sketch_to_photo.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "name": "sketch-to-photo.ipynb",
      "provenance": [],
      "collapsed_sections": [
        "WVNpSZ82dRtr",
        "lgloWW1pdYBo",
        "3eKtodqwdu8L",
        "r0PeVLLwlF5Q",
        "eOsktznvnbAn",
        "sSTemp1iwt6K",
        "jFeM1x5ywz0z",
        "kVliJmOKjPRR",
        "kFW18ep-jHcr",
        "QgKn98431JUn",
        "pdyU38I_jLQw",
        "HRcary84w5oM",
        "Ytu17F_Pw__h",
        "k0HvbdS8Cs9k",
        "MRYZuwXLFmvE",
        "6Sij7bY3wft7",
        "DjTC9-tDwkoV"
      ],
      "toc_visible": true,
      "machine_shape": "hm"
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WVNpSZ82dRtr"
      },
      "source": [
        "# Imports"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2sT1s1APDJBE"
      },
      "source": [
        "import scipy\r\n",
        "import tensorflow as tf\r\n",
        "import os\r\n",
        "import sys\r\n",
        "import time\r\n",
        "import numpy as np\r\n",
        "from numpy import cov\r\n",
        "from numpy import trace\r\n",
        "from numpy import iscomplexobj\r\n",
        "from numpy import asarray\r\n",
        "import glob\r\n",
        "from tqdm import tqdm\r\n",
        "import cv2 as cv\r\n",
        "from IPython import display\r\n",
        "from google.colab import files\r\n",
        "import tensorflow as tf\r\n",
        "from matplotlib import pyplot as plt\r\n",
        "from tensorflow.keras import Sequential, Model\r\n",
        "from keras.optimizers import Adam\r\n",
        "from tensorflow.keras.layers import \\\r\n",
        "            Input, Dense, Reshape, Flatten, Dropout, \\\r\n",
        "            Concatenate, concatenate, BatchNormalization, Activation, \\\r\n",
        "            ZeroPadding2D, LeakyReLU,ReLU,UpSampling2D, \\\r\n",
        "            Conv2D,Conv2DTranspose\r\n",
        "from scipy.linalg import sqrtm\r\n",
        "from keras.applications.inception_v3 import InceptionV3\r\n",
        "from keras.applications.inception_v3 import preprocess_input\r\n",
        "from keras.applications.xception import Xception\r\n",
        "from keras.applications.mobilenet_v2 import MobileNetV2\r\n",
        "from keras.applications.resnet_v2 import ResNet50V2\r\n",
        "from skimage.transform import resize"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "lgloWW1pdYBo"
      },
      "source": [
        "# Download and Inspect Dataset"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Wwre9Bgqel5j"
      },
      "source": [
        "The data is downloaded from the CUHK Face Sketch Database (CUFS) website and save in **sketch_to_photo** folder.\r\n",
        "The folder structure inside is the following:\r\n",
        "*   train\r\n",
        "> * photos\r\n",
        "> * sketches\r\n",
        "*   test\r\n",
        "> * photos\r\n",
        "> * sketches\r\n",
        "\r\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "rmebSLn4N1kf"
      },
      "source": [
        "path = '/content/sketch_to_photo'\r\n",
        "train_sketch_url = 'http://mmlab.ie.cuhk.edu.hk/archive/sketchdatabase/CUHK/training_88/Cropped_Images/CUHK_training_cropped_sketches.zip'\r\n",
        "train_photo_url = 'http://mmlab.ie.cuhk.edu.hk/archive/sketchdatabase/CUHK/training_88/Cropped_Images/CUHK_training_cropped_photos.zip'\r\n",
        "test_sketch_url = 'http://mmlab.ie.cuhk.edu.hk/archive/sketchdatabase/CUHK/testing_100/Cropped_Images/CUHK_testing_cropped_sketches.zip'\r\n",
        "test_photo_url = 'http://mmlab.ie.cuhk.edu.hk/archive/sketchdatabase/CUHK/testing_100/Cropped_Images/CUHK_testing_cropped_photos.zip'"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "F_ZteRapRFsJ",
        "outputId": "5815e9cb-5358-4fd2-df0c-c90aff87669d"
      },
      "source": [
        "!mkdir sketch_to_photo"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "mkdir: cannot create directory ‘sketch_to_photo’: File exists\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "lRcIqhcBUlJk"
      },
      "source": [
        "# cd /content/sketch_to_photo/"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "K9Cy_yWHTRYw"
      },
      "source": [
        "path_to_train_sketch_zip = tf.keras.utils.get_file(path+'/train/CUHK_training_cropped_sketches.zip',\r\n",
        "                                      origin=train_sketch_url,archive_format='zip',\r\n",
        "                                       cache_subdir=path+'/train',\r\n",
        "                                      extract=True)\r\n",
        "\r\n",
        "path_to_train_photo_zip = tf.keras.utils.get_file(path+'/train/CUHK_training_cropped_photos.zip',\r\n",
        "                                      origin=train_photo_url,archive_format='zip',\r\n",
        "                                       cache_subdir=path+'/train',\r\n",
        "                                      extract=True)\r\n",
        "\r\n",
        "path_to_test_sketch_zip = tf.keras.utils.get_file(path+'/test/CUHK_testing_cropped_sketches.zip',\r\n",
        "                                       cache_subdir=path+'/test',\r\n",
        "                                      origin=test_sketch_url,\r\n",
        "                                      extract=True)\r\n",
        "\r\n",
        "path_to_test_photo_zip = tf.keras.utils.get_file(path+'/test/CUHK_testing_cropped_photos.zip',\r\n",
        "                                      origin=test_photo_url,\r\n",
        "                                       cache_subdir=path+'/test',\r\n",
        "                                      extract=True)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AHJASt_eVUFz"
      },
      "source": [
        "path_to_train_sketch = os.path.join(os.path.dirname(path_to_train_sketch_zip), 'sketches/')\r\n",
        "path_to_train_photo = os.path.join(os.path.dirname(path_to_train_photo_zip), 'photos/')\r\n",
        "path_to_test_sketch = os.path.join(os.path.dirname(path_to_test_sketch_zip), 'sketches/')\r\n",
        "path_to_test_photo = os.path.join(os.path.dirname(path_to_test_photo_zip), 'photos/')"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o3Sq3jMDYo6A"
      },
      "source": [
        "def load_sketch_photo(sketch_image_file, original_image):\r\n",
        "  sketch = tf.io.read_file(sketch_image_file)\r\n",
        "  sketch = tf.image.decode_jpeg(sketch,channels=3)\r\n",
        "\r\n",
        "  photo = tf.io.read_file(original_image)\r\n",
        "  photo = tf.image.decode_jpeg(photo,channels=3)\r\n",
        "  \r\n",
        "  sketch = tf.cast(sketch, tf.float32)\r\n",
        "  photo = tf.cast(photo, tf.float32)\r\n",
        "\r\n",
        "  return sketch, photo"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dgT6gqFBUZEv",
        "outputId": "8e9cf5f0-e21e-4868-c791-6ca4718fb1a8"
      },
      "source": [
        "example_sketch, example_photo = load_sketch_photo(path_to_train_sketch+'F2-005-01-sz1.jpg', path_to_train_photo+'f-005-01.jpg')\r\n",
        "fig, (ax1, ax2) = plt.subplots(1, 2)\r\n",
        "ax1.imshow((example_sketch/127.5-1) * 0.5 + 0.5)\r\n",
