03_M04L02_Lab_Develop GANs using Keras__M.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<p style=\"text-align:center\">\n",
    "    <a href=\"https://skills.network\" target=\"_blank\">\n",
    "    <img src=\"https://cf-courses-data.s3.us.cloud-object-storage.appdomain.cloud/assets/logos/SN_web_lightmode.png\" width=\"200\" alt=\"Skills Network Logo\"  />\n",
    "    </a>\n",
    "</p>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# **Lab: Develop GANs Using Keras**\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Estimated time needed: **30** minutes\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this lab, you will gain practical experience in developing generative adversarial networks (GANs) using Keras. You will learn to preprocess data, construct the generator and discriminator models, combine them to create the GAN, train the GAN, and evaluate its performance. GANs are powerful tools for generating synthetic data, and this lab provides a solid foundation for further exploration and experimentation with more advanced GAN architectures and applications.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Objectives\n",
    "\n",
    "After completing this lab, you will be able to: \n",
    "- Build, train, and evaluate GANs using Keras \n",
    "- Explore GAN architectures, data preprocessing, model training, and performance evaluation \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Step-by-step instructions \n",
    "\n",
    "### Step 1: Data preprocessing \n",
    "\n",
    "#### Objective: \n",
    "- Load and preprocess the MNIST dataset for training a GAN. \n",
    "\n",
    "#### Instructions: \n",
    "1. Load the MNIST dataset: \n",
    "   - Use Keras to load the MNIST dataset.\n",
    "   - Normalize the image pixel values to the range [-1, 1].\n",
    "\n",
    "2. Reshape the data:\n",
    "   - Expand the dimensions of the images to match the input shape required by the GAN (28×28×1).\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "%%capture\n",
    "!pip install tensorflow-cpu==2.16.2\n",
    "\n",
    "# Suppress warnings and set environment variables\n",
    "import os\n",
    "os.environ['TF_CPP_MIN_LOG_LEVEL'] = '2'\n",
    "os.environ['TF_ENABLE_ONEDNN_OPTS'] = '0'\n",
    "\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(60000, 28, 28, 1)\n"
     ]
    }
   ],
   "source": [
    "import numpy as np\n",
    "from tensorflow.keras.datasets import mnist\n",
    "import warnings\n",
    "\n",
    "# Suppress all Python warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "\n",
    "# Load the MNIST dataset\n",
    "(x_train, _), (_, _) = mnist.load_data()\n",
    "\n",
    "# Normalize the pixel values to the range [-1, 1]\n",
    "x_train = x_train.astype('float32') / 127.5 - 1.\n",
    "x_train = np.expand_dims(x_train, axis=-1)\n",
    "\n",
    "# Print the shape of the data\n",
    "print(x_train.shape)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Explanation: \n",
    "\n",
    "This step prepares the MNIST dataset for training by normalizing the pixel values to the range [-1, 1] and reshaping the images to have a single color channel. Normalization helps in faster convergence during training, and reshaping is required because the input layer of our GAN expects a three-dimensional tensor. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 2: Building the generator model \n",
    "\n",
    "#### Objective: \n",
    "- Construct the generator model for the GAN using the Keras functional API. \n",
    "\n",
    "#### Instructions: \n",
    "1. Define the generator. \n",
    "   - Create a Sequential model. \n",
    "   - Add Dense, LeakyReLU, BatchNormalization, and Reshape layers to build the generator.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "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_5\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"sequential_5\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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_18 (<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\">256</span>)            │        <span style=\"color: #00af00; text-decoration-color: #00af00\">25,856</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_13 (<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\">256</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ batch_normalization_7           │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</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",
       "│ dense_19 (<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\">512</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">131,584</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_14 (<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\">512</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ batch_normalization_8           │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>)            │         <span style=\"color: #00af00; text-decoration-color: #00af00\">2,048</span> │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">BatchNormalization</span>)            │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_20 (<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\">1024</span>)           │       <span style=\"color: #00af00; text-decoration-color: #00af00\">525,312</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_15 (<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\">1024</span>)           │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ batch_normalization_9           │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1024</span>)           │         <span style=\"color: #00af00; text-decoration-color: #00af00\">4,096</span> │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">BatchNormalization</span>)            │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_21 (<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\">784</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">803,600</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ reshape_2 (<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\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)      │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
       "</pre>\n"
      ],
      "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_18 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │        \u001b[38;5;34m25,856\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_13 (\u001b[38;5;33mLeakyReLU\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ batch_normalization_7           │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │         \u001b[38;5;34m1,024\u001b[0m │\n",
       "│ (\u001b[38;5;33mBatchNormalization\u001b[0m)            │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_19 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │       \u001b[38;5;34m131,584\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_14 (\u001b[38;5;33mLeakyReLU\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ batch_normalization_8           │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │         \u001b[38;5;34m2,048\u001b[0m │\n",
