FED_Train_Validation.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "cea65667",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "5d1bc629",
   "metadata": {},
   "source": [
    "# Getting Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "a49c3924",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>emotion</th>\n",
       "      <th>pixels</th>\n",
       "      <th>Usage</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>70 80 82 72 58 58 60 63 54 58 60 48 89 115 121...</td>\n",
       "      <td>Training</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>151 150 147 155 148 133 111 140 170 174 182 15...</td>\n",
       "      <td>Training</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>231 212 156 164 174 138 161 173 182 200 106 38...</td>\n",
       "      <td>Training</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>24 32 36 30 32 23 19 20 30 41 21 22 32 34 21 1...</td>\n",
       "      <td>Training</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>6</td>\n",
       "      <td>4 0 0 0 0 0 0 0 0 0 0 0 3 15 23 28 48 50 58 84...</td>\n",
       "      <td>Training</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   emotion                                             pixels     Usage\n",
       "0        0  70 80 82 72 58 58 60 63 54 58 60 48 89 115 121...  Training\n",
       "1        0  151 150 147 155 148 133 111 140 170 174 182 15...  Training\n",
       "2        2  231 212 156 164 174 138 161 173 182 200 106 38...  Training\n",
       "3        4  24 32 36 30 32 23 19 20 30 41 21 22 32 34 21 1...  Training\n",
       "4        6  4 0 0 0 0 0 0 0 0 0 0 0 3 15 23 28 48 50 58 84...  Training"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import scipy\n",
    "import pandas as pd\n",
    "df = pd.read_csv('./fer2013/fer2013.csv')\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "0bcc96c1",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 35887 entries, 0 to 35886\n",
      "Data columns (total 3 columns):\n",
      " #   Column   Non-Null Count  Dtype \n",
      "---  ------   --------------  ----- \n",
      " 0   emotion  35887 non-null  int64 \n",
      " 1   pixels   35887 non-null  object\n",
      " 2   Usage    35887 non-null  object\n",
      "dtypes: int64(1), object(2)\n",
      "memory usage: 841.2+ KB\n"
     ]
    }
   ],
   "source": [
    "df.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "41eb0c05",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Training       28709\n",
       "PublicTest      3589\n",
       "PrivateTest     3589\n",
       "Name: Usage, dtype: int64"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "num_classes = 7\n",
    "width = 48\n",
    "height = 48\n",
    "emotion_labels = [\"Angry\", \"Disgust\", \"Fear\", \"Happy\", \"Sad\", \"Surprise\", \"Neutral\"]\n",
    "classes = np.array((\"Angry\", \"Disgust\", \"Fear\", \"Happy\", \"Sad\", \"Surprise\", \"Neutral\"))\n",
    "df.Usage.value_counts() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "ed17d859",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(48, 48)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "k = np.array(list(map(int,df.iloc[0,1].split(\" \"))),dtype='uint8').reshape((48,48))\n",
    "k.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "138140c7",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "11579973",
   "metadata": {},
   "source": [
    "# Preparing Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "eb7e2ead",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = []\n",
    "y_train = []\n",
    "X_test = []\n",
    "y_test = []\n",
    "for index, row in df.iterrows():\n",
    "    k = row['pixels'].split(\" \")\n",
    "    if row['Usage'] == 'Training':\n",
    "        X_train.append(np.array(k))\n",
    "        y_train.append(row['emotion'])\n",
    "    elif row['Usage'] == 'PublicTest':\n",
    "        X_test.append(np.array(k))\n",
