train.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sys\n",
    "import os\n",
    "from os import listdir\n",
    "import random\n",
    "import shutil\n",
    "from matplotlib import pyplot\n",
    "from matplotlib.image import imread\n",
    "import numpy as np\n",
    "from numpy import asarray\n",
    "from numpy import save\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from keras.preprocessing.image import ImageDataGenerator\n",
    "from keras.applications import densenet, inception_v3, mobilenet_v2\n",
    "from keras.preprocessing.image import load_img\n",
    "from keras.preprocessing.image import img_to_array\n",
    "from keras.utils import to_categorical\n",
    "from keras.models import Sequential, Model, load_model\n",
    "from keras.layers import Conv2D, MaxPooling2D\n",
    "from keras.layers import Activation, Dropout, Flatten, Dense\n",
    "from keras.optimizers import SGD\n",
    "from keras.callbacks import EarlyStopping, ReduceLROnPlateau, ModelCheckpoint, Callback\n",
    "from keras import regularizers\n",
    "from keras import backend as K"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "K.set_learning_phase(1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "img_width, img_height = 224, 224\n",
    "epochs = 10\n",
    "batch_size = 30"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Found 30 images belonging to 2 classes.\n",
      "Found 20 images belonging to 2 classes.\n"
     ]
    }
   ],
   "source": [
    "datasetfolder = '/home/parkar-sl/Desktop/Deep/TF2/datasets/covid19/'\n",
    "\n",
    "# create data generator\n",
    "# This involves first defining an instance of the ImageDataGenerator\n",
    "# that will scale the pixel values to the range of 0-1.\n",
    "datagen = ImageDataGenerator(rescale=1.0/255.0)\n",
    "\n",
    "# prepare iterators\n",
    "train_it = datagen.flow_from_directory(datasetfolder + 'train/',\n",
    "    class_mode='categorical', batch_size=batch_size, target_size=(224, 224))\n",
    "test_it = datagen.flow_from_directory(datasetfolder + 'val/',\n",
    "    class_mode='categorical', batch_size=batch_size, target_size=(224, 224))\n",
    "\n",
    "n_classes = train_it.num_classes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "def build_model():\n",
    "    base_model = densenet.DenseNet121(input_shape=(224, 224, 3),\n",
    "                      weights='imagenet',\n",
    "                      include_top=False,\n",
    "                      pooling='avg')\n",
    "    for layer in base_model.layers:\n",
    "        layer.trainable = True\n",
    "\n",
    "    x = base_model.output\n",
    "    predictions = Dense(n_classes, activation='softmax')(x)\n",
    "    model = Model(inputs=base_model.input, outputs=predictions)\n",
    "\n",
    "    return model\n",
    "\n",
    "def build_inceptionv3():\n",
    "    base_model = inception_v3.InceptionV3(input_shape=(224, 224, 3),\n",
    "                      weights='imagenet',\n",
    "                      include_top=False,\n",
    "                      pooling='avg')\n",
    "    for layer in base_model.layers:\n",
    "        layer.trainable = True\n",
    "        \n",
    "    x = base_model.output\n",
    "    predictions = Dense(n_classes, activation='softmax')(x)\n",
    "    model = Model(inputs=base_model.input, outputs=predictions)\n",
    "\n",
    "    return model\n",
    "\n",
