CNN_MNIST.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "authorship_tag": "ABX9TyOWAnwPtgFMG9vg1j45eXEW"
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    },
    "widgets": {
      "application/vnd.jupyter.widget-state+json": {
        "504de3a1a5204f899f54c216300f6aec": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HBoxModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HBoxModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HBoxView",
            "box_style": "",
            "children": [
              "IPY_MODEL_36a882544f8d4e638858089a65d2f86a",
              "IPY_MODEL_52b6e314572d4792b8823631499b22db",
              "IPY_MODEL_4c31aabb7df34e679ea72fb41af26fbf"
            ],
            "layout": "IPY_MODEL_b76e858f0ac244d09abde8e74bdbe6aa"
          }
        },
        "36a882544f8d4e638858089a65d2f86a": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_9fb8295c350f49c6977080fb2ad221c9",
            "placeholder": "​",
            "style": "IPY_MODEL_3cc93da3a57148ba9a4a5aee4c0baa5d",
            "value": "Dl Completed...: 100%"
          }
        },
        "52b6e314572d4792b8823631499b22db": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "FloatProgressModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "FloatProgressModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "ProgressView",
            "bar_style": "success",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_f470b9b679864dce9c12e61d0bacc19a",
            "max": 5,
            "min": 0,
            "orientation": "horizontal",
            "style": "IPY_MODEL_446b4b24ae9641d5bc128474becade20",
            "value": 5
          }
        },
        "4c31aabb7df34e679ea72fb41af26fbf": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "HTMLModel",
          "model_module_version": "1.5.0",
          "state": {
            "_dom_classes": [],
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "HTMLModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/controls",
            "_view_module_version": "1.5.0",
            "_view_name": "HTMLView",
            "description": "",
            "description_tooltip": null,
            "layout": "IPY_MODEL_44ef6fcacd394f64b2932744e879cccf",
            "placeholder": "​",
            "style": "IPY_MODEL_e53ce710dd5e474cb266fb9bd10c3be6",
            "value": " 5/5 [00:00&lt;00:00, 14.95 file/s]"
          }
        },
        "b76e858f0ac244d09abde8e74bdbe6aa": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "9fb8295c350f49c6977080fb2ad221c9": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "3cc93da3a57148ba9a4a5aee4c0baa5d": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        },
        "f470b9b679864dce9c12e61d0bacc19a": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "446b4b24ae9641d5bc128474becade20": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "ProgressStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "ProgressStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "bar_color": null,
            "description_width": ""
          }
        },
        "44ef6fcacd394f64b2932744e879cccf": {
          "model_module": "@jupyter-widgets/base",
          "model_name": "LayoutModel",
          "model_module_version": "1.2.0",
          "state": {
            "_model_module": "@jupyter-widgets/base",
            "_model_module_version": "1.2.0",
            "_model_name": "LayoutModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "LayoutView",
            "align_content": null,
            "align_items": null,
            "align_self": null,
            "border": null,
            "bottom": null,
            "display": null,
            "flex": null,
            "flex_flow": null,
            "grid_area": null,
            "grid_auto_columns": null,
            "grid_auto_flow": null,
            "grid_auto_rows": null,
            "grid_column": null,
            "grid_gap": null,
            "grid_row": null,
            "grid_template_areas": null,
            "grid_template_columns": null,
            "grid_template_rows": null,
            "height": null,
            "justify_content": null,
