InceptionV3.ipynb

{
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/eordog/erdos_artemis/blob/main/InceptionV3.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "First, we start by importing all necessary packages"
      ],
      "metadata": {
        "id": "Pay8IaO5t3QN"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "J08neNEas2wt"
      },
      "outputs": [],
      "source": [
        "from tensorflow.keras.layers import Input, Lambda, Dense, Flatten,Dropout\n",
        "from tensorflow.keras.models import Model\n",
        "from keras import layers\n",
        "from tensorflow.keras.preprocessing import image\n",
        "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
        "from tensorflow.keras.models import Sequential\n",
        "from sklearn.model_selection import train_test_split\n",
        "import tensorflow as tf\n",
        "import numpy as np\n",
        "import pandas as pd\n",
        "import os\n",
        "import cv2\n",
        "import matplotlib.pyplot as plt\n",
        "from random import shuffle\n",
        "from tqdm import tqdm\n",
        "from sklearn.metrics import classification_report,confusion_matrix\n",
        "import seaborn as sns"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We determine the labels that will be used for classifying different types of tumors. \"Back\" refers to an MRI scan taken from the back of a patient, \"top\" is taken from above and \"side\" is from one side of the patient."
      ],
      "metadata": {
        "id": "88gYStD4uCMN"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "CJMdff9QcUO3"
      },
      "outputs": [],
      "source": [
        "labels = ['glioma_back', 'glioma_side', 'glioma_top', 'meningioma_back', 'meningioma_side', 'meningioma_top', \n",
        "           'no_back', 'no_side', 'no_top', 'pituitary_back', 'pituitary_side', 'pituitary_top']"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We go through both training and testing set and merge them first, then convert the images into arrays. "
      ],
      "metadata": {
        "id": "0lQcRBAMuGUU"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "DYGiPtC6cZzj",
        "outputId": "bd3a71c5-a511-49bf-9dde-4613288552bf"
      },
      "outputs": [
        {
          "name": "stderr",
          "output_type": "stream",
          "text": [
            "100%|██████████| 276/276 [00:08<00:00, 34.49it/s]\n",
            "100%|██████████| 273/273 [00:05<00:00, 47.55it/s] \n",
            "100%|██████████| 267/267 [00:05<00:00, 46.83it/s] \n",
            "100%|██████████| 255/255 [00:05<00:00, 46.43it/s] \n",
            "100%|██████████| 271/271 [00:05<00:00, 49.89it/s] \n",
            "100%|██████████| 291/291 [00:05<00:00, 50.92it/s] \n",
            "100%|██████████| 36/36 [00:25<00:00,  1.40it/s]\n",
            "100%|██████████| 90/90 [01:03<00:00,  1.43it/s]\n",
            "100%|██████████| 267/267 [00:04<00:00, 54.55it/s] \n",
            "100%|██████████| 280/280 [00:05<00:00, 50.40it/s] \n",
            "100%|██████████| 288/288 [00:05<00:00, 50.18it/s] \n",
            "100%|██████████| 259/259 [00:05<00:00, 48.85it/s] \n",
            "100%|██████████| 22/22 [00:15<00:00,  1.43it/s]\n",
            "100%|██████████| 12/12 [00:08<00:00,  1.40it/s]\n",
            "100%|██████████| 66/66 [00:45<00:00,  1.45it/s]\n",
            "100%|██████████| 29/29 [00:20<00:00,  1.42it/s]\n",
            "100%|██████████| 16/16 [00:11<00:00,  1.40it/s]\n",
            "100%|██████████| 69/69 [00:51<00:00,  1.35it/s]\n",
            "100%|██████████| 4/4 [00:02<00:00,  1.78it/s]\n",
            "100%|██████████| 5/5 [00:03<00:00,  1.40it/s]\n",
            "100%|██████████| 96/96 [00:56<00:00,  1.69it/s]\n",
            "100%|██████████| 42/42 [00:30<00:00,  1.40it/s]\n",
            "100%|██████████| 22/22 [00:13<00:00,  1.59it/s]\n",
            "100%|██████████| 10/10 [00:06<00:00,  1.53it/s]\n"
          ]
        }
      ],
      "source": [
        "X_train = []\n",
        "y_train = []\n",
        "image_size = 150\n",
        "for i in labels:\n",
        "    folderPath = os.path.join('/content/drive/MyDrive/Colab_Notebooks/Brain_Tumor_Data_12_Categories','Training',i)\n",
        "    for j in tqdm(os.listdir(folderPath)):\n",
        "        img = cv2.imread(os.path.join(folderPath,j))\n",
        "        img = cv2.resize(img,(image_size, image_size))\n",
        "        X_train.append(img)\n",
        "        y_train.append(i)\n",
        "        \n",
        "for i in labels:\n",
        "    folderPath = os.path.join('/content/drive/MyDrive/Colab_Notebooks/Brain_Tumor_Data_12_Categories','Testing',i)\n",
        "    for j in tqdm(os.listdir(folderPath)):\n",
        "        img = cv2.imread(os.path.join(folderPath,j))\n",
        "        img = cv2.resize(img,(image_size,image_size))\n",
        "        X_train.append(img)\n",
        "        y_train.append(i)\n",
        "        \n",
        "X_train = np.array(X_train)\n",
        "y_train = np.array(y_train)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We can see some sample images from each label."
      ],
      "metadata": {
        "id": "EtAzQcWxuLSU"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 253
        },
        "id": "b-NlPiOJcpsp",
        "outputId": "6fcae928-478d-456c-bf4c-4009db35555e"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1440x1440 with 12 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "k=0\n",
        "fig, ax = plt.subplots(1,12,figsize=(20,20))\n",
        "fig.text(s='Sample Image From Each Label',size=18,fontweight='bold',\n",
        "             fontname='monospace',y=0.62,x=0.4,alpha=0.8)\n",
        "for i in labels:\n",
        "    j=0\n",
        "    while True :\n",
        "        if y_train[j]==i:\n",
        "            ax[k].imshow(X_train[j])\n",
        "            ax[k].set_title(y_train[j])\n",
        "            ax[k].axis('off')\n",
        "            k+=1\n",
        "            break\n",
        "        j+=1"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We perform a train-test split, reserving 10% of the training set to test our model."
      ],
      "metadata": {
        "id": "C97hbfAHuRVu"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "esT6lRoReY1a"
      },
      "outputs": [],
      "source": [
        "X_train,X_test,y_train,y_test = train_test_split(X_train,y_train, test_size=0.1,random_state=607)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Apply one hot encoding for all the labels."