       "│ (\u001b[38;5;33mBatchNormalization\u001b[0m)            │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_20 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1024\u001b[0m)           │       \u001b[38;5;34m525,312\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_15 (\u001b[38;5;33mLeakyReLU\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1024\u001b[0m)           │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ batch_normalization_9           │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1024\u001b[0m)           │         \u001b[38;5;34m4,096\u001b[0m │\n",
       "│ (\u001b[38;5;33mBatchNormalization\u001b[0m)            │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_21 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m784\u001b[0m)            │       \u001b[38;5;34m803,600\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ reshape_2 (\u001b[38;5;33mReshape\u001b[0m)             │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m1\u001b[0m)      │             \u001b[38;5;34m0\u001b[0m │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">1,493,520</span> (5.70 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m1,493,520\u001b[0m (5.70 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">1,489,936</span> (5.68 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m1,489,936\u001b[0m (5.68 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">3,584</span> (14.00 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m3,584\u001b[0m (14.00 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from tensorflow.keras.models import Sequential \n",
    "from tensorflow.keras.layers import Dense, LeakyReLU, BatchNormalization, Reshape \n",
    "\n",
    "# Define the generator model \n",
    "def build_generator(): \n",
    "    model = Sequential() \n",
    "    model.add(Dense(256, input_dim=100)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(BatchNormalization(momentum=0.8)) \n",
    "    model.add(Dense(512)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(BatchNormalization(momentum=0.8)) \n",
    "    model.add(Dense(1024)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(BatchNormalization(momentum=0.8)) \n",
    "    model.add(Dense(28 * 28 * 1, activation='tanh')) \n",
    "    model.add(Reshape((28, 28, 1))) \n",
    "    return model \n",
    "\n",
    "# Build the generator \n",
    "generator = build_generator() \n",
    "generator.summary()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Explanation: \n",
    "\n",
    "This step involves building the generator model for the GAN. The generator takes a random noise vector as an input and generates a synthetic image. The model uses Dense, LeakyReLU, BatchNormalization, and Reshape layers to achieve this. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 3: Building the discriminator model \n",
    "\n",
    "#### Objective: \n",
    "- Construct the discriminator model for the GAN using the Keras functional API. \n",
    "\n",
    "#### Instructions: \n",
    "1. Define the discriminator. \n",
    "   - Create a Sequential model. \n",
    "   - Add Flatten, Dense, and LeakyReLU layers to build the discriminator. \n",
    "\n",
    "2. Compile the discriminator. \n",
    "   - Compile the model using binary cross-entropy loss and the Adam optimizer. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "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_6\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"sequential_6\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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",
       "│ flatten_3 (<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\">784</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_22 (<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\">512</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">401,920</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_16 (<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\">512</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_23 (<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\">256</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">131,328</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_17 (<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\">256</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_24 (<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\">257</span> │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
       "</pre>\n"
      ],
      "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",
       "│ flatten_3 (\u001b[38;5;33mFlatten\u001b[0m)             │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m784\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_22 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │       \u001b[38;5;34m401,920\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_16 (\u001b[38;5;33mLeakyReLU\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_23 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │       \u001b[38;5;34m131,328\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_17 (\u001b[38;5;33mLeakyReLU\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_24 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)              │           \u001b[38;5;34m257\u001b[0m │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">533,505</span> (2.04 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m533,505\u001b[0m (2.04 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">533,505</span> (2.04 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m533,505\u001b[0m (2.04 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from tensorflow.keras.layers import Flatten \n",
    "from tensorflow.keras.models import Sequential \n",
    "from tensorflow.keras.layers import Dense, LeakyReLU\n",
    "\n",
    "# Define the discriminator model \n",
    "def build_discriminator(): \n",
    "    model = Sequential() \n",
    "    model.add(Flatten(input_shape=(28, 28, 1))) \n",
    "    model.add(Dense(512)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(Dense(256)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(Dense(1, activation='sigmoid')) \n",
    "    return model \n",
    "\n",
    "# Build and compile the discriminator \n",
    "discriminator = build_discriminator() \n",
    "discriminator.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy']) \n",
    "discriminator.summary()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Explanation: \n",
    "\n",
    "This exercise involves building the discriminator model for the GAN. The discriminator takes an image as an input and outputs a probability indicating whether the image is real or fake. The model uses Flatten, Dense, and LeakyReLU layers to achieve this. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 4: Building the GAN Model \n",
    "\n",
    "#### Objective: \n",
    "- Combine the generator and discriminator to create the GAN model using the Keras functional API. \n",
    "\n",
    "#### Instructions: \n",
    "1. Define the GAN. \n",
    "   - Create an input layer for the noise vector. \n",
    "   - Pass the noise vector through the generator to produce a synthetic image. \n",