            "justify_items": null,
            "left": null,
            "margin": null,
            "max_height": null,
            "max_width": null,
            "min_height": null,
            "min_width": null,
            "object_fit": null,
            "object_position": null,
            "order": null,
            "overflow": null,
            "overflow_x": null,
            "overflow_y": null,
            "padding": null,
            "right": null,
            "top": null,
            "visibility": null,
            "width": null
          }
        },
        "e53ce710dd5e474cb266fb9bd10c3be6": {
          "model_module": "@jupyter-widgets/controls",
          "model_name": "DescriptionStyleModel",
          "model_module_version": "1.5.0",
          "state": {
            "_model_module": "@jupyter-widgets/controls",
            "_model_module_version": "1.5.0",
            "_model_name": "DescriptionStyleModel",
            "_view_count": null,
            "_view_module": "@jupyter-widgets/base",
            "_view_module_version": "1.2.0",
            "_view_name": "StyleView",
            "description_width": ""
          }
        }
      }
    }
  },
  "cells": [
    {
      "cell_type": "code",
      "execution_count": 19,
      "metadata": {
        "id": "-2pumT0GPc52"
      },
      "outputs": [],
      "source": [
        "# Importing the relevant packages\n",
        "import tensorflow as tf\n",
        "import tensorflow_datasets as tfds\n",
        "\n",
        "import matplotlib.pyplot as plt\n",
        "import numpy as np"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Downloading and preprocessing the data"
      ],
      "metadata": {
        "id": "wLvxBWYZP2Rc"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Defining some constants/hyperparameters\n",
        "BUFFER_SIZE = 70_000 # for reshuffling\n",
        "BATCH_SIZE = 128\n",
        "NUM_EPOCHS = 20"
      ],
      "metadata": {
        "id": "M-zwOO7zP4Jm"
      },
      "execution_count": 2,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Downloading the MNIST dataset\n",
        "mnist_dataset, mnist_info = tfds.load(name='mnist', with_info=True, as_supervised=True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 84,
          "referenced_widgets": [
            "504de3a1a5204f899f54c216300f6aec",
            "36a882544f8d4e638858089a65d2f86a",
            "52b6e314572d4792b8823631499b22db",
            "4c31aabb7df34e679ea72fb41af26fbf",
            "b76e858f0ac244d09abde8e74bdbe6aa",
            "9fb8295c350f49c6977080fb2ad221c9",
            "3cc93da3a57148ba9a4a5aee4c0baa5d",
            "f470b9b679864dce9c12e61d0bacc19a",
            "446b4b24ae9641d5bc128474becade20",
            "44ef6fcacd394f64b2932744e879cccf",
            "e53ce710dd5e474cb266fb9bd10c3be6"
          ]
        },
        "id": "s0RigLl7QW-N",
        "outputId": "f37fd970-1f3e-4928-ffae-22427a3df106"
      },
      "execution_count": 3,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Downloading and preparing dataset 11.06 MiB (download: 11.06 MiB, generated: 21.00 MiB, total: 32.06 MiB) to /root/tensorflow_datasets/mnist/3.0.1...\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "Dl Completed...:   0%|          | 0/5 [00:00<?, ? file/s]"
            ],
            "application/vnd.jupyter.widget-view+json": {
              "version_major": 2,
              "version_minor": 0,
              "model_id": "504de3a1a5204f899f54c216300f6aec"
            }
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Dataset mnist downloaded and prepared to /root/tensorflow_datasets/mnist/3.0.1. Subsequent calls will reuse this data.\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Extracting the train and test datasets\n",
        "mnist_train, mnist_test = mnist_dataset['train'], mnist_dataset['test']"
      ],
      "metadata": {
        "id": "CbyaPZhbQoc1"
      },
      "execution_count": 4,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Creating a function to scale our image data\n",
        "def scale(image, label):\n",
        "    image = tf.cast(image, tf.float32)\n",
        "    image /= 255.\n",
        "\n",