    "def build_mobilenetv2():\n",
    "    base_model = mobilenet_v2.MobileNetV2(input_shape=(224, 224, 3),\n",
    "                      weights='imagenet',\n",
    "                      include_top=False,\n",
    "                      pooling='avg')\n",
    "    for layer in base_model.layers:\n",
    "        layer.trainable = True\n",
    "\n",
    "    x = base_model.output\n",
    "    predictions = Dense(n_classes, activation='softmax')(x)\n",
    "    model = Model(inputs=base_model.input, outputs=predictions)\n",
    "\n",
    "    return model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "model = build_model()\n",
    "inceptionv3model = build_inceptionv3()\n",
    "mobilenetv2model = build_mobilenetv2()\n",
    "\n",
    "sgd = SGD(lr=0.01, decay=1e-6, momentum=0.9)\n",
    "model.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['acc', 'mse'])\n",
    "inceptionv3model.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['acc', 'mse'])\n",
    "mobilenetv2model.compile(loss='categorical_crossentropy', optimizer=sgd, metrics=['acc', 'mse'])\n",
    "\n",
    "\n",
    "early_stop = EarlyStopping(monitor='val_loss', patience=8, verbose=1, min_delta=1e-4)\n",
    "reduce_lr = ReduceLROnPlateau(monitor='val_loss', factor=0.1, patience=4, verbose=1, min_delta=1e-4)\n",
    "filepathdn = \"covid19_densenet_{epoch:02d}_{val_acc:.2f}.hdf5\"\n",
    "filepathin = \"covid19_inception_{epoch:02d}_{val_acc:.2f}.hdf5\"\n",
    "filepathmn = \"covid19_mobilenet_{epoch:02d}_{val_acc:.2f}.hdf5\"\n",
    "\n",
    "checkpointdn = ModelCheckpoint(filepathdn, monitor='val_acc', verbose=1, save_best_only=True, mode='max')\n",
    "checkpointin = ModelCheckpoint(filepathin, monitor='val_acc', verbose=1, save_best_only=True, mode='max')\n",
    "checkpointmn = ModelCheckpoint(filepathmn, monitor='val_acc', verbose=1, save_best_only=True, mode='max')\n",
    "\n",
    "callbacks_list_dn = [early_stop, reduce_lr, checkpointdn]\n",
    "callbacks_list_in = [early_stop, reduce_lr, checkpointin]\n",
    "callbacks_list_mn = [early_stop, reduce_lr, checkpointmn]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "1/1 [==============================] - 26s 26s/step - loss: 0.7021 - acc: 0.5667 - mse: 0.2520 - val_loss: 0.5738 - val_acc: 0.7500 - val_mse: 0.1959\n",
      "\n",
      "Epoch 00001: val_acc improved from -inf to 0.75000, saving model to covid19_densenet_01_0.75.hdf5\n",
      "Epoch 2/10\n",
      "1/1 [==============================] - 8s 8s/step - loss: 0.2090 - acc: 1.0000 - mse: 0.0433 - val_loss: 0.3462 - val_acc: 0.9000 - val_mse: 0.0994\n",
      "\n",
      "Epoch 00002: val_acc improved from 0.75000 to 0.90000, saving model to covid19_densenet_02_0.90.hdf5\n",
      "Epoch 3/10\n",
      "1/1 [==============================] - 8s 8s/step - loss: 0.0569 - acc: 1.0000 - mse: 0.0055 - val_loss: 0.2206 - val_acc: 0.9500 - val_mse: 0.0529\n",
      "\n",
      "Epoch 00003: val_acc improved from 0.90000 to 0.95000, saving model to covid19_densenet_03_0.95.hdf5\n",
      "Epoch 4/10\n",
      "1/1 [==============================] - 8s 8s/step - loss: 0.0206 - acc: 1.0000 - mse: 8.3422e-04 - val_loss: 0.1727 - val_acc: 0.9500 - val_mse: 0.0416\n",
      "\n",
      "Epoch 00004: val_acc did not improve from 0.95000\n",
      "Epoch 5/10\n",
      "1/1 [==============================] - 8s 8s/step - loss: 0.0090 - acc: 1.0000 - mse: 1.4333e-04 - val_loss: 0.1465 - val_acc: 0.9500 - val_mse: 0.0369\n",