    "        y_test.append(row['emotion'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "3321202f",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['70', '80', '82', ..., '106', '109', '82'], dtype='<U3')"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train[0] #consists pixel’s number is in the form of a string, converting it into numbers is easy by typecasting."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "36247294",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x1eded247a90>"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.imshow(np.array(X_train[0], dtype = 'uint8').reshape(48,48,1), cmap = 'gray') #plotting 48x48 image from pixels in selected sample"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "ccc41bac",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train = np.array(X_train, dtype = 'uint8')\n",
    "y_train = np.array(y_train, dtype = 'uint8')\n",
    "X_test = np.array(X_test, dtype = 'uint8')\n",
    "y_test = np.array(y_test, dtype = 'uint8')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "66516447",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(28709, 48, 48, 1)"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "X_train = X_train.reshape(X_train.shape[0], 48, 48, 1)\n",
    "X_test = X_test.reshape(X_test.shape[0], 48, 48, 1)\n",
    "X_train.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "e73c667c",
   "metadata": {},
   "outputs": [],
   "source": [
    "import tensorflow.keras\n",
    "from tensorflow.keras.utils import to_categorical\n",
    "y_train= to_categorical(y_train, num_classes=7)\n",
    "y_test = to_categorical(y_test, num_classes=7)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bc34fc7b",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "e3bc0630",
   "metadata": {},
   "source": [
    "# Image Augmentation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "e7bd1107",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.preprocessing.image import ImageDataGenerator \n",
    "datagen = ImageDataGenerator( \n",
    "    rescale=1./255,\n",
    "    rotation_range = 10,\n",
    "    horizontal_flip = True,\n",
    "    width_shift_range=0.1,\n",
    "    height_shift_range=0.1,\n",
    "    fill_mode = 'nearest')\n",
    "\n",
    "testgen = ImageDataGenerator( \n",
    "    rescale=1./255\n",
    "    )\n",
    "datagen.fit(X_train)\n",
    "batch_size = 64 "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "331aac55",
   "metadata": {},
   "source": [
    "## Fitting the generator to data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "b4b36d20",
   "metadata": {},
   "outputs": [],
   "source": [
    "train_flow = datagen.flow(X_train, y_train, batch_size=batch_size) \n",
    "test_flow = testgen.flow(X_test, y_test, batch_size=batch_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "43a0be53",
   "metadata": {},
   "source": [
    "## Plotting random images of emotions after normalizing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "427fcd3e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 9 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "for X_batch, y_batch in datagen.flow(X_train, y_train, batch_size=9):\n",
    "    for i in range(0, 9): \n",
    "        plt.axis('off') \n",
    "        plt.subplot(330 + 1 + i) \n",
    "        plt.imshow(X_batch[i].reshape(48, 48), cmap=plt.get_cmap('gray'))\n",
    "    plt.axis('off') \n",
    "    plt.show() \n",
    "    break"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f84c629c",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "4e3913b7",
   "metadata": {},
   "source": [
    "# Building Model using CNN"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "944eb84a",
   "metadata": {},
   "outputs": [],
   "source": [
    "from tensorflow.keras.utils import plot_model\n",
    "from tensorflow.keras.models import Model\n",
    "from tensorflow.keras.layers import Input, Dense, Flatten, Dropout, BatchNormalization\n",
    "from tensorflow.keras.layers import Conv2D\n",
    "from tensorflow.keras.layers import MaxPooling2D\n",
    "from tensorflow.keras.layers import concatenate\n",
    "from tensorflow.keras.optimizers import Adam, SGD\n",
    "from tensorflow.keras.regularizers import l1, l2\n",
    "from matplotlib import pyplot as plt\n",
    "from sklearn.metrics import confusion_matrix"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "1caa5f69",