        "    return image, label"
      ],
      "metadata": {
        "id": "RPlS79z9Q-Jg"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Scaling the data\n",
        "train_and_validation_data = mnist_train.map(scale)\n",
        "test_data = mnist_test.map(scale)"
      ],
      "metadata": {
        "id": "N2lyUspGRP1h"
      },
      "execution_count": 6,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Defining the size of the validation set\n",
        "num_validation_samples = 0.1 * mnist_info.splits['train'].num_examples\n",
        "num_validation_samples = tf.cast(num_validation_samples, tf.int64)"
      ],
      "metadata": {
        "id": "9dZla1c0RqtZ"
      },
      "execution_count": 7,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Defining the size of the test set\n",
        "num_test_samples = mnist_info.splits['test'].num_examples\n",
        "num_test_samples = tf.cast(num_test_samples, tf.int64)"
      ],
      "metadata": {
        "id": "-jCLBuIKRxC4"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Reshuffling the dataset\n",
        "train_and_validation_data = train_and_validation_data.shuffle(BUFFER_SIZE)"
      ],
      "metadata": {
        "id": "CqaaKbYJR97H"
      },
      "execution_count": 9,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Splitting the dataset into training + validation\n",
        "train_data = train_and_validation_data.skip(num_validation_samples)\n",
        "validation_data = train_and_validation_data.take(num_validation_samples)"
      ],
      "metadata": {
        "id": "aoNjOshVSN8r"
      },
      "execution_count": 11,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Batching the data\n",
        "train_data = train_data.batch(BATCH_SIZE)\n",
        "validation_data = validation_data.batch(num_validation_samples)\n",
        "test_data = test_data.batch(num_test_samples)"
      ],
      "metadata": {
        "id": "7-Ku1DNwSew-"
      },
      "execution_count": 12,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Creating the model and training it"
      ],
      "metadata": {
        "id": "UNUxHz8FSu1m"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Outlining the model/architecture of our CNN\n",
        "model = tf.keras.Sequential([\n",
        "    tf.keras.layers.Conv2D(50, 5, activation='relu', input_shape=(28, 28, 1)),\n",
        "    tf.keras.layers.MaxPooling2D(pool_size=(2,2)),\n",
        "    tf.keras.layers.Conv2D(50, 3, activation='relu'),\n",
        "    tf.keras.layers.MaxPooling2D(pool_size=(2,2)),\n",
        "    tf.keras.layers.Flatten(),\n",
        "    tf.keras.layers.Dense(10)\n",
        "])"
      ],
      "metadata": {
        "id": "8s8M-yhkTH1N"
      },
      "execution_count": 13,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# A brief summary of the model and parameters\n",
        "model.summary(line_length = 75)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "o_d98i-EUGl5",
        "outputId": "9222a34b-3d5b-4b79-ca1e-677e4c0f340c"
      },
      "execution_count": 14,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Model: \"sequential\"\n",
            "___________________________________________________________________________\n",
            " Layer (type)                    Output Shape                  Param #     \n",
            "===========================================================================\n",
            " conv2d (Conv2D)                 (None, 24, 24, 50)            1300        \n",
            "                                                                           \n",
            " max_pooling2d (MaxPooling2D)    (None, 12, 12, 50)            0           \n",
            "                                                                           \n",
            " conv2d_1 (Conv2D)               (None, 10, 10, 50)            22550       \n",
            "                                                                           \n",
            " max_pooling2d_1 (MaxPooling2D)  (None, 5, 5, 50)              0           \n",
            "                                                                           \n",