      "\n",
      "Epoch 00005: val_acc did not improve from 0.95000\n",
      "Epoch 6/10\n",
      "1/1 [==============================] - 8s 8s/step - loss: 0.0048 - acc: 1.0000 - mse: 4.0102e-05 - val_loss: 0.1323 - val_acc: 0.9500 - val_mse: 0.0356\n",
      "\n",
      "Epoch 00006: val_acc did not improve from 0.95000\n",
      "Epoch 7/10\n",
      "1/1 [==============================] - 8s 8s/step - loss: 0.0031 - acc: 1.0000 - mse: 1.6760e-05 - val_loss: 0.1209 - val_acc: 0.9500 - val_mse: 0.0340\n",
      "\n",
      "Epoch 00007: val_acc did not improve from 0.95000\n",
      "Epoch 8/10\n",
      "1/1 [==============================] - 8s 8s/step - loss: 0.0022 - acc: 1.0000 - mse: 9.3993e-06 - val_loss: 0.1151 - val_acc: 0.9500 - val_mse: 0.0336\n",
      "\n",
      "Epoch 00008: val_acc did not improve from 0.95000\n",
      "Epoch 9/10\n",
      "1/1 [==============================] - 8s 8s/step - loss: 0.0017 - acc: 1.0000 - mse: 5.9098e-06 - val_loss: 0.1115 - val_acc: 0.9500 - val_mse: 0.0336\n",
      "\n",
      "Epoch 00009: val_acc did not improve from 0.95000\n",
      "Epoch 10/10\n",
      "1/1 [==============================] - 8s 8s/step - loss: 0.0013 - acc: 1.0000 - mse: 4.0766e-06 - val_loss: 0.1078 - val_acc: 0.9500 - val_mse: 0.0331\n",
      "\n",
      "Epoch 00010: val_acc did not improve from 0.95000\n"
     ]
    }
   ],
   "source": [
    "model_history = model.fit_generator(train_it,\n",
    "                                    epochs=epochs,\n",
    "                                    validation_data=test_it,\n",
    "                                    validation_steps=test_it.samples // batch_size,\n",
    "                                    callbacks=callbacks_list_dn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "1/1 [==============================] - 13s 13s/step - loss: 0.6941 - acc: 0.4667 - mse: 0.2491 - val_loss: 0.6605 - val_acc: 0.5500 - val_mse: 0.2351\n",
      "\n",
      "Epoch 00001: val_acc improved from -inf to 0.55000, saving model to covid19_inception_01_0.55.hdf5\n",
      "Epoch 2/10\n",
      "1/1 [==============================] - 3s 3s/step - loss: 0.4668 - acc: 0.8333 - mse: 0.1432 - val_loss: 0.5840 - val_acc: 0.6000 - val_mse: 0.2015\n",
      "\n",
      "Epoch 00002: val_acc improved from 0.55000 to 0.60000, saving model to covid19_inception_02_0.60.hdf5\n",
      "Epoch 3/10\n",
      "1/1 [==============================] - 3s 3s/step - loss: 0.1574 - acc: 1.0000 - mse: 0.0249 - val_loss: 0.4463 - val_acc: 0.8500 - val_mse: 0.1414\n",
      "\n",
      "Epoch 00003: val_acc improved from 0.60000 to 0.85000, saving model to covid19_inception_03_0.85.hdf5\n",
      "Epoch 4/10\n",
      "1/1 [==============================] - 3s 3s/step - loss: 0.0608 - acc: 1.0000 - mse: 0.0045 - val_loss: 0.3494 - val_acc: 0.8500 - val_mse: 0.1025\n",
      "\n",
      "Epoch 00004: val_acc did not improve from 0.85000\n",
      "Epoch 5/10\n",
      "1/1 [==============================] - 3s 3s/step - loss: 0.0287 - acc: 1.0000 - mse: 0.0010 - val_loss: 0.2811 - val_acc: 0.8500 - val_mse: 0.0765\n",
      "\n",
      "Epoch 00005: val_acc did not improve from 0.85000\n",
      "Epoch 6/10\n",
      "1/1 [==============================] - 3s 3s/step - loss: 0.0161 - acc: 1.0000 - mse: 3.3191e-04 - val_loss: 0.2391 - val_acc: 0.9500 - val_mse: 0.0619\n",