    "   - Pass the synthetic image through the discriminator to get the classification. \n",
    "   - Compile the GAN using binary cross-entropy loss and the Adam optimizer. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "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: \"functional_61\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"functional_61\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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",
       "│ input_layer_8 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">InputLayer</span>)      │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">100</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ sequential_5 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Sequential</span>)       │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)      │     <span style=\"color: #00af00; text-decoration-color: #00af00\">1,493,520</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ sequential_6 (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">Sequential</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\">533,505</span> │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
       "</pre>\n"
      ],
      "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",
       "│ input_layer_8 (\u001b[38;5;33mInputLayer\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m100\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ sequential_5 (\u001b[38;5;33mSequential\u001b[0m)       │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m1\u001b[0m)      │     \u001b[38;5;34m1,493,520\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ sequential_6 (\u001b[38;5;33mSequential\u001b[0m)       │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)              │       \u001b[38;5;34m533,505\u001b[0m │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">2,027,025</span> (7.73 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m2,027,025\u001b[0m (7.73 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">1,489,936</span> (5.68 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m1,489,936\u001b[0m (5.68 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">537,089</span> (2.05 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m537,089\u001b[0m (2.05 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from tensorflow.keras.layers import Input \n",
    "from tensorflow.keras.models import Model \n",
    "\n",
    "# Create the GAN by stacking the generator and the discriminator \n",
    "def build_gan(generator, discriminator): \n",
    "    discriminator.trainable = False \n",
    "    gan_input = Input(shape=(100,)) \n",
    "    generated_image = generator(gan_input) \n",
    "    gan_output = discriminator(generated_image) \n",
    "    gan = Model(gan_input, gan_output) \n",
    "    gan.compile(loss='binary_crossentropy', optimizer='adam') \n",
    "    return gan \n",
    "\n",
    "# Build the GAN \n",
    "gan = build_gan(generator, discriminator) \n",
    "gan.summary()\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Explanation: \n",
    "\n",
    "This step involves combining the generator and discriminator models to create the GAN. The GAN takes a noise vector as an input, generates a synthetic image using the generator, and classifies the image using the discriminator. The discriminator is set to non-trainable when compiling the GAN to ensure that only the generator is updated during the adversarial training. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 5: Training the GAN \n",
    "\n",
    "#### Objective: \n",
    "- Train the GAN on the MNIST dataset. \n",
    "\n",
    "#### Instructions: \n",
    "1. Define training parameters. \n",
    "   - Set the batch size, number of epochs, and sample interval. \n",
    " \n",
    "2. Train the discriminator. \n",
    "   - Sample a batch of real images from the dataset. \n",
    "   - Generate a batch of synthetic images from the generator. \n",
    "   - Train the discriminator on both real and generated images. \n",
    " \n",
    "3. Train the generator. \n",
    "   - Generate a batch of noise vectors. \n",
    "   - Train the GAN to improve the generator’s ability to fool the discriminator. \n",
    " \n",
    "4. Print the progress: \n",
    "   - Print the discriminator and generator losses at regular intervals.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "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_7\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"sequential_7\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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",
       "│ flatten_4 (<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\">784</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_25 (<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\">512</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">401,920</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_18 (<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\">512</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_26 (<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\">256</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">131,328</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_19 (<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\">256</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_27 (<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\">257</span> │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
       "</pre>\n"
      ],
      "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",
       "│ flatten_4 (\u001b[38;5;33mFlatten\u001b[0m)             │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m784\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_25 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │       \u001b[38;5;34m401,920\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_18 (\u001b[38;5;33mLeakyReLU\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_26 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │       \u001b[38;5;34m131,328\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_19 (\u001b[38;5;33mLeakyReLU\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_27 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)              │           \u001b[38;5;34m257\u001b[0m │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">533,505</span> (2.04 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m533,505\u001b[0m (2.04 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">533,505</span> (2.04 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m533,505\u001b[0m (2.04 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Define and compile the discriminator model\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, LeakyReLU, Flatten\n",
    "\n",
    "def build_discriminator():\n",
    "    model = Sequential()\n",
    "    model.add(Flatten(input_shape=(28, 28, 1)))\n",
    "    model.add(Dense(512))\n",
    "    model.add(LeakyReLU(alpha=0.2))\n",
    "    model.add(Dense(256))\n",
    "    model.add(LeakyReLU(alpha=0.2))\n",
    "    model.add(Dense(1, activation='sigmoid'))\n",
    "    return model\n",
    "\n",
    "# Build and recompile the discriminator\n",
