Vgg16-fine-tunning.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "_cell_guid": "b1076dfc-b9ad-4769-8c92-a6c4dae69d19",
    "_uuid": "8f2839f25d086af736a60e9eeb907d3b93b6e0e5"
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "Using TensorFlow backend.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO]: Tensorflow version2.1.0\n",
      "[INFO]: GPU usage/device:GPU:0\n"
     ]
    }
   ],
   "source": [
    "#Import tensorflow and keras library\n",
    "import tensorflow as tf\n",
    "import keras_preprocessing\n",
    "from tensorflow.keras.preprocessing import image\n",
    "import pickle\n",
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.callbacks import TensorBoard\n",
    "from keras.models import Sequential\n",
    "from keras.layers import Convolution2D,MaxPooling2D,Flatten,Dense,Dropout\n",
    "from keras.layers.normalization import BatchNormalization\n",
    "from keras.optimizers import Adam\n",
    "import keras\n",
    "from tensorflow.keras.callbacks import ModelCheckpoint\n",
    "\n",
    "print(\"[INFO]: Tensorflow version{}\".format(tf.__version__))\n",
    "state_gpu = tf.test.gpu_device_name()\n",
    "print(\"[INFO]: GPU usage{0}\".format(state_gpu))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "ROT_RANGE = 10\n",
    "\n",
    "# this is the augmentation configuration we will use for training\n",
    "train_gen = ImageDataGenerator(\n",
    "rescale = 1./255,\n",
    "rotation_range = ROT_RANGE,\n",
    "width_shift_range=0.2,\n",
    "height_shift_range=0.2,\n",
    "shear_range=0.2,\n",
    "zoom_range=0.2,\n",
    "horizontal_flip=True,\n",
    "fill_mode='nearest')\n",
    "\n",
    "valid_gen = ImageDataGenerator(rescale = 1./255)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "TRAINING_DIR = '../input/tomato/New Plant Diseases Dataset(Augmented)/train/' \n",
    "VALIDATION_DIR = '../input/tomato/New Plant Diseases Dataset(Augmented)/valid/'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 18345 images belonging to 10 classes.\n",
      "Found 4585 images belonging to 10 classes.\n"
     ]
    }
   ],
   "source": [
    "TARGET_SIZE = (224,224)\n",
    "TRAIN_BATCH_SIZE = 128\n",
    "VALID_BATCH_SIZE = 32\n",
    "SEED = 42\n",
    "\n",
    "#Data Iterator\n",
    "train_data = train_gen.flow_from_directory(\n",
    "TRAINING_DIR,\n",
    "target_size = TARGET_SIZE,\n",
    "class_mode = 'categorical',\n",
    "color_mode = \"rgb\",\n",
    "batch_size = TRAIN_BATCH_SIZE,\n",
    "shuffle = True,\n",
    "seed = SEED\n",
    ")\n",
    "\n",
    "valid_data = valid_gen.flow_from_directory(\n",
    "VALIDATION_DIR,\n",
    "target_size = TARGET_SIZE,\n",
    "class_mode = 'categorical',\n",
    "color_mode = \"rgb\",\n",
    "batch_size = VALID_BATCH_SIZE\n",
    ")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "from keras.applications.vgg16 import VGG16\n",
    "base_model_weights_path = '/kaggle/input/vgg16/vgg16_weights_tf_dim_ordering_tf_kernels_notop.h5'\n",
    "base_model = VGG16(weights=base_model_weights_path, include_top=False, input_shape=(224, 224, 3))\n",
    "\n",
    "x = keras.layers.Flatten() (base_model.output)\n",
    "x = keras.layers.Dense(256, activation=\"relu\")(x)\n",
    "x = keras.layers.Dropout(0.25)(x)\n",
    "output = keras.layers.Dense(10, activation='softmax')(x)\n",
    "model = keras.models.Model(inputs=base_model.input, outputs=output)\n",
    "\n",
    "# The newly added layers are initialized with random values.\n",
    "# Make sure based model remain unchanged until newly added layers weights get reasonable values.\n",
    "for layer in base_model.layers:\n",
    "    layer.trainable = False"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input_1 (InputLayer)         (None, 224, 224, 3)       0         \n",
      "_________________________________________________________________\n",
      "block1_conv1 (Conv2D)        (None, 224, 224, 64)      1792      \n",
      "_________________________________________________________________\n",