      ],
      "metadata": {
        "id": "HqDDswrauZER"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "v5ujo6pAeiBj"
      },
      "outputs": [],
      "source": [
        "y_train_new = []\n",
        "for i in y_train:\n",
        "    y_train_new.append(labels.index(i))\n",
        "y_train = y_train_new\n",
        "y_train = tf.keras.utils.to_categorical(y_train)\n",
        "\n",
        "\n",
        "y_test_new = []\n",
        "for i in y_test:\n",
        "    y_test_new.append(labels.index(i))\n",
        "y_test = y_test_new\n",
        "y_test = tf.keras.utils.to_categorical(y_test)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Here we use a pre-trained CNN model called InceptionV3. The weights are transferred from ImageNet. To make some modifications in output layers, we have include_top=False. Some further details about the model can be found in the Keras documentation; https://keras.io/api/applications/inceptionv3/"
      ],
      "metadata": {
        "id": "nqrs55RRupaS"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "XUIIFg1EAW45",
        "outputId": "c357bb4f-dd2c-41e2-f368-6577fb33495a"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Downloading data from https://storage.googleapis.com/tensorflow/keras-applications/inception_v3/inception_v3_weights_tf_dim_ordering_tf_kernels_notop.h5\n",
            "87916544/87910968 [==============================] - 1s 0us/step\n",
            "87924736/87910968 [==============================] - 1s 0us/step\n"
          ]
        }
      ],
      "source": [
        "incv3 = tf.keras.applications.InceptionV3(\n",
        "    include_top=False,\n",
        "    weights=\"imagenet\",\n",
        "    input_tensor=None,\n",
        "    input_shape=None,\n",
        "    pooling=None,\n",
        "    classifier_activation=\"softmax\",\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We add an extra layer of GlobalAveragePooling2D which works very similar as MaxPooling2D layer and add another DropOut layer which randomly ignores certain nodes. This helps improve the model's accuracy."
      ],
      "metadata": {
        "id": "QSRlX_Y9vLHt"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "C9-WCZSCBzN_"
      },
      "outputs": [],
      "source": [
        "model = incv3.output\n",
        "model = tf.keras.layers.GlobalAveragePooling2D()(model)\n",
        "model = tf.keras.layers.Dropout(rate=0.5)(model)\n",
        "model = tf.keras.layers.Dense(12,activation='softmax')(model)\n",
        "model = tf.keras.models.Model(inputs=incv3.input, outputs = model)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We can see all the layers in the model by model.summary()"
      ],
      "metadata": {
        "id": "sfogy_2ivfcI"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "nIfukVHiB3L_",
        "outputId": "690f5e92-eba9-4477-d25a-843a1519a673"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Model: \"model\"\n",
            "__________________________________________________________________________________________________\n",
            " Layer (type)                   Output Shape         Param #     Connected to                     \n",
            "==================================================================================================\n",
            " input_1 (InputLayer)           [(None, None, None,  0           []                               \n",
            "                                 3)]                                                              \n",
            "                                                                                                  \n",
            " conv2d (Conv2D)                (None, None, None,   864         ['input_1[0][0]']                \n",
            "                                32)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization (BatchNorm  (None, None, None,   96         ['conv2d[0][0]']                 \n",
            " alization)                     32)                                                               \n",
            "                                                                                                  \n",
            " activation (Activation)        (None, None, None,   0           ['batch_normalization[0][0]']    \n",
            "                                32)                                                               \n",
            "                                                                                                  \n",
            " conv2d_1 (Conv2D)              (None, None, None,   9216        ['activation[0][0]']             \n",
            "                                32)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_1 (BatchNo  (None, None, None,   96         ['conv2d_1[0][0]']               \n",
            " rmalization)                   32)                                                               \n",
            "                                                                                                  \n",
            " activation_1 (Activation)      (None, None, None,   0           ['batch_normalization_1[0][0]']  \n",
            "                                32)                                                               \n",
            "                                                                                                  \n",
            " conv2d_2 (Conv2D)              (None, None, None,   18432       ['activation_1[0][0]']           \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_2 (BatchNo  (None, None, None,   192        ['conv2d_2[0][0]']               \n",
            " rmalization)                   64)                                                               \n",
            "                                                                                                  \n",
            " activation_2 (Activation)      (None, None, None,   0           ['batch_normalization_2[0][0]']  \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " max_pooling2d (MaxPooling2D)   (None, None, None,   0           ['activation_2[0][0]']           \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " conv2d_3 (Conv2D)              (None, None, None,   5120        ['max_pooling2d[0][0]']          \n",
            "                                80)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_3 (BatchNo  (None, None, None,   240        ['conv2d_3[0][0]']               \n",
            " rmalization)                   80)                                                               \n",
            "                                                                                                  \n",
            " activation_3 (Activation)      (None, None, None,   0           ['batch_normalization_3[0][0]']  \n",
            "                                80)                                                               \n",
            "                                                                                                  \n",
            " conv2d_4 (Conv2D)              (None, None, None,   138240      ['activation_3[0][0]']           \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_4 (BatchNo  (None, None, None,   576        ['conv2d_4[0][0]']               \n",
            " rmalization)                   192)                                                              \n",
            "                                                                                                  \n",
            " activation_4 (Activation)      (None, None, None,   0           ['batch_normalization_4[0][0]']  \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " max_pooling2d_1 (MaxPooling2D)  (None, None, None,   0          ['activation_4[0][0]']           \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_8 (Conv2D)              (None, None, None,   12288       ['max_pooling2d_1[0][0]']        \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_8 (BatchNo  (None, None, None,   192        ['conv2d_8[0][0]']               \n",
            " rmalization)                   64)                                                               \n",
            "                                                                                                  \n",
            " activation_8 (Activation)      (None, None, None,   0           ['batch_normalization_8[0][0]']  \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " conv2d_6 (Conv2D)              (None, None, None,   9216        ['max_pooling2d_1[0][0]']        \n",
            "                                48)                                                               \n",
            "                                                                                                  \n",
            " conv2d_9 (Conv2D)              (None, None, None,   55296       ['activation_8[0][0]']           \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_6 (BatchNo  (None, None, None,   144        ['conv2d_6[0][0]']               \n",
            " rmalization)                   48)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_9 (BatchNo  (None, None, None,   288        ['conv2d_9[0][0]']               \n",
            " rmalization)                   96)                                                               \n",
            "                                                                                                  \n",
            " activation_6 (Activation)      (None, None, None,   0           ['batch_normalization_6[0][0]']  \n",
            "                                48)                                                               \n",
            "                                                                                                  \n",
            " activation_9 (Activation)      (None, None, None,   0           ['batch_normalization_9[0][0]']  \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " average_pooling2d (AveragePool  (None, None, None,   0          ['max_pooling2d_1[0][0]']        \n",
            " ing2D)                         192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_5 (Conv2D)              (None, None, None,   12288       ['max_pooling2d_1[0][0]']        \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " conv2d_7 (Conv2D)              (None, None, None,   76800       ['activation_6[0][0]']           \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " conv2d_10 (Conv2D)             (None, None, None,   82944       ['activation_9[0][0]']           \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " conv2d_11 (Conv2D)             (None, None, None,   6144        ['average_pooling2d[0][0]']      \n",
            "                                32)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_5 (BatchNo  (None, None, None,   192        ['conv2d_5[0][0]']               \n",
            " rmalization)                   64)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_7 (BatchNo  (None, None, None,   192        ['conv2d_7[0][0]']               \n",
            " rmalization)                   64)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_10 (BatchN  (None, None, None,   288        ['conv2d_10[0][0]']              \n",
            " ormalization)                  96)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_11 (BatchN  (None, None, None,   96         ['conv2d_11[0][0]']              \n",
            " ormalization)                  32)                                                               \n",
            "                                                                                                  \n",