        "ax2.imshow((example_photo/127.5-1) * 0.5 + 0.5)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<matplotlib.image.AxesImage at 0x7ff0212d0358>"
            ]
          },
          "metadata": {
            "tags": []
          },
          "execution_count": 220
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9PeM97drbG83",
        "outputId": "9c2461be-caf2-43a8-de1d-c60b81731742"
      },
      "source": [
        "print(example_sketch.shape)\r\n",
        "print(example_photo.shape)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "(250, 200, 3)\n",
            "(250, 200, 3)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "3eKtodqwdu8L"
      },
      "source": [
        "# Load Dataset"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "KXoYJoyEjFQo"
      },
      "source": [
        "def load_image(image_file):\r\n",
        "  image = tf.io.read_file(image_file)\r\n",
        "  image = tf.image.decode_jpeg(image,channels=3)  \r\n",
        "  image = tf.cast(image, tf.float32)\r\n",
        "  return image"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "jdlJkrMWdzy9"
      },
      "source": [
        "train_sketch = glob.glob(path_to_train_sketch+\"/*\")\r\n",
        "train_photo = glob.glob(path_to_train_photo+\"/*\")\r\n",
        "test_sketch = glob.glob(path_to_test_sketch+\"/*\")\r\n",
        "test_photo = glob.glob(path_to_test_photo+\"/*\")"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "wyG5bTcIeOAu"
      },
      "source": [
        "train_sketch = sorted(train_sketch)\r\n",
        "train_photo = sorted(train_photo)\r\n",
        "test_sketch = sorted(test_sketch)\r\n",
        "test_photo = sorted(test_photo)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qUhiEqqte9pw"
      },
      "source": [
        "train_data_sketch = tf.data.Dataset.list_files(train_sketch, shuffle=False).map(load_image)\r\n",
        "train_data_photo = tf.data.Dataset.list_files(train_photo, shuffle=False).map(load_image)\r\n",
        "test_data_sketch = tf.data.Dataset.list_files(test_sketch, shuffle=False).map(load_image)\r\n",
        "test_data_photo = tf.data.Dataset.list_files(test_photo, shuffle=False).map(load_image)\r\n",
        "\r\n",
        "train_dataset = tf.data.Dataset.zip((train_data_sketch, train_data_photo))\r\n",
        "test_dataset = tf.data.Dataset.zip((test_data_sketch, test_data_photo))"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "r0PeVLLwlF5Q"
      },
      "source": [
        "# Preprocessing"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6_m14-TSgOQc"
      },
      "source": [
        "The images are resized to (256*256) and pixels values normalized between [-1,1].\r\n",
        "\r\n",
        "Data augmentation is applied using some transformations in the images as random noise and flipping."
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "g-fheKNEjA57"
      },
      "source": [
        "def resize_image(input_image, real_image, height, width):\r\n",
        "  input_image = tf.image.resize(input_image, [height, width])\r\n",
        "  real_image = tf.image.resize(real_image, [height, width])\r\n",
        "\r\n",
        "  return input_image, real_image\r\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "w3xBcZ8JmETa"
      },
      "source": [
        "def normalize(input_image, real_image):\r\n",
        "  if not (tf.keras.backend.max(input_image))<=1:\r\n",
        "    input_image = (input_image / 127.5) - 1\r\n",
        "    # input_image = (input_image / 255)\r\n",
        "  if not (tf.keras.backend.max(real_image))<=1:\r\n",
        "    real_image = (real_image / 127.5) - 1\r\n",
        "    # real_image = (real_image / 255)\r\n",
        "\r\n",
        "  return input_image, real_image"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "FcZsMy8EnA1x"
      },
      "source": [
        "def rgb_to_gray(input_image, real_image):\r\n",
        "  # input_image = tf.image.rgb_to_grayscale(input_image)\r\n",
        "  real_image = tf.image.rgb_to_grayscale(real_image)\r\n",
        "\r\n",
        "  return input_image, real_image"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "2jDdlZ5loGIK"
      },
      "source": [
        "def saturation(input_image, real_image, lower, upper, seed):\r\n",
        "  # input_image = tf.image.random_saturation(input_image, lower, upper, seed=seed)\r\n",
        "  real_image = tf.image.random_saturation(real_image, lower, upper, seed=seed)\r\n",
        "\r\n",
        "  return input_image, real_image"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "BsCu55OUoVur"
      },
      "source": [
        "def brightness(input_image, real_image, delta, seed):\r\n",
        "  input_image = tf.image.random_brightness(input_image, delta, seed=seed)\r\n",
        "  real_image = tf.image.random_brightness(real_image, delta, seed=seed)\r\n",
        "\r\n",
        "  return input_image, real_image"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "M9fd-7hfPupr"
      },
      "source": [
        "def noise(input_image, real_image, mean, stddev):\r\n",
        "  # input_image = input_image + tf.random.normal(tf.shape(input_image), mean=mean, stddev=stddev)\r\n",
        "  # input_image = tf.image.convert_image_dtype(input_image, tf.float32)\r\n",
        "  real_image = real_image + tf.random.normal(tf.shape(real_image), mean=mean, stddev=stddev)\r\n",
        "  real_image = tf.image.convert_image_dtype(real_image, tf.float32)\r\n",
        "\r\n",
        "  return input_image, real_image"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "tRLjJuvRzTHe"
      },
      "source": [
        "def quality(input_image, real_image, seed, min_jpeg_quality, max_jpeg_quality):\r\n",
        "  # input_image = tf.image.random_jpeg_quality(input_image, min_jpeg_quality, max_jpeg_quality, seed=seed)\r\n",
        "  # input_image = tf.image.convert_image_dtype(input_image, tf.float32)\r\n",
        "  real_image = tf.image.random_jpeg_quality(real_image, min_jpeg_quality, max_jpeg_quality, seed=seed)\r\n",
        "  real_image = tf.image.convert_image_dtype(real_image, tf.float32)\r\n",
        "\r\n",
        "  return input_image, real_image"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ko4hsHD_mGS1"
      },
      "source": [
        "def transformations(input_image, real_image, training_data=True):\r\n",
        "  seed = 1\r\n",
        "\r\n",
        "  input_image, real_image = resize_image(input_image, real_image, 256, 256)\r\n",
        "  input_image, real_image = normalize(input_image, real_image)\r\n",
        "\r\n",
        "  if(training_data):\r\n",
        "    # input_image, real_image = saturation(input_image, real_image, 0.95, 1, seed)\r\n",
        "    input_image, real_image = brightness(input_image, real_image, 0.2, seed)\r\n",
        "    # input_image, real_image = quality(input_image, real_image, seed, 95, 100)\r\n",
        "    input_image, real_image = noise(input_image, real_image, 0.1, 0.1)\r\n",
        "    if tf.random.uniform(()) > 0.5:\r\n",