      "block1_conv2 (Conv2D)        (None, 224, 224, 64)      36928     \n",
      "_________________________________________________________________\n",
      "block1_pool (MaxPooling2D)   (None, 112, 112, 64)      0         \n",
      "_________________________________________________________________\n",
      "block2_conv1 (Conv2D)        (None, 112, 112, 128)     73856     \n",
      "_________________________________________________________________\n",
      "block2_conv2 (Conv2D)        (None, 112, 112, 128)     147584    \n",
      "_________________________________________________________________\n",
      "block2_pool (MaxPooling2D)   (None, 56, 56, 128)       0         \n",
      "_________________________________________________________________\n",
      "block3_conv1 (Conv2D)        (None, 56, 56, 256)       295168    \n",
      "_________________________________________________________________\n",
      "block3_conv2 (Conv2D)        (None, 56, 56, 256)       590080    \n",
      "_________________________________________________________________\n",
      "block3_conv3 (Conv2D)        (None, 56, 56, 256)       590080    \n",
      "_________________________________________________________________\n",
      "block3_pool (MaxPooling2D)   (None, 28, 28, 256)       0         \n",
      "_________________________________________________________________\n",
      "block4_conv1 (Conv2D)        (None, 28, 28, 512)       1180160   \n",
      "_________________________________________________________________\n",
      "block4_conv2 (Conv2D)        (None, 28, 28, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "block4_conv3 (Conv2D)        (None, 28, 28, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "block4_pool (MaxPooling2D)   (None, 14, 14, 512)       0         \n",
      "_________________________________________________________________\n",
      "block5_conv1 (Conv2D)        (None, 14, 14, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv2 (Conv2D)        (None, 14, 14, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "block5_conv3 (Conv2D)        (None, 14, 14, 512)       2359808   \n",
      "_________________________________________________________________\n",
      "block5_pool (MaxPooling2D)   (None, 7, 7, 512)         0         \n",
      "_________________________________________________________________\n",
      "flatten_1 (Flatten)          (None, 25088)             0         \n",
      "_________________________________________________________________\n",
      "dense_1 (Dense)              (None, 256)               6422784   \n",
      "_________________________________________________________________\n",
      "dropout_1 (Dropout)          (None, 256)               0         \n",
      "_________________________________________________________________\n",
      "dense_2 (Dense)              (None, 10)                2570      \n",
      "=================================================================\n",
      "Total params: 21,140,042\n",
      "Trainable params: 6,425,354\n",
      "Non-trainable params: 14,714,688\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "model.summary()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "LEARNING_RATE = 0.0001\n",
    "#LEARNING_RATE = 0.001\n",
    "\n",
    "#Optimizer\n",
    "opt = Adam(lr = LEARNING_RATE)\n",
    "model.compile(optimizer=opt, loss=keras.losses.categorical_crossentropy, metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/25\n",
      "143/143 [==============================] - 319s 2s/step - loss: 1.3667 - accuracy: 0.5452 - val_loss: 0.7110 - val_accuracy: 0.7314\n",
      "Epoch 2/25\n",
      "143/143 [==============================] - 282s 2s/step - loss: 0.8520 - accuracy: 0.7262 - val_loss: 0.7263 - val_accuracy: 0.8247\n",
      "Epoch 3/25\n",
      "143/143 [==============================] - 281s 2s/step - loss: 0.7173 - accuracy: 0.7692 - val_loss: 0.5757 - val_accuracy: 0.8322\n",
