_size, self.hidden_size) * 0.01\n",
        "        self.b1 = np.zeros((1, self.hidden_size))\n",
        "        self.W2 = np.random.randn(self.hidden_size, self.output_size) * 0.01\n",
        "        self.b2 = np.zeros((1, self.output_size))\n",
        "\n",
        "    def sigmoid(self, x):\n",
        "        return 1 / (1 + np.exp(-x))\n",
        "\n",
        "    def sigmoid_derivative(self, x):\n",
        "        return x * (1 - x)\n",
        "\n",
        "    def forward(self, X):\n",
        "        # Forward Propagation\n",
        "        self.z1 = np.dot(X, self.W1) + self.b1\n",
        "        self.a1 = self.sigmoid(self.z1)\n",
        "        self.z2 = np.dot(self.a1, self.W2) + self.b2\n",
        "        self.a2 = self.sigmoid(self.z2)\n",
        "        return self.a2\n",
        "\n",
        "    def backward(self, X, y, learning_rate=0.1):\n",
        "        # Compute Error\n",
        "        error = y - self.a2\n",
        "        d_output = error * self.sigmoid_derivative(self.a2)\n",
        "\n",
        "        # Compute Hidden Layer Error\n",
        "        hidden_error = np.dot(d_output, self.W2.T)\n",
        "        d_hidden = hidden_error * self.sigmoid_derivative(self.a1)\n",
        "\n",
        "        # Update Weights and Biases\n",
        "        self.W2 += np.dot(self.a1.T, d_output) * learning_rate\n",
        "        self.b2 += np.sum(d_output, axis=0, keepdims=True) * learning_rate\n",
        "        self.W1 += np.dot(X.T, d_hidden) * learning_rate\n",
        "        self.b1 += np.sum(d_hidden, axis=0, keepdims=True) * learning_rate\n"
      ],
      "metadata": {
        "id": "ehVrGAOaNjzg"
      },
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "VepjtgpKNokr"
      },
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "ef4bHpPANoq-"
      },
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "FVEdJVHeNpv5"
      },
      "execution_count": 25,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Step 5: Train the Neural Network"
      ],
      "metadata": {
        "id": "CZ73UXZ0No55"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "nn = NeuralNetwork(input_size=784, hidden_size=128, output_size=10)"
      ],
      "metadata": {
        "id": "olVdGyhXPChX"
      },
      "execution_count": 26,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "epochs = 50"
      ],
      "metadata": {
        "id": "jBl1I0LTPMM_"
      },
      "execution_count": 27,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "for epoch in range(epochs):\n",
        "    output = nn.forward(x_train)\n",
        "    nn.backward(x_train, y_train, learning_rate=0.1)\n",
        "    if epoch % 10 == 0:\n",
        "        loss = np.mean((y_train - output) ** 2)\n",
        "        print(f'Epoch {epoch}, Loss: {loss:.4f}')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "IwGE9sqGPRhE",
        "outputId": "6934843d-894a-4145-9fc2-d63874b83f03"
      },
      "execution_count": 28,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 0, Loss: 0.2485\n",
            "Epoch 10, Loss: 0.1000\n",
            "Epoch 20, Loss: 0.1000\n",
            "Epoch 30, Loss: 0.1000\n",
            "Epoch 40, Loss: 0.1000\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "6wA94CK5O-MR"
      },
      "execution_count": 28,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "lKhK_FCjLPsl"
      },
      "execution_count": 28,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Step 6: Model Evaluation"
      ],
      "metadata": {
        "id": "MGd0rG0qPYth"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "predictions = nn.forward(x_test)"
      ],
      "metadata": {
        "id": "LkgIoAs_Pbdl"
      },
      "execution_count": 29,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "predicted_labels = np.argmax(predictions, axis=1)"
      ],
      "metadata": {
        "id": "ihd4KOq3LPvH"
      },
      "execution_count": 30,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "true_labels = np.argmax(y_test, axis=1)"
      ],
      "metadata": {
        "id": "g45P8p6cLPza"
      },
      "execution_count": 31,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "accuracy = np.mean(predicted_labels == true_labels)"
      ],
      "metadata": {
        "id": "qXLJ2BS_LP41"
      },
      "execution_count": 32,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(f'Test Accuracy: {accuracy * 100:.2f}%')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "epsUxA4nLMMy",
        "outputId": "333cf3e0-264e-4629-89d0-d58b737dc48c"
      },
      "execution_count": 33,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test Accuracy: 9.80%\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "XQaaFSNGPokF"
      },
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "WTzCKkF-Popi"
      },
      "execution_count": 33,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Step 7: Save the Model"
      ],
      "metadata": {
        "id": "CGW9HJbGPrDG"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "np.savez('mnist_model.npz', W1=nn.W1, W2=nn.W2, b1=nn.b1, b2=nn.b2)"
      ],
      "metadata": {
        "id": "n3OXELfYPuLC"
      },
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "RNz9lY4KPx-I"
      },
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "tBKp6lmvPyM2"
      },
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Step 8: Load the Model"
      ],
      "metadata": {
        "id": "UN9szX9xPyaB"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def load_model():\n",
        "    data = np.load('mnist_model.npz')\n",
        "    nn.W1 = data['W1']\n",
        "    nn.W2 = data['W2']\n",
        "    nn.b1 = data['b1']\n",
        "    nn.b2 = data['b2']"
      ],
      "metadata": {
        "id": "JNPwT-GrLMRN"
      },
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "nls1GWIHLMU0"
      },
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "mwWlmiV2LMYR"
      },
      "execution_count": 35,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Step 9: Visualize Predictions and Results"
      ],
      "metadata": {
        "id": "SK4M2XRIQC3E"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "fig, axes = plt.subplots(2, 5, figsize=(10, 5))\n",
        "axes = axes.ravel()\n",
        "for i in range(10):\n",
        "    axes[i].imshow(x_test[i].reshape(28, 28), cmap='gray')\n",