        "      input_image = tf.image.flip_left_right(input_image)\r\n",
        "      real_image = tf.image.flip_left_right(real_image)\r\n",
        "    if tf.random.uniform(()) > 0.5:\r\n",
        "      input_image = tf.image.flip_up_down(input_image)\r\n",
        "      real_image = tf.image.flip_up_down(real_image)\r\n",
        "\r\n",
        "\r\n",
        "  return input_image, real_image"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iMq4mz0FmPrA",
        "outputId": "1d1034d0-7c42-47e8-979c-b79df37a0bb9"
      },
      "source": [
        "plt.figure(figsize=(6, 6))\r\n",
        "for i in range(2):\r\n",
        "  transformed_example_sketch, transformed_example_photo = transformations(example_sketch, example_photo)\r\n",
        "  fig, (ax1, ax2) = plt.subplots(1, 2)\r\n",
        "  ax1.axis('off')\r\n",
        "  ax2.axis('off')\r\n",
        "  ax1.imshow(transformed_example_sketch * 0.5 + 0.5)\r\n",
        "  ax2.imshow(transformed_example_photo * 0.5 + 0.5)\r\n",
        "plt.show()"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n",
            "Clipping input data to the valid range for imshow with RGB data ([0..1] for floats or [0..255] for integers).\n"
          ],
          "name": "stderr"
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x432 with 0 Axes>"
            ]
          },
          "metadata": {
            "tags": []
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        },
        {
          "output_type": "display_data",
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 2 Axes>"
            ]
          },
          "metadata": {
            "tags": [],
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "JVnEe0zJtUf8"
      },
      "source": [
        "BATCH_SIZE = 1\r\n",
        "BUFFER_SIZE = 400\r\n",
        "\r\n",
        "apply_train = lambda x, y: transformations(x, y)\r\n",
        "apply_test = lambda x, y: transformations(x, y, training_data=False)\r\n",
        "train_dataset = train_dataset.map(apply_test).concatenate(train_dataset.map(apply_train))\r\n",
        "test_dataset = test_dataset.map(apply_test)\r\n",
        "\r\n",
        "train_dataset = train_dataset.shuffle(BUFFER_SIZE)\r\n",
        "train_dataset = train_dataset.batch(BATCH_SIZE)\r\n",
        "test_dataset = test_dataset.batch(BATCH_SIZE)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "eOsktznvnbAn"
      },
      "source": [
        "# Generative Adversarial Network"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-j2AX8lahZpE"
      },
      "source": [
        "Three generators model are defined using the one from Pix2Pix as starting point.\r\n",
        "\r\n",
        "https://www.tensorflow.org/tutorials/generative/pix2pix\r\n",
        "\r\n",
        "\r\n",
        "*   Pix2Pix Generator\r\n",
        "*   Pix2Pix with Xception\r\n",
        "*   Pix2Pix with MobileNet\r\n",
        "*   Pix2Pix with Resnet50\r\n",
        "\r\n",
        "As discriminator is used the same as in Pix2Pix\r\n",
        "\r\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "sSTemp1iwt6K"
      },
      "source": [
        "## Utils"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "IKjHQVMIE8xE"
      },
      "source": [
        "class ModelUtils:\r\n",
        "\r\n",
        "  @staticmethod\r\n",
        "  def generator_loss(disc_generated_output, gen_output, target, loss_object, LAMBDA=100):\r\n",
        "    gan_loss = loss_object(tf.ones_like(disc_generated_output), disc_generated_output)\r\n",
        "    # mean absolute error\r\n",
        "    l1_loss = tf.reduce_mean(tf.abs(target - gen_output))\r\n",
        "    total_gen_loss = gan_loss + (LAMBDA * l1_loss)\r\n",
        "    return total_gen_loss, gan_loss, l1_loss\r\n",
        "\r\n",
        "  @staticmethod\r\n",
        "  def discriminator_loss(disc_real_output, disc_generated_output, loss_object):\r\n",
        "    real_loss = loss_object(tf.ones_like(disc_real_output), disc_real_output)\r\n",
        "    generated_loss = loss_object(tf.zeros_like(disc_generated_output), disc_generated_output)\r\n",
        "    total_disc_loss = real_loss + generated_loss\r\n",
        "    return total_disc_loss\r\n",
        "\r\n",
        "  @staticmethod\r\n",
        "  def generate_images(generator, test_input, tar, training=True, show_images=False):\r\n",
        "    prediction = generator(test_input, training=training)\r\n",
        "    \r\n",
        "    if show_images:\r\n",
        "      plt.figure(figsize=(15,15))\r\n",
        "\r\n",
        "      display_list = [test_input[0], tar[0], prediction[0]]\r\n",
        "      title = ['Input Image', 'Ground Truth', 'Predicted Image']\r\n",
        "\r\n",
        "      for i in range(3):\r\n",
        "        plt.subplot(1, 3, i+1)\r\n",
        "        plt.title(title[i])\r\n",
        "        # getting the pixel values between [0, 1] to plot it.\r\n",
        "        plt.imshow(display_list[i] * 0.5 + 0.5)\r\n",
        "        plt.axis('off')\r\n",
        "      plt.show()\r\n",
        "\r\n",
        "  @staticmethod\r\n",
        "  def save_model(model, path, name, model_type=\"generator\"):\r\n",
        "    tf.saved_model.save(model, os.path.join(path, name, model_type))\r\n",
        "\r\n",
        "  @staticmethod\r\n",
        "  def load_model(path, name, model_type=\"generator\"):\r\n",
        "    return tf.saved_model.load(os.path.join(path, name, model_type))\r\n",
        "\r\n",
        "  @staticmethod\r\n",
        "  def downsample(filters, size, apply_batchnorm=True):\r\n",
        "    initializer = tf.random_normal_initializer(0., 0.02) \r\n",
        "    result = Sequential()\r\n",
        "    result.add(Conv2D(filters, size, strides=2, padding='same',\r\n",
        "                              kernel_initializer=initializer, use_bias=False))\r\n",
        "    if apply_batchnorm:\r\n",
        "      result.add(BatchNormalization())\r\n",
        "    result.add(LeakyReLU())\r\n",
        "    return result\r\n",
        "\r\n",
        "  @staticmethod\r\n",
        "  def upsample(filters, size, apply_dropout=False):\r\n",
        "    initializer = tf.random_normal_initializer(0., 0.02)\r\n",
        "    result = Sequential()\r\n",
        "    result.add(Conv2DTranspose(filters, size, strides=2,\r\n",
        "                                      padding='same',\r\n",
        "                                      kernel_initializer=initializer,\r\n",
        "                                      use_bias=False))\r\n",
        "    result.add(BatchNormalization())\r\n",
        "    if apply_dropout:\r\n",
        "        result.add(Dropout(0.5))\r\n",
        "    result.add(ReLU())\r\n",
        "    return result\r\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jFeM1x5ywz0z"
      },
      "source": [
        "## Generator"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kVliJmOKjPRR"
      },
      "source": [
        "### SimpleGenerator"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "WKq3dv0ojSXX"
      },
      "source": [
        "class SimpleGenerator:\r\n",
