AlexNet_Jigsaw.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "8voK3z9TXAb5"
      },
      "source": [
        "Importing all the required libraries and packages"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "qZY2jdwJazZT",
        "outputId": "b8a332fd-da58-4e8b-c5e9-f5ca27aea07c"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Collecting tensorflow-addons==0.8.3\n",
            "  Downloading tensorflow_addons-0.8.3-cp37-cp37m-manylinux2010_x86_64.whl (1.0 MB)\n",
            "\u001b[?25l\r\u001b[K     |▎                               | 10 kB 28.5 MB/s eta 0:00:01\r\u001b[K     |▋                               | 20 kB 21.1 MB/s eta 0:00:01\r\u001b[K     |█                               | 30 kB 10.1 MB/s eta 0:00:01\r\u001b[K     |█▎                              | 40 kB 4.4 MB/s eta 0:00:01\r\u001b[K     |█▋                              | 51 kB 4.5 MB/s eta 0:00:01\r\u001b[K     |██                              | 61 kB 5.4 MB/s eta 0:00:01\r\u001b[K     |██▎                             | 71 kB 5.5 MB/s eta 0:00:01\r\u001b[K     |██▋                             | 81 kB 4.3 MB/s eta 0:00:01\r\u001b[K     |███                             | 92 kB 4.7 MB/s eta 0:00:01\r\u001b[K     |███▎                            | 102 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███▋                            | 112 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████                            | 122 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████▎                           | 133 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████▋                           | 143 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████▉                           | 153 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████▏                          | 163 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████▌                          | 174 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████▉                          | 184 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████▏                         | 194 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████▌                         | 204 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████▉                         | 215 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████▏                        | 225 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████▌                        | 235 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████▉                        | 245 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████▏                       | 256 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████▌                       | 266 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████▉                       | 276 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████▏                      | 286 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████▍                      | 296 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████▊                      | 307 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████                      | 317 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████▍                     | 327 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████▊                     | 337 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████                     | 348 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████▍                    | 358 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████▊                    | 368 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████                    | 378 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████▍                   | 389 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████▊                   | 399 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████                   | 409 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████▍                  | 419 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████▊                  | 430 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████                  | 440 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████▎                 | 450 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████▋                 | 460 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████                 | 471 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████▎                | 481 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████▋                | 491 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████                | 501 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████▎               | 512 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████▋               | 522 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████               | 532 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████▎              | 542 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████▋              | 552 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████              | 563 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████▎             | 573 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████▋             | 583 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████▉             | 593 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████▏            | 604 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████▌            | 614 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████▉            | 624 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████▏           | 634 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████▌           | 645 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████▉           | 655 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████████▏          | 665 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████████▌          | 675 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████████▉          | 686 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████████▏         | 696 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████████▌         | 706 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████████▉         | 716 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████████▏        | 727 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████████▍        | 737 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████████▊        | 747 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████████        | 757 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████████▍       | 768 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████████▊       | 778 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████       | 788 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████▍      | 798 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████▊      | 808 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████      | 819 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████▍     | 829 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████▊     | 839 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████     | 849 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████▍    | 860 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████▊    | 870 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████    | 880 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████▎   | 890 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████▋   | 901 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████   | 911 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████▎  | 921 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |█████████████████████████████▋  | 931 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████  | 942 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████▎ | 952 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |██████████████████████████████▋ | 962 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████ | 972 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████▎| 983 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |███████████████████████████████▋| 993 kB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 1.0 MB 5.2 MB/s eta 0:00:01\r\u001b[K     |████████████████████████████████| 1.0 MB 5.2 MB/s \n",
            "\u001b[?25hRequirement already satisfied: typeguard in /usr/local/lib/python3.7/dist-packages (from tensorflow-addons==0.8.3) (2.7.1)\n",
            "Installing collected packages: tensorflow-addons\n",
            "Successfully installed tensorflow-addons-0.8.3\n"
          ]
        }
      ],
      "source": [
        "!pip install tensorflow-addons==0.8.3\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "tdVBD6Jnr_Ue",
        "outputId": "189f2f7a-dab4-499e-d4c7-7060193c67d8"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "TensorFlow version: 2.8.0\n"
          ]
        }
      ],
      "source": [
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "from tensorflow import keras\n",
        "import tensorflow_addons as tfa\n",
        "import matplotlib.pyplot as plt\n",
        "import random\n",
        "import math\n",
        "print(\"TensorFlow version:\", tf.__version__)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Defining Jigsaw function."
      ],
      "metadata": {
        "id": "HjTR_OkU4mbR"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Vkeg6U0NcymN"
      },
      "outputs": [],
      "source": [
        "def apply_jigsaw(arr, destinations):\n",
        "    \"\"\"Move cells of an image similar to a jigsaw puzzle.\n",
        "    This function will split the image into ``rows x cols`` cells and\n",
        "    move each cell to the target index given in `destinations`.\n",
        "    Added in 0.4.0.\n",
        "    **Supported dtypes**:\n",
        "        * ``uint8``: yes; fully tested\n",
        "        * ``uint16``: yes; fully tested\n",
        "        * ``uint32``: yes; fully tested\n",
        "        * ``uint64``: yes; fully tested\n",
        "        * ``int8``: yes; fully tested\n",
        "        * ``int16``: yes; fully tested\n",
        "        * ``int32``: yes; fully tested\n",
        "        * ``int64``: yes; fully tested\n",
        "        * ``float16``: yes; fully tested\n",
        "        * ``float32``: yes; fully tested\n",
        "        * ``float64``: yes; fully tested\n",
        "        * ``float128``: yes; fully tested\n",
        "        * ``bool``: yes; fully tested\n",
        "    Parameters\n",
        "    ----------\n",
        "    arr : ndarray\n",
        "        Array with at least two dimensions denoting height and width.\n",
        "    destinations : ndarray\n",
        "        2-dimensional array containing for each cell the id of the destination\n",
        "        cell. The order is expected to a flattened c-order, i.e. row by row.\n",
        "        The height of the image must be evenly divisible by the number of\n",
        "        rows in this array. Analogous for the width and columns.\n",
        "    Returns\n",
        "    -------\n",
        "    ndarray\n",
        "        Modified image with cells moved according to `destinations`.\n",
        "    \"\"\"\n",
        "    # pylint complains about unravel_index() here\n",
        "    # pylint: disable=unbalanced-tuple-unpacking\n",
        "\n",
        "    nb_rows, nb_cols = destinations.shape[0:2]\n",
        "\n",
        "    assert arr.ndim >= 2, (\n",
        "        \"Expected array with at least two dimensions, but got %d with \"\n",
        "        \"shape %s.\" % (arr.ndim, arr.sahape))\n",
        "    assert (arr.shape[0] % nb_rows) == 0, (\n",
        "        \"Expected image height to by divisible by number of rows, but got \"\n",
        "        \"height %d and %d rows. Use cropping or padding to modify the image \"\n",
        "        \"height or change the number of rows.\" % (arr.shape[0], nb_rows)\n",
        "    )\n",
        "    assert (arr.shape[1] % nb_cols) == 0, (\n",
        "        \"Expected image width to by divisible by number of columns, but got \"\n",
        "        \"width %d and %d columns. Use cropping or padding to modify the image \"\n",
        "        \"width or change the number of columns.\" % (arr.shape[1], nb_cols)\n",
        "    )\n",
        "\n",
        "    cell_height = arr.shape[0] // nb_rows\n",
        "    cell_width = arr.shape[1] // nb_cols\n",
        "\n",
        "    dest_rows, dest_cols = np.unravel_index(\n",
        "        destinations.flatten(), (nb_rows, nb_cols))\n",
        "\n",
        "    result = np.zeros_like(arr)\n",
        "    i = 0\n",
        "    for source_row in np.arange(nb_rows):\n",
        "        for source_col in np.arange(nb_cols):\n",
        "            # TODO vectorize coords computation\n",
        "            dest_row, dest_col = dest_rows[i], dest_cols[i]\n",
        "\n",
        "            source_y1 = source_row * cell_height\n",
        "            source_y2 = source_y1 + cell_height\n",
        "            source_x1 = source_col * cell_width\n",
        "            source_x2 = source_x1 + cell_width\n",
        "\n",
        "            dest_y1 = dest_row * cell_height\n",
        "            dest_y2 = dest_y1 + cell_height\n",
        "            dest_x1 = dest_col * cell_width\n",
        "            dest_x2 = dest_x1 + cell_width\n",
        "\n",
        "            source = arr[source_y1:source_y2, source_x1:source_x2]\n",
        "            result[dest_y1:dest_y2, dest_x1:dest_x2] = source\n",
        "\n",
        "            i += 1\n",
        "\n",
        "    return result"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "_q0Fmmv8c9Ni"
      },
      "outputs": [],
      "source": [
        "def generate_jigsaw_destinations(nb_rows, nb_cols, max_steps, seed,\n",
        "                                 connectivity=4):\n",
        "    \"\"\"Generate a destination pattern for :func:`apply_jigsaw`.\n",
        "    Added in 0.4.0.\n",
        "    Parameters\n",
        "    ----------\n",
        "    nb_rows : int\n",
        "        Number of rows to split the image into.\n",
        "    nb_cols : int\n",
        "        Number of columns to split the image into.\n",
        "    max_steps : int\n",
        "        Maximum number of cells that each cell may be moved.\n",
        "    seed : None or int or imgaug.random.RNG or numpy.random.Generator or numpy.random.BitGenerator or numpy.random.SeedSequence or numpy.random.RandomState\n",
        "        Seed value or alternatively RNG to use.\n",
        "        If ``None`` the global RNG will be used.\n",
        "    connectivity : int, optional\n",
        "        Whether a diagonal move of a cell counts as one step\n",
        "        (``connectivity=8``) or two steps (``connectivity=4``).\n",
        "    Returns\n",
        "    -------\n",
        "    ndarray\n",
        "        2-dimensional array containing for each cell the id of the target\n",
        "        cell.\n",
        "    \"\"\"\n",
        "    assert connectivity in (4, 8), (\n",
        "        \"Expected connectivity of 4 or 8, got %d.\" % (connectivity,))\n",
        "    random_state = np.random.default_rng(42)\n",
        "    steps = random_state.integers(0, max_steps, size=(nb_rows, nb_cols),\n",
        "                                  endpoint=True)\n",
        "    directions = random_state.integers(0, connectivity,\n",
        "                                       size=(nb_rows, nb_cols, max_steps),\n",
        "                                       endpoint=False)\n",
        "    destinations = np.arange(nb_rows*nb_cols).reshape((nb_rows, nb_cols))\n",
        "\n",
        "    for step in np.arange(max_steps):\n",
        "        directions_step = directions[:, :, step]\n",
        "\n",
        "        for y in np.arange(nb_rows):\n",
        "            for x in np.arange(nb_cols):\n",
        "                if steps[y, x] > 0:\n",
        "                    y_target, x_target = {\n",
        "                        0: (y-1, x+0),\n",
        "                        1: (y+0, x+1),\n",
        "                        2: (y+1, x+0),\n",
        "                        3: (y+0, x-1),\n",
        "                        4: (y-1, x-1),\n",
        "                        5: (y-1, x+1),\n",
        "                        6: (y+1, x+1),\n",
        "                        7: (y+1, x-1)\n",
        "                    }[directions_step[y, x]]\n",
        "                    y_target = max(min(y_target, nb_rows-1), 0)\n",
        "                    x_target = max(min(x_target, nb_cols-1), 0)\n",
        "\n",
        "                    target_steps = steps[y_target, x_target]\n",
        "                    if (y, x) != (y_target, x_target) and target_steps >= 1:\n",
        "                        source_dest = destinations[y, x]\n",
        "                        target_dest = destinations[y_target, x_target]\n",
        "                        destinations[y, x] = target_dest\n",
        "                        destinations[y_target, x_target] = source_dest\n",
        "\n",
        "                        steps[y, x] -= 1\n",
        "                        steps[y_target, x_target] -= 1\n",
        "\n",
        "    return destinations"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f8JYtDE6YrwE"
      },
      "source": [
        "Importing Fashion Dataset ."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "Ecuf3jgGsa8s",
        "outputId": "c491be05-6d8e-43f5-bdf0-d5cc19615530"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-labels-idx1-ubyte.gz\n",
            "32768/29515 [=================================] - 0s 0us/step\n",
            "40960/29515 [=========================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/train-images-idx3-ubyte.gz\n",
            "26427392/26421880 [==============================] - 0s 0us/step\n",
            "26435584/26421880 [==============================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-labels-idx1-ubyte.gz\n",
            "16384/5148 [===============================================================================================] - 0s 0us/step\n",
            "Downloading data from https://storage.googleapis.com/tensorflow/tf-keras-datasets/t10k-images-idx3-ubyte.gz\n",
            "4423680/4422102 [==============================] - 0s 0us/step\n",
            "4431872/4422102 [==============================] - 0s 0us/step\n"
          ]
        }
      ],
      "source": [
        "fashion_mnist = tf.keras.datasets.fashion_mnist\n",
        "\n",
        "(train_images, train_labels), (test_images, test_labels) = fashion_mnist.load_data()\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Th_91-_IG1dI"
      },
      "source": [
        "Adding one new dimension to train and test images for AlexNet model."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "eqhJ1m9rBGaJ"
      },
      "outputs": [],
      "source": [
        "train_images = train_images[..., np.newaxis]\n",
        "test_images = test_images[..., np.newaxis]"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jI51kNdKFWuX"
      },
      "source": [
        "Converting train and test labels into one hot encoded array because categoricalcrossentropy loss function demands it."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "os4J3TziBNrH"
      },
      "outputs": [],
      "source": [
        "train_labels = np.array(train_labels)\n",
        "new_labels= np.zeros((train_labels.size, train_labels.max()+1))\n",
        "new_labels[np.arange(train_labels.size),train_labels] = 1"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "8NOjYHazKtT-"
      },
      "outputs": [],
      "source": [
        "test_labels = np.array(test_labels)\n",
        "new_test_labels= np.zeros((test_labels.size, test_labels.max()+1))\n",
        "new_test_labels[np.arange(test_labels.size),test_labels] = 1"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "1XjhCGbDZlOb"
      },
      "source": [
        "AlexNet model for Unaugmented images."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "XsK3t3hssxJj"
      },
      "outputs": [],
      "source": [
        "model = tf.keras.models.Sequential([\n",
        "    keras.layers.Conv2D(filters=96, kernel_size=(11,11), strides=(4,4), activation='relu', input_shape=(28,28,1)),\n",
        "    keras.layers.BatchNormalization(),\n",
        "    keras.layers.MaxPool2D(pool_size=(3,3), strides=(2,2),padding=\"same\"),\n",
        "    keras.layers.Conv2D(filters=256, kernel_size=(5,5), strides=(1,1), activation='relu', padding=\"same\"),\n",
        "    keras.layers.BatchNormalization(),\n",
        "    keras.layers.MaxPool2D(pool_size=(3,3), strides=(2,2),padding=\"same\"),\n",
        "    keras.layers.Conv2D(filters=384, kernel_size=(3,3), strides=(1,1), activation='relu', padding=\"same\"),\n",
        "    keras.layers.BatchNormalization(),\n",
        "    keras.layers.Conv2D(filters=384, kernel_size=(3,3), strides=(1,1), activation='relu', padding=\"same\"),\n",
        "    keras.layers.BatchNormalization(),\n",
        "    keras.layers.Conv2D(filters=256, kernel_size=(3,3), strides=(1,1), activation='relu', padding=\"same\"),\n",
