2_7_1_Generate_Anime_Characters_using_GANs_.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": [
        "# 🚀 **CIFAR-10 Image Classification with GAN**\n",
        "\n",
        "This project demonstrates the process of image classification using the CIFAR-10 dataset along with Generative Adversarial Networks (GAN) to generate synthetic images. The dataset consists of 60,000 32x32 color images in 10 different classes, and the goal is to perform a variety of tasks including image classification, visualization, and generation of synthetic images through GANs.\n",
        "\n",
        "## 📚 **Project Overview**\n",
        "\n",
        "This project involves:\n",
        "1. **Loading and Preprocessing the CIFAR-10 dataset.**\n",
        "2. **Exploratory Data Analysis (EDA) and Visualization.**\n",
        "3. **Building and Training a Generative Adversarial Network (GAN):**\n",
        "   - **Generator Model**: Creates synthetic images.\n",
        "   - **Discriminator Model**: Distinguishes between real and fake images.\n",
        "\n",
        "---\n",
        "\n",
        "## 💻 **Key Steps in the Project**\n",
        "\n",
        "### 1. **Data Loading and Preprocessing**\n",
        "- The CIFAR-10 dataset is loaded and split into training and test sets.\n",
        "- Images are flattened into 1D arrays for easier analysis.\n",
        "- Labels are processed into DataFrames for better clarity and analysis.\n",
        "- Image pixel values are normalized to a range of [-1, 1] for compatibility with GANs.\n",
        "\n",
        "### 2. **Exploratory Data Analysis (EDA)**\n",
        "- The dataset is analyzed for missing values and statistical summary.\n",
        "- The distribution of the 10 classes in the dataset is displayed.\n",
        "- Sample images from each class are visualized to understand the image types and patterns.\n",
        "\n",
        "### 3. **Building the Generative Adversarial Network (GAN)**\n",
        "\n",
        "#### Generator Model:\n",
        "- A Sequential model with dense, reshape, batch normalization, and convolutional layers.\n",
        "- Uses `tanh` activation for generating images with pixel values between -1 and 1.\n",
        "\n",
        "#### Discriminator Model:\n",
        "- A Sequential model consisting of convolutional layers, LeakyReLU activations, dropout for regularization, and a dense layer with sigmoid activation for binary classification (real or fake).\n",
        "\n",
        "---\n",
        "\n",
        "## 🧑‍🏫 **Learning Objectives**\n",
        "- Gain an understanding of how GANs function and how to train them for generating images.\n",
        "- Learn how to process and manipulate image datasets for deep learning applications.\n",
        "- Build and evaluate both generator and discriminator models in a GAN setup.\n",
        "- Apply image normalization techniques for better model performance.\n",
        "\n",
        "---\n",
        "\n",
        "## 🔧 **Tools & Libraries Used**\n",
        "- **TensorFlow**: For building and training the deep learning models.\n",
        "- **Keras**: Used for creating the Sequential models.\n",
        "- **NumPy & Pandas**: For data manipulation and analysis.\n",
        "- **Matplotlib**: For visualizing data and displaying sample images.\n",
        "\n",
        "---\n",
        "\n",
        "## 📊 **Results and Future Directions**\n",
        "- After training the GAN, the generator model produces synthetic images that resemble those from the CIFAR-10 dataset.\n",
        "- Future improvements can include:\n",
        "  - Fine-tuning the GAN models for higher-quality images.\n",
        "  - Extending the project to use other types of neural networks for classification and image generation.\n",
        "  - Exploring advanced GAN architectures like Conditional GANs (cGANs) or DCGANs for improved results.\n",
        "\n",
        "---\n",
        "\n",
        "## 📅 **Conclusion**\n",
        "This project serves as an introductory exploration of GANs and image classification with the CIFAR-10 dataset. It provides hands-on experience with building neural networks and working with deep learning frameworks for image data.\n",
        "\n",
        "---\n",
        "\n",
        "## 📄 **References**\n",
        "- [CIFAR-10 Dataset](https://www.cs.toronto.edu/~kriz/cifar.html)\n",
        "- [TensorFlow GAN Tutorial](https://www.tensorflow.org/tutorials/generative/dcgan)"
      ],
      "metadata": {
        "id": "3YNuWs7HSOI6"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "8qqKHpHUSWnh"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "cKuyxqpdOPsl"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "zjRFmIjpOQaC"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "---"
      ],
      "metadata": {
        "id": "JNgBuOqHOBC1"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Import libraries"
      ],
      "metadata": {
        "id": "AMdvY2TPN3mm"
      },
      "execution_count": 99,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import numpy as np  # Import numpy for numerical operations"
      ],
      "metadata": {
        "id": "SvqSaYN1Jewf"
      },
      "execution_count": 42,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import pandas as pd  # Import pandas for data manipulation"
      ],
      "metadata": {
        "id": "5En2RHJ0KZ4W"
      },
      "execution_count": 43,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import matplotlib.pyplot as plt  # Import matplotlib for plotting graphs"
      ],
      "metadata": {
        "id": "Dgr4fmvWKZ6-"
      },
      "execution_count": 44,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "import tensorflow as tf  # Import tensorflow for machine learning models"
      ],
      "metadata": {
        "id": "UpHDImL7KZ82"
      },
      "execution_count": 45,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.datasets import cifar10  # Import CIFAR-10 dataset from Keras"
      ],
      "metadata": {
        "id": "TmFJEM2aKjNj"
      },
      "execution_count": 46,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras.models import Sequential  # Import Sequential model for Keras neural networks"
      ],
      "metadata": {
        "id": "YUrflBxYKjRB"
      },
      "execution_count": 47,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from tensorflow.keras import layers  # Import layers for adding layers to neural networks"