        "\r\n",
        "  @staticmethod\r\n",
        "  def build_model(input_shape, output_channels=3, initializer=tf.random_normal_initializer(0., 0.02)):\r\n",
        "    inputs = Input(shape=input_shape)\r\n",
        "\r\n",
        "    \r\n",
        "    down_stack = [\r\n",
        "      ModelUtils.downsample(64, 4, apply_batchnorm=False), # (bs, 128, 128, 64)\r\n",
        "      ModelUtils.downsample(128, 4), # (bs, 64, 64, 128)\r\n",
        "      ModelUtils.downsample(256, 4), # (bs, 32, 32, 256)\r\n",
        "      ModelUtils.downsample(512, 4), # (bs, 16, 16, 512)\r\n",
        "      ModelUtils.downsample(512, 4), # (bs, 8, 8, 512)\r\n",
        "      ModelUtils.downsample(512, 4), # (bs, 4, 4, 512)\r\n",
        "      ModelUtils.downsample(512, 4), # (bs, 2, 2, 512)\r\n",
        "      ModelUtils.downsample(512, 4), # (bs, 1, 1, 512)\r\n",
        "    ]\r\n",
        "    up_stack = [\r\n",
        "      ModelUtils.upsample(512, 4, apply_dropout=True), # (bs, 2, 2, 1024)\r\n",
        "      ModelUtils.upsample(512, 4, apply_dropout=True), # (bs, 4, 4, 1024)\r\n",
        "      ModelUtils.upsample(512, 4, apply_dropout=True), # (bs, 8, 8, 1024)\r\n",
        "      ModelUtils.upsample(512, 4), # (bs, 16, 16, 1024)\r\n",
        "      ModelUtils.upsample(256, 4), # (bs, 32, 32, 512)\r\n",
        "      ModelUtils.upsample(128, 4), # (bs, 64, 64, 256)\r\n",
        "      ModelUtils.upsample(64, 4), # (bs, 128, 128, 128)\r\n",
        "    ]\r\n",
        "\r\n",
        "    last = Conv2DTranspose(output_channels, 4,\r\n",
        "                                          strides=2,\r\n",
        "                                          padding='same',\r\n",
        "                                          kernel_initializer=initializer,\r\n",
        "                                          activation='tanh') # (bs, 256, 256, 3)\r\n",
        "\r\n",
        "    x = inputs\r\n",
        "    # Downsampling through the model\r\n",
        "    skips = []\r\n",
        "    for down in down_stack:\r\n",
        "      x = down(x)\r\n",
        "      skips.append(x)\r\n",
        "    skips = reversed(skips[:-1])\r\n",
        "\r\n",
        "    # Upsampling and establishing the skip connections\r\n",
        "    for up, skip in zip(up_stack, skips):\r\n",
        "      x = up(x)\r\n",
        "      x = Concatenate()([x, skip])\r\n",
        "    x = last(x)\r\n",
        "\r\n",
        "    return Model(inputs=inputs, outputs=x)\r\n",
        "\r\n",
        "\r\n",
        "class Discriminator:\r\n",
        "\r\n",
        "  @staticmethod\r\n",
        "  def build_model_boh(input_shape, initializer=tf.random_normal_initializer(0.0, 0.02)):\r\n",
        "    inputs = Input(shape=input_shape)\r\n",
        "    targets = Input(shape=input_shape)\r\n",
        "\r\n",
        "    x = concatenate([inputs, targets])\r\n",
        "\r\n",
        "\r\n",
        "\r\n",
        "  @staticmethod\r\n",
        "  def build_model_patchgan(input_shape, initializer=tf.random_normal_initializer(0., 0.02)):\r\n",
        "\r\n",
        "    inp = Input(shape=input_shape, name='input_image')\r\n",
        "    tar = Input(shape=input_shape, name='target_image')\r\n",
        "\r\n",
        "    x = concatenate([inp, tar]) # (bs, 256, 256, channels*2)\r\n",
        "\r\n",
        "    down1 = ModelUtils.downsample(64, 4, False)(x) # (bs, 128, 128, 64)\r\n",
        "    down2 = ModelUtils.downsample(128, 4)(down1) # (bs, 64, 64, 128)\r\n",
        "    down3 = ModelUtils.downsample(256, 4)(down2) # (bs, 32, 32, 256)\r\n",
        "\r\n",
        "    zero_pad1 = ZeroPadding2D()(down3) # (bs, 34, 34, 256)\r\n",
        "    conv = Conv2D(512, 4, strides=1,\r\n",
        "                                  kernel_initializer=initializer,\r\n",
        "                                  use_bias=False)(zero_pad1) # (bs, 31, 31, 512)\r\n",
        "\r\n",
        "    batchnorm1 = BatchNormalization()(conv)\r\n",
        "    leaky_relu = LeakyReLU()(batchnorm1)\r\n",
        "    zero_pad2 = ZeroPadding2D()(leaky_relu) # (bs, 33, 33, 512)\r\n",
        "    last = Conv2D(1, 4, strides=1, kernel_initializer=initializer)(zero_pad2) # (bs, 30, 30, 1)\r\n",
        "    last = Flatten()(last)\r\n",
        "    last = Dense(1 ,activation=\"sigmoid\")(last)\r\n",
        "\r\n",
        "    return Model(inputs=[inp, tar], outputs=last)\r\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kFW18ep-jHcr"
      },
      "source": [
        "### XceptionGenerator"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "q6sI8mjyE_zH"
      },
      "source": [
        "class XceptionGenerator:\r\n",
        "\r\n",
        "  @staticmethod\r\n",
        "  def build_model(input_shape, output_channels=3, initializer=tf.random_normal_initializer(0., 0.02)):\r\n",
        "\r\n",
        "    xception = Xception(weights='imagenet',include_top=False, input_shape=(256, 256, 3))\r\n",
        "\r\n",
        "    layer_names = [\r\n",
        "          'block13_sepconv2_bn',\r\n",
        "          'block13_pool'\r\n",
        "    ]\r\n",
        "    layers = [xception.get_layer(name).output for name in layer_names]\r\n",
        "    down_stack = Model(inputs=xception.input, outputs=layers)\r\n",
        "    down_stack.trainable = False\r\n",
        "\r\n",
        "    up_stack = [\r\n",
        "          ModelUtils.upsample(512, 4, apply_dropout=True),  # (bs, 8, 8, 1024)\r\n",
        "          ModelUtils.upsample(512, 4, apply_dropout=True)   # (bs, 16, 16, 1024)\r\n",
        "    ]\r\n",
        "\r\n",
        "    inputs = Input(shape=input_shape)\r\n",
        "    last = Conv2DTranspose(output_channels, 4,\r\n",
        "                                          strides=2,\r\n",
        "                                          padding='same',\r\n",
        "                                          kernel_initializer=initializer,\r\n",
        "                                          activation='tanh') # (bs, 256, 256, 3)\r\n",
        "\r\n",
        "    x = inputs\r\n",
        "    skips = down_stack(x)\r\n",
        "    x = skips[-1]\r\n",
        "    skips = reversed(skips[:-1])\r\n",
        "\r\n",
        "    # Upsampling and establishing the skip connections\r\n",
        "    for up, skip in zip(up_stack, skips):\r\n",
        "      x = up(x)\r\n",
        "      x = Concatenate()([x, skip])\r\n",
        "\r\n",
        "    # x = Conv2DTranspose(512, 4, strides=2, padding='same',\r\n",
        "    #                           kernel_initializer=initializer, use_bias=False)(x)\r\n",
        "    # x = Conv2DTranspose(256, 4, strides=2, padding='same',\r\n",
        "    #                           kernel_initializer=initializer, use_bias=False)(x)\r\n",
        "    # x = Conv2DTranspose(128, 4, strides=2, padding='same',\r\n",
        "    #                           kernel_initializer=initializer, use_bias=False)(x)\r\n",
        "    x = ModelUtils.upsample(512, 4)(x)\r\n",
        "    x = ModelUtils.upsample(256, 4)(x)\r\n",
