Update Ch01_GS_ComputerVision.ipynb

Author: alvarohernandezalt

Ch01_GS_ComputerVision.ipynb

@@ -259,11 +259,207 @@
         }
       ]
     },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "We defined the same architecture using both the Sequential and Functional APIs, which correspond to the symbolic or declarative way of implementing networks, and also a third tiem using an imperative approach. <br> To make clear that, in the end, the three networks are the same, no matter which approach we took, we trained and evaluated them on the ffamous MNIST dataset, obtaining a decent 98% accuracy on the test set."
+      ],
+      "metadata": {
+        "id": "5X27c4Q6gIE6"
+      }
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "If you are runnning this locally, for the next section of this notebook, you should need to instal `pillow`"
+      ],
+      "metadata": {
+        "id": "yHu72oYzlLfb"
+      }
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "# pip install pillow"
+      ],
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "E24krM9ik9mf",
+        "outputId": "84f1ee4a-5ea9-44a7-d642-49678f7586e1"
+      },
+      "execution_count": 11,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "Requirement already satisfied: pillow in /usr/local/lib/python3.10/dist-packages (9.4.0)\n"
+          ]
+        }
+      ]
+    },
+    {
+      "cell_type": "markdown",
+      "source": [
+        "### Loading images using the Keras API"
+      ],
+      "metadata": {
+        "id": "iq1Rp1XZlEYT"
+      }
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "# Import the necessary packages\n",
+        "\n",
+        "import glob\n",
+        "import os\n",
+        "import tarfile\n",
+        "import matplotlib.pyplot as plt\n",
+        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
+        "from tensorflow.keras.preprocessing.image import load_img, img_to_array\n",
+        "from tensorflow.keras.utils import get_file"
+      ],
+      "metadata": {
+        "id": "bVQnDzb4lhRJ"
+      },
+      "execution_count": 12,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "# Define the URL and destination of the CINIC-10 dataset, an alternative to the famous CIFAR-10 dataset\n",
+        "\n",
+        "DATASET_URL = 'https://datashare.is.ed.ac.uk/bitstream/handle/10283/3192/CINIC-10.tar.gz?sequence=4&isAllowed=y'\n",
+        "DATA_NAME = 'cinic10'\n",
+        "FILE_EXTENSION = 'tar.gz'\n",
+        "FILE_NAME = '.'.join([DATA_NAME, FILE_EXTENSION])"
+      ],
+      "metadata": {
+        "id": "RZ4DwQfAmtGW"
+      },
+      "execution_count": 13,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "# Download and decompress the data. By default, it will be stored in ~/.keras/datasets/<FILE_NAME>\n",
+        "\n",
+        "downloaded_file_location = get_file(origin=DATASET_URL, fname=FILE_NAME, extract=False)\n"
+      ],
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "7_rOUoZOn2A7",
+        "outputId": "958be2e5-4884-4924-8834-e63899467e47"
+      },
+      "execution_count": 14,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "Downloading data from https://datashare.is.ed.ac.uk/bitstream/handle/10283/3192/CINIC-10.tar.gz?sequence=4&isAllowed=y\n",
+            "687544992/687544992 [==============================] - 518s 1us/step\n"
+          ]
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "# Build the path to the data directory based on the location of the downloaded file\n",
+        "data_directory, _ = downloaded_file_location.rsplit(os.path.sep, maxsplit=1)\n",
+        "data_directory = os.path.sep.join([data_directory, DATA_NAME])"
+      ],
+      "metadata": {
+        "id": "39rcGo7UqQ7q"
+      },
+      "execution_count": 15,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "# Only extract the data if it hasn´t been extracted yet.\n",
+        "if not os.path.exists(data_directory) :\n",
+        "  tar = tarfile.open(downloaded_file_location)\n",
+        "  tar.extractall(data_directory)"
+      ],
+      "metadata": {
+        "id": "yuad3_COtRLW"
+      },
+      "execution_count": 16,
+      "outputs": []
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "# Load all the image paths and print the number of images\n",
+        "data_pattern = os.path.sep.join([data_directory, '*/*/*.png'])\n",
+        "image_paths = list(glob.glob(data_pattern))\n",
+        "print(f'There are {len(image_paths):,} images in the dataset')"
+      ],
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "9K5XBaYKt7NF",
+        "outputId": "54c06705-42e9-4f43-afef-a738c9504346"
+      },
+      "execution_count": 17,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "There are 270,000 images in the dataset\n"
+          ]
+        }
+      ]
+    },
+    {
+      "cell_type": "code",
+      "source": [
+        "# lOAD A SINGLE IMAGE FROM THE DATA SET AND PRINT ITS METADATA\n",
+        "\n",
+        "sample_image = load_img(image_paths[0])\n",
+        "print(f'Image type: {type(sample_image)}')\n",
+        "print(f'Image format: {sample_image.format}')\n",
+        "print(f'Image mode: {sample_image.mode}')\n",
+        "print(f'Image size: {sample_image.size}')"
+      ],
+      "metadata": {
+        "colab": {
+          "base_uri": "https://localhost:8080/"
+        },
+        "id": "VTogeqSK3H4C",
+        "outputId": "746700fb-73c1-4671-c31c-fdcbb49ee80f"
+      },
+      "execution_count": 19,
+      "outputs": [
+        {
+          "output_type": "stream",
+          "name": "stdout",
+          "text": [
+            "Image type: <class 'PIL.PngImagePlugin.PngImageFile'>\n",
+            "Image format: PNG\n",
+            "Image mode: RGB\n",
+            "Image size: (32, 32)\n"
+          ]
+        }
+      ]
+    },
     {
       "cell_type": "code",
       "source": [],
       "metadata": {
-        "id": "5X27c4Q6gIE6"
+        "id": "f55iZIpG8ZGP"
       },
       "execution_count": null,
       "outputs": []