Run_this_C4W3_Assignment.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "jupytext": {
      "encoding": "# -*- coding: utf-8 -*-"
    },
    "kernelspec": {
      "display_name": "Python 3",
      "language": "python",
      "name": "python3"
    },
    "colab": {
      "name": "Run_this_C4W3_Assignment.ipynb",
      "provenance": [],
      "collapsed_sections": []
    },
    "accelerator": "GPU"
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "prTKL3d2kGZE"
      },
      "source": [
        "# Week 3: Variational Autoencoders on Anime Faces\n",
        "\n",
        "For this exercise, you will train a Variational Autoencoder (VAE) using the [anime faces dataset by MckInsey666](https://github.com/bchao1/Anime-Face-Dataset). \n",
        "\n",
        "You will train the model using the techniques discussed in class. At the end, you should save your model and download it from Colab so that it can be submitted to the autograder for grading."
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nms7__aqDj48"
      },
      "source": [
        "***Important:*** *This colab notebook has read-only access so you won't be able to save your changes. If you want to save your work periodically, please click `File -> Save a Copy in Drive` to create a copy in your account, then work from there.*  "
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2Qxq9uZAk3Lh"
      },
      "source": [
        "## Imports"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "MooRFGEeI1zb"
      },
      "source": [
        "import tensorflow as tf\n",
        "import tensorflow_datasets as tfds\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np\n",
        "\n",
        "import os\n",
        "import zipfile\n",
        "import urllib.request\n",
        "import random\n",
        "from IPython import display"
      ],
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wL9rq-0uk7nS"
      },
      "source": [
        "## Parameters"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pjhN6GgfmUfx"
      },
      "source": [
        "# set a random seed\n",
        "np.random.seed(51)\n",
        "\n",
        "# parameters for building the model and training\n",
        "BATCH_SIZE=2000\n",
        "LATENT_DIM=512\n",
        "IMAGE_SIZE=64"
      ],
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "hXTdjxmolDBo"
      },
      "source": [
        "## Download the Dataset\n",
        "\n",
        "You will download the Anime Faces dataset and save it to a local directory."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qxKW6Q88KHcL"
      },
      "source": [
        "# make the data directory\n",
        "try:\n",
        "  os.mkdir('/tmp/anime')\n",
        "except OSError:\n",
        "  pass\n",
        "\n",
        "# download the zipped dataset to the data directory\n",
        "data_url = \"https://storage.googleapis.com/laurencemoroney-blog.appspot.com/Resources/anime-faces.zip\"\n",
        "data_file_name = \"animefaces.zip\"\n",
        "download_dir = '/tmp/anime/'\n",
        "urllib.request.urlretrieve(data_url, data_file_name)\n",
        "\n",
        "# extract the zip file\n",
        "zip_ref = zipfile.ZipFile(data_file_name, 'r')\n",
        "zip_ref.extractall(download_dir)\n",
        "zip_ref.close()"
      ],
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kD6WCIlclWaA"
      },
      "source": [
        "## Prepare the Dataset"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kbaVpD18ggOX"
      },
      "source": [
        "Next is preparing the data for training and validation. We've provided you some utilities below."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "NTlx97U_JDPB"
      },
      "source": [
        "# Data Preparation Utilities\n",
        "\n",
        "def get_dataset_slice_paths(image_dir):\n",
        "  '''returns a list of paths to the image files'''\n",
        "  image_file_list = os.listdir(image_dir)\n",
        "  image_paths = [os.path.join(image_dir, fname) for fname in image_file_list]\n",
        "\n",
        "  return image_paths\n",
        "\n",
        "\n",
        "def map_image(image_filename):\n",
        "  '''preprocesses the images'''\n",
        "  img_raw = tf.io.read_file(image_filename)\n",
        "  image = tf.image.decode_jpeg(img_raw)\n",
        "\n",
        "  image = tf.cast(image, dtype=tf.float32)\n",
        "  image = tf.image.resize(image, (IMAGE_SIZE, IMAGE_SIZE))\n",
        "  image = image / 255.0  \n",
        "  image = tf.reshape(image, shape=(IMAGE_SIZE, IMAGE_SIZE, 3,))\n",
        "\n",
        "  return image"
      ],
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "0uFon6vdhMhi"
      },
      "source": [
        "You will use the functions above to generate the train and validation sets."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "iGoCJ6DPJHL8",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "8ca50e83-274e-4140-9659-081880626fa9"
      },
      "source": [
        "# get the list containing the image paths\n",
        "paths = get_dataset_slice_paths(\"/tmp/anime/images/\")\n",
        "\n",
        "# shuffle the paths\n",
        "random.shuffle(paths)\n",
        "\n",
        "# split the paths list into to training (80%) and validation sets(20%).\n",
        "paths_len = len(paths)\n",
        "train_paths_len = int(paths_len * 0.8)\n",