      ],
      "metadata": {
        "id": "JLLFem4bKjTq"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "---"
      ],
      "metadata": {
        "id": "lZX47JX7Ku5C"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "LHV0xHC2Kv4i"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Ixy3kBYOKv7a"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "n1VWyaIcKv-D"
      },
      "execution_count": 48,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "---"
      ],
      "metadata": {
        "id": "5Yq8IaM8Kwne"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Load CIFAR-10 dataset"
      ],
      "metadata": {
        "id": "VSrkotiHKxxO"
      },
      "execution_count": 49,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "(x_train, y_train), (x_test, y_test) = cifar10.load_data()  # Load training and test data"
      ],
      "metadata": {
        "id": "IMDErkbKKjWj"
      },
      "execution_count": 50,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "2fEbVNY3KjY_"
      },
      "execution_count": 50,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Vpriuiv_K22u"
      },
      "execution_count": 50,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Print dataset shapes"
      ],
      "metadata": {
        "id": "9995hVv7K26_"
      },
      "execution_count": 51,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(f\"x_train shape: {x_train.shape}\")  # Print the shape of the training images"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DijrwNxXK2-B",
        "outputId": "867691d9-a407-49c7-dc98-cb9fcaedd1ed"
      },
      "execution_count": 52,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "x_train shape: (50000, 32, 32, 3)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "mQwwoe5QK9IH"
      },
      "execution_count": 52,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(f\"y_train shape: {y_train.shape}\")  # Print the shape of the training labels"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ai_zjXyhK9LW",
        "outputId": "94ff4917-ebae-4127-8916-f082dcedf7ba"
      },
      "execution_count": 53,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "y_train shape: (50000, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "QkwnzswzK9Ot"
      },
      "execution_count": 53,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(f\"x_test shape: {x_test.shape}\")  # Print the shape of the test images"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BfG16Cv8K9Ra",
        "outputId": "382ec58a-46bb-4ada-a36c-15a44cbc0020"
      },
      "execution_count": 54,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "x_test shape: (10000, 32, 32, 3)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "a8UeHSQYK9Tu"
      },
      "execution_count": 54,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(f\"y_test shape: {y_test.shape}\")  # Print the shape of the test labels"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "QPRKkkwLK9WQ",
        "outputId": "f2c65af1-2bc1-49cd-f112-a25a55ffc960"
      },
      "execution_count": 55,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "y_test shape: (10000, 1)\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "leXXxxW5LFaP"
      },
      "execution_count": 55,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "z1W8AcfQLFfp"
      },
      "execution_count": 55,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Convert to DataFrame for easy analysis (flattening images)"
      ],
      "metadata": {
        "id": "db1pvSV1LFlg"
      },
      "execution_count": 56,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "x_train_flat = x_train.reshape(x_train.shape[0], -1)  # Flatten training images into 1D arrays"
      ],
      "metadata": {
        "id": "uMRg26smLFqB"
      },
      "execution_count": 57,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "x_test_flat = x_test.reshape(x_test.shape[0], -1)  # Flatten test images into 1D arrays"
      ],
      "metadata": {
        "id": "tYcMo13DLFs6"
      },
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "tPm92vJ6LL-3"
      },
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "EzHy-gIILMET"
      },
      "execution_count": 58,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Create a DataFrame for the labels (y_train, y_test)"
      ],
      "metadata": {
        "id": "pXpPNd9gLMNv"
      },
      "execution_count": 59,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "y_train_df = pd.DataFrame(y_train, columns=[\"label\"])  # Convert training labels to DataFrame"
      ],
      "metadata": {
        "id": "s8CTV1AbLMQu"
      },
      "execution_count": 60,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "y_test_df = pd.DataFrame(y_test, columns=[\"label\"])  # Convert test labels to DataFrame"
      ],
      "metadata": {
        "id": "xljDs8DZLMTh"
      },
      "execution_count": 61,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "CiInC7BxLUQH"
      },
      "execution_count": 61,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "nCQKE1acLUTJ"
      },
      "execution_count": 61,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Show the first few records in the dataset"
      ],
      "metadata": {
        "id": "m-nZONo0LUV5"
      },
      "execution_count": 62,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nFirst 5 records of x_train flattened:\")\n",
        "print(x_train_flat[:5])  # Display first 5 flattened training images"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "1oGUkf0iLUYq",
        "outputId": "39895a01-d2e0-43be-cbac-440e21150fd9"
      },
      "execution_count": 63,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "First 5 records of x_train flattened:\n",
            "[[ 59  62  63 ... 123  92  72]\n",
            " [154 177 187 ... 143 133 144]\n",
            " [255 255 255 ...  80  86  84]\n",
            " [ 28  25  10 ...  72  65  46]\n",
            " [170 180 198 ...  73  77  80]]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "YZr5GWYgLMVJ"
      },
      "execution_count": 63,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "SPTLLRyvLFvg"
      },
      "execution_count": 63,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nFirst 5 records of y_train labels:\")\n",
        "print(y_train_df.head())  # Display first 5 labels from the training set"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "a-WXQ-GILF5_",