        "    keras.layers.BatchNormalization(),\n",
        "    keras.layers.MaxPool2D(pool_size=(3,3), strides=(2,2), padding=\"same\"),\n",
        "    keras.layers.Flatten(),\n",
        "    keras.layers.Dense(4096, activation='relu'),\n",
        "    keras.layers.Dropout(0.5),\n",
        "    keras.layers.Dense(4096, activation='relu'),\n",
        "    keras.layers.Dropout(0.5),\n",
        "    keras.layers.Dense(10, activation='softmax')\n",
        "])\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "fOGcf2JmZ1Yn"
      },
      "source": [
        "AlexNet model for Augmented images."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "1XhtA7Hvsy9m"
      },
      "outputs": [],
      "source": [
        "model_jig = tf.keras.models.Sequential([\n",
        "    keras.layers.Conv2D(filters=96, kernel_size=(11,11), strides=(4,4), activation='relu', input_shape=(28,28,1)),\n",
        "    keras.layers.BatchNormalization(),\n",
        "    keras.layers.MaxPool2D(pool_size=(3,3), strides=(2,2),padding=\"same\"),\n",
        "    keras.layers.Conv2D(filters=256, kernel_size=(5,5), strides=(1,1), activation='relu', padding=\"same\"),\n",
        "    keras.layers.BatchNormalization(),\n",
        "    keras.layers.MaxPool2D(pool_size=(3,3), strides=(2,2),padding=\"same\"),\n",
        "    keras.layers.Conv2D(filters=384, kernel_size=(3,3), strides=(1,1), activation='relu', padding=\"same\"),\n",
        "    keras.layers.BatchNormalization(),\n",
        "    keras.layers.Conv2D(filters=384, kernel_size=(3,3), strides=(1,1), activation='relu', padding=\"same\"),\n",
        "    keras.layers.BatchNormalization(),\n",
        "    keras.layers.Conv2D(filters=256, kernel_size=(3,3), strides=(1,1), activation='relu', padding=\"same\"),\n",
        "    keras.layers.BatchNormalization(),\n",
        "    keras.layers.MaxPool2D(pool_size=(3,3), strides=(2,2), padding=\"same\"),\n",
        "    keras.layers.Flatten(),\n",
        "    keras.layers.Dense(4096, activation='relu'),\n",
        "    keras.layers.Dropout(0.5),\n",
        "    keras.layers.Dense(4096, activation='relu'),\n",
        "    keras.layers.Dropout(0.5),\n",
        "    keras.layers.Dense(10, activation='softmax')\n",
        "])\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ldFs0qXFaLNY"
      },
      "source": [
        "Applying Jigsaw operation on imported train dataset and storing it."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Q77ij1PoYTRm"
      },
      "outputs": [],
      "source": [
        "def factor(r,c):\n",
        "\n",
        " ''' Takes integer row and column\n",
        "     returns list of factors of row and col, including the number\n",
        " '''\n",
        " row=[]\n",
        " col=[]\n",
        " \n",
        " i = 1\n",
        " while(i <= int(math.sqrt(r))):\n",
        "      if ( r % i == 0):\n",
        "        if (int( r/i) == i):\n",
        "          row.append(i)\n",
        "        else:\n",
        "          if (i == 1):\n",
        "            row.append(int(r/i))\n",
        "          else:\n",
        "            row.append(i)\n",
        "            row.append(int(r/i))\n",
        "      i = i + 1\n",
        "\n",
        " j = 1\n",
        " while(j <= int(math.sqrt(c))):\n",
        "      if ( c % j == 0):\n",
        "        if (int(c/j) == j):\n",
        "          col.append(j)\n",
        "        else:\n",
        "          if (j == 1):\n",
        "            col.append(int(c/j))\n",
        "          else:\n",
        "            col.append(j)\n",
        "            col.append(int(c/j))\n",
        "      j = j + 1\n",
        "\n",
        " return row,col     \n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "PypYE1TtdMzu"
      },
      "outputs": [],
      "source": [
        "destinations=[]\n",
        "roww=[2,7,4,14,28]\n",
        "coll=[2,7,4,14,28]\n",
        "r,c=factor(random.choice(roww),random.choice(coll))\n",
        "for i in r:\n",
        "  for j in c:\n",
        "    destinations.append(generate_jigsaw_destinations(i,j,5,42))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vhgL6R2Uidof",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "57c35c5d-2318-439b-e574-91ba07f4a7bb"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/ipykernel_launcher.py:1: VisibleDeprecationWarning: Creating an ndarray from ragged nested sequences (which is a list-or-tuple of lists-or-tuples-or ndarrays with different lengths or shapes) is deprecated. If you meant to do this, you must specify 'dtype=object' when creating the ndarray.\n",
            "  \"\"\"Entry point for launching an IPython kernel.\n"
          ]
        }
      ],
      "source": [
        "destinations = np.array(destinations)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "AMiKe-46il38"
      },
      "outputs": [],
      "source": [
        "des_shape=int(len(destinations))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0xWj0nzizySw",
        "outputId": "142b2d29-5947-4712-d9d9-a55e3473d9b4"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "5\n"
          ]
        }
      ],
      "source": [
        "print(des_shape)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "3SY-0F3f1CJp"
      },
      "outputs": [],
      "source": [
        "x_train_jigg=[]\n",
        "train_labels_jigg=[]\n",
        "\n",
        "for img in train_images[:40000]:\n",
        "  destination_no= random.randint(0,des_shape-1)\n",
        "  x_train_jigg.append(apply_jigsaw(img,destinations[destination_no]))\n",
        "\n",
        "x_train_jigg=np.array(x_train_jigg)   \n",
        "x_train_jig=np.concatenate((x_train_jigg,train_images))  \n",
        "\n",
        "for label in new_labels[:40000]:\n",
        "  train_labels_jigg.append(np.array(label))\n",
        "\n",
        "\n",
        "train_labels_jigg=np.array(train_labels_jigg)\n",
        "train_labels_jig=np.concatenate((train_labels_jigg,new_labels)) \n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "LzWla4-Cegj9"
      },
      "source": [
        "Number of images before Data Augmentation."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nhtzIROzWIUf",
        "outputId": "47cb2a70-5c95-4e01-e4a6-fc8aa44d5866"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(60000, 28, 28, 1)"
            ]
          },
          "metadata": {},
          "execution_count": 17
        }
      ],
      "source": [
        "train_images.shape"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ANSdk34AeovL"
      },
      "source": [
        "Number of images after Data Augmentation."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dqzoUW8dVkl9",
        "outputId": "27812ec0-303a-4e66-e60b-cc559865886c"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(100000, 28, 28, 1)"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ],
      "source": [
        "x_train_jig.shape"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "BVPCsgu0adms"
      },
      "source": [
        "Defining a loss function for training using losses.CategoricalCrossentropy, which takes one hot encoded array and returns a scalar loss for each example."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "lxgJGKejs7Oh"
      },
      "outputs": [],
      "source": [
        "loss_fn = tf.keras.losses.CategoricalCrossentropy()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "jhng_pPIat7H"
      },
      "source": [
        "Compiling Neural Network Model for Unaugmented images."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "QFou8rzStYRE"
      },
      "outputs": [],
      "source": [
        "model.compile(optimizer='adam',\n",
        "              loss=loss_fn,\n",
        "              metrics=['accuracy',tf.keras.metrics.Recall(),tf.keras.metrics.Precision(),tfa.metrics.F1Score(num_classes=2, average=\"micro\"),tf.keras.metrics.SpecificityAtSensitivity(0.5)])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "98_WQGpLie5b"
      },
      "source": [
        "Compiling Neural Network Model for Augmented images."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "xgJnSBqhtb7R"
      },
      "outputs": [],
      "source": [
        "model_jig.compile(optimizer='adam',\n",
        "              loss=loss_fn,\n",
        "              metrics=['accuracy',tf.keras.metrics.Recall(),tf.keras.metrics.Precision(),tfa.metrics.F1Score(num_classes=2, average=\"micro\"),tf.keras.metrics.SpecificityAtSensitivity(0.5)])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "kPDYwERka9HA"
      },
      "source": [
        "Using the Model.fit method to adjust the model parameters and minimize the loss for unaugmented images model."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "myMXxAWkthp7",
        "outputId": "8ac33499-17f3-4f78-95a9-688957d955c6"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "1872/1875 [============================>.] - ETA: 0s - loss: 0.6961 - accuracy: 0.7690 - recall: 0.7051 - precision: 0.8367 - f1_score: 0.7690 - specificity_at_sensitivity: 0.9962"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/engine/training.py:2034: UserWarning: Metric F1Score implements a `reset_states()` method; rename it to `reset_state()` (without the final \"s\"). The name `reset_states()` has been deprecated to improve API consistency.\n",
            "  m.reset_state()\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r1875/1875 [==============================] - 43s 15ms/step - loss: 0.6959 - accuracy: 0.7690 - recall: 0.7051 - precision: 0.8367 - f1_score: 0.7690 - specificity_at_sensitivity: 0.9962 - val_loss: 0.6288 - val_accuracy: 0.7833 - val_recall: 0.7555 - val_precision: 0.8191 - val_f1_score: 0.7833 - val_specificity_at_sensitivity: 0.9968\n",
            "Epoch 2/10\n",
            "1875/1875 [==============================] - 25s 14ms/step - loss: 0.5684 - accuracy: 0.8230 - recall: 0.7765 - precision: 0.8685 - f1_score: 0.8230 - specificity_at_sensitivity: 0.9982 - val_loss: 0.4753 - val_accuracy: 0.8415 - val_recall: 0.8110 - val_precision: 0.8735 - val_f1_score: 0.8415 - val_specificity_at_sensitivity: 0.9987\n",
            "Epoch 3/10\n",
            "1875/1875 [==============================] - 26s 14ms/step - loss: 0.4767 - accuracy: 0.8461 - recall: 0.8080 - precision: 0.8840 - f1_score: 0.8461 - specificity_at_sensitivity: 0.9989 - val_loss: 0.4424 - val_accuracy: 0.8380 - val_recall: 0.8067 - val_precision: 0.8739 - val_f1_score: 0.8380 - val_specificity_at_sensitivity: 0.9994\n",
            "Epoch 4/10\n",
            "1875/1875 [==============================] - 26s 14ms/step - loss: 0.4178 - accuracy: 0.8565 - recall: 0.8251 - precision: 0.8903 - f1_score: 0.8565 - specificity_at_sensitivity: 0.9992 - val_loss: 0.4438 - val_accuracy: 0.8363 - val_recall: 0.8133 - val_precision: 0.8638 - val_f1_score: 0.8363 - val_specificity_at_sensitivity: 0.9991\n",
            "Epoch 5/10\n",
            "1875/1875 [==============================] - 25s 13ms/step - loss: 0.3924 - accuracy: 0.8661 - recall: 0.8368 - precision: 0.8953 - f1_score: 0.8661 - specificity_at_sensitivity: 0.9994 - val_loss: 0.4014 - val_accuracy: 0.8661 - val_recall: 0.8481 - val_precision: 0.8883 - val_f1_score: 0.8661 - val_specificity_at_sensitivity: 0.9991\n",
            "Epoch 6/10\n",
            "1875/1875 [==============================] - 25s 14ms/step - loss: 0.3542 - accuracy: 0.8745 - recall: 0.8505 - precision: 0.9004 - f1_score: 0.8745 - specificity_at_sensitivity: 0.9995 - val_loss: 0.3905 - val_accuracy: 0.8566 - val_recall: 0.8300 - val_precision: 0.8890 - val_f1_score: 0.8566 - val_specificity_at_sensitivity: 0.9995\n",
            "Epoch 7/10\n",
            "1875/1875 [==============================] - 25s 13ms/step - loss: 0.3329 - accuracy: 0.8830 - recall: 0.8607 - precision: 0.9062 - f1_score: 0.8830 - specificity_at_sensitivity: 0.9996 - val_loss: 0.3930 - val_accuracy: 0.8467 - val_recall: 0.8203 - val_precision: 0.8742 - val_f1_score: 0.8467 - val_specificity_at_sensitivity: 0.9996\n",
            "Epoch 8/10\n",
            "1875/1875 [==============================] - 28s 15ms/step - loss: 0.3302 - accuracy: 0.8862 - recall: 0.8651 - precision: 0.9068 - f1_score: 0.8862 - specificity_at_sensitivity: 0.9996 - val_loss: 0.3682 - val_accuracy: 0.8738 - val_recall: 0.8480 - val_precision: 0.8992 - val_f1_score: 0.8738 - val_specificity_at_sensitivity: 0.9996\n",
            "Epoch 9/10\n",
            "1875/1875 [==============================] - 26s 14ms/step - loss: 0.3165 - accuracy: 0.8903 - recall: 0.8715 - precision: 0.9108 - f1_score: 0.8903 - specificity_at_sensitivity: 0.9997 - val_loss: 0.3554 - val_accuracy: 0.8759 - val_recall: 0.8547 - val_precision: 0.9018 - val_f1_score: 0.8759 - val_specificity_at_sensitivity: 0.9996\n",
            "Epoch 10/10\n",
            "1875/1875 [==============================] - 26s 14ms/step - loss: 0.2866 - accuracy: 0.8980 - recall: 0.8825 - precision: 0.9161 - f1_score: 0.8980 - specificity_at_sensitivity: 0.9997 - val_loss: 0.4159 - val_accuracy: 0.8783 - val_recall: 0.8583 - val_precision: 0.9024 - val_f1_score: 0.8783 - val_specificity_at_sensitivity: 0.9995\n"
          ]
        }
      ],
      "source": [
        "history = model.fit(train_images, new_labels, epochs=10, validation_data=(test_images,new_test_labels))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "wraqP4MzbOhX"
      },
      "source": [
        "The Model.evaluate method checks the models performance on \"Test-set\""
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "l0TXu47Gtkm2",
        "outputId": "6bd2707f-809b-4a97-c09f-5d92815aa008"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            " 22/313 [=>............................] - ETA: 2s - loss: 0.3463 - accuracy: 0.8750 - recall: 0.8551 - precision: 0.9053 - f1_score: 0.8750 - specificity_at_sensitivity: 0.9994"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/engine/training.py:2034: UserWarning: Metric F1Score implements a `reset_states()` method; rename it to `reset_state()` (without the final \"s\"). The name `reset_states()` has been deprecated to improve API consistency.\n",
            "  m.reset_state()\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 2s 7ms/step - loss: 0.4159 - accuracy: 0.8783 - recall: 0.8583 - precision: 0.9024 - f1_score: 0.8783 - specificity_at_sensitivity: 0.9995\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.41590264439582825,\n",
              " 0.8783000111579895,\n",
              " 0.858299970626831,\n",
              " 0.9024287462234497,\n",
              " 0.8782999515533447,\n",
              " 0.9995333552360535]"
            ]
          },
          "metadata": {},
          "execution_count": 23
        }
      ],
      "source": [
        "model.evaluate(test_images,new_test_labels)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "ufHj61zqbbD7"
      },
      "source": [
        "Using the Model.fit method to adjust the model parameters and minimize the loss for Augmented images model."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "oPsWxzTPtnlW",
        "outputId": "b4843064-9f1e-40d7-ab18-f81cce0cd362"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/10\n",
            "3125/3125 [==============================] - ETA: 0s - loss: 0.7023 - accuracy: 0.7655 - recall_1: 0.6972 - precision_1: 0.8338 - f1_score: 0.7655 - specificity_at_sensitivity_1: 0.9960"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/engine/training.py:2034: UserWarning: Metric F1Score implements a `reset_states()` method; rename it to `reset_state()` (without the final \"s\"). The name `reset_states()` has been deprecated to improve API consistency.\n",
            "  m.reset_state()\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\b\r3125/3125 [==============================] - 43s 13ms/step - loss: 0.7023 - accuracy: 0.7655 - recall_1: 0.6972 - precision_1: 0.8338 - f1_score: 0.7655 - specificity_at_sensitivity_1: 0.9960 - val_loss: 0.5919 - val_accuracy: 0.8018 - val_recall_1: 0.7735 - val_precision_1: 0.8403 - val_f1_score: 0.8018 - val_specificity_at_sensitivity_1: 0.9978\n",
            "Epoch 2/10\n",
            "3125/3125 [==============================] - 42s 13ms/step - loss: 0.5311 - accuracy: 0.8249 - recall_1: 0.7777 - precision_1: 0.8682 - f1_score: 0.8249 - specificity_at_sensitivity_1: 0.9983 - val_loss: 0.4097 - val_accuracy: 0.8511 - val_recall_1: 0.8244 - val_precision_1: 0.8844 - val_f1_score: 0.8511 - val_specificity_at_sensitivity_1: 0.9992\n",
            "Epoch 3/10\n",
            "3125/3125 [==============================] - 41s 13ms/step - loss: 0.4444 - accuracy: 0.8474 - recall_1: 0.8117 - precision_1: 0.8808 - f1_score: 0.8474 - specificity_at_sensitivity_1: 0.9990 - val_loss: 0.4363 - val_accuracy: 0.8454 - val_recall_1: 0.8175 - val_precision_1: 0.8767 - val_f1_score: 0.8454 - val_specificity_at_sensitivity_1: 0.9991\n",
            "Epoch 4/10\n",
            "3125/3125 [==============================] - 41s 13ms/step - loss: 0.4064 - accuracy: 0.8610 - recall_1: 0.8317 - precision_1: 0.8902 - f1_score: 0.8610 - specificity_at_sensitivity_1: 0.9992 - val_loss: 0.3886 - val_accuracy: 0.8593 - val_recall_1: 0.8364 - val_precision_1: 0.8877 - val_f1_score: 0.8593 - val_specificity_at_sensitivity_1: 0.9995\n",
            "Epoch 5/10\n",
            "3125/3125 [==============================] - 41s 13ms/step - loss: 0.3842 - accuracy: 0.8659 - recall_1: 0.8401 - precision_1: 0.8927 - f1_score: 0.8659 - specificity_at_sensitivity_1: 0.9994 - val_loss: 0.5244 - val_accuracy: 0.8460 - val_recall_1: 0.8122 - val_precision_1: 0.8818 - val_f1_score: 0.8460 - val_specificity_at_sensitivity_1: 0.9987\n",
            "Epoch 6/10\n",
            "3125/3125 [==============================] - 48s 15ms/step - loss: 0.3389 - accuracy: 0.8796 - recall_1: 0.8582 - precision_1: 0.9008 - f1_score: 0.8796 - specificity_at_sensitivity_1: 0.9996 - val_loss: 0.3491 - val_accuracy: 0.8741 - val_recall_1: 0.8582 - val_precision_1: 0.8927 - val_f1_score: 0.8741 - val_specificity_at_sensitivity_1: 0.9996\n",
            "Epoch 7/10\n",
            "3125/3125 [==============================] - 41s 13ms/step - loss: 0.3262 - accuracy: 0.8838 - recall_1: 0.8649 - precision_1: 0.9044 - f1_score: 0.8838 - specificity_at_sensitivity_1: 0.9996 - val_loss: 0.3516 - val_accuracy: 0.8750 - val_recall_1: 0.8486 - val_precision_1: 0.9048 - val_f1_score: 0.8750 - val_specificity_at_sensitivity_1: 0.9996\n",
            "Epoch 8/10\n",
            "3125/3125 [==============================] - 43s 14ms/step - loss: 0.3038 - accuracy: 0.8912 - recall_1: 0.8745 - precision_1: 0.9099 - f1_score: 0.8912 - specificity_at_sensitivity_1: 0.9997 - val_loss: 0.3716 - val_accuracy: 0.8757 - val_recall_1: 0.8625 - val_precision_1: 0.8915 - val_f1_score: 0.8757 - val_specificity_at_sensitivity_1: 0.9992\n",
            "Epoch 9/10\n",
            "3125/3125 [==============================] - 48s 15ms/step - loss: 0.2901 - accuracy: 0.8969 - recall_1: 0.8811 - precision_1: 0.9133 - f1_score: 0.8969 - specificity_at_sensitivity_1: 0.9997 - val_loss: 0.3240 - val_accuracy: 0.8837 - val_recall_1: 0.8711 - val_precision_1: 0.9016 - val_f1_score: 0.8837 - val_specificity_at_sensitivity_1: 0.9997\n",
            "Epoch 10/10\n",
            "3125/3125 [==============================] - 46s 15ms/step - loss: 0.2689 - accuracy: 0.9032 - recall_1: 0.8890 - precision_1: 0.9183 - f1_score: 0.9032 - specificity_at_sensitivity_1: 0.9998 - val_loss: 0.3342 - val_accuracy: 0.8814 - val_recall_1: 0.8652 - val_precision_1: 0.9016 - val_f1_score: 0.8814 - val_specificity_at_sensitivity_1: 0.9996\n"
          ]
        }
      ],
      "source": [
        "history_jig = model_jig.fit(x_train_jig,train_labels_jig, epochs=10, validation_data=(test_images,new_test_labels))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "esn22qUmtqwn",
        "outputId": "6dfb5f1e-01d8-43ce-868b-46f6301b100f"
      },
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            " 22/313 [=>............................] - ETA: 2s - loss: 0.3635 - accuracy: 0.8849 - recall_1: 0.8636 - precision_1: 0.9075 - f1_score: 0.8849 - specificity_at_sensitivity_1: 0.9992"
          ]
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "/usr/local/lib/python3.7/dist-packages/keras/engine/training.py:2034: UserWarning: Metric F1Score implements a `reset_states()` method; rename it to `reset_state()` (without the final \"s\"). The name `reset_states()` has been deprecated to improve API consistency.\n",
            "  m.reset_state()\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "313/313 [==============================] - 3s 9ms/step - loss: 0.3342 - accuracy: 0.8814 - recall_1: 0.8652 - precision_1: 0.9016 - f1_score: 0.8814 - specificity_at_sensitivity_1: 0.9996\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "[0.3341737985610962,\n",
              " 0.8813999891281128,\n",
              " 0.8651999831199646,\n",
              " 0.9016256928443909,\n",
              " 0.8813999891281128,\n",
              " 0.9996111392974854]"
            ]
          },
          "metadata": {},
          "execution_count": 25
        }
      ],
      "source": [
        "model_jig.evaluate(test_images,new_test_labels)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Plotting graphs of different metrics."