    "discriminator = build_discriminator()\n",
    "discriminator.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\n",
    "discriminator.summary()\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step \n",
      "0 [D loss: 0.9583353996276855] [D accuracy: 12.5%] [G loss: 0.8986772298812866]\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 33ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 23ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 33ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 38ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 22ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 44ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 65ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 42ms/step\n",
      "10 [D loss: 0.19486072659492493] [D accuracy: 89.96549844741821%] [G loss: 0.3276514708995819]\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 34ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 35ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 28ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 34ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 35ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 32ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 26ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 34ms/step\n",
      "20 [D loss: 0.10262157768011093] [D accuracy: 94.80346441268921%] [G loss: 0.1873997449874878]\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 23ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 38ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 37ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 31ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 35ms/step\n",
      "30 [D loss: 0.06927143037319183] [D accuracy: 96.49367332458496%] [G loss: 0.13083170354366302]\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 36ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 23ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 34ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 28ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 31ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 31ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 34ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 23ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 36ms/step\n",
      "40 [D loss: 0.05227656662464142] [D accuracy: 97.35419750213623%] [G loss: 0.10077070444822311]\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 28ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 26ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 28ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 44ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 32ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 25ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 26ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 30ms/step\n"
     ]
    }
   ],
   "source": [
    "# Training parameters \n",
    "\n",
    "batch_size = 64 \n",
    "epochs = 50\n",
    "# epochs = 10\n",
    "sample_interval = 10\n",
    "\n",
    "# Adversarial ground truths \n",
    "real = np.ones((batch_size, 1)) \n",
    "fake = np.zeros((batch_size, 1)) \n",
    "\n",
    "# Training loop \n",
    "for epoch in range(epochs): \n",
    "    # Train the discriminator \n",
    "    idx = np.random.randint(0, x_train.shape[0], batch_size) \n",
    "    real_images = x_train[idx] \n",
    "    noise = np.random.normal(0, 1, (batch_size, 100)) \n",
    "    generated_images = generator.predict(noise) \n",
    "    d_loss_real = discriminator.train_on_batch(real_images, real) \n",
    "    d_loss_fake = discriminator.train_on_batch(generated_images, fake) \n",
    "    d_loss = 0.5 * np.add(d_loss_real, d_loss_fake) \n",
    "\n",
    "    # Train the generator \n",
    "    noise = np.random.normal(0, 1, (batch_size, 100)) \n",
    "    g_loss = gan.train_on_batch(noise, real) \n",
    "\n",
    "    # Print the progress \n",
    "    if epoch % sample_interval == 0: \n",
    "        print(f\"{epoch} [D loss: {d_loss[0]}] [D accuracy: {100 * d_loss[1]}%] [G loss: {g_loss}]\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Explanation: \n",
    "\n",
    "This step involves training the GAN on the MNIST dataset. The training loop alternates between training the discriminator and the generator. The discriminator is trained on batches of real and generated images, whereas the generator is trained to improve its ability to fool the discriminator. The progress is printed at regular intervals to monitor the training process. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 6: Assessing the Quality of Generated Images \n",
    "\n",
    "### Objective: \n",
    "- Evaluate the performance of the trained GAN. \n",
    "\n",
    "### Evaluating the GAN\n",
    "\n",
    "After training the GAN, we need to assess the quality of the synthetic images generated by the generator. There are two main ways to evaluate the performance of GANs: qualitative assessment and quantitative assessment.\n",
    "\n",
    "### Qualitative Assessment: Visual Inspection\n",
    "\n",
    "Visual inspection is a straightforward method to assess the quality of images generated by a GAN. You can use the `sample_images` function provided in the lab to visualize a grid of generated images. During visual inspection, look for the following qualities:\n",
    "\n",
    "- **Clarity**: The images should be sharp and not blurry. Blurry images indicate that the generator is struggling to learn the patterns in the data.\n",
    "- **Coherence**: The generated images should have a coherent structure that resembles the original images in the dataset. For example, in the case of MNIST, the generated images should resemble handwritten digits with the correct number of strokes and shapes.\n",
    "- **Diversity**: There should be a variety of images generated by the GAN. If all images look similar, it might indicate that the generator is overfitting or has collapsed to a single mode.\n",
    "\n",
    "#### Instructions:\n",
    "\n",
    "1. Run the `sample_images` function after training the GAN to display a grid of generated images.\n",
    "2. Inspect the images for clarity, coherence, and diversity.\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Requirement already satisfied: matplotlib in /opt/conda/lib/python3.11/site-packages (3.9.2)\n",
      "Requirement already satisfied: contourpy>=1.0.1 in /opt/conda/lib/python3.11/site-packages (from matplotlib) (1.3.1)\n",
      "Requirement already satisfied: cycler>=0.10 in /opt/conda/lib/python3.11/site-packages (from matplotlib) (0.12.1)\n",
      "Requirement already satisfied: fonttools>=4.22.0 in /opt/conda/lib/python3.11/site-packages (from matplotlib) (4.55.0)\n",
      "Requirement already satisfied: kiwisolver>=1.3.1 in /opt/conda/lib/python3.11/site-packages (from matplotlib) (1.4.7)\n",
      "Requirement already satisfied: numpy>=1.23 in /opt/conda/lib/python3.11/site-packages (from matplotlib) (1.26.4)\n",
      "Requirement already satisfied: packaging>=20.0 in /opt/conda/lib/python3.11/site-packages (from matplotlib) (24.0)\n",