      "\n",
      "Epoch 00006: val_acc improved from 0.85000 to 0.95000, saving model to covid19_inception_06_0.95.hdf5\n",
      "Epoch 7/10\n",
      "1/1 [==============================] - 3s 3s/step - loss: 0.0096 - acc: 1.0000 - mse: 1.2390e-04 - val_loss: 0.1970 - val_acc: 0.9500 - val_mse: 0.0484\n",
      "\n",
      "Epoch 00007: val_acc did not improve from 0.95000\n",
      "Epoch 8/10\n",
      "1/1 [==============================] - 3s 3s/step - loss: 0.0059 - acc: 1.0000 - mse: 4.9386e-05 - val_loss: 0.1670 - val_acc: 0.9500 - val_mse: 0.0403\n",
      "\n",
      "Epoch 00008: val_acc did not improve from 0.95000\n",
      "Epoch 9/10\n",
      "1/1 [==============================] - 3s 3s/step - loss: 0.0038 - acc: 1.0000 - mse: 2.1568e-05 - val_loss: 0.1420 - val_acc: 1.0000 - val_mse: 0.0328\n",
      "\n",
      "Epoch 00009: val_acc improved from 0.95000 to 1.00000, saving model to covid19_inception_09_1.00.hdf5\n",
      "Epoch 10/10\n",
      "1/1 [==============================] - 3s 3s/step - loss: 0.0026 - acc: 1.0000 - mse: 1.0151e-05 - val_loss: 0.1228 - val_acc: 1.0000 - val_mse: 0.0277\n",
      "\n",
      "Epoch 00010: val_acc did not improve from 1.00000\n"
     ]
    }
   ],
   "source": [
    "inceptionv3model_history = inceptionv3model.fit_generator(train_it,\n",
    "                                    epochs=epochs,\n",
    "                                    validation_data=test_it,\n",
    "                                    validation_steps=test_it.samples // batch_size,\n",
    "                                    callbacks=callbacks_list_in)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/10\n",
      "1/1 [==============================] - 8s 8s/step - loss: 0.7079 - acc: 0.5667 - mse: 0.2551 - val_loss: 0.5295 - val_acc: 0.7500 - val_mse: 0.1791\n",
      "\n",
      "Epoch 00001: val_acc improved from -inf to 0.75000, saving model to covid19_mobilenet_01_0.75.hdf5\n",
      "Epoch 2/10\n",
      "1/1 [==============================] - 3s 3s/step - loss: 0.4114 - acc: 0.9333 - mse: 0.1230 - val_loss: 0.3905 - val_acc: 0.8500 - val_mse: 0.1211\n",
      "\n",
      "Epoch 00002: val_acc improved from 0.75000 to 0.85000, saving model to covid19_mobilenet_02_0.85.hdf5\n",
      "Epoch 3/10\n",
      "1/1 [==============================] - 3s 3s/step - loss: 0.0929 - acc: 1.0000 - mse: 0.0128 - val_loss: 0.3173 - val_acc: 0.8500 - val_mse: 0.0974\n",
      "\n",
      "Epoch 00003: val_acc did not improve from 0.85000\n",
      "Epoch 4/10\n",
      "1/1 [==============================] - 3s 3s/step - loss: 0.0244 - acc: 1.0000 - mse: 0.0011 - val_loss: 0.2993 - val_acc: 0.9000 - val_mse: 0.0930\n",
      "\n",
      "Epoch 00004: val_acc improved from 0.85000 to 0.90000, saving model to covid19_mobilenet_04_0.90.hdf5\n",
      "Epoch 5/10\n",
      "1/1 [==============================] - 3s 3s/step - loss: 0.0098 - acc: 1.0000 - mse: 1.8340e-04 - val_loss: 0.2853 - val_acc: 0.9500 - val_mse: 0.0888\n",
      "\n",
      "Epoch 00005: val_acc improved from 0.90000 to 0.95000, saving model to covid19_mobilenet_05_0.95.hdf5\n",
      "Epoch 6/10\n",
      "1/1 [==============================] - 3s 3s/step - loss: 0.0054 - acc: 1.0000 - mse: 6.0727e-05 - val_loss: 0.2780 - val_acc: 0.9500 - val_mse: 0.0863\n",
      "\n",
      "Epoch 00006: val_acc did not improve from 0.95000\n",