      "Epoch 4/25\n",
      "143/143 [==============================] - 281s 2s/step - loss: 0.6330 - accuracy: 0.7971 - val_loss: 0.4696 - val_accuracy: 0.8335\n",
      "Epoch 5/25\n",
      "143/143 [==============================] - 280s 2s/step - loss: 0.5798 - accuracy: 0.8120 - val_loss: 0.4253 - val_accuracy: 0.8537\n",
      "Epoch 6/25\n",
      "143/143 [==============================] - 280s 2s/step - loss: 0.5414 - accuracy: 0.8241 - val_loss: 0.2611 - val_accuracy: 0.8667\n",
      "Epoch 7/25\n",
      "143/143 [==============================] - 279s 2s/step - loss: 0.5081 - accuracy: 0.8328 - val_loss: 0.5806 - val_accuracy: 0.8476\n",
      "Epoch 8/25\n",
      "143/143 [==============================] - 276s 2s/step - loss: 0.4819 - accuracy: 0.8432 - val_loss: 0.4636 - val_accuracy: 0.8625\n",
      "Epoch 9/25\n",
      "143/143 [==============================] - 275s 2s/step - loss: 0.4588 - accuracy: 0.8495 - val_loss: 0.3196 - val_accuracy: 0.8711\n",
      "Epoch 10/25\n",
      "143/143 [==============================] - 274s 2s/step - loss: 0.4538 - accuracy: 0.8498 - val_loss: 0.4224 - val_accuracy: 0.8733\n",
      "Epoch 11/25\n",
      "143/143 [==============================] - 275s 2s/step - loss: 0.4194 - accuracy: 0.8610 - val_loss: 0.4529 - val_accuracy: 0.8981\n",
      "Epoch 12/25\n",
      "143/143 [==============================] - 274s 2s/step - loss: 0.4146 - accuracy: 0.8624 - val_loss: 0.3035 - val_accuracy: 0.8748\n",
      "Epoch 13/25\n",
      "143/143 [==============================] - 272s 2s/step - loss: 0.3988 - accuracy: 0.8680 - val_loss: 0.2441 - val_accuracy: 0.8836\n",
      "Epoch 14/25\n",
      "143/143 [==============================] - 273s 2s/step - loss: 0.3859 - accuracy: 0.8734 - val_loss: 0.3452 - val_accuracy: 0.8849\n",
      "Epoch 15/25\n",
      "143/143 [==============================] - 273s 2s/step - loss: 0.3777 - accuracy: 0.8732 - val_loss: 0.2317 - val_accuracy: 0.8722\n",
      "Epoch 16/25\n",
      "143/143 [==============================] - 273s 2s/step - loss: 0.3602 - accuracy: 0.8792 - val_loss: 0.2745 - val_accuracy: 0.8867\n",
      "Epoch 17/25\n",
      "143/143 [==============================] - 273s 2s/step - loss: 0.3532 - accuracy: 0.8835 - val_loss: 0.5543 - val_accuracy: 0.8875\n",
      "Epoch 18/25\n",
      "143/143 [==============================] - 274s 2s/step - loss: 0.3545 - accuracy: 0.8820 - val_loss: 0.3280 - val_accuracy: 0.8803\n",
      "Epoch 19/25\n",
      "143/143 [==============================] - 275s 2s/step - loss: 0.3443 - accuracy: 0.8853 - val_loss: 0.2181 - val_accuracy: 0.9023\n",
      "Epoch 20/25\n",
      "143/143 [==============================] - 272s 2s/step - loss: 0.3345 - accuracy: 0.8890 - val_loss: 0.1595 - val_accuracy: 0.9093\n",
      "Epoch 21/25\n",
      "143/143 [==============================] - 273s 2s/step - loss: 0.3293 - accuracy: 0.8899 - val_loss: 0.2461 - val_accuracy: 0.8864\n",
      "Epoch 22/25\n",
      "143/143 [==============================] - 273s 2s/step - loss: 0.3180 - accuracy: 0.8931 - val_loss: 0.3945 - val_accuracy: 0.8939\n",
      "Epoch 23/25\n",
      "143/143 [==============================] - 274s 2s/step - loss: 0.3203 - accuracy: 0.8919 - val_loss: 0.3014 - val_accuracy: 0.9012\n",
      "Epoch 24/25\n",
      "143/143 [==============================] - 273s 2s/step - loss: 0.3131 - accuracy: 0.8952 - val_loss: 0.5049 - val_accuracy: 0.8961\n",
      "Epoch 25/25\n",
      "143/143 [==============================] - 275s 2s/step - loss: 0.2985 - accuracy: 0.8988 - val_loss: 0.1816 - val_accuracy: 0.8939\n"
     ]
    }
   ],
   "source": [
    "TRAINING_NUM = train_data.n #or train_data.samples\n",
    "VALID_NUM = valid_data.n\n",
    "EPOCHS = 25\n",
    "\n",
    "STEP_SIZE_TRAIN = TRAINING_NUM // TRAIN_BATCH_SIZE \n",
    "STEP_SIZE_VALID = VALID_NUM // VALID_BATCH_SIZE\n",
    "\n",
    "# Fit model to get reasonable weights for newly added layers.\n",
