MMC Classification with Augmented Data 2D.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "467ed1d0",
   "metadata": {},
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings(\"ignore\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "5d9be511",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "2023-06-30 08:03:59.663458: I tensorflow/core/platform/cpu_feature_guard.cc:182] This TensorFlow binary is optimized to use available CPU instructions in performance-critical operations.\n",
      "To enable the following instructions: AVX2 FMA, in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
     ]
    }
   ],
   "source": [
    "import os\n",
    "import cv2\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "from sklearn.model_selection import train_test_split \n",
    "import tensorflow as tf\n",
    "from tensorflow.keras.models import Sequential, load_model\n",
    "from tensorflow.keras.layers import Conv2D, MaxPooling2D,BatchNormalization, Dense, Dropout, Flatten\n",
    "from tensorflow.keras.optimizers import Adam\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "from sklearn.metrics import accuracy_score, confusion_matrix, classification_report, ConfusionMatrixDisplay"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "4631b1c8",
   "metadata": {},
   "outputs": [],
   "source": [
    "f = np.load(\"Features2D/npz/Features2.npz\")\n",
    "f_noise = np.load(\"Features2D/npz/Features_noise.npz\")\n",
    "f_shift = np.load(\"Features2D/npz/Features_shift.npz\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "db963bab",
   "metadata": {},
   "outputs": [],
   "source": [
    "spec = f['spec'] #(400,1025,1099)\n",
    "mfcc = f['mfcc'] #(400,20,1099)\n",
    "mel = f['mel']#(400,128,1099)\n",
    "y = f['target']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "54482a18",
   "metadata": {},
   "outputs": [],
   "source": [
    "spec_noise = f_noise['spec'] #(400,1025,1099)\n",
    "mfcc_noise = f_noise['mfcc'] #(400,20,1099)\n",
    "mel_noise = f_noise['mel']#(400,128,1099)\n",
    "y_noise = f_noise['target']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "57eac514",
   "metadata": {},
   "outputs": [],
   "source": [
    "spec_shift = f_shift['spec'] #(400,1025,1099)\n",
    "mfcc_shift = f_shift['mfcc'] #(400,20,1099)\n",
    "mel_shift = f_shift['mel']#(400,128,1099)\n",
    "y_shift = f_shift['target']"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d9448b9d",
   "metadata": {},
   "source": [
    "## 3. Data Preprocessing\n",
    "### 3.1 Resize the data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "b0a48043",
   "metadata": {},
   "outputs": [],
   "source": [
    "mfcc_resized = np.empty((400, 120, 600))\n",
    "mfcc_noise_resized = np.empty((400, 120, 600))\n",
    "mfcc_shift_resized = np.empty((400, 120, 600))\n",
    "\n",
    "for i in range(0,400):\n",
    "    mfcc_resized[i] = cv2.resize(mfcc[i],(600,120))\n",
    "    mfcc_noise_resized[i] = cv2.resize(mfcc_noise[i],(600,120))\n",
    "    mfcc_shift_resized[i] = cv2.resize(mfcc_shift[i],(600,120))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "3fa96a8a",
   "metadata": {},
   "outputs": [],
   "source": [
    "spec_resized = np.empty((400, 300, 300))\n",
    "spec_noise_resized = np.empty((400, 300, 300))\n",
    "spec_shift_resized = np.empty((400, 300, 300))\n",
    "\n",
    "for i in range(0,400):\n",
    "    spec_resized[i] = cv2.resize(spec[i],(300,300))\n",
    "    spec_noise_resized[i] = cv2.resize(spec_noise[i],(300,300))\n",
    "    spec_shift_resized[i] = cv2.resize(spec_shift[i],(300,300))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "d782e4ae",
   "metadata": {},
   "outputs": [],
   "source": [
    "mel_resized = np.empty((400, 300, 400))\n",
    "mel_noise_resized = np.empty((400, 300, 400))\n",
    "mel_shift_resized = np.empty((400, 300, 400))\n",
    "\n",
    "for i in range(0,400):\n",
    "    mel_resized[i] = cv2.resize(mel[i],(400,300))\n",
    "    mel_noise_resized[i] = cv2.resize(mel_noise[i],(400,300))\n",
    "    mel_shift_resized[i] = cv2.resize(mel_shift[i],(400,300))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fe37f7e2",
   "metadata": {},
   "source": [
    "### 3.2 Train Test Split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "184ed362",
   "metadata": {},
   "outputs": [],
   "source": [
    "mfcc0_train,mfcc_test,y0_train,y_test = train_test_split(mfcc_resized,y, \n",
    "                                                         train_size = 0.7, \n",
    "                                                         random_state = 13, \n",
    "                                                         stratify= y)\n",
    "mfcc_train_agumented = np.concatenate((mfcc0_train,mfcc_noise_resized,mfcc_shift_resized))\n",
    "y1_train_agumented = np.concatenate((y0_train,y_noise,y_shift))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "a7c14703",
   "metadata": {},
   "outputs": [],
   "source": [
    "spec0_train,spec_test,y0_train,y_test = train_test_split(spec_resized,y, \n",
    "                                                   train_size = 0.7, \n",
    "                                                   random_state = 13, \n",
    "                                                   stratify= y)\n",
    "spec_train_agumented = np.concatenate((spec0_train,spec_noise_resized,spec_shift_resized))\n",
    "y2_train_agumented = np.concatenate((y0_train,y_noise,y_shift))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "d08ead48",
   "metadata": {},
   "outputs": [],
   "source": [
    "mel0_train,mel_test,y0_train,y_test = train_test_split(mel_resized,y, \n",
    "                                                       train_size = 0.7, \n",
    "                                                       random_state = 13, \n",
    "                                                       stratify= y)\n",
    "mel_train_agumented = np.concatenate((mel0_train,mel_noise_resized,mel_shift_resized))\n",
    "y3_train_agumented = np.concatenate((y0_train,y_noise,y_shift))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "d7aebd72",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.savez_compressed(\"Features2D/npz_new/mfcc_train_test_agumented.npz\",\n",
    "                    mfcc_train = mfcc_train_agumented, \n",
    "                    mfcc_test = mfcc_test, \n",
    "                    y_train = y1_train_agumented, \n",
    "                    y_test = y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "caa7a683",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.savez_compressed(\"Features2D/npz_new/spec_train_test_agumented.npz\",\n",
    "                    spec_train = spec_train_agumented, \n",
    "                    spec_test = spec_test, \n",
    "                    y_train = y2_train_agumented, \n",
    "                    y_test = y_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "2822e3c8",
   "metadata": {},
   "outputs": [],
   "source": [
    "np.savez_compressed(\"Features2D/npz_new/mel_train_test_agumented.npz\",\n",
    "                    mel_train = mel_train_agumented, \n",
    "                    mel_test = mel_test, \n",
    "                    y_train = y3_train_agumented, \n",
    "                    y_test = y_test)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "adb83584",
   "metadata": {},
   "source": [
    "### 3.2 Feature Scaling"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "9a40b74c",
   "metadata": {},
   "outputs": [],
   "source": [
    "spec_scaled = spec_resized / np.amax(spec_resized)\n",
    "mfcc_scaled = (mfcc_resized - np.mean(mfcc_resized))/np.std(mfcc_resized)\n",
    "mel_scaled = mel_resized / np.amax(mel_resized)\n",
    "\n",
    "spec_noise_scaled = spec_noise_resized / np.amax(spec_noise_resized)\n",
    "mfcc_noise_scaled = (mfcc_noise_resized - np.mean(mfcc_noise_resized))/np.std(mfcc_noise_resized)\n",
    "mel_noise_scaled = mel_noise_resized / np.amax(mel_noise_resized)\n",
    "\n",
    "spec_shift_scaled = spec_shift_resized / np.amax(spec_shift_resized)\n",
    "mfcc_shift_scaled = (mfcc_shift_resized - np.mean(mfcc_shift_resized))/np.std(mfcc_shift_resized)\n",
    "mel_shift_scaled = mel_shift_resized / np.amax(mel_shift_resized)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "158f5626",
   "metadata": {},
   "source": [
    "### 3.3. Reshape Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "095ac704",
   "metadata": {},
   "outputs": [],
   "source": [
    "N,row,col = mfcc_scaled.shape\n",
    "mfcc_reshaped = mfcc_scaled.reshape((N,row,col,1))\n",
    "N,row,col = mfcc_noise_scaled.shape\n",
    "mfcc_noise_reshaped = mfcc_noise_scaled.reshape((N,row,col,1))\n",
    "N,row,col = mfcc_shift_scaled.shape\n",
    "mfcc_shift_reshaped = mfcc_shift_scaled.reshape((N,row,col,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "3050cd2c",
   "metadata": {},
   "outputs": [],
   "source": [
    "N,row,col = spec_scaled.shape\n",
    "spec_reshaped = spec_scaled.reshape((N,row,col,1))\n",
    "N,row,col = spec_noise_scaled.shape\n",
    "spec_noise_reshaped = spec_noise_scaled.reshape((N,row,col,1))\n",
    "N,row,col = spec_shift_scaled.shape\n",
    "spec_shift_reshaped = spec_shift_scaled.reshape((N,row,col,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "b3c57cc2",
   "metadata": {},
   "outputs": [],
   "source": [
    "N,row,col = mel_scaled.shape\n",
    "mel_reshaped = mel_scaled.reshape((N,row,col,1))\n",
    "N,row,col = mel_noise_scaled.shape\n",
    "mel_noise_reshaped = mel_noise_scaled.reshape((N,row,col,1))\n",
    "N,row,col = mel_shift_scaled.shape\n",
    "mel_shift_reshaped = mel_shift_scaled.reshape((N,row,col,1))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "16e372df",
   "metadata": {},
   "source": [
    "## 4. Classification"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "a6ed8c63",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Train the model\n",
    "def modelBuilder3L(X_train,\n",
    "                   f1,k1,a1,\n",
    "                   mr1,mc1,sr1,sc1,\n",
    "                   f2,k2,a2,\n",
    "                   mr2,mc2,sr2,sc2,\n",
    "                   f3,k3,a3,\n",
    "                   mr3,m3,sr3,sc3,\n",
    "                   d1,dr1,da1,r1,\n",
    "                   d2,dr2,da2,r2,\n",
    "                   num):\n",
    "    \n",
    "    '''\n",
    "    args:\n",
    "    \n",
    "    X_train: training data\n",
    "    f1,k1,a1: num of filters, filter size and activation func of 1st conv2D layer\n",
    "    mr1,mc1,sr1,sc1: filter size and strides of 1st MaxPooling2D layer\n",
    "    f2,k2,a2: num of filters, filter size and activation func of 2nd conv2D layer\n",
    "    mr2,mc2,sr2,sc2: filter size and strides of 2nd MaxPooling2D layer\n",
    "    f3,k3,a3: num of filters, filter size and activation func of 3rd conv2D layer\n",
    "    mr3,sc3,sr3,sc3: filter size and strides of 3rd MaxPooling2D layer\n",
    "    d1,dr1,da1, r1: num of units, dropout, activation func and regularizer para of 1st fully connected layer\n",
    "    d2,dr2,da2, r2: num of units, dropout, activation func and regularizer para of 2nd fully connected layer\n",
    "    num: integer for distinguishing different model\n",
    "    \n",
    "    return:\n",
    "    model\n",
    "    \n",
    "    '''\n",
    "    \n",
    "    N,row,col,Nc = X_train.shape\n",
    "    \n",
    "    model = Sequential(name = \"CNN2D_\"+str(num))\n",
    "    #L1\n",
    "    model.add(Conv2D(f1,(k1,k1),activation = a1,input_shape = (row,col,Nc), padding = 'same', name = 'Conv2D_1'))\n",
    "    model.add(MaxPooling2D((mw1,mh1), strides = (sw1,sh1), padding = 'same', name = 'MaxPooling2D_1'))\n",
    "    model.add(BatchNormalization(name = \"BN1\"))\n",
    "    \n",
    "    #L2\n",
    "    model.add(Conv2D(f2,(k2,k2), activation = a2, padding = 'same',name = 'Conv2D_2'))\n",
    "    model.add(MaxPooling2D((mw2,mh2),strides = (sw2,sh2), padding = 'same', name = 'MaxPooling2D_2'))\n",
    "    model.add(BatchNormalization(name = \"BN2\"))\n",
    "\n",
    "    #L3\n",
    "    model.add(Conv2D(f3,(k3,k3), activation = a3, padding = 'same',name = 'Conv2D_3'))\n",
    "    model.add(MaxPooling2D((mw3,mh3), strides = (sw3,sh3), padding = 'same', name = 'MaxPooling2D_3'))\n",
    "    model.add(BatchNormalization(name = \"BN3\"))\n",
    "   \n",
    "    #Flatten layer\n",
    "    model.add(Flatten(name = 'Flatten'))\n",
    "    #FC4\n",
    "    model.add(Dense(d1, activation = da1, kernel_regularizer = tf.keras.regularizers.L2(r1), name='Dense1'))\n",
    "    model.add(Dropout(dr1, name = \"Dropout1\"))\n",
    "    #FC5\n",
    "    model.add(Dense(d2, activation = da2, kernel_regularizer = tf.keras.regularizers.L2(r2), name='Dense2'))\n",
    "    model.add(Dropout(dr2, name = \"Dropout2\"))\n",
    "    #Softmax layer\n",
    "    model.add(Dense(4, activation ='softmax', name='Softmax'))\n",
    "\n",
    "    optimizer = Adam(learning_rate=0.0001)\n",
    "    model.compile(optimizer= optimizer, loss= 'sparse_categorical_crossentropy', metrics=['accuracy'])\n",
    "\n",
    "    model.summary()\n",
    "    \n",
    "    return model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "14441af7",
   "metadata": {},
   "source": [
    "### 4.1 MFCC"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "87351dfe",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load MFCC data set\n",
    "mfcc_file = np.load(\"Features2D/npz_new/mfcc_train_test_agumented.npz\")\n",
    "\n",
    "mfcc_train = mfcc_file['mfcc_train']\n",
    "y1_train = mfcc_file['y_train']\n",
    "\n",
    "# validation set\n",
    "mfcc_train,mfcc_val,y1_train,y1_val = train_test_split(mfcc_train,y1_train, \n",
    "                                                       train_size = 0.8,  \n",
    "                                                       random_state = 13, \n",
    "                                                       stratify= y1_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "a4f24a0c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature Scaling\n",
    "mfcc_train_scaled = (mfcc_train - np.mean(mfcc_train))/np.std(mfcc_train)\n",
    "mfcc_val_scaled = (mfcc_val - np.mean(mfcc_val))/np.std(mfcc_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "064762a7",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Reshape the data\n",
    "N,row,col = mfcc_train_scaled.shape\n",
    "mfcc_train_reshaped = mfcc_train_scaled.reshape((N,row,col,1))\n",
    "N,row,col = mfcc_val_scaled.shape\n",
    "mfcc_val_reshaped = mfcc_val_scaled.reshape((N,row,col,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "52bd0099",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"CNN2D_1\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " Conv2D_1 (Conv2D)           (None, 120, 600, 8)       80        \n",
      "                                                                 \n",
      " MaxPooling2D_1 (MaxPooling2  (None, 60, 300, 8)       0         \n",
      " D)                                                              \n",
      "                                                                 \n",
      " BN1 (BatchNormalization)    (None, 60, 300, 8)        32        \n",
      "                                                                 \n",
      " Conv2D_2 (Conv2D)           (None, 60, 300, 8)        584       \n",
      "                                                                 \n",
      " MaxPooling2D_2 (MaxPooling2  (None, 30, 150, 8)       0         \n",
      " D)                                                              \n",
      "                                                                 \n",
      " BN2 (BatchNormalization)    (None, 30, 150, 8)        32        \n",
      "                                                                 \n",
      " Conv2D_3 (Conv2D)           (None, 30, 150, 16)       1168      \n",
      "                                                                 \n",
      " MaxPooling2D_3 (MaxPooling2  (None, 15, 75, 16)       0         \n",
      " D)                                                              \n",
      "                                                                 \n",
      " BN3 (BatchNormalization)    (None, 15, 75, 16)        64        \n",
      "                                                                 \n",
      " Flatten (Flatten)           (None, 18000)             0         \n",
      "                                                                 \n",
      " Dense1 (Dense)              (None, 64)                1152064   \n",
      "                                                                 \n",
      " Dropout1 (Dropout)          (None, 64)                0         \n",
      "                                                                 \n",
      " Dense2 (Dense)              (None, 32)                2080      \n",
      "                                                                 \n",
      " Dropout2 (Dropout)          (None, 32)                0         \n",
      "                                                                 \n",
      " Softmax (Dense)             (None, 4)                 132       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 1,156,236\n",
      "Trainable params: 1,156,172\n",
      "Non-trainable params: 64\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "f1,k1,a1 = 8,3,'relu'\n",
    "mw1,mh1,sw1,sh1 = 2,3,2,2\n",
    "f2,k2,a2 = 8,3,'relu'\n",
    "mw2,mh2,sw2,sh2 = 2,3,2,2\n",
    "f3,k3,a3 = 16,3,'relu'\n",
    "mw3,mh3,sw3,sh3 = 2,3,2,2\n",
    "d1,dr1,da1,r1 = 64,0.2,'relu',0.2\n",
    "d2,dr2,da2,r2 = 32,0.2,'relu',0.2\n",
    "num = 1\n",
    "\n",
    "model1 = modelBuilder3L(mfcc_train_reshaped,\n",
    "                        f1,k1,a1,\n",
    "                        mw1,mh1,sw1,sh1,\n",
    "                        f2,k2,a2,\n",
    "                        mw2,mh2,sw2,sh2,\n",
    "                        f3,k3,a3,\n",
    "                        mw3,mh3,sw3,sh3,\n",
    "                        d1,dr1,da1,r1,\n",
    "                        d2,dr2,da2,r2,\n",
    "                        num)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "0fd0055a",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/80\n",
      "44/44 [==============================] - 37s 735ms/step - loss: 30.8823 - accuracy: 0.3831 - val_loss: 26.2593 - val_accuracy: 0.2546\n",
      "Epoch 2/80\n",
      "44/44 [==============================] - 16s 372ms/step - loss: 22.7770 - accuracy: 0.4942 - val_loss: 19.8386 - val_accuracy: 0.4769\n",
      "Epoch 3/80\n",
      "44/44 [==============================] - 15s 333ms/step - loss: 17.4306 - accuracy: 0.6042 - val_loss: 15.7417 - val_accuracy: 0.5185\n",
      "Epoch 4/80\n",
      "44/44 [==============================] - 13s 293ms/step - loss: 14.0707 - accuracy: 0.6574 - val_loss: 13.2176 - val_accuracy: 0.4120\n",
      "Epoch 5/80\n",
      "44/44 [==============================] - 13s 302ms/step - loss: 11.8819 - accuracy: 0.6806 - val_loss: 11.4397 - val_accuracy: 0.4537\n",