            " flatten (Flatten)               (None, 1250)                  0           \n",
            "                                                                           \n",
            " dense (Dense)                   (None, 10)                    12510       \n",
            "                                                                           \n",
            "===========================================================================\n",
            "Total params: 36360 (142.03 KB)\n",
            "Trainable params: 36360 (142.03 KB)\n",
            "Non-trainable params: 0 (0.00 Byte)\n",
            "___________________________________________________________________________\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Defining the loss function\n",
        "# We use a loss calculation that automatically corrects for the missing softmax\n",
        "# That is the reason for 'from_logits=True'\n",
        "loss_fn = tf.keras.losses.SparseCategoricalCrossentropy(from_logits=True)"
      ],
      "metadata": {
        "id": "XF5wOGoIVJ9B"
      },
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Compiling the model with Adam optimizer and the categorical crossentropy as a loss function\n",
        "model.compile(optimizer='adam', loss=loss_fn, metrics=['accuracy'])"
      ],
      "metadata": {
        "id": "i2tlsnsFVeSg"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Defining early stopping to prevent overfitting\n",
        "early_stopping = tf.keras.callbacks.EarlyStopping(\n",
        "    monitor = 'val_loss',\n",
        "    mode = 'auto',\n",
        "    min_delta = 0,\n",
        "    patience = 2,\n",
        "    verbose = 0,\n",
        "    restore_best_weights = True\n",
        ")"
      ],
      "metadata": {
        "id": "mcJb_VJnVoEB"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# Train the network\n",
        "model.fit(\n",
        "    train_data,\n",
        "    epochs = NUM_EPOCHS,\n",
        "    callbacks = [early_stopping],\n",
        "    validation_data = validation_data,\n",
        "    verbose = 2\n",
        ")"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "uLDkl08iVyn1",
        "outputId": "6bf0e710-5f3a-476d-dcdb-8ebdf8699e2e"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Epoch 1/20\n",
            "422/422 - 72s - loss: 0.2626 - accuracy: 0.9249 - val_loss: 0.0828 - val_accuracy: 0.9750 - 72s/epoch - 170ms/step\n",
            "Epoch 2/20\n",
            "422/422 - 66s - loss: 0.0710 - accuracy: 0.9783 - val_loss: 0.0458 - val_accuracy: 0.9852 - 66s/epoch - 156ms/step\n",
            "Epoch 3/20\n",
            "422/422 - 69s - loss: 0.0539 - accuracy: 0.9833 - val_loss: 0.0483 - val_accuracy: 0.9823 - 69s/epoch - 164ms/step\n",
            "Epoch 4/20\n",
            "422/422 - 74s - loss: 0.0444 - accuracy: 0.9863 - val_loss: 0.0348 - val_accuracy: 0.9907 - 74s/epoch - 176ms/step\n",
            "Epoch 5/20\n",
            "422/422 - 80s - loss: 0.0373 - accuracy: 0.9891 - val_loss: 0.0293 - val_accuracy: 0.9915 - 80s/epoch - 189ms/step\n",
            "Epoch 6/20\n",
            "422/422 - 68s - loss: 0.0315 - accuracy: 0.9903 - val_loss: 0.0260 - val_accuracy: 0.9900 - 68s/epoch - 162ms/step\n",
            "Epoch 7/20\n",
            "422/422 - 67s - loss: 0.0285 - accuracy: 0.9912 - val_loss: 0.0299 - val_accuracy: 0.9895 - 67s/epoch - 158ms/step\n",
            "Epoch 8/20\n",
            "422/422 - 66s - loss: 0.0254 - accuracy: 0.9926 - val_loss: 0.0156 - val_accuracy: 0.9943 - 66s/epoch - 156ms/step\n",
            "Epoch 9/20\n",
            "422/422 - 66s - loss: 0.0218 - accuracy: 0.9931 - val_loss: 0.0192 - val_accuracy: 0.9938 - 66s/epoch - 157ms/step\n",
            "Epoch 10/20\n",
            "422/422 - 66s - loss: 0.0195 - accuracy: 0.9938 - val_loss: 0.0149 - val_accuracy: 0.9958 - 66s/epoch - 157ms/step\n",
            "Epoch 11/20\n",
            "422/422 - 71s - loss: 0.0166 - accuracy: 0.9949 - val_loss: 0.0143 - val_accuracy: 0.9955 - 71s/epoch - 169ms/step\n",
            "Epoch 12/20\n",
            "422/422 - 66s - loss: 0.0152 - accuracy: 0.9949 - val_loss: 0.0100 - val_accuracy: 0.9967 - 66s/epoch - 157ms/step\n",
