03_7CNN_with_image_augmentation.ipynb

{
 "cells": [
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "TfWEgbPgblp3"
   },
   "source": [
    "## Import relevant libraries "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "id": "Io5NCIV7bkl3"
   },
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")\n",
    "import os\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.layers import Input, Conv2D, BatchNormalization, MaxPooling2D, Dropout, GlobalAveragePooling2D, Dense\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.callbacks import EarlyStopping, ModelCheckpoint, ReduceLROnPlateau\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.metrics import confusion_matrix\n",
    "import numpy as np\n",
    "import seaborn as sns"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "cxKEDCgz6z0K"
   },
   "source": [
    "## Apply CNN Model with 7 Layers with Image Augmentation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "8vw6FVrd6qjz",
    "outputId": "d7eda847-199d-4a24-d543-b26598d57b19"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 66788 images belonging to 4 classes.\n",
      "Found 16696 images belonging to 4 classes.\n"
     ]
    }
   ],
   "source": [
    "# Reference: https://www.tensorflow.org/api_docs/python/tf/keras/preprocessing/image/ImageDataGenerator \n",
    "# Reference: https://stackoverflow.com/questions/53037510/can-flow-from-directory-get-train-and-validation-data-from-the-same-directory-in\n",
    "\n",
    "# generate batches of tensor image data with real-time data augmentation\n",
    "\n",
    "datagen=ImageDataGenerator(rescale=1./255,\n",
    "        shear_range=0.2,\n",
    "        zoom_range=0.2,\n",
    "        horizontal_flip=True,\n",
    "        validation_split=0.2)\n",
    "\n",
    "train_generator = datagen.flow_from_directory(\n",
    "    \"/OCT2017 /train\",\n",
    "    target_size = (224, 224),\n",
    "    batch_size=128,\n",
    "    class_mode = \"categorical\",\n",
    "    subset=\"training\")\n",
    "\n",
    "validation_generator = datagen.flow_from_directory(\n",
    "    \"/OCT2017 /train\",\n",
    "    target_size = (224, 224),\n",
    "    batch_size=128,\n",
    "    class_mode = \"categorical\",\n",
    "    subset=\"validation\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "id": "zHazHC097SNO"
   },
   "outputs": [],
   "source": [
    "tf.keras.backend.clear_session()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "id": "B0cmWrbs7STe"
   },
   "outputs": [],
   "source": [
    "# Reference: https://keras.io/api/layers/convolution_layers/convolution2d/\n",
    "# Reference: https://keras.io/api/layers/normalization_layers/batch_normalization/\n",
    "# Reference: https://keras.io/api/layers/pooling_layers/max_pooling2d/\n",
    "# Reference: https://keras.io/api/layers/regularization_layers/dropout/\n",
    "# Reference: https://keras.io/api/layers/pooling_layers/global_average_pooling2d/\n",
    "# Reference: https://keras.io/api/layers/core_layers/dense/\n",
    "\n",
    "\n",
    "input_shape = (224,224,3)\n",
    "\n",
    "# Input Layer\n",
    "Input_Layer = Input(shape=(input_shape))\n",
    "\n",
    "# Convolutional Layer\n",
    "First_Conv2D = Conv2D(32, kernel_size = (3, 3), activation=\"relu\",padding = 'same')(Input_Layer)\n",
    "# Batch Normalization\n",
    "First_BatchNormalization = BatchNormalization()(First_Conv2D)\n",
    "\n",
    "# Convolutional Layer\n",
    "Second_Conv2D = Conv2D(32, kernel_size = (3, 3), activation=\"relu\",padding = 'same')(First_BatchNormalization)\n",
    "# Batch Normalization\n",
    "Second_BatchNormalization = BatchNormalization()(Second_Conv2D)\n",
    "\n",
    "# MaxPool Layer\n",
    "maxpool = MaxPooling2D(pool_size=(2,2))(Second_BatchNormalization)\n",
    "# Dropout Layer\n",
    "dropout = Dropout(rate=0.2)(maxpool)\n",
    "\n",