        "    x = ModelUtils.upsample(128, 4)(x)\r\n",
        "    x = last(x)\r\n",
        "\r\n",
        "    return Model(inputs=inputs, outputs=x)\r\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "QgKn98431JUn"
      },
      "source": [
        "### ResnetGenerator\r\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "1RJZhM0-1Pxi"
      },
      "source": [
        "class ResnetGenerator:\r\n",
        "\r\n",
        "  @staticmethod\r\n",
        "  def build_model(input_shape, output_channels=3, initializer=tf.random_normal_initializer(0., 0.02)):\r\n",
        "\r\n",
        "    resnet = ResNet50V2(weights='imagenet',include_top=False, input_shape=(256, 256, 3))\r\n",
        "\r\n",
        "    layer_names = [\r\n",
        "          'conv1_conv',   # (128 128 64)\r\n",
        "          'conv3_block1_preact_relu', # (32 32 256)\r\n",
        "          'conv4_block1_preact_relu', # (16 16 512)\r\n",
        "          # 'conv5_block1_2_relu', # (8 8 512)\r\n",
        "          'conv5_block3_2_relu', # (8 8 512)  \r\n",
        "    ]\r\n",
        "    layers = [resnet.get_layer(name).output for name in layer_names]\r\n",
        "    down_stack = Model(inputs=resnet.input, outputs=layers)\r\n",
        "    down_stack.trainable = False\r\n",
        "\r\n",
        "    up_stack = [\r\n",
        "          # ModelUtils.upsample(512, 4, apply_dropout=True),  # (bs, 8, 8, 1024)\r\n",
        "          ModelUtils.upsample(512, 4, apply_dropout=True),   # (bs, 16, 16, 1024)\r\n",
        "          # ModelUtils.upsample(512, 4, apply_dropout=True),   # (bs, 16, 16, 1024)\r\n",
        "          # ModelUtils.upsample(512, 4), # (bs, 16, 16, 1024)\r\n",
        "          ModelUtils.upsample(256, 4), # (bs, 32, 32, 512)\r\n",
        "          # ModelUtils.upsample(128, 4), # (bs, 64, 64, 256)\r\n",
        "          # ModelUtils.upsample(64, 4), # (bs, 128, 128, 128)\r\n",
        "    ]\r\n",
        "\r\n",
        "    inputs = Input(shape=input_shape)\r\n",
        "    # x = inputs\r\n",
        "    # down_stack = down_stack(x)\r\n",
        "    # down_stack = ModelUtils.downsample(512, 4)(down_stack[-1])\r\n",
        "    # down_stack = ModelUtils.downsample(512, 4)(down_stack)\r\n",
        "    \r\n",
        "    last = Conv2DTranspose(output_channels, 4,\r\n",
        "                                          strides=2,\r\n",
        "                                          padding='same',\r\n",
        "                                          kernel_initializer=initializer,\r\n",
        "                                          activation='tanh') # (bs, 256, 256, 3)\r\n",
        "\r\n",
        "    x = inputs\r\n",
        "    skips = down_stack(x)\r\n",
        "    x = skips[-1]\r\n",
        "    skips = reversed(skips[:-1])\r\n",
        "\r\n",
        "    # Upsampling and establishing the skip connections\r\n",
        "    # set_trace()\r\n",
        "    for up, skip in zip(up_stack, skips):\r\n",
        "      x = up(x)\r\n",
        "      x = Concatenate()([x, skip])\r\n",
        "    \r\n",
        "    x = ModelUtils.upsample(128, 4)(x)\r\n",
        "    x = ModelUtils.upsample(64, 4)(x)\r\n",
        "    x = Concatenate()([x, list(skips)[0]])\r\n",
        "    x = last(x)\r\n",
        "    return Model(inputs=inputs, outputs=x)\r\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pdyU38I_jLQw"
      },
      "source": [
        "### MobileNetGenerator"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "qCnKW939w5zf"
      },
      "source": [
        "class MobileNetGenerator:\r\n",
        "\r\n",
        "  @staticmethod\r\n",
        "  def build_model(input_shape, output_channels=3, initializer=tf.random_normal_initializer(0., 0.02)):\r\n",
        "\r\n",
        "    base_model = MobileNetV2(input_shape=[256, 256, 3], include_top=False)\r\n",
        "\r\n",
        "    # Use the activations of these layers\r\n",
        "    layer_names = [\r\n",
        "        'block_1_expand_relu',   # 64x64\r\n",
        "        'block_3_expand_relu',   # 32x32\r\n",
        "        'block_6_expand_relu',   # 16x16\r\n",
        "        'block_13_expand_relu',  # 8x8\r\n",
        "        'block_16_project',      # 4x4\r\n",
        "    ]\r\n",
        "    layers = [base_model.get_layer(name).output for name in layer_names]\r\n",
        "\r\n",
        "    # Create the feature extraction model\r\n",
        "    down_stack = Model(inputs=base_model.input, outputs=layers)\r\n",
        "    down_stack.trainable = False\r\n",
        "\r\n",
        "    up_stack = [\r\n",
        "        ModelUtils.upsample(512, 3),  # 4x4 -> 8x8\r\n",
        "        ModelUtils.upsample(256, 3),  # 8x8 -> 16x16\r\n",
        "        ModelUtils.upsample(128, 3),  # 16x16 -> 32x32\r\n",
        "        ModelUtils.upsample(64, 3),   # 32x32 -> 64x64\r\n",
        "    ] \r\n",
        "\r\n",
        "\r\n",
        "    inputs = Input(shape=input_shape)\r\n",
        "\r\n",
        "    last = Conv2DTranspose(output_channels, 4,\r\n",
        "                                          strides=2,\r\n",
        "                                          padding='same',\r\n",
        "                                          kernel_initializer=initializer,\r\n",
        "                                          activation='tanh') # (bs, 256, 256, 3)\r\n",
        "\r\n",
        "    x = inputs\r\n",
        "    skips = down_stack(x)\r\n",
        "    x = skips[-1]\r\n",
        "    skips = reversed(skips[:-1])\r\n",
        "\r\n",
        "    # Upsampling and establishing the skip connections\r\n",
        "    for up, skip in zip(up_stack, skips):\r\n",
        "      x = up(x)\r\n",
        "      x = Concatenate()([x, skip])\r\n",
        "\r\n",
        "    x = last(x)\r\n",
        "\r\n",
        "    return Model(inputs=inputs, outputs=x)\r\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "HRcary84w5oM"
      },
      "source": [
        "## Discriminator"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "7O4BYqD1w7mL"
      },
      "source": [
        "class Discriminator:\r\n",
        "  \r\n",
        "  @staticmethod\r\n",
        "  def build_model(input_shape, initializer=tf.random_normal_initializer(0., 0.02)):\r\n",
        "\r\n",
        "    inp = Input(shape=input_shape, name='input_image')\r\n",
        "    tar = Input(shape=input_shape, name='target_image')\r\n",
        "\r\n",
        "    x = concatenate([inp, tar]) # (bs, 256, 256, channels*2)\r\n",
        "\r\n",
        "    down1 = ModelUtils.downsample(64, 4, False)(x) # (bs, 128, 128, 64)\r\n",
        "    down2 = ModelUtils.downsample(128, 4)(down1) # (bs, 64, 64, 128)\r\n",
        "    down3 = ModelUtils.downsample(256, 4)(down2) # (bs, 32, 32, 256)\r\n",
        "\r\n",
        "    zero_pad1 = ZeroPadding2D()(down3) # (bs, 34, 34, 256)\r\n",
        "    conv = Conv2D(512, 4, strides=1,\r\n",
        "                                  kernel_initializer=initializer,\r\n",
        "                                  use_bias=False)(zero_pad1) # (bs, 31, 31, 512)\r\n",
