04 - Inception_model.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "rkVdJTgVqasW"
      },
      "source": [
        "<a href=\"https://githubtocolab.com/luismateoh/proyecto-ai4eng/blob/main/04%20-%20Inception_model.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open in Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "FBwLzU87ohQm"
      },
      "source": [
        "\n",
        "\n",
        "#  Modelo 04\n",
        "\n",
        "**Autor:** [luismateoh](https://github.com/luismateoh)<br>\n",
        "**Día de creación:** 2022/09/21<br>\n",
        "**Descripción:** Modelo Inception-v3<br>\n",
        "**Url Dataset:** https://data.mendeley.com/datasets/snkd93bnjr/1\n",
        "\n",
        "Se sugiere cambiar el tipo de entorno de ejecución a GPU"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "OsRiDuIhohQz"
      },
      "source": [
        "\n",
        "\n",
        "## Introducción sobre la base de datos\n",
        "\n",
        "El conjunto de datos contiene un total de 17.092 imágenes de células normales individuales, que se adquirieron utilizando el analizador CellaVision DM96 en el Laboratorio Central del Hospital Clínic de Barcelona. El conjunto de datos está organizado en los siguientes ocho grupos: neutrófilos, eosinófilos, basófilos, linfocitos, monocitos, granulocitos inmaduros (promielocitos, mielocitos y metamielocitos), eritroblastos y plaquetas o trombocitos. El tamaño de las imágenes es de 360 ​​x 363 píxeles, en formato JPG, y fueron anotadas por clínicos patólogos expertos. Las imágenes fueron capturadas de individuos sin infección, enfermedad hematológica u oncológica y libres de cualquier tratamiento farmacológico en el momento de la extracción de sangre. Este conjunto de datos etiquetados de alta calidad se puede usar para entrenar y probar modelos de aprendizaje automático y aprendizaje profundo para reconocer diferentes tipos de células sanguíneas periféricas normales. Hasta donde sabemos, este es el primer conjunto disponible públicamente con un gran número de células sanguíneas periféricas normales, por lo que se espera que sea un conjunto de datos canónico para la evaluación comparativa del modelo.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "b0h1ikw6ohQ1"
      },
      "source": [
        "## Configuración del entorno y dependencias.\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uRPOPynmT_cE"
      },
      "source": [
        "Se instalan las dependencias que no estan disponibles en colab, se actualiza matplotlib, y se debe reiniciar la maquina virtual."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ZPqgTsGQPm-0",
        "outputId": "62f379ea-53cc-435c-ce23-3063a20b288b"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Requirement already satisfied: imagesize in /usr/local/lib/python3.7/dist-packages (1.4.1)\n",
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Requirement already satisfied: visualkeras in /usr/local/lib/python3.7/dist-packages (0.0.2)\n",
            "Requirement already satisfied: aggdraw>=1.3.11 in /usr/local/lib/python3.7/dist-packages (from visualkeras) (1.3.15)\n",
            "Requirement already satisfied: pillow>=6.2.0 in /usr/local/lib/python3.7/dist-packages (from visualkeras) (7.1.2)\n",
            "Requirement already satisfied: numpy>=1.18.1 in /usr/local/lib/python3.7/dist-packages (from visualkeras) (1.21.6)\n",
            "Looking in indexes: https://pypi.org/simple, https://us-python.pkg.dev/colab-wheels/public/simple/\n",
            "Requirement already satisfied: matplotlib in /usr/local/lib/python3.7/dist-packages (3.5.3)\n",
            "Requirement already satisfied: fonttools>=4.22.0 in /usr/local/lib/python3.7/dist-packages (from matplotlib) (4.37.1)\n",
            "Requirement already satisfied: python-dateutil>=2.7 in /usr/local/lib/python3.7/dist-packages (from matplotlib) (2.8.2)\n",
            "Requirement already satisfied: pyparsing>=2.2.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib) (3.0.9)\n",
            "Requirement already satisfied: numpy>=1.17 in /usr/local/lib/python3.7/dist-packages (from matplotlib) (1.21.6)\n",
            "Requirement already satisfied: cycler>=0.10 in /usr/local/lib/python3.7/dist-packages (from matplotlib) (0.11.0)\n",
            "Requirement already satisfied: pillow>=6.2.0 in /usr/local/lib/python3.7/dist-packages (from matplotlib) (7.1.2)\n",
            "Requirement already satisfied: kiwisolver>=1.0.1 in /usr/local/lib/python3.7/dist-packages (from matplotlib) (1.4.4)\n",
            "Requirement already satisfied: packaging>=20.0 in /usr/local/lib/python3.7/dist-packages (from matplotlib) (21.3)\n",
            "Requirement already satisfied: typing-extensions in /usr/local/lib/python3.7/dist-packages (from kiwisolver>=1.0.1->matplotlib) (4.1.1)\n",
            "Requirement already satisfied: six>=1.5 in /usr/local/lib/python3.7/dist-packages (from python-dateutil>=2.7->matplotlib) (1.15.0)\n"
          ]
        }
      ],
      "source": [
        "!pip install imagesize\n",
        "!pip install visualkeras\n",
        "!pip install matplotlib --upgrade"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "XGfeyrwKT_cG"
      },
      "source": [
        "Se importan las dependencias necesarias para ejecutar el notebook."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "WT74-UZyohQ2"
      },
      "outputs": [],
      "source": [
        "import os\n",
        "import os.path\n",
        "import imagesize\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import seaborn as sns\n",