        "outputId": "f48fdd44-46f7-43f7-8bc5-7097557436a9"
      },
      "execution_count": 64,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "First 5 records of y_train labels:\n",
            "   label\n",
            "0      6\n",
            "1      9\n",
            "2      9\n",
            "3      4\n",
            "4      1\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "EsyN5_2yLbVv"
      },
      "execution_count": 64,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "qNfGXUXHLbbB"
      },
      "execution_count": 64,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Show dataset info (memory usage, number of non-null entries)"
      ],
      "metadata": {
        "id": "rUDQ_9OuLbeR"
      },
      "execution_count": 65,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "x_train_df = pd.DataFrame(x_train_flat)  # Convert flattened images to DataFrame for summary"
      ],
      "metadata": {
        "id": "QidM1yB3K9ge"
      },
      "execution_count": 66,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nSummary of x_train flattened:\")\n",
        "print(x_train_df.info())  # Display info about the training images (e.g., memory usage, non-null count)\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9LyMHclUKjb8",
        "outputId": "1482b821-cc21-4422-fe84-6986b40b726b"
      },
      "execution_count": 67,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Summary of x_train flattened:\n",
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 50000 entries, 0 to 49999\n",
            "Columns: 3072 entries, 0 to 3071\n",
            "dtypes: uint8(3072)\n",
            "memory usage: 146.5 MB\n",
            "None\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "nrpwZDSQKZ_T"
      },
      "execution_count": 67,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "JH_RCltCKaBs"
      },
      "execution_count": 67,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nSummary of y_train labels:\")\n",
        "print(y_train_df.info())  # Display info about the training labels"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tO39ruqSKaGP",
        "outputId": "92c40ba8-4e76-4e0f-a157-190a67685a3e"
      },
      "execution_count": 68,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Summary of y_train labels:\n",
            "<class 'pandas.core.frame.DataFrame'>\n",
            "RangeIndex: 50000 entries, 0 to 49999\n",
            "Data columns (total 1 columns):\n",
            " #   Column  Non-Null Count  Dtype\n",
            "---  ------  --------------  -----\n",
            " 0   label   50000 non-null  uint8\n",
            "dtypes: uint8(1)\n",
            "memory usage: 49.0 KB\n",
            "None\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "K9nyzUKOKaM9"
      },
      "execution_count": 68,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "eZUB38eIKadW"
      },
      "execution_count": 68,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Show basic statistics (mean, min, max, std)"
      ],
      "metadata": {
        "id": "QBVY189KKagJ"
      },
      "execution_count": 69,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nSummary statistics of x_train flattened:\")\n",
        "print(x_train_df.describe())  # Display descriptive statistics (mean, std, min, max) for training images\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "HQKl9tBXKaiX",
        "outputId": "1c4eb185-f144-4841-c1d5-f591f3873f3b"
      },
      "execution_count": 70,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Summary statistics of x_train flattened:\n",
            "               0            1             2            3             4     \\\n",
            "count  50000.000000  50000.00000  50000.000000  50000.00000  50000.000000   \n",
            "mean     130.710740    136.05614    132.553800    130.14036    135.442380   \n",
            "std       73.412873     72.90798     80.449751     72.44259     71.901316   \n",
            "min        0.000000      0.00000      0.000000      0.00000      0.000000   \n",
            "25%       71.000000     77.00000     61.000000     71.00000     78.000000   \n",
            "50%      128.000000    135.00000    127.000000    127.00000    135.000000   \n",
            "75%      189.000000    195.00000    207.000000    188.00000    193.000000   \n",
            "max      255.000000    255.00000    255.000000    255.00000    255.000000   \n",
            "\n",
            "               5             6            7             8             9     \\\n",
            "count  50000.000000  50000.000000  50000.00000  50000.000000  50000.000000   \n",
            "mean     131.853580    131.050440    136.24616    132.581440    131.568860   \n",
            "std       79.598048     72.240546     71.67471     79.434675     72.016555   \n",
            "min        0.000000      0.000000      0.00000      0.000000      0.000000   \n",
            "25%       61.000000     73.000000     79.00000     62.000000     73.000000   \n",
            "50%      127.000000    129.000000    136.00000    128.000000    130.000000   \n",
            "75%      206.000000    188.000000    194.00000    206.000000    188.000000   \n",
            "max      255.000000    255.000000    255.00000    255.000000    255.000000   \n",
            "\n",
            "       ...          3062          3063          3064          3065  \\\n",
            "count  ...  50000.000000  50000.000000  50000.000000  50000.000000   \n",
            "mean   ...    113.877800    126.778360    125.697080    113.830580   \n",
            "std    ...     64.738943     63.954763     61.591233     64.894603   \n",
            "min    ...      0.000000      0.000000      0.000000      0.000000   \n",
            "25%    ...     64.000000     79.000000     81.000000     64.000000   \n",
            "50%    ...    106.000000    124.000000    123.000000    106.000000   \n",
            "75%    ...    157.000000    172.000000    167.000000    157.000000   \n",
            "max    ...    255.000000    255.000000    255.000000    255.000000   \n",
            "\n",
            "               3066          3067          3068          3069          3070  \\\n",
            "count  50000.000000  50000.000000  50000.000000  50000.000000  50000.000000   \n",
            "mean     126.564960    125.617420    113.906240    126.639080    125.850300   \n",
            "std       64.178427     61.831646     65.212671     64.926283     62.649286   \n",