      "Epoch 6/80\n",
      "44/44 [==============================] - 13s 291ms/step - loss: 10.3029 - accuracy: 0.7384 - val_loss: 10.1623 - val_accuracy: 0.5324\n",
      "Epoch 7/80\n",
      "44/44 [==============================] - 14s 307ms/step - loss: 9.1828 - accuracy: 0.7627 - val_loss: 9.1224 - val_accuracy: 0.5417\n",
      "Epoch 8/80\n",
      "44/44 [==============================] - 14s 323ms/step - loss: 8.3396 - accuracy: 0.8056 - val_loss: 8.2976 - val_accuracy: 0.6204\n",
      "Epoch 9/80\n",
      "44/44 [==============================] - 15s 341ms/step - loss: 7.6582 - accuracy: 0.8426 - val_loss: 7.7323 - val_accuracy: 0.6204\n",
      "Epoch 10/80\n",
      "44/44 [==============================] - 16s 373ms/step - loss: 7.1020 - accuracy: 0.8495 - val_loss: 7.0264 - val_accuracy: 0.7963\n",
      "Epoch 11/80\n",
      "44/44 [==============================] - 15s 336ms/step - loss: 6.6071 - accuracy: 0.8785 - val_loss: 6.5467 - val_accuracy: 0.8241\n",
      "Epoch 12/80\n",
      "44/44 [==============================] - 14s 307ms/step - loss: 6.1900 - accuracy: 0.8900 - val_loss: 6.1271 - val_accuracy: 0.8380\n",
      "Epoch 13/80\n",
      "44/44 [==============================] - 13s 294ms/step - loss: 5.8646 - accuracy: 0.8819 - val_loss: 5.8271 - val_accuracy: 0.8056\n",
      "Epoch 14/80\n",
      "44/44 [==============================] - 22s 494ms/step - loss: 5.4924 - accuracy: 0.9120 - val_loss: 5.4660 - val_accuracy: 0.8241\n",
      "Epoch 15/80\n",
      "44/44 [==============================] - 15s 347ms/step - loss: 5.2292 - accuracy: 0.8877 - val_loss: 5.1691 - val_accuracy: 0.8472\n",
      "Epoch 16/80\n",
      "44/44 [==============================] - 14s 308ms/step - loss: 4.8879 - accuracy: 0.9201 - val_loss: 4.9256 - val_accuracy: 0.8148\n",
      "Epoch 17/80\n",
      "44/44 [==============================] - 14s 319ms/step - loss: 4.6271 - accuracy: 0.9271 - val_loss: 4.6188 - val_accuracy: 0.8750\n",
      "Epoch 18/80\n",
      "44/44 [==============================] - 14s 308ms/step - loss: 4.3505 - accuracy: 0.9398 - val_loss: 4.4307 - val_accuracy: 0.8333\n",
      "Epoch 19/80\n",
      "44/44 [==============================] - 13s 305ms/step - loss: 4.1226 - accuracy: 0.9410 - val_loss: 4.1810 - val_accuracy: 0.8704\n",
      "Epoch 20/80\n",
      "44/44 [==============================] - 14s 311ms/step - loss: 3.9074 - accuracy: 0.9433 - val_loss: 3.9675 - val_accuracy: 0.8611\n",
      "Epoch 21/80\n",
      "44/44 [==============================] - 14s 308ms/step - loss: 3.7119 - accuracy: 0.9479 - val_loss: 3.8299 - val_accuracy: 0.8241\n",
      "Epoch 22/80\n",
      "44/44 [==============================] - 13s 306ms/step - loss: 3.5446 - accuracy: 0.9317 - val_loss: 3.6428 - val_accuracy: 0.8472\n",
      "Epoch 23/80\n",
      "44/44 [==============================] - 13s 304ms/step - loss: 3.3661 - accuracy: 0.9444 - val_loss: 3.4602 - val_accuracy: 0.8472\n",
      "Epoch 24/80\n",
      "44/44 [==============================] - 14s 315ms/step - loss: 3.1895 - accuracy: 0.9491 - val_loss: 3.2764 - val_accuracy: 0.8657\n",
      "Epoch 25/80\n",
      "44/44 [==============================] - 15s 335ms/step - loss: 3.0167 - accuracy: 0.9595 - val_loss: 3.1354 - val_accuracy: 0.8704\n",
      "Epoch 26/80\n",
      "44/44 [==============================] - 14s 313ms/step - loss: 2.8620 - accuracy: 0.9641 - val_loss: 2.9853 - val_accuracy: 0.8565\n",
      "Epoch 27/80\n",
      "44/44 [==============================] - 14s 327ms/step - loss: 2.7039 - accuracy: 0.9653 - val_loss: 2.8524 - val_accuracy: 0.8565\n",
      "Epoch 28/80\n",
      "44/44 [==============================] - 13s 291ms/step - loss: 2.5476 - accuracy: 0.9792 - val_loss: 2.7381 - val_accuracy: 0.8519\n",
      "Epoch 29/80\n",
      "44/44 [==============================] - 13s 294ms/step - loss: 2.4305 - accuracy: 0.9664 - val_loss: 2.6582 - val_accuracy: 0.8194\n",
      "Epoch 30/80\n",
      "44/44 [==============================] - 13s 294ms/step - loss: 2.3400 - accuracy: 0.9618 - val_loss: 2.4723 - val_accuracy: 0.8704\n",
      "Epoch 31/80\n",
      "44/44 [==============================] - 13s 298ms/step - loss: 2.1960 - accuracy: 0.9734 - val_loss: 2.3500 - val_accuracy: 0.8519\n",
      "Epoch 32/80\n",
      "44/44 [==============================] - 13s 295ms/step - loss: 2.0750 - accuracy: 0.9699 - val_loss: 2.2862 - val_accuracy: 0.8657\n",
      "Epoch 33/80\n",
      "44/44 [==============================] - 13s 298ms/step - loss: 1.9860 - accuracy: 0.9676 - val_loss: 2.1546 - val_accuracy: 0.8472\n",
      "Epoch 34/80\n",
      "44/44 [==============================] - 13s 297ms/step - loss: 1.9195 - accuracy: 0.9630 - val_loss: 2.1481 - val_accuracy: 0.8380\n",
      "Epoch 35/80\n",
      "44/44 [==============================] - 13s 297ms/step - loss: 1.8153 - accuracy: 0.9664 - val_loss: 1.9893 - val_accuracy: 0.8565\n",
      "Epoch 36/80\n",
      "44/44 [==============================] - 13s 305ms/step - loss: 1.7053 - accuracy: 0.9769 - val_loss: 1.9132 - val_accuracy: 0.8565\n",
      "Epoch 37/80\n",
      "44/44 [==============================] - 13s 297ms/step - loss: 1.6177 - accuracy: 0.9745 - val_loss: 1.8320 - val_accuracy: 0.8657\n",
      "Epoch 38/80\n",
      "44/44 [==============================] - 14s 311ms/step - loss: 1.5746 - accuracy: 0.9583 - val_loss: 1.7905 - val_accuracy: 0.8519\n",
      "Epoch 39/80\n",
      "44/44 [==============================] - 14s 309ms/step - loss: 1.4944 - accuracy: 0.9641 - val_loss: 1.6409 - val_accuracy: 0.8750\n",
      "Epoch 40/80\n",
      "44/44 [==============================] - 13s 301ms/step - loss: 1.4153 - accuracy: 0.9734 - val_loss: 1.6353 - val_accuracy: 0.8750\n",
      "Epoch 41/80\n",
      "44/44 [==============================] - 13s 306ms/step - loss: 1.3862 - accuracy: 0.9537 - val_loss: 1.5606 - val_accuracy: 0.8472\n",
      "Epoch 42/80\n",
      "44/44 [==============================] - 13s 303ms/step - loss: 1.2917 - accuracy: 0.9769 - val_loss: 1.4788 - val_accuracy: 0.8611\n",
      "Epoch 43/80\n",
      "44/44 [==============================] - 15s 340ms/step - loss: 1.2300 - accuracy: 0.9711 - val_loss: 1.4504 - val_accuracy: 0.8611\n",
      "Epoch 44/80\n",
      "44/44 [==============================] - 13s 305ms/step - loss: 1.1667 - accuracy: 0.9769 - val_loss: 1.3779 - val_accuracy: 0.8565\n",
      "Epoch 45/80\n",
      "44/44 [==============================] - 13s 304ms/step - loss: 1.1305 - accuracy: 0.9641 - val_loss: 1.3245 - val_accuracy: 0.8426\n",
      "Epoch 46/80\n",
      "44/44 [==============================] - 13s 302ms/step - loss: 1.1184 - accuracy: 0.9641 - val_loss: 1.3342 - val_accuracy: 0.8472\n",
      "Epoch 47/80\n",
      "44/44 [==============================] - 13s 303ms/step - loss: 1.0535 - accuracy: 0.9676 - val_loss: 1.2380 - val_accuracy: 0.8565\n",
      "Epoch 48/80\n",
      "44/44 [==============================] - 14s 307ms/step - loss: 0.9949 - accuracy: 0.9641 - val_loss: 1.1919 - val_accuracy: 0.8704\n",
      "Epoch 49/80\n",
      "44/44 [==============================] - 15s 341ms/step - loss: 0.9428 - accuracy: 0.9780 - val_loss: 1.1532 - val_accuracy: 0.8519\n",
      "Epoch 50/80\n",
      "44/44 [==============================] - 14s 321ms/step - loss: 0.8977 - accuracy: 0.9745 - val_loss: 1.1181 - val_accuracy: 0.8565\n",
      "Epoch 51/80\n",
      "44/44 [==============================] - 15s 345ms/step - loss: 0.8538 - accuracy: 0.9803 - val_loss: 1.1114 - val_accuracy: 0.8472\n",
      "Epoch 52/80\n",
      "44/44 [==============================] - 13s 302ms/step - loss: 0.8358 - accuracy: 0.9757 - val_loss: 1.0409 - val_accuracy: 0.8472\n",
      "Epoch 53/80\n",
      "44/44 [==============================] - 14s 312ms/step - loss: 0.7961 - accuracy: 0.9734 - val_loss: 1.0366 - val_accuracy: 0.8657\n",
      "Epoch 54/80\n",
      "44/44 [==============================] - 14s 314ms/step - loss: 0.7698 - accuracy: 0.9792 - val_loss: 1.0082 - val_accuracy: 0.8657\n",
      "Epoch 55/80\n",
      "44/44 [==============================] - 14s 312ms/step - loss: 0.7484 - accuracy: 0.9757 - val_loss: 0.9580 - val_accuracy: 0.8519\n",
      "Epoch 56/80\n",
      "44/44 [==============================] - 13s 292ms/step - loss: 0.7085 - accuracy: 0.9838 - val_loss: 0.9056 - val_accuracy: 0.8472\n",
      "Epoch 57/80\n",
      "44/44 [==============================] - 13s 289ms/step - loss: 0.7087 - accuracy: 0.9688 - val_loss: 0.9336 - val_accuracy: 0.8750\n",
      "Epoch 58/80\n",
      "44/44 [==============================] - 15s 347ms/step - loss: 0.6714 - accuracy: 0.9803 - val_loss: 0.9021 - val_accuracy: 0.8565\n",
      "Epoch 59/80\n",
      "44/44 [==============================] - 15s 335ms/step - loss: 0.6789 - accuracy: 0.9711 - val_loss: 0.8900 - val_accuracy: 0.8611\n",
      "Epoch 60/80\n",
      "44/44 [==============================] - 17s 373ms/step - loss: 0.6601 - accuracy: 0.9757 - val_loss: 0.8740 - val_accuracy: 0.8657\n",
      "Epoch 61/80\n",
      "44/44 [==============================] - 14s 327ms/step - loss: 0.6231 - accuracy: 0.9815 - val_loss: 0.8308 - val_accuracy: 0.8704\n",
      "Epoch 62/80\n",
      "44/44 [==============================] - 14s 324ms/step - loss: 0.6244 - accuracy: 0.9722 - val_loss: 0.8650 - val_accuracy: 0.8611\n",
      "Epoch 63/80\n",
      "44/44 [==============================] - 15s 341ms/step - loss: 0.5975 - accuracy: 0.9792 - val_loss: 0.8192 - val_accuracy: 0.8704\n",
      "Epoch 64/80\n",
      "44/44 [==============================] - 16s 353ms/step - loss: 0.5655 - accuracy: 0.9919 - val_loss: 0.8183 - val_accuracy: 0.8657\n",
      "Epoch 65/80\n",
      "44/44 [==============================] - 15s 341ms/step - loss: 0.5386 - accuracy: 0.9803 - val_loss: 0.8044 - val_accuracy: 0.8472\n",
      "Epoch 66/80\n",
      "44/44 [==============================] - 17s 390ms/step - loss: 0.5359 - accuracy: 0.9838 - val_loss: 0.7681 - val_accuracy: 0.8380\n",
      "Epoch 67/80\n",
      "44/44 [==============================] - 17s 380ms/step - loss: 0.5133 - accuracy: 0.9815 - val_loss: 0.8147 - val_accuracy: 0.8380\n",
      "Epoch 68/80\n",
      "44/44 [==============================] - 17s 394ms/step - loss: 0.5156 - accuracy: 0.9780 - val_loss: 0.8583 - val_accuracy: 0.8472\n",
      "Epoch 69/80\n",
      "44/44 [==============================] - 15s 331ms/step - loss: 0.5469 - accuracy: 0.9711 - val_loss: 0.7515 - val_accuracy: 0.8704\n",
      "Epoch 70/80\n",
      "44/44 [==============================] - 16s 369ms/step - loss: 0.5303 - accuracy: 0.9803 - val_loss: 0.7594 - val_accuracy: 0.8611\n",
      "Epoch 71/80\n",
      "44/44 [==============================] - 16s 361ms/step - loss: 0.5095 - accuracy: 0.9745 - val_loss: 0.7273 - val_accuracy: 0.8796\n",
      "Epoch 72/80\n",
      "44/44 [==============================] - 16s 372ms/step - loss: 0.4710 - accuracy: 0.9815 - val_loss: 0.6917 - val_accuracy: 0.8843\n",
      "Epoch 73/80\n",
      "44/44 [==============================] - 15s 333ms/step - loss: 0.4685 - accuracy: 0.9838 - val_loss: 0.6792 - val_accuracy: 0.8657\n",
      "Epoch 74/80\n",
      "44/44 [==============================] - 14s 308ms/step - loss: 0.4469 - accuracy: 0.9815 - val_loss: 0.7318 - val_accuracy: 0.8472\n",
      "Epoch 75/80\n",
      "44/44 [==============================] - 14s 307ms/step - loss: 0.4591 - accuracy: 0.9734 - val_loss: 0.6755 - val_accuracy: 0.8750\n",
      "Epoch 76/80\n",
      "44/44 [==============================] - 13s 296ms/step - loss: 0.4911 - accuracy: 0.9688 - val_loss: 0.7749 - val_accuracy: 0.8426\n",
      "Epoch 77/80\n",
      "44/44 [==============================] - 14s 317ms/step - loss: 0.4666 - accuracy: 0.9792 - val_loss: 0.6914 - val_accuracy: 0.8611\n",
      "Epoch 78/80\n",
      "44/44 [==============================] - 14s 324ms/step - loss: 0.4681 - accuracy: 0.9838 - val_loss: 0.7102 - val_accuracy: 0.8704\n",
      "Epoch 79/80\n",
      "44/44 [==============================] - 14s 313ms/step - loss: 0.4365 - accuracy: 0.9861 - val_loss: 0.6754 - val_accuracy: 0.8565\n",
      "Epoch 80/80\n",
      "44/44 [==============================] - 13s 300ms/step - loss: 0.4315 - accuracy: 0.9850 - val_loss: 0.6820 - val_accuracy: 0.8657\n"
     ]
    }
   ],
   "source": [
    "history1 = model1.fit(mfcc_train_reshaped,y1_train,validation_data=(mfcc_val_reshaped,y1_val),batch_size=20,epochs=80)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "54a70c0d",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'loss')"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Train and Validation accuracy and loss\n",
    "f = plt.figure(figsize = (12,4))\n",
    "\n",
    "plt.subplot(1,2,1)\n",
    "plt.title('train validation accuracy')\n",
    "plt.plot(history1.history['accuracy'])\n",
    "plt.plot(history1.history['val_accuracy'])\n",
    "plt.legend(['train','validation'])\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('accuracy')\n",
    "\n",
    "\n",
    "plt.subplot(1,2,2)\n",
    "plt.title('train validation loss')\n",
    "plt.plot(history1.history['loss'])\n",
    "plt.plot(history1.history['val_loss'])\n",
    "plt.legend(['train','validation'])\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('loss')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "1f2037e5",
   "metadata": {},
   "outputs": [],
   "source": [
    "mfcc_test = mfcc_file['mfcc_test']\n",
    "y_test_mfcc = mfcc_file['y_test']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "86e7da56",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature Scaling\n",
    "mfcc_test_scaled = (mfcc_test - np.mean(mfcc_test))/np.std(mfcc_test)\n",
    "# Reshape the data\n",
    "N,row,col = mfcc_test_scaled.shape\n",
    "mfcc_test_reshaped = mfcc_test_scaled.reshape((N,row,col,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "b327d080",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4/4 [==============================] - 1s 149ms/step\n"
     ]
    }
   ],
   "source": [
    "y1_prob = model1.predict(mfcc_test_reshaped)\n",
    "y1_pred = np.argmax(y1_prob, axis = -1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "id": "ac44691f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Accuracy\n",
    "acc_mfcc = round(accuracy_score(y1_pred,y_test_mfcc),2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "id": "ca5050e7",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "       Angry       0.91      0.97      0.94        30\n",
      "       Happy       0.97      1.00      0.98        30\n",
      "     Relaxed       0.96      0.87      0.91        30\n",
      "         Sad       0.90      0.90      0.90        30\n",
      "\n",
      "    accuracy                           0.93       120\n",
      "   macro avg       0.93      0.93      0.93       120\n",
      "weighted avg       0.93      0.93      0.93       120\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Confusion Matrix\n",
    "cm_mfcc = confusion_matrix(y_test_mfcc,y1_pred)\n",
    "cm_display_mfcc = ConfusionMatrixDisplay(confusion_matrix = cm_mfcc, display_labels = ['Angry', 'Happy', 'Relaxed', 'Sad'])\n",
    "print(classification_report(y_test_mfcc, y1_pred, target_names = ['Angry', 'Happy', 'Relaxed', 'Sad']))\n",
    "cm_display_mfcc.plot()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "id": "87ef3dbe",
   "metadata": {},
   "outputs": [],
   "source": [
    "model1.save(os.getcwd() + \"/Features2D/Model/Conv2D_mfcc_agumented.h5\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7ca332de",
   "metadata": {},
   "source": [
    "### 4.2 Spectrogram "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "id": "aee04d80",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Load Spectrogram data set\n",
    "spec_file = np.load(\"Features2D/npz_new/spec_train_test_agumented.npz\")\n",
    "\n",
    "spec_train = spec_file['spec_train']\n",
    "y2_train = spec_file['y_train']\n",
    "\n",
    "spec_train,spec_val,y2_train,y2_val = train_test_split(spec_train,y2_train, \n",
    "                                                       train_size = 0.8,\n",
    "                                                       random_state = 13, \n",
    "                                                       stratify= y2_train)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "id": "0ebc91cf",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature Scaling\n",
    "spec_train_scaled = spec_train / np.amax(spec_train)\n",
    "spec_val_scaled = spec_val / np.amax(spec_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "id": "5f1f4078",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Reshape the data\n",
    "N,row,col = spec_train_scaled.shape\n",
    "spec_train_reshaped = spec_train_scaled.reshape((N,row,col,1))\n",
    "N,row,col = spec_val_scaled.shape\n",
    "spec_val_reshaped = spec_val_scaled.reshape((N,row,col,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "id": "e25a3e5e",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"CNN2D_2\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " Conv2D_1 (Conv2D)           (None, 300, 300, 5)       50        \n",
      "                                                                 \n",
      " MaxPooling2D_1 (MaxPooling2  (None, 150, 150, 5)      0         \n",
      " D)                                                              \n",
      "                                                                 \n",
      " BN1 (BatchNormalization)    (None, 150, 150, 5)       20        \n",
      "                                                                 \n",
      " Conv2D_2 (Conv2D)           (None, 150, 150, 8)       368       \n",
      "                                                                 \n",
      " MaxPooling2D_2 (MaxPooling2  (None, 75, 75, 8)        0         \n",
      " D)                                                              \n",
      "                                                                 \n",
      " BN2 (BatchNormalization)    (None, 75, 75, 8)         32        \n",
      "                                                                 \n",
      " Conv2D_3 (Conv2D)           (None, 75, 75, 16)        1168      \n",
      "                                                                 \n",
      " MaxPooling2D_3 (MaxPooling2  (None, 38, 38, 16)       0         \n",
      " D)                                                              \n",
      "                                                                 \n",
      " BN3 (BatchNormalization)    (None, 38, 38, 16)        64        \n",
      "                                                                 \n",
      " Flatten (Flatten)           (None, 23104)             0         \n",
      "                                                                 \n",
      " Dense1 (Dense)              (None, 64)                1478720   \n",
      "                                                                 \n",