   "metadata": {},
   "outputs": [],
   "source": [
    "def FER_Model(input_shape=(48,48,1)):\n",
    "    #first input model\n",
    "    visible = Input(shape=input_shape, name='input')\n",
    "    \n",
    "    num_classes = 7\n",
    "    #1-st block\n",
    "    conv1_1 = Conv2D(64, kernel_size=3, activation='relu', padding='same', name = 'conv1_1')(visible)\n",
    "    conv1_1 = BatchNormalization()(conv1_1)\n",
    "    conv1_2 = Conv2D(64, kernel_size=3, activation='relu', padding='same', name = 'conv1_2')(conv1_1)\n",
    "    conv1_2 = BatchNormalization()(conv1_2)\n",
    "    pool1_1 = MaxPooling2D(pool_size=(2,2), name = 'pool1_1')(conv1_2)\n",
    "    drop1_1 = Dropout(0.3, name = 'drop1_1')(pool1_1)#the 2-nd block\n",
    "    conv2_1 = Conv2D(128, kernel_size=3, activation='relu', padding='same', name = 'conv2_1')(drop1_1)\n",
    "    conv2_1 = BatchNormalization()(conv2_1)\n",
    "    conv2_2 = Conv2D(128, kernel_size=3, activation='relu', padding='same', name = 'conv2_2')(conv2_1)\n",
    "    conv2_2 = BatchNormalization()(conv2_2)\n",
    "    conv2_3 = Conv2D(128, kernel_size=3, activation='relu', padding='same', name = 'conv2_3')(conv2_2)\n",
    "    conv2_2 = BatchNormalization()(conv2_3)\n",
    "    pool2_1 = MaxPooling2D(pool_size=(2,2), name = 'pool2_1')(conv2_3)\n",
    "    drop2_1 = Dropout(0.3, name = 'drop2_1')(pool2_1)#the 3-rd block\n",
    "    conv3_1 = Conv2D(256, kernel_size=3, activation='relu', padding='same', name = 'conv3_1')(drop2_1)\n",
    "    conv3_1 = BatchNormalization()(conv3_1)\n",
    "    conv3_2 = Conv2D(256, kernel_size=3, activation='relu', padding='same', name = 'conv3_2')(conv3_1)\n",
    "    conv3_2 = BatchNormalization()(conv3_2)\n",
    "    conv3_3 = Conv2D(256, kernel_size=3, activation='relu', padding='same', name = 'conv3_3')(conv3_2)\n",
    "    conv3_3 = BatchNormalization()(conv3_3)\n",
    "    conv3_4 = Conv2D(256, kernel_size=3, activation='relu', padding='same', name = 'conv3_4')(conv3_3)\n",
    "    conv3_4 = BatchNormalization()(conv3_4)\n",
    "    pool3_1 = MaxPooling2D(pool_size=(2,2), name = 'pool3_1')(conv3_4)\n",
    "    drop3_1 = Dropout(0.3, name = 'drop3_1')(pool3_1)#the 4-th block\n",
    "    conv4_1 = Conv2D(256, kernel_size=3, activation='relu', padding='same', name = 'conv4_1')(drop3_1)\n",
    "    conv4_1 = BatchNormalization()(conv4_1)\n",
    "    conv4_2 = Conv2D(256, kernel_size=3, activation='relu', padding='same', name = 'conv4_2')(conv4_1)\n",
    "    conv4_2 = BatchNormalization()(conv4_2)\n",
    "    conv4_3 = Conv2D(256, kernel_size=3, activation='relu', padding='same', name = 'conv4_3')(conv4_2)\n",
    "    conv4_3 = BatchNormalization()(conv4_3)\n",
    "    conv4_4 = Conv2D(256, kernel_size=3, activation='relu', padding='same', name = 'conv4_4')(conv4_3)\n",
    "    conv4_4 = BatchNormalization()(conv4_4)\n",
    "    pool4_1 = MaxPooling2D(pool_size=(2,2), name = 'pool4_1')(conv4_4)\n",
    "    drop4_1 = Dropout(0.3, name = 'drop4_1')(pool4_1)\n",
    "    \n",
    "    #5-th block\n",
    "    conv5_1 = Conv2D(512, kernel_size=3, activation='relu', padding='same', name = 'conv5_1')(drop4_1)\n",
    "    conv5_1 = BatchNormalization()(conv5_1)\n",
    "    conv5_2 = Conv2D(512, kernel_size=3, activation='relu', padding='same', name = 'conv5_2')(conv5_1)\n",
    "    conv5_2 = BatchNormalization()(conv5_2)\n",
    "    conv5_3 = Conv2D(512, kernel_size=3, activation='relu', padding='same', name = 'conv5_3')(conv5_2)\n",
    "    conv5_3 = BatchNormalization()(conv5_3)\n",
    "    conv5_4 = Conv2D(512, kernel_size=3, activation='relu', padding='same', name = 'conv5_4')(conv5_3)\n",
    "    conv5_3 = BatchNormalization()(conv5_3)\n",
    "    pool5_1 = MaxPooling2D(pool_size=(2,2), name = 'pool5_1')(conv5_4)\n",
    "    drop5_1 = Dropout(0.3, name = 'drop5_1')(pool5_1)#Flatten and output\n",
    "    flatten = Flatten(name = 'flatten')(drop5_1)\n",
    "    ouput = Dense(num_classes, activation='softmax', name = 'output')(flatten)# create model \n",
    "    model = Model(inputs =visible, outputs = ouput)\n",
    "    \n",