    "# Convolutional Layer\n",
    "Third_Conv2D = Conv2D(64, kernel_size = (3, 3), activation=\"relu\",padding = 'same')(dropout)\n",
    "# Batch Normalization\n",
    "Third_BatchNormalization = BatchNormalization()(Third_Conv2D)\n",
    "\n",
    "# Convolutional Layer\n",
    "Fourth_Conv2D = Conv2D(64, kernel_size = (3, 3), activation=\"relu\",padding = 'same')(Third_BatchNormalization)\n",
    "# Batch Normalization\n",
    "Fourth_BatchNormalization = BatchNormalization()(Fourth_Conv2D)\n",
    "\n",
    "# MaxPool Layer\n",
    "maxpool = MaxPooling2D(pool_size=(2,2))(Fourth_BatchNormalization)\n",
    "# Dropout Layer\n",
    "dropout = Dropout(rate=0.2)(maxpool)\n",
    "\n",
    "# Convolutional Layer\n",
    "Fifth_Conv2D = Conv2D(128, kernel_size = (3, 3), activation=\"relu\",padding = 'same')(dropout)\n",
    "# Batch Normalization\n",
    "Fifth_BatchNormalization = BatchNormalization()(Fifth_Conv2D)\n",
    "\n",
    "# Convolutional Layer\n",
    "Sixth_Conv2D = Conv2D(128, kernel_size = (3, 3), activation=\"relu\",padding = 'same')(Fifth_BatchNormalization)\n",
    "# Batch Normalization\n",
    "Sixth_BatchNormalization = BatchNormalization()(Sixth_Conv2D)\n",
    "\n",
    "# MaxPool Layer\n",
    "maxpool = MaxPooling2D(pool_size=(2,2))(Sixth_BatchNormalization)\n",
    "# Dropout Layer\n",
    "dropout = Dropout(rate=0.2)(maxpool)\n",
    "\n",
    "# Convolutional Layer\n",
    "Seventh_Conv2D = Conv2D(256, kernel_size = (3, 3), activation=\"relu\",padding = 'same')(dropout)\n",
    "# Batch Normalization\n",
    "Seventh_BatchNormalization = BatchNormalization()(Seventh_Conv2D)\n",
    "\n",
    "# GlobalAveragePooling Layer\n",
    "global_average_pool = GlobalAveragePooling2D()(Seventh_BatchNormalization)\n",
    "\n",
    "# Dense Layer\n",
    "Output_Layer = Dense(4, activation=\"softmax\")(global_average_pool)\n",
    "\n",
    "# Creating a model\n",
    "model2 = Model(inputs=Input_Layer,outputs=Output_Layer)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "KTTAwgxS7SWd",
    "outputId": "27305c0c-6dba-4af6-869a-4dcf464c754e"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " input_1 (InputLayer)        [(None, 224, 224, 3)]     0         \n",
      "                                                                 \n",
      " conv2d (Conv2D)             (None, 224, 224, 32)      896       \n",
      "                                                                 \n",
      " batch_normalization (BatchN  (None, 224, 224, 32)     128       \n",
      " ormalization)                                                   \n",
      "                                                                 \n",
      " conv2d_1 (Conv2D)           (None, 224, 224, 32)      9248      \n",
      "                                                                 \n",
      " batch_normalization_1 (Batc  (None, 224, 224, 32)     128       \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " max_pooling2d (MaxPooling2D  (None, 112, 112, 32)     0         \n",
      " )                                                               \n",
      "                                                                 \n",
      " dropout (Dropout)           (None, 112, 112, 32)      0         \n",
      "                                                                 \n",
      " conv2d_2 (Conv2D)           (None, 112, 112, 64)      18496     \n",
      "                                                                 \n",
      " batch_normalization_2 (Batc  (None, 112, 112, 64)     256       \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " conv2d_3 (Conv2D)           (None, 112, 112, 64)      36928     \n",
      "                                                                 \n",
      " batch_normalization_3 (Batc  (None, 112, 112, 64)     256       \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " max_pooling2d_1 (MaxPooling  (None, 56, 56, 64)       0         \n",