      ],
      "metadata": {
        "id": "PmiT7tGa47iU"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "RLOQRzyPrnoA",
        "outputId": "a816c676-aa7b-4364-95df-ff2ab59316d8"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.plot(history.history['accuracy'])\n",
        "plt.plot(history_jig.history['accuracy'])\n",
        "plt.title('Train accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['model','model_jig'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "7dtRvka5jCAU",
        "outputId": "b08321d9-1670-4392-ad96-7dbc99bd7b1d"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.plot(history.history['val_accuracy'])\n",
        "plt.plot(history_jig.history['val_accuracy'])\n",
        "plt.title('Test accuracy')\n",
        "plt.ylabel('accuracy')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['model','model_jig'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "k-DpYOg8urMx",
        "outputId": "26a435be-4f72-4bea-d393-5e467d2f633a"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.plot(history.history['precision'])\n",
        "plt.plot(history_jig.history['precision_1'])\n",
        "plt.title('Train Precision')\n",
        "plt.ylabel('Precision')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['model','model_jig'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "71bqF2u4jurQ",
        "outputId": "181d78c5-b83e-492e-ac3d-04a6188202ec"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.plot(history.history['val_precision'])\n",
        "plt.plot(history_jig.history['val_precision_1'])\n",
        "plt.title('Test Precision')\n",
        "plt.ylabel('Precision')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['model','model_jig'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "Akf-NperkVxt",
        "outputId": "0e7132e6-3594-408c-ced1-86b9dd43c735"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.plot(history.history['loss'])\n",
        "plt.plot(history_jig.history['loss'])\n",
        "plt.title('Train loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['model', 'model_jig'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "erSmgderkrIg",
        "outputId": "6a3c1162-325e-489d-a495-ab75cd46429b"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYgAAAEWCAYAAAB8LwAVAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nO3deVzUdf7A8debWxBBxQtRQQENFe8jNcu8KK3szqw0223bzv3Vtlu/3V9t7dVe7dZmh5XVqmlll2ltt5Zaed8nKhriCQoqN3x+f3wHHXWAQWbmOwzv5+MxD2a+55tR5j2fW4wxKKWUUmcLsjsApZRS/kkThFJKKZc0QSillHJJE4RSSimXNEEopZRySROEUkoplzRBKKWUckkThGq0ROSE06NSRIqcXk86j+stEpGf1LA/UUSMiITUL3KlfEP/o6pGyxjTtOq5iGQBPzHGfGFfREr5Fy1BKHUWEQkSkUdEZKeI5IrI2yLSwrEvQkRmObYfE5EVItJGRP4IXAQ85yiBPFfHe8aLyHwRyRORTBH5qdO+gSKyUkQKROSgiDxdUyyefC9U46YlCKXOdR8wAbgYOAw8C0wDJgKTgRigA1AC9AaKjDG/EZGhwCxjzCvncc+5wEYgHugGfC4iO40xXwHPAM8YY2aKSFOgh+Mcl7Gcx72VcklLEEqd6y7gN8aYbGNMCfA74DpH20EZ0BJINsZUGGNWGWMK6nMzEekADAV+bYwpNsasBV4BbnMcUgYki0icMeaEMeZ7p+0ejUUpZ5oglDpXJ+B9R7XNMWALUAG0AWYCnwJzRSRHRP4qIqH1vF88kGeMOe60bQ/Q3vH8DiAV2OqoRhrv2O6NWJQ6RROEUuf6EbjMGBPr9IgwxuwzxpQZY54wxqQBQ4DxnP6mf75TI+cALUQk2mlbR2AfgDFmhzFmItAa+AswT0SiaolFqXrTBKHUuV4E/iginQBEpJWIXOV4PkJEeopIMFCAVc1T6TjvINDZjeuHOxqYI0QkAisRLAP+7NiWjlVqmOW45y0i0soYUwkcc1yjspZYlKo3TRBKnesZYD7wmYgcB74HBjn2tQXmYX0gbwEWY1X1VJ13nYgcFZFna7j+CazG5KrHpVgN4IlYpYn3gcedutxmAJtE5ITjHjcZY4pqiUWpehNdMEgppZQrWoJQSinlkiYIpZRSLmmCUEop5ZImCKWUUi4FzFQbcXFxJjEx0e4wlFKqQVm1atURY0wrV/sCJkEkJiaycuVKu8NQSqkGRUT2VLdPq5iUUkq5pAlCKaWUS5oglFJKuRQwbRCulJWVkZ2dTXFxsd2hBISIiAgSEhIIDdUJQ5VqDAI6QWRnZxMdHU1iYiIiYnc4DZoxhtzcXLKzs0lKSrI7HKWUDwR0FVNxcTEtW7bU5OABIkLLli21NKZUIxLQCQLQ5OBB+l4q1bgEfIKoTXlFJQcLiikqrbA7FKWU8iuNPkEAHDpewrHCUrvDqFViYiJHjhyp9zFKKeWORp8gQoKDiA4P4VhRGbo2hlJKndboEwRATGQoZRWVFHqhmikrK4tu3boxZcoUUlNTmTRpEl988QVDhw4lJSWF5cuXk5eXx4QJE0hPT2fw4MGsX78egNzcXMaMGUP37t35yU9+ckYCmzVrFgMHDqR379787Gc/o6JCq8iUUp4V0N1cnT3x0SY25xRUu/9kaTmhQUGEhbifM9Pim/H4Fd1rPS4zM5N33nmHGTNmMGDAAN58802WLFnC/Pnz+dOf/kSHDh3o06cPH3zwAV999RW33XYba9eu5YknnmDYsGE89thjLFy4kFdffRWALVu28NZbb7F06VJCQ0O5++67mT17NrfdpuvVK6U8p9EkiNqEBAnllYYwL1w7KSmJnj17AtC9e3dGjhyJiNCzZ0+ysrLYs2cP7777LgCXXnopubm5FBQU8M033/Dee+8BMG7cOJo3bw7Al19+yapVqxgwYAAARUVFtG7d2guRK6Uas0aTIGr7pn+ssJS9eYV0btWUpuGefVvCw8NPPQ8KCjr1OigoiPLy8jqPTDbGMHnyZP785z97NE6llHKmbRAO0RGhBImQX1jm83tfdNFFzJ49G4BFixYRFxdHs2bNGD58OG+++SYAn3zyCUePHgVg5MiRzJs3j0OHDgGQl5fHnj3VztirlFLnpdGUIGoTHCRER4SQX1RGfGyETweF/e53v2Pq1Kmkp6cTGRnJG2+8AcDjjz/OxIkT6d69O0OGDKFjx44ApKWl8Yc//IExY8ZQWVlJaGgo06ZNo1OnTj6LWSkV+CRQunb279/fnL1g0JYtW7jgggvcvkZ+USl7cgvpHBdF0widkM6Vur6nSin/JiKrjDH9Xe3TKiYn0eFWNdOxIt9XMymllL/RBOEkKEho1iSU/KIyKgOkZKWUUudLE8RZYpuEUlFpOFlSbncoSillK00QZ2kaEUJwkHDMht5MSinlTzRBnCVIhGYRoRQUazWTUqpx0wThQkykVc10olirmZRSjZcmCBeahjuqmbQ3k1KqEfNqghCRDBHZJiKZIvJINcfcICKbRWSTiLzptH2yiOxwPCZ7M86zBYkQ0ySUgqIyKiv9p5rJE+tBDBkyBICcnByuu+46j8anlAosXhtJLSLBwDRgNJANrBCR+caYzU7HpACPAkONMUdFpLVjewvgcaA/YIBVjnOPeives8U0CSXvZCnHS8qIaeKNKfzssWzZMgDi4+OZN2+ezdEopfyZN6faGAhkGmN2AYjIXOAqYLPTMT8FplV98BtjDjm2jwU+N8bkOc79HMgA5px3NJ88Agc2uH14UwxdSisIDhIICXZ9UNuecNlTNV4nKyuLjIwMBg8ezLJlyxgwYAC33347jz/+OIcOHWL27NkkJyczdepUdu3aRWRkJNOnTyc9PZ3c3FwmTpzIvn37uPDCC89ZD+LZZ5+ltLSUQYMG8fzzzxMcXE2czr9X06acOHGCrKwsxo8fz8aNGyksLGTKlCls3LiRrl27kpOTw7Rp0+jf3+XgSqVUI+HNKqb2wI9Or7Md25ylAqkislREvheRjDqci4jcKSIrRWTl4cOHPRg6CEJIkFBRaTDUr5opMzOThx56iK1bt7J169ZT60H8/e9/509/+hOPP/44ffr0Yf369fzpT386ta5D1XoQmzZt4uqrr2bv3r3AmetBrF27luDg4FOT/Z2P559/nubNm7N582Z+//vfs2rVqnr9vkqpwGD3ZH0hQApwCZAAfCMiPd092RgzHZgO1lxMNR5cyzd9V8pKytl1+AQdW0QSG3n+1Uz+vh7EkiVLeOCBBwDo0aMH6enp530tpVTg8GaC2Ad0cHqd4NjmLBv4wRhTBuwWke1YCWMfVtJwPneR1yKtRlRYMKHBQeQXldUrQeh6EEqphsibVUwrgBQRSRKRMOAmYP5Zx3yAIxGISBxWldMu4FNgjIg0F5HmwBjHNp8SR2+m48XlVHixN5Pd60EMHTqUt99+G4DNmzezYYP7bTVKqcDltRKEMaZcRO7F+mAPBmYYYzaJyJPASmPMfE4ngs1ABfCwMSYXQER+j5VkAJ6sarD2uIoyOHEQmjSHsKhzdsc0CeXIiRIKistoXo9SRE18uR6Eq3Uu7r77biZPnkxaWhrdunWje/fuxMTEePaXVEo1OLoeRGWF1bspqhXEnNMOjjGGrQeO0yQ0mMS4cxNIQ5Kbm0vfvn3PKW1UVFRQVlZGREQEO3fuZNSoUWzbto2wsHMToq4HoVRgqWk9CLsbqe0XFGyVHEoKcNFR6lQ1U+7JUsorKwkJapiDz3Nycrjkkkv45S9/ec6+wsJCRowYQVlZGcYYnn/+eZfJQSnVuGiCAIhoBgU5UF4KIed+MMZGOqqZisppEeX/H5y5ubmMHDnynO3fffcdLVu2PGd7dHQ0Z5e+lFIq4BOEMab29aXDY4AcqxQREnfO7iahwYSFWL2ZGkKCaNmyJWvXrvX4dQOlOlIp5Z6GWV/ipoiICHJzc2v/YAsJh+AwKC5wubuqmulEcTnlFZVeiNT/GWPIzc0lIiLC7lCUUj4S0CWIhIQEsrOzcWuUddExKM2Bg8XgosRRWl7JoeMllBwJJSo8oN+2akVERJCQkGB3GEopHwnoT7rQ0FCSkpLcO3jbJzD/JrjtQ+h8yTm7jTFc+o/FxMdGMPsngz0ap1JK+aOArmKqk6ThVjXTjs9d7hYRxqe347uduRw+XuLj4JRSyvc0QVQJi4JOQyDzi2oPGZ8eT6WB/27c78PAlFLKHpognCWPhsNb4dhel7u7to0mpXVTPlqvCUIpFfg0QThLGWP9rKaaCaxSxIqsPA4WFPsoKKWUsocmCGdxKRDbseZqpl7tMAYWailCKRXgNEE4E7GqmXYthnLXDdFdWjXlgnbNWLhBE4RSKrBpgjhbymgoOwl7llV7yPj0dqzac5ScY0U+DEwppXxLE8TZqrq71lDNdEV6PKDVTEqpwKYJ4mxhUdBpaI0N1R1bRpKeEMOC9Tk+DEwppXxLE4QrKaPhyDY4Wv0qbeN6tmNddj57cwt9GJhSSvmOJghXqrq7ZlZfihiX3g6ABRu0FKGUCkyaIFxpmQyxnWBH9e0QCc0j6dMxlgXrtB1CKRWYNEG4ImJVM+3+ptrurmANmtu8v4Bdh0/4MDillPINTRDVSa69u+u4nu0QgQXam0kpFYA0QVQn6SIIDq+xN1PbmAgGdGqhvZmUUgFJE0R1wqIgcWiNDdVgTb2x/eAJth887qPAlFLKNzRB1CR5NBzZDkezqj3ksh7tCBJYsE5LEUqpwOLVBCEiGSKyTUQyReQRF/uniMhhEVnrePzEaV+F0/b53oyzWimjrZ81VDO1ig5ncOeWLFi/v/a1r5VSqgHxWoIQkWBgGnAZkAZMFJE0F4e+ZYzp7Xi84rS9yGn7ld6Ks0Ytk6F5Yo3TboDVm2nXkZNs3l/gm7iUUsoHvFmCGAhkGmN2GWNKgbnAVV68n+dVze66+xsoq379h4webQkOEu3NpJQKKN5MEO2BH51eZzu2ne1aEVkvIvNEpIPT9ggRWSki34vIBFc3EJE7HcesPHz4sAdDd5IyGsoKYW/13V1bRIUxNDmOhVrNpJQKIHY3Un8EJBpj0oHPgTec9nUyxvQHbgb+JSJdzj7ZGDPdGNPfGNO/VatW3okwsfburmBNAb43r5AN+/K9E4dSSvmYNxPEPsC5RJDg2HaKMSbXGFM1VPkVoJ/Tvn2On7uARUAfL8ZavbBISBxWa4IYm9aW0GCtZlJKBQ5vJogVQIqIJIlIGHATcEZvJBFp5/TySmCLY3tzEQl3PI8DhgKbvRhrzVJGQ+6OGru7xkSGMjyllVYzKaUChtcShDGmHLgX+BTrg/9tY8wmEXlSRKp6Jd0vIptEZB1wPzDFsf0CYKVj+9fAU8YY+xJEcu3dXcGa4XXfsSJW7z3mg6CUUsq7Qrx5cWPMx8DHZ217zOn5o8CjLs5bBvT0Zmx10rILNE+yEsTAn1Z72Oi0NoSFBLFgfQ79OjX3YYBKKeV5djdSNwzOs7vW0N01OiKUS1Jb8fGG/VRWajWTUqph0wThruTRUF4Ee5bWeNj4XvEcLChhRVaejwJTSinv0AThrsRhbnV3HdmtNRGhQdqbSSnV4GmCcFdYpDUFeC2zu0aFhzCyWxs+2bif8opKHwWnlFKepwmiLpJHQ24m5O2u8bDx6e04cqKUH3ZrNZNSquHSBFEXVbO71jJ534hurYkKC9aFhJRSDZomiLo41d31sxoPiwgNZlRaGz7ZeIAyrWZSSjVQmiDqKmUM7P62xu6uYE0BfqywjKWZR3wUmFJKeZYmiLpKqeruuqTGw4anxhEdEaK9mZRSDZYmiLpKHAYhEbCj5naI8JBgxqS15dNNBygt12ompVTDowmirkKbWFOA19IOATC+VzuOF5fz7Q4vrVWhvG/xX2HV63ZHoZQtNEGcj5TRkLcT8nbVeNiw5DhimoRqNVNDdfwgLHoKFv8NdIZe1QhpgjgfyaOsn7VUM4UGB5HRvS2fbz5IcVmFDwJTHrX+LTAVUJANBzfZHY1SPqcJ4ny07AItOrtdzXSipJxF27SaqUExBtbMhFbdrNfb/2tvPErZQBPE+UoZA1nfQllRjYdd2LklLaPCdNBcQ5O9Ao5shwvvhfi+sP1TuyNSyuc0QZyv5NFQXgxZNc/uGhIcREaPtny55RCFpeU+Ck7V25qZEBoF3SdAaoaVME7qmBbVuGiCOF+JQ63urrVM3gfWoLmisgq+2nrIB4Gpeis5ARvfg+5XQ3g0pI4FjFtVikoFEk0Q56sO3V0HJrWgVXQ4C9Zpb6YGYfOHUHoC+t5qvW7XC6LbaTuEanQ0QdRHyhirq2vuzhoPCw4SxvVsx9fbDnGiRKuZ/N6aWdAyGToMsl6LWP/WmV9Beam9sSnlQ5og6iPF0d21ltldwZoCvKS8ki82H/RyUKpejmTC3mXQ5xYrMVTpehmUHrf2KdVIaIKojxadoUWXWleZA+jbsTntYiK0N5O/WzsLJBh6TTxze9LFVpuT9mZSjYgmiPpys7trkKOaafH2w+QXlfkoOFUnFeWwdo71bxrd9sx9YZGQNBy2faKjqlWjoQmivlJGObq71jy7K8D4XvGUVRg+23TAB4GpOsv8Ak4csKqXXEkdC0d3W6sKKtUIaIKor07DIKSJW9VMvRJi6NCiic7N5K/WzISoVo5urS6kOLZv+8R3MSllI68mCBHJEJFtIpIpIo+42D9FRA6LyFrH4ydO+yaLyA7HY7I346yX0AhIusit8RAiwrie8SzNPMLRk9obxq+cOGx1Y+11EwSHuj4mtgO06aHtEKrR8FqCEJFgYBpwGZAGTBSRNBeHvmWM6e14vOI4twXwODAIGAg8LiLNvRVrvbnZ3RWs3kzllYZPtZrJv6yfC5Xl0OfWmo9LHQt7v4Oio76JSykbebMEMRDINMbsMsaUAnOBq9w8dyzwuTEmzxhzFPgcyPBSnPV3anbX2ksR3eObkRQXpdVM/sQYa+xDwkBo1bXmY1MzrBleM7/0TWxK2cibCaI98KPT62zHtrNdKyLrRWSeiHSoy7kicqeIrBSRlYcP2zhbaoska2CV29VM7Vi28whHTpT4IDhVq32r4