    "history = model.fit_generator(generator = train_data,\n",
    "                             steps_per_epoch = STEP_SIZE_TRAIN,\n",
    "                             validation_data = valid_data,\n",
    "                             validation_steps = STEP_SIZE_VALID,\n",
    "                             epochs = EPOCHS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.figure( figsize = (15,8)) \n",
    "    \n",
    "plt.subplot(221)  \n",
    "# Accuracy \n",
    "plt.plot(model.history.history['accuracy'])\n",
    "plt.plot(model.history.history['val_accuracy'])\n",
    "plt.title('Model Accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Training set', 'Validation set'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "\n",
    "# Loss\n",
    "plt.figure( figsize = (15,8)) \n",
    "plt.subplot(222)  \n",
    "plt.plot(model.history.history['loss'])\n",
    "plt.plot(model.history.history['val_loss'])\n",
    "plt.title('Model Loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Training set', 'Validation set'], loc='upper left')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] Best Validation Accuracy: % 90.92905521392822\n"
     ]
    }
   ],
   "source": [
    "best_val_acc = max(history.history['val_accuracy'])\n",
    "print(\"[INFO] Best Validation Accuracy: %\",best_val_acc*100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "for layer in base_model.layers:\n",
    "    layer.trainable = True\n",
    "\n",
    "# compile the model with a SGD/momentum optimizer\n",
    "# and a very slow learning rate (This ensures the base model weights do not change a lot)\n",
    "model.compile(loss='categorical_crossentropy',\n",
    "              optimizer=keras.optimizers.SGD(lr=1e-4, momentum=0.9),\n",
    "              metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/25\n",
      "143/143 [==============================] - 321s 2s/step - loss: 0.2553 - accuracy: 0.9137 - val_loss: 0.2969 - val_accuracy: 0.9294\n",
      "Epoch 2/25\n",
      "143/143 [==============================] - 313s 2s/step - loss: 0.1931 - accuracy: 0.9330 - val_loss: 0.1213 - val_accuracy: 0.9284\n",
      "Epoch 3/25\n",
      "143/143 [==============================] - 312s 2s/step - loss: 0.1715 - accuracy: 0.9399 - val_loss: 0.0751 - val_accuracy: 0.9471\n",
      "Epoch 4/25\n",
      "143/143 [==============================] - 312s 2s/step - loss: 0.1472 - accuracy: 0.9489 - val_loss: 0.1280 - val_accuracy: 0.9433\n",
      "Epoch 5/25\n",
      "143/143 [==============================] - 320s 2s/step - loss: 0.1272 - accuracy: 0.9573 - val_loss: 0.3136 - val_accuracy: 0.9354\n",
      "Epoch 6/25\n",
      "143/143 [==============================] - 311s 2s/step - loss: 0.1164 - accuracy: 0.9601 - val_loss: 0.0438 - val_accuracy: 0.9572\n",
      "Epoch 7/25\n",
      "143/143 [==============================] - 313s 2s/step - loss: 0.1143 - accuracy: 0.9599 - val_loss: 0.0725 - val_accuracy: 0.9600\n",
      "Epoch 8/25\n",
      "143/143 [==============================] - 312s 2s/step - loss: 0.1025 - accuracy: 0.9653 - val_loss: 0.1306 - val_accuracy: 0.9635\n",
      "Epoch 9/25\n",
      "143/143 [==============================] - 319s 2s/step - loss: 0.0937 - accuracy: 0.9674 - val_loss: 0.0192 - val_accuracy: 0.9611\n",
      "Epoch 10/25\n",
      "143/143 [==============================] - 326s 2s/step - loss: 0.0815 - accuracy: 0.9722 - val_loss: 0.2753 - val_accuracy: 0.9677\n",
      "Epoch 11/25\n",
      "143/143 [==============================] - 314s 2s/step - loss: 0.0831 - accuracy: 0.9722 - val_loss: 0.0835 - val_accuracy: 0.9728\n",
      "Epoch 12/25\n",
      "143/143 [==============================] - 311s 2s/step - loss: 0.0823 - accuracy: 0.9721 - val_loss: 0.0878 - val_accuracy: 0.9598\n",
      "Epoch 13/25\n",
      "143/143 [==============================] - 316s 2s/step - loss: 0.0743 - accuracy: 0.9755 - val_loss: 0.1708 - val_accuracy: 0.9723\n",