      " 2D)                                                             \n",
      "                                                                 \n",
      " dropout_1 (Dropout)         (None, 56, 56, 64)        0         \n",
      "                                                                 \n",
      " conv2d_4 (Conv2D)           (None, 56, 56, 128)       73856     \n",
      "                                                                 \n",
      " batch_normalization_4 (Batc  (None, 56, 56, 128)      512       \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " conv2d_5 (Conv2D)           (None, 56, 56, 128)       147584    \n",
      "                                                                 \n",
      " batch_normalization_5 (Batc  (None, 56, 56, 128)      512       \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " max_pooling2d_2 (MaxPooling  (None, 28, 28, 128)      0         \n",
      " 2D)                                                             \n",
      "                                                                 \n",
      " dropout_2 (Dropout)         (None, 28, 28, 128)       0         \n",
      "                                                                 \n",
      " conv2d_6 (Conv2D)           (None, 28, 28, 256)       295168    \n",
      "                                                                 \n",
      " batch_normalization_6 (Batc  (None, 28, 28, 256)      1024      \n",
      " hNormalization)                                                 \n",
      "                                                                 \n",
      " global_average_pooling2d (G  (None, 256)              0         \n",
      " lobalAveragePooling2D)                                          \n",
      "                                                                 \n",
      " dense (Dense)               (None, 4)                 1028      \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 586,020\n",
      "Trainable params: 584,612\n",
      "Non-trainable params: 1,408\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model2.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "cU89aRBWv0zB",
    "outputId": "00438720-651b-4589-ad93-ae30c8d124be"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.4440 - accuracy: 0.8405\n",
      "Epoch 1: val_accuracy improved from -inf to 0.73664, saving model to model2_with_augmentation_save/weights-01-0.7366.hdf5\n",
      "521/521 [==============================] - 1241s 2s/step - loss: 0.4440 - accuracy: 0.8405 - val_loss: 0.8730 - val_accuracy: 0.7366\n",
      "Epoch 2/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.2231 - accuracy: 0.9235\n",
      "Epoch 2: val_accuracy improved from 0.73664 to 0.92651, saving model to model2_with_augmentation_save/weights-02-0.9265.hdf5\n",
      "521/521 [==============================] - 1159s 2s/step - loss: 0.2231 - accuracy: 0.9235 - val_loss: 0.2186 - val_accuracy: 0.9265\n",
      "Epoch 3/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.1888 - accuracy: 0.9362\n",
      "Epoch 3: val_accuracy improved from 0.92651 to 0.93256, saving model to model2_with_augmentation_save/weights-03-0.9326.hdf5\n",
      "521/521 [==============================] - 1146s 2s/step - loss: 0.1888 - accuracy: 0.9362 - val_loss: 0.1922 - val_accuracy: 0.9326\n",
      "Epoch 4/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.1686 - accuracy: 0.9422\n",
      "Epoch 4: val_accuracy did not improve from 0.93256\n",
      "521/521 [==============================] - 1114s 2s/step - loss: 0.1686 - accuracy: 0.9422 - val_loss: 0.2739 - val_accuracy: 0.9118\n",
      "Epoch 5/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.1576 - accuracy: 0.9464\n",
      "Epoch 5: val_accuracy did not improve from 0.93256\n",