        "from pathlib import Path\n",
        "from collections import Counter\n",
        "from ast import literal_eval\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "import tensorflow as tf\n",
        "\n",
        "import keras \n",
        "from keras import layers\n",
        "from keras.models import Sequential\n",
        "from tensorflow.keras import Input\n",
        "from keras.layers import Rescaling\n",
        "from keras.layers import Conv2D\n",
        "from keras.layers import MaxPooling2D\n",
        "from keras.layers import Flatten\n",
        "from keras.layers import Dense\n",
        "from keras.layers import Dropout\n",
        "from keras.layers import BatchNormalization\n",
        "\n",
        "\n",
        "import visualkeras\n",
        "\n",
        "from sklearn.metrics import classification_report\n",
        "from sklearn.metrics import confusion_matrix"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "E9kNOy_eohQ5"
      },
      "source": [
        "## Carga de la base de datos\n",
        "\n",
        "### Descarga de datos sin procesar\n",
        "\n",
        "Primero, descarguemos el archivo ZIP de 265M de los datos sin procesar:\n",
        "\n",
        "Si la descarga falla o el link esta roto, revisar la pagina https://data.mendeley.com/datasets/snkd93bnjr/1 para obtener el enlace actualizado. \n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "_UoC-7fbohQ7",
        "outputId": "a8e060e0-0c9e-4014-dbe7-bdc30ffa37f7"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current\n",
            "                                 Dload  Upload   Total   Spent    Left  Speed\n",
            "100  265M  100  265M    0     0  11.7M      0  0:00:22  0:00:22 --:--:-- 13.6M\n"
          ]
        }
      ],
      "source": [
        "!curl -O https://prod-dcd-datasets-cache-zipfiles.s3.eu-west-1.amazonaws.com/snkd93bnjr-1.zip\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "FGO8DAv_ohQ8",
        "outputId": "384f0862-7a36-4df8-a82b-30deeed63f87"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "__MACOSX  PBC_dataset_normal_DIB  sample_data\n"
          ]
        }
      ],
      "source": [
        "!unzip -q snkd93bnjr-1.zip\n",
        "!unzip -q PBC_dataset_normal_DIB.zip\n",
        "!rm -f snkd93bnjr-1.zip PBC_dataset_normal_DIB.zip\n",
        "!ls\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "YgESie9MohQ9"
      },
      "source": [
        "Ahora tenemos una carpeta `PBC_dataset_normal_DIB` que contiene ocho subcarpetas. Cada subcarpeta contiene archivos de imagen para cada categoría\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "9z0zLlUVohQ-",
        "outputId": "1a4dfdee-521e-4194-fd63-6b8e8864b083"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "basophil    erythroblast  lymphocyte  neutrophil\n",
            "eosinophil  ig\t\t  monocyte    platelet\n"
          ]
        }
      ],
      "source": [
        "!ls PBC_dataset_normal_DIB  \n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "uKr5EgyWohQ_"
      },
      "source": [
        "### Filtrar imágenes corruptas\n",
        "\n",
        "Cuando se trabaja con una gran cantidad de datos de imágenes del mundo real, las imágenes corruptas son comunes.\n",
        "Filtramos las imágenes mal codificadas que no tengan la cadena \"JFIF\" en su cabecera.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YXWi7rMqohRA",
        "outputId": "537d7fa6-53c0-4094-be84-c71e85f0c81f"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Deleted 1 images\n"
          ]
        }
      ],
      "source": [
        "\n",
        "\n",
        "num_skipped = 0\n",
        "for folder_name in (\"basophil\", \"erythroblast\", \"lymphocyte\", \"neutrophil\", \"eosinophil\", \"ig\", \"monocyte\", \"platelet\"):\n",
        "    folder_path = os.path.join(\"PBC_dataset_normal_DIB\", folder_name)\n",
        "    for fname in os.listdir(folder_path):\n",
        "        fpath = os.path.join(folder_path, fname)\n",
        "        try:\n",
        "            fobj = open(fpath, \"rb\")\n",
        "            is_jfif = tf.compat.as_bytes(\"JFIF\") in fobj.peek(10)\n",
        "        finally:\n",
        "            fobj.close()\n",
        "\n",
        "        if not is_jfif:\n",
        "            num_skipped += 1\n",
        "            # Delete corrupted image\n",
        "            os.remove(fpath)\n",
        "\n",
        "print(\"Deleted %d images\" % num_skipped)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "A-QK5FTYohRB"
      },
      "source": [
        "## Generación del `Dataset`\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "PR1j8gavZMKC"
      },
      "source": [
        "Creamos primero un dataset con el 80% de las imagenes para el entrenamiento, y el 20 restante quedarán de validación, que a su vez partiremos para obtener un tercer conjuto de datos de test."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "P2v7G6UCohRC",
        "outputId": "08605650-6356-4ec0-ee9e-1985a7b57994"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Firt Dataset partition\n",
            "Found 17092 files belonging to 8 classes.\n",
            "Using 13674 files for training.\n",
            "Found 17092 files belonging to 8 classes.\n",
            "Using 3418 files for validation.\n",
            "\n",
            "Second Dataset partition\n",
            "Train dataset: 13674 images.\n",
            "Validation dataset: 3258 images.\n",
            "Test dataset: 160 images.\n"