      "Epoch 7/10\n"
     ]
    }
   ],
   "source": [
    "mobilenetv2model_history2 = mobilenetv2model.fit_generator(train_it,\n",
    "                                    epochs=epochs,\n",
    "                                    validation_data=test_it,\n",
    "                                    validation_steps=test_it.samples // batch_size,\n",
    "                                    callbacks=callbacks_list_mn)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "print(\"model history: \", model_history.history.keys())\n",
    "\n",
    "plt.figure(0)\n",
    "plt.plot(model_history.history['acc'], 'r')\n",
    "plt.plot(model_history.history['val_acc'], 'g')\n",
    "plt.xticks(np.arange(0, 20, 1.0))\n",
    "plt.rcParams['figure.figsize'] = (8, 6)\n",
    "plt.xlabel(\"Num of Epochs\")\n",
    "plt.ylabel(\"Accuracy\")\n",
    "plt.title(\"Training Accuracy vs Validation Accuracy\")\n",
    "plt.legend(['train', 'validation'])\n",
    "\n",
    "plt.figure(1)\n",
    "plt.plot(model_history.history['loss'], 'r')\n",
    "plt.plot(model_history.history['val_loss'], 'g')\n",
    "plt.xticks(np.arange(0, 20, 1.0))\n",
    "plt.rcParams['figure.figsize'] = (8, 6)\n",
    "plt.xlabel(\"Num of Epochs\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.title(\"Training Loss vs Validation Loss\")\n",
    "plt.legend(['train', 'validation'])\n",
    "\n",
    "plt.figure(2)\n",
    "plt.plot(model_history.history['mse'], 'r')\n",
    "plt.plot(model_history.history['val_mse'], 'g')\n",
    "plt.xticks(np.arange(0, 20, 1.0))\n",
    "plt.rcParams['figure.figsize'] = (8, 6)\n",
    "plt.xlabel(\"Num of Epochs\")\n",
    "plt.ylabel(\"MSE\")\n",
    "plt.title(\"Training Loss vs Validation Loss\")\n",
    "plt.legend(['train', 'validation'])\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(\"model history: \", inceptionv3model_history.history.keys())\n",
    "\n",
    "plt.figure(0)\n",
    "plt.plot(inceptionv3model_history.history['acc'], 'r')\n",
    "plt.plot(inceptionv3model_history.history['val_acc'], 'g')\n",
    "plt.xticks(np.arange(0, 10, 1.0))\n",
    "plt.rcParams['figure.figsize'] = (8, 6)\n",
    "plt.xlabel(\"Num of Epochs\")\n",
    "plt.ylabel(\"Accuracy\")\n",
    "plt.title(\"Training Accuracy vs Validation Accuracy\")\n",
    "plt.legend(['train', 'validation'])\n",
    "\n",
    "plt.figure(1)\n",
    "plt.plot(inceptionv3model_history.history['loss'], 'r')\n",
    "plt.plot(inceptionv3model_history.history['val_loss'], 'g')\n",
    "plt.xticks(np.arange(0, 10, 1.0))\n",
    "plt.rcParams['figure.figsize'] = (8, 6)\n",
    "plt.xlabel(\"Num of Epochs\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.title(\"Training Loss vs Validation Loss\")\n",
    "plt.legend(['train', 'validation'])\n",
    "\n",
    "plt.figure(2)\n",
    "plt.plot(inceptionv3model_history.history['mse'], 'r')\n",
    "plt.plot(inceptionv3model_history.history['val_mse'], 'g')\n",
    "plt.xticks(np.arange(0, 10, 1.0))\n",
    "plt.rcParams['figure.figsize'] = (8, 6)\n",
    "plt.xlabel(\"Num of Epochs\")\n",
    "plt.ylabel(\"MSE\")\n",
    "plt.title(\"Training Loss vs Validation Loss\")\n",
    "plt.legend(['train', 'validation'])\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(\"model history: \", mobilenetv2model_history2.history.keys())\n",
    "\n",
    "plt.figure(0)\n",
    "plt.plot(mobilenetv2model_history.history['acc'], 'r')\n",