      "521/521 [==============================] - 1118s 2s/step - loss: 0.1576 - accuracy: 0.9464 - val_loss: 0.2601 - val_accuracy: 0.9079\n",
      "Epoch 6/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.1512 - accuracy: 0.9479\n",
      "Epoch 6: val_accuracy did not improve from 0.93256\n",
      "521/521 [==============================] - 1113s 2s/step - loss: 0.1512 - accuracy: 0.9479 - val_loss: 0.2147 - val_accuracy: 0.9313\n",
      "Epoch 7/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.1391 - accuracy: 0.9522\n",
      "Epoch 7: val_accuracy improved from 0.93256 to 0.94005, saving model to model2_with_augmentation_save/weights-07-0.9400.hdf5\n",
      "521/521 [==============================] - 1126s 2s/step - loss: 0.1391 - accuracy: 0.9522 - val_loss: 0.1801 - val_accuracy: 0.9400\n",
      "Epoch 8/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.1360 - accuracy: 0.9542\n",
      "Epoch 8: val_accuracy did not improve from 0.94005\n",
      "521/521 [==============================] - 1131s 2s/step - loss: 0.1360 - accuracy: 0.9542 - val_loss: 0.2948 - val_accuracy: 0.9039\n",
      "Epoch 9/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.1276 - accuracy: 0.9559\n",
      "Epoch 9: val_accuracy improved from 0.94005 to 0.94017, saving model to model2_with_augmentation_save/weights-09-0.9402.hdf5\n",
      "521/521 [==============================] - 1101s 2s/step - loss: 0.1276 - accuracy: 0.9559 - val_loss: 0.1793 - val_accuracy: 0.9402\n",
      "Epoch 10/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.1234 - accuracy: 0.9583\n",
      "Epoch 10: val_accuracy improved from 0.94017 to 0.95214, saving model to model2_with_augmentation_save/weights-10-0.9521.hdf5\n",
      "521/521 [==============================] - 1093s 2s/step - loss: 0.1234 - accuracy: 0.9583 - val_loss: 0.1431 - val_accuracy: 0.9521\n",
      "Epoch 11/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.1168 - accuracy: 0.9599\n",
      "Epoch 11: val_accuracy did not improve from 0.95214\n",
      "521/521 [==============================] - 1068s 2s/step - loss: 0.1168 - accuracy: 0.9599 - val_loss: 0.1667 - val_accuracy: 0.9417\n",
      "Epoch 12/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.1152 - accuracy: 0.9611\n",
      "Epoch 12: val_accuracy did not improve from 0.95214\n",
      "521/521 [==============================] - 1076s 2s/step - loss: 0.1152 - accuracy: 0.9611 - val_loss: 0.3727 - val_accuracy: 0.8917\n",
      "Epoch 13/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.1121 - accuracy: 0.9617\n",
      "Epoch 13: val_accuracy did not improve from 0.95214\n",
      "521/521 [==============================] - 1078s 2s/step - loss: 0.1121 - accuracy: 0.9617 - val_loss: 0.1580 - val_accuracy: 0.9463\n",
      "Epoch 14/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.1057 - accuracy: 0.9637\n",
      "Epoch 14: val_accuracy did not improve from 0.95214\n",
      "521/521 [==============================] - 1075s 2s/step - loss: 0.1057 - accuracy: 0.9637 - val_loss: 0.1709 - val_accuracy: 0.9447\n",
      "Epoch 15/50\n",
      "521/521 [==============================] - ETA: 0s - loss: 0.1051 - accuracy: 0.9636\n",
      "Epoch 15: val_accuracy did not improve from 0.95214\n",
      "521/521 [==============================] - 1101s 2s/step - loss: 0.1051 - accuracy: 0.9636 - val_loss: 0.4228 - val_accuracy: 0.8816\n",
      "Epoch 15: early stopping\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<keras.callbacks.History at 0x7f974fa3b9d0>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "earlystop = EarlyStopping(monitor='val_accuracy', min_delta=0, patience=5, verbose=1)\n",
    "\n",
    "# Creating model2_with_augmentation_save folder before running the code.\n",
    "filepath=\"model2_with_augmentation_save/weights-{epoch:02d}-{val_accuracy:.4f}.hdf5\"\n",