          ]
        }
      ],
      "source": [
        "\n",
        "#Para el modelo que usaremos debe ser de (227,227)\n",
        "image_size = (224, 224)\n",
        "batch_size = 32\n",
        "\n",
        "print(\"Firt Dataset partition\")\n",
        "train_ds = tf.keras.preprocessing.image_dataset_from_directory(\n",
        "    \"PBC_dataset_normal_DIB\",\n",
        "    validation_split=0.20,\n",
        "    subset=\"training\",\n",
        "    seed=2022,\n",
        "    image_size=image_size,\n",
        "    batch_size=batch_size,\n",
        "    label_mode='categorical'\n",
        ")\n",
        "val_ds = tf.keras.preprocessing.image_dataset_from_directory(\n",
        "    \"PBC_dataset_normal_DIB\",\n",
        "    labels='inferred',\n",
        "    validation_split=0.20,\n",
        "    subset=\"validation\",\n",
        "    seed=2022,\n",
        "    image_size=image_size,\n",
        "    batch_size=batch_size,\n",
        "    label_mode='categorical'\n",
        ")\n",
        "\n",
        "\n",
        "print(\"\\nSecond Dataset partition\")\n",
        "\n",
        "test_dataset = val_ds.take(5)\n",
        "val_ds = val_ds.skip(5)\n",
        "\n",
        "print(f\"Train dataset: {len(np.concatenate([i for x, i in train_ds], axis=0))} images.\")\n",
        "print(f\"Validation dataset: {len(np.concatenate([i for x, i in val_ds], axis=0))} images.\")\n",
        "print(f\"Test dataset: {len(np.concatenate([i for x, i in test_dataset], axis=0))} images.\")"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "K0gWqmpdohRE"
      },
      "source": [
        "## Usando el aumento de datos de imagen\n",
        "\n",
        "Cuando no tiene un conjunto de datos de imagen grande, es una buena práctica\n",
        "introduce la diversidad de la muestra mediante la aplicación de transformaciones aleatorias pero realistas a la\n",
        "imágenes de entrenamiento, como volteo horizontal aleatorio o pequeñas rotaciones aleatorias. Este\n",
        "ayuda a exponer el modelo a diferentes aspectos de los datos de entrenamiento mientras se disminuye sobreajuste.\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "Ke_wLv5ZohRF"
      },
      "outputs": [],
      "source": [
        "data_augmentation = keras.Sequential(\n",
        "    [\n",
        "        layers.RandomFlip(\"horizontal_and_vertical\"),\n",
        "        layers.RandomRotation(0.2),\n",
        "        layers.RandomZoom(.3, .3)\n",
        "\n",
        "    ]\n",
        ")\n",
        "\n",
        "augmented_train_ds = train_ds.map(\n",
        "  lambda x, y: (data_augmentation(x, training=True), y))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "vZhXZ4I9cPoi"
      },
      "source": [
        "## Configurar el conjunto de datos para el rendimiento\n",
        "\n",
        "Asegurémonos de usar la captación previa almacenada en búfer para que podamos obtener datos del disco sin tener bloqueos de I/O:\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "62n-RAUgcPaR"
      },
      "outputs": [],
      "source": [
        "train_ds = train_ds.prefetch(buffer_size=32)\n",
        "val_ds = val_ds.prefetch(buffer_size=32)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "EOurend8ohRJ"
      },
      "source": [
        "## Construir un modelo\n",
        "\n",
        "Construiremos una versión de Inception-v3\n",
        "\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "YFdK2wTLvrV7",
        "outputId": "4e21936a-7c31-49db-dc3a-7ad61a66c9db"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"InceptionV3\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                   Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            " input_1 (InputLayer)           [(None, 224, 224, 3  0           []                               \n",
            "                                )]                                                                \n",
            "                                                                                                  \n",
            " sequential (Sequential)        (None, 224, 224, 3)  0           ['input_1[0][0]']                \n",
            "                                                                                                  \n",
            " rescaling (Rescaling)          (None, 224, 224, 3)  0           ['sequential[0][0]']             \n",
            "                                                                                                  \n",
            " conv2d (Conv2D)                (None, 112, 112, 32  896         ['rescaling[0][0]']              \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " batch_normalization (BatchNorm  (None, 112, 112, 32  128        ['conv2d[0][0]']                 \n",
            " alization)                     )                                                                 \n",
            "                                                                                                  \n",
            " activation (Activation)        (None, 112, 112, 32  0           ['batch_normalization[0][0]']    \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " conv2d_1 (Conv2D)              (None, 112, 112, 64  18496       ['activation[0][0]']             \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " batch_normalization_1 (BatchNo  (None, 112, 112, 64  256        ['conv2d_1[0][0]']               \n",
            " rmalization)                   )                                                                 \n",