            "min        0.000000      0.000000      0.000000      0.000000      0.000000   \n",
            "25%       79.000000     81.000000     64.000000     78.000000     80.000000   \n",
            "50%      124.000000    123.000000    106.000000    124.000000    123.000000   \n",
            "75%      172.000000    167.000000    157.000000    172.000000    168.000000   \n",
            "max      255.000000    255.000000    255.000000    255.000000    255.000000   \n",
            "\n",
            "               3071  \n",
            "count  50000.000000  \n",
            "mean     114.381860  \n",
            "std       66.077526  \n",
            "min        0.000000  \n",
            "25%       63.000000  \n",
            "50%      106.000000  \n",
            "75%      158.000000  \n",
            "max      255.000000  \n",
            "\n",
            "[8 rows x 3072 columns]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "yviJ9xqCLr03"
      },
      "execution_count": 70,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Sb0YXl7zLr93"
      },
      "execution_count": 70,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nSummary statistics of y_train labels:\")\n",
        "print(y_train_df.describe())  # Display descriptive statistics for training labels"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uHRiArKTLsEC",
        "outputId": "fdad7744-f923-4061-8cdb-3484dc9c24f4"
      },
      "execution_count": 71,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Summary statistics of y_train labels:\n",
            "             label\n",
            "count  50000.00000\n",
            "mean       4.50000\n",
            "std        2.87231\n",
            "min        0.00000\n",
            "25%        2.00000\n",
            "50%        4.50000\n",
            "75%        7.00000\n",
            "max        9.00000\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "h22lts8LLsHb"
      },
      "execution_count": 71,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "NZ8VNKXpLsLN"
      },
      "execution_count": 71,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Check for missing values in x_train and y_train"
      ],
      "metadata": {
        "id": "j46z_XzWLsPz"
      },
      "execution_count": 72,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "missing_train_images = x_train_df.isnull().sum().sum()  # Check for missing values in training images\n"
      ],
      "metadata": {
        "id": "dN8W6XxJLsRa"
      },
      "execution_count": 73,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "missing_train_labels = y_train_df.isnull().sum().sum()  # Check for missing values in training labels\n"
      ],
      "metadata": {
        "id": "IXx-yc9yLsWP"
      },
      "execution_count": 74,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(f\"\\nMissing values in x_train: {missing_train_images}\")  # Print the number of missing values in images\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XDHZK3CXKakJ",
        "outputId": "5285abc4-f57c-47d7-c90f-ea08b48826bc"
      },
      "execution_count": 75,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Missing values in x_train: 0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "MIpyk7XPKama"
      },
      "execution_count": 75,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(f\"Missing values in y_train: {missing_train_labels}\")  # Print the number of missing values in labels\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dune1_WPJeze",
        "outputId": "b9e52c70-c4f3-4d2b-efd1-836729bed747"
      },
      "execution_count": 76,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Missing values in y_train: 0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "9D12oYRMJe2C"
      },
      "execution_count": 76,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "3W14XFbYJe4v"
      },
      "execution_count": 76,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Check class distribution"
      ],
      "metadata": {
        "id": "_MYj5RIDJe7b"
      },
      "execution_count": 77,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "unique_classes, counts = np.unique(y_train, return_counts=True)  # Get unique class labels and their counts\n"
      ],
      "metadata": {
        "id": "bmtVac56MGwf"
      },
      "execution_count": 78,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nClass labels:\", unique_classes)  # Print unique class labels"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "TqMnCGiqMGzb",
        "outputId": "88621e4a-7d82-47e6-e8fc-04c329a2a622"
      },
      "execution_count": 79,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Class labels: [0 1 2 3 4 5 6 7 8 9]\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"Class distribution:\", dict(zip(unique_classes, counts)))  # Print the distribution of classes in the training set\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "2c6G9iwaMG6e",
        "outputId": "01ceb9e9-1beb-4fb5-df5c-cd75b895e9ee"
      },
      "execution_count": 80,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Class distribution: {0: 5000, 1: 5000, 2: 5000, 3: 5000, 4: 5000, 5: 5000, 6: 5000, 7: 5000, 8: 5000, 9: 5000}\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "11P2vDrtMHCI"
      },
      "execution_count": 80,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "RYqOybK1MHFh"
      },
      "execution_count": 80,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Display some sample images from each class"
      ],
      "metadata": {
        "id": "1kdeV_7yMHJE"
      },
      "execution_count": 81,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "class_names = ['airplane', 'automobile', 'bird', 'cat', 'deer',  # List of class names\n",
        "               'dog', 'frog', 'horse', 'ship', 'truck']  # List of class names\n"
      ],
      "metadata": {
        "id": "ACnDrqVPMHMi"
      },
      "execution_count": 82,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "OW2BOpmsMSgi"
      },
      "execution_count": 82,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def plot_sample_images(x, y, class_names):  # Function to plot sample images from each class\n",
        "    plt.figure(figsize=(12, 12))  # Create a figure with a size of 12x12 inches\n",
        "    for i in range(10):  # Loop over 10 classes\n",