            "Epoch 13/20\n",
            "422/422 - 67s - loss: 0.0140 - accuracy: 0.9957 - val_loss: 0.0107 - val_accuracy: 0.9972 - 67s/epoch - 159ms/step\n",
            "Epoch 14/20\n",
            "422/422 - 66s - loss: 0.0116 - accuracy: 0.9963 - val_loss: 0.0119 - val_accuracy: 0.9965 - 66s/epoch - 157ms/step\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<keras.src.callbacks.History at 0x7908d5745ff0>"
            ]
          },
          "metadata": {},
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "## Testing our model"
      ],
      "metadata": {
        "id": "k4E3v63EWYLy"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "# Testing our model\n",
        "test_loss, test_accuracy = model.evaluate(test_data)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "dA7lDrYbWmBy",
        "outputId": "ac6b56fe-58b5-4dbc-e6e0-5de5b0b6e1b8"
      },
      "execution_count": 20,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 4s 4s/step - loss: 0.0339 - accuracy: 0.9903\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Printing the test results\n",
        "print('Test loss: {0:.4f}. Test accuracy: {1:.2f}%'.format(test_loss, test_accuracy*100.))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "sdW4Pt4GWrjV",
        "outputId": "8336c184-44fa-4b6f-ebc0-1543f59b2ca0"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Test loss: 0.0339. Test accuracy: 99.03%\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Split the test_data into 2 arrays, containing the images and the corresponding labels\n",
        "for images, labels in test_data.take(1):\n",
        "    images_test = images.numpy()\n",
        "    labels_test = labels.numpy()\n",
        "\n",
        "# Reshape the images into 28x28 form, suitable for matplotlib (original dimensions: 28x28x1)\n",
        "images_plot = np.reshape(images_test, (10000,28,28))"
      ],
      "metadata": {
        "id": "URvVI-qmW1xN"
      },
      "execution_count": 22,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# The image to be displayed and tested\n",
        "i = 502\n",
        "\n",
        "# Plot the image\n",
        "plt.figure(figsize=(2, 2))\n",
        "plt.axis('off')\n",
        "plt.imshow(images_plot[i-1], cmap=\"gray\", aspect='auto')\n",
        "plt.show()\n",
        "\n",
        "# Print the correct label for the image\n",
        "print(\"Label: {}\".format(labels_test[i-1]))"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 208
        },
        "id": "ENaHjEtYW4NC",
        "outputId": "fa87179f-9290-41f1-c32b-a1187992a9a1"
      },
      "execution_count": 23,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 200x200 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Label: 5\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Obtain the model's predictions (logits)\n",
        "predictions = model.predict(images_test[i-1:i])\n",
        "\n",
        "# Convert those predictions into probabilities\n",
        "probabilities = tf.nn.softmax(predictions).numpy()\n",
        "\n",
        "# Convert the probabilities into percentages\n",
        "probabilities = probabilities*100\n",
        "\n",
        "# Create a bar chart to plot the probabilities for each class\n",
        "plt.figure(figsize=(12,5))\n",
        "plt.bar(x=[1,2,3,4,5,6,7,8,9,10], height=probabilities[0], tick_label=[\"0\",\"1\",\"2\",\"3\",\"4\",\"5\",\"6\",\"7\",\"8\",\"9\"])\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 463
        },
        "id": "Va4VfmYBW806",
        "outputId": "325a1ce5-007e-4141-d42b-84e87f754776"
      },
      "execution_count": 27,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "1/1 [==============================] - 0s 26ms/step\n"
          ]