            "                                                                                                  \n",
            " activation_1 (Activation)      (None, 112, 112, 64  0           ['batch_normalization_1[0][0]']  \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " activation_2 (Activation)      (None, 112, 112, 64  0           ['activation_1[0][0]']           \n",
            "                                )                                                                 \n",
            "                                                                                                  \n",
            " separable_conv2d (SeparableCon  (None, 112, 112, 12  8896       ['activation_2[0][0]']           \n",
            " v2D)                           8)                                                                \n",
            "                                                                                                  \n",
            " batch_normalization_2 (BatchNo  (None, 112, 112, 12  512        ['separable_conv2d[0][0]']       \n",
            " rmalization)                   8)                                                                \n",
            "                                                                                                  \n",
            " activation_3 (Activation)      (None, 112, 112, 12  0           ['batch_normalization_2[0][0]']  \n",
            "                                8)                                                                \n",
            "                                                                                                  \n",
            " separable_conv2d_1 (SeparableC  (None, 112, 112, 12  17664      ['activation_3[0][0]']           \n",
            " onv2D)                         8)                                                                \n",
            "                                                                                                  \n",
            " batch_normalization_3 (BatchNo  (None, 112, 112, 12  512        ['separable_conv2d_1[0][0]']     \n",
            " rmalization)                   8)                                                                \n",
            "                                                                                                  \n",
            " max_pooling2d (MaxPooling2D)   (None, 56, 56, 128)  0           ['batch_normalization_3[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_2 (Conv2D)              (None, 56, 56, 128)  8320        ['activation_1[0][0]']           \n",
            "                                                                                                  \n",
            " add (Add)                      (None, 56, 56, 128)  0           ['max_pooling2d[0][0]',          \n",
            "                                                                  'conv2d_2[0][0]']               \n",
            "                                                                                                  \n",
            " activation_4 (Activation)      (None, 56, 56, 128)  0           ['add[0][0]']                    \n",
            "                                                                                                  \n",
            " separable_conv2d_2 (SeparableC  (None, 56, 56, 256)  34176      ['activation_4[0][0]']           \n",
            " onv2D)                                                                                           \n",
            "                                                                                                  \n",
            " batch_normalization_4 (BatchNo  (None, 56, 56, 256)  1024       ['separable_conv2d_2[0][0]']     \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_5 (Activation)      (None, 56, 56, 256)  0           ['batch_normalization_4[0][0]']  \n",
            "                                                                                                  \n",
            " separable_conv2d_3 (SeparableC  (None, 56, 56, 256)  68096      ['activation_5[0][0]']           \n",
            " onv2D)                                                                                           \n",
            "                                                                                                  \n",
            " batch_normalization_5 (BatchNo  (None, 56, 56, 256)  1024       ['separable_conv2d_3[0][0]']     \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " max_pooling2d_1 (MaxPooling2D)  (None, 28, 28, 256)  0          ['batch_normalization_5[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_3 (Conv2D)              (None, 28, 28, 256)  33024       ['add[0][0]']                    \n",
            "                                                                                                  \n",
            " add_1 (Add)                    (None, 28, 28, 256)  0           ['max_pooling2d_1[0][0]',        \n",
            "                                                                  'conv2d_3[0][0]']               \n",
            "                                                                                                  \n",
            " activation_6 (Activation)      (None, 28, 28, 256)  0           ['add_1[0][0]']                  \n",
            "                                                                                                  \n",
            " separable_conv2d_4 (SeparableC  (None, 28, 28, 512)  133888     ['activation_6[0][0]']           \n",