      " Dropout1 (Dropout)          (None, 64)                0         \n",
      "                                                                 \n",
      " Dense2 (Dense)              (None, 32)                2080      \n",
      "                                                                 \n",
      " Dropout2 (Dropout)          (None, 32)                0         \n",
      "                                                                 \n",
      " Softmax (Dense)             (None, 4)                 132       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 1,482,634\n",
      "Trainable params: 1,482,576\n",
      "Non-trainable params: 58\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "f1,k1,a1 = 5,3,'relu'\n",
    "mw1,mh1,sw1,sh1 = 3,3,2,2\n",
    "f2,k2,a2 = 8,3,'relu'\n",
    "mw2,mh2,sw2,sh2 = 3,3,2,2\n",
    "f3,k3,a3 = 16,3,'relu'\n",
    "mw3,mh3,sw3,sh3 = 3,3,2,2\n",
    "d1,dr1,da1,r1 = 64,0.2,'relu',0.2\n",
    "d2,dr2,da2,r2 = 32,0.2,'relu',0.2\n",
    "num = 2\n",
    "model2 = modelBuilder3L(spec_train_reshaped,\n",
    "                        f1,k1,a1,\n",
    "                        mw1,mh1,sw1,sh1,\n",
    "                        f2,k2,a2,\n",
    "                        mw2,mh2,sw2,sh2,\n",
    "                        f3,k3,a3,\n",
    "                        mw3,mh3,sw3,sh3,\n",
    "                        d1,dr1,da1,r1,\n",
    "                        d2,dr2,da2,r2,\n",
    "                        num)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "id": "9a503680",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/80\n",
      "44/44 [==============================] - 26s 524ms/step - loss: 32.7881 - accuracy: 0.4537 - val_loss: 29.9270 - val_accuracy: 0.2546\n",
      "Epoch 2/80\n",
      "44/44 [==============================] - 23s 526ms/step - loss: 26.8439 - accuracy: 0.6238 - val_loss: 24.7061 - val_accuracy: 0.3287\n",
      "Epoch 3/80\n",
      "44/44 [==============================] - 21s 466ms/step - loss: 21.9600 - accuracy: 0.6933 - val_loss: 20.4431 - val_accuracy: 0.3519\n",
      "Epoch 4/80\n",
      "44/44 [==============================] - 20s 461ms/step - loss: 18.0310 - accuracy: 0.8090 - val_loss: 17.2313 - val_accuracy: 0.3380\n",
      "Epoch 5/80\n",
      "44/44 [==============================] - 20s 451ms/step - loss: 15.1121 - accuracy: 0.8090 - val_loss: 14.6813 - val_accuracy: 0.3056\n",
      "Epoch 6/80\n",
      "44/44 [==============================] - 20s 446ms/step - loss: 12.8502 - accuracy: 0.8484 - val_loss: 12.6879 - val_accuracy: 0.3426\n",
      "Epoch 7/80\n",
      "44/44 [==============================] - 20s 445ms/step - loss: 11.1097 - accuracy: 0.8646 - val_loss: 11.4661 - val_accuracy: 0.4259\n",
      "Epoch 8/80\n",
      "44/44 [==============================] - 20s 454ms/step - loss: 9.7345 - accuracy: 0.8924 - val_loss: 10.2984 - val_accuracy: 0.3981\n",
      "Epoch 9/80\n",
      "44/44 [==============================] - 20s 445ms/step - loss: 8.6448 - accuracy: 0.9178 - val_loss: 9.0333 - val_accuracy: 0.5741\n",
      "Epoch 10/80\n",
      "44/44 [==============================] - 20s 450ms/step - loss: 7.7489 - accuracy: 0.9444 - val_loss: 8.0752 - val_accuracy: 0.6250\n",
      "Epoch 11/80\n",
      "44/44 [==============================] - 20s 447ms/step - loss: 7.0209 - accuracy: 0.9375 - val_loss: 7.4097 - val_accuracy: 0.6389\n",
      "Epoch 12/80\n",
      "44/44 [==============================] - 21s 474ms/step - loss: 6.4025 - accuracy: 0.9514 - val_loss: 6.7313 - val_accuracy: 0.7130\n",
      "Epoch 13/80\n",
      "44/44 [==============================] - 20s 446ms/step - loss: 5.8926 - accuracy: 0.9641 - val_loss: 6.2611 - val_accuracy: 0.7500\n",
      "Epoch 14/80\n",
      "44/44 [==============================] - 20s 452ms/step - loss: 5.4480 - accuracy: 0.9572 - val_loss: 5.8068 - val_accuracy: 0.7593\n",
      "Epoch 15/80\n",
      "44/44 [==============================] - 20s 445ms/step - loss: 5.0609 - accuracy: 0.9780 - val_loss: 5.3331 - val_accuracy: 0.7870\n",
      "Epoch 16/80\n",
      "44/44 [==============================] - 19s 444ms/step - loss: 4.7340 - accuracy: 0.9664 - val_loss: 4.9012 - val_accuracy: 0.8194\n",
      "Epoch 17/80\n",
      "44/44 [==============================] - 20s 450ms/step - loss: 4.4847 - accuracy: 0.9606 - val_loss: 4.7781 - val_accuracy: 0.8148\n",
      "Epoch 18/80\n",
      "44/44 [==============================] - 20s 457ms/step - loss: 4.2368 - accuracy: 0.9676 - val_loss: 4.5421 - val_accuracy: 0.7685\n",
      "Epoch 19/80\n",
      "44/44 [==============================] - 20s 451ms/step - loss: 3.9545 - accuracy: 0.9780 - val_loss: 4.1666 - val_accuracy: 0.8333\n",
      "Epoch 20/80\n",
      "44/44 [==============================] - 20s 446ms/step - loss: 3.7219 - accuracy: 0.9769 - val_loss: 3.9499 - val_accuracy: 0.8102\n",
      "Epoch 21/80\n",
      "44/44 [==============================] - 20s 451ms/step - loss: 3.5512 - accuracy: 0.9676 - val_loss: 3.7233 - val_accuracy: 0.8472\n",
      "Epoch 22/80\n",
      "44/44 [==============================] - 19s 442ms/step - loss: 3.3395 - accuracy: 0.9815 - val_loss: 3.5874 - val_accuracy: 0.8009\n",
      "Epoch 23/80\n",
      "44/44 [==============================] - 20s 456ms/step - loss: 3.1566 - accuracy: 0.9838 - val_loss: 3.3881 - val_accuracy: 0.8102\n",
      "Epoch 24/80\n",
      "44/44 [==============================] - 20s 450ms/step - loss: 2.9848 - accuracy: 0.9826 - val_loss: 3.2399 - val_accuracy: 0.8426\n",
      "Epoch 25/80\n",
      "44/44 [==============================] - 20s 454ms/step - loss: 2.8467 - accuracy: 0.9711 - val_loss: 3.1438 - val_accuracy: 0.8241\n",
      "Epoch 26/80\n",
      "44/44 [==============================] - 20s 447ms/step - loss: 2.7146 - accuracy: 0.9850 - val_loss: 2.9963 - val_accuracy: 0.8148\n",
      "Epoch 27/80\n",
      "44/44 [==============================] - 19s 441ms/step - loss: 2.6004 - accuracy: 0.9676 - val_loss: 2.8515 - val_accuracy: 0.8565\n",
      "Epoch 28/80\n",
      "44/44 [==============================] - 20s 446ms/step - loss: 2.4578 - accuracy: 0.9850 - val_loss: 2.6797 - val_accuracy: 0.8333\n",
      "Epoch 29/80\n",
      "44/44 [==============================] - 20s 446ms/step - loss: 2.3196 - accuracy: 0.9861 - val_loss: 2.6146 - val_accuracy: 0.8194\n",
      "Epoch 30/80\n",
      "44/44 [==============================] - 19s 443ms/step - loss: 2.2217 - accuracy: 0.9769 - val_loss: 2.4651 - val_accuracy: 0.8287\n",
      "Epoch 31/80\n",
      "44/44 [==============================] - 20s 447ms/step - loss: 2.1217 - accuracy: 0.9792 - val_loss: 2.3536 - val_accuracy: 0.8565\n",
      "Epoch 32/80\n",
      "44/44 [==============================] - 20s 445ms/step - loss: 1.9882 - accuracy: 0.9850 - val_loss: 2.2530 - val_accuracy: 0.8287\n",
      "Epoch 33/80\n",
      "44/44 [==============================] - 20s 449ms/step - loss: 1.8993 - accuracy: 0.9780 - val_loss: 2.1314 - val_accuracy: 0.8426\n",
      "Epoch 34/80\n",
      "44/44 [==============================] - 20s 444ms/step - loss: 1.8082 - accuracy: 0.9826 - val_loss: 2.1538 - val_accuracy: 0.8009\n",
      "Epoch 35/80\n",
      "44/44 [==============================] - 20s 452ms/step - loss: 1.7393 - accuracy: 0.9815 - val_loss: 1.9860 - val_accuracy: 0.8565\n",
      "Epoch 36/80\n",
      "44/44 [==============================] - 21s 470ms/step - loss: 1.6406 - accuracy: 0.9861 - val_loss: 1.8685 - val_accuracy: 0.8380\n",
      "Epoch 37/80\n",
      "44/44 [==============================] - 21s 484ms/step - loss: 1.5412 - accuracy: 0.9861 - val_loss: 1.8005 - val_accuracy: 0.8565\n",
      "Epoch 38/80\n",
      "44/44 [==============================] - 20s 459ms/step - loss: 1.5135 - accuracy: 0.9676 - val_loss: 1.7631 - val_accuracy: 0.8380\n",
      "Epoch 39/80\n",
      "44/44 [==============================] - 20s 450ms/step - loss: 1.4332 - accuracy: 0.9792 - val_loss: 1.7479 - val_accuracy: 0.8194\n",
      "Epoch 40/80\n",
      "44/44 [==============================] - 19s 437ms/step - loss: 1.3704 - accuracy: 0.9803 - val_loss: 1.6060 - val_accuracy: 0.8472\n",
      "Epoch 41/80\n",
      "44/44 [==============================] - 20s 455ms/step - loss: 1.2877 - accuracy: 0.9803 - val_loss: 1.5979 - val_accuracy: 0.8287\n",
      "Epoch 42/80\n",
      "44/44 [==============================] - 20s 449ms/step - loss: 1.2244 - accuracy: 0.9873 - val_loss: 1.4878 - val_accuracy: 0.8565\n",
      "Epoch 43/80\n",
      "44/44 [==============================] - 23s 528ms/step - loss: 1.1680 - accuracy: 0.9861 - val_loss: 1.5298 - val_accuracy: 0.7685\n",
      "Epoch 44/80\n",
      "44/44 [==============================] - 22s 498ms/step - loss: 1.1128 - accuracy: 0.9826 - val_loss: 1.3556 - val_accuracy: 0.8704\n",
      "Epoch 45/80\n",
      "44/44 [==============================] - 20s 445ms/step - loss: 1.1388 - accuracy: 0.9699 - val_loss: 1.3936 - val_accuracy: 0.8426\n",
      "Epoch 46/80\n",
      "44/44 [==============================] - 19s 442ms/step - loss: 1.0500 - accuracy: 0.9861 - val_loss: 1.3078 - val_accuracy: 0.8565\n",
      "Epoch 47/80\n",
      "44/44 [==============================] - 19s 442ms/step - loss: 0.9854 - accuracy: 0.9838 - val_loss: 1.2249 - val_accuracy: 0.8565\n",
      "Epoch 48/80\n",
      "44/44 [==============================] - 20s 449ms/step - loss: 0.9968 - accuracy: 0.9769 - val_loss: 1.3113 - val_accuracy: 0.8194\n",
      "Epoch 49/80\n",
      "44/44 [==============================] - 19s 441ms/step - loss: 0.9381 - accuracy: 0.9838 - val_loss: 1.1686 - val_accuracy: 0.8565\n",
      "Epoch 50/80\n",
      "44/44 [==============================] - 24s 537ms/step - loss: 0.8811 - accuracy: 0.9850 - val_loss: 1.1091 - val_accuracy: 0.8657\n",
      "Epoch 51/80\n",
      "44/44 [==============================] - 21s 481ms/step - loss: 0.8273 - accuracy: 0.9861 - val_loss: 1.1272 - val_accuracy: 0.8472\n",
      "Epoch 52/80\n",
      "44/44 [==============================] - 26s 586ms/step - loss: 0.7916 - accuracy: 0.9861 - val_loss: 1.0684 - val_accuracy: 0.8472\n",
      "Epoch 53/80\n",
      "44/44 [==============================] - 21s 469ms/step - loss: 0.7634 - accuracy: 0.9896 - val_loss: 1.0558 - val_accuracy: 0.8519\n",
      "Epoch 54/80\n",
      "44/44 [==============================] - 20s 457ms/step - loss: 0.7204 - accuracy: 0.9896 - val_loss: 1.0021 - val_accuracy: 0.8380\n",
      "Epoch 55/80\n",
      "44/44 [==============================] - 20s 457ms/step - loss: 0.6865 - accuracy: 0.9907 - val_loss: 1.0070 - val_accuracy: 0.8472\n",
      "Epoch 56/80\n",
      "44/44 [==============================] - 20s 455ms/step - loss: 0.6703 - accuracy: 0.9861 - val_loss: 0.9349 - val_accuracy: 0.8519\n",
      "Epoch 57/80\n",
      "44/44 [==============================] - 21s 478ms/step - loss: 0.7127 - accuracy: 0.9803 - val_loss: 1.0677 - val_accuracy: 0.8241\n",
      "Epoch 58/80\n",
      "44/44 [==============================] - 20s 462ms/step - loss: 0.6644 - accuracy: 0.9838 - val_loss: 0.9140 - val_accuracy: 0.8565\n",
      "Epoch 59/80\n",
      "44/44 [==============================] - 20s 447ms/step - loss: 0.6209 - accuracy: 0.9884 - val_loss: 0.8988 - val_accuracy: 0.8426\n",
      "Epoch 60/80\n",
      "44/44 [==============================] - 20s 453ms/step - loss: 0.6031 - accuracy: 0.9884 - val_loss: 0.9014 - val_accuracy: 0.8333\n",
      "Epoch 61/80\n",
      "44/44 [==============================] - 19s 443ms/step - loss: 0.5635 - accuracy: 0.9919 - val_loss: 0.8172 - val_accuracy: 0.8472\n",
      "Epoch 62/80\n",
      "44/44 [==============================] - 20s 449ms/step - loss: 0.5255 - accuracy: 0.9977 - val_loss: 0.7742 - val_accuracy: 0.8611\n",
      "Epoch 63/80\n",
      "44/44 [==============================] - 20s 451ms/step - loss: 0.5242 - accuracy: 0.9884 - val_loss: 0.8739 - val_accuracy: 0.8148\n",
      "Epoch 64/80\n",
      "44/44 [==============================] - 20s 446ms/step - loss: 0.5254 - accuracy: 0.9884 - val_loss: 0.8502 - val_accuracy: 0.8519\n",
      "Epoch 65/80\n",
      "44/44 [==============================] - 20s 451ms/step - loss: 0.5263 - accuracy: 0.9896 - val_loss: 0.8317 - val_accuracy: 0.8519\n",
      "Epoch 66/80\n",
      "44/44 [==============================] - 20s 463ms/step - loss: 0.4821 - accuracy: 0.9931 - val_loss: 0.7532 - val_accuracy: 0.8565\n",
      "Epoch 67/80\n",
      "44/44 [==============================] - 22s 492ms/step - loss: 0.4780 - accuracy: 0.9850 - val_loss: 0.7665 - val_accuracy: 0.8472\n",
      "Epoch 68/80\n",
      "44/44 [==============================] - 20s 460ms/step - loss: 0.5124 - accuracy: 0.9838 - val_loss: 0.7671 - val_accuracy: 0.8565\n",
      "Epoch 69/80\n",
      "44/44 [==============================] - 20s 453ms/step - loss: 0.4718 - accuracy: 0.9931 - val_loss: 0.8274 - val_accuracy: 0.8287\n",
      "Epoch 70/80\n",
      "44/44 [==============================] - 20s 444ms/step - loss: 0.4445 - accuracy: 0.9907 - val_loss: 0.7528 - val_accuracy: 0.8380\n",
      "Epoch 71/80\n",
      "44/44 [==============================] - 19s 439ms/step - loss: 0.4377 - accuracy: 0.9850 - val_loss: 0.6949 - val_accuracy: 0.8704\n",
      "Epoch 72/80\n",
      "44/44 [==============================] - 20s 449ms/step - loss: 0.4446 - accuracy: 0.9896 - val_loss: 0.7827 - val_accuracy: 0.8426\n",
      "Epoch 73/80\n",
      "44/44 [==============================] - 20s 443ms/step - loss: 0.4334 - accuracy: 0.9931 - val_loss: 0.7121 - val_accuracy: 0.8519\n",
      "Epoch 74/80\n",
      "44/44 [==============================] - 20s 459ms/step - loss: 0.4151 - accuracy: 0.9954 - val_loss: 0.6651 - val_accuracy: 0.8704\n",
      "Epoch 75/80\n",
      "44/44 [==============================] - 20s 457ms/step - loss: 0.4144 - accuracy: 0.9919 - val_loss: 0.6535 - val_accuracy: 0.8611\n",
      "Epoch 76/80\n",
      "44/44 [==============================] - 20s 451ms/step - loss: 0.4118 - accuracy: 0.9884 - val_loss: 0.7293 - val_accuracy: 0.8426\n",
      "Epoch 77/80\n",
      "44/44 [==============================] - 20s 457ms/step - loss: 0.3961 - accuracy: 0.9942 - val_loss: 0.6335 - val_accuracy: 0.8843\n",
      "Epoch 78/80\n",
      "44/44 [==============================] - 20s 454ms/step - loss: 0.3948 - accuracy: 0.9884 - val_loss: 0.7043 - val_accuracy: 0.8565\n",
      "Epoch 79/80\n",
      "44/44 [==============================] - 20s 461ms/step - loss: 0.3924 - accuracy: 0.9815 - val_loss: 0.7071 - val_accuracy: 0.8519\n",
      "Epoch 80/80\n",
      "44/44 [==============================] - 20s 448ms/step - loss: 0.3997 - accuracy: 0.9884 - val_loss: 0.6878 - val_accuracy: 0.8657\n"
     ]
    }
   ],
   "source": [
    "history2 = model2.fit(spec_train_reshaped,y2_train,validation_data=(spec_val_reshaped,y2_val),batch_size=20,epochs=80)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "id": "0a4637d3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'loss')"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA+kAAAGJCAYAAAD2VnIMAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjcuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/bCgiHAAAACXBIWXMAAA9hAAAPYQGoP6dpAADCHElEQVR4nOzdd3xTZfvH8U/SXbqAlrZAoey9ZMlGRREQcKOoCCo4wMXj4nH7qPxcuEBRH0H0EUUUcYAgIqBs2XtvaEsLdO/m/P44TUpoC6UtJC3f9+uVV9KTc5IraSG5zn3d120xDMNARERERERERFzO6uoARERERERERMSkJF1ERERERETETShJFxEREREREXETStJFRERERERE3ISSdBERERERERE3oSRdRERERERExE0oSRcRERERERFxE0rSRURERERERNyEknQRERERERERN6EkXSRfdHQ0w4cPd3UYxXrppZewWCxO20oa8xdffIHFYuHAgQPlFs+BAwewWCx88cUX5faYIiIiFY2+P5wfV35/GD58ONHR0Rf9eUXOl5J0qTCWL1/OSy+9RGJioqtDuaRMnz6d9957z9VhiIiIlIq+P7iGvj+IlJ6nqwMQKanly5fz8ssvM3z4cEJCQsr98Xfu3InVWrHOW12MmKdPn86WLVt47LHHnLbXrVuXjIwMvLy8Lujzi4iIlIW+PxSm7w8i7k1JulRKNpuN7OxsfH19S3yMj4/PBYzownBlzBaL5bze30uVYRhkZmbi5+fn6lBEROQc9P3hwtP3B5Fzq1in/eSS9dJLL/Hkk08CUK9ePSwWi9McKYvFwpgxY/j6669p0aIFPj4+zJs3D4C3336brl27Ur16dfz8/Gjfvj3ff/99oec4c36WfR7WsmXLGDt2LGFhYVSpUoUbbriB+Pj4s8b79ttvY7FYOHjwYKH7xo0bh7e3N6dOnQLg77//5pZbbqFOnTr4+PgQFRXF448/TkZGxjnfl6LmlG3dupUrr7wSPz8/ateuzauvvorNZit07E8//cSAAQOoWbMmPj4+NGjQgP/85z/k5eU59unduzdz5szh4MGDjvfcPperuDllf/75Jz169KBKlSqEhIQwePBgtm/f7rSPfX7cnj17HCMbwcHBjBgxgvT09HO+7vN5z3bs2MGtt95KWFgYfn5+NGnShGeffdZpn6NHj3Lvvfc63ot69erx4IMPkp2d7RTvmYqaqxcdHc11113H/Pnz6dChA35+fnzyyScATJ06lSuvvJIaNWrg4+ND8+bN+fjjj4t8jb/99hu9evUiMDCQoKAgOnbsyPTp0wF48cUX8fLyKvLvcNSoUYSEhJCZmXnO91FEpLLT94eiXarfH4qSlpbGv/71L6KiovDx8aFJkya8/fbbGIbhtN+CBQvo3r07ISEhBAQE0KRJE/7973877fPhhx/SokUL/P39qVq1Kh06dHB8doucD42kS4Vw4403smvXLr755hveffddQkNDAQgLC3Ps8+eff/Ldd98xZswYQkNDHR8G77//PoMGDeKOO+4gOzubb7/9lltuuYVff/2VAQMGnPO5H374YapWrcqLL77IgQMHeO+99xgzZgwzZswo9phbb72Vp556iu+++87x5cDuu+++45prrqFq1aoAzJw5k/T0dB588EGqV6/O6tWr+fDDDzly5AgzZ848r/cpNjaWK664gtzcXJ555hmqVKnCp59+WuQo7hdffEFAQABjx44lICCAP//8kxdeeIHk5GTeeustAJ599lmSkpI4cuQI7777LgABAQHFPv8ff/xBv379qF+/Pi+99BIZGRl8+OGHdOvWjXXr1hVq1nLrrbdSr149xo8fz7p16/jvf/9LjRo1eOONN876Okv6nm3atIkePXrg5eXFqFGjiI6OZu/evfzyyy+89tprABw7doxOnTqRmJjIqFGjaNq0KUePHuX7778nPT0db2/vEr33p9u5cye33347999/PyNHjqRJkyYAfPzxx7Ro0YJBgwbh6enJL7/8wkMPPYTNZmP06NGO47/44gvuueceWrRowbhx4wgJCWH9+vXMmzePoUOHctddd/HKK68wY8YMxowZ4zguOzub77//nptuukmjFCIi6PtDSV0q3x/OZBgGgwYNYtGiRdx77720bduW+fPn8+STT3L06FFH7Fu3buW6666jdevWvPLKK/j4+LBnzx6WLVvmeKzPPvuMRx55hJtvvplHH32UzMxMNm3axKpVqxg6dOh5xSWCIVJBvPXWWwZg7N+/v9B9gGG1Wo2tW7cWui89Pd3p5+zsbKNly5bGlVde6bS9bt26xt133+34eerUqQZg9OnTx7DZbI7tjz/+uOHh4WEkJiaeNd4uXboY7du3d9q2evVqAzC+/PLLYuMzDMMYP368YbFYjIMHDzq2vfjii8aZ/2TPjPmxxx4zAGPVqlWObcePHzeCg4MLvXdFPe/9999v+Pv7G5mZmY5tAwYMMOrWrVto3/379xuAMXXqVMe2tm3bGjVq1DBOnDjh2LZx40bDarUaw4YNK/Ra7rnnHqfHvOGGG4zq1asXeq4zlfQ969mzpxEYGOi0zTAMp9/nsGHDDKvVavzzzz+FHtO+X1HvvWEU/I2c/r7WrVvXAIx58+aVKO6+ffsa9evXd/ycmJhoBAYGGp07dzYyMjKKjbtLly5G586dne6fNWuWARiLFi0q9DwiIpcqfX/Q9we7u+++2ymm2bNnG4Dx6quvOu138803GxaLxdizZ49hGIbx7rvvGoARHx9f7GMPHjzYaNGixTljECkJlbtLpdGrVy+aN29eaPvpZ4FPnTpFUlISPXr0YN26dSV63FGjRjmVOvfo0YO8vLwiS9FON2TIENauXcvevXsd22bMmIGPjw+DBw8uMr60tDQSEhLo2rUrhmGwfv36EsVoN3fuXC6//HI6derk2BYWFsYdd9xRaN/TnzclJYWEhAR69OhBeno6O3bsOK/nBYiJiWHDhg0MHz6catWqOba3bt2aq6++mrlz5xY65oEHHnD6uUePHpw4cYLk5OSzPldJ3rP4+Hj++usv7rnnHurUqeN0vP33abPZmD17NgMHDqRDhw6FnqeoEveSqFevHn379j1r3ElJSSQkJNCrVy/27dtHUlISYJbTpaSk8MwzzxQaDT89nmHDhrFq1Sqnv6+vv/6aqKgoevXqVaq4RUQuRfr+cOl8fzjT3Llz8fDw4JFHHnHa/q9//QvDMPjtt98AHA0Hf/rppyKnANj3OXLkCP/88895xSBSFCXpUmnUq1evyO2//vorl19+Ob6+vlSrVo2wsDA+/vhjR1J0LmcmePYyM/ucsOLccsstWK1WR1mbYRjMnDmTfv36ERQU5Njv0KFDjg+mgIAAwsLCHElWSWO0O3jwII0aNSq03V5ufbqtW7dyww03EBwcTFBQEGFhYdx5552lel77cxf3XM2aNSMhIYG0tDSn7aV9b0vynu3btw+Ali1bFvs48fHxJCcnn3Wf0ijub3HZsmX06dPHMd8uLCzMMZ/NHrf9S9m5YhoyZAg+Pj58/fXXjuN//fVX7rjjjlKfXBARuRTp+8Ol8/2hqOeuWbMmgYGBhZ739NiGDBlCt27duO+++wgPD+e2227ju+++c0rYn376aQICAujUqRONGjVi9OjRTuXwIudDSbpUGkXNm/r7778ZNGgQvr6+fPTRR8ydO5cFCxYwdOjQQg1BiuPh4VHk9nMdX7NmTXr06MF3330HwMqVKzl06BBDhgxx7JOXl8fVV1/NnDlzePrpp5k9ezYLFixwNFMp7mxtWSUmJtKrVy82btzIK6+8wi+//MKCBQscc7ku1POeqTTvrSves+KS3tOb5JyuqL/FvXv3ctVVV5GQkMCECROYM2cOCxYs4PHHHwfOP+6qVaty3XXXOZL077//nqysLMcXJRERKRl9fyi5ivz9oSz8/Pz466+/+OOPP