        "        idx = np.where(y == i)[0][0]  # Get the index of the first image in the class\n",
        "        plt.subplot(5, 5, i+1)  # Create a subplot for each image\n",
        "        plt.imshow(x[idx])  # Display the image at the given index\n",
        "        plt.title(class_names[i])  # Set the title of the subplot to the class name\n",
        "        plt.axis('off')  # Hide axes for cleaner visualization\n",
        "    plt.show()  # Display the plot"
      ],
      "metadata": {
        "id": "iqz-imTBMSjf"
      },
      "execution_count": 83,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "RZNhkXMQMYlo"
      },
      "execution_count": 83,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nSample images from each class:\")\n",
        "plot_sample_images(x_train, y_train, class_names)  # Plot sample images from the training set"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 392
        },
        "id": "ZCcWMUYOMYpC",
        "outputId": "dd05c0db-f747-4d5e-b5ce-df39aeaf4a97"
      },
      "execution_count": 84,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Sample images from each class:\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1200x1200 with 10 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "FC_zGNgmMYr8"
      },
      "execution_count": 84,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "EjQ2CiubMYxp"
      },
      "execution_count": 84,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Summary statistics of pixel values"
      ],
      "metadata": {
        "id": "yiqRDOH4MY07"
      },
      "execution_count": 85,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "print(\"\\nSummary statistics of pixel values:\")\n",
        "print(f\"Data type: {x_train.dtype}\")  # Print the data type of the training images\n",
        "print(f\"Max pixel value in training set: {x_train.max()}\")  # Print the maximum pixel value in the training images\n",
        "print(f\"Min pixel value in training set: {x_train.min()}\")  # Print the minimum pixel value in the training images\n"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "3ktvadWYMY4j",
        "outputId": "33c91036-7487-42fd-d2ac-0f7874d26fe8"
      },
      "execution_count": 86,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "Summary statistics of pixel values:\n",
            "Data type: uint8\n",
            "Max pixel value in training set: 255\n",
            "Min pixel value in training set: 0\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "-QpdMRuxMqU6"
      },
      "execution_count": 86,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "PS_wDklGMqYP"
      },
      "execution_count": 86,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Normalize the images to [-1, 1] for GAN"
      ],
      "metadata": {
        "id": "Zl4YiCDSMsRS"
      },
      "execution_count": 87,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "x_train = x_train / 127.5 - 1  # Normalize the pixel values of training images to the range [-1, 1]"
      ],
      "metadata": {
        "id": "3pClGpiNMqdl"
      },
      "execution_count": 88,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "x_test = x_test / 127.5 - 1  # Normalize the pixel values of test images to the range [-1, 1]"
      ],
      "metadata": {
        "id": "Mu2_skXhMZAC"
      },
      "execution_count": 89,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "---"
      ],
      "metadata": {
        "id": "32WNoXKVPjO0"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "mAHDoMBnMSlR"
      },
      "execution_count": 89,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "6x1OuUiKMSqe"
      },
      "execution_count": 89,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "cikljOoGPoEk"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "---"
      ],
      "metadata": {
        "id": "8NlrECnBPofp"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Build Generator Model"
      ],
      "metadata": {
        "id": "uT222D5wMSs0"
      },
      "execution_count": 90,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def build_generator():  # Function to build the generator model for the GAN\n",
        "    model = Sequential()  # Initialize a Sequential model\n",
        "    model.add(layers.Dense(8 * 8 * 256, activation=\"relu\", input_shape=(100,)))  # Add a dense layer with ReLU activation\n",
        "    model.add(layers.Reshape((8, 8, 256)))  # Reshape the output into a 3D tensor\n",
        "    model.add(layers.BatchNormalization())  # Add batch normalization for stabilizing training\n",
        "    model.add(layers.UpSampling2D())  # Upsample the image size\n",
        "    model.add(layers.Conv2D(128, kernel_size=3, padding=\"same\", activation=\"relu\"))  # Add convolutional layer\n",
        "    model.add(layers.BatchNormalization())  # Add batch normalization\n",
        "    model.add(layers.UpSampling2D())  # Upsample the image size again\n",
        "    model.add(layers.Conv2D(64, kernel_size=3, padding=\"same\", activation=\"relu\"))  # Add another convolutional layer\n",
        "    model.add(layers.BatchNormalization())  # Add batch normalization\n",
        "    model.add(layers.Conv2D(3, kernel_size=3, padding=\"same\", activation=\"tanh\"))  # Add final convolutional layer with tanh activation\n",
        "    return model  # Return the constructed generator model"
      ],
      "metadata": {
        "id": "90ZCeYWJMSxh"
      },
      "execution_count": 91,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "HtGBVc0WM20o"
      },
      "execution_count": 91,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "generator = build_generator()  # Build the generator model"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "kpzDzMPUM23s",
        "outputId": "fea5f655-7cdd-431c-8408-9e2b0a334984"
      },
      "execution_count": 92,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.10/dist-packages/keras/src/layers/core/dense.py:87: 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"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "generator.summary()  # Print the summary of the generator model"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 508
        },
        "id": "6rBx0dxpM29e",
        "outputId": "1221fbf1-0850-436c-fd66-472a83bdca81"
      },
      "execution_count": 93,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1mModel: \"sequential_48\"\u001b[0m\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"sequential_48\"</span>\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓\n",