        "\n",
        "train_paths = paths[:train_paths_len]\n",
        "val_paths = paths[train_paths_len:]\n",
        "\n",
        "# load the training image paths into tensors, create batches and shuffle\n",
        "training_dataset = tf.data.Dataset.from_tensor_slices((train_paths))  #filename\n",
        "training_dataset = training_dataset.map(map_image)\n",
        "training_dataset = training_dataset.shuffle(1000).batch(BATCH_SIZE)\n",
        "\n",
        "# load the validation image paths into tensors and create batches\n",
        "validation_dataset = tf.data.Dataset.from_tensor_slices((val_paths))\n",
        "validation_dataset = validation_dataset.map(map_image)\n",
        "validation_dataset = validation_dataset.batch(BATCH_SIZE)\n",
        "\n",
        "\n",
        "print(f'number of batches in the training set: {len(training_dataset)}')\n",
        "print(f'number of batches in the validation set: {len(validation_dataset)}')"
      ],
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "number of batches in the training set: 26\n",
            "number of batches in the validation set: 7\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "72ZRga9vlonx"
      },
      "source": [
        "## Display Utilities\n",
        "\n",
        "We've also provided some utilities to help in visualizing the data."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jC1cpLViJLIu"
      },
      "source": [
        "def display_faces(dataset, size=9):\n",
        "  '''Takes a sample from a dataset batch and plots it in a grid.'''\n",
        "  dataset = dataset.unbatch().take(size)\n",
        "  n_cols = 3\n",
        "  n_rows = size//n_cols + 1\n",
        "  plt.figure(figsize=(5, 5))\n",
        "  i = 0\n",
        "  for image in dataset:\n",
        "    i += 1\n",
        "    disp_img = np.reshape(image, (64,64,3))\n",
        "    plt.subplot(n_rows, n_cols, i)\n",
        "    plt.xticks([])\n",
        "    plt.yticks([])\n",
        "    plt.imshow(disp_img)\n",
        "\n",
        "\n",
        "def display_one_row(disp_images, offset, shape=(28, 28)):\n",
        "  '''Displays a row of images.'''\n",
        "  for idx, image in enumerate(disp_images):\n",
        "    plt.subplot(3, 10, offset + idx + 1)\n",
        "    plt.xticks([])\n",
        "    plt.yticks([])\n",
        "    image = np.reshape(image, shape)\n",
        "    plt.imshow(image)\n",
        "\n",
        "\n",
        "def display_results(disp_input_images, disp_predicted):\n",
        "  '''Displays input and predicted images.'''\n",
        "  plt.figure(figsize=(15, 5))\n",
        "  display_one_row(disp_input_images, 0, shape=(IMAGE_SIZE,IMAGE_SIZE,3))\n",
        "  display_one_row(disp_predicted, 20, shape=(IMAGE_SIZE,IMAGE_SIZE,3))\n"
      ],
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "2brROh6qLJbs"
      },
      "source": [
        "Let's see some of the anime faces from the validation dataset."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "5eZsrZtqJOzv",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 250
        },
        "outputId": "ea54d340-162f-49fd-d639-9cb4a2cd06e2"
      },
      "source": [
        "display_faces(validation_dataset, size=12)"
      ],
      "execution_count": 8,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 360x360 with 12 Axes>"
            ]
          },
          "metadata": {}
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LSBtdCVim9aC"
      },
      "source": [
        "## Build the Model"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WQvzWaNqLrB1"
      },
      "source": [
        "You will be building your VAE in the following sections. Recall that this will follow and encoder-decoder architecture and can be summarized by the figure below.\n",
        "\n",
        "<img src=\"https://drive.google.com/uc?export=view&id=1YAZAeMGEJ1KgieYk1ju-S9DoshpMREeC\" width=\"60%\" height=\"60%\"/>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "KHNxIUUS9ng9"
      },
      "source": [
        "### Sampling Class\n",
        "\n",
        "You will start with the custom layer to provide the Gaussian noise input along with the mean (mu) and standard deviation (sigma) of the encoder's output. Recall the equation to combine these:\n",
        "\n",
        "$$z = \\mu + e^{0.5\\sigma} * \\epsilon  $$\n",
        "\n",
        "where $\\mu$ = mean, $\\sigma$ = standard deviation, and $\\epsilon$ = random sample"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "L-3qk6ZBm0Fl"
      },
      "source": [
        "class Sampling(tf.keras.layers.Layer):\n",
        "  def call(self, inputs):\n",
        "    \"\"\"Generates a random sample and combines with the encoder output\n",
        "    \n",
        "    Args:\n",
        "      inputs -- output tensor from the encoder\n",
        "\n",
        "    Returns:\n",
        "      `inputs` tensors combined with a random sample\n",
        "    \"\"\"\n",
        "    ### START CODE HERE ###\n",
        "    mu, sigma = inputs\n",
        "    batch = tf.shape(mu)[0]\n",
        "    dim = tf.shape(mu)[1]\n",
        "    epsilon = tf.random.normal( shape=(batch, dim))\n",
        "    z = mu + tf.math.exp(0.5*sigma) *epsilon\n",
        "    ### END CODE HERE ###\n",
        "    return  z / tf.math.reduce_max(z)"
      ],
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tZjCSa7Y-Gvk"