            " activation_5 (Activation)      (None, None, None,   0           ['batch_normalization_5[0][0]']  \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " activation_7 (Activation)      (None, None, None,   0           ['batch_normalization_7[0][0]']  \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " activation_10 (Activation)     (None, None, None,   0           ['batch_normalization_10[0][0]'] \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " activation_11 (Activation)     (None, None, None,   0           ['batch_normalization_11[0][0]'] \n",
            "                                32)                                                               \n",
            "                                                                                                  \n",
            " mixed0 (Concatenate)           (None, None, None,   0           ['activation_5[0][0]',           \n",
            "                                256)                              'activation_7[0][0]',           \n",
            "                                                                  'activation_10[0][0]',          \n",
            "                                                                  'activation_11[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_15 (Conv2D)             (None, None, None,   16384       ['mixed0[0][0]']                 \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_15 (BatchN  (None, None, None,   192        ['conv2d_15[0][0]']              \n",
            " ormalization)                  64)                                                               \n",
            "                                                                                                  \n",
            " activation_15 (Activation)     (None, None, None,   0           ['batch_normalization_15[0][0]'] \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " conv2d_13 (Conv2D)             (None, None, None,   12288       ['mixed0[0][0]']                 \n",
            "                                48)                                                               \n",
            "                                                                                                  \n",
            " conv2d_16 (Conv2D)             (None, None, None,   55296       ['activation_15[0][0]']          \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_13 (BatchN  (None, None, None,   144        ['conv2d_13[0][0]']              \n",
            " ormalization)                  48)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_16 (BatchN  (None, None, None,   288        ['conv2d_16[0][0]']              \n",
            " ormalization)                  96)                                                               \n",
            "                                                                                                  \n",
            " activation_13 (Activation)     (None, None, None,   0           ['batch_normalization_13[0][0]'] \n",
            "                                48)                                                               \n",
            "                                                                                                  \n",
            " activation_16 (Activation)     (None, None, None,   0           ['batch_normalization_16[0][0]'] \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " average_pooling2d_1 (AveragePo  (None, None, None,   0          ['mixed0[0][0]']                 \n",
            " oling2D)                       256)                                                              \n",
            "                                                                                                  \n",
            " conv2d_12 (Conv2D)             (None, None, None,   16384       ['mixed0[0][0]']                 \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " conv2d_14 (Conv2D)             (None, None, None,   76800       ['activation_13[0][0]']          \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " conv2d_17 (Conv2D)             (None, None, None,   82944       ['activation_16[0][0]']          \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " conv2d_18 (Conv2D)             (None, None, None,   16384       ['average_pooling2d_1[0][0]']    \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_12 (BatchN  (None, None, None,   192        ['conv2d_12[0][0]']              \n",
            " ormalization)                  64)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_14 (BatchN  (None, None, None,   192        ['conv2d_14[0][0]']              \n",
            " ormalization)                  64)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_17 (BatchN  (None, None, None,   288        ['conv2d_17[0][0]']              \n",
            " ormalization)                  96)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_18 (BatchN  (None, None, None,   192        ['conv2d_18[0][0]']              \n",
            " ormalization)                  64)                                                               \n",
            "                                                                                                  \n",
            " activation_12 (Activation)     (None, None, None,   0           ['batch_normalization_12[0][0]'] \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " activation_14 (Activation)     (None, None, None,   0           ['batch_normalization_14[0][0]'] \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " activation_17 (Activation)     (None, None, None,   0           ['batch_normalization_17[0][0]'] \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " activation_18 (Activation)     (None, None, None,   0           ['batch_normalization_18[0][0]'] \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " mixed1 (Concatenate)           (None, None, None,   0           ['activation_12[0][0]',          \n",
            "                                288)                              'activation_14[0][0]',          \n",
            "                                                                  'activation_17[0][0]',          \n",
            "                                                                  'activation_18[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_22 (Conv2D)             (None, None, None,   18432       ['mixed1[0][0]']                 \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_22 (BatchN  (None, None, None,   192        ['conv2d_22[0][0]']              \n",
            " ormalization)                  64)                                                               \n",
            "                                                                                                  \n",
            " activation_22 (Activation)     (None, None, None,   0           ['batch_normalization_22[0][0]'] \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " conv2d_20 (Conv2D)             (None, None, None,   13824       ['mixed1[0][0]']                 \n",
            "                                48)                                                               \n",
            "                                                                                                  \n",
            " conv2d_23 (Conv2D)             (None, None, None,   55296       ['activation_22[0][0]']          \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_20 (BatchN  (None, None, None,   144        ['conv2d_20[0][0]']              \n",
            " ormalization)                  48)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_23 (BatchN  (None, None, None,   288        ['conv2d_23[0][0]']              \n",
            " ormalization)                  96)                                                               \n",
            "                                                                                                  \n",
            " activation_20 (Activation)     (None, None, None,   0           ['batch_normalization_20[0][0]'] \n",
            "                                48)                                                               \n",
            "                                                                                                  \n",
            " activation_23 (Activation)     (None, None, None,   0           ['batch_normalization_23[0][0]'] \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " average_pooling2d_2 (AveragePo  (None, None, None,   0          ['mixed1[0][0]']                 \n",
            " oling2D)                       288)                                                              \n",
            "                                                                                                  \n",
            " conv2d_19 (Conv2D)             (None, None, None,   18432       ['mixed1[0][0]']                 \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " conv2d_21 (Conv2D)             (None, None, None,   76800       ['activation_20[0][0]']          \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " conv2d_24 (Conv2D)             (None, None, None,   82944       ['activation_23[0][0]']          \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " conv2d_25 (Conv2D)             (None, None, None,   18432       ['average_pooling2d_2[0][0]']    \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_19 (BatchN  (None, None, None,   192        ['conv2d_19[0][0]']              \n",
            " ormalization)                  64)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_21 (BatchN  (None, None, None,   192        ['conv2d_21[0][0]']              \n",
            " ormalization)                  64)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_24 (BatchN  (None, None, None,   288        ['conv2d_24[0][0]']              \n",
            " ormalization)                  96)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_25 (BatchN  (None, None, None,   192        ['conv2d_25[0][0]']              \n",
            " ormalization)                  64)                                                               \n",
            "                                                                                                  \n",
            " activation_19 (Activation)     (None, None, None,   0           ['batch_normalization_19[0][0]'] \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " activation_21 (Activation)     (None, None, None,   0           ['batch_normalization_21[0][0]'] \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " activation_24 (Activation)     (None, None, None,   0           ['batch_normalization_24[0][0]'] \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " activation_25 (Activation)     (None, None, None,   0           ['batch_normalization_25[0][0]'] \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " mixed2 (Concatenate)           (None, None, None,   0           ['activation_19[0][0]',          \n",
            "                                288)                              'activation_21[0][0]',          \n",
            "                                                                  'activation_24[0][0]',          \n",
            "                                                                  'activation_25[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_27 (Conv2D)             (None, None, None,   18432       ['mixed2[0][0]']                 \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_27 (BatchN  (None, None, None,   192        ['conv2d_27[0][0]']              \n",
            " ormalization)                  64)                                                               \n",
            "                                                                                                  \n",