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1200x500 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAA9EAAAGtCAYAAAART/fgAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAAAnbUlEQVR4nO3df5RXBZ3/8dfwYwZWmCFIZpgExDIhFVM0HLEfa2wcjnnkSKYeOovi5m47msBWC7uZWipom5qpmB4WLWNN28Sso6xNhesuKGJ0NDfUMpkNZ9x2Y0ZpGVjm8/1jz863Sc07MMOHGR+Pc+45zr33cz/vuRnHJ/dzP7eiVCqVAgAAALyhQeUeAAAAAPoLEQ0AAAAFiWgAAAAoSEQDAABAQSIaAAAAChLRAAAAUJCIBgAAgIJENAAAABQ0pNwD/L7Ozs5s27YtI0eOTEVFRbnHAQAAYIArlUp5+eWXU19fn0GD/vC15gMuordt25bx48eXewwAAADeZJqbm3PIIYf8wX16FNF79uzJZZddljvvvDMtLS2pr6/Pueeem89+9rNdV41LpVIuvfTS3Hbbbdm+fXtmzJiRFStW5PDDDy/0HiNHjuwavrq6uifjAQAAQI+1t7dn/PjxXT36h/Qooq+++uqsWLEid9xxR4488sg8/vjjOe+881JTU5NPfvKTSZJrrrkmN9xwQ+64445MmjQpl1xySWbNmpWnn346w4YNe8P3+L8Yr66uFtEAAADsN0VuKa4olUqlogf88Ic/nNra2qxcubJr3dy5czN8+PDceeedKZVKqa+vz1/91V/lU5/6VJKkra0ttbW1uf3223P22We/4Xu0t7enpqYmbW1tIhoAAIA+15MO7dG3c5900klpamrKM888kyT5yU9+kkceeSSzZ89Okjz//PNpaWnJzJkzu15TU1OT6dOnZ/369a95zI6OjrS3t3dbAAAA4EDUo49zL1myJO3t7Zk8eXIGDx6cPXv25Morr8y8efOSJC0tLUmS2trabq+rra3t2vb7li1blssvv3xvZgcAAID9qkdXou++++584xvfyOrVq/PEE0/kjjvuyN/93d/ljjvu2OsBli5dmra2tq6lubl5r48FAAAAfalHV6I//elPZ8mSJV33Nh999NF54YUXsmzZssyfPz91dXVJktbW1owbN67rda2trXn3u9/9msesqqpKVVXVXo4PAAAA+0+PrkT/9re/fdWDpwcPHpzOzs4kyaRJk1JXV5empqau7e3t7Xn00UfT0NDQC+MCAABA+fToSvRpp52WK6+8MhMmTMiRRx6ZH//4x7n22muzYMGCJP/7deALFy7MFVdckcMPP7zrEVf19fWZM2dOX8wPAAAA+02PIvorX/lKLrnkkvzlX/5lXnrppdTX1+fP//zP87nPfa5rn8985jPZsWNHLrjggmzfvj0nn3xyHnzwwULPiAYAAIADWY+eE70/eE40AAAA+1OfPScaAAAA3sxENAAAABQkogEAAKAgEQ0AAAAFiWgAAAAoSEQDAABAQSIaAAAAChpS7gEAgP3r0CXfK/cIB5xfLj+13CMA0E+4Eg0AAAAFiWgAAAAoSEQDAABAQSIaAAAAChLRAAAAUJCIBgAAgIJENAAAABQkogEAAKAgEQ0AAAAFiWgAAAAoSEQDAABAQSIaAAAAChLRAAAAUJCIBgAAgIJENAAAABQkogEAAKAgEQ0AAAAFiWgAAAAoSEQDAABAQSIaAAAAChLRAAAAUJCIBgAAgIJENAAAABQkogEAAKAgEQ0AAAAFiWgAAAAoSEQDAABAQSIaAAAAChLRAAAAUFCPIvrQQw9NRUXFq5bGxsYkyc6dO9PY2JgxY8ZkxIgRmTt3blpbW/tkcAAAANjfehTRGzduzIsvvti1PPTQQ0mSM888M0myaNGi3H///bnnnnuybt26bNu2LWeccUbvTw0AAABlMKQnOx988MHdfl6+fHne/va35/3vf3/a2tqycuXKrF69OqecckqSZNWqVZkyZUo2bNiQE0888TWP2dHRkY6Ojq6f29vbe/o7AAAAwH6x1/dE79q1K3feeWcWLFiQioqKbNq0Kbt3787MmTO79pk8eXImTJiQ9evXv+5xli1blpqamq5l/PjxezsSAAAA9Km9jug1a9Zk+/btOffcc5MkLS0tqayszKhRo7rtV1tbm5aWltc9ztKlS9PW1ta1NDc37+1IAAAA0Kd69HHu37Vy5crMnj079fX1+zRAVVVVqqqq9ukYAAAAsD/sVUS/8MIL+f73v59vf/vbXevq6uqya9eubN++vdvV6NbW1tTV1e3zoAAAAFBue/Vx7lWrVmXs2LE59dRTu9ZNmzYtQ4cOTVNTU9e6LVu2ZOvWrWloaNj3SQEAAKDMenwlurOzM6tWrcr8+fMzZMj/f3lNTU3OP//8LF68OKNHj051dXUuuuiiNDQ0vO43cwMAAEB/0uOI/v73v5+tW7dmwYIFr9p23XXXZdCgQZk7d246Ojoya9as3Hzzzb0yKAAAAJRbRalUKpV7iN/V3t6empqatLW1pbq6utzjAMCAc+iS75V7hAPOL5ef+sY7ATBg9aRD9/oRVwAAAPBmI6IBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKKjHEf2rX/0qH/vYxzJmzJgMHz48Rx99dB5//PGu7aVSKZ/73Ocybty4DB8+PDNnzsyzzz7bq0MDAABAOfQoon/zm99kxowZGTp0aB544IE8/fTT+dKXvpS3vOUtXftcc801ueGGG3LLLbfk0UcfzUEHHZRZs2Zl586dvT48AAAA7E9DerLz1VdfnfHjx2fVqlVd6yZNmtT1z6VSKddff30++9nP5vTTT0+SfO1rX0ttbW3WrFmTs88+u5fGBgAAgP2vR1eiv/Od7+T444/PmWeembFjx+bYY4/Nbbfd1rX9+eefT0tLS2bOnNm1rqamJtOnT8/69etf85gdHR1pb2/vtgAAAMCBqEcR/Ytf/CIrVqzI4YcfnrVr1+YTn/hEPvnJT+aOO+5IkrS0tCRJamtru72utra2a9vvW7ZsWWpqarqW8ePH783vAQAAAH2uRxHd2dmZ4447LldddVWOPfbYXHDBBfn4xz+eW265Za8HWLp0adra2rqW5ubmvT4WAAAA9KUeRfS4cePyrne9q9u6KVOmZOvWrUmSurq6JElra2u3fVpbW7u2/b6qqqpUV1d3WwAAAOBA1KOInjFjRrZs2dJt3TPPPJOJEycm+d8vGaurq0tTU1PX9vb29jz66KNpaGjohXEBAACgfHr07dyLFi3KSSedlKuuuiof/ehH89hjj+XWW2/NrbfemiSpqKjIwoULc8UVV+Twww/PpEmTcskll6S+vj5z5szpi/kBAABgv+lRRJ9wwgm59957s3Tp0nz+85/PpEmTcv3112fevHld+3zmM5/Jjh07csEFF2T79u05+eST8+CDD2bYsGG9PjwAAADsTxWlUqlU7iF+V3t7e2pqatLW1ub+aADoA4cu+V65Rzjg/HL5qeUeAYAy6kmH9uieaAAAAHgzE9EAAABQkIgGAACAgkQ0AAAAFCSiAQAAoCARDQAAAAWJaAAAAChIRAMAAEBBIhoAAAAKEtEAAABQkIgGAACAgkQ0AAAAFCSiAQAAoCARDQAAAAWJaAAAAChIRAMAAEBBIhoAAAAKEtEAAABQkIgGAACAgkQ0AAAAFCSiAQAAoCARDQAAAAWJaAAAAChIRAMAAEBBIhoAAAAKEtEAAABQkIgGAACAgkQ0AAAAFCSiAQAAoCARDQAAAAWJaAAAAChIRAMAAEBBIhoAAAAKEtEAAABQkIgGAACAgkQ0AAAAFCSiAQAAoKAeRfRll12WioqKbsvkyZO7tu/cuTONjY0ZM2ZMRowYkblz56a1tbXXhwYAAIBy6PGV6COPPDIvvvhi1/LII490bVu0aFHuv//+3HPPPVm3bl22bduWM844o1cHBgAAgHIZ0uMXDBmSurq6V61va2vLypUrs3r16pxyyilJklWrVmXKlCnZsGFDTjzxxNc8XkdHRzo6Orp+bm9v7+lIAAAAsF/0+Er0s88+m/r6+hx22GGZN29etm7dmiTZtGlTdu/enZkzZ3btO3ny5EyYMCHr169/3eMtW7YsNTU1Xcv48eP34tcAAACAvtejiJ4+fXpuv/32PPjgg1mxYkWef/75vPe9783LL7+clpaWVFZWZtSoUd1eU1tbm5aWltc95tKlS9PW1ta1NDc379UvAgAAAH2tRx/nnj17dtc/T506NdOnT8/EiRNz9913Z/jw4Xs1QFVVVaqqqvbqtQAAALA/7dMjrkaNGpV3vvOdee6551JXV5ddu3Zl+/bt3fZpbW19zXuoAQAAoL/Zp4h+5ZVX8vOf/zzjxo3LtGnTMnTo0DQ1NXVt37JlS7Zu3ZqGhoZ9HhQAAADKrUcf5/7Upz6V0047LRMnTsy2bdty6aWXZvDgwTnnnHNSU1OT888/P4sXL87o