      "Requirement already satisfied: pillow>=8 in /opt/conda/lib/python3.11/site-packages (from matplotlib) (11.0.0)\n",
      "Requirement already satisfied: pyparsing>=2.3.1 in /opt/conda/lib/python3.11/site-packages (from matplotlib) (3.2.0)\n",
      "Requirement already satisfied: python-dateutil>=2.7 in /opt/conda/lib/python3.11/site-packages (from matplotlib) (2.9.0)\n",
      "Requirement already satisfied: six>=1.5 in /opt/conda/lib/python3.11/site-packages (from python-dateutil>=2.7->matplotlib) (1.16.0)\n",
      "\u001b[1m1/1\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 128ms/step\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x1000 with 25 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "!pip install matplotlib\n",
    "import matplotlib.pyplot as plt \n",
    "\n",
    "def sample_images(generator, epoch, num_images=25): \n",
    "    noise = np.random.normal(0, 1, (num_images, 100)) \n",
    "    generated_images = generator.predict(noise) \n",
    "    generated_images = 0.5 * generated_images + 0.5  # Rescale to [0, 1] \n",
    "    fig, axs = plt.subplots(5, 5, figsize=(10, 10)) \n",
    "    count = 0 \n",
    "\n",
    "    for i in range(5): \n",
    "        for j in range(5): \n",
    "            axs[i, j].imshow(generated_images[count, :, :, 0], cmap='gray') \n",
    "            axs[i, j].axis('off') \n",
    "            count += 1 \n",
    "    plt.show() \n",
    "\n",
    "# Sample images at the end of training \n",
    "sample_images(generator, epochs)\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "By using visual inspection, you can quickly identify any issues with the generated images. If you notice blurriness, lack of structure, or lack of diversity, consider adjusting the model architecture or training parameters.\n",
    "\n",
    "### 2. Quantitative Assessment: Metrics\n",
    "\n",
    "While visual inspection provides an intuitive understanding of the GAN’s performance, it can be subjective. To objectively evaluate GAN performance, you can use quantitative metrics such as:\n",
    "\n",
    "- **Inception Score (IS)**: This score measures both the quality and diversity of generated images by using a pre-trained classifier (such as Inception-v3) to predict the class of each image. A higher score indicates that the images are both high-quality and diverse. However, IS is not very effective for simple datasets like MNIST; it’s more suitable for complex datasets.\n",
    "\n",
    "- **Fréchet Inception Distance (FID)**: This metric calculates the distance between the distributions of generated images and real images. A lower FID score indicates that the generated images are more similar to real images. FID is commonly used and considered a reliable metric for evaluating GAN performance.\n",
    "\n",
    "- **Discriminator Accuracy**: During training, if the discriminator's accuracy is around 50%, it suggests that the generator is producing realistic images that are hard to distinguish from real ones. This metric is easy to implement and provides quick feedback on the training progress.\n",
    "\n",
    "### Instructions:\n",
    "\n",
    "You can use the provided code snippet to calculate the discriminator's accuracy on both real and fake images.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 29ms/step\n",
      "Discriminator Accuracy on Real Images: 100.00%\n",
      "Discriminator Accuracy on Fake Images: 100.00%\n"
     ]
    }
   ],
   "source": [
    "# Calculate and print the discriminator accuracy on real vs. fake images\n",
    "noise = np.random.normal(0, 1, (batch_size, 100))\n",
    "generated_images = generator.predict(noise)\n",
    "\n",
    "# Evaluate the discriminator on real images\n",
    "real_images = x_train[np.random.randint(0, x_train.shape[0], batch_size)]\n",
    "d_loss_real = discriminator.evaluate(real_images, np.ones((batch_size, 1)), verbose=0)\n",
    "\n",
    "# Evaluate the discriminator on fake images\n",
    "d_loss_fake = discriminator.evaluate(generated_images, np.zeros((batch_size, 1)), verbose=0)\n",
    "\n",
    "print(f\"Discriminator Accuracy on Real Images: {d_loss_real[1] * 100:.2f}%\")\n",
    "print(f\"Discriminator Accuracy on Fake Images: {d_loss_fake[1] * 100:.2f}%\")\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If the discriminator’s accuracy is around 50%, it implies the generator has learned to produce realistic images. Higher or lower accuracy may suggest that either the generator is not producing convincing images, or the discriminator is overfitting.\n",
    "\n",
    "### 3. Combining Qualitative and Quantitative Assessments\n",
    "\n",
    "For a comprehensive evaluation of the GAN:\n",
    "\n",
    "1. **Start with visual inspection** to get a quick sense of image quality. If the images look blurry or too similar, it might indicate problems with the training process.\n",
    "\n",
    "2. **Use quantitative metrics** like FID or discriminator accuracy to provide objective evidence of the GAN’s performance.\n",
    "\n",
    "3. **Monitor training progress** by visualizing the generator and discriminator losses over time. This helps in understanding if the GAN is suffering from instability or if one model is overpowering the other.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "#### Explanation: \n",
    "\n",
    "By combining qualitative inspection and quantitative metrics, you can effectively assess the quality of images generated by GANs. This dual approach provides a more robust evaluation, ensuring that the generated data is not only visually plausible but also statistically similar to the real data. Experiment with different architectures, learning rates, and training parameters to improve the GAN’s performance further.\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Exercises\n",
    "\n",
    "### Exercise 1: Modify the generator’s architecture \n",
    "\n",
    "#### Objective: \n",
    "- Experiment with adding more layers to the generator to understand how the depth of the network impacts the quality of generated images. \n",
    "\n",
    "#### Instructions: \n",
    "1. Add one more Dense layer with 2048 units to the generator model. \n",
    "2. Rebuild the generator and print the summary. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "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_8\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"sequential_8\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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_28 (<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\">256</span>)            │        <span style=\"color: #00af00; text-decoration-color: #00af00\">25,856</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_20 (<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\">256</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ batch_normalization_10          │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</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",