            " activation_27 (Activation)     (None, None, None,   0           ['batch_normalization_27[0][0]'] \n",
            "                                64)                                                               \n",
            "                                                                                                  \n",
            " conv2d_28 (Conv2D)             (None, None, None,   55296       ['activation_27[0][0]']          \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_28 (BatchN  (None, None, None,   288        ['conv2d_28[0][0]']              \n",
            " ormalization)                  96)                                                               \n",
            "                                                                                                  \n",
            " activation_28 (Activation)     (None, None, None,   0           ['batch_normalization_28[0][0]'] \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " conv2d_26 (Conv2D)             (None, None, None,   995328      ['mixed2[0][0]']                 \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " conv2d_29 (Conv2D)             (None, None, None,   82944       ['activation_28[0][0]']          \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " batch_normalization_26 (BatchN  (None, None, None,   1152       ['conv2d_26[0][0]']              \n",
            " ormalization)                  384)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_29 (BatchN  (None, None, None,   288        ['conv2d_29[0][0]']              \n",
            " ormalization)                  96)                                                               \n",
            "                                                                                                  \n",
            " activation_26 (Activation)     (None, None, None,   0           ['batch_normalization_26[0][0]'] \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " activation_29 (Activation)     (None, None, None,   0           ['batch_normalization_29[0][0]'] \n",
            "                                96)                                                               \n",
            "                                                                                                  \n",
            " max_pooling2d_2 (MaxPooling2D)  (None, None, None,   0          ['mixed2[0][0]']                 \n",
            "                                288)                                                              \n",
            "                                                                                                  \n",
            " mixed3 (Concatenate)           (None, None, None,   0           ['activation_26[0][0]',          \n",
            "                                768)                              'activation_29[0][0]',          \n",
            "                                                                  'max_pooling2d_2[0][0]']        \n",
            "                                                                                                  \n",
            " conv2d_34 (Conv2D)             (None, None, None,   98304       ['mixed3[0][0]']                 \n",
            "                                128)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_34 (BatchN  (None, None, None,   384        ['conv2d_34[0][0]']              \n",
            " ormalization)                  128)                                                              \n",
            "                                                                                                  \n",
            " activation_34 (Activation)     (None, None, None,   0           ['batch_normalization_34[0][0]'] \n",
            "                                128)                                                              \n",
            "                                                                                                  \n",
            " conv2d_35 (Conv2D)             (None, None, None,   114688      ['activation_34[0][0]']          \n",
            "                                128)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_35 (BatchN  (None, None, None,   384        ['conv2d_35[0][0]']              \n",
            " ormalization)                  128)                                                              \n",
            "                                                                                                  \n",
            " activation_35 (Activation)     (None, None, None,   0           ['batch_normalization_35[0][0]'] \n",
            "                                128)                                                              \n",
            "                                                                                                  \n",
            " conv2d_31 (Conv2D)             (None, None, None,   98304       ['mixed3[0][0]']                 \n",
            "                                128)                                                              \n",
            "                                                                                                  \n",
            " conv2d_36 (Conv2D)             (None, None, None,   114688      ['activation_35[0][0]']          \n",
            "                                128)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_31 (BatchN  (None, None, None,   384        ['conv2d_31[0][0]']              \n",
            " ormalization)                  128)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_36 (BatchN  (None, None, None,   384        ['conv2d_36[0][0]']              \n",
            " ormalization)                  128)                                                              \n",
            "                                                                                                  \n",
            " activation_31 (Activation)     (None, None, None,   0           ['batch_normalization_31[0][0]'] \n",
            "                                128)                                                              \n",
            "                                                                                                  \n",
            " activation_36 (Activation)     (None, None, None,   0           ['batch_normalization_36[0][0]'] \n",
            "                                128)                                                              \n",
            "                                                                                                  \n",
            " conv2d_32 (Conv2D)             (None, None, None,   114688      ['activation_31[0][0]']          \n",
            "                                128)                                                              \n",
            "                                                                                                  \n",
            " conv2d_37 (Conv2D)             (None, None, None,   114688      ['activation_36[0][0]']          \n",
            "                                128)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_32 (BatchN  (None, None, None,   384        ['conv2d_32[0][0]']              \n",
            " ormalization)                  128)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_37 (BatchN  (None, None, None,   384        ['conv2d_37[0][0]']              \n",
            " ormalization)                  128)                                                              \n",
            "                                                                                                  \n",
            " activation_32 (Activation)     (None, None, None,   0           ['batch_normalization_32[0][0]'] \n",
            "                                128)                                                              \n",
            "                                                                                                  \n",
            " activation_37 (Activation)     (None, None, None,   0           ['batch_normalization_37[0][0]'] \n",
            "                                128)                                                              \n",
            "                                                                                                  \n",
            " average_pooling2d_3 (AveragePo  (None, None, None,   0          ['mixed3[0][0]']                 \n",
            " oling2D)                       768)                                                              \n",
            "                                                                                                  \n",
            " conv2d_30 (Conv2D)             (None, None, None,   147456      ['mixed3[0][0]']                 \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_33 (Conv2D)             (None, None, None,   172032      ['activation_32[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_38 (Conv2D)             (None, None, None,   172032      ['activation_37[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_39 (Conv2D)             (None, None, None,   147456      ['average_pooling2d_3[0][0]']    \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_30 (BatchN  (None, None, None,   576        ['conv2d_30[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_33 (BatchN  (None, None, None,   576        ['conv2d_33[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_38 (BatchN  (None, None, None,   576        ['conv2d_38[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_39 (BatchN  (None, None, None,   576        ['conv2d_39[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " activation_30 (Activation)     (None, None, None,   0           ['batch_normalization_30[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_33 (Activation)     (None, None, None,   0           ['batch_normalization_33[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_38 (Activation)     (None, None, None,   0           ['batch_normalization_38[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_39 (Activation)     (None, None, None,   0           ['batch_normalization_39[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " mixed4 (Concatenate)           (None, None, None,   0           ['activation_30[0][0]',          \n",
            "                                768)                              'activation_33[0][0]',          \n",
            "                                                                  'activation_38[0][0]',          \n",
            "                                                                  'activation_39[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_44 (Conv2D)             (None, None, None,   122880      ['mixed4[0][0]']                 \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_44 (BatchN  (None, None, None,   480        ['conv2d_44[0][0]']              \n",
            " ormalization)                  160)                                                              \n",
            "                                                                                                  \n",
            " activation_44 (Activation)     (None, None, None,   0           ['batch_normalization_44[0][0]'] \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " conv2d_45 (Conv2D)             (None, None, None,   179200      ['activation_44[0][0]']          \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_45 (BatchN  (None, None, None,   480        ['conv2d_45[0][0]']              \n",
            " ormalization)                  160)                                                              \n",