0aNTXV2diy66KA0NDa/7zdwAAADQn/Qoov/93/8955xzTv7zP/8zBx98cE4++eRs2LAhBx98cJLkuuuuy6BBgzJ37tx0dHRk1qxZufnmm/tkcAAAANjfKkqlUqncQ/yu9vb21NTUpK2tLdXV1eUeBwAGnEOXfK/cIxxwfrn81HKPAEAZ9aRD9+meaAAAAHgzEdEAAABQkIgGAACAgkQ0AAAAFCSiAQAAoCARDQAAAAWJaAAAAChIRAMAAEBBIhoAAAAKEtEAAABQkIgGAACAgkQ0AAAAFCSiAQAAoCARDQAAAAWJaAAAAChIRAMAAEBBIhoAAAAKEtEAAABQkIgGAACAgkQ0AAAAFCSiAQAAoCARDQAAAAWJaAAAAChIRAMAAEBBIhoAAAAKEtEAAABQkIgGAACAgkQ0AAAAFCSiAQAAoCARDQAAAAWJaAAAAChIRAMAAEBBIhoAAAAKEtEAAABQkIgGAACAgkQ0AAAAFLRPEb18+fJUVFRk4cKFXet27tyZxsbGjBkzJiNGjMjcuXPT2tq6r3MCAABA2e11RG/cuDFf/epXM3Xq1G7rFy1alPvvvz/33HNP1q1bl23btuWMM87Y50EBAACg3PYqol955ZXMmzcvt912W97ylrd0rW9ra8vKlStz7bXX5pRTTsm0adOyatWq/Ou//ms2bNjQa0MDAABAOexVRDc2NubUU0/NzJkzu63ftGlTdu/e3W395MmTM2HChKxfv/41j9XR0ZH29vZuCwAAAByIhvT0BXfddVeeeOKJbNy48VXbWlpaUllZmVGjRnVbX1tbm5aWltc83rJly3L55Zf3dAwAAADY73p0Jbq5uTkXX3xxvvGNb2TYsGG9MsDSpUvT1tbWtTQ3N/fKcQEAAKC39SiiN23alJdeeinHHXdchgwZkiFDhmTdunW54YYbMmTIkNTW1mbXrl3Zvn17t9e1tramrq7uNY9ZVVWV6urqbgsAAAAciHr0ce4PfvCDefLJJ7utO++88zJ58uT89V//dcaPH5+hQ4emqakpc+fOTZJs2bIlW7duTUNDQ+9NDQAAAGXQo4geOXJkjjrqqG7rDjrooIwZM6Zr/fnnn5/Fixdn9OjRqa6uzkUXXZSGhoaceOKJvTc1AAAAlEGPv1jsjVx33XUZNGhQ5s6dm46OjsyaNSs333xzb78NAAAA7HcVpVKpVO4hfld7e3tqamrS1tbm/mgA6AOHLvleuUc44Pxy+anlHgGAMupJh+7Vc6IBAADgzUhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCehTRK1asyNSpU1NdXZ3q6uo0NDTkgQce6Nq+c+fONDY2ZsyYMRkxYkTmzp2b1tbWXh8aAAAAyqFHEX3IIYdk+fLl2bRpUx5//PGccsopOf300/PTn/40SbJo0aLcf//9ueeee7Ju3bps27YtZ5xxRp8MDgAAAPtbRalUKu3LAUaPHp0vfvGL+chHPpKDDz44q1evzkc+8pEkyc9+9rNMmTIl69evz4knnljoeO3t7ampqUlbW1uqq6v3ZTQA4DUcuuR75R7hgPPL5aeWewQAyqgnHbrX90Tv2bMnd911V3bs2JGGhoZs2rQpu3fvzsyZM7v2mTx5ciZMmJD169e/7nE6OjrS3t7ebQEAAIADUY8j+sknn8yIESNSVVWVv/iLv8i9996bd73rXWlpaUllZWVGjRrVbf/a2tq0tLS87vGWLVuWmpqarmX8+PE9/iUAAABgf+hxRB9xxBHZvHlzHn300XziE5/I/Pnz8/TTT+/1AEuXLk1bW1vX0tzcvNfHAgAAgL40pKcvqKyszDve8Y4kybRp07Jx48Z8+ctfzllnnZVdu3Zl+/bt3a5Gt7a2pq6u7nWPV1VVlaqqqp5PDgAAAPvZPj8nurOzMx0dHZk2bVqGDh2apqamrm1btmzJ1q1b09DQsK9vAwAAAGXXoyvRS5cuzezZszNhwoS8/PLLWb16dX70ox9l7dq1qampyfnnn5/Fixdn9OjRqa6uzkUXXZSGhobC38wNAAAAB7IeRfRLL72UP/3TP82LL76YmpqaTJ06NWvXrs2f/MmfJEmuu+66DBo0KHPnzk1HR0dmzZqVm2++uU8GBwAAgP1tn58T3ds8JxoA+pbnRL+a50QDvLntl+dEAwAAwJuNiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUNKTcAwAAAPvfoUu+V+4RDji/XH5quUegH3AlGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAArqUUQvW7YsJ5xwQkaOHJmxY8dmzpw52bJlS7d9du7cmcbGxowZMyYjRozI3Llz09ra2qtDAwAAQDn0KKLXrVuXxsbGbNiwIQ899FB2796dD33oQ9mxY0fXPosWLcr999+fe+65J+vWrcu2bdtyxhln9PrgAAAAsL8N6cnODz74YLefb7/99owdOzabNm3K+973vrS1tWXlypVZvXp1TjnllCTJqlWrMmXKlGzYsCEnnnjiq47Z0dGRjo6Orp/b29v35vcAAACAPrdP90S3tbUlSUaPHp0k2bRpU3bv3p2ZM2d27TN58uRMmDAh69evf81jLFu2LDU1NV3L+PHj92UkAAAA6DN7HdGdnZ1ZuHBhZsyYkaOOOipJ0tLSksrKyowaNarbvrW1tWlpaXnN4yxdujRtbW1dS3Nz896OBAAAAH2qRx/n/l2NjY156qmn8sgjj+zTAFVVVamqqtqnYwAAAMD+sFdXoi+88MJ897vfzQ9/+MMccsghXevr6uqya9eubN++vdv+ra2tqaur26dBAQAAoNx6FNGlUikXXnhh7r333vzgBz/IpEmTum2fNm1ahg4dmqampq51W7ZsydatW9PQ0NA7EwMAAECZ9Ojj3I2NjVm9enXuu+++jBw5sus+55qamgwfPjw1NTU5//zzs3jx4owePTrV1dW56KKL0tDQ8JrfzA0AAAD9SY8iesWKFUmSD3zgA93Wr1q1Kueee26S5LrrrsugQYMyd+7cdHR0ZNasWbn55pt7ZVgAAAAopx5FdKlUesN9hg0blptuuik33XTTXg8FAAAAB6J9ek40AAAAvJmIaAAAAChIRAMAAEBBIhoAAAAKEtEAAABQkIgGAACAgkQ0AAAAFCSiAQAAoCARDQAAAAWJaAAAAChIRAMAAEBBIhoAAAAKEtEAAABQkIgGAACAgkQ0AAAAFCSiAQAAoCARDQAAAAWJaAAAAChIRAMAAEBBIhoAAAAKEtEAAABQkIgGAACAgkQ0AAAAFCSiAQAAoCARDQAAAAWJaAAAAChIRAMAAEBBIhoAAAAKEtEAAABQkIgGAACAgkQ0AAAAFCSiAQAAoCARDQAAAAWJaAAAAChIRAMAAEBBIhoAAAAK6nFEP/zwwznttNNSX1+fioqKrFmzptv2UqmUz33ucxk3blyGDx+emTNn5tlnn+2teQEAAKBsehzRO3bsyDHHHJObbrrpNbdfc801ueGGG3LLLbfk0UcfzUEHHZRZs2Zl586d+zwsAAAAlNOQnr5g9uzZmT179mtuK5VKuf766/PZz342p59+epLka1/7Wmpra7NmzZqcffbZ+zYtAAAAlFGv3hP9/PPPp6WlJTNnzuxaV1NTk+nTp2f9+vWv+ZqOjo60t7d3WwAAAOBA1KsR3dLSkiSpra3ttr62trZr2+9btmxZampqupbx48f35kgAAADQa8r+7dxLly5NW1tb19Lc3FzukQAAAOA19WpE19XVJUlaW1u7rW9tbe3a9vuqqqpSXV3dbQEAAIADUa9G9KRJk1JXV5empqaude3t7Xn00UfT0NDQm28FAAAA+12Pv537lVdeyXPPPdf18/PPP5/Nmzdn9OjRmTBhQhYuXJgrrrgihx9+eCZNmpRLLrkk9fX1mTNnTm/ODQAAAPtdjyP68ccfzx//8R93/bx48eIkyfz583P77bfnM5/5THbs2JELLrgg27dvz8knn5wHH3www4YN672pAQAAoAx6HNEf+MAHUiqVXnd7RUVFPv/5z+fzn//8Pg0GAAAAB5qyfzs3AAAA9BciGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAUJKIBAACgIBENAAAABYloAAAAKEhEAwAAQEEiGgAAAAoS0QAAAFCQiAYAAICCRDQAAAAU1GcRfdNNN+XQQw/NsGHDMn369Dz22GN99VYAAACwX/RJRH/zm9/M4sWLc+mll+aJJ57IMccck1mzZuWll17qi7cDAACA/WJIXxz02muvzcc//vGcd955SZJbbrkl3/ve9/L3f//3WbJkSbd9Ozo60tHR0fVzW1tbkqS9vb0vRgOAN73Ojt+We4QDjv/u4M3InwWv5s+CN6//+9++VCq94b4VpSJ79cCuXbvyR3/0R/nWt76VOXPmdK2fP39+tm/fnvvuu6/b/pdddlkuv/zy3hwBAAAAeqy5uTmHHHLIH9yn169E//rXv86ePXtSW1vbbX1tbW1+9rOfvWr/pUuXZvHixV0/d3Z25r/+678yZsyYVFRU9PZ4A1J7e3vGjx+f5ubmVFdXl3ucAcN57TvObd9wXvuG89o3nNe+49z2Dee1bzivfce57ZlSqZSXX3459fX1b7hvn3ycuyeqqqpSVVXVbd2oUaPKM0w/V11d7f8gfcB57TvObd9wXvuG89o3nNe+49z2Dee1bzivfce5La6mpqbQfr3+xWJvfetbM3jw4LS2tnZb39ramrq6ut5+OwAAANhvej2iKysrM23atDQ1NXWt6+zsTFNTUxoaGnr77QAAAGC/6ZOPcy9evDjz58/P8ccfn/e85z25/vrrs2PHjq5v66Z3VVVV5dJLL33Vx+LZN85r33Fu+4bz2jec177hvPYd57ZvOK99w3ntO85t3+n1b+f+PzfeeGO++MUvpqWlJe9+97tzww03ZPr06X3xVgAAALBf9FlEAwAAwEDT6/dEAwAAwEAlogEAAKAgEQ0AAAAFiWgAAAAoSEQPADfddFMOPfTQDBs2LNOnT89jjz1W7pH6vYcffjinnXZa6uvrU1FRkTVr1pR7pH5v2bJlOeGEEzJy5MiMHTs2c+bMyZYtW8o91oCwYsWKTJ06NdXV1amurk5DQ0MeeOCBco814CxfvjwVFRVZuHBhuUfp1y677LJUVFR0WyZPnlzusQaEX/3qV/nYxz6WMWPGZPjw4Tn66KPz+OOPl3usfu/QQw991b+zFRUVaWxsLPdo/dqePXtyySWXZNKkSRk+fHje/va35wtf+EJ85/G+e/nll7Nw4cJMnDgxw4cPz0knnZSNGzeWe6wBRUT3c9/85jezePHiXHrppXniiSdyzDHHZNasWXnppZfKPVq/tmPHjhxzzDG56aabyj3KgLFu3bo0NjZmw4YNeeihh7J79+586EMfyo4dO8o9Wr93yCGHZPny5dm0aVMef/zxnHLKKTn99NPz05/+tNyjDRgbN27MV7/61UydOrXcowwIRx55ZF588cWu5ZFHHin3SP3eb37zm8yYMSNDhw7NAw88kKeffjpf+tKX8pa3vKXco/V7Gzdu7Pbv60MPPZQkOfPMM8s8Wf929dVXZ8WKFbnxxhvzb//2b7n66qtzzTXX5Ctf+Uq5R+v3/uzP/iwPPfRQvv71r+fJJ5/Mhz70ocycOTO/+tWvyj3agOERV/3c9OnTc8IJJ+TGG29MknR2dmb8+PG56KKLsmTJkjJPNzBUVFTk3nvvzZw5c8o9yoDyH//xHxk7dmzWrVuX973vfeUeZ8AZPXp0vvjFL+b8888v9yj93iuvvJLjjjsuN998c6644oq8+93vzvXXX1/usfqtyy67LGvWrMnmzZvLPcqAsmTJkvzLv/xL/vmf/7ncowx4CxcuzHe/+908++yzqaioKPc4/daHP/zh1NbWZuXKlV3r5s6dm+HDh+fOO+8s42T923//939n5MiRue+++3Lqqad2rZ82bVpmz56dK664oozTDRyuRPdju3btyqZNmzJz5syudYMGDcrMmTOzfv36Mk4Gb6ytrS3J/8YevWfPnj256667smPHjjQ0NJR7nAGhsbExp556arc/a9k3zz77bOrr63PYYYdl3rx52bp1a7lH6ve+853v5Pjjj8+ZZ56ZsWPH5thjj81tt91W7rEGnF27duXOO+/MggULBPQ+Oumkk9LU1JRnnnkmSfKTn/wkjzzySGbPnl3myfq3//mf/8mePXsybNiwbuuHDx/uUz+9aEi5B2Dv/frXv86ePXtSW1vbbX1tbW1+9rOflWkqeGOdnZ1ZuHBhZsyYkaOOOqrc4wwITz75ZBoaGrJz586MGDEi9957b971rneVe6x+76677soTTzzhXrJeNH369Nx+++054ogj8uKLL+byyy/Pe9/73jz11FMZOXJkucfrt37xi19kxYoVWbx4cf7mb/4mGzduzCc/+clUVlZm/vz55R5vwFizZk22b9+ec889t9yj9HtLlixJe3t7Jk+enMGDB2fPnj258sorM2/evHKP1q+NHDkyDQ0N+cIXvpApU6aktrY2//AP/5D169fnHe94R7nHGzBENLDfNTY25qmnnvI3or3oiCOOyObNm9PW1pZvfetbmT9/ftatWyek90Fzc3MuvvjiPPTQQ6/6G3323u9eZZo6dWqmT5+eiRMn5u6773b7wT7o7OzM8ccfn6uuuipJcuyxx+app57KLbfcIqJ70cqVKzN79uzU19eXe5R+7+677843vvGNrF69OkceeWQ2b96chQsXpr6+3r+z++jrX/96FixYkLe97W0ZPHhwjjvuuJxzzjnZtGlTuUcbMER0P/bWt741gwcPTmtra7f1ra2tqaurK9NU8IddeOGF+e53v5uHH344hxxySLnHGTAqKyu7/oZ52rRp2bhxY7785S/nq1/9apkn6782bdqUl156Kccdd1zXuj179uThhx/OjTfemI6OjgwePLiMEw4Mo0aNyjvf+c4899xz5R6lXxs3btyr/tJsypQp+cd//McyTTTwvPDCC/n+97+fb3/72+UeZUD49Kc/nSVLluTss89Okhx99NF54YUXsmzZMhG9j97+9rdn3bp12bFjR9rb2zNu3LicddZZOeyww8o92oDhnuh+rLKyMtOmTUtTU1PXus7OzjQ1NbkXkgNOqVTKhRdemHvvvTc/+MEPMmnSpHKPNKB1dnamo6Oj3GP0ax/84Afz5JNPZvPmzV3L8ccfn3nz5mXz5s0Cupe88sor+fnPf55x48aVe5R+bcaMGa96bOAzzzyTiRMnlmmigWfVqlUZO3Zsty9rYu/99re/zaBB3VNk8ODB6ezsLNNEA89BBx2UcePG5Te/+U3Wrl2b008/vdwjDRiuRPdzixcvzvz583P88cfnPe95T66//vrs2LEj5513XrlH69deeeWVbldFnn/++WzevDmjR4/OhAkTyjhZ/9XY2JjVq1fnvvvuy8iRI9PS0pIkqampyfDhw8s8Xf+2dOnSzJ49OxMmTMjLL7+c1atX50c/+lHWrl1b7tH6tZEjR77qnv2DDjooY8aMcS//PvjUpz6V0047LRMnTsy2bdty6aWXZvDgwTnnnHPKPVq/tmjRopx00km56qqr8tGPfjSPPfZYbr311tx6663lHm1A6OzszKpVqzJ//vwMGeI/n3vDaaedliuvvDITJkzIkUcemR//+Me59tprs2DBgnKP1u+tXbs2pVIpRxxxRJ577rl8+tOfzuTJk/VBbyrR733lK18pTZgwoVRZWVl6z3veU9qwYUO5R+r3fvjDH5aSvGqZP39+uUfrt17rfCYprVq1qtyj9XsLFiwoTZw4sVRZWVk6+OCDSx/84AdL//RP/1TusQak97///aWLL7643GP0a2eddVZp3LhxpcrKytLb3va20llnnVV67rnnyj3WgHD//feXjjrqqFJVVVVp8uTJpVtvvbXcIw0Ya9euLSUpbdmypdyjDBjt7e2liy++uDRhwoTSsGHDSocddljpb//2b0sdHR3lHq3f++Y3v1k67LDDSpWVlaW6urpSY2Njafv27eUea0DxnGgAAAAoyD3RAAAAUJCIBgAAgIJENAAAABQkogEAAKAgEQ0AAAAFiWgAAAAoSEQDAABAQSIaAAAAChLRAAAAUJCIBgAAgIJENAAAABT0/wBIBFtEtPqnDgAAAABJRU5ErkJggg==\n"
          },
          "metadata": {}
        }
      ]
    }
  ]
}