      },
      "source": [
        "### Encoder Layers\n",
        "\n",
        "Next, please use the Functional API to stack the encoder layers and output `mu`, `sigma` and the shape of the features before flattening. We expect you to use 3 convolutional layers (instead of 2 in the ungraded lab) but feel free to revise as you see fit. Another hint is to use `1024` units in the Dense layer before you get mu and sigma (we used `20` for it in the ungraded lab).\n",
        "\n",
        "*Note: If you did Week 4 before Week 3, please do not use LeakyReLU activations yet for this particular assignment. The grader for Week 3 does not support LeakyReLU yet. This will be updated but for now, you can use `relu` and `sigmoid` just like in the ungraded lab.*"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7VSVYjDim4Dk"
      },
      "source": [
        "def encoder_layers(inputs, latent_dim):\n",
        "  \"\"\"Defines the encoder's layers.\n",
        "  Args:\n",
        "    inputs -- batch from the dataset\n",
        "    latent_dim -- dimensionality of the latent space\n",
        "\n",
        "  Returns:\n",
        "    mu -- learned mean\n",
        "    sigma -- learned standard deviation\n",
        "    batch_3.shape -- shape of the features before flattening\n",
        "  \"\"\"\n",
        "  ### START CODE HERE ###\n",
        "  x = tf.keras.layers.Conv2D(filters=32, kernel_size=3, strides=2, padding='same', activation='relu')(inputs)\n",
        "  x = tf.keras.layers.BatchNormalization()(x)\n",
        "\n",
        "  x = tf.keras.layers.Conv2D(filters=64, kernel_size=3, strides=2, padding='same', activation='relu')(x)\n",
        "  x = tf.keras.layers.BatchNormalization()(x)\n",
        "\n",
        "  x = tf.keras.layers.Conv2D(filters=128, kernel_size=3, strides=2, padding='same', activation='relu')(x)\n",
        "  x = tf.keras.layers.BatchNormalization()(x)\n",
        "\n",
        "  batch_3 = x\n",
        "  \n",
        "  x = tf.keras.layers.Flatten()(x)\n",
        "  x = tf.keras.layers.Dense(units=1024, activation='relu')(x)\n",
        "\n",
        "  mu = tf.keras.layers.Dense(units=LATENT_DIM, activation='relu')(x)\n",
        "  sigma = tf.keras.layers.Dense(units=LATENT_DIM, activation='relu')(x)\n",
        "  \n",
        "  ### END CODE HERE ###\n",
        "\n",
        "  # revise `batch_3.shape` here if you opted not to use 3 Conv2D layers\n",
        "  print('shape:', batch_3.shape)\n",
        "  return mu, sigma, batch_3.shape"
      ],
      "execution_count": 10,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZOy7wPPY-g-N"
      },
      "source": [
        "### Encoder Model\n",
        "\n",
        "You will feed the output from the above function to the `Sampling layer` you defined earlier. That will have the latent representations that can be fed to the decoder network later. Please complete the function below to build the encoder network with the `Sampling` layer."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "w8Y-wLFym60N"
      },
      "source": [
        "def encoder_model(latent_dim, input_shape):\n",
        "  \"\"\"Defines the encoder model with the Sampling layer\n",
        "  Args:\n",
        "    latent_dim -- dimensionality of the latent space\n",
        "    input_shape -- shape of the dataset batch\n",
        "\n",
        "  Returns:\n",
        "    model -- the encoder model\n",
        "    conv_shape -- shape of the features before flattening\n",
        "  \"\"\"\n",
        "  ### START CODE HERE ###\n",
        "  inputs = tf.keras.layers.Input(shape=input_shape)\n",
        "  mu, sigma, conv_shape = encoder_layers(inputs, LATENT_DIM)\n",
        "  z = Sampling()((mu, sigma))\n",
        "  model = tf.keras.models.Model(inputs=inputs , outputs=[mu,sigma,z])\n",
        "  ### END CODE HERE ###\n",
        "  model.summary()\n",
        "  return model, conv_shape"
      ],
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "encoder_model(LATENT_DIM, (64,64,3))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iUX1BhA85hY5",
        "outputId": "148e10d0-3c86-4eac-8869-a0e66c9b7b80"
      },
      "execution_count": 12,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "shape: (None, 8, 8, 128)\n",
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                   Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            " input_1 (InputLayer)           [(None, 64, 64, 3)]  0           []                               \n",
            "                                                                                                  \n",
            " conv2d (Conv2D)                (None, 32, 32, 32)   896         ['input_1[0][0]']                \n",
            "                                                                                                  \n",
            " batch_normalization (BatchNorm  (None, 32, 32, 32)  128         ['conv2d[0][0]']                 \n",
            " alization)                                                                                       \n",
            "                                                                                                  \n",
            " conv2d_1 (Conv2D)              (None, 16, 16, 64)   18496       ['batch_normalization[0][0]']    \n",