       "│ dense_29 (<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\">512</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">131,584</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_21 (<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\">512</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ batch_normalization_11          │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">512</span>)            │         <span style=\"color: #00af00; text-decoration-color: #00af00\">2,048</span> │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">BatchNormalization</span>)            │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_30 (<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\">1024</span>)           │       <span style=\"color: #00af00; text-decoration-color: #00af00\">525,312</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_22 (<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\">1024</span>)           │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ batch_normalization_12          │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1024</span>)           │         <span style=\"color: #00af00; text-decoration-color: #00af00\">4,096</span> │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">BatchNormalization</span>)            │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_31 (<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\">2048</span>)           │     <span style=\"color: #00af00; text-decoration-color: #00af00\">2,099,200</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_23 (<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\">2048</span>)           │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ batch_normalization_13          │ (<span style=\"color: #00d7ff; text-decoration-color: #00d7ff\">None</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">2048</span>)           │         <span style=\"color: #00af00; text-decoration-color: #00af00\">8,192</span> │\n",
       "│ (<span style=\"color: #0087ff; text-decoration-color: #0087ff\">BatchNormalization</span>)            │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_32 (<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\">784</span>)            │     <span style=\"color: #00af00; text-decoration-color: #00af00\">1,606,416</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ reshape_3 (<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\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">28</span>, <span style=\"color: #00af00; text-decoration-color: #00af00\">1</span>)      │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
       "</pre>\n"
      ],
      "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_28 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │        \u001b[38;5;34m25,856\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_20 (\u001b[38;5;33mLeakyReLU\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ batch_normalization_10          │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │         \u001b[38;5;34m1,024\u001b[0m │\n",
       "│ (\u001b[38;5;33mBatchNormalization\u001b[0m)            │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_29 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │       \u001b[38;5;34m131,584\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_21 (\u001b[38;5;33mLeakyReLU\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ batch_normalization_11          │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │         \u001b[38;5;34m2,048\u001b[0m │\n",
       "│ (\u001b[38;5;33mBatchNormalization\u001b[0m)            │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_30 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1024\u001b[0m)           │       \u001b[38;5;34m525,312\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_22 (\u001b[38;5;33mLeakyReLU\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1024\u001b[0m)           │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ batch_normalization_12          │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1024\u001b[0m)           │         \u001b[38;5;34m4,096\u001b[0m │\n",
       "│ (\u001b[38;5;33mBatchNormalization\u001b[0m)            │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_31 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m2048\u001b[0m)           │     \u001b[38;5;34m2,099,200\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_23 (\u001b[38;5;33mLeakyReLU\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m2048\u001b[0m)           │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ batch_normalization_13          │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m2048\u001b[0m)           │         \u001b[38;5;34m8,192\u001b[0m │\n",
       "│ (\u001b[38;5;33mBatchNormalization\u001b[0m)            │                        │               │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_32 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m784\u001b[0m)            │     \u001b[38;5;34m1,606,416\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ reshape_3 (\u001b[38;5;33mReshape\u001b[0m)             │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m28\u001b[0m, \u001b[38;5;34m1\u001b[0m)      │             \u001b[38;5;34m0\u001b[0m │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">4,403,728</span> (16.80 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m4,403,728\u001b[0m (16.80 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">4,396,048</span> (16.77 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m4,396,048\u001b[0m (16.77 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">7,680</span> (30.00 KB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m7,680\u001b[0m (30.00 KB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Write your code here\n",
    "\n",
    "# Modify the generator model by adding an additional Dense layer \n",
    "\n",
    "def build_generator(): \n",
    "    model = Sequential() \n",
    "    model.add(Dense(256, input_dim=100)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(BatchNormalization(momentum=0.8)) \n",
    "    model.add(Dense(512)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(BatchNormalization(momentum=0.8)) \n",
    "    model.add(Dense(1024)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(BatchNormalization(momentum=0.8)) \n",
    "    model.add(Dense(2048))  # New layer added \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(BatchNormalization(momentum=0.8)) \n",
    "    model.add(Dense(28 * 28 * 1, activation='tanh')) \n",
    "    model.add(Reshape((28, 28, 1))) \n",
    "    return model \n",
    "\n",
    "# Rebuild the generator \n",
    "generator = build_generator() \n",
    "generator.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<details>\n",
    "    <summary>Click here for Solution</summary>\n",
    "\n",
    "```python\n",
    "# Modify the generator model by adding an additional Dense layer \n",
    "\n",
    "def build_generator(): \n",
    "    model = Sequential() \n",
    "    model.add(Dense(256, input_dim=100)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(BatchNormalization(momentum=0.8)) \n",
    "    model.add(Dense(512)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(BatchNormalization(momentum=0.8)) \n",
    "    model.add(Dense(1024)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(BatchNormalization(momentum=0.8)) \n",