    "plt.plot(mobilenetv2model_history.history['val_acc'], 'g')\n",
    "plt.xticks(np.arange(0, 10, 1.0))\n",
    "plt.rcParams['figure.figsize'] = (8, 6)\n",
    "plt.xlabel(\"Num of Epochs\")\n",
    "plt.ylabel(\"Accuracy\")\n",
    "plt.title(\"Training Accuracy vs Validation Accuracy\")\n",
    "plt.legend(['train', 'validation'])\n",
    "\n",
    "plt.figure(1)\n",
    "plt.plot(mobilenetv2model_history.history['loss'], 'r')\n",
    "plt.plot(mobilenetv2model_history.history['val_loss'], 'g')\n",
    "plt.xticks(np.arange(0, 10, 1.0))\n",
    "plt.rcParams['figure.figsize'] = (8, 6)\n",
    "plt.xlabel(\"Num of Epochs\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.title(\"Training Loss vs Validation Loss\")\n",
    "plt.legend(['train', 'validation'])\n",
    "\n",
    "plt.figure(2)\n",
    "plt.plot(mobilenetv2model_history.history['mse'], 'r')\n",
    "plt.plot(mobilenetv2model_history.history['val_mse'], 'g')\n",
    "plt.xticks(np.arange(0, 10, 1.0))\n",
    "plt.rcParams['figure.figsize'] = (8, 6)\n",
    "plt.xlabel(\"Num of Epochs\")\n",
    "plt.ylabel(\"MSE\")\n",
    "plt.title(\"Training Loss vs Validation Loss\")\n",
    "plt.legend(['train', 'validation'])\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "import yagmail"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "print(\"model history: \", model_history.history.keys())\n",
    "\n",
    "plt.figure(0)\n",
    "plt.plot(model_history.history['acc'], 'r')\n",
    "plt.plot(model_history.history['val_acc'], 'g')\n",
    "plt.xticks(np.arange(0, 16, 1.0))\n",
    "plt.rcParams['figure.figsize'] = (8, 6)\n",
    "plt.xlabel(\"Num of Epochs\")\n",
    "plt.ylabel(\"Accuracy\")\n",
    "plt.title(\"Training Accuracy vs Validation Accuracy\")\n",
    "plt.legend(['train', 'validation'])\n",
    "plt.savefig('acc.png')\n",
    "\n",
    "plt.figure(1)\n",
    "plt.plot(model_history.history['loss'], 'r')\n",
    "plt.plot(model_history.history['val_loss'], 'g')\n",
    "plt.xticks(np.arange(0, 16, 1.0))\n",
    "plt.rcParams['figure.figsize'] = (8, 6)\n",
    "plt.xlabel(\"Num of Epochs\")\n",
    "plt.ylabel(\"Loss\")\n",
    "plt.title(\"Training Loss vs Validation Loss\")\n",
    "plt.legend(['train', 'validation'])\n",
    "plt.savefig('loss.png')\n",
    "\n",
    "plt.figure(2)\n",
    "plt.plot(model_history.history['mse'], 'r')\n",
    "plt.plot(model_history.history['val_mse'], 'g')\n",
    "plt.xticks(np.arange(0, 16, 1.0))\n",
    "plt.rcParams['figure.figsize'] = (8, 6)\n",
    "plt.xlabel(\"Num of Epochs\")\n",
    "plt.ylabel(\"MSE\")\n",
    "plt.title(\"Training MSE vs Validation MSE\")\n",
    "plt.legend(['train', 'validation'])\n",
    "plt.savefig('mse.png')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "yag = yagmail.SMTP(\"shblrb\", \"@SaincTelette70\")\n",
    "contents = ['Hello,\\n', '\\nTraining has ended successfully! Here are the results:', \n",
    "            str(model_history.history), '\\nBest regards, \\nYour Laptop', \n",
    "            yagmail.inline(\"acc.png\"),\n",
    "            yagmail.inline(\"loss.png\"),\n",
    "            yagmail.inline(\"mse.png\")]\n",
    "yag.send('sohaib-l@hotmail.fr', 'Training Ended - NTU49_revised', contents)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "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": "3.6.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}