7jrrrvYtGkTQ4YM4eqrr3Z8F2jWrBk7d+7k22+/pXv37vzwww90796dF1988YLEJJWbGsdJhVGa0cEffvgBX19f5s+f77TcyNSpU8sztGINGTKEhx56iJ07dzJjxgz8/f0ZOHCg4/7Nmzeza9cupk2bxrBhwxzbFyxYUKrnq1u3Lrt37y60fefOnU4/L168mBMnTjBr1ix69uzp2L5///5Cx5b0fa9bt26RzwVmh/XQ0FCqVKlSosc6m5K+Z/Xr1wdgy5YtxT5WWFgYQUFBZ90HCs7QJyYmOq2xe66SxdP98ssvZGVl8fPPPzuNACxatMhpvwYNGjjibtiw4Vkfc9iwYQwePJh//vmHr7/+mnbt2tGiRYsSxyQicinQ94dzuxS+PxT33H/88QcpKSlOo+n2sn17bABWq5WrrrqKq666igkTJvD666/z7LPPsmjRIvr06QNAlSpVGDJkCEOGDCE7O5sbb7yR1157jXHjxqmhq5wXjaRLhWH/DzoxMbHEx3h4eGCxWJxGPA8cOMDs2bPLObqi3XTTTXh4ePDNN98wc+ZMrrvuOqcPGvuZ4NPP/BqGwfvvv1+q5+vfvz8rV65k9erVjm3x8fGO0dazPW92djYfffRRocesUqVKicrXIiMjadu2LdOmTXP6HW3ZsoXff/+d/v37n+/LKVJJ37OwsDB69uzJlClTOHTokNN99mOtVivXX389v/zyC2vWrCn0XPb97InzX3/95bgvLS2NadOmlSnupKSkQl/4rrnmGgIDAxk/fnyhZdTOHCHo168foaGhvPHGGyxZskSj6CIiRdD3h3O7FL4/FKV///7k5eUxceJEp+3vvvsuFouFfv36AXDy5MlCx7Zt2xaArKwsAE6cOOF0v7e3N82bN8cwDHJyci5A9FKZaSRdKoz27dsD5rIet912G15eXgwcOPCsZ1cHDBjAhAkTuPbaaxk6dCjHjx9n0qRJNGzYkE2bNl3wmGvUqMEVV1zBhAkTSElJcSpVA2jatCkNGjTgiSee4OjRowQFBfHDDz+c95wqu6eeeoqvvvqKa6+9lkcffdSxhErdunWdXm/Xrl2pWrUqd999N4888ggWi4WvvvqqyDKx9u3bM2PGDMaOHUvHjh0JCAhwOpt/urfeeot+/frRpUsX7r33XscSKsHBwbz00kulek1nOp/37IMPPqB79+5cdtlljBo1inr16nHgwAHmzJnDhg0bAHj99df5/fff6dWrF6NGjaJZs2bExMQwc+ZMli5dSkhICNdccw116tTh3nvv5cknn8TDw4MpU6YQFhZW6ARAca655hq8vb0ZOHAg999/P6mpqXz22WfUqFGDmJgYx35BQUG8++673HfffXTs2JGhQ4dStWpVNm7cSHp6utOJAS8vL2677TYmTpyIh4cHt99+e9neXBGRSkjfH87tUvj+UJSBAwdyxRVX8Oyzz3LgwAHatGnD77//zk8//cRjjz3mOEn/yiuv8NdffzFgwADq1q3L8ePH+eijj6hduzbdu3cHzM/5iIgIunXrRnh4ONu3b2fixIkMGDCg0Jx3kXO6SF3kRcrFf/7zH6NWrVqG1Wp1WhIEMEaPHl3kMZ9//rnRqFEjw8fHx2jatKkxderUEi1HYl9C5cyluRYtWnRey1x99tlnBmAEBgYWWlLLMAxj27ZtRp8+fYyAgAAjNDTUGDlypLFx48ZCy5OUJGbDMIxNmzYZvXr1Mnx9fY1atWoZ//nPf4zPP/+80BIqy5YtMy6//HLDz8/PqFmzpvHUU08Z8+fPL/TaUlNTjaFDhxohISEG4Fi6pKglVAzDMP744w+jW7duhp+fnxEUFGQMHDjQ2LZtm9M+9tdy5lImRS1pVpSSvmeGYRhbtmwxbrjhBiMkJMTw9fU1mjRpYjz//PNO+xw8eNAYNmyYERYWZvj4+Bj169c3Ro8ebWRlZTn2Wbt2rdG5c2fD29vbqFOnjjFhwoRil2AbMGBAkXH//PPPRuvWrQ1fX18jOjraeOONN4wpU6YU+Zp//vlno2vXro73sVOnTsY333xT6DHty/Jcc801Z33PREQuZfr+oO8PhlF4CTbDMIyUlBTj8ccfN2rWrGl4eXkZjRo1Mt566y2n5fMWLlxoDB482KhZs6bh7e1t1KxZ07j99tuNXbt2Ofb55JNPjJ49exrVq1c3fHx8jAYNGhhPPvmkkZSUdNaYRIpiMYwL1GFBREQuuI0bN9K2bVu+/PJL7rrrLleHIyIiIiJlpDnpIiIV2GeffUZAQAA33nijq0MRERERkXKgOekiIhXQL7/8wrZt2/j0008ZM2bMBet8KyIiIiIXl8rdRUQqoOjoaOLi4ujbty9fffWVmtKIiIiIVBJK0kVERERERETchOaki4iIiIiIiLgJJekiIiIiIiIibuKSaxxns9k4duwYgYGBWCwWV4cjIiKCYRikpKRQs2ZNrFadPy8P+rwXERF3cj6f9Zdckn7s2DGioqJcHYaIiEghhw8fpnbt2q4Oo1LQ572IiLijknzWX3JJur0D8uHDhwkKCnJxNCIiIpCcnExUVJS69Jcjfd6LiIg7OZ/P+ksuSbeXvAUFBelDW0RE3IrKssuPPu9FRMQdleSzXhPfRERERERERNyEknQRERERERERN6EkXURERERERMRNXHJz0kVERERERNxBXl4eOTk5rg5DyomXlxceHh5lfhwl6SIiIiIiIhdZamoqR44cwTAMV4ci5cRisVC7dm0CAgLK9DhK0kVERERERC6ivLw8jhw5gr+/P2FhYVrdoxIwDIP4+HiOHDlCo0aNyjSi7tIk/a+//uKtt95i7dq1xMTE8OOPP3L99def9ZjFixczduxYtm7dSlRUFM899xzDhw+/KPGKiIiIiIiUVU5ODoZhEBYWhp+fn6vDkXISFhbGgQMHyMnJKVOS7tLGcWlpabRp04ZJkyaVaP/9+/czYMAArrjiCjZs2MBjjz3Gfffdx/z58y9wpCIiIiIiIuVLI+iVS3n9Pl06kt6vXz/69etX4v0nT55MvXr1eOeddwBo1qwZS5cu5d1336Vv374XKkwRERERERGRi6JCLcG2YsUK+vTp47Stb9++rFixothjsrKySE5OdrqIiFwKYpIyWHfolBrSiJRBQmoWf2yLY8mueFeHIiIil4gKlaTHxsYSHh7utC08PJzk5GQyMjKKPGb8+PEEBwc7LlFRURcjVBERl9l8JIlHvllP9zcWceNHy3n02w0kZ1bO5V3iU7L4e3c8/1t5kJX7TpCda3N1SFLJrD+UyH1frmHC7ztdHYqISKUSHR3Ne++95+ow3FKl7+4+btw4xo4d6/g5OTlZibqIVDq5eTYW7Yzns7/3sXr/Scd2iwV+3niMtQdP8f5tbekQXe28Hzs1K5dV+06w5Wgyl9evRuf61c/r+NX7T/L37njqVPOnWWQQDWsE4Ot1/s1UMrLzWLX/BCv2nmDrsWR2xCaTkJrttI+/twed61Wje6MwejQKpVGNgCLnh+Xk2Zi7OYYpS/dzPCWL69vVYnjXaMKDfM87LqncwgJ9APOEkIjIpa537960bdu2XJLrf/75hypVqpQ9qEqoQiXpERERxMXFOW2Li4sjKCio2K6IPj4++Pj4XIzwRKSSS0rPYWtMErYzBmuz8/JISM0mITWLhBTz2tvTSv9WEfRoFIaXR/kXLRmGwcET6fy9J4Glu+NZvvcEKZm5AHhaLQxsU5N7u9cjO8/Go9+u5/DJDG79ZAUPX9mIh69siGcRMRmGQXJGLvGpWRxPzmTNwVMs3Z3AukOnyLUVlMzfeFktnhvQnGpVvM8aY3aujbd/38mnf+1z2m61QL3QKlxWpyrDukTTqnZwsa9x67FkluyKZ+nuBNYePEV2nvObb7FAvepViKrmz9ZjSSSkZrNoZzyLdpqlyTUCfejeMJTujULp3jAUHy8Pvl19iC+WHyAmKdPxOB8v3st//97HwDY1ua97fZrXDDrrawOw2Qwyc/Pw965QH6VynhxJemoWhmGoyZOIyFkYhkFeXh6enuf+bAwLC7sIEVVMFeqbRZcuXZg7d67TtgULFtClSxcXRSQiRTEMg5X7TnI8JZPQAJ/8izdV/b2xWivOF9zsXBvrDpmJ6t97Eth8JBHbeUzv/n7tEapV8ea61pFc364W7aJCzusL/sETacxef4w/d8SRnp3ndF9KZi6xyZlO20L8vbitYx3u7lqXyOCCE5dzH+nBiz9tZdb6o7y/cDffrTlMgE/Bf/8GkJaVy4nU7EJJsF2dav40qhHAnzuPM2vdURbtOM6/+zfj5va1i3xNe+NTefTb9Ww5avYB6dMsnLSsXLbHJpOYnsPe+DT2xqcxc+0ROterxsge9bmyaQ2sVguHT6Yze/1RZm84yt74NKfHrRXiR/eGoVxWN4SmEUE0Dg/Ez9sclbfZDHbEprB0Tzx/705g9f6THE/JYtb6o8xafxQAb0+royQ+NMCH4V3rUj8sgC+WHWD1gZPMWneUWeuO0qNRKE9c04Q2USGFXpthGCzZFc8b83ZyWZ0QXruhVZHvmVQOoQHmyaicPIOkjBxC/M9+ckpEpDQMwyAjJ+/cO14Afl4eJfp+Mnz4cJYsWcKSJUt4//33AZg6dSojRoxg7ty5PPfcc2zevJnff/+dqKgoxo4dy8qVK0lLS6NZs2aMHz/eqb9YdHQ0jz32GI899hhgdkb/7LPPmDNnDvPnz6dWrVq88847DBo06IK8bnfm0iQ9NTWVPXv2OH7ev38/GzZsoFq1atSpU4dx48Zx9OhRvvzySwAeeOABJk6cyFNPPcU999zDn3/+yXfffcecOXNc9RKkgjAMg+TMXKp4exQ5gng2aw6YX/T7tYwo9j8wm81g1vqjrD90yhxNzR9VTcnM5Zrm4Tx3XXOnpOhcsS7bc4JDJ9O5tmXEOUcrz/Y46w8nMnv9UeZtiXWMstp5Wi3865rGDO9Wr1SPX5w1B07yf7/tYM3BU4Xu87BaaBIe6BjV7FSvWonKnqcu28+kRXuIquZP04hAmkYE0TQikGY1gwjy9Trrsdm5No4lZuT/XrKIT80mISXL8fOJ/N/VibRscvOcM/CcPJvTCDKYyaq/t3PMnh4WqlfJPxkR6E1YgA9HEzP4ZWMMCalZfLniIF+uOEj9sCrc060eN11W25FYnulkWjZzNh3jx/VHWXco8ayvzcvDQrs6VemRP1LcunYIHkWcBAn09WLCkLb0bBzGc7O3OI0gFyXQx5PQQB+aRpi/qx4Nw6hT3R+AdYdO8e9Zm9kRm8KT32/ih3VH6NMsvOBkTKA36w8l8sov28jIySPE34s3bmpN3xYRgPl3eTwli23Hkvlpw1F+3RTDqv0nWbX/JPXDqlDN39vpb8fH00rPxmbpeveGodQLrVLsv0Or1ULzmkE0rxnEqJ4NyMzJY+3BU/y9O4Gle+LZeiyZ7FwbTcIDubdHPQa3rYmPp/l76N8qko2HE/ns7338tiWWv3cn8PfuBAa0iuSJvk2oF2qW4204nMgbv+1gxb4TAMQmZTCuf7MS//uWisfH04MQfy8S03M4npKlJF1ELoiMnDyav+CaZaW3vdK3RFVh77//Prt27aJly5a88sorAGzduhWAZ555hrfffpv69etTtWpVDh8+TP/+/Xnttdfw8fHhyy+/ZODAgezcuZM6deoU+xwvv/wyb775Jm+99RYffvghd9xxBwcPHqRatfOfrleRWQwXtv1dvHgxV1xxRaHtd999N1988QXDhw/nwIEDLF682OmYxx9/nG3btlG7dm2ef/55hg8fXuLnTE5OJjg4mKSkJIKCzl3OKBXTrrgUNhxKZHtsMjtiUtgRm8yp9BwsFqjm7+1IJupWr8KYKxpSM6To6RLzt8by4P/WYjPgjZtaMaRj0f+pfPbXPl6bu73YeOpW9+f929rRtohRObus3Dx+2RjDf//ex47YFMBMlh7o3YAR3aIL/ee5MzaFnzceJSElm+oB5muqHuBNtSrerDlwitkbjnLwRPpZ3ydPq4XZo7vRslbR5cbnY3dcCm/M28kf280pKb5eVtrUDuFkmpkEn0ov3LjM29NKx+iqjLmiEV0aFD3PefHO44z44h+K+p/K28PKHZfXYcwVDake4DytJSs3j+mrDjHxzz2cSMsufHAJVa/iTbf8JLhHo1CnEepzyc2zsWzvCceJEvsZ8qr+Xtx5eV3u6lKXYD8v1h1MZNmewqP1Vgt0axjKoDY1qVXV+Xm9PKw0jwyiynkmhyfTstkRW3iVC39vT0Lz/47OdeIkJ8/G50v3894fu8jMKb5RW9cG1Zlwa1sigouf5x2TlMEXyw4wffUhx4kkiwW6NQjl+na16NsinMBznIgpqZNp2cSnZNE4vOh56naHT6bz7h+7+HH9UQzD/HcypGMUp9Kzmbs5FjD/9oZ1qcvoKxpStZQn0k6nz6byV57v6dUTlrD7eCpf39eZbg1DyylCEbmUZWZmsn//furVq4evry/p2blun6RD4Tnp9nxu9uzZDB48+KzHtmzZkgceeIAxY8YARY+kP/fcc/znP/8BIC0tjYCAAH777Teuvfba0r24i+zM3+vpzudzyaVJuivoi1DlN2nRHt6aX/IuvBFBvky7pxNNIgKdti/bk8CIqf84yn99vaz8NLp7of3WHTrFrZNXkGszuL1TFM0igxyjiqlZOTw/eytHEzPwtFp4/OrGPNCrgWO0MzMnjz3HU1myK55pyw9wPL8xkb+3B5HBvo5S37BAHx7r04hejcOYuzmGH9cfY3vMuZcT9PPy4NqWEQxuW5MGYQFO970+dzu/bYmlYY0Afn24e5GJWXJmDot2HCcuOdOsDkjJIj41i9Qs51F5m81g89EkbIY5Wn5rhyge69PIqQlXTp6N4ylZrDlwkqW7E1i6J8ExouvtYWXyXZdxZVPn1RsOnkhj4IdLSc7M5abLatO7SRg7HCdeUjiaaK7qEODjyf0963Nvj3r4enrwy6ZjvP37Tg6fzHC8D2GBPo6TGaEBPoQFeBMa6OP4uXqAN95nVFl4WC1EBPmWS4l+alYuM9ccZsqy/Y64vD2seFgthcrbWtYK4vq2tRjUpiY13LiR2aET6Xy96iDHkjI5kZrlqCKxGQYP9GrAqB71S/zepWbl8vOGY2Tn5nFty8izJvYXy/aYZN6ct8Mxvx3MEwg3tqvN41c3onZV/3J7Ln02lb/yfE+HfraS5XtP8N6QtlzfrlY5RSgil7Izk7mKUO4OxSfpR44coVatgv8fU1NTeemll5gzZw4xMTHk5uaSkZHBv/71L958802g6CT9u+++45ZbbnE8TnBwMB9++CHDhg0rnxd7gZVXkq76PCk3m44k8vxPW7m+bU2Gd40uU3OdE6lZPP/TFjYdSaK6PaHKT6a6NQwtdtT1fysPOhL0zvWq0bJWsFkWHRlEvdAqpGfnFZQ9p2Tx0eK97Dmeyi2Tl/PZsA6OrtUbDicy8ss1ZOfZuLZFBOk5efy1K57R09fx85hujrONSek5PDx9Pbk2g+taR/L6Da0Kve72j1bj3z9uZs6mGN6av5PFO48THuTLjtgU9iekkXdaOXV4kA/Du9ZjaKc6BPp6OiWbz/64xelxvTwsXNGkBi1rBZsjhKlZJKSYZdu1q/pxfdtaXN08vNiR1tdvaMXag6fYczyV//ttBy8NauF0/8ETaQybsvqco/Gnu7ZFBE/0bULDGgGF7vPysFIrxI9abWsxuG0tDMNgb3wab87bwe/b4rj/q7V8fEd7+jQ3E/W0rFxGfbmW5Mxc2tUJ4fUbW+Lj6cHANjUdj/n37njemLeDLUeTeWfBLqatOEhogLejEqFGoA+P9WnMrR1qn/c0h/IW4OPJiG71GNYlmvlbY/ns732sP5QIeeac124NQx0Nzs5ntN6V6lT3Z1z/ZuXyWAE+ngztXHz5mys0iwxi6ohOrNx3gg8W7ibQ15PHr25M0wgl0ZcadXgXkQvNYrFU6EakZ3Zpf+KJJ1iwYAFvv/02DRs2xM/Pj5tvvpns7LNXN3p5OVfPWSwWbGd27L0EVNy/BHEr2bk2xn63kT3HU9l4OJHDJzN4bkCzUo1Abj2WxKgv1zpGSY+cynC6f+KiPQzrUpd/92/mNPr788ZjPP+TmciOuaIhT/RtUuixq/h4Or5sAVzZtAb3TlvD2oOnuGvKat4f0pYGNQIYPnU16dl5dG8Yyvu3tyUlM5f+7//NnuOpvPjTVt66pQ2GYfDk9xs5mphB3er+jL+xcIIOEOznxcTb29G7cRgv/ryVfw44z9UO8feieWQQN7evzXWta+LtWZBMDm5bi34tI5m+6iAf/LmHk2nZdIquxuB2NRnQKrJMcyOrVvHmzZtbM3zqP3yx/ABXNatBj0Zml81NRxIZMfUfTqRlExnsS+d61fKnCJgnSoJ8PQu91jrV/AtVGZyNxWKhYY0AJt1xGY99u4E5m2N48Ou1TBx6Gdc0D+epHzaxMy6FsEAfJt/Z3jF3+HQ9GoXRrUEov26O4Z3fd3LwRDoJqVmOaQL3dKtX7NxvV/GwWujfKpL+rSLZEZuMYUCT8MAK1VDvUnN5/epcfp7LzknlUiP/c+N4ytn7OYiIVHbe3t7k5Z17xH/ZsmUMHz6cG264ATBH1g8cOHCBo6s8lKRLufjs733sOZ5KFW8P0rLzmLJsP8dTMnnn1jZFJlfF+XnjMZ76fiOZOTaiq/vz4sAWZOfZHEtb7Y1P5eeNx/hyxUFzdOv2djSNCGLRjuOMnbEBw4A7L6/Dv65pXKLnC/H35uv7OvPIN+v5fVscD01fR7Cf2SCobVQIn9xlJoc+AR58cHs7hn62kplrj9ClQXWSM3L4fVsc3h5WJt5+2VnnzVosFm7pEEXH6Gp8+89hQvy9HCP8NQJ9zlp14O1pZXi3etzeuQ4Z2Xnl2rSod5MaDOtSly9XHOSJmRuZ/1hPNh5J4sH/rSU9O48WNYOYOqIjNQIvXOmxl4eV929ri9Vq4ZeNxxj99Tr6toxgzqYYPK0WPrrjsrOuXW21WhjUpibXtojg+7VHOJ6Syd1dostlnvCFphFZkYpBI+kiIqbo6GhWrVrFgQMHCAgIKHaUu1GjRsyaNYuBAwdisVh4/vnnL8kR8dJSki5ldvBEGh8s3A3Aqze0xIKFJ7/fyK+bYjiZls0nd7U/Z+OnPJvBm/N38MkScz3lXo3D+OC2dgT7Fz7uxstq8cTMTeyKS2XQxGWM6BrNtBUHyLUZDGpTk1cGtTyvUntfLw8+vrM9L/y0ha9XHSIxPYcm4YF8MaKjU6n45fWr8+hVjXn3j108N3uLoxP4v/s3LXad5zNFh1bhmX5NSxzb6Xw8Pc7rhEdJjevXjKV7EtgXn8Zdn69me0wyuTaD7g1DmXxX+4vStdrTw8q7t7bBwwKzNxxjzqYYAF4c2JyO0SXr5untaXW7cmkRqRxOXytdRORS9sQTT3D33XfTvHlzMjIymDp1apH7TZgwgXvuuYeuXbsSGhrK008/TXLyufspiUmN46RMDMNg+NR/WLIrnm4Nq/O/eztjsVhYujuB+79aQ1p2Hs0ig3jnljY0r1n0+73u0Clen7PdsezSA70a8GTfJkUuJWWXkJrFkzM3OjV0uqJJGJ8O64BXKeceG4bBF8sPsObgKV64rnmRo7d5NoM7/7vKsfzSNc3D+eSu9mWaf+8ONh1J5MaPljuWGxvctiZv3dzGqfT+YsizGTyVv6zXbR2jip1CIFLZ6LOp/JXne7p0dwJ3fr6KxuEB/P54r3KKUEQuZWdrMCYVV3k1jnNtJyWpEDJz8nhz3g4++2sfGdnOc1DmbI5hya54vD2s/GdwwQh290ahzLi/C6EBPmyPSab/B39z539XsWjnceznhfbGp/LAV2u58aPlrDl4Cl8vKx/c3o5n+jU9a4IOEBrgw5ThHXlpYHN8vax0a1idj+5oX+oEHcyS9BHd6jFpaPHl1R5WC+/f1pb6oVVoFhnEWze3qRRJZOvaITzTryneHlbu71Wfd29te9ETdDDf37dvac3Sp69Qgi4ibqNGkH1OukbSRUTkwlO5u5xVZk4eo75ay1+7zBHrz/7e5+iWnZ6Txyu/bAPgwd4NqH/GEl8tawXz40NdeWPeDn7bEsvSPeayW41qBNCiZhC/bIohz2ZgtcDN7WvzWJ/Gxa5XXhSLxeKYq+3tYb1oCV2NIF9+f7wnVoulUjX7uq9HfYZ1iXZJcn46i8VSrktbiYiUVViAmaQnpueQlZt3QaYeiYiI2ClJl2JlZOcx8ss1LN2TgJ+XB9WqeHM0MYN//7iZ/y7dR3T1KhxPyaJeaBUe7N2gyMeIqubPxKGXcfhkOl8sP8CMfw6z+3gqu4+nAtCnWThPXduExuEl7wp+Jld8WXL1cl4XiqsTdBERdxTs54WXh4WcPIMTqdnndUJZRETkfClJlyKlZ+dy37Q1LN97An9vD74Y0Yk2UcF8vfIQH/65m33xaeyLTwPgtetbOi2FVpSoav48f11zHu3TiBmrD7MvIZUbL6td4qZgIiIirmK1WggN8CEmKZP4lCwl6SIickEpSZdC0rJyueeLf1i1/yRVvD2Ydk8nOuQn0/d0r8ctHWrz2V/7+HLlQW5oV4uuDUNL/NhBvl6M7Fn/QoUuIiJyQdQINJN0zUsXEZELTUm6OMnIzmPE1H9YfeAkAT6eTLunE+3rVnXaJ9DXi7HXNGHsNU1cFKWIiMjFpbXSRUTkYlGSLg7ZuTYe/Hotqw+cJNDXky/v6US7OlXPfaCIiEglpyRdREQuFnWJuoRk5uSxLz7VsQTa6fJsBmO/28DinfH4eln5YkRHJegiIiL57B3e41MzXRyJiIhUdhpJv4Tc/9ValuyKp2uD6jx9bVPaRIUAYBgGL/y0hV83xeDlYeGTuzrQvq4auomIiNiFBfkCGkkXEZELTyPpl4jV+0+yJH+t8+V7TzB40jJGf72O/QlpvP37Tr5edQiLBd4d0pZejcNcHK2IiFRWH3/8Ma1btyYoKIigoCC6dOnCb7/95rg/MzOT0aNHU716dQICArjpppuIi4tzYcQm+0i6GseJiJRedHQ07733nuNni8XC7Nmzi93/wIEDWCwWNmzYUKbnLa/HuVg0kn6JmLhoDwADWkXi6+XBrPVHmLM5ht+2xGDLr35/7fpWXNe6pgujFBGRyq527dr83//9H40aNcIwDKZNm8bgwYNZv349LVq04PHHH2fOnDnMnDmT4OBgxowZw4033siyZctcGrfmpIuIlL+YmBiqVi3fKbbDhw8nMTHRKfmPiooiJiaG0NCSr0rlSkrSLwEbDyfy1654PKwWnr62KXWq+zOyZz3enLeTP3ccB+Cpa5swtHMdF0cqIiKV3cCBA51+fu211/j4449ZuXIltWvX5vPPP2f69OlceeWVAEydOpVmzZqxcuVKLr/8cleEDJhLsIGZpBuGgcVicVksIiKVRURExEV5Hg8Pj4v2XOVB5e6XAPso+uC2NalT3R+AphFBTBnekR8e7MqU4R14sFcDV4YoIiKXoLy8PL799lvS0tLo0qULa9euJScnhz59+jj2adq0KXXq1GHFihVnfaysrCySk5OdLuUidgvMfYqIjRPN58m1kZKVWz6PLSJiZxiQneaaSxFNpYvy6aefUrNmTWw2m9P2wYMHc88997B3714GDx5MeHg4AQEBdOzYkT/++OOsj3lmufvq1atp164dvr6+dOjQgfXr1zvtn5eXx7333ku9evXw8/OjSZMmvP/++477X3rpJaZNm8ZPP/2ExWLBYrGwePHiIsvdlyxZQqdOnfDx8SEyMpJnnnmG3NyC/9979+7NI488wlNPPUW1atWIiIjgpZdeKtF7VVYaSa/ktscks2BbHBYLPNS7YaH7z1wDXURE5ELbvHkzXbp0ITMzk4CAAH788UeaN2/Ohg0b8Pb2JiQkxGn/8PBwYmNjz/qY48eP5+WXXy7/YFNiYfUneIW3JNC3DSmZuRxPziLI16v8n0tELl056fC6i6ad/vsYeFc552633HILDz/8MIsWLeKqq64C4OTJk8ybN4+5c+eSmppK//79ee211/Dx8eHLL79k4MCB7Ny5kzp1zl2xm5qaynXXXcfVV1/N//73P/bv38+jjz7qtI/NZqN27drMnDmT6tWrs3z5ckaNGkVkZCS33norTzzxBNu3byc5OZmpU6cCUK1aNY4dO+b0OEePHqV///4MHz6cL7/8kh07djBy5Eh8fX2dEvFp06YxduxYVq1axYoVKxg+fDjdunXj6quvPufrKQsl6ZWcfRS9f6tIGtYIcHE0IiIi0KRJEzZs2EBSUhLff/89d999N0uWLCnTY44bN46xY8c6fk5OTiYqKqqsoUJQ/pfm5GOEBfqQkplLfEqWPlNF5JJTtWpV+vXrx/Tp0x1J+vfff09oaChXXHEFVquVNm3aOPb/z3/+w48//sjPP//MmDFjzvn406dPx2az8fnnn+Pr60uLFi04cuQIDz74oGMfLy8vpxOy9erVY8WKFXz33XfceuutBAQE4OfnR1ZW1lnL2z/66COioqKYOHEiFouFpk2bcuzYMZ5++mleeOEFrFaz4Lx169a8+OKLADRq1IiJEyeycOFCJelSenuOpzJ3cwwAY64oPIouIiLiCt7e3jRsaH4utW/fnn/++Yf333+fIUOGkJ2dTWJiotNoelxc3DnnEvr4+ODj41P+wQZFmtcZJ6kZAfviIT5VzeNEpJx5+Zsj2q567hK64447GDlyJB999BE+Pj58/fXX3HbbbVitVlJTU3nppZeYM2cOMTEx5ObmkpGRwaFDh0r02Nu3b6d169b4+vo6tnXp0qXQfpMmTWLKlCkcOnSIjIwMsrOzadu2bYlfg/25unTp4tRfpFu3bqSmpnLkyBHHyH/r1q2djouMjOT48ePn9VyloSS9Evto8R4MA65uHk6zyCBXhyMiIlIkm81GVlYW7du3x8vLi4ULF3LTTTcBsHPnTg4dOlTkF7WLwjfE/AKbk04jvxSWYlWHdxEpfxZLiUrOXW3gwIEYhsGcOXPo2LEjf//9N++++y4ATzzxBAsWLODtt9+mYcOG+Pn5cfPNN5OdnV1uz//tt9/yxBNP8M4779ClSxcCAwN56623WLVqVbk9x+m8vJynNlkslkJz8i8EJemV1KET6fy0wTwbp1F0ERFxF+PGjaNfv37UqVOHlJQUpk+fzuLFi5k/fz7BwcHce++9jB07lmrVqhEUFMTDDz9Mly5dXNfZ3WIxS95P7CHaKwmoyvGUTNfEIiLiYr6+vtx44418/fXX7NmzhyZNmnDZZZcBsGzZMoYPH84NN9wAmHPMDxw4UOLHbtasGV999RWZmZmO0fSVK1c67bNs2TK6du3KQw895Ni2d+9ep328vb3Jy8s753P98MMPTqt1LFu2jMDAQGrXrl3imC8UdXevpD5esoc8m0HPxmG0iQpxdTgiIiIAHD9+nGHDhtGkSROuuuoq/vnnH+bPn++Y3/fuu+9y3XXXcdNNN9GzZ08iIiKYNWuWa4MONEvea3mcArRWuohc2u644w7mzJnDlClTuOOOOxzbGzVqxKxZs9iwYQMbN25k6NCh5zXqPHToUCwWCyNHjmTbtm3MnTuXt99+22mfRo0asWbNGubPn8+uXbt4/vnn+eeff5z2iY6OZtOmTezcuZOEhARycnIKPddDDz3E4cOHefjhh9mxYwc//fQTL774ImPHjnXMR3cljaRXQodOpDNzzREAHrlSo+giIuI+Pv/887Pe7+vry6RJk5g0adJFiqgEgmoBUIOTQH0l6SJySbvyyiupVq0aO3fuZOjQoY7tEyZM4J577qFr166Ehoby9NNPn9dymAEBAfzyyy888MADtGvXjubNm/PGG284pj8B3H///axfv54hQ4ZgsVi4/fbbeeihh/jtt98c+4wcOZLFixfToUMHUlNTWbRoEdHR0U7PVatWLebOncuTTz5JmzZtqFatGvfeey/PPfdc6d+YcmQxjBIujFdJJCcnExwcTFJSEkFBlXOe9pMzNzJz7RF6NArlq3s7uzocERE5h0vhs+liK9f39I+XYOm7HG08jG6brqVpRCDzHutZLnGKyKUpMzOT/fv3U69ePadGaVKxne33ej6fS64fy5dydSAhjVnrjwLw+NWNXRyNiIhIJZA/kh6YbXb01Ui6iIhcSC5P0idNmkR0dDS+vr507tyZ1atXF7tvTk4Or7zyCg0aNMDX15c2bdowb968ixit+/vgz93k2QyuaBLGZXWqujocERGRii9/Trpfppmkn0zPJifvwnf3FRGRS5NLk/QZM2YwduxYXnzxRdatW0ebNm3o27dvsWvPPffcc3zyySd8+OGHbNu2jQceeIAbbriB9evXX+TI3dPe+FRmaxRdRESkfAXVBMAzLQYPqwXDgJNp5bekkIiIyOlcmqRPmDCBkSNHMmLECJo3b87kyZPx9/dnypQpRe7/1Vdf8e9//5v+/ftTv359HnzwQfr3788777xzkSN3Tx8s3I3NgD7NatC6doirwxEREakc8svdLalxhFfxAFTyLiIiF47LkvTs7GzWrl1Lnz59CoKxWunTpw8rVqwo8pisrKxCE/D9/PxYunRpsc+TlZVFcnKy06Uy2h2Xws8bzXXRH+ujUXQREZFyUyUMrJ5g2GhUJR1Qki4i5eMS6+Fd6ZXX79NlSXpCQgJ5eXmEh4c7bQ8PDyc2NrbIY/r27cuECRPYvXs3NpuNBQsWMGvWLGJiYop9nvHjxxMcHOy4REVFlevrcBfvL9yNYUDfFuG0rBXs6nBEREQqD6vVMS+9oW8SAMdTMl0ZkYhUcB4eZlVOdramzlQm9t+n/fdbWhVqnfT333+fkSNH0rRpUywWCw0aNGDEiBHFlscDjBs3jrFjxzp+Tk5OrnSJ+s7YFOZsNk9UaBRdRETkAgiMhKTD1PVKAsI0ki4iZeLp6Ym/vz/x8fF4eXlhtbq8n7eUkc1mIz4+Hn9/fzw9y5ZmuyxJDw0NxcPDg7i4OKftcXFxREREFHlMWFgYs2fPJjMzkxMnTlCzZk2eeeYZ6tevX+zz+Pj44OPjU66xu5tJi/ZgGNC/VQTNIrW+roiISLnLbx5X2yMRULm7iJSNxWIhMjKS/fv3c/DgQVeHI+XEarVSp04dLBZLmR7HZUm6t7c37du3Z+HChVx//fWAefZh4cKFjBkz5qzH+vr6UqtWLXJycvjhhx+49dZbL0LE7unwyXR+3WTORX+od0MXRyMiIlJJ5TePq8FJAOJTlaSLSNl4e3vTqFEjlbxXIt7e3uVSFeHScvexY8dy991306FDBzp16sR7771HWloaI0aMAGDYsGHUqlWL8ePHA7Bq1SqOHj1K27ZtOXr0KC+99BI2m42nnnrKlS/DpT77ex82A3o0CtVcdBGR83V4NcRshA73mvOORYoTZM5Jr5YXD8DxZCXpIlJ2Vqu1UGNsEZcm6UOGDCE+Pp4XXniB2NhY2rZty7x58xzN5A4dOuR0JiIzM5PnnnuOffv2ERAQQP/+/fnqq68ICQlx0StwrROpWXy35jAAD/Zq4OJoREQqGJsNvhsGKTHgEwhtbnN1ROLO8svdA7OPAxpJFxGRC8fljePGjBlTbHn74sWLnX7u1asX27ZtuwhRVQzTlh8gM8dG69rBdGlQ3dXhiIhULMfWmQk6wLL3ofUQKOMcMqnEAs0k3TcjP0nXnHQREblAVNtXQaVl5TJthdlk4oFeDcrcnEBE5JKz49eC28e3we7fXReLuL/8kXTPtFjAID07j7SsXNfGJCIilZKS9Arq238Ok5SRQ73QKvRtUXQ3fBEROYsdc83r0PylK5e+57JQpALIXyfdkpdFLe90AI5rNF1ERC4AJekVUE6ejc//3gfAyB718bBqFF0qKMOAQyshO83VkYg7OLHXvJwvw4CDyyEzueTHJOyBhJ1g9YJbvzKvDy03G8mJFMXTG6qEAdDMPwVQybuIiFwYStIroJ83HONYUiahAT7ceFktV4cjUnp7FsKUvjBzhKsjEVdLPwmfXgGTe0BK7Pkdu/BlmNoPPu5qJuslsXOOeR3dHWo0Neejw/mNphvGeYUplUB+yXsDX/OEkJJ0ERG5EJSkVzCGYfDJX+ZI0z3do/H18nBxRCJlcGydeb17PhzbcHGeMzsdln0Ae/64OM8nJbNpBmQlQU4arPy45McdWFaQWCcdhi8GwJ+vQd455grbS92bDjCvuz1iXu+cA/E7z/28uVnw9S2wZVbJY5WKL795XLR3EgDxKZmujEZERCopJekVzKx1R9kVl0qAjyd3dK7r6nBEyibxYMHtZe9f+OeL2QSf9oYFz8P028yfxfUMA9Z+UfDzmimQmXTu4zKT4ccHAANa3Qpt7wDDBn+9aY6snzpQ9HGp8XB4lXm7ST/zOqwJNMlP2Jd9cPbnzcuFH+6FPQvgl8fMKgC5NOSPpNe0ngI0J11ERC4MJekVyJajSTw7ezMA9/esT7Cfl4sjEimjU6cl6dtmw8l9F+Z5bDZY8RH89ypzHjIWsOXArFGQU0lHwgwD/noLvhwMGYmujubsDq+C+B3g5Q/VGkBWMqyZeu7jfnsakg5BSB0Y8A5c/xHc9Dn4BMOR1fBxd3Ok/Uy7fgMMiGwLwbULtnd/zLzeNAOSjhb9nIYBvzwK238BD28Y8hX4VzvPFywVVn6SXgPzxIySdBERuRCUpFcQJ9Oyuf+rtWTm2OjdJIyHrmjo6pAubbY8yMspv8crz0QxL/fcpb5gNmu7kHNq83LO/fiJh8zrgAhzBHT5xOL3TUswE6fTL1kp544j9ThMvwXmj4O8bGjcD0avhio1IH47LHyl5K/pYsjJKJ/fy19vwZ+vwr7F5gmQ8pCZXLLYcrPP73Hto+gtb4QeY83bKz82S8qLs+0n2DgdLFa44VPwDTK3t7oZHlwKUZdDdgr8cB9knHI+9sxSd7uoTlCnq3kCZ+VHhZ/TMOD352DD/8znvXkq1O91fq9VKrb8JL1qbjwAccmV9CSfiIi4lJL0CiA3z8bD36zjaGIGdav78/6Qduro7ko5GfBhe/ikpzm/uayWfQCvRZgjc2WVcQrebW42Yztb4r9lFoyPMhOOCyHjFEzsYMZRnLxcSDpi3r46P1He8LVZinw6m80sKX6rgfnaTr+82QDithX/HLlZ8N8+5vxzT19ztPX2byCsMQyeZO6zcpKZyLqDhN3wZn1z9Lssf1urPoVFrxX8XJb59xmnzCR6an/4vyhzznfi4aL3TUuAb4bC+Fqwc17JH3/rj+bt9iPMsvXAmpAaCxu/LfqYlFjzbwKg22NQt4vz/SF14K5Z5qh8yjGY86+C+7LTYN8i8/aZSToUjKav+gRmDoedvxWcdPj7HViRfyJp0ERodl3JXqNUHvlJelCO+f9UTJKSdBERKX9K0iuAt+bvZNmeE/h7e/DpXR0I9leZu0vtWwyn9sPxbWaCVxaZyfDX24BhjujabGV7vP1/Q2ocHF0Df/6n6H0SD5sJjpFnJhy7F5TtOYvyz3/N+cCHVxU/XzflmBmD1Qta3QI1L4PcTFj9ScE+hgG/Pwtr80ufPbwLLhYr5GXBui+Lj2P3AnPee5UwGLUYOt4HlvwTXI2vgQ73mrdnP1R4tLUoycfM5PN8R4ptNjj8z9lPKICZGOakw/4lZoJYmmqNTd/Bb0+at5sNMq/3/XX26oqYjbDhG+fL2i/g2zvg7cZmeffB/LLxg8tgcrfCDdP2/ml2V985x6xYWPhyyUbdN31n/t7DW0Kt9uYyV10eMu9b/oFZtXI6mw1+Gg0ZJyGiNfQeV/TjeleBGz8Diwds+QE2f18QZ24mhNSFGs0LH9foGnNuui3HPHnwzW3wThPzvbD/m+o7Htrdce7XJpVPfuM434w4AI4lZmCoy7+IiJQzJelu7peNx/jkL3Oe7ls3t6FJRKCLIxJ2zCm4vfQ9SIkr/WOt/cLsaA2QsCt/rmwZ2JthgZmA71vifL/NBrMfNJ/T08/c9tNoSDtRtuc9XU4GrJxc8HNx617b56OHRIHVWjCCufozyEo1b//1dkHZ8fWT4fn4gstt083tW2cVTuTstuQnZm1ugxrNCt9/zX/M0dbkozDniaIfIzMZ1n8N0wbBhObwzRD4bljJEujjO+CPl+H9NvB5H7Np3enz8E+XnW7OhQbzBMTu+ebv6nxO3Oycl99IDeh0P9zyBfhVM3/fR9cUfUxKLPz3apj9gPPll0dhx69mwl2jOfR5Ge6ZD7U6mE3dvh8Bs0ebJ2HmPwtf3WCeIAptAt6B5kms3b+fPd7TG8a1H15wAqX9cPANhhN7nP+9pcTB1zeblQEePmYS7uld/OPXbg+9njJv/zrWrNw4vdTdUkRFksUCt30N9/8FXcZAQLh5QmDHr+b9PZ8qOIkgl56gSACs2SlUIYP07DySM0owvUhEROQ8KEl3YzFJGTz1vdl9+oFeDRjQOtLFEbkhW56ZKJd0beTyeL5d+WW8/tUhOxUWjy/dY+VmFSSgoU3M66XvlW0+8qEV5nXVeub17Aedm4atnAQH/jYbdI1cCGFNzcTq10dL/rz7FsOSt4ofTd7wNaQnFPx8Yk/R+9nno4fkr1LQ9DozYc5MhHXTzGR90avmfdf+H7S93fn4BleBb4gZ/4GlhR8/K8UsVQZoeXPRMXhXgRs/zR9t/R5m3GWOqtsv3wyFtxvBTw+Zo9sY5r67foOfxhSdQBsGbJxhrvf9UWdYOsFsbgbmyP+KYqovtv5oNkyrGg23fwtWT9g8E357qmS/mwPLYObdZnVC6yHme2b1gAZXmPcXV/K+ZZYZV0C4+Z6efun6CDywDB5aYZ5EqXM53DMPejwBWMy52W83LigB73CvWbHQYbj587nWHD/yj5nMe/qZ1RR2PoFm1QPAsvfM179rvjlSv3ehOXXh+o/M9c3Ppce/zBH6rCTzBIb932+T/sUfY7FAZBvo+xqM3Q53/QiX3W2eqLji3+d+Tqm8fALNxoRAU3/zZOLRxAxXRiQiIpWQknQ39svGY2Tk5NGmdjBP9m3i6nDc09Yf4Y8XzS7dF6Pk8Mg/kBZvfkm75Qtz27ppcHz7+T/Wpu8gJQYCI+HOH8yRwSOrCxLt85WdbpYtgznKXK2+OUI8N3+EOHZLQZO0vq9DeAszQbV6mfPhN35z7ufISjFHkRe96jzn2S4vF5Z/aN72Nb/IcrKYkXT78mshdcxrq0fBWtWL3yiIu9czcPmDhY/39Ibmg83bm2cWvn/HXLOsuXpDM+EqTu0O0DO/PHz7z+ZJBvtl5xzzMUIbw5XPw6ObzFFWiwds+tZsRnf6313GKbNM/cdRELvJTLSb9Df/Vm7PHyVf92XRlQv2EeXL7obGfeGGTwAL/PMZLHq9+PjBXGP+m9vMWBv3M+fbW/P/e2/Yx7zes7DoY+3VBj3+Zc7jPv1yzX8goqXz/h5ecNXzMHwOBNU2y8L9qsFt38B1E8DbHy5/yPy7OrQcDq8uPu7TG8b5hTjf1/kB89/E0bVmqfn0W82TP+EtYdQSs0FcSXh4mY3lvPzNE1QZJ8GvKtTpcu5jIf9Ex5Uw6APzREVRo+9yackfTW9axWxcGZOkJF1ERMqXknQ3NndzLAA3t6+tRnHFsSdnSYchfuf5Hbt1ttlULHZzyY+xl942vgbq9YRmA82u5L8/f37PbbMVrAt++UNmybd9pLio0UebDeaNM0uKs9OKfsxj68CWayb9NZoVzMfdPBM2TDdPZORlQ+NrzXJiMJPXK/Ln9M59qvh1pe3WTitYv3rZ+4WXt9r+k/kYftXMUmE490h61boF21rfZo7o2qcAdLofej9TfDz2RG37z4U7gduTz5Y3nzux6vWU2Qisz8vOl2teNRPC0auh5xNmrE36wfUfm8etmgxL3jRvH1hmLvm1bbaZnF/xLDyx22xU1+IGM/GObAO5GbD6U+fnj9tqnqCxepprfdtf24C3zdt/vWmeYClqXnnCbvjfTeYofN3ucMtUMzG1a3CleX1sfeGTAyf2mkmwxWrGeD6iu5ld1AdNNEfam542Mh1UE9oMMW8XN5qekVgwr93+93i6gBoF87535v+76/wg3LewZCPopwttaP4u7RpfCx6e5/cYInb5zeMa+pr/Tx3TSLqIiJQzJelu6mhiBhsOJ2KxQN8WEa4Oxz2ln3QeHdxbzEhhUWz5ifWRf2DmiJJ10jaMgiTdXirb52UzsdqzwGxIVVI758KJ3eaIvD1B6fIwYDHnIp/eYMwwzJLnlR+Zz2GfU3sm+wh8ncvNpLR2BzOxBLPs/fhW8A+FQR86J63dHitYrurHB4qf352bXVCqXa0+YJj7ZyYXxGlPyDqNgohW5u2SlrsDePmaI7oAbW43S7bPlmDX7WYu35aZ5Py3kHai4PdRkhFXqwdcdpc5Unr6pevDULNt4RjaDIF+b5m3F78O39wO066D5CPme3Pv72bif/r62RaL+V6D2Rzv9JMta6eZ1036Q2B4wfaO98GV+R34/34HvujvPKc96Qh8eb05whzZxjwh4OXnHGtghDn6jFHQ1dzOniTX62kmxefLr6r5vgUW8X9U10cBi5lgF3UCbfNM84RFjeZQu2PRj9/1EbMio0oY3PE99Ps/82+kNDrcY06pAPNvS6S08pvH1fHMT9LV4V1ERMqZknQ3NW+LOYresW41agSV8ktpZbf9Z7PU1u58lpna92fBPOETu82S+XNJ2GWWblu9CkqIqzeAjiPN2/OfKz7BPZ1hmPNsATreW7C+c2hDc2QeCkbZwSx1/uezgp/to4pnOpTfNC7q8oJtPZ80u6bbDfqgcDJm9YAbPwHvADPRP/25T7d5ptmRPSAC7v3DTK6TDsFvT5v371tklnh7+plJevWG5vYT+4qeiuBoHFfXeXunUfDoRnO02nqO/6KsHmapNBSMnIM5mm3LNRPX0EZnf4zS6jzKHC0H86SLYYO2d8L9f5tzoIvSfLDZLyDjFKz7ytyWnW6WzkPRI8o9n4Qb/ws+QWZjwMndzU7laQlmgp58BKo3gjtnFfwtnanhVeb16ScyDKPgPTt9Pnh5CWtcsMTZsg+c7zuyBv6eYN4+vWHcmarVg0c2wONbodHVZYvHYoFbv4THt2ltcymb/JH0SKu5coVG0kVEpLwpSXdTv22OAaBfq0t4FD1hN8RsKv5++5JK9lGxg8vNzuIlYZ8LW7uTeb3603Mn+fZR9Pq9nJOhXk+Zo33Ht5pl5edycLk5gu/hU3iutb3D+ZbvzaXSVkwyS52hIIHbvaBwabfNVjD3t85pSbqHF9z0X7NBXLfHil4XGsxmZf3eMG8ver1gbvvpj+8oz38QqlTPb7hmhY3TYdtPBaPolw0z7w+pa96fk2Z2ED9dbraZ8EPBnHQ7i8WMp6Rzf+0j5Tt/K+gKv+UH87q4hnHlpeeT0Otps+HdzVPh+kngE1D8/lYPc3QezGZreTnme5eZZL4P9a8o+rjWt8ADf5t/r1nJ8MO9MKmzeYIpqDYMmw1VQot/3gb5SfrehQUnTOK2QvwOczk7+whzebNXDmyaAUlHzZNYf70Fn19j/v6r1jM775+NfzXw9CmfeKweEFyrfB5LLl35c9Kr5ZkNMmMSNZIuIiLlS0m6G4pNymTNQXPN5mtbXqJJel4OTOkLn11hJhNnSj5W0NH7in9DUC2zaZZ9LeezSYkt6Po98H1z3jMULCdVnJ35ZeZndoX2r2YuywTmUlTnmuNuH0VvO7TwqHat9hDdwxwF/u4umJ/fSfrK52HAu+YodnaquR766eK3m/O4varklzafpnoDGL0Krn757HG1vcNM1mw55vz1nNO+eO6aBwk7zdHcDiPMbXUuh+6Pm7dnjza7n1s8oMtoc5und8Eo+ZnN45KPmCPPnr6lK7M+Xc3LzGQvJ938vSYdKej2bx9lv1AsFvPv75F1JX+utkPN8u2kw+bJhNMbxp2tcqBqNIz4zTwpYLGaJe7+oWaCHlz77M9Z53KzcVpqHMRtMbfZR9EbXVO4aVt5iepoTkmw5ZhrjE8bCH++anagb3mT2Qne3mBQpKIIMk/0BGXHA+ruLiIi5U9Juhuav9UcdbysTgiRwX7n2LuSit8B6SfMZLWopmxbfwQMs7Q7pE7R5bzF2fC1+bhRnSG8OfR5yezenRoLvz5WdGl2Sqw5+g1FL93UaZQZS1YSfHVj8WuDH1ljrh1tsRaMqJ7JPpp+bL153WWMOU/bajWblkHhknf7fPSojqVviGWxmCctqtQw3397J3goOLHQ4R7npKrXMxDR2pzPDmbidXojOEfJ+xnz0h3z0euUvVu2xVIwmr7l+/x51gbU6Xru5NUVvPzMzuVgrqF+eKV5cqPdnec+1sPTPCkwfK5ZsXD3zyUr5/f0Meedg1kxYhiw2V5tcFPpXkdJ2UfTN35jnkTzDjDXvL/p8wt3ckDkQsovd/fJMD+r45IzybNdhNVFRETkkqEk3Q3NzS9179/qEl4X/diGgtt7FxYuRbeXutuTswYlTNJttoImXfbycW///KXIPM3S400zCh9nH3mv1d5R6ujE0xuGzoDwVpB2HL663hzttzMM+Odz+CK/3Lz5YHOEuygNripYMqzdnWZXansiay9L3jHXeY3uouajl0aVUBicv+b1yknmmugHV5hzoT28C5fne3qbXeQ9fQELdHvU+X77azzzpIVjPvoZpe6lZS9r3/OHucQZQKsLnHyWRcd7zWTVXvLfpF/RzdeKU7eL2QAwvEXJjzn938jh1WY/Ae8As9P5hdTo6oImgjUvg/v/Mlcy0FJml6zx48fTsWNHAgMDqVGjBtdffz07dzo3F+zduzcWi8Xp8sADD7go4jPkN47zyDiBnzWXXJtBfErWOQ4SEREpOSXpbiY+JYvVB8yS60u21B0gZoN57Z0/v/f35wuasp3Yay43ZvGA5teb2+r3Nn9O2GnO5S7O/sXm+tw+wQXHAtRsV7DU19wnC0rp7YordT+dX4i5tnS1+uZI8Vc3mOXzaSfMdZ7njDVL8htcCf3fKf5xLBYY8rW5tvbAD5yTmXo9wDvQHPW3j7QDHFppXtcpY5IO5lJh7fNL2n980OxeDubc/6ISyRpN4Z75MPzXwmtqO0bSz0jSi+rsXhY1mppl/rZcc5621ROan+eSYheTX1XnJnH29/tCslebHFoJ6/JPVDUdYJ6kupAsFrjjBxjyP7PrfXEnp+SSsWTJEkaPHs3KlStZsGABOTk5XHPNNaSlOS8vOXLkSGJiYhyXN99800URn8G/mtlTBGgRYK4MckxrpYuISDlSku5m5m+NxTCgTe1gale9wF+eLwbDMBPVokrIz8Y+kt7nJfANgePbYP3/zG32pmD1e0NAmHnbL8RccgzOvhSbff5vmyGFk5Nuj5sl0lnJ8MV1+etS50BWCuxbYu5TXOM1u4AacNdsc63y+B3w5SD4uKtZnu7hDX1fNxOWKtXP/jghUea61VYP5+2ePgXJlr3kPemoOSpqsRa8B2V1zavmyYaUY7D/L8BiLodVnJptIbp74e3V6pvXZ85JTyznkXRwXmqt/hXnfo9drctocz358JbQoJiGceWpWn1zXrstx5zyARe+sZ5dYLi5csHp67fLJWvevHkMHz6cFi1a0KZNG7744gsOHTrE2rVrnfbz9/cnIiLCcQkKKmb1govNYnFUVDULMKf6qMO7iIiUJyXpbua3Lfau7pWk1H3lR/BmPZjYAZa8CSf3n/uYvNyC5lb1e5uNsgAWvWZ27z6z1N3uXCXvqccLOrQXtdSVhyfcMRPa3QUY5rrUU/qaZep5WWaSE9b03PFXrWsm6n5VzSZyqbHmnPf7FpqJ2bmWFTsXR8l7/ms5nD+KHtEKfALL9th2PgFmGbsl/yRBs4HmEnHnyz6SfnKf8/J09pH0quU0kg7Oc6tLsja6qwXVhEc3wL0LCp+MuRAsloJ/I2CeILgYJwdEziEpyVxvvFq1ak7bv/76a0JDQ2nZsiXjxo0jPT39rI+TlZVFcnKy0+WCyW8e18DHfA4l6SIiUp6UpLuRE6lZrNxnlrr3qwyl7nm5sPxD8/aJPWaS/UFbc/mlNVOLX1M8YadZFu4daC5t1fE+s3t3ahz8eL95v4dP4WWj7GuX71tiPveZ7A3jancsfi6vT4A5J/uWaWaDtKNrC9ZQb9K/5PNoazSFO38w56h3GgWjlkBk65Idey6NrjbLueN3mGXk5TUf/Uy1O8C1/wehTQrWAz9fwbXNCoK8bLObud3pjePKS0gd6PygeWLHvt68u/MNvvDl5qez/xsBsy+CRrbFxWw2G4899hjdunWjZcuC6TJDhw7lf//7H4sWLWLcuHF89dVX3Hnn2Zsrjh8/nuDgYMclKirqwgWe3zyujlciAMe0DJuIiJSjUraBlgthwbY48mwGLWoGUbd6FVeHU3a7f4eUGPCvbpZPb55pNiI7vMq8eAeY6z+fyV7qHtnaHHW2epvLh303DHb8at7X+BrntcrBLLn2qwoZp+DoGuf52UU1jDubFtebSeqsUQXLup2r1P1MtdrDg0vPvd/58gsxl7Xav8ScK2/v7F4e89HP1HmUeSktq4dZgWA/oVA12lzaLcWsGCEkujyiLNDv/8r38Sqbej3A6mWWvFeEagOp9EaPHs2WLVtYutT5/8pRowr+32nVqhWRkZFcddVV7N27lwYNiu5rMG7cOMaOHev4OTk5+cIl6oFmtVuExVwuVSPpIiJSnjSS7iIpmTmsOXDS6TJr/VGgEnV1t8//bnuHuTb0XT/C2O0FZcm75hV9nL1pXGTbgm3NBkGdLgU/FzWX1uphNmWDwiXvu+bBqf3mOt8tSthQLLg23P0L9HvTXGrs9Od3NXsVwabvCqYGXIgkvTyc2Twu6Yh57VXFbMAkF49PINwwGa7+j3miR8SFxowZw6+//sqiRYuoXfvsyyV27twZgD179hS7j4+PD0FBQU6XCya/3L16nrlWekySRtJFRKT8uDxJnzRpEtHR0fj6+tK5c2dWr1591v3fe+89mjRpgp+fH1FRUTz++ONkZla8D8cbP1rOzZNXOF1W769Epe6Jh2HPAvP2ZXcXbA+MMMvXAfb+WXTJu30kvWbbgm0WC1zzGmAxR8sb9y36eR3z0vOXbLPlwd8T4Lu7zJ/b3Abe51GlYPWAzvfDFePca8ko+3rpsZvAsJml3vnll27nzOZxiQfM66p13es9vVS0uhm6PaL3XlzGMAzGjBnDjz/+yJ9//km9evXOecyGDRsAiIx0k5PY+VN1gjLNk+saSRcRkfLk0nL3GTNmMHbsWCZPnkznzp1577336Nu3Lzt37qRGjRqF9p8+fTrPPPMMU6ZMoWvXruzatYvhw4djsViYMGGCC15B6aRn57L7eCoAdav7c/pX5V6Nw6gfFuCawMrT+v+ZyWN0j8INx2p3NOebZ5w0R81rtS+4Ly/XbLYGziPpALXbm0s4+QSCl1/Rz2sfST+23nyceePgwN/mtuaD4crny/rK3ENIFES0NpN0KP/56OXJMZKePwJ2Ieaji0iFMXr0aKZPn85PP/1EYGAgsbGxAAQHB+Pn58fevXuZPn06/fv3p3r16mzatInHH3+cnj170rp1OfX2KKuq0QD4pJi9Nk6kZZOZk4ev10VoAikiIpWeS5P0CRMmMHLkSEaMMNcInjx5MnPmzGHKlCk888wzhfZfvnw53bp1Y+jQoQBER0dz++23s2rVqosad1nZy+ICfDxZ8mQl7K6clwvrvzJvF9lF3Qvq9zLnl+/50zlJT9gFuRnmfPXqRXQTj+p09ucOijSXtIrbAp/0AiMPvPzNkvV2d1au0cOm1xUk6e5a6g4F62Lby91PXYDl10Skwvj4448B6N27t9P2qVOnMnz4cLy9vfnjjz947733SEtLIyoqiptuuonnnnvOBdEWI39lCmtWEuFeGcTl+BGTlEm90ErQT0ZERFzOZeXu2dnZrF27lj59CroNW61W+vTpw4oVK4o8pmvXrqxdu9ZREr9v3z7mzp1L//79i32ei7okSwnF5HeBjQj2dXEkF8iePyD5qLnEU3Fdtu1rfZ+5prl9PnpE69IvVWYfTTfyzNH4+/+Gy+6qXAk6QNPT/u7dOknPP9mSeBBys08bSS/H5ddEpMIwDKPIy/DhwwGIiopiyZIlnDhxgszMTHbv3s2bb77pPuukgzltqopZ8dcu0FxCLkYl7yIiUk5cNpKekJBAXl4e4eHhTtvDw8PZsWNHkccMHTqUhIQEunfvjmEY5Obm8sADD/Dvf/+72OcZP348L7/8crnGXlYxSeYHeWRlTdIdDeOGgqdP0fvY544fXg2ZSeZSVFD0fPTzddnd5nz3RtdA73Hg6V36x3Jn4S3N6oC8HAhr5upoihcQblZGZKfCqQNmsg4aSReRiq1qNKQdp7nfSeYRwVEl6SIiUk5c3jjufCxevJjXX3+djz76iHXr1jFr1izmzJnDf/7zn2KPGTduHElJSY7L4cOHi933YrGXu1fKJD3pKOyeb94+21JnVetC9UbmaPe+JQXbi+rsfr5CG8KDy6DPi5U3QQezMmDwJLjx09JXHVwMFotz8zj7SHpVjaSLSAWWPy+9oWcCoA7vIiJSflw2kh4aGoqHhwdxcXFO2+Pi4oiIKLq7+fPPP89dd93FffeZ3cFbtWpFWloao0aN4tlnn8VaRKLi4+ODj08xo7kuUpCkF9P8rCKzN4yr2x1CG51934ZXwYndZsl780FmJ3Z707iyjKSL+6ne0Jw/H7sZ0swlizSSLiIVWv6JxtqW44A6vIuISPlx2fCbt7c37du3Z+HCgjnJNpuNhQsX0qVL0etRp6enF0rEPTzMTqqGYVy4YMtZpS13t+XBui/N2+3vPvu+AA3z+xHsWQiGYTaNy0k3188uqmmcVFz25nF7F5nXPsHmUnoiIhVV/kh6jdwYAI5pJF1ERMqJS7u7jx07lrvvvpsOHTrQqVMnRydXe7f3YcOGUatWLcaPHw/AwIEDmTBhAu3ataNz587s2bOH559/noEDBzqS9YrA3jguMqSSjaTvWQjJR8A3BJoNOvf+dbuBhw8kHYaE3QXz0SNbm+uTS+VhP+lyxGz6qFF0Eanw8pP04KxjgEbSRUSk/Lg0SR8yZAjx8fG88MILxMbG0rZtW+bNm+doJnfo0CGnkfPnnnsOi8XCc889x9GjRwkLC2PgwIG89tprrnoJpWIfSa9Z2UbSl39gXre7E7xK8Nq8/aFuV9i3yCx5P3XA3F6W+ejinqrlj6Tbcs1rzUcXkYouP0n3TTuKFRsxiRkYhoGlsq0kIiIiF51Lk3SAMWPGMGbMmCLvW7x4sdPPnp6evPjii7z44osXIbILIy0rl+RMM1GpVEuwHV0LB/4Gqydc/mDJj2t4lZmk7/kDslLNbZqPXvnYy93tNJIuIhVdYCR4eGPJyyaSExzNDiM5I5dgfy9XRyYiIhWcG7eErpzsTeMCfTwJ9K1EH+RL3zOvW90CwbVLfpx9KbYDS83GYqCR9MrIvxr4VSv4WWuki0hFZ/WA4CgAWvifAuBYkkreRUSk7JSkX2T2UvdKNYqesAe2/2Le7vbo+R1boxkE1oTczIKmcefqCi8V0+mj6RpJF5HKIL/kvYXfSUDz0kVEpHwoSb/IHMuvVaamccs/AAxofK2ZdJ8PiwUaXlnwc0QrNY2rrE7v2K8kXUQqA/ta6V4nAHV4FxGR8qEk/SJzdHYPqiQj6SmxsPEb83a3x0r3GPal2EDz0SuzahpJF5FKJj9Jj9Ja6SIiUo6UpF9kscn5a6SHVJIkfeXHkJcNUZ2hbtHr259T/d5gyf9T1Hz0yste7u5XFXyDXBuLiEh5yE/Sw/PMtdJjlKSLiEg5UJJ+kR2zj6RXhjnpmUmwZop5u7Sj6GAmbS1uMNdXb3BFeUQm7qjO5eAdCA2uPPe+IiIVQf5ykiGZ9rXSVe4uIiJl5/Il2C41sfY56cGVYE76mqmQlQxhTc356GVx438BQ/PRK7OgmvDkbvCsBCeoRETAMZLuk30SfzLV3V1ERMqFRtIvMvsHeM2KXu6ek2GWugN0fQSsZfxTslqVoF8KvPzMZoEiIpWBb7BZDYY5Lz02KZM8m+HioEREpKJTkn4RpWblkpKZC0BERR5Jj98J/70aUmMhqJa5NrqIiMilKH80PdoaT67NICE1y7XxiIhIhack/SKKzR9FD/T1JMCnAs40MAxzDvonvSBuM/hXhxsmg6e3qyMTERFxjfwkvbmvuVb6UTWPExGRMqqAmWLFVaGbxqWfhJ8fhh2/mj/Xv8JM0AMjXBuXiIiIK+Un6Y28EyAtf6lVrTIpIiJloCT9IqqwTeMyk+DTXpB4CKxe0OcluPyhss9DFxERqehCzA7vda3xgNZKFxGRslOSfhHZm8ZVuJH0rT+aCXpQLbj9G4hs4+qIRERE3INjrfRYAI6cSndhMCIiUhloKPQiqrAj6Zu/N687jVSCLiIicrr8JD0kOwYLNg6cUJIuIiJloyT9IjqWVAHnpCcfgwNLzdstb3JtLCIiIu4muDZYPPC0ZRFGEgdOpLk6IhERqeCUpF9E9u7ukRVpjfStPwIGRHWGEHXCERERceLhZSbqQB1LHEdOZZCTZ3NxUCIiUpEpSb+IYhIrYLm7vdS95c2ujUNERMRdVTWbxzXwOkGezeDIKTWPExGR0lOSfpGkZOaQkpULVKBy9xN74dg6sFihxfWujkZERMQ95c9Lb+l/CoADCSp5FxGR0lOSfpHYm8YF+XpSxaeCNNXfMsu8rt8bAmq4NBQRERG3lZ+kN/RMAGC/knQRESkDJekXybGK1tndMGDzTPO2St1FRESKl5+k17bEAXBQzeNERKQMlKRfJBWuaVzcFkjYCR4+0Ow6V0cjIiLivvKT9OrZMQDs1zJsIiJSBkrSL5JjiRVs+TV7w7hGV4NvsGtjERERcWdV6wHgn3UcH7I1J11ERMpESfpFEluRyt1ttoL56K1U6i4iInJWflXBOxCA2pZ4jpxKJztXy7CJiEjpKEm/SI7ll7tHVISR9COrIekQeAdA42tdHY2IiIh7s1hOax53ApsBR06p5F1EREpHSfpFYh9Jr1kRRtLtpe5NrwOvChCviIiIq+Wvld46IH8ZNjWPExGRUlKSfpHE2Mvd3b1xnM0G22abt1XqLiIi5Wz8+PF07NiRwMBAatSowfXXX8/OnTud9snMzGT06NFUr16dgIAAbrrpJuLi4lwUcQlVM+elN/c+DsCBBI2ki4hI6ShJvwhSMnNIzcoFKkDjuFP7IS0ePH3N9dFFRETK0ZIlSxg9ejQrV65kwYIF5OTkcM0115CWVjDy/Pjjj/PLL78wc+ZMlixZwrFjx7jxxhtdGHUJhDUFoL5xCNBIuoiIlJ5bJOmTJk0iOjoaX19fOnfuzOrVq4vdt3fv3lgslkKXAQMGXMSIz499FD3Yzwt/b08XR3MOx7eZ12FNwMPLtbGIiEilM2/ePIYPH06LFi1o06YNX3zxBYcOHWLt2rUAJCUl8fnnnzNhwgSuvPJK2rdvz9SpU1m+fDkrV650cfRnUaM5AOGZ+wHYrw7vIiJSSi5P0mfMmMHYsWN58cUXWbduHW3atKFv374cP368yP1nzZpFTEyM47JlyxY8PDy45ZZbLnLkJXcsMX+NdHcfRQeIy0/Sa7RwbRwiInJJSEpKAqBatWoArF27lpycHPr06ePYp2nTptSpU4cVK1YU+zhZWVkkJyc7XS6qsKaABd/sk4SSxEGtlS4iIqXk8iR9woQJjBw5khEjRtC8eXMmT56Mv78/U6ZMKXL/atWqERER4bgsWLAAf39/t07SC5ZfqwBJun0kvUYz18YhIiKVns1m47HHHqNbt260bNkSgNjYWLy9vQkJCXHaNzw8nNjY2GIfa/z48QQHBzsuUVFRFzL0wrz9oVp9AJpYD2kZNhERKbVSJemLFi0qlyfPzs5m7dq1TmfLrVYrffr0OevZ8tN9/vnn3HbbbVSpUqXI+11+Zh04lp+kR1SEzu72JD28uWvjEBGRSm/06NFs2bKFb7/9tsyPNW7cOJKSkhyXw4cPl0OE5yn/s7OV51FsBhzWMmwiIlIKpUrSr732Who0aMCrr75apg/BhIQE8vLyCA8Pd9p+rrPldqtXr2bLli3cd999xe7j8jPrQGz+Guk13X0kPScTTuw1b9dQki4iIhfOmDFj+PXXX1m0aBG1a9d2bI+IiCA7O5vExESn/ePi4oiIiCj28Xx8fAgKCnK6XHT5U8Uu8z0GwAHNSxcRkVIoVZJ+9OhRxowZw/fff0/9+vXp27cv3333HdnZ2eUd31l9/vnntGrVik6dOhW7jzucWY9xjKS7eZKesAuMPPANgcBIV0cjIiKVkGEYjBkzhh9//JE///yTevXqOd3fvn17vLy8WLhwoWPbzp07OXToEF26dLnY4Z6f/JH0Jhbzu8YBzUsXEZFSKFWSHhoayuOPP86GDRtYtWoVjRs35qGHHqJmzZo88sgjbNy4scSP4+HhUWjt03OdLQdIS0vj22+/5d577z3rfu5wZv3wSfNDulaIm5e7O0rdW4DF4tpYRESkUho9ejT/+9//mD59OoGBgcTGxhIbG0tGhll1FhwczL333svYsWNZtGgRa9euZcSIEXTp0oXLL7/cxdGfQ34VWs2cA1ixaSRdRERKpcyN4y677DLGjRvHmDFjSE1NZcqUKbRv354ePXqwdevWsx7r7e1N+/btnc6W22w2Fi5ceM6z5TNnziQrK4s777yzrC/hgkrNyuVgfpLeJCLQxdGcQ1z+70tN40RE5AL5+OOPSUpKonfv3kRGRjouM2bMcOzz7rvvct1113HTTTfRs2dPIiIimDVrlgujLqFq9cHTFy9bFnUscVorXURESqXUSXpOTg7ff/89/fv3p27dusyfP5+JEycSFxfHnj17qFu3bok6ro8dO5bPPvuMadOmsX37dh588EHS0tIYMWIEAMOGDWPcuHGFjvv888+5/vrrqV69emlfwkWxMzYZw4DwIB+qB/i4OpyzO77dvNZ8dBERuUAMwyjyMnz4cMc+vr6+TJo0iZMnT5KWlsasWbPOWWHnFqweENYEgCaWI0rSRUSkVDxLc9DDDz/MN998g2EY3HXXXbz55puOpVMAqlSpwttvv03NmjXP+VhDhgwhPj6eF154gdjYWNq2bcu8efMczeQOHTqE1ep8LmHnzp0sXbqU33//vTThX1TbYlIAaBbpggY25+v0cncRERE5fzVaQMxGmlgOs+BUBtm5Nrw9Xb7irYiIVCClStK3bdvGhx9+yI033oiPT9Gjw6GhoSVeqm3MmDGMGTOmyPsWL15caFuTJk0wDKPE8brStmPmkm/N3T1Jz0iE5KPm7bCmLg1FRESkwspvHtfC8zC2LDh0Mp2GNQJcHJSIiFQkpUrST59DXuwDe3rSq1ev0jx8pbItJj9Jr+nmSbq91D2oNviFuDQUERGRCit/ylhzD/PE98ETaUrSRUTkvJSq/mr8+PFMmTKl0PYpU6bwxhtvlDmoyiLPZrAz1kzS3b7c/biaxomIiJRZfpJey3YMH7LZrw7vIiJynkqVpH/yySc0bVq4JLpFixZMnjy5zEFVFvsT0sjMseHn5UF09SquDufs7CPp4WoaJyIiUmqBEeBXFSs2GlqOqnmciIict1Il6bGxsURGRhbaHhYWRkxMTJmDqiy255e6N4kIxMPq5uuOx+U3jauhpnEiIiKlZrE4PkubWg5z8ES6iwMSEZGKplRJelRUFMuWLSu0fdmyZSXq6H6pqDDz0Q1D5e4iIiLlJb8qrYn1sMrdRUTkvJWqcdzIkSN57LHHyMnJ4corrwTMZnJPPfUU//rXv8o1wIqswnR2T4mBzCSwFKzvKiIiIqWUPy+9ieUwxxIzyMrNw8fTw8VBiYhIRVGqJP3JJ5/kxIkTPPTQQ2RnZwPg6+vL008/zbhx48o1wIrMXu7u9k3j7KXu1RuCZ9FL6omIiEgJheeXu1sPYzPg8MkMdXgXEZESK1WSbrFYeOONN3j++efZvn07fn5+NGrUqNg10y9FCalZHE/JwmKBphGBrg7n7I7b56Or1F1ERKTMwszmuuGWUwSTyoEELcMmIiIlV6o56XYBAQF07NiRli1bKkE/g30UPbp6Far4lOpcyMVjT9LD1TRORESkzHyDILgOYDaP23081cUBiYhIRVLq7HHNmjV89913HDp0yFHybjdr1qwyB1bRVZj56ABxahonIiJSrsKbQ9IhmlgPsfVYkqujERGRCqRUI+nffvstXbt2Zfv27fz444/k5OSwdetW/vzzT4KDg8s7xgqpwnR2t+VB/E7zdg2tkS4iIlIu8j9Tm1oOO07ci4iIlESpkvTXX3+dd999l19++QVvb2/ef/99duzYwa233kqdOnXKO8YKqaBpnJvPRz+5D/KywNMPqtZzdTQiIiKVQ/4UsibWw+xLSCM1K9fFAYmISEVRqiR97969DBgwAABvb2/S0tKwWCw8/vjjfPrpp+UaYEWUmZPH3nhzXdTmkW5eWeAodW8K1jK1KBARkUpu2rRpzJkzx/HzU089RUhICF27duXgwYMujMwN2UfSrUcAw3HyXkRE5FxKlZVVrVqVlJQUAGrVqsWWLVsASExMJD09vfyiq6B2x6WSZzOo6u9FeJCbN9Q7vt28rqGmcSIicnavv/46fn5+AKxYsYJJkybx5ptvEhoayuOPP+7i6NxM9YZg9aIKGdQigS1HNS9dRERKplSN43r27MmCBQto1aoVt9xyC48++ih//vknCxYs4KqrrirvGCucbTHmB3HzmkFYLBYXR3MOx/NH0sM1H11ERM7u8OHDNGzYEIDZs2dz0003MWrUKLp160bv3r1dG5y78fSG0EZwfBtNrIfZqnnpIiJSQqVK0idOnEhmZiYAzz77LF5eXixfvpybbrqJ5557rlwDrIgqTGf39JNw+B/ztjq7i4jIOQQEBHDixAnq1KnD77//ztixYwHw9fUlIyPDxdG5oYhWcHwbrSz7+V1JuoiIlNB5J+m5ubn8+uuv9O3bFwCr1cozzzxT7oFVZNtjzKkAzdw5Sc9Khem3QmosBEZC7U6ujkhERNzc1VdfzX333Ue7du3YtWsX/fv3B2Dr1q1ER0e7Njh3VKsDbJrBZdbdTIpLISs3Dx9PD1dHJSIibu6856R7enrywAMPOEbSxZnNZrj/8mu5WTDjTjjyD/hVhbt+BJ8AV0clIiJubtKkSXTp0oX4+Hh++OEHqlevDsDatWu5/fbbXRydG4rqCEA7j73k2fLYFZvq4oBERKQiKFW5e6dOndiwYQN169Yt73gqvCOnMkjNysXbw0qDMDdMfG158MN9sG8ReFWBO75XqbuIiJRISEgIEydOLLT95ZdfdkE0FUB4S/D0Iyg3jfqWGLYeS6JVbTdf9UVERFyuVEn6Qw89xNixYzl8+DDt27enSpUqTve3bt26XIKriOyj6I3CA/DycLMlzQwDfnkUtv8MHt5w29dQu4OroxIRkQpi3rx5BAQE0L17d8AcWf/ss89o3rw5kyZNomrVqi6O0M14eEHNdnBoOZdZd7PlWBdXRyQiIhVAqbLI2267jf379/PII4/QrVs32rZtS7t27RzXlzJHqbs7zkf/+21Y/xVYrHDT59DgCldHJCIiFciTTz5JcrL5Obd582b+9a9/0b9/f/bv3+9oIidnyD8Z3s6yWx3eRUSkREo1kr5///7yjqPS2B1nNo1r6m5JumHAP5+bt/u9Cc0HuTYeERGpcPbv30/z5uaSnT/88APXXXcdr7/+OuvWrXM0kZMzRJmNWdtZ9/BKTDJ5NgMPq5svzyoiIi5VqiRdc9GLdyI1G4CIIF8XR3KGE3shJcYsc293p6ujERGRCsjb25v09HQA/vjjD4YNGwZAtWrVHCPscobaZvO4JpYjeGSnsS8+lUbhgS4OSkRE3FmpkvQvv/zyrPfbP7QvRafSzSS9ahUvF0dyhgN/m9e1O4KXn2tjERGRCql79+6MHTuWbt26sXr1ambMmAHArl27qF27toujc1OBERBcB2vSIdpY97L1WLKSdBEROatSJemPPvqo0885OTmkp6fj7e2Nv7+/knSgqr+3iyM5gz1Jj+7h2jhERKTCmjhxIg899BDff/89H3/8MbVq1QLgt99+49prr3VxdG6sdgdIOkQ7yx62HE3i+na1XB2RiIi4sVIl6adOnSq0bffu3Tz44IM8+eSTZQ6qojIMg1PpOYCbJemGAfvzk/R6StJFRKR06tSpw6+//lpo+7vvvuuCaCqQqE6wdRaXWXfzXzWPExGRcyhVkl6URo0a8X//93/ceeed7Nixo7wetkJJzswlz2YAEOLvRuXuCbsg7Th4+EAtLbkmIiKll5eXx+zZs9m+fTsALVq0YNCgQXh4eLg4MjeWPy+9nXU3W48lYhgGFouax4mISNHKdSFvT09Pjh07dl7HTJo0iejoaHx9fencuTOrV68+6/6JiYmMHj2ayMhIfHx8aNy4MXPnzi1L2OUmMb/U3d/bA18vN/qysv8v8zqqE3i5WUM7ERGpMPbs2UOzZs0YNmwYs2bNYtasWdx55520aNGCvXv3ujo89xXRGsPDh2qWVKplHeHIqQxXRyQiIm6sVCPpP//8s9PPhmEQExPDxIkT6datW4kfZ8aMGYwdO5bJkyfTuXNn3nvvPfr27cvOnTupUaNGof2zs7O5+uqrqVGjBt9//z21atXi4MGDhISElOZllLuTaW4+H71eT9fGISIiFdojjzxCgwYNWLlyJdWqVQPgxIkT3HnnnTzyyCPMmTPHxRG6KU9vLJFt4Mhq2ln2sPVYElHV/F0dlYiIuKlSJenXX3+9088Wi4WwsDCuvPJK3nnnnRI/zoQJExg5ciQjRowAYPLkycyZM4cpU6bwzDPPFNp/ypQpnDx5kuXLl+PlZZaTR0dHl+YlXBCJ9vno7tTZ3WaDA0vN22oaJyIiZbBkyRKnBB2gevXq/N///d95naS/JEV1giOrucy6my1Hk7m2ZaSrIxIRETdVqnJ3m83mdMnLyyM2Npbp06cTGVmyD53s7GzWrl1Lnz59CoKxWunTpw8rVqwo8piff/6ZLl26MHr0aMLDw2nZsiWvv/46eXl5xT5PVlYWycnJTpcLxS07u8dvh/QT4OUPtdq7OhoREanAfHx8SElJKbQ9NTUVb283+uxzR7XNnjDtrOZIuoiISHHKdU76+UhISCAvL4/w8HCn7eHh4cTGxhZ5zL59+/j+++/Jy8tj7ty5PP/887zzzju8+uqrxT7P+PHjCQ4OdlyioqLK9XWczi3L3e2j6FGdwdON4hIRkQrnuuuuY9SoUaxatQrDMDAMg5UrV/LAAw8waNCgEj/OX3/9xcCBA6lZsyYWi4XZs2c73T98+HAsFovTpcIv8Va7EwBNLYfYe/S4i4MRERF3Vqok/aabbuKNN94otP3NN9/klltuKXNQxbHZbNSoUYNPP/2U9u3bM2TIEJ599lkmT55c7DHjxo0jKSnJcTl8+PAFi89R7u5Ond3tTeO09JqIiJTRBx98QIMGDejSpQu+vr74+vrStWtXGjZsyHvvvVfix0lLS6NNmzZMmjSp2H2uvfZaYmJiHJdvvvmmHF6BCwXXwhYYiafFRmTaDo6nZLo6IhERcVOlmpP+119/8dJLLxXa3q9fvxLPSQ8NDcXDw4O4uDin7XFxcURERBR5TGRkJF5eXk7LvDRr1ozY2Fiys7OLLLXz8fHBx8enRDGV1Ul7uXsVNxmxttng4DLzdrSaxomISNmEhITw008/sWfPHscSbM2aNaNhw4bn9Tj9+vWjX79+Z93Hx8en2O8DRcnKyiIrK8vx84Wc3lZa1qhOsO0nLrPuZtPhJPo014orIiJSWKmS9OLmnnl5eZX4Q9Hb25v27duzcOFCRyM6m83GwoULGTNmTJHHdOvWjenTp2Oz2bBazSKAXbt2ERkZ6RZz4RLdbU563BbIOAXeAVCzraujERGRCmjs2LFnvX/RokWO2xMmTCi35128eDE1atSgatWqXHnllbz66qtUr1692P3Hjx/Pyy+/XG7Pf0HU7gjbfqKddTer9p+gT/Pwcx8jIiKXnFIl6a1atWLGjBm88MILTtu//fZbmjdvXuLHGTt2LHfffTcdOnSgU6dOvPfee6SlpTm6vQ8bNoxatWoxfvx4AB588EEmTpzIo48+ysMPP8zu3bt5/fXXeeSRR0rzMsqdY066u4yk25deq9MFPNyoBF9ERCqM9evXl2g/i8VSbs957bXXcuONN1KvXj327t3Lv//9b/r168eKFSucqulON27cOKcTCsnJyRe0D02p5M9Lb2fdzcR9J1wcjIiIuKtSJenPP/88N954I3v37uXKK68EYOHChXzzzTfMnDmzxI8zZMgQ4uPjeeGFF4iNjaVt27bMmzfP0Uzu0KFDjhFzgKioKObPn8/jjz9O69atqVWrFo8++ihPP/10aV5GuXO7Oen785P06O6ujUNERCqs00fKL5bbbrvNcbtVq1a0bt2aBg0asHjxYq666qoij7mY09tKLbINhtWLMFsyScd2k5x5OUG+bvKdQURE3EapkvSBAwcye/ZsXn/9db7//nv8/Pxo3bo1f/zxB7169TqvxxozZkyx5e2LFy8utK1Lly6sXLmyNGFfcG61BJstDw4uN2+raZyIiFRg9evXJzQ0lD179hSbpFcIXr5YaraFI//QybqdNQdOcmVTlbyLiIizUiXpAAMGDGDAgAHlGUuFZhgGp9LyR9Ldodw9ZiNkJYFPEES0cXU0IiIipXbkyBFOnDhBZGSkq0MpuwZXwpF/6GXdyKp9StJFRKSwUi3B9s8//7Bq1apC21etWsWaNWvKHFRFlJ6dR3aeDXCTcnf7+uh1u4JHqc/FiIiIlLvU1FQ2bNjAhg0bANi/fz8bNmzg0KFDpKam8uSTT7Jy5UoOHDjAwoULGTx4MA0bNqRv376uDbw8NLwagB7Wzazeq/XSRUSksFIl6aNHjy5yvfGjR48yevToMgdVEdmbxvl4WvHzKrqpzUUVu8m8jurs2jhERETOsGbNGtq1a0e7du0As5Fsu3bteOGFF/Dw8GDTpk0MGjSIxo0bc++999K+fXv+/vtv959zXhK1LsPmE0KwJR3PmLWkZOa4OiIREXEzpRpi3bZtG5dddlmh7e3atWPbtm1lDqoiKmga512uHW5LLT2/a2xgJSgNFBGRSqV3794YhlHs/fPnz7+I0VxkVg+sDa+ErbPoYd3ImgOnuKJpDVdHJSIibqRUI+k+Pj7ExcUV2h4TE4On56VZWn0y3c2WX0s/aV77VXVtHCIiIuKskVny3su6iZX7tRSbiIg4K1WSfs011zBu3DiSkpIc2xITE/n3v//N1VdfXW7BVSSJjs7ubjAfHSDjlHntX821cYiIiIizBubytW2s+9i+Z5+LgxEREXdTqmHvt99+m549e1K3bl3HfLINGzYQHh7OV199Va4BVhSn0txo+TUoSNL9lKSLiIi4lcAIssNa4h2/hdC4paRm9SXA59KsRBQRkcJKNZJeq1YtNm3axJtvvknz5s1p374977//Pps3byYqKqq8Y6wQTtrnpFdxg5H0vBzISjZvayRdRETE7Xg3uQaAHpaNrDlw0sXRiIiIOyn1adsqVarQvXt36tSpQ3a2OYr822+/ATBo0KDyia4CKSh3d4ORdPsoOhbwDXZpKCIiIlKEhn1g6QR6Wjfx2d4EejdR8zgRETGVKknft28fN9xwA5s3b8ZisWAYhlNH87y8vHILsKI46U7l7vamcb7BYHWD5eBERETEWVQncjwDqJ6bQsKuVdC/uasjEhERN1GqcvdHH32UevXqcfz4cfz9/dmyZQtLliyhQ4cOLF68uJxDrBgS3ancXU3jRERE3JuHFzl1ewJQK2EZaVm5Lg5IRETcRamS9BUrVvDKK68QGhqK1WrFw8OD7t27M378eB555JHyjrFCsI+kh7jDSHqGffk1JekiIiLuyr95XwBzvfSDp86xt4iIXCpKlaTn5eURGBgIQGhoKMeOHQOgbt267Ny5s/yiq0Dsc9KruUOSbi9310i6iIiI+2pwFQDtLLtZv2u/i4MRERF3Uao56S1btmTjxo3Uq1ePzp078+abb+Lt7c2nn35K/fr1yzvGCuGUvdzdHZJ0jaSLiIi4v5AokgIaEJy6l9xdC4FOro5IRETcQKlG0p977jlsNhsAr7zyCvv376dHjx7MnTuXDz74oFwDrAgyc/LIyDGb5bnFnHT7SLpfVdfGISIiImdlaXQ1AHVPrdC8dBERAUo5kt63b1/H7YYNG7Jjxw5OnjxJ1apVnbq8XypO5Ze6e1otBPiUelW78qPGcSIiIhVCUMtrYf1kelo3smTncfq3runqkERExMVKNZJelGrVql2SCTqctvxaFW/3eA8yNJIuIiJSIdTtSrbVj3BLIntX/+bqaERExA2UW5J+KXMsv+bvBqXuAOkaSRcREakQPH1IbHILAG0Pf0lm/vQ5ERG5dClJLwdutfwaaCRdRESkAgntM5Y8rPSwbGDjmqWuDkdERFxMSXo5cKvl16BgTrq6u4uIiLg9a/V6bK96BQCeKye6OBoREXE1JenlwLH8mjt0djcMrZMuIiJSwRhdHwGgTdIf5Jw44NpgRETEpZSklwNH4zh3GEnPSYe8LPO2RtJFREQqhObte7GKVnhi4/iC91wdjoiIuJCS9HJgL3d3iyTdPoru4Q3eVVwbi4iIiJSIh9XC1nrDAQjd9W3B57mIiFxylKSXg5P55e4h7tDd/fSmce6wHJyIiIiUSIPLB7HNVhcfWwa2fz53dTgiIuIiStLLgaNxXBU3GElX0zgREZEKqUuDUL60DgYgb8XHkJPh4ohERMQVlKSXA7dagk1N40RERCokb08ruU0Hc8QIxSvzBGz8xtUhiYiIC7hFkj5p0iSio6Px9fWlc+fOrF69uth9v/jiCywWi9PF19f3IkZbWGJ+ubt7jKRrjXQREZGK6ppWtflvbn8AjOUfgi3PxRGJiMjF5vIkfcaMGYwdO5YXX3yRdevW0aZNG/r27cvx48eLPSYoKIiYmBjH5eDBgxcxYmfZuTZSs3IBqOoOc9LT88vdNZIuIiJS4fRsHMYvHldx0gjAcnIfbP3R1SGJiMhF5vIkfcKECYwcOZIRI0bQvHlzJk+ejL+/P1OmTCn2GIvFQkREhOMSHh5+ESN2Zp+PbrVAkK8bJOmOOekaSRcREalofL08uLxpHabmXmtu+PsdsNlcG5SIiFxULk3Ss7OzWbt2LX369HFss1qt9OnThxUrVhR7XGpqKnXr1iUqKorBgwezdevWYvfNysoiOTnZ6VKeTjk6u3tjtbpBN3VHubtG0kVERCqifi0jmJbXl1T84fg22PWbq0MSEZGLyKVJekJCAnl5eYVGwsPDw4mNjS3ymCZNmjBlyhR++ukn/ve//2Gz2ejatStHjhwpcv/x48cTHBzsuERFRZXrayhoGucGo+igxnEiIiIV3BVNapDrHcS03PxBjL/eBsNwbVAiInLRuLzc/Xx16dKFYcOG0bZtW3r16sWsWbMICwvjk08+KXL/cePGkZSU5LgcPny4XONxLL/mDp3dQSPpIiIiFVwVH08Gt63J57n9ybb4wLF1sPdPV4clIiIXiUuT9NDQUDw8PIiLi3PaHhcXR0RERIkew8vLi3bt2rFnz54i7/fx8SEoKMjpUp5OprvR8mugkXQREZFK4I7OdTlJENNzrzA3/P2OawMSEZGLxqVJure3N+3bt2fhwoWObTabjYULF9KlS5cSPUZeXh6bN28mMjLyQoV5VgXLr7lJubsax4mIiFR4LWsF0zYqhMk5A8izeMHBZXBwuavDEhGRi8Dl5e5jx47ls88+Y9q0aWzfvp0HH3yQtLQ0RowYAcCwYcMYN26cY/9XXnmF33//nX379rFu3TruvPNODh48yH333eeS+E/lz0mv6g4j6TYbZCaat1XuLiIiUqHdeXldYqnOL9b80fS/3nZtQCIiclF4ujqAIUOGEB8fzwsvvEBsbCxt27Zl3rx5jmZyhw4dwmotOJdw6tQpRo4cSWxsLFWrVqV9+/YsX76c5s2buyR+typ3z0wEI3+ZFo2ki4iIVGjXtY7kP79u4530fgzyXYh170I4ug5qXebq0ERE5AJy+Ug6wJgxYzh48CBZWVmsWrWKzp07O+5bvHgxX3zxhePnd99917FvbGwsc+bMoV27di6I2uRW5e72UnfvQPB0g5MGIiIiRfjrr78YOHAgNWvWxGKxMHv2bKf7DcPghRdeIDIyEj8/P/r06cPu3btdE6wL+Xp5cHP72hw2wlnpr7npIiKXCrdI0iuygiXY3CAptjeN0yi6iIi4sbS0NNq0acOkSZOKvP/NN9/kgw8+YPLkyaxatYoqVarQt29fMjMzL3KkrndH5zoAvHCqr7lhxxw4sdeFEYmIyIWmJL2MHEuwVXGDJN0+ku6vJF1ERNxXv379ePXVV7nhhhsK3WcYBu+99x7PPfccgwcPpnXr1nz55ZccO3as0Ij7paB+WADdGlZnj60We0O6AQas/NjVYYmIyAWkJL2MTjoax7lDubvWSBcRkYpt//79xMbG0qdPH8e24OBgOnfuzIoVK4o9Lisri+TkZKdLZXFn57oAvJ18lblhw9cF1XMiIlLpKEkvg9w8G8mZuYCbdHfXGukiIlLBxcbGAjgayNqFh4c77ivK+PHjCQ4OdlyioqIuaJwXU5/m4dQI9OG39CYkBTWBnHRYN83VYYmIyAWiJL0MEjNyHLeD/TSSLiIi4irjxo0jKSnJcTl8+LCrQyo3Xh5WbusYBVj4yjLA3LjqE8jNdmlcIiJyYShJLwP7fPQgX088PdzgrbTPSVfjOBERqaAiIiIAiIuLc9oeFxfnuK8oPj4+BAUFOV0qk6Gd6+LtYeWDuDZk+4ZBSgxsm+3qsERE5AJwg8yy4jrlWH7NDUrdQeXuIiJS4dWrV4+IiAgWLlzo2JacnMyqVavo0qWLCyNzrYhgX4Z2rkM2Xsz06GduXDERDMO1gYmISLlTkl4GbrX8GqjcXUREKoTU1FQ2bNjAhg0bALNZ3IYNGzh06BAWi4XHHnuMV199lZ9//pnNmzczbNgwatasyfXXX+/SuF3tod4N8PWy8vaJruR5+ELMRji4zNVhiYhIOVOSXgZutfwaaCRdREQqhDVr1tCuXTvatWsHwNixY2nXrh0vvPACAE899RQPP/wwo0aNomPHjqSmpjJv3jx8fX1dGbbL1Qjy5e4u0ZwiiHmeV5gbVxS91ryIiFRcStLL4GSaWe4e4g7Lr8Fpc9KVpIuIiPvq3bs3hmEUunzxxRcAWCwWXnnlFWJjY8nMzOSPP/6gcePGrg3aTTzQqwEBPp68Y1+ObedvkLDHtUGJiEi5UpJeBvaRdLdYfg1OS9JDXBqGiIiIXBhVq3hzT/d67DNqssKzI2DAig9dHZaIiJQjJellUKuqH53qVaNhjQBXh2Iuw5Kdat5WubuIiEildW/3egT7eTEhLb+B3LovIW6ra4MSEZFyoyS9DIZ1iea7+7twe6c6rg6loGmcxQo+wa6NRURERC6YYD8vRvWszz9GUxZ7dAXDBr89rU7vIiKVhJL0ysLeNM6vKlj1axUREanMhneNJjTAm2fThpBr9YEDf8PWH10dloiIlANlc5WFmsaJiIhcMqr4ePJg74YcJYzPGWxu/P15yE5zbWAiIlJmStIri4zTRtJFRESk0rvr8ro0rBHAhPT+nPKKgOQjsPQ9V4clIiJlpCS9stAa6SIiIpcUb08r429sRRbePJN2m7lx2ftwcr9rAxMRkTJRkl5ZOEbSlaSLiIhcKjpGV+P2TlHMt3VkrUdryMuC359zdVgiIlIGStIrC42ki4iIXJKeubYZoQG+PJ1+JzY8YMevsHuBq8MSEZFSUpJeWTgax2lOuoiIyKUk2N+LFwY2Z49Rm2m2a8yNs0bCib2uDUxEREpFSXpFk3ocpg6ATTOdtytJFxERuWQNbB1Jr8ZhvJF9K3u8GpvfC6bfWvD9QEREKgwl6RXN1h/h4FKY+y/ISinYrnJ3ERGRS5bFYuHV61uClx+3pzxGmm8EnNgD3w2DvBxXhyciIudBSXpFk7DLvM5MgrXTCrarcZyIiMglLaqaP2Ovbkw8Idye+jh5XlVg/18w519gGK4OT0RESkhJekWTsLvg9opJkJtt3tZIuoiIyCXvvu716d0kjE25UTxrfQzDYoV108zvDCIiUiEoSa9o7Em61RNSjsHmmebZcc1JFxERueRZrRbevbUttUL8+DapBTOqPWDe8ftzsPM31wYnIiIloiS9IslKMRNzgK6PmNfL3je32/Lnm6ncXURE5JJWtYo3H995Gd4eVp452o2tNW8GDPjhPojb6urwRETkHJSkVyQn9pjXVWpA98fAJwgSdsLGb8ztnr7g7e+y8ERERMQ9tK4dwouDmgMWbtg/iKTwLpCdCt/cBmkJrg5PRETOwi2S9EmTJhEdHY2vry+dO3dm9erVJTru22+/xWKxcP3111/YAN2FvdQ9tBH4BkOHe8yfl7xpXmsUXURERPIN7VSHGy+rRbbhyY0Jo8gNjobEQzDjroKeNiIi4nZcnqTPmDGDsWPH8uKLL7Ju3TratGlD3759OX78+FmPO3DgAE888QQ9evS4SJG6AXtn99BG5vXlD4KHN6TnnxFX0zgRERHJZ7FYeO36VjSNCGRvmg8P2p7C8A6EQ8thzlh1fBcRcVMuT9InTJjAyJEjGTFiBM2bN2fy5Mn4+/szZcqUYo/Jy8vjjjvu4OWXX6Z+/foXMVoXc4ykNzavAyOgze0F96tpnIiIiJzGz9uDz4Z1oEagDwviQ3i9ypNmx/f1X8HKj10dnoiIFMGlSXp2djZr166lT58+jm1Wq5U+ffqwYsWKYo975ZVXqFGjBvfee+85nyMrK4vk5GSnS4VlT9KrNyrY1vURwGLeVpIuIiIiZ4iq5s+0ezoR6OvJZzEN+a7q/eYdvz8La4ofFBEREddwaZKekJBAXl4e4eHhTtvDw8OJjY0t8pilS5fy+eef89lnn5XoOcaPH09wcLDjEhUVVea4XcKWV9A4LvS0JD20ITQbaN4OCC98nIiIiFzymkUG8d9hHfD2tPL0se6sqDYYDBv8+jj89jTk5bo6RBERyefycvfzkZKSwl133cVnn31GaGhoiY4ZN24cSUlJjsvhw4cvcJQXSNJhyMsCDx8IqeN8X783odP95hx1ERERkSJ0rl+dibe3w2qxcPuxW/k7Kn8N9VWTYfqtkJnk2gBFRAQAT1c+eWhoKB4eHsTFxTltj4uLIyIiotD+e/fu5cCBAwwcONCxzWazAeDp6cnOnTtp0KCB0zE+Pj74+PhcgOgvMkepewOwejjfFxQJ/d+8+DGJiIhIhXJNiwhev6EVz8zazF27e/J+67oM2vcKlr0L4b9Xw9Bvodol1O9HRMQNuXQk3dvbm/bt27Nw4ULHNpvNxsKFC+nSpUuh/Zs2bcrmzZvZsGGD4zLo/9u78/ioqruP45/ZsyckIQuQkLDIviibgGvFUrXUpQpaVBAfrFWsiviIpSpupS1qcbetW92KVsVHRaWKQBUBgYKKQtgJS0ICIZmsk2TmPH/cMBgB2UJmhnzfr9d5zeTOmTu/M5dw8rvn3HN/8QvOPvtsVq5cGblT2Q/HD1d2FxERETkKlw/M5s7zugJw89fteaz9Y5j4NrArD/7+E9g4P7QBioi0cCEdSQeYOHEiY8aMoX///gwcOJAZM2ZQWVnJNddcA8DVV19N27ZtmTZtGlFRUfTs2bPR+5OSkgD2237C+eHK7iIiIiJH6ddndiQx2sXvZn3DX76NJb/zX5ge9yfsBf+Fly+Bnz5gXUZns4U6VBGRFifkSfqoUaMoLi7m7rvvprCwkL59+/LRRx8FF5PLz8/Hbo+oS+ePjwOt7C4iIiJylC4fmE1yrJub/rmCt9b5Kci+mxd7vIb72zdgzp1Q+DX8/C/gig51qCIiLYrNGGNCHURz8nq9JCYmUlZWRkJCQqjDOXzTO0NlEYyfB21PCXU0IiLShCK2bwpj+k4P35ebSviffyzFW1NP59axvN53Jcmf3wfGD21OhlGvQmLbUIcpIhLRjqRf0hB1JKgutRJ00DXpIiIi0qQG5ibzxvWDSU/wsK64krM+68bKM5+H6FawYwX87SzYtizUYYqItBhK0iPB3vujx2eCJz60sYiIiMgJp2tGAv9342mcnJ2Et6aei+e4eLHni5i07tZAwQvnwzdvhjpMEZEWQUl6JNDK7iIiInKcZSRGMfO6U7ny1GyMgamfVTIh+o/Udfwp+H3w1rUw/4/Qsq6UFBFpdkrSI4FWdhcREZFm4HE6eOCiXvz50t64nXZm51Xw04JfU9hjvFVh/jQrWa+rDm2gIiInMCXpkWDvSLpWdhcRkRZg6tSp2Gy2RqVr166hDqtFGdk/i7euH0LbpGg2lfgY8t+zmdPhTozdCaveghfOg9L8UIcpInJCUpIeCYIj6UrSRUSkZejRowcFBQXB8vnnn4c6pBanV7tEPrj5dC45pS0BA7/+rheTY+7D70m0FpT76xmw7uNQhykicsJRkh7u/PVQstF6runuIiLSQjidTjIyMoIlNTU11CG1SInRLh4Z2ZdnrjyF5Fg3r+/KYVjF/eyM7w7Ve+DVy+DTByHgD3WoIiInDCXp4a50CwTqwBkNCbpHqYiItAzr1q2jTZs2dOjQgdGjR5Of/+NTq30+H16vt1GRpvOznpl8dMvpDOuWxiZ/KqcX38E7zp8BBv7zZ3j5YijO06JyIiJNwBnqAOQQgiu7dwK7zqmItCR+v5+6urpQhyFNwOVy4XA4Qh1GxBg0aBAvvvgiXbp0oaCggHvvvZfTTz+dVatWER9/4FuRTps2jXvvvbeZI21Z0uKj+PvV/Xln5XamfbCGW8qvZp69A3/2PIdn0wJ4cqB1b/W2/aHdAGjXH3JOA6cn1KGLiEQUmzEt65Sn1+slMTGRsrIyEhISQh3OoS18DD6+C3r+Ei59PtTRiEgzMMZQWFhIaWlpqEORJpSUlERGRgY2m22/1yKub2pmpaWltG/fnkceeYRrr732gHV8Ph8+ny/4s9frJSsrS9/pcVLpq+fJeet59rNNZAfyudf1EgMda3GZ2sYVY1Kh3xjoPw4S24UmWBGRMHAkfb1G0pvDv39vLaxy6QuQ3v3I3quV3UVanL0JelpaGjExMQdM6iRyGGOoqqqiqKgIgMzMzBBHFHmSkpI46aSTWL9+/UHreDwePB6N2DaXWI+T//1ZV0YNyOKB2asZ/V07XNTT1ZbPZRkFDE/cStquL7FVFMJnD8Pnf4Eu58PA6yD3DND/ayIiB6Uk/Xjbthy+eNx6/vLFcO0caJVz+O/Xyu4iLYrf7w8m6CkpKaEOR5pIdHQ0AEVFRaSlpWnq+xGqqKhgw4YNXHXVVaEORX6gfUosf7+6P8s2l/DsZ5v493dOvinowN0F0KX1Nfz+5E0MKZmFY8tnsOZ9q3QeDiNmQEKbUIcvIhKWdJHz8WQM/HuK9dzmgIpCeOlCKC88/H3s3puka2V3kZZg7zXoMTExIY5EmtreY6p1Bg5t0qRJLFiwgM2bN/PFF19w8cUX43A4uOKKK0IdmhxE/5xknrmqH/Mnnc01Q3OIdTvIK67mqkUZDNh2My/2/SdVfa4BhxvWzYEnT4WVr2mhORGRA1CSfjytfg/yF1krs//PJ9YI+p7N8PIl1m1LDqVyN1Tttp6ndDyekYpImNEU9xOPjunh27ZtG1dccQVdunRh5MiRpKSksHjxYlq3bh3q0OQQslNiuGdEDxb97hymnN+NtknRlFTWMnWxoc+yn/KnnL9Tk9YHfGXwzm/gtVHgLQh12CIiYUVJ+vFSXwuf3GM9H3ITtD0FrnoH4jKg6Ft4dSTUVv74PgpWWo+JWeCOPZ7RioiIhI2ZM2eyY8cOfD4f27ZtY+bMmXTsqJPVkSQhysX4Mzqw4PazeGr0KfRr34o6v+Hpb1303DqJd1PHE7DvHVUfCB/eAYWrQh22iEhYUJJ+vCx7Dko2QmwaDP2ttS05F66aBVFJsO1LeP0qCAQOvo9v37YeO51z3MMVEQknOTk5zJgxI9RhiMgxcjrsnN8rk7d+M4R3bhzK8B7p1BsHv912Nj+reYAtUV3B54Ulz8AzQ+GvZ8LSZ6G6NNShi4iEjJL046F6Dyz4k/X8J1PA8717uqZ3h9FvgisGNsyFjZ8eeB/1PvjuPet5z0uPb7wiIk3grLPO4pZbbmmSfS1dupTrrruuSfYlIuGhb1YSf72qP7N/exrDe6SzNtCOs0t/z9i6/2VpzOkE7C5rFuHs2+DhLjDrN7D1S123LiItjpL04+E/D1mJelp3OPkAK9FmDYBTrraeL3/xwPtY/4l1vVZ8G2g/5LiFKiLSXIwx1NfXH1bd1q1ba/E8kRNUjzaJ/PWq/rx/02mc0z2T+f6+XFbyG/pXPc6jznHsju0I9TXw1Wvw3Lnw9FD48u8aXReRFkNJelMr2QhL/mo9/+n9YD/IbXZOGWM95n144NXev3nTeux5ycH3ISItgjGGqtr6kBRzmCNYY8eOZcGCBTz66KPYbDZsNhsvvvgiNpuNDz/8kH79+uHxePj888/ZsGEDF154Ienp6cTFxTFgwAA++eSTRvv74XR3m83Gs88+y8UXX0xMTAydO3fm3XffbcqvWUSaWc+2ifz96v7Mn3QW40/PJRCTwl8qhtFv9338su4+5kcPo97usdby+WAS/LkD/O0smDMF1syGqpJQN0FE5LjQfdKb2uczIFAHHX8CnYYdvF56d8gaBFuXwMpX4fTb9r3mq7CSd4Cevzyu4YpI+Kuu89P97jkh+ezv7htOjPvQXcWjjz7K2rVr6dmzJ/fddx8A3377LQCTJ0/moYceokOHDrRq1YqtW7dy/vnn8+CDD+LxeHjppZcYMWIEeXl5ZGdnH/Qz7r33Xv785z8zffp0Hn/8cUaPHs2WLVtITk5umsaKSEjkpMYy5YLu3PbTLrz/dQEvL97C8q2dGLunEwmM5BLH54zxzCc3kA87Vlhl0RPWm9ucDL0ugx6XQEJmaBsiItJElKQ3tW1LrceBh3EtZb+xVpK+/B8w9FawN0xsyPsA6qshuYPV+YiIhLnExETcbjcxMTFkZGQAsGbNGgDuu+8+zj333GDd5ORk+vTpE/z5/vvvZ9asWbz77rtMmDDhoJ8xduzY4H2y//CHP/DYY4/x5Zdf8rOf/ex4NElEmlmUy8Gl/dpxab92bC2p4pPVO5m7uohXNsbzYtXPyGQ3A+xruCBhI4Oca0mq2LAvaZ8zBXLPsBL2rhdAjE7eiUjkUpLelOprYdda63l6z0PX734RfDgZSrfApvnW6Dvsm+re6zLQfXVFWrxol4Pv7hsess8+Vv3792/0c0VFBVOnTmX27NkUFBRQX19PdXU1+fn5P7qf3r17B5/HxsaSkJBAUVHRMccnIuEnKzmGa4bmcs3QXLw1dczPK+b/Vmxn9tpU3i0dCkAbp5ffZq7mp/7/kFyyAjYtsMq7N0Fmb+hwFuSeCdmDwa01LkQkcihJb0q710OgHjwJkNju0PXdMdBnFHz5N2sBuY4/sa6v2jDXel2ruosI1vXYhzPlPFzFxsY2+nnSpEl8/PHHPPTQQ3Tq1Ino6GguvfRSamtrf3Q/Lper0c82m43Aj93GUkROCAlRLn7Rpw2/6NOGovIa3lmxnX8t28a6Ipi8dRCTGURH5y4mtF7JT+o/J7F8LRR8ZZWFj4LDDa27Wgv6pnVrKN2tv9U0GCIiYShy/+oLR0XfWY9p3Q7/P/1+Y60kfc1sqCiCNe9biX5GL2h90nELVUSkqbndbvx+/yHrLVy4kLFjx3LxxRcD1sj65s2bj3N0InIiSIuP4rozOjL+9A58s72M2d8U8NGqQjbsTuXWgmHAMDLtpVyZvoVzo1bToXwZzoodUPi1Vb4vMcsaae9wpjVVPj4jJG0SEfkhJelNKZikdz/896T3gHYDrGvZV74K6xpWOO51WdPHJyJyHOXk5LBkyRI2b95MXFzcQUe5O3fuzNtvv82IESOw2WzcddddGhEXkSNis9no3S6J3u2SmPyzrqwuKOejVQV8uKqQdUUwvSCJ6fQBRnFm6yrOTyvhlOgC2tdvwV2SZ12eWLYVVr5iFbBG23PPsEr7obquXURCRkl6U9p5FEk6WKPp25Zat27bezu2Hpc0aWgiIsfbpEmTGDNmDN27d6e6upoXXnjhgPUeeeQRxo0bx5AhQ0hNTeWOO+7A6/U2c7QicqKw2Wx0b5NA9zYJTPxpFzbvquST1Tv5ZPVOlm7ew4LiWBYUxwJZ2GwD6ZIez+m9YhgWt5GevpXEbv8cCr6G4jVW+fJvgM26rj33TOhyHrQbCA792SwizcNmDvcmuCcIr9dLYmIiZWVlJCQkNO3OZ/S2FoEbOxtyTjv899VWwsNdwdfwR2r2YBj3UdPGJiIRoaamhk2bNpGbm0tUVFSow5Em9GPH9rj2TS2UvlMBKK2qZcHaYhZvLGHJpt1sLK7cr052cgxnZDkYHrOO7r6vSC5ajG1XXuNK0clw0nArYc89E6KTmqcBInLCOJJ+KSxOCT755JNMnz6dwsJC+vTpw+OPP87AgQMPWPftt9/mD3/4A+vXr6euro7OnTtz2223cdVVVzVz1D/gK7cSdDjykXR3LPQeCUuftX7WvdFFREREjllSjJsL+7blwr5tASgqr+HLTSV8uamEZZv3sLrQS35JFa+UwCtkABm4necxNK2e8+PXMdC/gna7PsNRXQJf/dMqAJ5Ea+G5xLbWY+tu1rXtqSdpMToROWYhT9Jff/11Jk6cyDPPPMOgQYOYMWMGw4cPJy8vj7S0tP3qJycnM2XKFLp27Yrb7eb999/nmmuuIS0tjeHDQ3OLIgCKrPsBE5dxdNcw9RtrJel2J/S4uElDExERERFr4bmf927Dz3u3AcBbU8eK/FKWbS5h+ZY9fLO9jPKaeubtsDOPLkAXHFzGeYlbGBn3NafULCauMh98ZVBUBkXfNv6A+DbWrd86nAXZp0JStpJ2ETliIZ/uPmjQIAYMGMATTzwBQCAQICsri5tuuonJkycf1j5OOeUULrjgAu6///5D1j1u09+W/wPe+611G7WrZh3dPr55Ezzx1nQqEWmRNN39xKXp7s1L36kcjUDAkF9Sxdfby/hmWynLtuzh621l+AP7/lyOoYZ+SZWcke7j5MRKOnv2kLh7JWxZBH5f4x1Gt4LMPpDZ13rM6A3JuWB3NGu7RCT0Ima6e21tLcuXL+fOO+8MbrPb7QwbNoxFixYd8v3GGD799FPy8vL405/+dMA6Pp8Pn2/ff5jHbXGio1nZ/Yd66b7oIiIiIqFit9vISY0lJzWWX/SxRtsrfPUs21zCoo27WbxhN99sh89Ko/isdN/7WsefTq+M2zg7dhP9/V/Rvmwp0SWrsVXvgY3zrbKXM9q6XW96d0jvCW1OsRapc0U3Z1NFJIyFNEnftWsXfr+f9PT0RtvT09NZs2bNQd9XVlZG27Zt8fl8OBwOnnrqKc4999wD1p02bRr33ntvk8Z9QE2RpIuIiIhIWInzODmrSxpndbEuwyyvqWPZlj0saViM7pttZRSX+/i03MenpALnAOfgsdVxVtIuTo/bTm/7ZrJr15FYvh5bfTXs+K9V9rI7rYS93QDI6AUmAHXVUFdlPTpc1hT6dgM0Ci/SAoT8mvSjER8fz8qVK6moqGDu3LlMnDiRDh06cNZZZ+1X984772TixInBn71eL1lZWU0fVPD2a92aft8iIiIiEhbio1yc3SWNsxuS9qraevIKy1m7s5w1DY95heXsqoA5ezKZsycT6A+AnQCDEksZnrqbflE76FC/gZhdX2GrLIaClVY5mPnTrFXmO58LJ/3MusRSq8yLnJBCmqSnpqbicDjYuXNno+07d+4kIyPjoO+z2+106tQJgL59+7J69WqmTZt2wCTd4/Hg8XiaNO79VBRB1S7ABq27Ht/PEhEREZGwEeN2cnJ2K07ObtVoe3G5j7zCctYUellTWM53O7ysKfSyqCyZRWXJQGfgTNxOG6e2quTM2C30ta2jnX8b0dHRxMbG4/DEgCvW+jtz/VyoLoGvX7cKNkjtDG37WVPm254C6T00bV7kBBDSJN3tdtOvXz/mzp3LRRddBFgLx82dO5cJEyYc9n4CgUCj686b3d6p7skdwB0TujhEREREJCy0jvfQOt7DaZ1Tg9sqfPX8d8selm4uYenmElZuLaWmLsB/imP4T3E3YN+MTJsNslrF0Cktjo6tY8k67Xf08K8hp+QzkrbNxbF7Hexaa5W9t4YDiG0NiVmQlGU9xqVbI+5RiRCVZD13xYDDDU6P9eiKVnIvEkZCPt194sSJjBkzhv79+zNw4EBmzJhBZWUl11xzDQBXX301bdu2Zdq0aYB1jXn//v3p2LEjPp+PDz74gJdffpmnn346dI3QVHcREREROYQ4j5MzTmrNGSe1BsAfMGzfU82G4go2FFewvqihFFdQWlVHfkkV+SVVfBpcqskGnAGcQceYas5N2s6p7s109q8jrfxbXDUlUFlsle9f8344UjpBzmmQc7pV4tMP/R4ROS5CnqSPGjWK4uJi7r77bgoLC+nbty8fffRRcDG5/Px87HZ7sH5lZSU33HAD27ZZU4G6du3KK6+8wqhRo0LVhH0j6ek9QheDiEiEy8nJ4ZZbbuGWW24BwGazMWvWrOBMqx/avHkzubm5rFixgr59+x715zbVfkREjpTDbiM7JYbslBjO7poW3G6MYXdlLet2Wgn7puJKtpdWsb20mm17qimtqmNDVTQbqjrxDJ2AYYAhkUr6JZVzSkIF3WPKaO/YTRJleOrKcdV5cdaWYfeVYauvgfpa65ZxJmB96O71Vln+ovVzq1xr9N3utBarszutEfj07vtuK9cqF773dzq1VdaUfFcMxCQ307cocuIJeZIOMGHChINOb58/f36jnx944AEeeOCBZojqCGhldxGRJldQUECrVq0OXfEIjB07ltLSUt55553gtqysLAoKCkhNTT34G0VEmpHNZiM1zkNqnIfBHVP2e73CV8/mXZWsKSwnr+Ga9zWF5RSX2/i0NI5PS398/8mxbnJSYshJiSU3xUOneD+da78jo2QZsQWLsBV+A3s2HfjN6z/e99yTAIntoLrUSs7raxoaYIesQdDlfOh6AaR0PKrvQaSlCoskPaIFAlDUMAdJSbqISJP5sQVEm5LD4Wi2zxIRaQpxHic92ybSs21io+0llbWsKfSSV1jesGhdOSWVtVT66qnw1eOrDwTrlVTW8t/80u+9OxY4E7vtTDrG1TI4dgft4h1kxLnISHCSHusk3VVF1O7voOArKFwFPu++waq97E4I1EP+Iqt8fJe1sHJyByuZrynd9xidDKmdIKWztQheSifrMaFd4xH6vfz1ULoFqnZbM1jdsU32nYqEEyXpx6p0C9RVgsNj/ecjItLUjLHulRsKrhhr9aJD+Nvf/sbUqVPZtm1bo0uULrzwQlJSUpgyZQoTJ05k8eLFVFZW0q1bN6ZNm8awYcMOus8fTnf/8ssv+fWvf83q1avp2bMnU6ZMaVTf7/dz3XXX8emnn1JYWEh2djY33HADN998MwBTp07lH//4R3DfAPPmzSMnJ2e/6e4LFizg9ttv56uvviI5OZkxY8bwwAMP4HRa3eZZZ51F7969iYqK4tlnn8XtdnP99dczderUw/paRUSOh+RYN0M6pjKk44FnBtX7A1T46tm2p5otu6vYvLuSTbsq2bK7kh2lNez01lAfMKwrd7OuPAcKf7iHJOKjsslMvIh2bVz0jtpJjqeC6IRUEpLTSExJp3VKKin+IuxrP4K82bD5cyheY5UfqqsC7zbYOL/x9r1/V6d0tEbqy7ZZU/FLNoK/1qpjd1kr2+c2XEOfNVCL38kJQ0n6sdp79rD1SeDQ1ykix0FdFfyhTWg++3c7Dmuk4rLLLuOmm25i3rx5nHPOOQCUlJTw0Ucf8cEHH1BRUcH555/Pgw8+iMfj4aWXXmLEiBHk5eWRnZ19yP1XVFTw85//nHPPPZdXXnmFTZs2BZPvvQKBAO3ateNf//oXKSkpfPHFF1x33XVkZmYycuRIJk2axOrVq/F6vbzwwgsAJCcns2PHjkb72b59O+effz5jx47lpZdeYs2aNYwfP56oqKhGSfg//vEPJk6cyJIlS1i0aBFjx45l6NChnHvuuYdsj4hIKDgddpJi3CTFuPcbhQcIBAy7Kn0UlNawvbTaSuKLK4PJ/K6KWspr6imvqWDtTvgUN5AMBLAyeiurd9httI7rSFrC7bRvdwuD+Yp4U06piaPExFLij6GkPposTwW9ooroYCsgo24b8RWbsO/ZZF0rX7zaKvs1IsqaZl9ZBFsXW+U/0wGbtbJ9QibEt7EePQlWUl9XDfU+azq+M6qhTkNJyARPovV3vMPdUFzgjj/waP6Rqq20rtWPa33s+5IWQ1nlsQpej65F40Sk5WrVqhXnnXcer732WjBJf/PNN0lNTeXss8/GbrfTp0+fYP3777+fWbNm8e677x7WLTdfe+01AoEAzz33HFFRUfTo0YNt27bxm9/8JljH5XJx7733Bn/Ozc1l0aJFvPHGG4wcOZK4uDiio6Px+Xw/Or39qaeeIisriyeeeAKbzUbXrl3ZsWMHd9xxB3fffXdwpkDv3r255557AOjcuTNPPPEEc+fOVZIuIhHLbreRFh9FWnwUfbKS9nu9wldPYVk1O0prKGh4LCyroai8hqJyHzu9PnZX+vAHDIXeGgq9NXwNvEfng3xiFJAK7LtktHWMg96J5fSK3sVJzkIybSVUeNLYHdWeYnc2Ja40jLHRwbmLLjUraFe6jMSdS3BWFlqJe2WRNR3/mL8MJ8SmWavcx6VDbCrYHNZCe8ZYjzYbJLSBpPbQqr316IyCrUsgfzHkfwEFX4PxQ2oX6PgTq+QM1VR9+VFK0o/V3tuvpet6dBE5Tlwx1oh2qD77MI0ePZrx48fz1FNP4fF4ePXVV7n88sux2+1UVFQwdepUZs+eTUFBAfX19VRXV5Ofn39Y+169enVwevlegwcP3q/ek08+yfPPP09+fj7V1dXU1tYe8Yrtq1evZvDgwcEp8QBDhw6loqKCbdu2BUf+e/fu3eh9mZmZFBUVHdFniYhEkjiPk05p8XRKiz9onTp/gN0VtVbi7vVRVO6jqLwGf8AQ43YS63EQ43YS5bJT5PWxvriCDUUVbCiuZFeFj+IqP3OrYphLNvDDmVaVwPcXtOvYUEaSTDknxZRzUrSXXLeXLGcpSc5a3FExREfHEB0TS0xMLLH2WtyVhdjKC6C8EMoLoLbCut7dXwuBOmvXgXoo32GVprArzypLnram6qd2brhPvQecbuvRZgeMdRJg72NcesN1+w0lucPBp/X762DXOmsQsTTfOmHgjrH6cleM9T5XtPW5zobH6FZNsxK/MdYaBZ6Ew7pMTn6ckvRjpZXdReR4s9ki4oz7iBEjMMYwe/ZsBgwYwGeffcZf/vIXACZNmsTHH3/MQw89RKdOnYiOjubSSy+ltra2yT5/5syZTJo0iYcffpjBgwcTHx/P9OnTWbJkSZN9xve5XK5GP9tsNgKBwHH5LBGRSOFy2MlIjCIjMerQlX+grLqOHaXVVimrYUdpNUVeH3abNVXf5bDhtNsJGENxhY+dZdZofZHXR4k/gcVVCSyuanvIz3HYu5MQ5SQh2kVClIu4OCexHusEQqzbQZI7QJqzikx7GansoVVgDwmBUjxOGx6XE5fThc1ut5L6su3WGlV7tkDZVitRTusO2adisk+lJnMgtY5YEgoWYts4D9Z/CmX5+y+4dySikqzEOjoZYlKsxHv3eijO23eS4Yj2l2gl/61yrceoRKttwVK371Z9ewX8UFlsneTwbrdOePhrreR/76yCVjnW84Q2+y4viM+wZimUbrFOKOxaa5XaSkjMgqRs671J2dalCO64w0/6jYGKneDdAQltIS7tx9/rr2+YGREgeFLEZrdOXoT4RIOS9GNR77N+IUBJuoi0eFFRUVxyySW8+uqrrF+/ni5dunDKKacAsHDhQsaOHcvFF18MWNeYb968+bD33a1bN15++WVqamqCo+mLFy9uVGfhwoUMGTKEG264Ibhtw4YNjeq43W78fv8hP+utt97CGBMcTV+4cCHx8fG0a9fusGMWEZEjkxjtIjHaRbfMhCN6nzGGPVV1FJXXsNPrY6e3hiKv9by43EdxhfVYVF5DTV0Af8Cqv6fqcBLaxIayj8NuIyHKSVyUE49zIB6nHbfTjqc1uKmlqMbBnq9r2bO4jtr6bwFw2t2kxI0gJeaXdM/YRQfXbpI8hkSXIcEVIMEVIMppw+mwNxQnLgfE1Owk2rsJR8kG2L0OasqslfFrSoGN+4frjrdm+CZ3sJLruiorAa6rtp7X11g5zN7r9GvLrX3uWGGVY1VfffCFAvfaeweAw+GMgphU63KD2FRr0MLualg7wGldglCxE0o2wZ7N1ufvFd3KytFad7VOGFQWWTMM9myxHqtLfvxznVH7Zh5c86F1sqGZKEk/FrvWWf/APInNetBERMLV6NGj+fnPf863337LlVdeGdzeuXNn3n77bUaMGIHNZuOuu+46olHnX/3qV0yZMoXx48dz5513snnzZh566KFGdTp37sxLL73EnDlzyM3N5eWXX2bp0qXk5uYG6+Tk5DBnzhzy8vJISUkhMXH/hZNuuOEGZsyYwU033cSECRPIy8vjnnvuYeLEiY1WrhcRkfBgs9lIjnWTHOum64/cUdMYg68+QFl1Hd7qOrw1dZRV11FeU0+lz09VrXWruoqaesqq69hTVWsl85W17KmqxVtTjz9gjjDJt9QHTMMJBB/f4QEON3dIB3oT43aQFOWkXaKPdu4KMlxVpDsrSbVXkGCvoTwmi92xnamIatNoFNhgvtf+/fcejY92FJHh306KbzsJ1dtwBmqw7V1Az+HG5nBRZ2zU1gfw1QeorTf4/AFMdDKOpLZEJbcjJjWb+OR0XNXFVrK8t5Tmg7fAGnEvL7QWBQzUWwnw3lvvpZ4EnnhrJkJp/r5EurbcOqng3WaVw2GzWwsIVhZD9R7YstAqR6q+xio1pQ37dRz5Po6BkvRjUdSw4mR695BPiRARCQc/+clPSE5OJi8vj1/96lfB7Y888gjjxo1jyJAhpKamcscdd+D1eg97v3Fxcbz33ntcf/31nHzyyXTv3p0//elP/PKXvwzW+fWvf82KFSsYNWoUNpuNK664ghtuuIEPP/wwWGf8+PHMnz+f/v37U1FREbwF2/e1bduWDz74gNtvv50+ffqQnJzMtddey+9///uj/2JERCTkbDYbUS4HUS4H6QlHPh3fGENVrZ/ymnq8NVZy76v3U1sfsIo/QMBYMwKSY9wkxbhoFevGabdRUlnL7opadlX62F1Ry57KWkqqGh4bTgJU+vz46v34GpLhmjo/Fb56jIGqWj9VtX52eOFL4oC4A0RYDuQd5bfTqqH0PIr3Vjd8bh4ep504j5MYTw6x7k5Eux047Tbs0TacsZBgKvDgY7cthXpjw19iCOyyzh44Hf1wOew4o2w429qId9SSjJdkm5ekQBmJpgyP8eEI1GE39dhNPQ5TT5U7mTJPW8qis6iIygCHGw+1pNXm07p6IymVG4ivKcAXlUpVTFuqYtpSHduOmqg0jN3ZMGvOhs1uw2YMpq4a6muw1dWAvwZbfQ2nORNpzhv82Yw50DmVE5fX6yUxMZGysjISEo5sKs1+PpkKn/8F+l8LP3+kSeITkZatpqaGTZs2kZub22iRNIl8P3Zsm7RvOoE8+eSTTJ8+ncLCQvr06cPjjz/OwIEDD+u9+k5F5EQQCBjKG0b2S6utkf2y6oZSVRucCRAwxlp0vuF9xlhjiHuHEfeOJ9oatuz9uabOT0nDbIGShlJT799v1H3vFP/4KBcJ0U5i3U5q6vyUVtdRWmXNSjiRs8olvzvnqE7sfN+R9EsaST8W8W2g3QBoe0qoIxERETmhvP7660ycOJFnnnmGQYMGMWPGDIYPH05eXh5paWmhDk9EpFnY7TYSY1wkxrjI5vDvuNIUjDEEDASMwWm3NbrryQ/5AwZvdR2VtdalA9ZjPVW1fgIBg9+Y4KUCxlhJv91uw2Gz4Wi4kqzOb6gPBKjzG+r8AeoaZhRYxR9cT+CHMRoa7oqHFa8xhjq/aTTDobbeusTOYIInE77/Hsy+SwMcdpsVn82Gs+G5y9G8l7spST8Wg66zioiIiDSpRx55hPHjx3PNNdcA8MwzzzB79myef/55Jk+eHOLoREROfDabDYcNHBz6sl6H3UarWDetYt3NENmJTyvgiIiISFipra1l+fLlDBs2LLjNbrczbNgwFi1adMD3+Hw+vF5voyIiIhKJlKSLiIhIWNm1axd+v5/09PRG29PT0yksLDzge6ZNm0ZiYmKwZGVlNUeoIiIiTU5JuohIGGpha3q2CDqmx9edd95JWVlZsGzdujXUIYmIiBwVXZMuIhJGXC4XAFVVVURHN+fNPuR4q6qqAvYdYzm41NRUHA4HO3fubLR9586dZGQc+CbIHo8Hj8fTHOGJiIgcV0rSRUTCiMPhICkpiaKiIgBiYmJ+dDVVCX/GGKqqqigqKiIpKQmHwxHqkMKe2+2mX79+zJ07l4suugiAQCDA3LlzmTBhQmiDExEROc6UpIuIhJm9I4V7E3U5MSQlJR10FFj2N3HiRMaMGUP//v0ZOHAgM2bMoLKyMrjau4iIyIlKSbqISJix2WxkZmaSlpZGXV1dqMORJuByuTSCfoRGjRpFcXExd999N4WFhfTt25ePPvpov8XkRERETjRK0kVEwpTD4VBiJy3ahAkTNL1dRERaHK3uLiIiIiIiIhImlKSLiIiIiIiIhAkl6SIiIiIiIiJhosVdk26MAcDr9YY4EhEREcvePmlvHyXHTv29iIiEkyPp61tckl5eXg5AVlZWiCMRERFprLy8nMTExFCHcUJQfy8iIuHocPp6m2lhp+0DgQA7duwgPj4em812TPvyer1kZWWxdetWEhISmijC5qU2hF6kxw9qQziI9PihZbfBGEN5eTlt2rTBbteVaE1B/f0+kR4/qA3hINLjB7UhHER6/NA8fX2LG0m32+20a9euSfeZkJAQsf/I9lIbQi/S4we1IRxEevzQctugEfSmpf5+f5EeP6gN4SDS4we1IRxEevxwfPt6na4XERERERERCRNK0kVERERERETChJL0Y+DxeLjnnnvweDyhDuWoqQ2hF+nxg9oQDiI9flAbJHxF+nGN9PhBbQgHkR4/qA3hINLjh+ZpQ4tbOE5EREREREQkXGkkXURERERERCRMKEkXERERERERCRNK0kVERERERETChJJ0ERERERERkTChJP0YPPnkk+Tk5BAVFcWgQYP48ssvQx3SQf3nP/9hxIgRtGnTBpvNxjvvvNPodWMMd999N5mZmURHRzNs2DDWrVsXmmAPYNq0aQwYMID4+HjS0tK46KKLyMvLa1SnpqaGG2+8kZSUFOLi4vjlL3/Jzp07QxTx/p5++ml69+5NQkICCQkJDB48mA8//DD4erjH/0N//OMfsdls3HLLLcFt4d6GqVOnYrPZGpWuXbsGXw/3+Pfavn07V155JSkpKURHR9OrVy+WLVsWfD2cf59zcnL2OwY2m40bb7wRiIxj4Pf7ueuuu8jNzSU6OpqOHTty//338/11WMP5GMiRUV/ffNTXhx/19aETyX09RH5/H/K+3shRmTlzpnG73eb555833377rRk/frxJSkoyO3fuDHVoB/TBBx+YKVOmmLffftsAZtasWY1e/+Mf/2gSExPNO++8Y7766ivzi1/8wuTm5prq6urQBPwDw4cPNy+88IJZtWqVWblypTn//PNNdna2qaioCNa5/vrrTVZWlpk7d65ZtmyZOfXUU82QIUNCGHVj7777rpk9e7ZZu3atycvLM7/73e+My+Uyq1atMsaEf/zf9+WXX5qcnBzTu3dvc/PNNwe3h3sb7rnnHtOjRw9TUFAQLMXFxcHXwz1+Y4wpKSkx7du3N2PHjjVLliwxGzduNHPmzDHr168P1gnn3+eioqJG3//HH39sADNv3jxjTGQcgwcffNCkpKSY999/32zatMn861//MnFxcebRRx8N1gnnYyCHT31981JfH17U14dOpPf1xkR+fx/qvl5J+lEaOHCgufHGG4M/+/1+06ZNGzNt2rQQRnV4fthxBwIBk5GRYaZPnx7cVlpaajwej/nnP/8ZgggPraioyABmwYIFxhgrXpfLZf71r38F66xevdoAZtGiRaEK85BatWplnn322YiKv7y83HTu3Nl8/PHH5swzzwx23JHQhnvuucf06dPngK9FQvzGGHPHHXeY00477aCvR9rv880332w6duxoAoFAxByDCy64wIwbN67RtksuucSMHj3aGBN5x0AOTn19aKmvDx319aF1ovX1xkRefx/qvl7T3Y9CbW0ty5cvZ9iwYcFtdrudYcOGsWjRohBGdnQ2bdpEYWFho/YkJiYyaNCgsG1PWVkZAMnJyQAsX76curq6Rm3o2rUr2dnZYdkGv9/PzJkzqaysZPDgwREV/4033sgFF1zQKFaInGOwbt062rRpQ4cOHRg9ejT5+flA5MT/7rvv0r9/fy677DLS0tI4+eST+fvf/x58PZJ+n2tra3nllVcYN24cNpstYo7BkCFDmDt3LmvXrgXgq6++4vPPP+e8884DIusYyMGprw899fWho74+tE6kvh4is78PdV/vPOY9tEC7du3C7/eTnp7eaHt6ejpr1qwJUVRHr7CwEOCA7dn7WjgJBALccsstDB06lJ49ewJWG9xuN0lJSY3qhlsbvvnmGwYPHkxNTQ1xcXHMmjWL7t27s3LlyoiIf+bMmfz3v/9l6dKl+70WCcdg0KBBvPjii3Tp0oWCggLuvfdeTj/9dFatWhUR8QNs3LiRp59+mokTJ/K73/2OpUuX8tvf/ha3282YMWMi6vf5nXfeobS0lLFjxwKR8W8IYPLkyXi9Xrp27YrD4cDv9/Pggw8yevRoIPL+T5UDU18fWurrQ0d9feidSH09RGZ/H+q+Xkm6RJwbb7yRVatW8fnnn4c6lCPWpUsXVq5cSVlZGW+++SZjxoxhwYIFoQ7rsGzdupWbb76Zjz/+mKioqFCHc1T2nv0E6N27N4MGDaJ9+/a88cYbREdHhzCywxcIBOjfvz9/+MMfADj55JNZtWoVzzzzDGPGjAlxdEfmueee47zzzqNNmzahDuWIvPHGG7z66qu89tpr9OjRg5UrV3LLLbfQpk2biDsGIuFKfX1oqK8PDydSXw+R2d+Huq/XdPejkJqaisPh2G8Fwp07d5KRkRGiqI7e3pgjoT0TJkzg/fffZ968ebRr1y64PSMjg9raWkpLSxvVD7c2uN1uOnXqRL9+/Zg2bRp9+vTh0UcfjYj4ly9fTlFREaeccgpOpxOn08mCBQt47LHHcDqdpKenh30bfigpKYmTTjqJ9evXR8QxAMjMzKR79+6NtnXr1i04lS9Sfp+3bNnCJ598wv/8z/8Et0XKMbj99tuZPHkyl19+Ob169eKqq67i1ltvZdq0aUDkHAP5cerrQ0d9feiorw8PJ0pfD5Hb34e6r1eSfhTcbjf9+vVj7ty5wW2BQIC5c+cyePDgEEZ2dHJzc8nIyGjUHq/Xy5IlS8KmPcYYJkyYwKxZs/j000/Jzc1t9Hq/fv1wuVyN2pCXl0d+fn7YtOFAAoEAPp8vIuI/55xz+Oabb1i5cmWw9O/fn9GjRwefh3sbfqiiooINGzaQmZkZEccAYOjQofvdkmjt2rW0b98eiIzfZ4AXXniBtLQ0LrjgguC2SDkGVVVV2O2Nu0+Hw0EgEAAi5xjIj1Nf3/zU14ee+vrwcKL09RC5/X3I+/pjXnquhZo5c6bxeDzmxRdfNN9995257rrrTFJSkiksLAx1aAdUXl5uVqxYYVasWGEA88gjj5gVK1aYLVu2GGOsWwgkJSWZ//u//zNff/21ufDCC8PqNg6/+c1vTGJiopk/f36j2zlUVVUF61x//fUmOzvbfPrpp2bZsmVm8ODBZvDgwSGMurHJkyebBQsWmE2bNpmvv/7aTJ482dhsNvPvf//bGBP+8R/I91d8NSb823DbbbeZ+fPnm02bNpmFCxeaYcOGmdTUVFNUVGSMCf/4jbFuieN0Os2DDz5o1q1bZ1599VUTExNjXnnllWCdcP999vv9Jjs729xxxx37vRYJx2DMmDGmbdu2wduyvP322yY1NdX87//+b7BOuB8DOTzq65uX+vrwpL6++Z0Ifb0xkd3fh7qvV5J+DB5//HGTnZ1t3G63GThwoFm8eHGoQzqoefPmGWC/MmbMGGOMdRuBu+66y6SnpxuPx2POOecck5eXF9qgv+dAsQPmhRdeCNaprq42N9xwg2nVqpWJiYkxF198sSkoKAhd0D8wbtw40759e+N2u03r1q3NOeecE+y0jQn/+A/khx13uLdh1KhRJjMz07jdbtO2bVszatSoRvccDff493rvvfdMz549jcfjMV27djV/+9vfGr0e7r/Pc+bMMcABY4qEY+D1es3NN99ssrOzTVRUlOnQoYOZMmWK8fl8wTrhfgzk8Kmvbz7q68OT+vrQiPS+3pjI7u9D3dfbjDHm2MfjRURERERERORY6Zp0ERERERERkTChJF1EREREREQkTChJFxEREREREQkTStJFREREREREwoSSdBEREREREZEwoSRdREREREREJEwoSRcREREREREJE0rSRURERERERMKEknQRaVbz58/HZrNRWloa6lBERETkOFF/L3L0lKSLiIiIiIiIhAkl6SIiIiIiIiJhQkm6SAsTCASYNm0aubm5REdH06dPH958801g39S02bNn07t3b6Kiojj11FNZtWpVo3289dZb9OjRA4/HQ05ODg8//HCj130+H3fccQdZWVl4PB46derEc88916jO8uXL6d+/PzExMQwZMoS8vLzj23AREZEWRP29SORSki7SwkybNo2XXnqJZ555hm+//ZZbb72VK6+8kgULFgTr3H777Tz88MMsXbqU1q1bM2LECOrq6gCrsx05ciSXX34533zzDVOnTuWuu+7ixRdfDL7/6quv5p///CePPfYYq1ev5q9//StxcXGN4pgyZQoPP/wwy5Ytw+l0Mm7cuGZpv4iISEug/l4kghkRaTFqampMTEyM+eKLLxptv/baa80VV1xh5s2bZwAzc+bM4Gu7d+820dHR5vXXXzfGGPOrX/3KnHvuuY3ef/vtt5vu3bsbY4zJy8szgPn4448PGMPez/jkk0+C22bPnm0AU11d3STtFBERacnU34tENo2ki7Qg69evp6qqinPPPZe4uLhgeemll9iwYUOw3uDBg4PPk5OT6dKlC6tXrwZg9erVDB06tNF+hw4dyrp16/D7/axcuRKHw8GZZ575o7H07t07+DwzMxOAoqKiY26jiIhIS6f+XiSyOUMdgIg0n4qKCgBmz55N27ZtG73m8XgaddxHKzo6+rDquVyu4HObzQZY18+JiIjIsVF/LxLZNJIu0oJ0794dj8dDfn4+nTp1alSysrKC9RYvXhx8vmfPHtauXUu3bt0A6NatGwsXLmy034ULF3LSSSfhcDjo1asXgUCg0TVvIiIi0nzU34tENo2ki7Qg8fHxTJo0iVtvvZVAIMBpp51GWVkZCxcuJCEhgfbt2wNw3333kZKSQnp6OlOmTCE1NZWLLroIgNtuu40BAwZw//33M2rUKBYtWsQTTzzBU089BUBOTg5jxoxh3LhxPPbYY/Tp04ctW7ZQVFTEyJEjQ9V0ERGRFkP9vUiEC/VF8SLSvAKBgJkxY4bp0qWLcblcpnXr1mb48OFmwYIFwUVe3nvvPdOjRw/jdrvNwIEDzVdffdVoH2+++abp3r27cblcJjs720yfPr3R69XV1ebWW281mZmZxu12m06dOpnnn3/eGLNvIZk9e/YE669YscIAZtOmTce7+SIiIi2C+nuRyGUzxphQniQQkfAxf/58zj77bPbs2UNSUlKowxEREZHjQP29SHjTNekiIiIiIiIiYUJJuoiIiIiIiEiY0HR3ERERERERkTChkXQRERERERGRMKEkXURERERERCRMKEkXERERERERCRNK0kVERERERETChJJ0ERERERERkTChJF1EREREREQkTChJFxEREREREQkTStJFREREREREwsT/AyrVIH5sTECsAAAAAElFTkSuQmCC",
      "text/plain": [
       "<Figure size 1200x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Train and Validation accuracy and loss\n",
    "f = plt.figure(figsize = (12,4))\n",
    "\n",
    "plt.subplot(1,2,1)\n",
    "plt.title('train validation accuracy')\n",
    "plt.plot(history2.history['accuracy'])\n",
    "plt.plot(history2.history['val_accuracy'])\n",
    "plt.legend(['train','validation'])\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('accuracy')\n",
    "\n",
    "\n",
    "plt.subplot(1,2,2)\n",
    "plt.title('train validation loss')\n",
    "plt.plot(history2.history['loss'])\n",
    "plt.plot(history2.history['val_loss'])\n",
    "plt.legend(['train','validation'])\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('loss')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "id": "68603b8c",
   "metadata": {},
   "outputs": [],
   "source": [
    "spec_test = spec_file['spec_test']\n",
    "y_test_spec = spec_file['y_test']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "id": "e42f2a3c",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature Scaling\n",
    "spec_test_scaled = spec_test / np.amax(spec_test)\n",
    "\n",
    "# Reshape the data\n",
    "N,row,col = spec_test_scaled.shape\n",
    "spec_test_reshaped = spec_test_scaled.reshape((N,row,col,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "id": "ad1e99d4",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4/4 [==============================] - 1s 193ms/step\n"
     ]
    }
   ],
   "source": [
    "y2_prob = model2.predict(spec_test_reshaped)\n",
    "y2_pred = np.argmax(y2_prob, axis = -1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "id": "993754a4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.92"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Accuracy\n",
    "acc_spec = round(accuracy_score(y2_pred,y_test_spec),2)\n",
    "acc_spec"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "id": "b222d33c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "       Angry       0.93      0.93      0.93        30\n",
      "       Happy       0.94      1.00      0.97        30\n",
      "     Relaxed       0.90      0.90      0.90        30\n",
      "         Sad       0.93      0.87      0.90        30\n",
      "\n",
      "    accuracy                           0.93       120\n",
      "   macro avg       0.92      0.93      0.92       120\n",
      "weighted avg       0.92      0.93      0.92       120\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Confusion Matrix\n",
    "cm_spec = confusion_matrix(y_test_spec,y2_pred)\n",
    "cm_display_spec = ConfusionMatrixDisplay(confusion_matrix = cm_spec, display_labels = ['Angry', 'Happy', 'Relaxed', 'Sad'])\n",
    "print(classification_report(y_test_spec, y2_pred, target_names = ['Angry', 'Happy', 'Relaxed', 'Sad']))\n",
    "cm_display_spec.plot()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "id": "a1fde966",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save model\n",
    "model2.save(os.getcwd() + \"/Features2D/Model/Conv2D_spec_agumented.h5\") "
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8c0f5456",
   "metadata": {},
   "source": [
    "### 4.3 Mel Spectrogram"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "id": "23b2c9a1",
   "metadata": {},
   "outputs": [],
   "source": [
    "mel_file = np.load(\"Features2D/npz_new/mel_train_test_agumented.npz\")\n",
    "\n",
    "mel_train = mel_file['mel_train']\n",
    "y3_train = mel_file['y_train']\n",
    "mel_train,mel_val,y3_train,y3_val = train_test_split(mel_train,y3_train, \n",
    "                                                     train_size = 0.8, \n",
    "                                                     random_state = 13, \n",
    "                                                     stratify= y3_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "id": "c9377980",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature Scaling\n",
    "mel_train_scaled = (mel_train - np.mean(mel_train))/np.std(mel_train)\n",
    "mel_val_scaled = (mel_val - np.mean(mel_val))/np.std(mel_val)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "id": "0d5aa99d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Reshape the data\n",
    "N,row,col = mel_train_scaled.shape\n",
    "mel_train_reshaped = mel_train_scaled.reshape((N,row,col,1))\n",
    "\n",
    "N,row,col = mel_val_scaled.shape\n",
    "mel_val_reshaped = mel_val_scaled.reshape((N,row,col,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "id": "bfa600a6",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Model: \"CNN2D_3\"\n",
      "_________________________________________________________________\n",
      " Layer (type)                Output Shape              Param #   \n",
      "=================================================================\n",
      " Conv2D_1 (Conv2D)           (None, 300, 400, 8)       80        \n",
      "                                                                 \n",
      " MaxPooling2D_1 (MaxPooling2  (None, 150, 200, 8)      0         \n",
      " D)                                                              \n",
      "                                                                 \n",
      " BN1 (BatchNormalization)    (None, 150, 200, 8)       32        \n",
      "                                                                 \n",
      " Conv2D_2 (Conv2D)           (None, 150, 200, 8)       584       \n",
      "                                                                 \n",
      " MaxPooling2D_2 (MaxPooling2  (None, 75, 100, 8)       0         \n",
      " D)                                                              \n",
      "                                                                 \n",
      " BN2 (BatchNormalization)    (None, 75, 100, 8)        32        \n",
      "                                                                 \n",
      " Conv2D_3 (Conv2D)           (None, 75, 100, 16)       1168      \n",
      "                                                                 \n",
      " MaxPooling2D_3 (MaxPooling2  (None, 38, 50, 16)       0         \n",
      " D)                                                              \n",
      "                                                                 \n",
      " BN3 (BatchNormalization)    (None, 38, 50, 16)        64        \n",
      "                                                                 \n",
      " Flatten (Flatten)           (None, 30400)             0         \n",
      "                                                                 \n",
      " Dense1 (Dense)              (None, 64)                1945664   \n",
      "                                                                 \n",
      " Dropout1 (Dropout)          (None, 64)                0         \n",
      "                                                                 \n",
      " Dense2 (Dense)              (None, 32)                2080      \n",
      "                                                                 \n",
      " Dropout2 (Dropout)          (None, 32)                0         \n",
      "                                                                 \n",
      " Softmax (Dense)             (None, 4)                 132       \n",
      "                                                                 \n",
      "=================================================================\n",
      "Total params: 1,949,836\n",
      "Trainable params: 1,949,772\n",
      "Non-trainable params: 64\n",
      "_________________________________________________________________\n"
     ]
    }
   ],
   "source": [
    "f1,k1,a1 = 8,3,'relu'\n",
    "mw1,mh1,sw1,sh1 = 2,3,2,2\n",
    "f2,k2,a2 = 8,3,'relu'\n",
    "mw2,mh2,sw2,sh2 = 2,3,2,2\n",
    "f3,k3,a3 = 16,3,'relu'\n",
    "mw3,mh3,sw3,sh3 = 2,3,2,2\n",
    "d1,dr1,da1,r1 = 64,0.2,'relu',0.2\n",
    "d2,dr2,da2,r2 = 32,0.2,'relu',0.2\n",
    "num = 3\n",
    "model3 = modelBuilder3L(mel_train_reshaped,\n",
    "                        f1,k1,a1,\n",
    "                        mw1,mh1,sw1,sh1,\n",
    "                        f2,k2,a2,\n",
    "                        mw2,mh2,sw2,sh2,\n",
    "                        f3,k3,a3,\n",
    "                        mw3,mh3,sw3,sh3,\n",
    "                        d1,dr1,da1,r1,\n",
    "                        d2,dr2,da2,r2,\n",
    "                        num)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "id": "c359f915",
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Epoch 1/80\n",
      "44/44 [==============================] - 31s 565ms/step - loss: 32.1929 - accuracy: 0.4572 - val_loss: 28.1732 - val_accuracy: 0.5694\n",
      "Epoch 2/80\n",
      "44/44 [==============================] - 23s 517ms/step - loss: 24.4580 - accuracy: 0.6898 - val_loss: 21.8202 - val_accuracy: 0.3519\n",
      "Epoch 3/80\n",
      "44/44 [==============================] - 23s 518ms/step - loss: 18.9122 - accuracy: 0.7951 - val_loss: 17.6477 - val_accuracy: 0.3565\n",
      "Epoch 4/80\n",
      "44/44 [==============================] - 24s 543ms/step - loss: 15.1917 - accuracy: 0.8507 - val_loss: 15.0429 - val_accuracy: 0.3704\n",
      "Epoch 5/80\n",
      "44/44 [==============================] - 23s 514ms/step - loss: 12.6021 - accuracy: 0.8947 - val_loss: 13.1956 - val_accuracy: 0.3426\n",
      "Epoch 6/80\n",
      "44/44 [==============================] - 22s 494ms/step - loss: 10.8048 - accuracy: 0.9178 - val_loss: 11.1936 - val_accuracy: 0.4259\n",
      "Epoch 7/80\n",
      "44/44 [==============================] - 22s 507ms/step - loss: 9.5246 - accuracy: 0.9236 - val_loss: 9.6247 - val_accuracy: 0.6019\n",
      "Epoch 8/80\n",
      "44/44 [==============================] - 22s 503ms/step - loss: 8.5068 - accuracy: 0.9363 - val_loss: 9.0258 - val_accuracy: 0.5509\n",
      "Epoch 9/80\n",
      "44/44 [==============================] - 22s 505ms/step - loss: 7.6835 - accuracy: 0.9456 - val_loss: 7.8506 - val_accuracy: 0.7130\n",
      "Epoch 10/80\n",
      "44/44 [==============================] - 22s 504ms/step - loss: 7.0229 - accuracy: 0.9387 - val_loss: 6.9284 - val_accuracy: 0.8657\n",
      "Epoch 11/80\n",
      "44/44 [==============================] - 21s 484ms/step - loss: 6.4897 - accuracy: 0.9421 - val_loss: 6.3767 - val_accuracy: 0.8750\n",
      "Epoch 12/80\n",
      "44/44 [==============================] - 22s 496ms/step - loss: 6.0024 - accuracy: 0.9606 - val_loss: 5.9256 - val_accuracy: 0.9120\n",
      "Epoch 13/80\n",
      "44/44 [==============================] - 22s 504ms/step - loss: 5.5941 - accuracy: 0.9583 - val_loss: 5.5916 - val_accuracy: 0.9028\n",
      "Epoch 14/80\n",
      "44/44 [==============================] - 22s 502ms/step - loss: 5.2254 - accuracy: 0.9711 - val_loss: 5.2169 - val_accuracy: 0.8889\n",
      "Epoch 15/80\n",
      "44/44 [==============================] - 24s 552ms/step - loss: 4.9785 - accuracy: 0.9410 - val_loss: 4.8788 - val_accuracy: 0.9213\n",
      "Epoch 16/80\n",
      "44/44 [==============================] - 22s 501ms/step - loss: 4.7202 - accuracy: 0.9572 - val_loss: 4.6924 - val_accuracy: 0.9167\n",
      "Epoch 17/80\n",
      "44/44 [==============================] - 22s 489ms/step - loss: 4.4690 - accuracy: 0.9572 - val_loss: 4.4031 - val_accuracy: 0.9352\n",
      "Epoch 18/80\n",
      "44/44 [==============================] - 22s 497ms/step - loss: 4.1586 - accuracy: 0.9780 - val_loss: 4.1441 - val_accuracy: 0.9167\n",
      "Epoch 19/80\n",
      "44/44 [==============================] - 22s 502ms/step - loss: 3.9084 - accuracy: 0.9722 - val_loss: 3.9071 - val_accuracy: 0.9352\n",
      "Epoch 20/80\n",
      "44/44 [==============================] - 21s 482ms/step - loss: 3.7573 - accuracy: 0.9641 - val_loss: 3.8004 - val_accuracy: 0.8981\n",
      "Epoch 21/80\n",
      "44/44 [==============================] - 22s 511ms/step - loss: 3.5397 - accuracy: 0.9757 - val_loss: 3.5336 - val_accuracy: 0.9306\n",
      "Epoch 22/80\n",
      "44/44 [==============================] - 23s 515ms/step - loss: 3.3613 - accuracy: 0.9803 - val_loss: 3.3415 - val_accuracy: 0.9352\n",
      "Epoch 23/80\n",
      "44/44 [==============================] - 21s 477ms/step - loss: 3.1577 - accuracy: 0.9838 - val_loss: 3.1700 - val_accuracy: 0.9306\n",
      "Epoch 24/80\n",
      "44/44 [==============================] - 22s 492ms/step - loss: 2.9620 - accuracy: 0.9884 - val_loss: 2.9981 - val_accuracy: 0.9306\n",
      "Epoch 25/80\n",
      "44/44 [==============================] - 21s 484ms/step - loss: 2.8128 - accuracy: 0.9780 - val_loss: 2.8830 - val_accuracy: 0.9074\n",
      "Epoch 26/80\n",
      "44/44 [==============================] - 22s 493ms/step - loss: 2.6884 - accuracy: 0.9745 - val_loss: 2.7307 - val_accuracy: 0.9259\n",
      "Epoch 27/80\n",
      "44/44 [==============================] - 22s 498ms/step - loss: 2.5448 - accuracy: 0.9757 - val_loss: 2.6204 - val_accuracy: 0.9306\n",
      "Epoch 28/80\n",
      "44/44 [==============================] - 22s 491ms/step - loss: 2.4469 - accuracy: 0.9780 - val_loss: 2.4660 - val_accuracy: 0.9491\n",
      "Epoch 29/80\n",
      "44/44 [==============================] - 22s 502ms/step - loss: 2.3122 - accuracy: 0.9803 - val_loss: 2.3624 - val_accuracy: 0.9398\n",
      "Epoch 30/80\n",
      "44/44 [==============================] - 23s 516ms/step - loss: 2.1816 - accuracy: 0.9815 - val_loss: 2.1975 - val_accuracy: 0.9583\n",
      "Epoch 31/80\n",
      "44/44 [==============================] - 22s 498ms/step - loss: 2.0911 - accuracy: 0.9838 - val_loss: 2.1861 - val_accuracy: 0.9074\n",
      "Epoch 32/80\n",
      "44/44 [==============================] - 26s 594ms/step - loss: 2.0002 - accuracy: 0.9699 - val_loss: 2.0588 - val_accuracy: 0.9120\n",
      "Epoch 33/80\n",
      "44/44 [==============================] - 27s 611ms/step - loss: 1.8993 - accuracy: 0.9780 - val_loss: 1.9352 - val_accuracy: 0.9398\n",
      "Epoch 34/80\n",
      "44/44 [==============================] - 25s 560ms/step - loss: 1.8010 - accuracy: 0.9803 - val_loss: 1.8325 - val_accuracy: 0.9352\n",
      "Epoch 35/80\n",