            " onv2D)                                                                                           \n",
            "                                                                                                  \n",
            " batch_normalization_6 (BatchNo  (None, 28, 28, 512)  2048       ['separable_conv2d_4[0][0]']     \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_7 (Activation)      (None, 28, 28, 512)  0           ['batch_normalization_6[0][0]']  \n",
            "                                                                                                  \n",
            " separable_conv2d_5 (SeparableC  (None, 28, 28, 512)  267264     ['activation_7[0][0]']           \n",
            " onv2D)                                                                                           \n",
            "                                                                                                  \n",
            " batch_normalization_7 (BatchNo  (None, 28, 28, 512)  2048       ['separable_conv2d_5[0][0]']     \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " max_pooling2d_2 (MaxPooling2D)  (None, 14, 14, 512)  0          ['batch_normalization_7[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_4 (Conv2D)              (None, 14, 14, 512)  131584      ['add_1[0][0]']                  \n",
            "                                                                                                  \n",
            " add_2 (Add)                    (None, 14, 14, 512)  0           ['max_pooling2d_2[0][0]',        \n",
            "                                                                  'conv2d_4[0][0]']               \n",
            "                                                                                                  \n",
            " activation_8 (Activation)      (None, 14, 14, 512)  0           ['add_2[0][0]']                  \n",
            "                                                                                                  \n",
            " separable_conv2d_6 (SeparableC  (None, 14, 14, 728)  378072     ['activation_8[0][0]']           \n",
            " onv2D)                                                                                           \n",
            "                                                                                                  \n",
            " batch_normalization_8 (BatchNo  (None, 14, 14, 728)  2912       ['separable_conv2d_6[0][0]']     \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " activation_9 (Activation)      (None, 14, 14, 728)  0           ['batch_normalization_8[0][0]']  \n",
            "                                                                                                  \n",
            " separable_conv2d_7 (SeparableC  (None, 14, 14, 728)  537264     ['activation_9[0][0]']           \n",
            " onv2D)                                                                                           \n",
            "                                                                                                  \n",
            " batch_normalization_9 (BatchNo  (None, 14, 14, 728)  2912       ['separable_conv2d_7[0][0]']     \n",
            " rmalization)                                                                                     \n",
            "                                                                                                  \n",
            " max_pooling2d_3 (MaxPooling2D)  (None, 7, 7, 728)   0           ['batch_normalization_9[0][0]']  \n",
            "                                                                                                  \n",
            " conv2d_5 (Conv2D)              (None, 7, 7, 728)    373464      ['add_2[0][0]']                  \n",
            "                                                                                                  \n",
            " add_3 (Add)                    (None, 7, 7, 728)    0           ['max_pooling2d_3[0][0]',        \n",
            "                                                                  'conv2d_5[0][0]']               \n",
            "                                                                                                  \n",
            " separable_conv2d_8 (SeparableC  (None, 7, 7, 1024)  753048      ['add_3[0][0]']                  \n",
            " onv2D)                                                                                           \n",
            "                                                                                                  \n",
            " batch_normalization_10 (BatchN  (None, 7, 7, 1024)  4096        ['separable_conv2d_8[0][0]']     \n",
            " ormalization)                                                                                    \n",
            "                                                                                                  \n",
            " activation_10 (Activation)     (None, 7, 7, 1024)   0           ['batch_normalization_10[0][0]'] \n",
            "                                                                                                  \n",
            " global_average_pooling2d (Glob  (None, 1024)        0           ['activation_10[0][0]']          \n",
            " alAveragePooling2D)                                                                              \n",