        "    axes[i].set_title(f'Pred: {predicted_labels[i]}')\n",
        "    axes[i].axis('off')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 321
        },
        "id": "UqA2nxMdQGAg",
        "outputId": "34b0aa72-7403-4916-80ae-10b871723693"
      },
      "execution_count": 36,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1000x500 with 10 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Usnr7QbyLMcY"
      },
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "QpdMH4WiLMhn"
      },
      "execution_count": 36,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Step 10: Confusion Matrix"
      ],
      "metadata": {
        "id": "h1_8hzU3QTML"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "conf_matrix = confusion_matrix(true_labels, predicted_labels)"
      ],
      "metadata": {
        "id": "VgBRLXChQX1f"
      },
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "sns.heatmap(conf_matrix, annot=True, fmt='d', cmap='Blues')\n",
        "plt.xlabel('Predicted')\n",
        "plt.ylabel('Actual')\n",
        "plt.title('Confusion Matrix')\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 472
        },
        "id": "lHK26ldiLMkV",
        "outputId": "52ea07d8-cb59-4f6f-d3d2-c3c464ce4a9d"
      },
      "execution_count": 38,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 640x480 with 2 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiQAAAHHCAYAAACPy0PBAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy81sbWrAAAACXBIWXMAAA9hAAAPYQGoP6dpAABvbklEQVR4nO3dd3xTVR8G8CfpSPeg0IXs0cHeFJRZtuwhglA2QlG2UHZBLFQBQbbKsIAgCAgoy6IUpJRZNmVTVlu66Urb5L5/IHmNLaQjtzeG5+vnfj5vz70550nklV/POfdGJgiCACIiIiIJyaUOQERERMSChIiIiCTHgoSIiIgkx4KEiIiIJMeChIiIiCTHgoSIiIgkx4KEiIiIJMeChIiIiCTHgoSIiIgkx4KESES3b99G+/btYW9vD5lMhr179+q1/wcPHkAmk2HTpk167fe/rFWrVmjVqpXUMYiokFiQkNG7e/cuRo8ejcqVK8PCwgJ2dnZo3rw5li9fjszMTFHH9vPzw5UrV7Bw4UKEhISgYcOGoo5XkoYMGQKZTAY7O7t8P8fbt29DJpNBJpPhq6++KnT/T58+xbx58xAZGamHtERk6EylDkAkpl9//RV9+/aFQqHA4MGDUbNmTWRnZ+PkyZOYOnUqrl27hvXr14sydmZmJsLDwzFz5kyMGzdOlDEqVKiAzMxMmJmZidK/LqampsjIyMD+/fvRr18/rXNbt26FhYUFsrKyitT306dPERgYiIoVK6Ju3boFft2RI0eKNB4RSYsFCRmt+/fvo3///qhQoQKOHTsGNzc3zTl/f3/cuXMHv/76q2jjP3/+HADg4OAg2hgymQwWFhai9a+LQqFA8+bN8eOPP+YpSLZt24YuXbrg559/LpEsGRkZsLKygrm5eYmMR0T6xSUbMlrBwcFIS0vD999/r1WMvFK1alWMHz9e83Nubi4WLFiAKlWqQKFQoGLFipgxYwaUSqXW6ypWrIj3338fJ0+eROPGjWFhYYHKlSvjhx9+0Fwzb948VKhQAQAwdepUyGQyVKxYEcDLpY5X//uf5s2bB5lMptV29OhRvPvuu3BwcICNjQ08PDwwY8YMzfnX7SE5duwY3nvvPVhbW8PBwQHdu3fHjRs38h3vzp07GDJkCBwcHGBvb4+hQ4ciIyPj9R/svwwYMAAHDx5EcnKypu3s2bO4ffs2BgwYkOf6xMRETJkyBbVq1YKNjQ3s7OzQqVMnXLp0SXPNn3/+iUaNGgEAhg4dqln6efU+W7VqhZo1a+L8+fNo0aIFrKysNJ/Lv/eQ+Pn5wcLCIs/779ChAxwdHfH06dMCv1ciEg8LEjJa+/fvR+XKldGsWbMCXT9ixAjMmTMH9evXx7Jly9CyZUsEBQWhf//+ea69c+cO+vTpg3bt2mHJkiVwdHTEkCFDcO3aNQBAr169sGzZMgDAhx9+iJCQEHz99deFyn/t2jW8//77UCqVmD9/PpYsWYJu3brhr7/+euPrfv/9d3To0AFxcXGYN28eJk2ahFOnTqF58+Z48OBBnuv79euHFy9eICgoCP369cOmTZsQGBhY4Jy9evWCTCbD7t27NW3btm2Dp6cn6tevn+f6e/fuYe/evXj//fexdOlSTJ06FVeuXEHLli01xYGXlxfmz58PABg1ahRCQkIQEhKCFi1aaPpJSEhAp06dULduXXz99ddo3bp1vvmWL1+OMmXKwM/PDyqVCgCwbt06HDlyBN988w3c3d0L/F6JSEQCkRFKSUkRAAjdu3cv0PWRkZECAGHEiBFa7VOmTBEACMeOHdO0VahQQQAghIWFadri4uIEhUIhTJ48WdN2//59AYDw5ZdfavXp5+cnVKhQIU+GuXPnCv/8v+SyZcsEAMLz589fm/vVGBs3btS01a1bV3B2dhYSEhI0bZcuXRLkcrkwePDgPOMNGzZMq8+ePXsKTk5Orx3zn+/D2tpaEARB6NOnj9C2bVtBEARBpVIJrq6uQmBgYL6fQVZWlqBSqfK8D4VCIcyfP1/Tdvbs2Tzv7ZWWLVsKAIS1a9fme65ly5ZabYcPHxYACJ9//rlw7949wcbGRujRo4fO90hEJYczJGSUUlNTAQC2trYFuv63334DAEyaNEmrffLkyQCQZ6+Jt7c33nvvPc3PZcqUgYeHB+7du1fkzP/2au/JL7/8ArVaXaDXPHv2DJGRkRgyZAhKlSqlaa9duzbatWuneZ//9PHHH2v9/N577yEhIUHzGRbEgAED8OeffyImJgbHjh1DTExMvss1wMt9J3L5y//0qFQqJCQkaJajLly4UOAxFQoFhg4dWqBr27dvj9GjR2P+/Pno1asXLCwssG7dugKPRUTiY0FCRsnOzg4A8OLFiwJd//DhQ8jlclStWlWr3dXVFQ4ODnj48KFWe/ny5fP04ejoiKSkpCImzuuDDz5A8+bNMWLECLi4uKB///746aef3licvMrp4eGR55yXlxfi4+ORnp6u1f7v9+Lo6AgAhXovnTt3hq2tLXbs2IGtW7eiUaNGeT7LV9RqNZYtW4Zq1apBoVCgdOnSKFOmDC5fvoyUlJQCj1m2bNlCbWD96quvUKpUKURGRmLFihVwdnYu8GuJSHwsSMgo2dnZwd3dHVevXi3U6/69qfR1TExM8m0XBKHIY7za3/CKpaUlwsLC8Pvvv2PQoEG4fPkyPvjgA7Rr1y7PtcVRnPfyikKhQK9evbB582bs2bPntbMjAPDFF19g0qRJaNGiBbZs2YLDhw/j6NGjqFGjRoFngoCXn09hXLx4EXFxcQCAK1euFOq1RCQ+FiRktN5//33cvXsX4eHhOq+tUKEC1Go1bt++rdUeGxuL5ORkzR0z+uDo6Kh1R8or/56FAQC5XI62bdti6dKluH79OhYuXIhjx47hjz/+yLfvVzmjoqLynLt58yZKly4Na2vr4r2B1xgwYAAuXryIFy9e5LsR+JVdu3ahdevW+P7779G/f3+0b98evr6+eT6TghaHBZGeno6hQ4fC29sbo0aNQnBwMM6ePau3/omo+FiQkNH67LPPYG1tjREjRiA2NjbP+bt372L58uUAXi45AMhzJ8zSpUsBAF26dNFbripVqiAlJQWXL1/WtD179gx79uzRui4xMTHPa189IOzftyK/4ubmhrp162Lz5s1af8FfvXoVR44c0bxPMbRu3RoLFizAypUr4erq+trrTExM8sy+7Ny5E0+ePNFqe1U45Ve8Fda0adMQHR2NzZs3Y+nSpahYsSL8/Pxe+zkSUcnjg9HIaFWpUgXbtm3DBx98AC8vL60ntZ46dQo7d+7EkCFDAAB16tSBn58f1q9fj+TkZLRs2RJnzpzB5s2b0aNHj9feUloU/fv3x7Rp09CzZ098+umnyMjIwJo1a1C9enWtTZ3z589HWFgYunTpggoVKiAuLg6rV6/GO++8g3ffffe1/X/55Zfo1KkTfHx8MHz4cGRmZuKbb76Bvb095s2bp7f38W9yuRyzZs3Sed3777+P+fPnY+jQoWjWrBmuXLmCrVu3onLlylrXValSBQ4ODli7di1sbW1hbW2NJk2aoFKlSoXKdezYMaxevRpz587V3Ia8ceNGtGrVCrNnz0ZwcHCh+iMikUh8lw+R6G7duiWMHDlSqFixomBubi7Y2toKzZs3F7755hshKytLc11OTo4QGBgoVKpUSTAzMxPKlSsnBAQEaF0jCC9v++3SpUuecf59u+nrbvsVBEE4cuSIULNmTcHc3Fzw8PAQtmzZkue239DQUKF79+6Cu7u7YG5uLri7uwsffvihcOvWrTxj/PvW2N9//11o3ry5YGlpKdjZ2Qldu3YVrl+/rnXNq/H+fVvxxo0bBQDC/fv3X/uZCoL2bb+v87rbfidPniy4ubkJlpaWQvPmzYXw8PB8b9f95ZdfBG9vb8HU1FTrfbZs2VKoUaNGvmP+s5/U1FShQoUKQv369YWcnByt6yZOnCjI5XIhPDz8je+BiEqGTBAKsXONiIiISATcQ0JERESSY0FCREREkmNBQkRERJJjQUJERESSY0FCREREkmNBQkRERJJjQUJERESSM8ontVr32Sh1hDdK2F6wr0wnIiLxWZTA34SW9cbppZ/Miyv10o8h4gwJERERSc4oZ0iIiIgMioy//+vCgoSIiEhsMpnUCQweCxIiIiKxcYZEJ35CREREJDnOkBAREYmNSzY6sSAhIiISG5dsdOInRERERJLjDAkREZHYuGSjEwsSIiIisXHJRid+QkRERCQ5zpAQERGJjUs2Or21MyQ2FqYIHtIYN9b0RfzWQQhd2AX1q5TWnLe2MMWS4U1xa10/xG8dhHPLemJ4ew+tPhRmJlg6oimiN36I2JCPsHVKazjbW5TYe9i+bSs6tWuDRvVqYWD/vrhy+XKJjV0QzFd0hpwNYL7iMORsAPOJRibXz2HEjPvdvcGqMe+idR13jFgRhsaT9yL00hMcmNMBbqWsAACL/BqjXd2yGL4iDPUn7MGqX69h6fCm6NywnKaPxUMao3ODchi05E90mHsQbo5W2Da1TYnkP3TwN3wVHITRY/2xfeceeHh4Yszo4UhISCiR8XVhPuPMBjCfsWYDmI+k9VYWJBbmJujRtAJmhZzDXzdicS/mBb74KRL3YlIxsr0nAKCphzO2Hr+DE9diEP08DRt/v4UrDxLRsGoZAICdlRn82lTD9M1ncPzqM0TeS8DHq07Cx9MFjaqVEf09hGzeiF59+qFHz96oUrUqZs0NhIWFBfbu/ln0sQuC+YwzG8B8xpoNYD5RyWT6OYyYpAVJfHw8goOD0bNnT/j4+MDHxwc9e/bEl19+iefPn4s2rqlcBlMTOZQ5Kq32zGwVfLycAQCno+LQpWE5zYxJixquqOpuj9BLTwAA9SqXhrmZCf64/Ezz+ltPUxD9PA1NPMQtSHKys3Hj+jU09WmmaZPL5WjatBkuX7oo6tgFwXxFZ8jZAOYrDkPOBjCf6Lhko5Nk7+7s2bOoXr06VqxYAXt7e7Ro0QItWrSAvb09VqxYAU9PT5w7d06UsdOycnE6Kg7T+tSBq6Ml5HIZ+r9XGU2ql4Grw8sCZPL3p3HzcQrurP8Aydv9sHdWe