    "    #summary layers\n",
    "    print(model.summary())\n",
    "    \n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2692a05a",
   "metadata": {},
   "source": [
    "## Compiling the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "ebb7c833",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"model_1\"\n",
      "_________________________________________________________________\n",
      "Layer (type)                 Output Shape              Param #   \n",
      "=================================================================\n",
      "input (InputLayer)           [(None, 48, 48, 1)]       0         \n",
      "_________________________________________________________________\n",
      "conv1_1 (Conv2D)             (None, 48, 48, 64)        640       \n",
      "_________________________________________________________________\n",
      "batch_normalization_17 (Batc (None, 48, 48, 64)        256       \n",
      "_________________________________________________________________\n",
      "conv1_2 (Conv2D)             (None, 48, 48, 64)        36928     \n",
      "_________________________________________________________________\n",
      "batch_normalization_18 (Batc (None, 48, 48, 64)        256       \n",
      "_________________________________________________________________\n",
      "pool1_1 (MaxPooling2D)       (None, 24, 24, 64)        0         \n",
      "_________________________________________________________________\n",
      "drop1_1 (Dropout)            (None, 24, 24, 64)        0         \n",
      "_________________________________________________________________\n",
      "conv2_1 (Conv2D)             (None, 24, 24, 128)       73856     \n",
      "_________________________________________________________________\n",
      "batch_normalization_19 (Batc (None, 24, 24, 128)       512       \n",
      "_________________________________________________________________\n",
      "conv2_2 (Conv2D)             (None, 24, 24, 128)       147584    \n",
      "_________________________________________________________________\n",
      "batch_normalization_20 (Batc (None, 24, 24, 128)       512       \n",
      "_________________________________________________________________\n",
      "conv2_3 (Conv2D)             (None, 24, 24, 128)       147584    \n",
      "_________________________________________________________________\n",
      "pool2_1 (MaxPooling2D)       (None, 12, 12, 128)       0         \n",
      "_________________________________________________________________\n",
      "drop2_1 (Dropout)            (None, 12, 12, 128)       0         \n",
      "_________________________________________________________________\n",
      "conv3_1 (Conv2D)             (None, 12, 12, 256)       295168    \n",
      "_________________________________________________________________\n",
      "batch_normalization_22 (Batc (None, 12, 12, 256)       1024      \n",
      "_________________________________________________________________\n",
      "conv3_2 (Conv2D)             (None, 12, 12, 256)       590080    \n",
      "_________________________________________________________________\n",
      "batch_normalization_23 (Batc (None, 12, 12, 256)       1024      \n",
      "_________________________________________________________________\n",
      "conv3_3 (Conv2D)             (None, 12, 12, 256)       590080    \n",
      "_________________________________________________________________\n",
      "batch_normalization_24 (Batc (None, 12, 12, 256)       1024      \n",
      "_________________________________________________________________\n",
      "conv3_4 (Conv2D)             (None, 12, 12, 256)       590080    \n",
      "_________________________________________________________________\n",
      "batch_normalization_25 (Batc (None, 12, 12, 256)       1024      \n",
      "_________________________________________________________________\n",
      "pool3_1 (MaxPooling2D)       (None, 6, 6, 256)         0         \n",
      "_________________________________________________________________\n",
      "drop3_1 (Dropout)            (None, 6, 6, 256)         0         \n",