            "                                                                                                  \n",
            " batch_normalization_1 (BatchNo  (None, 16, 16, 64)  256         ['conv2d_1[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " conv2d_2 (Conv2D)              (None, 8, 8, 128)    73856       ['batch_normalization_1[0][0]']  \n",
            "                                                                                                  \n",
            " batch_normalization_2 (BatchNo  (None, 8, 8, 128)   512         ['conv2d_2[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " flatten (Flatten)              (None, 8192)         0           ['batch_normalization_2[0][0]']  \n",
            "                                                                                                  \n",
            " dense (Dense)                  (None, 1024)         8389632     ['flatten[0][0]']                \n",
            "                                                                                                  \n",
            " dense_1 (Dense)                (None, 512)          524800      ['dense[0][0]']                  \n",
            "                                                                                                  \n",
            " dense_2 (Dense)                (None, 512)          524800      ['dense[0][0]']                  \n",
            "                                                                                                  \n",
            " sampling (Sampling)            (None, 512)          0           ['dense_1[0][0]',                \n",
            "                                                                  'dense_2[0][0]']                \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 9,533,376\n",
            "Trainable params: 9,532,928\n",
            "Non-trainable params: 448\n",
            "__________________________________________________________________________________________________\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(<keras.engine.functional.Functional at 0x7ff2131d6250>,\n",
              " TensorShape([None, 8, 8, 128]))"
            ]
          },
          "metadata": {},
          "execution_count": 12
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "A9ENB-6a-0R5"
      },
      "source": [
        "### Decoder Layers\n",
        "\n",
        "Next, you will define the decoder layers. This will expand the latent representations back to the original image dimensions. After training your VAE model, you can use this decoder model to generate new data by feeding random inputs."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qlTjAzgsm9Vn"
      },
      "source": [
        "def decoder_layers(inputs, conv_shape):\n",
        "  \"\"\"Defines the decoder layers.\n",
        "  Args:\n",
        "    inputs -- output of the encoder \n",
        "    conv_shape -- shape of the features before flattening\n",
        "\n",
        "  Returns:\n",
        "    tensor containing the decoded output\n",
        "  \"\"\"\n",
        "  ### START CODE HERE ###\n",
        "  flat_layers = conv_shape[1]*conv_shape[2]*conv_shape[3]\n",
        "  x = tf.keras.layers.Dense(units=flat_layers, activation='relu' )(inputs)\n",
        "\n",
        "  x = tf.keras.layers.Reshape((conv_shape[1], conv_shape[2], conv_shape[3]),\n",
        "                              input_shape=(flat_layers,))(x)\n",
        "  x = tf.keras.layers.Conv2DTranspose(filters=128, kernel_size=3, strides=2, padding='same', activation='relu')(x)\n",
        "  x = tf.keras.layers.BatchNormalization()(x)\n",
        "\n",
        "  x = tf.keras.layers.Conv2DTranspose(filters=64, kernel_size=3, strides=2, padding='same', activation='relu')(x)\n",
        "  x = tf.keras.layers.BatchNormalization()(x)\n",
        "\n",
        "  x = tf.keras.layers.Conv2DTranspose(filters=32, kernel_size=3, strides=2, padding='same',activation='relu')(x)\n",
        "  x = tf.keras.layers.BatchNormalization()(x)\n",
        "\n",
        "  x = tf.keras.layers.Conv2DTranspose(filters=3, kernel_size=3, strides=1, padding='same', activation='sigmoid')(x)\n",
        "  \n",
        "  \n",
        "  ### END CODE HERE ###\n",
        "  return x"
      ],
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "cfLLz84r_MlN"
      },
      "source": [
        "### Decoder Model\n",
        "\n",
        "Please complete the function below to output the decoder model."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "sUgTyqNFm_jR"
      },
      "source": [
        "def decoder_model(latent_dim, conv_shape):\n",
        "  \"\"\"Defines the decoder model.\n",
        "  Args:\n",
        "    latent_dim -- dimensionality of the latent space\n",
        "    conv_shape -- shape of the features before flattening\n",
        "\n",
        "  Returns:\n",
        "    model -- the decoder model\n",
        "  \"\"\"\n",
        "  ### START CODE HERE ###\n",
        "  # set the inputs to the shape of the latent space\n",
        "  inputs = tf.keras.layers.Input(shape=(latent_dim,))\n",
        "  # get the output of the decoder layers\n",
        "  outputs = decoder_layers(inputs, conv_shape)\n",
        "  model = tf.keras.models.Model(inputs= inputs, outputs= outputs)\n",
        "  ### END CODE HERE ###\n",