PDW6hunnbXvB5FxWs2kGgW7G6k/AhKNMelYpYQ36nKyMWa6Maa/MaZ/q1atvBKg25JHw+5vobSw1kPH92pHpYFPNmo1k19Y/R8IjbTmXqpNULBjVPXnVrdYpQKYNxPEPqCD0+sEx7ZTjDG5xpiqr9GvAP3cPdfvpIyGihK3urt2bRNNcuumLFing+ZsV3rSmpgvbQJENHPvnNSxVhtE9grvxqaUzbyZIFYAKSKSJCJhwE3AfOcDRKSd08srgS2O558CY0SkuaNxeoxjm//qNNT6FupmNdP49HYsz8rjYEGxD4JT1do835pCo28tjdPOulwKQSE6eZ8KeF5LEMaYcuBerA/2LcDbxphNIvKkiFzpOOx+EdkkIuuA+4EpjnPzgN9jJZkVwJOObf4rNOL07K5ujLQdnx6PMfDxBm2sttWamdaUKR0vdP+ciGbWFwJNECrAebUNwhjzsTEm1RjTxRjzR8e2x4wx8x3PHzXGdDfG9DLGjDDGbHU6d4YxJtnxeM2bcXpMymg4muVWd9fk1k3p1jZaezPZKXcn7Fl67sR87kjNsBq283Z7Jzal/IDdjdSBJWW09dONaiaAK3rFs2rPUXKO1TyPk/KStbNBgqDXzXU/t2q0tS4ipAKYJghPap4ILVPcGg8B1qA5gIVaivC9inJY+6bV+6xZu9qPP1vLLhCXqtVMKqBpgvC0lNFWTyY3urt2ahlFz/YxOgW4HXZ+Bcf3uzf2oTqpY61/65LjnotLKT/iVoIQkQdEpJlYXhWR1SIyxtvBNUinurt+69bh49PbsS47n725tScU5UFrZloD3lLrMUA/NQMqSmHXIo+FpZQ/cbcEMdUYU4DV3bQ5cCvwlNeiasiquru6Wc00zlHNtGCDliJ85uQRa0bWXjdBSNj5X6fDIIiI0WomFbDcTRBVXTwuB2YaYzY5bVPOQsKthWUyP3eru2tC80j6dIzVdghfWv8WVJbVr3oJrFlfk0fB9s+gstIzsSnlR9xNEKtE5DOsBPGpiEQD+hdRneRRju6u7i0sMz49nk05Bew+ctK7cSkraa+eCe37Q+sL6n+91Aw4eQhy1tT/Wkr5GXcTxB3AI8AAY0whEArc7rWoGrqq7q7uVjP1bIcIOvWGL+SshsNb6l96qJI8yuoqq9VMKgC5myAuBLYZY46JyC3Ab4F874XVwDVPtLpAujkeom1MBAM6tdBBc76weqa1AmCPazxzvcgWVluEJggVgNxNEC8AhSLSC3gI2An8x2tRBYLk0ZC11JoMzg3j0tux7eBxZizZTUVl7W0X6jyUFsLGd6H7BKtx2VNSM+DAeijQEqAKLO4miHJjjMFa8Oc5Y8w0INp7YQWAlFFWd9fd7nV3vbZfAhelxPHkgs1c+8Iytuwv8HKAjdCWj6CkwHPVS1WqusrqGhEqwLibII6LyKNY3VsXikgQVjuEqk4dZncFaBoewn+mDuRfN/Zmb14hV/x7CU99spWi0govB9qIrJkJzZOsfxtPatUVYjtpglABx90EcSNQgjUe4gDW+gx/81pUgSAkHJIuthqq3ejuCtY04BP6tOfLBy/m6j7teXHxTsb+6xu+3WHjanmBIm+XNXjxfCbmq42IVYrYtQjKdF4tFTjcShCOpDAbiBGR8UCxMUbbIGqTMgqO7XG7u2uV5lFh/O36Xsz56WBCgoRbX13O/7y1llxdovT8rX3TMTHfRO9cP3UslBfB7m+8c32lbODuVBs3AMuB64EbgB9E5DpvBhYQkqu6u57fjJ8XdmnJxw9cxP2XJrNgfQ4jn17M2yt/xLhZIlEOlRVWgugyEmJcLYvuAYnDIDRKezOpgOJuFdNvsMZATDbG3AYMBP7Pe2EFiOadIK6r2+MhXIkIDebBMV35+P6LSGndlF/NW8/El79n1+ETHgw0wO38Ggr2eb5x2llIOHQZYbVDaAJXAcLdBBFkjDnk9Dq3Duc2bimjrUVp3OzuWu1l2kTz1p0X8udrerIpp4CMZ77l2S93UFquA9prtWYmRLaErpd79z6pGVYiOrjRu/dRykfc/ZD/r4h8KiJTRGQKsBD42HthBZDkUdaMnx6omw4KEiYO7MiXD13MmLQ2PP35di5/9ltWZPn3aqy2OpkLWxdC+o31m5jPHVWLCGk1kwoQ7jZSPwxMB9Idj+nGmF97M7CA0WmIVTddj2qms7WOjuC5m/vy2u0DKCqt4PoXv+PR9zaQX1jmsXsEjA1ve2ZiPnc0bQ3t+2l3VxUw3K4mMsa8a4x50PF435tBBZSQcOh8sduzu9bFiK6t+fzB4fz0oiTeWrGXkU8v5qN1OdqIXaVqYr74vtCmu2/umZoB2SvhhHZNVg1fjQlCRI6LSIGLx3ER0aG+7koeBcf2wpEdHr90ZFgIvxmXxvx7h9EuJoL75qxh6usryD6qCxCRswYObfJN6aFK6ljA6FrVKiDUmCCMMdHGmGYuHtHGmGa+CrLBS6lfd1d39Ggfwwf3DOWx8Wn8sDuP0U9/w8vf7KK8ohE3Yq+ZBSER0ONa392zbTpEt9N2CBUQtCeSL8R2hFbd3J5243wFBwlThyXx+YMXMzS5JX/8eAtXTVvKhuxGOPFuWRFsmAdpV0GTWN/dV8QqRez8CspLfXdfpbzAqwlCRDJEZJuIZIrIIzUcd62IGBHp73idKCJFIrLW8XjRm3H6RPIo2LMMSrw/fqF9bBNevq0/L0zqy+HjJVw1bQlPfrSZkyXlXr+339jyEZTk+7Z6qUpqBpSesLo3K9WAeS1BiEgwMA24DEgDJopImovjooEHgB/O2rXTGNPb8bjLW3H6TMpoq7trlnuzu9aXiHBZz3Z88dDF3DyoI68t282Yf37Dl1sO+uT+tlsz05pAr9Mw39876WKrakt7M6kGzpsliIFApjFmlzGmFJiLNV342X4P/AUo9mIs9ut4IYQ19XnjZbOIUP4woSfz7hpC0/AQ7nhjJXfPXsWhggB+u49mWeNO+twKQTbUooZFWkli+yc6qlo1aN7862kP/Oj0Otux7RQR6Qt0MMYsdHF+koisEZHFInKRqxuIyJ0islJEVh4+7OfdCk/N7vqFLR8a/To156P7hvHw2K58seUQI/+xmFnf76EyEBcnWjMbEOjtpYn53JE61kpUXui5ppSv2NZI7VhT4mmsFerOth/oaIzpAzwIvCki5/SaMsZMN8b0N8b0b9WqlXcD9oSUUZC/F45st+X2YSFB3DMimc9+MZz0DjH89oONXP/Sd2w/eNyWeLzi1MR8l0JMgn1x6KhqFQC8mSD2AR2cXic4tlWJBnoAi0QkCxgMzBeR/saYEmNMLoAxZhXWEqepXozVN+o5u6unJMZFMeuOQTx9Qy92HT7BuGe/5e+fbqO4LAAWJ9q1CAqyoe+t9sYRkwBtemqCUA1aiBevvQJIEZEkrMRwE3Bz1U5jTD4QV/VaRBYBvzTGrBSRVkCeMaZCRDoDKcAuL8bqG7EdoNUF1rQbQ+6zNRQR4Zq+CVzStTV/XLiF577OZMH6HP50dU+GJMfVfoFqlFdUUlpRSWm500/H87JyQ2lFBSXlp7eXVVjbTh9nTj1Pbt2Uy3u2ReqywM+amdCkufcn5nNH6lhY8k8ozIPIFnZHo1SdeS1BGGPKReRe4FMgGJhhjNkkIk8CK40x82s4fTjwpIiUAZXAXcaYwJiRLmUU/PCS1d01vKnd0dAiKox/3NCLa/u253/f38DNr/zAiK6tiI4Ipczpg76kvPL061Mf+Kf3lTr2e7pJY+LADjxxZQ/CQtwo7BbmWRPz9Z9qtfnYLTUDvv27NSaipy6fohoeb5YgMMZ8zFmzvhpjHqvm2Eucnr8LvOvN2GyTPBqW/dvqZdPND77lOgxJjuO/vxjOtK8z+WDtPoJFCAsJIiwkiNDgIMKCg2gaHkJ4lON1iLWt6piws7aFOu0Ld7rGqeNrOCc8JIjgIOFfX2xn2tc72XHwBC/c0o9W0bV86E4pNoYAAB/xSURBVG94x+pKbMfYB1fa94XIOKuaSROEaoC8miCUC87dXf0oQYC1ONFDY7ry0JiudocCwMNju9GtbTMenreOK59bwvRb+9MzIab6E9bMhHa9oW1P3wVZk6Bgq5pp60KoKIdg/XNTDYtOteFrIWHQ+RLItKe7a0NzRa943v35EIJEuO7FZXy4dp/rA3PWwoEN/lN6qJI6FoqPQfZyuyNRqs40QdgheRTk/wiHt9kdyZkqK2HzfPjid1DqP7PBdo+P4cN7h9IrIZYH5q7lqU+2UnF2Y8eaWRAc7n9VOZ1HQFCo9mZSDZImCDtUze7q5cn73GYMbP0Ypg+Ht2+1et7MnGA1+vqJuKbhzPrJICYN6siLi3dyxxsryC9yLJBUVmwtDJR2pdWDyZ9ENIPEobBNE4RqeDRB2CEmAVqn2T4eAmOsLrcvj4C5E611s6+eDte/bq2l8NrlUJBjb4xOwkKC+OPVPfnDhB4s2XGEq6ctZefhE7B1ARTbNDGfO1Iz4Mg2yGv4PbVV46IJwi7Jo2DPd1BiwyhmY6yul6+OhtnXQWEuXDUN7lkBvW6E7lfDLe9Cfja8Osbvpou4ZXAn3vzpYPKLypjw3FLylrxqTameONzu0Fw7NapaFxFSDYsmCLukjLbWSt79jW/vu/tbq2Qw82oo2A9XPAP3rrK+fTv3skkaDrcvhPJimDEW9q3ybZy1GJjUgg/vHUq/mOPEHviO5bGXY+oyoM6XWnSGuFRth1ANjiYIu3QY7NvZXfd8B6+PhzfGw9HdcPnf4f7V0G+K1bPKlXa9YOqnVpyvX2GVOvxIQvNIXk7fBgL/s/UCHpi7lqJSP50uJHUsZC2xp8So1HnSBGGXqu6u3p7d9ccVVmnhtQyr11TGU3D/Whj4U/dGG7fsAnd8Bi2SYPYNsPE978VaV5WVhG6Yg3S+hJvHDuOj9Tlc/9Iyco4V2R3ZuVIvs0qMO7+2OxKl3KYJwk4po62J5Q5v9fy1962G2dfDq6Ng/zoY8wd4YB0M/jmERtTtWtFtYcpCSBgA86bC8pc9H+/52L0Y8n9E+t7KPSOSeeW2/mQdKeTK55awIst/emAB0GEQRMToIkKqQdEEYadTs7t6sLvr/vUw52arZ1L2Chj5ODyw3pocMCzy/K/bJBZufQ+6XgYf/xK+/rP9A/3WzISIWOg6DoCRF7Thg3uGEB0Rys0vf8+c5Xvtjc9ZcIj1773jU2u8iVINgCYIO8W091x314Ob4a1b4aWLYM8SGPFbKzFc9KDnJgUMbQI3zITet8Dip6xEUWlTnX9hHmxZAOk3nFEiSm4dzQd3D+XCLnE8+t4GHvtwI2UVfvKBnJoBJw9Dzmq7I1HKLTo5jN1SRsN3z1uNl+HRdT//8Hbrw3rje1Zj8sW/hsF3W9/4vSE4BK56DqLiYOm/4OQRuGa672dP3fguVJRYy4qeJSYylNemDOCv/93KS9/sYtuB4zw/qS8tm9o8w2vySJAgqzdTQn97Y1HKDVqCsFuyo7vrrsV1Oy93J7z3M3h+kDVK96IH4RfrYcT/ei85VBGB0U9Y7RqbP7DaOnzdO2f1f6BtOrRLd7k7OEh49PIL+OeNvVjz4zGufG4pm3MKfBvj2SJbWL3XtLuraiA0Qdit42AIi3Z/2o2jWfDBPfDcANj8IVx4r5UYRj7m+0VphtwHE160um++Pt4qTfjC/nVwYL3L0sPZru6TwLy7LqSi0nDtC8tYuH6/DwKsQepYa1LB/GomHVTKj2iCsFtwKHS+2GqorqnR99iP8NED8O9+1roHg35m9Uoa83uruscuvSfCxDlWF9oZY+HoHu/fc83sOk3Ml54Qy/z7hnJBu2jueXM1//hsG5WeXtnIXV0vs37u0N5MyjOWZh7hm+2HvXJtTRD+IGUMFOyDQ1vO3VeQAwsfgmf7wNo3od/t8MBayPgzRLfxfayupI6F2z60GmBnjLUazL2lrBjWvwUXjK9Tial1dARz7hzMDf0T+PdXmdw5cxXHi8u8F2d14lKheaJ2d1Uese3Ace6auYq//NfFDMceoAnCHySPsn46VzMdPwif/Bqe6Q2rXremwrhvNYz7OzSLtyXMGnUcBLc76tZfy4C933vnPtsWWusrnMfEfOEhwfzl2nSeuLI7X287xDXPLyPryEkvBFkDEas3065FfjWlump4Dh0vZurrK4gIC2b6bf0JDvL8VDOaIPxBTHto3d2qZjpxGD79DTzTyxqQln493LcKrvgXxHawO9KatUmzRl1HtYL/TPDOt+Q1syCmAyRdcl6niwiThyQyc+pADp8o4crnlniteF6t1LHWHFe+nodLBYzC0nJ+8sZK8k6WMmPyANrHNvHKfTRB+IuU0bBnmZUYvn8euk+Ae1dYs6w2T7Q7OvfFdrTmb2rdDeZMhLVzPHftY3utqSp6T4Kg+v3XHZIcx/x7hhEf24Qpry3nlW93YXw18K/TUKtLsvZmUuehotLwwNy1bNiXz78n9ql5Gd560gThL7pPsNYw7noZ3P0DXP2iNQ9SQxQVB5M/gsRh8MFdsOzfnrnu2jmAgd43e+RyHVtG8u7PhzA6rQ1/WLiFh95ZR3GZDwb+hYRDlxFWCcvu0eiqwfnjwi18vvkgj41PY1Sad9shNUH4i/g+8NtDcN2r0CrV7mjqLzwaJr0DaRPgs9/C54/V78OwshLWzoKki6F5J4+FGRUewguT+vE/o1J5b/U+bpz+PQfyiz12/WqlZsDxHKvLq1JuemNZFjOW7mbKkERuH5rk9ftpgvAn/rqewfkKCYfrZkD/O2DpM/DhvVBRfn7XyvrGqmLqe5tnYwSCgoQHRqXw4i392HHwOFc+t4TVe496/D5nSBkDiPZmUm77cstBnvhoE6MuaMP/jU/zyT29miBEJENEtolIpog8UsNx14qIEZH+TtsedZy3TUTGejNO5UVBwTDuH3DJo1YJ4O1boew8puNeM8uaDbXbOM/H6JDRoy3v3T2E8NAgbnrpe95Z+aPX7kXT1tC+n7ZDKLds3JfPfXPWkBbfjGcn9vZKjyVXvJYgRCQYmAZcBqQBE0XknLQnItHAA8APTtvSgJuA7kAG8LzjeqohEoFLHrEWKdr2Ccy8BoqOuX9+0VHYPB96Xm9NGOhF3do2Y/49w+if2JyH563niY82Ue6tyf5SM6yV+k4c8s71VUDIOVbE1NdXENsklBmTBxAZ5rsp9LxZghgIZBpjdhljSoG5wFUujvs98BfAueL3KmCuMabEGLMbyHRcTzVkA39qVTllr4DXx8HxA+6dt2FetRPzeUPzqDD+M3Ugtw9N5LWlWQz9y1f8bOZKnl+UydLMIxR4aoBd6ljAeHa6dxVQjheXMfX1FRSWVjDj9gG0blbHtVzqyZupqD3gXEbPBgY5HyAifYEOxpiFIvLwWed+f9a57b0VqPKhHtdAk+YwdxK8OgZufb/23lprZkGbntYSqD4SEhzE41d0Z2BiCz7eeID12cf4dNPBU/s7x0WRnhBDekIsvTrE0j2+GRGhdSzktu0J0fGw/RPoM8nDv4Fq6MorKrn3zTXsOHSC16YMoFvbZj6PwbbpvkUkCHgamFKPa9wJ3AnQsWNHzwSmvK/LCJjykTUL7IyxMGkexPd2feyBDbB/LWT8xZZG/Mt6tuOynu0AOFZYyvrsfNZnH2Ptj/ks25nLB2tzAAgJElLbRNOrgyNpJMSS2qYpIcE1FNJFrFLEhnegvMT3U6Yrv2WM4bH5m1i8/TB/vqYnw1Nb2RKHNxPEPsB56G+CY1uVaKAHsEisP/y2wHwRudKNcwEwxkwHpgP0799fO5Q3JO37WQPqZl5tzQQ78U1IGn7ucWtmQ3CYtTCQzWIjwxie2uqMP9YD+cWsyz7G+uxjrPsxn4Xr9zNnuVVwjggNont8DOkJMfRylDQSW0YizokuNQNWvQZ7lkKXS339Kyk/9fK3u3jzh73cdXEXJg6078uveGv0qIiEANuBkVgf7iuAm40xm6o5fhHwS2PMShHpDryJ1e4QD3wJpBhjqh3F1L9/f7Ny5UrP/hLK+wpyrEbrvJ1w7auQduXpfeUl8I+u0PkSuP51mwKsG2MMWbmFpxLG+uxjbMzJp7jMauhuFhFCekLs6eqptqG0fTEN6TcFLvuLvcErv/DJhv38fPZqxqW349839SHIyz2WRGSVMcblClZeK0EYY8pF5F7gUyAYmGGM2SQiTwIrjTHzazh3k4i8DWwGyoF7akoOqgFrFg+3fwxv3gjvTIZxT0P/26192z62ejCdx8R8dhERkuKiSIqL4qreVrNZeUUl2w+esJKGo4pq+je7KHfMvjmzSXe6rfqQuaF3kN6xOb0SYoiNDPNZzGUVlRSWVHCytJzC0nJOOp4XlVZwsrSCwpLyUz/LKg2X9WjLBe18Xx/eGKzee5RfvLWWvh1j+cf1vbyeHGrjtRKEr2kJooErLbQSxI7P4NLfwkW/hNnXwaGt1oJIQYHVy7m4rIJNOQWszz5G1IaZ3HDgH4ws+Rs7jZVUOraIpFeHWHo5Sho92jejSWgwRWUVnCypcHx4n/5AP+On84f6qe1n7iusOr+kgtLz6MZ7abfW3DOiC/06+XiRqgC2N7eQq59fSlR4CO/fPcRnS+TWVILQBKH8R0WZNdp6/VzodTOsmwPDH4ZLf2N3ZN6Vvw/+mUbxiMdZ3f421mYfY72jeirHMe1HVbNFXf5co8KCiQwPsX6GhRAVfvpnk1Cn187HVXN8ZFgIkWHBFJZU8MZ3Wby2dDdHC8sYlNSCu0ckMzwl7sy2FVUn+YVlXPPCUo6cKOW9u4fQpVVTn91bE4RqOCor4fP/g++es17fvxZaeH/OGdu9OMxaenbqJ2dsPny8xGrH2FdARWWlyw/wSBcf6BEhwV6tnigsLWfO8h95+ZtdHCgopkf7ZtxzSTJju7e1vVqkoSktr2TyjOWs3JPHzDsGMbhzS5/eXxOEaliMgRWvWGtcj3jU7mh846s/wLf/gId3+n5t8XooKa/ggzX7eGHRTrJyC+nSKoqfX5LMVb3jCa2pi68CrE4ND72zjvdW7+OfN/bi6j4JPo+hpgSh/4LK/4hYo64bS3IAq7urqYTML+2OpE7CQ4K5cUBHvnzoEv49sQ9hIcH88p11XPK3RbyxLMs306c3YM9+mcl7q/fxi1EptiSH2miCUMofxPeFyLgGO3lfcJBwRa94Pr5/GK9NGUC7mAgen7+JYX/5imlfZ3puepIA8v6abP75xXau6dueB0am2B2OS7aNpFZKOQkKskZVb11gTYke3DD/NEWEEd1aM6Jba5bvzmPa15n87dNtvLhoJ7cN6cTtQ5OI81HvHH/2w65cfj1vA4M7t+Cpa9L9toFfSxBK+YvUDCjOhx9/qP3YBmBgUgvemDqQBfcN46LUOJ5ftJNhf/mK383fxL5j5zHle4DYefgEd85cRUKLJrx0S3/CQvz3Y9h/I1OqsekyAoJCG2w1U3V6tI/h+Un9+OLBi7kiPZ5Z3+/h4r9+zcPvrGPn4RN2h+dTuSdKuP21FYQECa9PGUhMZKjdIdVIE4RS/iI82lrHO0BXmevSqil/u74Xi381glsGd+Kj9TmMenoxd89excZ9+XaH53XFZRXcOXMVBwuKeXlyfzq2jLQ7pFppglDKn6RmwJFtkLfL7ki8pn1sE353ZXeW/PpS7r6kC99uP8L4fy9h8ozlLN+dZ3d4XlFZaXVnXbXnKP+8sTd9Oza3OyS3aIJQyp+kjrF+Bmgpwllc03AeHtuNpY9eyq8yurJxXz43vPQd172wjK+3HiJQxmgB/O2zbSxcv59HL+vG5Y7p4xsCTRBK+ZMWnSGua8C1Q9SkWUQod1+SzJJfX8oTV3Znf34xt7++gsufXcJH63KoqGzYiWLu8r28sGgnEwd25M7hne0Op040QSjlb1LHQtZSKC6wOxKfahIWzOQhiSx6+BL+fn0vSssruG/OGkY9vZi3VuyltLzSWst8xavw4T1w/GDtF7XZtzsO85sPNjI8tRW/v6q733ZnrY5OtaGUv8laCq9fDjf8B9JcLePeOFRWGj7bfIDnv9pB1IHvuS1iCaP5npDKEkCgeSdrydoW/vmtfNuB41z3wjLaN2/CO3ddSHSEf/ZY0qk2lGpIOgyCiNhG0Q5Rk6DjOWTkzuJDcx9zwv7IcLOSOaUXMUmeYk76q5QXHsO8Ohb2r7M71HMcKijm9teW0yQsmBlTBvhtcqhNwxyuqVQgCw6BlNFWgqistEZZNxblJbDtE1gzE3Z+BaYSSbwIRvyGqAuuIC2niEVf7+TR5Yd4RX7DrLCniJmewafp/6RDnzH0TIghPMTetUMKS8u5442VHCsq4+2fXUh8bBNb46kPTRBK+aPUDNjwDuSshgSXpf/AcnATrJkF6+ZCUR40aw8XPQS9J50x3Xu/Tk14dUoLDh0vZvnuPGZv68L1Wx7g8rX3cv+Ke1gUNJi+HZszMKkFgzq3oE+H5jQJ813CqKg03D9nLZty8nn5tv70aB/js3t7gyYIpfxRl0tBgq1v04GaIIrzYcM8KzHkrLZGkXcbB31udYwqr/6DvXV0BOPT4xmfHg+F31A+83pe3P8sCzpG8NLJYfz7qx088yWEBgu9EmIdCaMl/To1p2m49z72/rBwM19sOcgTV3Zn5AVtvHYfX9FGaqX81WuXWz2Zfr7E7kg8p7IS9iy1qpA2fwjlxdC6O/S9FXreAFHnuVhO6Ul4+zbI/AIu/S0FAx5g1Z5jfL87l+W789iQnU95pSE4SOgR38xKGEktGZDYwmPTXby+dDe/+2gzU4cm8dgVaR65pi/U1EitJQil/FXqWPj8McjPhhj/WyugTvL3wbo3rdLC0SwIbwa9b7ZKC/F9Tq+per7ComDiXPjgbvjqDzQ7eYQRY//MiG6tAatdYPWeY/ywO5cfdufxxnd7ePnb3YhAt7bNGJTUgkFJLRiQ1OK8Zpv9YvNBnlywmdFpbfjNuAvq97v4ES1BKOWvDm+DaQNh3NMw4A67o6m78lLY9rGVFHZ+aS2IlHiRlRQuuALCvDAXUWUlfPYb+P556HEdTHgBQsLOOay4rIJ1Px7jh915LN+dx6o9RylyLG7UpVUUgzq3dCSNlrSNiajxlhv35XP9i9+R0qYpc+8cTGRYw/rerUuOKtUQGQPP9oG4VJj0tt3RuO/gZqsKaf1bUJgL0fHQZ5JVYvDFmAVjYMk/4csnrLacG2ZCeNMaTyktr2RjTj4/7Mpj+e5cVmYd5XhJOQCdWkYyMLHFqaSR0LzJqQFvOceKmDBtKaHBQbx/zxBaR9ecTPyRJgilGqpPHoFVr8GvdnvnG7enFOfDxndh9UynBufLHQ3Ol9bY4Ow1q/8DHz1gVWHd/E6d2jcqKg1b9hfw/S6rDWN5Vh7HCq1V8eJjIhiY1IKBSS35z3dZ7DtaxLyfD6Fr22gv/SLepQlCqYZq59cwcwJMfAu6ZtgdzZmMgawlVhXS5g+hvAhap1lJIf0GiIqzO0LYsgDmTYXYjtao69gO53WZykrDjkMnTrVh/LArjyMnSggJEl67fQAXpbTycOC+Y1uCEJEM4BkgGHjFGPPUWfvvAu4BKoATwJ3GmM0ikghsAbY5Dv3eGHNXTffSBKECUnkp/DUJel4HVzxjdzSWghxYOxvWzIaju60G5x7XWj2R4vvWv8HZ07KWwpybrPU2bnkPWner9yWNMew+cpLSikq6tW3mgSDtY0uCEJFgYDswGsgGVgATjTGbnY5pZowpcDy/ErjbGJPhSBALjDE93L2fJggVsN66FbJXwINbPPfhW1EGpSes7qGlJ63nJWe9dvX8eI5VajjV4HwLXHClf1d/ARzYALOutUZqT3oHOgy0OyK/YVc314FApjFmlyOIucBVwKkEUZUcHKKAwKjvUsqTUjNgy3zI/BJi2p/7oV1yvJoPdufXZ/2sKHX//iFNrG6kYVEQ0QyGPWg1OLfs4r3f2dPa9oSpn8Ksa+CNK+HGmdZ0JqpG3kwQ7YEfnV5nA4POPkhE7gEeBMKAS512JYnIGqAA+K0x5lsX594J3AnQsWNHz0WulD9JGQMSBLOvrf3Y0MjTH+Zh0Y4P9RhoFg9hTZ32OZ6HNz3ztavndjQwe0OLJEeSuNaqcrrqeeh1o91R+TXbO+waY6YB00TkZuC3wGRgP9DRGJMrIv2AD0Sk+1klDowx04HpYFUx+Th0pXyjaSurWqRgv4sPd6cP89DIwPkw95amrWHKQph7M7x/JxQegQvvsTsqv+XNBLEPcO4ykODYVp25wAsAxpgSoMTxfJWI7ARSAW1kUI1T8ii7IwgcEc1g0jx476fw6f/CycMw8nH/a1z3A96cR3gFkCIiSSISBtwEzHc+QERSnF6OA3Y4trdyNHIjIp2BFCBwV3FXSvlWaARc/zr0u90aVDf/Xqgotzsqv+O1EoQxplxE7gU+xermOsMYs0lEngRWGmPmA/eKyCigDDiKVb0EMBx4UkTKgErgLmNMnrdiVUo1QkHBMP6fVrXT4r9AYR5cNwNCG+76DZ6mA+WUUuqH6fDJr6DjhTBxDjSJtTsin9ElR5VSqiaD7oRrX7HGm7w+Do4fsDsiv6AJQimlwBqtPultyNsNr46B3J12R2Q7TRBKKVWly6Uw5SNrQOGMsZCz1u6IbKUJQimlnLXvZw2oC4mA18fDrsV2R+SaMdbiS5s/hE0feOUWtg+UU0opvxOXAnd8BjOvgdnXWe0TaVfZF09lhVXltX8dHFhn/dy/zppmHaBND+g+weO31QShlFKuNIuH2z+2puV4ezKMfxr6T/X+fSvK4PDW00lg/3prssGyk9b+4HBo0x26XwPtekG7dGtdby/QBKGUUtWJbAG3fgDvTIEF/wMnj8Dwhz036rqsyFqBb/9aR+lgPRzcdHoyxdAoKwH0vdVKBm3ToVVXCA71zP1roQlCKaVqEhYJN82G+ffB13+EE4fgsr9CUB2bcEuOWyUB55LB4a1grLWwiYi1ksCguxwlg97WEq11vY8HaYJQSqnaBIdas79GxcGyf1trbV/9EoSEuT6+MM8pEThKBrmZp/c3bWMlgW6XO5JBL4jp4HfzQWmCUEopdwQFwZg/QFQr+PwxKDpqrStRevLMZLB/HeQ7rXQQ29GqGkq/6XSbQXRb+36POtAEoZRSdTH0AYiMs6qc/pYM5cWOHQItk6HDIBj409NtBpEtbA23PjRBKKVUXfWZZJUCtnwErbo5kkEPa93rAKIJQimlzkfySOsRwHQktVJKKZc0QSillHJJE4RSSimXNEEopZRySROEUkoplzRBKKWUckkThFJKKZc0QSillHJJjDF2x+ARInIY2FOPS8QBRzwUTkOn78WZ9P04k74fpwXCe9HJGNPK1Y6ASRD1JSIrjTH97Y7DH+h7cSZ9P86k78dpgf5eaBWTUkoplzRBKKWUckkTxGnT7Q7Aj+h7cSZ9P86k78dpAf1eaBuEUkopl7QEoZRSyiVNEEoppVxq9AlCRDJEZJuIZIrII3bHYycR6SAiX4vIZhHZJCIP2B2T3UQkWETWiMgCu2Oxm4jEisg8EdkqIltE5EK7Y7KTiPyP4+9ko4jMEZEIu2PytEadIEQkGJgGXAakARNFJM3eqGxVDjxkjEkDBgP3NPL3A+ABYIvdQfiJZ4D/GmO6Ab1oxO+LiLQH7gf6G2N6AMHATfZG5XmNOkEAA4FMY8wuY0wpMBe4yuaYbGOM2W+MWe14fhzrA6C9vVHZR0QSgHHAK3bHYjcRiQGGA68CGGNKjTHH7I3KdiFAExEJASKBHJvj8bjGniDaAz86vc6mEX8gOhORRKAP8IO9kdjqX8CvgEq7A/EDScBh4DVHldsrIhJld1B2McbsA/4O7AX2A/nGmM/sjcrzGnuCUC6ISFPgXeAXxpgCu+Oxg4iMBw4ZY1bZHYufCAH6Ai8YY/oAJ4FG22YnIs2xahuSgHggSkRusTcqz2vsCWIf0MHpdYJjW6MlIqFYyWG2MeY9u+Ox0VDgShHJwqp6vFREZtkbkq2ygWxjTFWJch5WwmisRgG7jTGHjTFlwHvAEJtj8rjGniBWACkikiQiYViNTPNtjsk2IiJYdcxbjDFP2x2PnYwxjxpjEowxiVj/L74yxgTcN0R3GWMOAD+KSFfHppHAZhtDstteYLCIRDr+bkYSgI32IXYHYCdjTLmI3At8itULYYYxZpPNYdlpKHArsEFE1jq2/a8x5mMbY1L+4z5gtuPL1C7gdpvjsY0x5gcRmQesxur9t4YAnHZDp9pQSinlUmOvYlJKKVUNTRBKKaVc0gShlFLKJU0QSimlXNIEoZRSyiVNEEr5ARG5RGeMVf5GE4RSSimXNEEoVQcicouILBeRtSLykmO9iBMi8k/H2gBfikgrx7G9ReR7EVkvIu875u9BRJJF5AsRWSciq0Wki+PyTZ3WW5jtGKGrlG00QSjlJhG5ALgRGGqM6Q1UAJOAKGClMaY7sBh43HHKf4BfG2PSgQ1O22cD04wxvbDm79nv2N4H+AXW2iSdsUa2K2WbRj3VhlJ1NBLoB6xwfLlvAhzCmg78Lccxs4D3HOsnxBpjFju2vwG8IyLRQHtjzPsAxphiAMf1lhtjsh2v1wKJwBLv/1pKuaYJQin3CfCGMebRMzaK/N9Zx53v/DUlTs8r0L9PZTOtYlLKfV8C14lIawARaSEinbD+jq5zHHMzsMQYkw8cFZGLHNtvBRY7VurLFpEJjmuEi0ikT38Lpdyk31CUcpMxZrOI/Bb4TESCgDLgHqzFcwY69h3CaqcAmAy86EgAzrOf3gq8JCJPOq5xvQ9/DaXcprO5KlVPInLCGNPU7jiU8jStYlJKKeWSliCUUkq5pCUIpZRSLmmCUEop5ZImCKWUUi5pglBKKeWSJgillFIu/T9237M+xiScJgAAAABJRU5ErkJggg==\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.plot(history.history['val_loss'])\n",
        "plt.plot(history_jig.history['val_loss'])\n",
        "plt.title('Test Loss')\n",
        "plt.ylabel('loss')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['model', 'model_jig'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "BFP2pO95kw71",
        "outputId": "4985b81a-fd85-46c1-ff07-7dd01c33008e"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.plot(history.history['recall'])\n",
        "plt.plot(history_jig.history['recall_1'])\n",
        "plt.title('Train recall')\n",
        "plt.ylabel('recall')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['model', 'model_jig'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "v3YI7I7ukxyh",
        "outputId": "c6b22824-d984-4ef0-e2e4-12eccb3ddd86"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.plot(history.history['val_recall'])\n",
        "plt.plot(history_jig.history['val_recall_1'])\n",
        "plt.title('Test recall')\n",
        "plt.ylabel('recall')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['model', 'model_jig'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "0BUq-GgHkyuy",
        "outputId": "efc3e196-e969-4449-bbc9-01611837bb78"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.plot(history.history['f1_score'])\n",
        "plt.plot(history_jig.history['f1_score'])\n",
        "plt.title('Train f1_score')\n",
        "plt.ylabel('f1_score')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['model', 'model_jig'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "HyeDPujfkzlI",
        "outputId": "559c0e7c-2e00-47d4-d695-10a55900978b"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.plot(history.history['val_f1_score'])\n",
        "plt.plot(history_jig.history['val_f1_score'])\n",
        "plt.title('Test f1_score')\n",
        "plt.ylabel('f1_score')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['model', 'model_jig'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "nIUaxJLDk9oJ",
        "outputId": "e191ec12-1f47-405d-fd1a-effbe93a46a9"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZYAAAEWCAYAAABFSLFOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdeXhV1bn48e+bOSEJZA6QQMAwSGQOgwPK4IBTqeitUrWordxba7Xea6u23traWuu9tv3Vqr11oGIVcWoVWwUVRUBUCAhIQCAMkQAZSAKZSXLy/v7YO3AMgeRADich7+d5zpN91p7efdDznrXW3muJqmKMMcZ0lKBAB2CMMeb0YonFGGNMh7LEYowxpkNZYjHGGNOhLLEYY4zpUJZYjDHGdChLLCbgROQdEZkd6DjaQ0SqRGSguxwpIm+JyEEReVVErheRd9txjJ+KyDP+j7bzE5FJIrLlOOv7uZ95cDuO9X8i8t8dG6E5EWLPsZgTISJVXm+jgEOAx33/76r64qmP6tQSkRuBHwLnqGrjCR4jA9gJhJ7oMdp5nqXAC6raqROaiOwCvqeq75/kcSbjXG9aR8RlfBMS6ABM16Sq0c3Lx/syEJEQf35hBlh/YOtpfH3GnBhVtZe9TuoF7AIudJcnAwXAPUAh8DcgDvgnUAKUu8tpXvsvxUlMADcBK4BH3W13Apce59z3AHuASmALMM0t/wXwGvCyu24tMNJrvz7A625MO4E7vNYFAz8Ftrv7rgHS3XUKZAK/BOqBBqAK+G5z7F7HyQLeA8qAIuCnXrG94C5/5R6zyn1d4G4/3Os4yUANkHScz+GYnzHwEE5tss49x+PHOY4AfwCKgQrgC+Asd124++/ylXs9/wdEtvh3/y93333AzV7HvQzY5H6ee4C7vfdzl/8GNAG1bpw/ATLczycEuBbIaRHvXcBCd/k54NdAD/cYTV6fax/3M0zw2neM+3mFBvr/odPtZX0sxh9SgXicX/RzcPry/uq+74fzP/3jx9l/Ak6SSAT+B3hWRKTlRiIyBLgdGKeqMcAlOEmu2QzgVTeW+cAbIhIqIkHAW8B6oC8wDfiRiFzi7vefwCycL8NY4BacL6XDVPUB4DfAy6oararPtogtBngfWITzpZYJLGnlWs93//Zyj/MRsAC4wWubWcASVS1pZf9mx/yMVfVnwHLgdvcctx/nOBe7MQ0GegLfAkrddb91y0e519MX+LnXvqnuPn1xEu0TIhLnrnsWp4k0BjgL+KDliVX1RpykdaUb5/+02OQtYIiIDPIq+zbOv633caqBS4G97nGiVXUvzg+Yb3lteiOwQFUbjvN5mBNgicX4QxPwgKoeUtVaVS1V1ddVtUZVK3F+QV9wnP3zVfVpVfUA84DeQEor23lwfkUPE5FQVd2lqtu91q9R1dfcL47fAxHARGAczq//B1W1XlV3AE8D17n7fQ+4X1W3qGO9qpbimyuAQlX9narWqWqlqn7Wzn3nAbO8kumNOL/mj+kEPuNjaQBigKE4fbCbVXWfG8sc4C5VLXPP8RuOfGbN+z6oqg2q+jZOTWGI17phIhKrquWqutbXwFS1BngTJ9HiJpihwMJ2HmIebsJ2bwaYRRufqzkxlliMP5Soal3zGxGJEpG/iEi+iFQAy4Bex7nTp7B5wf0yAYhuuZGq5gE/wmlaKhaRBSLSx2uT3V7bNuE01fTB+VXfR0QONL9wmr6ak1c6TjPYyTjhY7gJqAaYLCJDcWoHx/3yPIHP+Fjn/gCnpvMEzmf6lIjEAkk4N2ms8frMFrnlzUr16/1NNRz5d7sapwaYLyIficjZvsTlZT5uYsGprbzh9d9IW97ESW4DgIuAg6q66gTjMMdhicX4Q8tbDf8L55frBFWN5Ujzz1HNWz6fSHW+qp6HkywUeMRrdXrzgtv8lQbsxUk4O1W1l9crRlUvczffDZxxkqHtBga25xKOUd786/pG4DXvRH0MbX3G7b79U1UfU9WxwDCcpq8fA/txmteyvD6znup1E0cbx1ytqjNw+oveAF451qZtHOo9IElERuEkmPnH2O6o47if4Ssc+VyttuInlljMqRCD86V0QETigQc64qAiMkREpopIOE7HdHOHbbOxIjJTREJwajaHgE+BVUCliNzjPosSLCJnicg4d79ngF+JyCBxjBCRBB/D+yfQW0R+JCLhIhIjIhNa2a7EjbllEnoBuArnS/D5dpyvrc+4qJVzHEVExonIBBEJBapxPtcmt8b3NPAHEUl2t+3r1S91vGOGuc/49HSbJSv4+r9Tu+N0938V+F+cvrP3jnOcBBHp2aL8eZybLL6BJRa/scRiToX/B0Ti/Or9FKcJpSOE43Qo78dpPksG7vNa/ybOnUTlOL9QZ7rt/x6cPpBROHeE7cdJJs1fQr/H+WX7Ls6X4LNu/O3m9kFcBFzpxrYNmNLKdjU4/SEfu01ME93y3Th3silOx3tb2vqM/whcIyLlIvLYcY4Ti5NAyoF8nI77/3XX3QPkAZ+6zW3vc6QPpS03Arvc/f4DuP4Y2z0M3O9+FncfY5v5wIXAq3qMW71V9UvgJWCHe6w+bvnHOEltrarmtzN24yN7QNKclkTkF0Cmqt7Q1radlYjMxbmz6f5Ax3I6EZEPgPnayR8W7crsAUljOiH3ifyZwOjARnJ6cZs7x+Dcim78xJrCjOlkRORXwEbgf1V1p1f5T91xs1q+3vHx+JOOcZyqtvfuukRkHk7z3Y/cpkrjJ9YUZowxpkNZjcUYY0yH6vZ9LImJiZqRkRHoMIwxpktZs2bNflVNam1dt08sGRkZ5OTkBDoMY4zpUkTkmLdrW1OYMcaYDmWJxRhjTIeyxGKMMaZDdfs+ltY0NDRQUFBAXV1b4/6ZtkRERJCWlkZoaGigQzHGnCKWWFpRUFBATEwMGRkZtDK/lGknVaW0tJSCggIGDBgQ6HCMMaeINYW1oq6ujoSEBEsqJ0lESEhIsJqfMd2MJZZjsKTSMexzNKb7saYwY4w5nahCQy3UV8GhSvdv1ZH33mWZ0yB9fIeHYImlG2h+CDQxMfGktjHG+EmTp+0kUF/pta4KDlW02N5rG/W077wRPS2xGGNMp6cKdQegsgiqvF6VhVBVDFWFUHvg60mhoaZ9xw4KQcOiITwaDYuBMOdvU3RvNDSaprAeeEJjaArtgSe0B54Q529jSA88odE0hkbTGBJFQ3APGkOi6N2rB62OyXKSLLF0Urt27WL69OlMnDiRlStXMm7cOG6++WYeeOABiouLefHFF8nMzOSWW25hx44dREVF8dRTTzFixAhKS0uZNWsWe/bs4eyzz8Z7BOsXXniBxx57jPr6eiZMmMCTTz5JcHBwAK/UmC7C0wDVJV9PEEclD/ev59BRuzcGhXMwOJ799KJco6mmFzVEUk0EVSGR1GgElURSpZFUEUFlUwRVGkGFRlDRFEGlhlPbFEJTzYn0WzbhTIZa8bXS/75iGN89r+Pv2LTE0oZfvpXLpr0VbW/og2F9Ynngyqw2t8vLy+PVV19l7ty5jBs3jvnz57NixQoWLlzIb37zG9LT0xk9ejRvvPEGH3zwAd/5zndYt24dv/zlLznvvPP4+c9/zr/+9S+effZZADZv3szLL7/Mxx9/TGhoKLfddhsvvvgi3/nOdzr0+ozpUg5VOomistBJFoeXWySPmlKcmaJbiIxHo1Ooj0ziQNxoinr2Ir8+hq01UWw8GMlX9dEUaxyVRBLfI5zM5GgSeoQRFCQEiRAsEBQkBIsQHCQEBQm9REhoXh8EQSKHtzmybYv9RAgSDh/j69t+fT8RZ3lwSoxfPlJLLJ3YgAEDGD58OABZWVlMmzYNEWH48OHs2rWL/Px8Xn/9dQCmTp1KaWkpFRUVLFu2jL///e8AXH755cTFxQGwZMkS1qxZw7hx4wCora0lOTk5AFdmjI+aPNB4CBrrwFPv/G2sb/H+kPPyHDqyfPi9u33dwaOTR0P10ecLCoXoFIhOhrj+kD4OolPR6BTKguLYeSiarVVRbCgP48v99eQVV1F1qPHw7onRYWQmRzP4jBimJkeTmRzD4JRoEqLDT+GHFjiWWNrQnpqFv4SHH/mPMCgo6PD7oKAgGhsbfX6aXVWZPXs2Dz/8cIfGacxhnkao2e82DRUf+Vtb1sqXffNyOxJDU2Pb526P8Fg3YaRA3zFHlmNSnSQSnQoxqTSF92TPwUPkFVexrbiSrUVVbPuqiryiSqrrPUA9UE9idDiDkqO5ekxfMlNiGJwczaCUGOJ7hHVMvF2UJZYubNKkSbz44ov893//N0uXLiUxMZHY2FjOP/985s+fz/33388777xDeXk5ANOmTWPGjBncddddJCcnU1ZWRmVlJf379w/wlZhOrakJasu9+hKKobq4RfIoOW5zUb2E0xgUhicoHE9QKE1B4WhwGBocASFhEBKOhMQQFBGOhEQSHBZOcGgkIWHhhIRHERQSDs2v4DAIifB677Wu5fvgFvu1eK6qqUnZc6DWSR4FVWwrqmJb8SbyiquoqT9yZ1VSTDiDU6L5t+x0BqVEMyg5hkHJ0cR18wRyLJZYurBf/OIX3HLLLYwYMYKoqCjmzZsHwAMPPMCsWbPIysrinHPOoV+/fgAMGzaMX//611x88cU0NTURGhrKE088YYmlO1J1blc9nBi8kkNzWXWxu1zc+u2rIRHur/wUiB9Abe9sCuqj2VIdxbryMNaWhlGsPSmjJwlxvahvbKKm3kNtvYfGJt+mRA8LCSIqLJio0GAiw4KJCgtx/zqvyNCQI8thwUSFQWSYh6jQQ0SFNRIRVkdUaDCVdY1sK65iW1El24qryCuuorbhyLWlxIYzKDmGa8elO8kjJZpBydH0irIE4otuP+d9dna2tpzoa/PmzZx55pkBiuj0Y59nAHgaoXAD7P382ImjlTuXCAqBHslHEkZ00pHmouhkd10KGp1EfmUwq/PLWb2rjNW7ytm53+mrCA8JYnS/XozPiGfcgHhG94sjOvzrv2HrG5uorfdQ09B4ONnU1HuoqW88stzgoba+5XoPte4+R8rdfRqc94cam9r8eFJjI47UPFKiGZwSTWZSDD2jbLDU9hKRNaqa3do6q7EYczrwNMC+9bBrBeR/DF996tRIABDokegkhx5JkJDplTi8k0gKRPSCoKNHevI0KZv3VZDjJpHVu3ZRXOkkpp6RoYzLiOO6cemMGxDPWX16EhZy/NGiwkKCCAsJoicd/0XuaVJqG1okKTcJRYYFMyglmtgISyD+ZInFmK6osd6pjeSvgF0fw+7PnAftABIGwVkzof950G8CxPSBYN/+V69r8LB+9wFy8stZtbOMtfnlVLp3PfXpGcHZZyQwLiOe8QPiyUyKJiio84wJFxwkRIeHHFVLMqeO3z95EZkO/BEIBp5R1d+2WN8fmAskAWXADapa4K57BLjc3fRXqvqyWz4VeBQIA9YA31XVRhGZDLwJ7HT3+buqPujHyzPm1Gg8BHvWOEkkfwXsXnXkae2koTDiWsg4D/qfCzEpPh/+YG0Da/Ld2sjOMjYUHKTe4zQpDU6J5spRfQ43bfXtFdmRV2ZOQ35NLCISDDwBXAQUAKtFZKGqbvLa7FHgeVWd5yaMh4EbReRyYAwwCggHlorIO0AVMA+YpqpbReRBYDbwrHu85ap6hT+vyxi/a6iDgtVOs9auFc5yozv9QMpZMPpGyDjXSSQ9fB/frfBgHat2lZGzq4xVO8vYUlSJKoQECcPTenLzuRlkZ8ST3T/O7nwyPvN3jWU8kKeqOwBEZAEwA/BOLMOA/3SXPwTe8CpfpqqNQKOIbACmu9vUq+pWd7v3gPs4kliM6Xrqa6BglVsj+dhJJJ56QCB1OGTf4iSR/udAVLxPh1ZVtpdUu53szmt3WS0AUWHBjO0fx2XDe5OdEcfo9Dgiw2yIH3Ny/J1Y+gK7vd4XABNabLMemInTXHYVECMiCW75AyLyOyAKmIKTkPYDISKSrao5wDVAutfxzhaR9cBe4G5VzW0ZlIjMAeYAh2/FNeaUOlTl9Ivkf+wkkz1roKkBJAh6j4Txc5ymrX4TITKu3YdtalJ2l9fwZWElWwor2bjnIGvyyymtrgcgoUcY4zLiuemcAYzLiGNY71hCgm1aJtOxOkPv1t3A4yJyE7AM2AN4VPVdERkHrARKgE/cchWR64A/iEg48C7QfCP6WqC/qlaJyGU4tZ9BLU+oqk8BT4Fzu7Ffr84YgLoKJ5HsWuG89q1zniaXYOgzGs6+DTImQfoEiIht1yHLq+vdBFLBl4WVfFlYydaiysMP9olA//goJg9JZvyAOLIz4hmY2MMmXzN+5+/Esoev1ybS3LLDVHUvTo0FEYkGrlbVA+66h4CH3HXzga1u+SfAJLf8YmCwW17hddy3ReRJEUlU1f1+ubouoiPmYznnnHNYuXIle/fu5Y477uC1117zV7inh9oDzi2/+c2JZD1okzMGVd8xcO6dTtNW+gQIjz7uoQ41esgrrmKLWwvZ7CaTooojz6HERYUyJDWGb2WnMzQ1hqG9YxmcEk1UWGf47Wi6G3//V7caGCQiA3ASynXAt703EJFEoExVm3D6Sua65cFAL1UtFZERwAic2gkikqyqxW6N5R6OJJ9UoMit1YzHmXq51M/X2C2sXLkSgD59+lhSaU2Tx7n9N28J5L0Pe3KcRBIcBmnjYNLdTmd72ngIi2r1EKrO8CJf7qtkS5FTA/lyXwU79lfjcZ9UDwsOIjM5mnMzExmaGsOQ1FjOTI0hKSbcaiKm0/BrYnFvAb4dWIxzu/FcVc117+TKUdWFwGTgYRFRnKawH7i7hwLL3f9ZKnBuQ24eie7HInIFTuL4s6p+4JZfA3xfRBqBWuA6PdmhBd65Fwq/OKlDHCV1OFz62+Nu0tnmY4mOjqaqqopdu3ZxxRVXsHHjRmpqarjpppvYuHEjQ4YMYe/evTzxxBNkZ7f6MO7pp7LwSCLZ8aEznhbi1Egm3Q0Dzoe0bAg9+vbcg7UNbg3EqxmrsPLwsyIAaXGRDE2N4ZKsVIakxnBm7xgyEnpYn4jp9PxeT1bVt4G3W5T93Gv5NeCon8CqWodzZ1hrx/wx8ONWyh8HHj/JkDuNzj4fy5NPPklcXBybNm1i48aNjBo1qiMvv/NpPOQ0b21f4iSUoo1OeXQKDL7UmT984BTokXB4lwZPEzsKK/nSTSBb3FrI3oN1h7eJjQhhaGosV43py5DUGIamxjA4JYYYezrcdFHWANuWNmoW/tTZ52NZsWIFd955JwBnnXUWI0aMOOFjdVplO9xayRLYucyZuyMo1Llb68JfQOaFznMlInialC8LK1i7YRdrvzrA5n0VbC+posHj1BhDg4UzkqIZPyCeIamxbl9IDKmxEdaMZU4rllg6MZuPJQAOVTmd7XnvO69ydxCHuAwYNctJJBmTIDyayroGPv/qAGu+2Maa/HI+/6rcnasDkmPCyeoTy5ShyU4CSY1lQGKPNsfQMuZ0YImlCwv0fCznnnsur7zyClOmTGHTpk188UUH90WdCqpQlOskke1LIP8T53mS0Cinj2TibZA5DY0fSEF5LTn5Zax5Zyc5u8oPP60eJDA0NZaZY9LIzohjTL840uIirRZiui1LLF3YqZyPpbUvydtuu43Zs2czbNgwhg4dSlZWFj179uzYi/SHmjKns725iauq0ClPzoKJ34fMadT3mUBucR1r8stZ8045OflLKHFH840OD2F0v15MPyuV7P7xjEzvaf0hxnix+VhsPpY2lZaWMmbMGPLz879W7vF4aGhoICIigu3bt3PhhReyZcsWwsK+PrZUwD/PJo/zZPvhW4HXAOoMEX/GFMi8kPLek1hTFkFOfjlr88tZX3Dg8Lwe6fGRjO0Xx9iMeMb2i2NIagzBnWg0X2MCweZjMSds7969TJ48mbvvvvuodTU1NUyZMoWGhgZUlSeffPKopBIwFXu9bgVeCnUHnOFS+o5FL7iHgoSz+aSuP6vzK1jzQTk7Spw7vEKDhaw+PblhYn+y+8cxtn8cybERgb0WY7oYSyzmsNLSUqZNm3ZU+SeffEJCQsJR5TExMbSs7QXU/jxY9wJsXQzF7jinMb1pHHwZO3udzUeNWXyyt4k1y8s5UFMNbKJXVChj+8Vxzdg0xvaLY2R6LyJCbRBGY06GJZZjUNVu1/makJDAunXrOvSYfm9qbaiFTQth7TxnQEcJ5lDa2ew468d81DSCd4rjyc2pcOdYL+SMpB5cPCyF7P7xjOkfxxlJNnaWMR3NEksrIiIiKC0tJSEhwb50ToKqUlpaSkSEH5qSinJhzTzY8DLUHUDjBrBx6I94sGA0q7c5HenhIUGMTAvm1vMHkt0/jtH94oi3uUWM8TtLLK1IS0ujoKCAkpKSQIfS5UVERJCWltYxBztUBbl/dxLKnhwIDqNp6BUsi76cBzbEkb+ujmG9Y7n/8r6M7R9HVjvmXjfGdDxLLK0IDQ1lwIABgQ7DgPOcyd7PnaauL16H+kpIHELjRQ/xpk7iDx+XUlBey1l9w3j6yrO48Mxkq2UaE2CWWEznVHcQNrziJJTCLyAkErKuon7UDbxS2Jc/f7SDPQcKGJHWk19+I4upQy2hGNNZWGIxnYeqMxnWmnmQ+w9orHVGgr7sUerOvJpXN1bw5ILt7DuYy6j0Xvz6qrOYPDjJEooxnYwlFhN4NWWw/iVY+zyUfAlh0TDyWhgzm7qkESxYvZv/+9PnFFbUMbZ/HI9cPYJJgxItoRjTSVliMYHR1AS7ljtNXZvfAk899M2Gb/wJsmZSK5HMX/UVf3luKcWVhxifEc/vvjWSc86wO/WM6ewssZhTq7II1r3o1E7Kd0JETxh7M4ydDSlZ1NQ38uKnX/GXZTvYX3WIiQPj+eN1ozn7jKMf0DTGdE6WWIz/NXlg+wew5jnYugiaGqH/eTDlp3DmlRAaSfWhRl74aDtPLdtBaXU952Ym8MTU0UwYaAnFmK7G74lFRKYDf8SZmvgZVf1ti/X9cea5TwLKcKYgLnDXPQJc7m76K1V92S2fCjwKhAFrgO+60yCLe67LgBrgJlVd6+dLNMdysAA+f8F5HdwNUYnOMPRjZkNiJgBVhxp5fmkezyzfSVl1PZMGJXLntEFkZ8QHOHhjzInya2IRkWDgCeAioABYLSILVXWT12aPAs+r6jw3YTwM3CgilwNjgFFAOLBURN4BqoB5wDRV3SoiDwKzgWeBS4FB7msC8Gf3rzlVPA3OWF1r5zkDQKo6Iwhf/GsYchmEOE++V9Y1MG/lLp5ZsZMDNQ1cMDiJO6YNYmz/uABfgDHmZPm7xjIeyFPVHQAisgCYAXgnlmHAf7rLHwJveJUvU9VGoFFENgDT3W3qVXWru917wH04iWUGTpJS4FMR6SUivVV1n9+u0DgOFsDqZ53+k6oiiOkNk/4LRt8IcUfmezlY28BzH+/i2RU7qKhrZOrQZO6YNohR6b0CGLwxpiP5O7H0BXZ7vS/g6BrEemAmThPWVUCMiCS45Q+IyO+AKGAKTkLaD4SISLaq5gDXAOnHOV9f4GuJRUTmAHOAw5NgmZNQ/CU8dxnUlsOgS5yO+MyLIPjIf14HaxqY+/FO5n68k8q6Ri48M4U7pw1ieFoXmBjMGOOTdicWEbkS+JeqNnVwDHcDj4vITcAyYA/gUdV3RWQcsBIoAT5xy1VErgP+ICLhwLuAx5cTqupTwFPgTPTVYVfSHZVuh+dnQFAI3PYZJA3+2uoDNfU8u2Inz328i8pDjVySlcIPpw7irL6WUIw5XflSY7kW+H8i8jowV1W/bMc+ezhSmwBIc8sOU9W9ODUWRCQauFpVD7jrHgIectfNB7a65Z8Ak9zyi4Hmb7M2z2c60IHdTlLx1MPNb38tqZRV1/PM8h3MW7mL6noPlw1P5YdTB3Fm79gABmyMORXanVhU9QYRiQVmAc+JiAJ/BV5S1cpj7LYaGCQiA3C+4K8Dvu29gYgkAmVuTeg+nDvEmjv+e6lqqYiMAEbg1E4QkWRVLXZrLPfgJh9gIXC725czATho/St+UlnkJJW6Cpi9EJKdqYdLqw7x1PId/O2TfGobPFw+vDc/nDqIIakxAQ7YGHOq+NTHoqoVIvIaEAn8CKdP5Mci8piq/qmV7RtF5HZgMc7txnNVNde9kytHVRcCk4GH3US1DPiBu3sosNx9yroC5zbkRnfdj0XkCiAI+LOqfuCWv41zq3Eezu3GN/tyfaadasrgb9+EykK48R/QZxT7qw7x1DInoRxq9HDlyD7cPiWTQSmWUIzpbqS9M/yJyAzgJiATeB6Y59YaooBNqprhryD9KTs7WzvV9LqdXd1BmPcNKN4M178KAy9gR0kVs57+lJLKQ8wY1Zfbp2ZyRlJ0oCM1xviRiKxR1ezW1vlSY5kJ/EFVl3kXqmqNiHz3ZAI0XUR9Nbz4LWf2xutehIEXsL2killPfYqnSXnrh+eR1cc65Y3p7nyZXq+wZVJxn4xHVZd0aFSm82mogwXfhoJVcPUzMPgS8oqdpNKkyoI5Ey2pGGMA3xLLRa2UXdpRgZhOzNMAr86GHUthxpOQ9U0nqTz9KU0KL9060fpSjDGHtdkUJiLfB24DznCffm8WA3zsr8BMJ9Hkgb/f6gweefnvYNQs8oorue6pzwBYMGcCmcmWVIwxR7Snj2U+8A7OGF73epVXqmqZX6IynUNTEyz8oTOb40W/gnHfY1tRJbOe/gwRp6aSmWyd9MaYr2tPYlFV3SUiP2i5QkTiLbmcplThnZ84Y39Nvg/OvYOtRZV8++lPERFLKsaYY2pvjeUKnOHpFfCevk+BgX6IywSSKrz/AKx+Gs75IVxwD1uLKpn11KcEBwkvzZlotxMbY46pzcSiqle4fwf4PxzTKSx7FD7+I2TfAhf9ii1FVXz76U8JCXZqKgMtqRhjjqPdd4WJyEIRmeU+EGlOV588AR/+GkbOgst+x5dFlcxyk8qCOWdbUjHGtMmX241/hzPw42YReU1ErhGRCD/FZQIh56+w+KcwbAZ843E2F1Xx7ac/Iyw4iAVzzmZAYo9AR2iM6QJ8GYTyI+Ajd3DIqcCtOANG2nC1p4P1L8M/74JBF8PMZxKya0MAACAASURBVNhUVMP1z3xKRGgwL906kQxLKsaYdvJpEEoRiQSuxBlCfwzOFMGmq9u0EN74PgyYBN96nk3FdYeTyoI5E+mfYEnFGNN+vkz09QrOVMOLgMeBj/ww6Zc51ba9B6/dAn3HwnUvkVtSz/XPfEZUaDAvWVIxxpwAX2oszwKzVNWn2RpNJ7ZzObx8gzOXyvWvsnG/h+uf+Yzo8BBeunUi/RLsPg1jjO/aM6TLVHe+kx7ADHd+lMNU9e9+is340+7VMP9aiMuAG99gY5kcTioL5kwkPd6SijHmxLSnxnIB8AFO30pLClhi6Wr2bYAXr4boZPjOm3xRHsINz1pSMcZ0jPY8IPmAu/igqu70XudOOWy6kpItzuyPYTEweyEbDkZwwzOfEhsZyku3WlIxxpw8X55jeb2Vstfa2klEpovIFhHJE5F7W1nfX0SWiMgGEVkqImle6x4RkY3u61qv8mkislZE1onIChHJdMtvEpESt3ydiHzPh+s7/ZXtcGZ/lGAnqVTFcsMznxEbGWo1FWNMh2lPH8tQIAvoKSIzvVbFAsd9QNJ95uUJnLlcCoDVIrJQVTd5bfYo8LyqzhORqTijKN8oIpfj3NI8CggHlorIO6paAfwZmKGqm0XkNuB+nGmTAV5W1dvbuq5u52ABzJsBnkNw09usr0nghmc/o1eUU1NJi7OkYozpGO3pYxmCMwhlL77ez1KJ85Dk8YwH8lR1B4CILABmAN6JZRjwn+7yh8AbXuXLVLURaHTngpkOvILTt9P8YGZPYG87rqP7qixyaip1B2D2QtbV9+HGZz4jrkcYL82ZSN9ekYGO0BhzGmlPH8ubwJsicraqfuLj8fsCu73eFwATWmyzHpgJ/BG4CogRkQS3/AER+R0QBUzhSEL6HvC2iNQCFcBEr+NdLSLnA1uBu1TV+/wAiMgcYA5Av379fLykLqamzOlTqdwHN/6Dzxsz+M6zTlJZMGcifSypGGM6WJt9LCLyE3fx2yLyWMtXB8RwN3CBiHyOcwfaHsCjqu8CbwMrgZeAT4DmZ2juAi5T1TTgr8Dv3fK3gAxVHQG8xzFGBlDVp1Q1W1Wzk5KSOuASOqm6CnhhJpRuh1kvsZYh3PjsKuKjw3j53y2pGGP8oz1NYZvdvzkncPw9QLrX+zS37DBV3YtTY0FEooGrVfWAu+4h4CF33Xxgq4gkASNV9TP3EC/jjAaAqpZ6HfoZ4H9OIObTQ301zP8WFH4B177AmuCRzH52FYnRTvNX756WVIwx/tGeprC33L+Hf/2LSBAQ7XakH89qYJB7W/Ie4Drg294biEgiUOYOD3MfzsCWzR3/vVS1VERGACOAd93deorIYFXdinNjwGZ3n96qus/d5hscSYrdS0MdLLgedn8GVz/LmogJzJ67iqSYcF66dSKpPW1QamOM//gyVth84D9wmqNWA7Ei8kdV/d9j7aOqjSJyO7AYCAbmqmquiDwI5KjqQmAy8LCIKLAMaJ4CORRY7j7pXwHc4HbkIyK3Aq+LSBNQDtzi7nOHiHwDaATKOHKnWPfhaYDXboYdH8KMJ8mJnszsZ1eRHBthScUYc0qIqrZvQ5F1qjpKRK7HuQ34XmCN25/RZWVnZ2tOzom08nVCTR54/XuQ+3e47FFWJ1/NTXNXkRIbwUtzJpISa0nFGNMxRGSNqma3ts6XByRDRSQU+CawUFUbcG77NZ1BUxMsvMNJKhf+klVJVzN77ipSelpSMcacWr4klr8Au3AGo1wmIv1xmqhMoKnConth3Qtw/k/4rM+N3PTXVfTuGcGCWy2pGGNOLV9mkHwM8L69OF9EpnR8SMZnSx6EVX+Bs2/ns/7/zs3Praa3W1NJjrGkYow5tXzpvA8HrgYyWuz3YAfHZHyRMxdW/B7G3synmXdx83M59I2LZP6tEyypGGMCwpeJvt4EDgJrgEP+Ccf4bM086DOaT878Gbc8l0NaXCTzb51IUkx4oCMzxnRTviSWNFWd7rdIjO+q98O+9Xw14k5unpdDelyUJRVjTMD50nm/UkSG+y0S47vtHwLKf32eSL/4KF6aY0nFGBN4vtRYzgNuEpGdOE1hAmhXf46lS9u+hEqJoTBqKP+4dSKJ0ZZUjDGB50tiudRvURjfqdKUt4SPGs/im2f3s6RijOk02t0Upqr5OANKTnWXa3zZ33Swoo0EVRfzUdNwLslKDXQ0xhhzWLsTg4g8ANyDM1AkOGN5veCPoEw75C0BYFvMeLL6xLaxsTHGnDq+1DiuwhkxuBoOD3cf44+gTNsat73PFk0n+6ws3IE6jTGmU/AlsdSrM2KlAohID/+EZNpUX43s/pSPPCOYfpY1gxljOhdfEssrIvIXoJc7bP37wNP+Ccsc164VBDc1sD5sLGP6xQU6GmOM+Rpfxgp7VEQuwhl4cgjwc1V9z2+RmWNq3Po+DRpGQtYFBAVZM5gxpnPxZaywHsAHqvqeiAwBhohIqDt8vjmFDm15j9VNZ3Lh8P6BDsUYY47iS1PYMiBcRPrizDF/I/CcP4Iyx1GeT4/KnawKHsnEgQmBjsYYY47iS2IRVa0BZgJ/VtV/A7La3ElkuohsEZE8Ebm3lfX9RWSJiGwQkaUikua17hER2ei+rvUqnyYia0VknYisEJFMtzxcRF52z/WZiGT4cH1dgmfb+wDowGmEhdhjRMaYzsenxCIiZwPXA/9yy4Lb2CEYeALnqf1hwCwRGdZis0eB592hYR4EHnb3vRxnCuRRwATgbhFpfmDjz8D1qjoKmA/c75Z/FyhX1UzgD8AjPlxfl3Dgi0Xs1XhGjZkQ6FCMMaZVviSWO3EejvyHquaKyEDgwzb2GQ/kqeoOVa0HFgAzWmwzDPjAXf7Qa/0wYJmqNqpqNbABaB5dWYHmJNMT2OsuzwDmucuvAdPkdHrIw9NAjz0fs1JHcv7g5EBHY4wxrfJlSJdlqvoNVX3Efb9DVe9oXi8if2plt77Abq/3BW6Zt/U4zWvgPIQZIyIJbvl0EYkSkURgCs6QMgDfA94WkQKcvp7ftjyfqjbizB9zVEeEiMwRkRwRySkpKWnfB9AJNO1eTURTNeV9JhEZdtzKojHGBExHNtKfe4L73Q1cICKfAxcAewCPqr4LvA2sBF4CPgE87j53AZepahrwV+D3vpxQVZ9S1WxVzU5KSjrBsE+9os/fxqNC3zE2HqgxpvPyd+/vHo7UMgDS3LLDVHWvqs5U1dHAz9yyA+7fh1R1lKpehDNM/1YRSQJGqupn7iFeBs5peT4RCcFpJiv1y5UFgOYtYYNmct6IQYEOxRhjjsnfiWU1MEhEBohIGHAdsNB7AxFJFJHmOO4D5rrlwW6TGCIyAhgBvAuUAz1FZLC7z0XAZnd5ITDbXb4G57kb9cuVnWJaXUpq9Wby4yYSGxEa6HCMMeaYfJmPpS1HdZKraqOI3A4sxrmDbK7b8f8gkKOqC4HJwMMiojjPyvzA3T0UWO72vVcAN7j9JrhDyrwuIk04ieYWd59ngb+JSB5QhpPITgt71r5DGkqPYRcHOhRjjDkuae8PehH5N1V99VhlInKTqj7X8SH6V3Z2tubk5AQ6jDZtfPIG0ouW0PBfeSTG2vifxpjAEpE1qprd2jpfmsLuO15ZV0wqXYYqqSUfszlyjCUVY0yn12ZTmIhcClwG9BWRx7xWxQKN/grMHFGwdS1pWsaXA6YGOhRjjGlTe/pY9gI5OJN8rfEqr8S57df42Ver3iINyDyn5bOlxhjT+bSZWFR1PbBeRObbSMaBEfXVUr4KTqdfemagQzHGmDb50seSISKvicgmEdnR/PJbZAaAfftLObN+I2WpkwIdijHGtIsvieWvOIM/NuIMr/I88II/gjJHbFy5iHBpIHn0ZYEOxRhj2sWXxBKpqktwblHOV9VfAJf7JyzTrGHLe9QTSp8R0wIdijHGtIsvD0gecp+Q3+Y+9LgHiPZPWAagrLqeQZWr2Bs3hoywqECHY4wx7eLrsPlRwB3AWOAGjgyfYvzg4zXrGBS0h/ChFwU6FGOMabd211hUdbW7WAXc3HK9iPxJVX/YUYEZ2L/+HQBSx1j/ijGm6+gMw+abVlTWNZBa8jEVoUlIcstJN40xpvOySdM7qaWb93GOfEF9/8lwGk2CaYw5/Vli6aS2rP2InlJD/Eib1MsY07V0ZGKxn9UdpK7BQ+RXS2lCCDpjSqDDMcYYn7Q7sYjI8DY2+eNJxmJcK7bt52xdR1XCCIiKD3Q4xhjjE19qLE+KyCoRuU1EerZcacPmd5yPNmxjZNB2m9TLGNMltTuxqOok4HqcOeXXiMh8EbEHLDpYg6eJui1LCEYJHnRhoMMxxhif+dTHoqrbgPuBe4ALgMdE5EsRmXmsfURkuohsEZE8Ebm3lfX9RWSJiGwQkaUikua17hER2ei+rvUqXy4i69zXXhF5wy2fLCIHvdb93Jfr6wxW7SxjbMNaGkJjoG+rk7MZY0yn1u4HJEVkBM6DkZcD7wFXqupaEekDfAL8vZV9goEngIuAAmC1iCxU1U1emz0KPK+q80RkKvAwcKOIXA6MAUYB4cBSEXlHVSvc2lPzOV4H3vQ63nJVvaK919XZLPpiH7cFf0HQwAsg2JcRd4wxpnPwpcbyJ2AtMFJVf6CqawFUdS9OLaY144E8Vd2hqvXAAqDlbFXDgA/c5Q+91g8Dlqlqo6pWAxuA6d47ikgsMBV4w4fr6LSampStuTn0llJrBjPGdFm+JJZ/qOrfVLW2uUBE7gRQ1b8dY5++wG6v9wVumbf1QHNT2lVAjIgkuOXTRSRKRBJxhupPb7HvN4ElqlrhVXa2iKwXkXdEJKu1oERkjojkiEhOSUnJMS/4VPt89wHOqs1x3mTaaMbGmK7Jl8TynVbKbuqAGO4GLhCRz3H6bfYAHlV9F3gbWAm8hNPc5mmx7yx3XbO1QH9VHYlTw2q1JqOqT6lqtqpmJyUldcAldIzFuYVcELwBT3wm9OoX6HCMMeaEtJlYRGSWiLwFDBCRhV6vD4GyNnbfw9drGWlu2WGquldVZ6rqaOBnbtkB9+9DqjpKVS/CeQBzq1dciThNbf/yOlaFqla5y28Doe52nZ6q8sEX+UwM/tKawYwxXVp7eodXAvuAROB3XuWVOP0ex7MaGCQiA3ASynXAt703cL/4y1S1CbgPmOuWBwO9VLXUvXFgBPCu167XAP9U1TqvY6UCRaqqIjIeJ3GWtuMaA27zvkp6H/ycsLB6OMOawYwxXVebiUVV84F84GxfD66qje6kYIuBYGCuquaKyINAjqouBCYDD4uIAsuAH7i7hwLLxRmAsQK4QVUbvQ5/HfDbFqe8Bvi+iDQCtcB1qqq+xh0Ii3ILuSBoAxochmTYQNHGmK5L2vreFZEVqnqeiFQC3hsLoKoa688A/S07O1tzcnICHQaX/GEZz1TfTnp6BsxeGOhwjDHmuERkjaq2+rBde2os57l/Yzo6MOPYUVLFwaJdpEfkQ+Z3Ax2OMcacFF8GoZwoIjFe72NEZIJ/wupeFucWMSn4C+eN9a8YY7o4X243/jPOtMTNqt0yc5IW5xYyo8dmiE6FlFYfvTHGmC7Dl8Qi3h3h7l1cNubISdp3sJYNu8vIbloPZ0y12SKNMV2eL4llh4jcISKh7utOYIe/Ausu3s0tYrjsIKKxwp62N8acFnxJLP8BnIPzPEoBMAGY44+gupNFGwuZGfslIDDQZos0xnR97W7KUtVinGdHTAcpq67ns52lPJqYC/EjoUdCoEMyxpiT1mZiEZGfqOr/iMif+PpzLACo6h1+iawbeH9TET20hj5VG2HUjwIdjjHGdIj21Fia504J/FOEp5lFuYV8I3YrUu+x24yNMaeN9iSWa4F/4ozb9Uc/x9NtVNY1sGLbfhb03gIVMZA+PtAhGWNMh2hP5/1Yd5bIW0QkTkTivV/+DvB09eGWEuo9HrJqc2DA+RAcGuiQjDGmQ7SnxvJ/wBJgILAGZ4ywZuqWGx8t3lhIdnQp4dV7IPM/Ax2OMcZ0mDZrLKr6mKqeiTMy8UBVHeD1sqRyAuoaPHy4pZibkrc7Bda/Yow5jbTnrrBYd+rfn7XW9KWqbU32ZVpYvm0/NfUezmE9xA+E+AGBDskYYzpMe5rC5gNX4DSDKdYUdtIWbSwkIUKJK1kFo64PdDjGGNOh2jNs/hXuX/tZ3QEaPE28v7mI/+hXhHxVY8O4GGNOO74Mm3+ViPT0et9LRL7pn7BOX5/tKONgbQPTI3MhKBQyJgU6JGOM6VC+jBX2gKoebH6jqgeAB9raSUSmi8gWEckTkXtbWd9fRJaIyAYRWSoiaV7rHhGRje7rWq/y5SKyzn3tFZE33HIRkcfcc20QkTE+XN8psTi3kMjQYPqXfwr9JkJ4dKBDMsaYDuVLYmlt2+M2pYlIMPAEcCkwDJglIsNabPYo8LyqjgAeBB52970cGAOMwhnw8m4RiQVQ1UmqOkpVRwGfAH93j3UpMMh9zaGTzRfT1KTO3CuZQQQV5zrD5BtjzGnGl8SSIyK/F5Ez3NfvcTr0j2c8kKeqO1S1HlgAzGixzTDgA3f5Q6/1w4BlqtqoqtXABmC6945uopkKvOEWzcBJUqqqnwK9RKS3D9foV5/vPkBx5SG+FbfNKbD+FWPMaciXxPJDoB54GSdB1AE/aGOfvsBur/cFbpm39cBMd/kqIEZEEtzy6SISJSKJwBQgvcW+3wSWuLdDt/d8iMgcEckRkZySkpI2LqHjLM4tJCw4iOF1OdAjCVKGn7JzG2PMqeLLsPnVwL0i0sNd7ih3A4+LyE3AMpz5Xjyq+q6IjANWAiU4TV6eFvvOAp7x9YSq+hTwFEB2dvZRIzb7g6qyaGMh550RR+iujyDzQgjyJa8bY0zX4MtdYeeIyCZgs/t+pIg82cZue/h6LSPNLTtMVfeq6kxVHQ38zC074P59yO1LuQjn+ZmtXvEk4jS1/cuX8wXK5n2VfFVWw3Xp5VBTak/bG2NOW778ZP4DcAlQCqCq64Hz29hnNTBIRAaISBjORGELvTcQkUQRaY7jPmCuWx7sNokhIiOAEcC7XrteA/xTVeu8yhYC33HvDpsIHFTVfT5co98syi0kSOA8We8UWMe9MeY01e6mMABV3S3i/eD9UU1TLbdvFJHbgcVAMM54Y7ki8iCQo6oLgcnAwyKiOE1hzf02ocBy93wVwA2q2uh1+OuA37Y45dvAZUAeUAPc7Mv1+dPijYWMHxBP1FcfQeoIiE4KdEjGGOMXviSW3SJyDqAiEgrcidssdjyq+jbOF7532c+9ll8DXmtlvzqcO8OOddzJrZQpbd9QcMrtKKliS1Elv760H3y0Cs75YaBDMsYYv/GlKew/cL60+wJ7cZ4v6XRf4p3R4twiAC7rsQ2aGq1/xRhzWvPlrrD9gI2YeAIW5RYyMq0n8YWvQ1g0pE8IdEjGGOM3vtwVNlBE3hKREhEpFpE3RcRGNm7D3gO1rN99gEuyUiBviTM2WEhYoMMyxhi/8aUpbD7wCtAb6AO8Crzkj6BOJ+/mFgJwRVotHMi3p+2NMac9XxJLlKr+zR1ipVFVXwAi/BXY6WJRbiGDU6LpV/apU2C3GRtjTnO+JJZ3ROReEclwRyT+CfC2iMS3NrOkgdKqQ6zaWcb0rFTYvgTiMiDhjECHZYwxfuXL7cbfcv/+O87MkeA8DX8dNpNkq97fXESTwvQz4+H55TDyukCHZIwxfudLjeUeYKQ7k+RfcQaJvFpVB6iqJZVWLM4tIj0+kjMbNkFDtfWvGGO6BV8Sy/2qWiEi5+EMVf8MnWy+k86ksq6BFdv2Mz0rFdn+AQSF2GyRxphuwZfE0jx8y+XA06r6L8Dumz2GD7eUUO9pYvpZbv9K+gSIiA10WMYY43e+JJY9IvIX4FqcTvtwH/fvVhZvLCQpJpzRcQ1Q+IXdDWaM6TZ8SQzfwhlM8hJ3WPt44Md+iaqLq2vw8OGWYi7JSiFo54dOofWvGGO6CV+GdKnhyNzyuMPRd4oh6Tub5dv2U1PvYXpWb/jiTxCVCKkjAx2WMcacEtaU5QeLNhbSMzKUCQN6wfYP4IwpNlukMabbsG+7DtbgaeL9zUVMOzOZ0JJcqC6x0YyNMd2KJZYO9tmOMg7WNjhP2+ctcQqt494Y041YYulgi3L3ERkazPmDk5xmsJThEJMS6LCMMeaU8XtiEZHpIrJFRPJE5N5W1vcXkSUiskFElopImte6R0Rko/u61qtcROQhEdkqIptF5A63fLKIHBSRde7r5y3P509NTcri3CKmDE0ioqkWvvoUMq22YozpXnya895XIhIMPAFcBBQAq0Vkoapu8trsUeB5VZ0nIlOBh4EbReRyYAzOTJXhwFIReUdVK4CbgHRgqKo2iUiy1/GWq+oV/ryuY/l8dzkllYe4JCsVdi2HpgbrXzHGdDv+rrGMB/JUdYeq1gMLgBktthkGfOAuf+i1fhiwzB2ivxrYAEx3130feFBVmwBUtdiP19BuizYWEhYcxNShyU7/SmgU9JsY6LCMMeaU8ndi6Qvs9npf4JZ5Ww/MdJevAmJEJMEtny4iUSKSCEzBqaUAnAFcKyI5IvKOiAzyOt7ZIrLeLc9qLSgRmePum1NSUnJyV+hSVRblFnJuZgIxEaHOMC4ZkyAkvEOOb4wxXUVn6Ly/G7hARD4HLgD2AB5VfRd4G1iJM1PlJxwZrywcqFPVbOBpYK5bvhbor6ojgT8Bb7R2QlV9SlWzVTU7KSmpQy5i074KdpfVOmODle2Esh32tL0xplvyd2LZw5FaBkCaW3aYqu5V1ZmqOhr4mVt2wP37kKqOUtWLcOZ+2eruVsCRUQD+AYxwt69Q1Sp3+W0g1K3t+N3ijYUECVx4ZopTWwHrXzHGdEv+TiyrgUEiMkBEwnAmBVvovYGIJIpIcxz34dY+RCTYbRJDREbgJI933e3ewGkaA6eWs9XdLlVExF0ej3N9pX66tq9ZnFvE+AHxJESHQ94H0LOfzRZpjOmW/HpXmKo2isjtOINXBgNzVTVXRB4EclR1ITAZeFhEFFgG/MDdPRRY7uaJCuAGVW101/0WeFFE7gKqgO+55dcA3xeRRqAWuE5