        "\r\n",
        "    batchnorm1 = BatchNormalization()(conv)\r\n",
        "    leaky_relu = LeakyReLU()(batchnorm1)\r\n",
        "    zero_pad2 = ZeroPadding2D()(leaky_relu) # (bs, 33, 33, 512)\r\n",
        "    last = Conv2D(1, 4, strides=1, kernel_initializer=initializer)(zero_pad2) # (bs, 30, 30, 1)\r\n",
        "\r\n",
        "    return Model(inputs=[inp, tar], outputs=last)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ytu17F_Pw__h"
      },
      "source": [
        "## GAN"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "AfuyTVWIFCkW"
      },
      "source": [
        "class GANModels:\r\n",
        "\r\n",
        "  def __init__(self, gen_optimizer, disc_optimizer, model_name, generator, discriminator, log_dir='logs', model_dir='models', out_channels=3):\r\n",
        "    self.generator_model = generator\r\n",
        "    self.discriminator_model = discriminator\r\n",
        "    self.generator_optimizer = gen_optimizer\r\n",
        "    self.discriminator_optimizer = disc_optimizer\r\n",
        "    self.output_channels = out_channels\r\n",
        "    self.model_name = model_name\r\n",
        "    self.summary_writer = tf.summary.create_file_writer(os.path.join(log_dir, model_name))\r\n",
        "    self.log_dir = log_dir\r\n",
        "    self.model_dir = model_dir\r\n",
        "    self.total_epochs = 0\r\n",
        "    self.checkpoint = tf.train.Checkpoint(generator_optimizer=self.generator_optimizer,\r\n",
        "                                    discriminator_optimizer=self.discriminator_optimizer,\r\n",
        "                                    generator=self.generator_model,\r\n",
        "                                    discriminator=self.discriminator_model)\r\n",
        "\r\n",
        "  def get_generator_model(self):\r\n",
        "    return self.generator_model\r\n",
        "\r\n",
        "  def get_discriminator_model(self):\r\n",
        "    return self.discriminator_model\r\n",
        "\r\n",
        "  def get_optimizers(self):\r\n",
        "    return self.generator_optimizer, self.discriminator_optimizer\r\n",
        "\r\n",
        "  def get_model_name(self):\r\n",
        "    return self.model_name\r\n",
        "\r\n",
        "  def save_model(self):\r\n",
        "    tf.saved_model.save(self.generator_model, self.model_dir +\"/\"+self.model_name+\"/generator\")\r\n",
        "    tf.saved_model.save(self.discriminator_model, self.model_dir +\"/\"+self.model_name+\"/discriminator\")\r\n",
        "\r\n",
        "  def load_model(self):\r\n",
        "    self.generator_model = tf.saved_model.load(self.model_dir +\"/\"+self.model_name+\"/generator\")\r\n",
        "    self.discriminator_model = tf.saved_model.load(self.model_dir +\"/\"+self.model_name+\"/discriminator\")\r\n",
        "\r\n",
        "  @tf.function\r\n",
        "  def train_step(self, input_image, target, epoch):\r\n",
        "    loss_object = tf.keras.losses.BinaryCrossentropy(from_logits=True)\r\n",
        "    \r\n",
        "    with tf.GradientTape() as gen_tape, tf.GradientTape() as disc_tape:\r\n",
        "      gen_output = self.generator_model(input_image, training=True)\r\n",
        "      disc_real_output = self.discriminator_model([input_image, target], training=True)\r\n",
        "      disc_generated_output = self.discriminator_model([input_image, gen_output], training=True)      \r\n",
        "      gen_total_loss, gen_gan_loss, gen_l1_loss = ModelUtils.generator_loss(disc_generated_output, gen_output, target, loss_object)      \r\n",
        "      disc_loss = ModelUtils.discriminator_loss(disc_real_output, disc_generated_output, loss_object)\r\n",
        "\r\n",
        "    discriminator_gradients = disc_tape.gradient(disc_loss, self.discriminator_model.trainable_variables)\r\n",
        "    discriminator_optimizer.apply_gradients(zip(discriminator_gradients, self.discriminator_model.trainable_variables))\r\n",
        "\r\n",
        "    generator_gradients = gen_tape.gradient(gen_total_loss, self.generator_model.trainable_variables)    \r\n",
        "    generator_optimizer.apply_gradients(zip(generator_gradients, self.generator_model.trainable_variables))\r\n",
        "\r\n",
        "\r\n",
        "    with self.summary_writer.as_default():\r\n",
        "      tf.summary.scalar('gen_total_loss', gen_total_loss, step=epoch)\r\n",
        "      tf.summary.scalar('gen_gan_loss', gen_gan_loss, step=epoch)\r\n",
        "      tf.summary.scalar('gen_l1_loss', gen_l1_loss, step=epoch)\r\n",
        "      tf.summary.scalar('disc_loss', disc_loss, step=epoch)\r\n",
        "\r\n",
        "    return gen_output\r\n",
        "\r\n",
        "  \r\n",
        "  def fit(self, train_ds, epochs, test_ds, checkpoint_prefix, show_progress=False):\r\n",
        "    for epoch in tqdm(range(epochs)):\r\n",
        "\r\n",
        "      display.clear_output(wait=True)\r\n",
        "\r\n",
        "      start = time.time()\r\n",
        "      \r\n",
        "      if show_progress:\r\n",
        "        for example_input, example_target in test_ds.take(1):\r\n",
        "          ModelUtils.generate_images(self.generator_model, example_input, example_target, \r\n",
        "                                     training=False, show_images=show_progress)\r\n",
        "      print(\"Epoch: \", epoch)\r\n",
        "\r\n",
        "      for (input_image, target) in train_ds:\r\n",
        "        #print(\"Working on a batch...\")\r\n",
        "        generated = self.train_step(input_image, target, epoch)\r\n",
        "        \r\n",
        "      # saving (checkpoint) the model every  x epochs\r\n",
        "      if (epoch + 1) % 20 == 0:\r\n",
        "        self.checkpoint.save(file_prefix = checkpoint_prefix)\r\n",
        "\r\n",
        "      print('\\nTime taken for epoch {} is {} sec\\n'.format(epoch + 1, time.time()-start))\r\n",
        "\r\n",
        "    # saving (checkpoint) the model just at the end\r\n",
        "    self.checkpoint.save(file_prefix = checkpoint_prefix)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "tlVM4ZncI9og"
      },
      "source": [
        "# Create and Train GAN"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OYIvoUlFjDUX"
      },
      "source": [
        "Three GANs model versions are created and trained using the generators presented previously.\r\n",
        "\r\n",
        "*careful with running out of RAM,   :)*"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "is88YgVbFJLX"
      },
      "source": [
        "input_shape = [256,256,3]\r\n",
        "generator_optimizer = tf.keras.optimizers.Adam(2e-4, beta_1=0.5)\r\n",
        "discriminator_optimizer = tf.keras.optimizers.Adam(2e-4, beta_1=0.5)\r\n",
        "EPOCHS = 400\r\n",
        "checkpoint_prefix = \"/content/sketch_to_photo/checkpoints/\"\r\n",
        "log_dir = \"/content/sketch_to_photo/logs/\"\r\n",
        "model_dir = \"/content/sketch_to_photo/models/\""
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "VB4uAblqaF9C"
      },
      "source": [