            "                                                                                                  \n",
            " dropout (Dropout)              (None, 1024)         0           ['global_average_pooling2d[0][0]'\n",
            "                                                                 ]                                \n",
            "                                                                                                  \n",
            " dense (Dense)                  (None, 8)            8200        ['dropout[0][0]']                \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 2,789,824\n",
            "Trainable params: 2,781,088\n",
            "Non-trainable params: 8,736\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "data_augmentation = keras.Sequential(\n",
        "    [\n",
        "        layers.RandomFlip(\"horizontal_and_vertical\"),\n",
        "        layers.RandomRotation(0.2),\n",
        "        layers.RandomZoom(.3, .3)\n",
        "\n",
        "    ]\n",
        ")\n",
        "\n",
        "\n",
        "def make_model(input_shape, num_classes):\n",
        "    inputs = keras.Input(shape=input_shape)\n",
        "    # Image augmentation block\n",
        "    x = data_augmentation(inputs)\n",
        "\n",
        "    # Entry block\n",
        "    x = layers.Rescaling(1.0 / 255)(x)\n",
        "    x = layers.Conv2D(32, 3, strides=2, padding=\"same\")(x)\n",
        "    x = layers.BatchNormalization()(x)\n",
        "    x = layers.Activation(\"relu\")(x)\n",
        "\n",
        "    x = layers.Conv2D(64, 3, padding=\"same\")(x)\n",
        "    x = layers.BatchNormalization()(x)\n",
        "    x = layers.Activation(\"relu\")(x)\n",
        "\n",
        "    previous_block_activation = x  # Set aside residual\n",
        "\n",
        "    for size in [128, 256, 512, 728]:\n",
        "        x = layers.Activation(\"relu\")(x)\n",
        "        x = layers.SeparableConv2D(size, 3, padding=\"same\")(x)\n",
        "        x = layers.BatchNormalization()(x)\n",
        "\n",
        "        x = layers.Activation(\"relu\")(x)\n",
        "        x = layers.SeparableConv2D(size, 3, padding=\"same\")(x)\n",
        "        x = layers.BatchNormalization()(x)\n",
        "\n",
        "        x = layers.MaxPooling2D(3, strides=2, padding=\"same\")(x)\n",
        "\n",
        "        # Project residual\n",
        "        residual = layers.Conv2D(size, 1, strides=2, padding=\"same\")(\n",
        "            previous_block_activation\n",
        "        )\n",
        "        x = layers.add([x, residual])  # Add back residual\n",
        "        previous_block_activation = x  # Set aside next residual\n",
        "\n",
        "    x = layers.SeparableConv2D(1024, 3, padding=\"same\")(x)\n",
        "    x = layers.BatchNormalization()(x)\n",
        "    x = layers.Activation(\"relu\")(x)\n",
        "\n",
        "    x = layers.GlobalAveragePooling2D()(x)\n",
        "    if num_classes == 2:\n",
        "        activation = \"sigmoid\"\n",
        "        units = 1\n",
        "    else:\n",
        "        activation = \"softmax\"\n",
        "        units = num_classes\n",
        "\n",
        "    x = layers.Dropout(0.5)(x)\n",
        "    outputs = layers.Dense(units, activation=activation)(x)\n",
        "    return keras.Model(inputs, outputs)\n",
        "\n",
        "\n",
        "model = make_model(input_shape=(224,224) + (3,), num_classes=8)\n",
        "model._name=\"InceptionV3\"\n",
        "model.summary()\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 662
        },
        "id": "udyKtAYGfWlf",
        "outputId": "8c91fd56-2923-4a22-c56a-889f0b2d1227"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<PIL.Image.Image image mode=RGBA size=2859x1231 at 0x7FE3F0604D90>"
            ]
          },
          "execution_count": 10,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "visualkeras.layered_view(model,legend=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "y5vGWUpDohRK"
      },
      "source": [
        "## Entrenar el modelo\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "rDUTPTQXohRL",
        "outputId": "5d217f76-3071-419b-9dcb-2c861a53561e"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/25\n",
            "428/428 [==============================] - 184s 391ms/step - loss: 0.4421 - accuracy: 0.8455 - val_loss: 11.2530 - val_accuracy: 0.1344\n",
            "Epoch 2/25\n",
            "428/428 [==============================] - 167s 389ms/step - loss: 0.2548 - accuracy: 0.9137 - val_loss: 1.5222 - val_accuracy: 0.6234\n",
            "Epoch 3/25\n",
            "428/428 [==============================] - 166s 388ms/step - loss: 0.1876 - accuracy: 0.9379 - val_loss: 3.6204 - val_accuracy: 0.5356\n",
            "Epoch 4/25\n",
            "428/428 [==============================] - 166s 388ms/step - loss: 0.1673 - accuracy: 0.9470 - val_loss: 2.1748 - val_accuracy: 0.5147\n",
            "Epoch 5/25\n",
            "428/428 [==============================] - 166s 387ms/step - loss: 0.1427 - accuracy: 0.9519 - val_loss: 1.2622 - val_accuracy: 0.5930\n",
            "Epoch 6/25\n",
            "428/428 [==============================] - 166s 387ms/step - loss: 0.1291 - accuracy: 0.9574 - val_loss: 0.3376 - val_accuracy: 0.8950\n",
            "Epoch 7/25\n",
            "428/428 [==============================] - 166s 387ms/step - loss: 0.1198 - accuracy: 0.9602 - val_loss: 3.7467 - val_accuracy: 0.4681\n",
            "Epoch 8/25\n",