      "Epoch 14/25\n",
      "143/143 [==============================] - 314s 2s/step - loss: 0.0657 - accuracy: 0.9772 - val_loss: 0.0967 - val_accuracy: 0.9730\n",
      "Epoch 15/25\n",
      "143/143 [==============================] - 317s 2s/step - loss: 0.0687 - accuracy: 0.9750 - val_loss: 0.0104 - val_accuracy: 0.9701\n",
      "Epoch 16/25\n",
      "143/143 [==============================] - 319s 2s/step - loss: 0.0615 - accuracy: 0.9790 - val_loss: 0.1180 - val_accuracy: 0.9743\n",
      "Epoch 17/25\n",
      "143/143 [==============================] - 309s 2s/step - loss: 0.0587 - accuracy: 0.9795 - val_loss: 0.0660 - val_accuracy: 0.9655\n",
      "Epoch 18/25\n",
      "143/143 [==============================] - 316s 2s/step - loss: 0.0559 - accuracy: 0.9807 - val_loss: 0.0671 - val_accuracy: 0.9802\n",
      "Epoch 19/25\n",
      "143/143 [==============================] - 312s 2s/step - loss: 0.0530 - accuracy: 0.9810 - val_loss: 0.0227 - val_accuracy: 0.9776\n",
      "Epoch 20/25\n",
      "143/143 [==============================] - 316s 2s/step - loss: 0.0512 - accuracy: 0.9822 - val_loss: 0.0302 - val_accuracy: 0.9646\n",
      "Epoch 21/25\n",
      "143/143 [==============================] - 312s 2s/step - loss: 0.0539 - accuracy: 0.9811 - val_loss: 0.0615 - val_accuracy: 0.9754\n",
      "Epoch 22/25\n",
      "143/143 [==============================] - 308s 2s/step - loss: 0.0495 - accuracy: 0.9833 - val_loss: 0.0137 - val_accuracy: 0.9756\n",
      "Epoch 23/25\n",
      "143/143 [==============================] - 310s 2s/step - loss: 0.0443 - accuracy: 0.9847 - val_loss: 0.0124 - val_accuracy: 0.9855\n",
      "Epoch 24/25\n",
      "143/143 [==============================] - 311s 2s/step - loss: 0.0392 - accuracy: 0.9865 - val_loss: 0.3706 - val_accuracy: 0.9673\n",
      "Epoch 25/25\n",
      "143/143 [==============================] - 305s 2s/step - loss: 0.0421 - accuracy: 0.9853 - val_loss: 0.0184 - val_accuracy: 0.9796\n"
     ]
    }
   ],
   "source": [
    "TRAINING_NUM = train_data.n #or train_data.samples\n",
    "VALID_NUM = valid_data.n\n",
    "EPOCHS = 25\n",
    "\n",
    "STEP_SIZE_TRAIN = TRAINING_NUM // TRAIN_BATCH_SIZE \n",
    "STEP_SIZE_VALID = VALID_NUM // VALID_BATCH_SIZE\n",
    "\n",
    "# Fit model to get reasonable weights for newly added layers.\n",
    "history = model.fit_generator(generator = train_data,\n",
    "                             steps_per_epoch = STEP_SIZE_TRAIN,\n",
    "                             validation_data = valid_data,\n",
    "                             validation_steps = STEP_SIZE_VALID,\n",
    "                             epochs = EPOCHS)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[INFO] Best Validation Accuracy: % 98.55040907859802\n"
     ]
    }
   ],
   "source": [
    "best_val_acc = max(history.history['val_accuracy'])\n",
    "print(\"[INFO] Best Validation Accuracy: %\",best_val_acc*100)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x576 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "\n",
    "plt.figure( figsize = (15,8)) \n",
    "    \n",
    "plt.subplot(221)  \n",
    "# Accuracy \n",
    "plt.plot(model.history.history['accuracy'])\n",
    "plt.plot(model.history.history['val_accuracy'])\n",
    "plt.title('Model Accuracy')\n",
    "plt.ylabel('Accuracy')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Training set', 'Validation set'], loc='upper left')\n",
    "plt.show()\n",
    "\n",
    "\n",
    "# Loss\n",
    "plt.figure( figsize = (15,8)) \n",
    "plt.subplot(222)  \n",
    "plt.plot(model.history.history['loss'])\n",
    "plt.plot(model.history.history['val_loss'])\n",
    "plt.title('Model Loss')\n",
    "plt.ylabel('Loss')\n",
    "plt.xlabel('Epoch')\n",
    "plt.legend(['Training set', 'Validation set'], loc='upper left')\n",
    "plt.show()"
   ]
  }
 ],
 "metadata": {
  "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": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}