            "                                                                                                  \n",
            " activation_45 (Activation)     (None, None, None,   0           ['batch_normalization_45[0][0]'] \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " conv2d_41 (Conv2D)             (None, None, None,   122880      ['mixed4[0][0]']                 \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " conv2d_46 (Conv2D)             (None, None, None,   179200      ['activation_45[0][0]']          \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_41 (BatchN  (None, None, None,   480        ['conv2d_41[0][0]']              \n",
            " ormalization)                  160)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_46 (BatchN  (None, None, None,   480        ['conv2d_46[0][0]']              \n",
            " ormalization)                  160)                                                              \n",
            "                                                                                                  \n",
            " activation_41 (Activation)     (None, None, None,   0           ['batch_normalization_41[0][0]'] \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " activation_46 (Activation)     (None, None, None,   0           ['batch_normalization_46[0][0]'] \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " conv2d_42 (Conv2D)             (None, None, None,   179200      ['activation_41[0][0]']          \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " conv2d_47 (Conv2D)             (None, None, None,   179200      ['activation_46[0][0]']          \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_42 (BatchN  (None, None, None,   480        ['conv2d_42[0][0]']              \n",
            " ormalization)                  160)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_47 (BatchN  (None, None, None,   480        ['conv2d_47[0][0]']              \n",
            " ormalization)                  160)                                                              \n",
            "                                                                                                  \n",
            " activation_42 (Activation)     (None, None, None,   0           ['batch_normalization_42[0][0]'] \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " activation_47 (Activation)     (None, None, None,   0           ['batch_normalization_47[0][0]'] \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " average_pooling2d_4 (AveragePo  (None, None, None,   0          ['mixed4[0][0]']                 \n",
            " oling2D)                       768)                                                              \n",
            "                                                                                                  \n",
            " conv2d_40 (Conv2D)             (None, None, None,   147456      ['mixed4[0][0]']                 \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_43 (Conv2D)             (None, None, None,   215040      ['activation_42[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_48 (Conv2D)             (None, None, None,   215040      ['activation_47[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_49 (Conv2D)             (None, None, None,   147456      ['average_pooling2d_4[0][0]']    \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_40 (BatchN  (None, None, None,   576        ['conv2d_40[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_43 (BatchN  (None, None, None,   576        ['conv2d_43[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_48 (BatchN  (None, None, None,   576        ['conv2d_48[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_49 (BatchN  (None, None, None,   576        ['conv2d_49[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " activation_40 (Activation)     (None, None, None,   0           ['batch_normalization_40[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_43 (Activation)     (None, None, None,   0           ['batch_normalization_43[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_48 (Activation)     (None, None, None,   0           ['batch_normalization_48[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_49 (Activation)     (None, None, None,   0           ['batch_normalization_49[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " mixed5 (Concatenate)           (None, None, None,   0           ['activation_40[0][0]',          \n",
            "                                768)                              'activation_43[0][0]',          \n",
            "                                                                  'activation_48[0][0]',          \n",
            "                                                                  'activation_49[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_54 (Conv2D)             (None, None, None,   122880      ['mixed5[0][0]']                 \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_54 (BatchN  (None, None, None,   480        ['conv2d_54[0][0]']              \n",
            " ormalization)                  160)                                                              \n",
            "                                                                                                  \n",
            " activation_54 (Activation)     (None, None, None,   0           ['batch_normalization_54[0][0]'] \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " conv2d_55 (Conv2D)             (None, None, None,   179200      ['activation_54[0][0]']          \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_55 (BatchN  (None, None, None,   480        ['conv2d_55[0][0]']              \n",
            " ormalization)                  160)                                                              \n",
            "                                                                                                  \n",
            " activation_55 (Activation)     (None, None, None,   0           ['batch_normalization_55[0][0]'] \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " conv2d_51 (Conv2D)             (None, None, None,   122880      ['mixed5[0][0]']                 \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " conv2d_56 (Conv2D)             (None, None, None,   179200      ['activation_55[0][0]']          \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_51 (BatchN  (None, None, None,   480        ['conv2d_51[0][0]']              \n",
            " ormalization)                  160)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_56 (BatchN  (None, None, None,   480        ['conv2d_56[0][0]']              \n",
            " ormalization)                  160)                                                              \n",
            "                                                                                                  \n",
            " activation_51 (Activation)     (None, None, None,   0           ['batch_normalization_51[0][0]'] \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " activation_56 (Activation)     (None, None, None,   0           ['batch_normalization_56[0][0]'] \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " conv2d_52 (Conv2D)             (None, None, None,   179200      ['activation_51[0][0]']          \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " conv2d_57 (Conv2D)             (None, None, None,   179200      ['activation_56[0][0]']          \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_52 (BatchN  (None, None, None,   480        ['conv2d_52[0][0]']              \n",
            " ormalization)                  160)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_57 (BatchN  (None, None, None,   480        ['conv2d_57[0][0]']              \n",
            " ormalization)                  160)                                                              \n",
            "                                                                                                  \n",
            " activation_52 (Activation)     (None, None, None,   0           ['batch_normalization_52[0][0]'] \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " activation_57 (Activation)     (None, None, None,   0           ['batch_normalization_57[0][0]'] \n",
            "                                160)                                                              \n",
            "                                                                                                  \n",
            " average_pooling2d_5 (AveragePo  (None, None, None,   0          ['mixed5[0][0]']                 \n",
            " oling2D)                       768)                                                              \n",
            "                                                                                                  \n",
            " conv2d_50 (Conv2D)             (None, None, None,   147456      ['mixed5[0][0]']                 \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_53 (Conv2D)             (None, None, None,   215040      ['activation_52[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_58 (Conv2D)             (None, None, None,   215040      ['activation_57[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_59 (Conv2D)             (None, None, None,   147456      ['average_pooling2d_5[0][0]']    \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_50 (BatchN  (None, None, None,   576        ['conv2d_50[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_53 (BatchN  (None, None, None,   576        ['conv2d_53[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_58 (BatchN  (None, None, None,   576        ['conv2d_58[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_59 (BatchN  (None, None, None,   576        ['conv2d_59[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " activation_50 (Activation)     (None, None, None,   0           ['batch_normalization_50[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_53 (Activation)     (None, None, None,   0           ['batch_normalization_53[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_58 (Activation)     (None, None, None,   0           ['batch_normalization_58[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_59 (Activation)     (None, None, None,   0           ['batch_normalization_59[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " mixed6 (Concatenate)           (None, None, None,   0           ['activation_50[0][0]',          \n",
            "                                768)                              'activation_53[0][0]',          \n",
            "                                                                  'activation_58[0][0]',          \n",