      "44/44 [==============================] - 25s 567ms/step - loss: 1.6989 - accuracy: 0.9861 - val_loss: 1.7378 - val_accuracy: 0.9444\n",
      "Epoch 36/80\n",
      "44/44 [==============================] - 22s 502ms/step - loss: 1.5890 - accuracy: 0.9907 - val_loss: 1.6338 - val_accuracy: 0.9537\n",
      "Epoch 37/80\n",
      "44/44 [==============================] - 22s 497ms/step - loss: 1.4936 - accuracy: 0.9919 - val_loss: 1.5846 - val_accuracy: 0.9306\n",
      "Epoch 38/80\n",
      "44/44 [==============================] - 21s 488ms/step - loss: 1.4440 - accuracy: 0.9861 - val_loss: 1.5419 - val_accuracy: 0.9259\n",
      "Epoch 39/80\n",
      "44/44 [==============================] - 24s 555ms/step - loss: 1.3786 - accuracy: 0.9896 - val_loss: 1.5611 - val_accuracy: 0.8889\n",
      "Epoch 40/80\n",
      "44/44 [==============================] - 23s 514ms/step - loss: 1.3203 - accuracy: 0.9734 - val_loss: 1.3819 - val_accuracy: 0.9259\n",
      "Epoch 41/80\n",
      "44/44 [==============================] - 21s 476ms/step - loss: 1.2590 - accuracy: 0.9838 - val_loss: 1.3316 - val_accuracy: 0.9306\n",
      "Epoch 42/80\n",
      "44/44 [==============================] - 22s 503ms/step - loss: 1.2024 - accuracy: 0.9861 - val_loss: 1.3956 - val_accuracy: 0.8704\n",
      "Epoch 43/80\n",
      "44/44 [==============================] - 22s 511ms/step - loss: 1.1434 - accuracy: 0.9815 - val_loss: 1.2418 - val_accuracy: 0.9259\n",
      "Epoch 44/80\n",
      "44/44 [==============================] - 23s 499ms/step - loss: 1.0723 - accuracy: 0.9907 - val_loss: 1.1762 - val_accuracy: 0.9120\n",
      "Epoch 45/80\n",
      "44/44 [==============================] - 26s 602ms/step - loss: 1.0007 - accuracy: 0.9873 - val_loss: 1.0676 - val_accuracy: 0.9583\n",
      "Epoch 46/80\n",
      "44/44 [==============================] - 23s 523ms/step - loss: 0.9351 - accuracy: 0.9977 - val_loss: 1.0303 - val_accuracy: 0.9259\n",
      "Epoch 47/80\n",
      "44/44 [==============================] - 23s 521ms/step - loss: 0.8885 - accuracy: 0.9884 - val_loss: 0.9859 - val_accuracy: 0.9306\n",
      "Epoch 48/80\n",
      "44/44 [==============================] - 22s 501ms/step - loss: 0.8526 - accuracy: 0.9873 - val_loss: 1.0640 - val_accuracy: 0.8704\n",
      "Epoch 49/80\n",
      "44/44 [==============================] - 22s 497ms/step - loss: 1.0439 - accuracy: 0.9387 - val_loss: 1.1210 - val_accuracy: 0.9213\n",
      "Epoch 50/80\n",
      "44/44 [==============================] - 22s 502ms/step - loss: 0.9275 - accuracy: 0.9838 - val_loss: 0.9876 - val_accuracy: 0.9120\n",
      "Epoch 51/80\n",
      "44/44 [==============================] - 22s 503ms/step - loss: 0.7981 - accuracy: 0.9942 - val_loss: 0.9125 - val_accuracy: 0.9167\n",
      "Epoch 52/80\n",
      "44/44 [==============================] - 22s 507ms/step - loss: 0.7280 - accuracy: 0.9931 - val_loss: 0.8126 - val_accuracy: 0.9491\n",
      "Epoch 53/80\n",
      "44/44 [==============================] - 23s 514ms/step - loss: 0.7117 - accuracy: 0.9803 - val_loss: 0.8269 - val_accuracy: 0.9167\n",
      "Epoch 54/80\n",
      "44/44 [==============================] - 22s 509ms/step - loss: 0.6658 - accuracy: 0.9931 - val_loss: 0.7577 - val_accuracy: 0.9352\n",
      "Epoch 55/80\n",
      "44/44 [==============================] - 23s 511ms/step - loss: 0.6283 - accuracy: 0.9907 - val_loss: 0.7400 - val_accuracy: 0.9444\n",
      "Epoch 56/80\n",
      "44/44 [==============================] - 22s 507ms/step - loss: 0.6124 - accuracy: 0.9884 - val_loss: 0.6972 - val_accuracy: 0.9583\n",
      "Epoch 57/80\n",
      "44/44 [==============================] - 22s 494ms/step - loss: 0.6063 - accuracy: 0.9884 - val_loss: 0.7115 - val_accuracy: 0.9259\n",
      "Epoch 58/80\n",
      "44/44 [==============================] - 24s 544ms/step - loss: 0.5579 - accuracy: 0.9919 - val_loss: 0.6704 - val_accuracy: 0.9167\n",
      "Epoch 59/80\n",
      "44/44 [==============================] - 25s 559ms/step - loss: 0.5450 - accuracy: 0.9919 - val_loss: 0.6297 - val_accuracy: 0.9583\n",
      "Epoch 60/80\n",
      "44/44 [==============================] - 26s 587ms/step - loss: 0.5298 - accuracy: 0.9896 - val_loss: 0.6091 - val_accuracy: 0.9444\n",
      "Epoch 61/80\n",
      "44/44 [==============================] - 22s 506ms/step - loss: 0.5314 - accuracy: 0.9896 - val_loss: 0.6479 - val_accuracy: 0.9213\n",
      "Epoch 62/80\n",
      "44/44 [==============================] - 20s 463ms/step - loss: 0.5307 - accuracy: 0.9850 - val_loss: 0.7234 - val_accuracy: 0.8843\n",
      "Epoch 63/80\n",
      "44/44 [==============================] - 21s 484ms/step - loss: 0.5039 - accuracy: 0.9884 - val_loss: 0.5756 - val_accuracy: 0.9537\n",
      "Epoch 64/80\n",
      "44/44 [==============================] - 21s 471ms/step - loss: 0.5058 - accuracy: 0.9861 - val_loss: 0.6223 - val_accuracy: 0.9398\n",
      "Epoch 65/80\n",
      "44/44 [==============================] - 21s 478ms/step - loss: 0.4487 - accuracy: 0.9907 - val_loss: 0.5572 - val_accuracy: 0.9213\n",
      "Epoch 66/80\n",
      "44/44 [==============================] - 21s 480ms/step - loss: 0.4226 - accuracy: 0.9942 - val_loss: 0.5321 - val_accuracy: 0.9306\n",
      "Epoch 67/80\n",
      "44/44 [==============================] - 21s 472ms/step - loss: 0.4280 - accuracy: 0.9896 - val_loss: 0.5093 - val_accuracy: 0.9537\n",
      "Epoch 68/80\n",
      "44/44 [==============================] - 21s 468ms/step - loss: 0.4066 - accuracy: 0.9884 - val_loss: 0.5781 - val_accuracy: 0.9259\n",
      "Epoch 69/80\n",
      "44/44 [==============================] - 21s 473ms/step - loss: 0.6592 - accuracy: 0.9560 - val_loss: 0.6850 - val_accuracy: 0.9583\n",
      "Epoch 70/80\n",
      "44/44 [==============================] - 21s 469ms/step - loss: 0.5113 - accuracy: 0.9907 - val_loss: 0.5653 - val_accuracy: 0.9491\n",
      "Epoch 71/80\n",
      "44/44 [==============================] - 21s 473ms/step - loss: 0.4280 - accuracy: 0.9931 - val_loss: 0.5086 - val_accuracy: 0.9444\n",
      "Epoch 72/80\n",
      "44/44 [==============================] - 21s 470ms/step - loss: 0.3850 - accuracy: 0.9965 - val_loss: 0.4998 - val_accuracy: 0.9352\n",
      "Epoch 73/80\n",
      "44/44 [==============================] - 21s 472ms/step - loss: 0.3702 - accuracy: 0.9861 - val_loss: 0.5333 - val_accuracy: 0.9259\n",
      "Epoch 74/80\n",
      "44/44 [==============================] - 23s 530ms/step - loss: 0.3708 - accuracy: 0.9884 - val_loss: 0.4989 - val_accuracy: 0.9306\n",
      "Epoch 75/80\n",
      "44/44 [==============================] - 23s 531ms/step - loss: 0.3665 - accuracy: 0.9838 - val_loss: 0.5011 - val_accuracy: 0.9120\n",
      "Epoch 76/80\n",
      "44/44 [==============================] - 22s 511ms/step - loss: 0.3402 - accuracy: 0.9931 - val_loss: 0.4438 - val_accuracy: 0.9537\n",
      "Epoch 77/80\n",
      "44/44 [==============================] - 25s 566ms/step - loss: 0.3179 - accuracy: 0.9977 - val_loss: 0.4400 - val_accuracy: 0.9259\n",
      "Epoch 78/80\n",
      "44/44 [==============================] - 23s 516ms/step - loss: 0.3454 - accuracy: 0.9931 - val_loss: 0.4703 - val_accuracy: 0.9306\n",
      "Epoch 79/80\n",
      "44/44 [==============================] - 21s 472ms/step - loss: 0.3607 - accuracy: 0.9931 - val_loss: 0.4599 - val_accuracy: 0.9444\n",
      "Epoch 80/80\n",
      "44/44 [==============================] - 20s 460ms/step - loss: 0.3394 - accuracy: 0.9861 - val_loss: 0.4370 - val_accuracy: 0.9444\n"
     ]
    }
   ],
   "source": [
    "history3 = model3.fit(mel_train_reshaped,y3_train,validation_data=(mel_val_reshaped,y3_val),batch_size=20,epochs=80)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "id": "0dacccd0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0, 0.5, 'loss')"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 1200x400 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Train and Validation accuracy and loss\n",
    "f = plt.figure(figsize = (12,4))\n",
    "\n",
    "plt.subplot(1,2,1)\n",
    "plt.title('train validation accuracy')\n",
    "plt.plot(history3.history['accuracy'])\n",
    "plt.plot(history3.history['val_accuracy'])\n",
    "plt.legend(['train','validation'])\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('accuracy')\n",
    "\n",
    "\n",
    "plt.subplot(1,2,2)\n",
    "plt.title('train validation loss')\n",
    "plt.plot(history3.history['loss'])\n",
    "plt.plot(history3.history['val_loss'])\n",
    "plt.legend(['train','validation'])\n",
    "plt.xlabel('epoch')\n",
    "plt.ylabel('loss')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "id": "8f1bc1b5",
   "metadata": {},
   "outputs": [],
   "source": [
    "model3.save(os.getcwd() + \"/Features2D/Model/Conv2D_mel_agumented.h5\") "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "id": "e5985931",
   "metadata": {},
   "outputs": [],
   "source": [
    "mel_test = mel_file['mel_test']\n",
    "y_test_mel = mel_file['y_test']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "id": "aab860b1",
   "metadata": {},
   "outputs": [],
   "source": [
    "mel_test_scaled = (mel_test - np.mean(mel_test))/np.std(mel_test)\n",
    "\n",
    "# Reshape the data\n",
    "N,row,col = mel_test_scaled.shape\n",
    "mel_test_reshaped = mel_test_scaled.reshape((N,row,col,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "id": "dc2e7218",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4/4 [==============================] - 1s 221ms/step\n"
     ]
    }
   ],
   "source": [
    "y3_prob = model3.predict(mel_test_reshaped)\n",
    "y3_pred = np.argmax(y3_prob, axis = -1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "id": "e6e0d2f8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.98"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Accuracy\n",
    "acc_mel = round(accuracy_score(y3_pred,y_test_mel),2)\n",
    "acc_mel"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "id": "0b95e2b8",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "       Angry       1.00      0.97      0.98        30\n",
      "       Happy       0.97      1.00      0.98        30\n",
      "     Relaxed       0.97      0.97      0.97        30\n",
      "         Sad       0.97      0.97      0.97        30\n",
      "\n",
      "    accuracy                           0.97       120\n",
      "   macro avg       0.98      0.98      0.97       120\n",
      "weighted avg       0.98      0.97      0.97       120\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Confusion Matrix\n",
    "cm_mel = confusion_matrix(y_test_mel,y3_pred)\n",
    "cm_display_mel = ConfusionMatrixDisplay(confusion_matrix = cm_mel, display_labels = ['Angry', 'Happy', 'Relaxed', 'Sad'])\n",
    "print(classification_report(y_test_mel, y3_pred, target_names = ['Angry', 'Happy', 'Relaxed', 'Sad']))\n",
    "cm_display_mel.plot()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "afd38ef4",
   "metadata": {},
   "source": [
    "## 5. Ensenble 3 Models"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "id": "91423e4e",
   "metadata": {},
   "outputs": [],
   "source": [
    "mfcc_file_ense = np.load(os.getcwd() + \"/Features2D/npz_new/mfcc_train_test_agumented.npz\")\n",
    "mfcc_ense_train = mfcc_file_ense['mfcc_train']\n",
    "mfcc_ense_test =mfcc_file_ense['mfcc_test']\n",
    "y_ense_train = mfcc_file_ense['y_train']\n",
    "y_ense_test = mfcc_file_ense['y_test']\n",
    "model_mfcc = load_model(os.getcwd() + \"/Features2D/Model/Conv2D_mfcc_agumented.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "id": "c708b319",
   "metadata": {},
   "outputs": [],
   "source": [
    "spec_file_ense = np.load(os.getcwd() + \"/Features2D/npz_new/spec_train_test_agumented.npz\")\n",
    "spec_ense_train = spec_file_ense['spec_train']\n",
    "spec_ense_test = spec_file_ense['spec_test']\n",
    "y_ense_train = spec_file_ense['y_train']\n",
    "y_ense_test = spec_file_ense['y_test']\n",
    "model_spec = load_model(os.getcwd() + \"/Features2D/Model/Conv2D_spec_agumented.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "id": "ce67f297",
   "metadata": {},
   "outputs": [],
   "source": [
    "mel_file_ense = np.load(os.getcwd() + \"/Features2D/npz_new/mel_train_test_agumented.npz\")\n",
    "mel_ense_train = mel_file_ense['mel_train']\n",
    "mel_ense_test =mel_file_ense['mel_test']\n",
    "y_ense_train = mel_file_ense['y_train']\n",
    "y_ense_test = mel_file_ense['y_test']\n",
    "model_mel = load_model(os.getcwd() + \"/Features2D/Model/Conv2D_mel_agumented.h5\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "id": "25336c77",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature Scaling\n",
    "mfcc_ense_train_scaled = (mfcc_ense_train - np.mean(mfcc_ense_train))/np.std(mfcc_ense_train)\n",
    "spec_ense_train_scaled = spec_ense_train / np.amax(spec_ense_train)\n",
    "mel_ense_train_scaled = (mel_ense_train - np.mean(mel_ense_train))/np.std(mel_ense_train)\n",
    "\n",
    "mfcc_ense_test_scaled = (mfcc_ense_test - np.mean(mfcc_ense_test))/np.std(mfcc_ense_test)\n",
    "spec_ense_test_scaled = spec_ense_test / np.amax(spec_ense_test)\n",
    "mel_ense_test_scaled = (mel_ense_test - np.mean(mel_ense_test))/np.std(mfcc_ense_test)\n",
    "\n",
    "\n",
    "# Reshape the data\n",
    "N,row,col = mfcc_ense_train_scaled.shape\n",
    "mfcc_ense_train_reshaped = mfcc_ense_train_scaled.reshape((N,row,col,1))\n",
    "N,row,col = spec_ense_train_scaled.shape\n",
    "spec_ense_train_reshaped = spec_ense_train_scaled.reshape((N,row,col,1))\n",
    "N,row,col = mel_ense_train_scaled.shape\n",
    "mel_ense_train_reshaped = mel_ense_train_scaled.reshape((N,row,col,1))\n",
    "\n",
    "N,row,col = mfcc_ense_test_scaled.shape\n",
    "mfcc_ense_test_reshaped = mfcc_ense_test_scaled.reshape((N,row,col,1))\n",
    "N,row,col = spec_ense_test_scaled.shape\n",
    "spec_ense_test_reshaped = spec_ense_test_scaled.reshape((N,row,col,1))\n",
    "N,row,col = mel_ense_test_scaled.shape\n",
    "mel_ense_test_reshaped = mel_ense_test_scaled.reshape((N,row,col,1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "id": "f48d577e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "34/34 [==============================] - 7s 168ms/step\n",
      "34/34 [==============================] - 15s 408ms/step\n",
      "34/34 [==============================] - 12s 317ms/step\n"
     ]
    }
   ],
   "source": [
    "# Train 3 models\n",
    "y_ense_prob1 = model_mfcc.predict(mfcc_ense_train_reshaped)\n",
    "y_ense_pred1 = np.argmax(y_ense_prob1, axis= -1)\n",
    "\n",
    "y_ense_prob2 = model_spec.predict(spec_ense_train_reshaped)\n",
    "y_ense_pred2 = np.argmax(y_ense_prob2, axis= -1)\n",
    "\n",
    "y_ense_prob3 = model_mel.predict(mel_ense_train_reshaped)\n",
    "y_ense_pred3 = np.argmax(y_ense_prob3, axis= -1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "id": "2329d66d",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Vote for most accuracte prediction\n",
    "def get_majority(pred) :\n",
    "    N = len(pred[0]) # num of examples\n",
    "    vote = [] \n",
    "    for i in range(N) :\n",
    "        elements = [j[i] for j in pred] # 3 Prediction results  using 3 different model\n",
    "        elements = np.array(elements)\n",
    "        uniq, freq = np.unique(elements, return_counts= True) # Count the occurence of unqiue result\n",
    "        vote.append(uniq[np.argmax(freq)])# Choose the prediction with most occurence\n",
    "    \n",
    "    vote = np.array(vote)\n",
    "    return vote"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "id": "6bea2894",
   "metadata": {},
   "outputs": [],
   "source": [
    "# 3 different predictions\n",
    "train_pred = [y_ense_pred1, y_ense_pred2, y_ense_pred3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "id": "64d17b12",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_train_pred = get_majority(train_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "id": "5ab661e6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9833"
      ]
     },
     "execution_count": 84,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Model Evaluation\n",
    "train_corr = len(y_train_pred) - np.count_nonzero(y_train_pred - y_ense_train)\n",
    "train_acc = np.round((train_corr/ len(y_train_pred)),4)\n",
    "train_acc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "id": "99f79be5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "4/4 [==============================] - 1s 248ms/step\n",
      "4/4 [==============================] - 1s 169ms/step\n",
      "4/4 [==============================] - 1s 340ms/step\n"
     ]
    }
   ],
   "source": [
    "# Test ensembled models\n",
    "y_test_prob1 = model_spec.predict(spec_ense_test_reshaped)\n",
    "y_test_pred1 = np.argmax(y_test_prob1, axis= -1)\n",
    "\n",
    "y_test_prob2 = model_mfcc.predict(mfcc_ense_test_reshaped)\n",
    "y_test_pred2 = np.argmax(y_test_prob2, axis= -1)\n",
    "\n",
    "y_test_prob3 = model_mel.predict(mel_ense_test_reshaped)\n",
    "y_test_pred3 = np.argmax(y_test_prob3, axis= -1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "id": "434ad020",
   "metadata": {},
   "outputs": [],
   "source": [
    "test_pred = [y_test_pred1, y_test_pred2, y_test_pred3]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "id": "f3e6a64c",
   "metadata": {},
   "outputs": [],
   "source": [
    "y_test_pred = get_majority(test_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "id": "50a2ffdc",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.9333"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Model Evaluation\n",
    "test_corr = len(y_test_pred) - np.count_nonzero(y_test_pred - y_ense_test)\n",
    "test_acc = np.round((test_corr/ len(y_test_pred)),4)\n",
    "test_acc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "id": "beafc40c",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "              precision    recall  f1-score   support\n",
      "\n",
      "       Angry       0.91      0.97      0.94        30\n",
      "       Happy       0.97      1.00      0.98        30\n",
      "     Relaxed       0.88      0.97      0.92        30\n",
      "         Sad       1.00      0.80      0.89        30\n",
      "\n",
      "    accuracy                           0.93       120\n",
      "   macro avg       0.94      0.93      0.93       120\n",
      "weighted avg       0.94      0.93      0.93       120\n",
      "\n"
     ]
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 2 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Confusion Matrix\n",
    "cm_ense = confusion_matrix(y_ense_test,y_test_pred)\n",
    "cm_display_ense = ConfusionMatrixDisplay(confusion_matrix = cm_ense, display_labels = ['Angry', 'Happy', 'Relaxed', 'Sad'])\n",
    "print(classification_report(y_ense_test, y_test_pred, target_names = ['Angry', 'Happy', 'Relaxed', 'Sad']))\n",
    "cm_display_ense.plot()\n",
    "plt.show()"
   ]
  }
 ],
 "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": "3.10.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}