              "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                        \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape               \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m        Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩\n",
              "│ dense_38 (\u001b[38;5;33mDense\u001b[0m)                     │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m16384\u001b[0m)               │       \u001b[38;5;34m1,654,784\u001b[0m │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ reshape_21 (\u001b[38;5;33mReshape\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m8\u001b[0m, \u001b[38;5;34m8\u001b[0m, \u001b[38;5;34m256\u001b[0m)           │               \u001b[38;5;34m0\u001b[0m │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ batch_normalization_39               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m8\u001b[0m, \u001b[38;5;34m8\u001b[0m, \u001b[38;5;34m256\u001b[0m)           │           \u001b[38;5;34m1,024\u001b[0m │\n",
              "│ (\u001b[38;5;33mBatchNormalization\u001b[0m)                 │                             │                 │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ up_sampling2d_27 (\u001b[38;5;33mUpSampling2D\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m16\u001b[0m, \u001b[38;5;34m16\u001b[0m, \u001b[38;5;34m256\u001b[0m)         │               \u001b[38;5;34m0\u001b[0m │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ conv2d_83 (\u001b[38;5;33mConv2D\u001b[0m)                   │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m16\u001b[0m, \u001b[38;5;34m16\u001b[0m, \u001b[38;5;34m128\u001b[0m)         │         \u001b[38;5;34m295,040\u001b[0m │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ batch_normalization_40               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m16\u001b[0m, \u001b[38;5;34m16\u001b[0m, \u001b[38;5;34m128\u001b[0m)         │             \u001b[38;5;34m512\u001b[0m │\n",
              "│ (\u001b[38;5;33mBatchNormalization\u001b[0m)                 │                             │                 │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ up_sampling2d_28 (\u001b[38;5;33mUpSampling2D\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m128\u001b[0m)         │               \u001b[38;5;34m0\u001b[0m │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ conv2d_84 (\u001b[38;5;33mConv2D\u001b[0m)                   │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m64\u001b[0m)          │          \u001b[38;5;34m73,792\u001b[0m │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ batch_normalization_41               │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m64\u001b[0m)          │             \u001b[38;5;34m256\u001b[0m │\n",
              "│ (\u001b[38;5;33mBatchNormalization\u001b[0m)                 │                             │                 │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ conv2d_85 (\u001b[38;5;33mConv2D\u001b[0m)                   │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m3\u001b[0m)           │           \u001b[38;5;34m1,731\u001b[0m │\n",
              "└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓\n",
              "┃<span style=\"font-weight: bold\"> Layer (type)                         </span>┃<span style=\"font-weight: bold\"> Output Shape                </span>┃<span style=\"font-weight: bold\">         Param # </span>┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩\n",
              "│ dense_38 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                     │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16384</span>)               │       <span style=\"color: #00af00; text-decoration-color: #00af00\">1,654,784</span> │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ reshape_21 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Reshape</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">8</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">8</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)           │               <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ batch_normalization_39               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">8</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">8</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)           │           <span style=\"color: #00af00; text-decoration-color: #00af00\">1,024</span> │\n",
              "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">BatchNormalization</span>)                 │                             │                 │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ up_sampling2d_27 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">UpSampling2D</span>)      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span>)         │               <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ conv2d_83 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)                   │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">128</span>)         │         <span style=\"color: #00af00; text-decoration-color: #00af00\">295,040</span> │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ batch_normalization_40               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">128</span>)         │             <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span> │\n",
              "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">BatchNormalization</span>)                 │                             │                 │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ up_sampling2d_28 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">UpSampling2D</span>)      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">128</span>)         │               <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ conv2d_84 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)                   │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">64</span>)          │          <span style=\"color: #00af00; text-decoration-color: #00af00\">73,792</span> │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ batch_normalization_41               │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">64</span>)          │             <span style=\"color: #00af00; text-decoration-color: #00af00\">256</span> │\n",