      "_________________________________________________________________\n",
      "conv4_1 (Conv2D)             (None, 6, 6, 256)         590080    \n",
      "_________________________________________________________________\n",
      "batch_normalization_26 (Batc (None, 6, 6, 256)         1024      \n",
      "_________________________________________________________________\n",
      "conv4_2 (Conv2D)             (None, 6, 6, 256)         590080    \n",
      "_________________________________________________________________\n",
      "batch_normalization_27 (Batc (None, 6, 6, 256)         1024      \n",
      "_________________________________________________________________\n",
      "conv4_3 (Conv2D)             (None, 6, 6, 256)         590080    \n",
      "_________________________________________________________________\n",
      "batch_normalization_28 (Batc (None, 6, 6, 256)         1024      \n",
      "_________________________________________________________________\n",
      "conv4_4 (Conv2D)             (None, 6, 6, 256)         590080    \n",
      "_________________________________________________________________\n",
      "batch_normalization_29 (Batc (None, 6, 6, 256)         1024      \n",
      "_________________________________________________________________\n",
      "pool4_1 (MaxPooling2D)       (None, 3, 3, 256)         0         \n",
      "_________________________________________________________________\n",
      "drop4_1 (Dropout)            (None, 3, 3, 256)         0         \n",
      "_________________________________________________________________\n",
      "conv5_1 (Conv2D)             (None, 3, 3, 512)         1180160   \n",
      "_________________________________________________________________\n",
      "batch_normalization_30 (Batc (None, 3, 3, 512)         2048      \n",
      "_________________________________________________________________\n",
      "conv5_2 (Conv2D)             (None, 3, 3, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "batch_normalization_31 (Batc (None, 3, 3, 512)         2048      \n",
      "_________________________________________________________________\n",
      "conv5_3 (Conv2D)             (None, 3, 3, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "batch_normalization_32 (Batc (None, 3, 3, 512)         2048      \n",
      "_________________________________________________________________\n",
      "conv5_4 (Conv2D)             (None, 3, 3, 512)         2359808   \n",
      "_________________________________________________________________\n",
      "pool5_1 (MaxPooling2D)       (None, 1, 1, 512)         0         \n",
      "_________________________________________________________________\n",
      "drop5_1 (Dropout)            (None, 1, 1, 512)         0         \n",
      "_________________________________________________________________\n",
      "flatten (Flatten)            (None, 512)               0         \n",
      "_________________________________________________________________\n",
      "output (Dense)               (None, 7)                 3591      \n",
      "=================================================================\n",
      "Total params: 13,111,367\n",
      "Trainable params: 13,103,431\n",
      "Non-trainable params: 7,936\n",
      "_________________________________________________________________\n",
      "None\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\assas\\.conda\\envs\\tf_gpu_env\\lib\\site-packages\\tensorflow\\python\\keras\\optimizer_v2\\optimizer_v2.py:374: UserWarning: The `lr` argument is deprecated, use `learning_rate` instead.\n",
      "  warnings.warn(\n"
     ]
    }
   ],
   "source": [
    "model = FER_Model()\n",
    "opt = Adam(lr=0.0001, decay=1e-6)\n",
    "model.compile(loss='categorical_crossentropy', optimizer=opt, metrics=['accuracy'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c0937324",
   "metadata": {},
   "source": [