    "checkpoint = ModelCheckpoint(filepath=filepath, monitor='val_accuracy',verbose=1, save_best_only=True, mode='max')\n",
    "\n",
    "# Callbacks\n",
    "optimizer = tf.keras.optimizers.Adam()\n",
    "model2.compile(optimizer=optimizer,loss=\"categorical_crossentropy\",metrics=[\"accuracy\"])\n",
    "\n",
    "# Creating a list with all the callbacks we want\n",
    "callback_list = [earlystop, checkpoint]\n",
    "\n",
    "train_samples = 66788\n",
    "validation_samples = 16696\n",
    "batch_size = 128\n",
    "model2.fit(train_generator, validation_data= validation_generator, steps_per_epoch=train_samples//batch_size, callbacks=callback_list, epochs=50)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 231
    },
    "id": "FD9Bd4bH7ScN",
    "outputId": "da0d815b-c767-4de8-81eb-1a4b72b2e362"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 2160x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot training & validation acc values\n",
    "plt.figure(figsize=(30, 5))\n",
    "plt.subplot(121)\n",
    "plt.plot(model2.history.history['accuracy'])\n",
    "plt.plot(model2.history.history['val_accuracy'])\n",
    "plt.title('Model Accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Test'], loc='upper left')\n",
    "\n",
    "# Plot training & validation loss values\n",
    "plt.subplot(122)\n",
    "plt.plot(model2.history.history['loss'])\n",
    "plt.plot(model2.history.history['val_loss'])\n",
    "plt.title('Model loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Train', 'Test'], loc='upper left')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/"
    },
    "id": "4b0UPNxUBx66",
    "outputId": "d6d962ff-23a4-4eb1-bce0-8623b38edabe"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 968 images belonging to 4 classes.\n",
      "8/8 [==============================] - 13s 2s/step - loss: 0.1103 - accuracy: 0.9638\n",
      "Accuracy on Test Dataset:96.38429880142212, Loss on Test Dataset: 0.11026882380247116\n"
     ]
    }
   ],
   "source": [
    "# Accuracy on test dataset\n",
    "\n",
    "test_generator = datagen.flow_from_directory(\n",
    "    \"/OCT2017 /test\",\n",
    "    target_size = (224, 224),\n",
    "    batch_size=128,\n",
    "    class_mode = \"categorical\",\n",
    "    shuffle = False)\n",
    "test_samples = 968\n",
    "batch_size = 128\n",
    "\n",
    "result = model2.evaluate(test_generator, steps = test_samples // batch_size+1, verbose = 1)\n",
    "print(\"Accuracy on Test Dataset:{0}, Loss on Test Dataset: {1}\".format(result[1]*100,result[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 295
    },
    "id": "VY-mqqNbAhtz",
    "outputId": "5baee339-f1c1-42f0-bf47-27b0f0b8f25a"
   },
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Plot Confusion Matrix\n",
    "\n",
    "# predicted values\n",
    "y_pred = model2.predict(test_generator, test_samples // batch_size+1)\n",
    "predicted_values = np.argmax(y_pred, axis=1)\n",
    "\n",
    "confusionmatrix = confusion_matrix(test_generator.classes,predicted_values)\n",
    "plt.title(\"7 Layer CNN without Image Augmentation\")\n",
    "sns.heatmap(confusionmatrix, annot=True, cmap=\"YlGnBu\")\n",
    "plt.xlabel('Predicted Values')\n",
    "plt.ylabel('Actual Values')\n",
    "plt.show()"
   ]
  },
  {
   "attachments": {},
   "cell_type": "markdown",
   "metadata": {
    "id": "-ayh0Db9SKlG"
   },
   "source": [
    "**Observations**\n",
    "\n",
    "We are getting an accuracy of 0.96 on the test dataset using 7 Layers CNN Model with Image Augmentation."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "id": "uNm7-2o9AsAQ"
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "collapsed_sections": [],
   "machine_shape": "hm",
   "name": "03 7CNN with Image Augmentation.ipynb",
   "provenance": []
  },
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}