    "    model.add(Dense(2048))  # New layer added \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(BatchNormalization(momentum=0.8)) \n",
    "    model.add(Dense(28 * 28 * 1, activation='tanh')) \n",
    "    model.add(Reshape((28, 28, 1))) \n",
    "    return model \n",
    "\n",
    "# Rebuild the generator \n",
    "generator = build_generator() \n",
    "generator.summary() \n",
    "```\n",
    "\n",
    "</details>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Exercise 2 - Adjust the discriminator’s learning rate \n",
    "\n",
    "#### Objective: \n",
    "- Explore how changing the learning rate of the discriminator’s optimizer affects training stability and model performance. \n",
    "\n",
    "#### Instructions: \n",
    "1. Change the learning rate of the Adam optimizer for the discriminator to 0.0002. \n",
    "2. Rebuild and compile the discriminator with the new learning rate. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "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_9\"</span>\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1mModel: \"sequential_9\"\u001b[0m\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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",
       "│ flatten_5 (<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\">784</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_33 (<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\">512</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">401,920</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_24 (<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\">512</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_34 (<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\">256</span>)            │       <span style=\"color: #00af00; text-decoration-color: #00af00\">131,328</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_25 (<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\">256</span>)            │             <span style=\"color: #00af00; text-decoration-color: #00af00\">0</span> │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_35 (<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\">257</span> │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n",
       "</pre>\n"
      ],
      "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",
       "│ flatten_5 (\u001b[38;5;33mFlatten\u001b[0m)             │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m784\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_33 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │       \u001b[38;5;34m401,920\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_24 (\u001b[38;5;33mLeakyReLU\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m512\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_34 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │       \u001b[38;5;34m131,328\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ leaky_re_lu_25 (\u001b[38;5;33mLeakyReLU\u001b[0m)      │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m256\u001b[0m)            │             \u001b[38;5;34m0\u001b[0m │\n",
       "├─────────────────────────────────┼────────────────────────┼───────────────┤\n",
       "│ dense_35 (\u001b[38;5;33mDense\u001b[0m)                │ (\u001b[38;5;45mNone\u001b[0m, \u001b[38;5;34m1\u001b[0m)              │           \u001b[38;5;34m257\u001b[0m │\n",
       "└─────────────────────────────────┴────────────────────────┴───────────────┘\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">533,505</span> (2.04 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Total params: \u001b[0m\u001b[38;5;34m533,505\u001b[0m (2.04 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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\">533,505</span> (2.04 MB)\n",
       "</pre>\n"
      ],
      "text/plain": [
       "\u001b[1m Trainable params: \u001b[0m\u001b[38;5;34m533,505\u001b[0m (2.04 MB)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    },
    {
     "data": {
      "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"
      ],
      "text/plain": [
       "\u001b[1m Non-trainable params: \u001b[0m\u001b[38;5;34m0\u001b[0m (0.00 B)\n"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Write your code here\n",
    "\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Flatten, LeakyReLU\n",
    "\n",
    "def build_discriminator(): \n",
    "    model = Sequential() \n",
    "    model.add(Flatten(input_shape=(28, 28, 1))) \n",
    "    model.add(Dense(512)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(Dense(256)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(Dense(1, activation='sigmoid')) \n",
    "    return model \n",
    "\n",
    "# Rebuild and compile the discriminator with a lower learning rate \n",
    "discriminator = build_discriminator() \n",
    "discriminator.compile(loss='binary_crossentropy', \n",
    "                      optimizer=tf.keras.optimizers.Adam(learning_rate=0.0002), \n",
    "                      metrics=['accuracy']) \n",
    "\n",
    "discriminator.summary()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<details>\n",
    "    <summary>Click here for Solution</summary>\n",
    "\n",
    "```python\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.models import Sequential\n",
    "from tensorflow.keras.layers import Dense, Flatten, LeakyReLU\n",
    "\n",
    "def build_discriminator(): \n",
    "    model = Sequential() \n",
    "    model.add(Flatten(input_shape=(28, 28, 1))) \n",
    "    model.add(Dense(512)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(Dense(256)) \n",
    "    model.add(LeakyReLU(alpha=0.2)) \n",
    "    model.add(Dense(1, activation='sigmoid')) \n",
    "    return model \n",
    "\n",
    "# Rebuild and compile the discriminator with a lower learning rate \n",
    "discriminator = build_discriminator() \n",
    "discriminator.compile(loss='binary_crossentropy', \n",
    "                      optimizer=tf.keras.optimizers.Adam(learning_rate=0.0002), \n",
    "                      metrics=['accuracy']) \n",
    "\n",
    "discriminator.summary()\n",
    "\n",
    "```\n",
    "\n",
    "</details>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Exercise 3 - Visualize training progress \n",
    "\n",
    "#### Objective: \n",
    "- Visualize the loss of the discriminator and generator during training to monitor the training process.  \n",
    "\n",
    "#### Instructions: \n",
    "1. Modify the training loop to store the discriminator and generator losses. \n",
    "2. Plot the losses after training to visualize the training progress. \n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 59ms/step \n",
      "0 [D loss: 2.1259984970092773] [D accuracy: 25.0] [G loss: 0.08216238766908646]\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 64ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 46ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 50ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 50ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 51ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 41ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 32ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 45ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 60ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 50ms/step\n",