        "  model.summary()\n",
        "  return model"
      ],
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "decoder_model(LATENT_DIM, conv_shape = (None,8,8,128))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "PMD22rPjmKjN",
        "outputId": "9a883e23-d625-43ea-e9ca-05ef8127daba"
      },
      "execution_count": 15,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"model_1\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_2 (InputLayer)        [(None, 512)]             0         \n",
            "                                                                 \n",
            " dense_3 (Dense)             (None, 8192)              4202496   \n",
            "                                                                 \n",
            " reshape (Reshape)           (None, 8, 8, 128)         0         \n",
            "                                                                 \n",
            " conv2d_transpose (Conv2DTra  (None, 16, 16, 128)      147584    \n",
            " nspose)                                                         \n",
            "                                                                 \n",
            " batch_normalization_3 (Batc  (None, 16, 16, 128)      512       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv2d_transpose_1 (Conv2DT  (None, 32, 32, 64)       73792     \n",
            " ranspose)                                                       \n",
            "                                                                 \n",
            " batch_normalization_4 (Batc  (None, 32, 32, 64)       256       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv2d_transpose_2 (Conv2DT  (None, 64, 64, 32)       18464     \n",
            " ranspose)                                                       \n",
            "                                                                 \n",
            " batch_normalization_5 (Batc  (None, 64, 64, 32)       128       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv2d_transpose_3 (Conv2DT  (None, 64, 64, 3)        867       \n",
            " ranspose)                                                       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 4,444,099\n",
            "Trainable params: 4,443,651\n",
            "Non-trainable params: 448\n",
            "_________________________________________________________________\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.engine.functional.Functional at 0x7ff197dedd10>"
            ]
          },
          "metadata": {},
          "execution_count": 15
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ps0yuE1d_cQc"
      },
      "source": [
        "### Kullback–Leibler Divergence\n",
        "\n",
        "Next, you will define the function to compute the [Kullback–Leibler Divergence](https://arxiv.org/abs/2002.07514) loss. This will be used to improve the generative capability of the model. This code is already given.\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tngFmDDwnDn-"
      },
      "source": [
        "def kl_reconstruction_loss(inputs, outputs, mu, sigma):\n",
        "  \"\"\" Computes the Kullback-Leibler Divergence (KLD)\n",
        "  Args:\n",
        "    inputs -- batch from the dataset\n",
        "    outputs -- output of the Sampling layer\n",
        "    mu -- mean\n",
        "    sigma -- standard deviation\n",
        "\n",
        "  Returns:\n",
        "    KLD loss\n",
        "  \"\"\"\n",
        "  kl_loss = 1 + sigma - tf.square(mu) - tf.math.exp(sigma)\n",
        "  return tf.reduce_mean(kl_loss) * -0.5"
      ],
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Pi1I431I_og7"
      },
      "source": [
        "### Putting it all together\n",
        "\n",
        "Please define the whole VAE model. Remember to use `model.add_loss()` to add the KL reconstruction loss. This will be accessed and added to the loss later in the training loop."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "cuPHg28JnGCp"
      },
      "source": [
        "def vae_model(encoder, decoder, input_shape):\n",
        "  \"\"\"Defines the VAE model\n",
        "  Args:\n",
        "    encoder -- the encoder model\n",
        "    decoder -- the decoder model\n",
        "    input_shape -- shape of the dataset batch\n",
        "\n",
        "  Returns:\n",
        "    the complete VAE model\n",
        "  \"\"\"\n",
        "  ### START CODE HERE ###\n",
        "  inputs = tf.keras.layers.Input(shape=input_shape)\n",
        "\n",
        "  # get mu, sigma, and z from the encoder output\n",
        "  mu, sigma, z = encoder(inputs)\n",
        "  \n",
        "  # get reconstructed output from the decoder\n",
        "  reconstructed = decoder(z)\n",
        "\n",
        "  # define the inputs and outputs of the VAE\n",
        "  model = tf.keras.Model(inputs=inputs, outputs=reconstructed)\n",
        "\n",
        "  \n",
        "  ### END CODE HERE ###\n",
        "  return model"
      ],
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "P_lbWSKbALf-"
      },
      "source": [
        "Next, please define a helper function to return the encoder, decoder, and vae models you just defined.\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hnPo0Pr3nIts",
        "lines_to_next_cell": 2
      },
      "source": [
        "def get_models(input_shape, latent_dim):\n",