    "# Training the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "180cec27",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\assas\\.conda\\envs\\tf_gpu_env\\lib\\site-packages\\tensorflow\\python\\keras\\engine\\training.py:1969: UserWarning: `Model.fit_generator` is deprecated and will be removed in a future version. Please use `Model.fit`, which supports generators.\n",
      "  warnings.warn('`Model.fit_generator` is deprecated and '\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/50\n",
      "448/448 [==============================] - 97s 188ms/step - loss: 2.0199 - accuracy: 0.2179 - val_loss: 1.8766 - val_accuracy: 0.2159\n",
      "Epoch 2/50\n",
      "448/448 [==============================] - 81s 180ms/step - loss: 1.7800 - accuracy: 0.2580 - val_loss: 1.8325 - val_accuracy: 0.2705\n",
      "Epoch 3/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 1.7288 - accuracy: 0.3007 - val_loss: 1.7450 - val_accuracy: 0.3363\n",
      "Epoch 4/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 1.6638 - accuracy: 0.3390 - val_loss: 1.6538 - val_accuracy: 0.3597\n",
      "Epoch 5/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 1.5943 - accuracy: 0.3734 - val_loss: 1.5519 - val_accuracy: 0.4082\n",
      "Epoch 6/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 1.5274 - accuracy: 0.4029 - val_loss: 1.4992 - val_accuracy: 0.4221\n",
      "Epoch 7/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 1.4680 - accuracy: 0.4295 - val_loss: 1.4837 - val_accuracy: 0.4489\n",
      "Epoch 8/50\n",
      "448/448 [==============================] - 81s 180ms/step - loss: 1.4054 - accuracy: 0.4555 - val_loss: 1.3686 - val_accuracy: 0.4784\n",
      "Epoch 9/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 1.3477 - accuracy: 0.4846 - val_loss: 1.2894 - val_accuracy: 0.5183\n",
      "Epoch 10/50\n",
      "448/448 [==============================] - 81s 182ms/step - loss: 1.2969 - accuracy: 0.5043 - val_loss: 1.2349 - val_accuracy: 0.5411\n",
      "Epoch 11/50\n",
      "448/448 [==============================] - 81s 182ms/step - loss: 1.2482 - accuracy: 0.5224 - val_loss: 1.1808 - val_accuracy: 0.5570\n",
      "Epoch 12/50\n",
      "448/448 [==============================] - 82s 182ms/step - loss: 1.2169 - accuracy: 0.5364 - val_loss: 1.1931 - val_accuracy: 0.5559\n",
      "Epoch 13/50\n",
      "448/448 [==============================] - 82s 182ms/step - loss: 1.1886 - accuracy: 0.5468 - val_loss: 1.1918 - val_accuracy: 0.5587\n",
      "Epoch 14/50\n",
      "448/448 [==============================] - 82s 182ms/step - loss: 1.1607 - accuracy: 0.5589 - val_loss: 1.1279 - val_accuracy: 0.5821\n",
      "Epoch 15/50\n",
      "448/448 [==============================] - 81s 182ms/step - loss: 1.1282 - accuracy: 0.5727 - val_loss: 1.1941 - val_accuracy: 0.5573\n",
      "Epoch 16/50\n",
      "448/448 [==============================] - 82s 182ms/step - loss: 1.1077 - accuracy: 0.5822 - val_loss: 1.1170 - val_accuracy: 0.5779\n",
      "Epoch 17/50\n",
      "448/448 [==============================] - 81s 182ms/step - loss: 1.0859 - accuracy: 0.5922 - val_loss: 1.0858 - val_accuracy: 0.6010\n",
      "Epoch 18/50\n",
      "448/448 [==============================] - 81s 182ms/step - loss: 1.0673 - accuracy: 0.5978 - val_loss: 1.0980 - val_accuracy: 0.5857\n",
      "Epoch 19/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 1.0473 - accuracy: 0.6022 - val_loss: 1.0638 - val_accuracy: 0.6057\n",
      "Epoch 20/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 1.0275 - accuracy: 0.6113 - val_loss: 1.0537 - val_accuracy: 0.6021\n",
      "Epoch 21/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 1.0117 - accuracy: 0.6200 - val_loss: 1.0440 - val_accuracy: 0.6160\n",
      "Epoch 22/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.9953 - accuracy: 0.6245 - val_loss: 1.0470 - val_accuracy: 0.6197\n",
      "Epoch 23/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.9830 - accuracy: 0.6334 - val_loss: 1.0926 - val_accuracy: 0.6032\n",
      "Epoch 24/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.9657 - accuracy: 0.6377 - val_loss: 1.0615 - val_accuracy: 0.6024\n",
      "Epoch 25/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.9574 - accuracy: 0.6419 - val_loss: 1.0256 - val_accuracy: 0.6300\n",