0z6Lhx/3YgFALg4WEKZo0JKRrZW33HJmXD5uw+xJCUnQaVSwcnJSavdyckJ8fHxoo5dEMxXdIacDWC+4jDkbADziY4zJDpJdpfNJ598gr59+2Lt2rWQ/etDFgQBH3/8MT755BOEh4e/sR+lUgmlUqn9elUOZCZmb3zdiBVhWDP2Xdz9tj9yVWpE3kvAzr/uo27ll3/Yx3T2RqNqZdAn6Hc8ik9Dcy9XLB3hg2eJGfjjyrM39k1ERESFI1lBcunSJWzatClPMQIAMpkMEydORL169XT2ExQUhMDAQK02U69uMPfu8cbX3Y99gY5zD8JKYQo7SzPEJGdi88RWeBD7AhbmJpj3YX30//IYDl94DAC4+jAJtSuWwvhuNfHHlWeITc6EwswE9lbmWrMkzg6WiE3O0P0BFIOjgyNMTEzybORKSEhA6dKlX/OqksN8RWfI2QDmKw5DzgYwn+iMfLlFHyT7hFxdXXHmzJnXnj9z5gxcXFx09hMQEICUlBStw8yjS4FzZChzEZOcCQdrc/jWdceBs9EwM5HD3MwEgiBoXatSC5DLXxZQF+/FIztHhVa13DTnq7nboXwZG0REibf/BQDMzM3h5V0DEaf/P3ukVqsRERGO2nV0F3FiY76iM+RsAPMVhyFnA5hPdNxDopNkMyRTpkzBqFGjcP78ebRt21ZTfMTGxiI0NBTffvstvvrqK539KBQKKBQKrTZdyzUA4FvHHTKZDLeepqCKqx0WDmqIW09SEPLHbeSqBIRde4aFgxohM1uF6OdpeM/bFQNaVsH0zS+LqNSMHGw+dhuLhjRGUpoSqZk5WDK8KU5HxeHsbXELEgAY5DcUs2dMQ40aNVGzVm1sCdmMzMxM9OjZS/SxC4L5jDMbwHzGmg1gPpKWZAWJv78/SpcujWXLlmH16tVQqV7e8WJiYoIGDRpg06ZN6Nevn2jj21mZI3BgA5R1skZSmhJ7Tz9E4I/nkat6OSsyZNlxBA5ogA2ftoCjjQLR8WkI/PECvjsSpelj2qYzUAsCtk5pA4WZHL9feoqJ3755z4u+dOzUGUmJiVi9cgXi45/Dw9MLq9d9BycDmbpkPuPMBjCfsWYDmE9UcuPekKoPMuHf6xISyMnJ0eySLl26NMzMdM9wvIl1n436iCWahO1DpY5ARER/syiBX80t2yzUSz+Zx2bqpR9DZBDfZWNmZgY3NzfdFxIREZFRMoiChIiIyKgZ+TNE9IEFCRERkdiM/A4ZfeAnRERERJLjDAkREZHYuGSjEwsSIiIisXHJRicWJERERGLjDIlOLNmIiIhIcpwhISIiEhuXbHRiQUJERCQ2LtnoxJKNiIiIJMcZEiIiIrFxyUYnFiRERERi45KNTkZZkKjvnpc6gg78tl8iIqJ/MsqChIiIyKBwyUYnFiRERERiY0GiEz8hIiIiIxUWFoauXbvC3d0dMpkMe/fu1TovCALmzJkDNzc3WFpawtfXF7dv39a6JjExEQMHDoSdnR0cHBwwfPhwpKWlaV1z+fJlvPfee7CwsEC5cuUQHBxc6KwsSIiIiMQmk+nnKKT09HTUqVMHq1atyvd8cHAwVqxYgbVr1yIiIgLW1tbo0KEDsrKyNNcMHDgQ165dw9GjR3HgwAGEhYVh1KhRmvOpqalo3749KlSogPPnz+PLL7/EvHnzsH79+sJ9RIIgCIV+hwbOst44qSO8UdLZlVJHICKiv1mUwOYFy+7r9NJP5i+ji/xamUyGPXv2oEePHgBezo64u7tj8uTJmDJlCgAgJSUFLi4u2LRpE/r3748bN27A29sbZ8+eRcOGDQEAhw4dQufOnfH48WO4u7tjzZo1mDlzJmJiYmBubg4AmD59Ovbu3YubN28WOB9nSIiIiMSmpxkSpVKJ1NRUrUOpVBYp0v379xETEwNfX19Nm729PZo0aYLw8HAAQHh4OBwcHDTFCAD4+vpCLpcjIiJCc02LFi00xQgAdOjQAVFRUUhKSipwHhYkRERE/xFBQUGwt7fXOoKCgorUV0xMDADAxcVFq93FxUVzLiYmBs7OzlrnTU1NUapUKa1r8uvjn2MUBO+yISIiEpue7rIJCAjApEmTtNoUCoVe+pYaCxIiIiKx6elJrQqFQm8FiKurKwAgNjYWbm5umvbY2FjUrVtXc01cXJzW63Jzc5GYmKh5vaurK2JjY7WuefXzq2sKgks2REREb6FKlSrB1dUVoaGhmrbU1FRERETAx8cHAODj44Pk5GScP///J6AfO3YMarUaTZo00VwTFhaGnJwczTVHjx6Fh4cHHB0dC5yHBQkREZHIZDKZXo7CSktLQ2RkJCIjIwG83MgaGRmJ6OhoyGQyTJgwAZ9//jn27duHK1euYPDgwXB3d9fciePl5YWOHTti5MiROHPmDP766y+MGzcO/fv3h7u7OwBgwIABMDc3x/Dhw3Ht2jXs2LEDy5cvz7O0pAuXbIiIiERWlGJCH86dO4fWrVtrfn5VJPj5+WHTpk347LPPkJ6ejlGjRiE5ORnvvvsuDh06BAsLC81rtm7dinHjxqFt27aQy+Xo3bs3VqxYoTlvb2+PI0eOwN/fHw0aNEDp0qUxZ84crWeVFMRb8RyS5vWrYOJgX9T3Lg+3MvboN3E99v95WXO+e5s6GNHnXdTzKg8nB2s0+SAIl2890erjm5n90aaJB9zK2CMtU4nTl+5j1vJfcOvB/9fNMi/mfb7I4OkbsfOw9pf96es5JNu3bcXmjd8jPv45qnt4YvqM2ahVu7Ze+tYH5is6Q84GMF9xGHI24O3MVxLPIbHus1Ev/aTvMt4vZ30rlmysLRW4cusJJgTtyPe8laU5TkXexawVe1/bx8UbjzBq3hbU7fU5uo1dBZlMhgOr/SGXa1e9I+eEoKJvgObY98clfb4VjUMHf8NXwUEYPdYf23fugYeHJ8aMHo6EhARRxiss5jPObADzGWs2gPlEJdPTYcTeioLkyF/XEbj6APb9cTnf8z/+ehZB6w/h2Omo1/axYfdf+OvCXUQ/S0TkzccIXLUf5dxKoYK7k9Z1KS8yEZvwQnMos3P1+l5eCdm8Eb369EOPnr1RpWpVzJobCAsLC+zd/bMo4xUW8xlnNoD5jDUbwHxikmoPyX/JW1GQ6JuVhTkGd2uK+4/j8ThG+yl0Xwf0w6Nji3AiZAoGd28qyvg52dm4cf0amvo007TJ5XI0bdoMly9dFGXMwmC+ojPkbADzFYchZwOYj6Rn0AXJo0ePMGzYMKljaIzq+x6e/7UECeFL0b65N7qMWYmcXJXmfODqA/josw14f8xK7A2NxPKADzD2w5Z6z5GUnASVSgUnJ+3ZGScnJ8THx+t9vMJivqIz5GwA8xWHIWcDmE9snCHRzaDvsklMTMTmzZuxYcOG116jVCrzPMdfUKsgk5voPc/2g2cRGnETrqXtMGGwL7YsHoY2Q5dqlmUWfXtIc+2lqMewslRg4mBfrP7xuN6zEBHRf4exFxP6IGlBsm/fvjeev3fvns4+goKCEBgYqNVm4tIIZm6Ni5UtP6lpWUhNy8Ld6Oc4c/kBnoUFo3ubOvjp0Pl8rz975QFmjOoEczNTZOfoby+Jo4MjTExM8mzkSkhIQOnSpfU2TlExX9EZcjaA+YrDkLMBzCc2FiS6Sbpk06NHD/Ts2RM9evTI9yjIQ1UCAgKQkpKidZi6NBA9u0wmgwwymJu9vqar7fEOElPS9VqMAICZuTm8vGsg4nS4pk2tViMiIhy169TT61hFwXxFZ8jZAOYrDkPOBjAfSU/SGRI3NzesXr0a3bt3z/d8ZGQkGjR4c3GR33P9/71cY21pjirlymh+rljWCbWrl0VSagYexSTB0c4K5Vwd4eZsDwCoXvHltxTGJqQiNuEFKpZ1Qp8ODRAafgPxSWko6+KAyUPbI1OZg8MnrwEAOreoCWcnW5y5/ABZ2Tlo29QTnw1vj69/CIUYBvkNxewZ01CjRk3UrFUbW0I2IzMzEz169hJlvMJiPuPMBjCfsWYDmE9UnCDRSdKCpEGDBjh//vxrCxKZTAZ9PLetvncFHPluvObn4Cm9AQAh+05j1Nwt6NKyFr6dP0hzPmTxy420n6/9DQvX/QZldi6a16uCcQNawdHOCnEJL3Dywh20HrIEz5PSAAA5uSqM7tcCwZN7QyaT4e6j55i2ZDc27D5V7Pz56dipM5ISE7F65QrExz+Hh6cXVq/7Dk4GMnXJfMaZDWA+Y80GMJ+YuGSjm6RPaj1x4gTS09PRsWPHfM+np6fj3LlzaNmycHeq/PtJrYZGX09qJSKi4iuJJ7U6DNyil36St36kl34MkaQzJO+9994bz1tbWxe6GCEiIjI0nCHRzaBv+yUiIjIGLEh0M+gHoxEREdHbgTMkREREIuMMiW4sSIiIiMTGekQnLtkQERGR5DhDQkREJDIu2ejGgoSIiEhkLEh0Y0FCREQkMhYkunEPCREREUmOMyRERERi4wSJTixIiIiIRMYlG924ZENERESSM84ZEtcqUicgIiLS4AyJbsZZkBARERkQFiS6ccmGiIiIJMcZEiIiIpFxhkQ3FiRERERiYz2iE5dsiIiISHKcISEiIhIZl2x0Y0FCREQkMhYkurEgISIiEhkLEt24h4SIiIgkxxkSIiIisXGCRCcWJERERCLjko1uXLIhIiIiyb0VBUnzmmWxa1533NsyEpkHJ6KrT94v35s9yAf3to5C4t5P8OsXvVHF3UHr/M653XBr83Ak/fIJ7m0dhe+ndIRbKWvN+fdqvYOf5nTDva2jEL9nHE6vHIj+rT1FfV/bt21Fp3Zt0KheLQzs3xdXLl8WdbzCYr6iM+RsAPMVhyFnA5hPLDKZTC+HMXsrChJrCzNcufccE1Yfy/f85L4NMbZbXXz6ze9oMeFHpGflYP/nvaAwM9FcE3bpET4K+hV1Rm7CgM/3o7KbPbbNfF9zvqm3G67ef44Bn+9Ho7EhCDl6Hd9N7oBOjSuJ8p4OHfwNXwUHYfRYf2zfuQceHp4YM3o4EhISRBmvsJjPOLMBzGes2QDmExMLEt3eioLkyLkHCPzhFPadupvvef8e9bF4+xkcOH0PVx/EY8RXh+DmZI1uzf4/k/LN3os4czMG0XEvcPrGM3z101k09nSDqcnLj/DLHWcxPyQcp288w/1nKVj1y0UcOf8A3ZtXFeU9hWzeiF59+qFHz96oUrUqZs0NhIWFBfbu/lmU8QqL+YwzG8B8xpoNYD6S1ltRkLxJRVd7uJWyxrGL0Zq21IxsnI2KQRNP93xf42ijQP/Wnjh94ylyVerX9m1vrUDSiyy9Z87JzsaN69fQ1KeZpk0ul6Np02a4fOmi3scrLOYrOkPOBjBfcRhyNoD5xMYZEt0kL0gyMzNx8uRJXL9+Pc+5rKws/PDDD6KO7+poBQCIS8rQao9LyoDL3+de+XzYu4jfMw5Pd45FOWdb9A3c99p+e79XHQ2qu+CHI3nfV3ElJSdBpVLByclJq93JyQnx8fF6H6+wmK/oDDkbwHzFYcjZAOYTnUxPhxGTtCC5desWvLy80KJFC9SqVQstW7bEs2fPNOdTUlIwdOjQN/ahVCqRmpqqdQjqXFHyLtt1Dk3HbUGXGT9DpRbw3ZQO+V7XovY7WDepPcYu/x03og1/bZOIiEhqkhYk06ZNQ82aNREXF4eoqCjY2tqiefPmiI6O1v3ivwUFBcHe3l7ryL37e4FfH/P3zIjzv2ZDnB2tEPuvWZOE1CzceZKMYxejMXjRb+jUuDKaeLppXfNurbL4eV53fLb+OLaF3ihwjsJwdHCEiYlJno1cCQkJKF26tChjFgbzFZ0hZwOYrzgMORvAfGLjko1ukhYkp06dQlBQEEqXLo2qVati//796NChA9577z3cu3evQH0EBAQgJSVF6zCt4lvgDA9iUvAsMR2t65bTtNlamaORhysibj597evkf//BMP/HnTjv1XoHewJ7YNaGk9hw8EqBMxSWmbk5vLxrIOJ0uKZNrVYjIiIctevUE23cgmK+ojPkbADzFYchZwOYT2wsSHST9EmtmZmZMDX9fwSZTIY1a9Zg3LhxaNmyJbZt26azD4VCAYVCodUmk2u/LWsLM63nilR0sUPtymWQ9CILj56/wKq9FzCtfxPceZKMB7EpmDuoGZ4lpGvuymnk4YoG1V1w6tpTJKdloZKbA+YOaoa7T5MRcfPlElOL2u9gd2APrNp7EXv/uq3Zf5Kdo0JSmrJIn8+bDPIbitkzpqFGjZqoWas2toRsRmZmJnr07KX3sYqC+YwzG8B8xpoNYD4xGXktoReSFiSenp44d+4cvLy8tNpXrlwJAOjWrZtexqlfzQVHgvtqfg4e3QoAEHL0GkYtPYIlO8/BysIMKz/1hYONAqeuPUW32buhzFEBADKUOejerCpmfeQDawszxCSm48j5B1gcFIHsv6/5yNcb1hZm+Kx/Y3zWv7FmrLDLj9Bh2i69vI9/6tipM5ISE7F65QrExz+Hh6cXVq/7Dk4GMnXJfMaZDWA+Y80GMB9JSyYIgiDV4EFBQThx4gR+++23fM+PHTsWa9euhVr9+ltr82PZaZk+4okmaf9EqSMQEdHfLErgV/NqUw/ppZ/bX3bUSz+GSNKCRCwsSIiIqKBKoiCp/pl+CpJbwcZbkEj+HBIiIiIiSfeQEBERvQ2M/Q4ZfWBBQkREJDLWI7pxyYaIiIgkxxkSIiIikcnlnCLRhQUJERGRyLhkoxuXbIiIiIyQSqXC7NmzUalSJVhaWqJKlSpYsGAB/vm0D0EQMGfOHLi5ucHS0hK+vr64ffu2Vj+JiYkYOHAg7Ozs4ODggOHDhyMtLU3veVmQEBERiUyK77JZvHgx1qxZg5UrV+LGjRtYvHgxgoOD8c0332iuCQ4OxooVK7B27VpERETA2toaHTp0QFZWluaagQMH4tq1azh69CgOHDiAsLAwjBo1Sm+fzStcsiEiIhKZFEs2p06dQvfu3dGlSxcAQMWKFfHjjz/izJkzAF7Ojnz99deYNWsWunfvDgD44Ycf4OLigr1796J///64ceMGDh06hLNnz6Jhw4YAgG+++QadO3fGV199BXd3d73l5QwJERGRyPQ1Q6JUKpGamqp1KJX5f4Frs2bNEBoailu3bgEALl26hJMnT6JTp04AgPv37yMmJga+vr6a19jb26NJkyYID3/5rcrh4eFwcHDQFCMA4OvrC7lcjoiICL1+RixIiIiI/iOCgoJgb2+vdQQFBeV77fTp09G/f394enrCzMwM9erVw4QJEzBw4EAAQExMDADAxcVF63UuLi6aczExMXB2dtY6b2pqilKlSmmu0Rcu2RAREYlMX09qDQgIwKRJk7TaFApFvtf+9NNP2Lp1K7Zt24YaNWogMjISEyZMgLu7O/z8/PSSR5+MsyAxs5A6ARERkYa+9pAoFIrXFiD/NnXqVM0sCQDUqlULDx8+RFBQEPz8/ODq6goAiI2NhZubm+Z1sbGxqFu3LgDA1dUVcXFxWv3m5uYiMTFR83p94ZINERGREcrIyIBcrv3XvImJCdRqNQCgUqVKcHV1RWhoqOZ8amoqIiIi4OPjAwDw8fFBcnIyzp8/r7nm2LFjUKvVaNKkiV7zGucMCRERkQGR4sv1unbtioULF6J8+fKoUaMGLl68iKVLl2LYsGGaTBMmTMDnn3+OatWqoVKlSpg9ezbc3d3Ro0cPAICXlxc6duyIkSNHYu3atcjJycG4cePQv39/vd5hA7AgISIiEp0Ut/1+8803mD17NsaOHYu4uDi4u7tj9OjRmDNnjuaazz77DOnp6Rg1ahSSk5Px7rvv4tChQ7Cw+P/Wh61bt2LcuHFo27Yt5HI5evfujRUrVug9r0z45yPbjIRltzVSR3ijpN1jpI5ARER/syiBX83rzz+ml34uzGmjl34MEWdIiIiIRCbFks1/DQsSIiIikbEe0Y132RAREZHkOENCREQkMi7Z6MaChIiISGSsR3RjQUJERCQyzpDoxj0kREREJDnOkBAREYmMEyS6vRUzJM1ruGHXrE64t3EwMveNQdcmFfNcM3tAI9zbNBiJO0fi1/ldUcXNXuv8Z33r44/FPZGwcwSebRuW7zjlSttg9+zOSNg5Ag9/GIIvhvjARC7en8Lt27aiU7s2aFSvFgb274srly+LNlZRMF/RGXI2gPmKw5CzAcwnFplMppfDmL0VBYm1wgxX7idgwroT+Z6f3Ksuxr5fC5+uCUOLqT8jXZmD/YHvQ2FmornG3NQEu/+6i28PXsu3D7lcht1zOsPcTI7Wn+3ByK+P4aO2HpgzsLEo7+nQwd/wVXAQRo/1x/ade+Dh4Ykxo4cjISFBlPEKi/mMMxvAfMaaDWA+ktZbUZAcuRCNwK1nsO/0/XzP+3erjcU/nceBiAe4+iARI5Ydg1spK3RrWklzzec/nsU3+y7j6oPEfPvwrVsOXuUcMWxpKC7fT8CRC9GYv/UMRneuATNT/X/MIZs3oleffujRszeqVK2KWXMDYWFhgb27f9b7WEXBfMaZDWA+Y80GMJ+YZDL9HMbsrShI3qSiiy3cSlnj2KXHmrbUjGycvRWHJh4uBe6niacLrj5MRFxypqbt6MVHsLdWwLt8Kb1mzsnOxo3r19DUp5mmTS6Xo2nTZrh86aJexyoK5is6Q84GMF9xGHI2gPnExiUb3SQvSG7cuIGNGzfi5s2bAICbN29izJgxGDZsGI4d0/1lREqlEqmpqVqHoMop8PiujlYAoFVIvPw5Ay5/nysIF0crxCVnaPeR9LJPF4eC91MQSclJUKlUcHJy0mp3cnJCfHy8XscqCuYrOkPOBjBfcRhyNoD5SHqSFiSHDh1C3bp1MWXKFNSrVw+HDh1CixYtcOfOHTx8+BDt27fXWZQEBQXB3t5e68i9c7iE3gEREZFuXLLRTdKCZP78+Zg6dSoSEhKwceNGDBgwACNHjsTRo0cRGhqKqVOnYtGiRW/sIyAgACkpKVqHadUOBc4Qk/RyVsPZwVKr3dnBCrFJGfm9JF+xSRlw/tdMiLPjyz5jkwveT0E4OjjCxMQkz0auhIQElC5dWq9jFQXzFZ0hZwOYrzgMORvAfGLjko1ukhYk165dw5AhQwAA/fr1w4sXL9CnTx/N+YEDB+Kyjlu6FAoF7OzstA6ZiVmBMzyIfYFnieloXecdTZutpRkaVXdGRFRsgfuJuBmLmhVKoYz9/wubtnXfQUq6Ejei898IW1Rm5ubw8q6BiNPhmja1Wo2IiHDUrlNPr2MVBfMVnSFnA5ivOAw5G8B8JD3JH4z2quKTy+WwsLCAvf3/n/9ha2uLlJSUYo9hbWGq9VyRii52qF3JCUkvlHgUn4ZV+y5jWr8GuPM0BQ9iUzF3YGM8S8zQuiunXGkbONoqUK6MDUzkMtSu9HId8+6zFKRn5eL3yEe48SgJ309sg5mbTsPF0QpzBzbBut+uITtXXez38G+D/IZi9oxpqFGjJmrWqo0tIZuRmZmJHj176X2somA+48wGMJ+xZgOYT0zGPruhD5IWJBUrVsTt27dRpUoVAEB4eDjKly+vOR8dHQ03N7dij1O/qjOOfNFd83PwiOYAgJDQmxi1/A8s2R0JKwszrPRvCQdrc5y6HoNu8w5AmaPSvGb2wEYY1NZT83PE8n4AgPYzfsGJq0+hVgvoveA3LB/TAn9+2RPpWbnYeiwK87eeKXb+/HTs1BlJiYlYvXIF4uOfw8PTC6vXfQcnA5m6ZD7jzAYwn7FmA5hPTKxHdJMJgiBINfjatWtRrlw5dOnSJd/zM2bMQFxcHL777rtC9WvZbY0+4okmafcYqSMQEdHfLErgV/NWX5/SSz9/Tmim+6L/KElnSD7++OM3nv/iiy9KKAkRERFJSfI9JERERMaOSza6sSAhIiISGTe16ib5k1qJiIiIOENCREQkMk6Q6MaChIiISGRyViQ6ccmGiIiIJMcZEiIiIpFxgkQ3FiREREQi4102urEgISIiEpmc9YhO3ENCREREkuMMCRERkci4ZKMbCxIiIiKRsR7RzSgLErm5hdQRiIiIqBCMsiAhIiIyJDJwikQXFiREREQi4102uvEuGyIiIpIcZ0iIiIhExrtsdGNBQkREJDLWI7pxyYaIiIgkxxkSIiIikck5RaITCxIiIiKRsR7RjQUJERGRyLipVTfuISEiIiLJcYaEiIhIZJwg0Y0FCRERkci4qVW3t3bJxsbCFMFDGuPGmr6I3zoIoQu7oH6V0prz1hamWDK8KW6t64f4rYNwbllPDG/voTnvaGOOr4Y1wcXlvRC/dRBurumLL4c1gZ2VWYm9h+3btqJTuzZoVK8WBvbviyuXL5fY2AXBfEVnyNkA5isOQ84GMB9J560tSFaNeRet67hjxIowNJ68F6GXnuDAnA5wK2UFAFjk1xjt6pbF8BVhqD9hD1b9eg1LhzdF54blAABujlZwK2WFGT+cRaNJezF61Um0q1sWq8e8WyL5Dx38DV8FB2H0WH9s37kHHh6eGDN6OBISEkpkfF2YzzizAcxnrNkA5hOTTE+HMTO4gkQQBNHHsDA3QY+mFTAr5Bz+uhGLezEv8MVPkbgXk4qR7T0BAE09nLH1+B2cuBaD6Odp2Pj7LVx5kIiGVcsAAK4/SsbAr/7AwfOPcD/2BY5ffYbAHy+gc8NyMCmBb1EK2bwRvfr0Q4+evVGlalXMmhsICwsL7N39s+hjFwTzGWc2gPmMNRvAfGKSyWR6OYyZwRUkCoUCN27cEHUMU7kMpiZyKHNUWu2Z2Sr4eDkDAE5HxaFLw3KaGZMWNVxR1d0eoZeevLZfOyszpGbkQKUWt6jKyc7GjevX0NSnmaZNLpejadNmuHzpoqhjFwTzFZ0hZwOYrzgMORvAfCQ9yTa1Tpo0Kd92lUqFRYsWwcnJCQCwdOnSN/ajVCqhVCq12gRVDmQmr9/LkZaVi9NRcZjWpw5uPk5GXEoW+jWvhCbVy+BuzAsAwOTvT2Plx81xZ/0HyMlVQy0IGLf2L/x1IzbfPp1sFZjepy42/h71xrz6kJScBJVKpfmMNBmcnHD//j3Rx9eF+YrOkLMBzFcchpwNYD6xlcDE+X9egQqSffv2FbjDbt26Fei6r7/+GnXq1IGDg4NWuyAIuHHjBqytrQs0PRUUFITAwECtNlOvbjD37vHG141YEYY1Y9/F3W/7I1elRuS9BOz86z7qVn75h31MZ280qlYGfYJ+x6P4NDT3csXSET54lpiBP6480+rL1tIMP89oh5uPk7HwJ1bqRESkzdiXW/ShQAVJjx49CtSZTCaDSqXSfSGAL774AuvXr8eSJUvQpk0bTbuZmRk2bdoEb2/vAvUTEBCQZ7bF1W+7ztfdj32BjnMPwkphCjtLM8QkZ2LzxFZ4EPsCFuYmmPdhffT/8hgOX3gMALj6MAm1K5bC+G41tQoSGwtT7J3VHmmZOegffAy5KvH3wDg6OMLExCTPRq6EhASULl36Na8qOcxXdIacDWC+4jDkbADzkfQKtIdErVYX6ChoMQIA06dPx44dOzBmzBhMmTIFOTk5RXoDCoUCdnZ2Wseblmv+LUOZi5jkTDhYm8O3rjsOnI2GmYkc5mYmeTbYqtQC5P+Yd7O1NMO+2R2QnatC30W/59mTIhYzc3N4eddAxOlwTZtarUZERDhq16lXIhnehPmKzpCzAcxXHIacDWA+sclk+jmMmaQPRmvUqBHOnz8Pf39/NGzYEFu3bi2xaS3fOu6QyWS49TQFVVztsHBQQ9x6koKQP24jVyUg7NozLBzUCJnZKkQ/T8N73q4Y0LIKpm8+A+BVMdIeVgpTDA8Og52VOexe7n/F89QsqEXe2DrIbyhmz5iGGjVqomat2tgSshmZmZno0bOXqOMWFPMZZzaA+Yw1G8B8YuKSjW5FKkjS09Nx/PhxREdHIzs7W+vcp59+Wqi+bGxssHnzZmzfvh2+vr6FmmUpDjsrcwQObICyTtZISlNi7+mHCPzxvGbJZciy4wgc0AAbPm0BRxsFouPTEPjjBXx35OWm1bqVndC4+ss7cq6u6qPVt9eYnYh+niZq/o6dOiMpMRGrV65AfPxzeHh6YfW67+BkIFOXzGec2QDmM9ZsAPOJSapNrU+ePMG0adNw8OBBZGRkoGrVqti4cSMaNmwI4OW+zblz5+Lbb79FcnIymjdvjjVr1qBatWqaPhITE/HJJ59g//79kMvl6N27N5YvXw4bGxu9ZpUJhXzwx8WLF9G5c2dkZGQgPT0dpUqVQnx8PKysrODs7Ix794q+2/nx48c4f/48fH19YW1tXeR+rPtsLPJrS0LC9qFSRyAior9ZlMBawZAf9fNE2U0f1i7wtUlJSahXrx5at26NMWPGoEyZMrh9+zaqVKmCKlWqAAAWL16MoKAgbN68GZUqVcLs2bNx5coVXL9+HRYWFgCATp064dmzZ1i3bh1ycnIwdOhQNGrUCNu2bdPLe3ql0AVJq1atUL16daxduxb29va4dOkSzMzM8NFHH2H8+PHo1Uv6qTMWJEREVFAlUZAM3X5FL/1s7F+rwNdOnz4df/31F06cOJHveUEQ4O7ujsmTJ2PKlCkAgJSUFLi4uGDTpk3o378/bty4AW9vb5w9e1Yzq3Lo0CF07twZjx8/hru7e/Hf1N8K/WC0yMhITJ48GXK5HCYmJlAqlShXrhyCg4MxY8YMvQUjIiIyFvp6dLxSqURqaqrW8e9ncb2yb98+NGzYEH379oWzszPq1auHb7/9VnP+/v37iImJga+vr6bN3t4eTZo0QXj4y83D4eHhcHBw0BQjAODr6wu5XI6IiAi9fDavFLogMTMzg1z+8mXOzs6Ijo4G8PJNPHr0SK/hiIiI6P+CgoJgb2+vdQQFBeV77b179zT7QQ4fPowxY8bg008/xebNmwEAMTExAAAXFxet17m4uGjOxcTEwNnZWeu8qakpSpUqpblGXwo9UVWvXj2cPXsW1apVQ8uWLTFnzhzEx8cjJCQENWvW1Gs4IiIiYyDX0102+T17S6FQ5HutWq1Gw4YN8cUXXwB4+ff31atXsXbtWvj5+ekljz4Veobkiy++gJubGwBg4cKFcHR0xJgxY/D8+XOsX79e7wGJiIj+6/T1HJL8nr31uoLEzc0tz0NGvby8NCsbrq6uAIDYWO2vRImNjdWcc3V1RVxcnNb53NxcJCYmaq7Rl0LPkPxzHcnZ2RmHDh3SayAiIiIqvubNmyMqSvv71W7duoUKFSoAACpVqgRXV1eEhoaibt26AIDU1FRERERgzJgxAAAfHx8kJyfj/PnzaNCgAQDg2LFjUKvVaNKkiV7zSvpgNCIioreBFA9GmzhxIpo1a4YvvvgC/fr1w5kzZ7B+/XrNaoZMJsOECRPw+eefo1q1aprbft3d3TVfGePl5YWOHTti5MiRWLt2LXJycjBu3Dj0799fr3fYAEUoSCpVqvTGD7Y4zyEhIiIyRlI8qLVRo0bYs2cPAgICMH/+fFSqVAlff/01Bg4cqLnms88+Q3p6OkaNGoXk5GS8++67OHTokOYZJACwdetWjBs3Dm3bttU8GG3FihV6z1vo55AsX75c6+ecnBxcvHgRhw4dwtSpUzF9+nS9BiwKPoeEiIgKqiSeQzJ61zW99LOuTw299GOICv2vYfz48fm2r1q1CufOnSt2ICIiImOjr7tsjFmh77J5nU6dOuHnn3/WV3dERERGg9/2q5veJqp27dqFUqVK6as7IiIio8Fv+9WtSA9G++cHKwgCYmJi8Pz5c6xevVqv4YiIiOjtUOiCpHv37loFiVwuR5kyZdCqVSt4enrqNVxR2Ze2lzoCERGRht72RxixQhck8+bNEyEGERGR8eKSjW6FLtpMTEzyPEYWABISEmBiYqKXUERERPR2KfQMyeseW6JUKmFubl7sQERERMZGzgkSnQpckLx6KptMJsN3330HGxsbzTmVSoWwsDCD2UNCRERkSFiQ6FbggmTZsmUAXs6QrF27Vmt5xtzcHBUrVsTatWv1n5CIiIiMXoELkvv37wMAWrdujd27d8PR0VG0UERERMaEm1p1K/Qekj/++EOMHEREREaLSza6Ffoum969e2Px4sV52oODg9G3b1+9hCIiIqK3S6ELkrCwMHTu3DlPe6dOnRAWFqaXUERERMaE32WjW6GXbNLS0vK9vdfMzAypqal6CUVERGRM+G2/uhV6hqRWrVrYsWNHnvbt27fD29tbL6GIiIiMiVxPhzEr9AzJ7Nmz0atXL9y9exdt2rQBAISGhmLbtm3YtWuX3gMSERGR8St0wdW1a1fs3bsXd+7cwdixYzF58mQ8efIEx44dQ9WqVcXIqHdyGTC1qzdOf94Bd1d0x6kF7TGhs/ZD3UrbKrDMrwEuLOqEuyu6YesnzVHJ2Vpz3sHKDJ9/UAcn5rXD3RXdcfaLjljQrzZsLQpd4xXZ9m1b0aldGzSqVwsD+/fFlcuXS2zsgmC+ojPkbADzFYchZwOYTyzcQ6JbkWaAunTpgr/++gvp6em4d+8e+vXrhylTpqBOnTr6zicK/w4e8GtZCTO3X0LLwKNYuOcqxravhuGtq2iu2TCmKSqUtsbQNafRfuExPE7IwI7x78HS/OUD4VwcLOFib4H5P19Bm/m/Y8Lm82hVwwVLBjcokfdw6OBv+Co4CKPH+mP7zj3w8PDEmNHDkZCQUCLj68J8xpkNYD5jzQYwn5jkMpleDmNW5CWpsLAw+Pn5wd3dHUuWLEGbNm1w+vRpfWYTTcPKpXD40jOEXo3B44QM/HrhKY5fj0Pdii8f9lbZ2QYNKzth+raLuPQwCXdj0zD9x4uwMJOjZ6NyAICop6kYuT4CR6/E4GF8Ov6Keo7Fv1xHu1quMCmBG85DNm9Erz790KNnb1SpWhWz5gbCwsICe3f/LPrYBcF8xpkNYD5jzQYwH0mrUAVJTEwMFi1ahGrVqqFv376ws7ODUqnE3r17sWjRIjRq1EisnHp17l4i3vUsg8rOL7+Px7usPRpXdcKxa7EAAHPTlx+LMketeY0gANm5ajSq6vTafu0szZCWlQuVOv8vINSXnOxs3Lh+DU19mmna5HI5mjZthsuXLoo6dkEwX9EZcjaA+YrDkLMBzCc2LtnoVuAND127dkVYWBi6dOmCr7/+Gh07doSJiYlev78mPT0dP/30E+7cuQM3Nzd8+OGHcHJ6fQEAvPyWYaVSqdUmqHIgMzF77WtWHo6CrYUpwua1g0oQYCKTYdEv17DnzCMAwJ2YF3ickIGAnjUwbetFZChzMaptNbiXsoKLnUW+fZayNseEzp7YcvJ+Id914SUlJ0GlUuX5bJycnHD//j3Rx9eF+YrOkLMBzFcchpwNYD6x8UmtuhW4IDl48CA+/fRTjBkzBtWqVdPL4N7e3jh58iRKlSqFR48eoUWLFkhKSkL16tVx9+5dLFiwAKdPn0alSpVe20dQUBACAwO12mwa9INtww9e+5puDd5Br8bl4L/hLKKepqJGOXsE9q2N2JQs7DwdjVy1gOHrTmPpoPq4sbQrclVqnLj5HKFXY5DfnykbC1P8MK4Zbj1LxZL9N4r6cRAREb21Crxkc/LkSbx48QINGjRAkyZNsHLlSsTHxxdr8Js3byI3NxcAEBAQAHd3dzx8+BBnzpzBw4cPUbt2bcycOfONfQQEBCAlJUXrsKnX642vmd2rJlYevoVfzj3Gzaep+DniEb4NvYNPOnporrkSnYx2C4/BY8I+1J32GwZ+8xccrc0RHZ+u1Ze1whTbPmmO9KxcDF97GrkiL9cAgKODI0xMTPJs5EpISEDp0qVFH18X5is6Q84GMF9xGHI2gPnExk2tuhW4IGnatCm+/fZbPHv2DKNHj8b27dvh7u4OtVqNo0eP4sWLF8UKEh4ejnnz5sHe3h4AYGNjg8DAQJw8efKNr1MoFLCzs9M63rRcAwAW5iZQC9qFg0ot5Ls+9yIrF4lp2ajkbI06FRxx+NIzzTkbC1P8OL45slVqDFkdDmWuOm8HIjAzN4eXdw1EnA7XtKnVakREhKN2nXolkuFNmK/oDDkbwHzFYcjZAOYTG/eQ6Fboh2ZYW1tj2LBhGDZsGKKiovD9999j0aJFmD59Otq1a4d9+/YVqr9XX8mclZUFNzc3rXNly5bF8+fPCxtRp6NXYvBpJ088ScxE1LNU1CzngNG+1bD91APNNe/XL4uENCWeJGbAq6w95verjUORT3H8RhyAv4uRT9+FpbkJPtlwGjaWprCxfPlxJrxQQuyJkkF+QzF7xjTUqFETNWvVxpaQzcjMzESPnm+eHSopzGec2QDmM9ZsAPORtIr1FC8PDw8EBwcjKCgI+/fvx4YNGwrdR9u2bWFqaorU1FRERUWhZs2amnMPHz7Uuam1KGZtv4TPunkj6MO6cLJVIDYlEyEn7mPZr//f/+Fib4F5fWqhtJ0F4v7eW/L1b/8/X6u8AxpULgUACP+8g1b/jWcewuOEDL3n/qeOnTojKTERq1euQHz8c3h4emH1uu/gZCBTl8xnnNkA5jPWbADziYmbWnWTCYIg/qaH1/j3ZtSmTZuiQ4f//+U+depUPH78GD/++GOh+nX/eLde8onl3kpW80REhqIkHrD9RehdvfQzo20V3Rf9R5Xcc87zMXfu3Dee//LLL0soCRERkXg4Q6KbsX95IBEREf0HSDpDQkRE9DbgDIluLEiIiIhEJjP2e3b1gEs2REREJDnOkBAREYmMSza6sSAhIiISGVdsdOOSDREREUmOMyREREQiM/YvxtMHFiREREQi4x4S3bhkQ0RERJLjDAkREZHIuGKjGwsSIiIikcnBikQXoyxI0lPSpY5ARESkwRkS3biHhIiIiCRnlDMkREREhoR32ejGgoSIiEhkfA6JblyyISIiIslxhoSIiEhknCDRjQUJERGRyLhkoxuXbIiIiEhynCEhIiISGSdIdGNBQkREJDIuR+jGz4iIiIgkxxkSIiIikcm4ZqMTCxIiIiKRsRzR7a1dsrGxMEXQ4Ia4sqInYjZ/iCOBHVC/spPm/OqPmyHlx0Fax8/T22j1UcXVFtsmt8K99X3x6PsPcGhuB7zn7VJi72H7tq3o1K4NGtWrhYH9++LK5cslNnZBMF/RGXI2gPmKw5CzAcwnFrlMppejOBYtWgSZTIYJEyZo2rKysuDv7w8nJyfY2Nigd+/eiI2N1XpddHQ0unTpAisrKzg7O2Pq1KnIzc0tVpb8vLUFyTejfNC6lhtGr/4LzT47gGOXn2HvTF+4OVpqrjka+QTVPt6pOYZ/c1Krj58+awNTExm6fn4ULWf+hqvRSdgxtQ2c7S1Ez3/o4G/4KjgIo8f6Y/vOPfDw8MSY0cORkJAg+tgFwXzGmQ1gPmPNBjCfMTt79izWrVuH2rVra7VPnDgR+/fvx86dO3H8+HE8ffoUvXr10pxXqVTo0qULsrOzcerUKWzevBmbNm3CnDlz9J7xrSxILMxM0K1xeczZdgGnbsbhXuwLLPr5Mu7HvMDwdh6a65Q5asSlZGmO5PRszblStgpUdbPDsl+u4Vp0Mu7FvMC8Hy/A2sIU3uUcRH8PIZs3oleffujRszeqVK2KWXMDYWFhgb27fxZ97IJgPuPMBjCfsWYDmE9MMj0dRZGWloaBAwfi22+/haOjo6Y9JSUF33//PZYuXYo2bdqgQYMG2LhxI06dOoXTp08DAI4cOYLr169jy5YtqFu3Ljp16oQFCxZg1apVyM7Oft2QRfJWFiSmJjKYmsihzFZptWdmq9DUo4zm53e9XXBnbV+cW9INS4c1hqONueZc4gslbj1JwYctKsNKYQoTuQxD21ZHXEomIu8nipo/JzsbN65fQ1OfZpo2uVyOpk2b4fKli6KOXRDMV3SGnA1gvuIw5GwA84lNJtPPoVQqkZqaqnUolco3ju3v748uXbrA19dXq/38+fPIycnRavf09ET58uURHh4OAAgPD0etWrXg4vL/7QgdOnRAamoqrl27psdPSOKC5MKFC7h//77m55CQEDRv3hzlypXDu+++i+3bt+vsI79/OYIq542vScvKRcStOEztVQuujpaQy2To924lNK5eGq4OL5dsQi89xcdr/kK3hUcx98eLaO7lgp+ntdVaw+v+xe+oXbEUnmzoj7gfBsC/ixd6LwrVmkkRQ1JyElQqFZycnLTanZycEB8fL+rYBcF8RWfI2QDmKw5DzgYw339FUFAQ7O3ttY6goKDXXr99+3ZcuHAh32tiYmJgbm4OBwcHrXYXFxfExMRorvlnMfLq/Ktz+iRpQTJ06FDcvXsXAPDdd99h9OjRaNiwIWbOnIlGjRph5MiR2LBhwxv7yO9fjvL6fp1jj171F2QyGaJW98HzkAH4uIMndp16ALXw8vzP4Q9w8PxjXH+UjF/PPcIHX/6BBlVLa21a/WpoYzxPyULHwMNoM+sgfj33CNuntIaLg+VrRiUioreRTCbTyxEQEICUlBStIyAgIN8xHz16hPHjx2Pr1q2wsBB/b2NxSXrb7+3bt1GtWjUAwOrVq7F8+XKMHDlSc75Ro0ZYuHAhhg0b9to+AgICMGnSJK22d0bs0jn2/bg0dJl/BFYKU9hamiE2ORMbP30PD+Je5Hv9g7g0xKdmobKrLY5fi0HLGq7oWL8sKoz4CS8yX87ITN5wBq1rumFAi8pYtk+/U1n/5OjgCBMTkzwbuRISElC6dGnRxi0o5is6Q84GMF9xGHI2gPnEpq/f/hUKBRQKRYGuPX/+POLi4lC/fn1Nm0qlQlhYGFauXInDhw8jOzsbycnJWrMksbGxcHV1BQC4urrizJkzWv2+ugvn1TX6IukMiZWVlWaq7cmTJ2jcuLHW+SZNmmgt6eRHoVDAzs5O65CZmBU4Q4YyF7HJmXCwNkeb2u747dzjfK9zL2WFUjYKxCRnAgAsFS9rOfWrKZW/qQVB9G91NDM3h5d3DUScDv//uGo1IiLCUbtOPVHHLgjmKzpDzgYwX3EYcjaA+YxR27ZtceXKFURGRmqOhg0bYuDAgZr/bWZmhtDQUM1roqKiEB0dDR8fHwCAj48Prly5gri4OM01R48ehZ2dHby9vfWaV9IZkk6dOmHNmjX47rvv0LJlS+zatQt16tTRnP/pp59QtWpVUcZuW9sNkMlw52kqKrvaYv6A+rj9NAVbjt+BtcIU03vXxi9nohGXnIlKLi/P34t9gdBLTwEAZ24/R3J6NtaOaYbFu68gMzsXQ9pUQwVnGxy++ESUzP80yG8oZs+Yhho1aqJmrdrYErIZmZmZ6NGzl+4XlwDmM85sAPMZazaA+cQkxZNabW1tUbNmTa02a2trODk5adqHDx+OSZMmoVSpUrCzs8Mnn3wCHx8fNG3aFADQvn17eHt7Y9CgQQgODkZMTAxmzZoFf3//As/UFJSkBcnixYvRvHlztGzZEg0bNsSSJUvw559/wsvLC1FRUTh9+jT27Nkjyth2VuaY278e3EtZISlNiX1norFgRyRyVQJM5QJqlHfEhy2qwN7aDM+SMvHH5Wf4fGcksnPVAF7eZdN7UShm96uH/bPawdREhpuPU/DhV3/ianSSKJn/qWOnzkhKTMTqlSsQH/8cHp5eWL3uOzgZyNQl8xlnNoD5jDUbwHxiMtQntS5btgxyuRy9e/eGUqlEhw4dsHr1as15ExMTHDhwAGPGjIGPjw+sra3h5+eH+fPn6z2LTBAEQfdl4klOTsaiRYuwf/9+3Lt3D2q1Gm5ubmjevDkmTpyIhg0bFrpP+w9DREiqP7Ehg6SOQEREf7MogV/Nd0Y+1Us/feu666UfQyR5QSIGFiRERFRQJVGQ7Lr0TC/99Knjppd+DBG/XI+IiEhkb+VTSAuJBQkREZHIpNjU+l/Doo2IiIgkxxkSIiIikXF+RDcWJERERCLjio1uXLIhIiIiyXGGhIiISGRyLtroxIKEiIhIZFyy0Y1LNkRERCQ5zpAQERGJTMYlG51YkBAREYmMSza6ccmGiIiIJGeUMyTZKYlSRyAiItLgXTa6GWVBQkREZEi4ZKMbCxIiIiKRsSDRjXtIiIiISHKcISEiIhIZb/vVjQUJERGRyOSsR3Tikg0RERFJjjMkREREIuOSjW4sSIiIiETGu2x045INERERSY4zJERERCLjko1uLEiIiIhExrtsdOOSDREREUnurShImtd0x665XXEvZDgyfxuPrj6V81wz+6OmuLdlBBL3+OPXhT1Rxd1Bc668sy3WjPfFjQ1DkLjHH9e+98OsgU1hZqr98fnWL4/jS/shbtcYRP84Ej/O7ILyzraiva/t27aiU7s2aFSvFgb274srly+LNlZRMF/RGXI2gPmKw5CzAcwnFpme/jFmb0VBYm1hhiv34zFh9Z/5np/cpwHGdquLT1ceQ4uJO5CelYP9C3pAYWYCAPAoVwpyuQzjvjmG+mNC8Nn6MIzoXAvz/Zpp+qjgYoedc7riz0uP0WTcNnSbtRdOdhbYPut9Ud7ToYO/4avgIIwe64/tO/fAw8MTY0YPR0JCgijjFRbzGWc2gPmMNRvAfGKSyfRzGLO3oiA5cu4hAn8Ix77wu/me9+9RD4u3n8GB0/dw9UE8Riw5Ajcna3TzqQIAOHr+IUYvO4rQi9F4EJOKXyPuY/nu8+jerKqmj/pVnWEil2HeD6dwPyYFkXef4+ufL6BO5TIwNdH/xxyyeSN69emHHj17o0rVqpg1NxAWFhbYu/tnvY9VFMxnnNkA5jPWbADziUmmp8OYvRUFyZtUdLWDWylrHIuM1rSlZmTjbFQMmni5vvZ1dtYKJKZlaX6+cCcOakHA4HY1IJfLYGdljgFtPXEsMhq5KrVeM+dkZ+PG9Wto6vP/GRq5XI6mTZvh8qWLeh2rKJiv6Aw5G8B8xWHI2QDmI+lJWpB88sknOHHiRLH6UCqVSE1N1ToEVW6BX+/qaA0AiEvK0GqPS86Ay9/n/q2ymz3GdK2D73+7oml7GJuK92fuRaBfM6T8Mg6xu8agrJMNPgr6rQjv6s2SkpOgUqng5OSk1e7k5IT4+Hi9j1dYzFd0hpwNYL7iMORsAPOJTS6T6eUwZpIWJKtWrUKrVq1QvXp1LF68GDExMYXuIygoCPb29lpH7r2jIqR9yd3JGvsW9MDuk7ex8fA1TbuLoxVWj2+LraHX8e747fD9bCeyc9XYNqOLaFmIiOi/gUs2ukm+ZHPkyBF07twZX331FcqXL4/u3bvjwIEDUKsLtswREBCAlJQUrcO0crsCjx+TlA4AcHa00mp3drBC7N/nXnErZY1Di3rj9I1n8F8RqnVu9Pu1kZqejZkb/sKle8/x19WnGPblYbSpVx6NPV6/9FMUjg6OMDExybORKyEhAaVLl9brWEXBfEVnyNkA5isOQ84GMB9JT/KCpFatWvj666/x9OlTbNmyBUqlEj169EC5cuUwc+ZM3Llz542vVygUsLOz0zpkJgV/3tuDmFQ8S0xH6zrlNG22luZo5OGKiBv/n7Fxd7LG4cW9cfF2HEYtOwpB0O7HSmEG9b8aVX8XVXI9PxHHzNwcXt41EHE6XNOmVqsRERGO2nXq6XWsomC+ojPkbADzFYchZwOYT3ScItHJYJ7UamZmhn79+qFfv36Ijo7Ghg0bsGnTJixatAgqlapYfVtbmKGKu73m54ou9qhduTSSXijx6PkLrNp7EdP6N8adp8l4EJuKuYN88CwhXXNXjruTNQ4v6oPouFQEfH8CZewtNX3F/r335ODZ+/ikRz0EfNgYPx2/BVtLMwT6NcPD2FRE3o0rVv78DPIbitkzpqFGjZqoWas2toRsRmZmJnr07KX3sYqC+YwzG8B8xpoNYD4xGfszRPTBYAqSfypfvjzmzZuHuXPn4vfffy92f/WrOePI4j6an4NHtQAAhBy9jlHLjmLJrvOwsjDDyk/awsFGgVPXnqLbnL1Q5rwshNrUK4+qZR1QtawD7oaM0OrbsvNyAMDxS48xJPgQJvZpgEl9GiBDmYuIG8/QbfZeZGUXr6DKT8dOnZGUmIjVK1cgPv45PDy9sHrdd3AykKlL5jPObADzGWs2gPlIWjJB+PfiQ8mpVKkSzp07l2fXdHG9KhIMVdK+8VJHICKiv1mUwK/mZ+6l6KWfxpXtdV/0HyXpDMn9+/elHJ6IiKhEcMFGN8k3tRIREREZ5B4SIiIio8IpEp1YkBAREYmMd9noxoKEiIhIZEb+1He94B4SIiIikhxnSIiIiETGCRLdWJAQERGJjRWJTlyyISIiIslxhoSIiEhkvMtGNxYkREREIuNdNrpxyYaIiIgkxxkSIiIikXGCRDejLEhMbIz32xCJiOg/iBWJTlyyISIiIskZ5QwJERGRIeFdNrqxICEiIhIZ77LRjUs2REREIpPp6SiMoKAgNGrUCLa2tnB2dkaPHj0QFRWldU1WVhb8/f3h5OQEGxsb9O7dG7GxsVrXREdHo0uXLrCysoKzszOmTp2K3NzcQqbRjQUJERGRETp+/Dj8/f1x+vRpHD16FDk5OWjfvj3S09M110ycOBH79+/Hzp07cfz4cTx9+hS9evXSnFepVOjSpQuys7Nx6tQpbN68GZs2bcKcOXP0nlcmCIKg914lZtNvk9QR3ih+2xCpIxAR0d8sSmDzwtUnaXrpp2ZZmyK/9vnz53B2dsbx48fRokULpKSkoEyZMti2bRv69OkDALh58ya8vLwQHh6Opk2b4uDBg3j//ffx9OlTuLi4AADWrl2LadOm4fnz5zA3N9fL+wI4Q0JERCQ6mZ7+KY6UlBQAQKlSpQAA58+fR05ODnx9fTXXeHp6onz58ggPDwcAhIeHo1atWppiBAA6dOiA1NRUXLt2rVh5/o2bWomIiP4jlEollEqlVptCoYBCoXjj69RqNSZMmIDmzZujZs2aAICYmBiYm5vDwcFB61oXFxfExMRorvlnMfLq/Ktz+sQZEiIiIpHJZPo5goKCYG9vr3UEBQXpHN/f3x9Xr17F9u3bS+DdFg1nSIiIiESmr7t+AwICMGnSJK02XbMj48aNw4EDBxAWFoZ33nlH0+7q6ors7GwkJydrzZLExsbC1dVVc82ZM2e0+nt1F86ra/SFMyRERET/EQqFAnZ2dlrH6woSQRAwbtw47NmzB8eOHUOlSpW0zjdo0ABmZmYIDQ3VtEVFRSE6Oho+Pj4AAB8fH1y5cgVxcXGaa44ePQo7Ozt4e3vr9b29tQWJjYUpFvs1xvVVffB8y0f4fUFn1K/ipDmf9tOQfI/xXWvk6cvcVI5Twd2Q9tMQ1KpQqsTew/ZtW9GpXRs0qlcLA/v3xZXLl0ts7IJgvqIz5GwA8xWHIWcDmE80EjyIxN/fH1u2bMG2bdtga2uLmJgYxMTEIDMzEwBgb2+P4cOHY9KkSfjjjz9w/vx5DB06FD4+PmjatCkAoH379vD29sagQYNw6dIlHD58GLNmzYK/v7/OmZnCemsLklUfN0eb2m4YufIEmkz+BccuP8X+2R3g5mgFAKg8cofW8fHqk1CrBfwS8TBPX59/1BDPEjNKNP+hg7/hq+AgjB7rj+0798DDwxNjRg9HQkJCieZ4HeYzzmwA8xlrNoD5xCTFXTZr1qxBSkoKWrVqBTc3N82xY8cOzTXLli3D+++/j969e6NFixZwdXXF7t27NedNTExw4MABmJiYwMfHBx999BEGDx6M+fPn6+2zeeWtfA6JhZkJYn4YiA+Cj+Hwxcea9hOL3sfRi08wf8fFPK/5cWob2FqY4v0FR7Ta29Uti0WDG2Hgkj9wbllP+EzdhysPE984vj6eQzKwf1/UqFkLM2a9fDiNWq1G+7Yt8eGAQRg+clSx+y8u5jPObADzGWs24O3NVxLPIbn5TD+/tHq6WemlH0P0Vs6QmJrIYGoihzJHpdWema2Cj6dLnuud7S3Qsd472Hzsdp72laObYcTKE8jIVuV5nVhysrNx4/o1NPVppmmTy+Vo2rQZLl/KW0yVNOYrOkPOBjBfcRhyNoD5xKavu2yM2VtZkKRl5eJ0VBym9a4DV0dLyGUyfPBeZTSpXgYujpZ5rh/QsipeZOVg35lorfa1Y9/F90ejcPFeyU4XJiUnQaVSwcnJSavdyckJ8fHxJZolP8xXdIacDWC+4jDkbADziU2K77L5r5G8IFm5ciUGDx6suTc6JCQE3t7e8PT0xIwZM3R+gY9SqURqaqrWIahydI47cuUJyGTAnXUfIHHbIIzp5IWdf92HoM67gjW4dTX8dOKe1ozKmE5esLE0w1d7rhTyHRMR0VuHFYlOkj6H5PPPP0dwcDDat2+PiRMn4uHDh/jyyy8xceJEyOVyLFu2DGZmZggMDHxtH0FBQXnOm3l3h3mNHm8c+37sC3ScdwhWClPYWpohNjkTmye0xP24F1rXNfN0RvWy9hj89Z9a7S1quqFJ9TJI3DZIq/3Eovex4+Q9jF51UvcHUESODo4wMTHJs5ErISEBpUuXFm3cgmK+ojPkbADzFYchZwOYj6Qn6QzJpk2bsGnTJuzatQuHDh3CzJkzsXz5csycORMBAQFYt24dtm3b9sY+AgICkJKSonWYeXYpcIYMZS5ikzPhYG2OtnXK4tezj7TOD25THRfuxuPqwySt9qkbIuAzdR+affby6BX0OwDA7+vjCPzxQoHHLwozc3N4eddAxOlwTZtarUZERDhq16kn6tgFwXxFZ8jZAOYrDkPOBjCf2Azhu2wMnaQzJE+fPkXDhg0BAHXq1IFcLkfdunU15+vXr4+nT5++sY/8nuEvMzHTOXbbOu6QQYbbT1NQ2dUWCwc1wq0nKQj58/8bV20tzdCzaQXMCDmX5/WPE9KBfxTqaVkvl5buxbzA0xK4BXiQ31DMnjENNWrURM1atbElZDMyMzPRo2cv3S8uAcxnnNkA5jPWbADzicnYN6Tqg6QFiaurK65fv47y5cvj9u3bUKlUuH79OmrUePnwsWvXrsHZ2VmUse2tzDHvw/oo62SNpDQlfol4iMAfLyBX9f89JH2aVYJMJsPOk/dEyVAcHTt1RlJiIlavXIH4+Ofw8PTC6nXfwclApi6ZzzizAcxnrNkA5iNpSfocktmzZ2PdunXo3r07QkND8cEHH2Dbtm0ICAiATCbDwoUL0adPHyxdurRQ/ep6DonU9PEcEiIi0o+SeA7J3bhMvfRTxTnvnaDGQtIZksDAQFhaWiI8PBwjR47E9OnTUadOHXz22WfIyMhA165dsWDBAikjEhERFR+XbHR6K5/UKjXOkBARGY4SmSF5rqcZkjKcISEiIqIiMvY7ZPSBBQkREZHIeJeNbpI/qZWIiIiIMyREREQi4wSJbixIiIiIxMaKRCcWJERERCLjplbduIeEiIiIJMcZEiIiIpHxLhvdWJAQERGJjPWIblyyISIiIslxhoSIiEhkXLLRjQUJERGR6FiR6GKUBYkqSz9fYkREREQlwygLEiIiIkPCJRvdWJAQERGJjPWIbrzLhoiIiCTHGRIiIiKRcclGNxYkREREIuN32ejGgoSIiEhsrEd04h4SIiIikhxnSIiIiETGCRLdWJAQERGJjJtadeOSDREREUmOMyREREQi4102ur0VMyTNa7hh16xOuLdxMDL3jUHXJhXzXDN7QCPc2zQYiTtH4tf5XVHFzV7rvKONAhsntUXs9uF4tm0Y1nzSCtYW2vVc7+ZVcPrrvkjYOQJR332EiT3riviugO3btqJTuzZoVK8WBvbviyuXL4s6XmExX9EZcjaA+YrDkLMBzCcamZ4OI/ZWFCTWCjNcuZ+ACetO5Ht+cq+6GPt+LXy6Jgwtpv6MdGUO9ge+D4WZieaajZN94VW+FN6fsx+9F/yGd2u4YZV/K8359vXLY+Pktvju0DU0GLcD49eG4ZNutfFxl5qivKdDB3/DV8FBGD3WH9t37oGHhyfGjB6OhIQEUcYrLOYzzmwA8xlrNoD5SFpvRUFy5EI0Areewb7T9/M979+tNhb/dB4HIh7g6oNEjFh2DG6lrNCtaSUAgMc7DujQoDzGrvwTZ2/F4dSNGExafxJ936sKt1JWAIABratjf8QDfHfoOh7EvsChc9H4ctcFTO5VT5T3FLJ5I3r16YcePXujStWqmDU3EBYWFti7+2dRxiss5jPObADzGWs2gPnExAkS3d6KguRNKrrYwq2UNY5deqxpS83IxtlbcWji4QIAaOLpiqQ0JS7cea655ljkY6gFAY2qv7xGYWaCrOxcrb4zs1V4p4wNyjvb6jVzTnY2bly/hqY+zTRtcrkcTZs2w+VLF/U6VlEwX9EZcjaA+YrDkLMBzCc2mUw/hzGTtCB59uwZ5syZgzZt2sDLyws1atRA165d8f3330OlUpVIBlfHlzMcccmZWu1xyRlw+fuci6MVnv/rvEotIPGFUnPN0YvR6O5TGa1ql4VMBlR1t8f4HnUAAG5/X6MvSclJUKlUcHJy0mp3cnJCfHy8XscqCuYrOkPOBjBfcRhyNoD5SHqS3WVz7tw5+Pr6omrVqrC0tMTt27cxYMAAZGdnY8qUKdiwYQMOHToEW9s3zy4olUoolUqtNkGVA5mJmZjx89hw+AYqu9pj9+zOMDOVIzUjG6v2X8HsAY2gFoQSzUJERIaFd9noJtkMyYQJEzBx4kScO3cOJ06cwKZNm3Dr1i1s374d9+7dQ0ZGBmbNmqWzn6CgINjb22sduXcOFzhHTFIGAMDZwVKr3dnBCrF/n4tNykCZf503kctQylahuQYAZm0+jdIffAeP4VtQ0W8zzt2KAwDcj0ktcJ6CcHRwhImJSZ6NXAkJCShdurRexyoK5is6Q84GMF9xGHI2gPnExiUb3SQrSC5cuIBBgwZpfh4wYAAuXLiA2NhYODo6Ijg4GLt27dLZT0BAAFJSUrQO06odCpzjQewLPEtMR+s672jabC3N0Ki6MyKiYgEAETdj4GijQL0q//9D36p2WchlMpy9FavVn1ot4GliOnJy1ejXoipO34hBfGpWgfMUhJm5Oby8ayDidPg/xlUjIiIcteuIs4m2MJiv6Aw5G8B8xWHI2QDmI+lJtmTj7OyMZ8+eoXLlygCA2NhY5Obmws7ODgBQrVo1JCYm6uxHoVBAoVBotf17ucbawlTruSIVXexQu5ITkl4o8Sg+Dav2Xca0fg1w52kKHsSmYu7AxniWmKG5KyfqcTIOn4/GqnGt8OnqMJiZyrFs9HvYeeIOniW+nCFxsrVAz+aVEXblKSzMTTC4rSd6Na+C9jN+KfqH9AaD/IZi9oxpqFGjJmrWqo0tIZuRmZmJHj17iTJeYTGfcWYDmM9YswHMR9KSrCDp0aMHPv74Y3z55ZdQKBRYsGABWrZsCUvLl0sjUVFRKFu2rF7Gql/VGUe+6K75OXhEcwBASOhNjFr+B5bsjoSVhRlW+reEg7U5Tl2PQbd5B6DM+f/G2qFLfsey0e/htwVdoRYE7A2/h8nrT2qN81EbDwQNbQaZDIi4GYsOM3/BudtxenkP/9axU2ckJSZi9coViI9/Dg9PL6xe9x2cDGTqkvmMMxvAfMaaDWA+MRn7cos+yARBmh2XaWlpGD58OHbv3g2VSgUfHx9s2bIFlSq9fPbHkSNHkJKSgr59+xa6b8tua/QdV6+Sdo+ROgIREf3NogR+NU/JVOulH3tL431ah2QFyStZWVnIzc2FjY2N3vpkQUJERAXFgsQwSP7lehYWFlJHICIiEhWXbHSTvCAhIiIydqxHdDPeuR8iIiL6z+AMCRERkdg4RaITCxIiIiKR8dHxunHJhoiIiCTHGRIiIiKR8S4b3ViQEBERiYz1iG5csiEiIhKbTE9HEaxatQoVK1aEhYUFmjRpgjNnzhTrrYiFBQkREZGR2rFjByZNmoS5c+fiwoULqFOnDjp06IC4OHG+Z604WJAQERGJTKanfwpr6dKlGDlyJIYOHQpvb2+sXbsWVlZW2LBhgwjvsnhYkBAREYlMJtPPURjZ2dk4f/48fH19NW1yuRy+vr4IDw/X8zssPm5qJSIi+o9QKpVQKpVabQqFAgqFIs+18fHxUKlUcHFx0Wp3cXHBzZs3Rc1ZJAK9UVZWljB37lwhKytL6ij5MuR8hpxNEJivuAw5nyFnEwTmKw5DzlYS5s6dKwDQOubOnZvvtU+ePBEACKdOndJqnzp1qtC4ceMSSFs4MkEQBEkrIgOXmpoKe3t7pKSkwM7OTuo4eRhyPkPOBjBfcRlyPkPOBjBfcRhytpJQmBmS7OxsWFlZYdeuXejRo4em3c/PD8nJyfjll1/Ejlso3ENCRET0H6FQKGBnZ6d15FeMAIC5uTkaNGiA0NBQTZtarUZoaCh8fHxKKnKBcQ8JERGRkZo0aRL8/PzQsGFDNG7cGF9//TXS09MxdOhQqaPlwYKEiIjISH3wwQd4/vw55syZg5iYGNStWxeHDh3Ks9HVELAg0UGhUGDu3LmvnRKTmiHnM+RsAPMVlyHnM+RsAPMVhyFnM1Tjxo3DuHHjpI6hEze1EhERkeS4qZWIiIgkx4KEiIiIJMeChIiIiCTHgoSIiIgkx4JEh1WrVqFixYqwsLBAkyZNcObMGakjAQDCwsLQtWtXuLu7QyaTYe/evVJH0ggKCkKjRo1ga2sLZ2dn9OjRA1FRUVLH0lizZg1q166teaiQj48PDh48KHWsfC1atAgymQwTJkyQOgoAYN68eZDJZFqHp6en1LG0PHnyBB999BGcnJxgaWmJWrVq4dy5c1LHAgBUrFgxz+cnk8ng7+8vdTSoVCrMnj0blSpVgqWlJapUqYIFCxbAkO57ePHiBSZMmIAKFSrA0tISzZo1w9mzZ6WORXrCguQNduzYgUmTJmHu3Lm4cOEC6tSpgw4dOiAuLk7qaEhPT0edOnWwatUqqaPkcfz4cfj7++P06dM4evQocnJy0L59e6Snp0sdDQDwzjvvYNGiRTh//jzOnTuHNm3aoHv37rh27ZrU0bScPXsW69atQ+3ataWOoqVGjRp49uyZ5jh58qTUkTSSkpLQvHlzmJmZ4eDBg7h+/TqWLFkCR0dHqaMBePnv9J+f3dGjRwEAffv2lTgZsHjxYqxZswYrV67EjRs3sHjxYgQHB+Obb76ROprGiBEjcPToUYSEhODKlSto3749fH198eTJE6mjkT5I+k06Bq5x48aCv7+/5meVSiW4u7sLQUFBEqbKC4CwZ88eqWO8VlxcnABAOH78uNRRXsvR0VH47rvvpI6h8eLFC6FatWrC0aNHhZYtWwrjx4+XOpIgCC+/2KtOnTpSx3itadOmCe+++67UMQps/PjxQpUqVQS1Wi11FKFLly7CsGHDtNp69eolDBw4UKJE2jIyMgQTExPhwIEDWu3169cXZs6cKVEq0ifOkLxGdnY2zp8/D19fX02bXC6Hr68vwsPDJUz235OSkgIAKFWqlMRJ8lKpVNi+fTvS09MN6rsd/P390aVLF60/f4bi9u3bcHd3R+XKlTFw4EBER0dLHUlj3759aNiwIfr27QtnZ2fUq1cP3377rdSx8pWdnY0tW7Zg2LBhkMlkUsdBs2bNEBoailu3bgEALl26hJMnT6JTp04SJ3spNzcXKpUKFhYWWu2WlpYGNUtHRccntb5GfHw8VCpVnsfruri44ObNmxKl+u9Rq9WYMGECmjdvjpo1a0odR+PKlSvw8fFBVlYWbGxssGfPHnh7e0sdCwCwfft2XLhwwSDXxps0aYJNmzbBw8MDz549Q2BgIN577z1cvXoVtra2UsfDvXv3sGbNGkyaNAkzZszA2bNn8emnn8Lc3Bx+fn5Sx9Oyd+9eJCcnY8iQIVJHAQBMnz4dqamp8PT0hImJCVQqFRYuXIiBAwdKHQ0AYGtrCx8fHyxYsABeXl5wcXHBjz/+iPDwcFStWlXqeKQHLEhIVP7+/rh69arB/Qbj4eGByMhIpKSkYNeuXfDz88Px48clL0oePXqE8ePH4+jRo3l+EzQE//xtuXbt2mjSpAkqVKiAn376CcOHD5cw2UtqtRoNGzbEF198AQCoV68erl69irVr1xpcQfL999+jU6dOcHd3lzoKAOCnn37C1q1bsW3bNtSoUQORkZGYMGEC3N3dDeazCwkJwbBhw1C2bFmYmJigfv36+PDDD3H+/Hmpo5EesCB5jdKlS8PExASxsbFa7bGxsXB1dZUo1X/LuHHjcODAAYSFheGdd96ROo4Wc3NzzW9VDRo0wNmzZ7F8+XKsW7dO0lznz59HXFwc6tevr2lTqVQICwvDypUroVQqYWJiImFCbQ4ODqhevTru3LkjdRQAgJubW56i0svLCz///LNEifL38OFD/P7779i9e7fUUTSmTp2K6dOno3///gCAWrVq4eHDhwgKCjKYgqRKlSo4fvw40tPTkZqaCjc3N3zwwQeoXLmy1NFID7iH5DXMzc3RoEEDhIaGatrUajVCQ0MNaq+BIRIEAePGjcOePXtw7NgxVKpUSepIOqnVaiiVSqljoG3btrhy5QoiIyM1R8OGDTFw4EBERkYaVDECAGlpabh79y7c3NykjgIAaN68eZ5bzG/duoUKFSpIlCh/GzduhLOzM7p06SJ1FI2MjAzI5dp/JZiYmECtVkuU6PWsra3h5uaGpKQkHD58GN27d5c6EukBZ0jeYNKkSfDz80PDhg3RuHFjfP3110hPT8fQoUOljoa0tDSt30rv37+PyMhIlCpVCuXLl5cw2ctlmm3btuGXX36Bra0tYmJiAAD29vawtLSUNBsABAQEoFOnTihfvjxevHiBbdu24c8//8Thw4eljgZbW9s8e22sra3h5ORkEHtwpkyZgq5du6JChQp4+vQp5s6dCxMTE3z44YdSRwMATJw4Ec2aNcMXX3yBfv364cyZM1i/fj3Wr18vdTQNtVqNjRs3ws/PD6amhvOf4K5du2LhwoUoX748atSogYsXL2Lp0qUYNmyY1NE0Dh8+DEEQ4OHhgTt37mDq1Knw9PQ0iP8mkx5IfZuPofvmm2+E8uXLC+bm5kLjxo2F06dPSx1JEARB+OOPPwQAeQ4/Pz+po+WbC4CwceNGqaMJgiAIw4YNEypUqCCYm5sLZcqUEdq2bSscOXJE6livZUi3/X7wwQeCm5ubYG5uLpQtW1b44IMPhDt37kgdS8v+/fuFmjVrCgqFQvD09BTWr18vdSQthw8fFgAIUVFRUkfRkpqaKowfP14oX768YGFhIVSuXFmYOXOmoFQqpY6msWPHDqFy5cqCubm54OrqKvj7+wvJyclSxyI9kQmCAT2Gj4iIiN5K3ENCREREkmNBQkRERJJjQUJERESSY0FCREREkmNBQkRERJJjQUJERESSY0FCREREkmNBQmSEhgwZgh49emh+btWqFSZMmFDiOf7880/IZDIkJyeX+NhE9N/CgoSoBA0ZMgQymQwymUzzBX/z589Hbm6uqOPu3r0bCxYsKNC1LCKISAqG80UKRG+Jjh07YuPGjVAqlfjtt9/g7+8PMzMzBAQEaF2XnZ0Nc3NzvYxZqlQpvfRDRCQWzpAQlTCFQgFXV1dUqFABY8aMga+vL/bt26dZZlm4cCHc3d3h4eEBAHj06BH69esHBwcHlCpVCt27d8eDBw80/alUKkyaNAkODg5wcnLCZ599hn9/I8S/l2yUSiWmTZuGcuXKQaFQoGrVqvj+++/x4MEDtG7dGgDg6OgImUyGIUOGAHj5pXBBQUGoVKkSLC0tUadOHezatUtrnN9++w3Vq1eHpaUlWrdurZWTiOhNWJAQSczS0hLZ2dkAgNDQUERFReHo0aM4cOAAcnJy0KFDB9ja2uLEiRP466+/YGNjg44dO2pes2TJEmzatAkbNmzAyZMnkZiYiD179rxxzMGDB+PHH3/EihUrcOPGDaxbtw42NjYoV64cfv75ZwBAVFQUnj17huXLlwMAgoKC8MMPP2Dt2rW4du0aJk6ciI8++gjHjx8H8LJw6tWrF7p27YrIyEiMGDEC06dPF+tjIyJjI/GX+xG9Vfz8/ITu3bsLgiAIarVaOHr0qKBQKIQpU6YIfn5+gouLi9a3q4aEhAgeHh6CWq3WtCmVSsHS0lI4fPiwIAiC4ObmJgQHB2vO5+TkCO+8845mHEHQ/sbgqKgoAYBw9OjRfDO++ibppKQkTVtWVpZgZWUlnDp1Suva4cOHCx9++KEgCIIQEBAgeHt7a52fNm1anr6IiPLDPSREJezAgQOwsbFBTk4O1Go1BgwYgHnz5sHf3x+1atXS2jdy6dIl3LlzB7a2tlp9ZGVl4e7du0hJScGzZ8/QpEkTzTlTU1M0bNgwz7LNK5GRkTAxMUHLli0LnPnOnTvIyMhAu3bttNqzs7NRr149AMCNGze0cgCAj49PgccgorcbCxKiEta6dWusWbMG5ubmcHd3h6np//9vaG1trXVtWloaGjRogK1bt+bpp0yZMkUa39LSstCvSUtLAwD8+uuvKFu2rNY5hUJRpBxERP/EgoSohFlbW6Nq1aoFurZ+/frYsWMHnJ2dYWdnl+81bm5uiIiIQIsWLQAAubm5OH/+POrXr5/v9bVq1YJarcbx48fh6+ub5/yrGRqVSqVp8/b2hkKhQHR09GtnVry8vLBv3z6tttOnT+t+k0RE4KZWIoM2cOBAlC5dGt27d8eJEydw//59/Pnnn/j000/x+PFjAMD48eOxaNEi7N27Fzdv3sTYsWPf+AyRihUrws/PD8OGDcPevXs1ff70008AgAoVKkAmk+HAgQN4/vw50tLSYGtriylTpmDixInYvHkz7t69iwsXLuCbb77B5s2bAQAff/wxbt++jalTpyIqKgrbtm3Dpk2bxP6IiMhIsCAhMmBWVlYICwtD+fLl0atXL3h5eWH48OHIysrSzJhMnjwZgwYNgp+fH3x8fGBra4uePXu+sd81a9agT58+GDt2LDw9PTFy5Eikp6cDAMqWLYvAwEBMnz4dLi4uGDduHABgwYIFmD17NoKCguDl5YWOHTvi119/RaVKlQAA5cuXx88//4y9e/eiTp06WLt2Lb744gsRPx0iMiYy4XU734iIiIhKCGdIiIiISHIsSIiIiEhyLEiIiIhIcixIiIiISHIsSIiIiEhyLEiIiIhIcixIiIiISHIsSIiIiEhyLEiIiIhIcixIiIiISHIsSIiIiEhyLEiIiIhIcv8DZlTGSz/nQH8AAAAASUVORK5CYII=\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "SKsUtVZPQerE"
      },
      "execution_count": 38,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "xoXCSp7iQe-9"
      },
      "execution_count": 38,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "jXvy4GA7QffQ"
      },
      "execution_count": 38,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "# Step 11: Interactive Dashboard with Dash"
      ],
      "metadata": {
        "id": "3EIA4SgCQf-t"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "app = dash.Dash(__name__)"
      ],
      "metadata": {
        "id": "ZleXv0w8QhKL"
      },
      "execution_count": 39,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "server = app.server"
      ],
      "metadata": {
        "id": "5-G2LEGWLMmJ"
      },
      "execution_count": 40,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "8-ojP7w6LMoD"
      },
      "execution_count": 40,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "app.layout = html.Div([\n",
        "    html.H1(\"MNIST Digit Classifier\"),\n",
        "    dcc.Upload(\n",
        "        id='upload-image',\n",
        "        children=html.Button('Upload Image'),\n",
        "        multiple=False\n",
        "    ),\n",
        "    html.Div(id='output-image-upload'),\n",
        "    html.H3(\"Prediction: \"),\n",
        "    html.Div(id='prediction-output')\n",
        "])"
      ],
      "metadata": {
        "id": "rZp6KY2rLMqd"
      },
      "execution_count": 41,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "B_xE6HdYQrP8"
      },
      "execution_count": 41,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Image Preprocessing Function"
      ],
      "metadata": {
        "id": "JQNvUCx3Qrnf"
      },
      "execution_count": 42,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def preprocess_image(contents):\n",
        "    content_string = contents.split(',')[1]\n",
        "    decoded = base64.b64decode(content_string)\n",
        "    image = Image.open(io.BytesIO(decoded)).convert('L')\n",
        "    image = image.resize((28, 28))\n",
        "    image_array = np.array(image).reshape(1, 784) / 255.0\n",
        "    return image_array\n"
      ],
      "metadata": {
        "id": "VzjZzNBALMue"
      },
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "dJsr-BnvLMw6"
      },
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Callback Function for Upload and Prediction"
      ],
      "metadata": {
        "id": "BRbMBbdBLMze"
      },
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "@app.callback(\n",
        "    [Output('output-image-upload', 'children'),\n",
        "     Output('prediction-output', 'children')],\n",
        "    [Input('upload-image', 'contents')]\n",
        ")\n",
        "def update_output(contents):\n",
        "    if contents is None:\n",
        "        return None, \"No image uploaded\"\n",
        "\n",
        "    image_array = preprocess_image(contents)\n",
        "    prediction = nn.forward(image_array)\n",
        "    predicted_digit = np.argmax(prediction)\n",
        "\n",
        "    return html.Img(src=contents, style={'height': '100px'}), f'Predicted Digit: {predicted_digit}'\n"
      ],
      "metadata": {
        "id": "sH7pdCMSLM8_"
      },
      "execution_count": 45,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "mCHqQnKcQ_oj"
      },
      "execution_count": 45,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Run the Dash App"
      ],
      "metadata": {
        "id": "CheW1EeAQ_0Q"
      },
      "execution_count": 46,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "if __name__ == '__main__':\n",
        "    app.run_server(debug=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 671
        },
        "id": "K41kE4lJRBHR",
        "outputId": "11cc30e8-8cb5-4714-e0d7-ba7c8531c926"
      },
      "execution_count": 47,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<IPython.core.display.Javascript object>"
            ],
            "application/javascript": [
              "(async (port, path, width, height, cache, element) => {\n",
              "    if (!google.colab.kernel.accessAllowed && !cache) {\n",
              "      return;\n",
              "    }\n",
              "    element.appendChild(document.createTextNode(''));\n",
              "    const url = await google.colab.kernel.proxyPort(port, {cache});\n",
              "    const iframe = document.createElement('iframe');\n",
              "    iframe.src = new URL(path, url).toString();\n",
              "    iframe.height = height;\n",
              "    iframe.width = width;\n",
              "    iframe.style.border = 0;\n",
              "    iframe.allow = [\n",
              "        'accelerometer',\n",
              "        'autoplay',\n",
              "        'camera',\n",
              "        'clipboard-read',\n",
              "        'clipboard-write',\n",
              "        'gyroscope',\n",
              "        'magnetometer',\n",
              "        'microphone',\n",
              "        'serial',\n",
              "        'usb',\n",
              "        'xr-spatial-tracking',\n",
              "    ].join('; ');\n",
              "    element.appendChild(iframe);\n",
              "  })(8050, \"/\", \"100%\", 650, false, window.element)"
            ]
          },
          "metadata": {}
        }
      ]
    }
  ]
}