        "  \"\"\"Returns the encoder, decoder, and vae models\"\"\"\n",
        "  ### START CODE HERE ###\n",
        "  encoder, conv_shape = encoder_model(latent_dim=latent_dim, input_shape=input_shape)\n",
        "  decoder = decoder_model(latent_dim=latent_dim, conv_shape=conv_shape)\n",
        "  vae = vae_model(encoder, decoder, input_shape=input_shape)  \n",
        "\n",
        "  ### END CODE HERE ###\n",
        "  return encoder, decoder, vae"
      ],
      "execution_count": 18,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wJsdzZTPVgOn"
      },
      "source": [
        "Let's use the function above to get the models we need in the training loop.\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IHdr3CUznL5Z",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "ba8adcc7-09b0-4c9b-8a1e-a4eeca603e4f"
      },
      "source": [
        "encoder, decoder, vae = get_models(input_shape=(64,64,3,), latent_dim=LATENT_DIM)"
      ],
      "execution_count": 19,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "shape: (None, 8, 8, 128)\n",
            "Model: \"model_2\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                   Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            " input_3 (InputLayer)           [(None, 64, 64, 3)]  0           []                               \n",
            "                                                                                                  \n",
            " conv2d_3 (Conv2D)              (None, 32, 32, 32)   896         ['input_3[0][0]']                \n",
            "                                                                                                  \n",
            " batch_normalization_6 (BatchNo  (None, 32, 32, 32)  128         ['conv2d_3[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " conv2d_4 (Conv2D)              (None, 16, 16, 64)   18496       ['batch_normalization_6[0][0]']  \n",
            "                                                                                                  \n",
            " batch_normalization_7 (BatchNo  (None, 16, 16, 64)  256         ['conv2d_4[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " conv2d_5 (Conv2D)              (None, 8, 8, 128)    73856       ['batch_normalization_7[0][0]']  \n",
            "                                                                                                  \n",
            " batch_normalization_8 (BatchNo  (None, 8, 8, 128)   512         ['conv2d_5[0][0]']               \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " flatten_1 (Flatten)            (None, 8192)         0           ['batch_normalization_8[0][0]']  \n",
            "                                                                                                  \n",
            " dense_4 (Dense)                (None, 1024)         8389632     ['flatten_1[0][0]']              \n",
            "                                                                                                  \n",
            " dense_5 (Dense)                (None, 512)          524800      ['dense_4[0][0]']                \n",
            "                                                                                                  \n",
            " dense_6 (Dense)                (None, 512)          524800      ['dense_4[0][0]']                \n",
            "                                                                                                  \n",
            " sampling_1 (Sampling)          (None, 512)          0           ['dense_5[0][0]',                \n",
            "                                                                  'dense_6[0][0]']                \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 9,533,376\n",
            "Trainable params: 9,532,928\n",
            "Non-trainable params: 448\n",
            "__________________________________________________________________________________________________\n",
            "Model: \"model_3\"\n",
            "_________________________________________________________________\n",
            " Layer (type)                Output Shape              Param #   \n",
            "=================================================================\n",
            " input_4 (InputLayer)        [(None, 512)]             0         \n",
            "                                                                 \n",
            " dense_7 (Dense)             (None, 8192)              4202496   \n",
            "                                                                 \n",
            " reshape_1 (Reshape)         (None, 8, 8, 128)         0         \n",
            "                                                                 \n",
            " conv2d_transpose_4 (Conv2DT  (None, 16, 16, 128)      147584    \n",
            " ranspose)                                                       \n",
            "                                                                 \n",
            " batch_normalization_9 (Batc  (None, 16, 16, 128)      512       \n",
            " hNormalization)                                                 \n",
            "                                                                 \n",
            " conv2d_transpose_5 (Conv2DT  (None, 32, 32, 64)       73792     \n",
            " ranspose)                                                       \n",
            "                                                                 \n",
            " batch_normalization_10 (Bat  (None, 32, 32, 64)       256       \n",
            " chNormalization)                                                \n",
            "                                                                 \n",
            " conv2d_transpose_6 (Conv2DT  (None, 64, 64, 32)       18464     \n",
            " ranspose)                                                       \n",
            "                                                                 \n",
            " batch_normalization_11 (Bat  (None, 64, 64, 32)       128       \n",
            " chNormalization)                                                \n",
            "                                                                 \n",
            " conv2d_transpose_7 (Conv2DT  (None, 64, 64, 3)        867       \n",
            " ranspose)                                                       \n",
            "                                                                 \n",
            "=================================================================\n",
            "Total params: 4,444,099\n",
            "Trainable params: 4,443,651\n",
            "Non-trainable params: 448\n",
            "_________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "N6IwN5vlAb5w"
      },
      "source": [
        "## Train the Model\n",
        "\n",
        "You will now configure the model for training. We defined some losses, the optimizer, and the loss metric below but you can experiment with others if you like.\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dHPwSmZFnQ_2"
      },
      "source": [
        "optimizer = tf.keras.optimizers.Adam(learning_rate=0.002)\n",
        "loss_metric = tf.keras.metrics.Mean()\n",
        "mse_loss = tf.keras.losses.MeanSquaredError()\n",
        "bce_loss = tf.keras.losses.BinaryCrossentropy()"
      ],
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AWRzFxYkAvXH"
      },
      "source": [
        "You will generate 16 images in a 4x4 grid to show\n",
        "progress of image generation. We've defined a utility function for that below."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "DGe445j0nTmf"
      },
      "source": [
        "def generate_and_save_images(model, epoch, step, test_input):\n",
        "  \"\"\"Helper function to plot our 16 images\n",
        "\n",
        "  Args:\n",
        "\n",
        "  model -- the decoder model\n",
        "  epoch -- current epoch number during training\n",
        "  step -- current step number during training\n",
        "  test_input -- random tensor with shape (16, LATENT_DIM)\n",
        "  \"\"\"\n",
        "  predictions = model.predict(test_input)\n",
        "\n",
        "  fig = plt.figure(figsize=(4,4))\n",
        "\n",
        "  for i in range(predictions.shape[0]):\n",
        "      plt.subplot(4, 4, i+1)\n",
        "      img = predictions[i, :, :, :] * 255\n",
        "      img = img.astype('int32')\n",
        "      plt.imshow(img)\n",
        "      plt.axis('off')\n",
        "\n",
        "  # tight_layout minimizes the overlap between 2 sub-plots\n",
        "  fig.suptitle(\"epoch: {}, step: {}\".format(epoch, step))\n",
        "  plt.savefig('image_at_epoch_{:04d}_step{:04d}.png'.format(epoch, step))\n",
        "  plt.show()"
      ],
      "execution_count": 21,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ZgJfazr6A_py"
      },
      "source": [
        "You can now start the training loop. You are asked to select the number of epochs and to complete the subection on updating the weights. The general steps are:\n",
        "\n",
        "* feed a training batch to the VAE model\n",
        "* compute the reconstruction loss (hint: use the **mse_loss** defined above instead of `bce_loss` in the ungraded lab, then multiply by the flattened dimensions of the image (i.e. 64 x 64 x 3)\n",
        "* add the KLD regularization loss to the total loss (you can access the `losses` property of the `vae` model)\n",
        "* get the gradients\n",
        "* use the optimizer to update the weights\n",
        "\n",
        "\n",
        "When training your VAE, you might notice that there’s not a lot of variation in the faces. But don’t let that deter you! We’ll test based on how well it does in reconstructing the original faces, and not how well it does in creating new faces.\n",
        "\n",
        "The training will also take a long time (more than 30 minutes) and that is to be expected. If you used the mean loss metric suggested above, train the model until that is down to around 320 before submitting.\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "hvL1bHXJnajM",
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 329
        },
        "outputId": "f748a583-aa35-4ca4-83a9-0f1916ca8e34"
      },
      "source": [
        "# Training loop. Display generated images each epoch\n",
        "\n",
        "### START CODE HERE ###\n",
        "epochs = 50\n",
        "### END CODE HERE ###\n",
        "\n",
        "random_vector_for_generation = tf.random.normal(shape=[16, LATENT_DIM])\n",
        "generate_and_save_images(decoder, 0, 0, random_vector_for_generation)\n",