Vm2e79JsdJVVsKarkF1cOA08D7FwGw68GJ3ZjjOlW/JpY4HCT1Nstyn7utfwa8For+9Xh3BnW2jEP4MwL07L8ceDxkwzZZ4tziwC4OCsVdq+C+kprBjPGdFudofO+y1uUW8jI9F706RXp9K9IMAy8INBhGWNMQFhiOUl7D9SyfvcBZ2wwcJ5fSRsHET0DG5gxxgSIJZaT9G5uIQCXZKVA9X7Yt95uMzbGdGuWWE7SotxChqTEMDApGrZ/CKj1rxhjujVLLCehtOoQq3aWObUVcPpXIuOgz6jABmaMMQFkieUkvL+5iCaFS85KBVVnmPyBUyAoONChGWNMwFhiOQmLNhaSHh/JsN6xULQRqoqsf8UY0+1ZYjlBFXUNfJxXyvSsVETEZos0xhiXJZYT9OGXxdR7mpxBJ8HpX0keBrF9AhuYMcYEmCWWE5QYHc43RvZhdHoc1Fc7s0VabcUYY/w/pMvp6tzMRM7NdAdO3rUCPPXWv2KMMViNpWPkLYGQSOh3TqAjMcaYgLPE0hG2L4GMcyE0ItCRGGNMwFliOVnl+VCaZ0/bG2OMyxLLyWqeLdL6V4wxBrDEcvLylkBsGiQODnQkxhjTKVhiORnNs0VmTrXZIo0xxuX3xCIi00Vki4jkici9razvLyJLRGSDiCwVkTSvdY+IyEb3da1XuYjIQyKyVUQ2i8gdXuWPuefaICJj/HpxBTlw6P+3d/+xXld1HMefL0FJwKICt7owIaOMGoLdGeVKk9gwW5SzaQaVW2trWOpqlf1Ym1uzP1w//mCmGQ0nUxdBYw2TpYa5UrkCakI1opILNGglhM4UePXH5xjfi7Th9fPlXO59Pba7+/me7+fz+b4/Z/d7399zzud7zr6Mr0REdOjq91gkjQKWAPOAfmC9pNW2N3fsdiNwm+1lki4EbgAWSboYOAeYBYwBfi3pbtv7gE8DU4CzbB+SdHo510XA9PLzLuCm8rs7/nwv6KSsFhkR0aHbLZZzga22t9l+HrgTWHDEPjOA+8r2/R3PzwAesH3A9jPA48D88tzngOttHwKwvbuUL6BJUrb9EDBB0hu6cWFAM77S09tMlR8REUD3E0sPsL3jcX8p6/QYcEnZ/ihwmqTXl/L5ksZKmgi8n6aVAnAmcJmkPkl3S5r+Ml4PSZ8tx/bt2bNncFf27D9h58bcDRYRcYShMHj/JeB8SRuB84EdwEHba4E1wG+BO4DfAQfLMWOA52z3Aj8Clr6cF7R9i+1e272TJk0aXNTbslpkRMTRdDux7OBwKwNgcin7H9s7bV9iezbw9VL2dPn9bduzbM8DBPypHNYPrCzbq4CZx/p6rRnzanjLfOjp7v0BEREnmm4nlvXAdEnTJJ0CXA6s7txB0kRJL8ZxHaX1IWlU6RJD0kya5LG27Pdzmq4xaFo5Lyac1cAny91hc4C9tnd15cqmz4Mr7spqkRERR+jqXWG2D0i6CrgHGAUstf2kpOuBPturgQuAGyQZeABYXA4/GfiNmu+H7AMW2j5QnvsOsFzStcB+4DOlfA3wQWAr8CxwZTevLyIiXkq2a8dQVW9vr/v6+mqHERFxQpH0aBnnfomhMHgfERHDSBJLRES0KoklIiJalcQSERGtSmKJiIhWJbFERESrRvztxpL2AH8b5OETgX+0GM6JLvUxUOrjsNTFQMOhPs6wfdQ5sUZ8YnklJPX9v/u4R6LUx0Cpj8NSFwMN9/pIV1hERLQqiSUiIlqVxPLK3FI7gCEm9TFQ6uOw1MVAw7o+MsYSERGtSoslIiJalcQSERGtSmIZJEnzJf1R0lZJX60dT02Spki6X9JmSU9Kurp2TLWVheo2SvpF7VhqkzRBVIL0EgAAA7ZJREFU0gpJf5C0RdK7a8dUi6Rry3vk95LukPSq2jF1QxLLIEgaBSwBLgJmAB+XNKNuVFUdAL5oewYwB1g8wusD4GpgS+0ghogfAL+0fRZwNiO0XiT1AF8Aem2/g2bxw8vrRtUdSSyDcy6w1fY2288DdwILKsdUje1dtjeU7X/T/OPoqRtVPZImAxcDt9aOpTZJrwHeB/wYwPbztp+uG1VVo4FTJY0GxgI7K8fTFUksg9MDbO943M8I/kfaSdJUYDbwcN1Iqvo+8GXgUO1AhoBpwB7gJ6Vr8FZJ42oHVYPtHcCNwFPALmCv7bV1o+qOJJZojaTxwM+Aa2zvqx1PDZI+BOy2/WjtWIaI0cA5wE22ZwPPACNyTFLSa2l6NqYBbwTGSVpYN6ruSGIZnB3AlI7Hk0vZiCXpZJqkstz2ytrxVHQe8GFJf6XpIr1Q0u11Q6qqH+i3/WILdgVNohmJPgD8xfYe2y8AK4H3VI6pK5JYBmc9MF3SNEmn0AzAra4cUzWSRNOHvsX2d2vHU5Pt62xPtj2V5u/iPtvD8lPpsbD9d2C7pLeWornA5ooh1fQUMEfS2PKemcswvZFhdO0ATkS2D0i6CriH5s6OpbafrBxWTecBi4AnJG0qZV+zvaZiTDF0fB5YXj6EbQOurBxPFbYflrQC2EBzJ+VGhunULpnSJSIiWpWusIiIaFUSS0REtCqJJSIiWpXEEhERrUpiiYiIViWxRJzAJF2QGZRjqEliiYiIViWxRBwHkhZKekTSJkk3l/Va9kv6Xlmf415Jk8q+syQ9JOlxSavKHFNIerOkX0l6TNIGSWeW04/vWO9keflWd0Q1SSwRXSbpbcBlwHm2ZwEHgU8A44A+228H1gHfKofcBnzF9kzgiY7y5cAS22fTzDG1q5TPBq6hWRvoTTQzIURUkyldIrpvLvBOYH1pTJwK7KaZVv+uss/twMqyfskE2+tK+TLgp5JOA3psrwKw/RxAOd8jtvvL403AVODB7l9WxNElsUR0n4Bltq8bUCh984j9Bju/0n86tg+S93VUlq6wiO67F7hU0ukAkl4n6Qya99+lZZ8rgAdt7wX+Jem9pXwRsK6szNkv6SPlHGMkjT2uVxFxjPLJJqLLbG+W9A1graSTgBeAxTSLXp1bnttNMw4D8CnghyVxdM4GvAi4WdL15RwfO46XEXHMMrtxRCWS9tseXzuOiLalKywiIlqVFktERLQqLZaIiGhVEktERLQqiSUiIlqVxBIREa1KYomIiFb9FzJ5nAsl0/qwAAAAAElFTkSuQmCC\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.plot(history.history['specificity_at_sensitivity'])\n",
        "plt.plot(history_jig.history['specificity_at_sensitivity_1'])\n",
        "plt.title('Train specificity_at_sensitivity')\n",
        "plt.ylabel('specificity_at_sensitivity')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['model', 'model_jig'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 295
        },
        "id": "dBoMOK1Tk_p1",
        "outputId": "31ecf3d6-9b95-412a-d824-fe4769259cb1"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ],
      "source": [
        "plt.plot(history.history['val_specificity_at_sensitivity'])\n",
        "plt.plot(history_jig.history['val_specificity_at_sensitivity_1'])\n",
        "plt.title('Test specificity_at_sensitivity')\n",
        "plt.ylabel('specificity_at_sensitivity')\n",
        "plt.xlabel('epoch')\n",
        "plt.legend(['model', 'model_jig'], loc='upper left')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Making Predictions for first 25 test images.\n"
      ],
      "metadata": {
        "id": "mntPfTt95D88"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "yCCsLSX6nCka"
      },
      "outputs": [],
      "source": [
        "predictions = model_jig.predict(test_images[:25])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "0dmGiDdcoYdo",
        "outputId": "b0283ba8-1b33-4fb9-ace3-89c48182854d"
      },
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(25, 10)"
            ]
          },
          "metadata": {},
          "execution_count": 39
        }
      ],
      "source": [
        "predictions.shape"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Storing the predicted values in class_id."
      ],
      "metadata": {
        "id": "FsutqQGp5bnp"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "8uV7yC2in4ML"
      },
      "outputs": [],
      "source": [
        "class_id = np.argmax(predictions,axis= 1)"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "vW4ZOLq1pmAg"
      },
      "outputs": [],
      "source": [
        "class_names = ['T-shirt/top', 'Trouser', 'Pullover', 'Dress', 'Coat',\n",
        "               'Sandal', 'Shirt', 'Sneaker', 'Bag', 'Ankle boot']"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "aj70LaziqDDH"
      },
      "outputs": [],
      "source": [
        "class_predicted_name=[]\n",
        "for id in class_id:\n",
        "    class_predicted_name.append(class_names[id])"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 713
        },
        "id": "InwKCabrvf5u",
        "outputId": "74b13572-f409-42f8-df4e-d24ffd9208dc"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 25 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "plt.figure(figsize=(10,10))\n",
        "for i in range(25):\n",
        "    plt.subplot(5,5,i+1)\n",
        "    plt.tight_layout()\n",
        "    plt.xticks([])\n",
        "    plt.yticks([])\n",
        "    plt.grid(False)\n",
        "    plt.imshow(test_images[i].squeeze(), cmap=plt.cm.binary)\n",
        "    plt.xlabel('actual label : '+ class_names[test_labels[i]]+'\\n'+'predicted label : '+ class_predicted_name[i])\n",
        "    \n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "UPm6xr8jbjn2"
      },
      "source": [
        "Example of augmentation variation in few images."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "TXjgGDkwANKy"
      },
      "outputs": [],
      "source": [
        "row = [2,4,7,14,28]\n",
        "col = [2,4,7,14,28]\n",
        "dest=[]\n",
        "for i in row:\n",
        "  for j in col:\n",
        "    dest.append(generate_jigsaw_destinations(i,j,5,42))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "SnaJwPK9vjPc"
      },
      "outputs": [],
      "source": [
        "all_jig_pattern=[]\n",
        "for i in range(25):\n",
        "  all_jig_pattern.append(apply_jigsaw(train_images[6],dest[i]))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 729
        },
        "id": "BS19QP_hn9Uq",
        "outputId": "233a2e23-70e8-4331-ff49-6d91996e1673"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 25 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "ro=[2,2,2,2,2,4,4,4,4,4,7,7,7,7,7,14,14,14,14,14,28,28,28,28,28]\n",
        "co=[2,4,7,14,28,2,4,7,14,28,2,4,7,14,28,2,4,7,14,28,2,4,7,14,28]\n",
        "plt.figure(figsize=(10,10))\n",
        "for i in range(25):\n",
        "    plt.subplot(5,5,i+1)\n",
        "    plt.tight_layout()\n",
        "    plt.xticks([])\n",
        "    plt.yticks([])\n",
        "    plt.grid(False)\n",
        "    plt.imshow(all_jig_pattern[i].squeeze(), cmap=plt.cm.binary)\n",
        "    plt.xlabel(class_names[train_labels[6]]+'\\n'+'nb_rows :'+str(ro[i])+'\\n'+'nb_col :'+str(co[i]))\n",
        "    \n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "tp-lmIIE1cds"
      },
      "outputs": [],
      "source": [
        "all_jig_pattern1=[]\n",
        "for i in range(25):\n",
        "  all_jig_pattern1.append(apply_jigsaw(train_images[20],dest[i]))"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 729
        },
        "id": "Gacit05o1g7F",
        "outputId": "6e4154c3-6aee-45f3-dbc9-f8a1ddb8f87a"
      },
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x720 with 25 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        }
      ],
      "source": [
        "ro=[2,2,2,2,2,4,4,4,4,4,7,7,7,7,7,14,14,14,14,14,28,28,28,28,28]\n",
        "co=[2,4,7,14,28,2,4,7,14,28,2,4,7,14,28,2,4,7,14,28,2,4,7,14,28]\n",
        "plt.figure(figsize=(10,10))\n",
        "for i in range(25):\n",
        "    plt.subplot(5,5,i+1)\n",
        "    plt.tight_layout()\n",
        "    plt.xticks([])\n",
        "    plt.yticks([])\n",
        "    plt.grid(False)\n",
        "    plt.imshow(all_jig_pattern1[i].squeeze(), cmap=plt.cm.binary)\n",
        "    plt.xlabel(class_names[train_labels[20]]+'\\n'+'nb_rows :'+str(ro[i])+'\\n'+'nb_col :'+str(co[i]))\n",
        "    \n",
        "plt.show()"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "provenance": []
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}