              "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">BatchNormalization</span>)                 │                             │                 │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ conv2d_85 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)                   │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">3</span>)           │           <span style=\"color: #00af00; text-decoration-color: #00af00\">1,731</span> │\n",
              "└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m2,027,139\u001b[0m (7.73 MB)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">2,027,139</span> (7.73 MB)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m2,026,243\u001b[0m (7.73 MB)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">2,026,243</span> (7.73 MB)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m896\u001b[0m (3.50 KB)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">896</span> (3.50 KB)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "---"
      ],
      "metadata": {
        "id": "Yb80lFJJPwsD"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "KDT8HjxFM3AX"
      },
      "execution_count": 93,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "qCHENtcHM3Gd"
      },
      "execution_count": 93,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "MvHZO390PvuC"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "---"
      ],
      "metadata": {
        "id": "BhK9PGkDPyNi"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Build Discriminator Model"
      ],
      "metadata": {
        "id": "385f4BHLM3R4"
      },
      "execution_count": 94,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def build_discriminator():  # Function to build the discriminator model for the GAN\n",
        "    model = Sequential()  # Initialize a Sequential model\n",
        "    model.add(layers.Conv2D(64, kernel_size=3, strides=2, padding=\"same\", input_shape=(64, 64, 3)))  # Add convolutional layer\n",
        "    model.add(layers.LeakyReLU(alpha=0.2))  # Add LeakyReLU activation\n",
        "    model.add(layers.Dropout(0.3))  # Add dropout for regularization\n",
        "    model.add(layers.Conv2D(128, kernel_size=3, strides=2, padding=\"same\"))  # Add another convolutional layer\n",
        "    model.add(layers.LeakyReLU(alpha=0.2))  # Add another LeakyReLU activation\n",
        "    model.add(layers.Dropout(0.3))  # Add dropout for regularization\n",
        "    model.add(layers.Flatten())  # Flatten the output for the fully connected layer\n",
        "    model.add(layers.Dense(1, activation=\"sigmoid\"))  # Add a dense layer with sigmoid activation for binary classification\n",
        "    return model  # Return the constructed discriminator model"
      ],
      "metadata": {
        "id": "Hcyknrp1MS9v"
      },
      "execution_count": 95,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "Hxx9yiE4NBtY"
      },
      "execution_count": 95,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "discriminator = build_discriminator()  # Build the discriminator model"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CKDH1pi9NBxB",
        "outputId": "b55f823a-a56f-485c-cee9-f210395e4afb"
      },
      "execution_count": 96,
      "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",
            "/usr/local/lib/python3.10/dist-packages/keras/src/layers/activations/leaky_relu.py:41: UserWarning: Argument `alpha` is deprecated. Use `negative_slope` instead.\n",
            "  warnings.warn(\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "discriminator.summary()  # Print the summary of the discriminator model"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 393
        },
        "id": "6amCmgDYNB0M",
        "outputId": "1c2a6d41-7d2a-4a3c-e014-6bd94eec095d"
      },
      "execution_count": 97,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1mModel: \"sequential_49\"\u001b[0m\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\">Model: \"sequential_49\"</span>\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓\n",
              "┃\u001b[1m \u001b[0m\u001b[1mLayer (type)                        \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1mOutput Shape               \u001b[0m\u001b[1m \u001b[0m┃\u001b[1m \u001b[0m\u001b[1m        Param #\u001b[0m\u001b[1m \u001b[0m┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩\n",
              "│ conv2d_86 (\u001b[38;5;33mConv2D\u001b[0m)                   │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m64\u001b[0m)          │           \u001b[38;5;34m1,792\u001b[0m │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ leaky_re_lu_91 (\u001b[38;5;33mLeakyReLU\u001b[0m)           │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m64\u001b[0m)          │               \u001b[38;5;34m0\u001b[0m │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ dropout_39 (\u001b[38;5;33mDropout\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m32\u001b[0m, \u001b[38;5;34m64\u001b[0m)          │               \u001b[38;5;34m0\u001b[0m │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ conv2d_87 (\u001b[38;5;33mConv2D\u001b[0m)                   │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m16\u001b[0m, \u001b[38;5;34m16\u001b[0m, \u001b[38;5;34m128\u001b[0m)         │          \u001b[38;5;34m73,856\u001b[0m │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ leaky_re_lu_92 (\u001b[38;5;33mLeakyReLU\u001b[0m)           │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m16\u001b[0m, \u001b[38;5;34m16\u001b[0m, \u001b[38;5;34m128\u001b[0m)         │               \u001b[38;5;34m0\u001b[0m │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ dropout_40 (\u001b[38;5;33mDropout\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m16\u001b[0m, \u001b[38;5;34m16\u001b[0m, \u001b[38;5;34m128\u001b[0m)         │               \u001b[38;5;34m0\u001b[0m │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ flatten_16 (\u001b[38;5;33mFlatten\u001b[0m)                 │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m32768\u001b[0m)               │               \u001b[38;5;34m0\u001b[0m │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ dense_39 (\u001b[38;5;33mDense\u001b[0m)                     │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)                   │          \u001b[38;5;34m32,769\u001b[0m │\n",