        "\n",
        "for epoch in range(epochs):\n",
        "  print('Start of epoch %d' % (epoch,))\n",
        "\n",
        "  # Iterate over the batches of the dataset.\n",
        "  for step, x_batch_train in enumerate(training_dataset):\n",
        "    with tf.GradientTape() as tape:\n",
        "      ### START CODE HERE ### \n",
        "      predictions = vae(x_batch_train)\n",
        "      flattened_inputs = tf.reshape(x_batch_train, shape=[-1])\n",
        "      flattened_outputs = tf.reshape(predictions, shape=[-1])\n",
        "      loss = mse_loss(flattened_inputs, flattened_outputs) * 784\n",
        "      # Compute reconstruction loss\n",
        "      loss += sum(vae.losses)   \n",
        "        \n",
        "\n",
        "    grads = tape.gradient(loss, vae.trainable_weights)\n",
        "    optimizer.apply_gradients(zip(grads, vae.trainable_weights))\n",
        "    ### END CODE HERE\n",
        "    \n",
        "    loss_metric(loss)\n",
        "\n",
        "    if step % 10 == 0:\n",
        "      display.clear_output(wait=False)    \n",
        "      generate_and_save_images(decoder, epoch, step, random_vector_for_generation)\n",
        "    print('Epoch: %s step: %s mean loss = %s' % (epoch, step, loss_metric.result().numpy()))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 288x288 with 16 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          }
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch: 43 step: 20 mean loss = 31.213568\n",
            "Epoch: 43 step: 21 mean loss = 31.20966\n",
            "Epoch: 43 step: 22 mean loss = 31.205063\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "N5wfzGfABny6"
      },
      "source": [
        "# Plot Reconstructed Images\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BnQQlWZHaj90"
      },
      "source": [
        "As mentioned, your model will be graded on how well it is able to reconstruct images (not generate new ones). You can get a glimpse of how it is doing with the code block below. It feeds in a batch from the test set and plots a row of input (top) and output (bottom) images. Don't worry if the outputs are a blurry. It will look something like below:\n",
        "\n",
        "<img src=\"https://drive.google.com/uc?export=view&id=1OPMbZOxX9fx8tK6CGVbrMaQdgyOiQJIC\" width=\"75%\" height=\"60%\"/>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "TfIbqTIKSXEe"
      },
      "source": [
        "test_dataset = validation_dataset.take(1)\n",
        "output_samples = []\n",
        "\n",
        "for input_image in tfds.as_numpy(test_dataset):\n",
        "      output_samples = input_image\n",
        "\n",
        "idxs = np.random.choice(64, size=10)\n",
        "\n",
        "vae_predicted = vae.predict(test_dataset)\n",
        "display_results(output_samples[idxs], vae_predicted[idxs])"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9YKUOCA5BtAA"
      },
      "source": [
        "# Plot Generated Images\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ylxL9z15ctsy"
      },
      "source": [
        "Using the default parameters, it can take a long time to train your model well enough to generate good fake anime faces. In case you decide to experiment, we provided the code block below to display an 8x8 gallery of fake data generated from your model. Here is a sample gallery generated after 50 epochs.\n",
        "\n",
        "<img src=\"https://drive.google.com/uc?export=view&id=1QwElgfg5TY6vCgI1FK6vdI8Bo6UZKfuX\" width=\"75%\" height=\"60%\"/>"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "zCpTybvGSS6L"
      },
      "source": [
        "def plot_images(rows, cols, images, title):\n",
        "    '''Displays images in a grid.'''\n",
        "    grid = np.zeros(shape=(rows*64, cols*64, 3))\n",
        "    for row in range(rows):\n",
        "        for col in range(cols):\n",
        "            grid[row*64:(row+1)*64, col*64:(col+1)*64, :] = images[row*cols + col]\n",
        "\n",
        "    plt.figure(figsize=(12,12))       \n",
        "    plt.imshow(grid)\n",
        "    plt.title(title)\n",
        "    plt.show()\n",
        "\n",
        "# initialize random inputs\n",
        "test_vector_for_generation = tf.random.normal(shape=[64, LATENT_DIM])\n",
        "\n",
        "# get predictions from the decoder model\n",
        "predictions= decoder.predict(test_vector_for_generation)\n",
        "\n",
        "# plot the predictions\n",
        "plot_images(8,8,predictions,'Generated Images')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "J4IixoasCfoR"
      },
      "source": [
        "### Save the Model\n",
        "\n",
        "Once your satisfied with the results, please save and download the model. Afterwards, please go back to the Coursera submission portal to upload your h5 file to the autograder."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "A9E8qwDAVMPs"
      },
      "source": [
        "vae.save(\"anime.h5\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# vae.evaluate(validation_dataset)\n",
        "vae.save(\"/tmp/anime.h5\")"
      ],
      "metadata": {
        "id": "cMsRU6uXQg1k"
      },
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        ""
      ],
      "metadata": {
        "id": "th7OjVCtQovW"
      },
      "execution_count": null,
      "outputs": []
    }
  ]
}