            "                                                                  'activation_59[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_64 (Conv2D)             (None, None, None,   147456      ['mixed6[0][0]']                 \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_64 (BatchN  (None, None, None,   576        ['conv2d_64[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " activation_64 (Activation)     (None, None, None,   0           ['batch_normalization_64[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_65 (Conv2D)             (None, None, None,   258048      ['activation_64[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_65 (BatchN  (None, None, None,   576        ['conv2d_65[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " activation_65 (Activation)     (None, None, None,   0           ['batch_normalization_65[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_61 (Conv2D)             (None, None, None,   147456      ['mixed6[0][0]']                 \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_66 (Conv2D)             (None, None, None,   258048      ['activation_65[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_61 (BatchN  (None, None, None,   576        ['conv2d_61[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_66 (BatchN  (None, None, None,   576        ['conv2d_66[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " activation_61 (Activation)     (None, None, None,   0           ['batch_normalization_61[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_66 (Activation)     (None, None, None,   0           ['batch_normalization_66[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_62 (Conv2D)             (None, None, None,   258048      ['activation_61[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_67 (Conv2D)             (None, None, None,   258048      ['activation_66[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_62 (BatchN  (None, None, None,   576        ['conv2d_62[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_67 (BatchN  (None, None, None,   576        ['conv2d_67[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " activation_62 (Activation)     (None, None, None,   0           ['batch_normalization_62[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_67 (Activation)     (None, None, None,   0           ['batch_normalization_67[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " average_pooling2d_6 (AveragePo  (None, None, None,   0          ['mixed6[0][0]']                 \n",
            " oling2D)                       768)                                                              \n",
            "                                                                                                  \n",
            " conv2d_60 (Conv2D)             (None, None, None,   147456      ['mixed6[0][0]']                 \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_63 (Conv2D)             (None, None, None,   258048      ['activation_62[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_68 (Conv2D)             (None, None, None,   258048      ['activation_67[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_69 (Conv2D)             (None, None, None,   147456      ['average_pooling2d_6[0][0]']    \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_60 (BatchN  (None, None, None,   576        ['conv2d_60[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_63 (BatchN  (None, None, None,   576        ['conv2d_63[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_68 (BatchN  (None, None, None,   576        ['conv2d_68[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_69 (BatchN  (None, None, None,   576        ['conv2d_69[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " activation_60 (Activation)     (None, None, None,   0           ['batch_normalization_60[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_63 (Activation)     (None, None, None,   0           ['batch_normalization_63[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_68 (Activation)     (None, None, None,   0           ['batch_normalization_68[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_69 (Activation)     (None, None, None,   0           ['batch_normalization_69[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " mixed7 (Concatenate)           (None, None, None,   0           ['activation_60[0][0]',          \n",
            "                                768)                              'activation_63[0][0]',          \n",
            "                                                                  'activation_68[0][0]',          \n",
            "                                                                  'activation_69[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_72 (Conv2D)             (None, None, None,   147456      ['mixed7[0][0]']                 \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_72 (BatchN  (None, None, None,   576        ['conv2d_72[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " activation_72 (Activation)     (None, None, None,   0           ['batch_normalization_72[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_73 (Conv2D)             (None, None, None,   258048      ['activation_72[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_73 (BatchN  (None, None, None,   576        ['conv2d_73[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " activation_73 (Activation)     (None, None, None,   0           ['batch_normalization_73[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_70 (Conv2D)             (None, None, None,   147456      ['mixed7[0][0]']                 \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_74 (Conv2D)             (None, None, None,   258048      ['activation_73[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_70 (BatchN  (None, None, None,   576        ['conv2d_70[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_74 (BatchN  (None, None, None,   576        ['conv2d_74[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " activation_70 (Activation)     (None, None, None,   0           ['batch_normalization_70[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " activation_74 (Activation)     (None, None, None,   0           ['batch_normalization_74[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " conv2d_71 (Conv2D)             (None, None, None,   552960      ['activation_70[0][0]']          \n",
            "                                320)                                                              \n",
            "                                                                                                  \n",
            " conv2d_75 (Conv2D)             (None, None, None,   331776      ['activation_74[0][0]']          \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_71 (BatchN  (None, None, None,   960        ['conv2d_71[0][0]']              \n",
            " ormalization)                  320)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_75 (BatchN  (None, None, None,   576        ['conv2d_75[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " activation_71 (Activation)     (None, None, None,   0           ['batch_normalization_71[0][0]'] \n",
            "                                320)                                                              \n",
            "                                                                                                  \n",
            " activation_75 (Activation)     (None, None, None,   0           ['batch_normalization_75[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " max_pooling2d_3 (MaxPooling2D)  (None, None, None,   0          ['mixed7[0][0]']                 \n",
            "                                768)                                                              \n",
            "                                                                                                  \n",
            " mixed8 (Concatenate)           (None, None, None,   0           ['activation_71[0][0]',          \n",
            "                                1280)                             'activation_75[0][0]',          \n",
            "                                                                  'max_pooling2d_3[0][0]']        \n",
            "                                                                                                  \n",
            " conv2d_80 (Conv2D)             (None, None, None,   573440      ['mixed8[0][0]']                 \n",
            "                                448)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_80 (BatchN  (None, None, None,   1344       ['conv2d_80[0][0]']              \n",
            " ormalization)                  448)                                                              \n",
            "                                                                                                  \n",
            " activation_80 (Activation)     (None, None, None,   0           ['batch_normalization_80[0][0]'] \n",
            "                                448)                                                              \n",
            "                                                                                                  \n",
            " conv2d_77 (Conv2D)             (None, None, None,   491520      ['mixed8[0][0]']                 \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " conv2d_81 (Conv2D)             (None, None, None,   1548288     ['activation_80[0][0]']          \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_77 (BatchN  (None, None, None,   1152       ['conv2d_77[0][0]']              \n",
            " ormalization)                  384)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_81 (BatchN  (None, None, None,   1152       ['conv2d_81[0][0]']              \n",
            " ormalization)                  384)                                                              \n",
            "                                                                                                  \n",
            " activation_77 (Activation)     (None, None, None,   0           ['batch_normalization_77[0][0]'] \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " activation_81 (Activation)     (None, None, None,   0           ['batch_normalization_81[0][0]'] \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " conv2d_78 (Conv2D)             (None, None, None,   442368      ['activation_77[0][0]']          \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " conv2d_79 (Conv2D)             (None, None, None,   442368      ['activation_77[0][0]']          \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " conv2d_82 (Conv2D)             (None, None, None,   442368      ['activation_81[0][0]']          \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " conv2d_83 (Conv2D)             (None, None, None,   442368      ['activation_81[0][0]']          \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " average_pooling2d_7 (AveragePo  (None, None, None,   0          ['mixed8[0][0]']                 \n",
            " oling2D)                       1280)                                                             \n",
            "                                                                                                  \n",
            " conv2d_76 (Conv2D)             (None, None, None,   409600      ['mixed8[0][0]']                 \n",
            "                                320)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_78 (BatchN  (None, None, None,   1152       ['conv2d_78[0][0]']              \n",
            " ormalization)                  384)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_79 (BatchN  (None, None, None,   1152       ['conv2d_79[0][0]']              \n",
            " ormalization)                  384)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_82 (BatchN  (None, None, None,   1152       ['conv2d_82[0][0]']              \n",
            " ormalization)                  384)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_83 (BatchN  (None, None, None,   1152       ['conv2d_83[0][0]']              \n",
            " ormalization)                  384)                                                              \n",
            "                                                                                                  \n",
            " conv2d_84 (Conv2D)             (None, None, None,   245760      ['average_pooling2d_7[0][0]']    \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_76 (BatchN  (None, None, None,   960        ['conv2d_76[0][0]']              \n",
            " ormalization)                  320)                                                              \n",