      "Epoch 26/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.9407 - accuracy: 0.6469 - val_loss: 1.0164 - val_accuracy: 0.6230\n",
      "Epoch 27/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.9286 - accuracy: 0.6515 - val_loss: 1.0271 - val_accuracy: 0.6286\n",
      "Epoch 28/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.9096 - accuracy: 0.6606 - val_loss: 1.0310 - val_accuracy: 0.6252\n",
      "Epoch 29/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.8989 - accuracy: 0.6628 - val_loss: 1.0076 - val_accuracy: 0.6317\n",
      "Epoch 30/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.8904 - accuracy: 0.6701 - val_loss: 0.9476 - val_accuracy: 0.6520\n",
      "Epoch 31/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.8728 - accuracy: 0.6710 - val_loss: 1.0437 - val_accuracy: 0.6247\n",
      "Epoch 32/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.8604 - accuracy: 0.6750 - val_loss: 0.9955 - val_accuracy: 0.6353\n",
      "Epoch 33/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.8538 - accuracy: 0.6788 - val_loss: 0.9933 - val_accuracy: 0.6339\n",
      "Epoch 34/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.8391 - accuracy: 0.6850 - val_loss: 0.9995 - val_accuracy: 0.6562\n",
      "Epoch 35/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.8281 - accuracy: 0.6903 - val_loss: 1.0157 - val_accuracy: 0.6386\n",
      "Epoch 36/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.8125 - accuracy: 0.6953 - val_loss: 0.9675 - val_accuracy: 0.6439\n",
      "Epoch 37/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.7979 - accuracy: 0.7027 - val_loss: 0.9996 - val_accuracy: 0.6489\n",
      "Epoch 38/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.7936 - accuracy: 0.7028 - val_loss: 0.9541 - val_accuracy: 0.6637\n",
      "Epoch 39/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.7777 - accuracy: 0.7071 - val_loss: 0.9735 - val_accuracy: 0.6587\n",
      "Epoch 40/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.7702 - accuracy: 0.7115 - val_loss: 1.0062 - val_accuracy: 0.6514\n",
      "Epoch 41/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.7587 - accuracy: 0.7167 - val_loss: 0.9696 - val_accuracy: 0.6637\n",
      "Epoch 42/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.7505 - accuracy: 0.7214 - val_loss: 0.9551 - val_accuracy: 0.6684\n",
      "Epoch 43/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.7389 - accuracy: 0.7262 - val_loss: 0.9941 - val_accuracy: 0.6531\n",
      "Epoch 44/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.7255 - accuracy: 0.7323 - val_loss: 0.9417 - val_accuracy: 0.6659\n",
      "Epoch 45/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.7204 - accuracy: 0.7315 - val_loss: 0.9787 - val_accuracy: 0.6659\n",
      "Epoch 46/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.7103 - accuracy: 0.7337 - val_loss: 0.9484 - val_accuracy: 0.6729\n",
      "Epoch 47/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.6967 - accuracy: 0.7408 - val_loss: 1.0062 - val_accuracy: 0.6612\n",
      "Epoch 48/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.6855 - accuracy: 0.7457 - val_loss: 0.9548 - val_accuracy: 0.6709\n",
      "Epoch 49/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.6765 - accuracy: 0.7467 - val_loss: 1.0035 - val_accuracy: 0.6615\n",
      "Epoch 50/50\n",
      "448/448 [==============================] - 81s 181ms/step - loss: 0.6726 - accuracy: 0.7475 - val_loss: 0.9760 - val_accuracy: 0.6695\n"
     ]
    }
   ],
   "source": [
    "num_epochs = 50\n",
    "history = model.fit_generator(train_flow, \n",
    "                    steps_per_epoch=len(X_train) / batch_size, \n",
    "                    epochs=num_epochs,  \n",
    "                    verbose=1,  \n",
    "                    validation_data=test_flow,\n",
    "                    validation_steps=len(X_test) / batch_size)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b0603a6b",
   "metadata": {},
   "source": [
    "# Validation"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "7bbc3190",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    },