        "Train modelResnet50"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Ki3zOPzJNPV3"
      },
      "source": [
        "generatorResnet = ResnetGenerator.build_model(input_shape)\r\n",
        "discriminatorResnet = Discriminator.build_model(input_shape)\r\n",
        "modelResnet = GANModels(generator_optimizer, discriminator_optimizer, \"modelResnet\", \r\n",
        "                           generatorResnet, discriminatorResnet, log_dir, model_dir)\r\n",
        "modelResnet.fit(train_dataset, 6, test_dataset, checkpoint_prefix, show_progress=False)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UUaMvrtOaN7D"
      },
      "source": [
        "Train modelMobileNet"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "gthg1YEGFg95"
      },
      "source": [
        "generatorMobileNet = MobileNetGenerator.build_model(input_shape)\r\n",
        "discriminatorMobileNet = Discriminator.build_model(input_shape)\r\n",
        "modelMobileNet = GANModels(generator_optimizer, discriminator_optimizer, \"modelMobileNet\", \r\n",
        "                           generatorMobileNet, discriminatorMobileNet, log_dir, model_dir)\r\n",
        "modelMobileNet.fit(train_dataset, EPOCHS, test_dataset, checkpoint_prefix, show_progress=False)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "9SXMXsCFaRRB"
      },
      "source": [
        "Train modelXception"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "xTXbIZ6BuFP4"
      },
      "source": [
        "generatorXception = XceptionGenerator.build_model(input_shape)\r\n",
        "discriminatorXception = Discriminator.build_model(input_shape)\r\n",
        "modelXception = GANModels(generator_optimizer, discriminator_optimizer, \"modelXception\", \r\n",
        "                          generatorXception, discriminatorXception, log_dir, model_dir)\r\n",
        "modelXception.fit(train_dataset, EPOCHS, test_dataset, checkpoint_prefix, show_progress=False)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "AbFD9PdwaThp"
      },
      "source": [
        "Train Pix2Pix"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "dLK1fz7J1_sy"
      },
      "source": [
        "generatorSimple = SimpleGenerator.build_model(input_shape)\r\n",
        "discriminatorSimple = Discriminator.build_model(input_shape)\r\n",
        "modelSimple = GANModels(generator_optimizer, discriminator_optimizer, \"modelSimple\", \r\n",
        "                        generatorSimple, discriminatorSimple, log_dir, model_dir)\r\n",
        "modelSimple.fit(train_dataset, EPOCHS, test_dataset, checkpoint_prefix, show_progress=False)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Dil1LktXWmEF"
      },
      "source": [
        "# tf.keras.utils.plot_model(generatorMobileNet, show_shapes=True, dpi=64)\r\n",
        "# !rm -r /content/sketch_to_photo/logs/"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WzrEmNGwNKlA"
      },
      "source": [
        "# Test"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jEMQt4zkjvGz"
      },
      "source": [
        "Some images from the test set are taken for showing some results.\r\n",
        "The model must be specified.\r\n",
        "\r\n",
        "\r\n",
        "*   modelSimple\r\n",
        "*   modelMobileNet\r\n",
        "*   modelXception\r\n",
        "*   modelResnet50\r\n",
        "\r\n",
        "\r\n",
        "\r\n",
        "```\r\n",
        "results(model, test_dataset)\r\n",
        "```\r\n",
        "\r\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "pj-28VJYMOdJ"
      },
      "source": [
        "def results(model, test_dataset):\r\n",
        "  for inp, tar in test_dataset.take(20):\r\n",
        "    ModelUtils.generate_images(model.get_generator_model(), inp, tar, False, True)\r\n",
        "\r\n",
        "results(modelResnet, test_dataset)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pUOSdWTY3Llo"
      },
      "source": [
        "# Results"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "k0HvbdS8Cs9k"
      },
      "source": [
        "## TensorBoard"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "eLWvkQzi3HV5"
      },
      "source": [
        "# Loading tensorboard\r\n",
        "%load_ext tensorboard\r\n",
        "%reload_ext tensorboard\r\n",
        "%tensorboard --logdir /content/sketch_to_photo/logs"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "MRYZuwXLFmvE"
      },
      "source": [
        "## MAE"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "P9H0oSljFqFY",
        "outputId": "fc74ede8-fd21-4cbe-e75c-4b1386b2fe87"
      },
      "source": [
        "def compute_mae(G, test_dataset):\r\n",
        "  mae = 0.0\r\n",
        "  count = 0\r\n",
        "\r\n",
        "  for input, target in test_dataset:\r\n",
        "    count += 1\r\n",
        "    generated = G(input)\r\n",
        "    mae += tf.reduce_mean(tf.abs(tf.subtract(target, generated)))\r\n",
        "  \r\n",
        "  return mae.numpy(), mae.numpy()/count\r\n",
        "\r\n",
        "# print(\"MAE modelMobileNet: \", compute_mae(modelMobileNet.get_generator_model(), test_dataset))\r\n",
        "# print(\"MAE modelXception: \", compute_mae(modelXception.get_generator_model(), test_dataset))\r\n",
        "# print(\"MAE modelSimple: \", compute_mae(modelSimple.get_generator_model(), test_dataset))\r\n",
        "print(\"MAE modelResnet: \", compute_mae(modelResnet.get_generator_model(), test_dataset))"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "MAE modelResnet:  (19.737858, 0.19737857818603516)\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LbaEsD2hC39P"
      },
      "source": [
        "## Frechet Inception Distance"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "6Sij7bY3wft7"
      },
      "source": [
        "### Functions"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "4mDaK0ynC-DF"
      },
      "source": [
        "# scale an array of images to a new size\r\n",
        "def scale_images(images, new_shape):\r\n",
        "\timages_list = list()\r\n",
        "\tfor image in images:\r\n",
        "\t\t# resize with nearest neighbor interpolation\r\n",
        "\t\tnew_image = resize(image, new_shape, 0)\r\n",
        "\t\t# store\r\n",
        "\t\timages_list.append(new_image)\r\n",
        "\treturn asarray(images_list)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "ih9OcIEGD7XP"
      },
      "source": [
        "# calculate frechet inception distance\r\n",
        "def calculate_fid(model, images1, images2):\r\n",
        "\t# calculate activations\r\n",
        "  act1 = model.predict(images1)\r\n",
        "  act2 = model.predict(images2)\r\n",
        "\t# calculate mean and covariance statistics\r\n",
        "  mu1, sigma1 = act1.mean(axis=0), cov(act1, rowvar=False)\r\n",
        "  mu2, sigma2 = act2.mean(axis=0), cov(act2, rowvar=False)\r\n",
        "\t# calculate sum squared difference between means\r\n",
        "  ssdiff = np.sum((mu1 - mu2)**2.0)\r\n",
        "\t# calculate sqrt of product between cov\r\n",
        "  covmean = sqrtm(sigma1.dot(sigma2))\r\n",