            "                                                                                                  \n",
            " activation_78 (Activation)     (None, None, None,   0           ['batch_normalization_78[0][0]'] \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " activation_79 (Activation)     (None, None, None,   0           ['batch_normalization_79[0][0]'] \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " activation_82 (Activation)     (None, None, None,   0           ['batch_normalization_82[0][0]'] \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " activation_83 (Activation)     (None, None, None,   0           ['batch_normalization_83[0][0]'] \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_84 (BatchN  (None, None, None,   576        ['conv2d_84[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " activation_76 (Activation)     (None, None, None,   0           ['batch_normalization_76[0][0]'] \n",
            "                                320)                                                              \n",
            "                                                                                                  \n",
            " mixed9_0 (Concatenate)         (None, None, None,   0           ['activation_78[0][0]',          \n",
            "                                768)                              'activation_79[0][0]']          \n",
            "                                                                                                  \n",
            " concatenate (Concatenate)      (None, None, None,   0           ['activation_82[0][0]',          \n",
            "                                768)                              'activation_83[0][0]']          \n",
            "                                                                                                  \n",
            " activation_84 (Activation)     (None, None, None,   0           ['batch_normalization_84[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " mixed9 (Concatenate)           (None, None, None,   0           ['activation_76[0][0]',          \n",
            "                                2048)                             'mixed9_0[0][0]',               \n",
            "                                                                  'concatenate[0][0]',            \n",
            "                                                                  'activation_84[0][0]']          \n",
            "                                                                                                  \n",
            " conv2d_89 (Conv2D)             (None, None, None,   917504      ['mixed9[0][0]']                 \n",
            "                                448)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_89 (BatchN  (None, None, None,   1344       ['conv2d_89[0][0]']              \n",
            " ormalization)                  448)                                                              \n",
            "                                                                                                  \n",
            " activation_89 (Activation)     (None, None, None,   0           ['batch_normalization_89[0][0]'] \n",
            "                                448)                                                              \n",
            "                                                                                                  \n",
            " conv2d_86 (Conv2D)             (None, None, None,   786432      ['mixed9[0][0]']                 \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " conv2d_90 (Conv2D)             (None, None, None,   1548288     ['activation_89[0][0]']          \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_86 (BatchN  (None, None, None,   1152       ['conv2d_86[0][0]']              \n",
            " ormalization)                  384)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_90 (BatchN  (None, None, None,   1152       ['conv2d_90[0][0]']              \n",
            " ormalization)                  384)                                                              \n",
            "                                                                                                  \n",
            " activation_86 (Activation)     (None, None, None,   0           ['batch_normalization_86[0][0]'] \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " activation_90 (Activation)     (None, None, None,   0           ['batch_normalization_90[0][0]'] \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " conv2d_87 (Conv2D)             (None, None, None,   442368      ['activation_86[0][0]']          \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " conv2d_88 (Conv2D)             (None, None, None,   442368      ['activation_86[0][0]']          \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " conv2d_91 (Conv2D)             (None, None, None,   442368      ['activation_90[0][0]']          \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " conv2d_92 (Conv2D)             (None, None, None,   442368      ['activation_90[0][0]']          \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " average_pooling2d_8 (AveragePo  (None, None, None,   0          ['mixed9[0][0]']                 \n",
            " oling2D)                       2048)                                                             \n",
            "                                                                                                  \n",
            " conv2d_85 (Conv2D)             (None, None, None,   655360      ['mixed9[0][0]']                 \n",
            "                                320)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_87 (BatchN  (None, None, None,   1152       ['conv2d_87[0][0]']              \n",
            " ormalization)                  384)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_88 (BatchN  (None, None, None,   1152       ['conv2d_88[0][0]']              \n",
            " ormalization)                  384)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_91 (BatchN  (None, None, None,   1152       ['conv2d_91[0][0]']              \n",
            " ormalization)                  384)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_92 (BatchN  (None, None, None,   1152       ['conv2d_92[0][0]']              \n",
            " ormalization)                  384)                                                              \n",
            "                                                                                                  \n",
            " conv2d_93 (Conv2D)             (None, None, None,   393216      ['average_pooling2d_8[0][0]']    \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_85 (BatchN  (None, None, None,   960        ['conv2d_85[0][0]']              \n",
            " ormalization)                  320)                                                              \n",
            "                                                                                                  \n",
            " activation_87 (Activation)     (None, None, None,   0           ['batch_normalization_87[0][0]'] \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " activation_88 (Activation)     (None, None, None,   0           ['batch_normalization_88[0][0]'] \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " activation_91 (Activation)     (None, None, None,   0           ['batch_normalization_91[0][0]'] \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " activation_92 (Activation)     (None, None, None,   0           ['batch_normalization_92[0][0]'] \n",
            "                                384)                                                              \n",
            "                                                                                                  \n",
            " batch_normalization_93 (BatchN  (None, None, None,   576        ['conv2d_93[0][0]']              \n",
            " ormalization)                  192)                                                              \n",
            "                                                                                                  \n",
            " activation_85 (Activation)     (None, None, None,   0           ['batch_normalization_85[0][0]'] \n",
            "                                320)                                                              \n",
            "                                                                                                  \n",
            " mixed9_1 (Concatenate)         (None, None, None,   0           ['activation_87[0][0]',          \n",
            "                                768)                              'activation_88[0][0]']          \n",
            "                                                                                                  \n",
            " concatenate_1 (Concatenate)    (None, None, None,   0           ['activation_91[0][0]',          \n",
            "                                768)                              'activation_92[0][0]']          \n",
            "                                                                                                  \n",
            " activation_93 (Activation)     (None, None, None,   0           ['batch_normalization_93[0][0]'] \n",
            "                                192)                                                              \n",
            "                                                                                                  \n",
            " mixed10 (Concatenate)          (None, None, None,   0           ['activation_85[0][0]',          \n",
            "                                2048)                             'mixed9_1[0][0]',               \n",
            "                                                                  'concatenate_1[0][0]',          \n",
            "                                                                  'activation_93[0][0]']          \n",
            "                                                                                                  \n",
            " global_average_pooling2d (Glob  (None, 2048)        0           ['mixed10[0][0]']                \n",
            " alAveragePooling2D)                                                                              \n",
            "                                                                                                  \n",
            " dropout (Dropout)              (None, 2048)         0           ['global_average_pooling2d[0][0]'\n",
            "                                                                 ]                                \n",
            "                                                                                                  \n",
            " dense (Dense)                  (None, 12)           24588       ['dropout[0][0]']                \n",
            "                                                                                                  \n",
            "==================================================================================================\n",
            "Total params: 21,827,372\n",
            "Trainable params: 21,792,940\n",
            "Non-trainable params: 34,432\n",
            "__________________________________________________________________________________________________\n"
          ]
        }
      ],
      "source": [
        "model.summary()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We compile the model using the optimizer and loss functions given below."