      "10 [D loss: 0.34541231393814087] [D accuracy: 94.91003751754761] [G loss: 0.06948293000459671]\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 46ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 41ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 63ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 51ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 53ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 67ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 49ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 50ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 50ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 51ms/step\n",
      "20 [D loss: 0.20797055959701538] [D accuracy: 97.36407399177551] [G loss: 0.06025505065917969]\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 31ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 43ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 51ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 51ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 41ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 46ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 52ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 46ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 40ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 52ms/step\n",
      "30 [D loss: 0.1456626057624817] [D accuracy: 98.22142720222473] [G loss: 0.053290072828531265]\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 51ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 48ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 49ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 54ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 48ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 59ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 52ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 57ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 50ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 59ms/step\n",
      "40 [D loss: 0.1115155816078186] [D accuracy: 98.65792989730835] [G loss: 0.04782131686806679]\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 73ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 59ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 45ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 50ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 49ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 51ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 49ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 59ms/step\n",
      "\u001b[1m2/2\u001b[0m \u001b[32m━━━━━━━━━━━━━━━━━━━━\u001b[0m\u001b[37m\u001b[0m \u001b[1m0s\u001b[0m 37ms/step\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1000x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Write your code here\n",
    "\n",
    "# Initialize lists to store losses \n",
    "d_losses = [] \n",
    "g_losses = [] \n",
    "\n",
    "# Training loop with loss storage \n",
    "for epoch in range(epochs): \n",
    "    idx = np.random.randint(0, x_train.shape[0], batch_size) \n",
    "    real_images = x_train[idx] \n",
    "    noise = np.random.normal(0, 1, (batch_size, 100)) \n",
    "    generated_images = generator.predict(noise) \n",
    "    d_loss_real = discriminator.train_on_batch(real_images, real) \n",
    "    d_loss_fake = discriminator.train_on_batch(generated_images, fake) \n",
    "    d_loss = 0.5 * np.add(d_loss_real, d_loss_fake) \n",
    "    d_losses.append(d_loss[0]) \n",
    "  \n",
    "    noise = np.random.normal(0, 1, (batch_size, 100)) \n",
    "    g_loss = gan.train_on_batch(noise, real) \n",
    "    g_losses.append(g_loss) \n",
    "  \n",
    "    if epoch % sample_interval == 0: \n",
    "        print(f\"{epoch} [D loss: {d_loss[0]}] [D accuracy: {100 * d_loss[1]}] [G loss: {g_loss}]\") \n",
    "  \n",
    "# Plot the training losses \n",
    "plt.figure(figsize=(10, 5)) \n",
    "plt.plot(d_losses, label='Discriminator Loss') \n",
    "plt.plot(g_losses, label='Generator Loss') \n",
    "plt.xlabel('Epoch') \n",
    "plt.ylabel('Loss') \n",
    "plt.title('Training Losses') \n",
    "plt.legend() \n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<details>\n",
    "    <summary>Click here for Solution</summary>\n",
    "\n",
    "```python\n",
    "# Initialize lists to store losses \n",
    "d_losses = [] \n",
    "g_losses = [] \n",
    "\n",
    " \n",
    "# Training loop with loss storage \n",
    "for epoch in range(epochs): \n",
    "    idx = np.random.randint(0, x_train.shape[0], batch_size) \n",
    "    real_images = x_train[idx] \n",
    "    noise = np.random.normal(0, 1, (batch_size, 100)) \n",
    "    generated_images = generator.predict(noise) \n",
    "    d_loss_real = discriminator.train_on_batch(real_images, real) \n",
    "    d_loss_fake = discriminator.train_on_batch(generated_images, fake) \n",
    "    d_loss = 0.5 * np.add(d_loss_real, d_loss_fake) \n",
    "    d_losses.append(d_loss[0]) \n",
    "  \n",
    "\n",
    "    noise = np.random.normal(0, 1, (batch_size, 100)) \n",
    "    g_loss = gan.train_on_batch(noise, real) \n",
    "    g_losses.append(g_loss) \n",
    "\n",
    "  \n",
    "    if epoch % sample_interval == 0: \n",
    "        print(f\"{epoch} [D loss: {d_loss[0]}] [D accuracy: {100 * d_loss[1]}] [G loss: {g_loss}]\") \n",
    "  \n",
    "# Plot the training losses \n",
    "plt.figure(figsize=(10, 5)) \n",
    "plt.plot(d_losses, label='Discriminator Loss') \n",
    "plt.plot(g_losses, label='Generator Loss') \n",
    "plt.xlabel('Epoch') \n",
    "plt.ylabel('Loss') \n",
    "plt.title('Training Losses') \n",
    "plt.legend() \n",
    "plt.show() \n",
    "```\n",
    "\n",
    "</details>\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Summary \n",
    "By completing these exercises, students will:  \n",
    "1. Understand the impact of adding more layers to the generator on the quality of generated images. \n",
    "2. Learn how adjusting the learning rate of the discriminator’s optimizer can affect training stability and performance. \n",
    "3. Visualize the training progress by plotting the losses of the discriminator and generator. \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Conclusion: \n",
    "\n",
    "In this lab, you have gained practical experience in developing GANs using Keras. By completing these exercises, you have developed a strong understanding of the key concepts and techniques involved in working with GANs. Continue experimenting with different architectures, datasets, and training strategies to deepen your knowledge and skills in this exciting area of machine learning. Thank you for participating in this lab, and happy coding! \n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Copyright © IBM Corporation. All rights reserved.\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.9"
  },
  "prev_pub_hash": "a15aba3b24bff4b757cc330b887b99e24759a4bc72375ddd66d988acbcabf860"
 },
 "nbformat": 4,
 "nbformat_minor": 4
}