            "428/428 [==============================] - 166s 387ms/step - loss: 0.1148 - accuracy: 0.9607 - val_loss: 0.4283 - val_accuracy: 0.8797\n",
            "Epoch 9/25\n",
            "428/428 [==============================] - 166s 387ms/step - loss: 0.1094 - accuracy: 0.9639 - val_loss: 0.2666 - val_accuracy: 0.9116\n",
            "Epoch 10/25\n",
            "428/428 [==============================] - 166s 386ms/step - loss: 0.1095 - accuracy: 0.9643 - val_loss: 0.6933 - val_accuracy: 0.7953\n",
            "Epoch 11/25\n",
            "428/428 [==============================] - 165s 385ms/step - loss: 0.1108 - accuracy: 0.9642 - val_loss: 0.1967 - val_accuracy: 0.9365\n",
            "Epoch 12/25\n",
            "428/428 [==============================] - 166s 386ms/step - loss: 0.0953 - accuracy: 0.9688 - val_loss: 0.2906 - val_accuracy: 0.9095\n",
            "Epoch 13/25\n",
            "428/428 [==============================] - 165s 386ms/step - loss: 0.0948 - accuracy: 0.9684 - val_loss: 0.2093 - val_accuracy: 0.9309\n",
            "Epoch 14/25\n",
            "428/428 [==============================] - 165s 386ms/step - loss: 0.0856 - accuracy: 0.9710 - val_loss: 0.5894 - val_accuracy: 0.8432\n",
            "Epoch 15/25\n",
            "428/428 [==============================] - 165s 386ms/step - loss: 0.0913 - accuracy: 0.9711 - val_loss: 0.1791 - val_accuracy: 0.9515\n",
            "Epoch 16/25\n",
            "428/428 [==============================] - 165s 385ms/step - loss: 0.0859 - accuracy: 0.9725 - val_loss: 0.8980 - val_accuracy: 0.7990\n",
            "Epoch 17/25\n",
            "428/428 [==============================] - 166s 387ms/step - loss: 0.0913 - accuracy: 0.9714 - val_loss: 0.5005 - val_accuracy: 0.8407\n",
            "Epoch 18/25\n",
            "428/428 [==============================] - 165s 386ms/step - loss: 0.0798 - accuracy: 0.9738 - val_loss: 1.9439 - val_accuracy: 0.6136\n",
            "Epoch 19/25\n",
            "428/428 [==============================] - 166s 386ms/step - loss: 0.0772 - accuracy: 0.9754 - val_loss: 0.1402 - val_accuracy: 0.9527\n",
            "Epoch 20/25\n",
            "428/428 [==============================] - 166s 387ms/step - loss: 0.0797 - accuracy: 0.9732 - val_loss: 0.7392 - val_accuracy: 0.8263\n",
            "Epoch 21/25\n",
            "428/428 [==============================] - 166s 386ms/step - loss: 0.0671 - accuracy: 0.9789 - val_loss: 0.0994 - val_accuracy: 0.9672\n",
            "Epoch 22/25\n",
            "428/428 [==============================] - 165s 385ms/step - loss: 0.0769 - accuracy: 0.9757 - val_loss: 0.1641 - val_accuracy: 0.9451\n",
            "Epoch 23/25\n",
            "428/428 [==============================] - 166s 386ms/step - loss: 0.0808 - accuracy: 0.9738 - val_loss: 1.9992 - val_accuracy: 0.5387\n",
            "Epoch 24/25\n",
            "428/428 [==============================] - 166s 387ms/step - loss: 0.0760 - accuracy: 0.9751 - val_loss: 4.4364 - val_accuracy: 0.3717\n",
            "Epoch 25/25\n",
            "428/428 [==============================] - 166s 387ms/step - loss: 0.0742 - accuracy: 0.9765 - val_loss: 0.3849 - val_accuracy: 0.8929\n"
          ]
        }
      ],
      "source": [
        "epochs = 25\n",
        "\n",
        "callbacks = [\n",
        "    keras.callbacks.ModelCheckpoint(\"save_at_{epoch}.h5\"),\n",
        "]\n",
        "\n",
        "model.compile(optimizer='Adam', loss='categorical_crossentropy', metrics = ['accuracy'])\n",
        "hist = model.fit(train_ds, \n",
        "                 epochs=epochs, \n",
        "                 callbacks=callbacks, \n",
        "                 validation_data=val_ds)\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "nOH3YYfDohRL"
      },
      "source": [
        "## Evaluación y metricas ejecutando inferencia sobre nuevos datos"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "pmI3ZDJ7iOkm"
      },
      "source": [
        "### Resumen de precisión y función de perdida del modelo"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 367
        },
        "id": "yXTLJXY8iHRo",
        "outputId": "59820b14-b13d-4819-a3d2-e00b862c9971"
      },
      "outputs": [
        {
          "data": {
            "text/plain": [
              "<matplotlib.legend.Legend at 0x7fe3865c56d0>"
            ]
          },
          "execution_count": 12,
          "metadata": {},
          "output_type": "execute_result"
        },
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 1440x360 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "fig, (ax1, ax2) = plt.subplots(1, 2, sharex=False, sharey=False)\n",
        "fig.set_size_inches(20,5)\n",
        "\n",
        "ax1.plot(hist.history['accuracy'], label='train')\n",
        "ax1.plot(hist.history['val_accuracy'],label='test')\n",
        "ax1.set_title('model accuracy')\n",
        "ax1.set_ylabel('accuracy')\n",
        "ax1.set_xlabel('epoch')\n",
        "ax1.legend(loc='upper left')\n",
        "\n",
        "ax2.plot(hist.history['loss'], label='train')\n",
        "ax2.plot(hist.history['val_loss'],label='test')\n",
        "ax2.set_title('model loss')\n",
        "ax2.set_ylabel('loss')\n",
        "ax2.set_xlabel('epoch')\n",
        "ax2.legend(loc='upper right')\n"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "i-jJGUQMyFfL"
      },
      "source": [
        "### Clasificación de datos nuevos\n",
        "Tenga en cuenta que el aumento y la eliminación de datos están inactivos en el momento de la inferencia."