      ],
      "metadata": {
        "id": "flRMAhwgvtvq"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "2VZzCTknCXcb"
      },
      "outputs": [],
      "source": [
        "model.compile(\n",
        "  loss='categorical_crossentropy',\n",
        "  optimizer=\"adam\",\n",
        "  metrics=['accuracy']\n",
        ")"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We fit the model on the training set, while separating 20% for validation. "
      ],
      "metadata": {
        "id": "bzxkBSRuvvj0"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "background_save": true,
          "base_uri": "https://localhost:8080/"
        },
        "id": "M1chsmaWCcGk",
        "outputId": "ef10569f-ad6a-4972-f723-f194915e5294"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "Epoch 1/50\n",
            "83/83 [==============================] - 495s 6s/step - loss: 0.9739 - accuracy: 0.7002 - val_loss: 488.0224 - val_accuracy: 0.1160\n",
            "Epoch 2/50\n",
            "83/83 [==============================] - 487s 6s/step - loss: 0.4939 - accuracy: 0.8505 - val_loss: 1.9107 - val_accuracy: 0.4266\n",
            "Epoch 3/50\n",
            "83/83 [==============================] - 487s 6s/step - loss: 0.4692 - accuracy: 0.8508 - val_loss: 0.9870 - val_accuracy: 0.6928\n",
            "Epoch 4/50\n",
            "83/83 [==============================] - 489s 6s/step - loss: 0.2310 - accuracy: 0.9326 - val_loss: 0.6047 - val_accuracy: 0.8294\n",
            "Epoch 5/50\n",
            "83/83 [==============================] - 492s 6s/step - loss: 0.2307 - accuracy: 0.9250 - val_loss: 0.7571 - val_accuracy: 0.8328\n",
            "Epoch 6/50\n",
            "83/83 [==============================] - 490s 6s/step - loss: 0.2200 - accuracy: 0.9384 - val_loss: 0.9387 - val_accuracy: 0.7440\n",
            "Epoch 7/50\n",
            "83/83 [==============================] - 490s 6s/step - loss: 0.1865 - accuracy: 0.9460 - val_loss: 0.3637 - val_accuracy: 0.9147\n",
            "Epoch 8/50\n",
            "29/83 [=========>....................] - ETA: 5:14 - loss: 0.1882 - accuracy: 0.9494"
          ]
        }
      ],
      "source": [
        "history = model.fit(X_train, y_train,\n",
        "  validation_split = 0.2,\n",
        "  epochs=12,\n",
        "  batch_size=32,shuffle=True)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We can plot to see how accuracies change in the validation and training set as the model trains. This will help to make necessary modifications in the model."
      ],
      "metadata": {
        "id": "USnIqijyv7o7"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "-CMHYYXH__3F",
        "outputId": "535160aa-73ac-40f5-dc42-96e8dfab181b"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plt.plot(history.history['accuracy'], label='train acc')\n",
        "\n",
        "plt.plot(history.history['val_accuracy'], label='val acc')\n",
        "\n",
        "plt.legend()\n",
        "\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Similarly, we plot the loss function."
      ],
      "metadata": {
        "id": "Rwkv9Z16wAx0"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 265
        },
        "id": "ZYeGCMWdAOF9",
        "outputId": "00b0e302-2599-4e17-efc6-9f3ab0bdad18"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 432x288 with 1 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "plt.plot(history.history['loss'], label='train loss')\n",
        "plt.plot(history.history['val_loss'], label='val loss')\n",
        "plt.legend()\n",
        "plt.show()"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Now, we can evaluate our model on the test set. We see that the model achieves an accuracy of 93.8%."
      ],
      "metadata": {
        "id": "JlRl5DUUwGBZ"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "ASewBCVxAYK7",
        "outputId": "97f6e35e-dca3-480c-8f1c-31762eef5777"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "11/11 [==============================] - 1s 43ms/step - loss: 0.2531 - accuracy: 0.9385\n"
          ]
        },
        {
          "data": {
            "text/plain": [
              "[0.2531375586986542, 0.9384615421295166]"
            ]
          },
          "execution_count": 23,
          "metadata": {},
          "output_type": "execute_result"
        }
      ],
      "source": [
        "model.evaluate(X_test, y_test,batch_size=32)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We can get the predicted outcome on the rows using argmax."
      ],
      "metadata": {
        "id": "PsWqhobYwSJK"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "id": "H0aNF_K4juVy"
      },
      "outputs": [],
      "source": [
        "pred = model.predict(X_test)\n",
        "pred = np.argmax(pred,axis=1)\n",
        "y_test_new = np.argmax(y_test,axis=1)"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "We can evalute the model using classification_report."
      ],
      "metadata": {
        "id": "ea3fEaXRwUJF"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "CuYxxbZSnwhR",
        "outputId": "24a515e3-d5c3-4b89-c346-02114b50d73c"
      },
      "outputs": [
        {
          "name": "stdout",
          "output_type": "stream",
          "text": [
            "              precision    recall  f1-score   support\n",
            "\n",
            "           0       1.00      0.92      0.96        26\n",
            "           1       1.00      0.93      0.96        27\n",
            "           2       0.88      0.90      0.89        31\n",
            "           3       0.94      0.94      0.94        33\n",
            "           4       0.87      0.96      0.92        28\n",
            "           5       0.86      0.92      0.89        39\n",
            "           6       1.00      0.67      0.80         6\n",
            "           7       1.00      0.91      0.95        11\n",
            "           8       0.98      1.00      0.99        42\n",
            "           9       0.92      1.00      0.96        22\n",
            "          10       0.96      0.92      0.94        26\n",
            "          11       1.00      0.94      0.97        34\n",
            "\n",
            "    accuracy                           0.94       325\n",
            "   macro avg       0.95      0.92      0.93       325\n",
            "weighted avg       0.94      0.94      0.94       325\n",
            "\n"
          ]
        }
      ],
      "source": [
        "print(classification_report(y_test_new,pred))"
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "Confusion matrix is a very useful tool which shows the misclassifications. "
      ],
      "metadata": {
        "id": "L43Rovf2wYz9"
      }
    },
    {
      "cell_type": "code",
      "execution_count": null,
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 560
        },
        "id": "gFxz8-dTpPCG",
        "outputId": "a1caae3a-cb15-4632-e3ec-2c259a3e429e"
      },
      "outputs": [
        {
          "data": {
            "image/png": "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\n",
            "text/plain": [
              "<Figure size 1008x504 with 2 Axes>"
            ]
          },
          "metadata": {},
          "output_type": "display_data"
        }
      ],
      "source": [
        "fig,ax=plt.subplots(1,1,figsize=(14,7))\n",
        "sns.heatmap(confusion_matrix(y_test_new,pred),ax=ax,xticklabels=labels,yticklabels=labels,annot=True\n",
        "        ,alpha=0.7,linewidths=2)\n",
        "fig.text(s='Heatmap of the Confusion Matrix',size=18,fontweight='bold',\n",
        "             fontname='monospace',y=0.95,x=0.28,alpha=0.8)\n",
        "\n",
        "plt.show()"
      ]
    }
  ],
  "metadata": {
    "accelerator": "GPU",
    "colab": {
      "collapsed_sections": [],
      "name": "InceptionV3.ipynb",
      "provenance": [],
      "include_colab_link": true
    },
    "kernelspec": {
      "display_name": "Python 3",
      "name": "python3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "nbformat": 4,
  "nbformat_minor": 0
}