        "\t# check and correct imaginary numbers from sqrt\r\n",
        "  if iscomplexobj(covmean):\r\n",
        "    covmean = covmean.real\r\n",
        "\t# calculate score\r\n",
        "  fid = ssdiff + trace(sigma1 + sigma2 - 2.0 * covmean)\r\n",
        "  return fid"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "XgKj-ZesEHk7"
      },
      "source": [
        "def save_output(model, test_dataset):\r\n",
        "  if not os.path.exists(\"sketch_to_photo/output/\"+model.get_model_name()):\r\n",
        "    os.makedirs(\"sketch_to_photo/output/\"+model.get_model_name()+\"/real\")\r\n",
        "    os.makedirs(\"sketch_to_photo/output/\"+model.get_model_name()+\"/generated\")\r\n",
        "    os.makedirs(\"sketch_to_photo/output/\"+model.get_model_name()+\"/sketch\")\r\n",
        "  for index, (inp, tar) in enumerate(tqdm(test_dataset.take(len(test_dataset))), start=1):\r\n",
        "    generated_photo = model.get_generator_model()(inp, training=False)[0]\r\n",
        "    tar = np.squeeze(tar)\r\n",
        "    inp = np.squeeze(inp)\r\n",
        "    tf.keras.preprocessing.image.save_img(\"/content/sketch_to_photo/output/\"+str(model.get_model_name())+\"/real/\"+str(index)+\".jpg\", \r\n",
        "                                        tar, data_format=None, file_format=None, scale=True)\r\n",
        "    tf.keras.preprocessing.image.save_img(\"/content/sketch_to_photo/output/\"+str(model.get_model_name())+\"/generated/\"+str(index)+\".jpg\", \r\n",
        "                                        generated_photo, data_format=None, file_format=None, scale=True)\r\n",
        "    tf.keras.preprocessing.image.save_img(\"/content/sketch_to_photo/output/\"+str(model.get_model_name())+\"/sketch/\"+str(index)+\".jpg\", \r\n",
        "                                        inp, data_format=None, file_format=None, scale=True)\r\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "S21OQTQ8lBWD"
      },
      "source": [
        "def numpyimage(sketch_image_file, original_image):\r\n",
        "  image1 = cv.imread(sketch_image_file)\r\n",
        "  image2 = cv.imread(original_image)\r\n",
        "  return image1, image2\r\n",
        "\r\n",
        "def save_join(model, left, right):\r\n",
        "  if left == \"sketch\":\r\n",
        "    left = sorted(glob.glob(\"/content/sketch_to_photo/output/\"+model.get_model_name()+\"/sketch/*\"))\r\n",
        "  elif left == \"real\":\r\n",
        "    letf = sorted(glob.glob(\"/content/sketch_to_photo/output/\"+model.get_model_name()+\"/real/*\"))\r\n",
        "  elif left == \"generated\":\r\n",
        "    left = sorted(glob.glob(\"/content/sketch_to_photo/output/\"+model.get_model_name()+\"/generated/*\"))\r\n",
        "\r\n",
        "  if right==\"sketch\":\r\n",
        "    right = sorted(glob.glob(\"/content/sketch_to_photo/output/\"+model.get_model_name()+\"/sketch/*\"))\r\n",
        "  elif right == \"real\":\r\n",
        "    right = sorted(glob.glob(\"/content/sketch_to_photo/output/\"+model.get_model_name()+\"/real/*\"))\r\n",
        "  elif right == \"generated\":\r\n",
        "    right = sorted(glob.glob(\"/content/sketch_to_photo/output/\"+model.get_model_name()+\"/generated/*\"))\r\n",
        "  \r\n",
        "  out_dir = \"/content/sketch_to_photo/output/\"+model.get_model_name()+\"/join/\"\r\n",
        "  if not os.path.exists(out_dir):\r\n",
        "    os.makedirs(out_dir)\r\n",
        "\r\n",
        "  for i, (r, g) in enumerate(zip(left, right)):\r\n",
        "    im1,im2 = numpyimage(r, g)\r\n",
        "    im_v = cv.hconcat([im1, im2])\r\n",
        "    cv.imwrite(out_dir+str(i)+\"_merged.jpg\", im_v)"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "3wF8GkwVEKsk"
      },
      "source": [
        "def load_output(model):\r\n",
        "  real_photos = sorted(glob.glob(\"/content/sketch_to_photo/output/\"+model.get_model_name()+\"/real/*\"))\r\n",
        "  generated_photos = sorted(glob.glob(\"/content/sketch_to_photo/output/\"+model.get_model_name()+\"/generated/*\"))\r\n",
        "  real_photos = tf.data.Dataset.list_files(real_photos, shuffle=False).map(load_image)\r\n",
        "  generated_photos = tf.data.Dataset.list_files(generated_photos, shuffle=False).map(load_image)\r\n",
        "\r\n",
        "  return real_photos, generated_photos"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "Va4OzpnfhRLA"
      },
      "source": [
        "def fid_measure_for_data(model_inception, real_photos, generated_photos):\r\n",
        "  real_photos = scale_images(real_photos, (299,299,3))\r\n",
        "  generated_photos = scale_images(generated_photos, (299,299,3))\r\n",
        "  real_photos = preprocess_input(real_photos)\r\n",
        "  generated_photos = preprocess_input(generated_photos)\r\n",
        "  return round(calculate_fid(model_inception, real_photos, generated_photos), 3)\r\n"
      ],
      "execution_count": null,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "DjTC9-tDwkoV"
      },
      "source": [
        "### Calculate FID"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "WTWMF6xNwCYB"
      },
      "source": [
        "Save output files for a model\r\n",
        "\r\n",
        "\r\n",
        "*   modelSimple\r\n",
        "*   modelMobileNet\r\n",
        "*   modelXception\r\n",
        "\r\n",
        "\r\n"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "iTtjU1cUEB1o",
        "outputId": "d533ac70-ea8f-466e-d1a8-c9f6578b04f4"
      },
      "source": [
        "save_output(modelResnet, test_dataset)\r\n",
        "save_join(modelResnet, \"sketch\", \"generated\")"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "100%|██████████| 100/100 [00:01<00:00, 66.70it/s]\n"
          ],
          "name": "stderr"
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "82JiZ3AtwTmx"
      },
      "source": [
        "Load output files for a model and calculate FID measure\r\n",
        "\r\n",
        "\r\n",
        "*   modelSimple\r\n",
        "*   modelMobileNet\r\n",
        "*   modelXception"
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "BVcLciAuprLk",
        "outputId": "f46804ce-e50d-4e77-a912-26da17b01eaa"
      },
      "source": [
        "model_inception = InceptionV3(include_top=False, pooling='avg', input_shape=(299,299,3))\r\n",
        "\r\n",
        "real_photos, generated_photos = load_output(modelResnet)\r\n",
        "fid_measure = fid_measure_for_data(model_inception, real_photos, generated_photos)\r\n",
        "print(\"\\nFID: \", fid_measure)"
      ],
      "execution_count": null,
      "outputs": [
        {
          "output_type": "stream",
          "text": [
            "\n",
            "FID:  80.004\n"
          ],
          "name": "stdout"
        }
      ]
    },
    {
      "cell_type": "code",
      "metadata": {
        "id": "o_8Y15j4aBo7"
      },
      "source": [
        "# download all output files \r\n",
        "!zip -r /content/output.zip /content/sketch_to_photo/output/\r\n",
        "files.download(\"/content/output.zip\")"
      ],
      "execution_count": null,
      "outputs": []
    }
  ]
}