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "X1R4kGJugMDd"
      },
      "outputs": [],
      "source": [
        "labels_list = sorted(['erythroblast','eosinophil','monocyte','platelet','lymphocyte',\n",
        " 'neutrophil','ig','basophil'])\n",
        "\n",
        "Y_pred = np.array([])\n",
        "Y_labels =  np.array([])\n",
        "for x, y in test_dataset:\n",
        "  Y_pred = np.concatenate([Y_pred, np.argmax(model.predict(x), axis = -1)])\n",
        "  Y_labels = np.concatenate([Y_labels, np.argmax(y.numpy(), axis=-1)])"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "xE398-Ytgrno"
      },
      "source": [
        "### Matriz de confusión"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 755
        },
        "id": "QVCy7tF9ohRL",
        "outputId": "f9a92304-4e32-4888-a812-f3f0e4e37fd4"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 864x720 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        }
      ],
      "source": [
        "cf_matrix = confusion_matrix(Y_labels, Y_pred,  normalize=\"true\")\n",
        "plt.figure(figsize = (12,10))\n",
        "heatmap = sns.heatmap(cf_matrix, \n",
        "            annot=True, \n",
        "            xticklabels = labels_list, \n",
        "            yticklabels = labels_list,\n",
        "            cmap=plt.cm.Greens\n",
        "            )\n",
        "plt.tight_layout()\n",
        "heatmap.set_ylabel('True label', fontsize=10)\n",
        "heatmap.set_xlabel('Predicted Label', fontsize=10)\n",
        "\n",
        "plt.title('Confusion Matrix')\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "loDIZhCbimY5"
      },
      "source": [
        "### Reporte de la clasificación"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "X_DgzWOlikTU",
        "outputId": "494fdca6-9687-4e71-cd02-dc8621489edf"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Classification Report\n",
            "\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "    basophil       0.89      1.00      0.94        16\n",
            "  eosinophil       0.97      1.00      0.98        29\n",
            "erythroblast       1.00      0.54      0.70        13\n",
            "          ig       0.56      0.95      0.70        20\n",
            "  lymphocyte       1.00      0.31      0.47        13\n",
            "    monocyte       0.25      0.23      0.24        13\n",
            "  neutrophil       0.92      0.92      0.92        39\n",
            "    platelet       1.00      0.94      0.97        17\n",
            "\n",
            "    accuracy                           0.81       160\n",
            "   macro avg       0.82      0.74      0.74       160\n",
            "weighted avg       0.85      0.81      0.80       160\n",
            "\n"
          ]
        }
      ],
      "source": [
        "print('Classification Report\\n')\n",
        "print(classification_report(Y_labels, Y_pred, target_names=labels_list))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "dCgnirTDOavh"
      },
      "source": [
        "Prueba de la mejor epoca\n"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "CgTchGrmOdgh",
        "outputId": "3b61b9d0-e4d7-49ac-e6a0-3633b8b8acb9"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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",
            "text/plain": [
              "<Figure size 864x720 with 2 Axes>"
            ]
          },
          "metadata": {
            "needs_background": "light"
          },
          "output_type": "display_data"
        },
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Classification Report\n",
            "\n",
            "              precision    recall  f1-score   support\n",
            "\n",
            "    basophil       1.00      1.00      1.00        17\n",
            "  eosinophil       1.00      1.00      1.00        28\n",
            "erythroblast       0.88      1.00      0.93         7\n",
            "          ig       0.96      1.00      0.98        26\n",
            "  lymphocyte       1.00      1.00      1.00        11\n",
            "    monocyte       1.00      1.00      1.00        14\n",
            "  neutrophil       1.00      0.94      0.97        35\n",
            "    platelet       1.00      1.00      1.00        22\n",
            "\n",
            "    accuracy                           0.99       160\n",
            "   macro avg       0.98      0.99      0.99       160\n",
            "weighted avg       0.99      0.99      0.99       160\n",
            "\n"
          ]
        }
      ],
      "source": [
        "from keras.models import load_model\n",
        "\n",
        "path_model = 'save_at_21.h5'\n",
        "\n",
        "best_epoch = load_model(path_model)\n",
        "\n",
        "labels_list = sorted(['erythroblast','eosinophil','monocyte','platelet','lymphocyte',\n",
        " 'neutrophil','ig','basophil'])\n",
        "\n",
        "Y_pred = np.array([])\n",
        "Y_labels =  np.array([])\n",
        "for x, y in test_dataset:\n",
        "  Y_pred = np.concatenate([Y_pred, np.argmax(best_epoch.predict(x), axis = -1)])\n",
        "  Y_labels = np.concatenate([Y_labels, np.argmax(y.numpy(), axis=-1)])\n",
        "\n",
        "cf_matrix = confusion_matrix(Y_labels, Y_pred,  normalize=\"true\")\n",
        "plt.figure(figsize = (12,10))\n",
        "heatmap = sns.heatmap(cf_matrix, \n",
        "            annot=True, \n",
        "            xticklabels = labels_list, \n",
        "            yticklabels = labels_list,\n",
        "            cmap=plt.cm.Greens\n",
        "            )\n",
        "plt.tight_layout()\n",
        "heatmap.set_ylabel('True label', fontsize=10)\n",
        "heatmap.set_xlabel('Predicted Label', fontsize=10)\n",
        "\n",
        "plt.title('Confusion Matrix')\n",
        "plt.show()\n",
        "\n",
        "print('Classification Report\\n')\n",
        "print(classification_report(Y_labels, Y_pred, target_names=labels_list))"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "collapsed_sections": [],
      "provenance": []
    },
    "gpuClass": "standard",
    "kernelspec": {
      "display_name": "Python 3.11.0 64-bit",
      "language": "python",
      "name": "python3"
    },
    "language_info": {
      "codemirror_mode": {
        "name": "ipython",
        "version": 3
      },
      "file_extension": ".py",
      "mimetype": "text/x-python",
      "name": "python",
      "nbconvert_exporter": "python",
      "pygments_lexer": "ipython3",
      "version": "3.11.0"
    },
    "vscode": {
      "interpreter": {
        "hash": "c261aea317cc0286b3b3261fbba9abdec21eaa57589985bb7a274bf54d6cc0a7"
      }
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}