    {
     "data": {
      "text/plain": [
       "<Figure size 432x288 with 0 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "train_loss=history.history['loss']\n",
    "val_loss=history.history['val_loss']\n",
    "train_acc=history.history['accuracy']\n",
    "val_acc=history.history['val_accuracy']\n",
    "\n",
    "epochs = range(len(train_acc))\n",
    "\n",
    "plt.plot(epochs,train_loss,'r', label='train_loss')\n",
    "plt.plot(epochs,val_loss,'b', label='val_loss')\n",
    "plt.title('train_loss vs val_loss')\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('loss')\n",
    "plt.legend()\n",
    "plt.figure()\n",
    "\n",
    "plt.plot(epochs,train_acc,'r', label='train_acc')\n",
    "plt.plot(epochs,val_acc,'b', label='val_acc')\n",
    "plt.title('train_acc vs val_acc')\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('accuracy')\n",
    "plt.legend()\n",
    "plt.figure()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "08a790ac",
   "metadata": {},
   "outputs": [],
   "source": [
    "model.save('./working/Fer2013.h5') #backing up working model and weights"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "ba700242",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "113/113 [==============================] - 5s 35ms/step - loss: 0.9760 - accuracy: 0.6695 0s - loss: 0.9777 - \n",
      "Test Loss 0.9760403633117676\n",
      "Test Acc: 0.6695458292961121\n"
     ]
    }
   ],
   "source": [
    "loss = model.evaluate(X_test/255., y_test) \n",
    "print(\"Test Loss \" + str(loss[0]))\n",
    "print(\"Test Acc: \" + str(loss[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "50bf41cd",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "113/113 [==============================] - 3s 29ms/step\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "import itertools\n",
    "def plot_confusion_matrix(y_test, y_pred, classes,\n",
    "                          normalize=False,\n",
    "                          title='Unnormalized confusion matrix',\n",
    "                          cmap=plt.cm.Blues):\n",
    "    cm = confusion_matrix(y_test, y_pred)\n",
    "    \n",
    "    if normalize:\n",
    "        cm = np.round(cm.astype('float') / cm.sum(axis=1)[:, np.newaxis], 2)\n",
    "        \n",
    "    np.set_printoptions(precision=2)\n",
    "        \n",
    "    plt.imshow(cm, interpolation='nearest', cmap=cmap)\n",
    "    plt.title(title)\n",
    "    plt.colorbar()\n",
    "    tick_marks = np.arange(len(classes))\n",
    "    plt.xticks(tick_marks, classes, rotation=45)\n",
    "    plt.yticks(tick_marks, classes)\n",
    "\n",
    "    thresh = cm.min() + (cm.max() - cm.min()) / 2.\n",
    "    for i, j in itertools.product(range(cm.shape[0]), range(cm.shape[1])):\n",
    "        plt.text(j, i, cm[i, j],\n",
    "                 horizontalalignment=\"center\",\n",
    "                 color=\"white\" if cm[i, j] > thresh else \"black\")\n",
    "\n",
    "    plt.tight_layout()\n",
    "    plt.ylabel('True expression')\n",
    "    plt.xlabel('Predicted expression')\n",
    "    plt.show()\n",
    "y_pred_ = model.predict(X_test/255., verbose=1)\n",
    "y_pred = np.argmax(y_pred_, axis=1)\n",
    "t_te = np.argmax(y_test, axis=1)\n",
    "fig = plot_confusion_matrix(y_test=t_te, y_pred=y_pred,\n",
    "                      classes=classes,\n",
    "                      normalize=True,\n",
    "                      cmap=plt.cm.Greys,   title='Average accuracy: ' + str(np.sum(y_pred == t_te)/len(t_te)) + '\\n')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "058c629d",
   "metadata": {},
   "source": [
    "# Saving the model"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "62f3b6b2",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Saved model to disk\n"
     ]
    }
   ],
   "source": [
    "model_json = model.to_json()\n",
    "with open(\"model.json\", \"w\") as json_file:\n",
    "    json_file.write(model_json)\n",
    "model.save_weights(\"model.h5\")\n",
    "print(\"Saved model to disk\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b787837c",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}