              "└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\">┏━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┳━━━━━━━━━━━━━━━━━┓\n",
              "┃<span style=\"font-weight: bold\"> Layer (type)                         </span>┃<span style=\"font-weight: bold\"> Output Shape                </span>┃<span style=\"font-weight: bold\">         Param # </span>┃\n",
              "┡━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━━━━━━━━━━━━━╇━━━━━━━━━━━━━━━━━┩\n",
              "│ conv2d_86 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)                   │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">64</span>)          │           <span style=\"color: #00af00; text-decoration-color: #00af00\">1,792</span> │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ leaky_re_lu_91 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">LeakyReLU</span>)           │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">64</span>)          │               <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ dropout_39 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">64</span>)          │               <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ conv2d_87 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Conv2D</span>)                   │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">128</span>)         │          <span style=\"color: #00af00; text-decoration-color: #00af00\">73,856</span> │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ leaky_re_lu_92 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">LeakyReLU</span>)           │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">128</span>)         │               <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ dropout_40 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dropout</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">16</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">128</span>)         │               <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ flatten_16 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Flatten</span>)                 │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">32768</span>)               │               <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
              "├──────────────────────────────────────┼─────────────────────────────┼─────────────────┤\n",
              "│ dense_39 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Dense</span>)                     │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)                   │          <span style=\"color: #00af00; text-decoration-color: #00af00\">32,769</span> │\n",
              "└──────────────────────────────────────┴─────────────────────────────┴─────────────────┘\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m108,417\u001b[0m (423.50 KB)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Total params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">108,417</span> (423.50 KB)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m108,417\u001b[0m (423.50 KB)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">108,417</span> (423.50 KB)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
            ],
            "text/html": [
              "<pre style=\"white-space:pre;overflow-x:auto;line-height:normal;font-family:Menlo,'DejaVu Sans Mono',consolas,'Courier New',monospace\"><span style=\"font-weight: bold\"> Non-trainable params: </span><span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> (0.00 B)\n",
              "</pre>\n"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "---"
      ],
      "metadata": {
        "id": "6De0B4pdNw6g"
      }
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "UuXdfoPGNB5i"
      },
      "execution_count": 97,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [],
      "metadata": {
        "id": "9xDfw10lNB81"
      },
      "execution_count": 104,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "### 📅 **Conclusion**\n",
        "\n",
        "In this project, we successfully developed and trained a Generative Adversarial Network (GAN) to generate images resembling the CIFAR-10 dataset. The GAN consists of two primary components: the **generator** and the **discriminator**.\n",
        "\n",
        "- The **generator** is a neural network that learns to create synthetic images from random noise. Initially, the images are random, but through the adversarial training process, the generator improves its output to resemble real images from the CIFAR-10 dataset.\n",
        "  \n",
        "- The **discriminator** is a classifier that distinguishes between real images (from the dataset) and fake images (produced by the generator). It provides feedback to the generator to guide its learning.\n",
        "\n",
        "Through iterative training, both models improve: the generator becomes better at creating realistic images, and the discriminator becomes better at detecting fake ones. This **adversarial process** helps the generator produce increasingly convincing images that closely resemble the CIFAR-10 classes.\n",
        "\n",
        "The **training process** involves optimizing both models simultaneously, using the **binary cross-entropy loss** to measure how well the discriminator is distinguishing between real and fake images, and the generator’s success in fooling the discriminator. We also incorporated techniques such as **batch normalization** and **upsampling** to stabilize and improve the model’s learning process.\n",
        "\n",
        "Ultimately, the GAN framework demonstrated its potential for image generation tasks, and we observed the gradual improvement in the quality of generated images over time. The project highlights the power of GANs in producing high-fidelity images from random noise and emphasizes their practical applications in fields like computer vision, art generation, and synthetic data creation.\n",
        "\n",
        "This experiment also underlines the importance of both model architectures, the interplay between the generator and discriminator, and the overall adversarial learning approach in achieving realistic image generation. Future improvements can include experimenting with more complex architectures, increasing the training data, or fine-tuning the models further to generate higher-resolution images."
      ],
      "metadata": {
        "id": "DjNYU3meSapy"
      }
    }
  ]
}